The Alaska Earthquake March 27, 1964: Regional Effects This volume was published as separate chapters A—J GEOLOGICAL SURVEY PROFESSIONAL PAPER 543 UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director (A) (B) (C V (D) (E) (F) (G) (H) (I) (J) CONTENTS [Letters designate the separately published chapters] Slide-induced waves, seiching, and ground fracturing caused by the earthquake of March 27, 1964, at Kenai Lake, Alaska, by David S. McCulloch. Geomorphic effects of the earthquake of March 27, 1964, in the Martin-Bering Rivers area, Alaska, by Samuel J. Tuthill and Wilson M. Laird. Gravity survey and regional geology of the Prince William Sound, epicentral region, Alaska, by J. E. Case, D. F. Barnes, George Plafker, and. S. L. Robbins. Geologic effects of the March 1964 earthquake and associated seismic sea waves on Kodiak and nearby islands, Alaska, by George Plafker and Reuben Kachadoorian. Effects of the earthquake of March 27, 1964, in the Copper River Basin area, Alaska, by Oscar J. Ferrians, Jr. Ground breakage and associated effects in the Cook Inlet area. Alaska, resulting from the March 27, 1964, earthquake, by Helen L. Foster and Thor N. V. Karlstrom. Surface faults on Montague Island associated with the 1964 Alaska earthquake, by George Plafker. Erosion and deposition on a beach raised by the 10434 earthquake. Montague Island, Alaska, by M. J. Kirkby and Anne V. Kirkby. Tectonics of the March 27, 1964, Alaska earthquake, by George Plafker. EiTects of the Alaska earthquake of March 27, 1964, 011 shore processes and beach morphology, by Kirk W. Stanley. U.S. GOVERNMENT PRINTING OFFICE: 1968 Of 298*580 The Alaska Earthquake March 27,1964 27/7 Regional Effects ”/5 194‘? ”/4 ‘ , ; Baa, “Sunbeam iSPLAY \ xx \ K ’ Slide-induced waves, seiching and ground fracturing at Kenai lake GEOLOGICAL SURVEY PROFESSIONAL PAPER, 543-AA THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS Slide-Induced Waves, Seiching And Ground Fracturing Caused by the Earthquake Of March 27, 1964 At Kenai Lake, Alaska By DAVID S. MCCULLOCH GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—A UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1966 For sale by the Superintendent of Documents, US. Government Printing Office Washington, DC. 20402 THE ALASKA EARTHQUAKE SERIES The US. Geological Survey is publishing the results of investigations of the Alaska earthquake of March 27, 1964, in a series of six professional papers. Professional Paper 542 describes the effects of the earthquake on Alaskan com- munities. Professional Paper 543 describes the earthquake’s regional effects. Other professional papers will describe the effects on the hydrologic regimen; the effects on transporta- tion, communications, and utilities; and the history of the field investigations and reconstruction effort. Abstract _____________________ Introduction __________________ Location and physiographic setting _____________________ Bathymetry __________________ Slides and slide-generated waves- Lakeview slides ___________ Depositional units on the delta ___________ Description of slides- - _ Backfill waves ________ Far-shore waves _______ Generation of the slide-induced waves_ - . Index map ________________ 2. Map of depositional units on Lakeview delta __________ . Map of preslide bathymetry off Lakeview and Rocky Creek deltas ____________ . Map of postslide bathymetry ofl’ Lakeview and Rocky Creek deltas ____________ . Fathograms and profiles across Lakeview slide _____ . Map of areas of net erosion and net deposition ofi‘ Lakeview delta __________ . Map showing direction of travel and inshore limit of backfill wave at Lakeview delta ___________________ Page coupe» 03030310 (I) 1. Bathymetric map and profiles of Kenai Lake. 2. Map of wave damage and sliding at Kenai Lake. Page A2 6 CONTENTS Slides and slide-generated 10. 11. 12. . Sketches waves—Continued Lawing slide ______________ Delta materials _______ Description of the slide- Backfill wave _________ Accounts by local residents ___________ Ship Creek slide ----------- Delta materials ------- Waves _______________ Rocky Creek slides-_-_- _ - - - Delta materials ------- Description of slides_ - - Other slides --------------- ILLUSTRATIONS PLATES [Plates are in pocket] FIGURES showing backfill wave height during runup at Lakeview ------------- . Sequential sketches of the generation of the backfill wave at Lakeview ------- Volume, potential energy, scarp angle, and distance of debris travel of the major slides _____________ Map showing changes in the shoreline and direction, height, and inshore limit of the backfill wave at Lawing delta ------------ Map of Lawing showing areas of erosion and depo- sition, and magnitude and direction of waves ------- 1/ 750 Page A12 12 12 12 17 18 18 18 22 22 24 24 Page A9 10 12 13 14 Page Seiching ---------------------- A25 Seiche periods ------------- 25 Eyewitness accounts of seiching ---------------- 28 Tilting of the lake—the probable cause of seiching- 29 Seiche wave damage ------- 30 Spreading of delta sediments_--- 31 Ground fractures on deltas ----- 33 Causes of fracturing ------- 37 Conclusions ------------------- 39 References cited --------------- 40 Page 13. Fathogram and profile across Lawing slide ------------ A15 14. Backfill wave-damaged forest at Lawing ______________ 16 15. Sediment frozen to ice carried ashore by a backfill wave- 16 16. Block of frozen sediment carried ashore by a backfill wave ___________________ 17 17. Ship Creek delta showing areas of erosion and depo- sition, and the direction, magnitude, and runup height of waves __________ 19 18. Fathograms and profiles across and to the east of Ship Creek slide --------- 20, 21 V VI 19 20 21 22 . The slide scarp along the front of Rocky Creek delta- . Fathogram and profile across Rocky Creek slide _______ . Limnogram of seiching on Kenai Lake _____________ . Graph of the power spectral density function versus frequency for the seiche waves __________________ Page A22 23 26 27 CONTENTS 23. Sketch showing wave height during runup on Ship Creek delta _____________ 24. Sketch of a typical railroad bridge __________________ 25. A bridge bulkhead and rail- road tracks ______________ 26. Sheared angle bar bolts in railroad tracks ___________ 27. Split tie and extended guard timbers on railroad track__ Page A30 31 31 32 32 28. 29. 30. 31. 32. 33. Map of surface fracturing on Lakeview delta. _________ Map of surface fracturing on Rocky Creek delta _______ Scarp on Lakeview delta____ Scarp on side of Victory Creek __________________ A rotational slump _________ Fracture along railroad _____ Page A34 35 36 36 37 38 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS SLIDE-INDUCED WAVES, SEICHING, AND GROUND FRACTURING CAUSED BY THE EARTHQUAKE OF MARCH 27, 1964, AT KENAI LAKE, ALASKA The March 27, 1964, earthquake dis- lodged slides from nine deltas in Kenai Lake, south-central Alaska. Sliding re- moved protruding parts of deltas—often the youngest parts—and steepened delta fronts, increasing the chances of further sliding. Fathograms show that debris from large slides spread widely over the lake floor, some reaching the toe of the opposite shore; at one place debris traveled 5,000 feet over the horizontal lake floor. Slides generated two kinds of local waves: a backfill and far-shore wave. Backfill waves were formed by water Most of the loss of life and dam— age to property during the Alaska earthquake of 1964 was caused by waves that inundated coastal com— munities. Some of the waves were of the tsunami type, some Were seiches, and some were caused by submarine sliding including slides from the margins of deltas. Along the coastline of Prince William Sound, the deltas provide almost the only flat land for build- ing sites that is far enough from the steep fiord walls to be safe from avalanches yet close enough to sea level to be useful as harbors. Con- sequently, waves produced by slides along the delta margins as- By David S. McCulloch ABSTRACT that rushed toward the delta to fill the void left by the sinking slide mass, over- topped the slide scrap, and came ashore over the delta. Some backfill waves had runup heights of 30 feet and ran inland more than 300 feet, uprooting and breaking of large trees. Far- shore waves hit the shore opposite the slides. They were formed by slide de- bris that crossed the lake floor and forced water ahead of it, which then ran up the opposite slope, burst above the lake surface, and struck the shore. One far-shore wave had a runup height of 72 feet. INTRODUCTION sume great importance. Many residents of these coastal commu— nities realized immediately that some of the waves that washed over the deltas were caused by slides from the delta edges ( Grantz and others, 1964). During the months following the earthquake, geologists of the US. Geological Survey studying some of these communities (Henry Coulter, Reuben Kachadoorian, Richard Lemke) and geologists studying the shoreline of Prince William Sound (Lawrence Mayo, George Plafker) attempted to dis- tinguish between damage caused by tsunamis, by seiches' and by Kenai Lake was tilted and seiched; a power spectrum analysis of a limnw gram shows a wave having the period of the calculated uninodal seiche (36 minutes) and several shorter period waves. In constricted and shallow reaches, waves caused by seiching had 20- and 30-foot runup heights. Deep lateral spreading of sediments toward delta margins displaced deeply driven railroad—bridge piles, and set up stress fields in the surface sediments which resulted in the formation of many shear and some tension fractures on the surface of two deltas. waves produced by local subma- rine slides. To assist in making t h e s e distinctions, bathymetric contour maps were made in some areas of suspected sliding. Areas of erosion and deposition could be clearly outlined by comparing pre- earthquake and postearthquake bathymetry, and at one place (Kachadoorian, 1965) sliding has been related to a wave that swept over a delta. In a study of earthquake dam- age to The Alaska Railroad, the author and M. G. Bonilla exam- ined landslides from deltas in Kenai Lake. At three places, there was clear evidence that the slides A1 A2 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS had created local damaging waves. As these waves came ashore carry- ing broken pieces of lake ice as much as 2 feet thick they tore blocks of frozen sediment weigh- ing as much as 50 tons from the delta surface. Some waves were 30 feet high and ran inshore for 320 feet. Waves broke off large spruce trees or tore them out by the roots, and drove them inland like batter- ing rams. The author, assisted by Law— rence Mayo, made a bathymetric map of the lake and mapped the wave damage along its shoreline. This study showed that under— water areas of erosion and deposi— tion caused by sliding can be map- ped, that a characteristic wave pat- tern is caused by landslides, that some slides may occur without pro- ducing waves, and that certain parts of deltas may be more sus- ceptible to sliding than others. A seiche formed in the lake, but in most places the effects of the seiche could be clearly distinguished from those of the slide-induced waves. Several deltas were laced with a network of ground fractures caused by the earthquake. Some of this fracturing seems to be re- lated to the stress formed in the surface material of the delta by the lateral spreading of the underly- ing sediments. The author is grateful to his col- leagues in the US. Geological Sur- vey both for assistance in the field and for much helpful discussion. M. G. Bonilla collected some of the field data. David Dawdy helped with the analysis of a linmogram lent to the author by R. A. J ohn- son of the Chugach Electric Asso- ciation, and Richard Singleton of Stanford Research Institute made the computer program for this analysis available. John Ingram of Cooper Landing, Alaska, pro- vided lake-level data for Kenai Lake; Lee Gotch of The Alaska Railroad and Frank Buskie of the Alaska Department of Highways, assisted the author in surveys to establish the amount of tilting in the Kenai Lake basin. LOCATION AND PHYSIOGRAPHIC Kenai Lake lies in a narrow, gla- cially scoured trough near the cen- ter of the Kenai Peninsula in south-central Alaska (fig. 1). The lake is 60 miles south of Anchorage and 80 miles southwest of the cal- 160° 150° culated position of the epicenter of the 1964 earthquake. The lake is about 23 miles long and aver- ages 1% miles wide. Its deepest part is approximately 135 feet be— low sea level and the steep rock 140° 130° 60° 400 MILES 1.—Index map showing location of Kenai Lake. SETTING walls on either side reach altitudes of 3,000—4,000 feet; were the lake open to the sea it would be a fiord. Deltas have been built by creeks flowing down the steep valley walls, by rivers flowing into the eastern arm of the lake, and by a river that enters the lake at the junction of the two western arms. The lake is drained by Kenai River which flows out of the lake at its western end near the town of Cooper Landing, then westward into Skilak Lake and eventually into Cook Inlet at the town of Kenai. . In plan the lake has four straight segments, three of which join at abrupt angles. Although Martin, Johnson, and Grant (1915) showed no faults in this area, the rectilinear shape of the basin suggests that it is structural- ly controlled. SLIDE-INDUCED WAVES, SEICI—IING, GR/O‘UND FRACTURING AT KENAI LAKE A3 The underwater contours of the lake were plotted by making 85 traverses across the lake in a 14- foot skiff powered by an outboard motor. Soundings were made with a portable continuously recording fathometer (Triton, model F—712— A) . A transducer having a sound- ing pulse rate of 50 kc was placed in a few inches of water in the stern of the boat so that it would give the strongest return signal on the fathogram. A speed of approxi- mately 8 miles per hour gave the BATHYM ETRY best signal-to-noise ratio. Sound- ing traverses were run between points on shore identifiable on aerial photographs. Sidewise drift was kept to a minimum by lining up trees with the point on shore at the end of the traverse. The locations of the sounding traverses, the bathymetric map constructed from them, and some typical cross sections of the lake are shown on plate 1. The lake basin is a flat-floored trench bounded in places by eX- tremely steep rock walls. The flat— ness of the floor is interrupted only by deltas and a single bedrock is- land. The lake is 570 feet deep or 135 feet below sea level at its deep- est point, which is about 21/2 miles east of Porcupine Island (pl. 1, sec- tion 0—0") . Subbottom reflec- tions recorded on the fathograms show the lake floor to be underlain by horizontal layers of unconsoli- dated lacustrine sediment that form a smooth bottom surface hav- ing a sharp break in slope at the junction with the basin walls. SLIDES AND SLIDE-GENERATED' WAVES The earthquake triggered slides from nine deltas in Kenai Lake (pl. 2). Some of these slides caused two kinds of destructive local waves. One kind, which will be called a “backfill wave,” was formed by water that rushed to— ward the delta to fill the void left by the sliding mass, overtopped the scarp, and ran inland, inundating the edge of the delta. The other kind of wave, which Will be called the “far-shore wave,” hit the shore opposite the delta. The fact that the far-shore wave struck the shoreline adjacent to the area on which the debris from the slides was deposited suggests that the movement of the slide debris out across the lake floor was directly related to the movement of the water which caused the far-shore wave. Waves produced by the sliding of material that was in large part or wholly subaerial are well known and have been described, for ex- ample, by Miller (1960), Wiegel 796—520 0—66—2 (1964, p. 85—87), and Kiersch (1964). However, waves induced by sliding material that was mostly submerged are neither common nor well described. The single ex- ample known to the author of a slide-generated wave of the back- fill type is the destructive wave that resulted from the earthquake- triggered slide at Port Royal, Jamaica, in 1692 (Heath, 1748; Link, 1960). Similarly the author has found only one example in the literature of a disturbance of water related to the movement of slide debris across the bottom of a body of water which may be simi- lar to the suggested generative process of the far-shore wave (Heim, 1932, p. 42). Slides from deltas also occurred without producing detectable waves. One such slide carried away 260 feet of railway roadbed on the delta of Rocky Creek. Being ice-laden, the Kenai Lake waves probably caused more de— struction than waves of similar heights in Prince William Sound. Between the time of the earth- quake and the time that the wave heights were measured the lake rose 51/2 feet. Wave heights cor- rected to the lake level at the time of the earthquake are shown on plate 2. This rise in lake level obscured the evidence of minor wave action along the eastern arm of the lake and as far west as Cooper Landing that was clearly visible on aerial photographs taken soon after the earthquake. LAKEVIEW SLIDES At Lakeview there were two slides from the delta built by Vic- tory Creek (also called Victor and Vicory Creek) into Kenai Lake— a north-facing and a larger west- facing slide. These slides and the waves they generated are described first because the relationship be- tween the slides and their waves is clearer here than elsewhere. A4 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS DEPOSITIONAL UNITS ON THE DELTA At least three depositional units can be recognized on the Lakeview delta surface on an aerial photo- graph taken in 1951 (fig. 2). The oldest of these, unit 1, is the high— est; it has a somewhat irregular surface and is covered with mature spruce trees. The surface has been so modified since deposition that stream channels which must have been present at one time are no longer visible. Unit 2, the next younger unit, is somewhat lower and is covered by low brush; there are many abandoned stream chan— nels. On a photograph probably taken in 1911 (Martin and others, 1915, pl. 27A), unit 2 was bare of vegetation and its .surface was laced with distributary channels. Unit 3, the youngest unit, was de— posited on the northern edge of the delta; its deposition seems to have started just prior to 1911. When photographed in 1951, this unit was still being formed at the mouth of the creek and was still bare of vegetation. The slides removed some of units 1 and 2, and nearly all of unit 3. All three units, as exposed in the major slide scarps or in scarps of small slumps on the surface of the delta, consist primarily of sandy gravel of subangular to sub- round platy pebbles and cobbles, and some boulders of metasedi- ment. On the major slide scarps the exposed beds are thick and lensing, and contain only an 0c- casional thin sand lens of short lateral extent. DESCRIPTION OF SLIDES The location of the major slide scarps (fig. 2) suggests that the position of the scarps is not gov- erned in any obvious way by the boundaries of the three depos- 1500 FEET EXPLANATION V ’/ %A Unit 3 Lowest surface. Continuous with modern flood plain. Deposition started shortly before 1911. Bare of vegetation in 1 Intermediate surface. Covered with abandoned channels. Bare of vegetation in 1911. Low brush cover in 1951 Highest surface. No old stream channels visible. Covered with mature spruce 2.-——Distribution of depositional units on Lakeview delta. tional units. The north-facing scarp cuts across all three units and the west-facing scarp cuts across two units. Sliding seems to have been localized in areas that protrude from the edge of the delta. Such areas are probably more susceptible to sliding than adjacent areas because they are the parts of the delta having the largest amount of material bound- ed by the shortest possible surface- of rupture. The curved surface of rupture along which the sliding occurred (fig. 5) suggests that the west-fac- ing slide was a rotational slump (Varnes, 1958, p. 21). It is well known that earthquakes can cause rotational slumps. Terzaghi (1950, p. 89-91) and Taylor (1948, p. 452) have shown that horizontal accelerations in the direction of the free slope increase the shearing stress along the potential sliding surface of rotational slumps. The lateral spreading of the delta sedi- ments suggests that there may have been a considerable loss of strength in the sediments. This would un- doubtedly have reduced the resist- ance to failure along a potential surface of rupture and would have promoted the sliding. The location of the debris from the Lakeview slides can be ap- proximated by comparing the pre- sliding and postsliding bathyme— try (figs. 3 and 4) . The presliding bathymetry was reconstructed by extending contours from areas in which there was no evidence for sliding. These 0 o n t o u r s were drawn by assuming that the delta front was a smooth curve and that 3.—Pres1ide bathymetry off Lakeview and Rocky Creek deltas. Contours solid in areas in which no sliding is thought to have occurred and dashed where reconstructed in areas of ero- sion and deposition. ——) the lake floor was as flat here as elsewhere. The west-facing slide cut an arcuate scallop from the delta. The sliding took place along an irregular curved surface, its lower part extending below the lake floor. The scarp has a slope of about 44° in the upper 15 feet, 28° to a depth of 75 feet, and finally about 11° to a depth of 285 feet. It then rises in a reverse slope at an angle of about 13°. During deposition the slide ma- terial must have spread laterally because there is a great difference between the volume of material re- moved from the delta and the ma- terial deposited at the base of the 4.——Postslide bathymetry off Lakeview and Rocky Creek deltas showing di- rection (arrow), magnitude of dam- age (on arrow shaft), and runup heights (at arrow point, in feet) of slide-induced waves. ——) SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KENAI LAKE A5 Lakeview delta Rocky Creek delta 1000 o 1000 2000 3000 FEET I | | I | I I | | | I I I I DIFFERENT CONTOUR INTERVALS ARE USED (Limit of irregular battom topography \N Rocky Creek delta 0 1000 2000 3000 FEET LLJ I | I | I | | I I I I DIFFERENT CONTOUR INTERVALS ARE USED A6 THE ALASKA EARTHQUAKE, MARCH 27, 19642 REGIONAL EFFECTS FATHOGRAM RECORDED ON LINE 94 E W 0' —————— \ 0' 100' 100, 200' _____ 200V 300: —————————— 3m! 400’ 400' 0' 500’ 1000' 1500' 2000' 2500' 3000' 3500' 4000' 4500' 5000' 5400I PROFILE ALONG FATHOMETER LINE 94 W \ 0' \\\\1Preinde profile 100' \\ 200' _‘** 300' 400' 500’ 500' 0’ 500' 1 000’ 1 500' 2000' 2500' 3000’ 3500’ 4000’ 4500' PROFILE ALONG FATHOMETER LINE 72 Note: Because of the large vertical exaggeration on the fathog’rams, these profiles. which have no vertical exaggeration, are also shown 5.—Fathograms and profiles of Lakeview delta slides. Lines of profiles are shown on figure 4. SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KENAI LAKE A7 scarp (fig. 5). The extent of lat- eral spreading can be estimated two ways: first, by noting the area of deposition indicated by the pre- earthquake and postearthquake bathymetry; and second, by out- lining the areas of rough bottom topography shown on the fath- ometer profiles (fig. 6). The boundary of the irregular topogra— phy coincides approximately with the western edge of the deposi— tional area, but the eastern bound- ary lies within the zone of erosion. This relationship suggests that the lower part of the slide scarps are covered by slide material. BACKFILL WAVES Powerful backfill waves rushed back up onto the delta surface af- ter each of the slides at Lakeview. . The wave direction was determined by noting which sides of the trees had been debarked, how trees had fallen, how the brush had been combed, and, in one place, by the direction in which a house had been carried from its foundations (fig. 7, next page). The waves broke off some spruce trees near the ground and uprooted others—some as much as 21/2 feet in diameter at the base. Trees that had traveled far— thest in the waves were shorn of their limbs and stripped of their bark; others that had traveled shorter distances retained most of their limbs and had been only par- tially stripped of bark. A roughly rectangular block about 3 by 15 by 20 feet, composed of bedded sandy gravel with clasts up to boulder size, was found 40 feet from the scarp, approximately halfway between lines 2 and 3 on figure 7. Assuming a density of 1.75 (50 percent of the total as saturated void space with a density of 1.0 and 50 percent of the total as sediment with a density of 2.5), this block would have a weight of about 50 tons. This block of sedi- ment, like the extensively battered trees, was probably torn from the surface of the slide block by the in- coming waves, for no place was found on the delta surface from which the block could have come. Although the block did not appear to be frozen when examined on the 8th of August, it probably was fro- zen at the time it was carried in by the waves; without interstitial ice the block would undoubtedly have distintegrated. La keview delta $/ em __/ 1000 o 1000 20 l | I | 3000 FEET Limit of irregular bottom topography Rocky Creek delta 6.—Areas of net erosion and net deposition off Lakeview and Rocky Creek deltas as indicated by difierences between presliding and postsliding bathymetry. A8 THE ALASKA EARTHQUAKE, MARCH 27, 19642 REGIONAL EFFECTS House and foundation from APPROXIMATE SCALE 500 O 500 I I | | | I I I 1000 FEET which it was carried by wave \“~....__/1951 shoreline 7.—MaIp of Lakeview delta showing direction of travel and inshore limit of slide-induced bIackfill waves. The five traverses on which the height of wave (damage was measured are shown in figure 8. A log house was carried by the backfill wave for more than 200 feet from the concrete foundation to which its sill had been bolted (fig. 7). The walls were demol— ished but the roof structure hav— ing a braced triangular cross sec— tion, although distorted, remained intact. The roof and the remains of the walls were deposited by the wave in a tangle of uprooted and broken spruce trees. The ground was snow covered at the time of the earthquake and the wave washed away the snow, leav- ing a clear record of its inshore limit. This inshore limit is drawn on figure 7, as shown on aerial photographs taken a few days after the earthquake. The debris transported by the waves de- creased in size inshore—the far- thest inshore deposit consisting of small pockets of sand and pebbles, twigs, and small turf blocks. Many blocks of ice, some to which sand and pebbles were frozen, were carried to the inshore limit of the wave-washed area. To determine the altitude of the crest (runup) of the backfill wave as it ran up on shore, the height of the damage to trees was re- corded. Four traverses were run from the toe of the scarp in the direction of wave travel on the delta (figs. 7 and 8). Horizontal distances were determined by pace, and altitudes were measured with a stadia rod and hand level. Al- titudes were also recorded for the base of the damaged trees and for the heads and toes of small scarps. The lake level was 437 .0 feet when the traverses were run, and these altitudes have been adjusted in fig- ure 8 to the altitude of the lake at the time of the earthquake (431.6 ft) . The height of the wave above the lake surface may have been somewhat greater or less than in— dicated by these altitudes, for the lake might have been seiching at the time the wave was formed. The crest of the backfill wave was about 10 feet above the delta sur- face as the wave crossed the scarp. The height of the wave decreased inshore but the altitude of the highest damage rose and fell some- what with the changing slope of the delta surface. FAR-SHORE WAVES The far-shore wave hit the shore across the lake from the delta. In the middle of the wave-washed area the runup reached 25—35 feet above lake level and decreased to 11—13 feet to the sides. The orientation of the damage indicates that there was considerable varia- tion in the direction of the wave travel. In the most severely dam- aged area the waves tore vegeta- tion from the bedrock shoreline, and broke off large spruce trees or stripped off their bark and lower limbs. GENERATION OF THE SLIDE- INDUCED WAVES Because the relationship be- tween slides and waves is clearer at the Lakeview delta than else- where, the Lakeview slides will be used as the model for the discus— sion of slide-generated waves. It is believed that the waves were generated as described below (fig. 9, p. A10). The west-facing slide seems to have occurred first for on the point of land between the two scarps, where damage to trees indicates that the area was washed over by waves from both slides, the debris SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KENAI LAKE A9 30’ 20' 10' 0. | 250' 200’ 150’ 100’ 0’ Line 2 30’ 30, 20’ 20; 10’ or I l 300’ 250’ 30’ 20’ 10’ 250’ 30’ 20’ 10’ 8.—Upper limit of backflll wave damage along traverses run parallel to wave travel direction on Lakeview delta. Location of traverses shown on figure 7. A10 THE ALASKA EARTHQUAKE, MARCH ‘M’U/lg‘ (11%“ 27, 19642 REGIONAL EFFECTS 9.-—Sequential sketches of the slide-generated backfill wave that ran back up onto Lakeview delta. A, Preslide delta. B, Start of slumping and water rushing down into the void left by the sinking slide. high water just 01f the slide scarp. D, High water ran inland over the delta, radiating away from the high-water welt. As shown in D, the second slide near the mouth of the creek occurred after the larger slide. is alined as it would have been by a wave from the north-facing scarp. Thus the north-facing slide occurred after the west-fac- ing slide. As the western slide mass submerged, water rushed into the void behind it. The inrush of water must have taken place almost simultaneously with the downward movement of the block, as shown by the extensively damaged trees and blocks of sediment found above the scarp that had been torn from the slide areas. Being com- posed of granular material the slide mass was probably broken by progressive slumping as it moved. As the whole slide mass finally submerged, water rushed in from all but the scarp side. The level of the inrushing water would have been below the general lake level because it must have run down slope to fill the void. This prob- ably was why no evidence was found for damage caused by water rushing into the slide area. The water rushing toward the delta formed a wave that overran the edge of the delta. The direction in which the wave traveled over (7, Inward-rushing water formed a welt of the delta surface (figs. 7 and 9D) shows that the water was spread- ing radially from an area that lay just off the scarp, suggesting that the inward-rushing water formed a welt of high water in that locae tion (fig. 90). This high-water welt probably resulted from a com- bination of (1) the convergence of the inward-rushing water and (2) the upward deflection of this water by the slide scarp. In model studies, in which waves were impulsively generated (J ohn- son and Bermal, 1949; Wiegel, 1955; Prins, 1957 ) waves of the SLIDE-INDUCED WAVES, SEICI—IING, GROUND FRACTURING AT KENAI LAKE A11' backfill type have either not been observed or not described. How- ever, the models were generally de- signed for the study 0 ‘ gravity waves that traveled awayf from an initial disturbance. ‘ The far-shore wave may have been formed in the following way: As the landslide block moved downward into the lake, it dis- placed water ahead of it. Once set in motion by the landsliding, some of the sediment probably be- came entrained in the water. En— trainment could have been rapid, because the sediment lacked clay- sized material and was therefore noncohesive. Furthermore, being largely below lake level before slid- ing, it was water saturated. The moving water, part of which con- tained entrained debris, ran across the lake floor to the toe of the far shore, ran up the sloping far side of the lake basin, and broke above the surface of the lake as a far- shore wave. This relationship be— tween the far-shore wave and the moving water that probably as- sisted in the transportation of the debris is suggested by the fact that the far-shore wave occurred only along the shore that was adjacent to the area of the lake floor on which debris was deposited (fig. 4). Thus, the far-shore wave may have been produced by water forced ahead of the entrained de- bris, in much the same way that violent winds were forced ahead of subaerial landslides (W i t k i n d , 1964) and forced ahead of the giant wave produced by the over- topping of Vaiont Dam (Kiersch, 1964). This mechanism was sug- gested by Heim (1924,p. 22) : * * * When a wind from a landslide can sweep down an entire forest and can carry people and cattle several hundred meters through the air, so a subaqueous slide of the larger kind on a steep slope must be accompanied by a tidal wave (Flutwelle) that can be propagated over 796—520 0—dB6———-—3 and beyond the area of deposition, can tear out sediments, form ripple marks and disturb the benthos. The orientation of damage caused by the far-shore wave shows that there was considerable variation in the direction in which the water was traveling as it came onshore. This variation in direction raises the possibility that the movement of the slide debris was not unidi- rectional on the lake floor, and that if one had the opportunity to study current—oriented features in the slide debris one might find some variation in their directions. A possible alternative cause of the far—shore wave is that it was produced near the slide and then traveled across the lake as a wave and hit the opposite shore. Slides have been known to cause such waves (Jones and others, 1961), but, for several reasons, this alter- native seems unlikely for the Lake- view slide. The correspondence between the edge of the deposi- tional area on the lake floor and the wave-washed area would have to be a fortuitous rather than a causa- tive relationship. The fact that this same relationship occurs in three other areas suggests that it is causative. Furthermore, if the far-shore wave had been generated near the slide area, one might ex- pect that, allowing for refraction, the directions of wave travel would converge toward the point of ori- gin. However, the directions do not seem to show any such pattern; rather, they show a wide variation in direction of travel. A disturbance of water possibly analogous to the mechanism pro- posed for the far—shore wave was caused by a slump described by Heim in 1932 (p. 42). In 1875, in the town of Horgan on Lake Zurich, Switzerland, a part of the shoreline subsided. The floor of the lake, which had been at a depth of 135 meters, was raised 1—2 meters, and on the opposite shore the water became disturbed or turbid (triibte). Kuenen (1950, p. 49), apparently describing ob- servations by Heim, said that the turbulent water Suddenly appeared as boiling masses of muddy water. Kuenen concluded that, because the slump could not rise much above the deepest part of the lake and because sounding showed that the main mass of sediment spread out on the deeper part of the lake floor, a watery turbidity current must have been developed that had suflicient momentum to cross the lake floor and be carried up the opposite slope to the surface. Al- though Kuenen attributes the up- welling at Horgan to a turbidity current, there is no direct evidence that such a current was created by the Lakeview slide. The coarse- ness of the sediments involved in the Lakeview slide and the fact that the sheet of debris has an abrupt front and a relief of about 60 feet (fig. 5) argue against trans- portation by a turbidity current. The wide lateral spreading of the slide debris is more likely to be due to some degree of entrainment of the sediment into water par- tially set in motion by the displace? ment of the slide mass. Even a small degree of entrainment would have reduced internal friction; this reduced friction would have resulted in a lower angle of repose, which would have allowed the sediments to move down a more gentle slope. In figure 10 (next page) values are given for the volume, potential energy, angle of slope, and distance traveled for the Lakeview and the three other slides The volumes and potential energies are calcu- lated for a vertical slice having a thickness of 1 foot as measured from the reconstructed preslide bathymetry. A12 THE ALASKA EARTHQUAKE, MARCH 27, 19642 Water level REGIONAL EFFECTS lake floor Normalized Volume Potential potential ener- Scarp slope Depth Lake-floor Distance of Slide (cu ft) energy gy (potential (angle A, in (Z, in feet) slope (angle debris travel (10*? ft lbs) energy+vol- degrees) B, in degrees) (D, in feet) urne) Lakeview ______________________________________________ 226,875 1 17,285.97 76, 191 10. 5 220 3 1, 280 Lawing .......... 74, 125 9, 390. 15 126, 679 20 400 2 2, 100 Ship Creek ...... __- 198, 250 28, 079. 59 148, 372 25 520 0 4, 750 Rocky Creek ........................................... 66, 375 6, 998. 55 105, 439 25 380 2. 5 600 1 Because some material was moved upward by rotation at the Lakeview slide, the potential energy of the material raised to the level of the lake floor has been subtracted from the potential energy of the material above the level of the lake floor. 10.—The volume, potential energy, scarp slope, and the distance traveled by the slide debris of the four major slides. The volume and potential energy are calculated for a vertical slice having a thickness of 1 foot as measured from, reconstructed preslide bathymetric contours along fathogram line 94 (fig. 5), profiles A—A’ (figs. 13, 18), and fathogram line 89 (fig. 20). LAWING SLIDE DELTA MATERIALS The Lawing delta is at the northeast end of the lake (pl. 2). Trial River and Ptarmigan Creek flow across the delta, but the shape suggests that the delta was built principally by Ptarmigan Creek. The slide removed the most re- cently deposited material at the mouth of Trail River. The area had a cover of low brush and formed a slight bulge on the shore- line of the delta (fig. 11), accord- ing to 1951 aerial photographs. The sediments exposed along the scarp and in blocks carried onto the delta surface by the wave con- sist of stratified sandy pebble gravel and pebbly sand. The pebbles are tabular and subangular to subround. Although cobbles are present, the sediment is gen- erally finer than in the Lakeview delta. DESCRIPTION OF THE SLIDE The amount of erosion caused by the sliding and the distribution of the slide debris can be approxi- mated, as they were at Lakeview, by comparing presliding and post- sliding bathymetry. The post- slide bathymetry ofi Lawing (fig. 12, p. A14) shows a reentrant in the contours downslope from the slide scarp probably caused by the sliding. Farther downslope, and extending nearly to the deepest part of the lake, is an area of irreg— ular topography. A cross section on which the postslide and recon- structed preslide bathymetry are compared shows that the slide re- moved a block of sediment above the scarp as much as 80 feet thick (fig. 13, p. A15). Before the slide, the steepest part of the delta had a slope of about 22°; after the slide it was about 245°. Thus here, as at Lakeview, sliding locally steepened the front of the delta. At Lakeview the amount of deposition on the lower part of the lake changed the bottom bathymetry enough that the dep- ositional area could be reason- ably well delineated by comparing postslide and preslide contours. At Lawing, however, the amount of material involved in the sliding was considerably less (fig. 10), and only a suggestion of a depositional area is indicated by the slight southward displacement of the 420- and 440-foot contours (fig. 12). If there were no other rea- son to believe that there had been deposition in the area, this evidence would be inconclusive, but as at Lakeview, the probable deposition- al area lies Within the area of ir- regular topography. Isolated topographic highs in this area sug- gest that it is unlike the adjacent smooth lake floor; and, as at Lake- view, there was wave damage on the shore opposite the scarp, in front of the depositional area. A comparison of areas of erosion and deposition and a comparison of cross sections of preslide and post- slide bathymetry (figs. 12 and 13) indicate that there was consider- able lateral spreading of the slide debris as it moved down and out over the nearly level lake floor. The fact that slide debris traveled about 3,000 feet on a slope of about 1.5° suggests that there may have been some entrainment of the de- bris in the water that was set in motion by the sliding. BACKFILL WAVE The two waves at Lawing were similar in pattern to the waves pro- duced by the sliding at Lakeview. SLIDE-INDUCED 'HHLH.‘ .1“ Approximate location of 3 boathouse and shed destroyed by wave APPROXIMATE SCALE 500 o 500 L i . | I 5' -Break knfiléfie dong former? fiverbank (\tnshore [Minot-wave ' Ribbons ' house WAVES, SEICHING, GR‘OUND FRACTURING AT KENAI LAKE Fractured ground 1000 FEET l 11.—Map of Lawing area showing the slide scam and former shoreline, and the direction (arrow), magnitude of damage (on arrow shaft), and runup height, in feet (at arrow base), of the backfill wave. The inshore limit of the wave is indicated by the dotted line. Houses of two eyewitnesses are noted. A13 ' The wave that struck and over- ran the edge of Lawing delta spread radially as it came on shore as did the backfill wave at Lake- view. The wave directions (fig. 11) indicate that as the wave traveled to the west and south from the scarp it was refracted by the shallow water along the delta. The wave was carrying blocks of ice 16—20 inches thick (as reported by Mr. Gibbons, a resident at Law- ing), and as it came on shore it broke off large spruce and cotton- wood trees (fig. 14,1). A16). Some of the blocks of ice that had been frozen to the beach carried sedi— ment on their undersides (fig. 15, p. A16). Frozen blocks of sedi- ment were also carried ashore by the wave; the largest of these was a flat-topped, pyramidal-shaped block approximately 14 feet wide at the base and about 41/2 feet high (fig. 16, p. A17). This block still had a considerable amount of in- terstitial ice when seen on the 15th of May. The wave also carried driftwood, cobbles, pebbles, and unfrozen turf. A flower garden that lay just at the edge of the wave—washed area was covered by several inches of sand. The residents at Lawing were fortunate in that the wave stopped just short of two occupied houses; the only buildings destroyed were a shed and boat house. Along the western edge of the wave-washed area the w ave stopped at the foot of a break in slope. In the central part of the area, where wave heights were greatest, the wave ran back up the valley of Trail River for a distance of about 800 feet and overflowed the low West riverbank. Aerial photographs taken by the US. Army the day of the earthquake show a pile of broken ice, higher than the level of the river, filling the channel from bank to bank at the inshore limit of the Wave. A14 Area of probable erosion Probable distribution of slide debris 500 520 Bathymetric contours 1951 shoreline and reconstructed preslide bathymetric contours THE ALASKA EARTHQUAKE, MARCH 27, 1964! EXPLANATION Limit of, irregular bottom topography 21° /\_/L\/ Direction of wave travel, magnitude of damage (number on shaft), and runup height, in feet (at arrow point) Inshore limit of wave REGIONAL EFFECTS GA ‘9; ‘9 1000 0 1000 2000 FEET I | | | l I l l l l I I I DIFFERENT CONTOUR INTERVALS ARE USED 12.~Probable distribution of erosion and deposition that resulted from the slide and the direction, magnitude, runup heights, and inshore limit of waves. Profiles along A—A' in figure 13. SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KENAI LAKE A15 0' . 0' .i 100' f 100' A J V d 200' a- ’ 200' 300’ 3007 FATHOGRAM RECORDED ON LINE 64 O’ \ 100, \\\ . . , \/Presl|de profile 200 \\ 300' \ x _ 400’ Area of irregular topography 4000’ 4500’ 5000’ 2000' 2500' 3000’ 3500' PROFILES ALONG LINE A-A’ Note: Because of the large vertical exaggeration on the fathogram, this profile, which has no vertical ex— aggeration, isalso shown Line of profiles: is shown on figure 12. Note that when the depth on the 0’ 500’ 1000’ 1500’ 13.—Fathogram and profiles of Lawing slide. fathogram exceeds 300 feet, the recording is shifted upward and the 300-foot—depth is placed at the original zero depth. A16 THE ALASKA EARTHQUAKE, MARCH 27, 19642 REGIONAL EFFECTS 14.—Area washed over by the backfill wave at Lawing. The wave traveled from left to right, breaking down and debarking the trees. Brush in the background was bent to the right. The only branches left on the spruce tree beside the man are on the lee side. The pile of gravel in the left foreground was probably frozen at the time it was carried ashore by the wave. 15.—A large block of lake ice to which sediment was frozen when carried ashore by the wave. SLIDE-INDUCED WAVE-S, SEICHING, GRJOUN'D FRACTURING AT KENAI LAKE A17 The estimated runup heights along the shore are shown in figure 11. A maximum of about 30 feet occurred in the area behind the scarp, and heights decreased irreg- ularly away from the scarp. An abrupt break in slope bounding a former course of Trail River (fig. 10) may have locally increased runup heights, because runup heights increase with the steepness of the bank hit by the wave (U.S. Army Corps Engineers, 1961, p. 89). ACCOUNTS BY LOCAL RESIDENTS To the author’s knowledge the first person to study the Lawing slide and its associated wave was Hadley Roberts, of the U.S. Forest Service, whose house is just be- yond the inshore limit of the wave on the Lawing delta (fig. 11). Mr. Roberts deduced from the radiat- ing pattern of damage to trees that the slide had caused the wave. He drew a sketch map showing the preslide and postslide shorelines, the direction of wave travel, and the inshore limit of the wave. The map of this same area prepared by the present writer includes addi- tional observations and is con- structed from preearthquake and postearthquake a e r i a 1 photo- graphs, but it differs only slightly- from the map made by Mr. Roberts. Fortunately, there were two eye- witnesses to some of the wave ac— tion at Lawing—Mr. Frank Gib- bons and Mrs. Hadley Roberts. Both live on Lawing delta and both were at home at the time of the earthquake (houses shown on fig. 11). Mr. Gibbons was in his living room at the time of the earthquake. The first tremor, a strong north- south motion, knocked him to his knees in front of a large window facing west over the lake. There 16.—Large block of bedded sandy pebble gravel carried ashore by the wave at Lawing. The slide scarp forms the boundary of the beach to the right; debarked trees in the background. was a pause in the shaking and then the motion shifted to east- west. During the east-west shak- ing he saw a wave about one-fourth mile from shore. It is not clear if he saw the wave form or if it had already formed when he first saw it. As the wave came toward him the ice broke across its rising front “like a patchwork quilt.” The wave hit the shore and ran up around his house. As the water receded, Mr. and Mrs. Gibbons tried to leave their house, but the ground was shaking so Violently that they couldn’t get out. A sec- ond smaller wave ran up on shore about a minute after the large wave. Then they left the house. The following is a part of a let- ter written by Mrs. Roberts. On March 27, 1964, at 5 :36 pm. I was standing at the stove getting ready to put a lid on a pan of potatoes. The kids (Carol and Bruce) were near me and as the earthquake started their eyes popped open wider and they came to me. As it got worse we huddled together. I think at this time I remember hearing cupboard doors banging and dishes crashing to the floor. I could hear the house squeaking'and groaning and we were swaying so very hard that I de- cided to get out. I noticed that the propane was still burning under the potatoes so I turned it off. The crash- ing dishes were between me and "the back door so we stepped out the garage door. The kids were both crying and screaming at this time and I really can’t remember too many details. In the garage Carol screamed that she had lost her shoe as we were going out the door and I kept thinking I had to get back in there and get her shoe before we could go out in the snow. All this time the car was swaying tremendously and also sliding back and forth on the con- crete floor. I was afraid of being crushed between it and the wall so was really watching it closely. Carol has reminded me that while we were in the garage we saw a gallon bottle of milk tip over and we were standing among the broken pieces of glass and in the milk part of the time. This bothered Carol since she was missing a shoe, and is no doubt why she remembered it and I didn’t. After a while in the garage; and I haven’t ‘the slightest idea how long we were finere—dt started letting up a little so stepped back into the kitchen just iri time to see some big wave action on the lake * * *. That ice on the lake was breaking up so fast, and oscillating so and then all of a sudden I could see lots of black A18 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS muddy water and it was at a much higher level than normal; in fact, it was almost up level with the bank. At this time I thought I saw a life jacket floating by but quickly decided that it couldn’t be (but it was, as the boat house had just been ripped to shreds). Right behind the jacket was a huge wall of water and I noticed that the splashes were going ever so high. Then remembering what had happened at the Madison Canyon Earthquake (near Yellowstone), in 1959, that we had just visited this past summer, I decided to get away from the waters edge. I also saw a snowslide across the lake and from the looks of the flying snow it 'iooked like a real big one. The wave was going parallel to the beach and our house, and was going south. We walked past the fireplace (the kitchen way was too littered to walk through) over to the basement door, and we must have walked through de- bris but I really can’t remember doing it. The kids’ boots were in the base ment so I had them stand at the head of the stairs while I ran down for them. I remember them screaming and crying for having been left alone and I think Carol even started down the steps after me. But I hurried and ran back up the steps two at a time. The house was still rocking some and that was scaring them. When I first saw the wave action I think I said something like “we'd better get out of here before that water gets us”. This scared Carol, too, I think, as she kept saying that we were going to die. I got their boots and coats on and went to the front room closet for my coat and when I passed the window I looked out and didn‘t see any sign of a wave but did notice the water level was much higher and flowing quite close to the house. We went out the back door and I carried Bruce and Carol and ran and we beat it for the bunk- house where we found Freddy. The bunkhouse isn’t any higher, but is cer- tainly a lot further away from the waters edge. Both observers noted a pause after the initial shaking, and Mr. Gibbons indicated that the direc- tion of shaking changed after the pause. Also after the pause Mrs. Roberts saw “wave action” and oscillation in the lake and rising water. During the second period of shaking both saw a large wave; it is not certain, however, that both saw the same wave. The wave came ashore around Mr. Gibbon’s house from an estimated distance of a quarter of a mile to the west. Mrs. Roberts saw what may have been the same wave as a “high wall of muddy water” running south along the edge of the delta from the direction of Mr. Gibbon’s house. If this “wall” was the slide- induced wave, it suggests that the slide occurred during the latter part of the shaking after a seiche had been initiated in the lake. SHIP CREEK SLIDE DELTA MATERIALS A slide at Ship Creek delta re- moved a low protruding strip of sediment about 1,400 feet long and 250 feet wide from the delta edge. On 1951 aerial photographs the slide area was seen to be bare of trees and was probably composed of the youngest delta sediments (fig. 17). Sliding also is suggested at the eastern edge of the delta by the radiating pattern formed by incoming waves immediately in- shore of a reentrant in the bathy- metric contours. No scarp was found here, however, and if sliding did occur, it involved material that was entirely submerged. The sediment in the delta was not ex- amined, but because Ship Creek has a high gradient having a fall of 3,000 feet per mile, the sedi- ment is probably relatively coarse. In front of and somewhat to the east of the delta the topography of the lake floor is slightly irregular. Comparison of fathometer profiles across this area and across the flat lake floor about 5,000 feet to the east shows that the slide debris can be readily recognized (fig. 18, p. A20—A21). In the slide area the normal horizontal bedding shown by subbottom reflections has been destroyed, and if the lake floor was as flat here before the slide as it is now to the east and west, then the sliding can be as- sumed to have formed 50 feet of re- lief. How much of the relief is due to deposition and how much to erosion is difficult to determine from a comparison of reconstruct- ed presliding and postsliding bathymetry because there is good reason to suspect that erosion did occur. The fact that subbottom re- flections were recorded in the un- disturbed areas by the high-fre- quency, low—power signal from the fathometer suggests that the bot- tom sediment is very soft. Soft sediments are typical of fiords where there is quiet-water sedimen- tation of fine sediment (Bennett and Savin, 1963). Richard Mal- loy of the US. Coast and Geodetic Survey examined the records and said that he had seen similar rec- ords made with a comparable fathometer where the sediment was so soft that a sounding lead easily penetrated it and recorded depths several feet deeper than the fath- ometer. Such soft sediment should have been easily eroded. Some of the slide debris may have traveled as much as 7,000 feet, or more than twice the distance traveled by the material in the Lawing slide. This distance is not surprising, however, for not only was more material involved in the slide at Ship Creek, but the lake here is nearly 100 feet deeper. Thus the available energy and the resulting momentum of the slide debris would have been much greater at Ship Creek (fig. 10‘). WAVES There was a considerable amount of wave damage in the Ship Creek delta reach of the lake (pl. 2). On the western margin of the delta, damage was clearly due to a wave produced by seicheing (see p. A30) , but the damage on the rest of the delta and on the far shore was A19 SLIDE-INDUCED WAVES, SEICHING, GR'OUND FRACTURING AT KENAI LAKE .wH gnaw dc 852% 3 Jo>§> mo 0393a ac 85A $353 «0 mange: @555 was . ENE: $9256 95? mo $296.5 mo 2—5 _ 0:: . menu—33>: :2: «Ed 305 «0 5am ‘i 352“ Bob: 93 33 E Jsmmw: 95:5 93 “Swag :o 33:55 wMaEmc me 33:55: 49:»: 933 MO #55025 LVIV @595 no on: one 6:.» ma whrwfl E 859% was "Emmy: @mwfiwc owEEMxS . €58un $5 wax .noflwwoawu HEN nowmomo mo mumps MEBoam 336 mefio Saw Scum 35ml: 0mm: mm< w4<>mm§i KJOPZOU hzwmmuma _ mm“. 0009 000m OOON Good 0 000— \/l\_\/\ m _ m— .EQEMOQS 5333 31395 mo #55 953:8 £52553 wEin coauiamcovou HE“ 3:32.? 32 353:8 oEwEfi—umm own new $53 as: we 553.5% £32m .8395 2930.5 «a u2< ZO_._. 0: gm: 52:3 .3503 $9: .mEEMOEE m5 :0 :ozwgmmmmxo _wo_to> wufi. 05 we omzmoom ”302 mm mZ_._ KWEEOICE GZOA< 57.0mm .0000 60mm boom ‘oomv \ooow .oomm .ooom ‘oomm boow .oomH bog .oom _o .005 . _ _ _ _ _ _ _ _ _ .OON .oow bOm “\ .oo¢ ‘ .oom ‘ ‘ ‘oom 1“ .ooH ‘ ‘ .o x, mm mz: mmhmzorbfi. ozo._< MEEOWE oo.Moo¢m boom ooomv .ooov .oomm .ooom .oomm boom 603 be? .oom ~o . _ _ _ _ q . _ _ _ .000 how J \‘ 63 \“ ‘ 68 a ‘ .ooN ‘ boa , ‘ .o HFQ m2... @292 mmjiomn. con .oomw .oooo Koomm .ooom oomv ‘ooov ooomm boom .oomN boom ‘oom H .009 .oom .0 ~ _ _ _ _ _ _ _ q _ _ . _ _ bow “ ‘ 11 ‘1 w‘ “““““““ “ z ‘ 1 1 “1 ‘\1‘ 1 w‘ “ 60m x>[‘ ‘ l ulll? / i: “““ .WWM TIE?! “0 «m3 om mm 502.: ma 5:; Eamhuoaou .miwwElmtnwt 235 r!“- /o/// ‘ ‘ . x5311: ‘ ‘ /O/ .oom 2:95 0332.“. \z//.,/ 1 62 ‘\ //v// .o /i¢II LN bow ‘ bow bom : .oov ‘oom boN .ooH A22 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS probably caused by slide-generated waves. If these waves are analo- gous to those caused by the Lake- view slides, the wave that over- topped the scarp was a backfill wave and the wave that caused the severe damage opposite the delta and extending to the east was a far-shore wave. Where the far- shore wave hit a steep bank, it had a maximum runup height of 72 feet—the highest along the entire lake. The wave damage caused by the far-shore wave was similar to that at Lakeview in that generally the intensity of the damage de- creased toward the margins of the wave-washed area, and the wave directions were not uniform. The shoreline hit by the far- shore wave extends to the east and West of the area of deposition, as shown on figure 17. The deposi- tional area is drawn conservatively however, and debris might occur farther west—nearly to the fatho— gram that shows no irregular to— pography. The position of the western edge of the area of prob- able erosion also suggests this pos- sibility. Control is somewhat bet- ter on the eastern side of the de- positional area, and debris prob- ably does not extend much farther east. Despite these small discre- pancies, the general relationship between debris and the far-shore wave established at the Lakeview slide seems also to exist here. ROCKY CREEK SLIDES The slides that occurred at the delta of Rocky Creek (also called Boulder Creek) just north of Lakeview (pl. 2) differed from those previously described. Many small slumps trimmed back the shore of the delta by as much as 180 feet (fig. 29). One slump car- ried away 260 feet of track and roadbed of The Alaska Railroad. At one location where the center- line of the track formerly lay about 30 feet above the lake, sounding shows that there is now 9 feet of water (fig. 19). This sliding also differed from the Lawing and Lakeview slides in that there was no evidence of waves, either on the delta or on the opposite shore. ~ DELTA MATERIALS Rocky Creek has a steep gra- dient, falling about 4,500 feet in 2 miles. The sediments exposed along the scarps are therefore coarse, poorly sorted sand and sandy gravel. Most of the appar- ent dips visible on the scarps were estimated to be less than 10°. In mid-May and early August the face of the scarp was dry, except for a few wet lenticular sandy silt beds of short lateral extent and about 6 inches thick (fig. 19). The gravel is composed primarily of angular and subangular tabular pebbles and some cobbles of slaty shale. About 150 feet upslope from the scarp of the slide that carried away the rail line, 8 feet of very sandy gravel is exposed in a shallow borrow pit. The coarse sediment seen in these out- crops is probably representative of the delta as a whole. 19.——Scarps along the front of Rocky Creek delta produced by many small slump— type slides. The gently dipping coarse sediment is clearly visible. With the exception of the sandy lens at left, most of the sediment is dry. 0’ 100’ 200’ 300’ 500' E \ \\[Preslide profile SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KENAI LAKE A23 0' 0’ 100' 100' 200' 200' 300. 300' FATHOGRAM RECORDED ON LINE 89 100’ 500’ 1000’ 1500’ 2000’ 2500’ 3000’ 3500' 4000’ 4490' PROFILE ALONG FATHOMETER LINE 89 Note: Because of the large vertical exaggeration on the fathogram, this profile, which has no vertical exag- geration; is also shown 20,—Fathogram and profile at Rocky Creek delta. Line of profiles shown on figure 4. A24 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS DESCRIPTION OF SLIDES A comparison of the areas of net erosion and net deposition (fig. 6) and a comparison of the profiles shown in figure 20 (p. A23) sug- gest that there was some lateral spreading of the slide debris on the lake floor during deposition. Just how much spreading occurred is difficult to say, for, unlike the edges of the Lawing and Lakeview slides, the edge of the depositional areas farthest from the scarp at Rocky Creek does not coincide with the limit of irregular bottom topography. However, the recon- structed preearthquake bathym— etry off the Rocky Creek delta may be in error. At any rate, be- cause the discrepancy occurs at the edge of the area where the sheet of slide debris was probably thin, it is not important in terms of the amount of material involved. Although it is difficult to locate the position of the outer margin of the depositional area, an outer limit defined either by the edge of the area of irregular bottom topography or by the difference between the preslide and postslide bathymetry suggests that the slide debris traveled a shOrter distance than that of the Lakeview slides. The contrast with the Lawing slide is even greater. At Lawing the slide debris moved down a slope of about 20°, traveled. for about 2,100 feet over the bottom, and reached a final depth of about 450 feet. At Rocky Creek the sedi- ment moved down a somewhat steeper slope (25°), reached ap- proximately the same final depth (400 ft), yet traveled only about 600 feet over the lake floor. Another important difference be- tween the slides is that at Lawng there were both backfill and far- shore waves, but at Rocky Creek there was no evidence that the sliding generated waves. A prob- able explanation is that at Lawing the slide acted as a single mass and that the volume of the ma- terial and the speed of the sliding were sufficient to deliver enough energy to the water to carry the debris a long way from the scarp and to create a wave on the far shore. At Rocky Creek, although a given amount of slide debris had about the same potential energy as at Lawing (fig. 10) and although the total amount of slide material and hence the total potential energy was considerably greater, the sliding took place as a series of small slumps. Furthermore, the slumping occurred over a longer period of time, and at any given moment the total amount of ma- terial moving downslope was not large enough to deliver sufficient energy to the water to carry the slide debris any great distance across the lake floor or to create a wave on the far shore. Sliding steepened the delta front, especially in the upper part (fig. 20); in the upper 100 feet the angle of the slope has been changed from about 22° to about 30°. Unless the slope is regraded, chances for subsequent sliding have been increased. OTHER SLIDES Several other slides occurred in the lake. One small slide removed a point of land on the edge of a delta on the south shore of the lake 2 miles east of Cooper Landing (pl. 2). The slide half submerged a boathouse and carried away a stone—filled wooden crib 60 feet long, 5 feet wide, and 8 feet high that served as a pier. Another slide probably occurred on the southeast side of Quartz Creek delta, as suggested by an area of irregular bottom topography. No scarp was formed but the outer- most 200 feet of the point was sub- merged during the earthquake. Soundings showed that the point had dropped approximately 12—14 feet. The fact that the trees in the submerged area were not tilted in- dicates that there was no apprecia- ble rotation of the delta sediments during submergence. Two small scarps were found 3 miles south- east of Quartz Creek; both were slumps in delta deposits. None of these slides caused detected waves. Small slides occurred along the front of Meadow Creek delta about 2 miles west of Lawing. Two fathograms off the delta show minor topographic irregularities having local relief of about 3 feet, but show no horizontal bedding as in immediately adjacent fatho- grams. The fact that the area of irregular topography lies in a lake- ward bow of the 500- and 520-foot contours suggests that deposition has taken place. Sliding may have caused the waves that struck the delta and the far shore. Sliding also occurred from the small delta of Porcupine Creek, west of the mouth of Snow River. A scarp was visible along the mouth of the creek, but to the north the delta edge had been sub- merged during the earthquake and no scarp could be seen. Two cot- tages on the northwest side of the delta were partially submerged. An area of irregular bottom topo- graphy lies just offshore. Wave damage occurred in the area, the wave probably having been in— duced by the sliding. Fracturing of the delta surface that accom- panied the sliding severely dam- aged an asphalt-paved parking lot belonging to the U.S. Forest Serv- 1ce. SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KEENAI LAKE A25 Seiches are periodic oscillations in lakes, bays, or rivers that are commonly produced by changes in wind stress or atmospheric pres- sure, and less commonly by earth- quakes. The oscillations take the form of standing waves. The periods of the standing waves may be the natural period of the basin, harmonics of the natural period, or, in a complicated basin, they may be controlled by some segment of the basin. The earthquake pro- duced complex seiching in Kenai Lake. In areas of abrupt shallow- ing and where the water was con- stricted, some of the waves dam- aged the shoreline. SEICHE PERIODS A record of most of the seiching was made by a continuously re- cording water—level gauge at the Chugach Electric Association power station near Porcupine Is- land (pl. 2), the record (limno- gram) of the seiching is repro- duced photographically at natural scale in figure 21 (next page) . The recording device is a 12—inch (inside diameter) vertical pipe which is located inside the power- plant at the edge of the lake. The bottom of the pipe is in the lake, the top is at an altitude of 443.5 feet, and at 426 feet there is a 11/2- inch opening into the pipe. A float rides in the water within the pipe, and the movement of a wire at- tached to the float is converted to an electric signal by a Leeds and Northrup 7948 transmitter. The signal is recorded by a General Electric-type H.L. pen recorder on paper driven at a rate of 1_ inch per hour. (Specifications cour- SEICHING tesy of R. A. Johnson, Chugach Electric Association.) The lake-level recorder stopped about 5:30 p.m., the time of the main shock when the power was cut at the plant, and did not start recording until 8:06 p.m.; thus about 21/2 hours of the initial seich- ' ing was not recorded. The limno- gram shows a long—period oscilla- tion and superimposed higher fre— quency waves having periods of only a few minutes. It is extreme- ly difficult to tell the number and frequency of the waves that oc- curred by looking at this record. However, it is possible to deter- mine the frequency and the relative amounts of energy of the waves at the water-level gauge by a power spectrum analysis of the limnogram. [The power spectrum analysis used (Blackman and Tu- key, 1959) is a type of Fourier analysis that considers the record- ed events to be aperiodic and treats data points as statistical samples based on the premise that even if an event is periodic, data points se- lected from the record do not accu- rately represent the event. The data points were taken at one-half- minute intervals for the first 8.5 hours of record using a X4 photo- graphic enlargement of the limno- gram.] The analysis was done by a computer that plotted the rela- tive amounts of energy against the frequency of all waves from 0 to 64 cycles per hour (fig. 22). [Amounts of energy are shown as the “power spectral density func- tion” defined by Blackman and Tukey (1959, p. 176) as “a value of a function (or the entire function) .whose intergral over any fre- quency interval represents the con- tribution to the variance from that frequency interval”.] Because the total amount of energy dissipated at any frequency as given by this analysis depends in part upon a comparison of the amplitude of the wave at this frequency with the amplitudes of all other waves measured by the lake—level record- er and because seiching produces standing waves with maximum amplitudes between nodes and minimum amplitudes at nodes, the amount of energy dissipated at any given frequency depends in part on the location of the lake-level re— corder. Thus, the closer the lake- level recorder is to a node, the lower the value of the power- spectral density function for a wave oscillating around that node. The power spectrum analysis showed that the complex record of seiching was produced by at least nine distinct waves having periods between 1.40 and 36.36 minutes. The periods of the standing waves formed by seiching are de- termined by the size and shape of the basin (for a thorough discus- sion of seiching, see Hutchinson, 1957, p. 299—333). The natural period of a basin can be approxi- mated by using the Merian formula (Hutchinson, 1957, p. 301) for an elongate rectangular tank having vertical sides: T = 7% where T is the period, Z is lake length, 9 is the gravitational force, and the average depth. For the total lake, 5 (402.5 ft) was determined by averaging the depths at 187 points spaced one-eighth mile REGIONAL EFFECTS . 0 THE ALASKA EARTHQUAKE, MARCH 27, 1964 A26 .wMa‘H 35M an 359.833 “€333th 23 an 60383 238% no EEMOEE‘HIAN mmnox mom odm ream 9mm 9% gm 2N QR 3N Qom mam Qmm mew 3% 9mm 0.8 @333 I came 33 98.. manor m.- o.- 3N 0.8 now 0.8 3: QB fl: 3: DD 0.: fl: o9 mg: 0.2 min m .N w . . I A .I i . h! kl I ed? J 1" i/dn I II. i 23 33 wmao: 3: 0.: mi 0.2 3H oi n: Q: m? 0.2 mm od 3 9w 3. o.“ 333 I I I I odme ‘1.le 1. L.K £. Saar? E: 1L1 ? Mg I“ ‘ 4‘ l‘ I‘ \ I r! _ . name 33 manor 3 9m 0m o.m me oé Wm Qm ma o.~ 3 o; m. o 39 r r r I I I I D P o.Nm¢ I 1.... if: 1 RE: 335% 9.32 88' J1 I . . 33 I m u _ _ _ 9:3 .1333 NI A3C] nunv SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KENAI LAKE A27 82.25 I 11.53 min 68.48 — 54.71 — m— 36.36 min 40.94 - POWER SPECTRAL DENSITY FUNCTION 4.05 min 27.17 — 9.43 min 24.0 min 13.40 5.40 min 6.06 min —0.37 0.00 8.00 16.00 .E E [x “l N 32.00 FREQUENCY, IN CYCLES PER HOUR 24.00 40.00 1.40 min 56.00 48.00 64.00 22.—~Computed power spectral density function versus the frequency of the seiching recorded on the limnogram at the hydro- electric powerplant. indicated for the larger waves apart along a line 23.375 miles long drawn halfway between the lake shores. he calculated period of the lake is 36.35 minutes; thus, the observed wave having a period of 36.36 minutes was a uninodal seiche at the natural period of the basin. The near agreement of the calculated and observed periods re- flects the fact that despite its two sharp bends, the steep-walled gen- erally flat-floored lake behaved like a rectangular tank. The relationship of the shape of the lake basin to the periods of the other waves is not as easily estab- lished. In the ideal rectangular tank, seiches can form as har- monics of the natural basin. They may oscillate around 2, 3, 4 . . . n nodes, their periods being related to that of the uninodal seiche by Tn =;1i (—9331,, where n is the number of nodes. Assuming that Kenai Lake acted as a rectangular tank, the calculated periods (in minutes) of the harmonic waves would be n2: 18.14, m = 12.11, n4= Each peak represents a wave of that frequency, and the periods of these waves (in minutes) are 9.08, n5=7.27, n6=6.05, and n7= 5.12. Although the observed pe- riods (fig. 22) are longer than these predicted values, departures in the shape of a lake basin from a rectangular tank g e n e r a l l y lengthen the periods of the har- nomic waves (Hutchinson, 1957, p. 307) ; thus, these waves may have been harmonics. However, the seiches could have formed as oscil- lations in the straight segments of the rectilinear basin or for the short-period waves, as transverse A28 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS seiches (the calculated period for a transverse seiche west of the power station is 2.25 minutes, and a wave having a 2.27-minute pe- riod was recorded on the limno- gram). It is, therefore, unjusti- fiable to assume, without further study, that the observed waves were harmonics of the whole lake basin. EYEWITNESS ACCOUNTS OF SEICHING The seiching was seen by people at Cooper Landing on the Kenai River, at the west end of the lake, at Lawing, and near Snow River. As might be expected by looking at the limnogram (fig. 21) these peo- ple saw only the short—period waves. If the amplitude of the long-period wave had been con— siderably greater than that of the shorter period waves, the long-pe- riod wave might have been rec- ognized by the observers. Perhaps it would have been possible to de- tect the long-period seiche by not— ing changes in the runup height of successive waves, but without the limnogram all waves having pe- riods of more than 4 minutes would have been undetected. Mr. I. P. Cooke, chief engineer of The Alaska Railroad, _ was alongside the Kenai River near Cooper Landing at the time of the earthquake. He said that he saw the Kenai River reverse its di- rection and flow back toward Kenai Lake. Mr. John Ingram of Cooper Landing was on the Snug Harbor ‘ Road (along the south side of the lake) near its junction with the main highway west of the mouth of Kenai River. When the earth- quake started, he checked his par— ents’ house near the lakeshore to be sure that they were all right. He went to his car to drive to his home at Cooper Landing but had trouble driving back because the slope of the moving ground kept changing and stalling the car, and he had to wait until cracks that were opening and closing were closed to permit his passing. He said that as some of the cracks closed they spouted sand into the air. When he reached the high- way junction, he looked to the right toward Kenai River and saw the highway bridge and then no- ticed a ball of asphalt (he later ex- plained he meant a ground wave) moving toward the bridge. He turned toward the bridge and saw that it had been knocked from its pilings and was lying in the water. He estimates that it took him about 4 minutes to reach this spot after the start of the shaking. While looking at the bridge, Mr. Ingram saw the lake water begin to recede from the highway embankment. He estimates that the lake was lowered about 40 feet (soundings suggest closer to 15 feet), and he could see dead logs on the lake bottom. During with- drawal, he saw Kenai River flow back into the lake as a stretch of rapids. He said that the reverse flow extended as far downstream as Cooper Landing. When he saw the depth to which the water withdrew, he became alarmed for his children’s safety because his is the first house west of the highway, beside Kenai River. He immedi- ately went home and got his chil- dren out of the house and into the car. He returned to the house to shut off his propane tank, came out, and then went back to shut off his oil tank. He met a highly agitated neighbor and told her “to get hold of herself because the shaking might not stop.” Then he drove his children away from the house westward on the high- way but was stopped by a slump that had dropped the roadbed along a 3-foot high scarp. When they stopped and got out of the car the ground was still shaking so violently that the trees were swinging sharply. Mr. Ingram estimated that the time between waves, from high water to the fol— lowing high water, was about 4 minutes. He noted that on the following day the ice in the lake at the bridge site moved back and forth as this wave continued to affect the lake water. The seiching at the west end of the lake was also seen about 25 minutes after the earthquake by Mr. J. W. Moorcroft of Moose Pass. He estimated that the time from high water to high water was 3—4 minutes, and that the total change in water level was about 20 feet. He said that the rising water “looked like a wave, but not a wall of water.” He also esti— mated that for 150 yards water ran backward from the Kenai River into the lake. Because wave damage was very slight at the western end of the lake, field evidence for the height of the waves was difficult to find. Hand leveling from the lake sur- face to the upper limit of drift- wood and silt on the south shore near the highway crossing indi- cated that some waves had runup heights of at least 7 feet. Seiching was seen at Lawing by Mrs. Hadley Roberts and Mr. Frank Gibbons. Their full de- scriptions were given in the discus- sion of the slide-generated waves at Lawing. The highest probable wave formed by seiching is de- scribed by Mrs. Roberts as having been “almost up level with the bank.” This level would indicate a height of about 10 feet. Neither observer estimated the time be- tween successive waves. Seiching at the eastern end of the lake was seen ‘by Mr. Thrall from his home just west of the Snow River delta. He did not see the lake during the earthquake but SLIDE-INDUCED WAVES, SEICHING, GROUND FRACT'URING AT KENAI LAKE A29 watched it for about an hour after the shaking stopped. Mr. Thrall said that the water went in and out like a wave and he guessed, but said he was not sure, that it was about a minute between successive highs. The water carried ice up onto the edge of Snow River delta and piled it into the river mouth. From the high-water mark indi- cated on the bank by Mr. Thrall, the wave had a height of 81/2 feet. TILTING OF THE LAKE— THE PROBABLE CAUSE OF SEICHING Seiches caused by earthquakes have been known for a long time; the 1755 Lisbon earthquake pro- duced seiches in lakes in Scotland, England, Germany, and the Low Countries (Hutchinson, 1957, p. 300). Similarly the 1964 Alaska earthquake caused seiches as far away as Michigan (Waller and others, 1965). However, as quoted by Hutchinson (1957, p. 328), G. Chrystal pointed out that gen- erally the vibrations caused by earthquakes have too short a period to cause a resonant response by a lake; even as a result of the Lisbon earthquake, Loch Lomond responded with a trinodal or quad- rinodal rather than a uninodal seiche. Hebgen Lake, Mont, ex- perienced seiching that resulted from the 1957 earthquake. The seiching was attributed to tilting of the lake basin (Myers and Hamilton, 1964). Wiegel and Camotim (1962) succeeded in pro— ducing seiches by suddenly lower- ing one end of a model of Hebgen Lake basin. Their analysis showed that the observed wave was the first harmonic of a seiche that formed in one arm of the lake. As shown below, Kenai Lake basin was so tilted that its western end was lowered about 3 feet more than its eastern end. Although the earthquake might have produced short-period seiches within the lake, the uninodal seiche may have resulted from the tilting of the lake. Kenai Lake basin was tilted. The tilt was measured to establish the deformation of the land sur- face in the area affected by the earthquake. Because Kenai Lake was the only long lake that had bench marks near its ends prior to the earthquake, it was the only lake on which tilting could be readily established. Eighteen days after the earth- quake, temporary bench marks were placed at the ends of the lake—one on bedrock on the north shore 21/2 miles east of Cooper Landing, the other on the rail of The Alaska Railroad line on the east shore one-third mile south of Rocky Creek. The bench marks were installed on a calm day while the lake was frozen; thus distor— tion of the water surface due to wind stress was probably minimal. The US. Army supplied helicop- ter transportation, and both bench marks were placed within 1% hours. Temporary bench marks were used because the ground was snow covered and the preearthquake bench marks could not be found. In a later survey, run to relate the preearthquake and postearth— quake ‘bench marks, only the hori- zontal hole drilled in a vertical rock out could be found for the preearthquake bench mark south of Rocky Creek (US. Coast and Geodetic first order bench mark Z11) ; the center of the hole was taken as the position of the bench mark. At the west end of the lake the original bench mark (US. Army Engineers third order bench mark MR4) had been replaced by the Alaska Department of High- ways; the postearthquake bench mark was related to this replace- ment. The postearthquake survey indicated that the western pre- earthquake bench mark had been lowered 3.0 feet with respect to the bench mark on the eastern end of the lake. These measurements combined with the postearthquake releveling survey by the US. Coast and Geodetic Survey along the eastern shore of the lake which shows that Lawing was lowered 0.430 feet more than Primrose (pl. 2) indicate that the true dip of the tilted surface is N. 72° W. at 1 foot in 5.4 miles. Both the amount and the direc— tion of tilt as given by this calcula- tion are somewhat questionable. The accuracy is not known for the altitude of either the original third order bench mark or its replace- ment. Furthermore, bends in the isobases of land-level changes (Plafker, 1965) suggest that tilt- ing was accompanied by local warping; thus, Kenai Lake basin may not have been tilted as a plane. The 3—foot westward tilt indicated by this survey is, how— ever, so great that probably some tilting did occur. A comparable tilt in approximately the same di- rection is suggested by the isobases drawn by Plafker (1965, p. 1677, fig. 2). According to these iso- bases the eastern end of the lake (on the basis of the US. Coast and Geodetic Survey releveling) was lowered 4.6 feet, and the western end of the lake lies in the trough of maximum land depression, which was lowered as much as 8 feet to the northeast and south- west of the lake. A westward tilt of the lake basin also is suggested by an eyewitness (Mr. Hadley, Roberts), who said that the water at Lawing on the eastern end of the lake was 2 feet lower after the earthquake. A30 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS 30’ 20’ 10’ 1636M» .r .5515 250’ 200’ 150’ Line 2 23.—Upper limit of wave damage caused by seiching along two traverses run parallel to wave travel direction on the west side of Ship Creek delta. SEICHE WAVE DAMAGE Slide—generated waves did not cause all the wave damage shown in plate 2, and as noted earlier there was additional minor wave action visible along the entire shoreline after the earthquake. Much of this minor wave action resulted from seiching. Although seiching produces standing waves, these waves become translationary where the water shallows or the lake is constricted. One constricted area is the nar- row shallow gut between Porcu- pine Island and the north shore of the lake. High water rushed through the gut from the arm to the northwest and from the arm to the east. Ice scars on spruce trees indicate that the water reached a maximum height of 20 feet on the low point on the north side of Porcupine Island. On the north shore of the lake by Porcu- pine Island, some aspens and bushes were bent by the water, and the ground was covered with a thin layer of sand and silt. On the south shore of the lake opposite Porcupine Island, where the lake is not constricted and the water is deep, Mr. Stanley Weller, an em- ployee of the Chugach Electric As- sociation, said that the water re— moved snow within 2 feet below a catwalk on the north side of the powerhouse that stands at the lake- shore. This level would indicate a maximum runup height of about 11 feet, or only about half the height. of the crest of the wave that washed through the con- stricted gut north of Porcupine Island. Waves formed by the seiching were the probable cause of two areas of wave damage on the south shore of the lake, about 2 miles southeast of Quartz Creek, where waves came ashore over an abrupt shallowing (pls. 1 and 2). All the trees at the water’s edge were knocked down by the waves though most were still attached to their roots. The largest tree was about 11/2 feet in diameter. Some turf was stripped from the bed- rock, and cobbles, pebbles, and sand were thrown up on the wave- Locations of traverses are shown on figure 17. washed area. The upper limit of wave damage at the shoreline was about 30 feet above the lake, yet despite the relatively gentle slope of the ground, the wave damage extended only about 100 feet inshore. In several other areas of wave damage where there is no evidence that the waves were generated by slides, the waves are assumed to have been caused by seiching. On the west shore of Ship Creek delta, water ran inland for at least 366 feet and had initial runup heights as much as 30 feet (fig. 23). The travel directions indicate that the wave came from the west and was refracted along the curved delta face. This powerful wave seems rather large for a seiche, but there is no evidence that it was generated by a slide. No slide scarp was visible along the shoreline, and al- though the small topographic high at the toe of the delta might sug- gest that sliding had occurred, fathograms showed undisturbed horizontal bedding beyond and to either side of the high. SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KENAI LAKE A31 SPREADING OF DELTA SEDIMENTS The long period of shaking that accompanied the earthquake mobi- lized wet unconsolidated sediments ranging in composition from silt to coarse gravel. On some deltas the sediment spread radially down- slope and toward delta margins. At Kenai Lake radial spreading is well shown on Lakeview and Rocky Creek deltas by damage in- flicted upon the railroad bridges. (Construction of these bridges is shown on fig. 24.) The bridges were placed under tension when the spreading of the sedi- ments beneath the bridges in- creased the distance between bents. Some tension was released as stringers skidded over the tops of the bents. Spreading also moved Guard timber _ Rail 4"X8” , 7 , XTie ‘ 1-. 8 stringers : i Pile cap ' Bulkhead 9"x17'l / _ planks s. H (WV . ',. 4 X12 Fillers 4"><8” ._ Sway brace Bent ,- V‘ 5 piles nifkgir'rqxx \\ "E‘uwfi’ 24,; fl . 25 (right).—Northeast corner of the railroad bridge on Rocky Creek delta. The stringers rammed the bulkhead and drove the bulkhead timbers away from the fillers during the early part of the shaking. As shaking con- tinued, the underlying sediments of the delta spread laterally and put the bridge and rail line in tension; this tension pulled the stringers part way off the pile cap and sheared the bolts in the track connectors (angle bars). the bulkheads apart; bulkheads were moved 4 and 8 inches away from the ends of the stringers at Lakeview and 31/2 and 4 inches at Rocky Creek (fig. 25). The track was also put in ten- sion, and the rails were pulled apart endwise on the bridges and adjacent embankments. Expan— sion joints were opened to their 24 (left).—Parts of a typical wooden railroad bridge. The ties lie across heavy wooden “stringers,” which in turn rest upon horizontal timbers called “pile caps” that are fastened by iron drift pins to the tops of five supporting “piles.” Each group of five piles and its pile cap is called a “bent.” A re- taining wall at each end of the bridge is called a “bulkhead” and is made by nailing horizontal 4- by 12-inch planks to vertical 4- by 8-inch wooden “fillers.” The fillers in turn are bolted to the pile cap of the end bents. “Guard timbers” (4— by 8-inch planks) are lag bolted to the ends of the ties on the top of the bridge. A32 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS 26.—Tension on Rocky Creek railroad bridge relieved by shearing the bolts in an angle bar and pulling the track apart. 27.—Tensi0n on Lakeview delta railroad bridge pulled guard timbers apart and split a tie. SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KENAI LAKE A33 maximum width, and in several places the tensional force was great enough to shear the bolts in the track connectors (angle bars) (fig. 26). The largest pull-apart was 17 inches at the south end of the Lakeview bridge, and two pull- aparts of 9 inches each were meas- ured on the Rocky Creek bridge. Tension also pulled a guard timber endwise on the Lakeview bridge and split the tie to which it was lag bolted (fig. 27). The most deeply driven piles in the end bents of the Lakeview bridge had penetration depths of 15 and 20 feet, and of the Rocky Creek bridge 19 and 27 feet. There was no indication that these piles had been tilted while being carried horizontally by the spread— ing sediment. This lack of tilting suggests that sediment as deep or deeper than the piles was involved in the lateral spreading. There is some evidence that the end of Rocky Creek bridge was given compressive "blows during the earthquake. This evidence consists of multiple dents made by a pebble as it was caught between the bridge and the bulkhead and then fell free, only to be caught again by the successive compres- sion. Four, and possibly five, dents made by the pebble suggest that there were at least this many blows. These compressive blows were so strong that the stringers rammed , the bulkhead timbers away from the fillers (fig. 25). Although pulsating compressive blows were given to the Rocky Creek bridge, the bridge was ulti- mately put in tension. Thus there is evidence for transient compres- sion on the bridge only during the initial period of the shaking. However, there is no reason to be— lieve that compressive pulses did not continue throughout the dura- tion of the shaking—while the ground was spreading laterally. Deep horizontal movement of unconsolidated sediment during the earthquake was not restricted to deltas. Evidence of such move- ment was observed at many places by the author and M. G. Bonilla during a 2-month study of earth- quake damage to the rail belt. Unconsolidated sediment sufl’ered a general loss of strength that re- duced its capacity to maintain the local, normally stable topographic relief. As a result, in areas under- lain by unconsolidated sediment, almost invariably, deep ground movement occurred toward free faces—stream valleys and breaks in slope. This movement was es— pecially well shown by the damage to railway bridges at stream val- leys. As the ground moved to- ward the streams, it carried piles driven to as much as 30 feet below the ground surface. The stream- ward movement of the ground de- creased ‘bank-to-bank widths by as much as 6 feet, and as a. result bridges were arched upward or deflected horizontally by the com- pression. Multiple pebble dents on the ends of some of the bridges show that compressive blows were also given to bridges ultimately put in compression by converging stream banks. The strength of the unconsolidated sediment of the valley bottoms was so reduced by the earthquake that the hori- zontally deflected bridges dragged deeply driven piling several feet sidewise through the sediment. GROUND FRACTURES ON DELTAS Ground fractures produced by the earthquake on Lakeview and Rocky Creek deltas are of two principal types: ( 1) arcuate frac- tures bounding rotational slump blocks and (2) interconnected fractures forming networks of parallelograms produced by lat- eral spreading of the sediment. The parallelogram—shaped net- work is the predominant fracture pattern; arcuate fractures also oc- curred on both deltas, but were larger and more extensive on Lake- view delta. Figures 28 and 29 (p. A34, A35) show these fractures as mapped on the ground and on low— level aerial photographs. Because the upper few feet of the ground was frozen at the time of the earth- quake, the fractures were clearly defined (in some places fractures cut twigs and leaves frozen to the surface) and relative movement between adjacent blocks was read- ily established. The arcuate fractures on Lake— view delta occur on the points of land north and south of the west- facing scarp. The fractures form irregular scarps bounding blocks that have been downdropped and moved lakeward. Downdropping was as much as 5 feet along some scarps, and some blocks have un- dergone slight backward rotation (figs. 30 and 31, p. A36). The lower end of Victory Creek was diverted into one of these arcuate fractures. Although movement of the slump blocks was minor, continued shaking might well have triggered additional landslides. Landslides bounded by the arcuate scarps A34 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS \ 1951 shoreline/\\\ \ Major slide scarp \ Heavy lines indicate arcuate slump \\ scarps. Bar and ball on down- \ dropped side = . _ | \ \ \ \ o - 3-inch gap, 18 inches deep, $52,? 1 inch down on north \ / 4x44] 0 ”’0 N ‘96 679 "a, 6—inch gap, 1/1 inch down APPROXIMATE SCALE 0" north 500 0 500 1000 FEET | | | | ‘ ' 28.—Slide scarps and surface fracturing on Lakeview delta. SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KENAI LAKE A35 SKA RAILROAD NCHORAGE HIGHWAY 3 inches down on west 2—foot gap, 18 inches deep, ' ' 2 feet deep, 5 inches down on west Beach 0 500 1 000 F E ET Lilli] l 29,—Preslide shoreline, slump scarps, and ground fracturing on Rocky Creek delta. A36 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS 30.—Five—foot high scarp along the hack of a rotational slump on Lakeview delta. Kenai Lake lies to the left. The slide-induced wave deposited the small pile of sediment on the dirt road displaced by the seam and debarked the trees. 31.—High scarp (4% ft) of a rotational slump on the south side of Victory Creek on Lakeview delta. SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KENAI LAKE A37 would have removed the remain- ing protruding areas from the delta margin. The location of the arcuate slump scarps corroborates the proposition that protruding areas are less stable than adjacent parts of the delta because they con- tain the greatest amount of mass bounded by the shortest potential surface of rupture. The second principal type of fracturing occurred over most. of the surface of Rocky Creek delta and in two areas on Lakeview del- ta. On the upper parts of Rocky Creek delta the fractures are 1—11/2 feet deep. The blocks appear to have been separated horizontally by downslope movement, and there is no evidence of rotation. Down- slope the fractures become longer and more interconnected. Further downslope the horizontal and ver- tical displacements between blocks increase progressively. The blocks fringing the delta’s edge (fig. 32) were downdropped as much as sev- eral feet. In plan View, most of the interconnected fractures form parallelograms having acute an- gles of about 60°. The long axis of the parallelogram lies normal to the slope and approximately parallel to the adjacent face of the delta. There are also occasional short fractures that lie parallel to the shoreline. A similar network of fractures bounding parallelogram-shaped blocks occurs on the lower slopes of Lakeview delta near Victory Creek and on the south shore (fig. 28). As at Rocky Creek delta, the long axes lie normal to the slope and, on the south shore, parallel to the delta face. The movement be- tween blocks was primarily hori- zontal, and in some places it was large. Mr. Bruce Cannon, en- gineer of structures for The Alas- ka Railroad, measured gaps as great as 4 feet near Victory Creek while the ground was frozen. 32.—Slump block having some rotation and horizontal separation at the edge of Rocky Greek delta. CAUSES OF FRACTURING The parallelogram pattern of fracturing probably involved only a shallow surface layer, for down- dropping was significant only on blocks adjacent to the faces of the deltas. Considering the deep ground movement on both of these deltas and the shallow fracturing, it seems likely that this fracturing was produced by stress developed in the somewhat more brittle sur- face layer by the spreading of the underlying sediment. Further- more, the surface fracture patterns are consistent with the stress dis- tribution that would be set. up by deep ground movement. The relationship between the probable stress field and the re- sulting fracture pattern is appar— ent on both the Rocky Creek and Lakeview deltas. Spreading of the underlying sediment should form a stress field in the overlying layer with components acting at right angles—one parallel and one normal to the delta face. Such a stress field should produce shear fractures that lie at some inter- mediate position between the axes of maximum and minimum strain. These are probably the fractures that intersect at 60° and form the parallelogram-shaped blocks. This stress distribution should also form tension fractures normal to the axis of maximum strain; indeed, the occasional fractures on Rocky Creek delta that were formed parallel to the shoreline are prob- ably tension fractures. On Rocky Creek delta the frac- ture pattern suggests that deep ground movement was generally radial (fig. 29). On Lakeview delta it appears that the principal movement of the unconsolidated sediment, and hence the direction of maximum strain in the surface material lies toward the west-fac- ing scarp that bounded the major slide (fig. 28). This fracture pat- tern also occurs on the south side of the Lakeview delta, and again it has the same orientation to the probable radial spreading that would have occurred downslope and toward the face of the delta. A third type of fracturing is found on both deltas; it occurs along the railway line, sometimes in the marginal drainage ditch and sometimes in the adjacent bank A38 THE ALASKA EARTHQUAKE, MARCH 27, 1964:, REGIONAL EFFECTS 33.—Ground fracture bounding the railroad line on Rocky Creek delta just north of the slide that carried away part of the rail line. SLIDE-INDUCED WAVES, SEICHING, GROUND FRACTURING AT KENAI LAKE A39 (fig. 33). There is usually hori- zontal separation along the frac- ture. At Rocky Creek delta, where the rail line lies near the delta face, The earthquake triggered many subaqueous slides from the deltas in Kenai Lake. The slides gen- erally removed protruding areas of the delta margins. Most of the projecting areas were composed of the youngest delta sediments, but one included older delta sediments. Projecting areas are probably the least stable parts of deltas, for they contain the largest amount of mass bounded by the shortest potential surface of rupture. Therefore, slides on projections may be pre- dictable hazards in other areas. Two kinds of destructive local waves were caused by the slides from deltas: 1. A backfill wave, formed by wa— ter that rushed toward the delta to fill the void left by the sink- ing slide mass, overtopped the slide scarp, and inundated the margin of the delta. Some backfill waves had crests as high as 30 feet above the lake level and ran inshore several hun- dred feet, uprooting and break- ing off large spruce trees. 2. A far-shore wave struck the shore opposite the slides. The wave was formed when water displaced by the slide mass, and possibly carrying some slide de- bris, flowed across the lake floor to the toe of the far shore and forced water to run up the far side of the like basin and burst above the surface of the lake. Where one of these bursts of water struck a steep bank, it ran up as much as 72 feet above lake level. These waves were also highly destructive to the forest, there was also some minor down- dropping along the fracture. The linear irregularity in the surface may have oriented the stress local- CONCLUSIONS and in places they broke ofi large trees and stripped the vegetation from the bedrock shore. Slide debris ranging in composi- tion from pebbly sand to sandy cobble gravel became entrained in water set in motion by the sliding and was carried radially down- slope away from the landslide scarp and out over the nearly hori- zontal lake floor. Fathograms show that the debris from one slide was transported 5,000 feet or more from the delta. Generally, the larger the slide mass and the deeper the water into which the slide mass moved, the greater was the momentum of the mass; thus the slide debris was carried far— ther, and the far-shore wave was stronger. Fathograms suggest that the soft lake-floor sediment in the path of the slide debris was eroded. Variations of wave-travel directions in areas struck by the far-shore waves exceed those which might haVe been produced by re- fraction, and suggest that the movement of the water-entrained slide debris was not unidirectional. Not all slides into the lake gen— erated waves. At Rocky Creek, where sliding took place as a series of small slumps along the whole margin of the delta, no evidence was found for wave action on either the delta or the opposite shore. As might be expected, the slide debris from these small slumps was not carried far from the delta. In several areas, sliding has steepened the slopes on the delta ly to produce the fracturing or may simply have made the surface layer weaker here than in adjacent areas. fronts and thus has increased the future hazard of slides. In sub- aerial slides, although the surface of rupture may be steeper than the original slope, slide debris usually comes to rest at the base of the rupture surface and thus reduces the chances for major sliding. In the subaqueous slides, however, the debris has largely been carried away from the surface of rupture; in the places where sliding has steepened the slope of the delta front, the chances for future sliding may have been increased. Waves formed by seiching in Kenai Lake during the earthquake oscillated at the calculated unino- dal natural period of the basin and at eight or more shorter periods. It was not determined if the shorter period waves were har- monics of the whole basin or if they were seiches controlled by segments of the rectilinear basin. The waves were in general only about 5—6 feet high, but in several places where the lake shallows abruptly or is constricted the waves were considerably higher. In one place on the shallow edge of a delta the waves had initial runup heights of 30 feet and ran inshore as much as 360 feet. Re- leveling of preearthquake bench marks along the lake suggests that the western end of the lake was lowered about 3 feet below the east- ern end. The tilting was prob- ably responsible for the uninodal and multinodal seiches. During the earthquake the un- consolidated sediments in the deltas spread laterally toward the margins of the deltas. Horizontal A40 THE ALASKA EARTHQUAKE, MARCH 27, 1964: displacement of piling on two rail- road bridges shows that the spreading involved material as deep or deeper than 30 feet. Mul- tiple dents made by a pebble caught between timbers at the end of one bridge are evidence for transient compression during the shaking. The spreading of the unconsolidated sediments pro- duced stress in the upper few feet of material on the delta surface. The pattern of the surface frac- tures formed in response to the spreading was controlled by the distribution of stress. The damage wrought by the earthquake at Kenai Lake was 10- cally severe, but was relatively minor compared to the potential Bennett, L. C., and Savin, S. M., 1963, The natural history of Hardan- gerfjord. Studies of the sediments of parts of the Ytre Samlafjord with the continuous seismic pro- files : Sarsia, v. 14, p. 79—94. Blackman, R. B., and Tukey, J. W., 1959, The measurement of power spectra : New York, Dover Publications, Inc., 190 p. Grantz, Arthur, Plafker, George, and Kachadoorian, Reuben, 1964, Alas- ka’s Good Friday earthquake, March 27, 1964, a preliminary geo- logic evaluation: U.S. Geol. Survey Circ. 491, 35 p. Heath, Rev. E., 1748, A full account of the late dreadfull earthquake at Port Royal in Jamaica, m A true and particular relation of the dreadful earthquake which hap- pen’d at Lima * * * 1746: London, printed for T. Osborne in Gray’s Inn, p. 327—341. Heim, Albert, 1932, Bergsturz und Men- schenleben: Vierteljahrsschrift Na- turf. Gesell. Ziirich, v. 77, no. 20, p. 1—218. , Heim, Arnold, 1924, Uber submarine denudation und chemische sedi— mente: Geol. Rundschau, v. 15, no. 1, p. 1—47. earthquake hazards of this area. Fortunately, the lake was ex- tremely low at the time of the earthquake. Had the lake level been 10 feet higher as it was 6 months later, the effects of seiches and slide-generated waves would have been greatly magnified. In- stead of coming only 360 feet in- shore at Lawing the waves would have traveled at least 600 feet in- land. Houses which this time were just as the inshore limit of the slide-generated waves would have been under 10 feet of water that was driving ‘blocks of sediment and trunks of large trees inland. Un- doubtedly such waves would have demolished the houses and would probably have taken the lives of REFERENCES CITED Hutchinson, G. E., 1957, A treatise on limnology: New York, John Wiley & Sons, v. 1, 1,015 p. Johnson, J. W., and Bermel, K. J., 1949, Impulsive waves in shallow water as generated by falling weights: Am. Geophys. Union Trans, v. 30, n0. 2, p. 223—230. Jones, F. 0., Embody, D. R., and Peter- son, W. L., 1961, Landslides along the Columbia River valley, north- eastern Washington: U.S. Geol. Survey Prof. Paper 367, 98 p. [1962] Kachadoorian, Reuben, 1965, Effects of the March 27, 1964, earthquake on Whittier, Alaska: U.S. Geol. Sur- vey Prof. Paper 542—B, p. B1—B21. Kiersch, G. A., 1964, Vaiont reservoir disaster: Civil Eng, v. 34, no. 3, p. 32—39. Kuenen, Ph. H., 1950, Turbidity cur- ents of high density : Internat. Geol. Cong, 18th, London, 1948, Rept, pt. 8, p. 44—52. _ Link, M. C., 1960, Exploring the drowned city of Port Royal: Natl. Geog. Mag., no. 117, p. 151—183. Martin, G. C., Johnson, B. L., and Grant, U. S., 1915, Geology and mineral resources of Kenai Peninsula, Alas— ka: U.S. Geol. Survey Bull. 587, 243 p. REGIONAL EFFECTS, several people. In the future de- velopment of this area, and in similar physiographic situations in other earthquake-prone areas, it would be judicious to keep in mind what happened at Kenai Lake during the 1964 earthquake. Building sites on deltas that are not carried away by landslides may be subjected to extremely de- structive waves, to severe ground cracking, or to submergence. Areas across the lake from a delta may be struck by a wave produced by a landslide from the delta. Building sites at constrictions in the lake or inshore from abruptly shallowing areas may be subjected to destructive waves caused by seiching. Meyers, W. B., and Hamilton, Warren, 1964, Deformation accompanying the Hebgen Lake earthquake of August 17, 1959, in The Hebgen Lake, Montana, earthquake of Au- gust 17, 1959: U.S. Geol. Survey Prof. Paper 435, p. 55—98. Miller, D. J ., 1960, Giant waves in Litu- ya Bay, Alaska: U.S. Geol. Survey Prof. Paper 354—0, p. 51—86. Plafker, George, 1965, Tectonic deforma- tion associated with the 1964 Alas- kan earthquake: Science, v. 148, no. 3678, p. 167 5—1687. Prins, J. E., 1957, Water waves due to a local disturbance: Coastal Eng. Conf., 6th, Proc., p. 147—162. Richter, C. F., 1958, Elementary seis- mology: San Francisco, Calif., W. H. Freeman and Co., 768 p. Taylor, D. W., 1948, Fundamentals of soil mechanics: New York, John Wiley & Sons, 700 p. Terzaghi, Karl, 1950, Mechanism of landslides, in Application of geology to engineering practice, Berkey Volume: New York, Geol. Soc. America, p. 83—123. U.S. Army Corps Engineers, Beach Ero- sion Board, 1961, Shore protection and design: U.S. Army Corps Engi- neers, Beach Erosion Board, Tech. Rept. 4, 242 p. SLIDE-INDUCED WAVES, SEICHING, GR‘OUN'D FRACTURING Varnes, D. J ., 1958, Landslide types and processes, in Eckel, E. B., ed., Land- slides and engineering practice: National Research Council High- way Research Board Spec. Rept. 29, NAS—NRO Pub. 544, p. 20—47. Waller, R. M., Thomas, H. E., and Vor- his, R. 0., 1965, Effects of the Good Friday earthquake on water sup- plies: Am. Water Works Assoc. Jour., v. 57, no. 2, p. 123—131. Wiegel, R. L., 1955, Laboratory studies of gravity waves generated by the movement of a submerged body: Am. Geophys. Union Trans, v. 36, no. 5, p. 759—774. 1964, Oceanographical engineer- ing: Englewood Cliffs, N.J., Pren- tice-Hall, Inc., 532 p. AT KENAI LAKE A41 Wiegel, R. L., and Camotim, Data, 1962, Model study of oscillations of Heb- gen Lake: Seismol. Soc. America Bu11., v. 52, no. 2, p. 273—277. Witkind, I. J ., 1964, Events of the night of August 17, 1959, The human story, in The Hebgen Lake, Mon- tana, earthquake of August 17, 1959; U.S. Geol. Survey Prof. Pa- per 435, p. 1—4. U.Si GOVERNMENT PRINTING OFFICE: 1966 O~796-520 N m 0 w l H mm m A m4, m N w 0 0 0 DEW p A I m m 3 6mm m x mm E BATHYMETRIC MAP AND PROFILES OF KENAI LAKE, ALASKA PROFESSIONAL PAPER 545—A UNITED STATES DEPARTMENT OF THE INTERIOR PLATE 2 GEOLOGICAL SURVEY EXPLANATION Area of slide deposition Outline is dashed where area is indefinite 30 1 4 and runup height of waves, in feet Magnitude of damage, shown here Direction, magnitude of damage (at base of shaft), was by George Plafker and L. R. Mayo. Damage increases from 1 through and other illustrations, 1; as follows: assigned on the basis ofa number- ing system similar to one developed irection of Brush combed in d 1. bs broken trees. Small lim and minor ice scarring on wave travel. hts only afew feet 2. Trees and limbs as much as 6 lg Runup he inches in diameter broken. Small trees uprooted. Runup reached 20feet on steep slopes 3. Trees and limbs as much as 1 in diameter broken. Exten- sive ice scarring. Boulders and foot small blocks offrozen sediment carried on shore. Runup reached a maximum of about 30feet 4. Some turf stripped from bed- rock. Large limbs and trees as 1/2 much as 2 bro ken; feet in diameter large trees uprooted. Very large blocks offrozen sedi— ment carried on shore by wave. Runup he ights of 35feet common, maximum was 72feet t of waves i 1m Inshore 1 Path of snow and rock avalanche hown me S me in shorel Change by darkl R z w (SEWARm {24"} Small slide Submerged Small slide scarp Small slide scarp Chugach Electric Association hydroelectric power station GEOLOGICAL SURVEY. WASHINGTON. D. C.A-1966—G66069 INTERIORi Base from U.S. Geological Survey Seward B—7 and Seward B—8 quadrangles Data collected by David S. McCulloch and L. R. Mayo, 1964 AND THE CHANGES AND RUNUP HEIGHTS OF WAVES, MAGNITUDE , 9 SHOWING THE DIRECTION ALASKA, 9 MAP OF KENAI LAKE IN THE SHORELINE THAT RESULTED FROM SLIDING DURING THE MARCH 27 EARTHQUAKE SCALE 1:63 360 5 MILES 5 KILOMETERS INTERVAL 100 FEET CONTOUR The Alaska Earthquake March 27; 1964 Kodiak Martm Bering Rivers Area L G E o L o G I'C AL 5 UR V E Y, I: R510 FE 88 IO NPTA L , A P ER 5 45:3 - B g1, n ,. Mm.» 3m.» . . V . . . . a A , . THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS Geomorphic Effects of the Earthquake of March 27, 1964 In the Martin-Bering Rivers Area, Alaska By SAMUEL J. T‘UTHILL and WILSON M. LAIRD GEOLOGICAL SURVEY PROFESSIONAL PAPER 543-B UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON: 1966 For sale by the Superintendent of Documents, US. Government Printing Office Washington, D.C. 20402 - Price 30 cents (paper cover) THE ALASKA EARTHQUAKE SERIES The US. Geological Survey is publishing the results of investigations of the earthquake in a series of six Professional Papers. Professional Paper 543 describes the regional effects of the earthquake. Other Professional Papers describe the effects of the earthquake on communities; the effects on hydrology; the effects on transporta- tion, communications, and utilities; and the history of the field investigations and reconstruction effort. Page Abstract ________________________________________ B1 Introduction _________________ 1 Acknowledgments ,,,,,,,,,,,, 2 Geomorphic effects of the Alaska earthquake ________________________________ Earthquake-induced ground fractures ____________________________ 3 Mudvent deposits ________________ 8 Earthquake-fountain craters ________________________________ 11 Subsidence craters 11 Mudcones ______________________________ 13 Page 1. Index map showing location of Martin-Bering Rivers area ,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,, 2. Map of earthquake-induced ground fractures, middle Martin River valley ______________ 3. E arthquake—induced g r o u n d fracture, upper Martin River valley _________________________ 4. Histograms and percent-fre- quency curves of earth- quake-induced ground frac- tures _____________________________________ 5. Map, frequency diagram, his- tograms, and percent-fre- quency curves of earth- quake-induced ground frac- tures in the middle Martin River valley ,,,,,,,,,,,,,,,,,,,,,,,,,, B2 7 Geomorphic effects—Con. 8, 9. 10. 11. 12. 13. 14. 15. 16. CONTENTS Page Avalanches ____________________________ B13 Snow avalanches __ Rock avalanches _. Subaqueous landslides ________ 19 Turbidity changes in lakes on the Martin River glacier _________________________________ 21 Filling of ice-walled sinkholes ____________________________ 21 Gravel-coated snow cones _____ 22 Lake-ice fracture __________________ 24 ILLUSTRATIONS FIGURES Page Histogram and cumulative curve of total fracture dis- placement ______________________________ B8 Earthquake-induced ground fractures, Copper River delta _____________________________________ Mudvent deposits ____________________ 9, 10 Cross section of a mudvent deposit __________________________________ 10 Earthquake-fountain crater __ 11 Subsidence crater in upper Martin River ________________________ 12 Subsidence craters in Copper River delta ___________________________ 12 Mud cones ,,,,,,,,,,,,,,,,,,,,,, 13 Camp and Sioux glacier slides 14 Map of Martin-Bering Rivers area showing generalized distribution of avalanches- 15 Geomorphic effects—Con. Pan Evidence of regional uplift ___________________________________ B26 Faulting _______________________________ 26 Effects of the earthquake upon animal populations __________________ 26 Migratory fish ..... 26 Land snails __________________ 27 Fur-bearing animals ____________ 27 Testimony of the only resident of the area ________ 27 References cited __________________________ 28 Page 17. Map showing avalanches in the Chugach Mountains area ______________________________________ B16 18. Avalanche debris and source areas of Camp slide ______________ 17 19-21. Bathymetric maps: 19. Tokun Lake ______________________ 18 20. Northeast part of Lake Charlotte ________________________ 20 21. Northern part of Kushtaka Lake ____________ 21 22. Lakes in terminus of Martin River glacier ________________________ 22 23. Gravel-covered snow cones ______ 23 24. Circular bodies of unsorted gravel with pebble rims-..» 23 25. Partially melted snow cones 1 24 26. Idealized diagram of sequence of events suggested as an explanation of the forma- tion of snow cones ________________ 25 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS GEOMORPHIC EFFECTS OF THE EARTHQUAKE OF MARCH 27 , 1964, IN THE MARTIN-BERING RIVERS AREA, ALASKA By SAMUEL J. TUTHILL1 and WILSON M. LAIRD 2 The Alaska earthquake of March 27, 1964, caused widespread geomorphic changes in the Martin-Bering Rivers area—900 square miles of uninhabited mountains, alluvial flatlands, and marshes north of the Gulf of Alaska, and east of the Copper River. This area is at lat 60°30’ N. and long 144°22’ W., 32 miles east of Cordova, and approxi- mately 130 miles east-southeast of the epicenter of the earthquake. ABSTRACT The geomorphic effects observed were: (1) earthquake-induced ground fractures, (2) mudvent deposits, (3) “earthquake-fountain” craters, (4) subsidence, (5) mudcones, (6) ava- lanches, (7) subaqueous landslides, (8) turbidity changes in ice-basined lakes on the Martin River glacier, (9) filling of ice-walled sinkholes, (10) gravel- coated snow cones, (11) lake ice frac- tures, and (12) uplift accompanied the earthquake. In addition to geomorphic effects, the earthquake afl’ected the animal populations of the area. These include migratory fish, terrestrial mollusks, fur-bearing animals, and man. The Alaska earthquake clearly de- lineated areas of alluvial fill, snow and rock avalanche corridors, and deltas of the deeper lakes as unsuitable for fu- ture construction. The Martin-Bering Rivers area lies between lat 60°15’ N. and 60°45’ N. and long 144°00’ W. and 144°45’ W. Our discus- sion is confined to the region shown on the Cordova B-l (1951), B-2 (1950), 0-1 (1959), and 0-2 (1959) quadrangles of the US. Geological Survey topo- graphic series at a scale of 1 :63,- 360. The western edge of the area is 32 miles east of the town of Cordova. These quadrangles include approximately 900 square miles of uninhabited, marshy, mountainous, and allu- 1 Assistant Professor of Geology, Muskingum College, New Concord, Ohio. 2 North Dakota State Geologist and Professor of Geology, University of North Dakota, Grand Forks, N. Dak. INTRODUCTION vial flatlands, most of Which is in the Chugach National Forest. The area is approximately 130 miles east-southeast of the epi- center of the Alaska earthquake. In the summers of 1962 and 1963, with the financial support of the National Science Founda- tion, a research team from the University of North Dakota, of which we were a part, studied the limnology, malacology, and gla- cial geology of the Martin River area. The observations and maps made during these two summers provided the preearthquake data for evaluation of the geomorphic effects after the earthquake. In 1964, fieldwork was carried on between June 2 and July 10, and was again financed by the National Science Foundation. The Bering River area has been the subject of intensive geo- logic study in past years because coal and petroleum exist in the Tertiary rocks of that region. Martin (1904, 1905a, 1906, 1907, 1908, and 1921), C. A. Fisher (1910), C. A. Fisher and Calvert (1914), W. L. Fisher (1912), Miller, Rossman, and Hickcox (1945), Miller (1951), Barnes (1951), Kachadoorian (1955), Payne (1955), and Miller, Payne, and Gryc ( 1959) have reported on the stratigraphy, economic geology, engineering geology, and structural geology of the area. Bl BZ Publications by the University of North Dakota parties include reports by Reid and Clayton (1963), Clayton (1964), Tuthill (1963), Tuthill and Laird (1964), and Reid and Callender (1965). Grantz, Plafker, and Kacha— doorian (1964) and Post (1965) have discussed earthquake ef- fects in the Martin-Bering Rivers area and Remnitz and Marshall (1966) have discussed the effects of the Alaskan earthquake on the unlithified sediments of the Cop- per River delta to the west of the area of this report. There are several physio- graphic units and two geologic provinces in the Martin-Bering Rivers area. The region south of the Martin River valley and east of the Ragged Mountains (figs. 1, 16) is underlain by Ter- tiary rocks. These Tertiary foot— hills consist of sedimentary and slightly metamorphosed rocks deformed into complex struc- tural features which have east- northeast trending axes. The northeastern one-eighth of the area has been glaciated; valley cross sections are not much dif- ferent in this area, however, from those beyond the limit of glaciation. The Ragged Mountains lie be- tween the Copper River delta and the Tertiary foothills east of Martin Lake and consist of a thrust block of Tertiary(?) ig- neous and metamorphic rocks. The Copper River delta is a lowland west of the Ragged Mountains underlain by allu- vium, and covered by intertidal marsh, alder, and muskeg growths. The Bering River valley is a low-lying alluvial plain extending from the southwest shore of Kushtaka Lake and the Bering Lake area to the sea. ALASKA EARTHQUAKE, MARCH 27, 1964 1.—-Index map showing location of the Martin-Bering Rivers area. The Martin River valley is a flat-lying valley train over which the main channels of the anasta- mosing Martin River migrate during peak ablation seasons. A terminal moraine of the Martin River glacier extends across the valley from the mountain north of Tokun Lake to the Chugach Mountains in sec. 20, T. 16 S., R. 6 E. (Cordova B-l). About 6 square miles of alluvial plain and dead-ice moraine lie between the terminal moraine and the present ice front of the glacier. This area is designated as the upper Martin River valley. That part of the valley between the terminal moraine and the point where the Martin River makes a sharp turn to the west in the area of the confluence of the To- kun Lake outlet stream and the Martin River (sec. 9, T. 17 S., R. 6 E., Cordova B-2) is desig- nated as the middle Martin River valley. The part of the valley west of this is referred to as the lower Martin River valley. The Martin River valley is presumably the trace of the Saint Elias-Chugach fault. However, the actual nature and detailed position of this fault is unknown. The Chugach Mountains are a thrust block of late Mesozoic sedimentary, igneous, and meta- morphic rocks. In the area cov- ered by this report, the maxi- mum altitude is 7,713 feet above sea level (1959). Glaciers occupy most of the valleys; all the val- leys which penetrate the sea- ward ridge of the Chugach Mountains are presently glaci- ated or have contained glaciers. These valleys are steep and U- shaped in cross section. Heavy precipitation in the area is caused by marine air cur- rents that are forced to drop their moisture as they rise over the Ragged Mountains, the Ter- tiary foothills, and the Chugach Mountains. Precipitation aver- ages about 145 inches per year (U.S. Dept. of Agriculture, 1941). Periods of clear weather of a week or more are rare during the summer months, and field- work is usually hampered by rapidly fluctuating lake and river levels. The growth of Pacific Coastal forest in the Prince William Sound area (Heusser, 1960, p. 50) is a direct response to the moist climate. Sitka spruce is found in much of the better- drained area below tree line (ap- proximately 1,000—1,200). Moun- tain muskeg covers the foothills above and beyond the glacial de- posits. Alder growths cover the moraines, the older outwash de- posits, and much of the debris- covered terminal surface of the glaciers. ACKNOWLEDGMENTS We are deeply indebted to the National Science Foundation for financial support. We wish to express our grati- tude to Mr. Theodore F. Freers of the North Dakota Geological Survey who was a member of the field party during the first two weeks of June 1964. He participated in the collection of data in the Sioux glacier area GEOMORPHIC EFFECTS, MARTIN-BERING RIVERS AREA B3 and shares responsibility with Tuthill for the map of earth- quake-induced ground fractures. Several residents of Cordova greatly facilitated our field oper- ations. Foremost among these were Mr. Karl Barth, Mr. James Osborne, and Mr. Kenneth Smith. GEOMORPHIC EFFECTS OF THE ALASKA EARTHQUAKE Twelve geomorphic effects of the Alaska earthquake were ob- served and (or) studied during the 1964 field season. These were : (1) earthquake-induced ground fractures, (2) mudvent deposits, (3) “earthquake-foun- tain” craters, (4) subsidence, (5) mudcones, (6) avalanches, (7) subaqueous landslides, (8) turbidity changes in ice-basined lakes, (9) filling of ice-walled sinkholes, (10) gravel-coated 2.-—Earthquake-induced ground fractures in the middle Martin River snow cones, (11) lake ice frac- ture, and (12) effects of regional uplift. EARTHQUAKE-INDUCED GROUND FRACTURES The most commonly observed geomorphic effect of the Alaska earthquake in the Martin-Bering Rivers area was earthquake-in- duced ground fracture (figs. 2, 3). Such features, frequently re— ported as a result of earthquakes, usually occur in unlithified sedi- ments and are oriented parallel to streambanks. Accounts of ground fractures comparable to those seen in the Martin-Bering Rivers area are given by: Davi- son (1931), Silgado (1951), Levin (1940), Small (1948), Oakeshott (1954), and Ulrich (1936). In the Martin River valley, 1,461 fractures were measured valley. Slump blocks moved toward right. Note vertical displacement, especially beside pool, left. ALASKA EARTHQUAKE, MARCH 27, 1964 B4 50:3 Hahn Efiafi “23: one 3 @2302“ 652M guacaméxanvfinumlfi £893. $8355 smzonfi M56 somaowm Pwafiwpw mfi was Eogxfim mouzuogm 9595 mgmmqsta 3:: $.63 :36 ZOC.MTS 60°15‘ "M 1 0 1 2 3 4 MILES L__.,_L,,AI¥J¥4741 16.—Generalized distribution of avalanches caused by the March 27, 1964, earth- quake in the Martin-Bering Rivers area. Avalanches are shown in black. Inset area is shown in detail on the next page. 315 ALASKA EARTHQUAKE, MARCH 27, 1964 B16 .msmshwoaonn ~33,» manna—$3339 an. @3269 mm mod: E30: 23 58:3 «93‘ €me .38de nozm 59.32 on» .30: 233502 Saunas 05 E mononufi><|§fi V \ oo,\ ”‘7“ HIN\\\W j < Waxes m.___>_ H \ écfi . \ ooox ,fifi _ mm 29 _Omaow .m 2F GEOMORPHIC EFFECTS. MARTIN-BERING RIVERS AREA 317 rested. Rock debris of the medial moraine beneath the avalanche snow could be distinguished from fresh rock-avalanche debris by its darker color. ROCK AVALANCHES The source areas of the rock avalanches could not be ex- amined because of the previously mentioned fall of rock through- out the month of June. While impressive, these later rockfalls delivered only minor quantities of material to the valley floors when compared with the massive deposits which came down dur- ing the earthquake. Observation through binoculars and an ex- amination of the avalanche de- bris indicated that the source area was intensely jointed gran- odiorite. Rock debris on the Sioux gla- cier made up approximately one- fourth of the terminal thickness of the avalanche body. The area of the slide was approximately 1.2 square miles and the esti- mated average thickness was 10 feet. Thus approximately 33 x 106 cubic yards of material (9><106 cubic yards of rock and the balance snow) was added to the surface of the glacier. The rock debris on the Sioux glacier slide consisted of an assortment of bedrock rubble ranging in grain size from fine sand to blocks having median diameters to 20 feet. The Camp slide had a stratig- raphy similar to that of the Sioux glacier slide, but vegeta- tive debris was included in the avalanche because forested areas were involved (fig. 18). The Sioux glacier and Camp slides were probably air-layer lubri- cated phenomena of the type de- scribed by Shreve (1965, p. 151). As figure 17 shows, the rock avalanches were concentrated in h- n » «1;. an...“ «it 18.—Avalanche debris of Camp slide viewed from distal margin. Arrows in- dicate source areas of rock debris. B18 ALASKA EARTHQUAKE, MARCH 27, 1964 O METERS O METERS 20 CI C METERSO D 1% VERTICAL EXAGGERATION APPROXIMATELY X 6 O METERS 2O EXPLANATION Lake bottom 1n 1963 0 1/2 1 KILOMETER I I I I I l 4I Lake bottom in 1964 BATHYMETRIC CONTOURS ARE IN METERS 19.——Map of Tokun Lake showing bathymetry and profiles of bottom configurations in 1963 and 1964. GEOMORPHIC EFFECTS, MARTIN—BERING RiVERs AREA B19 the Chugach Mountains near Sioux glacier. Unusually severe snow avalanching, but surpris- ingly little rock avalanching, oc- curred in the Tertiary foothills and the Ragged Mountains. Slopes in the Ragged Mountains are as steep as those in the Chu- gach Mountains, but valley con- figurations are different. The Chugach Mountains have been extensively glaciated by valley glaciers and all the valleys lead- ing into the Martin River valley are U-shaped. The foothills have been recently glaciated only northeast of Kushtaka, Tokun, and Charlotte Lakes (fig. 17). Slopes are much more gentle in the foothills than in either the Chugach or the Ragged Moun- tains. The Martin River valley fol- lows the presumed trace of the Saint Elias-Chugach fault. It is possible that a complexly frac- tured zone in the trace of the fault acted as a damper to the earthquake shock waves. Thus the smaller number of rock ava- lanches in the Ragged Mountains and the foothills could perhaps be explained by the absence of as severe a trigger mechanism. Local faulting in the area of Sioux glacier is also a possibility, but we observed no evidence to support this idea. DUST A secondary effect of the rock avalanches throughout the Chu- gach Mountains was the deposi- tion of a dust coat on the gla- ciers. The distribution of these dust coats as observed in the field and from the air during June and early July is indicated on figure 16. SUBAQUEOUS LANDSLIDES Lakes in the Martin-Bering Rivers area have been divided into two types on the basis of their bathymetric configuration. Lakes of the first type have slopes of about 10° or more around most of their shorelines and have average depths of 70— 120 feet. Their bottoms are re- latively flat. Tokun Lake, Lake Charlotte, and Kushtaka Lake are examples of this type. The second type of lake is extremely shallow throughout, depths rare- ly exceeding 6 feet. These lakes have been filled by outwash sedi- ments from either the Martin or the Bering River. Bering Lake and Martin Lake are examples. Little Martin Lake is interme- diate between the two types. The western half of the lake is filled by outwash sediments from the Martin River to within 6—15 feet of its surface, and the eastern half attains depths of 50 feet. Slopes in the eastern part of the lake are as high as 18°. Subaqueous landslides oc- curred in the deltaic sediments of all lakes of the first type. The most detailed study was made of Tokun Lake (fig. 19). The west- ern shore of Tokun Lake consists of outwash sediments and in 1963 had a very low gradient. The north and south shores are steep and are composed of angular boulders and cobbles. Three streams enter Tokun Lake and have formed deltas having steep distal faces. All these deltas slid —presumably triggered by the earthquake. The largest slide in Tokun Lake involved the sedi- ments deposited by Tokun Creek in the eastern quarter of the lake. The landslides generated waves which fractured lake ice and modified the shores by over- turning many flagstones along the southwestern shore. Algae which normally grow on the up- per surface of submerged rocks were found on the lower side of these rocks, and dead aquatic in- sect pupae and larvae were ob- served on their upper surface. Lake ice and water set in motion by the subaqueous landslides in the opposite (eastern) end of the lake probably overturned the flagstones. Lake Charlotte has six inlet streams. The largest of these is the melt-water distributary from the Lake Charlotte lobe of the Martin River glacier at the northeastern end of the lake; a large subaqueous slide in its delta lengthened the shoreline 80 feet and increased the area of the lake by approximately 5.1 x 104 square feet (2.1 percent of the 1962 area). In 1962 the shoreline was 29,385 feet in length. The sliding of the delta front of the outwash plain at the north- east end of the lake increased the shoreline by 165 feet. The sliding of other deltas built out into the lake reduced the shoreline by 120 feet so that the 1964 shoreline measured 29,430 feet. The bathy- metric changes which occurred in the northeast part of the lake are shown in figure 20 (next page). At the extreme southeastern end of the shore a delta of angu— lar gravel and sand slid. Mature Sitka spruce trees remained up- right and rooted in the proximal part of the subaqueous slide— the upper parts of their trunks remaining above the surface of the lake. B20 ALASKA EARTHQUAKE, MARCH 27, 1964 EXPLANATION N ‘ 5 Bathymetric contour in 1964 Bathymetric contour in 1962 \0 6 Shoreline in 1962 Shoreline in 1964 A A’ O METERS 10 —20 VERTICAL EXAGGERATION X 5 100 O 100 200 METERS | | l | 1 | l J BATHYMETRlC CONTOURS ARE IN METERS 20,—Bathymetric map and profiles showing bottom configurations in northeast part of Lake Charlotte in 1962 and 1964. GEOMORPHIC EFFECTS, MARTIN-BERING RIVERS AREA B21 A wave generated by this slide caused lake ice to strike the boles of trees and scuff and strip off the bark 3—6 feet above the scarp of the slump, and approximately 11 feet above the July 6 water level. All the other deltas in Lake Charlotte slid. The one formed by the stream which drains the lake at an altitude of 319 feet, north of Lake Charlotte, modi- fied the bathymetry of the cen- tral part of the lake. Kushtaka Lake, like Lake Charlotte is formed behind a terminal moraine‘and is fed by melt-water streams from the Kushtaka lobe of the Martin River glacier. In late July 1963 the delta formed by this drainage had slopes (based on depths 100 meters from shore) as high as 11°. The entire front of this delta slid, and slopes were changed to about 40° by this one subaqueous landslide. Figure 21 shows the bathymetric modifica- tions of the northern end of Kushtaka Lake and the front of the delta. The shallow lakes of the second type exhibited some settling of gravel bars and deltas, but showed no indication of subaque~ ous slides. The lack of steep free slopes on the distal faces of these bars and deltas no doubt explains their stability even during what must have been rather severe shaking. TURBIDITY CHANGES IN LAKES ON THE MARTIN RIVER GLACIER The present ice margin of the Martin River glacier is mantled by superglacial drift over most of its area. The distal half mile or so in sees. 22, 25—27, 35, and 36, T. 16 S., R. 6 E.; secs. 30—34, T. 16 S., R. 7 E.; and parts of Shoreline in 1963 Lake bottom in 1963 Lake bottom in 1964 B 0 /_ O METERS 20 20 4O 4O VERTlCAL EXAGGERAT‘ON X 18 O l 2 KILOMETERS L_|__I_L;l—_# EATHYMETRIC CONTOURS ARE IN METERS 21.—Bathymetric map and profiles showing bottom configurations of the northern part of Kushtaka Lake in 1963 and 1964. the Kushtaka Lake lobe are cov- ered with spruce and alder growths. The forested area is studded with ice-basin lakes that have become insulated by drift, to a greater or lesser degree, from the underlying glacier ice. The water of some of these lakes had become clear and temperate by 1962 and contained aquatic vegetation and animals. When surface ice melted from these lakes in June 1964, several were observed, from the air, to have become turbid (fig. 22, next page). The insulating drift on the east shore of Black Lake (NEIA, NEl/l, sec. 35 and NW1/4., NW1/1, sec. 36, T. 16 S., R. 7 E.) was only 6 feet thick in 1963. The exposure of the underlying dead glacier ice by slumping of the drift mantle caused turbid melt- water to enter many of the form- erly clear lakes along the glacial margin. The turbidity increase, is esti- mated at no more than 50—100 ppm (parts per million) of sus- pended material. Though this in- crease is well below the turbidity levels of ice-walled sinkhole lakes and proglacial lakes in the area, it will have an unfavorable effect upon both the flora and fauna of the lakes. If drift accumulates and retards melting of the ice, little permanent damage to plants and animals is likely, but several seasons of such turbid conditions would return the lakes to a low state of ecologic com- plexity—a state equivalent to that at Kushtaka Lake, Lake Charlotte, and the unnamed lake at an altitude of 319 feet north of Lake Charlotte (Tuthill, 1963). FILLING OF ICE-WALLED SINKHOLES In 1962 several ice-walled sinkhole lakes on the Lake Char- lotte lobe of the Martin River glacier were drained of water (Reid and Clayton, 1963). Four of these lake basins were empty in July of 1963, but in June 1964 two were partially filled with water. No ice had formed on the surface of these two lakes as it had on the adjacent sinkhole lakes which were full of water in July 1963. This lack of winter ice indicates that the filling of the lake basins had occurred in late spring because water would not have been available until ablation had begun; had block- age of the subglacial drainage occurred in the fall, the lake would have been ice covered. These lake basins probably were sealed during the earth- B22 22.—Lakes in dead ice and at the margin of active ice of the quake by readjustment of rock debris in the moulins by which the lakes were drained in 1962. The lakes drained subglacially on July 7 and caused floods on the narrow outwash plain at the head of Lake Charlotte. GRAVEL-COATED SNOW CONES Nine conical mounds of allu- vium were observed near the out- let of Martin Lake in the lower Martin River valley where there had been no mounds in July 1963 (fig. 23). The largest cone was 51/; feet high and had a basal diameter of 15 feet. In addition, three 8-inch-thick flat circular bodies of gravel having 3- to 5- inch-wide peripheral rims of well- ALASKA EARTHQUAKE, MARCH 27, 1964 sorted large pebbles (fig. 24) were observed; apparently they were residues of melted-out snow cones. All the conical mounds had circular aprons rimmed by bands of well-sorted large peb- bles. The conical mounds contained central cores of snow under a layer of unsorted sand and gravel of a type common to this part of the Martin River valley (fig. 25, p. B24). The mounds were on the north- eastern bank of the outlet of Martin Lake in NW% sec. 13, T. 17 S., R. 4 E. (Cordova B-2 quadrangle). The opposite bank is formed by talus from the Ragged Mountains. Snow of an eroded avalanche stood approxi- mately 15 feet above the south- Martin River glacier. Black Lake in lower right and the two lakes in the extreme lower left were clear water bodies in 1962 and 1963, and contained well-established aquatic flora and fauna. Icebergs in the marginal lake were slightly more numerous than in 1962 and 1963. eastern bank, opposite the snow cones. The gravel surrounding the snow cones was littered with avalanche debris such as broken tree trunks and branches, and had vague lineations normal to the drainages of the area (that is, deposited from the south- west). Earthquake—induced ground fractures and mudvent deposits crossed the apron of one cone and the alluvium between the various cones. This feature and the occurrence of an unusual number of avalanches and rock- falls in the general area indicate that the mechanism which formed these gravel-coated snow cones was the result of the Alaska earthquake. GEOMORPHIC EFFECTS, MARTIN-BERING RIVERS AREA B23 23.—Gravel—covered snow cones. Arrows indicate flood-plain debris covered by gravel only on the southwest sides. This feature supports the idea that gravel was deposited in a northeast direction. Avalanche snow in background. Geologist is standing on the northeast bank of Martin Lake outlet stream. 24.—Circular bodies of unsorted gravel rimmed with large well-sorted pebbles. Note raised “stone net.” Snow cone behind geologist has a circular apron which is also rimmed by large pebbles. Arrows indicate pebble rim. B24 ALASKA EARTHQUAKE, MARCH 27, 1964 25.—Partially melted snow cone. Arrow at A indicates place where gravel coating has been removed to show snow in- The sequence of events out- lined below seems to best explain the conditions observed in the field (fig. 26) : 1. The earthquake shock dis- lodged snow on the north- east flank of the Ragged Mountains and caused an avalanche which flowed to- ward the north down a couloir. 2. The avalanche struck the Martin Lake outlet stream, which may or may not have been frozen over. 3. Large amounts of water, sand, and gravel were splashed out of the river- bed and onto a preexisting snow drift on the northeast bank of the river. 4. The avalanche-driven water and alluvium eroded the snow drift into several dis- crete units which were cov— ered by gravel and sand. terior. B indicates fresh avalanche debris. 5. The sand and gravel retarded ablation of the snow inter- iors. 6. The largest pebbles rolled down the sides of the cones when the cones were at their maximum size and formed the rim of well- sorted large pebbles. The pebble rims were easily distinguished in the field at the end of June 1964, but they prob- ably will not be preserved. LAKE-ICE FRACTURE All the lakes in the area had fractured surface ice on June 2, 1964. Mr. Lester New of Cordova (see p. 2) stated that the ice of Bering Lake stacked up against the north shore. Geolo- gists who flew over the area shortly after the earthquake re- ported that fracturing of lake ice was general in the area. The movement of ice on Bering Lake was evidently the result of uplift of the region. A seiche type of wave must have formed which caused the piling up of ice blocks. 26.——Idealized diagram showing se- quence of events suggested as the mechanism of formation for the gravel—covered snow cones in the Mar- tin River valley (not to scale). A, Snoweovered Ragged Mountains and snow drift on northeast side of Martin Lake outlet stream. B, Earthquake triggers large avalanche which splashed water and alluvium from bed of Martin Lake outlet onto s'now drift. C, Gravel-covered remnants of the snow drift melted slowly. Because of their higher momentum, large pebbles were dislodged from slopes of cone and collected at the position of greatest basal diameter. B25 GEOMORPHIC EFFECTS, MARTIN—BERING RIVERS AREA N 7 \‘ ,‘7 A < J rxq > (“(q VK‘VVAV EXPLANATION Water and Bedrock and stream-gravel splash Alluvium Snow talus B26 EVIDENCE OF REGIONAL UPLIFT Despite search for evidence that the lakes of the area had tilted, none was found. Clearly, the entire Bering Lake area had been raised—a rim of newly ex- posed land could be seen around the entire lake and the area of the lake was greatly reduced. In 1962 and 1963 Bering Lake was intertidal, in that water-level fluctuations of approximately 18 inches occurred between tides. Large mudflats in the southern and eastern parts of the former lake basin are now above water level (fig. 16). In the Martin River valley the streams have shifted their chan- nels to a greater or lesser degree. In the middle of the upper Martin River valley the main channel shifted to a former flood channel north of the 1963 channel. In early June, flow of water was divided between the new and the old channels. As the season pro- gressed, the new channel was eroded sufficiently to command most of the flow. ALASKA EARTHQUAKE, MARCH 27, 1964 In the upper and middle Mar- tin River valley, only this one diversion could be identified be- cause the variations in river stage during a summer are ex— tremely great. Subglacial drain- age of ice-walled sinkhole lakes on the Martin River glacier and variation in amounts of water produced by melting or rain from day to day makes flash flooding a common occurrence. No data exist by which normal river levels or channels can be deter- mined, and it is our opinion that variations are so great that such data would not be useful. In June 1964 the distributary streams in the lower Martin River valley were cutting new channels in the centers of the 1963 channels. The clear-water streams from Tokun Lake and Martin Lake were not invaded by glacial melt water from the Martin River as they had been in 1963. By July 6, 1964, no melt- water had entered Little Martin Lake, where usually the water became turbid during the melting season because of the mixing of melt water from the Martin River with the clear lake water. The drainage modifications may not have been wholly due to uplift. The spring season ar- rived much later in 1964 than in the previous two years. Lake ice and snow in the forests remained fully three weeks longer in 1964 than it had in 1962 or 1963. Thus the ablation of snow and glacier ice probably did not reach its peak until after we had left the field. FAULTING No evidence was found that the Chugach-Saint Elias, Ragged Mountain, or any of the previous- ly mapped structures in the Ter- tiary foothills (Martin, 1908; Miller, 1951) had been reacti- vated by the earthquake. The concentration of rock avalanches in the area of Sioux glacier may have been associated with local surface faulting, but the bedrock source areas of these avalanches could not be examined closely. That the area was involved in a large regional uplift is apparent from changes in mean tide levels, but the inland extent of this up- lift has not been determined. EFFECTS OF THE EARTHQUAKE UPON ANIMAL POPULATIONS Several reactions to the earth- quake by the animal population were either observed or reported to us. The absolute assignment of the observed phenomena to earthquake effects is not possible in all places because the patterns of animal behavior in this area are not well known and because our observations were no more than casual. MIGRATORY FISH Salmon is an important eco- nomic resource in the Martin- Bering Rivers area. In June and July, red or sockeye salmon spawn in great numbers in Lake Charlotte, in the unnamed lake at an altitude of 319 feet north of Lake Charlotte, and in Deadwood Lake. Their seasonal occupancy of Tokun, Little Martin, Martin, Kushtaka, and Bering Lakes has long been known. In 1963 salmon in the red phase were observed entering Tokun Lake. No “bright” (silv— ery) fish were seen. In 1964 the fact that all but one of the sev- eral thousand fish observed in the lake were “bright” indicates that they had just arrived from salt water and had not taken a long time to migrate to the spawning beds. This shorter migration time gave the fish more time for the spawn and should have in- sured the optimum chance of a successful spawn. We suggest that the clearing of the water of the outlet streams farther down valley by the downcutting of the Martin River—probably a result of the GEOMORPHIC EFFECTS, MARTIN-BERING RIVERS AREA 327 uplift of the area—was respon- sible for the more rapid progress of fish migrating from the sea. Thus the effects of the Alaska earthquake may have helped rather than hurt future salmon crops in this area. LAND SNAILS Mr. Rae Baxter of the Alaska Department of Fish and Game at Cordova stated (oral commun., 1964) that large numbers of land snails were trapped in earth- quake-induced ground fractures in the area of Mirror and Martin Sloughs on the Copper River delta. The snails evidently fell in- to the fractures and could not climb out because of the granular nature of the sediments and the unstable sides of the fractures. In the vegetated delta of Lake Charlotte and Kushtaka Lake the land-snail fauna were destroyed by the subaqueous landslides. The geomorphic adjustments to the earthquake probably had a sig- nificant detrimental effect upon land-snail populations through- out the Martin-Bering Rivers area. Because the vegetation of the lake deltas in the area was largely grasses and other low plants, the greatest population density of land snails observed in 1962 and 1963 was in these very areas around the lakes— areas which were most af- fected by earthquake-triggered subaqueous landslides. FUR-BEARING ANIMALS Mr. Lester New of Cordova, who was trapping in the Bering Lake area at the time of the earthquake, states (written com- mun., Apr. 10, 1965), * * * * it [the earthquake] raised the country [the Bering Lake area] so much it pushed the beaver right up out of the water, where there was water left. Hundreds of big five- beaver houses pushed right up out of the water, their entrance ways as much as four feet above the water line. The beaver are gone now. Whether they were killed or left the area I won’t know until I go in to town May 15 [1965]. Then I will check with the Fish and Game Dept. to find out what has really happened to the beaver. I think they were killed. I have had a very good season [1965] on other furs such as mink, otter, and wolverine, they apparently were not affected by the quake * * *. While we were camped in the upper Martin River valley in early June a mature beaver wan- dered into our camp. It was in- vestigating the many distribu- tary stream channels of the up- per Martin River as if looking for a place to build a dam. When it came within a few yards of our camp we noticed its tail was cut through in three places. At the time, we assumed that it had been driven out of a lodge by its elders, and we attached no im- portance to its appearance in the valley. In light of Mr. New’s statement, the migration of this individual may have been associ- ated with the earthquake rather than with the normal pattern of behavior of a beaver-lodge popu- lation. TESTIMONY OF THE ONLY RESIDENT OF THE AREA As far as we discovered, Mr. New was the only resident of the area at the time of the earth- quake. In reply to a request for his recollections, Mr. New sent us the information given below. He states in a covering letter that these recollections were taken from the notebook which he customarily keeps while in “the bush.” The following ex- cerpt, slightly edited, is from a note written shortly after the earthquake: At 5:32 pm. the cabin and every- thing in it all but exploded. It hit very hard. The cabin raunched side- ways several feet [to the east (oral commun., July 1 1965)] the first jolt, then the second seemed to raunch it back the other way. Everything was flying that was loose or could be loosened. Utensils, gas lamps, frying pans, everything was in the air. The first thought that came to my mind was, ‘This is what I have been dread- ing, that ominous feeling of doom I had had all day long. Now just what in the world is happening?’ Glancing out the window I looked into solid ice, the lake had busted and stacked ice 8. lot higher than my cabin and it was moving in. I jumped for the door and was thrown down by the gyrations of the cabin. It was much rougher than being on a seine boat in very rough water. When I did get on my feet, I kicked the door open, figuring on the ice coming on into the cabin. Outside there wasn’t a breath of air stirring, yet all the trees were gripped in a frenzy, as though they were in a hur- ricane gale. The ice had come against the shore and now it was piling up off shore like levees or break waters. Looking across the lake over to Ham- ilton Mt. to the east everything was in movement. Hamilton Mt. to the east, Tokun Ridge to the north, and the Ragged Mt. Range to the west. Looking south at the Chilkat Mt. Range, everything seemed to be mov- ing. All the Sawtooth Range seemed to be hinged and was in one great up and down movement. There were huge rock slides and all the canyons were filled with rock and ice. Then it be- came deathly quiet, except for the booming of giant boulders moving off the high mountain ranges. For several minutes this continued, then all was quiet. B28 ALASKA EARTHQUAKE, MARCH 27, 1964 Barnes, F. F., 1951, A review of the geology and coal resources of the Bering River coal field, Alaska: U.S. Geol. Survey Circ. 146, 11 p. Bramhall, E. H., 1938, The central Alaska earthquake of July 22, 1937: Seismol. Soc. America Bull., V. 28, no. 2, p. 71—75. Broadhead, G. C., 1902, The New Ma- drid earthquake: Am. Geologist, v. 30, p. 76—87. Byerly, Perry, 1942, Seismology: New York, Prentice-Hall, 256 p. Clayton, Lee, 1964, Karst topography on stagnant glaciers: Jour. Glaci- ology, v. 5, no. 37, p. 107—112. Coombs, H. A., and Barksdale, J. D., 1942, The Olympic earthquake of November 13, 1939: Seismol. Soc. America Bull., v. 32, no. 1, p. 1—6. Davison, Charles, 1931, The Japanese earthquake of 1923: London, T. Murby and Co., 127 p. Dobrovolny, Ernest, and Lemke, R. W., 1961, Engineering geology and the Chilean earthquake of 1960: U.S. Geol. Survey Prof. Paper 424-0, p. C357—C359. Fisher, C. A., 1910, Report on the Cun- ningham coal property, Controller Bay region, Alaska: U.S. Cong. Joint Comm. to Investigate, Dept. Interior and Bur. Forestry Hear- ings, V. 2, p. 1073—1092. Fisher, C. A., and Calvert, W. R., 1914, Geology of the Bering River field and its relations to coal min- ing conditions: U.S. 63d Cong., 2d sess., House Doc. 876, p. 29—50. Fisher, W. L., 1912, Alaskan coal prob- lems: U.S. Bur. Mines Bull. 36, 32 p. Grantz, Arthur, Plafker, George, and Kachadoorian, Reuben, 1964, Alas- ka’s Good Friday earthquake, March 27, 1964, a preliminary ge- ologic evaluation: U.S. Geol. Sur- vey Circ. 491, 35 p. Heusser, C. J., 1960, Late—Pleistocene environments of north Pacific North America—an elaboration of late-glacial and postglacial cli- matic, physiographic, and biotic changes: Am. Geog. Soc. Spec. Pub. 35, 308 p. REFERENCES CITED Housner, G. W., 1958, The mechanism of sandblows: Seismol. Soc. Am. Bull., v. 48, no. 2, p. 155—161. Kachadoorian, Reuben, 1955, Engi- neering geology of the Katalla area: U.S. Geol. Survey Misc. Geol. Inv. Map I-308, scale 1:63, 360. Levin, S. B., 1940, The Salvador earth- quake of December 1936: Seismol. Soc. America Bull., v. 30, no. 4, p. 377—407. Macelwane, S. J., 1947, When the earth quakes: Milwaukee, Bruce Pub- lishing Co., 288 p. Martin, G. G., 1904, Petroleum fields of Alaska and the Bering River coal fields: U.S. Geol. Survey Bull. 225, p. 365—382. 1905a, The petroleum fields of the Pacific coast of Alaska, with an account of the Bering River coal deposits: U.S. Geol. Survey Bull. 250, 64 p. 1905b, Notes on the petroleum fields of Alaska; Bering River coal fields: U.S. Geol. Survey Bull. 259, p. 128—139. ——1906, The distribution and char- acter of the Bering River coal: U.S. Geol. Survey Bull. 284, p. 65—77. 1907, Petroleum at Controller Bay: U.S. Geol. Survey Bull. 314, p. 89—103. 1908, Geology and mineral re- sources of the Controller Bay re- gion, Alaska: U.S. Geol. Survey Bull. 335, 141 p. ——1921, Preliminary report of petroleum in Alaska: U.S. Geol. Survey Bull. 719, 83 p. Miller, D. J., 1951, Geology and oil possibilities of the Katalla district, Alaska; U.S. Geol. Survey open- file report, 66 p. Miller, D. J., Payne, T. G., and Gryc, George, 1959, Geology of possible petroleum provinces in Alaska, with an annotated bibliography, by E. H. Cobb: U.S. Geol. Survey Bull. 1094, 131 p. Miller, D. J., Rossman, D. L., and Hickcox, C. A., 1945. Geologic and topographic map and sections of the Katalla area, Alaska: U.S. Geol. Survey Mineral Inv. Prelim. Strategic Map, scale 1:31,680; (accompanies Preliminary report on petroleum possibilities in the Katalla area, Alaska, 18 p.) Morse, W. C., 1941, New Madrid earthquake craters: Seismol. Soc. America Bull., v. 31, no. 4, p. 309— 319. Oakeshott, G. B., 1954, Geologic set- ting and effects of Kern County [California] earthquakes: Seis— mol. Soc. America Bull., v. 44, no. 2B, app. A, p. 327—337. Payne, T. G., 1955, Mesozoic and Ceno- zoic tectonic elements of Alaska: U.S. Geol. Survey Misc. Geol. Inv. Map I-84, scale 1:5,000,000. Post, A. S., 1965, Alaskan glaciers— Recent observations in respect to the earthquake—advance theory: Science, v. 148, no. 3668, p. 366- 368. Reid, J. R., and Callender, Edward, 1965, Origin of debris-covered icebergs and mode of flow of ice into “Miller Lake,” Martin River Glacier, Alaska: J our. Glaciology, v. 5, p. 497-503. Reid, J. R., and Clayton, Lee, 1963, Observations of rapid waterlevel fluctuations in ice sink—hole lakes, Martin River Glacier, Alaska: Jour. Glaciology, v. 4, no. 35, p. 650—652. Reimnitz, Erk, and Marshall, N. F., 1966, Effects of the Alaska earth- quake and tsunami on Recent del- taic sediments: Jour. Geophys. Research. (In press.) Shepard, E. M., 1905, The New Madrid earthquake: Jour. Geology, v. 13, p. 45—62. Shreve, R. L., 1965, Air-layer lubrica- tion of large avalanches [abs.]: Geol. Soc. America, Ann. Mtg. Kansas City, Mo., 1965, Program, p. 151. Silgado F., Enrique, 1951, The An- cash, Peru, earthquake of Novem- ber 10, 1946: Seismol. Soc. Amer- ica Bull., v. 41, no. 2, p. 83—100. Slemmons, D. B., 1957, Geologic effects of the Dixie-Fairview Peak, Ne- GEOMORPHIC EFFECTS, MARTIN-BERING RIVERS AREA B29 vada, earthquakes of December 16, 1964: Seismol. Soc. America Bull., v. 4'7, no. 4, p. 353-377. Small, W. M., 1948, A short description of the general geology of the Do— minican Republic, with notes on the earthquake of August 4, 1946: Seismol. Soc. America Bull., v. 38, no. 1, p. 19—32. Steinbrugge, K. V., and Moran, D. F., 1956, Damage caused by the earth- quake of July 6 and August 23, 1954: Seismol. Soc. America Bull., v. 46, no. 1, p. 15—33. Tarr, R. S., and Martin, Lawrence, 1912, The earthquakes at Yakutat Bay, Alaska, in September 1899, with a preface by G. K. Gilbert: U.S. Geol. Survey Prof. Paper 69, 135 p. Tocher, Don, 1956, Movement on the Rainbow Fault [Nevada], in The Fallon—Stillwater earthquakes of July 6, 1954, and August 23, 1954: Seismol. Soc. America Bull., v. 46 no. 1, p. 10-14. Tuthill, S. J., 1963, Preliminary report on the molluscan fauna of the Martin River Glacier and associ- ated area (south-central Alaska): Veliger, v. 6, no. 2, p. 84—90. Tuthill, S. J., and Laird, W. M., 1964, Molluscan fauna of the Martin River Glacier, south-central Alaska [abs.]: Geol. Soc. Amer- ica, Cordilleran Sec., 60th Ann. Mtg, Seattle, 1964, Program, p. 64. Ulrich, F. P., 1936, Helena [Montana] earthquake: Seismol. Soc. America Bull., v. 26, no. 4, p. 323—339. US. Department of Agriculture, 1941, Climate and man: US. Dept. Agri— culture Yearbook 1941, 1248 p. 72? U. S. GOVERNMENT PRlNTING OFFICE: 1966—0 2l4-612 V The Alaska Earthquake March 27, 1964 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS Gravity Survey and Regional Geology of the Prince William Sound Epicentral Region, Alaska By J. E. CASE, D. F. BARNES, GEORGE PLAFKER, and S. L. ROBBINS GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—C UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON: 1966 For sale by the Superintendent of Documents, US. Government Printing Office Washington, D.C. 20402—Price 20 cents (paper cover) THE ALASKA EARTHQUAKE SERIES The US. Geological Survey is publishing the results of investigations of the earthquake in a series of six Professional Papers. Professional Paper 543 describes the regional effects of the earthquake. Other Profes- sional Papers describe the effects of the earthquake on communities; the effects on hydrology; the effects on transportation, communications, and utilities; and the history of the field investigations and reconstruction effort. CONTENTS Page Page Page Abstract _______________________ Cl Stratigraphy—Continued Gravity—Continued Introduction ____________________ 1 Unconsolidated deposits ______ CS Regional gravity gradient- - - _ C9 Stratigraphy ____________________ 1 Rock densities ______________ 5 Port Gravina gravity low- - - _ 9 Valdez Group _______________ 4 Structure _______________________ 5 Prince William Sound gravity Orca Group ________________ 4 Gravity ________________________ 6 high _____________________ 10 Granitic rocks ______________ 4 General features of the gravity References cited ________________ 11 map _____________________ 8 ILLUSTRATIONS FIGURES Page 1. Generalized geologic map of the Prince William Sound region ___________ CZ 2. Simple Bouguer anomaly map, Prince William Sound region ____________ 7 3. Gravity anomalies across Prince William Sound _______________________ 8 4. Interpretation of the gravity anomaly over Knight Island ______________ 10 TABLE Page 1. Densities of Prince William Sound rock units _________________________ C5 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS GRAVITY SURVEY AND REGIONAL GEOLOGY OF THE PRINCE WILLIAM SOUND EPICENTRAL REGION, ALASKA By J. E. Case, D. F. Barnes, George Plafker, and S. L. Robbins Sedimentary and volcanic rocks of Mesozoic and early Tertiary age form a roughly arcuate pattern in and around Prince William Sound, the epicentral region of the Alaska earthquake of 1964. These rocks include the Valdez Group, a predominantly slate and graywacke se- quence of Jurassic and Cretaceous age, and the Orca Group, a younger sequence of early Tertiary age. The Orca consists of a lower unit of dense—average 2.87 g per cm3 (grams per cubic centimeter)— pillow basalt and greenstone intercalated with sedimentary rocks and an upper unit of lithologically variable sandstone interbedded with siltstone or argillite. Densities of the elastic rocks in both the Valdez and Orca Groups average about 2.69 g per cm3. Granitic rocks of rela- tively low density (2.62 g per cm3) cut the Valdez and Orca Groups at several localities. Both the Valdez and the Orca Groups ABSTRACT were complexly folded and extensively faulted during at least three major episodes of deformation: an early period of Cretaceous or early Tertiary orogeny, a second orogeny that probably culmi- nated in late Eocene or early Oligocene time and was accompanied or closely followed by emplacement of granitic batholiths, and a third episode of deforma- tion that began in late Cenozoic time and continued intermittently to the present. About 500 gravity stations were estab- lished in the Prince William Sound region in conjunction with postearthquake geo- logic investigations. Simple Bouguer anomaly contours trend approximately parallel to the arcuate geologic structure around the sound. Bouguer anomalies decrease northward from +40 mgal (milligals) at the southwestern end of Montague Island to —70 mgal at College and Harriman Fiords. Most of this change may be interpreted as a regional Following the Alaska earth- quake of March 27, 1964, the US. Geological Survey made an inten- sive investigation of the effects of the earthquake in the Prince William Sound region where the epicenter was located. Concurrent INTRODUCTION investigations included studies of the regional geology, changes of land level, faulting, and effects of waves on shorelines which were summarized by Plafker and Mayo (1965). At the same time, this gravity survey, mentioned briefly The geologic map of Prince William Sound (fig. 1) has been generalized and modified from pub- lished small-scale geologic maps STRATIGRAPHY compiled by Grant and Higgins (1910) and Moffit (1954). Struc- tural data were gained from recon- naissance studies by George Plafker, gradient caused by thickening of the continental crust. Superimposed on the gradient is a prominent gravity high of as much as 65 mgal that extends from Elrington Island on the southwest, across ISnight and Glacier Islands to the Ellamar Peninsula and Valdez on the northeast. This high coincides with the wide belt of greenstone and pillow basalt of the Orca Group and largely reflects the high density of these volcanic rocks. A large low in the east-central part of the sound is inferred to have a composite origin, and results from the combined effects of low- density sedimentary and granitic rocks. The Prince William Sound gravity high extends southwest-northeast without major horizontal offset for more than 100 miles. Thus the belt of volcanic rocks causing the high constitutes a major virtually continuous, geologic element 0 south-central Alaska. by Barnes and Allen (1965), was conducted in order to establish the regional geologic and tectonic set- ting of the earthquake. Geologic sections of this report were prepared by Plafker and Case, the gravity sec- tions by Case, Barnes, and Robbins. L. R. Mayo, and J. E. Case in 1964. Some geologic details on Knight Island are based on recent mapping by Richter (1965). The ages of the Cl C2 ALASKA EARTHQUAKE, MARCH 27, 1964 50 . Bligh I T05 30 MILES I 1 Geology compiled from Moffit (1954) and Grant and Higgins (1910) supplemented with reconnaissance mapping by George Plafker, L. R. Mayo, and J. E. Case during 1964 1.—Generalized geologic map of the unAVL'l‘Y SURVEY, REGIONAL GEOLOGY, PRINCE WILLIAM SOUND EPICENTRAL REGION EXPLANATION sl ou Eli x ; Unconsolidated deposits QUATER- NARY UNCONFORMITY ” F Sedimentary rocks of Orca Group / Slightly metamorphosed and complexly ; # folded marine sandstone and dark- fl gray to black or reddish-brown hard Grani tic rock siltstone and argillite., Includes minor . . .s _ amounts of conglomerate, limestone, Granite and quartz diorite in stocks, and basaltic lava flows. Contains dikes, and S’Llls that intrude Valdez plant material of Tertiary age and J and Orca Groups nondiagnostic marine megafossils Eocene \l fl TERTIARY Volcanic rocks of Orca Group Gently to moderately folded, greenish- black slightly altered basaltic flows (greenstone) that commonly exhibit pillow structures. Intertongues with massive conglomerate and breccia and thin-bedded argillite and sandstone. Contains marine megafossils of prob- e able middle to late Eocene age J UNCONFORM/TY Valdez Group Tightly folded and metamorphosed marine graywacke and dark-gray to black slate containing minor amounts g of argillite and conglomerate 2 V JURASSICU) AND CRETACEOUS q (\T T74 VU”“¢\2 Glacier Contact Dashed where approximately located; dotted where inferred U Fault Dashed where approximately located; dotted where inferred. U, upthrown side 90 saw 50 Anticline Overturned anticline Minor folds ‘ Showing trace of axial plane, plunge of axis, and dip of limbs 50 75 90 _.J_ ‘0‘” —l— Upright Overturned Vertical goon upsfik Generalized strike and dip of beds _.15_ Strike and dip of foliation Prince William Sound region, Alaska. C4 Valdez and Orca Groups have been revised by reinterpretation of the collections from the Valdez Group and by studies of a new fossil collection from the Orea Group (Plafker and MaeNeil, 1966). The geologic units in Prince William Sound shown on figure 1 include: (1) the Valdez Group, a sequence of eugeosynclinal elastic rocks of Jurassic(?) and Cretaceous age; (2) the Orca Group, a sequence of early Tertiary age which can be subdivided, in a general way, into a lower predominantly volcanic unit and an upper predominantly sedi- mentary unit; (3) bodies of granitic rocks which intrude both the Valdez and Orca Groups; and (4) uncon- solidated continental and marine deposits of Quaternary age. VALDEZ GROUP The Valdez Group includes the oldest bedded rocks exposed in the Prince William Sound region. As defined by Schrader (1900, p. 408), the Valdez is a lithologically monot- onous sequence of slate, argillite, and metamorphosed graywacke that is exposed along the northern and western shores of the sound and also underlies much of the adjacent Chugach and Kenai Mountains. The sequence is characterized by thin to thick beds of light—gray or tan-weathering poorly sorted sand- stone of graywacke type which is rhythmically interbedded with thin- bedded, finely laminated dark—gray to black argillite and slate. In the vicinity of granitic intrusive rocks, the slate and argillite are locally altered to phyllite and the gray- wacke to semischist. Thin units of stretched pebble or cobble conglom- erate and lenses of altered basaltic flows or intrusives occur locally within the sequence. The sedimentary sequence as— signed to the Valdez Group is probably tens of thousands of feet thick. The thickness cannot be accurately measured because the ALASKA EARTHQUAKE, MARCH 27, 1964 sequence lacks key beds and the units are duplicated and interrupted by complex folding and faulting. Determinative fossils from the Valdez Group, and probably equiv- alent sequences elsewhere along the margin of the Gulf of Alaska, in- clude Inoceramus and Aucella that indicate a Jurassic(?) and Cretace- ous age for the Valdez (D. L. Jones, oral commun., Nov. 24, 1965). ORCA GROUP The predominantly volcanic unit that forms the lower part of the Orca Group crops out in a discon- tinuous belt 52 miles long and up to 12 miles wide that trends from the mainland east of Valdez Arm south- westward through Glacier, Lone, Knight, Bainbridge, Evans, and Elrington Islands (fig. 1). The unit consists principally of slightly altered greenish-black dense basaltic lava flows, pillow lavas, flow brec- cias, and coarser textured diabase intrusives which have collectively been termed “greenstone.” Most of the flows are characterized by strikingly well formed pillow struc— tures. The volcanic rocks inter— tongue With thick beds of conglom- eratic argillite, conglomerate, dark- gray siltstone or argillite, and graywacke sandstone. The thick— ness of the unit differs markedly along the length of the outcrop belt. Along the southeast shore of Valdez Arm it is several thousand feet thick (Capps and Johnson, 1915, p. 45). The greatest thickness of the unit probably occurs on and near Knight Island. The volcanic rocks dip 40°—80° W. across most of an outcrop belt as much as 11 miles wide. The breadth and dips in- dicate an apparent thickness of about 40,000 feet, but it is entirely possible that there has been signifi- cant repetition by folding and faulting. Marine megafossils col- lected during the 1964 reconnais- sance survey indicate that the volcanic unit is of early Tertiary (probably middle to late Eocene) age (Plafker and MacNeil, 1966). The predominantly sedimentary unit that comprises the upper part of the Orca Group includes the bedded rocks that occur especially on the eastern and southern shores of the sound and on the islands of the central part of the sound (fig. 1). In its gross aspects the unit is distinguished from the Valdez Group by a somewhat more variable lithology and a generally lower de- gree of metamorphism. The se- quence consists mainly of thin to thick beds of graywacke sandstone, but also includes light-colored ar— kosic, carbonaceous, tuffaceous, cal— careous and conglomeratic sand- stones. The sandstone commonly is interbedded with dark-gray to black or reddish-brown dense hard siltstone or argillite which locally contains abundant carbonized plant remains. Minor amounts of light— gray—weathering limestone occur as thin beds and concretions within the siltstone or argillite. Tabular and lenticular masses of basaltic lavas are locally interbedded with the clastic rocks, most notably on Hinchinbrook Island and the ad- jacent mainland to the east. The unit is at least several thousand feet thick, but complex folding and faulting preclude a reliable measure- ment. The only fossils diagnostic of age that have been obtained from the sedimentary unit are Alnus pollen that indicates a post-Mesozoic age (W. R. Evitt, written commun., Nov. 10, 1964). The unit overlies, and probably is in part interbedded with, the volcanic sequence of the Orea Group from which lower Tertiary f01~sils have been collected; GRANITIC ROCKS Granitic rocks intrude the Valdez Group at numerous localities in the northwestern part of Prince William Sound, and a single large intrusive mass cuts the Orca Group GRAVITY SURVEY, REGIONAL GEOLOGY, PRINCE WILLIAM SOUND EPICENTRAL REGION C5 on the mainland in the eastern part of the sound. The intrusive rocks consist mainly of pale-pink to pinkish—gray porphyritic biotite granite and light-gray hornblende- biotite quartz diorite. The granite body that cuts the Orca Group on the mainland in the eastern part of the sound can be no older than the rocks it intrudes (early Tertiary) and may be slightly younger. The remaining granitic bodies that in- trude the Valdez Group are litho- logically similar and probably be- long to the same epoch of intrusion, although the available data do not preclude a pre-Orca age. UN CONSO LIDATED DEPOSITS The Mesozoic and Tertiary rocks of the Prince William Sound region are overlain with marked angular unconformity by nearly flat-lying fluvial, glacial, and marine deposits of gravel, sand, mud, and till. The various types of these uncfinsofiE—fi dated deposits are not differentiated on figure 1. ROCK DENSITIES The densities of approximately 50 hand specimens representing the major lithologic units (except the unconsolidated sediments) were measured by comparing their weight in air and water (table 1). No effort was made to determine the effect of porosity on the measured densities, but tests of the influence of porosity on densities of similar rocks from other parts of Alaska TABLE 1.——Densities of Prince William Sound rock units Rock unit Specimens Densities (g per cm3) Minimum Maximum Average have shown that the effect is negligible. The heaviest rocks in the assem- blage are volcanic rocks of the Orca Group and the lightest are the granitic intrusive rocks. The sam- pling is not sufficiently representa— tive to establish definitely the lack of density contrast between the Orca sedimentary rocks and the older, more metamorphosed Valdez sedimentary rocks. However, seven slate samples from the combined groups had an average density of 2.74 g per cm3 (grams per cubic centimeter), whereas 17 argillite and graywacke samples averaged 2.67 g per cm3. Furthermore, four metagraywacke samples from the Valdez Group averaged 2.69 against 2.66 for the four less-metamorphosed graywacke samples from the Orca Group. Thus, overall densities of the two formations would depend on the relative abundance of these rock types. N0 samples were obtained from the Quaternary de- posits which probably have an average density of about 2 g per cm3. Granite _______________________________ 8 2.58 2 . 69 2.62 Sedimentary rocks of Orca Group ........ 16 2.63 2.75 2.69 Volcanic rocks of Orca Group ___________ 8 2.78 2.96 2.87 Sedimentary rocks of Valdez Group ______ 10 2.64 2.74 2.69 STRUCTURE The Valdez Group was intensely folded and faulted during a major period of orogeny which is tenta- tively placed in the time interval from Early Cretaceous to early Tertiary. The strike of bedding planes and the trend of fold axes are almost east-west in the region east of Valdez Arm, and the trend swings abruptly southwestward in the region west of the arm. Folds are characteristically tightly ap- pressed and overturned toward the south; drag folds and minor thrust faults are common. A second major period of orogeny that probably culminated in late Eocene or early Oligocene time (Plafker and MacNeil, 1966) com— plexly folded and faulted the Orca Group and intensified the deforma— tion of older rocks. It was accom— panied by, or closely followed by, emplacement of granitic batholiths. The strike of bedding planes and fold axes approximately parallels the structural trends of the older rocks, but they are notably diver— gent at many localities. The folds range from open to tightly ap- pressed and locally are overturned both towards the north and south. They are of small amplitude and lateral extent and are complicated by intricate drag folding and minor thrust faults. The most recent period of de- formation, which has continued intermittently since late Cenozoic time, is attested by the uplift of the adjacent Chugach and Kenai Moun- tains, by the frequent earthquakes in the region, and by recent surface faulting. All the rocks of the Prince William Sound region are intensely fractured and complexly faulted. The major lineaments show up on aerial photographs as conspicuous linear depressions and have been delineated in photogeological studit s CS by Condon and Cass (1958) and Condon (1965). Two preexisting faults that displace sedimentary rocks of the Orca Group on Mon- tague Island (fig. 1) were reacti- vated during the March 27 earth— quake (Plafker, 1965). The only faults thus far recognized in the region that juxtapose lithologically distinctive sequences are in the Valdez Arm area (fig. 1) where rocks of the Valdez Group have been thrust from the northeast over the younger rocks of the Orca Group (Capps and Johnson, 1915, p. 62— 63). Topographic and geologic evidence suggests the presence of major faults at other places where ALASKA EARTHQUAKE, MARCH 27, 1964 contact relationships are obscured by the lack of key beds, the presence of brecciation and shearing associ- ated with minor folding and fault- ing, and the extensive areas covered by water, unconsolidated deposits, or vegetation. According to Moffit (1954, p. 275), there are no known depositional contacts between the Valdez and Orca Groups anywhere in Prince William Sound. At least seven major shear zones as much as 1,000 feet wide transect the area mapped by Richter (1965, p. 14) on Knight Island. Richter reports that there is no good evi— dence to indicate the sense or motion Gravity was measured at ap— proximately 500 stations within the sound (fig. 2). Most of the meas- urements were made with LaCoste and Romberg portable geodetic meter G—17, but approximately 75 measurements were made with a World-Wide meter. The meas- urements are based on absolute gravity values of 981,954.90 mgal beneath bench mark “A71” at Cordova Post Office, 982,008.00 mgal at bench mark “H11” in Valdez, and 981,921.30 mgal at bench mark “L31” in Anchorage that were established by at least three postquake ties to the North American Standardization Pendu- lum station (Woollard and Rose, 1963) at Fairbanks. These values differ by as much as 0.5 mgal from previous values measured near the stations established by Thiel and others (1958) prior to the earth- quake. These deviations reflect small differences in location, adjust- ments in the North American gravity network, and changes caused by the Alaska earthquake (Barnes, 1966). During the survey, GRAVITY float-plane flights were used to establish 10 subsidiary base stations within the sound. Most of the remaining measurements were made at shoreline points reached by small-boat traverses. Base stat-ions were occupied at intervals of 1—4 days. A few inland stations were reached by automobile or aircraft trips. The drift and calibration characteristics of meter G—17 are such that the accuracy of 95 percent of observed gravity values should be better than 0.2 mgal. Contouring in the water-covered parts of Prince William Sound was helped by surface-ship gravity measure— ments made by the US. Coast and Geodetic Survey ship Surveyor. W. H. Bastian and L. R. Mayo assisted in the collection of the gravity data, and L. J. LaPointe made some of the computations. Most of the gravity stations were located on beaches or rocks along the shoreline. Elevation control was obtained by measuring the height of the station above water level and making a tide correction to obtain an elevation relative to along the shear zones, but left- lateral strike-slip movement of as much as half a mile is weakly suggested by the supposed disloca— tion of a coarse—grained gabbro body along one of the zones. On Knight Island the thick sequence of volcanic rocks dips westward beneath probably older rocks of the Valdez Group which crop out west of Knight Island Passage. Here, the structural re- lationships suggest the possibility that the two units may be separated by a westward-dipping overthrust concealed beneath Knight Island Passage. mean sea level. A few stations were occupied at tidal bench marks and a few at geodetic bench marks established prior to the earthquake. Elevations at bench marks with respect to mean sea level had changed, owing to the tectonic deformation that accompanied the earthquake, so corrections to post— earthquake mean sea level were required for these stations. Eleva— tions of all shoreline stations are probably correct to within 3 feet, and the elevations of the few sta— tions not on the seacoast should be correct to within 20 feet. Location control was provided by topo— graphic maps of the Geological Survey, scale 1:63,360. Simple Bouguer anomalies were computed with a reduction factor correspond- ing to a density of 2.67 g per cm“. Simple Bouguer anomaly values are probably accurate to within about 1 mgal. Terrain corrections were made for a few stations close to the line A—A’ (fig. 3) and ranged from 0.2 to 13.5 mgal for terrain through zone 0 of Hayford—Bowie (Swick, 1942). GRAVITY SURVEY, REGIONAL GEOLOGY, PRINCE WILLIAM SOUND EPICENTRAL REGION C7 147°- 146° Lake a , AGH’fJVJ «3“ a :W ; \k a; George E} - \ Elwik r, y‘MTVQg‘fl‘él/YNE‘Et‘X’Q 27,,Iww%\ ~ 5315) a 1 4:”- ‘w‘x; '6‘” FM" “Tm? he‘s; ’J‘cha * ‘ x f'"“‘ l’ '6 gay!“ {Rf ‘\ ‘.r(“.}’°"é<‘%vé‘3i F7" 67’?" «4° 5 ’ 5‘; ‘2‘?ka Xi‘a’fi‘ a “V if ( ” ' M ' / 157“” \\~r'““ Eff” H > ' 1514/ ‘\ ‘3 ”’ "32%” A9.3”25— X’Q /ELLAMAR/ /PENINSULA / E MA71i1¥ 2‘ j 1% EXPLANATION 0___.__ ml. 5. 60 ° -E:_~ Zero isobase +L/ . 0V / + Dashed where approximately located (after — /’ Plaflcer and Mayo, 1965, figure 8) Bainbridgel +25 Elrington I / / Simple Bouguer anomaly contours Dashed where approximately locatedl Contour interval 5 milligals. Anomaly contours were computed with a reduction/actor corresponding to a density of 2.67 g per cnry . Gravity station 0 10 20 30 MILES I x x . . I l l 2.——Simple Bouguer anomaly map, Prince William Sound region, Alaska. C8 BOUGUER ANOMALY, IN MILLIGALS ALASKA EARTHQUAKE, MARCH 27, 1964 \\Complete Bouguer anomaly ' \ / Assumed regional Bouguer anomaly I // _ 2 I —50—_><_./ :- —7o — — KENAI PENINSULA PORT KNIGHT ' ,A BLACKSTONE COCHRANE NELLIE ISLAND nght Green Montague A, SEA LEf/gcflo BAY BAY JUAN PASSAGE Island Island Island 8000 "9 Greenstone ~lv .— 16,000 1 2:888 o 10 20M|LES .— 20’000 Location of profile shown in figure 2 —— 3.——Gravity anomalies and profile across Prince William Sound. GENERAL FEATURES OF THE GRAVITY MAP The simple Bouguer anomaly contours indicate a general trend from positive values over the Alaskan Continental Shelf to nega- tive values on the mainland. Values are as high as +40 mgal at the southwestern part of Montague Island and decrease to —70 mgal at College and Harriman Fiords (fig. 2). Contours are crudely arcuate around Prince Willian Sound, roughly paralleling the dominant trend of the geologic structure. Superimposed on this regional pat- tern are the arcuate Prince William Sound gravity high, which is largest over Knight Island, and the broad gravity low covering Port Gravina and the southeastern part of the sound. The high coincides with a belt of greenstone and pillow basalt and is inferred to be caused by these rocks. The low is probably caused by a thick accumulation of recent sedimentary deposits, a gra— nitic intrusive rock, and, in some areas, terrain effects. In part the low is an “apparent” low due to the combined effect of the high to the northwest and the regional gradient. No other major correlations between the pattern of gravity anomalies and the mapped geologic features are apparent at this stage of inter- pretation. The new data provide a more complete gravity map of the Prince William Sound area than was available from the limited number of measurements made by Thiel and others (1959) which served as the basis for earlier Alaskan gravity maps (Woollard and others, 1960, fig. 2; Ivanhoe, 1961). These maps showed the high gravity values in the southern part of the sound and the low values in the Chugach Mountains. However, neither the magnitude nor the estent of the Prince William Sound high to the north and northeast was recog— nized, although Woollard and others correctly interpreted the anomaly as being caused by mafic rock (1960, p. 1029). Furthermore, we have measured lower gravity values in the Chugach Mountains than did Thiel’s group, so the gradient on the northwest side of Prince William Sound high is steeper, and causes a much larger gravity difference than had been previously indicated by Woollard and others (1960). Be- tween Knight Island and Harriman Fiord there is a total change of 135 mgal in 40 miles, in contrast to the change of about 60 mgal shown by Woollard and others. Between College Fiord and Kiniklik (near a low part of the Prince William Sound positive anomaly) there is a change of 70 mgal in 20 miles, in contrast to the change of about 50 mgal shown by Woollard and others. The zero isobase separating the areas of subsidence and uplift during the Alaska earthquake of March 27, 1964 (Plafker, 1965, fig. 2; Plafker and Mayo, 1965, fig. 8) approximately coincides with the steepest part of the gradient along the northwest side of the sound and locally parallels the gravity con- tours. This relationship could be purely coincidental, or it could be that the gradient represents, at least in part, a fault or hinge line at depth along which movement oc- curred during the earthquake. Press and Jackson (1965, p. 867) have postulated a near—vertical GRAVITY SURVEY, REGIONAL GEOLOGY, PRINCE WILLIAM SOUND EPICENTRAL REGION 09 fault which lies within 15 kilometers of the surface beneath the zero isobase and extends to a depth of 100—200 km. However, such a fault could explain only a part of the gradient on the north side of Knight Island, and its postulated configuration appears to be contrary to most of the available geologic and seismologic data (Plafker, 1965, p. 1686). The large steep gradient on the south side of the island suggests that the relatively dense greenstones are the real cause of the Knight Island anomaly. Such interpreta- tion, based on observed geology and measured density variations, seems preferable in View of the data presently available, and it forms the basis of the remaining discus- sion. However, future data may show that some of the Prince William Sound anomalies are caused by tectonic features that have not yet been recognized. REGIONAL GRAVITY GRADIENT In order to delineate the regional gravity field, which is presumed to reflect variations in crustal thick- ness and density, the local anomalies due to the greenstone belt must be removed by analytical or graphical methods. Conversely, in order to isolate the anomaly caused by the greenstone belt, a regional gravity gradient must be removed. Obvi— ously, a unique solution in removal of either the local anomaly or the regional is impossible, but reason— able approximations can be made for both the regional and local anomalies. Relationships between local ge- ology, simple Bouguer anomaly, and complete Bouguer anomaly along a profile extending from southeast of Montague Island to the head of Passage Canal are shown in figure 3. The complete Bouguer anomaly profile is approximate: values at stations near the line of prt file for which terrain corrections were made were projected to the line of profile. The complete Bouguer anomaly profile shows a regional gravity gradient that is nearly linear or slightly convex upward. For the following analy- sis, it is assumed that the regional gravity gradient is determined by the complete Bouguer gravity values at Montague Island and at the Kenai Peninsula. The assumed regional gradient between these points is shown by the long-dashed line (fig. 3). Data obtained from the US. Coast and Geodetic Survey ship Surveyor on the Continental Shelf and by Thiel and others (1959) at Middleton Island, about 50 miles southeast of Montague Island, suggest that approximately the same regional gradient extends to the continental slope. The gradient may be assumed to represent a gradual increase in crustal thick— ness from the continental margin to the northeast side of the sound. Woollard and others (1960) showed a similar gravity gradient and inferred increase in crustal thickness from about 24 km at Middleton Island to 33 km at Cordova and 37 km under the Chugach Mountains on their north— south profile through Middleton Island, Cordova, and Valdez. How- ever, from reinterpretation of seis— mic data obtained by Tatel and Tuve (1956) from shots fired in College Fiord, Woollard and others believe that the crust is thicker beneath Prince William Sound than the gravity data suggested. They derived a crustal thickness in Prince William Sound of about 50 km. The traveltime curves show high velocities east of Valdez, suggestive of mafic rocks where our new data show a possible extension of the Prince William Sound gravity high. Analysis of the new gravity data indicates that the regional gravity gradient is reasonably consistent with a change in crustal thickness from about 50 km near College Fiord (Woollard and others, 1960) to about 20 km near the continental margin, as measured by Shor (1962) southeast of Kodiak. PORT GRAVINA GRAVITY LOW The broad gravity low in the southeastern part of the Prince William Sound is best developed at the eastern end of Port Gravina. There it is in part related to the granitic pluton that extends from Port Gravina eastward into the Chugach Mountains (fig. 1), al- though the outcrop area is small in comparison to the area of the gravity anomaly. However, the anomaly probably also reflects a thick accumulation of poorly con- solidated sediments in the central and eastern parts of the sound. Its apparent extension into Montague Strait is influenced by terrain effects. Sample terrain corrections for two stations along the steep shoreline of the strait showed that the complete Bouguer anomalies at these stations are as much as 10 mgal higher than the simple Bouguer values contoured on the map. Corrections of this magnitude would tend to straighten the +15 and +20—mga1 contours in Monta— gue Strait and to eliminate much but not all of the tonguelike low shown by the contours on figure 2. The gravity low is also an apparent low imposed by the Prince William Sound high on the regional gravity gradient. The low which would remain in Montague Strait after terrain cor- rections and much of the low which occupies the central and south- eastern part of the sound in part represent thick, poorly consolidated glacial and marine sedimentary deposits. The average water depths in the central and eastern parts of the sound are shallow, about 300—— 800 feet, whereas water depths of 1,000—2,000 feet are common in the C10 northwestern part of the sound and in some of the narrower passages and fiords. If the sound were once glaciated to a depth equivalent to the present water depths in the flords, poorly consolidated material as much as LOGO—1,500 feet thick may be present at the site of the gravity 10w. However, sparker data obtained from the US. Coast and Geodetic Survey ship Surveyor (R. C. Malloy, oral commun, 1964, 1965) suggest that 500 feet of sedi— ment may be the maximum ac- cumulation in most of the south- eastern part of the sound. This thickness would account for not more than 10 mgal of the gravity low. The main cause of the gravity low in the southeastern part of the sound is thus presumed to be the granitic rocks which crop out near Port Gravina and which probably extend westward beneath the cover of sedimentary rocks. Although the one sample obtained from the Port Gravina pluton had a density as high as the average of the Valdez and Orca sedimentary rocks, sam- ples from plutons on the opposite side of the sound indicate that the average density of the Prince William Sound granite is about ALASKA EARTHQUAKE, MARCH 27, 1964 0.07 g per cm3 less than the den- sities of the Valdez and Orca sedi- mentary rocks. The absence of gravity lows over the intrusive bodies on the northwestern side of the sound suggests that they are smaller and that the Port Gravina intrusive is larger and possibly less dense. However, terrain corrections and precise contouring might also show small negative anomalies as- sociated with the intrusive rocks on the northwest. PRINCE WILLIAM SOUND GRAVITY HIGH When the assumed regional Bouguer anomaly values are sub— tracted from the observed Bouguer anomaly profile, a residual positive anomaly of as much as 67 mgal can be isolated over the Knight Island greenstone (fig. 4). The residual anomaly is asymmetrical: the gra- dient on the southeast flank is much steeper than on the northwest flank of the anomaly. The steepness of the gradient on the southeast flank indicates that most of the relatively dense rock mass causing the anom- aly probably lies at shallow depth— within the upper 10 km of the crust. The greenstones and pillow ba— salts in general dip steeply. Richter (1965) has postulated that the belt of greenstone cut by shear zones may represent a complex anticlinal structure whose axis follows the approximate centerline of Knight Island. The breadth of exposures on Knight Island is about 8 miles, or 42,000 feet, and the sequence probably extends downward at least an equivalent amount if the steep dips are maintained at depth. Available density measurements for rocks in Prince William Sound indicate that the average density of the greenstones is greater than that of the graywacke, slate, and granite with which the greenstones are in contact (table 1). The average density of the greenstone is 2.87 g per cm3, the slate and graywacke is 2.69 g per cm3, and the granitic rock is 2.60 g per cm3. Thus the density contrast between green- stone and the adjacent rocks is about 0.1—0.2 g per cm3. A simplified model of the green- stone mass on Knight Island has a computed gravitational effect which approximates the residual anomaly (fig. 4). In construction of the model and computation of its gravitational effects, the following \ Residual gravity anomaly GRAVITY ANOMALY, IN MILLIGALS A ox o o .nnl i l l l l KENAI PENINSULA poRT KNIGHT , , A BLACKSTONE COCHRANE NELLIE ISLAND Knight Green Montague A SEA LEVOEOLO, BAY BAY JUAN PASSAGE island Island lsljnd ,_ 8000,’ Graywacke, slate, and argiliite E 16,000, Graywacke, slate, argillite, Greenstone 24,000, and granitic(?) rocks An: +0.2 g per cm 0 10 20 MILES E 32,000 __.i ._ 40,000' Location of profile shown in figure 2 4,—Interpretation of the gravity anomaly over Knight Island. Ap is the density contrast. GRAVITY SURVEY, REGIONAL GEOLOGY, PRINCE WILLIAM SOUND EPICENTRAL REGION CII assumptions were made: (1) the source of the anomaly is two— dimensional, that is, its length is great with respect to its breadth; ('2) the average density contrast of the greenstone mass is about +0.2 g per cm3; (3) the width of the anomalous mass at the surface is about 40,000 feet; (4) it extends downward to a depth of about 40,000 feet below sea level, and (5) the contacts of the anomalous mass dip northwest, and the north- western contact dips more gently than the southeastern. Thus the model represents a mass whose breadth increases with depth. This configuration is consistent with Richter’s hypothesis that the vol- canic rocks form a major anticline. Gravitational effects of the sim- plified model were computed by the method described by Hubbert (1948). The computed effects ap- proximately coincide with the re- sidual anomaly, both in total ampli- tude and in gradient. Slight changes in density contrast or in dimensions of the mass would pro— duce an exact match, but these changes are not warranted because the dimensional assumptions and the regional gradient are more uncertain. The significant facts are: (1) a large greenstone mass is the most probable cause of the Knight Island anomaly; (2) its density is slightly greater than that of the adjacent rocks; (3) the mass probably extends to great depth, 40,000 feet or more below sea level; and (4) the gravity data indicate that faults may border both flanks of the Knight Island greenstone. The gravity highs form a contin- uous belt from the greenstone out- crops on Elrington and Latouche Islands, through the area of maxi- mum width of outcrop on Knight Island, to another high over the greenstone outcrops on Glacier Is- land, and then across Valdez Arm to the outcrops on Ellamar Penin- sula. No large greenstone outcrops have been mapped northeast of the Ellamar outcrops, and a Bouguer gravity measurement of — 1.3 mgal at the southeast end of Silver Lake suggests that the greenstones there are deeper and smaller. However, uncompleted surveys show positive gravity values 10 miles east of Barnes, D. F., 1966, Gravity changes during the Alaskan earthquake: Jour. Geophys. Research, v.‘ 71, no. 2, p. 451—456. Barnes, D. F., and Allen, R. V., 1965, Progress report on Alaskan gravity surveys [abs]: Alaskan Sci. Conf., 15th, College, Alaska, 1964, Proc., p. 165—166. Capps, S. R., and Johnson, B. L., 1915, The Ellamar district, Alaska: U.S. Geol. Survey Bull. 605, 125 p. Condon, W. H., 1965, Map of eastern Prince William Sound area, Alaska, showing fracture traces inferred from aerial photographs: U.S. Geol. Survey Misc. Geol. Inv. Map 1—453, scale 1:125,000. Condon, W. H., and Cass, J. T., 1958, Map of a part of the Prince William REFERENCES CITED Sound area, Alaska, showing linear geologic features as shown on aerial photographs: U.S. Geol. Survey Misc. Geol. Inv. Map I—273, scale 1:125,000. Grant, U. S., and Higgins, D. F., 1910, Reconnaissance of the geology and mineral resources of Prince William Sound, Alaska: U.S. Geol. Survey Bull. 443, 89 p. Hubbert, M. K., 1948, A line—integral method of computing the gravimetric effect of two—dimensional masses: Geophysics, v. 13, no. 2, p. 215—225. Ivanhoe, L. F., 1961, Bouguer gravity map of Alaska [abs]: Am. Assoc. Petroleum Geologists Bull., v. 45, no. 1, p. 128—129. Moffit, F. H., 1954, Geology of the Prince William Sound region, Alaska: U.S. Geol. Survey Bull. 989—E, p. 225—310. Valdez on the Richardson Highway, and 40—50 miles east of Valdez along the Tasnuna River up to its junction with the Copper River which could be continuous with the Prince William Sound high. If so, the incomplete data thus suggest that the belt of gravity highs may extend for 150 miles from Elrington Island at the coast to a point on the Copper River, where it is more than halfway across the Chugach Mountains. The computed profile (fig. 4) crosses the greenstone belt where the gravity anomaly as- sociated with it is largest. Al- though the magnitude of the high varies along the belt, the fact that there are no major horizontal offsets of the anomaly northeast of El- rington Island suggests that no major post—Eocene faults having strike—slip displacement measurea- ble in tens of miles obliquely cut the belt of gravity highs. The gravity survey has thus shown that the greenstone belt, which may be part of an ancient volcanic arc, consti- tutes one of the more continuous and significant geologic units along the Gulf of Alaska coastline. 1 Plafker, George, 1965, Tectonic deforma- tion associated with the 1964 Alaska earthquake: Science, v. 148, no. 3678, p. 1675—1687. Plafker, George, and MacNeil, F. S., 1966, Stratigraphic significance of Tertiary fossils from the Orca Group (Mesozoic), Prince William Sound, Alaska: U.S. Geol. Survey Prof. Paper 550-B, p. B62—B68. Plafker, George, and Mayo, L. R., 1965, Tectonic deformation, subaqueous slides, and destructive waves as- sociated with the Alaskan March 27, 1964, earthquake—an interim geo- logic evaluation: U.S. Geol. Survey open-file report, 21 p. Press, Frank, and Jackson, David, 1965, Alaskan earthquake, 27 March 1964 —Vertical extent of faulting and 012 elastic strain energy release: Science, v. 147, no. 3660, p. 867—868. Richter, D. H., 1965, Geology and mineral deposits of central Knight Island, Prince William Sound, Alaska: Alaska Div. Mines and Minerals Geol. Rept. 16, 37 p. Schrader, F. C., 1900, A reconnaissance of a part of Prince William Sound and the Copper River district, Alaska, in 1898: U.S. Geol. Survey Ann. Rept. 20, pt. 7, p. 341—423. Schrader, F. C., and Spencer, A. C., 1901, The geology and mineral resources of a portion of the Copper River dis- ALASKA EARTHQUAKE, MARCH 27, 1964 trict, Alaska: US. Geol. Survey Spec. Pub., 94 p. Shor, G. G., Jr., 1962, Seismic refraction studies off the coast of Alaska 1956— 1957: Seismol. Soc. America Bull., v. 52, no. 1, p. 37—57. Swick, C. H., 1942, Pendulum gravity measurements and isostatic reduc— tions: US. Coast and Geod. Survey Spec. Pub. 232, 82 p. Tatel, H. E., and Tuve, M. A., 1956, The earth’s crust, seismic studies: Car- negie Inst., Washington, Year Book 1955—1956, p. 81—85. Thiel, Edward, Ostenso, N. A., Bonini, W. E., and Woollard, G. P., 1958, Gravity measurements in Alaska: Woods Hole Oceanogr. Inst. Tech. Rept. Ref. 58—54, 104 p. 1959, Gravity measurements in Alaska: Arctic, v. 12, no. 2, p. 67—76. Woollard, G. P., Ostenso, N. A., and Thiel, Edward, 1960, Gravity anom— alies, crustal structure, and geology in Alaska: Jour. Geophys. Research, V. 65, no. 3, p. 1021—1037. Woollard, G. P., and Rose, J. C., 1963, International gravity measurements: Wisconsin Univ. Geophys. and Polar Research Center, 518 p. fl u.s. GOVERNMENT PRlNTING OFFICE: 1966—213-780 The Alaska Earthquake March 27, 1964 Kodiak and + w 3 Nearby Islands GEOLOGICAL S‘URV_E___Y PROFESSIONAL”"LPAPQR 543—0 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS Geologic Effects of the March 1964 Earthquake And Associated Seismic Sea Waves on Kodiak And Nearby Islands Alaska By GEORGE PLAFKER and REUBEN KACHADOORIAN GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—D UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1966 For sale by the Superintendent of Documents, US. Government Printing Office Washington, D. C. 20402 — Price 60 cents THE ALASKA EARTHQUAKE SERIES The U.S. Geological Survey is publishing the re— sults of investigations of the earthquake in a series of six Professional Papers. Professional Paper 543 describes the regional effects of the earthquake. Other Professional Papers describe the effects of the earthquake on communities; the effects on hy- drology; the effects on transportation, communica- tions, and utilities; and the history of the field in— vestigations and reconstruction effort. CONTENTS Page Page Page Abstract .......................................... D1 Surficial subsidence and asso- Vertical tectonic deformation—Con. Introduction .................................... 2 ciated ground cracks—Con. Regional setting of the changes D28 Purpose and scope of investiga- Alluvial and lacustrine Seismic sea waves .......................... 30 tion .......................................... 2 deposits .................................... D15 Chron01ogy ....... 30 Acknowledgements -------------------- 3 Artificial fills and embank— Crest heights ______________________________ 33 Geographic setting 3 ments ........................................ 15 Runup heights ____________________________ 34 Geologic setting --------------------- 6 Causes of surficial subsidence. 15 Periods ________________________________________ 35 The earthquake and Its after— Landslides ...................................... 18 Effects on shorelines and sea 5110ng """"""""" .' """""""""" 7 Distribution and nature .. .. 18 bottoms .................................... 37 Location and magnltude ........ 7 Causes ______________ 22 Ori . 37 Ground motion in the initial """"""""""""""" gin """"""""""""""""""""" shock ____________________________________ 7 Efiects on ground and surface Damage and casualties ............ 39 Surface waves ____________________________ 12 water ---------------------------------------- 23 Highways .................................... 39 Aftershocks ________________________________ 12 Observations ----- . .. 23 Fishing industry ........................ 42 Sounds __________________________________________ 13 Causes of fluctuations .............. 24 Cattle ranches ......... 43 Surficial subsidence and associ— Vertical tectonic deformation ...... 25 Logging industry 44 ated ground cracks ................ 14 Distribution ................................ 25 Casualties ................. 45 Beach deposits ............................ 15 Effects on shorelines ................ 27 References cited ............................ 45 ILLUSTRATIONS PLATES Facing page 1. Drowned stream estuary at Eagle Harbor along south of Ugak Bay ........................................... D10 2. Debris avalanche on the peninsula between Ugak and Kiliuda Bays ................. 11 3. Subsidence at Middle Bay ......... 30 4. Undermined trees on barrier beach in Izhut Bay, Afognak Island ................................................ 31 FIGURES Page Page Page 1. Physiographic map of south- 8. Submerged road at Lake 17. Map showing regional tec- central Alaska .................... D4 Rose Tead Lake ________________ D14 tonic deformation .............. D29 2- Physiographic map 0f KOdi' 9. Cracks and slumped delta at 18' Map showing wave damage ak and nearby islands ...... 5 Terror Lake ________________________ 16 and heights -------------------------- 31 3‘ Geologic sketch map """"""" 8 10. Sectioned extension crack.... 17 19' 111:2: wave crest, Womens 32 4' Mi? showmg aftershock (115' 11. Map showing landslide dis- 20. Breached barrier beach ........ 36 ributlon .............................. 9 . . . trlbution .............................. 19 21. Wave-damaged trees ............ 37 5' Mgfouigogyfonreigfigj 12. Rockslides ................................ 20 3: gas of gighway damage ...... 40 ' , ' ' _, 13. Reactivated debris slide on ' ri ge amage 41 33:12::1gnedMemlhm 10 Sitkinak Island .................. 20 24- ngagewiiimyed by 561mm 41 6. Kadiak Fisheries cannery 14- Earthflow on Sitinak Island-- 21 25. Roadway damage 'IIIIIILIIII 42 site ........................................ 11 15. Main scarp at head of Nar- 26. Ground vessels at Kodiak ...... 43 7. Piling at Kadiak Fisheries TOW Cape landslipe ------------ 21 27. Wave-deposited debris at cannery site ........................ 13 16. Isobase map ........................... 26 Beaty Ranch ...................... 44 V F999?!" TABLES Page Estimated property losses on Kodiak and nearby islands D3 Tidal constants at selected stations along the coasts of Kodiak and nearby islands ..................................................... 6 Tidal inundation resulting from tectonic and surficial subsidence 27 Heights and arrival times of seismic sea-wave crests at Womens Bay, Myrtle Creek, Terror River, and Uganik River on Kodiak Island ..... 34 THE ALASKA EARTHQUAKE, MARCH 27, I964: REGIONAL EFFECTS GEOLOGIC EFFECTS OF THE MARCH 1964 EARTHQUAKE AND ASSOCIATED SEISMIC SEA WAVES 0N KODIAK AND NEARBY ISLANDS, ALASKA By George Plafker and Reuben Kachadoorian Kodiak Island and the nearby islands constitute a mountainous landmass with an aggregate area of 4,900 square miles that lies at the western border of the Gulf of Alaska and from 20 to 40 miles off the Alaskan mainland. Igne- ous and metamorphic rocks underlie most of the area except for a narrow belt of moderately to poorly indurated rocks bordering the Gulf of Alaska coast and local accumulations of uncon- solidated alluvial and marine deposits along the streams and coast. The area is relatively undeveloped and is sparse— ly inhabited. About 4,800 of the 5,700 permanent residents in the area live in the city of Kodiak or at the Kodiak Naval Station. The great earthquake, which occur- red on March 27, 1964, at 5:36 pm. Alaska standard time (March 28, 1964, 0336 Greenwich mean time), and had a Richter magnitude of 8.4—8.5, was the most severe earthquake felt on Kodiak Island and its nearby islands in modern times. Although the epicenter lies in Prince William Sound 250 miles north— east of Kodiak—the principal city of the area, the areal distribution of the thousands of aftershocks that followed it, the local tectonic deformation, and the estimated source area of the sub- sequent seismic sea wave, all suggest that the Kodiak group of islands lay immediately adjacent to, and northwest of, the focal region from which the elastic seismic energy was radiated. The duration of strong ground motion in the area was estimated at 21/2—7 minutes. Locally, the tremors were pre- ceded by sounds audible to the human. ABSTRACT ear and were reportedly accompanied in several places by visible ground waves. Intensity and felt duration of the shocks during the main earthquake and aftershock sequence varied markedly within the area and were strongly in- fluenced by the local geologic environ- ment. Estimated Mercalli intensities in most areas underlain by unconsolidated Quaternary deposits ranged from VIII to as high as IX. In contrast, intensities in areas of upper Tertiary rock ranged from VII to VIII, and in areas of rel- atively well indurated lower Tertiary and Mesozoic rocks, from VI to VII. Local subsidence of as much as 10 feet was widespread in noncohesive granular deposits through compaction, flow, and sliding that resulted from vi— bratory loading during the earthquake. This phenomenon, which was largely restricted to saturated beach and allu— vial deposits or artificial fill, was locally accompanied by extensive cracking of the ground and attendant ejection of water and water—sediment mixtures. Numerous landslides, including a wide variety of rockfalls, rockslides, and flows along steep slopes, were trig— gered by the long—duration horizontal and vertical accelerations during the earthquake. The landslides are most numerous in a narrow belt along the southeast coast of Kodiak Island and the nearby offshore islands. Their abundance appears to be related to an area underlain predominantly by Terti- ary rocks. Temporary and permanent changes of level occurred after the earthquake in some wells, lakes, and streams throughout the area; ice was cracked, and the salinity of a few wells increas- ed. Permanent change of water level at some localities appears to be related to readjustments of fracture porosity by earthquake-induced movements of bed- rock blocks. Increased salinity of wells in coastal areas resulted from en- croachment of sea water into aquifiers after subsidence during the earth- quake, and to flooding of watersheds by seismic sea waves. Vertical displacements, both down- ward and upward, occurred throughout the area as a result of crustal warping along a northeast-trending axis. Most of Kodiak and all of Afognak, Shuyak, and adjacent islands are within a re- gional zone of subsidence whose trough plunges gently northeastward and ap- proximately coincides with the moun- tainous backbone of Kodiak Island. Subsidence in excess of 6 feet occurred throughout the northern part of the zone—a maximum subsidence of 61/2—7 feet having occurred on Marmot and, eastern Afognak Islands. Southeast of the axis of tectonic tilting, uplift of at least 21/2 feet occurred in a narrow zone that includes most of the south- easterly capes of Kodiak Island, the southeastern half of Sitkalidak Island, and Sitkinak Island. The uplift is in— ferred to extend offshore over much or all of the continental shelf adjacent to the Kodiak group of islands. Within the affected area, tectonic subsidence, which was locally augmented by sur- ficial subsidence of unconsolidated, D1 D2 deposits, caused widespread inunda- tion of shorelines and attendant dam- age to intertidal organisms, near- shore terrestrial vegetation, and salmon—spawning areas. The most devastating effect of the earthquake on Kodiak Island and near- by islands resulted from seismic sea waves that probably originated along a linear zone of differential uplift in the Gulf of Alaska. A train of at least seven seismic sea waves, having initial periods of 50—55 minutes, struck along all the southeast coast of the island group from 38 to 63 minutes after the earthquake. The southeast shores were repeatedly washed by destructive waves having runup heights along ex- posed coasts of perhaps as much as 40 feet above existing tide level, and of 8—20 feet along protected shores. Run- up heights of the waves were much less on the northwest and southwest sides of the islands, and no wave damage was incurred there. Locally, high- velocity currents that accompanied the waves caused intense erosion and re- distribution of unconsolidated natural ALASKA EARTHQUAKE, MARCH 27, 1964 and artificial shore deposits and of shallow sea-floor deposits. The Alaska earthquake was the greatest natural catastrophe to befall the Kodiak Island area in historic time. The combination of seismic shock and the earthquake—related tectonic de- formation and seismic sea waves took 18 lives, destroyed property worth about $45 million, and resulted in esti- mated losses of income to the fishing industry of an additional $5 million. Most of the damage and all of the loss of life were directly attributable to the seismic sea waves that crippled the city of Kodiak, wiped out the vil- lage of Kaguyak, and destroyed most of the village of Old Harbor and parts of the villages of Afognak and Uzinki. Bridges and segments of the highways in the vicinity of the city of Kodiak were washed out, and parts of the K0- diak Naval Station were inundated and damaged. Especially serious to all the damaged communities was the loss of fishing boats, seafood processing plants, and other waterfront installa— tions, which had been the mainstay of the economy. The great earthquake of March 27, 1964 was strongly felt throughout Kodiak Island and the nearby islands. Earth tremors triggered numerous landslides and avalanches, caused the ground to crack and subside in some areas of unconsolidated deposits, and affected water levels of streams, lakes, and wells. The earthquake was accompanied by regional tec- tonic warping that resulted in both extensive subsidence and local uplift of the land relative to sea level. It was followed after about half an hour by a train of at least seven destruc- tive large-amplitude seismic sea waves that repeatedly inundated low-lying segments of the shore. The loss of 18 lives and most of the property damage were due to seismic sea waves that battered the southeastern coast of the islands. Additional heavy losses INTRODUCTION resulted from inundation of coastal lowlands and waterfront installations as a result of the combined tectonic and local sub- sidence. Structural damage di- rectly attributable to seismic shock during the earthquake was light and was largely due to foun— dation failure of varying degrees. Total earthquake—related proper- ty damage in the area is esti- mated at about $45 million (table 1), and income losses to the fish- ing industry amount to an addi— tional $5 million. PURPOSE AND SCOPE OF INVESTIGATION This report presents the re- sults of a reconnaissance study of the earthquake effects over the vast uninhabited parts of the islands. The regional setting of the islands, and the effects of the earthquake at localities other than Additional heavy losses resulted from the combined regional tectonic and local surficial subsidence that oc- curred during the earthquake. Wide- spread shoreline flooding by high tides necessitated raising, protecting, or re- moving many installations otherwise undamaged by the earthquake or waves. Structural damage attributable to seismic shock during the earthquake was relatively light and was restricted to areas underlain by saturated uncon- solidated deposits. The chief structural failure in the area as a result of shak- ing was the collapse of part of a can- nery built on saturated beach deposits that were partially liquefied during the earthquake. Minor structural damage resulted from differential settlement and cracking of the ground on natural granular deposits and artificial fills. The overwhelming majority of struc- tures are constructed on indurated bed- rock; none of these sustained damage other than small losses resulting from shifting about and breakage of their contents. the urban areas are described. Detailed accounts of specific dam- age at the communities will be given in a separate volume of this series on earthquake effects on Kodiak and other communities on the Kodiak Islands. An effort was made to evaluate the local geologic factors that con- trol the distribution and charac- ter of the shock-induced effects because these factors have poten- tial significance in the design and construction of engineering works in seismically active areas. A re- connaissance of the tectonic changes in land level was made to delineate the areal extent and na- ture of these movements on the islands, and their relationship to the deformation that occurred elsewhere in south-central Alas- ka. These data are pertinent to any critical interpretation of the mechanics of the earthquake and TABLE 1.—Estimated property losses GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D3 on Kodiak and nearby islands Estimated Location Nature of damage replacement Source of data cost Kodiak City ............ Losses of private, $24,746,000 Tudor (1964). commercial, and public property. Afognak; village Losses of private 816,000 Bur. Indian Af- site abandoned. and public fairs (written property. commun., Sept. 25, 1964). Old Harbor ..... do 707,000 Do. Kaguyak; village .................. do .................. 321,000 Do. site abandoned. Uzinki ..... do .. 49,800 D0. Entire area ex- Commercial ................ 2,140,000 Alaska Dept. Fish cept city of and Game Kodiak. (1965). Entire area ............ Vessels damaged 2,466,500 Do. and lost. Kodiak Island Bridge and highway 3,359,000 Alaska Dept. High- exclusive of damage. ways (written Kodiak Naval commun., April Station. 1, 1964). Kodiak Naval Structures and 10,936,800 Tudor (1964). Station. equipment includ- ing bridges and highway. to the origin of the seismic sea waves that followed. Incidental to the geologic studies enumerated above, information was also ob- tained on the arrival times, runup heights, and characteristics of the seismic sea waves and the damage caused by them. The present report is based largely on studies made by Plaf- ker from float-equipped fixed- wing aircraft and helicopter dur- ing the periods July 14—20, 1964, and July 15—21, 1965; on studies made by Kachadoorian at Kodiak and vicinity July 14—18, 1964; and on observations on earth- quake effects made by G. W. Moore May 13—21, 1964. Moore also provided unpublished data on the geology of the islands. These observations were supplemented with studies of U.S. Coast and Geodetic Survey vertical aerial photography by the writers and W. H. Condon. Additional data was obtained from the numerous individuals and organizations list- ed in the acknowledgments. Kach- adoorian wrote the section on damage to the highways; the re- mainder of the report was written by Plafker. ACKNOWLEDGEMENTS We are greatly indebted to nu- merous individuals who provided us, in interviews and on form questionnaires, with eyewitness accounts of their experiences dur- ing and immediately after the earthquake. Personnel of the Fleet Weather Central, Kodiak Naval Station (Comdr. A. L. Dodson, Command- ing Officer) , obtained critical data on the seismic sea—wave sequence and the change in tide levels at the Naval base in Womens Bay and furnished helicopter transporta- tion to an otherwise inaccessible area of Narrow Cape. Dexter Lall and other biologists of the Alaska Department of Fish and Game provided unpublished data and nu- merous photographs of the earth— quake effects on coastal lakes and streams. June Brevdy of the U.S. Navy Radiological Laboratory furnished photographs, news- paper clippings, and published data on structural damage from seismic sea waves in the Kodiak area. Lt. Comdr. W. D. Barbee, of the U.S. Coast and Geodetic Survey, provided tide records and data on changes in land level at tide-gage stations on Kodiak Island. The U.S. Coast and Geo- detic Survey and the U.S. Army furnished postearthquake aerial photographs. Pierre St. Amand, of the U.S. Naval Ordnance Test Station, made available data ob- tained during an aerial recon- naissance of the area immediate- ly after the earthquake. Fred 0. Jones, consulting geologist, pro- vided photographs and data on earthquake effects in the ‘Terror River drainage basin. Pilots Bob Leonard and Al Cratty flew Plafker around the islands and freely shared with him their intimate knowledge of the Kodiak Islands area and of earthquake-induced changes that had occurred there. GEOGRAPHIC SETTING The group of islands described in this report, of which Kodiak lsland is the largest, lies at the western border of the Gulf of Alaska in the north Pacific Ocean between lat 56°30’ and 58°40’ N. and long 150°40’ and 154°50’W. (fig. 1). The group has an ag- gregate land area of 4,900 square miles, extends for a distance of 177 miles in a northeast direction, and is 67 miles wide at its great- est width. The largest islands of the group are shown on figure 2. ALASKA EARTHQUAKE, MARCH 27, 1964 D4 WRANGELL 100 MILES -central Alaska showing location of the Kodiak group of islands with respect to the epi- center of the March 27 earthquake and its zone of major aftershocks (lined pattern). Submarine contours in meters. 1. — Physiographic map of south 58° 57° GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA 154° 153° 152° I I l a CI id {Shuyak él an Shuyak Strait X j Marmot Island Raspberry I ‘ I’KJJX) ’ Kupreaan Strait‘ 0Q Uganikl h béon: 9'” \ q J Uzinki j 0‘3 Anton . Spruce . ‘3? Larsen ‘IsIand £9 Bay ‘ - w I 1&1 . A. , I. ‘ 3 W . " r 3 3w Chim‘ak ‘ l ‘ ‘ I . a” B” D Cape Chiniak M at” ‘4/pr0? WW g1) Y” fie/22w [’47 Q EXPLANATION Community X Cannery other than at communities shown + Cattle ranch [My , Sitkinak :3 2 Island : , ,, ,, ; ’ D ~ ~ ' 'J Government communications facility 0 10 20 30 MILES M 2.—Physiographic map of Kodiak and nearby islands. D5 D6 The population of the Kodiak group of islands is sparse, only about 3,600 permanent residents being reported in the 1960 cen- sus, in addition to 2,160 person— nel at the Kodiak Naval Station. The city of Kodiak and nearby Kodiak Naval Station have the only large concentrations of pop- ulation on the islands. The re- maining settlements are small fishing villages. Permanent pop- ulation of these settlements in 1960 was as follows: Community Population Afognak ................ 190 Akhiok .................. 7 2—7 5 Kaguyak ................ 36 Ka rluk .................... 129 Kodiak .................... 2,628 Larsen Bay ............ 7 2 Old Harbor ............ 193 Uzinki .................... 214 Of the total of some 5,700 per- sons living on these islands, more than 5,300 live within 20 miles of the city of Kodiak. Throughout the rest of the area the popula- tion is concentrated in the vil- lages with the exception of a few remote cattle ranches and Gov— ernment communications facili- ties. Several large canneries on the island employ numerous work— ers during the summer; at the time of the earthquake none of the canneries were in operation and their only occupants were the winter watchmen. Kodiak and the nearby islands are the structural extension of the Kenai-Chugach Mountains (fig. 1). This group of islands, and the Barren Islands that lie between the group and the Kenai Mountains, have been named the Kodiak Mountains section of the Pacific Border Ranges province (Wahrhaftig, 1965, p. 39). As a whole, the islands are mountain- ous, being lowest on the islands at the extreme north and south and highest in the middle (fig. 2). ALASKA EARTHQUAKE, MARCH 27, 1964 TABLE 2. — Tidal constants at selected stations along the coasts of Kodiak and nearby islands [Station: A, on the Gulf of Alaska; B, on Shelikof Strait. Predicted time and height of tides from U.S. Coast and Geodetic Survey (19648.). Local (Alaska standard) time; tide measurements in feet above lower low water] Annual tide Predicted tide, March 27—28, 1964 Station _ Low, evening, High, morning, (fig. 2) Dmmal Mean Highest March 27 March 23 range Time Height Time Height Perenosa Bay (A) ........... 11.3 5.9 13.4 7 :34 p.m. 0.1 1:29 a.m. 11.4 Izhut Bay (A) 8.9 4.5 11.0 7 :16 p.m. — .2 1 :13 a.m. 10.0 Kizhuyak Bay (A) 9.6 4.9 11.7 7 :06 p.m. — .2 1 :07 a.m. 9.7 Kodiak Harbor (A),.,..... 8.5 4.3 10.6 6 :52 p.m. — .3 0:57 a.m. 8.6 Ugak Bay (A) . 8.4 4.3 10.6 6 :35 p.m. —— .1 0:32 a.m. 8.6 Sitkalidak Island ( _ 8.3 4.4 10.6 6 :49 p.m. —— .1 0 :43 a.m. 8.6 Alitak Bay (AB) ....... 11.7 6.2 14.9 7:10 p.m. .1 0:59 p.m. 10.1 Uyak Bay (B) 13.8 7.3 13.8 7 :23 p.m. ——- .4 1 :24 a.m. 15.2 Uganik Bay (B) . 14.6 7.6 17.7 7 :22 p.m. -—< .3 1:25 a.m. 15.8 Malina Bay (B) ................. 14.5 7.7 17.7 7:24 p.m. —- .3 1 :26 a.m. 16.0 The western part of Kodiak Island has many broad alluvi- ated valleys and coastal lOW- lands that are underlain by thick glacial deposits. Sitkinak Island, at the south end of the group, has a maximum altitude of 1,470 feet. The surface of nearby Tugidak Island is a marine terrace that rises only about 100 feet above sea level. Shuyak Island and the northern part of Afognak Island are hilly lowlands. The coastline is extremely steep and irregular and is character- ized by many deep narrow fiords and rocky islets. Narrow marine terraces occur at intervals on the outer capes along the southeast coast of Kodiak and Sitkalidak Islands. Bars and spits have formed across the mouth of many of the bays and inlets. Numerous shallow lagoons and lakes occur behind low barrier beaches along the shore. There is a marked difference in the range of tides on the ocean side and on the Shelikof Straits side of the islands. Mean annual tide range is 13 to 141/; feet along Shelikof Strait in contrast to 8.2 to 11.3 feet at most places on the ocean coast. Furthermore, the difference in maximum annual tide height on the two sides of the island may be as much as 7 feet. These differences are signifi- cant to this study because they strongly influenced the amount and distribution of the damage caused by seismic sea waves and by tectonic subsidence along the coast. Table 2 shows the tidal constants at selected localities, and the time and height of high and low tide during the period when seismic sea waves struck the coast on the night of March 27. On March 27, 1964, the weath- er was mild; highest and lowest reported temperatures were 37° F. and 20° F.; there was little or no snow on the ground at sea level, but traces of snow fell dur- ing the day at Kodiak Naval Sta- tion and Larsen Bay (US. Weather Bureau, 1964). Winds were light with scattered high overcast. All the lakes were froz- en over. GEOLOGIC SETTING The oldest rocks in the area lie along the northwest coast of Ko- diak Island and consist chiefly of metavolcanic and marine sedi- mentary rocks of Triassic and Jurassic age that are cut by mafic intrusives. Younger Mesozoic GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D7 marine sedimentary rocks of prob- able Cretaceous age, complexly deformed and extensively intrud- ed by granitic rocks, underlie the axial part of the Kodiak Moun- tains. Northeast-trending belts of downfaulted lower Tertiary rocks make up the southeast side of Kodiak Island, Sitkalidak, Tugi- dak, and Sitkinak Islands. These belts are composed of a sequence of highly inclined and intensely folded sandstone, shale, conglom- erate, and altered volcanic rocks that are locally cut by felsic in- trusives. Overlying the older rocks are poorly consolidated up- per Tertiary marine sedimentary rocks that have been gently to moderately folded. Fold axes and major faults strike northeast, the northwest blocks of the faults generally being upthrown. Dis— tribution of the major lithologic units is shown in figure 3 (next page). On Kodiak and the nearby is- lands, unconsolidated deposits as a rule are thin and discontinuous. They consist mainly of glacial debris, alluvial and delta deposits, and beach deposits. Local thick accumulations of colluvium occur along the bases of the steeper slopes, particularly in areas un- derlain by Tertiary rocks. Most of the coast consists of rugged rocky bluffs ; extensive and well- developed beaches are uncommon. A significant proportion of the alluvial deposits in an east-west belt across the central part of the islands consists of virtually un— compacted volcanic ash and pum- ice deposited in the spring of 1912 during the eruption of Kat- mai Volcano, on the Alaska Pen- insula to the west (Capps, 1937, p. 170—171). The thickest and most exten— sive alluvial and glacial deposits are delineated on figure 3. Areas where bedrock is covered by thin deposits of soil, colluvium, glacial debris, stream gravel, marine beach deposits, or artificial fill are not shown. THE EARTHQUAKE AND ITS AFTERSHOCKS LOCATION AND MAGNITUDE The earthquake began at 5:35:12.15:0.25 p.m. Alaska standard time (S. T. Algermis— sen, US. Coast and Geodetic Sur- vey, oral commun., April 22, 1965); its hypocenter is 12—31 miles deep (20—50 km) ; and the epicenter is located at lat 61.1° N., long 147.7° W. in northern Prince William Sound, 250 miles northeast of the Kodiak group of islands (fig. 1). Estimates of Richter magnitude, based upon surface wave amplitudes (Ms) are in the range of 8.4 to 8.5 according to the US. Coast and Geodetic Survey. The focal re- gion, or zone of probable fault breakage during the earthquake, is approximately outlined by the spatial distribution of after- shocks. As shown by figure 1, this zone extends 400 miles south- westward from the epicenter in a broad belt that includes much of the continental shelf and south- eastern margin of the Kodiak group of islands. Perhaps related to the earth- quake was an apparent instantan- eous increase in the magnetic field by about 100 gamma on a continuously recording magnet- ometer in the city of Kodiak (Moore, 1964, p. 508). It is pos- sible, however, that the anomaly, which occurred 1 hour and 4 min- utes before the tremors were first felt, was caused by a local sur- face disturbance unrelated to the earthquake. During the first month after the earthquake, more than 7,500 aftershocks were recorded at seismograph stations in the focal region (S. T. Algermissen, oral commun., April 22, 1965). Loca- tions of the larger aftershocks (magnitudes of 5.0 or greater) with epicenters in the vicinity of the Kodiak group of islands are indicated on figure 4. Except for two aftershocks centered in the western part of the islands (p. D9) they all lie along the south- east coast of Kodiak, Sitkalidak, and Sitinak Islands and on the adjacent continental shelf. The largest aftershock recorded in the mapped area had a magnitude of 6.1. Focal depths range from about 9 to 19 miles (15—30 km). GROUND MOTION IN THE INITIAL SHOCK Strong ground motion that lasted from 21/2 to 7 minutes was experienced throughout the study area. However, marked variations in the duration, intensity, and na- ture of the shaking were reported in different parts'of the islands. According to most observers, the earthquake started with a gentle rolling motion for a period of 20 seconds to 1 minute, shook hard for about 21/; minutes, and then gradually subsided. The ground motion was generally described as rolling like sea waves, or as a strong horizontal oscillation. The jarring motion that was felt in D8 ALASKA EARTHQUAKE, MARCH 27, 1964 EXPLANATION UNCONSOLIDATED DEPOSITS Surficial deposits Include glacial, alluvial, and marine deposits, and artificial fill BEDROCK Sedimentary and volcanic rocks E Marine sedi— mentary rocks TERTIARY QUATERNARY Felsic and mafic intru- sive rocks CRETA- CEOUS (C7) Contact Metavolcanic and marine Fault sedimentary Northwest side upthrown L rocks TRIASSIC AND J URASSIC WWWW (i " 1315‘ 3.— Geologic sketch map of Kodiak and nearby islands. Geology generalized after Capps (1937) and from unpublished data by G. W. Moore. GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA g lzzéf'i‘nlukg :‘x E Widely felt aftershock 30 of April 14th 20 5.2 53: 2 _ 1—'5 20%)” 053 05.8050 832 23 E 25 33 5_-6 15 25 22 O 05.30ii E 20 15 ”Q) 023 35 25 30 72 (31 , 4.——Aftershock distribution (magnitude greater than 5.0) in t e 25 5.10 02 33 05.4 20 35 E» '_ ._‘_. 20 25 O 05.2 5-1 OH 30 3—3 15 E 5 00 L02 1 5 _. 05.2 15 35 05$ 20 5.1 D9 Kodiak Island area. Numbers indicate Richter magnitude (above) and focal depth in kilometers (below). Data after US. Coast and Geodetic Survey, Preliminary determination of epicenter cards, 1964. Me‘m Port Williams VI Cannery \X ngg‘ , 5» PortWakefieldv ‘ ‘5» y Cannery]: VI?VII A , :4 3» Karluk u, , y I,“ ' ’ Cape Chlnlak A \w VX—Yly ' 08. Station 7 ; VH‘VIII , ’ v)» x xx x _ , Saltery’; qye VII—VIII V ‘ Ran 'c h w Narrow Cape >y X Ranch ’ \ NKEMVW files-o \ / Port Hobron Ranch 'KALE AK S‘JYKENAK TELANL? Sitkinak Ranch x I’UGEUAK EELAN i") _,s< Eh». Us 5.—Reported ground motion, sounds, ground waves, and assigned Mercalli intensities on Kodiak and nearby islands. The directions of reported ground motion are shown by arrows; localities at which sound was reported are shown by solid circles, localities at which ground waves were reported by open circles, and assigned Mercalli intensities by Roman numerals. GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—D PLATE 1 DROWNED STREAM ESTUARY AT EAGLE HARBOR ALONG SOUTH SHORE OF UGAK BAY Extensive brown area of terrestrial vegetation has been killed by salt-water immersion after 4 feet of tectonic subsidence and an unknown amount of surficial subsidence. The gray ridge of beach gravel in the left part of view (arrows) has been eroded back and built up in adjustment to the new higher base level of deposition. The fact that the effects of inundation are scarcely visible along the steep bedrock shore in the background suggests that much of the subsidence at the estuary deposits was aurficial rather than tectonic. GEOLOGICAL SURVEY PROFESSIONAL PAPER 543-D PLATE 2 DEBRIS AVALANCHE ON THE PENINSULA BETWEEN UGAK AND KILIUDA BAYS A slide of Tertiary rocks from the 1,500-foot-high peak at upper right flowed into the uninhabited valley below at about 300-foot altitude where it spread out as a debris lobe roughly 1,500 feet across. The narrow streak of light-colored debris in the lower right corner is part of the slide mass that overflowed the near flank of the landslide scar. GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D11 6.——The Kadiak Fisheries cannery site in Shearwater Bay along the north shore of Kiliuda Bay. Cannery buildings in the foreground were split during the earthquake and the seaward parts of the structures were later washed away by seismic sea waves. The former locations of the buildings and dock are outlined by timber pile foundations in the beach and water. The entire beach is now inundated at high tide as a result of about 4 feet of tectonic subsidence and 2-10 feet of surficial subsidence of the beach deposits. The barge behind the cusp is a floating cannery that replaces the destroyed facility on land. Location of cannery shown on figure 5. some areas closer to the epicenter Was not experienced on Kodiak or the nearby islands during the in- itial shock, although it was re- ported in some of the larger af- tershocks. A local resident, Mr. Rudy Lorensen, timed the dura- tion of the strongest ground mo- tion as 21/2 minutes on a bedrock site near the city of Kodiak. Reports 'of the directions of vibrations show a wide variation. The apparent variations may be in part due to the difficulty in determining the direction from which a shock comes during such a distressing experience. In part they may result from a change in the source of the vibrations as the rupture propagated south- westward from its epicenter in Prince William Sound or to local variations in ground conditions. Figure 5 shows that the reported Vibration directions were mostly northeast-southwest to north- south, although in a few instances they were northwest—southeast. The intensity, or the destruc- tive potential, of the Vibrations in the area varied markedly from place to place and appeared to be closely controlled by the local geo- logic environment. The generali- zation that both the intensity and duration of earthquake Vibrations are enhanced in unconsolidated ground, particularly on water- soaked unconsolidated deposits (Gutenberg, 1957), is strikingly borne out by the distribution of vibration-induced damage in the area. The most severe shaking oc- curred in areas of thick saturated unconsolidated Quaternary de- posits throughout the islands. In these areas, structural damage resulted mainly from foundation failure attributable to partial liquefaction, differential settle- ment, and cracking of unconsoli- dated deposits and artificial fill. The sole structural failure re— ported in the area was the partial collapse of the main building of the Kadiak Fisheries cannery lo- cated on a beach cusp on Shear- water Bay along the north shore of Kiliuda Bay. The structure split in half when it was shifted off its piling foundation accord- ing to the cannery watchman, Mr. Sergei Pestrikoff. The piling was driven into beach gravel that was partially liquefied during the earthquake (fig. 6). The fallen section was later washed away by the seismic sea waves. D12 SURFACE WAVES Reliable data on the amplitude and period of the ground waves that accompanied the earthquake in the area are unavailable. Long- period surface waves set bodies of surface water, trees, radio an- tennas, and hanging objects such as lighting fixtures and wall pic— tures into violent oscillation throughout the islands but caused only minor damage to manmade structures. The period of an oscil— lating antenna pole about 50 feet high on a bedrock site was timed by one observer as “slightly less than one second.” Visible traveling undulations of the ground surface in areas un- derlain by unconsolidated depos- its were reported at the Villages of Uzinki and Afognak. The wavelength at Uzinki was de- scribed as 30 feet; estimates of amplitude could not be made. Similar progressive linear earth waves having estimated ampli- tudes as great as 3 feet were ob- served at numerous localities un- derlain by unconsolidated deposits in the part of south—central Alas- ka where the earthquake was strongly felt. Such waves have been frequently noted in compar- able geologic environments dur- ing other great earthquakes throughout the world. The waves may correspond to the large am- plitude “hydrodynamic” waves discovered by Leet (1946, p. 209— 211) on seismograph recordings of underground explosions. It is conceivable that such waves, which have particle orbits com- parable to those of water waves, may be propagated in ground that has become semifluid as a result of ground vibration. Elsewhere in areas underlain by consolidated deposits and in areas of poorly consolidated Ter- ALASKA EARTHQUAKE, MARCH 27, 1964 tiary rock, there was minor cracking of interior plaster, con— crete and concrete—block walls and floors, and a few concrete- block chimneys, and rupture of buried pipelines. In some of these areas slight damage was incur- red when heavy objects such as generators, appliances, and fur- niture were shifted about or over- turned. In areas underlain by un- consolidated deposits it was gen- erally difficult or impossible for people to stand or walk during the earthquake. Ground motion in areas of pre— Tertiary bedrock or bedrock man- tled by thin unconsolidated de- posits—areas where most com- munities and canneries of Kodiak and the adjacent islands are lo- cated-—was not strong enough to cause structural damage. Some small light objects or heavier ob- jects in precarious positions were thrown down, although in most places inanimate objects were not disturbed. People report that, al- though difiicult, it was possible to stand and move about during the earthquake. Estimates of the Mercalli in- tensity of ground shaking in the area, based on the reported ef- fects on people and inanimate ob- jects, are shown on figure 5. Ground-water effects, landsliding, and avalanching, which were widely distributed and bore no clearcut relationship to structural damage, were not considered as suitable criteria for intensity esti- mates. Except as noted, assigned intensity ratings are according to the abridged Modified Mercalli Intensity scale of 1931, in which the violence of ground motion is separated into 12 categories of increasing destructiveness from I to XII (Hodgson, 1964, p.58-59). With one exception, assigned Mercalli intensities range from‘ VI to VIII throughout the area. They are lowest in areas of pre- Tertiary bedrock or bedrock man— tled with thin well-drained un- consolidated deposits and highest in areas of poorly consolidated Tertiary rocks and thick uncon- solidated deposits. Significantly, the highest intensity experienced, which is estimated as VIII to IX on the scale, was at the Kadiak Fisheries cannery—the only large structure in the entire area that is founded wholly on unconsolid- ated, saturated, granular de- posits. AFTERSHOCKS Many of the numerous after- shocks that followed the main shock were felt as short, sharp, jarring shakes. They reportedly. occurred at intervals of a few minutes to an hour throughout the night of the earthquake and at increasingly longer intervals for a period of several weeks thereafter. One aftershock, at 12:55 p.m. on April 14th, was especially noted throughout the islands. Its epicenter was about 20 miles northwest of the city of Kodiak (fig. 4), the magnitude is given as 5.4, and the focal depth as about 19 miles (30 km). Oddly enough, although this af- tershock was the strongest one felt in the area, its magnitude was less than many of the other nearby aftershocks. It reportedly hit with a short jarring motion equal to the intensity of the main shock, after which it quickly sub- sided. Although it caused major consternation, it resulted in no noteworthy damage on Kodiak Island except for three broken pipelines at one cannery 0n the northwest coast of Kodiak Island, and a cracked roof at another cannery in Alitak Bay on the southwest coast of the island. GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D13 ' SOUNDS The initial shock was heard as a low-pitched rumble by two in- dividuals in the area. One of them reportedly heard the rumbling an estimated 5 seconds before the tremors were felt. During the weeks after the earthquake, many people heard aftershocks as deep rumbles a second or so before they were felt. Some individuals, both in the area and elsewhere in south-central Alaska, said they heard warning rumbling sounds before every felt aftershock. Richter (1958, p. 128) postulated that such noise may be produced during passage of the body waves, but that the shaking is felt only upon arrival of the slower mov- ing surface waves. SURFICIAL SUBSIDENCE AND ASSOCIATED GROUND CRACKS Vibratory loading of noncohe- sive granular deposits during the earthquake resulted in local sur- ficial subsidence through compac- tion, flow, and sliding. Areas of greatest subsidence were in thick, unconsolidated, saturated marine or lacustrine deposits and uncom- pacted manmade fills. Subsidence in some of these areas was ac- companied by cracking of the ground and ejection of water or water-sediment mixtures. By the time this reconnaissance study was initiated 4 months after the earthquake, however, much of the evidence of ground cracking was obliterated. Surficial subsidence of uncon- solidated beach and delta deposits along the coast was most clearly manifested in the relatively great- er amount of inundation of shore- line areas as compared to nearby rock outcrops. Virtually all areas underlain by such deposits, both in the zone of tectonic subsidence and slight uplift, show some indi- cation of salt water immersion in the form of dead brown terres- trial vegetation (pl. 1). BEACH DEPOSITS At the Kadiak Fisheries can- nery along the north shore of Kiliuda Bay, local subsidence of beach deposits was 2—10 feet more than that of the adjacent bedrock areas. The cannery is on a broad, roughly triangular cusp that juts out into Shearwa— ter Bay—a small bay on the north side of Kiliuda Bay (figs. 5, 6). Before the earthquake, the buildings shown in the foreground of figure 6 extended on piling slightly beyond the water’s edge, as shown in the photograph. Be- yond this was a pier 170 feet long. The piling had been driven to refusal, that is, 10—15 feet into the beach deposits. The piles were originally vertical and the tops were level. After the earthquake, measur- ed subsidence was 51/3 feet at the shipway piling to the right (north) of the cannery buildings. This subsidence, 2 feet more than that at a bedrock site 1 mile to the south, indicates surficial set- tlement of about 2 feet relative to bedrock. Furthermore, meas- urement of the difference in height of the tops of the piling that previously supported the de- stroyed part of the cannery indi- cates that an additional 7.8 feet of surficial settlement occurred towards the tip of the cusp (fig.7) . 7. — Timber pile foundation at Kadiak Fisheries cannery site. Pronounced tilting of unbroken piling is indicative of lateral motion of the piling through 10—15 feet of beach gravel. Piling tops formerly level are now 7.8 feet lower at the seaward end of the cannery site than in the foreground owing to local settle- ment of the distal part of the beach cusp. Bent drift pins in the piling tops indicate that the cannery moved southward (away from the observer) off the piling. D14 ALASKA EARTHQUAKE, MARCH 27, 1964 8.—Road near the head of Lake Rose Tead submerged by about 5 feet of surficial subsidence of inlet—stream deltaic deposits. Ugak Bay is visible in the background. Photograph by Alaska Department of Fish and Game. GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D15 The site is now submerged at high tides, and the facilities have been replaced by a floating can- nery. Ground cracks as much as a foot wide reportedly criss- crossed the cusp after the earth- quake. Mr. Edward Pestrikoff, the cannery watchman, saw cracks form in the ground during the earthquake and saw water ejected as high as 6 feet. Reduc- tion of shear strength of the un- consolidated deposits through par- tial liquefaction is suggested by the manner in which the founda- tion pilings shown in figure 7 were tilted without being broken when the cannery building col- lapsed. Elsewhere, local differential subsidence of unconsolidated beach deposits relative to nearby bed- rock amounted to an estimated 1—2 feet at Old Harbor and per- haps 5 feet near the McCord cat- tle ranch on Sitkalidak Island. ALLUVIAL AND LACUSTRIN E DEPOSITS Mr. Joe Beaty, a long-time cattle rancher at Narrow Cape, observed that during the earth- quake his truck and other heavy machinery settled as much as 2 inches into alluvial deposits on the flood plain of a small stream on the cape. He also noted much cracking at a bog between two lakes near the tip of the cape, and extrusion of gravel that had been buried beneath 20 feet of fine-grained surficial alluvial de- posits; thus the cracks evidently extended toat least that depth. Local subsidence of lacustrine delta deposits was noted at three localities by personnel of the Alaska Department of Fish and Game who carried out studies of the earthquake’s effects on sal- mon spawning and sport fishing (Alaska Dept. Fish and Game, 1965, p. 3—21). Several feet of subsidence occurred on deltas of three streams that enter Buskin Lake near the city of Kodiak and at the inlet stream deltas of two large lakes—Lake Rose Tead and Saltery Lake—along the north shore of Ugak Bay. Surfi- cial subsidence has resulted in lo- cal flooding of shoreline vegeta— tion of these lakes and inundation of a segment of road along the shore of Lake Rose Tead (fig. 8). Mr. Ron Hurst, a rancher at Saltery Lake, reports that several acres along the lakeshore sank as much as 3 feet, and large amounts of subsurface pumice was pump- ed up into Saltery Lake and caused excessive turbidity early in the summer of 1964. Slumped lake deltas, ground cracks, and spectacular sand-vent deposits were examined and pho- tographed by F. 0. Jones, con- sultant geologist to the Chugach Electric Association, in the up- per Terror River drainage, on Kodiak Island between Ugak and Viekoda Bays at about 1,250- 1,400 feet altitude. The inlet del- ta of Terror River at the head of Terror Lake slumped and sub- sided during the earthquake; near-shore vegetation there was put under 2—3 feet of water (fig. 9, next page). Lateral extension of the slumped delta margin caused it to be cut by a system of rectilinear cracks as much as 1 foot wide (fig. 9). Large vol- umes of light-colored fine- to me- dium-grained pumice were eject- ed to the surface from many of these cracks, and the pumice forms clastic dikes that fill the cracks below the surface (fig. 10 p. D17). Jones noted similar slumped and cracked deltas else- where along the margin of Terror Lake and saw from the air an extensive area of reticulate cracks and sand-vent deposits in a par- tially filled lake basin 3 miles up- stream from Terror Lake. Drowning of shoreline vegeta- tion was noted from the air along the margins of numerous lakes throughout the area. Undoubtedly shoreline subsidence, cracking of the ground, and extrusion of sedi— ment similar to that noted above also occurred in comparable geo- logic environments at the numer- ous other lakes on the island that were not examined in detail. ARTIFICIAL FILLS AND EMBANKMEN TS In addition to the subsidence of natural unconsolidated deposits, damage was caused to engineer- ing works by surficial subsidence and cracking of filled ground and by settling of engineered fills and breakwaters into underlying un- consolidated natural deposits. Most of these effects, which oc- curred in the Kodiak area and along the highway east of Kodiak, will be discussed separately in the section of this report that de- scribes effects on highways and in the companion report on dam- age to communities. The only other noteworthy examples of subsidence of artificial fill oc- curred at the US. Coast Guard Loran facility on Sitkinak Island where measured differential set- tlement of 47/8 inches occurred at one corner of a reinforced-con- crete structure that was reported- ly built on a 12—foot-thick fill. At Narrow Cape the approach fill of a ranch road on alluvial de- posits sank about 2 feet below the level of a bridge deck. CAUSES OF SURFICIAL SUBSIDEN CE From the foregoing descrip- tions it should be apparent that surficial subsidence of unconsoli- D16 ALASKA EARTHQUAKE, MARCH 27, 1964 defined by light—colored pumiceous ejecta. Photograph by F. 0. Jones. GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D17 10.—Sectioned extension crack 1 foot wide at head of Terror Lake showing light-colored fine— to medium-grained pumiceous sand ejecta at surface and subsurface c‘lastic dike. Photograph by F. 0. Jones. D18 dated deposits and attendant ground cracking occurred under a variety of geologic conditions over a vast area. Causes of the subsidence are probably diverse and are related to the physical properties of the materials com- posing each deposit, as well as to the configuration of the deposit. Subsidence of natural granular materials and uncompacted fills under Vibratory loading may re- sult from: (1) compaction, or volume reduction due to closer packing of the grains that com— pose the deposit; (2) varying ALASKA EARTHQUAKE, MARCH 27, 1964 degrees of liquefaction of satu- rated incoherent material with resultant lateral spreading through flow or sliding; (3) shear failure causing downward and outward movement of the slide mass; and (4) combinations of (1), (2), and (3). Details of the physical proper— ties and configuration of the af- fected materials before and after the earthquake are unavailable so that the specific cause, or causes, of subsidence are generally not known. Compaction of some of the deposits is suggested by sur- face ejection of water and water- sediment mixtures. Reduced shear strength through liquefaction is indicated at the Kadiak Fisher- ies cannery, where piling embed- ded 10—15 feet in beach deposits was dragged laterally and tilted to angles of as much as 20° with- out failure. Reduced bearing strength and compaction of some ground is suggested by the mass sinking of compacted fills, break— waters, structures, and heavy ob- jects into underlying alluvial de- posits and artificially filled ground. ‘ Long-duration horizontal and vertical accelerations during the earthquake triggered numerous landslides and avalanches on the slopes of Kodiak and the adja- cent islands. The landslides in- cluded a wide variety of falls, slides, and flows involving bed- rock, unconsolidated deposits, and snow in varying proportion.1 Landslides of rock and soil pre- sumably loosened by shaking dur- ing the initial shock were unusu- ally abundant in the months after the earthquake. Some of them may have been triggered by the larger aftershocks or by sum- mer thaw of frozen ground. None of the landslides or avalanches in the area caused property damage or human casualties. The following general descrip- tions of the distribution and na- ture of the landslides and aval— anches are based almost entirely on aerial observation, interpreta- tion of postearthquake aerial pho- tographs, and accounts by local lThe classification and nomenclature for landslides used herein correspond to those of Verna (1958). LANDSLIDES residents. Only a few of these features were examined on the ground. DISTRIBUTION AND NATURE Most of the landslides occurred along the southeast coast of Ko- diak Island and on the islands offshore from Kodiak Island. A few other landslides—mainly large rockfalls—occurred at widely scattered sites on Kodiak Island and on Afognak and Mar- mot Islands. The generalized ar- eal distribution of the larger land- slides that occurred during or after the earthquake and before August 1964 is shown in figure 11. Many of these slides were re- activated older slides. In some places—most notably along steep coastal bluffs—rockfalls are so closely spaced that a single sym- bol on figure 11 necessarily rep- resents several slides. No attempt was made to map the smaller rockfalls and soil slips, although, as might be expected, their fre- quency is roughly proportional to that of the large slides. Numerous snow avalanches and debris avalanches 0f snow-rock mixtures are visible on aerial photographs of the steep slopes at higher altitudes along the rug— ged backbone of Kodiak Island. They are not shown on figure 11 because the postearthquake aerial coverage was inadequate to per- mit photogeologic mapping of their distribution. Furthermore, most of the avalanches at high altitudes were obscured by new snowfall by the time this study was initiated 31/; months after the earthquake. In approximate order of abund- ance, the landslides in the area may be divided into the following general categories: (1) rock— slides and rockfalls in areas of consolidated rock, (2) debris slides and avalanches of Tertiary and Quaternary deposits, (3) earthflows and soil slips of un- consolidated surficial materials, and (4) rotational slumps in both consolidated rock and soil. Many of the landslides observed are complex features that probably GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA ' 5 x = hm» D19 11.—Generalized distribution of the larger landslides and soil slips triggered by the March 27, 1964, earthquake on Kodiak and nearby islands. Areas underlain by rocks of Tertiary or probable Tertiary age are shaded. D20 ALASKA EARTHQUAKE, MARCH 27, 1964 12,—Rockslides in tightly folded blocky sandstone along north shore of Ugak Bay, Kodiak Island. Bluff is about 1,000 feet high and is exposed to surf from Gulf of Alaska. 13. — Reactivated large debris slide in sedimentary rocks of middle Tertiary age along the south shore of Sitkinak Island. Note the characteristic steplike trans- verse scarps, displacing surface vegetation at head of the slide. A small rockfall covers part of the beach in lower left of photograph. Vertical height of slide mass is approximately 300 feet. involve varying combinations of falling, sliding, and fiowage, but in the descriptions that follow they have been classified on the basis of their most characteristic features as interpreted mainly from the aerial photographs and from aerial observation. Rockfalls and rockslides are locally very numerous along steep coastal bluffs (fig. 12). Common- ly, the bluffs are partially under- cut by wave erosion and have surf-cut sea caves near the high— tide line thatvlead to precipitous failures of large rock masses. The rock falls and slides are par- ticularly numerous where blocky massive sandstone and basalt of Tertiary age, or older igneous intrusive rocks, form the steep bluffs along the shore. They are relatively rare in areas underlain by the well-indurated bedded and slaty rocks of pre-Tertiary age that compose the central part of Kodiak Island and most of Afog- nak and Shuyak Islands (fig. 3). Many rockfalls, too small to be plotted on the landslide distribu- tion map, were seen in areasun- derlain by the blocky granitic rocks that compose much of the rugged drainage divide on Kodiak Island. . Debris slides and debris aval- anches were sporadically distrib- uted along the southeastern coast of Kodiak Island, on Sitkalidak Island, and on Sitkinak Island. They apparently occurred mostly in areas underlain by bedded rocks of Tertiary age or by Quaternary deposits. Such slides are especially numerous on deeply weathered, moderately steep slopes underlain by rocks of mid- dle to late Tertiary age. The de- bris slide shown in figure 13 is typical of many throughout the area. The slide illustrated on plate 2 is probably the largest that oc- curred in the area and is anoma- lous in that it spread as a rela- tively thin sheet over the valley below the slide source. The areal distribution and irregular surface characteristics of the slide debris suggest that it was emplaced by high-velocity viscous flow. GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D21 Earthflows and soil slips of un- consolidated materials occurred throughout the area, but were es- pecially common along the moder— ately steep slopes that were deep- ly weathered and mantled with thick deposits of one or more of the following: (1) glacial till, (2) volcanic ash, (3) colluvium. Typically, the flows and slips are manifested at the head as a series of transverse “wrinkles” or scarps stepped downslope, below which is a subtle bulge of debris (fig. 14). All gradations occur between the earthflows and soil slips that in— volve thin surficial layers of un- consolidated debris and the debris slides that are characterized by relatively coherent slide units. Rotational slumps were rather rare in the study area. This type of landslide was recognized only at Narrow Cape on Kodiak Island and along the southeast cape of Sitkalidak Island. Pierre St. Amand of the US. Naval Ord- nance Test Station and George W. Moore of the US. Geological Survey directed our attention to these features, which were stud- ied with special care because their long linear scarps bore a super- ficial resemblance to fault breaks. The Narrow Cape landslide in- volved moderately dipping, poor- ly indurated sedimentary rocks of late Tertiary age. It occurred on slopes of 15°—35° along a struc- turally controlled northeast-trend- ing linear valley that is incised 125—150 feet into a flat-lying ero- sion surface. The main scarp of the slide (fig. 15) broke at the top of the southeast valley wall along a distance of roughly 1,400 feet. The slide mass slumped down on the northwest side as much as 15 feet with headward tilting along a surface of rupture that intersected the surface on the valley floor 125—150 feet be- ‘3‘”, _ will“: 7‘ f ';‘”‘“:$;,. .%4.. {”59“ ~’:.. 3 14. — Earthflow resulting from soil slip or slump in deeply weathered Tertiary sedimentary deposits along the north shore of Sitkinak Island. The head of the slide is marked by numerous small scarps; the foot is the hummocky bulbous mass in the lower left corner. 15.——Main scarp at the head of a large rotational slump in Tertiary deposits on Narrow Cape, Kodiak Island. The landslide is approximately 1,400 feet long and 125—150 feet high. The rock mass at the left slid 8—15 feet downslope along a surface of rupture, the main scarp of which trends diagonally across the photograph. Note the headward tilt of the original surface on the upper part of the slide mass and draping of the vegetation mat over the crown of the main slide. D22 low. Open tension cracks and shallow grabens are abundant on the erosion surface headward from the main scarp. The slide toe is characterized by transverse pressure ridges and small over- thrusts in the surface mat of muskeg vegetation. The brush- covered hummocky topography of the lower part of the slide sur- face and the chaotic internal structure suggest that earlier gravity movement has taken place. At the southeast end of Sit- kalidak Island, a discontinuous zone of landslides occurs along the base of a moderately steep southeast-facing ridge that fronts on a broad, gently sloping terrace. The landslides have exposed a series of southeast-facing scarps that can be traced along the base of the ridge for about 11/2 miles. They die out both to the north- east and the southwest. As seen from the air, the freshly exposed scarps are estimated to be as much as 15 feet high and appear to be largely in unconsolidated de- posits. The slide masses consist of both headward—tilted blocks and irregular hummocky earthflows along the downslope sides of the scarps. The relatively straight and abrupt break in slope along which the sliding occurred on Sitkalidak Island strongly suggests the pos— sibility that the topographic fea- ture is controlled by a preexisting fault and that the line of slumps may have been related to renewed displacement along this inferred fault during the earthquake. For- tunately, it was feasible to check this possibility by measuring ver- tical displacement of the shore relative to sea level both to the northeast and southwest of the line of landslides. No changes in- dicative of vertical fault displace- ALASKA EARTHQUAKE, MARCH 27, 1964 ment during the earthquake were detected within the limits of pre- cision of the method (estimated to be ill/2 ft in this area). On the contrary, the measurements showed a slight northwest tilting of the island which is opposite in sense to the down-to—the-south— east movement along the landslide scarps. CAUSES Elastic ground vibrations dur- ing the 1964 earthquake trig- gered most of the fresh land- slides (and avalanches), or so loosened large masses of rock that they slid shortly thereafter. A contributory factor may have been ground accelerations related to the observed permanent verti- cal displacements, together with possible horizontal displacements, that occurred during the earth- quake. The eifect of transitory horizontal accelerations is to in- crease temporarily the shear stress on slope-forming mate- rials ; this increase may result in failure (Varnes, 1958, p. 43). In some instances slope failure may have occurred because of loss of shear strength through partial liquefaction of saturated granular materials below the layer of sea— sonal frost. Although the earthquake was clearly the trigger, it was not the ultimate cause of the slides. In a topographically rugged and geo- logically complex area such as Kodiak and the nearby islands, many rock masses are in a state of unstable equilibrium as a re- sult of local topography, lithol- ogy, structure, weathering, and water saturation. Even without earth shocks to act as a trigger, numerous landslides continually occur. Of particular interest is the concentration of earthquake- induced slides along the southeast coast of Kodiak Island and the nearby offshore islands (fig. 11), in contrast to their relative scarc- ity elsewhere. This distribution appears to be controlled primar- ily by local lithology and struc- ture but may also be related to proximity to the earthquake focal region. Steepness of slope prob- ably is not an important factor in controlling the overall distribu- tion, inasmuch as maximum slope angles in the area of few slides are, in general, equal to, or great- er than, those in the zone of many landslides. The strong coincidence of max- imum landslide frequency and outcrop areas of bedded Tertiary rock is shown in figure 11. It is also noteworthy that the land- slides in Tertiary rock are ex- ceedingly diverse and complex in character- and include all the types that were described above, whereas those in pre—Tertiary outcrop areas are almost exclu- sively rockfalls. ‘The predominance of landslid- ing in areas underlain by Terti— ary rock seems to be largely con- trolled by the physical properties of the rocks in at least five ways: (1) The Tertiary sequence locally contains inherently weak mate- rials such as clay, siltstone, and tuff along which sliding may oc- cur. ((2) The rocks, as a rule, are complexly folded and are broken by numerous discontinui- ties such as faults, joints, and bedding planes, which reduce shear strength. (3) The relative ease of weathering has resulted in local thick accumulations of landslide-prone soil and colluv- ium. (4) More rapid erosion has produced numerous steep-sided stream valleys and wave-cut cliffs, and attendant slope insta- bility has resulted from removal of lateral support. (5) Intensity of the tremors during the earth- GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D23 quake (fig. 5) suggests that the horizontal ground accelerations and amplitudes in some areas un- derlain by Tertiary rock were greater than in areas underlain by pre-Tertiary rock. Such dif— ferences in intensity would re- sult in proportionately larger transient increases in shear stress and attendant reductions in shear strength of slope-forming mate- rials in the Tertiary outcrop belt during the earthquake. In summary, the distribution of landsliding shown in figure 11 appears to be closely related to local geologic conditions, the slides being most abundant in areas underlain by the least competent Tertiary rocks, and relatively few and widely scattered throughout the remainder of the region. The five factors enumerated above apply especially to units of mid- dle and late Tertiary age. Rocks of early Tertiary age, as a rule, are well indurated, can be distin- guished from pre—‘Tertiary bed- ded rocks only with difficulty, and probably have physical properties and susceptibility to sliding that are intermediate between those of the late Tertiary and pre-Terti- ary sequences. Although it appears that land- slide frequency is largely con- trolled by the underlying rock units, the possibility must also be considered that proximity of the southeastern coast to known faults (fig. 3) or to the focal re- gion (as defined by aftershock distribution shown on figs. 1 and 4) may have resulted in a higher intensity of ground motion and attendant susceptibility to slid- ing. The limited available data, however, do not indicate that either proximity to faults or to the focal region had a discernible influence on landslide distribu- tion. Geologic study of the defor- mation in the area of highest landslide frequency failed to re- veal active faults having signifi- cant surface displacement along which landslides might be con- centrated. Furthermore, if the landsliding were controlled main- ly by proximity to the focal re- gion rather than by lithology and structure, there should be a pro- gressive decrease in landslide fre- quency northwestward, away from the focal region, instead of the abrupt change that occurs be- tween the belt of Tertiary rocks and the remainder of the area studied. EFFECTS ON GROUND AND SURFACE WATER Earthquake-induced fluctua- tions in level and (or) flow of some wells, lakes, and streams occurred at numerous localities throughout Kodiak and the adja- cent islands. Temporary and per- manent fluctuations of water level and flow that were recorded after the earthquake throughout much of Alaska, Canada, and con— terminous United States, Hawaii, and Puerto Rico (Waller and others, 1965) are to be detailed in a separate volume of this series of reports that deals with earth- quake-induced changes in the hy- drologic regime and, consequent- ly, will be described only briefly in the following paragraphs. OBSERVATIONS Reported effects of the earth- quake on the few water-supply wells on the islands varied mark— edly. Some wells went dry or de- creased yield and a few had in- creased yield, but most reportedly had no perceptible change. Many wells were muddied, and some along the coast became saline. Be- cause none of the wells on the islands were equipped with wa- ter-level recorders, all the avail- able data on earthquake—induced changes are, of necessity, based on observations by local residents reported in questionnaires and personal interviews. The only wells in the area known to have gone completely dry during the earthquake were shallow hand-dug wells on a small island near the Port Bailey can- nery on the south side of Kup- reanof Strait (fig. 2). A cannery worker, Mr. Rudy Lorenson, re- ported that after the earthquake the wells were dug 2 feet deeper without finding water. The wells were still dry 4 months later. Ranch wells at Narrow Cape (fig. 2) became muddy immediately after the earthquake. Although the yield of some wells reportedly increased, it is not knOwn wheth- er this increase was accompanied by a permanent change of level. Wells along the shore at Afognak Village were also muddied after the earthquake. These wells, and the school well at Old Harbor, became salty and unusable. There was no change in quality or level of water at the Coast Guard sta- tion near Cape Chiniak on Kodi- ak Island, or at the Federal Avia- tion Agency radio station at the north end of Shuyak Island. Observers throughout the area report that ice on Virtually all lakes and reservoirs at low alti- tudes was cracked during the D24 earthquake—particularly around the margins. At higher altitudes pilots noted cracking of ice on some large lakes, but many of the smaller lakes showed no percept- ible changes. A small cabin along the shore of a pond near Cape Chiniak was destroyed when seiching drove the ice cake over the bank and into the struc— ture. According to eyewitnesses at Afognak village, ice on the lakes began cracking as soon as the earthquake was felt. Marginal cracking of lake ice probably can be attributed to the inertial effect of the water and ice cakes in re- sisting horizontal ground oscilla- tions during the earthquake. In other parts of south-central Alas— ka it was noted that pressure ridges built up in the ice and snow along the shorelines in re- sponse to repeated pounding of the ground against the ice cakes. Although similar pressure ridges undoubtedly formed along lake- shores in the area, they were not specifically noted by local resi- dents. In general, lake levels were either unaffected by the earth— quake or were lowered. R. M. Waller has pointed out that er- roneous reports of lowered lake levels may have been made in places where earthquake-induced seiching stranded cracked ice along the shore, or where the tremors have caused collapse of peripheral ice and snow bridges that commonly form by normal lowering of lake levels during the winter. Two shallow lakes between 3 and 3% feet deep on Shuyak Island reportedly went dry during the earthquake. One filled up soon afterward but the other remained dry. Numerous shallow lakes on Kodiak Island that had gone dry during the earthquake were still dry or had lower water levels 4 months ALASKA EARTHQUAKE, MARCH 27, 1964 afterward. A cirque lake 2,000 feet long by 1,000 feet wide, at an altitude of 500 feet 5 miles northwest of Narrow Cape, re— portedly was lowered 20 feet dur- ing the earthquake, although nearby lakes showed no percept- ible change of level. A few local changes in stream— flow—involving both decreases and increases in volume—were reported in the area after the earthquake. Abnormally reduced water flow at two important sal- mon-rearing streams was noted by Dexter Lall of the Alaska De— partment of Fish and Game. A particularly interesting se- quence of changes was noted by Mr. Rudy Lorensen in a small stream that flows across Meso- zoic bedrock near the Port Bailey cannery on n o r t h e r n Kodiak Island. According to Mr. Loren- son, the creek flow was reduced to about half its normal volume after the earthquake. The re— maining flow was cut off entirely during the jarring aftershock of April 14th, the epicenter of which was located within 10 miles of the cannery (fig. 4). The stream re— mained dry for two months until another sharp aftershock in June started the normal volume of flow once again. The flow of streams at Afog- nak village and springs at a ranch on the north shore of Ugak ‘Bay on Kodiak Island reportedly in— creased for a period of one week after the earthquake. Four larger streams on which US. Geological Survey water-level recording gages were installed showed no significant flow changes following the earthquake. CAUSES OF FLUCTUATIONS Lack of detailed data on the pre— and postearthquake flow records of wells, streams, and springs for which earthquake-in- duced changes are reported pm- cludes analysis of the origin of these changes. ‘Elsewhere in the part of Alas- ka strongly affected by the earth- quake, where adequate data are available, the widespread lowered well and lake levels and reduced streamflow in granular materials has been attributed by H. E. Thomas of the US. Geological Survey to temporary increases in the porosity of saturated materi- als (in Grantz and others, 1964, p. 10). This mechanism could be the reason for lowered well levels in the islands, and, in conjunction with increased opportunity for water seepage from lakes and streams into earthquake-cracked frozen banks, and through crack- ed ice dams at lake outlets, could also explain some of the lowered lake and stream levels. In many places, however, low— ered water levels in wells, streams, and lakes appear to be permanent. The reported se- quence of abrupt changes in dis- charge of the small stream near Port Bailey suggests that, in areas underlain by bedrock, frac— ture porosity may be abruptly in- creased or decreased by slight shock-induced adjustments of bedrock blocks. Similar changes may have resulted from tectonic deformation within the area, es— pecially if significant horizontal extension accompanied the verti- cal subsidence. Retriangulation within the zone of subsidence near Anchorage has revealed hor- izontal extension roughly normal to the trend of the zone amount- ing to as much as 8x 10—5 or roughly one-half foot per mile (US. Coast and Geodetic Survey, 1965, p. 17). If comparable ex- tension occurred on Kodiak and the nearby islands, where sub- sidence is equal to or greater than that of the Anchorage area, near- GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D25 surface fracture porosity may hav e increased sufficiently to cause some of the reported changes in water regimen, and particularly in areas of fracture porosity. The permanently increased sa- linity of a few wells in highly permeable beach and alluvial de- posits at Afognak and the school well at Old Harbor probably was the result of encroachment of sea water into aquifers after tectonic subsidence. Temporarily increas- ed salinity of most of the shallow dug wells, however, resulted from flooding of watershed areas by the seismic sea waves. VERTICAL TECTONIC DEFORMA‘TION The March 27 earthquake was accompanied by tectonic defor- mation that resulted in vertical crustal movements, both upward and downward throughout the report area. Anomalous tide lev- els soon after the earthquake made coastal residents aware of the vertical displacement of the land. The isobase contours of figure 16 (next page) show the amount and direction of vertical displace- ment that accompanied the earth- quake. Directions and relative amounts of change were deter- mined at four localities from coupled pre- and postearthquake tide-gage readings, at about 25 localities from estimates made by local residents, and at 95 locali- ties from measurements of the displacement of the upper growth limit of sessile intertidal organ— isms along the seashore. Data- point locations are shown on figure 16. Determinations of vertical dis— placement at tide gages is prob- ably accurate to within a few tenths of a foot. Most of the es- timates made by fishermen, mar- iners, and other coastal residents who have had long experience with the local tides are probably correct to within a foot. The tech- nique used for estimating vertical displacements from the height of the upper growth limit of marine organisms above tidal datum planes has been outlined else- where (Plafker, 1965, p. 1675- 1679.) It will also be described in detail in a forthcoming volume on tectonic deformation, to be pre— pared for this series of publica- tions on the earthquake. Such measurements were carried out in 1964 and 1965 on a reconnais- sance basis in the area. Those based on difference in pre- and postearthquake altitudes of up- per growth limits of barnacles and algae are thought to be gen- erally accurate to about 1 foot. Most of the measurements, how- ever, are based only on the post- earthquake altitude of these or— ganisms above sea level; they contain inherent errors due to (1) deviation of sea level from predicted heights and (2) local variations in the growth limits of the organisms resulting from lo- cal factors of exposure, rock type, and water characteristics. They are estimated to be accurate to within 11/2 feet in sheltered bays on the eastern side of the islands. Along segments of the coast ex- posed to heavy surf or swells, and especially in areas of high and er— ratic tides such as the Shelikof Strait side of the islands (see table 2), the measurements may be in error by as much as 21/2 feet. In these areas, more reliance was placed on estimates made by coastal residents and on the measured difference in upper growth limits of pre- and post— earthquake organisms than on those measurements that require use of tide level as a datum. DISTRIBUTION Almost all of Kodiak Island, and all of Afognak, Shuyak, and Marmot Islands have undergone tectonic subsidence (fig. 16). The zone of subsidence consists of a broad trough that plunges gently northeastward along an axis that approximately coincides with the axis of the Kenai Mountains be- tween Alitak Bay and the city of Kodiak. The limbs of the trough are strongly asymmetric, the steepest limb beingsoutheast of the axis. Subsidence in excess of 6 feet has occurred in the vicinity of northern Kodiak Island, on Afognak Island, and on Shuyak Island. Maximum subsidence of 61/2—7 feet was measured on eastern Afognak Island and on Marmot Island. Uplift has occurred in a narrow zone that includes Narrow Cape on Kodiak Island, the southeast- ern half of Sitkalidak Island, all or part of Sitkinak Island, and some small islands off the south- eastern coast of Kodiak Island. Maximum uplift of 21/2 feet was measured on Sitkalidak Island. Mr. Joe Beaty (see p. D15)rep0rts that at the tip of Narrow Cape on Kodiak Island a surf-cut platform about 200 feet wide is now ex- posed at stages of tide that com- pletely inundated the platform before the earthquake. Uplift of at least 2 feet, and possibly in ex- cess of 3 feet, is suggested near this locality by postearthquake D26 ALASKA EARTHQUAKE, MARCH 27, 1964 N , 5_ a) Moimnéfi i Kodiak Fisheries ' ~‘ , “Cannery EXPLANATION _4_—..... Isobase contour Showing amount of uplift in feet. Dashed where approximate; dotted where inferred A Reoccupied US. Coast and Geodetic Survey tide gage Estimated precision : 1/; foot 0 Estimate by local residents or measurement of ,, , _ .. . difference between preearthquake and post ” earthquake growth limits of intertidal or ter- " '. restrial organisms ‘ Estimated precision :Ifoot arm: K! 5v anew: g: O 0 ...... +2 Measurement of postearthquake position of inter- tidal organisms relative to sea level Estimated precision i195 feet along east coast; :2% feet along west coast V ‘ ' ‘N‘sg‘ ;1 _ [:5 .' Little or no change in level'fi.‘ on Tugidak Island “no 16,—Areas affected by tectonic deformation on Kodiak and nearby islands. GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D27 incision of streams and the con— sequent partial draining of two small beach-barred lakes. How- ever, the outlet streams may have become incised through scouring action of the seismic sea waves that entered the lakes after the earthquake. At Sitkinak Island, another rancher, Mr. Hal Nelson, noted 1—2 feet of shoal- ing at the entrance to a lagoon immediately after the earth- quake—probably the result of a slight crustal uplift. Evidence of as much as 2 feet of uplift in the coastal area and small offshore islands between Narrow Cape, Sitkalidak Island, and Sitkinak Island suggests that the zone of uplift is probably continuous be- neath intervening water-covered areas. As far as could be determined during the brief reconnaissance of the area, changes of land level associated with the earthquake resulted from regional warping rather than faulting. A diligent search was made for surface faults, particularly in the vicinity of the zero isobase between the areas of uplift and subsidence, a zone roughly paralleling numer- ous known faults (fig. 3). No evi- dence of significant surface faulting was found, nor were there any anomalous abrupt changes in amounts of uplift or subsidence along the coast sug- gestive of measurable displace- ment along concealed faults. Past movements on the major faults in this zone have been mainly dip slip with the north- west blocks upthrown (fig. 3). Consequently, if any of these faults were reactivated during the earthquake only a reverse displacement could produce the observed pattern of deformation along the southeast coast of the islands. EFFECTS ON SHORELINES Tectonic subsidence, augment- ed locally by surficial subsidence of unconsolidated deposits, has had a profound effect on shoreline mOrphology, intertidal marine or- ganisms, and terrestrial vegeta- tion. Virtually all shorelines that subsided more than 3 feet show clear physical evidence of the change, and the changes were 10- cally noticeable where the subsi- dence was as little as 1 foot. The effects were most pro— nounced in areas of lowest mean annual tidal range (table 2) , inas- much as both the frequency and duration of shoreline immersion for a given amount of subsidence vary inversely with the tidal range. 'The conspicuous effect of subsidence is the fringe of dead brown terrestrial vegetation along the shore that has been killed by salt-water inundation at periods of high tides (pls. 1, 3). Beach berms and stream deltas in subsided areas are shifted landward and are building up to the new higher sea levels. The lower reaches of streams are now inundated by tides for distances of as much as 4,500 feet inland (table 3). The striking effect of subsidence in the estuarine part of Olds River at Kalsin Bay is il- lustrated by plate 3. Many former beach-barred lakes at stream mouths or bay heads have become tidal lagoons. Wave action at the higher sea levels is rapidly erod- ing shorelines composed of poorly consolidated deposits that are now brought within reach of the tides (pl. 4). An irreparable loss resulting from accelerated ero- sion of such deposits is the de- struction of many coastal archae- ological sites situated on them. Former reefs and low-lying islands along the coast are now submerged, and some tombolo— tied points or capes have become islands. Effects of the small and local- ized uplift along the shore are minor compared to those due to subsidence. At Narrow Cape a new surf-cut rock terrace 200 feet wide reportedly is exposed at zero tide, and nearby beach- barred lakes have been partially drained by incision of the outlet TABLE 3.— Tidal inundation resulting from tectonic and surficial subsidence Length of Approximate stream Stream Location tectonic newly inun~ (fig- 2) subsidence dated by (feet) tides (feet)1 Afognak Island Danger Bay Creeks ------ 14 miles northeast of 5% 865 Afognak. Marka Creek ---------------- 5 miles northeast of 4—5 2,600 Afognak. Back Bay Greeks ---------- 5 miles north of Afognak ...... 31/2—5 605 Afognak River -------------- 41/5 miles north-northeast 31/2—5 825 of Afognak. Kodiak Island Kiliuda Creeks .............. Kiliuda Bay ........................... 4—5 1,900 Eagle Harbor Creek.... Ugak Bay ....................... 4 850 Saltery Cove Creek .............................. do ....................... 5 570 Olds River ...................... Kalsin Bay ............................ 41/2 4,500 1 Alaska Department of Fish and Game (1965, p. 5). D28 streams. Elsewhere a few new reefs are exposed at stages of tide when they formerly were under water. Lowered tide levels in la- goons in the area with attendant reduced volume and velocity of diurnal flow through the outlets may eventually result in the bar- ring of the lagoon outlets and subsequent conversion of the la- goons to fresh-water lakes. REGIONAL SETTING OF THE CHANGES The 4,900-square-mile land area of Kodiak and the nearby islands within which vertical dis- placements were measured is but a part of a regional zone of tec— tonic deformation associated with the earthquake. An area of at least 70,000 square miles, and possibly 110,000 square miles or more of south-central Alaska, is involved (Plafker, 1965). The zone, which is more than 500 miles long and as much as 200 miles wide, is roughly parallel to the Gulf of Alaska coast from the Kodiak group of islands northeastward to Prince William Sound and thence eastward to about long 143° W. It consists of a major seaward zone of uplift bordered on the northwest and north by a major zone of subsi- dence (fig. 17). These two zones are separated by a line of zero land-level change that trends northeastward from Sitkinak Is— land along the seaward sides of Sitkalidak and Kodiak Islands (fig. 16). From Kodiak Island, the line trends northeastward to intersect the mainland between Seward and Prince William Sound. It then curves eastward ALASKA EARTHQUAKE, MARCH 27, 1964 through the western part of Prince William Sound to the vi- cinity of Valdez and crosses the Copper River valley about 50 miles above the mouth. In addition to most of the K0- diak group of islands, the zone of subsidence includes most of Cook Inlet, the Kenai Mountains, and the Chugach Mountains. The axis of maximum subsidence within this zone trends roughly northeastward along the crest of the Kodiak and Kenai Mountains and then bends eastward in the Chugach Mountains. Maximum recorded downwarping along this axis is 71/2 feet. In the northern part of the de- formed area, uplift in excess of 6 feet occurred over a wide area including part of Prince William Sound, the mainland east of the sound, and offshore islands on the continental shelf as far southwest as Middleton Island. Along sur— face faults on Montague Island, the uplift locally exceeded 30 feet (Plafker, 1965, fig. 7). South— west of Montague Island the sea bottom may have been uplifted more than 50 feet (Malloy, 1964, p. 1248), where pre- and post- earthquake bottom soundings show seaward continuations of new fault scarps along preexist- ing fault lines. Large-scale uplift of the continental shelf and slope southwest of Montague Island is inferred from the trend of iso— base contours in the northeastern part of the deformed area, and from the presence of the fringe of uplift along the outer coast of the Kodiak Island area (fig. 16). The minimum extent of this in— ferred offshore zone of uplift is thought to be roughly outlined by the earthquake focal region, or belt of major aftershocks, shown on figure 1. The major zones of vertical tectonic displacement and the belt of aftershocks, which lie within and parallel to the Aleutian Trench and the Aleutian volcanic arc, are inferred to have resulted from displacement along a fault or zone of faulting that dips at a moderate angle northward from the Aleutian Trench beneath the continent (Grantz and others, 1964, p. 2). Neither the orienta- tion nor the sense of movement on the primary fault (or faults), along which the earthquake pre- sumably occurred, is certain, in— asmuch as the primary fault or zone of faulting is not known to intersect the surface on land. The two known surface breaks, which occurred on preexisting faults on Montague Island along the axis of maximum uplift, are clearly within and subsidiary to the zone of regional uplift. The faulting on Montague Island is known to extend offshore south- westward at least 15 miles, and comparable fault displacements probably occurred along this trend or elsewhere on the sea floor of the continental shelf with- in the focal region of the earth- quake. Indeed, the direction of travel and reported arrival times of the initial wave crest in the area (p. D30—D39) suggest that the axis of maximum uplift (fig. 17), along which faulting occur- red on Montague Island, may ex- tend southwestward perhaps to the latitude of Sitkalidak Island (Plafker, 1965, p. 6; Plafker and Mayo, 1965, fig. 12). GEOLOGICAL SURVEY PROFESSIONAL PAPER 543-D PLATE 3 SUBSIDENCE AT MIDDLE BAY Extensive inundation at the head of Middle Bay resulting from 5 feet of tectonic subsidence and an unknown but substantial amount of surficial subsidence. Arrows indicate the approximate location of remnants of the Kodiak-Chiniak highway. Dashed line shows the approximate preearthquake mean high-water shoreline which has been shifted about half a, mile landward. GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—D PLATE 4 UNDERMINED TREES 0N BARRIER BEACH IN IZHUT BAY, AFOGNAK ISLAND Accelerated wave erosion in an area of about 5 feet of tectonic subsidence caused this damage. Foliage of trees along shore has turned brown from salt-water immersion of the roots. 156° GEOLOGIC EFFECTS ON THE 152° 150° 143° KODIAK ISLAND AREA ' 1329 1u6° 142° 15u° , I 62°* 60°' MONTAGUE ISLAND 58° I I 7‘ . I/ ‘- . I 6 ‘l a $3 I I $§> f . / I I / 1: I | | ,’ do § 1 , V. l I [I’flfl $$ 0 . , x I ' “ SITKALIDAK I I I" | 11;, ISLANDI I I I I «I» ‘flltr I I I '-/ I I | I I l I’ I TRI ITYIISLAN NDS I I,»-~.. ,2: I \l I I ___ .1” I K NI I x I 5y- ________ \fl ~~~\‘ O 50 100 1?0 200 KILOMETERS IF I I I | I I 0 I 50 100 I 150 MILES I J \ MIDDLETON ISLAND V EXPLANATION (CID Area of uplift or probable uplift Dashed where inferred #— E; Area of subsidence /7“~. Axis of uplift Dashed where inferred I? Axis of subsidence 96 Volcano Outer edge of continental shelf (600 feet) I I 17. — Setting of the observed and inferred vertical displacements of the Kodiak group of islands with respect to regional tectonic deformation in south- central Alaska. D30 By far the most catastrophic effect of the 1964 earthquake on Kodiak and the adjacent islands was the train of destructive seis- mic sea waves that first struck the coast about 38 minutes after the start of the earth tremors. The waves, repeatedly inundating low-lying coastal areas, caused 18 fatalities and extensive property damage along the southeast coast of the islands. This train of seis- mic sea waves resulted in the worst natural disaster to befall the islands in historic times. There is no authenticated record of previous destructive seismic sea waves along this same seg- ment of the Alaskan coast. According to a widely quoted account in the historical novel “Lord of Alaska” (Chevigny, 1942, p. 63), the original Russian settlement on Kodiak Island at the site of a native village in Three Saints Bay was wiped out by a “tidal wave” during a vio- lent earthquake in 1792 after which it was moved to the pres- ent location of Kodiak. The his- torian, Bancroft (1886, p. 324), states, however, that the decision to move the settlement was based solely on sound tactical and eco- nomic reasons. The original site continued to be inhabited by na- tives at least until 1805, at which time the Russian naval captain Lisiansky called at ‘Three Saints Bay in the ship N em (Bancroft, 1886, p. 434-435). CHRONOLOGY For a period of about half an hour after the earthquake, most residents were busily assessing the relatively light earthquake- induced damage and discussing ALASKA EARTHQUAKE, MARCH 27, 1964 SEISMIC SEA WAVES the terrifying shaking they had so recently endured. Few people recall having noticed any unusual movements of the sea in that in- terval of time. The tide was ebb- ing and nearly at low stage (table 2), the wind was calm, and there was ample daylight for reliable visual observations for about 2 hours after the earthquake, or until about 7:30 pm. The only reported noteworthy fluctuation of water level prior to arrival of the first seismic sea wave was a gradual rise and sub- sequent withdrawal of the sea. This fluctuation was noted by a fisherman, Mr. Jerry Tilley, who was aboard the 75-foot shrimp- fishing boat Fortress tied up at the city dock in Kodiak. Mr. Tilley states that the water level rose calmly from about 0 to 13 feet within 10 minutes after the earthquake was initially felt. Wa- ter level then receded gradually for an estimated 25 minutes to about —10 feet, after which the first wave moved in from the south as a large swell. This first, calm rise of water level has not been reported elsewhere, and it caused no damage whatever. The origin of this reported sea-level rise at Kodiak is uncertain. It could have been a local seiche generated largely by the tectonic tilting and possible horizontal dis— placement that occurred during the earthquake. However, the fact that it was not noticed at the nearby Naval station, which un— derwent approximately the same amount of subsidence, argues against this possibility. In any event, the limited data available on this initial sea—level rise sug- gest that it was a local phenome— non, rather than part of the train of long-period seismic sea waves that followed. At 6 :14 p.m., 38 minutes after the start of the earthquake, Fleet Weather Central, Kodiak Naval Station, received an ominous re- port of a “30—foot tsunami” at the Cape Chiniak station of the US. Coast Guard 15 miles to the southeast (fig. 18). An immedi- ate “tidal-wave” warning, broad- cast over the Armed Forces Ra- dio, resulted in the timely evacua- tion to high ground of most base and city personnel in the Kodiak area. Many individuals began im- mediately trying to fly amphibious aircraft from the harbor area, to pilot fishing boats to the relative safety of deep water, or to re- move vehicles and other valuable belongings to high ground, but with varying degrees of success. At about the time that the wave was observed at Cape Chin- iak, Mr. Joe Beaty (see p. D15) was on the roof of his ranch home at Narrow Cape repairing a ra- dio antenna broken during the earthquake. Looking seaward, he was astonished to see the sea withdraw a long distance from the shore; the withdrawal was immediately followed by a “wall of water” that moved in from the east and broke off shore. A wave then surged over low-lying parts of the ranch as much as three-quarters of a mile inland from the beach; it deposited vast quantities of driftwood but it stopped just short of the house and outbuildings. Seven head of cattle were drowned, contrary to one journalist’s account which claimed that the animals some- how sensed the oncoming wave in time to move to the safety of high ground. GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D31 /Navigation buoy in Shuyak X Strait shifted westward x iv 3:, vi . :E- ‘r ""<~'r§,\u‘§ ,~ ' 108 ‘ ' .‘Afognak Uzinki x * .‘5 ,_ U K}, 2 K" I l 2;» / 1i5+ A ”(‘33, <<\ 0‘ 4'2 O A \W ‘7‘ /<<\/< K i s» <0 r. , a , _ «$0 , ,éljearwale'rBay 66 65‘ , '1455, <<\ (3. A“ * 'C' ' p. ,. ' , «Old Havber x << 0* i j, ’ ’ ' X‘ 9‘12,’ Port Hobronivv ,2: 0c, \- . _, ’ / 1457/sz if; i /<\ 3% /No wave damage i w, 18. — Localities on Kodiak and nearby islands at which waves caused major property damage (black circles), shorelines showing physical evidence of wave damage (x), Shorelines reportedly inundated by runup (open circles), and U.S. Geo- logical Survey streamflow gages on which waves were recorded (triangles). Numbers indicate approximate highest runup in feet above existing tide level; plus sign following numeral indicates that tide stage is unknown and maximum runup shown is minimum value assuming coincidence of wave with high tide. Lined pattern offshore indicates inferred posi- tion of the northwest and southwest margins of the wave-:rest source. D32 ALASKA EARTHQUAKE, MARCH 27, 1964 Qfi\«w Q n w 19. — Seaplane ramp at Kodiak Naval Station along shore of Womens Bay showing approximate high-water level reached by first wave. Note calm water surface. Aircraft parked in front of hangar at left center were later moved to high ground and saved. Photograph by U.S. Navy. Shortly after the initial wave struck the outer coast from Cape Chiniak to Narrow Cape, inhabi- tants elsewhere along the south- east coast between Kaguyak and Afognak noticed rapid rises of the sea, accompanied in some places by dull rumbling sounds. By 6:35 p.m., 63 minutes after the start of the earthquake, the initial wave was cresting at the city of Kodiak, nearby Kodiak Naval Station, and other points along the shore of Chiniak Bay, and was inundating low-lying coastal areas. Everywhere except at Cape Chiniak and Narrow Cape, the first waves were de- scribed as a gentle gradual flood with swirls resembling riverflow that rose at estimated rates of as much as 3 feet per second. Figure 19 shows part of the airport at Kodiak Naval Station as the first wave crested. At Kodiak and Old Harbor, gradual withdrawals of about 10 feet reportedly preceded the initial wave, but no with- drawal in Womens Bay was ob- served at the Kodiak Naval Station. The initial wave was followed everywhere by a turbulent swirl- ing retreat of water from the shore, during which segments of the sea floor were bared far below the level of the extreme low tides. Much of the fishing fleet at Ko- diak was left grounded, but some boats were swept offshore and caught in whirlpool-like eddies or high-velocity currents that ac- companied the withdrawal. The water withdrew to about 10 feet below the postearthquake mean lower low water at the Kodiak Electric Association powerhouse in Kodiak. At the Kodiak Naval Station, the water withdrew to about 111/; feet below the post— earthquake mean lower low water. Approximately at 7:40 p.m., as darkness was falling and the tide beginning to flood, a second wave that was the largest and most destructive of the train in the Chiniak Bay area struck the coast. It swept away docks, houses, and vehicles along the shore and drove flotsam and fish— ing vessels inland as battering rams, causing much destruction of property previously undam- aged by the high water. The withdrawal that followed the second wave was extremely turbulent and rapid. Swirling currents with velocities variously estimated at 20—25 knots flowed through the straits near Kodiak. The water was choked with de- bris, and vessels that were un— able to steer against the swirling erratic currents were carried on its flood. The third wave, which struck the southeast coast of the area between 8 :30 and 9 :30 p.m., was slightly lower than the second wave in Chiniak Bay, but because GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D33 it came in on a rising tide it reached almost the same altitude along the shore as the previous wave, adding substantially to the damage. It was the highest and most destructive wave witnessed at the villages of Afognak, Old Harbor, and Uzinki; numerous homes and other structures were washed to sea, and fishing vessels were sunk or beached. At the widely spaced villages of Afog- nak and Old Harbor, the arrival time of the wave crest was re- corded on clocks that were stop- ped by the inundation almost si- multaneously—at 9:27 pm. and 9:28 p.m., respectively. By this time Virtually everyone along the southeast coast except those aboard boats and personnel at Fleet Weather Central had evac- uated to higher ground, and it was too dark in most places to permit reliable observations on the nature and sequence of waves. The instrumental records and the record kept at the Kodiak Naval Station show that waves continued to inundate the shore- line with progressively decreas- ing amplitude until early the next morning. However, the highest runup in some localities, includ- ing Womens Bay near Kodiak Harbor, coincided with high tide shortly after midnight (table 2). The greatest observed with— drawal of water occurred at the Kodiak Naval Station a few min- utes after midnight. At that time, water level fell some 24 feet from a plus 8-foot tide to 16 feet below mean lower low water. Along the Shelikof Strait side of the group of islands, no wave action or water motions were noted, aside from minor local tur- bulence, until about 90 minutes after the earthquake. Water fluc- tuations there consisted of rapid low-amplitude tide changes ac— companied by swift and erratic currents. The shoreline was inun- dated a few feet above normal preearthquake highest tide levels only by the waves that occurred between 11 :30 pm. and 1:00 a.m., within an hour of high tide (table 2). The last recorded an- omalous tide rise on the Shelikof Strait side of the island that could be attributed to wave action was at about 5:00 am. on the morning after the earthquake. CREST HEIGHTS Locations of coastal areas on Kodiak and nearby islands that were affected by the seismic sea waves are shown on figure 18. Also shown are maximum runup heights recorded at shore stations and the inferred source area of the waves. The sole operative tide gage on Kodiak and the nearby islands was at the Kodiak Naval Station in Womens Bay. It was put out of action by the earthquake trem- ors and was subsequently inun- dated by the waves; the mari- gram for March was lost. Obser- vations of water-level changes at the Naval Station were made by Fleet Weather Central personnel throughout the night of March 27 and the early hours of March 28. It is the most complete record available of the seismic sea waves in the area or elsewhere along the Gulf of Alaska coast in the region strongly affected by the earth- quake. Unique instrumental records of arrival times and maximum run- up heights of the higher waves were obtained from three US. Geological Survey streamflow gages at localities shown on fig— ure 18. The three gages were situ- ated at sites near stream mouths that subsided sufficiently during the earthquake to bring them within reach of the highest seis- mic sea waves; the two gages on the Shelikof Strait side of the island are low enough to record the higher astronomical tides since the earthquake. Arrival times, crest heights, and runup al— titudes of the waves at Womens Bay and at the three streamfiow gages are given in table 4 (next page). With the exception of the four localities enumerated above, all data on wave-arrival times were obtained through conversations or written communications with eyewitnesses. Most are little more than rough estimates of elapsed time beween start of the tremors and the wave-crest ar- rivals. Few observers had the op- portunity or presence of mind un- der the prevailing conditions to record the time at which the waves struck, and in any given locality opinions varied consider- ably. The maximum heights to Which the waves rose along the shore at localities other than Womens Bay and the sites of the streamflow gages could be readily ascertain- ed from the strand lines of sea- weed and driftwood, waterstains on buildings, and abraded bark or broken branches in trees and brush along the shore. Heights were measured by hand level and stadia rod above prevailing sea level and later reduced to mean lower low water. At most locali- ties measured heights were sub- stantially lower than those indi- cated by eyewitnesses or estimat- ed by some investigators who vis- ited the area after the earth- quake. Because much of the land area had been lowered or uplifted by the tectonic deformation that pre- sumably occurred during the earthquake (Plafker, 1965, p. 1680) and because the wave train struck the coast at stages of tide ranging from low to high, appro- D34 ALASKA EARTHQUAKE, MARCH 27, 1964 TABLE 4.——Heights (in feet) and arrival times of seismic sea—wave crests at Womens Bay, Myrtle Creek, Terror River, and U ganik River on Kodiak Island [Local (Alaska standard) time. MLLW, postearthquake mean lower low water] (KodiizljllxilziithLLion) Myrtle Creek2 Terror River3 Uganik River4 Wave Crest height Crest height Crest height Crest height Time Above Above Time Above Above Time Above Above Time Above Above “‘16 MLLW “de MLLW “‘19 MLLW “d" MLLW level level level level March 27 1 . 6:35 p.m. 10.8 10.6 6:45 p.m. 15.2 15.0 11:15 p.m. 8.0 18.2 11:00 p m 6.2 15.2 March 28 2 . 7 :40 p.m. 12.8 12.8 7:40 p.m. 18.6 18.8 1:15 a.m. 1.8 17.0 1 :15 a m 0.9 16.0 3 . 8:30— 8 :44 p.m. 8.6 9.7 8:30 p.m. 17.2 18.4 1250 a.m. 2.6 17.7 2 :32 a m 4.9 18.5 4 . 10:00 p.m. 9.4 13.7 9 :45 p.m. 14.3 18.4 3 :00 p.m. 8.6 21.1 ............................... 5 . 11:16— 11 :34 p.m. 11.6 18.8 11:15 p.m. 13.2 20.3 4:20 a.m. 9.2 18.2 ................................. March 28 6 .. 0:47 a.m. 3.2 11.6 2:40 a m 7.3 13.9 ................................. 7 .. 1:54 a.m. 6.4 13.8 ................................................................. 8 .. 2:20 a.m. 3.0 10.0 ................................................................... 9 . 2:58 a.m. 2.2 7.2 ................................................................................. 10 . 3:20 a.m. 1.6 5.6 ................................................................................. 1 Data from Fleet Weather Station, U.S. Navy, Kodiak. 2Data recorded on U.S. Geological Survey streamflow gage about 1 mile above preearthquake stream mouth. 3Data recorded on U.S. Geological Survey streamflow gage 0.7 mile above preearthquake stream mouth. 4 Data. priate corrections for vertical displacement and tide stage had to be made at each locality to obtain the actual runup heights shown in figure 18 and table 4. Where wave-crest arrival times were known, corrections were made by subtracting from the measured runup heights (1) the amount of subsidence, if any, at the shore station, and (2) the predicted tide height above lower low water at the time the wave crested. Where runup heights were measured from swash marks or other evidence left along the shore at unknown stages of tide, corrections could be made for vertical displacement but not for tide stage. In these localities, the height indicated on figure 18 is a minimum value (indicated by a plus sign following the numeral) that assumes coincidence of the highest wave with high tide on the morning of March 28 (table 2). Along the ocean side of the islands the highest waves appar- ently occurred between low and half tide, so the runup heights of the highest waves most probably were 4—8 feet higher than those shown on figure 18. RUNUP HEIGHTS In general, the waves were high and destructive only along the exposed ocean coast. They barely inundated shorelines above normal highest tide levels along recorded on U.S. Geological Survey streamflow gage 0.5 mile above preearthquake stream mouth. the Shelikof Strait side of the islands and the straits between the larger islands. The waves were too low to flood above the shoreline along the southwest shore of Kodiak Island or at Sit- kinak Island. As indicated by figure 18, max— imum runup heights of the waves around the islands ranged from a minimum of 5 feet or less to at least 311/2 feet. Large local varia- tions in runup heights occurred around the islands—apparently controlled by the highly irregular configuration of the coastline, dif— ferences in ofishore bottom topog- raphy, and length of wave travel path. ‘The highest recorded run- ups were along exposed beaches GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D35 and bluffs on southeastward-fac- ing shores of Kodiak and Sitkali- dak Islands, whereas runup heights in adjacent sheltered em- bayments and segments of coast protected by offshore reefs were substantially lower. The highest measured runup was at an uninhabited locality be- tween Cape Chiniak and Narrow Cape, where trees were damaged 311/2 feet above predicted astro- nomical high tide on the night of the earthquake; however, the highest wave probably coincided with the first large wave at low tide that was witnessed at Nar- row Cape and Cape Chiniak. If it did, runup above existing tide level would have been almost 40 feet. Runup of 241/; feet above the March 28 high-tide level was measured on an eastward-facing beach at Sitkalidak Island, and driftwood deposited on a terrace at the south end of the island is estimated to be at least 15 feet higher. Wave-deposited driftwood was also seen from the air at high altitudes on bluffs along the north coast of Marmot Island. In sheltered embayments along the ocean coast, runup heights ranged from 12 to 17 feet. The highest runup of 17+ feet occur- red at Kaguyak which is situated at the head of a large, deep bay that opens into the Gulf of Alas- ka. The rather low measured runup of 8 feet at Afognak vil- lage and at other localities on Afognak Island that face the open ocean are suggestive of rapid dissipation of wave energy with distance from the inferred offshore—source area, especially in areas of shoal waters and shallow reefs such as extend several miles offshore from Afognak village. On the Shelikof Strait side of the islands, maximum runup heights generally range from 5 to 6 feet. A lone exception is at the Terror River streamfiow gage site, where, as described in the following section, seiching caused anomalous water-level rises of as much as 10 feet above tide level at 3 :00 am. and 4:00 am. on the morning of March 28. Other than at the Terror River gage site, maximum recorded runups above tide level occurred during the first three waves. The highest runups above predicted postearthquake astronomical tides recorded at Kodiak Naval Station and at the Myrtle Creek streamflow gage (12.8 and 18.4 ft, respectively) occurred dur- ing the second wave. At Cape Chiniak and Narrow Cape the first wave (at least 311/3 and pos- sibly 40 ft) reportedly was the highest. At Afognak and Old Harbor the third wave (8 and 13 ft above tide level, respectively) was the highest. The maximum runup altitude in these areas, however, did not necessarily co- incide with the highest wave, in- asmuch as the maximum was the resultant of both wave height and the stage of postearthquake as— tronomical tide at any given 10- cality. As shown by table 2, the tides on which the seismic sea waves were superimposed ranged from slightly less than 0 to 16.0 feet. Consequently, the highest recorded runup at many locali— ties, including those stations in- dicated on table 4, coincided with wave arrivals close to midnight at about high tide, rather than with the larger early waves that arrived at lower tide stages. PERIODS The first three seismic sea waves as recorded at Womens Bay and the stream gage at Myrtle Creek had periods of 50—55 minutes (table 4). This period probably most nearly ap- proximates that of the waves at their source. Later irregularities in the interval between successive waves at these places and at the two stream gages on the Shelikof Strait side of the islands may re- flect superposition of local seiches upon the wave train. Such super- position may in part explain the inconsistencies in the reports by eyewitnesses at different locali- ties on the number and arrival times of wave crests. The anomalous increase in wave height recorded at the Ter- ror Bay streamflow gage (table 4) is suggestive of a longitudinal seiche excited in Terror Bay through approximate coincidence between the natural oscillation period of the bay and that of the seismic seawave train. That the anomalous high waves in the bay were purely of local origin is clearly demonstrated by the rec- ord of successive waves of declin- ing amplitude during the same period on the Uganik River gage and at inhabited areas nearby. The Terror River gage is at the head of a nearly rectangular fiord about 14 miles long, less than 1 mile wide, and with an average depth of roughly 270 feet. The natural period of oscillation for the fundamental mode of open- sided bays, such as Terror Bay, is given by the formula: Tn = _4l_ \/ yd where Tn is the natural period in seconds, l is the length of the bay in feet, g is the gravitational con- stant, and d is the water depth in feet (‘Ruttner, 1953, p. 43). The fundamental mode of oscillation is that in which the first node is situated at the mouth of the bay. By use of this formula and the dimensions and depth indicated above, the period for the funda- mental mode of water in the fiord is about 53 minutes, which coin- D36 ALASKA EARTHQUAKE, MARCH 27, 1964 20. -— Opening torn by seismic sea waves in the barrier beach of a large lake along the shore of Izhut Bay on Afognak Island. Channel scour and subsidence of at least 6 feet permit sea water to enter the basin at all stages of tide. Note partially submerged trees along inner margins of the beach spits, and small new rockslide along bluff in background. Photograph by Alaska Department of Fish and Game. GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D37 cides almost too closely with the period of 50—55 minutes indicated for the initial waves. Consequent- ly, the anomalously high wave recorded at the stream gage may have resulted from resonant am- plification of a longitudinal seiche that was excited by the train of seismic sea waves. EFFECTS ON SHOREILINES AND SEA BOTTOMS High-velocity currents associat- ed with the repeated ebb and flood of the waves resulted in ex- tensive damage through erosion of artificial fills and unconsolidat- ed deposits along the coast. The effect of the surges may be com— pared to erosion during extreme flood conditions of rivers. At sev- eral places, barrier beaches dam- ming coastal lakes were breached by the waves, and the lakes were thereby converted to lagoons, as illustrated in figure 20. Stream gravel was removed from the bed and banks of Olds River at the head of Kalsin Bay along a length of at least 400 feet of the estuary (Alaska Dept. Fish and Game, 1965, p. 6). Mud was eroded from tidal flats along the shore at Kodiak to a depth of 10 feet. Bottom changes off shore from Kodiak at a depth of 75 feet were reported by a scuba diver, Mr. Jerry Tilley, who noted that the channel had been scoured clean of mud and that thousands of clams were thereby exposed on the sea floor. The combination of swift cur- rents and the buoyant effect of large-amplitude waves moved several navigation and mooring buoys in the Chiniak Bay area, and dragged the navigation buoy in Shuyak Strait off station. Un- doubtedly the 20- to 25-knot cur- rents associated with the waves caused intensive scouring of sedi- ment from shorelines and narrow 21,—Shoreline damage between Narrow Cape and Cape Chiniak caused by driftwood-laden seismic sea wave. Spruce trees as much as 8 inches in diameter were broken off at about 80 feet altitude. Limbs were broken and trunks scarred to a measured altitude of 42 feet above lower low water. Crowns of downed trees point in direction of water movement. straits and deposition of the eroded material in deeper quieter water. Few such effects, however, have been documented along the shores of Kodiak and the nearby islands because of the general scarcity of preearthquake refer- ence marks in coastal areas and the complicating factor of wide— spread postearthquake subsi- dence. At uninhabited capes and em- bayments along the ocean coast, wave inundation was evidenced by newly deposited driftwood and seaweed above the level of high— est tides and by broken and scar- red shoreline vegetation in tim- bered areas. Figure 21 illustrates the effects on a stand of spruce trees near Narrow Cape of a driftwood—laden surge of water that ran up to an altitude of 42 feet above lower low water. ORIGIN The seismic sea waves clearly were generated off shore from the Kodiak group of islands with- in the Gulf of Alaska, as shown by the arrival times of the initial waves, the distribution of wave damage, and the orientation of damaged shorelines. The distance traveled by the waves can be de- termined where wave velocity along the propagation path and travel time are known. Because the wave lengths are long, seismic sea waves move as shallow water waves even in the deepest ocean. As such, their velocity is con- trolled by the depth and conforms closely to LaGrange’s equation (in Lamb, 1932, p. 257) : V = VW where g is the gravitational con- stant, and h is the water depth D38 along the travel path (as deter- mined from nautical charts). The travel time is taken as the elaps- ed time between arrival of the first wave crest at shore stations and the inferred time at which the waves were generated. Calculations of the distance travelled by the first waves to the few stations where reliable data are available on arrival times defines a source area that trends northeastward at a mini— mum distance of 28 miles off Cape Chiniak and 17 miles off- shore from Sitkalidak Island— assuming that the waves were generated at the time of the earthquake or shortly thereafter (fig. 18). A time delay between the beginning of the earthquake- and generation of the waves would modify the position of the inferred source by shifting it shoreward. Thus, if as much as 5 minutes elapsed between the be- ginning of the earthquake in Prince William Sound and the generation of the waves offshore from the Kodiak area, the infer- red source area would be 4—6 miles shoreward from the posi— tion shown on figure 18. The ab- sence of large waves at Sitkinak Island and in Alitak Bay suggests that the wave source did not ex- tend southwest of the general lat- itude of southern Sitkalidak Island. The anomalous lack of wave effects along the low-lying northeast shore of Sitkinak Island, which is ideally situated to receive heavy damage from northeasterly waves, further sug- gests that the waves were strong— ly directional—as might be ex- pected from a linear northeast- ' trending source area. Initial upward motion in the source area is suggested by (1) the distribution of measured land-level change along the ad— jacent coast and offshore islands, ALASKA EARTHQUAKE, MARCH 27, 1964 (2) the initial rise recorded on tide gages outside the immediate area affected by the earthquake (Van Dorn, 1964, p. 11), and (8) possibly by the unique at- mospheric waves (“seismic air waves”) recorded shortly after the earthquakes on microbara- graphs at both the University of California at Berkeley and the Scripps Institution of Oceanog- raphy at La Jolla (Van Dorn, 1964, p. 7; Bolt, 1964, p. 1096). The initial direction of water movement within the Kodiak islands area is less clear, how- ever, because there were no op- erative tide gages, and in many localities water movement during and after the earthquake was complicated by land-level changes along the coast, and perhaps also by local waves generated by the ground tremors or by subaerial and submarine landslides. With a few notable exceptions, however, most observers along the coast of the Kodiak group of islands and elsewhere along the coast of the Kenai Peninsula and Prince Wil— liam Sound area reported that the first strong surge of waves from the Gulf of Alaska was up- ward; this report tends to sup- port the inference of vertical dis- placement in the source area. The waves recorded on Kodiak and the nearby islands and those that struck the outer coast of the Kenai Peninsula (fig. 17; Plaf- ker and Mayo, 1965, p. 10, fig. 12) appear to have originated along a linear axis of maximum uplift that extends about 350 miles between Sitkalidak Island in the Kodiak group and Mon- tague Island in Prince William Sound. The discrepancy in the position of the inferred axis of maximum uplift offshore from the Kodiak Island area between figure 21 and that of a prelimin- ary report (Plafker and Mayo, 1965, fig. 12) results from a pre- vious error in calculating propa- gation velocities of the waves. The axis of the inferred wave source as shown in figure 17 and as computed independently by Van Dorn (1964, figs. 8, 10) are in close agreement. At the north end of this axis near Montague Island there is clear evidence for significant upward displacement in a 6-mile—wide segment of the sea floor. Both extreme crustal warping and large fault disloca- tion must have taken place. The southwest tip of Montague Island was elevated as much as 33 feet relative to sea level; there was at least 18 feet of vertical offset on land along the Patton Bay fault (Plafker, 1965, p. 1132; Plafker and Mayo, 1965, p. 7). Furthermore, the fault has been traced as a prominent scarp in the ocean floor for at least 15 miles southwest of Montague Is- land towards the Kodiak group of islands. As much as 50 feet of net uplift of the sea floor along this line has been inferred from com- parison of pre- and postearth- quake bathymetric surveys made by the US. Coast and Geodetic Surveys southwest of Montague Island (Malloy, 1964). The inferred source area shown on figure 18, from which destruc- tive waves radiated northwest- ward to batter the coast of the Kodiak group of islands could be a southwestern extension of the axis of uplift and faulting in the Montague Island area. As in that area, it may represent a relative- ly narrow strip of extreme differ- ential uplift and (or) faulting that is superimposed on the broad regional zone of uplift. The posi- tion of the inferred wave source is largely in an area of flat slopes and shoal water on the continen- tal shelf; the possibility is thus virtually precluded that the GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D39 waves result either from massive submarine landslides along the continental slope or from special resonant oscillations of water in the adjacent Aleutian Trench. The low—amplitude dipole-type wave that Van Dorn (1964) con- sidered to have been formed by tectonic warping of the broad zones of uplift and subsidence shown on figure 17 apparently did not adversely affect Kodiak or the nearby islands, although it may have contained much of the wave energy radiated out into the Pacific Ocean. N 0 negative wave phase was recorded within the zone of subsidence along the shores of the islands, nor was a positive phase reported at any The earthquake and attendant subsidence and seismic sea waves severed the main coastal high- ways on Kodiak Island and caused extensive damage to the fishing, cattle ranching, and logging in- dustries on Kodiak and the near— by islands. These industries con— stitute the main economic base for the region. The city of Kodiak also receives substantial revenue from the operation and mainte- nance of the Kodiak Naval Sta- tion that is about 6 miles south of the city. The effects of the earthquake on the highway routes and on the various industries throughout the Kodiak Island region are de- scribed in the following section. The nature and cause of the ex- tensive damage to public, private, commercial, and military proper- ty at population centers in the Kodiak region is discussed in a separate volume of this series of reports on earthquake effects locality other than the city of Ko- diak prior to the arrival of the initial wave crest. The tectonic warping did result, however, in almost immediate withdrawal of water from some tilted uplifted areas and in complementary rises of water level in depressed areas. The reported initial calm rise of water observed at Kodiak imme— diately after the earthquake may have been associated with this regional crustal tilting about the axis of zero land-level change. The amount of uplift at the inferred wave source offshore from the Kodiak group of islands is not known and probably can- not be determined because the area had not been covered by de- tailed preearthquake bathymetric surveys. That the vertical sea— floor displacements may be of the same order of magnitude as those in the Montague Island area is suggested by the similarity in the maximum runup heights of the seismic sea waves (35—40 ft) along physiographically compar- able segments of coast both on the Kenai Peninsula opposite Montague Island (Plafker and Mayo, 1965, p. 10) and on Ko— diak Island. Other factors, how- ever, such as rate of uplift, initial slope at the wave source, and energy loss along the propagation path preclude direct correlation of runup heights with displace— ment at the source. DAMAGE AND CASUALTIES on Alaskan communities. The estimated dollar value of damage by the earthquake is shown in table 1. HIGHWAYS There are approximately 125 miles of highway on Kodiak and the nearby islands; about 115 miles are on Kodiak Island—85 miles of which are rural routes (fig. 22, next page). This report will consider only the 85 miles of rural highway on Kodiak Island, because the remainder are in communities and will be dis- cussed in the companion chapter in Professional Paper 542. The earthquake caused approx— imately $5,036,000 worth of dam- age to the highway system on Kodiak Island—about $446,000 in the city of Kodiak and $1,677 ,- 000 in the Kodiak Naval Station (Tudor, 1964, fig. 3). Chiniak highway on the Naval Station was damaged from the station complex to Middle Bay, the south- ern boundary of the station. Most of the damage occurred on the 47-mile route to Chiniak and Sequel Point. The seismic sea waves destroyed two bridges at Womens Bay, four in Middle Bay, one across Mayflower Creek, one across the Kalsin River at Kalsin Bay, two on Twin Creek, and one at an unnamed creek at mile 33.7, about 0.8 mile east of Twin Creek. (Mile posts are reckoned north and south of the Buskin River—mile 0.0—which flows through the main complex of the Kodiak Naval Station.) Eyewitness accounts of the seismic sea waves indicate that the bridges were not destroyed by the incoming water of the first wave, but by the outgoing wave and by the more violent in- coming and outgoing subsequent ‘waves. Figures 23 and 24 (p. D41) show seismic sea—wave dam- age to two bridges. D40 ALASKA EARTHQUAKE, MARCH 27, 1964 152°30' 58°OO’ I EXPLANATION Highway net Dashed line indicates mad bed 89» that required relocation or (/05 replacemem / SL4 Iv . 0 Bridge destroyed by seismic Anton sea waves Larson Bay 0 Bridge replaced due to regional and local sub- Muller Point Sldence Spruce . . Cape Boundary of Naval KODIAK/ Reservation j C H I N I A K B A Y US NAVAL ./ . RESERVATION ,, $0“! . . N f ._ .__. ./ «i O o/\ O \a 1 Cape / Chlnlak 7 . Sequel Pomi 57°30' — Saltery Cove 1 Luke :3 Rose Tead “,1 00 ‘5 APPPOXlMAVE MEAN 4 9,09 mammary lash ’f“ Q9. Narrow 6) Cape «<1 P o 4 8 MILES A L, 1‘ , l 22. —- Earthquake damage to major highway routes on Kodiak Island. GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D41 w w ’, _ V ‘ _ . ‘ , '* 23. — Bridge at Womens Bay destroyed by seismic sea waves. Deck of bridge collapsed owing to destruction of piles. Compare with fig. 24. Photograph by U.S. Navy, March 30, 1964. 24.— Deck of highway bridge at Womens Bay washed away by seismic sea waves. Photograph by US. Navy, March 30, 1964. D42 ALASKA EARTHQUAKE, MARCH 27, 1964 25. — Section of roadway along Chiniak highway at Womens Bay destroyed by seismic sea waves. Photograph by U.S. Navy, March 30, 1964. In addition to those that were destroyed, five bridges on the Chiniak highway had to be re- placed either because they were underwater during high tide or because their approaches needed realinement. The bridges that re- quired replacement (indicated by open circles on fig. 22) are at Womens Bay, Middle Bay, and Kalsin Bay. The seismic sea waves destroy- ed at least 13 miles of roadway on the highway to Chiniak. About 3 miles of road was destroyed at Womens Bay (figs. 22 and 25), 3 miles at Middle Bay (pl. 3), 1 mile at Mayflower Creek, and 4 miles at Kalsin Bay. The roads in these areas were not only de- stroyed but had to be realined because regional and local sub- sidence was so great that the pre- earthquake alinement left the roadway under water during high tides. The Anton Larson highway spur, 9 miles long, starts near Buskin Lake on the Naval Reser- vation and runs west and then north to Anton Larson Bay. Dam- age occurred near the terminus of the road at the bay. One mile of roadway was destroyed and required replacing. In addition, two bridges near the bay required replacing. The Saltery Cove spur road originates at the Chiniak high- way at Middle Bay. It heads gen- erally southeastward for 14 miles to Saltery Cove on Ugak Bay. Damage on this spur occurred near its point of origin at Middle Bay and at the divide between the Saltery Cove and Middle Bay drainage basins. Two miles of roadway were destroyed and one bridge had to be replaced. The Narrow Cape spur road is 15 miles long; it originates at Kalsin Bay and heads south, where it skirts the east edge of Lake Rose Tead, then runs on to Pasagshak Bay. At the head of Pasagshak Bay, the road forks. One spur continues south about 2 miles to Pasagshak Point; the other spur heads east for about 5 miles to Narrow Cape. Two miles of roadway of the Narrow Cape spur was damaged by the earthquake and attendant subsidence. At Kalsin Bay 1 mile of roadway had to be replaced. Another mile of roadway re- quired replacement at Lake Rose Tead (fig. 8). In addition to 2 miles of roadway that was de- stroyed, three bridges had to be replaced on the Narrow Cape spur road. ‘Two bridges at Lake Rose Tead and one on a small creek that flows into the head of Pasagshak Bay required replace- ment. Along the 85 miles of rural highway on Kodiak Island, 18 miles of roadway, 22 bridges, and 1 culvert were either destroyed by seismic sea waves or had to be replaced because of regional and local subsidence. No bridges were destroyed by seismic shock even though some were on foundations of unconsolidated deposits. FISHING INDUSTRY The salmon, king crab, halibut, dungeness crab, shrimp, and clam fishing industry, upon which the economy of the area is almost en- GEOLOGIC ‘. ‘ 3" EFFECTS ON THE KODIAK ISLAND AREA D43 u . ‘ ‘ fl ,5 ,. W 26..—Chaotic condition of the commercial section of the city of Kodiak following inundation by seismic sea waves. The Ne. *‘ a. . 5‘.- .372», small-boat harbor, which was in the left background, contained an estimated 160 crab and salmon fishing boats when the waves struck. Photograph by US. Navy, March 30, 1964. tirely dependent, suffered devas— tating losses of processing facili— ties, vessels, and gear. Three can— neries at Kodiak and one at Uzinki were destroyed by seismic sea waves, and one other cannery and a cold-storage plant at Ko- diak were severely damaged. The Kadiak Fisheries cannery at Shearwater Bay was damaged by seismic shock during the earth- quake, after which it was virtu- ally destroyed by the waves. Esti- mated value of the structures and the finished products they con- tained is $2,610,000. The total number of lost or damaged fishing boats is not known. In the vicinity of Kodiak alone, 13 vessels were lost and 19 others were swamped or run aground (fig. 26). Vessels under- way in deep water were not dam- aged by the waves. One boat, a 38-foot seiner, went down with its crew of six when it either struck a wave—propelled log or was grounded between waves in shallow water near Spruce Cape, 8 miles northeast of Kodiak. Two other lives were lost on two fish- ing boats that sank in the Kodiak small—boat harbor. Replacement cost of vessels that were sunk, grounded, or otherwise damaged throughout the area has been esti- mated at $2,466,500. As a result of regional tectonic subsidence, three canneries and their facilities were made unus- able by inundation, and four others sustained costly damage. The lost and damaged cannery fa- cilities were valued at $1.3 mil- lion (Alaska Dept. Fish and Game, 1965, p. 21). Studies of salmon habitat by the Alaska Department of Fish D44 and Game (1965, p. 3) indicate that, because of subsidence, tides have made many previous inter— tidal salmon spawning areas nearly useless and have engulfed former fresh-water stream rear- ing areas. As shown by table 3, tidal inundation of streams ex- tends as much as 4,500 feet farther inland than before the earthquake. According to the De- partment of Fish and Game, at least two important salmon fish- ing areas did not produce in 1964, even though it was a bumper sal- mon year elsewhere. Studies have been initiated to determine wheth- er the poor catch was due to nor- mal seasonal variations in salmon runs, or was in some way related to the earthquake. Significant changes in currents and the time of tide changes have been reported by fishermen on the Shelikof Strait side of the islands since the earthquake. These changes are probably re- lated to changes in water depth resulting from the tectonic de- formation. The diiference in tides and currents has required modi- fication of fishing techniques and could have a detrimental long- term effect on the migration pat— tern of the salmon. In Sitkalidak and Raspberry Straits (fig. 2) where subsidence amounted to about 1 and 31/2 feet, respectively, local residents re- port not only higher high tides since the earthquake but also lower low tides. In these areas, clamming is reportedly better on minus tides than it was prior to the earthquake. These reports have not been confirmed or ex- plained by the fisheries authori- ties. Tectonic uplift and subsidence has changed the appearance of landmarks used by vessel opera- tors along the treacherous rocky coast and has altered water ALASKA EARTHQUAKE, MARCH 27, 1964 27. — Part of Beaty Ranch showing extent of wave inundation, as indicated by light-colored debris. Ranch buildings are a quarter of a mile from the shoreline. depths. These changes have re- portedly caused several vessels to be damaged or sunk since the earthquake by collision with sub- merged objects. CA'T‘TLE RANCHES The earthquake was a near- disaster to several of the small cattle ranchers in the area who were struggling to develop sound and profitable operations. Inundation by seismic sea waves caused extensive property damage at all the isolated cattle ranches along the southeast coast of Kodiak Island and on Sitkali— dak Island. According to a US. Department of Agriculture re- port (Oliver, 1964), 175 cattle were drowned, 3 ranch houses and 2025 outbuildings with their contents were swept away, and 25 miles of corral and range fenc- ing was lost. In addition, several hundred acres of land used for hay production was littered with wave—borne debris (fig. 27). Numerous beaches and deltas along the southeast coast of Ko- diak and Sitkalidak Islands are now inundated by tides, and salt water has ruined several hundred acres of the best beach rye lands which were used for winter graz- ing. LOGGING INDUSTRY Logging in the area is carried on by small operators and is al- most entirely for local consump- tion. One of the sawmills near Afognak village on southern Afognak Island was abandoned GEOLOGIC EFFECTS ON THE KODIAK ISLAND AREA D45 because seismic sea waves washed away most of the docking facili- ties and caused extensive damage to the mill. A second small mill near Uzinki sustained an un- known amount of wave damage to the dock. CASUALTIES In the Kodiak Island area the earthquake caused the loss of 18 lives through drowning, but the extent of wave damage and at- tendant loss of life would un- doubtedly have been much great-- er had not the largest waves co- incided with low, rather than high, stages of tide. Furthermore, the disaster would have been worse had the earthquake occur— red during the summer when the population of the islands is sub- stantially increased by large num- bers of fishermen and cannery workers, many of whom normally live and fish along shores. The toll of 18 dead, although relatively light, was to a consid— erable extent needless and was caused largely by public ignor- ance concerning the nature and destructive force of seismic sea waves. Especially serious is the prevailing mistaken belief that the first wave is necessarily the highest. As a consequence some individuals made the disastrous error of returning to low-lying coastal areas after the initial wave, and 10 of them were trap- ped and drowned by subsequent waves. Furthermore, the first wave did not reach any of the boat harbors until about an hour after the earthquake, and a local warning broadcast was issued half an hour earlier advising of the impending wave, yet few manned vessels were piloted to the safety of nearby deep water. All the vessels that were sunk, including three in which eight lives were lost, went down in shallow water. Clearly, a better method of disseminating information is re- quired as the key to minimizing loss of life from future seismic sea waves in coastal areas of the Kodiak group of islands, as well as elsewhere along coastal seg- ments of the Gulf of Alaska that may be threatened by similar waves in the future. Evacuation Alaska Department of Fish and Game, 1965, Post-earthquake fisheries evaluation; an interim report on the March 1964 earthquake effects on Alaska’s fishery resources: Juneau, Alaska, 72 p. Alaska Department of Health and Welfare, 1964, Preliminary report of earthquake damage to environ- mental health facilities and serv~ ‘lces in Alaska: Juneau, Alaska Dept. Health and Welfare, En— vironmental Health Br., 46 p. Bancroft, H. H., 1886, The history of Alaska: San Francisco, Calif, 45.. L. Bancroft and Co., 775 p. from the great 1964 Alaskan Bolt, B. A., 1964, Seismic air waves REFERENCES CITED earthquake: Nature, v. 202, no. 4937, p. 1095—1096. Capps, S. R., 1937, Kodiak and adja— cent islands [Alaska]: U.S. Geol. Survey Bull. 880-0, p. 111-184. Chevigny, Hector, 1942, Lord of Alaska; Baranov and the Russian adventure: New York, The Vik— ing Press, Inc., 320 p. Grantz, Arthur, Plafker, George, and Kachadoorian, Reuben, 1964, Alas- ka’s Good Friday earthquake, March 27, 1964: U.S. Geol. Sur— vey Circ. 491, 35 p. Gutenberg, Beno, 1957, Effects of ground on earthquake motion: Seismol. Soc. America Bull., v. 47, no. 3, p. 221—250. procedures for the populace and for boats should be carefully worked out in advance, and it should be made clear that the danger from such waves may per- sist for several hours after the initial wave strikes. The Seismic Sea Wave Warning System oper- ated by the US. Coast and Geo- detic Survey can provide advance warning where destructive waves are generated by distant earth- quakes in the Pacific Ocean basin. The present system, however, cannot assure warnings for loca- tions closer to the wave source than about 1,000 miles because at least an hour may be required to locate the earthquake epicenter, determine its magnitude, and broadcast a warning (Spaeth and Berkman, 1965, p. 4—7). Conse— quently, in earthquake-prone areas—such as the Gulf of Alaska and Alaska Peninsula— where any strongly felt earth- quake of long duration could be accompanied by destructive seis- mic sea waves, immediate evacu- ation of low-lying coastal areas is the only prudent course of action. Hodgson, J. H., 1964, Earthquakes and earth structure: New Jersey, Prentice-Hall, Inc., 166 p. Lamb, Horace, 1932, Hydrodynamics: 6th ed., - England, Cambridge Univ. Press, 738 p. Leet, L. D., 1946, Earth motion from the atomic bomb test [New Mex- ico, July 16, 1945]: Am. Scientist, v. 34, no. 2, p. 198—211. Malloy, R. J., 1964, Crustal uplift southwest of Montague Island, Alaska: Science, v. 146, no. 3647, p. 1048-1049. Moore, G. W., 1964, Magnetic dis- turbances preceding the Alaska 1964 earthquake: Nature, v. 203, no. 4944, p. 508-509. D46 Oliver, W. B., 1964, A shock to Alaska agriculture: Soil Conservation, V. 30, no. 2, p. 30-31. Plafker, George, 1965, Tectonic de— formation associated with the 1964 Alaska earthquake: Science, V. 148, no. 3678, p. 1675—1687. Plafker, George, and Mayo, L. R., 1965, Tectonic deformation, sub- aqueous slides and destructive waves associated with the Alas— kan March 27, 1964, earthquake —an interim geologic evaluation: U.S. Geol. Survey open—file re— port, 21 p. Richter, C. F., 1958, Elementary seisv mology: San Francisco, Calif., W. H. Freeman and Co., 768 p. Ruttner, Franz, 1953, Fundamentals of limnology: 2nd ed., Toronto, Canada, Toronto Univ. Press, 242 p. Spaeth, M. G., and Berkman, S. C., 1965, The tsunami of March 28, ALASKA EARTHQUAKE, MARCH 27, 1964 1964, as recorded at tide sta- tions: U.S. Coast and Geod. Sur— vey, 59 p. Tudor, W. J., 1964, Tsunami damage at Kodiak, Alaska, and Crescent City, California, from Alaskan earthquake of 27 March 1964: U.S. Naval Civil Eng. Lab., Tech. Note N—622, 124 p., Port Hue- neme, Calif. U.S. Coast and Geodetic Survey, 1964a, Prince William Sound, Alaskan earthquakes, March- April, 1964: U.S. Coast and Geod. Survey, Seismology Div. Prelim. Rept., 83 p. 1964b, Tide tables, high and low water predictions, 1964, West Coast North and South America including the Hawaiian Islands: U.S. Coast and Geod. Survey, 224 p. 1965, Assistance and recovery in Alaska, 1964: U.S. Coast and Geod. Survey, 45 p. U.S. Weather Bureau, 1964, Clima- tological data, Alaska, v. 50, no. 3, p. 42-58. Van Dorn, W. G., 1964, Source mech- anism of the tsunami of March 28, 1964, in Alaska: Coastal Eng. Conf., 9th, Lisbon 1964, Proc., p. 166-190. Varnes, D. J., 1958, Landslide types and processes, in Eckel, E. B., ed., Landslides and engineering: Natl. Acad. Sci—Natl. Research Council Pub. 544, Highway Research Board Spec. Rept. 29, p. 2047. Wahrhaftig, Clyde, 1965, The physio- graphic provinces of Alaska: U.S. Geol. Survey Prof. Paper 482, 52 p. Waller, R. M., Thomas, H. E., Vorhis, R. C., 1965, Effects of the Good Friday earthquake on water sup- plies: Am. Water Works Assoc. Jour., v. 57, no. 2, p. 123-131. {:u. 5. GOVERNMENT PRINTING OFFICE: 1966 —0 216-706 The Alaska Earthquake March 27, 1964 Regional Effects . Conner River Basin Area GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—5 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS Effects of the Earthquake Of March 27, 1964 In the Copper River Basin Area, Alaska By OSCAR J. FERRIANS, JR. A description and analysis of the earthquake- induced ground breakage, landsliding, regional uplift and subsidence, and damage. to manmade structures in the Copper River Basin GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—E UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1966 For sale by the Superintendent of Documents, US. Government Printing Office Washington, DC. 20402 THE ALASKA EARTHQUAKE SERIES The US. Geological Survey is publishing the re- sults of its investigations of the Alaska earthquake of March 27, 1964:, in a series of six professional papers. Professional Paper 543 describes the re— gional effects of the earthquake. In chapters of this volume already published, studies of slide-induced waves and seiching at Kenai Lake, geomorphic ef- fects in the Martin-Bering Rivers area, gravity sur- veys of the Prince William Sound region, and effects of the earthquake on Kodiak and nearby islands have been reported. Other professional papers in the series describe field investigations and reconstruction and the effects of the earthquake on communities, on the hydrologic regimen, and on transportation, utilities, and communications. Abstract _____________________ Introduction and acknowledg- ments ______________________ Geographic setting ____________ Physiography _____________ Climate __________________ Roads and settlements _____ Geologic setting _______________ Bedrock __________________ Major tectonic elements--- - Unconsolidated deposits-- _ - Permafrost and seasonal frost ------------------- The earthquake and its after- shocks --------------------- Seismic data -------------- Ground motion ----------- Sound ------------------- 1. Index map showing the loca- tion of the Copper River Basin area ______________ 2. Generalized geologic map of the Copper River Basin area ____________________ 3. Map showing major tectonic elements of south-central Alaska ----------------- 4. Map showing recent regional uplift and subsidence in south-central Alaska ----- 5. Aerial View of the northern end of Tonsina Lake--__ 6. Aerial view showing intensely fractured ice along the eastern shore of Tazlina Lake ___________________ 7. Map showing the distribution of ground cracks _________ 8. Sketch map of the Little Tonsina River landslide site-- 9. View of pressure ridges on Page E 1 @WWWWWNNN C5 “46503 Page E2 10 11 12 14 CONTENTS Geologic effects of the earth- quake ---------------------- Regional uplift and sub- sidence ----------------- Changes in ground— and surface-water conditions_ _ Ground water --------- Surface water _________ Cracking of lake and river ice-- Ground cracks and associ- ated landsliding --------- Terminology ---------- General distribution_ _ _ Geologic controls ______ Little Tonsina River site__ Other landslide sites . _ _ - Nelchina River outwash apron ______________ ILLUSTRATIONS FIGURES floor of Little Tonsina River valley ____________ 10. Large ground crack parallel- ing north side of small hill near mile 65, Richardson Highway --------------- 11. Photograph of pressure ridge on flat surface bordering Little Tonsina River ----- 12. Large ground crack formed in clearing along south shore of lake ____________ 13. Aerial view of the large braided flood plain of the Nelchina River ---------- 14. Photograph of typical ground crack in outwash apron of Nelchina River __________ 15. Vertical aerial view of Nel- china River delta ________ 16. Vertical aerial View of ground cracks along and parallel- ing hingeline on Nelchina River delta ------------- Page E8 Page E14 15 15 16 17 18 19 20 Geologic effects—Continued Ground cracks and associated landsliding—Continued Nelchina River delta- - - Origin --------------- Lateral extension- - Horizontal com— paction ________ Diflerential verti- cal compaction_ _ Snowslides, rockslides, and avalanches ------------- Effects of the earthquake on manmade structures --------- Damage to buildings and associated structures ----- Damage to highways and bridges ----------------- References -------------------- 17. Aerial view of north end of the large Nelchina River delta along the western shore of Tazlina Lake-__- 18. Aerial view of Kiana Creek delta along eastern shore of Tazlina Lake ------------ 19. Aerial view showing large in- undated area along eastern shore of Tazlina Lake-_-_ 20. Diagrammatic sections show— ing mechanism of forma- tion of basic types of ground cracks ----------- 21. View of pavement surface of the Richardson Highway- - 22. Little Tonsina River bridge shoved from right to left by landslide movement--_ 23. Small crack across aban- doned section of road on a low gravel terrace -------- v Page E17 23 23 23 24 24 25 25 25 27 Page E21 21 22 23 26 26 27 stag” THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS EFFECTS OF THE EARTHQUAKE OF MARCH 27, 1964, _IN THE COPPER RIVER BASIN AREA, ALASKA The Copper River Basin area is in south-central Alaska and covers 17,800 square miles. It includes most of the Copper River Basin and parts of the surrounding Alaska Range and the Tal- keetna, Chugach, and Wrangell Moun- tains. On March 27, 1964, shortly after 5: 36 p.m. Alaska standard time, a great earthquake having a Richter magnitude of about 8.5 struck south-central Alaska. Computations by the U.S. Coast and Geodetic Survey place the epicenter of the main shock at lat 61.1° N. and long 147.7° W., and the hypocenter, or actual point of origin, from 20 to 50 kilometers below the surface. The epicenter is near the western shore of Unakwik In- let in northern Prince William Sound; it is 30 miles from the closest point Within the area of study and 180 miles from the farthest point. Releveling data obtained in 1964 after the earthquake indicates that broad areas of south-central Alaska were warped by uplift and subsidence. The configuration of these areas generally parallels the trend of the major tectonic elements of the region. Presumably a large part of this change took place dur- ing and immediately after the 1964 earthquake. The water level in several wells in the area lowered appreciably, and the wa— ter in many became turbid; generally, however, within a few days after the earthquake the water level returned to normal and the suspended sediment set- tled out. Newspaper reports that the By Oscar J. Ferrians, Jr. ABSTRACT Copper River was completely dammed and Tazlina Lake drained proved erro- neous. The ice on most lakes was cracked, especially around the margins of the lakes where floating ice broke free from the ice frozen to the shore. Ice on Taz- lina, Klutina, and Tonsina Lakes was intensely fractured by waves generated by sublacustrine landslides ofi the fronts of deltas. These waves stranded large blocks of ice above water level along the shores. River ice was gen- erally cracked in the southern half of the area and was locally cracked in the northern half. In the area of study, the majority of the ground cracks occurred within a ra- dius of 100 miles from the epicenter of the earthquake. Ground cracks formed in flood plains of rivers, in deltas, and along the toes of alluvial fans. They also occurred locally in low terraces adjacent to flood plains, in highway and other fill material, along the mar- gins of lakes, along the faces of steep slopes of river bluffs and hillsides, and in areas cleared of vegetation for sev- eral years. The ground cracks were restricted to areas underlain by unconsolidated de- posits Where one or more of the follow- ing conditions existed: (1) permafrost was absent or deep lying, (2) the ground-water table was near the sur- face, (3) bedrock was relatively deep lying, and (4) slopes were steep. Be- cause the earthquake occurred in March, seasonal frost was present throughout the area. Despite the diversity of local condi- tions, the origin of most of the ground cracks can be explained by the following mechanisms: (1) lateral extension, caused by materials moving toward an unconfined face such as a lakeshore, river bluff, hillside, or terrace escarp— ment; (2) horizontal compaction, caused by repeated alternate compres- sion and dilation (in the horizontal direction) of materials in flat-lying areas where there are no unconfined faces; (3) differential vertical compac- tion, caused by the shaking of materials that vary laterally in thickness or character; and (4) combinations of the above. Snowslides, avalanches, and rock- slides were restricted to the moun- tainous areas surrounding the Copper River Basin. They were especially numerous in the Chugach Mountains which are closest to the. epicenter of the earthquake. The large amount of snow and rock debris that has cascaded onto the icefield and glaciers of these moun- tains, and, probably even more impor- tant, the overall disturbance to the ice field will afiect the regimen of the glaciers. Most of the damage to manmade structures occurred in the southern half of the area, and, primarily because of the sparsity of population and man- made structures, property damage was not great and no lives were lost. E1 E2 ALASKA EARTHQUAKE, MARCH 27, 1964 INTRODUCTION AND ACKNOWLEDGMENTS The Copper River Basin area, as defined in this report, covers about 17,800 square miles in south-central Alaska (fig. 1) ; it includes most of the Copper River Basin and parts of the surrounding mountains. ALASKA\\ Copper River _Basin area nichorag Wr\J Epicenter flu \ (fl GULF OF ALASKA up 0 200 400 MiLES K_J.AJ._.L_I 1.—Location of Copper River Basin area. Its boundary on the east is long 144°00’ W.; on the north, lat 63°- 00’ N.; on the west, long 147°40’ W.; and on the south, approxi- mately coincides with the axis of the Chugach Mountains. The primary purpose of this in- vestigation was to determine the effects of the great Alaska earth- quake of 1964: in the Copper River Basin area and the geologic factors that tend to control the distribu- tion and character of these effects. An understanding of these geo- logic factors is necessary to evalu- PHYSIOGRAPHY The Copper River Basin is bor- dered on the north by the Alaska Range, 0n the west by the Talk- eetna Mountains, on the south by the Chugach Mountains, and on ate properly the engineering significance of the earthquake. Because of the US. Geological Survey’s geologic mapping of the Copper River Basin, in progress since 1952, the general distribution and character of unconsolidated deposits in the area are well known. Without this work, ade- quate evaluations of the effects of the earthquake would not have been possible during the short time available for field investigation. Observations of the effects of the earthquake in the Copper River Basin were made between July 2 and July 23, 1964. During the 3%-month interval between the occurrence of the earthquake and the time when observations were made, undoubtedly many earth- quake-related features, especially subtle ones, were obliterated. Be- cause of this time lag a large part of the discussion about earthquake damage to manmade structures is based on eyewitness accounts; however, most of the discussion concerning the natural earth— quake-related features, as distin- guished from damage to life and property, is based on the author’s observations and on the study of postearthquake aerial photo- graphs. GEOGRAPHIC SE'I'I‘ING the east by the Wrangell Moun- tains. The basin floor is a plain of low relief into which the major streams have cut steep-walled valleys as much as 400 feet deep. However, Since most of the people live near the highway, a systematic study along the roads provided fairly comprehensive coverage of damage to manmade structures. In the vast, inaccessible areas away from the highways, the distribu- tion of earthquake-related features was determined by observation from light fixed—wing aircraft. At several sites where landings were possible, these features were ex- amined on the ground. I am greatly indebted to all of the individuals who provided eye- witness accounts of their experi- ences during and immediately after the earthquake. Rich Hous- ton and Cleo McMahan, both bush pilots, not only did an excellent job of chauffeuring, but also con- tributed their own time and energy to making critical observations pertinent to this study. James B. Small of the US. Coast and Geodetic Survey provided relevel- ing data, Robert M. Chapman of the US. Geological Survey made available an unpublished report which included his observations along the Richardson Highway made soon after the earthquake, and several individuals of the Alaska Department of Highways provided data concerning damage to the highways. the bordering mountains are ex- tremely rugged and they support numerous glaciers, which, along with their melt—water streams, have cut deep valleys opening into the basin. EFFECTS IN THE COPPER RIVER BASIN AREA E3 The greater part of the area of study is drained by the Copper River, which originates from large glaciers on the northern side of the Wrangell Mountains. From its source the river flows in a large arc around the northern, western, and southwestern sides of the Wrangell Mountains, and then southward through the Chugach Mountains into the Gulf of Alaska. Major tributaries to the Copper River in- clude the Chistochina, Sanford, Gakona, Gulkana, Tazlina, Klu- tina, Tonsina, and Chitina Rivers. The northwestern corner of the area is drained by the Susitna River which originates from a large glacier in the Alaska Range. From its source this river drains southward to the Copper River Basin, then makes a sharp turn and flows westward through moun- tainous terrain to the north end of the Susitna Lowland; from here it flows southward to Cook Inlet. Within the area of study the ma- jor tributaries to the Susitna River are the Maclaren and Oshetna Rivers. A small part of the southwestern corner of the area is drained by the Matanuska River which flows westward to Knik Arm of Knik Inlet. The basin floor is dotted with lakes; the three largest are Lake BEDROCK Mountainous areas around the margin of the Copper River Basin and prominent hills within the basin are underlain by altered volcanics and interbedded gray- wacke and slate, and by virtually unaltered sediments, all of which have been intruded locally by a 221—693 o—cc~—2 Louise, Crosswind Lake, and Ewan Lake. Tazlina, Klutina, Tonsina, and Paxson Lakes are large water bodies occupying deep valleys in the mountainous areas surround- ing the basin. CLIMATE The Copper River Basin area has an arctic continental weather regime with short warm summers and long cold winters. Mean an- nual temperature, as recorded at the Gulkana Federal Aviation Agency Station in the east—central part of the area of study, is 27°F, and mean annual precipitation is 12 inches. These climatic condi- tions are significant to this study because they are conducive to the formation and preservation of permafrost (perennially frozen ground)—and the presence or absence of permafrost is significant in relation to the distribution and character of earthquake effects. ROADS AND SETTLEMENTS Large segments of two major Alaska highways cross the Copper River Basin area. Both highways are paved two-lane roads which connect coastal cities in south- central Alaska with interior Alaska. The Glenn Highway, 314 GEOLOGIC SETTING wide variety of igneous rocks. Extensive areas are underlain also by a considerable thickness of un- altered basaltic and andesitic lava. These rocks range in age from middle Paleozoic to Pleisto- cene. The distribution of bed- rock within the area of study is shown on figure 2 (next page). miles long, extends from Anchor- age to Tok Junction on the Alaska Highway; the (Richardson High- way, 365 miles long, extends from the ice-free port of Valdez to Fair- banks. Locations along these highways are designated by the number of miles from the point of origin of the highway—along the Glenn Highway, miles from An- chorage, and along the Richardson Highway, miles from Valdez. Other roads within the area of study include the Edgerton High- way and the Lake Louise Road. The Edgerton Highway, a two- lane gravel road in the southeast— ern corner of the area, is 39- miles long and connects Chitina with the Richardson Highway. The Lake Louise Road, also a two-lane gravel road, leads north from mile 160, Glenn Highway, for approx- imately 19 miles to Lake Louise. The settlements within the area are concentrated along or near the major highways. Glennallen, the principal settlement of the area, with a population of around 200 people, is near the junction of the Glenn and Richardson Highways. Other settlements include Chitina, Copper Center, G‘rulkana, Gakona, Chistochina, and Paxson. Several roadhouses and some service enter- prises are located along the high- way between these settlements. MAJOR TECTONIC ELEMENTS There are six major tectonic ele— ments within the area of study (fig. 3, p. 6). From north to south, they consist of the Alaska Range geosyncline, the Talkeenta geanticline, the Matanuska geo— syncline, the seldovia geanticline, E4 ALASKA EARTHQUAKE, MARCH 27, 1964 144“ 63" 146° - 39's- '.';.'7';DV.‘&._ .; .2.&.;,...m.... .‘ ." .' “$7 ....7./ 3315‘... M .. , 62" Base from Anchorage, World (N. America) 1:1 000 000 scale, compiled by Army Map Service EXPLANATION '1.44° Coarse-grained unconsolidated deposits Fine- and coarse-grained unconsolidated Bedrock ranging in age from middle of Recent age 0 1o 20 30 I L | l l deposits of Pleistocene age Paleozoic to Pleistocene 40 MILES J 2.—Generalized geologic map of the Copper River Basin area. Geology compiled by Ferrians, Nichols, Williams, and Yehle; Grantz (pl. 23 in Andreason and others, 1964) ; and Coulter and Coulter (1962). EFFECTS IN THE COPPER RIVER BASIN AREA 150° 148° 146° I 144° We“ ,8 . Cach Est ° I 1 . o a . ¢ Hart! 0 50 100 MILES O ;a Standa AN KS I I l I I | ’ | a North Nena” ‘ {£3 , II o 55034 1 Nenana ' g. \ t a . 45 I ‘2‘ Auror / 9" River I ”‘ ‘ o 1 M A u " T: O’I \\ odg 4 Q‘b'g I 5050 b , ”Iv HARPER$ 0 “an; L South F3515. ' .65/5 572O‘MTVU_ Frankiino B‘ .0 4450- ~ . ’ 0 F 2 me R / 50-6 . 04 4< . . a“ .. “WM!" I m ' I ealy L ‘ ‘ \L ’= %B~\- I II 1? s ’01, ‘99 it. 65 3‘ ”V w ,. I o -' “‘ ,5: / 4% (gheod'ée aka t Lake ’ 712 I P51 ,. ' i’ ‘ 4 ,' Lake . M S I. m _Wns 16 . I I I5 leld Vfl ag 1 73 ‘ ss 3/“ _ unction 62° b . d;$ \a , acier I A!" Q} atme age ' or ' 81d akedI P ‘ I 2/5 Red Hea. 083‘ «9'5” RIQVZE WILLIA MQgi‘oma Ba . WK» ‘ SOUND I 0 ‘ anght I 2335252,,“ , (’3 inchinbrook I 9’ -5 I 411° 146° . Base from U.S. Geological Survey 1:2 500 000 a—Major tectonic elements of south-cwu'al Alaska (after Payne and Dutro, in Miller and others, 1958, pl. 2). diagonal lines. E5 64° 62° Geosynclinal area (troughs) shown by light stippling; geanticlinal areas (arches), by heavy stippling; and the Copper River Basin, by E6 and the Chugach Mountains geo— syncline (Payne and Dutro, in Miller, and others, 1959, pl. 2). In the central part of the area the Cenozoic Copper River Basin is superimposed on several of the other tectonic elements. UNCONSOLIDATED DEPOSITS In the mountainous areas sur- rounding the Copper River Basin the floors of the major valleys are underlain predominantly by un- consolidated sediments of Pleisto- cene and Recent age. These sedi- ments are chiefly outwash gravel deposited by glacier—fed streams and till deposited by glaciers which formerly covered the valley floors. Locally, significant thick- nesses of colluvial deposits are present. The greater part of the basin it- self is underlain by lacustrine, gla- cial, and fluvial sediments of Pleistocene and Recent age (Karl— strom and others, 1964). In the ALASKA EARTHQUAKE, MARCH 27, 1964 east—central part of the basin, these deposits are quite thick. The ex- act thickness is unknown; how- ever, as much as 400 feet is well exposed in bluffs along the deeply entrenched rivers, and near Glenn- allen a water well 502 feet deep bottomed in unconsolidated de- posits. In the western part and along the margin of the basin, these deposits generally overlie bedrock at relatively shallow depths. The distribution of un- consolidated deposits within the area of study is shown on figure 2. PERMAFROST AND SEASONAL FROST Permafrost and seasonal frost were important factors in deter- mining the distribution and char- acter of ground cracks and land- slides. Most fine-grained uncon- solidated deposits in the Copper River Basin are perennially frozen from 1 to 5 feet below the surface to depths of as much as 200 feet. In similar areas that have been cleared of vegetation for several years, such as those along the high- way net or in settled areas, the permafrost table (top of perma- frost) occurs from 10 to 20 feet be- low the surface. In contrast to the fine-grained deposits, the coarse-grained gravel deposits along the major streams generally are free of permafrost. Perma- frost generally is also absent near large deep lakes because of the enormous amount of heat held in these large bodies of water. In the Chugach Mountains south of the basin, permafrost occurs as iso- lated masses where local condi- tions are favorable for its forma- tion and preservation (Ferrians, 1965). Seasonal frost generally pene- trates to depths from 2 to 4 feet in poorly drained environments such as marshes; however, in well- drained environments such as gravel terraces, seasonal frost gen- erally penetrates from 10 to 20 feet. THE EARTHQUAKE AND ITS AFTERSHOCKS SEISMIC DATA Computations by the US. Coast and Geodetic Survey (1964, p. 30) placed the time of origin of the earthquake at shortly after 5:36 p.m‘. Alaska standard time, March 27, 1964, the epicenter at lat 61.05° N. and long 14750" W., and the hypocenter, or actual point of origin, at about 20 kilometers (about 121/2 miles) below the sur- face. Later computations, based on more data, place the epicenter at lat 61.5° N. and long 147.7° W. and the hypocenter at a depth between 20 and 50 kilometers. Estimates of the Richter magni- tude of the earthquake, from four United States stations, range from 8.4 to 8% and place it among the historically great earthquakes. The epicenter, which is near the west shore of Unakwit Inlet in northern Prince William Sound, is 30 miles from the closest point within the area of study and 180 miles from the farthest point. During the first month after the main shock more than 7,500 after- shocks were recorded at the five seismograph stations that were in- stalled in the epicentral area soon after the main shock occurred (Algermissen, 1965, p. 2). Press and Jackson (1965, p. 867) re- ported that approximately 12,000 aftershocks, having a magnitude equal to or greater than 3.5, oc- curred during a 69-day period after the main shock. Press and Jackson (1965, p. 868) EFFECTS IN THE COPPER RIVER BASIN AREA E7 estimated that a line distribution of 100 underground nuclear explo- sions, totalling 100 megatons each, would be necessary to equal the seismic energy released by the Alaska earthquake. GROUND MOTION The tremendous amount of energy released during the earth- quake caused severe ground shak- ing throughout most of the south- ern half of the area of study. However, locally there was varia- tion in the reported intensity and duration of shaking. Most observ— ers reported that the ground mo— tion lasted 3—5 minutes. Guten— berg (1956, 1957) demonstrated that within a small area the dura- tion and intensity of ground shak- ing can vary considerably depend— ing upon the character of the un- derlying materials. For example, water - saturated unconsolidated deposits will be affected more than similar unconsolidated deposits which are dry, and much more than hard crystalline bedrock. From Glennallen southward and westward the ground motion generally was described as having either a jarring effect or a strong rolling effect on most dwellings, whereas north of Glennallen most people reported a gentle rolling effect, like swells on the ocean. People reported seeing trees and telephone'and power poles being whipped back and forth by the ground motion. Many people reported seeing the ground undulate in waves like those generated on large bodies of water. Unfortunately, during a great earthquake it is difficult for individuals to make objective ob- servations. However, one ob— server estimated that at the junc- tion of the Glenn and Richardson Highways, 100 miles from the main epicenter, these waves were about 10 feet apart and about 3 feet high. Other reliable wit- nesses estimated that the ground waves near Slana, 165 miles from the main epicenter, were 50—60 feet apart and 18—20 inches high. All the witnesses stated that the waves moved from the southwest to the northeast. Even though many individuals have reported visible ground waves during major earthquakes, especially in flat-lying areas un— derlain by water-saturated uncon- solidated deposits, many seismolo- gists discount such occurrences. One of the main reasons for this skepticism is that ground waves recorded on seismographs have crests and troughs too small to be seen, commonly no more than 0.004 inch for large surface waves and hardly ever more than 0.020 inch (Leet and Leet, 1964, p. 79) ; fur- thermore, practically all of these waves travel too fast to be seen. Other explanations used to dis- count the existence of visible ground waves include optical, physiological, and psychological illusions. The opinion of Eiby (1957, p. 26) is typical of that of many seis- mologists. He presents argu- ments against the existence of visi- ble ground waves and then states, “However, something of the kind has so often been reported in good faith that it must be supposed that the violent shaking so effects a man as to produce an illusion of this kind. With this demonstra— tion of human fallibility it‘is as well to pass to the problem of in? strumental recording.” In spite of the difficulty of ex- plaining the origin of such phe- nomena, many workers have con- cluded that visible ground waves do exist. These include Dutton (1890, p. 267—268) , Oldham (1899, p. 4—41), the State Earthquake In- vestigation Commission (1908, v. 1, p. 380—381),Kn0tt (1908, p. 18), Fuller (1912, p. 57), Imamura (1937, p. 75), and India Geologi- cal Survey Officers and Roy (1939, p. 29—30). In a more recent publication, Richter (1958, p. 132) concludes that there is almost certainly a real phenomenon of visible progressing or standing waves on soft ground, but he thinks the effect may be as- sociated with earth lurching, with which it may be confused. The phenomenon of visible ground waves obviously poses a perplex- ing problem that needs further study, but the evidence in the Cop- per River basin area tends to for- tify the conclusion that visible ground waves do exist. SOUND A few people heard sounds which presumably were associated with the earthquake. Generally the sound was described as a rum- bling or roaring like distant thun- der. This type of phenomenon has occurred during many other g r e a t earthquakes. Thompson (1929) and Kingdon-Ward (1951, p. 130, and 1953, p. 172) have de- scribed unusually loud sounds heard during two such earth- quakes. ' Even though there is some dis- agreement on certain points, the perception of sound is fairly well explained by seismologists: the sound waves are produced directly by the transfer of elastic ane energy from the ground to the air (Richter, 1958, p. 128). Some earthquakes have trans- ferred enough energy to the air that sensitive barographs several hundred miles from the epicenter were affected. Gutenberg and Benioff (1939) first described this effect. E8 ALASKA EARTHQUAKE, MARCH 27, 1964 GEOLOGIC EFFECTS OF THE EARTHQUAKE REGIONAL UPLIFT AND SUBSIDENCE During the summer of 1964 the US. Coast and Geodetic Survey completed aproximately 900 miles of first-order leveling in Alaska areas affected by the earthquake. About 680 miles of this was re- leveling of previously established first-order leveling. Releveling was accomplished between Seward and Anchorage Via The Alaska Railroad, between Anchorage and Glennallen via the Glenn High- way, and between Valdez and a point 15 miles southeast of Fair- banks via the Richardson High- way. A large part of the resurveyed net was first leveled in 1922 and 1923; new stations were estab- lished and others releveled along parts of this net in 1943, 1944, and 1952. Examination of these data suggests that there have been minor changes in altitude prior to 1964; however, a large part of the change determined by the 1964 releveling must have occurred dur- ing and immediately after the earthquake. Changes in altitude of some stations, all on unconsoli- dated deposits, were obviously anomalous and undoubtedly repre- sented changes caused by local phenomena such as frost heaving, slumping, or thawing of perma- frost. These changes in altitude were not considered in evaluating regional uplift and subsidence. Recent regional changes in alti- tude in south-central Alaska are shown in figure 4, compiled pri- marily from releveling data sup- plemented with data from Grantz, Plafker, and Kachadoorian (1964, p. 4). Because of the lack of con— trol points, the location of the iso- base lines are only approximate. In spite of the sparsity of control points, it is obvious that a broad area, including a large segment of the Chugach Mountains, part of the Talkeetna Mountains, most of the Copper River Basin, and the Cook Inlet Lowland, did subside, and that large areas to the south and north of the subsided area were uplifted. The subsided area forms an arcuate trough sloping gently to the west and to the south- west. The maximum determined amount of subsidence was about 6 feet near Portage. The available data indicate that theluplifted area north of the trough forms a gentle arch having a maximum deter- mined uplift of almost 1 foot. Uplift to the south of the trough was much greater and was re- ported to be at least 33 feet on land (Plafker, 1965, p. 1679). The fact that the configuration of the subsided and uplifted areas generally parallels the trend of the major elements of the region (see fig. 3) suggests a genetic relation- ship. The southern two-thirds of the area of study is within the east- ern part of the trough, and the northern one-third of the area cov- ers a part of the southern edge of the uplifted arch north of the trough (fig. 4). CHANGES IN GROUND- AND SURFACE-WATER CONDI- TIONS GROUND WATER The water level in several wells in the area' lowered appreciably, and the water from many wells became turbid. Generally within a few days after the earthquake, the water level returned to normal and the suspended sediment set- tled out. Three wells in the Glen— nallen area were reported to have gone dry after the earthquake. Water from a spring on the north side of the Glenn Highway near mile 140 was reported to have been turbid and salty to the taste after the earthquake, but within a few days the condition of the wa- ter returned to normal. Ground water near the surface caused an increase in the intensity and duration of ground motion which, in turn, caused ground breakage. In many areas the ground water, along with fine- grained sediments, was ejected to the surface through cracks and was sprayed into the air. These phe- nomena are called “earthquake fountains” and have been reported in many areas during major earth- quakes, for example: the Charles- ton earthquake of 1886 (Dutton, 1904, p. 296—298) ; the California earthquake of 1906 (State Earth— quake Investigation Commission, 1908, p. 402—404) ; the Pleasant Valley, Nev., earthquake of 1915 (Jones, 1915, p. 196); the Japa- nese earthquake of 1923 (Ima- mura, 1937, p. 74) ; the Bihar- Nepal earthquake of 1934 (India Geological Survey officers and Roy, 1939, p. 33—34, 185—187) ; and the Alaska earthquake of 1958 (Davis and Sanders, 1960, p. 248— 250). SURFACE WATER Soon after the earthquake, news- papers reported that the Copper River was dry in its lower reaches and that Tazlina Lake had EFFECTS IN THE COPPER RIVER BASIN AREA 146° 144° E9 -—O %h “L . ”3‘23“" I ‘ «flaur/ w I" " :m La 1 1 "£1 HARPER Nb '6 5/5 / 50,60 KECHUMSTUK mu- 5720-M1 VET / 5300~ > 3.. 1': 1 100 MILES FranHlinO< i Fortymile 64° A ~‘ 62° , 142' Base from U.S. Geological Survey_ 1:2 500 000 4.——Areas of recent regional uplift (hachured) and subsidence (stippled). dashed where control is poor. Isobase lines are solid where control is good and Northernmost 0-isobase line represents approximate limit of detectable change in altitude. E10 ALASKA EARTHQUAKE, MARCH 27, 1964 v. ten/m 5.—View, to the northeast, of the northern end of Tonsina Lake showing intensely fractured ice and large‘blocks of ice stranded above water level along the shore. The lake is about half a mile Wide at this point. drained. Examinations by per- sonnel of the US Geological Sur- vey and the Alaska Department of Highways, several days after the earthquake, produced no evidence that the Copper River was, or had been, completely dammed. Water was flowing in the main channels of the Copper River and the water in Tazlina Lake was near its nor- mal level for that time of the year. Large waves were generated on Tazlina, Klutina, and Tonsina Lakes by sublacustrine landslides off the fronts of deltas. These waves stranded large blocks of ice above water level along the shores (fig. 5), and at the outlet of Taz— lina Lake, sediment and large blocks of ice were carried over the rapids and a short distance down the Tazlina River. A large snowslide near mile 50, Richardson Highway, dammed the Tiekel River, and ice jams near miles 74 and 78.5, Richardson Highway, dammed the Tonsina River. In both places, explosives were used to blast channels through the obstructions to pre- vent the river from flooding the highway. CRACKING OF LAKE AND RIVER ICE The earthquake caused the ice on most lakes to crack, especially around the margins of the lakes where the floating ice broke free from the ice frozen to the shore. In addition to the breakage along the shores of the lakes, a few linear cracks commonly formed at a tan- gent to the shorelines. In many places these linear cracks extended completely across the lakes. Tazlina, Klutina, and Tonsina Lakes were major exceptions to Photograph by U.S. Army, April 8, 1964. this general pattern. Ice on Klu- tina Lake and Tonsina Lake was highly fractured throughout (fig. 5). On Tazlina Lake the ice on the northern half of the lake was highly fractured (fig. 6), but ice on the southern half was fractured only around the margin, and a few linear cracks were formed across the lake. The greater intensity of the cracking of ice on these lakes was caused by water waves gener- ated by sublacustrine landsliding off the fronts of large deltas built out into the lakes. River ice cracked lecally throughout the area of investiga- tion. At several places along the Nelchina, Sanford, and Copper Rivers the cracks formed a sys- tematic reticulate pattern, easily recognized on aerial photographs. EFFECTS IN THE COPPER RIVER BASIN AREA E11 6.——View, to the east, showing intensely fractured ice along the eastern shore of Tazlina Lake, across from the large Nelchina River delta. stranded above water level along the shore are easily discernible, but the intense fracturing of ice away from the shore is obscured by the refreezing of the lake water and by a postearthquake snowfall. Large blocks of ice The segment of shoreline shown in this photograph is about half a mile long. Photograph by US. Army, April 8, 1964. GROUND CRACKS AND AS- SOCIATED LANDSLIDING TERMINOLOGY The terminology for ground cracks caused by earthquakes has been inconsistently used in the literature. The terms “fissure,” “fracture,” “furrow,” and “ground crack” have been used more or less synonymously. However, on the basis of priority the terms “fis— sure” and “fracture” should be re- served for two types of ground cracks of different origin. Old- ham (1899, p. 86) proposed the use of the term “fissure” for ground cracks formed by forces acting at the surface, and the term “frac- ture” for ground cracks formed by forces acting at depth, such as movement along a buried fault. In many places it would obviously be difficult to distinguish between the two types of ground cracks. 2211—693 0—66——3 For this reason and because of other inherent difliculties with this classification, I have used the gen- eral term “ground crack” which has no specific generic connotation. GENERAL DISTRIBUTION In the area of study, the ma— jority of the ground cracks oc— curred within a radius of 100 miles from the epicenter of the earth- quake. Ground cracks formed in the flood plains of rivers, in deltas, and along the toes of alluvial fans. They also occurred locally in low terraces adjacent to flood plains, in highway and other fill material, along the margins of lakes, along the face of steep slopes of river blufl's and hillsides, and in areas cleared of vegetation. The overall distribution pattern of ground cracks within the area of study indicates that the cracks were not localized in regional linear zones (fig. 7), and thus sug- gests that they were not caused by fault movement in bedrock beneath the unconsolidated d e p o s i t s . Rather, the distribution pattern indicates that local geologic factors controlled the distribution of the cracks. Nevertheless, the possi- bility that fault movement at depth did cause ground cracks to form locally can not be definitely eliminated. GEOLOGIC CONTROLS The ground cracks were re- stricted to areas underlain by un- consolidated deposits (see fig. 2) where one or more of the follow- ing conditions existed: (1) perma- frost was absent or deep lying, (2) the ground-water table was near the surface, (3) bedrock was rela- tively deep lying, and (4) slopes were steep. Because the earth- quake occurred in March, there If; Zmi. «L t >w ALASKA EARTHQUAKE, MARCH 27 , 1964 146° 4 4,. Rouncljop ”Wm » V Base from Anchorage, World (N. America) 1:1 000 000 scale, compiled by Army Map Service , ‘\\\ TA 25' WA , LAKE ( 2; Mame,” Mo 1‘ , 3 > / ~ _ 3 not-n A . EXPLANATlON 0 Area Where ground cracks were observed, July 1964 o 10 20 30 40 MILES l I | l l 7 .—Distribution of ground cracks in the Copper River Basin area. , xi? ‘ fobblckes x = ,9‘ l m. , M EFFECTS IN THE COPPER RIVER BASIN AREA E13 was a variable thickness of sea— sonal frost throughout the area. Most ground cracks occurred in coarse-grained unconsolidated de- posits, but some occurred in fine- grained deposits, generally in proximity to steep slopes, along the shores of lakes, or in areas cleared of vegetation for several years. Where well-drained unconsoli- dated deposits were thin, over— lying bedrock at shallow depths, grounds cracks generally did not form. Permafrost which extends from near the surface to depths as great as 200 feet is widespread in the Copper River Basin area, but gen- erally was absent in areas under- lain by coarse-grained deposits in which ground cracks formed (see figs. 2, 7). Permafrost generally was absent also near large deep lakes, and consequently, ground cracks occurred locally along the shores. In areas where ground cracks formed in fine-grained de- posits, the impervious permafrost table generally was from 10 to 20 feet below the surface; in fiat-ly- ing areas this situation permitted a water-saturated zone to exist be— tween the permafrost table and the base of the seasonal frost at the surface. The effect of permafrost on the distribution of ground cracks be- comes apparent if one considers that a thick layer of ice-rich per— mafrost behaves more like bedrock than like unconsolidated deposits and that, according to Gutenberg (1956, 1957), the intensity and duration of ground motion at the same distance from the epicenter may be 10 times as great in water- saturated unconsolidated deposits as in crystalline bedrock. Nat— urally, other conditions being equal, ground breakage will be most severe in areas receiving the greatest intensity of ground mo- tion for the longest period of time. When the earthquake occurred in the latter part of March, sea- sonal frost, which was near its annual maximum penetration, formed a thin, brittle layer extend- ing from the surface of the ground to variable depths, depending upon 1 o c a 1 conditions. The ground motion, where it was in- tense, broke this brittle layer, and thereby formed the ground cracks. At many places along slopes (un- confined faces), the seasonal frost probably prevented ground break- age, whereas in flat-lying areas un- derlain by permafrost-free water- saturated materials, the seasonal frost facilitated ground breakage. In most areas where severe ground breakage occurred, the ground—water table was no more than 15 feet from the surface, and generally just a few feet. In some places, large concentrations of va- dose water were perched on top of the impervious permafrost. In areas where the ground-water ta— ble was within a few feet of the surface, permafrost generally was absent; in areas where vadose wa- ter was perched on top of perma- frost, the permafrost table was rel- atively deep lying. Under both conditions the unconsolidated de- posits beneath the seasonal frost were saturated with water, and consequently, were. subjected to greater intensity and duration of ground motion than unconsoli- dated deposits perennially frozen or those not saturated with water. In many parts of the area, ground cracks formed on or near a steep slope, which provided an unconfined face for the lateral ex- tension of materials. These ground cracks generally paralleled the slope and occurred either on the face of the slope or behind it. Landforms with relatively steep slopes, along which ground cracks occurred, include hillsides, lake- shores, river bluffs, and terrace escarpments. LITTLE TONSINA RIVER SITE The Little Tonsina River land- slide site (fig. 8, next page) is in the southern part of the area, on the southwestern side of the Richardson Highway at mile 65. Landsliding occurred along the northwestern side of a small hill underlain by unconsolidated silt, sand, and gravel deposits. Lateral movement of segments of the hill toward the northwest caused sev- eral large ground cracks to form on and parallel to the face of the hill and pressure ridges to develop on the floor of the Little Tonsina River valley (fig. 9). One of the large ground cracks formed by the lateral extension of materials to- ward an unconfined face (hillside) is shown in figure 10, and a pres- sure ridge developed in front of the slide area is shown in figure 11. The horizontal forces that formed the ridge were transmitted through a layer of seasonally frozen peat and silty sand, approx- imately 2 feet thick, overlying water-saturated silty sand. Lo- cally, water-saturated silty sand was ejected to the surface through cracks in the seasonally frozen layer. OTHER LANDSLIDE SITES Other landslides include a large one at the mouth of Klutina Lake which caused the Klutina River to be diverted, several small ones along bluffs of the major rivers of the area, and numerous small slides marginal to lakes (see fig. 7). Ground cracks formed along margins of lakes by lateral move- ment of materials. An example is shown in figure 12 (p. 16). E14 ALASKA EARTHQUAKE, MARCH 27, 1964 Q ‘7 2’ s . \fb IR 0&0 “'6’ ”2672' \\\ ’1/ \\ \\ \\\\ s . st ‘ M||e 65 . \ a6}. \ I ’ 1, I Q . l ’5— Pressure rldges II I ‘1. I I /\Ground 7* I / cracks I HILL 0 500 FEET l____! 9.—View, to the northwest, showing pressure ridges on floor of Little Tonsina River valley near mile 65, Richardson Highway; pressure ridges (indicated by arrows) caused by landsliding on face of hill. EFFECTS IN THE COPPER RIVER BASIN AREA E15 10.——One of several large ground cracks paralleling northwest side of small hill near mile 65, Richardson Highway. Crack, formed in fine— to medium-grained sand, has maximum width of 3 feet and depth of about 5 feet. The lateral exten- sion of materials down the side of the hill caused the ground cracks and also caused pressure ridges to form on level ground several hundred feet in front of the base of hills 11.—-Pressure ridge, 6 feet high, on flat surface bordering Little Tonsina River’ near mile 65, Richardson Highway. Crest of ridge broken irregularly by upward flexing of frozen peat caused by landsliding on a nearby hillside. E16 ALASKA EARTHQUAKE, MARCH 27, 1964 12.—Large ground crack formed in clearing along south shore of lake about two- tenths of a mile north of mile 158, Glenn Highway. Maximum Width of crack is 4 feet and depth is 6 feet. This ground crack is one of several that generally parallel the lakeshore. EFFECTS IN THE COPPER RIVER BASIN AREA E17 13.—View, to the east, of the large braided flood plain of the Nelchina River 8 miles north of the terminus of the Nelchina Glacier. NELCHINA RIVER OUTWASH APRON The Nelchina River outwash apron, in the southwestern part of the area, is underlain by silty sand and gravel, is approximately 10 miles in length, and averages about 2 miles in width (fig. 13). The unvegetated parts of this outwash apron were literally crisscrossed with ground cracks that generally ranged from 1 to 2 feet in width and extended for great distances across the flood plain (fig. 14, next page). Measurement of the rather systematic reticulate pattern of ground breakage at several points suggested a preferred orientation of N. 35° W. and N. 50° E. Silty sand, locally including pebbles, and large quantities of water were ejected from many of the cracks during the earthquake. One of the most significant characteristics of this site was the absence of any unconfined face toward which the materials could move, and conse- quently, form ground cracks. An explanation for the origin of ground cracks in such flat—lying areas is given below on page E23. Extensive ground breakage, similar to that in the Nelchina River outwash apron, also oc— curred in the flood plains of the lower parts of the Chitina and Copper Rivers, and, locally, in flood plains of other streams (see fig. 7) . NELCHINA RIVER DELTA The Nelchina River delta (figs. 15—17, p. 19—21) is along the northwestern side of Tazlina Lake in the southwestern part of the area. The delta is approximately 21/2 miles wide at its Widest point and generally is underlain by silty sand and gravel. Undoubtedly, finer grained deltaic deposits are present at depth. All of these de- posits are free of permafrost, and the ground-water table is Within a few feet of the surface. Extensive ground breakage in the delta in- cluded a concentration of cracks along, and generally parallel to, a hingeline that crossed the delta. This hingeline is conspicuous in figure 15 because most of the snow cover on the downslope side has been removed by water and sedif ment ejected from ground cracks. Figure 16 shows ground cracks in the central part of the hingeline, and figure 17 shows cracks at the northern end of the delta. E18 ALASKA EARTHQUAKE, MARCH 27, 1964 14,—Typical ground crack in outwash *apron of Nelchina River 2 miles north of terminus of Nelchina Glacier; crack about 2 feet Wide at Widest point. Note how crack was deflected slightly in crossing swale in foreground. Standing water in swale marks the shallow ground-water table. Dark line behind man is trace of intersecting crack. EFFECTS IN THE COPPER RIVER BASIN AREA E19 ( ecove'red) ‘ 15,—The Nelchina River delta. Hingeline across the delta is conspicuous because-most of the snow has been removed from the (lownslope side by water and sediment ejected from ground cracks. The position and character of the water-delta interface or shoreline in 1948 compared with the postearthquake shoreline indicate that large segments of the delta front slid into the lake. Photograph by U.S. Coast and Geodetic Survey, April 10, 1964. E20 ALASKA EARTHQUAKE, MARCH 27, 1964 \\, 16.——Ground cracks along and paralleling hingeline on Nelchina River delta. Photograph by US. Geological Survey, October 1964. EFFECTS IN THE COPPER RIVER BASIN AREA E21 The available evidence suggests that the delta surface subsided— the greatest amount of subsidence occurring at the front of the delta and progressively lesser amounts upslope from the front. This dif- ferential subsidence resulted in a broad cambering of the delta sur— face which caused the concentra- tion of tension cracks along the hingeline. The position and character of the water-delta interface in 1948 as compared with the position and scalloped character of the post— earthquake interface (see fig. 15) indicate that not only was there subsidence but that large segments of the delta front slid into the lake. This sublacustrine landsliding generated large waves that se- verely fractured the ice on the . northern part of the lake and 17.—North end of the large Nelchina River delta along the western shore of Tazlina stranded large bIOCkS 0f Ice above Lake. Note the reticulate pattern of ground cracks which are as much as 6 feet water level along the shore. Large wide. quantities of sediment and ice were discharged through the outlet of the lake into the channel of Taz- lina River. Similar sublacustrine landsliding occurred off the fronts of deltas in Klutina Lake and in Tonsina Lake. The flooded delta of Kiana Creek along the eastern shore of Tazlina Lake is shown in figure 18. Subsidence, probably caused by compaction of unconsolidated sediments, and lateral movement of materials toward the lakeshore caused flooding of the delta and the formation of numerous ground cracks. A large area along the eastern shore of Tazlina Lake just south of the Kiana Creek delta was inundated (fig. 19). This area, which is part of an old alluvial- fan delta of Kiana Creek formed when the creek was a much larger stream, was subjected to changes 18.—View showing flooding of Kiana Creek delta along eastern shore of Tazlina similar to those that occurred at Lake- the Kiana Creek delta. E22 ALASKA EARTHQUAKE, MARCH 27, 1964 19.—Large inundated area along eastern shore of Tazlina Lake just south of the Kiana Creek delta. Lateral movement and vertical compaction of materials along lakeshore caused numerous ground cracks to form and the area to subside. Note normal beach (indicated by arrow) in the distance. EFFECTS IN THE COPPER RIVER BASIN AREA E23 LATERAL EXTENSION ——> <—— HORIZONTAL ‘COMPACTION DIFFERENTIAL VERTICAL COMPACTION Bedrock or other firm material 20.—Diagrammatic sections showing mechanism of formation Of basic types of ground cracks in unconsolidated deposits in the Copper River Basin area. ORIGIN As already explained, the ground cracks within the area of study were formed in unconsoli- dated deposits under a variety of geologic conditions. Despite this diversity of local conditions, the origin of most of the ground cracks can be explained by the fol- lowing mechanisms: (1) lateral extension, caused by materials moving toward an unconfined face such as a lakeshore, river bluff, hillside, or terrace escarpment; (2) horizontal compaction, caused by repeated alternate compression and dilation (in the horizontal direction) of materials in flat- lying areas where there are no un- confined faces; (3) differential vertical compactiOn, caused by the shaking Of materials that vary lat- erally in thickness or character; and (4) combinations of the above. Figure 20 shows the mechanism of formation of the three basic types of ground cracks. LATERAL EXTENSION Lateral-extension ground cracks occurred on and behind relatively steep slopes (fig. 20, section A). These steep slopes provided an unconfined face toward which ma- terials moved laterally, and in some places the movement was great enough to be considered landsliding. Generally the cracks formed parallel to the slope or unconfined face. HORIZONTAL COMPACTION The term “horizontal compac- tion” is used in this report for the earthquake-induced mechanism that causes unconsolidated mate- rials to lose volume in the horizon— E24 tal direction, and consequently, causes ground cracks to form (fig. 20, section B ) . In flat-lying areas where there are no unconfined faces toward which the materials can move laterally, the very pres- ence of numerous ground cracks, that extend for great distances indicates that the materials have lost volume in the horizontal direction. In March when the earthquake occurred, seasonal frost had pene— trated the surficial materials to a depth of several feet and formed a hard brittle surface layer. Be- cause of the extremely cold weath— er in the area, the frozen surficial materials were in a state of tension. This frozen layer probably was initially cracked by ground waves flexing the surface. The earthquake-induced ground waves must have subjected the water-saturated sediments beneath the seasonal frost to repeated alter: nate compression and dilation in the horizontal direction, the net result of these forces being hori- zontal compaction. Large quanti- ties of water and silt- and sand- sized material were ejected from the cracks, and sediment particles were rearranged so that they occu- pied less space. The high mobil- ity of the fine-grained materials ejected from the cracks suggests that they were liquefied sponta— neously by the ground motion. Terzaghi (1950, p. 100) describes spontaneous liquefaction as the transformation of fine sand or coarse silt into a liquid state. Similar mechanisms of forma— tion have been postulated for earthquake—induced ground cracks in South Carolina (Dutton, 1889, p. 267-268), in India (Oldham, 1899, p. 88—92), in south-central United States (Fuller, 1912, p. ALASKA EARTHQUAKE, MARCH 27, 1964 57), and in Japan (Imamura, 1937, p. 75). DIFFERENTIAL VERTICAL COMPACTION Without detailed data it is diffi- cult to determine the degree to which differential vertical com- paction contributed to the forma— tion of ground cracks throughout the area; nevertheless, in certain areas this mechanism appears to have been dominant in their for- mation. In a typical bedrock-confined valley underlain by unconsoli- dated deposits, the materials are much thinner along the margins of the valley wall than in the cen- ter of the valley. Where this con- dition exists and where the de- posits are susceptible to vertical compaction, the materials in the center of the valley subside more than the materials along the valley walls because they are thicker. Consequently, ground cracks form generally near and parallel to the valley wall. Ground cracks such as these were observed along the eastern margin of the outwash apron of Tazlina Glacier and lo- cally along the eastern margin of the outwash apron of Nelchina Glacier. Lateral variation in the texture and in void ratios of the unconsolidated deposits also can cause differential compaction. The broad cambering of the sur- face of the Nelchina River delta (fig. 15) and the concentration of ground cracks along and parallel to the hingeline (fig. 16) can be ex- plained by differential vertical compaction-caused by lateral vari— ation in the thickness of unconsoli— dated deposits. The wedge-shaped body of del— taic deposits was laid down over the lip of a valley wall, and the thicker deposits toward the front of the delta were compacted more than those toward the apex. The hingeline formed above and paral- lel to the lip of the buried valley walk—under conditions similar to those depicted in figure 20, sec— tion 0. SNOWSLIDES, ROCKSLIDES, AND AVALANCHES Within the area, snowslides, rockslides, and avalanches were re- stricted to the mountains sur- rounding the Copper River Basin. They were especially numerous in the Chugach Mountains which are closest to the epicenter of the earth- quake. Many glaciers emanate from an extensive icefield in the higher reaches of these rugged mountains. The remoteness of this vast mountainous region made it im- practical to make a comprehensive study of the slides. However, re- connaissance flights over the area, reports of slides along the Rich- ardson Highway where it crosses the Chugach Mountains, and study of postearthquake aerial photographs of the area indicate that slide activity was much greater than that in. past years when no earthquakes occurred. The large amount of snow and rock debris that has cascaded onto the icefield and glaciers, and, prob— ably even more important, the overall disturbance to the icefield will affect the regimen of the gla— ciers emanating from it. Accord- ing to Tarr and Martin (1912, p. 51—61), several glaciers in the Yakutat area advanced abnor- mally after the Yakutat earth- quake of 1899 because‘of earth- quake-induced changes in the regimen of the glaciers. If their hypothesis is true, it is not unrea— sonable to expect that some of the glaciers in the Chugach Mountains will advance an abnormal amount within the next few years. EFFECTS IN THE COPPER RIVER BASIN AREA E25 EFFECTS OF THE EARTHQUAKE ON MANMADE STRUCTURES DAMAGE TO BUILDINGS AND ASSOCIATED STRUC- TURES The greatest amount of damage to buildings, which were mostly of wood—frame or log construction, occurred within the southern half of the area, primarily because of proximity to the epicenter of the earthquake. Within this part of the area, breakage of fragile items inside buildings was widespread. However, major structural dam- age to buildings was restricted to localities where the intensity and duration of ground motion was greatest. The fellowing descriptions of damage to buildings and associ- ated structures at various sites Within the area were selected to give a representative picture of the range of damage: Sheep Mountain Inn, at mile 113, Glenn Highway, is on the toe of an alluvial fan at the base of Sheep Mountain. Several ground cracks, oriented in a general east- vVest direction andnparalleling the toe of the fan, formed in the cleared area where the buildings were located. The ground move- ment caused severe damage to un— derground pipes, a small cabin was shaken off its foundation, and other buildings shifted position. A diesel engine was shaken off a concrete slab to which it was bolted. Tazlina Glacier Lodge, at mile 156, Glenn Highway, had only minor damage. There was'consid- erable breakage inside, however, and most loose objects shifted position. Lee’s Guide Service, a short dis- tance north of the Glenn Highway at mile 158, was severly damaged. Several large crescent-shaped ground cracks formed in a cleared area adjacent and parallel to a lake. This area is underlain by gravelly silt. A small building was partly submerged in water when the ground upon which it was erected moved laterally to- ward the lake. When the ground cracks opened, water was ejected to the surface and flowed down the gentle slope to the lake. Glennallen, whose center is at mile 186, Glenn Highway, near the southern junction of the Glenn and Richardson Highways, is the largest community in the area. Breakage of fragile items inside buildings was widespread, but structural damage to buildings generally was not severe. A 5-unit motel, a barracks-type building, and a trailer house were shaken off their foundations. The foundation of the elementary school building also was severely damaged. S e v e r a 1 relatively small ground cracks formed in cleared areas and locally damaged structures. At the Copper Valley Electric Co. station, damage to the plant foundation, flooring, and equipment caused a power disrup- tion for a little more than 4 hours while repairs were being made and generating equipment was being checked. The Glennallen Road Camp of the Alaska Department of Highways, the largest installa- tion in Glennallen, sustained dam- age to underground sewers, steam and water lines, well casings, win- dows, and a boiler. Gulkana Airfield, at mile 120, Richardson Highway, received only minor damage, even to fragile items inside buildings. A few small cracks formed in a paved runway and in roads, and a water pipe broke under one house. Chistochina Lodge, at mile 238, Glenn Highway, had no signifi- cant damage; even glasses on a shelf did not tip over. Tsina Lodge, at mile 34.5, Rich- ardson Highway, in the southern part of the area of study had no major structural damage. How— ever, a few fragile items inside the building were broken, and many loose items were moved. _ Tonsina Lodge, at mlle 79, Rich- ardson Highway, had minor struc— tural damage; a fireplace and 2 chinmeys were cracked severely, windows were broken, walls cracked, and some difl'erential set- tlement occurred. Numerous fragile items inside the building were broken. DAMAGE TO HIGHWAYS AND BRIDGES Along the Glenn Highway from the Matanuska Glacier at the west- ern edge of the area of study to the junction of the Glenn and Richardson H i g h w a y s near Glennallen, several small cracks formed in the pavement, and at a few places minor slumping of roadcuts occurred. The majority of these cracks were less than 6 inches wide, and no major difi'eren- tial movement took place. From the junction north along the Glenn Highway, only a few small cracks formed in the pavementwnone that required major repair. E26 No damage was reported along the Lake Louise Road which ex- tends from about mile 160 on the Glenn Highway to the southern end of Lake Louise, a distance of 19 miles. Along the Richardson High- way, however, serious road dam- age occurred at several places be— tween mile 27 near the southern edge of the area of study and mile 91 near the junction of the Rich- ardson and Edgerton Highways. A large crack formed in the road— way at mile 28 near Worthington Glacier, and severe road damage occurred along a 22-mile segment between mile 64, 1 mile south of the Little Tonsina River bridge, and mile 86 near the Rock Creek bridge. Cracks as wide as 8 inches formed in the pavement, and 30- to 50-foot segments of the road- bed were moved differentially (fig. 21). This differential movement . was caused by .landsliding which occurred in unconsolidated depos- its on steep slopes near the high- way. The Little Tonsina River bridge also was damaged primar- ily by this landsliding (fig. 22). At mile 79.8 the slump of a long segment of the roadbed on a side- hill cut restricted vehicles to one- way traflic. At mile 86.1 just north of Rock Creek a small crack formed in an abandoned section of road, and therefore, the feature was well preserved. Because of the granular nature of the under- lying materials and because of the high ground-water table, the earthquake-induced ground mo- tion caused silt and fine sand, along with water, to be ejected from the crack onto the road sur- face (fig. 23). Only a few small cracks formed in the pavement and road fill along the Richardson Highway north of mile 91. ALASKA EARTHQUAKE, MARCH 27, 1964 . 21.—View, to the southeast, showing the pavement surface of the Richardson Highway just south of the Little Tonsina River bridge at mile 65. Note the 8-inch horizontal offset of the centerline; vertical displacement is as much as 3 inches. 22.—Lands1ide movement on a hillside approximately two-tenths of a mile away caused the Little Tonsina River bridge to be shoved from the right to the left. Minor slumping occurred in road fill to the left of the bridge. Immediately after the earthquake the pavement was buckled more than 3 feet high at the left end of the bridge. EFFECTS IN THE COPPER RIVER BASIN AREA E27 23.—Sma11 crack across abandoned section of road on a low gravel terrace adjacent to Rock Creek at mile 86.1, Richardson Highway. Note the fine sand and silt along both sides of the crack. Algermissen, S. T., 1965, Prince William Sound, Alaska earthquake of March 28, 1964, and aftershock sequence [abs]: Geol. Soc. America Spec. Paper 82, p. 2. Andreason, G. E., Grantz, Arthur, Zietz, Isidore, and Barnes, D. F., 1964, Geologic interpretation of magnetic and gravity data in the Copper River Basin, Alaska: U.S. Geol. Survey Prof. Paper 316—11, p. 135- 153, pls. 23, 24. Berg, Eduard, 1964, The Alaskan earth- quake of March 1964—a review: Alaska Univ., Geophys. Inst. Ann. Rept. 1963—64, p. 69—82. Ooulter, H. W., and Coulter, E. B., 1962, Preliminary geologic map of the Valdez-Tiekel belt, Alaska: U.S. Geol. Survey Misc. Geol. Inv. Map I—356, scale 1: 96,000. REFERENCES Davis, T. N., and Sanders, N. K., Alaska earthquake of July 10, 1958—Inten- sity distribution and field investiga— tion of northern epicentral region: Seismol. Soc. America Bull., v. 50, no. 2, p. 221—252. Dutton, C. E., 1889, The Charleston earthquake of August 31, 1886 : US Geological Survey 9th Ann. Rept, p. 203—528. 1904, Earthquakes in the light of the new seismology: New York, G. P. Putnam’s Sons, 314 p. Eiby, G. A., 1957, About earthquakes: New York, Harper & Bros, 168 p. Ferrians, O. J., Jr., 1965, Permafrost map of Alaska: US Geol. Survey Misc. Geol. Inv. Map I—445, scale 1 : 2,500,000. Fuller, M. L., 1912, The New Madrid earthquake: U.S. Geol. Survey Bull. 494, 119 p. Snowslides covered the Rich- ardson Highway near mile 38, mile 39, and mile 42, and a rockslide blocked the road near mile 44; a large snowslide near mile 50 dammed the Tiekel River, and be- cause of the danger of flooding the highway, explosives were used to blast a channel in the slide. Ice jams on the Tonsina River near mile 75 and mile 78.5 caused the river to rise and to erode highway fill near mile 76 and mile 77; ex- plosives were used to remedy this condition also. Between mile 66 and mile 73, accelerated surface drainage from the steep mountain slope paralleling the highway on the east caused large icings to form on the highway. The Edgerton Highway was damaged very little, except for a 200—foot landslide on the Lower Tonsina Hill, a short distance north of Lower Tonsina. The Lower Tonsina River bridge, which was shifted about 6 inches to the southeast, sustained consider- able structural damage. Grantz, Arthur, Plafker, George, and Kachadoorian, Reuben, 1964, Alaska’s Good Friday earthquake, March 27, 1964, a preliminary geologic evaluation : U.S. Geol. Sur- vey Circ. 491, 35 p. Gutenberg, Beno, 1956, Effects of ground on shaking in earthquakes: Am. Geophys. Union Trans, v. 37, no. 6, p. 757—760. 1957, Eflects of ground on earth— quake motion: Seismol. Soc. Amer- ica Bull., v. 47, no. 3, p. 221—250. Gutenberg, B., and Benihff, H., 1939, Waves and currents recorded by electromagnetic barographs: AM. Meteorol. Soc. Bull., v. 20, p. 421— 426. Imamura, Akitune, 1937, Theoretical and applied seismology: Tokyo, Maruzen 00., 358 p. E28 India Geological Survey Oflicers, and Roy, S. 0., 1939, The Bihar-Nepal earthquake of 1934: India Geol. Survey Mem., v. 73, 391 1). Jones, J. 0., 1915, The Pleasant Valley, Nevada, earthquake of October 2, 1915: Seismol. Soc. America Bu11., v. 5, p. 190—205. Karlstrom, T. N. V., Goulter, H. W., Fernald, A. T., Williams, J. R., Hop- kins, D. M., Péwé, T. L., Drewes, Harold, Muller, E. H., and Oondon, W. H., 1964, Surflcial geology of Alaska: US. Geol. Survey Misc. Geoil. Inv. Map I—357, scale 1: 1,584,000. Kingdom-Ward, F., 1951, Notes on the Assam earthquake: Nature, V. 167, p..130—131. 1953, The Assam earthquake of 1950: Geog. J0ur., v. 119, pt. 2, p. 169—182. Knott, C. G., 1908, The physics of earth- quake phenomena: Oxford, Claren- don Press, 283 p. ALASKA EARTHQUAKE, MARCH 27, 1964 Leet, L. D., and Leet, Florence, 1964, Earthquake—discoveries in seis- mology: New York, Dell Publishing 00., 224 D. Miller, D. J., Payne, T. G., and Gryc, George, 1959, Geology of possible petroleum provinces in Alaska: US. Geol. Survey Bull. 1094, 131 p., 6 pls. Oldham, R. D., 1899, Report on the great earthquake of 12th June 1897: India Geol. Survey Mem., v. 49, 379 p. Plafker, George, 1965, Tectonic defor- mation associated with the 1964 Alaska earthquake: Science, v. 148, no. 3678, p. 1675—1687. Press, Frank, and Jackson, David, 1965, Alaska earthquake, 27 March .1964—vertica1 extent of faulting and elastic strain energy release: Science, v. 147, no. 3660, p. 867—868. Richter, C. F., 1958, Elementary seis- mology: San Francisco, Calif., W. H. Freeman and 00., 768 p. State Earthquake Investigation Com mission, 1908, The California earth- quake of April 18, 1906: Carnegie Inst. Washington, v. 1, 451 p. Tarr, R. S., and Martin, L.M., 1912, The earthquake of Yakutat Bay, Alas- ka, in September, 1899: US. Geol. Survey Prof. Paper 69, 135 p. Terzaghi, Karl, 1950, Mechanism of landslides, in Application of geol- ogy to engineering practice (Berkey volume) : New York, Geol. Soc. America, p. 83—123. Thompson, A., 1929, Earthquake sound‘s heard at great distances: Nature, v. 124, p. 687—688. US. Coast and Geodetic Survey, 1964, Prince William Sound, Alaskan earthquakes, March—April 1964: US. Coast and Geod. Survey, Seis- mology Div. Prelim. Rept, 83 p. U.S. GOVERNMENT PRINTING OFFICE: 1966 0—221—693 he Alaska Earthquake March 27, 1964 ReEiunal Efects Ground Breakage in the Cook Inlet Area GEOLOGICAL SUR, THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS Ground Breakage and Associated Effects in the Cook Inlet Area, Alaska, Resulting from the March 27, 1964, Earthquake By HELEN L. FOSTER and THOR N. V. KARLSTROM A description of the ground cracks and deposits from ground-water eruptions and crustal changes, particularly in the Kenai Lowland GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—F UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1967 For sale by the Superintendent of Documents, US. Government Printing Office Washington, DC. 20402 THE ALASKA EARTHQUAKE SERIES The U.S. Geological Survey is publishing the re— sults of investigations of the earthquake in a series of six Professional Papers. Professional Paper 543 describes the regional effects of the earthquake. Other Professional Papers describe the history of the field investigations and reconstruction effort; the efl'ects of the earthquake on communities; the effects on hydrology; and the efi'ects on transportation, communications, and utilities. Page Abstract _____________________ F1 Introduction __________________ 1 Scope of report and source of data _____________________ 1 Acknowledgments ___________ 2 Regional geologic setting _____ 2 Geologic history and local geo- logic setting ______________ 2 Ground breakage ______________ 3 Types of breakage ___________ 3 CONTENTS Ground breakage—Con. Kenai Lowland ______________ Northeast-trending zone__ __ Coasts ___________________ Large lakes _______________ Southern interior __________ Northern Cook Inlet area- _ _ _ Northwest side of Cook Inlet_ Southeast side of Kachemak Bay ______________________ ILLUSTRATIONS PLATES [Plates are in pocket] Page F3 3 12 14 17 19 23 24 1. Map showing ground breakage resulting from the 1964 Alaska earthquake in the Kenai Lowland. 2. Map of Cook Inlet area. 3. Map of disruption on outwash delta of Tustumena Lake. Page 1. Index map _________________ v1 2. Azimuth-frequency diagram showing dominant trend of ground cracks in the Kenai Lowland _________________ F4 3, 4. Sketch maps of ground cracks in Kenai Low- land __________________ 5, 6 5—17. Photographs: 5. Tree split by ground crack ___________ 7 6. Typical large ground crack in Kenai Lowland _________ 8 7. Vertical displace- ment ___________ 9 8. Sand ridges depos- ited along ground crack ___________ 9 9. Large fissure _______ 10 10. Large ground crack with vertical dis- placement _______ 10 11. Microrelief features formed in ex- truded silt _______ 11 FIGURES 5—17. Photographs—Continued 12. Small vent which extruded watery silt _____________ 13. Detrital coal ex- truded with sand- 14. Ground cracks at Kasilof __________ 15. Collapse pits at Kasilof__-_______ 16. Outwash delta at head of Tustu- mena Lake a few days after the earthquake ______ 17. Crack between delta front and bedrock at Tustumena Lake ___________ 18. Diagram showingtype of ground disruption on the north shore of Tustumena Lake ______ 19. Diagram showing a type of slumping on south shore of Tustumena Lake _______ 20. Sketch map of Eklutna val- ley _____________________ Page F11 13 13 14 15 15 16 17 18 Page Causes of ground breakage _____ F24 Crustal changes in the Cook In- let region _________________ 25 Observed coastal changes _____ 25 Tilting of lake basins ________ 26 Pattern of subsidence ________ 27 Conclusions ___________________ 28 References ____________________ 28 Page 21—23. Photographs of Eklutna valley: 21. Sites of avalanche activity ________ F19 22. Dust cloud _______ 19 23. A crack in gravel outwash ________ 20 24. Photograph of failure of the delta front at Troublesome Creek, Lake George___ _ _ _ 21 25. Sketch map of Lake George area ____________________ 21 26—27. Photographs of Lake George: 26. Wall of silt punch- ed up on the emerged bottom _ 22 27. Silt squeezed up along cracks in the emerged bot- tom sediments- _ 22 28. Photograph of postearth— quake high-tide mark along Turnagain Arm sea blufis near Hope ______________ 25 V 164° 160° 156° 152° 148° 144° 140° 6A fl, —‘ \g/.\\ | . // \ ‘ 62,. 2/ /:3 :~» ‘é\ (, // 1‘\ a / 0 // l \ /_ \ H‘ an I w ‘ x \ am 4. ALASKA // /’o(” TALKEETN'A COPPER ‘\’K\ ‘\ 15> 62‘ ,’ // é l\MOLJNTAINS RIVER I/WRANGELR Z I / £3 \ ' \\ BASIN ’ 7 :16 I v “9— ,- .’ ’Q\ :\MOUNTAII~,S“> ‘ , (94- ]{Ov‘ ’ CHUGACH \ \~/“, 1 E3 ‘ l 6? A MOUNTAINS \“\ ~/ y. P5 nchorage ~\ 3’ / g \ a / o . J \ m ‘h \(M ’ ve/ \ 60‘ l 97 «3‘ ‘ o" -\ .. v-‘fi/ / /r SILK)" V " 476$ . ' EXPLANATION ‘GULF OF ‘ ALASKA * L u . . “ '=3 68 Eplcenter of March 27, 1964, Alaska earth- . quake 1 *6 Active volcano ____ 66b Fault Dashed where approximate I 5000 -— ' Submarine contour, in feet g Area of report ' ,0 30 60 MILES. 164° 160° 156° 152° 148° 144° 140°‘ 1.—~Index map showing location of Cook Inlet area. ‘ THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS GROUND BREAKAGE AND ASSOCIATED EFFECTS IN THE COOK INLET AREA, ALASKA, RESULTING FROM THE MARCH 27, 1964, EARTHQUAKE By Helen L. Foster and Thor N. V. Karlstrom The great 1964 Alaska earthquake caused considerable ground breakage in the Cook Inlet area of south-central Alaska. The breakage occurred largely in thick deposits of unconsolidated sedi- ments. The most important types of ground breakage were (1) fracturing- or cracking and the extrusion of sand and gravel with ground water along fractures in various types of landforms, and (2) slumping and lateral extension of unconfined faces, particularly along delta fronts. The principal concentration of ground breakage within the area cov- ered by this report was in a northeast- trending zone about 60 miles long and 6 miles wide in the northern part of the Kenai Lowland. The zone cut across diverse topography and stratigraphy. The great 1964 Alaska earth— quake oaused considerable ground breakage in the Cook Inlet area of south-central Alaska. Cook Inlet is the major marine reentrant of the Gulf of Alaska; it extends in- land between the Kenai and Chu- gach Mountains and the Alaska Range (fig. 1). It is bordered by extensive lowlands on the east, northwest, and north. The epicen- ter of the earthquake (fig. 1) was to the east and northeast of Cook Inlet and separated from it by ABSTRACT Cracks were as much as 30 feet across and 25 feet deep. Sand, gravel, and pieces of coal and lignite were extruded along many fissures. It is suggested that the disruption in this zone may be due to movement along a fault in the under- lying Tertiary rocks. The outwash deltas of Tustumena and Skilak Lakes in the Kenai Low- land, of Eklutna Lake and Lake George in the Chugach Mountains, of Bradley Lake in the Kenai Mountains, and at the outlet of upper Beluga Lake at the base of the Alaska Range showed much slumping, as did the delta of the Susitna River. Parts of the flood plains of the Skilak River, Fox River, and Eagle River were extensively cracked. A few avalanches and slumps oc- curred along the coast of Cook Inlet in scattered localities. Some tidal flats were cracked. However, in view of the many thick sections of unconsolidated sediments and the abundance of steep slopes, the cracking was perhaps less than might have been expected. Observations along the coasts indi- cated changes in sea level which, al- though caused partly by compaction of unconsolidated sediments, may largely be attributed to crustal deformation ac- companying the earthquake. Most of the Cook Inlet area was downwarped, although the northwest side of Cook Inlet may have been slightly unwarped. Maximum change in the Cook Inlet area was probably less than 6 feet. Little or no regional tilting was detected in the lake basins of Tustumena and Skilak Lakes. INTRODUCTION part of the Chugach Mountains and the Kenai Mountains. The greatest ground breakage outside of the Anchorage and Portage areas was in the Kenai Lowland, about 115 air miles west of the epicenter. SCOPE OF REPORT AND SOURCE OF DATA This report is concerned pri— marily with the ground breakage and related phenomena in the low- land areas of Cook Inlet—par- ticularly the more populated and accessible Kenai Lowland. De- scriptions of the damage in the Anchorage and Portage areas are available in previously published reports (Grantz and others, 1964; Hansen, 1965). Some observations of ground breakage in the north- western part of the Chugach Mountains are included here. Most of the information on earthquake effects presented in this report is based on field obser- vations made by Karlstrom F1 F2 between May 10 and July 15, 1964, and by Foster from May 10 to the end of May. Ground observations were made in the parts of the Kenai Lowland accessible by road and along beaches. Karlstrom made boat traverses along the shores of Tustumena Lake, Skilak Lake, and Lake George. The ground observations were supple- mented by reconnaissance from fixed-wing planes and by the study of aerial photographs made after the earthquake by the U.S. Army, the US. Coast and Geodetic Sur- vey, and the US. Geological Sur- vey. Localities near the head of Cook Inlet were checked primarily by road traverses. On the west side of Cook Inlet only one ground traverse was made near Tyonek. Other observations there were from fixed-wing planes. Ground breakage is difficult to detect from the air, especially in wooded areas. Consequently, some localities with ground breakage undoubtedly have been overlooked, particularly west of Cook Inlet. ACKNOWLEDGMENTS Special acknowledgment is due Mr. David M. Spencer, Mr. Robert Ward, Mr. Avery Thayer, and Mr. Will Troyer of the US. Fish and Wildlife Service in Kenai, Alaska, who provided essential logistic support and supplied personal ob- servations on changes in areas not visited by US. Geological Survey personnel. Thanks are also due Mr. Joe Magargl and Mr. George Calvin of Kasilof for their assist— ance during the boat traverse around the shores of Tustumena Lake, and to numerous members of the U.S. Geological Survey sta— tioned in Anchorge and Palmer, who supplied critical information based on personal observations of the earthquake effects. George Plafker, David McCulloch, Roger Waller, and Reuben Kachadoor- ALASKA EARTHQUAKE, MARCH 27, 1964 ian, all of the U.S. Geological Sur- vey, made their unpublished data available and gave assistance dur- ing the preparation of the manu- script. REGIONAL GEOLOGIC SETTING The Cook Inlet area is near the juncture of the western Pacific island arc system and the orogenic belts of the western part of North America. Just northeast of Cook Inlet the trends of the mountain ranges abruptly change from northwest to southwest and extend in broad subparallel arcs (fig. 1). Near the apices of these arcs the granite—cored Talkeetna Moun— tains separate the Chugach Moun— tains from the Alaska Range and form the divide between the head of Cook Inlet and the intermon- tane Copper River basin to the east. Several arcuate fault zones have been mapped in central and southern Alaska; these in part fol- low and in part transect the moun- tain-range structures. One of the faults, the Lake Clark-Castle Mountains fault, cuts the Alaska Range just north of Cook Inlet (Dutro, 1957), there having a northeasterly trend. A belt of high seismicity, coincident with the apical zone of the arcuate moun- tain structures, extends from Prince William Sound beyond Fairbanks north of the Alaska Range and includes the epicenter of the main shock of the 1964.- earthquake. GEOLOGIC HISTORY AND LOCAL GEOLOGIC SETTING The geologic structure of the Cook Inlet area is highly complex. The rocks of the region record a history of repeated geosynclinal sedimentation, deformation, and intrusion beginning in Paleozoic time and extending through the Tertiary. Near the end of the Cretaceous, there was downwarp- ing and subsidence of the Cook Inlet trough and, in the Tertiary, rocks that are now locally more than 15,000 feet thick were de- posited. By the end of Tertiary time, the major topographic ele- ments of the area were established. The subsequent geologic history has consisted largely of erosion and modification of the mountain- ous areas during repeated glacial- in'terglacial cycles and of partial filling of lowland areas and valleys with glacial drift and associated deposits (Karlstrom, 1964). Off- sets of surficial deposits, along pre- existing faults in bedrock, indicate continuing sporadic tectonism in the region through the Quater— nary and to the present. In the Cook Inlet area, it was primarily the thick unconsolidated deposits of Quaternary age that locally failed by fissuring, slumping, and subsidence during the 1964: Alaska earthquake. The Kenai Lowland part of the Cook Inlet area had the most ground breakage and therefore was examined in more detail; it is a broad low shelf 20 to 50 miles wide and 106 miles long. Most of the lowland is less than 400 feet above sea level, surfaces are flat to undulating, and local relief varies from a few feet to more than 200 feet. The Caribou Hills, a broad glaciated upland north of Homer, rise abruptly 1,000 to 2,000 feet above the general lowland surface. Remnants of this same upland sur— face occur as piedmont slopes adjacent to the Kenai Mountains between Skilak and Tustumena Lakes (Karlstrom, 1964, p. 12). Drainage in the Kenai Lowland is poorly integrated, and numer- ous lakes, marshes, and muskeg areas make up more than a third of the total surface. Two major lakes, Tustumena and Skilak, GROUND BREAKAGE, COOK INEUET AREA, ALASKA F3 occupy glacially scoured and mo- raine-dammed troughs and are drained respectively by the Kasilof and Kenai Rivers, which empty into Cook Inlet. Along the shore- line of Kenai Lowland are wave- cut cliffs that range in height from 800 feet in the Kachemak Bay area to less than 50 feet near the mouths of major drainage lines (Karl— strom, 1964, p. 12). Tertiary bed- rock is exposed locally beneath gla— cial deposits in the cliffs, especially along the northwest shore of TYPES OF BREAKAGE Ground failures in the Cook In— let area include most of the types that occurred elsewhere during the 1964 earthquake and that are de- scribed in detail from adjoining regions in other chapters of the US. Geological Survey 1964 Alaska earthquake series. Types of failure include (1) rock falls and avalanches resulting from dis- lodgement of bedrock, colluvium, and snow on steep rocky slopes; (2) landslides on moderate to gentle slopes with downslope mi— gration of surface materials giving rise to tensional cracking and fold- ing (pressure ridges) ; (3) slump- ing and lateral extension of sur- ficial deposits toward steep uncon- fined faces (river banks, lake shores, delta fronts, and sea blufl's) that cause tensional and rotational fracture patterns; (4) transverse and longitudinal tensional crack— ing of surface gravel in valley flood plains and valley trains; and (5) concentric and transverse cracking of frozen surface layers of muskeg and elevated tidal flats resulting largely from differential compac- 262—870 0—67—2 Kachemak Bay and along the Cook Inlet shoreline south of Kasilof. A complex of Quaternary mo- raines and associated drift and out- wash deposits of several ages cov- ers the greater part of the Kenai Lowland (pl. 1). Locally, the de- posits are more than 400 feet thick. The moraines have been modified to different degrees by erosion, by an irregular cover of loess, and by mantling with proglacial lake sedi— ments. Other proglacial lake sedi- GROUND BREAKAGE tion and downslope movements of underlying water-saturated fine- grained deposits. There was also extensive ground breakage and extrusion of sand and gravel along fractures which cut across the topography and so seem, at least in part, unrelated to it. KENAI LOWLAND N ORTHEAST-TRENDING ZONE Most localities of major ground breakage in the Kenai Lowland are roughly alined northeast- southwest in a zone 60 miles long and 6 miles wide which extends from Kasilof on the east shore of Cook Inlet to Chickaloon Bay on the south shore of Turnagain Arm (pls. 1, 2). The zone of ground breakage crosses an area of diverse land- forms underlain by Quaternary surficial deposits of differing origin and age (Karl‘strom, 1964). Included are elevated tidal flats of Recent age, end moraines of Knik (preclassical Wisconsin) age near Kasilof and Chickaloon Bay, and an interlobate moraine of Eklutna (Illinoian?) age in the central ments underlie terraced and chan— neled surfaces between major mor- ainal belts on coastal lowlands and in mountain valleys. The laws- trine sediments range in thickness from a few inches to at least 100 feet. Many of the streams have built large deltas; other smaller streams have deposited sediments to form alluvial fans and alluvial fan deltas. Thick and extensive elevated tidal-flat and beach de- posits occur in places along the coasts (pl. 1), particularly near Kenai and Kasilof. part of the zone (pl. 1). The mo- raines are mantled by a variable thickness of proglacial lake silt, sand, and gravel of Naptowne (classical Wisconsin) age. The proglacial lake deposits are thin- ner and coarser on the flanks of the moraine and become thicker and finer grained in the intermo- raine areas of the lowland. Inter- bedded glacial, glaciolacustrine, and glaciofluvial deposits, gener- ally hundreds of feet thick, under- lie the surface drift units and over- lie bedrock of Tertiary age. The intensity of ground crack— ing and ground—water eruption varies within the zone. The areas of broken ground are separated by large areas in which surface cracks are inconspicuous or absent. No consistent relationship is apparent between ground cracking and either the local topography or the underlying stratigraphy. Within the zone the ground—cracked lo- calities occur (1) on terraced mo- rainal slopes and crests underlain by thick sections of compact till mantled by thin sand and gravel deposits, (2) on channeled plains bordering moraines and underlain F4 ALASKA EARTHQUAKE, MARCH 27, 1964 Number of measurements 2.—Azimuth-frequency diagram showing dominant trend of ground cracks in the Kenai Lowland re— sulting from the 1964 Alaska earthquake. Includes measurements of 139 cracks in the central part of the northeast-trending zone of ground breakage in the Kenai Lowland. by thick sections of predominantly fine- to medium-grained sand, (3) in intermoraine depressions un- derlain by sand and silt, and (4) 0n elevated tidal flats underlain by eStuarine silt and sand. The cracks cut undeflected across small hills and depressions in morainal topography. The broken ground is character- ized by a mosaic of ground cracks that trend in many directions. L0- cally the cracks show topographic control by slumping on slopes, but in many places no such control is evident. In all localities observed on the ground, the northeast- trending cracks were the most per- sistent and in many places could be traced diagonally across slopes and ridges without apparent de- flections in trend. Measurements of the most conspicuous crack di- rections in the southern half of the zone suggest preferred frac- ture trends of N. 50°—55° E. (par- allel to the zone itself), N. 20°— 25° 13., N. 0°—5° W., N. 10°—15° W., N. 30°—35° W., N. 50°—55° W., and N. 75°—80° W. (fig. 2). Ex- amples of broken ground patterns within the zone are shown in fig- ures 3 and 4. In forested areas, cracking of the ground split nu- merous trees (fig. 5). Most of the cracks within the zone ranged in width from knife edge to a foot (fig. 6). They were generally very sharp, vertical, and straight with abrupt angular changes in direction and sharp angular intersections with other cracks (fig. 7). Major cracks had many smaller branch cracks that intersected the main cracks at sharp acute angles or sometimes at right angles. In places, major cracks abruptly changed direction at crack intersections. Vertical and horizontal displacements occurred along some cracks; most were less than a foot but a few were as much as 2 feet (fig. 7). Most displace- ments appeared to be primarily the result of adjustments in the sur- ficial materials by slumping and differential compaction, or of re— moval of underlying materials by ground-water eruption. F e a t u r e s associated with ground-water eruptions were com— mon throughout the zone. These features included ridges and mounds of sand alined along cracks (fig. 8) ; sand sheets as much as 2 feet thick, locally cov- ering acres of ground associated with the cracks (fig. 9) ;vents from which erupted sand either spread in all directions or was extruded primarily in one direction—and made a sand apron extending from the vent; collapse pits, as much as 30 feet in diameter; and fissures as much as 20 feet wide and 25 feet or more deep (figs. 9, 10), result— ing from the removal of material by copious outpouring of ground water. In a few places great out- pourings of sand were followed by small extrusions of watery silt which formed low (2 to 3 inches high) silt ridges and mounds and irregular microrelief features on the surface of the sand (figs. 11, 12). These silt extrusions must have occurred in the very last stages of the ground-water erup- tions and after the main shaking GROUND BRE’AKAGE., COOK INLET AREA, ALASKA 150° 52’40” 60°35’55” LAKE /\/| APPROXIMATE SCALE 100 200 FEET EXPLANATION Terraced and channeled sand plain Ground—water eruptive deposits of sand and silt with pebbles of elastic coal and lignite Trends of main ground crack sets Margin of well-drained ground 63) Crater-let which erupted ground water and sand B Letter indicates location of photograph 3.—Ground cracks (10c. 1, pl. 1), Kenai Lowland. F6 ALASKA EARTHQUAKE, MARCH 27, 1964 150“49’20” 60°39’ EXPLANATION Forested, terraced, and channeled moraines under- lain by variable thickness of stratified silt, sand, and gravel over till Muskeg underlain by 6—20 feet of peat and organic silt Ground-water eruptive deposits of sand and silt with pebbles of elastic coal and lignite D U Fault U, upthrown side; D, dowmhrow’n side O 50 100 FEET /\/\/\— M Open ground cracks 1-16 inches wide, 2—6 feet deep; locally displacements of 1—16 inches form graben (G) and horsts (H) Thin ground cracks cutting forest turf and peat deposits @239 Pressure ridges of forested turf and peat formed by landsliding Collapse pits and fissures associated with erupted sand deposits Open ground cracks with associated eruption of sand and ground water —— 250' Topographic contour B Letter indicates location of photograph 4.—Ground cracks (10c. 2, pl. 1), Kenai Lowland. GROUND BRE‘AKAGE‘, COIO-K INME’I‘ AREA, ALASKA 5.—Tree split by ground cracking (100. 2, pl. 1), Kenai Lowland. The ground crack which trends into the split in the tree (100. A, fig. 4) cuts across the top of a small knoll in moraine. F7 F8 ALASKA EARTHQUAKE, MARCH 27, 1964 (L—Typical large ground crack (Ice. 1, pl. 1), Kenai Lowland. Crack was 4 inches to 11A; feet wide; it cut a sand plain. Sand was not extruded along this crack (100. A, fig. 3). GROUND BRE‘AKAGE, COOK INLFET ARE‘A, ALASKA F9 8.——-Sand ridges deposited along a ground crack (100. 2. pl. 1) by eruption of ground water during the earthquake. A sand ridge on the right side of the picture intersects the main ridge at nearly right angles. Person has right leg in deep crack (loo. 0, fig. 4). 7 .—Vertical displacement (loo. 2, pl. 1). A scarp more than 200 feet long and 2 to 3 feet high was produced in a thick section of frozen peat and organic silt (10c. B, fig. 4). Crack at the base of the scarp was 6 inches to 1 foot Wide. An- other crack intersects scarp at right angles; this displacement probably re- sulted from diflerential compaction and downslope extension of silty deposits in a muskeg bordering a small lake. 9.———Large fissure at locality 2 (pl. 1). This fissure was 25 or more feet deep and as Wide as 20 feet in places. A thickness of 2 or more feet of sand which was extruded from the fissure can be seen around the tree. Photo- graph was taken at locality D (fig. 4). 10.——Large ground crack with vertical displacement (loo. 2, pl. 1). Thick de- posits of sand border the left side of the crack. Photograph was taken near E (fig. 4). GROUND BRE'AKAGE, COOK INEUEfl‘ AREA, ALASKA F11 11.—Microrelief features formed in ex- truded silt (10c. 2, pl. 1). Area pictured is about 48 inches wide) 3-inch-long jackknife is slightly above center of picture). The delicate silt mounds and ridges were formed in the last stages of ground-water eruption on the sur- face of previous sand extrusions. 12.—Small vent which extruded watery silt (loo. 2, pl. 1). J ackknife is 3 inches long. Near locality D (fig. 4), watery silt from small vents such as this was extruded in the last stages of eruption and spread out over the top of previ- ously erupted sand. 262—870 0—67—3 F12 had ceased, otherwise the small delicate microrelief features would not have been so well preserved. Although the greatest outpourings of sand were in relatively flat areas or depressions, some sand was ex- truded from cracks extending up the sides of small morainal hills. The deposits laid down by ground-water eruption within the zone are predominantly fine- to medium-grained sand. In most 10- calities the sand contains scattered pebbles and cobbles of detrital coal and lignite as much as 6 inches long (fig. 13). The largest amounts of sand and much of the coal were commonly concentrated at the in- tersection of two or more major cracks or at the intersection of a wide and a narrow one. Many, if not all, of the coal fragments car— ried to the surface during the earthquate by ground water could have been derived from near-sur- face deposits that contain detrital coal originally derived from un— derlying Tertiary coal beds. How— ever, derivation of some of the coal from the Tertiary coal beds at depths of several hundred feet, though unlikely, is not excluded as a possibility by present evidence. The quantity of sand deposited locally indicates copious ground— water discharge during the earth— quake and suggests the tapping of large reservoirs. Sand and silt de- posited on tree branches indicate that water was ejected out of some cracks to a height of at least 20 feet. Ground-water supplies in the general area of the central part of the zone are obtained from gravel aquifers under hydrostatic head. The aquifers are 100 to 200 feet below the surface and beneath the surface till units. It is probable that ground cracking penetrated through the surface till units at many places and tapped these deep ground-water reservoirs. ALASKA EARTHQUAKE, MARCH 27, 1964 In several, but not all, places where the northeast-trending zone crossed roads, cracks were found. The zone crossed the paved Ster- ling Highway east of Soldatna, and several cracks in the pavement were noted together with minor slumping in fill. Large cracks and large outpourings of sand occurred where the zone crossed the oil-well access road and several of the side roads connecting with it (pl. 1). Although this zone of ground breakage passed along the south- ern margin of the Swanson River oil field, no ground breakage was noted or reported from the field itself. Apparently the oil wells were not damaged or significantly disturbed by the earthquake. In the spring-fed Finger Lakes (about one-half mile southwest of loo. 2, pl. 1), there was a gradual lowering of water during the sum- mer following the earthquake, an occurrence which had not been ob- served in previous summers. It was suggested that the lowering might be due to fissures in the lake bot- toms or to a restriction of the flow of spring water as a result of the compaction of sediments (Alaska Dept. Fish and Game, 1965, p. 29). The lakes are in glacial moraine and the earthquake caused slump- ing on the sides of the moraines along the margins of the lakes. At several places within the zone, ancient collapse pits and linear troughs covered by mature forest vegetation record previous episodes of ground cracking and ground-water eruption. COASTS Much of the western coast of the Kenai Peninsula was exam— ined, and few changes attributable to the earthquake were found. Al— though steep high bluffs of uncon- solidated glacial deposits line much of the coast, there were no large landslides and very little slumping. At Kasilof, which is approxi— mately in line with the northeast— trending zone of ground breakage previously described (p. F3), cracks as much as 1 foot Wide were formed. The cracks were traceable for several hundred feet and cut sand dunes and beach deposits at the coast (fig. 14:). Crack walls were generally sharp——in fact, ex— ceptionally sharp for occurrence in loose sand. Some had small vertical displacements as much as 10 inches high. They had the usual pattern of many branches and cracks intersecting from several directions. Collapse occurred along some cracks where large amounts of sand had been extruded (fig. 15). Pressure ridges a few inches high and 2 to 3 inches across were noted in silt on tidal flats; some radiated from a central point in several directions for distances of 25 feet or more. At Chickaloon Bay a complex network of cracks extended sea- ward from the shore and perpen- dicular to it onto the tidal flats. Major crack systems also paral- leled the shore, but these were ob- served only from the air. Although the tidal flats at Kasilof and Chickaloon Bay were considerably fractured, other tidal flats in similar geologic and topographic situations had no ground breakage. For instance, no ground cracking was observed on the tidal flats at Kenai, Moose Point, Point Pos- session, Stariski, or in the Nin- ilchick area. On the mainland near Homer, ground fissures occurred in the bluff on which the U.S. Bureau of Land Management station is lo- cated. However, as elsewhere on the coast, ground breakage due to the earthquake was slight and less than might have been expected where steep slopes in materials GROUND BREAKAGE, COOK INLET AREA, ALASKA F13 14,—Ground cracks at Kasilof in ele- vated beach deposits. Cracks were as much as 1 foot Wide. Vertical displace- ment is evident along crack near man. Considerable elastic coal was brought up’from beneath the surface and de- posited With the sand. 13.——Detrita1 coal extruded with sand (loo. 1, pl. 1). COpiOus amounts of sand were extruded from a fissure cutting a gravel road (10c. B. fig. 3). Pieces of coal and lignite as long as 6 inches (some indicated by arrows) were ex- truded with the sand. F14 ALASKA EARTHQUAKE, MARCH 27, 15.—Collapse pits at Kasilof formed after eruptions of ground water and sand. Ptt in foreground is about 3 feet in diameter. Erupted sand covers the ground. A large crack extends from the pits in the middle background. susceptible to sliding abound (Waller, 1966, p. D6—D7) . Damage to Homer Spit, due largely to sub- marine sliding and flooding, has been described by Waller (1966). LARGE LAKES Tustumena Lake and Skilak Lake are the two largest lakes in the Kenai Lowland (pl. 1). Tustu- mena Lake is 25 miles long and 3 to 6 miles wide; Skilak Lake is 14 miles long and has a maximum width of about 4 miles. Both lakes occupy glacially scoured troughs that have been dammed by moraines. Ground cracking around these lakes was extensive in the large outwash deltas at the heads of the lakes. Damage elsewhere along the shores of the lakes was minor, al- though the lakes are surrounded primarily by thick unconsolidated deposits of Quaternary sediments. The several small occurrences of ground cracking which were ob- served along Tustumena and Skilak Lakes were mostly in un- stable slopes of alluvial fan deltas or at the tips of peninsulas or spits projecting into deep water. TUSTUMENA OUTWASH DELTA The outwash delta at the head of Tustumena Lake is 41/; miles wide; it partly buries two groups of ice-rounded bedrock hills that divide the outwash plain into two branches with separate delta fronts (pl. 3). During the earth- quake, both delta fronts failed by fracturing and slumping into the lake. Photographs obtained after the earthquake (fig. 16) indicate that a major zone of fissures oriented N. 5°—10° W. formed subparallel to the southern delta front about half a mile from the lake shore. This break defined the inland margin of the part of the delta most dis- rupted by the earthquake. Much sand was extruded along this frac- ture zone and flowed out upon the snow-covered surface (fig. 16). 1964 The fracture pattern in the dis— rupted part of the delta consisted of two sets of diagonal fractures formed on northwest and north— east trends and a less well de- veloped set formed about per- pendicular to the main fissured zone. It is believed that this frac- ture pattern formed during the earthquake when the lakeward un- confined face of the delta extended forward and pulled away from the main mass of the delta. McCulloch has described and explained sim- ilar crack patterns in Rocky Creek and Lakeview deltas at Kenai Lake (1966, p. A35). The ground cracks in the out- wash delta ranged in width from a fraction of an inch to 4 feet (fig. 17). Most cracks showed no verti— cal displacement, but a few indi- cated subsidence of several inches to 11/2 feet on the lake side. Little overall evidence of subsidence of the delta front was observed, and it is estimated to have been less than 10 feet. The southern edge of the out- wash delta is bordered by a low forested terrace (pl. 3, fig. 16). Cracks 3 to 4 feet wide and 2 to 3 feet deep spaced at intervals of 10 to 50 feet out through the forested area in a north-south direction along the lake edge. Many trees were split and ofi'set along cracks, the offset indicating a right-lateral ground displacement of 1 to 3 feet. The walls of the deeper cracks ex- posed the following terrace stra- tigraphy from top to bottom : vege- tation mat and organic silt about 1 foot thick over stratified pebble— to—boulder gravel containing lenses and beds of sand and silt. Sand, silt, and some gravel were brought to the surface by forceful ejection of ground water from the cracks. The extruded material spread out to form large sand sheets covering a few acres. Else- where sand ridges were deposited GROUND BRE’AKAGE, COOK INILE’I‘ AREA, ALASKA F15 16.—Outwash delta at head of Tu'stu- mena Lake a few days after the earth— quake. View is south (pl. 3). The ground was snow covered, and water and sand erupted along major cracks (dark-gray areas) partly removing the snow and covering the ground sur— face \Vi‘th sand. The cracked ice-covered lake surface is to the right. A major fissure system paralleling the delta front is seen to the left of center. Photograph by Avery Thayer, Fish and Wildlife Service, Kenai, Alaska. 17.—Crack between delta front and bedrock at Tustumena Lake (10c. A. pl. 3). A wedge of outwash deposits pulled away from the stable bedrock face during the earthquake shaking. F16 Old beach / // N Prequake /, ‘ T shoreline / / /// ALASKA EARTHQUAKE , Ground cracks APPROXl MATE SCALE 0 50 100 150 FEEY L__J_;J MARCH 27, 1964 Terrace gravel 18.—Diagram showing type of ground disruption on the north shore of Tusrtumena Lake (100. 3, pl. 1). along the margins of the cracks. Locally, boulders as much as 8 inches in long dimension were brought to the surface from the underlying gravel beds by the ejecting ground water. Such forceful ejection at this single locality may have been caused by conditions which dif- fered slightly from those in the delta to the north. Water moving through the outwash terrace under higher hydrostatic head (probably because the surface gradient of the terrace is steeper than that of the bordering outwash) may have been confined between more firmly frozen sediment layers. The firmer freezing could obtain because of the thicker organic silt layer, con- tinuous vegetation mat, and exten— sive forest cover. The difference in the pattern of cracking between the terrace and the outwash delta may reflect differences in the un— derlying bedrock topography. Be— cause of the general configuration of this valley, the bedrock slope under the terrace may be north- westward and that under the out- wash delta may be more to the west. Slippage of the lake front sediments over a northwest-slop— ing bedrock floor would explain the observed displacement. The northern delta front was fractured in a pattern similar to that of the southern front, but less severely. The diagonal and per— pendicular crack sets and the sub— parallel fissure zone separating the extended block from the inland part of the delta were present, but the patterns were not as strongly developed as in the southern front. A considerable amount of sand was extruded along many of the frac- tures. The fracturing was appar- ently caused by lateral spreading of the deltaic sediments toward the lake as a result of the earthquake shaking. OTHER TUSTUMENA LAKE GROUND CRACKING Local slumping at the fronts of alluvial-fan deltas occurred at the mouths of Indian Creek, Moose Creek, and Bear Creek (pl. 1). Slumping took place largely by v e r t i c a l displacement along ground cracks parallel to the lake margin. At Windy Point a block of beach and terrace gravel 50 by 100 feet slumped into the lake largely submerging spruce and birch trees 40 to 60 feet tall. Sub- parallel fractures 1 to 6 inches wide developed landward of the slump-block scarp. At the western end of Caribou Island the narrow tip of beach gravel cracked but did not slump. Ground cracking oc- curred over fairly large areas on shoreline terraces on opposite sides of the narrowest part of Tustu- mena Lake. Near Tanya Lake (loc. 3, pl. 1) , a northeast-trending set of ground cracks spaced 10 to 30 feet apart cut the beach deposits and a low forested gravel terrace. A graben 10 to 30 feet wide, about 100 feet long, and downdropped 4 to 6 feet formed and filled with water (fig. 18). Gravelly terrace deposits underlain by lake silt ap- parently broke into blocks along northeast fractures and slid lake- ward upon a slightly inclined slip plane in the silt (fig. 18). The for- ward translation and slumping of the lakeward blocks permitted the interior block to drop down. This mechanism is similar to that de- scribed for some of the landslides in the Anchorage area (Hansen, 1965). The submerged block with mature spruce trees indicated ver— tical slump of at least 20 feet at the shore. On the opposite shore of the lake (loo. 4, pl. 1), a section of a low forested terrace 500 feet long and 100 feet wide slumped toward the GROUND BREAKAGEI, COOK mm'r AREA, ALASKA F17 lake. The disrupted block is de— fined by an arcuate set of large fractures that intersect the lake margin at both ends; however, no displacement of the terrace is visi— ble where the fractures meet the lake shore. Tilted tree‘s occur on low ridges that appear to have formed by slight rotation along subordinate cracks in the block (fig. 19). Visible vertical displace- ments were slight, and modern beach deposits fronting the block were probably lowered only a few inches relative to the unaffected beaches to the north and south of the slumped area. SKILAK LAKE The entire outwash plain be- tween the front of Skilak Glacier and Skilak Lake was extensively fractured, and the delta front at the head of Skilak Lake slumped into the lake. The amount of slumping of the delta front below lake level is not known, but exami- nation of aerial photographs and ground observations suggest that a segment of the delta front about half a mile wide may have sub- sided sufficiently to be partially submerged during the low mean lake-level phase. Aspen and brush 10 to 20 feet high were standing in water 3 to 10 feet deep as much as 1,000 feet offshore from the new shore line. However, the fact that no freshly ice-scoured trees or other evidence of major water- Undisturbed terrace Disturbed terrace APPROXIMATE SCALE 0 25 50 75 FEET LL1_Ll_l—._J__._l level changes were observed any- where along the shores in the up- per part of the lake suggests that the amount of material that slumped from the delta front at the time of the earthquake was not large enough to generate large waves. Fracturing of the outwash plain drained a beaver—dammed lake along the south side of the valley about 2 miles from the head of the Skilak Lake (Avery Thayer, written commun, 1964). The beach fronting the Pipe Creek alluvial-fan delta (loo. 5, pl. 1) was cut by a system of inter— secting fractures one-half inch to 3 inches wide. The most con- spicuous cracks trended north or northeast, that is, transverse to rather than parallel to the shore. At the lower campground (loo. 6, pl. 1) near the outlet of Skilak Lake, ground cracks, mole tracks, and pressure ridges formed on a low forested terrace. The pressure ridges, 6 inches to 2 feet high, border the muskeg back of the campground and trend northwest- ward. The mole tracks developed by overthrusting (or underthrust- ing) of the forest litter layer; they are slightly offset along transect- ing cracks which trend northeast. This pattern suggests formation under compression, but the mech- anism is not clear. In the ab- sence of subsurface data, it can- not be determined whether the surface deformation resulted from Terrace gravel Beach sand and gravel Lake level 19.—Diagram showing a type of slumping on the south shore of T‘ustumena Lake (Ice. 4, pl. 1). stresses set up by slumping con- fined to the upper layers of the unconsolidated deposits or whether the deformation resulted from ac— commodation of the unconsoli- dated section to displacements at depth. The possibility of forma- tion by movement of the lake ice during the earthquake must also be considered. The north- to northeast-trending beaches at the lower end of the lake were fractured in two places. At locality 7 (pl. 1) the cracks are 3 inches to 1 foot wide and are con- fined to the modern beach gravel deposits. The largest crack was at the base of the scarp defining the inner edge of the beach. It was intersected by a few transverse cracks that extended diagonally across the beach toward the lake. At locality 8 (pl. 1) a north- to northeast-trending crack 5 to 15 inches wide cut diagonally across the beach and was traceable into higher ground a few hundred feet from the lake before being lost in dense forest. SOUTHERN INTERIOR Ground breakage in the interior of the Kenai Lowland outside of the northeast zone was limited to a few scattered localities (pl. 1) where the following types of dis- ruption were observed: 1. Landslides in unconsolidated deposits on moderate to steep slopes. Northeast of Homer. 2. Large avalanches. Anchor River. 3. Slumps in unconsolidated de- posits on or toward uncon- fined faces such as river banks, lake shores, and delta fronts. Chakok River, Fox River, Tustumena Lake, “Ski- lak Lake, and Stariski Creek. 4. Ground cracks in lowlands cov- ered by silty lake deposits, alluvium, or fluvioglacial de- F18 ALASKA EARTHQUAKE, MARCH 27 , 1964 Eklutmz 61°20'- APPROXIMATE SCALE 0 V2 1 MILE J i. 149°OO’ EXPLANATION Valley train Alluvial terrace and glacial moraine. Dark lobes are Recent mo- raines QR; Rock glacier or moraine Area of slope deposits fl Ground cracks produced by earthquake; asso- ciated locally with ground-water eruptive deposits (9 Collapse pit produced by the earthquake O. . Sites of unusually per- sistent avalanching 20.—Sketch map of Eklutna valley showing sites of persistent aval'anching and location of ground breakage. posits. Upland in headwaters of Bear Creek. 5. Ground cracks and pressure ridges in mu-skeg areas. Half a mile south of Clam Creek, Slikok Creek. 6. Cracks in flood plains and val- ley trains underlain by sand and gravel. Skilak River, F ox River. The cracking in the Fox River flood plain was particularly spec- tacular on the alluvial fan formed at the confluence between Sheep Creek and the Fox River. Copious amounts of sand erupted along an extensive mosaic of cracks. Along the channels of the Fox River, numerous alluvial deposits on slip-off slopes were broken along longitudinal and t r a n s v e r s e cracks. The alluvial flats on the in- side of meanders were cracked and channel banks slumped at nu- merous places. The cracking ap- peared to be concentrated along the axis of the Fox River valley. N0 flood-plain failures were ap— parent in tributaries heading in the Kenai Mountains. GROUND B-REIAKAGEZ, COOK mum AREA, ALASKA F19 NORTHERN COOK INLET AREA The major ground breakage in the northern Cook Inlet.area was in and around Anchorage and Portage. Ground cracks and land- slides were abundant and destruc- tive and have been described in detail by Hansen (1965), Mc- Culloch (1966), Kachadoorian (1965) and others. Ground break— age in this area also occurred in the Eklutna valley, around Lake George, in valleys in the western Chugach Mountains, and west of Anchorage including the delta of the Susitna River. At the head of Eklutna valley, unusual avalanche activity fol- lowed the earthquake. It began after the earthquake and con- tinued throughout the summer of 1964. Periodic observations made in the valley by U.S. Geological Survey personnel stationed in Anchorage indicate that it com- menced again after the spring thaw in 1965 and continued throughout that summer. As late as August 2, 1966, minor avalanche activity was reported (Ruth Schmidt, written commun., 1966). However, the total activity in the summer of 1966 was apparently considerably less than that in 1964, but more than before the earthquake. In the early part of the summer of 1964, rock debris cascaded nearly continually from three principal places on the steep high clifl's above Eklutna valley (fig. 20). Talus cones at the base of the cliffs were much enlarged (fig. 21). Dense dust clouds emanated from the avalanches and resulted in erroneous reports of volcanic eruptions when spotted by air- plane pilots (figs. 21, 22). Later in the summer, several of the ava- lanches became inactive, and by August 1965 activity was largely 21.—Sites of avalanche activity in Eklutna valley. These talus cones were present before the earthquake but have been enlarged by avalanche activity since the earthquake. The dust clouds are created by the avalanching. 22.——Dust cloud in Eklutna valley caused by avalanche activity. The avalanching has been persistent during the summers since the earthquake. 23.—Crack in gravel outwash in Eklutna valley. Crack was about 10 inches wide and split a tree stump. The two halves of the stump were offset about 14 inches. restricted to one avalanche chute at the head of the valley. Karl Gladys (oral commun., 1964) esti— mated that the normal avalanche activity, which usually lasts 2 to 3 weeks after spring thaw, con- stituted less than 2 percent of the avalanche activity during the sum- mer of 1964. The remainder is at- tributed to disruption caused by the earthquake along shear zones. Ground breakage was noted on the high terraces in the valley bottom below the northernmost avalanche source. One conspicuous set of cracks had a southeastward trend and crossed moraines, outwash, and colluvial deposits. An inter- secting set trends northeast. Ground breakage was also pre- sent on the floor of the valley at the head of Eklutna Lake. An eX~ tensive mosaic of cracks was present in the gravel outwash, and much sand and gravel erupted with the ground water during the earth- quake. Fissures were as wide as 11/2 feet. Horizontal offsets as much as 2 feet and vertical offsets of 1 foot were observed. Trees were split and offset along some cracks (fig. 23). Circular pits 41/2 feet in diameter and as much as 12 feet deep squirted sand and water (Karl Gladys, written commun., 1964). The main zone of cracking at the head of the lake covered an area of about 1 square mile. Addi- tional damage in the Eklutna area, particularly to the power-plant in- stallations at the lower end of Eklutna Lake, is described by Logan (1967). Within the Lake George basin of the Chugach Mountains 28 miles southeast of Palmer, damage resulting from the earthquake in- cluded failure along the fronts of deltas built at the mouths of two tributary valleys, and extensive cracking associated with eruptive deposits of emerged bottom sedi- ments at the head of the lake. The most extensive disruption oc- curred along the front of the Trou- blesome Creek delta where slump- ing caused lakeward rotational displacements of frozen gravel blocks as much as 20 feet wide and 100 feet long. Chaotic topography resulted (figs. 24, 25). The delta at the mouth of a creek entering the south end of upper Lake George failed by cracking along the margin of the delta front. Lateral extension, however, was slight, minor vertical or lateral displacements being evident along the cracks. A mosaic pattern of cracks de- veloped in the exposed lake-bottom sediments near the head of the lake (lake-bottom sediments were ex- posed at the time of the observa- tions because lake levels were low) . 24.—Failure of the delta front at Troublesome Creek, Lake George (loo. C, fig. 25). Man is standing on a large slumped and rotated block that is bounded front and back by large fissures. 61°15’ 148°30’ Troublesome Creek\ are 2 \g ( EXPLANATION Outwash Moraine of Recent age Emerged lake bottom sediments; line pattern indicates areas of severely cracked sedi- ments and eruptions of silt, sand, and water from cracks V Front of delta that failed during earthquake Earthquake cracks paralleling front of moraine J_L|_LL|_LL|_I_I_I_LJ.LLLLLLLLL Margin of bedrock areas A Letter indicates location of photograph 25.—Lake George area showing location of ground breakage. F22 ALASKA EARTHQUAKE, MARCH 27, 1964 26,—Wall of silt punched up on the emerged bottom of Lake George. The wall was 8 to 24 inches high, 30 feet long, and 6 inches Wide (100. B, fig. 25). 27.——Ridges of silt squeezed up along cracks in the emerged bottom sediments of Lake George (100. A. fig. 25). GROUND BREAKAGE, COOK LNTUET ARE'A, ALASKA F23 Lake silt was squeezed out of the underlying water-saturated sedi- ments and erupted through cracks in the seasonally frozen surface layer. At one locality a wall of frozen silt, 6 inches wide and about 30 feet long, was punched up 8 to 24 inches by pressures in the un— derlying unfrozen materials dur- ing the earthquake (fig. 26). The undisturbed structure of the fine laminated silts exposed in the wall indicated that the silt must have been rigid and frozen at the time of displacement. Disruption and minor dislocation of the wall along northeast-trending cracks indicate subsequent development of and minor displacements along cracks that cut transversely the northwest set of parallel fractures that de- fined the wall structure itself. At several locations, mounds of silt 1 to 2 feet high and 2 to 3 feet in diameter and showing internal flow structure formed at the inter- section of the cracks. Unfrozen silt was squeezed up as ridges along some of the cracks (fig. 27). On the north side of the moraine along the south side of lower Lake George, some sand eruption oc- curred at the contact of the mo- raine with the lake sediments through cracks transecting the moraine. Reconnaissance observations within the Chugach Mountains be- tween Knik Valley and Turnagain Arm revealed only two other local- ities of ground breakage: cracks cutting forest turf on thin allu- vium over bedrock near the mouth of south fork valley of Campbell Creek and cracks in thick muskeg and alluvial deposits in the middle part of Eagle River valley (pl. 2). The major crack set in Eagle River valley was roughly parallel to the axis of the valley, and the most conspicuous transverse set of cracks trended at right angles to the valley axis. No ground damage was observed upvalley or down- valley from these cracked zones. Ground observations were not made in the other valleys of the mountain range. West of Anchorage, tidal flats were cracked and slumping oc- curred at the mouth of Knik Arm 0n the west side. The delta of the Susitna R i v e r was severely cracked; parts of its front subsid- ed into Cook Inlet. In some small lakes in the Susitna River low- land, the sediments in the bottoms of the lakes were cracked. At the base of Mount Susitna along the projected trace of the Lake Clark- Castle Mountain fault, sand erup- tions occurred along the margins of lake terraces cut in till. To the northeast about 18 miles, also along the trace of the Lake Clark- Castle Mountain fault, there was disruption and slumping along a bluff face. NORTHWEST SIDE OF COOK INLET Limited observations on the northwest side of Cook Inlet indi- cate that ground breakage was relatively minor and confined mostly to deltas, tidal flats, and alluvial flats, and occurred in widely scattered localities. Ground breakage was noted in the follow- ing places (pl. 2) : 1. Beluga Lake. Beluga Lake is divided into upper and lower parts by deltaic outwash sediments deposited by drainage from Capps Gla- cier. This outwash delta was considerably fractured. At the head of upper Beluga Lake, terraces along outwash channels cut below the main surface level of the large outwash delta were locally fractured. At the head of lower Beluga Lake a large block slumped into the lake submerging trees. M in o r fracturing occurred at the outlet of lower Beluga Lake. 2. Inland from Trading Bay, 15 miles to the northeast. Ex- tensive fracturing occurred in a zone extending north- east more or less parallel to the shore of Trading Bay and along the approximate postulated trace of the Lake Clark—Castle M o u n t a i n fault. The zone is about 12 miles long. The dominant trend of the fractures is also northeast. Much sand was extruded along many of the fractures. 3. Northeast shore of Chakat- chama Lake. Cracks oc- curred in beach deposits along the shore of the lake (Gordon Giles, written com— mun., 1964). 4. Tyonek. Lake ice and muskeg were cracked. About 5 miles north of Tyonek, old slumps along the coast were reacti- vated, but displacement was less than 1 foot. 5. Trading Bay. On the coast 131/2 miles east of Kustatan, small slumps occurred in blufl's of glacial till (George Plafker, oral commun., 1966). 6. Redoubt Bay. Cracks were common along the Drift River from the mouth in- land 2 miles or more. All breakage was in alluvium. North from the mouth of Drift River, cracks associ- ated with extruded sand oc- curred along the coast for a distance of about 2 miles (George Plafker, oral com- mun., 1966). 7. Near the mouth of Tuxedni Channel. On the south side, cracks formed on steep slopes. Two months later, following spring thaw, land- slides occurred in the cracked terrain (George Plafker, oral commun., 1966). 8. North shore of Tuxedni Bay. Along the east side of Squarehead Cove, the bluff face sloughed off along the shore for about 1 mile (George Plafker, oral com- mun., 1966). 9. Kalgin Island. Slumps devel— oped in unconsolidated de— posits on the south end. 10. Augustine Island. Cracks more or less parallel to the coast were noted from the air on the north end and on the northeast side. All appeared to be in recent unconsoli- dated deposits of volcanic sand, gravel, and bouldery gravel. Some of the cracks ALASKA EARTHQUAKE, MARCH 27, 1964 were close to the shore, others were higher on the slope and subparallel to it. Slumps and small slides oc- curred in places along the cracks, and the surface was “stepped” downward toward the coast along some cracks (Bruce Reed and George Plafker, oral commun., 1966). SOUTHEAST SIDE OF KACHEMAK BAY Ground breakage at two places along the southeast side of Kache— mak Bay was noted by Plafker in a reconnaissance flight in 1964 (George Plafker, oral commun, 1966; pl. 2). Extensive fracturing resembling that of Skilak Lake occurred in the outwash delta at the head of Bradley Lake. Fine closely spaced cracks were noted in the tidal flats along the south shore of China Poot Bay. A rock- fall took place on a steep cliff nearby. The outwash flats along the glacial drainage which enters China Poot Bay from the south were considerably fissured. Ground breakage extended from the bay margin inland along the drain- age for about 2 miles. CAUSES OF GROUND BREAKAGE The northeast-trending zone of ground breakage in the Kenai Lowland may not be caused en— tirely by simple lateral movements or compaction in unconsolidated sediments. Several features indi- cate that additional factors were involved. For example, in many places there is no consistent re- lationship between the cracks and either the local topography or stratigraphy; the cracks are un- usually deep and are very per- sistent and have a marked linear trend. The following explanations for these features are suggested: (1) The disrupted zone may have re— sulted from movement triggered by the earthquake along a buried fault. (2) There may have been differential compaction of Qua- ternary sediments along a buried ridge of Tertiary rock. (3) Grav- ity sliding may have taken place along the interface between the unconsolidated Quaternary de- posits and the Tertiary bedrock. (4) A combination of these proc- esses could have been involved. The subsurface data available to the authors are insuflicient to define clearly the bedrock struc- tures beneath the Kenai Lowland, but some of the data suggest a buried fault; none preclude the possibility of such a fault. Field mapping in the Kenai Lowlands (Karlstrom, 1964) shows that the top of the Tertiary bedrock along the coast generally is above sea level south of, and below sea level north of, Kasilof. Thus there could be either an erosional break or a structural dislocation of the bed— rock floor where the inferred fault zone crosses the coast. In the Swanson River oil field, oil and gas have been produced from the Tertiary sedimentary rocks northwest of the zone of dis- rupted ground. To date, all wells drilled south of the zone have been dry. This change in subsurface conditions may be due to a fault, or it may indicate a facies change. By report, seismic profiles ob- tained from the northern part of the Kenai Lowland are difficult to interpret because of poor signal returns. They do suggest, however, an important subsurface disconti— nuity, either lithologic or struc- tural, that approximately coin— cides in position with the zone of broken ground developed during the 1964 earthquake. At one place along the trace of this disconti- nuity the seismic data can be inter- preted as recording a basement fault of northeast strike, a steep southeast dip, and a downdropped south block (C. E. Kirschner, oral commun, 1964). The unusual avalanche activity at the head of Eklutna valley fol- lowing the earthquake and con- tinuing into the summer of 1966 might also be related to move- ment along faults. The rocks in this area are highly sheared and crushed. If these rocks were made unstable by small movements along faults, the avalanche ac- tivity could have been set in mo- tion. Furthermore, the continued avalanche activity might be due to minor readjustments in the dis— turbed blocks following the main displacement. Although the bed- rock of this locality has not been mapped in detail, the fractured rock suggests that faults are pres- ent. If simple displacement of sheared blocks took place as a re- sult of the shaking of the earth- quake without faulting, it seems GROUND BREEAKAGE‘, COOK mum ARE’A, ALASKA F25 unlikely that the avalanching would have continued so long. The disruption in the valley floor at the head of Eklutna Lake and the damage to the powerplant facili- ties (Logan, 1967) also suggest that more than just the usual seis- mic efl'ects from a distant earth- quake were in operation here. On the northwest side of Cook Inlet northwest of Trading Bay, a concentration of cracks and sand extrusions in flood-plain deposits was noted near the projected trace of the Lake Clark—Castle Moun— tain fault. To the northeast of this locality along the trend of the fault, two other localities of dis- ruption were noted within the Cook Inlet area. This ground breakage may be indicative of movement along the Lake Clark- Castle Mountain fault, but other evidence of fault movement, par- ticularly in bedrock, has not been detected. The remainder of the ground breakage in the Cook Inlet area can be explained by the presence of thick sections of silt and water- logged sediments which are par- ticularly susceptible to disruption by seismic activity when in topo- graphically unfavorable situa- tions. Despite many such unfavor- able situations in the Kenai Low- land and the mountains bordering it, the overall ground breakage in this area was relatively minor. This stability may be explained, at least in part, by the fact that the ground was still firmly frozen at the time of the earthquake. CRUSTAL CHANGES IN THE COOK INLET AREA Crustal deformation that accom— panied the 19641 earthquake af- fected an area of 65,600 to 77,200 square miles in south-central Alaska. The deformation consisted of two major northeast-trending zones of uplift and subsidence be— tween the Aleutian Trench and the Aleutian volcanic arc (Plafker, 1965, p. 1686). Most of the Cook Inlet area is in the depressed area, but the region also includes areas ranging from those with no detect- able change to some on the north- west side of Cook Inlet that are slightly upwarped. However, maximum change in the Cook In- let area was small, probably not more than 6 feet, as compared with 50 feet to the southeast. OBSERVED COASTAL CHANGES Coastal observations in the Cook Inlet area indicate the following approximate changes in land-sea levels: 1. Along the south shore of Turn- ' again Arm near Hope, there was 4 to 6 feet of subsidence. Vegetated areas previously above storm beach levels were inundated during high tides following the earthquake. Houses in the low-lying parts of Hope were flooded. Inun— dation during high spring tides extended to just below the entrance of the General Store. High-water marks on forested slopes of a thick sec— tion of coarse alluvial fan east of Hope indicate inundations approximately 5 feet above the preearthquake high-tide mark (fig. 28). The coarse gravel substratum in the area and the absence of cracking or slumping along the coast suggest that most, if not all, of the recorded subsidence re- sulted from crustal subsid- ence and not from compac- tion or slumping of the sur- ficial deposits. 2. Along the south shore of Turn- again Arm in Chickaloon Bay, there was 3 to 4 feet of subsidence. Vegetated ele- vated tidal flats previously above the reach of high tide were inundated following the earthquake. The estimate of 3 to 4 feet of subsidence is based on observed changes in tidal levels in relation to the floor of a hunter’s cabin in 28.—Postearthquake high-tide mark along Turnagain Arm sea bluffs near Hope. Arrows indicate the pre- and postearthquake high-tide positions. The vertical dis tance of about 5 feet between these two positions approximates the amount of crustal subsidence that accompanied the 1964 earthquake. F26 the area that was for the first time partly inundated during the high tides (Avery Thayer, oral commun., 1965). How- ever, a thick section of fine— grained deposits underlies the tidal flats, so there is a possi— bility that part of the ob- served subsidence may have resulted from compaction. 3. Along the east coast of Cook Inlet from Point Possession to Anchor Point, observed subsidence ranged from a lit- tle less than a foot to none it all. Undercutting of the stabilized and vegetated col- luvial slopes along the sea bluffs was approximately the same as had been observed by Karlstrom during spring traverses in previous years. Vertical relations between storm beaches, elevated tidal- flat surfaces, and postearth- quake high-tide levels ap- peared comparable to those observed in previous years near Point Possession, Moose Point, Kenai, Kasilof, Ninil- chik, and Anchor Point. Lo— cal fishermen estimated a sub- sidence of about half a foot near Ninilchik on the basis of apparent changes in low tides. Careful observations on a pier gage at the cannery on the south side of the mouth of the Kenai River indicated that water levels during the spring high tides Ofollowing the earthquake did not quite at- tain the levels reached during the high tides of the preced— ing fall (Clarence Plat-t, oral commun, 1964). B e c a u s e high tides in this area are normally higher in fall than in spring, this observation is consistent with the geologic evidence of very little or no change in datum along the coast. A postearthquake re- ALASKA EARTHQUAKE, MARCH 27, 1964 survey of the highway from Kenai southwest to Ninilchik indicated no measurable change that could be attrib- uted to regional warping dur— ing the earthquake (Alaska Highway Dept. personnel, oral commun., 1964). Short- ‘term postearthquake tidal- gage records 5 u g g e s t a subsidence of 0.9 foot near Nikishka (Small, 1966, p. 20). 4. Along the south coast of the Kenai Lowland between An— chor Point and Homer Spit, regional tilting ranged from none to less than 1 foot of subsidence at Anchor Point, 2 feet at Homer, and 4 to 6 feet at the tip of Homer Spit. These figures are based on a rapid resurvey of the high- way af-ter the earthquake by the Alaska Highway Depart- ment. Of the 4 to 6 feet of subsidence at the end of Homer Spit—a low sand and gravel bar that extends 5 miles southeastward into Ka- chemak Bay from Homer—a minimum of 2 feet has been attributed to crustal subsid- ence and the remainder to differential compaction of the underlying unconsoli- d a t e (1 materials (Grantz, Plafker, and Kachadoorian, 1964, p. 9, 24). 5. Along the west shore of Cook Inlet from Point McKenzie to Kamishak Bay, there was 1 to 2 feet of subsidence along the slumped front of the Susitna Delta area ; no change to slightly down near Ty- onek; 1 to 2 feet of uplift along the coast of Kamishak Bay. The estimate of subsid— ence in the Susitna Delta area is based on observations of postearthquake tidal changes at a hunter’s cabin by person- nel of the Fish and Wildlife Service (Theron Smith, oral c0mmun., 1964). Probably most of this subsidence can be attributed to slumping and compaction of the delta front and adjoining elevated tidal flats between the Susitna River and McKenzie Point. This conclusion seems reason- able because (1) the coastal m a r gin was extensively cracked during the earth- quake and (2) changes in bathymetry of the bordering seaway (determined by US. Coast and Geodetic Surveys in the summers of 1963 and 1964) record a major slump of material along the coast, presumably resulting from the earthquake. No changes were observed along the coast near Tyonek except in one area where subsidence of storm beaches suggested a downdrop of about 1 foot. This area of obvious subsid- ence occurs at the foot of a section of coastal bluff where large preearthquake slump blocks moved locally during the quake. The anomalously low tide levels following the earthquake that were re- ported by residents Iliamna and Tuxedni Bays suggest an uplift along the Kamishak Bay coast of 1 to 2 feet. The estimates of land-sea changes along the south shore of Turnagain Arm are of the same order of magnitude as the changes determined by releveling bench marks in bedrock along the north shore of the arm. This agreement suggests that the estimates else- where in the area based on shore- line relations may also be approxi- mately correct. TILTING OF LAKE BASINS The large lake basins in the Cook Inlet area serve as natural GROUND BRE‘AKAGEi, cooK mmrr AREA, ALAISKA F27 tiltometers because their long di- rections are more or less parallel to the known direction of regional tilt. Regional tilting should be de- tectable in changes of water-plane levels. The positions of preearth- quake water levels were checked in a few of the lake basins by refer— ring to established bench marks. A preliminary resurvey of bench marks in relation to lake levels after the earthquake indicated southeastward tilting of Eklutna Lake ; the head of the lake was pos- sibly down 1 to 2 feet relative to its mouth (Russell Wayland, writ— ten commun., 1964). There was little or no tilting of Lake Chakat- chama (Gordon Giles, written commun., 1964); and little or no tilting of Bradley Lake near Homer (Marvin Slaughter, writ- ten commun., 1964). Air reconnaissance observations made of Tustumena Lake immedi- ately after the earthquake sug- gested southeastward tilting of the lak e basin. Outlet drainage through ‘the Kasilof River was re- ported to have decreased to such an extent that the day following the earthquake a galosh-shod biol- ogist was able to walk up its chan— nel (Alaska Dept. Fish and Game, 1965, p. 29). Forested areas at the lake head were reported to have been submerged. At the time of a 2-day boat traverse around the margins of the lake in June 1964, lake level was below the normal seasonal level (Joe Magargl, oral commun., 1964). It was about 5 feet below the average late autumn high-lake level, as marked by the inner edge of generally barren gravel beaches and by conspicuous high-water marks on rocky cliffs along the margin of the lake. Sub- merged trees were found to be as- sociated only with locally dis- turbed and slumped shoreline deposits; in no area could the sub- merged vegetation be related to regional tilting. Hand-leveling by Brunton compass at numerous lo- calities about the margin of the lake revealed no systematic differ- ences in the vertical interval be- tween lake level and high-water beaches and watermarks. Thus, if the lake basin had been tilted, the amount of tilting was apparently less than the accuracy of measure- ment by Brunton compass, which is estimated to be :1 foot. Observations made at the head of Tustumena Lake during the earthquake (Joe Secora, oral com- mun., 1964) indicate that the frozen surface of the lake cracked into a mosaic fracture pattern as the lake level rapidly rose and fell. Lake level receded about 2 feet and then rose; sloshing of the lake waters continued for about 2 hours. General lake level was judged to have been somewhat lower during the following day or two and then to have risen to about half a foot higher than the pre- earthquake level. Whether this rise resulted from belated tilting of the lake basin or from the nor- mal inflow during increased spring thaw is not clear. Mr. Seco‘ra’s account, however, is con- sistent with the later crude meas- urements on shoreline features Which suggest that tilting was probably less than 2 feet and con- ceivably less than 1 foot over a dis- tance of 25 miles between the mouth and head of the lake. Similar geologic observations and Brunton compass measure- ments made in June 1964 along the margins of Skilak Lake suggest the same general amount of reg- ional tilting during the earth- quake as on Tustumena Lake. Ex- tensive flooding of forests around the margins of Skilak Lake was observed by local pilots in Sep- tember 1964. A subsequent boat traverse around the lake margin by Avery Thayer (written com- mun., 1964) revealed unusually high lake levels but no changes that could be attributed to dif- ferential tilting. However, Kenai Lake, about 25 miles east of Skilak Lake in the Kenai Mountains, was in a zone of greater deformation and showed about 3 feet of tilt (McCulloch, 1966, p. A29). PATTERN OF SUBSIDENCE The combined geologic observa— tions and instrumental releveling data in the Cook Inlet area, to- gether with data from adjoining areas, indicates differential sub— sidence, maximum crustal depres- sion occurring along a linear zone between Portage and Aialik Bay or approximately coincident with the axis of the Kenai Mountains (Plafker, 1965). US. Coast and Geodetic Survey releveling indi- cates that maximum subsidence (nearly 6 feet) occurred near Portage and that the amount of subsidence diminishes progres- sively to about 2 feet northwest- ward towards Anchorage (US. Coast and Geodetic Survey, 1966, p. 122). These data indicate re- gional tilting to the southeast, ap- proximately at right angles to the axis of the Kenai Mountains. The gradient of regional tilt is about 0.1 foot per mile between Anchor Point and the Homer area, and in- creases to 0.2 foot per mile between Homer and Aialik Bay where the greatest subsidence (71/2 feet) measured in the region occurred (George Plafker, written com- mun., 1966). The releveling and geologic data are insufficient to ex— clude the possibility of some local vertical fault displacements, but, in general, subsidence appears to have resulted largely from down- warping or folding of the crustal rocks. The line of zero change separat— F28 ing the depressed area of Cook In— let from the more stable area to the north can be only approxi— mately located by the available Ground breakage in the Cook Inlet area occurred largely in un- consolidated deposits. It consisted of slumps toward unconfined faces such as river banks, delta fronts, and lake shores; ground cracks in lowland and valley bottom sur— faces; transverse and marginal cracks in modern and elevated tidal flats; and cracks in muskeg areas. Most of these breaks can be attributed to various processes and combinations of them such as dif- ferential compaction and down- slope movements of underlying water—saturated fine—grained sedi— ments. However, distribution of this type of damage was irregular, Alaska Department of Fish and Game, 1965, Post-earthquake fisheries eval- uation; an interim report on the March 1964 earthquake eflects on Alaska’s fishing resources: Juneau, Alaska, 72 p. Dutro, J. T., Jr., and Payne, T. G., 1957, Geologic map of Alaska: U.S. Geol. Survey, scale 1 : 2,500,000. Grantz, Arthur, Plafker, George, and Kachadoorian, Reuben, 1964, Alaska’s Good Friday earthquake, March 27, 1964, a preliminary geologic evalua- tion: U.S. Geol. Survey Circ. 491, 35 p. Hansen, W. R., 1965, Effects of the earthquake of March 27, 1964, at Anchorage, Alaska: U.S. Geol. Sur- vey Prof. Paper 542—A, p. A1—A68. Kachadoorian, Reuben, 1965, Effects of the earthquake of March 27, 1964, at Whittier, Alaska: U.S. Geol. Survey ALASKA EARTHQUAKE, MARCH 27, 1964 data. Apparently the line of zero change p a s s e s southwestward across the Susitna Lowlands, a few miles west of Tyonek «and into CONCLUSIONS and localities with seemingly simi- lar conditions were not necessarily similarly affected. Many areas, such as steep bluffs of unconsoli- dated materials along sea coast, were notable for their lack of visi- ble earthquake effects. Regional structural elements may have had some control over the location and type of damage incurred in the Cook Inlet area, particularly in the northeast-trending zone of ground breakage in the Kenai Lowland. Here, disruption in a thick se- quence of unCOnsolidated deposits may have been due to movement along a fault in the underlying Tertiary rocks. Faulting or adj ust- Cook Inlet along its western side (Thor Karlstrom, unpub. data, 1964; George Plafker, unpub. data, 1965). ments along previous faults also may have contributed to the ava— lanche activity in the Eklutna area and to the distribution of ground cracks along the west shore of Cook Inlet in line with. the Lake Clark-Castle Mountain fault. The distribution of ground breakage in the 1964 earthquake indicates that in the lowlands of the Cook Inlet area the most dam- age from future seismic activity can be expected on outwash deltas, lowland areas underlain by thick marine or lake silt, and in alluvium and outwash along the larger streams. REFERENCES CITED Prof. Paper 542—B, p. B1—B21. Karlstrom, T. N. V., 1964, Quaternary geology of the Kenai Lowland and glacial history of the Cook Inlet re- gion, Alaska: U.S. Geol. Survey Prof. Paper 443, 69 p. Logan, Malcolm, 1967, Effects of the March 27, 1964, Alaska earthquake on the Eklutna hydroelectric project, Anchorage, with a section on Televi- sion examination of earthquake dam- age to underground communication and electrical systems in Anchorage, by Lynne R. Barton: U.S. Geol. Sur- vey Prof. Paper 545—A, p. A1—A30. McCulloch, David S., 1966, Slide-in- duced waves, seiching and ground fracturing caused by the earthquake of March 27, 1964, at Kenai Lake, Alaska : U.S. Geol. Survey Prof. Paper 543—A, p. A1—A41. Plafker, George, 1965, Tectonic defor- mation associated with the 1964 Alaska earthquake: Science, v. 148, no. 3678, p. 1675—1687. Small, J. B., 1966, Alaskan surveys to determine crustal movement; Pt. 1, vertical bench mark displacement: U.S. Coast and Geodetic Survey, 24 p. U.S. Coast and Geodetic Survey, Wood, F. G., ed.-in-chief, 1966, The Prince William Sound, Alaska, earthquake of 1964 and aftershocks, VI: U.S. Coast and Geodetic Survey Publica- tion 10—3, 263 p. Waller, R. M., 1966, Effects of the earth- quake of March 27, 1964, in the Homer area, Alaska, with. a section on Beach changes on Homer Spit by Kirk M. Stanley: U.S. Geol. Survey Prof. Paper 542—D, p. D1—D28. ‘U.S. GOVERNMENT PRINTING OFFICE: 1967 0—262—870 UNITED STATES DEPARTMENT OF THE INTERIOR PROFESSIONAL PAPER SALE—F GEOLOGICAL SURVEY PLATE 1 R114 W. 30' 151°oo' R, 7 w. 30' 150°00' 152°oo' 61°00’ 61°00' R. IS W. ’4 o {4‘ a 4‘ ‘00 mom, EXPLANATION ’ ’ . . 1' ; ,ma ' . Scum //,// / fl / Moraine and associated drift deposits; commonly covered with 1—4 feet of silt (loess) Moraine and associated drift deposits mantled with variable thickness of lake deposits; mantle is pre- dominantly silt and sand with some gravel mprrs D Lil/{6" / > ,1 “/Ifm‘i‘] [111 ' ‘ as K . "e 1 hull/J7 ' A ' [II/mu ' 71/11» 11? // Lake deposits, thick; predominantly silt and sand with some gravel .. ‘ , ( _ 9’ . ; - , _ :, mun/g / ' > - ’ , 1 LII/f8 L} 0'4. 1,, z/V/ (r) 1151\101 ",1 0 (2:7 I j ‘, I'LL“ Fluvial deposits of sand and gravel; locally some silt. ” ' r " ' 1 ,, ' _ - - ' ' ‘ _ « , 1' , «LEE: C3 (Include floodplain, valley train, outwash, delta, ’ ' ' ' » ' ' " ' 1' and alluvial fan deposits) T ' Elevated tidal—flat and associated beach deposits; "* \ ,' «1 1' ' 1. i ' Ti 1 " 1 " ., ' " ,7,ng2 ' predommantly Sllt and sand; some gravel 1 1 . ~ 1 . ' , e 1 1 f _ . f , 1 , ~ $1M" 1/ x 77 \: / _ , _ .« \1 ' \1 4 {A (’2‘ I} + + + + + + + + _ ~ .. . . ~ . v t + + + 1 1 ‘ \,.. \ ' ' ‘ . 'W .' v’ " " ' ‘ , . . , 1;: ; ,,,,,,, + + + + 1‘ ‘ ‘1 - . ' ' + + + + Bedrock; hills and slopes discontinuously covered by glacial drift and alluvium TYPES OF GROUND BREAKAGE CAUSED '7 ml BY THE 1964 ALASKA EARTHQUAKE 1‘ Salama 7 ., p. . V , , , Ground cracks associated with eruptions of ground water and sand 0 Ground cracks I Cracks and (or) pressure ridges mostly in muskeg A T~ 5 » 1 Ground vcracks associated with slumping towards unconfined faces. (Includes mole tracks on the north side of the outlet of Skilak Lake) A Ground cracks with ground water and sand eruptions; ,1 , , . " " I" - _ - ' " 1; 1 ,_ ' ” H ”I" 1 associated with slumping towards unconfined faces ‘3' 1 ’ 3 ,,,,,, E] Ground cracks in sand and gravel of floodplains f or valley trains ,, ' 1 x 11 :1 . ’_ - ='~* . , ' . ~ ,. .. __‘\\fl 1 1 , . _ T‘4 N' ' 1" ' 1 C W V H v ' " -' . v ' . ' Y’Vinflpafls‘w /~ ( '_ 5 '_ _ ‘ ‘4 N. * Landslides 0r avalanches 3 Localities referred to in text Muskeg; generally underlain by 1‘20 feet of peat over organic silt; only largest areas shown 71951315313 TEN 15* 15' LI 1 N11171:! 11 1' ",II‘I Tustu ena outwash 1 , ' ' . ', ' ' ' _ _ ' " « . ,7 ,_ Hz-I [CU/N 1'\\ ...... de'ta ' ‘ , ‘ ' ' " " ' ' IS, 60°00' 60000' T 2 S. 35 I, 4 5- ‘ . , , p y I b J ‘ ' ‘ 1' 1‘ , , \‘ ‘ I ‘ 1' ”" :T » \ “ - ‘ ' ' Lain: U S Simeau'of' \ x T 6 5 Management ' ' Station? ___\ 1N \ W; Bluff Pomt \\ ‘- _ \ (Trialling moraine x ' 5214M 111mm? POEM " / /‘T:Jr'p’5 ./Cfuulsi may) ........... <-\ / ‘ ‘1 ”J /' Ch‘Tna Poof T' 7 S 5: 9b 1 ,,,,,,,,,, / ' ' 61:5.) L(Zigliién Isl Yu I Isl n \ /\ , . \L/ 1' T 8 S ‘ "‘ , 59°30 ‘Hesfeth 1 ‘ - -- , ' T 59°30' 152”00' 30r R113 w. R. 12 w; R. u w, 151°oo' R. 10 w‘ R. 9 w R 8 w, 30' R. 7 w‘ R, 6 w. R. 5 w. 150°OO' INTERIOR—GEOLOGICAL SURVEY. WASHINGTON. D.C.—1967—667277 Base from U.S. Geological Survey Kenai, 1958 Ground damage compiled by H. L. Foster and T. N. V. Karlstrom; and an unedlted advance print geology modified from Karlstrom,(1964) MAP SHOWING GROUND BREAKAGE RESULTING FROM THE 1964 ALASKA EARTHQUAKE IN THE KENAI LOWLAND AND ITS RELATION TO THE GEOLOGY SCALE 1:250 000 15 MILES I 5 5 O 5 10 15 KILOMETERS I—-—I l—I l—l l————-l | CONTOUR INTERVAL 200 FEET DOTTED LINES REPRESENT lOO FOOT CONTOURS DATUM IS MEAN SEA LEVEL DEPTH CURVES AND SOUNDINGS IN FEETgDATUM IS MEAN LOWER LOW WATER SHORELINE SHOWN REPRESENTS THE APPROXIMATE LINE OF MEAN HIGH WATER 1963 MAGNETIC DECLINATION AT SOUTH EDGE OF SHEET VARIES FROM 23°00’ TO 24°30’ EAST UNITED STATES DEPARTMENT OF THE INTERIOR PROFESSIONAL PAPER 543—17 GEOLOGICAL SURVEY PLATE 2 EXPLANATION § ‘_ -_ TYPES OF GROUND BREAKAGE* Ground cracks with associated eruptions of ground water and sand w / , ‘/\駢§§?~~m\zmofa we a? “t , &§_ “NM“xM 0 Ground cracks I Cracks and (or) pressure ridges in muskeg gigo‘ Z1«“3139"(L-{HoIrma ; ” J , Vomr ngorvzcl A .I .., Nomi} Ema! ,1 I [Shim . . . ' " Fim inland " O 9 Ground cracks assoc1ated With slumping to- A i \ ’ vfl.;g J i _ wards unconflned faces J5 , ‘ W ' ‘ , a ' Al. , p A L; Tab/“Tahiti PL] *‘4 Pom! flosscsilim . . . .‘a‘m‘lx — /“ Ground cracks w1th sand eruptlons; asso- 0.0L 2 _ ,7’ o w , , I ,,: Waco; 9 “ , - \J {3 vs Mr , _,_ . , .. , , O—le we 0' a ‘. C “—5 1L/ ! ,0 0 (233233" Va; ' 0% ciated with slumping towards unconfined faces El Ground cracks in sand and gravel of flood plains or valley trains a Landslides or avalanches .—5:1K Vertical tectonic displacement, in feet, along shore as estimated by Karlstrom (1964) .—2.6TP . [—2.3TP] Vertical tectonic displacement, in feet, meas- ured at standard U.S. Coast and Geodetic Survey tide-gage station; measurement given for 1964 and (or) [1965]; figure for Homer Spit is a tide-gage reading of 5.4 feet less estimated surficial compaction of unconsolidated sediments 9am ;'I+11/2L] / * nfizflza‘ehsud {few 3,” 3 ' éfisflm/ ‘ “arr: . —3.9T‘ . [—4.6T] Vertical tectonic displacement, in feet, meas- ured at temporary U.S. Coast and Geodetic Survey tide-gage station. Preearthquake observations were made between 1906 and 1957; postearthquake measurements were made in 1964 and [1965]; most of the dis- placement occurred during the earthquake , —5.es . {—1.231 Vertical tectonic displacement, in feet, meas- ured at U.S. Coast and Geodetic Survey bench mark. Preearthquake observations were made between 1922 and 1952; post- MEASUREMENTS OF TECTONIC earthquake measurements were made in LAN D.LEVEL CHANGE** 1964 and [1965]; most of the displacement occurred during the earthquake . —4.4B Vertical tectonic displacement, in feet, meas- ______ o W {X_ 4“,: .. ured along shore With tid? 1§V61 as a datum Isobase of tectonic land-level change; esti- / 3' _3 SEE _2 0,8“ :km ‘ in 1964; upper growth limit of barnacles mated accuracy :2 feet / <29 [ 3 8T] 3 9T ' i (K (Balams) ’ Q, ‘ " . — » T ”eke. p \ ______ / IV ,. b \ N . —4.ZBB Inferred fault AUG/saws fisiANli) (MFLX‘QNU‘, Vertical tectonic displacement, in feet, as measured difference between pre- and post- earthquake shoreline features in 1965; *Ground breakage in the “Zone of Extensive upper growth limit of barnacles (Balanus) Ground Breakage” is UOt ShOWIL Detailed location of ground breakage in this area is given on the map of the ground breakage /* , o +1%i%L o [+11/2L] in the Kenai Lowland / Vertical tectonic displacement, in feet, along MLand-level change data including location of / I shore as estimated by local resident; meas- zero isobase of tectonic land-level change urements given for 1964 and [1965] by George Plafker INTERIORMGEOLOGICAL SURVEY, WASHINGTON, D.C.~l967—667277 Base by U.S. Geological Survey, 1956 MAP OF THE COOK INLET AREA, ALASKA, SHOWING LOCATION OF GROUND BREAKAGE AND TECTONIC LAND-LEVEL CHANGES SCALE 1:1 000 000 lO 0 10 2O 3O 4O 50 MILES H H H H H . I fi. .7 : 10 O 10 20 3O 4O 50 KILOMETERS HHHHH. ' - . CONTOUR INTERVALS 50, 100,150 AND 300 FEET DATUM IS MEAN SEA LEVEL UNITED STATES DEPARTMENT OF THE INTERIOR GEOLOGICAL SURVEY PROFESSIONAL PAPER 545—1: _+ + ‘+ + \r + +1+ w+~w" w + ,‘+ , Area of enlargement B’ ,++ +,++ 4 4, + + + SCALE 1:63 360‘ T7. ‘ O 1 I I | I I CONTOUR INTERVAL lOO FEET DOTTED LINES REPRESENT 50-FOOT CONTOURS DATUM IS MEAN SEA LEVEL APPROXIMATE SCALE 0 1/2 I I I I I I gs ' Base from U.S. Geological Survey topographic quadrangle, 1951 24%, TRUE NORTH I )\ tr 0 z o C Lu 2 U ‘I E SESSWSIISZS MAPS OF DISRUPTION ON OUTWASH DELTA OF TUSTUMENA LAKE, ALASKA INTERIOR—GEOLOGICAL SURVEY, WASHINGTON, D.C.—1967‘GG7277 1 MILE PLATE 5 EXPLANATION Alluvium and beach deposits Sand, gravel, and silt Glacial outwash deposits Sand and gravel Fan delta deposits Sand and gravel (Shown only on A) Terrace deposits Organic sill 3—12lnches thick over sand and gravel Morainal deposits + + + + + + + + + +++ Bedrock Contact \f’: 3 ’4 Cracks produced by earthquake Barb shows direction of displace— ment along some cracks on terrace UJJ_LI_|_I_LLI_LLLLLL_I_I_L1 Margin of higher ground (Shown only on B) Area where snow was removed by sloshing of lake water after earthquake Area covered with sand and gravel resulting from ground-water eruptions along fissures during the earthquake (Shown only on B) A Letter indicates location of photograph Enlargement (B) shows out- wash delta front as sketched from aerial photographs taken a few days after the earthquake. (See fig. 16) Location of Tustumena delta can be seen on plate 1 The Alaska Earthquake March 27, 1964 Regionalmaa \abf' & K = .,,__ Surface Faults 4 on Montaguelsland GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—3} THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS Surface Faults On Montague Island Associated With the 1964 Alaska Earthquake By GEORGE PLAFKER A description and tectonic analysis of ground breakage and surface warping along two reactivated reverse faults GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—G UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1967 For sale by the Superintendent of Documents, US. Government Printing Office Washington, DC. 20402 THE ALASKA EARTHQUAKE SERIES The US. Geological Survey is publishing the re- sults of investigations of the Alaska earthquake of March 27, 1964, in a series of six professional papers. Professional Paper 543 describes the regional effects of the earthquake. Other professional papers describe the history of the field investigations and reconstruc- tion effort; the effects of the earthquake on communi— ties; the effects on the hydrologic regime; and the effects on transportation, communications, and utilities. Abstract _____________________ Introduction __________________ Regional tectonic setting _______ Geographic and geologic setting- Description of the faults ________ Patton Bay fault ____________ Beach and sea cliff near Neck Point ___________ From the sea cliff to Tor- tuous Creek __________ 1. Map showing regional setting of Montague Island _______________ 2. Map showing tectonic displacements on south— western Montague Is— land _________________ 3—5. Photographs: 3. Patton Bay fault in sea cliff near Jeanie Point ____________ 4. Patton Bay fault scarp crossing shore near Jeanie Point ___________ 5. Shear zone along Patton Bay fault__ Page G 1 «laptop—1 \1 13 Page G3 CONTENTS Description of the faults—Con. Patton Bay fault—Con. Tortuous Creek to the valley of Patton River _______ Patton River valley seg- ment ________________ Patton River valley to Purple Bluff __________ Submarine extension of Patton Bay fault _______________ ILLUSTRATIONS PLATES [Plates are in pocket] FIGURES 3—5. Photographs—Continued 6. Sketch map of tectonic cracks near Patton Bay fault ________________ 7-13. Photographs: 7. Tectonic crack in warped upthrown fault block _______ 8. Crack along joints in a graywacke bed__ 9. Crack in massive in- durated sandstone- 10. Fissure in glacial de- posits along trace of Patton Bay fault ____________ Page G16 21 22 26 1. Geologic map of part of the Patton Bay fault. 2. Geologic map of the Hanning Bay fault. Page G10 11 12 12 13 Description of the faults—Con. Hanning Bay fault __________ South of Fault Cove _______ At Fault Cove ____________ Fault Cove to Bay _________________ Horizontal displacements dicated by geodetic meas- urements _________________ Summary and conclusions ______ References ____________________ 7—13. Photographs—Continued 11. Warped muskeg a- long trace of Pat- ton Bay fault- _ _ _ 12. Warped and fissured muskeg along trace of Patton Bay fault ____________ 13. Tectonic fissures and landslides on ridge near Patton Bay fault ____________ 14—16. Aerial views: 14. Landslides along Patton Bay fault trace __________ v Hanning indi- Page G27 27 29 35 40 4 1 42 Page G14 15 15 17 14—16. Aerial views—Continued 15. Patton Bay fault trace showing off- set ____________ 16. Large landslide scars near Nellie Martin River--- 17. Sequential sketch of for- mation of landslides along the trace of a reverse fault __________ 18. Photograph of sheared rock along Patton Bay fault ________________ 19. Photograph and sketch of incised channel of Nellie Martin River--- 20—33. Photographs: 20. Tilted trees along Patton Bay fault trace __________ 21. Air view of tectonic fissures on ridge summit-r ______ 22. Tectonic fissures on ridge summit--- 23. Minor normal faults with down- hill side uplifted- 24. Vertical air view of reactivated faults at Purple Bluff ----------- Page G17 18 18 19 20 22 23 23 24 25 CONTENTS 20—33. Photographs—Continued 25. Air View of scarp at south end of Hanning Bay fault_- _________ 26. Hanning Bay fault scarp in glacial deposits ________ 27. Fissured and warp- ed scarp along Hanning Bay fault ----------- 28. Pre~ and post- earthquake ver- tical air views of Fault cove ------ 29. Air view of Han- ning Bay fault scarp at Fault Cove __________ 30. Hanning Bay fault scarp at south side of Fault Cove ___________ 31. Hanning Bay fault scarp at north side of Fault Cove- 32. Displaced storm beach along Han- ning Bay fault at north side of ' Fault Cove ----- Page G27 28 28 30 31 32 33 33 20—33. Photographs—Continued 33. Tectonic crack in reef near Han- ning Bay fault-- 34. Sketch map of tectonic cracks near Hanning Bay fault ____________ 35—39. Photographs: 35. Air. View of Han- ning Bay fault north of Fault Cove __________ 36. Toppled trees along Hanning Bay fault _ _________ 37. Ponded stream be- hind Hanning Bay fault scarp- 38. Air View of Han- ning Bay fault southwest from the bay ________ 39. Warped and fis- sured beach de- posits near north end of Hanning Bay fault _______ 40. Diagram illustrating the horizontal displace- ments indicated by re-triangulation ------ Page G34 35 36 37 38 38 39 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS SURFACE FAULTS ON MONTAGUE ISLAND ASSOCIATED WITH THE 1964 ALASKA EARTHQUAKE Two reverse faults on southwestern Montague Island in Prince William Sound were reactivated during the earthquake of March 27, 1964. New fault scarps, fissures, cracks, and flexures ap- peared in bedrock and unconsolidated surficial deposits along or near the fault traces. Average strike of the faults is between N. 37° E. and N. 47° E.; they dip northwest at angles ranging from 50° to 85°. The dominant motion was dip slip; the blocks northwest of the reactivated faults were relatively up- thrown, and both blocks were upthrown relative to sea level. No other earth- quake faults have been found on land. The Patton Bay fault on land is a complex system of en echelon strands marked by a series of spectacular land- slides along the scarp and (or) by a zone of fissures and flexures on the up- thrown block that locally is as much as 3,000 feet wide. The fault can be traced on land for 22 miles, and it has been mapped on the sea floor to the southwest of Montague Island an additional 17 miles. The maximum measured vertical component of slip is 20 to 23 feet and the maximum indicated dip slip is about The great earthquake of March 27, 1964, in south-central Alaska was accompanied by regional tec- tonic warping that caused both ex- tensive subsidence and uplift rela- tive to sea level. During the months after the earthquake geologists made a concerted search for evi- dence of surface faulting. A U.S. Geological Survey field party headed by the writer found two By George Plafker ABSTRACT 26 feet. A left-lateral strike-slip com- ponent of less than 2 feet occurs near the southern end of the fault on land where its strike changes from northeast to north. Indirect evidence from the seismic sea waves and aftershocks asso- ciated with the earthquake, and from the distribution of submarine scarps, suggests that the faulting on and near Montague Island occurred at the north- eastern end of a reactivated submarine fault system that may extend discon- tinuously for more than 300 miles from Montague Island to the area offshore of the southeast coast of Kodiak Island. The Hanning Bay fault is a minor rupture only 4 miles long that is marked by an exceptionally well defined almost continuous scarp. The maximum meas- ured vertical component of slip is 16% feet near the midpoint, and the indi- cated dip slip is about 20 feet. There is a maximum left-lateral strike-slip component of one-half foot near the southern end of the scarp. Warping and extension cracking occurred in bedrock near the midpoint on the upthrown block within about 1,000 feet of the fault scarp. The reverse faults on Montague Is- land and their postulated submarine extensions lie within a tectonically im- portant narrow zone of crustal attenu— ation and maximum uplift associated with the earthquake. However, there are no significant lithologic differences in the rock sequences across these faults to suggest that they form major tec- tonic boundaries. Their spatial distri- bution relative to the regional uplift associated with the earthquake, the earthquake focal region, and the epi- center of the main shock suggest that they are probably subsidiary features rather than the causative faults along which the earthquake originated. Approximately 70 percent of the new breakage along the Patton Bay and the Hanning Bay faults on Montague Is- land was along obvious preexisting active fault traces. The estimated ages of undisturbed trees on and near the fault trace indicate that no major dis- placement had occurred on these faults for at least 150 to 300 years before the 1964 earthquake. INTRODUCTION reverse faults on southwestern Montague Island in Prince Wil— liam Sound through combined air reconnaissance and studies of shoreline displacements. Although it is assumed that the fault dis- placements occurred during the main shock, or possibly during some of the subsequent large after- shocks, there were no eyewitnesses to fix the precise time of rupture. The faulting definitely occurred prior to March 31 (4 days after the earthquake), when a part of one scarp was photographed dur- ing a reconnaissance flight over the island. The faults could not be ex- amined on the ground until 2 months later, at the end of May. As far as could be determined, no movement occurred along any other surface faults on land dur- G1 G2 ing the earthquake. A diligent search was made for indications of renewed movement on known pre- existing faults, particularly in the vicinity of the zero isobase between the major zones of regional tec- tonic uplift and subsidence, but no measurable surface faulting was found. There were no anoma‘ lous abrupt changes in amounts of uplift .or subsidence along the coast or along level lines inland from the coast suggestive of displace- ment along concealed faults. All reported suspected surface fault- ing that was checked in the field turned out to be linear zones of landslides or surficial cracks in un- consolidated deposits. Although some of these lines of landsliding or cracking could possibly reflect fault movements at depth, no di- rect evidence for such movements is available. This report substantially en- larges upon preliminary summa- ries .of the nature and tectonic sig- nificance of the surface faulting that were based on data acquired during the 1964 field season (Plaf- ker, 1965; Plafker and Mayo, 1965). It presents details of the varied and unusually well ex- posed surface manifestations of the faults on Montague Island which have an important bearing on tectonic analyses of the earth- quake. The methods used in de- ALASKA EARTHQUAKE, MARCH 27, 1964 termining the location, displace— ment, and surface dip of the faults are fully described because de- tailed descriptions of overthrusts involving predominantly reverse or thrust movements are scarce compared to the literature on strike-slip or normal faulting. Such data are of increasing practi- cal importance to geologists and others who are concerned with identifying potentially a c t i v e faults and evaluating the possible effects of renewed activity on engi— neering works. Previous descriptions of the sur- face manifestations of overthrust faults are limited to studies of the great earthquake of 1931 centered on Hawke Bay, New Zealand (Henderson, 1933), the 1945 Mik- awa earthquake in southeastern Honshu, Japan (Tsuya, 1950), the 1952 Kern County, Calif, earth- quakes (Buwalda and St. Amand, 1955), and the great 1957 Gobi- Altai earthquake in Mongolia (Florensov and Solonenko, 1963). With the possible exception of some Gobi-Altai faults, the rela- tive displacements on the faults on Montague Island are considerably larger than any overthrusts previ- ously described. They are also unique in that absolute, as well as relative, displacements could be determined across these faults at four points where they intersect the shoreline. I am grateful to colleagues in the US. Geological Survey for assistance in the field and for much stimulating and helpful discus- sion. Field mapping of the faults was carried out by a boat-based and helicopter-supported U.S. Geologi- cal Survey party from May 30 to June 3, 1964, and by a helicopter- supported party from July 31 to August 7, 1965. L. R. Mayo and J. B. Case assisted in the 1964 fieldwork, and L. R. Mayo and M. G. Bonilla collected much of the data during the 1965 season. Additional critical observations on the northeastern segment of the Patton Bay fault were made by L. C. Cluff, of the firm Woodward- Clyde-Sherard and Associates, who visited the area briefly during the summer of 1966. The faults and related features were mapped on postearthquake vertical and oblique aerial photographs taken by the US. Army at scales of 1: 15,000 and 1 2 20,000 and on ver- tical photographs taken by the US. Coast and Geodetic Survey at scales ranging from 1:4,000 to 1: 20,000. Preearthquake vertical photographic coverage available for the entire island at a scale of 1:20,000 proved to be invaluable for determining the relationship of the fault movements to older geologic structures. REGIONAL TECTONIC SETTING The surface breaks on Monta- gue Island lie within the focal re- gion of the earthquake, as defined by aftershock distribution, and within a region of tectonic defor- mation associated with the earth— quake (fig. 1). The regional defor- mation has been described briefly in preliminary reports (Grantz and others, 1964; Plafker and Mayo, 1965; Plafker, 1965). Only the broad aspects of the regional tectonic movements will be out- lined here to provide a proper per- spective for the tectonic signifi— cance of these surface faults. A detailed description and analysis of the overall tectonic displacements will be given in a separate chapter on regional earthquake effects in this series. Tectonic movements, both verti- cal and horizontal, occurred over an area of at least 70,000 square miles, and possibly more than 110,- 000 square miles, of south-central Alaska. The deformed region, which is more than 500 miles long and as much as 200 miles wide, is roughly parallel to the Gulf of SURFACE FAULTS 0N MONTAGUE ISLAND 152' 150' 148° 1 G3 46" 142° 62° co ”“8 repo -q A.” INDEX MAP ° OF ALASKA 60° ‘MI 58" H ‘ SITKALIDAK l l ,I a... lSLANDI l / «\P~$ I I | / $0 | I ll / ‘3’ I l _. INDS l/ b , """" x E' " \3 ,' ’3” Epicenter of main shock Dashed line is boundary of zone of major aftershocks (M244) 56° o ______________________________ o 50 100 150 KI LOM ETERS Epicenter of major aftershock Outer edge of continental shelf (M 26.0) (600 feet) 0 , 50 I100 MILES I I I \ 60LETON ISLAND D Area of uplift or probable uplift Dashed where inferred @ Area of subsidence Axis of uplift Dashed where inferred _*__ Axis of subsidence 96 Volcano 1.—Setting of Montague Island with respect to regional tectonic deformation and seismicity that accompanied the March 27, 1964, earthquake. 260—834 0—67—2 G4 Alaska coast from the Kodiak group of islands northeastward to Prince William Sound and thence eastward to about long 143° W. It consists of a major seaward zone of uplift bordered on the north- west and north by a major zone of subsidence. These two zones are separated by a line of zero land- level change that trends northeast- ward to intersect the mainland be- tween Seward and Prince William Sound. It then curves eastward through the western part of Prince William Sound to the vicinity of Valdez and crosses the Copper River valley about 50 miles above the mouth of the river. The zone of subsidence includes most of the Kodiak group of is— lands, Cook Inlet, the Kenai Mountains, and the Chugach Mountains. The axis of maximum subsidence within this zone trends roughly northeastward along the ALASKA EARTHQUAKE, MARCH 27, 1964 crest of the Kodiak and Kenai Mountains and then bends east- ward in the Chugach Mountains. Maximum recorded downwarping is about 71/2 feet on the south coast of the Kenai Peninsula. In the northern part of the de- formed area, uplift that averages about 6 feet occurred over a wide zone including most of the Prince William Sound region, the main- land east of the sound, and off- shore islands as far southwest as Middleton Island at the edge of the Continental Shelf. Large-scale uplift of the Continental Shelf and slope southwest of Montague Island is inferred from the trend of isobase contours in the north- eastern part of the deformed zone and from the presence of a fringe of uplifted capes and points along the outer coast of the Kodiak Is- land area. The minimum extent of this inferred offshore zone of up- lift is thought to be roughly out- lined by the earthquake focal re- gion, or belt of major aftershocks, shown in figure 1. Montague Island lies within the earthquake focal region along the axis of maximum tectonic uplift. Combined surface faulting and re- gional warping have elevated the southwestern end of the island relative to sea level by as much as 38 feet (fig. 2). Southwest of Montague Island the sea bottom may have been uplifted more than 50 feet where pre- and postearth— quake bottom soundings show re- sidual vertical displacements at least as large as those on the adj a- cent land. As discussed later (p. 26), there is indirect evidence which suggests that the axis of maximum uplift and faulting may extend southwestward more than 300 miles to the area off the south- east coast of Kodiak Island. GEOGRAPHIC AND GEOLOGIC SETTING Montague Island is the most southerly of three long narrow is- lands that trend northeast across the south side of Prince William Sound (fig. 2). The island, which is 51 miles long by 4 to 12 miles wide, consists of a topographically rugged mountainous backbone ridge with average summit alti- tudes of 2,400 feet and a maximum altitude of 2,841 feet. A dense growth of timber and brush man- tles much of the valley floors, raised beaches, and lower moun- tain slopes below an altitude of 1,000 feet. Access through the vegetated areas is diflicult, and is mainly along streams, beaches, and bear trails. Open areas of mus- keg—a mat of Sphagnum moss sev- eral feet thick—sedge, and open- growth scrub occur wherever drainage is impeded by low and gently sloping topography or by impermeable soils. There are no permanent roads or inhabitants on the island although a few loggers and hunters usually live there intermittently during the summer months. No one was on Montague Island during the earthquake. The geology of Montague Island is known only in a general way from regional reconnaissance stud- ies (Grant and Higgins, 1910; Moflitt, 1954; Case and others, 1966; Plafker and MacNeil, 1966). The bedrock is part of the Orca Group, which consists of a thick sequence of well-indurated inter- bedded sandstone and siltstone, minor amounts of limestone as thin beds or lenses, and tabular to len- ticular masses of basaltic lava. The Orca is at least in part of early Tertiary (probably middle to late Eocene) age. The rocks were complexly folded and faulted in early Ter- tiary time. They generally strike northeast parallel to the trend of the island, but there are numer- ous local deviations from this trend. Folds on the island are characteristically overturned to- wards the northwest. with near- horizontal axes; folds in most of the remainder of Prince William Sound are overturned mainly to- ward the southeast. Faults asso- ciated with the folding are mainly southeast-dipping over- thrusts. Well-developed sets of steeply dipping conjugate shear joints trend at high angles to the strike of fold axes and faults in 60°OO’ : 59°45' SURFACE FAULTS 0N MONTAGUE ISLAND G5 148°00’ 147°30’ Fairbankso / / EXPLANATION U / , / Reverse fault / Dashed where approximately located; dotted / where inferredt U, npthrown aide o _ _ _ Isobase contour Showing amount of uplift, in feet. Dashed where approximate SiaIrA__.63 Relative horizontal displacement of triangulation station Arrow indicates direction; numeral is amount in feet (After Parkin, 1966) 10. 633*/ Danger I ‘I .16‘53 .16.SBB 1658* 16.539“ 'T'Bm‘ Vertical tectonic displacement, in feet, H‘s/4:; from measured difference between “EBB" / upper growth limit of preearthquake ,7; / barnacles and sea level (B) or difference between the upper growth limits of pre- and postearthquake barnacles (BB). Asterisk indicates 1965 measurement; all others are 1964. Displacements within areas shown on plates are not included on the figure . 24MM 24MM x \‘3 _ Approximate vertical displacement, in x A / feet, from difference between preearth- r," " quake and postearthquake water depths l/ (After R. J. Malloy,oral commun.,Apr. 30 aoMM // /, 1966) /// g ————— 300 —————— / ’1' Submarine contour , . , 30-42MM, / 0/, Shomng depth,infeet 30 / 0 ~ 24MM 2 , \\ m ( \ p. 03,. \ 0 J _J \\ E 3 (2)3 ‘ ‘~- - A gt :5 A’ Base from U.S. GeoI. Survey‘ Blyrng Sound 40’ < Postquake land girl”)..— 5 40’ and Seward 1250.000 Quadrangles 1953 I /——/”" \ 30' _,, 30, 20’ _‘ 20’ 10' 10' o _h,_ zwflwgngh__*___ 0 o 5 10 MILES o 5 10 MILES lam | I I o 5 10 15 KILOMETERS o 5 10 |__ .._l__; _i | I I CONTOUR INTERVAL 200 FEET DATUM Is MEAN SEA LEVEL PROFILE OF LAND-LEVEL CHANGE ALONG LINE A-A’ 15 KILOMETERS l 2.—Map and profile showing faults and horizontal and vertical tectonic displacements on southwestern Montague Island. G6 the' southwestern part of the island. Surface faults associated with the 1964 earthquake parallel or cut obliquely across the strike of the older bedrock folds and faults at a small angle and are ALASKA EARTHQUAKE, MARCH 27 , 1964 clearly discordant with the dip of these structures. Unconsolidated Quaternary gla- cial till of variable thickness ve- neers bedrock on low-lying parts of the island, and thick deposits of coarse alluvium occur along the lower reaches of the larger streams. Talus, landslide deposits, and frost rubble commonly conceal bedrock on and at the base of the steeper slopes. ‘ DESCRIPTION OF THE FAULTS The locations of mapped fault traces on Montague Island and the distribution of vertical uplift rela— tive to sea level are summarized in figure 2, and details of the features associated with the Patton Bay and Hanning Bay faults are shown on plates 1 and 2, respectively. The following general features are common to both the Patton Bay and Hanning Bay faults: (1) Both strike northeast almost parallel to the long axis of Montague Island; (2) both blocks are uplifted rela- tive to sea level, with the north- west blocks upthrown relative to the southeast blocks; (3) the faults are northwest-dipping r e v e r s e faults; (4) both faults may have small components of left-lateral strike-slip displacement near their southern limits of exposure amounting to less than one tenth the dip-slip component; and (5) the new breakage occurred along or near the traces of preexisting faults which appear to have under- gone earlier Holocene displace- ments. Measurement of the inclination of the fault plane and the dip-slip component of displacement is hampered by the general tendency (1) for the slip plane to be con- cealed by slumping or landsliding across the steep to overhanging fault scarps or (2) for bedrock displacements to be reflected at the surface by flexing and fissuring of overlying unconsolidated deposits and surface vegetation. Approxi- mate surface dips could be deter- mined from the inclination of sheared rock at several places along the fault scarps (pls. 1, 2) ; the dip of the Patton Bay fault is also visible in the sea clifl' where the fault intersects the coast. Absolute, as well as relative, ver— tical displacements could be deter- mined at shoreline intersections of the faults by measurement of shoreline changes relative to sea level. The isobase contours of fig- ure 2 show the amount and direc- tion of vertical displacement that accompanied the earthquake. Di- rections and relative amounts of change were determined at about 7 5 localities on southwestern Mon— tague Island from measurements of the displacement of the upper growth limit of sessile intertidal organisms along the seashore and from changes in the altitudes of storm beach berms. Measurements were made at 14 of these localities both in 1964 and 1965. Data—point locations and measured vertical displacements are shown in figure 2 and on plates 1 and 2. The technique used for measur- ing vertical displacements from the height of the upper growth limit of barnacles above tidal planes has been outlined elsewhere (Plafker, 1965, p. 1675—1679) and, consequently, will not be elabo- rated upon here. Those measure— ments based on differences in pre- and postearthquake altitudes of the upper growth limits of bar- nacles are thought to be generally accurate to about 1 foot. Many of the 1964 measurements, however, are based only on the postearth- quake altitude of these organisms above sea level; they contain in- herent errors due to deviation of sea level from predicted heights and to variations in the growth limits of these organisms resulting from local factors of exposure, rock type, and water characteris- tics. The measurements made on the eastern side of the islands are estimated to be accurate to within 11/2 feet in sheltered bays and to within 21/; feet on segments of the coast exposed to heavy surf and swells along the ocean side. At two localities on sandy stretches of beach where barnacles were not present, the vertical movements were determined from differences in altitude of the pre- and post- earthquake storm beach berms. The accuracy of these measure- ments is unknown, although they give results consistent with those obtained from barnacle—line meas- urements at nearby rocky shores. Submarine control for the iso- base contours southwest of Monta- gue Island is provided by compari- sons of pre— and postearthquake bathymetric charts. Because of navigational and other technical problems in carrying out such sur- veys, the inferred submarine dis- placements could locally be in error by 10 feet or more. Measurements of vertical dis- placement away from the shore are based mainly on the heights of slope breaks in surficial deposits which may or may not accurately reflect the vertical displacement within the underlying bedrock. In those rare instances where the dis- placement is apparently localized along a scarp in bedrock, the scarp height is taken as the net vertical displacement. Only the vertical component of surface displace- ment is shown on plates 1 and 2; dip-slip displacements may be as much as 25 percent greater, de- pending on the indicated dip of the fault at any given locality. PATTON BAY FAULT After the earthquake, the Patton Bay fault was traceable by a suc- cession of exceptionally interesting and complex features, such as fresh fault scarps, surface flexures, and gigantic landslides. From the place where the fault strikes out to sea 11/2 miles west of Neck Point, the rupture extends 11 miles northeastward to the valley of the Patton River (pl. 1). Northeast of Patton River, displacement is distributed across a broad zone of subparallel fissures as much as half a mile wide that can be traced an additional 11 miles as far as Pur- ple Blufi' on the southeast coast of the island. In the following account, the fault on land is described from southwest to northeast. Places re- ferred to, including the locations photographed and diagramed, are shown on the map of the surface ruptures (pl. 1). The northeastern segment of the fault is not shown on the detailed map, but is shown at a smaller scale in figure 2. Fig- ure 2 also shows the measured and inferred vertical displacements relative to sea level at and near the two faults and along a section across them. SURFACE FAULTS 0N MONTAGUE ISLAND BEACH AND SEA CLIFF NEAR NECK POINT The Patton Bay fault is best, ex- posed at its southwestern end at the beach and sea cliff 11/2 miles west of Neck Point. It is marked by a pronounced scarp 6 to 81/; feet high which trends north-south across the beach and the newly up- lifted surf-cut platform seaward from the beach (figs. 3, 4, follow— ing pages). Beach gravel draped across the scarp conceals the actual fault plane everywhere ex- cept at the one small outcrop at the toe of the beach prism where a 4-foot—wide zone of sheared silt- stone dips westward at an angle that averages about 85° (fig. 5). The near-vertical fissure exposed in the sea cliff is along the most probable continuation of the fault. Its steep dip, which is consistent with the fault attitude indicated by the sheared siltstone along the scarp, may represent only a near- surface steepening of the fault plane where it breaks to the sur- face along the most direct path. Previous movement could have oc- curred along the dipping faults exposed in the sea cliff (fig. 3), which are undoubtedly associated with this same line of faulting but which apparently were not re- activated during the earthquake. The position of the barnacle line along the shore in the vicinity of the fault provides an excellent means of determining the absolute vertical displacement of the blocks at and near the fault as well as the relative vertical displacement. The measurements indicate that both blocks are uplifted relative to sea. level and that the western block is upthrown 20 to 23 feet in relation to the eastern (pl. 1, section A—A’) . Uplift of the western block decreases progressively from 35 to 38 feet at a distance of 800 feet from the fault scarp to 21 to 231/2 feet at the scarp; uplift of the G7 eastern block is about 15 feet at the scarp. Only 6 to 81/2 feet of the displacement occurred along the fault scarp that cuts the beach gravel and the reef at the shore (fig. 3) ; the remainder of the off- set takes the form of a pronounced downwarping of the upthrown block within 800 feet of the fault. Numerous surface cracks as much as 170 feet long formed in bedrock on the reef part of the up- thrown block within 400 feet of the scarp. These newly opened cracks can be readily differentiated from preearthquake cracks because they expose fresh surfaces that do not have a white coating of calcareous algae. Locations, orientations, and displacements of the largest and most conspicuous of these cracks, which Vvere mapped by a pace-and- compass survey, are shown in fig- ure 6. Three typical cracks are illustrated in figures 7—9; the photographs were taken at the locations indicated in figure 6. No newly opened cracks were found near the fault in the downthrown block; close to the fault, however, this block is largely covered by beach deposits. The cracks all dip steeply or are vertical, and they appear to occur exclusively along bedding planes and preexisting joints. Displace- ment in the overwhelming major- ity of cracks is an extension perpendicular to the crack face ranging from less than 0.01 to 0.4 foot. Locally, the cracks are sub- stantially widened where loose narrow wedge-shaped rock slivers as much as 1.4 feet Wide have dropped down along them. Left- lateral displacements occurred along two of the cracks. One of the two trends northeast at an oblique angle to the fault scarp and was displaced 1.0 foot (fig. 9). The other crack, which is almost parallel to the scarp, had 0.4 foot of lateral ofl'set. G8 ALASKA EARTHQUAKE, MARCH 27, 1964 3.—Patton Bay fault (heavy solid and dotted line) at the beach and sea cliff near Jeanie Point. Note the vertical offset of contact between beach deposits (Qb) and talus (Qt). The man (circled) is standing at the base of a scarp 81/2 feet high. Light dotted lines are bedding traces in the Orca Group (To) ; heavy dashed lines are probable fault traces on which there apparently was no movement during the earthquake. Location shown in figure 6. SURFACE FAULTS ON MONTAGUE ISLAND G9 4.—View south from top of the sea cliff in figure 3. Arrows indicate fault trace which displaces the elevated beach prism and cuts obliquely across bedrock struc- tures in the surf-cut platform on the upthrown block. The lower part of the scarp was partially obliterated with newly deposited marine sand and gravel by the time this photograph was taken in August 1965. 5.—Near-vertical zone of sheared quartz-veined argillite and graywacke along the Pat- ton Bay fault trace at the shore. Location shown in figure 6. G10 _ ALASKA EARTHQUAKE, MARCH 27, 1964 57°48'15” 147°43'15" ‘ Jeanie Cove Postearthquake lower low _waterr O 100 200 FEET |_ | | Geology and planimetry by M. G. Bonilla and George Plafker 8/4/65 c 0 0 V Q ° v v . ,. Beach gravel Talus Orca Group D - 0 ' ‘ O . O I Q Mud, sand, and gravel U 80 0‘2 —————— TrL—M : —+— Contact Crack, showing dip and horizontal Vertical crack, showing %\orizontal separation, in feet separation in fe t U, upthrow'n side; D, downthrown side 31 OD Strike and dip of beds Ground photograph location and text figure number 6.—Sketch map of the seam and open cracks in the upthrown block of the Patton Bay fault at the shore near Jeanie Point. SURFACE FAULTS ON MONTAGUE ISLAND G11 Se , 73%. :x 4'. ‘ , 1 31:35; w , ”ti—Crack 0.2 to 0.4 foot Wide and at least 2 feet deep in the warped relatively up- thrown block of the Patton Bay fault. Crack follows preexisting bedding and joint planes. Location shown in figure 6. 260—834 0—-67——-3 8.~Looking down on vertical crack 0.01 to 0.015 foot wide along preexisting joints in a hard graywacke bed. Location shown in figure 6. 9.—Crack along bedding plane in massive indurated sandstone; 1 foot of left—lateral separation. Location shown in figure 6. One crack was found on the up— lifted wave-cut bedrock surface of the relatively upthrown block ap— proximately 1,000 feet southwest of the fault. It is as much as 0.2 foot wide and had possible right- lateral movement of less than 0.1 foot and dip-slip displacement of less than 0.2 foot. The open cracks that occur in the reef on the upthrown block are probably tension cracks associated with the downwarping near the fault. Left-lateral displacements observed on two of these cracks suggest a possible small strike-slip component of movement of less than 11/; feet on the Patton Bay fault at this locality. No conclusive evidence for horizontal offset of the preearthquake beach strand lines that cross the fault scarp was found on the ground, although an apparent slight tendency for left— lateral bending of these lines was noted on the large-scale vertical aerial photographs. FROM THE SEA CLIFF T0 TORTUOUS CREEK From the sea cliff to Tortuous Creek the Patton Bay fault trace is poorly defined by a discontinuous zone of fissures, warped surfaces, small landslides, and poorly eX— posed shear zones in rock (figs. 10— 13). The fault in this segment angles obliquely across the struc— tural and topographical grain of heavily timbered low rolling ter- rain with less than 700 feet of re— lief. I‘t gradually changes strike from due north at the shore to N. 37° E. at its northern end. In this segment the fault trace is anomalous in that it does not coin— cide with a major topographic or structural lineament. Features such as alined swampy muskegs subtly suggest that some prior small vertical displacement oc- curred in Holocene time. The prominent nortlnvest-trending lin- 10.—One of the larger fissures along the trace of the Patton Bay fault. The fissure. which is 3 feet wide and 7 feet deep, is in unconsolidated glacial till. eaments and imposing scarps along Deception Creek and upper Slide Creek immediately to the west of the 1964 rupture apparently de- fine a major fault trace along which large vertical up-to-the- northwest movements occurred previously but which shows no evidence of reactivation during this earthquake (pl. 1). The fault trace cannot be located precisely within a few hundred feet inland from the top of the sea cliff, because it is obscured by the headwall scarps of enormous in- cipient landslide blocks that occur all along the upper part of the cliff. It reappears some 500 feet inland from the cliff where it coin- cides with the linear edge of a mus- keg swamp. The fault trace here consists of a north-south trending zone of parallel surface fissures and flexures in glacial ‘till and Sphagnum peat. Individual fis— sures that parallel the trace are open as much as 3 feet (fig. 10) and have vertical offsets of as much as 5 feet with the west (downhill) sides upthrown rela— tive to the east sides. Net vertical displacement across the fissures is about 7 feet; the amount of verti— cal movement resulting from warping is not known. A possible left-lateral component of horizon— tal displacement is indicated by a 17-inch offset of segments of a log that had been frozen into the ground in a position almost normal to the trend of one of the larger surface breaks. A few gaping cracks that trend N. 150 XV. at an oblique angle to the fault trace are also suggestive of some left-lateral strike-slip movement. No evidence for lateral offsets was found any— where else along the fault trace. The fault continues for a mile as a line of poorly defined fissures and flexures through densely tim- bered terrain to the point where it intersects a muskeg bog about 500 feet long by 100 feet. wide that x 'as ALASKA EARTHQUAKE, MARCH 27, formerly flat and was crossed by a small meandering stream. Fault- ing has resulted in a pronounced eastward tilt of the eastern half of the muskeg and a slight anti— clinal flexing of the central part. The tilted segment is now partly inundated by a lake that is more than 8 feet deep where it is dammed against an erosional scarp that borders the meadow on the east; the lake shallows tovard the center of the muskeg (fig. 11). The anti- clinal flexure forms a low north— south trending ridge immediately west of the lake. On the flood plain of Strike Creek, the fault trace is vaguely defined by partially diverted drainage and tilted trees. Immedi- ately to the west of Strike Creek, where it cuts through a muskeg bog in a N. 35° E. direction, the fault trace is clearly visible as a broad flexure through .the form- erly horizontal bog surface and 1964 11 (above).—F0rmerly fiat muskeg that was warped and tilted eastward along the Patton Bay fault trace. The newly formed lake is more than 8 feet deep at the eastern (left) margin of the muskeg. Photograph by L. R. Mayo. 12 (above right) .—Warped and fissure‘d muskeg surface along the Patton Bay fault trace. West (left} side of this formerly flat meadow is uplifted 161/2 feet relative to the east side. Tilted trees in the background are along the flexure zone that marks the fault trace. Note the characteristic tilt of tree crowns toward the downthrown block. Arrows indicate fissures. 13 (bottom right).#l’art of the zone of tectonic fissures north of Tortuous Creek. Two of the larger laterally per- sistent fissures (arrows) that trend obliquely across the slopes are visible above the brush line. The cracks and scars below the brush line result from earthqualm—induced landslides and soil flows. Photograph by M. G. Bonilla. SURFACE FAULTS ON MONTAGUE ISLAND G16 the adjacent wooded area. A con- tinuous sharp flexure about 60 feet wide in the muskeg is broken by a series of open fissures parallel to the trend of the fault that dis- place the muskeg-covered surface and underlying stream gravel (fig. 12) . Relative vertical displacement across the part of the flexure in the bog is roughly 161/2 feet. This estimate of the vertical movement is minimal, because it includes only part of the zone of surface warp- ing associated with the fault. The total width of the zone, as indi- cated by the distribution of spruce trees which are tilted towards the downthrown block owing to warp— ing of the surface, is roughly 125 feet at this locality. From Strike Creek to Tortuous Creek the fault trends across heav- ily timbered terrain in which the trace is indicated by a vaguely de- fined discontinuous broad zone of tilted trees, ground fissures, and small landslides. These features were mapped from vertical and oblique photographs and from aerial observation. An outcrop consisting of highly sheared silt— stone was observed in the bank of Slide Creek at about the point where the fault trace would cross the creek, but no indications of new surface displacement were found in the densely vegetated val- ley bottom. Immediately to the north of Tortuous Creek, displacement is indicated by a zone of fissuring about half a mile wide that lies between the southern and central segments of the fault (pl. 1; fig. 13). The fissures range from a few hundred to more than 1,000 feet in length, and most have east-fac- ing scarps. The more continuous of these fissures are clearly tec- tonic features inasmuch as they ignore topography and trend obliquely across mountain slopes and even over ridge-s as high as ALASKA EARTHQUAKE, MARCH 27, 1,800 feet. However, apparent tec- tonic fissures that nearly parallel the contours of steep slopes are difficult to distinguish from scarps at the heads of incipient landslides. Scarps parallel to the slopes that show evidence of downhill move— ment of the downhill side are arbi- trarily mapped as landslide scarps; undoubtedly some repre- sent fissures along which sliding has subsequently developed. TORTUOUS CREEK TO THE VALLEY OF PATTON RIVER The 5%-mile-long segment of the Patton Bay fault between Tor- tuous Creek and the valley of Pat- ton River trends N. 35° E. and is the most continuous and impres- sive of all the segments. It lies at the base of a line of imposing ridge spurs which are part of the modi- fied scarp formed on the upthrown block of the Patton Bay fault. Along the ridge spurs the fault is marked by a spectacular series of gigantic landslides with headwall scarps as high as 300 feet, soil flows, and surface cracks asso- ciated with incipient slides on the slopes (figs. 14—16). The line of landslides along these spurs is broken only at the valleys of deeply incised streams that dissect the upthrown block intersecting the fault approximately at right angles. In some of these incised- stream valleys the actual fault trace could be seen, but in the inter- fluves the trace was concealed by the chaotic mass of crushed and pulverized landslide debris and a virtually impenetrable tangle of fallen brush and timber. The process by which landslide debris tends to bury the fault trace where it follows along the base of a prominent scarp is illustrated dia— grammatically in figure 17. This process of automatic self—conceal— ment undoubtedly occurs on most active overthrust faults and is 1964 probably one of the major reasons why faults of this type are com- monly difficult to recognize and map by surface geologic methods. These ridge spurs were espe- cially susceptible to mass gravita- tional movements through a com— bination of (1) steep slopes composed in part of unstable debris resulting from prior fault- ing; (2) bedding dips that tend to parallel the slopes; and (3) local zones of sheared and shattered bedrock along the fault plane and near it within the upthrown block. In view of this inherent. instability, it is not surprising that over- steepening resulting from renewed 14.—Looking southwest along the Pat- ton Bay fault in the vicinity of Jeanie Creek. Fault trace (dashed) is marked by the line of landslides at the base of the prominent scarp that is formed on the relatively upthrown block. -—> uplift along a steep or overhang- ing fault scarp at the base of the slopes, coupled with transient ac- celerations due to the elastic ground vibrations and tectonic dis- placements during the earthquake, should have caused the massive slope failures. The transient move- ments alone seem to have been suffi- cient to trigger landslides under comparable circumstances else— where in the area, such as along upper Slide Creek, even where renewed faulting did not occur. From Tortuous Creek to Jeanie Creek the fault trace is marked by a line of landslides at the base of an imposing eastward-facing slope that rises 1,500 feet at an average 15.—Patton Bay fault (dashed) look- ing northeast toward Jeanie Creek. The segment of the fault north of Jeanie Creek is offset about 0.4 mile to the east relative to the segment south of the creek. -> SURFACE FAULTS ON MONTAGUE ISLAND G17 EXPLANATION (h Bedrock Colluvium, alluvium, and glacial till SURFACE FAULTS ON MONTAGUE ISLAND 18.AShattered lentieular bondins of light—gray indurated sandstone in a matrix of sheared black argillite dipping 75° to 85° northwest along the Patton Bay fault at Jeanie Creek. 16 (above left)._——Lands1ide scars as high as 200 feet (arrows) on the rela- tively upthrown block of the Patton Bay fault (dashed) near Nellie Mar- tin River. The fault is exposed in the river channel where it dips about 50° northwest (right) and has at least 16 feet of vertical displacement. angle in excess of 30°. Sheared rock is exposed along the fault trace in an isolated bedrock out- crop 2 feet wide along the south bank of a small stream 0.8 mile north of Tortuous Creek. The out— crop consists of intensely frac— 17 (below left).—Sequential diagram illustrating stages in growth of land- slides along the base of a fault scarp of the reverse type-upon renewed dipslip displacement. A, preearthquake condi- tion; B, overhanging scarp formed near base of slope by renewed fault move- ment; 0, collapse of unstable slope above the fault scarp. tured argillite and graywacke that contain lentieular veinlets of quartz and thin seams of dark— gray clay gouge less than 1 inch wide dipping 65° to 75° NW7. Ex— cept for this one outcrop; the stream valley was so choked with slide debris that no information could be obtained on the displace- ment. Between the south and north forks of Jeanie Creek the fault is offset about 0.4 mile to the east (fig. 15). The base of the slope in this area is entirely concealed by landslide debris, so whether the two segments of the fault here are connected by an abrupt bend or whether they are offset en echelon in a right-handed sense is not known. Part of the fault zone is exposed in a 3-foot-wide outcrop in the bank of Jeanie Creek. The outcrop consists of shattered lentieular boudins of light—gray fine- to medium-grained arkosic sandstone in a matrix of sheared black argil— lite (fig. 18) that strikes N. 65° E. and dips 7 5° to 85° northwest. On the north side of the creek the fault passes through a terrace deposit of stream gravel which has been displaced into four gently sloping steps 10 to 20 feet wide separated by three fissured rises of compar- able width that are strongly tilted toward the downthrown (east) block. Minimum cumulative verti— cal displacement of 10 to 12 feet is represented by the rises that dis- place the terrace. G20 One mile farther north the fault plane is again exposed in the banks of Nellie Martin River; its trace is clearly defined where it crosses the broad alluvial flood plain of the river. The fault zone in the upthrown block along the river consists of 75 feet of intensely sheared siltstone containing veins and lentieular masses of quartz less than half an inch wide; bed- rock is concealed by stream de- posits in the downthrown block. The siltstone in this zone is so frac— tured throughout that it can be readily broken into chips by hand. The sheared rock trends N. 350 E. and dips northwest at angles rang— ing from 50° to near vertical. By August of 1965 the Nellie Martin River, which flows directly across the strike of the bedding and the fault, had incised its channel below the preearthquake level, through 4 feet of stream gravel and 12 feet of bedrock immediately upstream from the fault trace. The amount of incision suggests a minimum of 16 feet of relative vertical displace— ment across the fault at this local— ity (fig. 19). Both north and south of the river channel, the fault trace is marked by a well-defined flexure of the alluvial deposits and the usual as— sociated ground fissures and tilted trees. A profile perpendicular to the strike of the flexure zone about 300 feet north of Nellie Martin River is shown as B—B’ on plate 1. Open extension cracks as much as 1 foot wide parallel the upper part of the flexnre; steeply tilted and downed trees, with crowns pointing downslope, were found along its base. It may be assumed that, if there was no previous break in slope along the new flex— ure, the vertical displacement is the offset of the sloping alluvial fan surface across the flexure zone. ALASKA EARTHQUAKE, MARCH 27, 1964 19.—Incised channel of the Nellie Martin River immediately upstream from the gravel and 12 feet of bedrock at this locality between March 27, 1964, and August SURFACE FAULTS 0N MONTAGUE ISLAND earthquake / river bed ”XE: - Patton Bay fault scarp. The rejuvenated river has eroded about 4 feet of boulder 7, 1965. Photograph by L. R. Mayo. G21 As shown in the profile, the indi- cated maximum vertical displace- ment at this locality is about 20 feet. To the north, the fault trace is again concealed by landslides for a distance of almost 11/2 miles until it reappears as a zone of fissures and tilted trees in the alluvial fan de- posits 0f Braided Creek and the nearby streams. Individual fissures have as much as 4» feet of vertical displacement, the northwest sides invariably upthrown. At one point along this trace, the fault zone is defined by a 40- to 60-foot-Wide swath of trees tilted at a uniform angle of about 70", resulting from a southeastward surface tilt of 20°. Approximate vertical displace- ment across the zone, as calculated from its width and surface slope is 141/2 to 22 feet. North of Braided Creek the fault follows the contact between predominantly massive greenstone tulf along the base of the mountain front on the upthrown block and the alluvial flats on the down- thrown side. At the most northerly point on this segment, where it strikes into the alluvial valley of Patton River, ‘the fault is exposed in a dry creekbed as a 3-foot-wide zone of sheared, crumpled, and sericitizcd siltstone that strikes N. 10°E. and dips 67° northwest. On the flood plain immediately north of this outcrop, a vertical displace- ment of roughly 21 feet is indi- cated by an average 120 downslope tilt of the surface over a zone about 100 feet wide (determined from the tilting of trees). PATTON RIVER VALLEY SEGMENT The Patton River Valley seg— ment of the Patton Bay fault is unique in that it is entirely in an area of low relief underlain by thick deposits of unconsolidated G22 alluvium. A climax forest of Sitka spruce composed of imposing trees as much as 4 feet in base diameter and 140 feet in height occupies the entire valley floor except for the poorly drained swampy areas or active stream channels. Within this forest, the fault trace is de— fined by a remarkably persistent line of discontinuous- surface cracks and associated tilted trees that can be traced to the northern margin of the valley, a distance of nearly a mile. The cracks, which occur in peat moss and soil, are subparallel and trend N. 100 to 30° E. They are open 1 to 4 inches at the surface and show a few down-to—the south— east dip-slip displacements of 2 to 3 inches and no measurable strike— slip components. A minor set of cracks with a general N. 55° E. orientation was observed near the north end of the segment. All the cracks occurred in a well-defined zone a few hundred feet wide; no other newly opened cracks were found outside this zone in the same general area of the valley floor. Tilted trees are associated with the line of cracks, particularly south of Patton River (fig. 20), where average tree heights of more than 125 feet make tilts of as little as 5° from the vertical readily apparent. PATTON RIVER VALLEY TO PURPLE BLUFF The character of the surface dis— placements along the Patton Bay fault zone changes markedly t0 the north of Patton River 'alley. In this 11-mile-long segment. (pl. 1; figs. 21, 22), tectonic movements are strongly suggested by numer- ous peculiar subparallel north- east—trending fissures consisting of open cracks or normal faults with small vertical displacements that, for the most part, are unin- ALASKA EARTHQUAKE, MARCH 27, 1964 ‘a ,. “a“ - 20.—Looking southwest along the Patton Bay fault in the alluviated valley of Patton River. The fault trace is indicated by the tilted giant Sitka spruce trees and by ground fissures that roughly parallel the zone of tilted trees. fluenced by the topography. The southeastern margin of this zone of fissures is a conspicuous linea- ment, trending N. 30° E., similar to much of the Patton Bay fault to the south in that it forms a boundary between an imposing rugged ridge on the northwest and terrain of lower relief on the soutlr east, but different in not having a single continuous scarp or well- defined narrow zone of displace— ment. The character of the fissures ' in this segment suggests the possi— bility that the displacement here is taken up mainly by a broad anti— clinal flexure of the upthrown block and that the longitudinal zone of extension cracks and minor normal faults results from local tensile stress along the crest of the flexure. 21,—Part of the zone of tectonic fissures at 3,000 feet altitude on the ridge summit north of Patton River. The fissures, some of which are indicated by arrows, could be traced almost continuously to the northeast through the peak (center of photo- graph) and the notch on the skyline to the left of the peak. 22.—~Subparallel minor normal faults Within the area shown in figure 21. Note that some of these fissures coincide with pre- existing linear scarps. G24 The fissures were studied in de- tail only at one locality on the ridge immediately north of the Patton River valley (pl. 1, profile 0—0’) . Here the fissures occur over a zone about 1,500 feet wide and are generally subparallel. Their average strike is N. 20° E. parallel to the ridge crest, although some individual fissures and segments of nonlinear fissures trend at an oblique angle to the zone—between N. 80° W. and N. 70° E. Some of the fissures are several hundred feet long and as much as 17 inches wide. Most are nearly vertical with dip-slip displacement of as much as 21 inches (figs. 21, 22; pl. 1, profile 0—0’). The overall width of this zone of fissures is difficult to determine because cracks on the steep southeast-facing ridge slope cannot be readily differentiated from incipient landslides. The zone shown on plate 1, which is 3,000 feet wide, includes the entire area in which surface cracks of possible tectonic origin have been observed. That the fissures on the ridge are tectonic features is indicated by (1) the linearity and lateral continuity of the zone; (2) the fact that most of the fissures dis- place bedrock; (3) the occurrence of some fissures in which uplift on the downhill side, in relation to the uphill, precludes the possibility of landslide origin (fig. 23) ; and (4) coincidence of many of the fissures with preexisting linear scarps in rock along which previ- ous movements may have occurred. The fissure zone could be readily followed northward from the area shown on profile 0—6” of plate 1 across the highest peak on the ridge at an altitude of 2,200 feet. Beyond this point it is discontinu- ously evident from subparallel cracks and landslide scarps on the steep mountain slopes that were ALASKA EARTHQUAKE, MARCH 27 , 1964 13: 23.—Minor normal faults along a preexisting linear groove 011 a steep ridge slope. The downhill side is uplifted 16 inches. Patton River valley visible in the background. traced to Purple Bluff by aerial reconnaissance. Two localities in this zone of fis— sures that were examined on the ground by L. C. Cluff of )Vood— ward—Clyde—Sheraid and Asso- ciates in July 1966 are worthy of special mention. One is at the slope break near the base of the ridge 2 miles north of the Patton River Valley, where a southeast-facing scarp about half a mile long and -i feet high displaces vegetation and soil. The length of this scarp and evidence of surface displace- ment suggest that it may be a fault. The second locality is at the northern end of the zone, where renewed vertical surface displace- ment of no more than a few inches occurred along several preexisting lineaments that extend inland from Purple Bluff for a distance of about 1,600 feet (figs. 2, 24). Some of these fissures near the upper edge of the bluff may be along the heads of gigantic land- slide blocks, but most of them, in— cluding many with relatively uplifted downhill sides, seem to be SURFACE FAULTS 0N MONTAGUE ISLAND G25 24.—Vertical aerial photograph in the vicinity of Purple Blufi showing conspicuous lineaments (arrows) along the inferred northeastward extension of the Patton Bay fault. Renewed displacement occurred on some of these lineaments after the 1964 earthquake. A large earthquake-induced landslide may be seen at the base of the high blufi at bottom center. Photograph by U.S. Coast and Geodetic Survey, August 25, 1964. G26 of tectonic origin. The fissures strike about N. 30° E. and intersect the coast at an acute angle imme- diately north of Purple Bluff. No detailed studies of shoreline dis- placement were made across the zone of fissures at the coast; the available measurements of shore- line changes (fig. 2) suggest that the displacement here is no more than a few feet or that there may be no displacement at all. An alternative possibility is that the fault may once again be offset en echelon to the southeast and that it continues northeastward as a submarine feature. Although there is no direct evidence for such an offset, the precipitous linear coast of Montague Island north- east of Purple Bluff, a concentra— tion there of aftershock activity, and the occurrence of abrupt sub- marine scarps in the vicinity of Hinchinbrook Island (Van Huene and others, 1966) are suggestive of possible active faulting in this area. SUBMARINE EXTENSION OF PATTON BAY FAULT Subsequent to the discovery of the faults on Montague Island, the US. Coast and Geodetic Survey made hydrographic and seismic surveys of the ocean floor south- west of the island. These surveys have revealed a prominent south- east-facing bedrock scarp 40 to 90 feet high that is in line with, and inferred to be a continuation of, the Patton Bay fault (fig. 2). Comparison of detailed bottom soundings taken in 1927 with others taken after the earthquake , in 1964 indicates that the vertical uplift of the bottom on either side of this scarp is in the same sense as, and is comparable in amount to, that recorded on land for a dis- tance of at least 17 miles southwest of the point where the fault strikes out to sea near Neck Point (Mal- ALASKA EARTHQUAKE, MARCH 27, 1964 loy, 1965a, p. 1048—1049; 1965b, p. 22—26). The hydrographic sur- veys do not come closer than 61/2 nautical miles to the fault trace on land, so it is not known whether the onshore and offshore segments of the fault scarp are a continuous strand or perhaps are offset en echelon. Malloy (1965b, figs. 6, 7) be— lieves that a sharp 18-foot-high break in slope on the continuous sounding and seismic profiles closest to shore may represent the 1964 movement; an underwater photograph of What appears to be fresh breakage of the sea floor at this locality tends to support this interpretation. It is also significant that the vertical displacement across the submarine scarp in- dicated by the bathymetric surveys is compatible with the displace— ment of 20 to 23 feet measured across the scarp and warped upper plate along the nearby shore at the southern end of the fault. The fault probably continues farther southwestward, although the 1964 reconnaissance surveys did not track it as a continuous scarp beyond the limits of the pre- 1964 hydrographic survey, which was confined mainly to shoal waters. The lines of evidence dis- cussed below suggest that the Pat- ton Bay fault may be but the northern end of a fault, or of a system of discontinuous faults, that extends southwestward an ad- ditional 300 miles to the area off- shore from Sitkalidak Island in the Kodiak Island group. 1. That this postulated line of faulting is along a narrow zone of probable maximum submarine uplift is inferred from the spatial distribution and runup heights of the seis- mic sea waves associated with the earthquake (Van Dorn, 1964, fig. 6; Plafker and Mayo, 1965, p. 11; Plafker and Kachadoorian, 1966, p. 38—39). The waves clearly were generated on the Conti- nental Shelf within the zone of regional uplift by vertical displacements of the sea floor. The direction of travel and reported arrival times of the initial wave crest—the crest that struck the shores of the Kenai Peninsula nearest to Montague Island within 19 minutes and the southeast coast of Kodiak Island with- in 38 minutes after start of the earthquake—indicate that the wave crest was generated within the regional zone of uplift along one or more line sources in a narrow elongate belt that extends southwest- ward from Montague Island approximately along the axis of maximum uplift shown in figure 1. Furthermore, the similarities in maximum wave runup heights along physiographically compara- ble segments of coast, both on the Kenai Peninsula opposite Montague Island and on K0- diak Island, suggest that the vertical sea-floor displace— ments which generated the waves in these two areas may be of the same order of mag- nitude. 2. The inferred zone of fault dis- placement also has an espe- cially large concentration of features that seem to be a re- sult of recent subbottom tec- tonic movements (D. F. Barnes, unpub. data, 1966; Von Huene and others, 1966). These features include sharp- ly defined submarine fault scarps similar to those near Montague Island, subbottom discontinuities that displace possible Holocene deposits, and small folds in the Holo- cene deposits. SURFACE FAULTS 0N MONTAGUE ISLAND G27 25.——View southwest along Hanning Bay fault~ Northwest block (right) has overthrust the southeast block. Note warped sur— face and overhanging muskeg mat on the scarp in the foreground. Tree in the background has been tilted by the overthrusting. Location shown in figure 27. 3. The concentration of aftershock activity and release of seis- mic-wave energy which ap- parently occurred in this area could indicate postearth- quake fault adjustments (S. T. Algermissen, written com— mun., March 1965). Much of the seismic-wave energy released during the after- shock sequence was contained in the six recorded after- shocks shown in figure 1 with Richter magnitudes equal to or larger than 6.0 (as deter- mined by the U.S. Coast and Geodetic Survey). Of these, two were almost directly in the postulated fault zone, one was on its southwestward pro- jection in the Trinity Islands, two were northwest of the belt in positions where they could conceivably lie on the down-dip extension of north- west-dipping reverse faults, and only one was clearly un- related to this postulated belt of extreme uplift and fault- ing. HANNING BAY FAULT The Hanning Bay fault extends about 4 miles from the south shore of Hanning Bay almost to Mac- Leod Harbor (pl. 2; figs. 25—39). After the earthquake, the fault trace was exceptionally well ex- posed along nearly its entire length, either as a single steep scarp or as a narrow zone of sur- face warping. Descriptions of surface features along the fault are divided geographically into three segments as follows: (1) south of Fault Cove, (2) Fault Cove, and (3) Fault Cove to Han- ning Bay. Places referred to, in- cluding the locations of illustra— tive photographs and diagrams, are shown by the map and photo- mosaic on plate 2. Vertical dis- placements relative to sea level at and near the fault are shown on plate 2 and are contoured at a smaller scale in figure 2. SOUTH OF FAULT COVE The mile-long segment of the Hanning Bay fault south of Fault Cove trends N. 35° E. and is marked by a southeast-facing ,4 G28 ALASKA EARTHQUAKE, MARCH 27. 1964 26.~—Overhanging scarp of Hanning Bay fault in glacial till dipping 80° northwest. About 4 inches of left—lateral displacement is indicated by the rake of roots extending across the scam). Scale is 61/2 inches long. Location shown in figure 27. scarp in surficial deposits, tilted or toppled trees, and gaping exten— sion cracks. The scarp is clearly visible along the extreme south- ern part of the fault where it crosses open ground along a slight preexisting break in slope. The fault trace is marked by a scarp as much as 51/2 feet high; the sur- ticial muskeg layer and underly- ing glacial till on the northwest block has ridden up and over the southeastern block to form a prom— inent pressure ridge (fig. 25). Where visible beneath the surface vegetation, the overhanging scarp in till dips about 80° NW. At its southern end the scarp becomes progressively lower and finally disappears. No indication of fault displacement was found anywhere along the north shore of MacLeod Harbor—an area which was exam- ined in some detail. The displacement is entirely dip-slip except for about 4 inches of left-lateral strike-slip displace- ment near the south end of the fault. The lateral component of movement is indicated by rootlets that trend obliquely across the scarp (fig. 26), by the rake of stria- tions in the fault plane, and by the regular right-handed. en echelon arrangement of gaping fissures that follow the general course of the flexure but intersect it at an angle of about 15° (fig. 27). The absence of lateral displacements elsewhere along the fault, or of systematically offset drainage across the fault trace, demon— strates that the strike-slip compo- nent of displacement during this earthquake and during earlier movements was relatively minor. 27 (1eft).—Segment of Hanning Bay fault trace looking northeast showing surface flexture 3 to 5 feet high broken by gaping sub-parallel right-handed en echelon cracks, 3 to 6 feet apart. Gul- lies that intersect the scarp at right angles show no evidence of lateral off- set that would suggest strike-slip dis- placement. Arrow indicates location of figure 25. SURFACE FAULTS 0N MONTAGUE ISLAND North of the locality shown in figure 27, the fault trace trends in a N. 35° E. direction along the heavily timbered channel of a small stream, which drains into the beach-barred lake at the shore. For most of this distance the fault parallels the northwest edge of the stream channel except at one point where it cuts across a meander and through a small bedrock hill on the opposite side of the creek. The trace is clearly marked by a break in slope with associated ground cracks and a swath of tilted and downed trees whose crowns all point toward the relatively down- thrown block. In one locality the slope break is 81/2 feet high and 24 feet wide; right-handed en echelon tension cracks trending N. 15° E. at the top of the slope are as much as 1 foot wide and 61/2 feet deep. At the north end of this seg- ment, the fault. scarp forms the linear southwest shore of the beach-barred lake. An overhang— ing scarp that dips 80° northwest displaces glacially derived pebble- cobble gravel in a blue-gray silty- clay matrix. Striations in the soft clay are in the direction of dip. AT FAULT COVE By far the most spectacular seg- ment of the Hanning Bay fault is the part shown in figure 28 (next page). It crosses a former incon- spicuous embayment in the coast- line, which has been transformed by vertical displacement during the earthquake into a shallow cove (christened Fault Cove). The movement produced a single con- tinuous scarp 1.4 miles long that displaces bedrock and sediments of the former sea floor and beach deposits along the shore. The fault trace curves gradually from N. 35° E. to N. 45° E. as it crosses the cove. At extreme low tide it is exposed at the surface almost continuously except for the per- G29 manently submerged part at the cove entrance which, however, is readily visible from the air (fig. 29). The existence of a preearthquake submarine scarp in bedrock at this locality is clearly indicated by the fact that unconsolidated deposits were ponded on the shoreward side of the scarp, whereas bare rock is exposed almost everywhere else on the upthrown block. Along the south shore of Fault Cove, the scarp is a prominent southeast-facing break in slope about 81/2 feet high; marine sand and pebble-Cobble-boulder gravel that constitute the beach prism have slumped across the scarp and thereby concealed the actual fault surface (fig. 30). Linear strand lines on the uplifted beach, and bands of marine sand and gravel on the former submarine part of the beach that cross the fault at an oblique angle, show no evidence of lateral oflset. The part of the fault trace along the north shore of Fault Cove in the area seaward from the former beach prism is marked by an abrupt scarp in bedrock or in bed- rock mantled with a thin veneer of unconsolidated marine deposits. The scarp trends either parallel to the strike of bedding planes in the bedrock or intersects them at an acute angle of less than 10°. It varies in height from 121/2 to 131/2 feet, and along most of its length it consists of slumped bedrock and the overlying unconsolidated de— posits (fig. 31). Segments of the scarp that originally were vertical or overhanging have subsequently slumped back to the steep angle of repose shown in figure 31. In part, the slumping was accelerated by undercutting of the scarp toe and by removal of slump debris both by stream runoff diverted along the uphill-facing scarp and by cur- rent scouring. A shallow pond oc- G30 ALASKA EARTHQUAKE, MARCH 27, 1964 28.—Pre- and postearthquake high-tide shorelines at Fault Cove. The cove was formed by relative uplift of the former sea floor northwest of the fault scarp. The White color on'the elerated bedrock surface northwest of the fault results from dessi« cation of calcareous algae and other calcareous marine organisms. Upper photograph by US. Forest Service, June 8, 1959; lower photograph by U.S. Coast and Geodetic Survey, April 17, 1964. SURFACE FAULTS 0N MONTAGUE ISLAND G31 29.—Hanning Bay fault at Fault Cove approximately at low tide. The northwest (right) block is uplifted as much as 16% feet relative to the southeast block. Both blocks are uplifted relative to sea level. As a consequence, the cove is virtually dry at low tide, and the elevated beach and former offshore deposits are deeply dissected by streams. Veneer of unconsolidated deposits on the former submarine part of the downthrown block accumulated behind a preexisting fault scarp. G32 ALASKA EARTHQUAKE, MARCH 27, 1964 30.~View southwest along Hanning Bay fault. Scarp 81/2 feet high has displaced the elevated beach at south side of Fault (love, and beach deposits draped across the scarp conceal the fault plane. Comparative photographs show that modification of the scarp was negligible between May 30, 1964 (above), and August 4, 1965 (below). Bottom photograph by M. G. Bonilla. 31.—Hanning Bay fault scarp 131/2 feet high in bedrock at north side of Fault Cove. The fault dips at about 55° NW. (left), but slumping along the trace has formed a scam) that is nearly vertical or dips steeply southeast. The slumping process is accelerated by erosion along the base of the uphill—facing scarp. Location shown in figure 34. 32.—View northwest toward cobble~gravel storm beach at north side of Fault Cove which was displaced 161/3 feet vertically across the Hanning Bay fault scarp. Pond at right occupies a shallow tectonic depression. Photograph by M. G. Bonilla. G34 cupies a tectonic depression along the scarp base immediately below the elevated beach berm. At the one locality where the fault plane is exposed, it consists of a 11/2-foot-wide zone of sheared siltstone and intensely shattered graywacke containing thin seams of interstitial clay gouge and quartz veinlets. Planar elements within the shear zone dip 50°—60° northwest and cut obliquely across the prevailing southeast dips of the bedding. The maximum amount of meas— ured vertical displacement along the Hanning Bay fault is at the uplifted beach ridge shown in fig- ure 32; here the surface of the former beach berm was displaced 161/3 feet. Maximum dip—slip dis- placement at this locality is almost 20 feet. Bedrock in the upthrown block of the Hanning Bay fault in the area of maximum uplift was broken by numerous cracks that commonly trend at high angles to the fault trace. Individual cracks are as much as 200 feet long and 0.4 foot wide, and are open to depths of more than 7 feet (fig. 33). They are extension features without vertical or lateral dis- placement except for short sections where local caving has occurred. All but one of the cracks are in the upthrown block in a zone that ex- tends 650 feet northwest from the fault. scarp and 1,000 feet laterally along the scarp on the northeast side of Fault Cove. The largest and most continuous cracks are shown in figure 34. A careful search found no other newly opened cracks elsewhere in the ex- posed bedrock platform along the shore in either block except for one crack as much as 0.3 foot wide that opened along a preexisting fault in the uplifted surf—cut platform immediately south of Fault Cove. The relationship of these exten- ALASKA EARTHQUAKE, MARCH 27, 196.4 33.—Near-vertica1 extension crack on uplifted block of Hanning Bay fault at Fault Cove. Crack is 0.2 foot wide and at least 7 feet deep. Hole in foreground is caused by local caving at the intersection of two cracks. Location shown in figure 34. SURFACE FAULTS 0N MONTAGUE ISLAND 34.—Sketch map of the larger tectonic cracks in the upthrown block of the Hanning Bay fault at Fault Cove. Numerous smaller cracks in the same general area are not plotted. Location of photograph is shown on plate 2. Geology by M. G. Bonilla and George Plafker August 1, 1965. Base photograph by U.S. Coast and Geodetic Survey, July 28, 1964. sion cracks to the Hanning Bay fault is uncertain. One possibility is that the cracks result from local tensional stresses due to surface flexing at the culmination of a distinct anticlinal warp that paral- lels the fault trace within the up- thrown block in this area (fig. 2, profile A—A’) . FAULT COVE T0 HANNING BAY North of Fault Cove the appar- ent strike of the Hanning Bay fault swings to N. 50° E. as it crosses the 850—foot—high drainage divide between Fault Cove and Hanning Bay (fig. 35, next page). South of the divide, the fault trace follows along the northwest side of alined incised linear stream valleys that undoubtedly are con- trolled by a preexisting fault. The fault trace in this heavily tim— bered and topographically rugged segment is marked by a swath of fallen trees and brush, intermin- gled with debris from soil slips and small landslides (fie. 36). Ponds have formed where two small streams flow across the up- hill-facing scarp (fig. 37). Sound- ings made near the scarp at the downstream end of these ponds, indicate a minimum vertical fault displacement of 10 feet. North of the ponds, the fault scarp follows a peculiar course that is northwest of, and parallel to, the incised linear stream valley G36 ALASKA EARTHQUAKE, MARCH 27, 196.4 35,—Hanning Bay fault north of Fault Cove. The average 65°-NW. (left) dip of this segment of the fault causes the trace (dashed) to be deflected uphill (right) as it crosses the 850-foot-high ridge between Fault Cove and Hanning Bay. The pond, at the intersection of the fault scarp and the former beach ridge, occupies a closed tectonic depression in the downthrown block. G37 SURFACE FAULTS ON MONTAGUE ISLAND ath of tilted and fallen trees in timbered areas along the Hanning Bay fault. The upthrown block was overthruswt from right to left. 36.—A typical sw G38 ALASKA EARTHQUAKE, MARCH 27, 1964 37 .—Pond more than 10 feet deep formed where a stream flows across the uphill- faeing fault scarp (arrows). Streamflow is toward the observer. Rubble and tangle of fallen trees slid from the steep valley walls. 38.—Hanning Bay fault looking southwest from the bay. The fault trace on the ridge is marked by active landslides. SURFACE FAULTS ON MONTAGUE ISLAND G39 39.—Flexure and scarp with net vertical displacement of 7 feet in boulder-gravel beach along the Hanning Bay fault at south shore of Hanning Bay. Scarp 4 feet high in foreground is at the upper margin, and pond at right side of View is dammed against the base. The fault does not reappear on the north shore of Hanning Bay, which is visible in the background. that heads at the divide on the ridge (pl. 2). The actual rupture is well uphill from the valley bot- tom, and, at one locality, a zone of discontinuous scarps crosses a meadow 0n the ridge about 200 feet above the valley bottom. The main scarp is about 6 feet high in bed- rock and trends N. 35° to 40° E. along the edge of the meadow at the break in slope on the ridge. Discontinuous subparallel scarps or flexures with two sets of asso— ciated extension cracks trending N. 25° E. and N. 75° E. are exposed in an area about 150 feet wide on the ridge summit within the up- thrown block. From the meadow the fault can be traced northward to the drain- age divide as a line of scarps with associated landslides and tilted 0r fallen trees. The fault plane is ex- posed in a scarp 6 feet high on the northwest side of the prominent notch at the drainage divide (pl. 2; fig. 38). Its dip, as indicated by sheared argillite and brecciated soft sandstone, is 52° NW. This estimate is reasonably compatible with a dip of 65° for the fault plane at this locality, as calculated from the altitudes and relative positions on published topographic maps of the fault trace at the divide and at the shorelines of Fault Cove and Hanning Bay. (This calculation was not made on plate 2 because distortion of pho- tographs used in making the plani- metric map does not show these three points in their correct rela- tive horizontal positions.) From the drainage divide to the shore of Hanning Bay the fault trace is concealed beneath blocky talus derived from a large land- slide on the uphill side of the fault (pl. 2, fig. 38). The landslide pre- dates the earthquake, but a sub- stantial amount of debris with freshly exposed surfaces was ap- parently shaken down during the earthquake. Where the fault crosses the up- lifted boulder beach and former submarine platform along the shore of Hanning Bay, there is a pronounced flexure zone about 30 feet wide with a nearly vertical scarp as much as 5 feet high in boulder gravel at the head of the flexure (fig. 39). A small pond fills a depression in the downthrown G40 block along the northeastern mar- gin of this flexure, and the drain- age is diverted along the base of the scarp, which faces upslope. Net vertical displacement across the fault zone is uncertain because many boulders have slumped along the scarp and the top and bottom ALASKA EARTHQUAKE, MARCH 27, 1964 of the flexure are not clearly de- fined. Measurements of the height of the barnacle line on either side of the zone indicate that the north- west block is upthrown about 7 feet in relation to the southeast block (pl. 2). The fault apparently dies out somewhere beneath Hanning Bay. A detailed examination of the northeast shore along its projected strike did not reveal any of the features characteristics of the fault trace on land or any anomalous vertical displacements of the shoreline. HORIZONTAL DEISPLACEMENTS INDICATED BY GEODETIC A re-triangulation, by the US. Coast and Geodetic Survey, of part of the network of primary horizontal control stations in the vicinity of Montague Island sug— gests that both the horizontal shortening and left-lateral com- ponent of displacement in this area may be substantially greater than is indicated by the surface rup— tures (Parkin, 1966, fig. 4). Differences in displacement of two stations in the triangulation net—Stair (610) and Cape Cleare (553)—which straddle the Patton Bay fault as shown in figure 2, provide a geodetic check on the indicated fault movement obtained from field geologic studies. Station Cape Cleare is on the downthrown block 1 mile east of the fault trace; Station Stair is on the upthrown block 7 miles west of Station Cape Cleare and about 6 miles from the fault. Consequently, the indicated relative displacement between the two stations results from both movement on the fault and defor- mation of the crust between each of'the stations and the fault. Unfortunately, the stations are not tied together directly across Montague Island, but rather are connected through a third-order triangulation net (containing a number of geometrically weak figures), which extends around the MEASUREMENTS Stair 43}\\ Cape Cleare 15 58}\ / // 67 / 25 / 5 MILES I 5 KILOMETERS l___—l 40.—-0rientation of horizontal displacement vectors (solid arrows) at triangula- tion stations Stair and Cape Cleare relative to the average strike of the Patton Bay fault. Dashed arrows are the indicated components of horizontal displacement perpendicular to, and parallel to, the fault. Numerals are horizontal displacement in feet relative to station Fishhook (40 miles northeast of An— chorage), which was held fixed in the adjustment. Triangulation data from Parkin (1966, table 1). northeast end of the island. As a consequence, the original position of Station Cape Cleare relative to the remainder of the net is espe— cially uncertain. Added to this un~ certainty is the fact that during the earthquake all stations in this area experienced large differential vertical movements with resultant shifts in horizontal positions of the stations at high altitudes. Nevertheless, the general simi— larity in orientation of the two vectors and their striking differ- ence in magnitude are suggestive of substantial relative movements between them. As shown in figure 40, the prin- cipal vector of horizontal move- ment. derived from data is between N. 1 5° E. at an oblique the geodetic 0° W. and N. angle to the fault system, whereas the direction indicated by the p redominantly dip-slip displacements across the scarps is actually ab )ut N. 53° W., or approximately 11 average strike of th re-triangulation data The Patton Bay f( represented by a co of right-handed en e faults and associated has an average N. The fault at the surf 85° NW. near its sou ormal to the e faults. The further indi- lult on land is mplex system ahelon reverse flexures that 37° E. strike. ice dips about them end and 50° to 75° elsewhere along the scarp. Displacement on the fault is almost entirely dip slip, the north- west side upthrown relative to the southeast side. The maximum measured vertical component of slip is 20 to 23 feet and maximum indicated dip slip is about 26 feet. A left-lateral displacement com- ponent of less than 2 feet near the southern end of the fault is prob- ably a local phenomenon related to a change in strike of the fault that causes it to trend at a small oblique angle to the principal hor- izontal stress direction. The fault system on land for 22 miles aan be traced from shore to shore and is known to continue seaward to the southwest for at least 17 miles. Indirect evidence suggests that the fau lt system may extend southwestward on the sea floor more than 30 3 miles. The fault apparently dies out on its northwestern end but it could be offset en echelon in a right-handed sense and continue northeastward SURFACE FAULTS ON MONTAGUE ISLAND cate that the two stations moved 34 feet relatively toward one an— other in the direction of fault dip and that they were relatively dis- placed 21 feet in a left-lateral sense (fig. 40). The dip-slip short- ening could have been taken up by bending of the upthrown block in addition to the fault displacement. The strike-slip displacement, on the other hand, should have been offshore from Montague Island at least as far as Hinchinbrook Is- land. The Hanning Bay fault is a vir— tually continuous reverse fault with an average strike of N. 47° E., a total length of about 4 miles, and surface dips of 52° to 75° NW. Displacement is almost en— tirely dip slip except for a left- lateral strike-slip component of about a third of a foot near the southern limit of exposure. The maximum measured vertical com— ponent of slip is 16%, feet, and the maximum indicated dip slip is about 20 feet. The reverse faults on Montague Island and the postulated subma- rine extension of the Patton Bay Fault lie within a tectonically im- portant zone of crustal shortening and maximum known uplift as- sociated with the earthquake. The principal horizontal stress direc- tion is oriented roughly normal to the average strike of the faults in a general N. 53° W. to S. 53° E. direction. The displacements that occurred along these two faults during the earthquake are large, but there are four convincing indications that the faults are not the primary fea- tures along which the earthquake G41 readily observable had it occurred along the fault. Absence of a large lateral-slip component on the fault suggests that either the displace- ment was taken up mostly by warping between the fault and the stations or, more probably, that an error has been introduced into this part of the triangulation adjust— ment through a slight clockwise rotation of displacement vectors. SUMMARY AND CONCLUSIONS occurred: (1) The lithology of the rock sequences is not significantly different on the two sides of the faults, as it commonly is along faults that form major tectonic b o u n d a r i e s. (2) Displacement along these known and postulated faults is insufficient to explain the areal extent of regional uplift—— particularly the uplift seaward from the known and inferred trace. (3) Not enough movement occurred on these faults to account for the size and spatial distribu- tion of the earthquake focal re- gion, as defined by the after- shocks; this region extends in all directions far beyond the maxi- mum conceivable limits of the known and postulated surface faults. (4) The epicenter of the main shock is roughly 90 miles away from the fault traces and in a position where it is not likely to lie on the down-dip projection of the fault planes. Both the Patton Bay and the Hanning Bay faults on Montague Island are along prominent linear breaks in slope or linear stream valleys that, for the most part, were clearly Visible on aerial photographs taken before the earthquake (Condon and Cass, 1958). Only the southern segment of the Patton Bay fault, for a dis- G42 tance of 3.3 miles, does not follow a well-established prior fault trace but instead breaks away from a pronounced scarp along which much of the previous movements undoubtedly occurred. Net Quater- nary vertical displacement on and near the Patton Bay fault, as indi- Buwalda. J. P., and St. Amand, Pierre, 1955, Geological effects of the Arvin-Tehachapi earthquake: Cali- fornia Div. Mines Bull. 171, pt. 1, Geology, p. 41—56. Case, J. E., Barnes, D. F., Plafker, George, and Robbins, S. L., 1966, Gravity survey and regional geol- ogy of the Prince William Sound epicentral region, Alaska: U.S. Geol. Survey Prof. Paper 543-0, p. 01—012. Condon, W. H., and Cass, J. T., 1958, Map of a part of the Prince William Sound area, Alaska, showing linear geologic features as shown on aerial photographs: U.S. Geol. Sur- vey Misc. Geol. Inv. Map I—273, scale 1:125,000. Florensov, N. A., and Solonenko, V. P., eds., 1963, Gobi-Altayskoye zemle- tryasyeniye: Akad. Nauk SSSR, 391 p. Also, 1965, The Gobi-Altai earthquake: U.S. Dept. Commerce, 424 p. [English translation] Grant, U. S., and Higgins, D. F., 1910, Reconnaissance of the geology and mineral resources of Prince Wil- liam Sound, Alaska: U.S. Geol. Sur- vey Bull. 443, 89 p. Grantz, Arthur, Plafker, George, and ALASKA EARTHQUAKE, MARCH 27, 19614 cated by maximum topographic relief across it, could be 1,500 feet; it is probably no more than 100 feet on the Hanning Bay fault. Many of the preearthquake scarps along these faults are sharp postglacial topographic features. The straight spruce trees along the REFERENCES Kachadoorian, Reuben, 1964, Alas- ka’s Good Friday earthquake, March 27, 1964: U.S. Geol. Survey Circ. 491, 35 p. Henderson, J ., 1933, The geological aspects of the Hawke‘s Bay earth- quakes: New Zealand Jour. Sci, v. 15, no. 1, p. 38—75. Malloy, R. J., 1965a, Crustal uplift southwest of Montague Island, Alaska: Science, v. 146, no. 3647, p. 1048—1049. 1965b, Seafloor upheaval: Geo— Marine Technology, v. 1, p. 22—26. Moflitt, F. H., 1954, Geology of the Prince William Sound region, Alaska: U.S. Geol. Survey Bull. 989—E, p. 225—310. Parkin, E. J ., 1966, Horizontal displace- ments, pt. 2 of Alaskan surveys to determine crustal movement: U.S. Coast and Geodetic Survey, 11 p. Plafker, George, 1965, Tectonic defor- mation associated with the 1964 Alaska earthquake: Science, v. 148, 'no. 3678, p. 1675—1687. Plafker, George, and Kachadoorian, Reuben, 1966, Geologic effects of the March 27, 1964, earthquake and as- sociated seismic sea waves on Kodiak and nearby islands, Alaska : O faults, including many giants as much as 4 feet in base diameter that are probably 150 to 300 years old (J. Standerwick, U.S. Forest Serv- ice, oral commun., Oct. 4, 1966), indicate that no major displace- ment has occurred for at least that length of time. U.S. Geol. Survey Prof. Paper 543— D, p. D1—D46. Plafker, George, and MacNeil, F. S., 1966, Stratigraphic significance of Tertiary fossils from the Orca Group in the Prince William Sound region, Alaska: U.S. Geol. Survey Prof. Paper 550—B, p. B62—B68. Plafker, George, and Mayo, L. R., 1965, Tectonic deformation, subaqueous slides, and destructive waves asso- ciated with the Alaskan March 27, 1964, earthquake—an interim geo- logic evaluation: U.S. Geol. Survey open-file report, 21 p. Tsuya, Hiromichi, ed., 1950, The Fukui earthquake of June 28, 1948 : Tokyo, Report of the special committee for the study of the Fukui earthquake, 197 p. Van Dorn, W. G., 1964, Source mecha- nism of the tsunami of March 28, 1964, in Alaska : Coastal Eng. Conf., 9th, Lisbon 1964, Proc., p. 166—190. Von Huene, Roland, Shor, G. R., Jr., and St. Amand, Pierre, 1966, Active faults and structure of the conti- nental margin in the 1964 Alaskan aftershock area [abs]: Am. Geo- p’hys. Union, 47th Ann. Mtg., April 19—22, 1966, p. 176. UNITED STATES DEPARTMENT OF THE INTERIOR 59., GEOLOGICAL SURVEY V’ / wauj‘x» ”’ “I; I , It (6/ fl” . Ck 1x II , H ’\ ‘ i \l \ H mm ‘ III I II I ‘ ‘l ‘I. “ I‘lIWhIIfiV‘I? I:\ \V \I: IIIIIIIW W. I, \\,,Iw 'ittt‘ulx ‘ \\ \NII ‘,l‘¢~t\\\\ \\i\ ”All I. \\\\\\\\\\ s ., I..,II\\ w .. .II I, M I II. II‘IIII \ I \\\l '\ . ' l i N \l \\ M 1 III ~. _ / . w p, _, III \ ”:I‘ I ~ \IIIII \lIlI‘KIII "‘ Ill A“ I“ All} \\f\ "\7 I“ , Ni W ,tszj‘IIIIIIIII\\I\\\I\\IIIIIIII , \I , ,III\I\I‘ \ WW III .. . I? ‘I' WWI Ill‘ ':\‘\p ‘I‘llwlli \ \li \ I I \\\\\\\ \\ IIIII IIIII I ,\\\\ul\l\ I» I. \‘ \I\\\\\\ II :<,III‘I I. ‘ g -\\I . ‘ \IilllIilllt ‘\ \l\\ \ \\\\\\\ill‘l \llliIINIl III: 3‘ W I \L‘IliillIl \\ ‘ ”All i I NRIIIIVIKXII \\ \ III III \\\\\I\\I \IIIIIIIII ‘" 1.“.12‘\ “ III} Ill lift II“ i?“ . , \l‘ l News t, . ‘x I ' _ \: Ill, ilv ‘ If I ‘1 ' XI “hill \M xbhwlwgkilll \ “ will ’ ; NI A All , , W \ ”\‘ifl ll 4. I“ x’ l ‘\': I“ H‘i Ii " I WWIIIIIII, , \I ,, kl \ in fill I III I , ix \ ll“ id \‘\. II? III\ \ ’1 i“ \ ‘ \ \\fit I ' III I » I ‘ \ MANN \\ ' \\ .I , \ .2, “N“ i MI I\ K ,3 . , . \\\ :I ” I I I . I r INK IIN‘IIIIIW . .l.“ “I MAI ll} \\\l\\\\\\\lh\i\\‘\ ,l ‘ \ II» N W I . \‘I IilIlslIliilll‘l lliillt l\ \ Itlxllll \\\\\\\\\ \ \. ‘\ ll“ \l.‘ I II, M Q 09‘ :9 Base from U.S. Geological Survey 1:63.360 scale N, Blying Sound D—l, D—2, and D—3 sheets ,A APPRoXiMATE MEAN ‘ I ‘ I‘M Hub “1, ‘ 'I 1‘, ‘ I . \\ I\ ll I‘ it I I. I > ' A . , i I‘m a, M aka}, A x. I i m I, SCALE 1 :31 680 DECLINATION,1967 1 2 3 . l CONTOUR INTERVAL 100 FEET DATUM l8 147"30’ MEAN SEA LEVEL PRIOR TO EARTHQUAKE PATTON BAY FAU LT ZONE ,IIIIHW II \llll‘lll \‘\‘\l\‘ \\ \ ~\ W.“ i . g 3%, W's, fill“ II ,/}7 \ I ’ A l ,’ ‘3 T l t nil) \\ It All: ‘lw \/ 6 I. "I‘\\ ,4,” “I sit \I .. \\t\\l X‘er‘lgg‘: Hal :‘u .‘n in“: E‘,\,xk\\H AA. rim mm» mm H. ‘II M i, H”. 4mm .1 I... Il‘iéku tx\\:l\,t\\i‘i III mu. 4 MILES l 4 KILOMETERS o 100 200 300 400 500 1000 INDEX MAP l | l l . I 690 790 890 990 I FEET GEOLOGIC MAP OF PART OF THE PATTON BAY B B’ 100’ — _ 100! :__ ___[Postearthquake ground surface 90' ——:2:"-~-—~——“~::---- — 90' \\‘§_h.~&-._:: TTTTTTTTTTT ’3 80 —\\‘ — 80 :> \‘§5 70’ — “ ‘L — 70' if Projected slope of prequake/‘ A 2 A; 60’ — ground surface _ 60' 50' __ 2 _ 50' 50: _ Approximate vertical displacement — 50' O 40’-— \ \ To 1'3 — 40’ 40' —— — 40’ O \ I .’ x I \ \ Postearthquake ground surface < To 3 ._ a _ , , _ __ , I Scarp height 30 3O / 30 20' — [J x ', 81/2-6 ft To _ 20/ 20! _ / // _ 20’ 10,—; I ‘ : \ 10 10’ Sheared siltstone exposed in zone 10, / \ ‘ I \\ ,1 \VI — ' _ 75 ft ide al n B a'd d C k — '/ \ v , ‘ l/Approximate preearthquake‘ground surface 500 ftv‘;outh 2f gseCEifineline'r: To SEA LEVEL \ \ , V x , \ \ \ (POSTEARTHQUAKE MLLW) — 50, NW ' p — SEA LEVEL / \LV/ \\ l/ \ I \\ A \ / \ ‘ 10,—.— \"/ —— \\\—’I [TX \l \\ ’/ \\\ \l /I’ \\ \\\ //— 10' 10' "‘ — 10' l \ \,’ \ / \ "’ 20L Sheared graywacke and siltstone\\/\, ‘~~_—/ \\ _ 20’ 20, ~ —- 20’ Note: Bedrock structures shown are diagrammatic in zone 4 ft wide; dip 85° w. ‘\-/ 30' 30' 30' 30’ 0 100 200 FEET l FAULT, MONTAGUE ‘\\:\\\\ \\\ WIN l I II II I, I Ila \. m @% ./ /, INII ., ‘ I " ill ,\\{a\§“i‘\\\\ ‘ I‘M, I I, \ \IIllllllllIllIlI ‘l‘\ “I. i ‘\\I\\\ XXV. 1900' —i Projected cracks Snow 1800’ — 1700' —“ EXPLANATION 4—. _, Crack showing relative displacement 1600’ — 9” Width of surface crack, in inches 16V Vertical surface displacement, in inches , it, L /' 7/113. :70 . _ ', ,I’ a 3“' I {I u, ,\ fill, a \‘X‘ix 1 A\IIII \II IIIWN ‘I\\\ \\ ‘ I“, in \IlII‘l‘lI k2] \I\I\\ ,I‘ . \x \ ” \‘Il‘il‘ll. \ I III, \ \I I II ( \IIIlI, \\ ilk I “I I lIlIIIIlfiII I‘I ‘x'il ( 4 1H \\\ .i t i ‘ \ 1 x\ I It \\\\ll§il\i\\\\\ \V ll‘llIl \~. ‘I \I \, \\ \IWW \I \\\ \El\\l\\\l\\\\\\ \ It I \I II I: ‘1 \ I \\\ \\ n‘\i\'\,\ ,. \\ I. All II \\ ‘1 2‘ Kfi‘fi \. II ink \ its .\ \ , 21V,17H 15v, 3H\ /,/2H _, \I. \\ \\\.\\\\lI\\\I\l\\l:\§\\ \ \Illllhl l I \\ \ 4 fl 2/ "“E‘I‘Xt \' »\\I\ i , \ll‘xIIllllIl IIIII \i‘ \ a MIR”? \\ “if ”5&3 \\ I‘lll\\\:§xlsfil\\ I i I. I, NEE “all, is“ \ Kl \ \ \K ‘ ‘ , .‘ :ltIl". III I . . If; 5 x, H, E \m IIIIIIRIW “Ii all; All ‘ iii it Will I II ‘\ I I II WV Willi ,. l A \ iI i¥1§:\\\ I . III I I “I\ \\ II \ \\ “ T A?» I i ‘i ‘ 3 bill I i \ l \QIl‘sm “Al \‘I \W‘li , \ \ I I A ‘ 'u 1" \blfiéi‘lk‘ II , III 3 i l M ‘s H i \I III, \ \IlI lilies ‘ will“ 5.. x§§wxy II‘lI‘ : I W \ I \ .\ . \ AN II\\\\\\ \\I \ii‘lilkéf i‘x",x\\ kl; i X: \'\‘I{,:‘} "‘7 Al N5. IA “\ kl“ V ‘ ‘x Xx §\ Th“ 21 :Vilv‘ll\ l ‘ W A IN \IIIIIII I I III I\ III I ‘. II x I ‘\ \\ \ 1‘ lilii’l‘il” \ \ i I ll l ‘ ' ’l \\. ‘lilI\l\\ \\\\\\\\ \lIIIIII‘i II 6:9” Geology from field mapping by George Plafker, L. R. Mayo, M. G. Bonilla, and J. B. Case, 1964—65; data furnished by L. C. Cluff, 1966; photo interpretation by George Plafker 1V, 11/2H Cl — 1 700’ ‘ — 1600’ 1500’ 1500’ O 100 l ISLAND, ALASKA 200 l 300 l lNTERlOR—GEOLOGICAL SURVEV, WASHINGTON, D. C.—1967—G67283 400 l 500 FEET J ‘7 1900' — 1800' Recent Lower Tertiary PROFESSIONAL PAPER 545—G PLATE 1 EXPLANATION K31 Qal Alluvial mud, sand, and gravel QUATERNARY Marine beach deposits UNCONFORMITY TO Orca Group Undifferentiated graywacke sandstone, siltstone, green— stone, and conglomerate. Complexly folded and faulted. Commonly mantled with unconsolidated allupium, beach deposits, glacial till, and colluvium TERTIARY Contact Dashed where approximately located 67 #___.... 21 Reverse fault, showing direction and amount of surface dip Dashed where approximately located or concealed by landslides; dotted where inferredfrom discontinuous zone of surface cracks or flexures in alluvium. Number is approximate vertical separation in feet Zone of abundant fissures of tectonic origin Red lines are the longer cracks Prominent major linear feature interpreted as trace of thrust fault. No apparent displacement noted during the 1964 earthquake. 70 80 _J.._ W Inclined Vertical Contorted May be overturned May be overturned Strike and dip of beds ( T I: \ Larger earthquake-triggered landslides W Scarp at head of incipient landslide Area exposed by uplift associated with the 1964 earthquake X 16.58 X 16.588 X35.0DD* 16.58* 16.588* Vertical tectonic displacement, in feet, from meas- ured difference between upper growth limit of preearthquake barnacles and means highwater (B), the difference between upper growth limits of preearthquake and postearthquake barnacles (BB) and the difference in heights of preearthquake and postearthquake storm beach berms (DD). Asterisk indicates 1965 measurement, all others are 1964 21[ Oblique aerial photograph location and text figure num / / Vegetation UNITED STATES DEPARTMENT OF THE INTERIOR 60°OO' GEOLOGICAL SURVEY Recent Lower Tertiary MONTAGUE Approximate postearthquake ////TTHN_T‘*—"~'/ k) mean lower low water (M LLW) \\ ., .-"§-'7.:OBB* ££‘;WWWW6\ \ {V///7; ”z/éfl/fi/ 18 in. sheared graywacke )4. " , ,' , ’ ’y , /,/~/ *I (\l , 35.OBB* / \ ’3’ //777/}/C//?//7 and goug ’ ' 1' « V 'N’ " ’ / I 7 ' / l 4/ I ,4: 35 ‘ Q} ..,/ k (7/1» /, W i////” //" , >7 / ’1‘“ ‘53 "fw’ 32.3BB* \ (1)01,» / / / , . / ‘ / // \ To \ 06) / /,./” \\ \ e7 / / \\ i‘ .\ / \ \ \I TO /I \ ,,..--""\\ / i -‘ Fault Cove I ' / '\ .. \ I .. , i . , \ ,l A 32.28 \ / . \31.9BB‘* \- Bre ciated sandstone . A and siltstone’ ._./ A 7/77 "*— \\// . , \ 25.33 (€25 , 24.533* \ 24.33 \i \\ 27.0BB* Geology from field mapping by George Plafker. Planimetry and photomosaic from uncontrolled L. R. Mayo, M. G. Bonilla, and J. B. Case, 1964—65 U.S. Army vertical aerial photographs;July 27, 1964 scale approximately 1:15,000 0 1 2 MILES and photointerpretation by George Plafker l I l I 147.30, APPROXIMATE MEAN O 1000 2000 3000 4000 5000 10,000 FEET DECLINATION, 1967 I i I I i i i ‘ APPROXIMATE SCALE HANNING BAY V -l ‘ FAULT \Q ' B‘W Map area 4, \PATTON BAY Q FAULT / Maeleod Harbor B B 1500— 1500' A Al 50' Uplifted reef I‘ Beach deposits—*4 50' 40,— _ 1000’ —* 1000’ t x l , , Veneer of marine gravel 40' ( 3OI_ \\\ \1 I, //// 3O, , \\ \\ I I, ,”- ' ' 20 _ \\ \\ ,’ : ,’ 20, 500 —- 500 G U L F o F , \ X,’ ‘ i - 10 v \\ /,’ I my Uplifted reef Qs A L A s K A SEA LEVEL \ / x "—\—’I‘—/" Note: Bedrock structures shown are SEA LEVEL (POSTEARTHQUAKE MLLW) MLLW _. MLLW 0 5 10 MILES diagrammatic; they are based on a measured section alon th th h I of MaCIeOd Harbor g e nor 5 ore 100 200 300 40 500 INTERIOR—GEOLOGICAL SURVEY, WASHINGTON, D. C.—1967—GG7283 0 O FEET INDEX MAP | I I l I # OI I 10'00 I 20100 1 30'00 FEET THE HANNING BAY FAULT, MONTAGUE ISLAND, ALASKA GEOLOGIC MAP, SECTION, AND PHOTOMOSAIC OF PROFESSIONAL PAPER 545—G PLATE 2 EXPLANATION Marine shoreline deposits Beach deposits associated with the preearthqnake shore- line. Blank areas are predominantly boulder gravel and rock; patterned areas are mainly gravel and sand OUATERNARY Orca Group Graywacke sandstone and siltstone. Complexly folded and faulted. Locally mantled with alluvial deposits and glacial till. Dotted lines indicate trace of bedding; patterned areas are rock veneered with silt, sand, and gravel TERTIARY Contact Dashed where approximately located 55 24—? _ _. Reverse fault, showing direction and amount of surface dip and lateral displacement Dashed where concealed by water or surficial deposits. Number is approximate vertical separation, in feet A _’\ Fissures along the fault trace 65 65 .__|_ Inclined Overturned Strike and dip of beds 1E1?) Large landslides along the fault trace /, /, Area exposed by uplift associated with the 1964 earthquake 16.53 x x 16.533 1653* 16.SBB* Vertical tectonic displacement, in feet, from meas- ured difference between upper growth limit of preearthquake barnacles and mean high water (B) or the difference between the upper growth limits of preearthquake and postearthquake barnacles (BB). Asterisk indicates 1965 measurement; all others are 1964 Stream showing direction of flow; dashed where approximately located ' EXPLANATION OF SYMBOLS SHOWN ON THE PHOTOMOSAIC as] Oblique aerial photograph location and text figure number 30 —>— Ground photograph location and text figure number l..____..J Outline of vertical aerial photograph and text figure number The Alaska Earthquake March 27, 1964 Regional Effects g-y’ // f)? ‘ entitle Isfiff‘ i j‘s: ”mi MAR 20 nose LlBitew- . UNNERS‘J‘I (3? CL: Lanai-tin I Erosion and Deposition on a a. Raised Beach, Montague Island 0.5.5.1). . - , ~ , A GEOLOGICAL ,SURVEY PROFESSIONAL- PAPER 543-H x y. x.‘ L M. s 1 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS Erosion and Deposition on a Beach Raised by The 1964 Earthquake Montague Island, Alaska By M. J. KIRKBY and ANNE V. KIRKBY , A quantitative study of geomorpht'c modifications of uplifted coastal features GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—H UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1969 For sale by the Superintendent of Documents, U.S. Government Printing Office Washington, DC. 20402 — Price 60 cents (paper cover) THE ALASKA EARTHQUAKE SERIES The US. Geological Survey is publishing the re- sults of investigations of the Alaska earthquake of March 27, 1964, in a series of six professional papers. Professional Paper 543 describes the regional efiects of the earthquake. Other professional papers describe the history of the field investigations and recon- struction effort; the effects of the earthquake on com— munities; the efl'ects on the hydrologic regimen; and the effects on transportation, communications, and utilities. CONTENTS Page Page Abstract _____________________ H1 Erosion and deposition of bay- Erosion and deposition on up- Introduction __________________ 2 head deposits in MacLeod lifted beaches and rock plat- Physical setting _____________ 2 Harbor—Continued forms—Continued Methods ................... 4 Stream erosion—Continued MacLeod Harbor beaches—— Geology ...................... 4 Influence of underlying ma— Continued Bedrock ____________________ 4 terial on stream erosion_ _ H 10 Regrading of the beaches--- Recent unconsolidated deposits Streams in silt ____________ 11 Patton Bay beaches _________ in MacLeod Harbor ...... 5 Streams in sand ___________ 11 Stratigraphic evidence for Sand _____________________ 5 Stream in gravel __________ 14 sea-level changes ________ Silt ---------------------- 5 Lateral and vertical erosion- 14 Elevations of beach features- Gravel """ _ ' ‘ ' ‘j """"" 5 MOdel 0f erosion ---------- 15 Tectonic deformation of Beach deposrts 1n Patton Recession of knickpoints___ 15 beaches ________________ ‘ Bay ----- j: """"" 7 Other erosive agents in regrad- Obliteration of break in Erosron and deposrtion of bay- . . 15 l at to of beach head deposits in MacLeod mg Of deposits """"""" s ope. p "" Harbor ___________________ 7 Erosion and deposition on up- Regradmg of rock platforms- Formation of bay-head de- lifted beaches and rock plat— Regradlng of sand and posits ____________________ 7 forms -------------------- 16 gravel __________________ Stream erosion _______________ 8 MacLeod Harbor beaches- _ _ ._ 16 Conclusions ___________________ Data collected ____________ 8 Beach profiles _____________ 16 References cited _______________ ILLUSTRATIONS FIGURES Page 1. Index map of Montague Island ____________________________________________________ H3 2. Map of bay-head deposits in MacLeod Harbor ______________________________________ 6 3. Panoramic photograph of eroded bay-head deposits in MacLeod Harbor ________________ 9 4. Diagram showing relationship of volumes removed by fluvial erosion to elapsed time _____ 9 5. Graph of variations in erosion rates of streams in MacLeod Harbor ____________________ 10 6. Photograph of typical stream with silt banks in MacLeod Harbor _____________________ 11 7. Map, long profile, and cross sections of typical stream in silt in bay-head deposits of MacLeod Harbor ______________________________________________________________ 12 8. Planetable map, long profile, and cross sections of stream in sand, MacLeod Harbor _____ 13 9. Cross profile of gravel sheet and lateral migration of streams, MacLeod Harbor _________ 14 10. Graph showing inferred rates of lateral and of vertical erosion for streams, MacLeod Harbor _______________________________________________________________________ 14 11. Graph showing postearthquake degradation of silt, by rainwash and marine erosion, along a line of pilings _________________________________________________________________ 16 12. Sketch map of shoreline lithology and location of measured beach profiles _______________ 17 13. Graph of beach profiles around MacLeod Harbor ____________________________________ 18 14. Graph of beach profiles and stream profile through beach _____________________________ 18 15. Photograph of rock platform showing differential erosion _____________________________ 19 Page H19 20 22 26 28 33 36 38 39 41 VI CONTENTS Page. 16. Photograph of rock platform, MacLeod Harbor _____________________________________ H19 17. Map Of Patton Bay showing coastal features and deformation _________________________ 21 18, 19. Stratigraphic sections and beach profiles, Patton Bay _____________________________ 23, 24 20. Photograph showing cobbles of raised beach on truncated bedrock platform, Patton Bay__ 25 21. Stratigraphic section showing old raised beach, Patton Bay _________________________________ 25 22. Stratigraphic sections and profiles of Nellie Martin River _____________________________ 26 23. Graph showing distribution of elevations of beach features ____________________________ 27 24. Sketch of beach profiles in Patton Bay in relation to the preearthquake mean high water- 28 25. Graph showing variation in differences of elevation among coastal features in Patton Bay- 29 26. Sketch map of differences in elevation of raised beaches at the north end of Main Sandy Bay __________________________________________________________________________ 31 27. Sketch map and graph of differences in elevation of raised beaches at the south end of Main Sandy Bay ______________________________________________________________ 32 28. Sketch map of cliff degradation and accumulation of talus, Patton Bay _________________ 34 29. Sketch map of clifl" degradation by landslides and gullying in North Sandy Bay _________ 35 30. Idealized profile through two raised beaches _________________________________________ 36 31. Graph showing sequence of platform cutting in Patton Bay ___________________________ 36 32. Graph showing relationship between change of elevation of sea level and cross-sectional area- 37 33. Graph showing position of preearthquake storm beaches _______________________________ 39 TABLES P886 1. Summary of data for streams in MacLeod Harbor ____________________________________ H10 2. Relationship of width of platform regraded, cross—sectional area of material removed, and change of relative sea level _____________________________________________________ 37 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS EROSION AND DEPOSITION ON A BEACH RAISED BY THE 1964 EARTHQUAKE, MONTAGUE ISLAND, ALASKA By M. J. Kirkby1 and Anne V. Kirkby 2 During the 1964 Alaska earthquake, tectonic deformation uplifted the south- ern end of Montague Island as much as 33 feet or more. The uplifted shore- line is rapidly being modified by sub- aerial and marine processes. The new raised beach is formed in bedrock, sand, gravel, and deltaic bay~head deposits, and the efiect of each erosional process was measured in each material. Field- work was concentrated in two areas— MacLeod Harbor on the northwest side and Patton Bay on the southeast side of Montague Island. In the unconsoli- dated deltaic deposits of MacLeod Har- bor, 97 percent of the erosion up to June 1965, 15 months after the earthquake, was fluvial, 2.2 percent was by rain- wash, and only 0.8 percent was marine; 52 percent of the total available raised- beach material had already been re- moved. The volume removed by stream erosion was proportional to low-flow dis- charge raised to the power of 0.75 to 0.95, and this volume increased as the bed material became finer. Stream re- sponse to the relative fall in base level 1Formerly Research Associate, Dept. of Geography, The Johns Hopkins University and now Lecturer in Geography, Univer- sity of Bristol, England. ”Graduate student, Dept. of Geography, The Johns Hopkins University, Baltimore, Md. ABSTRACT was very rapid, most of the downcut- ting in unconsolidated materials occur- ring within 48 hours of the uplift for streams with low flows greater than 10 cubic feet per second. Since then, ero- sion by these streams has been predom- inantly lateral. Streams with lower dis- charges, in unconsolidated materials, still had knickpoints after 15 months. No response to uplift could be detected in stream courses above the former pre- earthquake sea level. Where the raised beach is in bedrock, it is being destroyed principally by marine action but at such a low rate that no appreciable erosion of bedrock was found 15 months after the earth- quake. A dated rock platform raised earlier has eroded at a mean rate of 0.49 foot per year. In this area the fac- tor limiting the rate of erosion was rock resistance rather than the transporting capacity of the waves. The break in slope between the top of the raised beach and the former seaclifl is being obliterated by debris which is accumulating at the base of the clifis and which is no longer being removed by the sea. Current cliff retreat by rock- fall, mudflows, and landslides was esti- mated at 0.7 to 2.0 feet per year, and in parts of Patton Bay the accumulation of debris has obliterated 78 percent of the original break in slope in 15 months. Evidence of two relative sea-level changes before 1964 was found in Pat- ton Bay. At a high stand of sea level lasting unti1 about 2000 B.P. (before present), an older raised beach was formed which, over a distance of 5 miles, shows 410 feet of deformation relative to the present sea level. Peat deposits exposed by the 1964 uplift also record a low sea level that lasted until at least 600 B.P. The 1964 raised beach was used to test the accuracy of identification of former sea-level elevations from raised beach features. The pre-1964 sea level could be accurately determined from the height of the former barnvacle line, so an independent check on high-water level was available. The most reliable topographic indicator was the eleva- tion of the break in slope at the top of a beach between a bedrock platform and a cliff. Even here, the former sea level could only be identified Within 5 feet. The breaks in slope at the top of gravel beaches were found to be poor indicators of former sea level. On Montague Island, evidence of former high sea levels appeared to be best preserved (1) as raised bedrock platforms on rocks of moderate resist- ance in slightly sheltered locations and (2) as raised storm beaches where the relief immediately inland was very low. H1 H2 On March 27, 1964, an' earth- quake shook parts of Alaska and caused great devastation in inhab- ited areas. However, maximum up- lift took place, and large bedrock faults were active, on Montague Island, an uninhabited island off the southern coast of Alaska. The southern end of Montague Island was uplifted as much as 33 feet or more during the earth- quake (fig. 1), and two active faults several miles long caused many landslides and broke or dis- placed many hundreds of trees (Plafker, 1967). Around the coast, marine deposits and a marine- abraded bedrock platform were lifted above sea level and exposed to subaerial processes. The uplift associated with the earthquake has provided physio- graphic conditions on a scale much larger than could be simulated in a laboratory and with the added ad- vantage that the features and changes were natural. The bay- head deposits of MacLeod Harbor (figs. 1, 2) on the northwest coast of the island provided an undis- sected surface, about 1 square mile in area and with a slope of only 20’, which had been suddenly uplifted 33 feet. On this surface, conse- quent drainage channels were ini- tiated and were rapidly eroded vertically and laterally; thus, a unique opportunity was provided to study the effects of rapid uplift on the mode and rate of fluvial processes and the resultant valley forms and long profiles. The earthquake also brought to view a bedrock platform and asso- ciated deposits, which allowed observation and measurement of features that are usually under water and therefore difficult to study. ALASKA EARTHQUAKE, MARCH 27, 1964 INTRODUCTION On both sides of the island the amount of subaerial modification of the 1964 raised beach could be measured. Especially important was the opportunity to study the area only 15 months after the earthquake. In the MacLeod Har- bor area, rivers in soft sediments have adjusted so swiftly that evi- dence of the postearthquake his- tory of the area will soon be re- moved by erosion. Subaerial degradation of the raised beaches and abandoned seaclifl's is so rapid that within a few years little of the original marine form will be left unchanged, and the growth of veg- etation, which has already begun on the beach, will hasten its obliteration. Fieldwork was done in 1965 at two locations, MacLeod Harbor and Patton Bay (fig. 1). The proj- ect was a reconnaissance and, al- though the study was concerned with postuplift changes in raised beaches, differences in emphasis re- sult from physiographic and tec- tonic differences between the two bays. The work was supported by a grant from the National Science Foundation, administered by The Johns Hopkins University. We would like to thank Professor M. G. Wolman of The Isaiah Bow— man Department of Geography, The Johns Hopkins University, for his continuous scientific and administrative help; and A. T. Hohl, Department of Geology, Princeton University, and Thomas Dunne, Department of Geog- raphy, The Johns Hopkins Uni- versity, for their assistance in carrying out the fieldwork. Mr. C. LaBounty, a longtime seasonal resident of Montague Island, pro- vided valuable data on pre— and postearthquake conditions in the bay-head deposits. PHYSICAL SETTING Montague Island, one of the outermost islands of Prince Wil— liam Sound, lies off the southern coast of Alaska at lat 147.5° W. and long 60° N. It is about 50 miles long, and has a maximum width of 15 miles and a minimum width of less than 5 miles. The island is rugged and mountainous, and a chain of peaks forms its back- bone at an average altitude of about 2,500 feet. Cliffs which occur almost all around the island make the coast dangerous. The 1964 earthquake elevated and ex— posed a flat bedrock platform which encircles the island; the former marine cliffs, now beyond the reach of the sea, were broke-n by a few inlets and sandy beaches that allowed access to the island. Travel across Montague Island is diflicult. It is uninhabited except for seasonal hunters and loggers; there are neither roads nor even paths, except those made by bears. The upper parts of the mountains are rocky and are snow-covered all year. Below the peaks the land is either steep and covered with alders or stunted conifers, or the slope becomes so gentle that water- logging prevents tree growth, and patches of wet peat interspersed with pools of water form muskeg. At altitudes below about 1,000 feet, the coniferous forest forms a dense cover down to sea level. Only where the slope is gentle enough for muskeg to form or where it is so steep that it is totally unstable is there any break in the forest. Most of the island has not been logged and is covered with apparently virgin forest contain- ing conifers more than 6 feet in EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H3 148° W 147° W I I 160° 150° 140° 130° N 60° T ! y0 Montague | “ Kodiak a K FIC OCEAN R0 , , ----- \\ D // ’ \ \ O 200 400 MILES / \ |—|__: Qg ' / \\\ / \ I \ l x I \ I \ I \ , 1 / A / I / / l / . / / z , / // // / / / / ,1 O / é / v. V . (9V / x ,’ ’0 0 0 l A? , é / éo _ / I} ’ EXPLANATION /“ UPLIFT, lN FEET 60° N , 7’ \ >30 fl \ ‘ l ’ \ , 20—30 Patton Bay ‘ \ 10—20 M a c L e o d H a r b B \ a ° ’ E /, <10 .« / / / v _ U ——————— ’ Fault active in 1964 earthquake Dashed where probable. U, upthrown side ’ ' 0 5 10 MILES / 34‘ L ____<__| l $2 ' 1.-——Index map of Montague Island showing amount of uplift (after Plafker and Mayo, 1965). A, area of figure 12; B, area of figure 17. 280-547 0 - 69 - 2 H4 diameter and over 400 years old. The trees are as much as 200 feet high, and there is a thick, usually thorny, undergrowth. The forest floor is covered with shallow ponds and bogs; large fallen tree trunks, many of them rotten, lie across one another in a. chaotic pattern and provide unstable footing. Even on the postearthquake raised beaches, plant growth is rapidly establishing itself; only 15 months after the earthquake, con- ifer seedlings were growing on surfaces that were formerly below sea level (G D. Hanna, 1965, oral commun) Most of the rivers on Montague Island are narrow, steep torrents that rush down the slopes in a ser- ies of rapids and waterfalls. Only the Nellie Martin River has achieved any substantial flood- plain development, and it flows through a wide valley, whose floor near the coast is covered with thick peat deposits that form vertical riverbanks as much as 10 feet high. METHODS Mapping was done with a plane— table and telescopic alidade. Beach and stream profiles were surveyed BEDROCK Little is known about the bed- rock of Montague Island; most ex- posures are covered by forest and muskeg or, at higher elevations, by scree and snow. The recent uplift has improved accessibility to the coast and has brought nearly con- tinuous exposures of bedrock into ALASKA EARTHQUAKE, MARCH 27, 1964 with a hand level or Abney cli— nometer, and distances were meas- ured by pacing. Preearthquake mean high water was assumed to be at the top of the abandoned bar— nacle line. (Plafker, 1965, p. 1—2). The level of postearthquake mean high water was obtained from the US. Coast and Geodetic Survey predicted tide tables. These agreed, within 1 foot, with the position of the new barnacle line where it was observed. Grain sizes of the MacLeod Harbor bay-head deposits were de- termined by counts (Wolman, 1954) of 100 pebbles in stream beds and by sieve analyses of fine materials. On beach profiles we measured the size of the largest stones in a section and estimated visually the percentage of material finer than 1 inch in diameter. Per- meabilities were determined ap- proximately by measuring the time for 1 inch of water to infil- trate; the water was contained by a metal ring. Discharge of streams in McLeod Harbor was calculated as Surface velocity in center of stream >< mean depth X width. This calculation gives values ap— proximately twice the true value, because the centerline surface ve- GEOLOGY view along the shore. These strata are thought to be of Tertiary age (Plafker and MacNeil, 1966, p. 66—67). A study of the coastal exposures indicates that the bedrock of the island consists of a sequence of in- terbedded thin layers of black ar- gillite, thick beds of graywackelike sandstone, and occasional beds of locity is approximately twice the mean velocity of the whole cross section. Discharges were measured at a time of relatively low flow (see table 1, p. H10) but were cor— rected for daily variations by multiplying by a standardizing factor: Discharge of stream 1 on June 20, 1965 Discharge of stream 1 on day of measurement To avoid confusion, the expres- sion “1964 sea level”, has not been used in this paper. Instead, the expressions “preearthquake sea level” and “postearthquake sea level” are used. A sea level higher than the post-earthquake sea level is called a raised sea level, and a sea level lower than the preearth- quake sea level is called a low sea level. However, inasmuch as the earthquake took place in 1964, it is correct, and gives rise to no am- biguity, to refer to the uplifted beach exposed by the earthquake as the 1964 raised beach. The changes in relative sea level are probably caused more by move- ment of the land than by absolute change of sea level, but in the ab- sence of positive evidence it is cus- tomary (Sparks, 1960, p. 211) to refer to the land as fixed in rela— tion to sea-level changes. conglomerate and volcanic rocks. The sedimentary rocks are uni- formly bedded and have flat paral- lel bedding planes. The coarser clastic rocks are of two types: (1) conglomerate having coarse round- ed quartz and rock-fragment clasts set in a gray sandstone matrix and (2) breccia of black angular argil- lite chips, also in a sandstone EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H5 matrix. The volcanic rocks were not seen in an outcrop, but hillside boulders and beach cobbles indi- cate that they consist of buff and green medium- to fine-grained ag- glomerate. In the Patton Bay region, light-gray volcanic inter- beds occur. Two major faults, both active in the 1964 earthquake, have been mapped on the island (Plafker, 1967, fig. 2). The faults trend roughly parallel to the island axis and have dips of 50°~85° NW. The Hanning Bay fault is well exposed for about 4 miles between Han- ning Bay and MacLeod Harbor; it has a maximum vertical displace- ment of 13 feet in bedrock and 17 feet in unconsolidated beach ma- terial (Plafker, 1967). Inland, the fault can be traced by landslides, fallen trees, and earth fissures sev- eral feet deep and 1 foot wide. The Patton Bay fault can be traced for about 22 miles, but for most of that distance it is con- cealed beneath landslide debris, its maximum measured vertical dis- placement is 20 to 23 feet (Plafker, 1967). This fault follows and ac- centuates a preexisting topo- graphic break in slope. Near the Nellie Martin River, the Patton Bay faultline is characterized by a linear scarp covered with debris, including fallen trees, and modi- fied by numerous landslides. The faultline scarp varies in height from 100 to 300 feet and has an average gradient of 40° made up of a series of irregular steps. At the top of the debris slope a clean exposure of rock 5 to 20 feet high slopes at an angle of 65° to 75°. There is no evidence of mineraliza- tion and no gouge. Surficial deposits comprise re— cently uplifted shallow-water ma- rine silt and sand, coarse beach sand and gravel, and fluvial gravel. Extensive peat and some glacial deposits are also present. RECENT UNCONSOLIDATED DEPOSITS IN MACLEOD HARBOR At the head of MacLeod Har- bor, uplift has exposed unconsoli- dated sediments which formerly occupied beach, intertidal, and subtidal zones. Three main sedi- ment types are recognized: sand, silt, and gravel. The areal distribu- tion of these types is shown in fig- ure 2. The deposits occupy a roughly rectangular basin that deepens seaward; they themselves also thicken seaward, reaching a maximum exposed thickness of 18 feet, but their total thickness is probably much greater. The sand occurs mainly on the south side of the bay and has a gradational contact with the silt. The silt occupies an area which was formerly occupied by quiet water in the deepest part of the bay head. The upper surfaces of the silt and sand were formerly partly intertidal and partly below spring low water level. SAND The sand of MacLeod Harbor is dark gray and uniform through- out the deposit. It stands in ver- tical cliffs as much as 20 feet high; for the most part it is well bedded, the thickness of individual beds ranging from 6 inches to 2 feet. The beds are continuous for a min- imum of 200 feet and show the in- ternal crossbedding that is char- acteristic of a mainly fluvial sedi- ment. Scour channels and filled chan- nels about 4 feet wide are common. Convolute bedding occurs at all levels in the sand, and some convo- lutions are truncated by overlying uncontorted beds as much as 12 feet below the present surface. The cause of the convolutions is not known, but the possibility that they were formed by thixotropic transformation p r 0 d u c e d by earthquakes prior to the 1964 earthquake best fits the known facts. Occasional well-rounded pebbles as much as 3 inches in diameter are found within the sand. Through- out the deposit there are discon- tinuous stringers up to 2 inches thick of small wood pieces, shells, seaweed, and pine needles, similar to present high-tide swash accumu- lations. These stringers probably indicate former positions of high tide. The permeability of the un- saturated sand is about 25 inches per hour. SILT The silt is black and uniform in composition, except for the per- centage of sand present, which in- creases toward the contact with the sand. The silt shows an apparent lack of bedding and internal struc- ture, although occasionally thin parallel bedding was suggested by difierential erosion along stream- beds. Clam shells (mostly M acona nasuta) that are found through— out the deposit are concentrated through erosion on the silt surface as a lag deposit. The permeability of the silt is about 0.01 inch per hour. GRAVEL Gravel deposits occur in three forms—as beaches around the in- land edge of MacLeod Harbor, as fans around the former mouths of streams entering the bay, and as lenses and discontinuous sheets at or near present sea level. The beach gravel consists mainly of local rock types and occasional exotic pebbles of granodiorite, granite, or hornblende gneiss. The gravel is very well rounded and well sorted and has only minor amounts of sand matrix. The mean size of the pebbles varies widely on the beaches around the bay; on some beaches more than 80 percent ALASKA EARTHQUAKE, MARCH 27, 1964 Campsne (Approximate) EXPLANATION _% l-_‘ | _ C—14 sample L“ W—1768 A; . . Sand Silt Contact Harbor 15— — — — Contours of land surface after postearth- quake downcutting, in feet above post» earthquake MHW Long-dashed where extrapolated ________ 20———— .... ..... Contours of preearthquake land surface, in feet above postearthquake MHW Dotted where extrapolated WV Eroded margin of preearthquake land surface 4%”...— Preearthquake storm beach Main stream Stream number SECTION 1 . Locality discussed in text 0 500 1000 FEET L____J_.._l Z—Bay-head deposits in MacLeod Harbor and June 1965 elevations of their upper (preearthquake) and lower (postearthquake) surfaces. EROSION AND DEPOSITION ON A RAISED BEACI-I, MONTAGUE ISLAND H7 are smaller than an inch in diame- ter, but on other beaches 80 per- cent are larger than 4 inches in diameter. On any individual beach the size range is much less, al- though generally pebble size de- creases from the storm beach to- ward the sea, and sorting becomes better in the same direction. Per- meabilizties are about 450 inches per hour. Behind the gravel beach around most of the bay is a storm beach which is characterized by little or no sand matrix, poorer sorting, coarser grain sizes, steeper gradi- ents on the upper surface both landward and seaward, and ex- tremelv high permeabilities. The river fan gravel deposits also consist of local rock types and vary in mechanical composition from their apexes in gaps through the storm beach to their seaward edges. At their apexes they are similar in composition to the beach gravel. The percentage of sand matrix increases and the size of the pebbles decreases toward the seaward edges of the fans. Near the lower part of the fans (section 1, fig. 2) the deposit is of interbedded gravel and sand, the gravel beds ranging in thickness from one pebble (1/2 inch) to 12 inches. Individual pebbles vary from 1/8 to 5 inches in diameter; many are disk shaped and most are well rounded but have low sphe- ricity. The gray-sand matrix makes up 40 percent of an individ- ual gravel bed; wood fragments and abraded logs are common. No graded bedding or cross bedding was observed. The upper and low- er contacts with adjacent sand beds are uneven, and the gravel lenses out, the thinnest beds having the least lateral continuity. The sand beds are discontinuous and from 1 to 10 inches thick; planar crossbeds are present in a few beds. The total section at this loca— tion (fig. 2) consists of 60 to 70 percent gravel beds and 30 to 40 percent sand beds, not counting sand matrix in the gravel beds. The gravel lenses and discontin- uous sheets occur only at or below 1965 mean sea level. They are found beneath the sand and silt deposits and have been partly eroded and redistributed by post- earthquake stream action. They form a more or less continuous sheet where the streams have been eroded down to 1965 sea level. Near the present high-water level EROSION AND DEPOSITION FORMATION OF BAY-HEAD DEPOSITS Both the surface composition and the topography of the pre- earthquake deposits at the head of MacLeod Harbor (fig. 2) seem consistent with conditions of nor— mal intertidal and shallow-water deposition by rivers at the head of an inlet. At their seaward limit the deposits seem to have formed a OF BAY-HEAD DEPOSITS steeper slope. The evidence for this statement, although indirect, is fairly conclusive: (1) the U.S. Geological Survey 1263,360 topo- graphic map and the U.S. Coast and Geodetic Survey charts show depths close inshore that indicate a steep offshore slope, and (2) be- fore the earthquake local fisher- men were able to anchor close in- shore in areas that are still close to where the streams flow into the sea, gravel lenses occur beneath the streambeds. These lenses are as much as 3 feet thick, and in places a sand bed lies between them and the gravel sheet on the present surface. The lower gravel lenses have a sand matrix and a maximum pebble size smaller than that of the present stream gravel. BEACH DEPOSITS IN PATTON BAY Bay-head deposits are not pres- ent in Patton Bay. Instead, uplift has merely widened the existing beach deposits or bedrock plat- forms. The distribution of the coastal forms is discussed on page H2O in relation to individual beach profiles. The coastline around Patton Bay consists mainly either (1) of wide gently sloping sand beaches with gravel storm beaches sur- mounted by large driftwood ac- cumulations at their inland edges or (2) of cliff-backed bedrock platforms with or without narrow sand or gravel beaches and drift- wood at their inward margins. Between zones of these two main coastal types, small areas of large boulders lie on bedrock. IN MACLEOD HARBOR the seaward edge of the silt. A steeper slope, therefore, very prob- ably existed at the seaward edge of the deposits, although whether its form was erosional or deposi- tional is unknown. It is also con- sidered probable that this slope was partly exposed by the uplift associated with the earthquake, be- cause major streams very rapidly trenched the newly uplifted de- H8 posits (C. LaBounty, oral com- mun., 1965) ; such trenching would be favored by the exposure of a steep slope by uplift above sea level. Figure 2 shows elevations of the bay-head deposits in June 1965. Surveyed contours of the remain- ing parts of the upper surface (in interfluve areas) have been inter- polated in stream-eroded areas to reconstruct the preearthquake topography of the deposits. Sur- veyed contours of areas of post- earthquake stream erosion have similarly been interpolated to esti- mate the final postearthquake to— pography of the bay-head depos- its. The upper surface could still be fairly well defined in June 1965, but the former seaward limit of the deposits is less definite. Four independent levels from in- terpolated 1965 mean high water to the preearthquake barnacle line give a mean measurement of 33 feet for the uplift. The mean tid- al range in MacLeod Harbor is 9.2 feet, and spring low tides were formerly 19 feet above the 1965 mean high water. The lowest ele- vations, which are still preserved on remnants of the preuplift sur- face, are 13 feet above the post- earthquake mean high water on sand and 9 feet above on silt. Thus, the top of the seaward steep slope apparently was at least 10 feet below spring low water (1963) on the silt and at least 6 feet below on the sand. The top of the steep slope may be close to the approximate position of the +10-foot contour of the upper surface shown in figure 2. The presence of a gravel sheet or sheets, both below the 1965 stream courses and in the outliers of silt and sand, suggests that a former stream system similar to the one which is now developing produced these gravel sheets. Since ALASKA EARTHQUAKE, MARCH 27, 1964 the gravel sheets could be traced down to the 1965 mean sea level, they very probably indicate a for- mer relative sea level that was at least as low as the 1965 one and that predates the finer deposits. Cores taken in the present inter- tidal zone (Barrett, 1966, p. 997) show a series of pebbly layers ex- tending down as low as 1 foot above mean lower low water. These sections seem to show deposition of 25 to 30 feet of sand in and close to the intertidal zone, and imply a fairly steady rise of relative sea level before the earthquake. A sim- ilar steady rise has been reported for other parts of Prince William Sound (Plafker and Rubin, 1967). Some indication of the rate of rise of relative sea level has been obtained by the dating of two wood samples. C—14 sample W— 17 68, from 4 feet above postearth- quake mean high water (section 2, fig. 2), was interbedded with sand that probably was deposited above extreme low water level (14 feet below the 1965 mean high water). The radiocarbon age of this sample is 820:200 B.P. 0—14 sample W— 1764, from a tree in the position of growth approximately at the level of the preearthquake barnacle line, has a radiocarbon age of 380:200 B.P. and indicates that mean high water was then at least 5 feet lower than the preearthquake mean high water. Rates of fall of relative sea level calculated from samples W— 1768 and W—1764 are 00331—0008 foot per year and more than 0014:0613, respectively. STREAM EROSION Bay-head silt, sand, and gravel were studied only at the head of MacLeod Harbor. Since the uplift of about 33 feet in March 1964, the area of intertidal and shallow— water deposits has been consider- ably dissected by rivers, few of which rise within the area that was formerly below sea level; the coastal perimeter of the uplifted deposits has been trimmed by wave action. By July 1965, about 50 per- cent of the available deposits had been regraded to a new level, ac- cordant with the present sea level. Most of the regrading has been done by lateral erosion of the largest rivers (fig. 3). Long profiles were made of 20 streams (numbered 1—17 and 20— 22, table 1), particular attention being paid to the position of knick- points. At least three cross sections were made of each stream in order to estimate the total volume re- moved from each by erosion. A grain-size analysis was made of samples from near the mouth of each stream, and a relative value of low—flow discharge was calcu- lated. Additional information about the earlier course of erosion was obtained from US. Coast and Geodetic Survey aerial photo- graphs taken 8 weeks and 21 weeks after the earthquake. Interpretation of the informa- tion collected has been in terms of the overall course of erosion by the rivers, including the recession of knickpoints. This part of the in— terpretation is essentially a study of fluvial processes in an unusual environment. The data have also been studied in relation to the modification of a raised platform. The rapidity with which such a raised platform is altered, and the agents responsible for it, are dis- cussed on pages H15—H16. DATA COLLECTED The banks of streams 1—9, 15, and 16 (table 1) are silt; the banks of streams 10—14 and 20—22 are sand; and the banks of stream 17 are gravel. Bed material is not in— frequently coarser than bank ma- t e r i a l . Low-flow discharges ranged from zero for the smallest EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H9 3.—Panorama of eroded bay-head deposits in MacLeod Harbor. streams in sand and 0.035 cfs (cu- 200 l | I l EXPLANATION bic feet per second) for the small- (187,000,000) est stream in silt (stream 6) to 301 —Tppervconfimmit cfs (stream 10). The volume of 180 __ _ material removed by all streams Estimatedtotal availame from the time of the earthquake (170,000,000 cu m to late June 1965 (fig. 4) was calculated from direct measure- 160 * (153,000,000) _ ments. For the larger streams “—w (streams 9, 10, 12, 21, and 22), es— . timates of volumes removed at 140 L — earlier stages were calculated from aerial photographs, which sand showed the width of erosion, and from planetable maps made in June 1965, which gave the depth of erosion. For these large streams, errors arise mainly from incorrect interpolation of the initial eleva- tion of the surface before stream erosion (fig. 2). This error in ini— tial elevation is unlikely to be more than 1 foot near stream mouths where thicknesses removed are 10 to 15 feet; it will be correspond- ingly smaller upstream where the rivers are narrower and thick- nesses of material removed are 0 to 5 feet. Measurements of widths of erosion, made from field meas- urements and from aerial photo- graphs, are considered to be much 20 _ more accurate than the depth measurements. For small streams, st 9 . . - o ream errors arlse mainly from the 1ntr1- , 40 cacy of variatlon vof Width and o lo 20 30 40 50 60 70 bifurcation, but all major tribu- TIME AFTER EARTHQUAKE, IN WEEKS taries, were included in our calcu— lations. An overall error of :10 rce is c nsidere a enerous . , _ pe _nt 0 d g 4.—Relatlonsh1p of volumes removed by fluvial erosion to time elapsed since the max1mum allowance (95 percent earthquake. Lower confidence limit 120 _ Silt O Gravel 100 —- From measurements 80 -- VOLUME REMOVED, IN MILLIONS OF CUBIC FEET 40— Stream 10 H10 ALASKA EARTHQUAKE, MARCH 27, 1964 TABLE l.—~—Summary of data for streams in MacLeod Harbor [’, no knickpoints apparent in stream] Total volume Total volume of Distance of Relative Cross- oi material material removed knickpoint Width at low-flow sectional removed from Size of bed (millions on it) Type of bank Stream number from mouth mouth discharge area at the time of the material ~———— material Remarks (feet) (feet) (cfs) mouth earthquake to D50 (in.) May 1964 Aug. 1964 (sq ft) June 1965 (8 weeks) (21 weeks) (cu ft) 1 _____________ 76 25 0. 37 97 5, 025 2. l _________________ Silt _______ 2 _____________ Many 22 1. 73 180 6, 160 1. 06-__- ____________ ___do _____ knick points 3 ______________ 75 15 . 67 103 6, 250 2. 64--__ ____________ __-do _____ 4 ______________ 66 23 . 051 128 6, 300 Silt _________________ - _ _do _____ 5 ______________ 63 18 . 035 78 5, 250 ___do__- ____________ __-do _____ 6 ______________ 255 32 . 35 198 41, 500 . 49-___ ____________ _--do _____ 7 ______________ 335 20 . 50 85 22, 400 . 74- __ _ ______________________ 8 ______________ 1, 150 72 5. 71 480 361, 000 1. 01____ ____________ __-do _____ 9 _____________ 0 190 73. 0 2, 600 3. 20><106 .96-___ 0. 81-_ 2. 17__ ___do ----- Gravel bed. 10 ____________ 0 1, 625 301. 0 25, 200 36. 7 X 106 1. 56- _ _ - 3. 54- _ 9. 50- _ Sanlfl and Do. s' t. 11 ____________ * 70 0 650 164, 000 Sand____ ____________ Sand-..__-- 12 ____________ 0 665 80. 8 9, 650 18. 0X106 1. 27____ 3. 18__ 12. 8__ -_-do _____ Do. 13 ____________ * 130 1. 10 1, 220 710, 000 Sand___- ____________ ___do _____ 14 ____________ * 30 0 120 19, 400 __-do___ ____________ --_do _____ l5 ____________ 50 ______ 0129 ____________________________________ Silt _______ Data in- complete. 16 ____________ 50 ______ . 0094 ____________________________________ -_-do _____ Do. 17 ____________ 110 39 1. 09 176 19, 600 2. 9 _________________ GraveL--- 20 ____________ * ______ 2. 16 ____________________________________ Sand ______ 21 ____________ 0 490 9. 30 6, 860 6. 04X 10‘ -------- . 92 2. 31 ___do _____ Gravel bed. 22 ____________ 0 1, 450 52. 3 21, 700 29. 0X 10° ________ 5. 20 24. 7 __-do _____ Do. Total ___________________________________ 94. 3X 105 ________ 13. 6 51. 5 __________ 100 confidence). Streams appear to erode at a more or less uniform rate until they approach the limits set by topography or by neighbor- ing streams, at which time their rate of erosion becomes much less. Stream 10, for example, is con— tinuing to erode at a uniform rate, Whereas stream 22 appears to have neared its limit. The overall re- lationship of volume removed to time elapsed since the earthquake, for these streams and for all the streams combined, is shown in figure 4. Volumes removed 'by erosion up to June 1965 have been calculated by summing cross-sectional areas removed along the length of the stream. Figure 5 and table 1 show volumes removed, low-flow dis- charge, bed and bank material, and other data. Each point in figure 5 has been distinguished as to bank material, and streams flowing between silt banks have been fur- ther distinguished as to bed—ma- terial size on the basis of a pebble count. 10— ,_. O / Zero discharge A .0 ._. I 0.01 — VOLUME REMOVED, IN MILLIONS OF CUBIC FEET EXPLANATION A Sand-banked stream X 2.6 Silt—banked stream 0 2.9 Gravel-banked stream (Stream 17) s Figures for silt and gravel streams show median grain size of bed material, in inches 1 l 0.001 I ‘ 0.01 0.1 1.0 10 100 1000 RELATIVE LOW-FLOW DISCHARGE. IN CUBIC FEET PER SECOND 5.—Variati0ns in erosion rate with low-flow discharge of streams in MacLeod Harbor. Arrow shows direction of increasingly coarse bed material. INFLUENCE OF UNDERLYING MATERIAL 0N STREAM EROSION Three major characteristics of the underlying material appar— ently control stream erosion. The first is the permeability of the deposit, which determines the drainage area required to produce surface runoff. For example, sur- face flow commonly appears on the silt immediately below its up- stream contact with a gravel bed, and results from the large differ- ence of permeability. Differences in permeability probably also cause large differences in the ratios of high—flow to low-flow discharge. The second control is the varia- tion in resistance to erosion. This EROSION AND DEPOSITION ON A RAISED BEACH, control might be expected to result in marked differences in drainage densities and networks. A dif— ference in networks is apparent (see fig. 7 for silt and fig. 8 for sand), but differences in drainage density are masked because many streams receive most of their water supply from streams outside the area of uplifted deposits. Their drainage density is therefore partly determined by the number and spacing of the feeder streams. The third control is the effect of bed armoring. Armoring is wide- spread in larger streams, mainly with gravel deposits that are eroded less readily than the bank materials. Extensive armoring is usually associated with streams that have almost reached their limit of possible downcutting, so it is not immediately clear, from the data in this area, whether the slow rate of downcutting is a cause or an effect of the armoring. Differences in erosion rates of streams flowing in silt and in sand banks may be clearly distinguished in figure 5. The distinction seems to depend mainly on the differ- ences in the ratio of high- to low- flow discharges, but the greater erodibility of sand may be a con— tributory cause. The silt channels can be distinguished by the size of the bed material except for stream 2, in which the small volume of material removed is probably due to the fact that the stream’s knick- point had already reached gravel and it could not cut back farther. Note that the plot of the gravel— banked stream 17 falls with silt- banked streams with the same bed-material size. The data may indicate that armoring of the bed with coarser material tends to re- duce the rate of erosion and that recession of the knickpoint is greatly slowed when it reaches the coarse gravel deposits which form the preearthquake storm beach. 280-547 0 - 69 - 3 MONTAGUE ISLAND H11 6.—Typical stream with silt banks (stream 8, table 1), MacLeod Harbor. Note the conspicuous white clam shells on the silt surface. STREAMS IN SILT Initial stream courses in silt seem to have been consequent and to follow the directions of greatest slope, as exemplified by the streams that have not migrated laterally and that still follow the reconstructed consequent direc— tions. The initial directions of larger streams, even though they have since moved laterally, also correspond to reconstructed conse- quent directions. Most of the streams in the silt are fed by run- off from the hillside along the north side of the harbor. A sharp division can be made between streams that can transport gravel effectively and those that cannot. Those that cannot carry gravel effectively are characterized by an abrupt ending of the chan- nel at its upper end, at the lower margin of the gravel beach or at a gravel fan. The great difference in permeability between silt and gravel emphasizes this feature. In progressively larger streams, the bed material tends to change (pro- vided suitable material is avail- able) from silt to clam shells to gravel. The amount of coarse armoring material and the ability to move it thus increase together, although figure 5 shows that the armor has a net retarding effect on erosion. Small streams in silt (fig. 6) generally have steep banks, narrow valleys, and very marked knick- points in their longitudinal pro- files (fig. 7). Streambank gradi- ents locally decline from the usual 60°—7 00 to 10°—20°, where the silt is saturated. This decline is asso- ciated with extensive slumping and earthflows. The saturation is usually caused by unchanneled seepage from the edge of the gravel beach. Figures 6 and 7 i1- lustrate typical silt valleys. STREAMS IN SAND The drainage areas of many smaller streams in sand (fig. 8) lie entirely within the bay-head sand area. These streams have no gravel or clam shells on their beds. Clearly defined channels are small and are confined to the seaward ends of the valleys. In the smallest H12 ALASKA EARTHQUAKE, MARCH 27, 1964 EXPLANATION E Upper silt surface E Lower silt surface I Knickpoint Lu X < C a 30 <— Preearthquako 200 Sq ft I MHW IE \ < Lu I- E7; 5 2° * B 2 LL 166 sq ft Z Lu —_ silt Surf > 3 \ ace 3 I 10 ‘\ \ < E K ‘ k ' t \\ A 2 me pom \ 198 sq ft 3 . t \ C < 0 | l | J 1kmckpomt| A > E 350 300 250 200 150 100 50 0 0 5 10 F E ET m DISTANCE, IN FEET |_l; NO VERTICAL EXAGGERATION 7.—Map (A), long profile (B), and cross sections (0) of typical stream in silt (stream 6, table 1) in bay-head deposits of MacLeod Harbor. Letters on profile show position of cross sections. Numbers in cross sections refer to cross-sectional area eroded. EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H13 Cross ‘ Cross [I section UDDer surface section l B Cross section 4% 1 Valley bottom APPROXIMATE NORTH — a l I U er surface 7/4513?“ DD Damp sand '1 I I: ,0, l . 1/41/5230 (90 r‘ u 77? Wet sand 0 50 100 FEET i l UJJ—LLU-LU .—.¥ Channel bank UPPER SURFACE Cross Cross Cross section section section Positiflj‘flf “J —— , ’ m 5 L o \\e*J b0“ \13 I O 50 100 FEET Postearthquake M HW LONGlTUDlNAL PROFILE (Following the most northerly tributary valley) A Position of upper surface —————————————————————— 7—- Valley bottom B _ _ _ flsiLiOMMErflrfice FEET — _ — h 20 w_bottom 10 C JEEI £13.“, i__l__l Valley bottom 0 O 10 20 FEET CROSS SECTIONS 8.—P1anetable map, long profile, and cross sections of a typical stream in sand (stream 13, table 1), MacLeod Harbor. H14 valleys, no channels at all are pres- sent. The downstream parts of the sand valleys are broad and have flat bottoms. These are joined, by sharp breaks in slope, to side slopes that are uniformly at the angle of repose. At the top of the side slopes there is a second break in slope, above which is an almost flat upper surface, which roughly corresponds to the original sur— face of the sand before and im‘ mediately after uplift. The whole flat section of the valley floor may be wet, even above the head and outside the banks of the channel; basal sapping of the side slopes ap- parently is chiefly responsible for valley widening. The upstream parts of the valleys are V-shaped in cross section and are dry on the valley bottom. The irregularity of the long profile suggests that there may be periods of infilling by wind between periods of fluvial erosion. Figure 8 illustrates a typical small stream valley in sand. Large streams, fed by water from the hillsides above the for- mer sea level, are gravel floored and have steep vertical banks in sand, or in sand and gravel in their upper reaches (but still below the preearthquake water level). Clear knickpoints are not present in the longitudinal profiles of either large or small streams flowing over sandbeds. STREAM IN GRAVEL One small stream (stream 17) flows down the hillside and then across a steep gravel beach di- rectly into the sea. The gravel had some finer matrix and was able to stand, at least for a short period, in vertical banks. The stream had a steep course and a knickpoint zone less well defined than those of the streams in silt (fig. 14, p. H18). It is debatable whether placement of this single example with the streams in silt (fig. 5) is normal. ALASKA EARTHQUAKE, MARCH 27, 1964 L|J x < 8 30 I ,_ g , 20 E ———————— L|J l— 115 '— 1“ face °l (Lego—s” Valley of stream 22 8 E 20 _ Former sgL—”’ / CL Z —— ——'T x x x X/X/‘X _ ~<- v— - ,xr I: ;’ Silt Valley of stream 10 ’ /7V} // 0 I 10 a , x Gravel sheet a: 7 < E // )7X7X7X/57an x r z Gravel she% 7* x x'Wx O l: 0 g 0 1000 2000 3000 4000 E HORIZONTAL DISTANCE, IN FEET Lu 0 8 weeks 21 weeks 65 weeks 0 8 weeks 21 weeks x’-—>¥——i> r ——> X ,Y-bx - ‘—‘>X x<———’ Lateral migration of stream 10 65 weeks Lateral migration of stream 22 9.——Cross profile of gravel sheet and lateral migration of streams 10 and 22 (table 1), MacLeod Harbor. Arrows are drawn to scale, and show distances and direction of migration. LATERAL AND VERTICAL EROSION Figure 9 shows gravel sheets that crop out approximately at the surface of the present major streams. One or more gravel sheets at a similar level may extend be- neath the remaining outliers of silt and sand. This preexisting gravel probably has some influence on the rate of downcutting of the streams. At one extreme, the control may be so complete that the present streams are simply being guided in their lateral erosion into hollows in the former gravel sheets and are thus prevented from further down- cutting. At the opposite extreme, the gravel sheets may only slow downcutting temporarily, without tending to shift the streams later- ally. The lateral migration of stream 10 shown in figure 9 might support either possibility, but the reversal of the direction of lateral eral migration by the gravel. Fig- ure 9 can thus be interpreted as a record of the rate of downcutting, which can be dated by reference to the lateral position of the streams in aerial photographs. The rates of lateral and vertical erosion deduced in this way (fig. 10) tend to support the generally held View that lateral erosion mainly takes place after most downcutting is completed. Ob- servations of Mr. C. LaBounty, (oral commun., 1965) indicate that the bulk of the vertical ero- TOTAL AMOUNT OF LATERAL CUTTING, IN FEET migration of stream 22 shows that 3% the stream does not simply move .2 E « into a hollow and stay there; if 33' _ streams did this, they would never fig change their direction of migra- E; — tion. Instead, it is concluded that F 8 0 IO 20 30 4O 50 60 70 lower pos1t10ns of the gravel sheet TIME mER EARTHQUAKE, 'N WEEKS represent later times and that the streams are steadily downcutting without rigid control of their lat- 10.——Inferred rates of lateral and of vertical erosion for streams 10 and 22, MacLeod Harbor. EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H15 sion in the major streams was very rapid, having been almost com- pleted within 48 hours, and that the lateral erosion has been pro- ceeding ever since. Erosion in un- consolidated sediments probably is limited by the rate at which ma- terial can be transported by the stream. Thus vertical erosion and lateral erosion compete for avail- able capacity, as figure 10 seems to show for the first period of 8 weeks. Vertical erosion will result in changes in stream gradient— perhaps the most independent control of transport rate in a given cross section—and gradient is itself controlled by the rate of sediment transport from up- stream. Material eroded from the bed will be redeposited if the slope is lowered too much, and the re— sult will be very little net vertical erosion. No such regulatory mechanism applies to lateral ero- sion, which may therefore continue unchecked until all available ma- terial is out of the possible mean- der belt of the river. Figure 10 illustrates this general course of events, but gives no information about possible mechanisms. MODEL OF EROSION The results of these erosional studies are valid only within a limited framework. It would for example be quite improper to ap- ply them to conditions in which the erosion was limited mainly by the resistance of the materials. Within the framework, however, the overall course of erosion fol- lowing a sudden relative lowering of base level may be approximated by an equation of the form: V=A-Q’(1—e‘“) where Q is the low-flow discharge, in cubic feet per second; V is the volume removed, in cubic feet; If is the time since the change of base level, in weeks; and A, r, and k are constants. For the MacLeod Harbor sedi- ments, the constants A, 7', and k have the following approximate values: Sand: A=1.9><10"; 7-20.75 ; k=0.012. Silt: A=1.7X105 ;7"=0.95 ; [620.012. The differences apparently due to lithology may partly reflect dif— ferences in hydrology rather than in erodibility (see p. H10—H11). Total volume removed seems to be the only consistent measure of the course of erosion. Recession of knickpoints (where present), cross-sectional area at the mouth of the stream, and width at the mouth of the stream all showed much more scatter in their relation to other variables. RECESSION OF KNICKPOINTS The knickpoints in the smaller streams in silt apparently are con- trolled mainly by slight variations in resistance of the material; the streams seem to pick out certain bedding planes which are not oth- erwise distinctive. The knickpoints recede until they finally become fixed at the junction with the gravel beach above; this junction has already occurred for stream 2. On the larger streams, the gra- dient of the stream above the former high-water mark and the gradient on the gravel spread be- low it seem to be more or less inde- pendent. No identifiable knick- points were observed in the long profiles, the only recognizable feature being a local steepening, probably associated with the former beach or storm beach. OTHER EROSIVE AGENTS IN REGRADING OF DEPOSITS The three principal types of ero- sional processes that modified the bay-head deposits were fluvial, marine, and slope processes. The most effective slope processes prob- ably were the slopewash, caused by raindrop impact, and erosion by wind. In the broad expanse of depos- its in MacLeod Harbor, fluvial ero- sion seems to be the most important process. The sea can only act along the perimeter, but rivers act over the whole surface. The volume of material removed by the rivers has been estimated from the planetable map of the deposits (fig. 2). Along the margin of the depos- its, marine erosion produces clifi's as high as 18 feet. Below these cliffs a narrow beach is being formed, coarse material accumu- lating at its top. Some intertidal sorting of material may also take place, but its total effect in Mac- Leod Harbor is thought to be slight. On the silt deposits, clam shells protect small sloping pillars of material; the silt around them is detached by raindrop impact and washed away by surface runoff. At the time of measurement, all these pillars sloped at about 30° to the horizontal in the same direction, which is presumed to be the angle and direction of the driving rain. Such an angle is probably not characteristic of all storms, so the heights measured probably repre- sent only the last major storm. The median vertical height of 50 of these pillars was 1.9 inches (0.16 ft). The total effect of this rainbeat since the time of uplift could be seen as follows: The shipworm (Ban/702a spp.) bores into wooden pilings underwater, but it cannot live within the silt; pilings become honeycombed with these boreholes and eventually break off within an inch of the surface of the silt; a line of old broken-ofl' pilings pro- j ecting out of the present silt beach therefore records the former sur- H16 face level very accurately, and the total erosion of the surface may be measured. The rarity of surface runoff on the sand and gravel is thought to minimize erosion on these deposits by rainbeat and rainwash. The profile of the silt along the line of pilings is shown in figure 11. Both the volume of material removed since the earthquake by surface wash above the top of the low seaclifl’ (an average thickness of 0.7 ft) and the cross-sectional area removed by marine erosion (175 square ft) can be calculated from figure 11. These computa- tions can be compared with the volume removed by fluvial erosion. Volumes removed by the three main processes responsible for the regrading of the deposits are: By rivers (see fig. 4) =87 mil- lion cu ft By rainwash=0.7 ft x area of silt deposits=2 million cu ft By the sea: 175 sq ft >< coast- al perimeter=7 00,000 cu ft. Wind action is locally evident on channel sides in the sand areas only. The total effect probably is a rearrangement rather than a net removal of material, because the ALASKA EARTHQUAKE, MARCH 27, 1964 110 I , 100 — — — Preearthquake MHW Part of profile enlarged below A | I l l 60 0 100 200 300 400 500 600 700 HORIZONTAL DISTANCE, IN FEET 80 I l l I ELEVATION ABOVE ARBITRARY DATUM, IN FEET I l . v l 1 Farther out,silt has i been regraded by intertidal streams ->i F°’I'her 5 “’face of , 7O _ N SIIt —i Ami; _____ \Estmhquak—e MHW Cliff with many/' ========== B small gullies Surface on Junemfl 50 l | l l | I I I | l 240 250 260 270 280 290 300 310 320 330 340 HORIZONTAL DISTANCE, IN FEET 11.——Postearthquake degradation of silt, by rainwash and marine erosion, along a line of pilings. A, Profile of entire beach; vertical exaggerationx 10. B, Part of A, without vertical exaggeration, showing retreat of seacliff. form of the upper surface of the sand is remarkably even and shows no hollows, such as might be ex- pected if deflation had occurred. It is concluded that, in MacLeod Harbor, the agents responsible for regrading are fluvial processes, rainwash, and marine action, in that order. However, a different order of importance might be ob- tained for, say, a narrow gravel beach. EROSION AND DEPOSITION 0N UPLIFTED BEACHES AND ROCK PLATFORMS Beach studies in MacLeod Har— bor and Patton Bay had three 0b- jectives. The first was to provide data on the detailed form of beaches and offshore platforms that are difficult to study under normal conditions; the second was to observe the rates of erosion and deposition on the raised beach as an aid to the general study of raised beaches; and the third was to clarify the tectonic history of the island. These three objectives are closely related and data were collected from (1) profiles perpen- dicular to the shoreline, (2) sur- veys of selected geomorphic fea- tures and of amounts of erosion and deposition, and (3) strati— graphic sections and their associ- ated radiocarbon dates. MACLEOD HARBOR BEACHES BEACH PROFILES Eight longitudinal profiles of preearthquake beaches and up- lifted sea floors were made in the MacLeod Harbor area (locations shown in fig. 12). Most profiles were on rock platforms, but some were on beaches that were com- posed wholly or partly of sand or gravel. Interpretation of the pro- files has been made in terms of the relief found on rock platforms, in terms of marine regrading of beaches, and in terms of knick- point recession and fluvial re- grading. The profiles seem to have gentler EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H17 mm' B H my ay EXPLANATION év e‘ 0 “—fl $4? [ . } v / '4 Silt River Beach profile and number 3 12,—Shoreline lithology and location of surveyed beach profiles between MacLeod Harbor and Fault Cove. H18 ALASKA EARTHQUAKE, MARCH 27, 1964 10 10 Most exposed beach 20 Least exposed beach 30~ \ a 2 ELEVATION RELATIVE TO PREEARTHQUAKE MHW, IN FEET 40 I I I I I I I I I O 100 200 300 400 500 600 700 800 900 1000 HORIZONTAL DISTANCE, IN FEET 13.—Beach profiles around MacLeod Harbor showing the eifect of increasing exposure to waves. Broken lines show effect of local modification by stream action. overall slopes below the preearth- 40 quake mean high water and broad— er rock platforms Where the beaches are more exposed to waves of long fetch. This relationship is shown by the series of five repre- sentative profiles in figure 13. How much of the differences in profiles is due to varying wave exposure and how much, for example, to varying initial slopes of the coast -PreearthquakeMHW~"“"-‘~--~-"~-"-'-6 ............ (M O I ELEVATION ABOVE POSTEARTHQUAKE MHW, IN FEET 8 I ._. o I is unknown- 0 a Postearthquake MHW Sediments from the upper parts (I no ” of both present and former beaches show a consistent change in grain 10 A I I I I I I size down slope; the finer material 4° 8°HORIZON1TTL DISTANf: m FEET2°° 24° is found in deeper water and on 0 lower gradients. In comparing ' Estimated preeanhquake spring low tide profiles 3 and 8, a given grain size 0! 14 (right).—Beach profiles 3 and 8 (figs. 12, 13), and stream profile 17 (table 1), MacLeod Harbor. A, Beach profile 3 and stream profile 17, vertical exaggeration X 4; B, parts of beach profiles 3 and 8 referred to preearth- quake spring low-tide level, no vertical exaggeration. ELEVATION BELOW 5 PREEARTHQUAKE SLWI IN FEET ._. u- HORIZONTAL DISTANCE, IN FEET EROSION AND DEPOSITION ON A RAISED on profile 3 is found on a steeper gradient and in shallower water than on profile 8. This difference may be partly explained by differ- ences in sizes of material availa- ble, but it fits in well with the fact that profile 8 is the more exposed (fig. 13). Profiles 3 and 8 seem to have a step composed of different mater- ials deposited over an earlier sur- face. The top of the step of both profiles is about 31/2 to 4 feet be- low the level of former spring low tides (fig. 14). The top of the steps seems to represent the shelf of deposits that was being built at the preearthquake sea level be- low the zone of effective wave transport. The relatively shallow depth compared with that for the bay—head sediments (about 10 ft below preearthquake spring low water) is consistent with the greater coarseness of the material in the beach profiles. On the more exposed beaches (profiles 1, 2, 5), the slope of the rock floor is prob- ably too gentle for the extension of sediments from the preearth- quake sea level to have produced a detectable step. The exposure of a former surface, on which this step rests, shows that at some previous time there must have been a sea level in the area at least as low as the postearthquake level. Such an earlier low sea level also explains the existence of gravel sheets be- neath the bay—head deposits, al- though no dateable material was found. A marked break in slope, similar in form to a wave-cut notch, was seen on the seaward side of an abandoned stack near profile 8, and in the rocky promontory at profile 7. This break is, at both places, about 16 feet below the pre- earthquake barnacle line. The break, which presumably was formed at or near a former high- tide level since it does not seem to BEACH, MONTAGUE ISLAND H19 mama 15.—Rock platform showing difierential erosion following the structure. 16.—Rock platform near beach profile 8 (fig. 12), MacLeod Harbor. be structural, may mark a former sea level that is between the pre- and postearthquake levels, al- though it presumably predates them. Bedrock platforms are notably smooth in overall outline, but in detail vary from highly irregular surfaces with as much as 5 feet of local relief to much smoother sur- faces with less than a foot of local relief. All the differences seem to emphasize lithological and struc- tural differences in the near-verti- cal beds (figs. 15, 16). REGRADING OF THE BEACHES The uplifted MacLeod Harbor beaches have been modified by both marine and fluvial processes, but slope processes seem to have had little erosional effect. H20 Marine processes act only in the zone below the storm beach. In the MacLeod Harbor area, storm beaches corresponding to the pre- earthquake sea level range from 9 feet (at the head of MacLeod Har- bor) to 12 feet (profile 4) above mean high water. There is no evi— dence of new storm beaches begin- ning to form, but, since uplift, the sea has regraded a zone between mean high water and 4 feet above it. In the short period since the earthquake, all unconsolidated ma- terial (but no bedrock) in this zone has been regraded. Marine re- grading probably will eventually extend up to the full elevation of a new storm beach. Marine processes have had a much greater effect than fluvial processes in the regrading of all beaches studied. Fluvial processes are effective only where sand or gravel occurs on the beach. Beach profiles 4 and 5 (fig. 13) are along streams, and stream 17 (table 1) is very close to profile 3. Comparison of profile 4 with profiles 1—3 and 8 (fig. 13) shows the typical almost-straight beach profile that results from re- grading by a competent stream flowing mainly through sand. The upper part of the course of the stream in profile 4 on the uplifted beach is erosional; lower down the stream spreads out into a fan and effectively regrades a width of sev- eral hundred feet of the beach by deposition. The low-flow discharge of the stream is 10.0 cfs. The stream in profile 5 has a low— flow discharge of only 0.03 cfs. This stream is competent to erode sand from an area underlain by cobbles but it cannot carry the cob- bles, so the fluvial regarding is limited to erosion and deposition of sand. Not far from the line of profile 5, the stream is divided into several small channels, each 4 to 10 feet wide and 1 to 2 feet deep, which do not appear to be cutting ALASKA EARTHQUAKE, MARCH 27, 1964 laterally; hence the area re- graded here is much smaller than that regraded by the stream in profile 4. Stream 17, cutting into beach profile 3, is a small stream with a low-flow discharge of 1.09 cfs, and a very steep course through uncon- solidated gravel (fig. 14). This stream illustrates the tendency for a channel of uniform gradient to develop across a beach. Lateral erosion is at a minimum here, ap— parently because this stream is still actively downcutting. The area regraded is therefore very small. On the beach, which is both nar- rower and steeper than the up— lifted bayhead, some resorting of beach deposits has already taken place, and streams flow at much steeper gradients than in the bay- head deposits. Thus stream 17 is steeper on the beach than it is farther upstream, whereas the stream in profile 4 has the same gradient above and below the pre- earthquake water level. The stream in profile 5, however, which falls down a steep gully onto the beach, is less steep in its beach section than immediately inland. The gradient of a stream across the beach, therefore, seems to be un— related to its gradient above the preearthquake beach; in other words, the relative drop in sea level caused. by the earthquake will not necessarily lead to the propaga- tion of a knickpoint inland along every stream course. The forma- tion of a knickpoint appears to be least likely in small streams flow- ing from steep hillsides. PATTON BAY BEACHES Seventeen beach profiles, on all types of material, were made in Patton Bay, across the island from MacLeod Harbor. These profiles show the relationships, both ver- tically and areally, between the following features, if they are present: postearthquake high—wa- ter mark, postearthquake storm beach, preearthquake high-water mark (the former barnacle line), preearthquake storm beach, and breaks in slope at the foot of clifi's corresponding to both the pre- earthquake sea level and to raised sea levels. Regrading of the up- lifted beaches was also studied. Regrading of sand and gravel was similar to that on the other side of the island. An estimate was also made of the rate at which rock platforms on the Patton Bay side have been regraded since the abandonment of the older uplifted shoreline. The many seaclifl's, some several hundred feet high, also provide an opportunity for exam- ining the way in which the break in slope at the foot of a cliff is obliterated by subaerial processes. Most of the coastline of Patton Bay now exhibits a rock platform several hundred feet wide, the landward edge of which is gener- ally covered by beach deposits (fig. 17). This platform, which seems to accord with the preearthquake sea level, is backed, except in the cen- ter of the bay, by a steep seaclifl' about 100 feet high. Above this sea cliff the land is gently rolling and rises 200 to 300 feet above sea level. In the center of the bay coastline, however, the cliff is absent, and instead there is a low step and a raised beach platform 80 to 500 feet wide, which is backed inland by slopes of 40° to 50° leading up to the main 2,500-foot backbone of the island. Sandy beaches occur in three main areas above the coastline studied. Just south of the center of the bay is the most extensive stretch of sand, called in this report “Main Sandy Bay.” The largest river on the island, the Ne]- lie Martin River, debouches at the south end of the bay. At this point the present beach is backed by two EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H21 ' ' 655' .v 1’.‘I'w/”'I,' ' I . ,g .;. 1’ - ‘ Q'Q Pz?or’ch Sandy Bay // K ’% C—lzzsaample Ply/”III”? ' ,1/[2/ P22 W—1770 Stearthqu I to: P21 ' 6/"), P19 P20 6 EXPLANATION PA TTON High ground BA Y Main Sandy Bay N z % g (D C Y“ P11 C-14 sample Nellie arti W'1766 Rive APPROXIMATE Rock platform 44/? Pointed ' VV . l ff 425 C 1 —H—O—H—O—O—O—O— Storm beach ____/5_._ Contour of amount of 1964 uplift, in feet 4———F’21 Location number -P11 Beach profile and number 0 1A 1/2 % 1 MILE 17.—Coasta1 features and deformation caused by the 1964 earthquake, Patton Bay H22 storm ridges, behind which is an extensive fiat area drained by the Nellie Martin River and its trib- utaries. There is a second sandy area called in this report “Bay Mouth Bar” at the southern limit of the coast studied, where a lagoon is dammed up behind two storm- beach ridges. The third sandy area is to the north of the center of the bay. This area, called “North Sandy Bay,” is backed by a cliff about 400 feet high, above which rise the main mountains of the island. There is also a small area of sand and gravel in the middle of the central rock platform, where two small rivers debouch across the platform. In several localities between a sand beach and an area of bare rock platform, a short sec- tion is covered by large boulders, 2 to 10 feet in diameter, lying on bedrock; their undersurfaces are angular but their upper surfaces are smooth. This contrast between smoothed upper surfaces and angular undersurfaces suggests that marine action has not been powerful enough to overturn them during the period required to smooth their upper surfaces by abrasion and solution. Smaller rounded cobbles, most of them as much as 6 inches in diameter, between the large boulders indi- cate that 6 inches is approximately the maximum size of stone that the waves were competent to transport. The main coastal features of the bay and the positions of the 17 beach profiles and other key loca- tions (indicated by series Nos. P1— 25) are shown in figure 17. STRATIGRAPHIC EVIDENCE FOR SEA-LEVEL CHANGES HIGH SEA LEVELS There is a continuous raised beach platform between Main Sandy Bay and North Sandy Bay that definitely represents a single former beach. Three profiles ALASKA EARTHQUAKE, MARCH 2.7, 1964 showing the raised beach deposits are described below. The position of beach profile P23 is shown in figure 17 . Figure 18 shows the profile and a strati- graphic sec-tion located on it. The material in the bottom 24 inches of the section is similar to the modern cobbles in the upper part of the beach and is definitely iden- tified as a gravel of the former beach. Its topographic situation might make it seem to be a part of the preearthquake beach, but the raised gravel is an integral part of the section above it. When sea level dropped from its raised level to the preearthquake level or low- er, vegetation would have covered the former beach material, the clifl" at the back of the beach would degrade, and the observed sequence of peat and angular gra- vel would be deposited. The date of the change of sea level must have been earlier than the deposi- tion of the lowest peat in the sec- tion, which has been dated by ra- diocarbon as 20701-200 B.P. (W—1770). The position of profile P19 is shown in figure 17. Sections D1 and D2 (fig. 19) are close to the bank of a small river, which has exposed a 4-foot—thick bed of angular to subrounded gravel in a finer ma.- trix overlying bedrock. Sections D1 and D2 are close together and have been correlated on the basis of the position of the top of the angular gravel bed. The gravel is thought to be a former alluvial fan of the river, deposited while the sea was at a higher level. In section D1 the level of the raised platform can be clearly seen in bedrock, and a small step in bedrock is exposed at the seaward edge of the vegetated platform. Section F (fig. 19) is close to this step, about 50 feet north of the river. In section F the material resting directly on bed— rock is much more uniformly rounded, has less fine matrix, and contains a wider variety of rock types than the angular gravels of sections D1 and D2. This rounded material is interpreted as being a true beach deposit, beyond the edge of the alluvial fan. At profile P24 (fig. 17), at the south end of North Sandy Bay, the raised beach platform is much narrower; it is the northernmost exposure where beach gravel is clearly resting on a nonstructural rock platform. The position of the break of slope at the foot of the cliff can be located very accurately here (figs. 20, 21). LOW SEA LEVELS Possible evidence of a sea level slightly lower than the preearth- quake level is shown by tree stumps, commonly found in the position of growth between pre- earthquake mean high water and the top of the preearthquake storm beach. These stumps all show some wave erosion. More substantial evi- dence of a former low sea level is found in peat layers in the banks of the Nellie Martin River. A continuous peat bed is exposed along the banks in the lower course of the river. The top of the bed slopes down toward the sea; it can be definitely identified as much as 1.8 feet below the postearthquake high-water mark and seems to con- tinue farther out to sea (fig. 22). Under present topographic condi- tions such peat is found behind storm beaches, that is, about 12 feet above high-water mark. It therefore seems probable that this peat layer was formed at a time when sea level was at least 14 feet lower than the postearthquake level at this locality, or 25 feet be- low the preearthquake level. A sample taken from 2 feet below the top of the peat in section B (fig. 22) has a radiocarbon date of 600:200 B.P. (W—1766). 60 50 40 30 20 ELEVATION, IN FEET 10 EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H23 Preearthquake MHW . a . .0 ' . n a ....:nf?KBeach material 'a'. ' ///////"""' Bedrock 100 150 ft at 37°00’ Degraded cliff Break of slope at head of raised beach 40.3 ft above preearthquake MHW Section SECTION Preearthauake DEPTH, Vegetation: spruce break of slope . 3 IN INCHES forest and grass In gravel - “ ' , Surface Black peat, roots not distinguishable :- ~—/.—to-l/2 in. angular ‘ gravel in gray clay ' ‘ matrix ‘lfDark-brown peat, H roots distinguish- Ji i able ir'Angular gravel in **** “J clay, as above . ivPeat, as above _* Angular gravel plus 10 percent rounded gravel in gray clay matrix ~—2»to 24~inrwell-round- ed boulders with some iron staining at the surface, in matrix of sand and s m all rounded “j pebbles as much i as 1/2 in. diameter i—Base of exposed section l l l 200 300 DISTANCE, IN FEET VERTICAL EXAGGERATION X5 C-I4 sample 1 W-1770 taken from this, +33 peat layer J 37 61 18.——Stratigraphic section and beach profile P23, Patton Bay. 400 H24 ALASKA EARTHQUAKE, ml 50 - ,_ uJ lJJ h. E 3' 40 ~ I 2 Id X < 3 o I E < 30 ~ LL] LAJ K O. LA] > o m ( z L g 20 L; Height of preearthquake a storm beach across river _l Lu MARCH 27, 1964 0 100 x Vestical cliff x Break of slope at too of raised beach 36° slope l Section F o o | 00 0 o 0 Marlne g'a/velo WM 07/ Sections D1 .5 and D; | l I l 200 300 HORIZONTAL DISTANCE. lN FEET SECTION Dz DEPTH BELOW SURFACE. SECT‘ON D1 in INCHES DEPTH 0 M Vegetation: grass, fern, BELOW SURFACE, 4 H salmonberry, spruce, alder IN INCHES “m“? _ 9 Ne Vegetation: grass, alder, spruce 10 9 . 0’33"": 5'“ and sand 11 .' i '. ea 11 e Y° Sandy clay 13 ' ' ' ' "Gray sandy silt 13 Gm A o A 0 Pure clay A ,' 1/2- to 4- in. gravel; angular AO‘ O to subrounded, all at -O A local black siltstone A Gravel in finer matrix , Q '" sandy clay matrix 0 A M SECTION F DEPTH BELOW SURFACE, IN INCHES Vegetation: dry mature spruce forest ’ Peaty humus with some mineral material at base -. . Brown-gray silty sand with occasional €03 lenses of well-rounded grit :21 4 ll l. lron»stained organic clay containing thin organic bands :Ql’£ Gray, mottled plastic clay with lenses of well-rounded 27 D _ grit and occasional subangular to subrounded O B siltstone pebbles " o . l . ° : Mainly rounded material from sand Size up 0 to 2 inc, and some large (6-50 in.) well- n O rounded boulders not of local rock, in a - , .. matrix of silt and some clay O . ' O 63 Bedrock 97 Base of section .0 o 0% 61 Bedrock 121 Base of section=river level Base of section 19.—Stratigraphic sections D1, D2, and F, and beach profile P19, Patton Bay. EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND 20.—Cobbles of raised beach resting on truncated bedrock platform, profile P24, Patton Bay. Marine beach deposit of rounded 6— to 24—in‘ boulders in matrix of well-sorted, well-rounded sand and gravel 7/ 1/4—to 4—in. rock fragments Folded beveled weathered bedrock \ Obscured Fine-grained W light—gray sandstone O 10 20 FEET ‘/ A 7 / Preearthquake _ _. beach gravel NO VERTICAL EXAGGERATlON 21.—Stratigraphic section showing old raised beach at P24, Patton Bay. H25 H26 ALASKA EARTHQUAKE , MARCH 27, 1964 SECTION A DEPTH BELOW SURFACE, IN FEET 0 1.6 Surface Beach gravel Peat Present river level 2221 - I 5.6 15-— C-14 sample W71766 Preearthquake MHW SECTION B DEPTH BELOW SURFACE, IN FEET 0 o o — Surface 3Q; lnterbedded beach sand and o gravel. Individual beds about 10 in. thick " Discontinuous stringers of small wood~fragments (swash debris) ~. . Graysiltysand with small woody ‘ fragments. Gradational lower contact. Sharp upper contact Peat containing wood fragments of all sizes, as much as 20 in. in diameter, all abraded. 3-in. clay layer in middle Present river level SECTION C DEPTH BELOW SURFACE. IN FEET 0 4.5 10.5 Surface \r it b t Storm beach gravel \§Q§\\\ \t \k l Beach gravel Light~brown sand with stringers of large wellrrounded boulders itté‘a/iliioil Present river level ELEVATION ABOVE POSTEARTHQUAKE MHW, IN FEET / , a u . u..,flg.,s no Rbiffle o N o o 1:. o o m o o , osltl ApprOX'ma’uip/ ” ’ / ’ Level OI “3° 0‘ u .u .fifi..§0au ks of “a“ on 029/ / / ' t #. er surface at time of measuremen Low.flow wat fl/ , °.'. . 7; ’ "Jo" ..,.7 /Raoids I i I I 800 1000 1200 1400 1600 HORIZONTAL DISTANCE, IN FEET I I i 1800 2000 2400 22.—Stratigraphic sections and long profile of Nellie Martin River. It is concluded that there was an earlier raised sea level, higher than the sea level just before the earth— quake, the exact elevation of which is discussed on pages H28 to H33. By 20701—200 B.P. the sea level was low enough relative to the land for a layer of peat to have grown upon the rock platform eroded at the higher sea level. A period of low sea level followed during which the sea was low enough for peat to form 1.8 feet below the 1965 high-water mark near the Nellie Martin River. The topographic position of the pres- ent peat layer behind storm beaches or higher indicates that the sea was locally at least 14 feet lower than at present relative to the land. The sea was at this low level at 600:200 B.P. Because the dated sample was near the top of the peat layer, it is assumed that relative sea level rose shortly after this time. ELEVATIONS OF BEACH FEATURES The sea level at which a raised beach was formed cannot be mea- sured directly; fossil beach fea- tures are among the best available indicators of the former sea level. The range of elevation of several beach features, and the relation- ship of each to the high-water mark, has been studied in Patton Bay, mainly with reference to fea- tures formed at the preearthquake sea level. From these data the reli- ability of each feature as an indi- cator of a former high-water mark has been assessed, and some con- clusions drawn about the causes of variation in elevation for each feature. The beach features chosen for this study were the break in slope between the foot of a cliff and a bedrock platform, the break in slope between a cliff and a gravel beach, and the top of the storm beach ridge. The elevation of a break in slope between a cliff and a bedrock plat- form could be measured in only three localities. The elevations of the break in slope above the rele- vant (preearthquake) mean high water, measured as the barnacle line, were +2.0 (P7), +0.6 (P4), and -0.8 (P5) feet. These figures suggest that, for the conditions found on Montague Island, the elevation of a break in slope in bedrock is within 2 feet of mean high-water level and that this fea- ture is the best indicator of a for- mer sea level. A gravel beach is a temporary EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND deposit whose thickness varies with time. A break in slope be- tween a cliff and a gravel beach is therefore likely to show a greater range in elevation between local- ities than one between a cliff and bedrock platform, and the lower limit of this range should be close to high-water mark. Seven meas- ured elevations of this feature (fig. 230) ranged from 0 to 16 feet above mean high water, with a mean value of +9.3 feet. At loca- tion P7 (fig. 17) the elevation of the break in slope increased by as much as 8 feet up small reentrants in the cliffs. It might be expected that the break in slope of a cliff with a gravel beach would give the most reliable evidence of high-wa- ter elevation; however, the pos‘ sible error of :8 feet shows that this feature is not a reliable indi— cator of this altitude. The difference in elevation be- tween the top of the preearthquake storm beach and the preearth- quake mean high-water level was measured at 11 points in Patton Bay. The values ranged from 6 to 21 feet above mean high water, with a mean of 12.0 feet (fig. 233) . These values are consistent with the idea that a gravel beach banked against a cliff is in effect an incomplete storm beach; there- fore, elevations of breaks in slope in gravel beaches should be lower than the elevations of storm beaches. In several locations a new storm beach is probably beginning to form in response to the postearth- quake sea level. Seven measure— ments show that the ridges formed are as much as 6 feet above mean high water (1965), with a mean value of 4.1 feet (fig. 23A). The present size and position of the ridges illustrates the initial stages in the formation of a mature storm ridge. The height of the top of a storm beach is dependent upon wave and shore environment, but it will ap- proach a maximum elevation which is stable for its own envi- ronment. Mean high-water level can be determined from the eleva- tion of the top of a storm beach H27 only within :8 feet, so this fea- ture is also a poor indicator of sea level. However, the difference in elevation between two mature par- allel storm beaches should be an accurate measure of the elevation diflerence in sea levels when the two ridges were formed. This dif- ference has been used in calculat- ing the elevation of the raised sea level in Main Sandy Bay. In MacLeod Harbor the level of mean high water could be deter- mined from the form of the beach profile, but no satisfactory corre- lation was found between beach form and water level in Patton Bay (fig. 24). The gradient of the beach, the depth from high-water mark to the seaward edge of the beach gravel, and grain size of the beach material were all found to be of little use for the precise loca- tion of sea level. In MacLeod Har- bor a consistent relationship was found between the preearthquake water level and the steep seaward edge of the gravel beach (fig. 14), but such a step is only to be ex- pected where the gradient of the 23,—Distribution of elevations of beach features; A, above postearthquake mean high water, of postearthquake storm beaches; B, above preearthquake mean high water, of preearthquake storm beaches; 0, above preearthquake mean high water, of the break in slope at heads of beaches. ,_ 3 LL 24.17 3 A 3 I 20.1— 2 LU E :> 16.1— o I E < 12.1— LLJ )— U) o n. m 8.1—— > o m < E 4.1— . E E 0.1 A _, 1 v 1 1 “J o 1 2 3 4 24.1 20.1 — 16.1 — ~ Heights of storm beaches in MacLeod Harbor 8.1 — 4.1-i 0-1 I 1 1 1 o 1 2 3 4 ELEVATION ABOVE PREEARTHQUAKE MHW, IN FEET L“. 24.1 “c: T E C E 201— Heights of break 2 in slope near g P7 < 16.1———-— D O I E 121 < '1 / _ LIJ LU o: o. m 8.1 - > O m «I z 4.1 — 9 1. § 51' 0-1 I 1 | “J o 1 2 3 4 NUMBER OF MEASUREMENTS IN EACH HEIGHT RANGE H28 ALASKA EARTHQUAKE, MARCH 27, 1964 FEET 20 15 U! 0 20 4O 60 80 100 FEET 3 P24 P23 P22 P12 P2 P24 P16 P11 Preearthquake MHW r6 24.——Beach profiles in Patton Bay in relation to the preearthquake mean high water, showing that the shape of the profile is not a reliable indicator of the position of the mean high water. See figure 17 for location of profiles. seaward-building gravel is ap- preciably lower than the angle of the rock platform on which it is built. The evidence from MacLeod Harbor (p. H19) suggests that these conditions are found only in areas well sheltered from wave attack. Because the whole of Pat- ton Bay is more exposed than any of the beaches in MacLeod Harbor, no steps similar to those in Mac— Leod Harbor are to be expected in Patton Bay. In summary, where a bedrock platform against the foot of a cliff is not covered by unconsolidated material, then the elevation of the wave-cut break in slope is consid- ered to be the best available topo- graphic indicator of mean high water, and the difference in eleva- tion between two breaks in slope formed at different sea levels is an excellent measure of the difference between the two sea levels. Where unconsolidated material lies at the foot of a seaclifl’, the shape of the underlying bedrock, the amount of gravel available, and random factors probably con- trol the height to which the gravel is accumulated. At one extreme there is no gravel, and at the other extreme a full storm beach is formed against the base of the cliff. The elevation of the break in slope between the gravel and the cliff therefore varies from near high-water mark to full storm- beach elevations. Such breaks in slope are not accurate indicators of former water levels. Storm—beach elevations increase with time to a maximum, which takes several years to reach. This maximum elevation is only con- stant for a given wave-exposure environment. Storm-beach eleva- tions are therefore poor indicators of the absolute water level, but the difference in elevation between two mature parallel storm beaches should provide a good measure of the difference of elevation between the two sea levels at which they were formed. TECTONIC DEFORMATION 0F BEACHES Elevations of beach features were measured at many localities over a 10-mile stretch of coastline in Patton Bay. The elevations of storm beaches, high-water marks, and breaks in slope were measured using the preearthquake high—wa- ter mark as a standard. In figure 25, each symbol represents a mea- sured point, and the lines show the interpolated elevations of each fea- ture between measured points, so each line shows the change of ele- vation of a beach feature along the coast, in relation to the preearth- quake high-water mark. The figure shows the elevations of pre- and postearthquake shoreline features and of raised shoreline features and the variation of height of these features along the coastline of Pat— ton Bay. All elevations were mea— EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H29 60 EXPLANATION I- 'G ® 3 Preearthquake break in slope Crest of postearthquake storm “' between foot of cliff and beach E gravel (G) or bedrock (B) 0 g g‘ 4° ” A Old raised beach MHW 30 I x .c E Crest of preearthquake storm Postearthquake MHW \ ‘3‘: beach 7; e < k 3 O E 20 A A O a: Pr G < Gearth iii Quake storm beach ./ \.G E /.G \G ,9 O Preearthquake MHW .‘B.'//- 3 g B _ x ® Postearth I'- 'x\\® _ quake Storm beach 5 l / X\ Postearthquake MHVII\®/® ® ® Lu X ‘5 X X X x ® '1 an / x X‘X x x a 20 —‘~: — o I- 2 . <>r >. Nellie Martin Mam Sandy Bay North Sandy Bay “J u ' r——J——‘—fi {—g d ’3 R'Ve' \m N m :22 M. a om M a a as are arm a 3:9“ a; was as a I II LII I III | [HM/II I |l||||lll| | 40 If I I | I I | I | I 0 1 2 4 5 6 7 8 9 10 11 SOUTH NORTH DISTANCE ALONG SHORELINE, IN MILES 25.—Variation in differences of elevation among coastal features in Patton Bay. sured relative to the preearthquake mean high water (which is taken as the top of the former barnacle zone). Postearthquake mean high water was estimated from US. Coast and Geodetic Survey tide tables. Where storm-beach crests and breaks in slope b et w ee 11 beaches or rock platforms and the cliffs behind them are present, their elevations are shown. Elevations are shown for fea- tures which correspond to pre— and postearthquake sea levels and also to a raised sea level. The ele— vation of the raised mean high water has been calculated in one of the following ways: (1) in Main Sandy Bay, as the differ— ence in elevations between the crest of the preearthquake storm beach and the crest of the raised storm beach, or (2) elsewhere in Patton Bay, as the difference in elevation between the top of the preearthquake barnacle zone (pre- earthquake mean high water) and the break in slope between the raised beach platform and the steep slope behind it (raised mean high water). Figure 25 shows the tectonic deformation of the shoreline from the time of the earlier raised beach until just be- fore the 1964 earthquake, and also the deformation caused by the 1964 earthquake. The deformation due to the 1964 earthquake is plot-ted in figure 17. These data are in general agreement with the smaller scale map by Plafker (1965, fig. 7), but they add detail for the Patton Bay area. The evidence for the raised sea level consists of: (1) a continuous platform from just north of loca- tion P24 (fig. 17) to the north end of Main Sandy Bay; (2) a con- tinuous raised storm—beach ridge across Main Sandy Bay; and (3) a platform at two points immedi- ately south of the Nellie Martin River, close to the southern end of the storm beach (locs. P9 and P10, fig. 17). The continuity of the landward edge of the platform and of the raised storm beach shows that they represent a single sea level throughout their extent. At Bay Mouth Bar an earlier storm beach within 2 feet of the elevation of the younger ridge has been preserved behind the younger ridge. This abandoned beach was presumably formed at a sea level almost at the preearthquake level, but no positive basis was found for correlating it with the raised beach deposits described above. Figure 17 shows not only the distance over which the raised beach can be detected but also the width of the platform which has has not yet been destroyed. At the north end of the coastline studied, the raised beach platform grad- ually narrows and finally disap- pears in North Sandy Bay where H30 the preearthquake seaclifi's are sev- eral hundred feet high. The ab- sence of the raised beach in this section can be attributed to the nature of the local rock, which offers very little resistance to ma- rine erosion; a preearthquake rock platform would have been eroded by marine action and its surface covered. with sand and gravel, in contrast to neighboring more re- sistant bare—rock platforms. The lack of resistance of the rock is also indicated by (1) the relatively gentle gradient of the cliffs (40°— 50°), (2) the tendency of the cliffs to degrade by landslides, mud- flows, and gullies rather than by rockfalls and talus accumulation (this tendency indicates a high clay content in the rock or in its primary weathering derivatives), and (3) the considerable amount of erosion of the cliffs since the earthquake (fig. 29, p. H35). The coastline north of North Sandy Bay was not examined in any detail. The south side of Box Point (north of the area of fig. 17) has continuous cliffs 50 to 100 feet high; here low raised platforms apparently have not been pre- served. The surface of the penin- sula is relatively low and flat at altitudes of 100 to 200 feet, but there is too much relief for a raised platform. It is therefore tenta- tively assumed that, within the area studied, only in the central part of Patton Bay was wave at- tack sufficiently attenuated and the rock sufficiently resistant to al- low the preservation of a raised beach platform. Since measurements of the rela- tive heights of the raised beaches are based on the elevations of breaks in slope and of storm beaches, it must be recognized, for the reasons discussed above, that some measurements may be in error by 5 to 10 feet. Nevertheless, the amount of the warping is much ALASKA EARTHQUAKE, MARCH 27, 1964 greater than the amount of error inherent in the measurements, so the general pattern of the uplift can be accepted. For example, near the northern end of the raised beach (fig. 25), its elevation above the preearthquake mean high wa- ter rises from 10.2 feet (location P17) to 45 feet (profile P22) in a distance of 1 mile. This deforma- tion is too great to be dismissed as an error of measurement, especi- ally inasmuch as the pattern of warping in this area is confirmed by the elevations of breaks in slope both in gravel and in bedrock. The relationship between the two measures of the uplift of the raised beaches—(1) the preearth- quake barnacle line to the break in slope of the raised beach and (2) the preearthquake storm beach to the raised storm beach—can be studied where the storm beaches abut against steep slopes at each end of Main Sandy Bay. Figure 26 shows the relationship at the northern end of the bay and figure 27 shows the relationship at the southern end. Both figures seem to show that, in the area of overlap, the difference in height between the two storm beaches tends to de- crease and the difference in height between the break in slope and the barnacle line tends to increase. Inasmuch as the storm ridges are curved, the exposure to waves obviously will not be the same on each ridge at the points where a section line drawn at right angles to the ridge intersects them (figs. 26, 27). The landward, older ridge will have been closer to the hillside, and therefore less exposed to waves. As a result, the landward ridge is lower for a given sea level, and the difference in height between the ridges is smaller than the difference between the sea lev— els at which the ridges were formed. Elevations of breaks of slope, on the other hand, increase up the inlets (p. H27), so the dif- ferences in elevation between the break in slope on the raised plat— form and the preearthquake high- water mark will tend to become progressively larger than the dif- ferences between the two sea levels in the zone where the steep hillside is farther from the shore. It is con- cluded that the true change of sea level is greater than that shown by the storm beaches and less than that shown by the breaks of slope. An interpolated value of the sea- level change is shown by the broken lines in figures 25, 2GB, and 278. South of the Nellie Martin River there are cliffs along the coast; no certain evidence of a raised beach was found, but with elevation dif- ferences as little as 5 feet above the preearthquake level, evidence would be very difficult to interpret in the absence of a continuous fea- ture. Possible evidence exists in the second storm-beach ridge at Bay Mouth Bar and at isolated locations at the mouths of streams where the cliff is absent. However, the primary purpose of this inves- tigation was not to locate possible raised beaches, but to use the evi- dence provided by a raised beach to support observations of changes to the beaches and platforms up- lifted by the earthquake. Comparison of the radiocarbon dates for all samples from Pat- ton Bay beaches indicates that observed data can be explained most simply by two relative sea— level changes before the 1964 earthquake. Each change has left abandoned shoreline features at various elevations (fig. 25). Data from elsewhere in Prince William Sound suggest that the 1,000 years before the earthquake was char- acterized by slow submergence of the land (Plafker and Rubin, 1967), and the data from Monta- gue Island are consistent with this EIROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H31 P16 MHW/break in slope=12.0 ft Z MHW/break in slope=14.4 ft P15 MHW/break in slope=21I7 ft P14 Raised storm beach/preearthquake storm beach:15.1 ft V Raised storm beach/preearthquake storm beach=21.8 ft D Cabins D: _________ \_-__ — — . W RaIsed storm beach/preearthquake storm beach=17.7 ft I EXPLANATION l: “‘5 c U 0 a ‘ II a MHW/break in slope=l4.4 ft. .0 :3 .9 Elevation difference between E I E named features 3 'i 3 (I) ’ 07 P16 8 i 4'2 .9 I S W “ I U‘ . I \ 0: I 5 Profile locatIon and number \ \\ l L \\ \\ I 3 \ \ \\ a, \ \ L \ \ : \ \ o. \\\\\\ \ A III CI Cabins P13 [:1 Raised storm beach/preearthquake storm beach=15.6 ft 0 200 400 600 800 FEET I I [ J EXPLANATION X Sea-level change obtained from difference in elevation between preearthquake MHW and raised sea-level break in slope Sea-level change obtained from difference in elevation between top of preearth- quake storm beach and top of raised storm beach N O Assumed actual sea-level change ESTIMATED SEA-LEVEL CHANGE BETWEEN PREEARTHQUAKE LEVEL AND RAISED SEA LEVEL. IN FEET 10 e B ‘D m V m H F. F. H a. N o. >- n. 3 0. O I I I I I I I I J 0 1000 2000 3000 NORTH DISTANCE ALONG SHORELINE, IN FEET SOUTH 26.—Relationship between measures of uplift of raised beaches at the north end of Main Sandy Bay. A, Differences in elevation of raised break in slope, raised storm beach, preearthquake storm beach, and preearthqwake mean high water. B, Measurements of sea-level change. H32 ESTIMATED SEA-LEVEL CHANGE BETWEEN PREEARTHQUAKE LEVEL AND RAISED SEA LEVEL, iN FEET Mi-iW/break in slope—“=83 ft 20— 10— oO_ ALASKA EARTHQUAKE, MARCH 27, 1964 / P11 Raised storm beach/ preearthquake storm beach:7.9 ft Gravel spits MHW axenbuueaxsod \ _ / / Raised storm beach/ preearthquake torm beach=4.8 ft A : Preearthquake stack 0 200 600 800 1000 FEET i l i B EXPLANATION x Sea-level change obtained from difference in elevation between preearthquake MHW and raised sea-level break in slope O Sea-level change obtained from difference in elevation between top of preearth- Assumed actual l I ha quake storm beach and top of raised N sea- eve C [$th P11 P10 P9 1 I | i 1 i 1000 2000 3000 DISTANCE MEASURED ALONG PREEARTHQUAKE STORM BEACH, lN FEET 27.—Relationship between measures of uplift of raised beaches at the south end of Main Sandy Bay. A, Sketch showing difference in elevation of raised break in slope, raised storm beach, preearthquake storm beach, and preearthquake mean high water. B, Measurements of sea-level change. EROSION AND DEPOSITION ON Sea level Location of Date relative to C—14 sample C—14 sample Basis for estimating mean high preearthquake (fig. 17) water level (feet) Sometime before 2070 +8 B.P. W—1770 P23 Raised beach peat. By 2070 B.P ____________ _<_ ——6 600 B.P ________________ _<_ —25 W—1776 P11 Nellie Martin peat. Preearthquake A.D. 0 ________ P11 Barnacle line. 1964. Postearthquake A.D. — 11 ________ P11 Estimate from U.S. 1965. Coast and Geodetic Survey tide tables. interpretation. If the central value for each radiocarbon date is used, then the sequence of events near the mouth of the Nellie Martin River seems to have been as shown in the table. These dates are consistent with relative sea-level changes at the same time and in the same direc- tion on both sides of the island, although the amount of vertical movement varies considerably from place to place. No clear evi- dence of other sea-level changes was found in the area studied. OBLITERATION OF BREAK IN SLOPE AT TOP OF BEACH One immediate effect of the 1964 earthquake was to initiate many landslides along the coast- line, but much mass wasting of the former cliffs has taken place since the earthquake. The break in slope at the top of the former beach is being rapidly obliterated by several processes—talus accum- ulation, landsl‘iding, gullying, and vegetation growth. The uplift has prevented the sea from removing mass waste from the base of the cliffs, so all the material that has been eroded from the cliffs since the earthquake can be measured. From these measurements the rates at which the different slope processes were operating were determined. OBLITERATION BY TALUS ACCUMULATION At profile P3 (fig. 17 ), talus ac- cumulations have already 0b- scured 85 percent of the break of slope at the foot of a vertical clifl'. The distribution of fallen debris is shown in figure 28. Some of this material no doubt fell at the time of the earthquake, but the major part probably has accumulated since the earthquake; the sound of falling stones was still frequently heard in 1965. The volume of talus material at profile P3 is estimated to be about 210,000 cu ft. The cliff from which it is derived is about 110 feet high and 580 feet long. The talus ac- cumulation therefore represents a mean cliff retreat of 2.4 feet in 15 months (to July 1965), assum- ing a 30-percent porosity in the talus. At this rate, if the total accumulation postdates the earth- quake and if the observed maxi- mum angle of rest of 22° is correct (fig. 28), the cliff will be completely regraded in 7] years. If any of the present accumulation occurred at the time of the earth- quake, then the time required for the cliff at profile P3 to become completely covered by its own talus will be as follows: Accumulation Time required at time of for total degrada- earthquake tion of clifi (percent) (years) 0 71 40 117 60 176 80 350 Unless almost all of the accumu- lation occurred at the time of the earthquake—and the continued falling of material makes this A RAISED BEACH, MONTAGUE ISLAND H33 seem improbable—then the cliff obviously will be completely ob- scured by its own talus in no more than a century or two. OBLITERATION BY LANDSLIDES AND GULLYING In North Sandy Bay near pro- file P24 (fig. 17 ), landslides and gullying seem to have been more important than simple talus ac- cumulation in obliterating the break in slope at the foot of the cliff. A horizontal distance of 2,940 feet of cliff was studied in ‘detail, and the distribution of fall- en material is shown in figure 29. In this total distance, 22 percent of the break in slope is virtually unobscured by debris, 17 percent of the break in slope is obscured by alluvial fan material, and 61 per- cent of the break in slope is ob- scured by landslide material. It was evident that much of the movement postdates the earth- quake; the stream in one gully was very rapidly transporting debris while studies were being made in July 1965, and material in some of the slides was still very wet and sticky. The extent of postearth- quake erosion is also shown by the very marked increase between May and August 1964 in fresh landslide scars that appeared on air photographs taken by the U.S. Coast and Geodetic Survey at those times. The estimated vol- umes of material transported are 670,000 cu ft in alluvial fans and 370,000 cu ft in landslides. The av- erage height of the cliff is 400 feet; so the mean distance of clifi re- treat since the earthquake has been 0.90 foot in 15 months. Landslides are far more effec- tive than gullying in degrading the cliff, because they act along its whole length. If landslides alone are taken into account, and if all the volume of landslide de- bris has accumulated since the H34 ALASKA EARTHQUAKE, MARCH 27, 1964 Storm beach fades out - Two newly fallen trees ¢ Dead tree %’ base EXPLANATION 21°50’ Angle of scree slope O 50 100 FEET l | | 28.—Degrading of vertical cliffs and accumulation of talus at profile P3, Patton Bay. EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND EXPLANATION Alluvial fan u_LJ_Ll_LLl_l_Ll_J_LLJ_Ll—l_l Landslide Cliff LLLLLLLL;LLL Position of cliff behind slide * Raised beach gravel .9 9: 9°4o' 9° 30: Angle of slope H35 100 150 200 FEET l 4 29.—Sketch showing degrading of clifl by landslides and gullying north of location P24 in North Sandy Bay. earthquake, then the time required for total degradation of the cliff would be 2,300 years. However, the gradient of the vegetated steep slope behind the raised beach is rarely less than 38°, so it is un- likely that the process of degrada- tion will proceed very far. At an early stage, vegetation will estab- lish itself on the slope, and further degradation will be exceedingly slow, except perhaps locally along streams or gullies. The striking difl’erence in be- havior between the resistant verti- cal clifl' and the nonresistant 50°- slope may help to explain some apparent anomalies of raised- beach preservation in other areas. For example, rapid accumulation of talus may destroy topographic evidence of a raised beach within a short period, but it is this same talus cover which protects the raised beach from subsequent weathering and erosion and so over the long term preserves the beach in the stratigraphic section. OBLITERATION BY VEGETATION Where the upper limit of a raised beach is not a well-defined clifl', but a low step leading to an- other raised beach, the junction be- tween the two beaches can be hidden by beach deposits and the growth of vegetation. It seems likely that this obliteration will occur extensively at the junction between the preearthquake beach and the raised beach, for example at the section shown in figure 19. Where the raised platform is well developed, the topographic step between the raised platform and the preearthquake beach is nearly always low, being nowhere more than a 10-foot step, even though H36 ALASKA EARTHQUAKE, MARCH 27, 1964 Beach material ”éfl/ 30.——Idea1ized profile through two raised beaches, with soil and vegetation omitted. Step between two rock platforms indicated by A. The development of the pre- earthquake platform at the ex- pense of the raised platform has been in progress only during the period that relative sea level was rising to preearthquake level, that is, since 600:200 B.P. or earlier. The probable sequence of sea levels and platform cutting is illustrated schematically in figure 31. Figure 313 shows two of the possible extreme cases if the gra- the difference between the two sea levels may have been 40 feet or more. Figure 30 shows such an idealized profile, with soil and veg— A etation layers omitted. Most of the \ topographic step between the two \ Raised sea level rock platforms, as visualized in figure 29, is obscured by beach de- posits, so the surface step is very low; a vegetation cover will soon make it extremely difficult to dis- tinguish between the two surfaces by topographic form alone. Two valid criteria may still be used: (1) there should be a marked sub- Case 1 surface step in the bedrock at A B of figure 31; and (2) if datable soil layers (for example peat) continue to form on the abandoned h beaches of Montague Island as they have in the past, then an age discontinuity should occur in the lowest layers at A. Regraded cross section: hox Preearthquake sea level REGRADING OF ROCK PLATFORMS The rock platforms accordant with the preearthquake sea level appear to have been formed by the regrading of a preexisting plat— form accordant with the raised- beach sea level. Regrading is well marked near location P21 (fig. 17) Case 2 Regraded cross section=1/2 how Preearthquake sea level where many bedrock knobs on the upper part of the preearthquake platform stand as outliers, clearly accordant with the raised-beach platform, and in places sur— rounded by preearthquake beach deposits. gradient than raised platform. 31.—Probable sequence of platform cutting in Patton Bay. A, Cutting of raised platform. B, Cutting of main platform during a time of rising sea level: case 1, main platform at same gradient as raised platform; case 2, main platform at lower EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H37 dients of the raised platform, main platform, and initial land slope are allowed to vary; it may be seen that the cross-sectional areas re- moved in forming the main plat- form range from a minimum of 1/zha: to a maximum of hm, where h is the difference between the elevations of the raised beach notch and the preearthquake notch and as is the width of the main platform, measured at right angles to the shoreline. In theory, the rate of regrading would depend on the resistance of the rock and on the thickness of rock to be removed and would de- crease as either factor increases. Rock resistance is an unknown variable, but the fall of relative sea level from the raised level to the preearthquake level is a meas- ure Of the thickness of the mate- rial to be removed; it is compared With platform width in table 2. Figure 32 shows the relation- ship of the cross-sectional area re- TABLE 2.—Relationshtp between change of sea level (h), width of platform regraded (x), and cross-sectional area of material removed (Azh-x) Site (fig. 17) h, in feet (raised level to preearth- quake level) moved by regrading (A), to the depth through which regrading has had to work. It can be ex- pressed by the best-fit relationship (fig. 32) A =2,000+360h. Dividing this expression through by h, the approximate width of the rock platform is given by the relationship 20 , ._. U! l 10 CROSS-SECTIONAL AREA OF MATERIAL REMOVED (A), IN THOUSANDS OF SQUARE FEET A = 2000 + 360k | l 0 10 20 CHANGE OF SEA LEVEL (IL), IN FEET 3O 40 50 32.—Relationship between change of sea level and cross-sectional area of material eroded during regrading of raised bedrock platform, Patton Bay. Points are plotted from data in table 2. z, in feet (base of raised beach to Azhd'), in square postearthquake feet mean high water) 43 400 17, 200 40 420 16, 800 45 400 18, 000 34 350 1 1, 900 21 320 6, 500 23 500 11, 500 17 550 9, 200 10 650 6, 500 13 450 5, 800 12 550 6, 600 6 750 4, 500 .. = 360 + 2,92. The mean rate of regrading can be found by dividing these two equations by the period during which relative sea level rose to the preearthquake level. Relative sea level was already rising at the time of formation of the low sea-level peat, dated at 600:20‘0 B.P. (p. H33). The rise of relative sea level had probably begun before the deposition of sample W—1768 in MacLeod Harbor at 800:200 B.P. The period of platform cut- ting has therefore been taken as 1,000 years (T), and the estimated mean rate of removal Of material has been A/T=2.0+0.36h sq ft per year :30 percent. The mean rate of Widening of the platform has similarly been m/T=0.36+2.0/h ft per year :30 percent. These empirical equations can be compared with two limiting ideal situations, the first a constant rate of removal of material, and the second a constant rate of lateral cutting. The first situation typi- cally would exist where the cliff is soft and is easily eroded, or the cliff is very high; the process is H38 limited by the capacity of the sea to carry away the eroded material. In the notation of the equations above, this situation gives: A / T = constant, 50/ T o: 1/h. The second situation is applicable to resistant cliffs where the limit- ing factor is the power of the waves to undercut the cliffs, and the transporting capacity of the water is more than adequate: A / T o: h w/ T = constant. Figure 32 shows that the data for Patton Bay more closely fits the second situation. The intercept of the area for zero net change of elevation is smaller than the h- dependent term and may be inter- preted as the amount of material removed from a coast where the sea returned to its original level and merely formed a steeper off- shore platform. A sufficiently great relative lowering of sea level presumably would overload the transporting capacity of the sys- tem and thus would approximate the first ideal situation, but so great a change was not observed in the field. In Patton Bay the resistance of the rocks seems to have been the main control of the erosion and re- grading of the rock platform to its preearthquake level, and the transporting capacity of the waves has been more than sufficient to remove the debris produced. REGRADING OF SAND AND GRAVEL Where the beach is in unconsol- idated materials, the whole area of the beach below the crest of the newly forming storm beach has apparently been regraded by ma- rine processes. It seems probable that the vertical range of this re- grading will increase as storms of greater magnitude build a higher storm-beach ridge. ALASKA EARTHQUAKE, MARCH 27, 1964 The upper limit of the future marine erosion and deposition of sand and gravel seems to be limited by the final storm-beach height, which should approximate the pre— earthquake storm-beach height, that is, 8 to 16 feet above high- water level. The horizontal extent of beach resorting up the bay is less clear. On a rock platform, the shape of the platform in cross sec— tion should determine the position of the storm beach. The beach pro- files (fig. 24) indicate that the base of the storm ridge forms from 4 to 6 feet below high-water level, or at about mean sea level. Main Sandy Bay is the best ex- ample of a beach of unconsolidated material, where waves could change the position of mean sea level in a horizontal direction by erosion or deposition. Figure 33 shows the presumed form of the raised beach before the preearth- quake sea level and the position of the preearthquake storm beach for profiles P11, P12, and P13. Wheth- er the effect of the sea—level change was erosional (P12 and P13) or depositional (P11) , the position of the new storm beach is immediate- ly at the foot of the old one, the horizontal distance between the two storm beaches being almost constant at 300 feet despite varia- tions in the amount of net relative sea-level change. The postearth- quake storm beaches in profiles P1 and P24 also are about 300 feet from the old storm ridges. In North Sandy Bay an esti- mate was made of the percentage of the area regraded by fluvial ac- tion (even if later covered with alluvial fan deposits). This esti- mate was made for the same sec- tion of beach as the estimates of obliteration of the break in slope by landslides and gullying (p. H33) . At the top of the beach, close to the cliff, 24 percent of the area had been regraded by fluvial ac- tion; just above the postearth— quake storm beach 33 percent had been regraded. The extent of re- grading depends on the position of the streams flowing off the hill- sides, and the lateral extent of re- grading will increase as the streams change course on the beach. The streams flowing into North Sandy Bay are small and drain very steep slopes. Regrad- ing consists mainly of erosion of the beach materials, followed by deposition of hillside material as alluvial fans, often above the original beach level. The overall course of regrading is illustrated by the beaches in MacLeod Harbor (profiles 3, 5, 8, fig. 13) and North Sandy Bay, and by the raised-beach section in Pat- ton Bay (profiles P19, P23, fig. 17). After a raised beach or rock platform that is backed by a steep hillside has been elevated, zones of marine and fluvial action are di- vided by a storm-beach ridge. On a beach, the storm ridge is formed about 300 feet away from the old storm ridge. On a rock platform, the storm ridge is formed with its seaward base at about mean sea level. This ridge remains until it is demolished by wave erosion of the rock, after which there will only be a gravel beach banked against a bedrock step as at A in figure 31. Below the storm-beach ridge a new equilibrium beach or bedrock plat- form will be formed by marine ac- tion; the exact profile of this ridge cannot be predicted (fig. 24). Above the storm beach, infilling by alluvial fan or colluvial mate- rial alternates with the develop- ment of soil and peat until the streams are able to establish semi- permanent courses and equilib- rium gradients through the new storm beach. EROSION AND DEPOSITION 30— Preearthquake storm beach ELEVATION RELATIVE TO PREEARTHQUAKE MHW, IN FEET .— O l Raised storm beach \ Raised MHW for P11 Preearthquake storm beaches ON A RAISED BEACH, MONTAGUE ISLAND H39 Raised storm beaches ProfiIe P12 Profi e P13 Raised MHW for P12 and P13 Profile P11 / Profile P12 _ - T FrgsTir—iigd’profile of / raised beach I’M e P13 10 ~— 20 I I I ' m I | I I 100 200 400 500 600 700 800 900 1000 HORIZONTAL DISTANCE, IN FEET 33.—Position of preearthquake storm beaches in relation to the presumed original surface of the raised beach. Profiles of original surface were obtained by transposing present profiles to raised high-water marks. Although most of the results of studies on Montague Island have only local relevance, some conclu- sions may be drawn which have a wider application. In MacLeod Harbor, the 33—foot uplift of the beach deposits and their subse- quent erosion provided a rare op- portunity to observe and measure the actual effect of a change in sea level on fluvial erosion. It was found that the streams flowing in sand and silt immediately cut a course of almost uniform gradient between the old and the new sea levels. This phase of the readjust- ment occurred within a matter of days for all streams flowing over unconsolidated materials and hav- ing low-flow discharges greater than about 1 cfs. Whether the CONCLUSIONS long-term adjustment of the long profile would be aggradation or headward recession of a knick- point could not be determined, but it is here suggested that the long- term adjustment may depend on whether the new course is less steep or steeper than the course above the former sea level. Thus, in mountainous regions, Where the streams have relatively steep gra- dients, a fall in sea level might lead to either aggradation or erosion; there would not neces- sarily be a relation between changes of sea level and the pres- ence of knickpoints. In lowland regions, however, where the lower courses of rivers have very low gradients, it is anticipated that knickpoints will develop and mi- grate upstream following a rela- tive fall in sea level. In addition to being an example of a dissected uplifted surface, the head of MacLeod Harbor is also an example of an area where fluvial processes are occurring so fast that changes may be readily studied. A mean rate of net sedi- ment removal of 1 cfs in a stream of low-flow discharge of 300 cfs may be unusually rapid, but the fact that the erosion rate varies in direct proportion to low—flow dis- charge may have wide application, inasmuch as it shows the efl'ect of discharge on the rate of erosion and the transport of uniform ma- terials. The relationships between the preearthquake sea level and the H40 beach features associated with it have been very clearly exposed by the uplift. Measurements over about 10 miles of coastline show that the height of storm beaches varies from 6 to 22 feet above high- water level, the mean being 12 feet. Where gravel was banked against bedrock cliffs at the top of a beach, the height of the break in slope between the gravel beach and the cliff varied from 2 to 15 feet above high-water level, the mean being 9 feet. Therefore, neither of these features can be depended upon to indicate the relative sea level with- in less than 5 to 10 feet unless the exposure and configuration of the shoreline is known in unusually great detail. The area studied was not large enough to serve as the basis for general deductions about where raised beaches are best preserved, but some observations were made. Gently sloping shores in unconsoli- dated materials seem well suited to preservation of raised storm beach- es. On rocky shores, such factors as rock resistance and wave exposure tend to influence the rates of for- mation and destruction of rock platforms in the same way, so there is no reason to assume that these factors influence preservation. In Patton Bay, however, the raised beach platform has in fact been best preserved where the rock seems to be most resistant (al- though not highly resistant), and the platform narrows dramatically and disappears where the rock be- comes less resistant. Other factors which should theoretically aid beach preservation are ( 1) a low gradient on the earlier shoreline, which favors more rapid forma- ALASKA EARTHQUAKE, MARCH 27, 196‘! tion of a platform at the old sea level, and (2) a large uplift where the lowered relative sea level ex- posed a larger area of platform and thus delayed its total destruc- tion. The influence of the slope of the former coast could not be clearly seen in the field, but broader platform remnants do seem to be associated with some of the areas of greater uplift. In Pat- ton Bay, therefore, rock platforms seem to be best preserved in an area (1) slightly protected from wave attack, (2) where the rock is mod- erately resist-ant and (3) where relatively great uplift has followed the formation of the platform. The 1964 deformation has pro- vided information about the way in which a raised beach is modified after uplift. Rates of widening of a rock platform by erosion have been calculated and have been shown to be roughly constant at 0.49 feet per year for cliffs of moderate height. The rate of reces- sion of 0.7—2.0 feet per year for cliffs on Montague Island should indicate the order of magnitude of recession in other areas. Storm beaches form on old platforms and their bases are roughly at mean sea level; they are converted to gravel banks against a step as the platform is eroded by the waves. Beaches of sand and gravel (other than bay-head deposits) seem to be regarded by marine ac- tion within 1 year, except for the upper part of the beach, where a new storm beach forms about 300 feet seaward of the previous one, provided the sea level drops at least 5 feet. The new storm beach may take many years to attain its full height. Behind a storm beach, there is a rapid fluvial regrading of beach material near the mouths of streams and a general slow i11- filling by alternate layers of vege- tation and alluvial and colluvial d e p o s i t s. Larger streams cut through the storm beach and drain part of the area behind it. Within a period estimated to be about 5 years, areas of extensive deposits, such as the bay-head sediments of MacLeod Harbor, will be regraded mainly by fluvial action before a new storm beach develops. The break in slope between the top of the beach and a cliff is al- most completely obscured after a few years unless the cliffs are par- ticularly resistant. Talus and land- slides may degrade some or all of the cliff, unless the rocks weather to fine-grained material and are stabilized at an early stage by vege— tation. Where the top of the beach is backed by another older beach, quite large (40 ft) differences in sea level were found to produce only low (5 ft) topographic steps. These steps could be so obscured by unconsolidated beach deposits and vegetation that a series of raised beach levels might appear on the surface to be one continuous slope. Raised beaches formed at a known date and having their ini- tial profiles almost perfectly pre- served are very rare, but they did exist on Montague Island immedi- ately after the 1964 earthquake. Thus an ideal opportunity was pro- vided to study the rates and proc- esses of destruction of raised beaches. Such studies of presently forming geomorphic features are an excellent basis for understand- ing and identifying similar fea- tures formed in the past. EROSION AND DEPOSITION ON A RAISED BEACH, MONTAGUE ISLAND H41 Barrett, P. J ., 1966, Effects of the 1964 Alaska earthquake on some shal- low-water sediments in Prince Wil- liam Sound, southeast Alaska: Jour. Sed. Petrology, v. 36, no. 4, p. 992—1006. Plafker, George, 1965, Tectonic defor- mation associated with the 1964 Alaska earthquake: Science, v. 148, no. 3678, p. 1675—1687. Plafker, George, 1967, Surface faults on Montague Island associated with the 1964 Alaska earthquake: U.S. Geol. Survey Prof. Paper 543—G, 42 p. REFERENCES CITED Plafker, George, and MacNeil, F. S., 1966, Stratigraphic significance of Tertiary fossils from the Orca Group in the Prince William Sound region, Alaska, in Geological Sur- vey research 1966: U.S. Geol. Sur- vey Prof. Paper 550-B, p. B62—B68. Plafker, George, and Mayo, L. R., 1965, Tectonic deformation, subaqueous slides, and destructive waves as- sociated with the Alaska March 27, 1964, earthquake—an interim geo- logic evaluation: U.S. Geol. Survey open file report, 21 p. Plafker, George, and Rubin, Meyer, 1967, Vertical tectonic displace- ments in south-central Alaska dur- ing and prior to the great 1964 earthquake: Jour. Geosciences [Osaka City Univ.], v. 10, art. 1—7, p. 1—14. Sparks, B. W. 1960, Geomorphology: London, England, Longmans, Green and 00., Ltd., 371 p. Wolman, M. G., 1954, A method of sam- ling coarse river-bed material : Am. Geophys. Union Trans, v. 35, no. 6, p. 951—956. U.S. GOVERNMENT PRINTING OFFICE: 1968 0—280-547 he Alaska Earthquake March 27, 1964 W _ [Tectonics U.s.an. ' , I * GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—1, E... , w‘ - , - .a-..w\ ”avg” _ _ ,N _ .._..m..,m. .mengm TECTONICS OF THE MARCH 27, 1964 ALASKA EARTHQUAKE with“: w ‘ uuu‘ Former sea floor at Cape Cleare, Montague Island, Prince William Sound, exposed by 26 feet of tectonic uplift. The surf-cut surface, which slopes gently from the base of the sea cliffs to the water, is about a quarter of a mile Wide. The White coating on the rocky surface consists mainly of the desiccated remains of calcareous algae and bryozoans. Photograph taken at about zero tide stage, May 30, 1964. THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS Tectonics of the March 27, 1964 Alaska Earthquake By GEORGE PLAFKER “But it is time that the geologist should, in some degree, overcome those first and natural impres- sions which induced the poets of old to select the rock as the emblem of firmness—the sea as the image of tuconstancy.” (Lyell, 1871;, P. 179) GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—1 UNITED STATES DEPARTMENT OF THE INTERIOR WALTER J. HICKEL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1969 For sale by the Superintendent of Documents, US. Government Printing Office Washington, DC. 20402 THE ALASKA EARTHQUAKE SERIES The US. Geological Survey is publishing the results of its investigations of the Alaska earthquake of March 27, 1964, in a series of six Professional Papers. Professional Paper 543 describes the re— gional effects of the earthquake. Eight chapters in this volume have already been published. Other Professional Papers in the series describe field in- vestigations and reconstruction and the effects of the earthquake on communities, on the hydrologic regimen, and on transportation, utilities, and communications. A selected bibliography and an in- dex for the entire series will be published in the con- cluding volume, Profession Paper 546. Abstract _____________________ Introduction __________________ Acknowledgments ___________ Seismicity ____________________ The main shock _____________ The aftershocks _____________ Focal mechanism studies _____ Deformation __________________ Regional vertical displace- ments __________________ Methods of measurement--- Distribution of land-level changes ________________ Geometry of the deforma- tion ___________________ Earthquake faults ___________ Montague Island faults- - _ _ Other possible earthquake faults on land __________ Horizontal displacements _____ Methods of measurement-- Amount and distribution of the displacements _______ Relationship to regional ver- tical displacements and surface faults ___________ Page I 1 2 ooumppm 00 20 24\ 25 25 26 26 26 29 30 CONTENTS Deformation—Continued Time and rate of the deforma- tion ___________________ Earthquake-related move- ments __________________ Preearthquake movements- _ Postearthquake movements- Effects of the tectonic displace- ments ____________________ Comparison with other earth- quakes ___________________ Tectonic setting _______________ The Aleutian Arc ____________ Major features ____________ Seismicity ________________ Inferred regional stress pat- tern ___________________ Summary of the pre-Holocene (Recent) tectonic history of south-central Alaska__ Post-Miocene deformation- Late Eocene to early Oligo- cene deformation ________ Late Cretaceous to early Tertiary deformation_ - _ - Middle(?) Jurassic to Early Cretaceous deformation_ _ ILLUSTRATIONS PLATES [Plates 1, 2 are in pocket] 1, 2. Maps showing ground deformation: 1. In south-central Alaska. 2. In the Prince William Sound region. Page I 31 31 32 32 33 42 44 44 44 45 49 50 50 51 52 53 Tectonic setting—Continued Major surface faults in south- central Alaska ------------ The preearthquake Holocene (Recent) record of verti- cal shoreline movements-- Long-term Holocene emer- gence and submergence- _ Short-term tectonic submer- gence __________________ Tectonic implications of the record of Holocene verti- cal movements __________ Mechanism of the earthquake___ General considerations _______ Thrust-fault model __________ Outline of the model _______ Major unresolved problems- Representation by disloca- tion theory _____________ Steep-fault model ___________ Outline of the model _______ Representation by disloca- tion theory- _ _ _ - ________ Summary and conclusions ______ References cited _______________ 3, 4. Photographs showing effects of the earthquake: Facing page 3. On intertidal organisms -------------------------- I 16 4. On shoreline features ---------------------------- l7 Frontispiece. Cape Cleare uplifted surface. FIGURES 4—7. Photographs—Continued 5. Measuring barnacle—line height --------- 6. Barnacle line defined by barnacles, Fucus, and lichens _________________________ 7. Barnacle line on ship hull -------------- 8. Map showing variations in mean tide levels ______ Page 1. Map showing setting of the earthquake _________ I 3 2. Map showing seismicity of the earthquake _______ 6 3. Isobase map _________________________________ 9 4-7. Photographs: 4. Zoned intertidal marine organisms, aerial view _______________________________ 13 9. Variation of tidal parameters with mean tide levels _____________________________________ VII Page 153 55 55 60 67 68 68 69 70 71 Page I 13 14 14 15 16 VIII 10. Graph showing relationship between high-tide curve and barnacle-line height _______________ 11—18. Photographs: 11 12 13. 14. 15. 16. 17. 18. . Postearthquake barnacle line on Kodiak Island ______________________________ . Drowned land plants in Harriman Fiord, Prince William Sound _______________ Wildflowers growing on uplifted surf-cut surface _____________________________ New storm beach on uplifted surf-cut platform ____________________________ Map showing source area of the seismic sea-wave crest _______________________ Map showing horizontal displacements--- Diagram showing relationship between adjustments 1 and 2 _________________ Relationship between horizontal and ver- tical displacements ___________________ 19—24. Photographs: 19 20 21 22 23 . Effect of tectonic subsidence on terrestrial vegetation, Resurrection Bay __________ . Uplifted surf-cut platform at Cape Cleare- . Bay-head deposits incised after uplift at MacLeod Harbor ____________________ . Barnacles and Littorina in trees __________ . Remains of kelp and desiccated calcareous organisms on uplifted surface at Cape Cleare ______________________________ . Remains of starfish in pool on uplifted sur- face at Cape Cleare __________________ OILJKODNJH CONTENTS Page I17 25. 26 27. 18 28. 18 29. 19 30. 31. 19 32. 22 33. 27 34. 29 35—39. 30 34 35 :2- 36 42. 43. 37 44. 37 TABLES . Tide-gage observations ____________________________________________________ . Seismic sea-wave data ____________________________________________________ . Horizontal displacements from triangulation _________________________________ . Comparative earthquake deformation _______________________________________ . Radiocarbon dates ________________________________________________________ Diagram illustrating effects of horizontal tectonic displacement on confined water bodies ________ Well records from subsided zone ________________ Photograph showing roadway damage from subsidence on Kodiak Island ________________ Photograph showing efiect of uplift at Orca Inlet canneries _____________________________ Map showing tectonic setting of the earthquake- _ Map showing earthquake epicenters, 1954~63__ - - _ Structure section of deformed early Cenozoic rocks- Structure section of deformed probable late Mesozoic rocks ____________________________ Map showing areas of preearthquake tectonic submergence and radiocarbon sites _________ Age-displacement graph of radiocarbon-dated material ---------------------------------- Photographs: 35. Drowned mountains, southern Kenai Pe- ninsula _____________________________ Uplifted marine terraces on Middleton Island ------------------------------ Raised beach at Controller Bay__-_ ______ 38. Drowned trees on Latouche Island _______ 39. Drowned tree stumps at Cape Suckling--- Alternative fault models for the earthquake-- Block diagram of thrust-fault model ___________ Sequence of deformation, thrust-fault model _____ Correlation of thrust fault-extension model with dislocation theory -------------------------- Correlation of steep fault model with dislocation theory ____________________________________ 36. 37. Page I 10 23 28 43 54 Page I 39 41 42 35 56 57 58 59 59 61 61 63 63 66 68 69 THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS TECTONICS OF THE MARCH 27, 1964, ALASKA EARTHQUAKE The March 27, 1964, earthquake was accompanied by crustal deformation—- including warping, horizontal distor- tion, and faulting—over probably more than 110,000 square miles of land and sea bottom in south-central Alaska. Regional uplift and subsidence occurred mainly in two nearly parallel elongate zones, together about 600 miles long and as much as 250 miles wide, that lie along the continental margin, From the earthquake epicenter in northern Prince William Sound, the deformation ex- tends eastward 190 miles almost to long 142° and southwestward slightly more than 400 miles to about long 155°. It extends across the two zones from the chain of active volcanoes in the Aleu- tian Range and Wrangell Mountains probably to the Aleutian Trench axis. Uplift that averages 6 feet over broad areas occurred mainly along the coast of the Gulf of Alaska, on the adjacent Continental Shelf, and probably on the continental slope. This uplift attained a measured maximum on land of 38 feet in a northwest-trending narrow belt less than 10 miles wide that is exposed on Montague Island in southwestern Prince William Sound. Two earth- quake faults exposed on Montague Island are subsidiary northwest—dip- ping reverse faults along which the northwest blocks were relatively dis- placed a maximum of 26 feet, and both blocks were upthrown relative to sea level. From Montague Island, the faults and related belt of maximum uplift may extend southwestward on the Continen- tal Shelf to the vicinity of the Kodiak group of islands. To the north and northwest of the zone of uplift, subsi- dence forms a broad asymmetrical downwarp centered over the Kodiak- Kenai-Chugach Mountains that aver- ages 21/2 feet and attains a measured maximum of 71/2 feet along the south- west coast of the Kenai Peninsula. Maximum indicated uplift in the Alaska and Aleutian Ranges to the north of the zone of subsidence was 11/2 feet. Re- By George Plafker ABSTRACT triangulation over roughly 25,000 square miles of the deformed region in and around Prince William Sound shows that vertical movements there were accompanied by horizontal dis- tortion, involving systematic shifts of about 64 feet in a relative seaward di- rection. Comparable horizontal move- ments are presumed to have affected those parts of the major zones of uplift and subsidence for which retriangula- tion data are unavailable. Regional vertical deformation gener- ated a train of destructive long-period seismic sea waves in the Gulf of Alaska as well as unique atmospheric and iono- spheric disturbances that were recorded at points far distant from Alaska. Warping resulted in permanent tilt of larger lake basins and temporary re- ductions in discharge of some major rivers. L'plift and subsidence relative to sea level caused profound modifica- tions in shoreline morphology with at- tendant catastrophic etfects on the near- shore biota and costly damage to coastal installations. Systematic horizontal movements of the land relative to bodies of confined or semiconfined water may have caused unexplained short- period waves—some of which were highly destructive—observed during or immediately after the earthquake at certain coastal localities and in Kenai Lake. Porosity increases, probably re- lated to horizontal displacements in the zone of subsidence, were reflected in lowered well-water levels and in losses of surface water. The primary fault, or zone of faults, along which the earthquake occurred is not exposed at the surface on land. Focal—mechanism studies, when con- sidered in conjunction with the pattern of deformation and seismicity, suggest that it was a complex thrust fault (megathrust) dipping at a gentle angle beneath the continental margin from the vicinity of the Aleutian Trench. Movement on the megathrust was ac- companied by subsidiary reverse fault- ing, and perhaps wrench faulting, with- in the upper plate. Aftershock distri- bution suggests movement on a segment of the megathrust, some 550—600 miles long and 110—180 miles wide, that under— lies most of the major zone of uplift and the seaward part of the major zone of subsidence. According to the postulated model, the observed and inferred tectonic displace- ments that accompanied the earth- quake resulted primarily from (1) rela- tive seaward displacement and uplift of the seaward part of the block by movement along the dipping mega- thrust and subsidiary faults that break through the upper plate to the surface, and (2) simultaneous elastic horizontal extension and vertical attenuation (subsidence) of the crustal slab behind the upper plate. Slight uplift inland from the major zones of defamation presumably was related to elastic strain changes resulting from the overthrust- ing; however, the data are insuflicient to permit conclusions regarding its cause. The belt of seismic activity and major zones of tectonic deformation associated with the 1964 earthquake, to a large extent, lie between and parallel to the Aleutian Volcanic Arc and the Aleu- tian Trench, and are probably geneti- cally related to the arc. Geologic data indicate that the earthquake-related tectonic movements were but the most recent pulse in an episode of deforma- tion that probably began in late Plio- cene time and has continued intermit- tently to the present. Evidence for pro- gressive coastal submergence in the de- formed region for several centuries pre- ceding the earthquake, in combination with transverse horizontal shortening indicated by the retriangulation data, suggests pre-earthquake strain directed at a gentle angle downward beneath the arc. The duration of strain accumu- lation in the epicentral region, as in- terpreted from the time interval during which the coastal submergence occurred, probably is 930—1360 years. 11 12 Among the most notable aspects of the 1964 Alaska earthquake was the great areal extent and amount of the tectonic movements that ac- companied it. From the epicenter in northern Prince William Sound, the region affected by tec- tonic deformation parallels the trends of the Aleutian Volcanic Are, the Aleutian Trench, and the Gulf of Alaska coast for about 600 miles (fig. 1). In south—central Alaska, where the northeastern end of the arc intersects the con- tinent at an oblique angle, the pat— tern of deformation can be ob— served in an exceptionally com- plete profile extending more than 200 miles northward from the sea- ward edge of the Continental Shelf across the northeastern end of the volcanic arc. Within this region, tectonic displacements on land include absolute vertical movements ranging from as much as 38 feet of uplift to 71/2 feet of subsidence, relative horizontal movements of about 64 feet, and dip-slip offset on reverse faults of as much as 26 feet. Fur- thermore, the available data indi- cate that these movements extend- ed over a large segment. of the ad- ljacent offshore area where they may have been as large, or larger, than those measured on land. This report substantially en-~ larges upon a preliminary suin- mary and interpretation of the tee- tonic movements that accom- panied the 1964 earthquake (Plaf— ker, 1965). It presents the avail— able data on the distribution and nature of the earthquake-related displacements and of the mani- fold, often disastrous, effects of these movements. Geologic, geo— detic, and seismologic data perti- nent to the tectonics that were available to the writer prior to ALASKA EARTHQUAKE, MARCH 27, 1964 INTRODUCTION completion of this report in July 1967 are summarized. Implici- tions of the data for the earth- quake mechanism are reviewed, and a tentative qualitative model is outlined, which attempts to eX-‘ plain most of the observations. ACKNOWLEDGMENTS Field mapping of the vertical shoreline displacements and sur- face faults was accomplished by a party headed by the writer from mid—May through August 1964 and during 1 month in 1965. Geo- logical Survey personnel included L. R. Mayo, J. B. Case, S. L. Rob— bins, and William Bastian during 1964, and Mayo and M. G. Bonilla during 1965. An especially large part of the fieldwork and data compilation were carried out by L. R. Mayo during both field sea— sons. G Dallas Hanna, marine bi— ologist of the California Academy of Sciences, spent 3 weeks with the party in Prince William Sound in 1964, during which time he studied effects of the earthquake on the ecology of the intertidal fauna and flora and provided invaluable advice on the use of sessile marine o r g a n i s m s for determining changes in land level. The US. Geological Survey re« search vessel, Don J. Miller, was used as a base of operations for work in Prince William Sound and Resurrection Bay during the 1964 season. Helicopters and fixed- wing aircraft were used during both field seasons for work along shorelines elsewhere. Outstanding logistical support of the field in- vestigations was provided—often under difficult circumstances—by the crew of 'the Don J. Miller, which consisted of 'Capt. John Stacey and cook—seaman John Muttart of the U.S. Geological Survey, and by bush pilots Jim Osborne, Glenn Wheeler, “Stinky” Myers, Al Cratty, Bob Leonard, Oren Hudson, and Bob Barnett. R. L. Detterman, Reuben Kaoh— adoorian, T. N. V. Karlstrom, G. W. Moore, and B. L. Reed, all of the Geological Survey, contri- buted data on changes in land level as determined by coastal resi- dents of Cook Inlet and the Kodiak group of islands. Many residents of Alaska gave helpful information on earthquake effects, including tectonic changes, in about 150 interviews with the writer and on numerous form questionnaires. Virtually all the seismologic, ge- odetic, and marine data incorpo— rated in this report were obtained by the US. Coast and Geodetic Survey as part of their massive investigations of this major seis- mic event. I am especially indebt— ed to the numerous individuals in the US. Coast and Geodetic Sur- vey who freely made available unpublished data and have dis- cussed their interpretations of these data with the writer on many occasions. Among these, special thanks for assistance are due S. T. Algermissen of the Seismology Divisidn, W. D. Barbee, R. J. Mal- loy, and E. W. Richards of the Office of Oceanography, and C. A. Whitten, E. J. Parkin and J. B. Small of the Geodesy Division. Postearthquake vertical aerial photographs were p r o v i d e d through the courtesy of Col. M. L. Falwell, US. Army, Fort Rich- ardson, and H. R. Cravat of the Photogrammetry Division, US. Coast and Geodetic Survey. R. A. Page, J r., of the Lamont Geologi- cal Observatory furnished a com- puter plot of the larger magnitude 150° TE CTONICS l60° 150c A :/ .‘ oiuilllmwlmwmmm ndreanof 'S \xi“ OCEAN \—--~ * Epicenter of the 1964 earthquake Approximate limit of human perceptibility Dashed where inferred 400 600 800 l | | 200 400 600 800 1000 MILES 1000 KILOMETERS EXPLANATION O A Shallow depth Intermediate depth (<70 km) (70—200 km) Large earthquake epicenters (M>7) for period 1904—52 (From Gutenberg and Richter, 1954) Aleutian Trench Approximate limit of landslides, avalanches, and ground cracks \\\\\8\\ Approximate area of major tectonic deformation Dashed where inferred 1.—Map of Alaska and adjacent areas showing the location of the 1964 earthquake, the area affected :by the earthquake HHHHHHHHHHHHHHH Aleutian volcanic arc ------- I so — - - - - — - - Approximate outer edge of continental shelf Depth in meters 3 epicenters of previous major earthquakes, belts of active volcanism, and the Aleutian Trench. 307—3634 0—70—2 I4 aftershock epicenters. William Stauder, S. J., provided data on focal mechanisms of the earth— quake and its aftershock sequence in advance of his own publication. D. F. Barnes and Roland von Huene of the Geological Survey made available unpublished re- sults of submarine geophysical and geological studies carried out THE MAIN SHOCK The 1964 earthquake was cen— tered in a sparsely inhabited, mountainous area of northern Prince William Sound in south- central Alaska. near the eastern end of the Aleutian Arc (fig. 1). Its epicenter was located by the US. Coast and Geodetic Survey (Wood, 1966, p. 62) at lat 61.06° N., long 147 .44° W. and its origin time was at about 5: 36 p.m., Fri- day, March 27, 1964, A.s.t. (03 : 36: 13.5, Saturday, March 28, 1964, G.m.t.). The hypocenter, or focus, could not be determined more closely than between 12 and 31 miles (20—50 km) in depth. Magnitude of the earthquake, based upon surface-wave ampli- tudes (M8) is estimated to have been about 8.4 (Pasadena seismo— graph station). Earthquake vibra- tions were felt over an area in ex- cess of a million square miles in Alaska and adjacent parts of Can- ada (fig. 1), and they caused wide- spread damage throughout an area of about 50,000 square miles in south-central Alaska. The mani- fold effects of the shaking, which have been described in numerous reports are concisely summarized by Hansen and Eckel (1966) in the introductory volume of this series of reports. ALASKA EARTHQUAKE, MARCH 27, 1'9'64 in south-central Alaska after the earthquake. I am grateful to colleagues in the Geological Survey for very helpful discussions of the various tectonic aspects of the earthquake during which many of the ideas incorporated in this report were generated. The manuscript has been improved through critical re- views by E. B. Eckel and G. D. Eberlein and reviews of portions by D. F. Barnes, R. R. Coats, R. O. Burford, and C. B. Raleigh. Vec- tor shifts for the readjustment of triangulation data were cal— culated by J. T. Minta and C. R. Lloyd of the Geological Survey’s Topographic Division. SEISMICITY Rupture along a fault of con- siderable length is suggested from (1) the exceptionally long dura- tion of strong ground motion, (2) the character of the radiated seis- mic surface and body waves, and (3) the extensive belt over which aftershocks were distributed. Eyewitness accounts indicate that both the amplitude and dura- tion of ground motion definiwa tended to be largest in areas of relatively unconsolidated satur- ated deposits and least in areas of crystalline or metamorphic rocks. Within the immediate area af- fected by the earthquake, the only known instrumental records of the duration of shaking were made on several automatic recording charts in a steam powerplaznt built on bed- rock at Whittier. On the clearest of these records, trace vibrations lasted for nearly 4 minutes. At other bedrock sites on Kodiak Is— land and on the Kenai Peninsula, where the motion was timed by db- servers wit-h pocket watches or clocks, it was 21/2—5 minutes. Anomalously short durations, ranging from only 15 seconds to 11/2 minutes, were reported by residents at Seldovia and in two nearby localities in areas under- lain by metamorphic rock at the southwestern tip of the Kenai Mountains (Plafker, Kachadoori- an, Eckel, and Mayo, 1969). According to most eyewitness accounts, the earthquake started without prior warning as a gentle rolling motion that lasted for a period of 20 seconds to 1 minute, shook hard for as much as 4 min- utes, and then gradually subsided. There were no foreshocks percep- tible to observers such as are known to immediately precede many great earthquakes. A few observers heard premonitory low rumbling sounds se a1 seconds before the earthquake was felt. The ground motion was variously described as a rolling wavelike motion, a strong horizontal accel- eration, or a hard jarring motion. A few individuals noted that the ground motion during the earth- quake periodically eased up for a short period and then resumed ., with increased Violence. Reports of the directions of vibrations vary Widely, and at some places, prevailing vibration directions reportedly changed dur— ing the earthqiiake. The majority of the reports indicate a tendency for ground oscillation to be grouped in the quadrants between northwest-southeast and north- east-southwest although it also was reported from all other quad— rants. A westward propagation of vibrations between Anchorage and Kenai is indicated by the fact that power failure due to shaking in the Anchorage area caused an overload of the circuits at the in— terconnected Bernice Lake power- plant near Kenai, 53 miles farther from the epicenter, some 15—20 seconds before the plant superin- tendent felt the tremors. On the basis of a study of seis- mic surface waves, Toksiiz, Ben- Menahem, and Harkrider (1965, p. 154) determined that the rup- ture propagated S. 50° W. fro-m the epicenter of the main shock for a distance of about 370 miles (600 km) at an average velocity of 3 km per sec. Furumoto (1967) found from analysis of Rayleigh waves recorded by the Kipapa, Hawaii, strain seismometer that the rupture. more probably propa- gated S. 30 : 5° W. for 500 miles (800 km) at this same velocity. A detailed study of P—phases by VVyss and Brune (1967) suggests that the rupture actually broke in a complex series of events at an average propagation velocity. of 3.5 km per my, and that each event was characterized by more or less distinct high-amplitude bursts or events. Their data further indicate that, although the rupture propa- gated initially in various azimu— thal directions, after an elapsed time of 44 seconds it continued only in a southwesterward direc- tion. THE AFTERSHOCKS The epicenters and depth distri- butions of 598 aftershocks with Richter magnitudes equal to or greater than 4.4 recorded through December 31, 1964, are shown on figure 2. Of these; the largest shock had a magnitude of 6.7, six were larger than 6.0, 127 were between 5.0 and 6.0, and the remainder were less than 5.0 (US. Coast and Geo— detic Survey, Preliminary Deter- mination of Epicenter cards). During this same period of time, thousands of smaller aftershocks were also recorded. TECTONICS The aftershock sequence dimin— ished rapidly in frequency and' in- tensity. All aftershocks with mag- nitudes larger than 6.0 recorded teleseismically by the US. Coast and Geodetic Survey, occurred within the first several hours and most of those that were strongly felt occurred within 3 weeks of the main event. Daily frequency of all shocks dropped rapidly from a high of about 120 the first day > after the earthquake to 30 in 5 days, 15 in 10 days, and a steadily decreasing number thereafter (Jordan, Lander, and Black, 1965, p. 1324). As noted by Jordan and associates, the number of after- shocks at the beginning of the series may be much larger than indicated, because the first after- shocks are commonly masked by the train of large amplitude waves generated by the main shock and these waves may have a duration of several hours at distant stations. The larger aftershocks 0154.4) were concentrated mainly in an arcuate belt 600 miles long by as much as 200 miles wide that roughly parallels the continental margin. From the epicenter of the main shock it extends 425 miles southwestward 'to the vicinity of the Trinity Islands and about 175 miles eastward nearly to long 142° W. (fig. 2). Most of the largest aftershocks in this belt (leé5.0) were situated over that part of the belt lying seaward from the zero isobase between the major zones of uplift and subsidence. Smaller aftershocks (.llé4.4) were spread over a larger area that extends in- land from the zero isobase beneath the coastal mountains that border the Gulf of Alaska, and the small- est aftershocks (1l[<4.4) were scattered over an even larger area. There was a distinct concentration of activity within the aftershock belt on the Continental Shelf at the southwestern end and in the area southeast of Prince William Sound 15 near the northeastern end. Burk (1965, p. 150) noted that the fairly sharp southwestern limit of the aftershock belt approximately coincides with a transverse struc— tural boundary on the Alaska Pe- ninsula, and he suggested the pos- sibility that a transverse fault marks the southwestern margin of the crustal block involved in the 1964 earthquake. Only about 3 per- cent of the aftershocks with mag- nitudes of 4.4 or more were outside the main belt of activity along the continental margin. Their epicen- ters were widely scattered, mainly in the Shelikof Strait—Cook Inlet areas and on the ocean floor sea- ward from the Aleutian Trench axis. Significantly, none of the aftershocks were centered as far inland as the chain of active vol- canoes in the Aleutian Range, southern Alaska Range, and the lVrangell Mountains. The possi- bility that they were directly re- lated to vulcanism is thus ruled out. Depth distribution of the after- shocks is less perfectly defined than their epicentral positions be- cause of ( 1) inherent errors in the determination of hypocenters in areas of uncertain crustal struc— ture and seismic velocity, and (2) the wide spacing of the seismo— graphs on which the shocks were recorded. Algermissen (1965) found that hypocenters of after— shocks with magnitudes greater than 5.0 were at depths between 3 and 25 miles (5—40 km) and they average about 12 miles (20 km). Only 25 of the earthquakes with magnitudes of 4.4 or more were deeper than 22 miles (35 km), the deepest ones being less than 56 miles (90 km). According to Page (1967; also unpub. data), the mi- croaftershocks—which had a spa— tial distribution similar to that of the aftershocks—were at depths of 22 miles (35 km) or less. He I6 62° ~ 58° 56" — ALASKA EARTHQUAKE, MARCH 27, 19164 150° 148° 146° 144° 142° l55° . 54° 152° I I x I I f l l I (:0 TALKEETNA MOUNTAINS _ e I ?‘ 2&4“, Q- {\n: ZEUNTEELL +3. “u' CHUGACHI @ as *5“ "’Ou Anchom e Ep nter u x ‘3,” NTA 1* ‘0‘ g f“ , ’+’Vs . \\ ~~ x + dig: GP'RI ~ I» ‘ xx x an ILLA ++ g \Q 00 SOUU + 6&2: ® \ Q + 9% 3 ? 7o 0 m I 60 x e . e HYPOCENTRAL DEPTH. I so w 3 IN KILOMETERS 1 En Tr nity El<20 | 401-2 I xyflslanisx X E] 21_35 :2 o’ ix 3645 30 L5 2 Chirikoflslandx x" + , 111 46—60 m2 " I 61*90 20 I; E“ a o 50 100 150 MILES x 2; l I I I I I’ I f m M 10 D 0 50 100 150 KILOMETERS “I Z .J o 100 o 110 MILES 8U BMARINE CONTOU RS IN FEET DISTANCE PERPENDICULAR T0 ZERO ISOBASE l l l l l | EX PLANATION + X A . I * <20 21 35 36 45 46 6° 61 90 Active or dormant volcano Epicenter of aftershock showing depth of focus, in kilometers Aftershock data after R. A. ——"""—-"e- -------- Page, Jr., Lamont geological Observatory of olumbia Approximate zero isobase be— . . . tween major zones of tectonic 115112;???) ( written commun.. uplift and subsidence 2.—Distributi0n and depth of aftershocks (4.4éME6.7) from March 27, to December 31, 1964. suggests that subcrustal earth- quakes in and near the aftershock region may have represented the normal seismicity of the region. Aftershock hypoeenters do not fall into any well-defined planar zone although there is a vague tendency towards a slight deepening of their lower limit beneath the continent. The proportion of hypocenters deeper than 12 miles relative to those shallower than 12 miles also shows an increase in the same di- rection (fig. 2). Large earthquakes at shallow and intermediate depths are thought by most geologists and ge- ophysicists to result from sudden rupture, or faulting, in strained rocks. Aftershocks which follow large earthquakes presumably rep— resent continuous adjustments of the strained volume of rock, or focal region, within which fault- ing occurred (Beniofi, 1951). According to this model, there- fore, faulting associated with the 1964 earthquake was largely con- fined to the part of the continental margin extending roughly 150—200 miles northward from the axis of the Aleutian Trench, it was lim- ited in depth to the crust, or per— haps the uppermost part of the mantle, and its lower limit may deepen slightly beneath the arc. FOCAL MECHANISM STUDIES Mechanism studies of the main shock, of a number of larger after- shocks, and of one preshock that occurred about 7 weeks prior to the earthquake provide data rele- vant to the fault orientation and sense of displacement at the earth— quake foci. Body—wave solutions define a pair of orthogonal planes at the focus, one of which presum- ably contains the active fault sur- face. Inherent in the. focal mech- anism studies are the basic as- sumptions of an elastic-rebound TE CTONICS source and initial displacements at the earthquake foci that approxi- mately reflect the regional stress field. Focal mechanism studies of P- waves for the main shock yield one well-defined nodal plane that strikes between N. 61° and 66° E. and dips 82°—85° SE. (Stauder and Bollinger, 1966. Algermissen, 1965; written commun., March 19, 1965). The alternative low-angle plane is restricted by the data to a plane that dips towards the northwest; Algermissen’s solu- tion suggests an inclination of about 8°. If the well-defined nodal plane is regarded as the fault plane, its strike was N. 61°—66° E. and the motion was predominantly dip-slip on a steep reverse fault with the southeast side relatively upthrown. Alternatively, if the steep nodal plane is considered to be normal to the motion, the fault would be a northwest-dipping thrust. Focal mechanism studies of the main shock alone cannot distinguish which of these two planes is the fault plane. Surface- wave studies that define the direc- tion of rupture may permit dis— tinction between the fault and auxiliary planes in cases where the strikes of the two planes differ significantly. However, as noted by Savage and Hastie (1966, p. 4900), surface waves do not per- mit a unique solution for the 1964 earthquake because the direction of rupture propagation is essen- tially the same for either plane. Stauder and Bollinger (1966) determined focal mechanisms, based on combined P-wave first- motion and S-wave polarization data, for a preshock and 25 after- shocks in the Kodiak Island and Prince William Sound areas. Most of these solutions have one near— vertical nodal plane that resembles the well-defined nodal plane for the main shock; the other dips I7 5°—15° to the northwest or north. Strike of the steep plane is be- tween N. 50° and 72° E. to the southwest of Prince William Sound; it is variable in the Prince William Sound area and nearly east-west to the east of the sound. This systematic variation in orien- tation of the steep nodal plane tends to follow a change in trend of tectonic features along the coastal belt. Four aftershock so- lutions in the Prince William Sound area and one located sea- ward from the Aleutian Trench off Kodiak Island, however, are anomalous in that they do not cor— respond to this general pattern. The preshock and all the after- shock fault-plane solutions are subject to the same ambiguity of interpretation as the main shock. Thus, on the basis of all the in- dividual solutions (except for the five apparently anomalous ones) the motion at the source of the earthquake and the related pre- and aftershocks may be ( 1) al- most entirely dip-slip on a steeply dipping reverse fault along which the seaward side is relatively up- thrown or (2) dip—slip on a northward-dipping thrust fault along which the landward block overrides the seaward block in an average S. 25° E. direction to the southwest of Prince William Sound and in a S. 10°—15° W. di- rection to the east of Prince Wil— liam Sound (Stauder and Bollin- ger, 1966, p. 5295). In considering the solutions in relation to one another, Stauder and Bollinger observe that in the first alterna- tive the faulting may consist of en echelon segments that follow a sinuous path roughly paralleling the curving trend of both the af— tershock belt and the zero isobase between the major zones of earth- quake-related vertical tectonic de- formation (fig. 2). In the second alternative the thrust plane has a 18 dip of less than 14° beneath the continent, displacement of the up- ! per plate is relatively seaward, and the direction of motion is roughly normal to the trend of ‘ the aftershock belt and the zones . of tectonic deformation. 1 The steep plane in solutions of the main shock, the preshock, and 1 most aftershocks differs in strike from the tectonic trends of the region, the orientation of the ‘ earthquake focal region, and the pattern of vertical displacements ‘associated with the earthquake. Crustal deformation, including both vertical and horizontal move- ments, associated with the 1964 Alaska earthquake was more ex- tensive than any known to have (been related to a single tectonic ‘event. Vertical movements oc- curred over an arcuate region that roughly parallels the continental lmargin for almost 600 miles from the southwestern tip of the Kodiak ‘group of islands northeastward through Prince William Sound and thence eastward to about long 142° W. (fig. 3). In a northwest to southeast direction, the deforma- tion extends at least 200 miles from the west shore of Cook Inlet to Middleton Island at the seaward edge of the Continental Shelf. In addition, crustal warping appears ‘to extend inland as far as the Alas- ‘ka Range and it may extend sea— ward to the axis of the Aleutian ‘Trench. ‘ Observable tectonic deforma— ‘tion involved ( 1) regional crustal warping, including both uplift and subsidence relative to sea level, in broad zones that roughly parallel the trend of the continental mar— gin, (2). systematic regional hori- zontal extension and shortening in ALASKA EARTHQUAKE, MARCH 27, 1964 These differences suggest to Stan- der and Bollinger (1966, p. 5293— 5294) that the steep plane corre- sponds to the auxiliary plane rather than the fault plane. They also interpret the spatial distribu- tion of foci for which mechanism studies were made and the nature of the inferred motions as being more compatible with thrusting on the shallowly dipping plane than to movement on the steep plane. Solutions for the main shock and most of the other shocks sug— gest a maximum-stress axis at the foci of these earthquakes oriented approximately normal, or at a small oblique angle, to major structural elements within the fo- cal region. Within the limitations of the data, this orientation is in reasonably good agreement with geodetic and geologic data cited in subsequent sections (p. I 49) that suggest dominantly tangen- tial compressive stress on land within the region afl'ected by tec- tonic movements during the earthquake. DEFORMATION a direction approximately trans- verse to that of the zones of warp- ing, (3) displacement across longi— tudinal reverse faults exposed on land and on the sea floor, and (4) possible displacement on at least one wrench fault. Evidence for these various movements, their manifold effects, and their tectonic significance are discussed below. REGIONAL VERTICAL DISPLACEMENTS Notable tectonic changes in land level during the 1964 earthquake occurred over an area of at least 70,000 square miles, and probably more than 110,000 square miles, of south-central Alaska. The areal distribution and approximate amount of the vertical displace— ments are summarized on figure 3. Plates 1 and 2 show data points where quantitative measurements of vertical displacement were made, as well as the method used and year of measurement. Also shown on the figure and plates are isobase contours, or lines of equal vertical displacement based on these data, and the approximate axes of the major upwarped and downwarped zones. The deforma— tion includes two broad major zones of warping, each about 600 miles long and as much as 130 miles wide. The seaward zone is one of uplift that includes a fringe of coast along the Gulf of Alaska, the adjacent Continental Shelf, and perhaps the continental slope; it is bordered to the northwest and west by a zone of subsidence. Slight up- lift also occurred in at least three areas extending inland from the major zone of subsidence as far northward as the Alaska Range. METHODS OF MEASUREMENT Quantitative information on vertical displacements along the coast (pls. 1, 2) comes mainly from (1) comparison of pre- and post- earthquake tide gage readings, (2) the position of the upper growth limit of certain sessile intertidal organisms relative to sea level, (3) differences in the pre- and post- earthquake positions of the upper growth limits of sessile intertidal organisms or the lower growth limit of terrestrial vegetation, (4) differences in the heights of pre- and postearthquake storm beaches, (5) estimates or measurements of changes in the position of shoreline TECTONICS 158° 156° 154° 152° 150° 148° 146° 14‘4" 149° 1 i l , I I l | EXPLANATION 64° ~ __53.._ ................ _‘ Isobase contour, showing uplift (+) or subsidence (—) in feet Dashed where approximately 10- ‘ sated; dotted where inferred ( Q . ‘14, ' M ch Ge ‘“‘—-—‘ __ - 015%“: $93" Approximate axis of maximum uplift ( “$1,;th ,/' / ./' —-—~——+——~— w Approximate axis of maximum subsidence \\\\\\\\\\\\\\\\‘\ .‘_ 159° ~ Possible zone of slight uplift . h (less than 2 ft) ' “mane“ .. [IV * ”a... G "a U. S. Coast and Geodetic Survey ELL 9* "s. first order level net Chitina LIA/721i, . N ale ans : 8 Active or dormant volcano it 3* .. 5* 1 1-35 60 ~ 1 ‘ 1° ‘01 35° .. .' A PATTON BAY ’ r ‘ .... ' FAULT x, 8’ <2“ ‘ Hoiner Middleton X ‘ ‘W ..' dep‘l’ia IV ) Island/.3 v ,3 . 'l. y f A ' 1 m / .~ ’ ° ° 7/ i” ‘c' 58° ~ f / M H6 3 P 100 150 MILES 56° ~ " I [ ChirikM'S‘and 50 100 150 KILOMETERS _ l SUBMARINE CONTOURS IN FEET l y s v 1 3.—Map showing the distribution of tectonic uplift and subsidence in south-central Alaska. IQ I10 markers by local residents, and (6) measured changes in the height of tidal bench marks rela- tive to sea level. For those offshore areas where detailed preearth- quake bathymetry was available, approximate vertical displace- ments were also obtained from comparison of pre- and postearth- quake depth 8 o u n din g s. The amount and distribution of the vertical displacements inland from the coast (pls. 1, 2) was precisely determined along the highway— railroad routes between Seward, Anchorage, Fairbanks, and Val- dez by comparison of pre- and postearthquake level lines tied to tidal bench marks. Isobase contours plotted on plates 1 and 2, which were derived from all the sources listed above, represent absolute changes in alti- tude related to the earthquake. The accuracy of the contouring varies greatly from place to place de- pending upon the type and amount of data, but, because it is essen— tially an averaging process, the contouring tends to be most ac- curate in areas where many data points are available, and least ac— curate in areas where data are . sparse. In general, contours shown as solid or dashed lines are esti— mated to be accurate at least to within :1 contour interval; at most places they are probably ac- curate to within half an interval. The various techniques used for determining vertical displacement, estimates of their relative preci— sion, and sources of data are out- lined below. COMPARISON OF PRE- AND POSTEARTHQUAKE TIDE-GAGE READINGS The directions and relative amounts of vertical displacement were determined from coupled pre- and postearthquake tide- gage readings made by the US. Coast and Geodetic Survey at two ALASKA EARTHQUAKE, MARCH 27, 1964 TABLE 1.—Land-level changes based on comparison of pre— and postearth- quake tide series at tide stations in south-central Alaska [After Small (1966, p. 19—22) and U.S. Coast and Geodetic Survey (unpub. data, April 22, 1966). Astensk indicates statlon is plotted in fig. 10; double asterisk indicates standard tide-gage station operational before and after the earthquake] ‘ Length of tide series Land Location move- | ment ‘ Preearthquake Postearthquake (feet) Prince William Sound *Cordova __________ 1950—1951 (2 May—Nov. 1964 +6. 2 years) May-Nov. 1965 +6. 2 *Port Gravina _____ July lg—Sept. 26, July 3—31, 1964 +4. 5 191 June l5—Aug. 19, May 16—June 16, +4. 2 1915 1965 Port Fidalgo ______ Aug. 24—26, 1915 June lS—July 17, +2. 4 1965 Valdez ___________ 1924—26 (3 years) May—July 1964 —. 9 July—Nov. 1965 1 ——3. 5 *Whittier __________ June 12—29, 1956 May—June 1964; —5. 4 0ct.—Dec. 1964 July—Nov. 1965 1 -5. 7 *Chenega Island--- July—Aug. 1957 July 7—Aug. 4, +4. 8 1964 Green Island ______ Au1.g1213—Sept. 20, August, 1965 +6. 6 19 *Port Chalmers- --_ July—Sept. 1933 May 19—30; July +10. 5 (3 months) 7—Aug. 6, 1964 July 19—Aug. 23, +10. 5 1965 *Sawmill Bay, May, 1927 (1 May 20—June 1; +7. 2 Evans Island___ month) June 14—July 7, 1964 July 29—Aug. 31, +7. 2 1965 Hogg Bay ________ June 15-July 21, July 20—Aug. 16, +5. 8 1927 1965 Renal Peninsula and Cook Inlet Day Harbor ______ May 20—July 13, May 7—June 30, - 5 1928 1965 **Seward ___________ 1926-38 (13 June 1964—Jan ——3. 6 years) 1965 Jan—Sept. 1965 ——3. 6 Aialik Bay -------- July 6—Sept. 21, May—June 1965 —4. 5 1912 *Two Arm Bay_ _ -_ July 20—Sept. 18, May 14—June 30, -—5. 4 1 2 1965 Chance Cove ----- May 14—July 25, May 15—June 30, ——6. 6 1093 1965 Shelter Cove ______ June—Aug 1927 July 1965 —5. 4 (3 months) *Port Dick ________ Aug. l—Sept. 26, May 21—June 30, —6. 2 1930 1965 **Homer ___________ Oct. 3—Nov. 17, June—Dec. 1964; —5. 4 1962 Sept—Nov 1 —-5. 8 1965 *Port Chatham- _ _ _ May 22—June 4, —4. 6 196 *Seldovia __________ April 30—Oct. 8, June—Oct.1964 +3. 9 1908 May 8—Sept. 29, April-July 1965 --3. 8 1910 Nikiski ___________ —. 9 See footnote at end of table. June 18—July 31, 1964 TECTONICS TABLE 1.——Land-level changes based on companion of pre- and postearth- quake tide series at tide stations in south-central Alaska—Continued Length of tide series Land Location move- ment Preearthquake Postearthquake (feet) Kenai Peninsula and Cook Inlet—Continued Anchorage ________ June—Sept. 1910; May-Oct. 1964 —2. 6 June—Oct., Dec. 1922; May-Nov. 1923; May— Sept. 1924; May—Oct. 1925 April—Oct. 1965 —2. 3 Kodiak group of islands Carry Inlet _______ July 1931 July 8—Aug. 7, —3. 2 1965 Redfox Bay _______ July—Aug. 1926 JugJ 615—Aug. 8, —3. 4 *Tonki Bay ________ Aug. 23—31, 1932 July 10—Aug. 7, —5. 2 1965 Nachalni Island-__ July 30—Sept. 8, June 14—July 10, —3. 9 1941 1965 Dolphin Point_ _ __ Sept. 10—Oct. 1, June 7—Ju1y 8, —2. 9 1941 1965 St. Paul Harbor___ June—Oct. 1964 —5. 5 May—Nov. 1965 —5. 0 *Ugak Bay ________ August 1932 June 12-July 10, —4. 2 1965 Port Hobron ------ June l—Sept. 19, July 26—Aug. 27, —. 7 1928 1965 Jap Bay __________ July 1931 July 28—Aug. 24, . 0 1965 I"Lazy Bay ________ May 16—Sept. 29, June 11—30, July —. 4 1929 l—Aug. 14, 1964 June l—Sept. 30, July 28~Aug. l5, —. 2 1930 Aug. 20—24, 1965 *Larsen Bay _______ Aug. 19—22, 1929 June 13—Aug. 31, —2. 5 1964 July 18—Aug. 17, —2. 4 1965 Uyak Bay ________ July 14—30, 1908 July 22—Aug. 17, —1. 9 1965 *Port O’Brien ------ June 22—July 30, July Aug. 1964 —3. 6 1929 June 8—July 19, —3. 6 1965 **Women’s Bay _____ 1950—59 (10 April—July 1964 —5. 6 years) Alaska Peninsula **Kukak Bay _______ July—Aug. 1949 July 17—Aug. 16, —. 5 (2 months) 1965 1 Change between 1964 and 1965 possibly due to surflcial compaction. operative standard stations, and 34 temporary stations within the region afl'ected by tectonic move- ments during the earthquake (table 1). Such measurements were made at localities where a 307—634 0—70—‘3 series of pre-earthquake tidal ob- servations had been made and where tidal bench marks had been established that were not destroy- ed during or before the earth- quake. Ill The Seward and Womens Bay (Kodiak) stations were equipped with standard automatic tide gages; readings could therefore be compared for series taken imme- diately before and after the earth- quake. These determinations of vertical displacement are probably accurate to within a few tenths of a foot. The only other standard station in the region, at Homer Spit in Kachemak Bay, could not be used directly to measure verti- cal tectonic displacement because it was inoperative at the time of the earthquake, and the tectonic subsidence was augmented by pro- nounced surficial settlement of the unconsolidated deposits that made up the spit. Because the amount of surficial settling was known from level lines to areas of relatively firm ground (Waller, 1966a, p. D13), net tectonic subsidence could be obtained by subtracting the surficial effect from total sub- sidence relative to sea level at the gage. This difference amounted to about 2.9 feet. Accuracy of the changes deter- mined at the temporary stations depends largely on the length of the preearthquake series of obser- vations, some of which were as short as 4 days, and also on the time interval between the pre- and postearthquake series, which at one station is 54 years and at many is more than 30 years (table 1). Positions of the preearthquake tidal datum planes at these sta— tions may not have been precisely determined originally, and (or) they may have changed from their original values because of relative land-level changes in the time in- terval between tide observations. At most of the temporary stations such errors are believed to be small because, where other sources of data such as estimates by local residents or measured differences between pre- and postearthquake ~112 shoreline features are available in the immediate vicinity, they tend to agree with the tide data within 0.5 foot or less. An exception is the 7 .O—foot uplift of the Sawmill Bay station on Evans Island, Prince William Sound, derived from the tidal observations( table 1). It is about 2 feet too low when compared with (1) the 90-foot estimated uplift of shoreline fea- tures whose preearthquake heights were precisely known to residents of the area and (2) the 8.9 feet of uplift indicated by differences in the pre- and postearthquake upper growth limit of barnacles (pl. 2). The preearthquake tidal observations at this particular sta- tion consist of a 1—month series taken in 1927, 37 years before the postearthquake series with which it was compared. Therefore, either about 2 feet of uplift occurred in the area during the 37 years prior to the earthquake or the tidal meas— urements are in error. Because Sawmill Bay had been continu- ously inhabited during this time interval and because there was no indication of preearthquakc changes in land level, it appears probable that the latter alternative is the correct one. THE UPPER GROWTH LIMIT or SESSILE) INTERTIDAL ORGANISMS RELATIVE TO SEA LEVEL Vertical displacements in coast— al areas were determined mainly from more than 800 measurements of the upper growth limits of in— tertidal sessile marine organisms relative to sea level along the long, intricately embayed rocky coast. In Prince William Sound and Resurrection Bay, where we used outboard-motor-powered skiffs for studying these changes during 1964, measurements were made continuously along the shore at spacings of about 1—5 miles except in those places where cliffs, heavy surf, or floe ice in fiords prevented ALASKA EARTHQUAKE, MARCH 27, 1964 boat landings or where the upper growth limits were not well-de- fined. During both the 1964 and 1965 seasons, measurements of vertical displacement were also made at localities shown on plates 1 and 2 that were accessible by light plane or helicopter along the ocean coast, on the offshore islands, and around the shores of the K0- diak group of islands. In measuring land-level change from the displacement of sessile marine organisms relative to sea level, the zonation of plants and animals between tide marks was used—a zonation that has long been recognized by marine ecol- ogists. The intertidal zone along the predominantly steep and rocky coastline of south—central Alaska is inhabited by certain species of organisms—notably barnacles, mussels, and algae—Whose verti- cal growth limits are usually well defined (pl. 3A, facing p. I 16; fig. 4). The zonation of intertidal organisms in the Prince William Sound region was studied in detail by a party headed by G Dallas Hanna of the California Academy of Sciences in 1965. T0 the writer’s knowledge, there were no pub- lished preearthquake data on the intertidal ecology for any part of the Gulf of Alaska coast. In particular, the common acorn b a r n a cl e, Balanus balanoides (Linnaeus), and closely similar forms such as B. glandular, are widely distributed on rocky shores and form a conspicuous band with a sharply defined, readily recog- nizable upper limit (figs. 5—7) The common olive-green rockweed (Fucus distz'chus), which has an upper growth limit near that of the barnacles, served as a useful datum for measuring land-level changes along shores where bar- nacles were absent or poorly de— veloped (pl. 33, facing p. 114). The upper limit of this zone, re- ferred to as the “barnacle line,” corresponds roughly to the top of the Balanoid or Midlittoral Zone of Stephenson and Stephenson (1949) ; to Zone 2, the High Tide Region, or the Upper Horizon of Ricketts and Calvin (1962); and the Upper Intertidal Zone of Rigg and Miller (1949). The barnacle line usually, but not always, ap- proximates the lower growth limit of the dark-gray—to-black encrust- ing lichen (Vewucam‘a) which commonly forms a black band in the splash zone immediately above the barnacles and Fume (pl. 3A, facing p. I16; fig. 4). The upper limit of all intertidal organisms depends mainly on the ability of immature individuals to survive prolonged exposure to air and on the tidal characteristics at any given locality (Kaye, 1964, p. 591—592). In referring to the sur- vival ability of barnacles, Kaye h’as termed this maximum expo— sure interval the “lethal limit.” He found experimentally that it was close to 150 hours for yearling bar- nacles and ranged to an absolute maximum of 192 hours for mature barnacles. Fum, which has a nearly identical upper growth range, must have approximately the same “lethal limit.” To a lesser extent the upper growth limit of barnacles and F ucus depends upon a number of other factors which, in parts of Prince William Sound, locally cause the barnacle line to deviate as much as 0.6 foot from its average height at any given locality. Wave action during the lowest annual neap tides and pro- tection from desiccation by shady locations tend to elevate the upper growth limit; exposure to fresh water near large streams or tide- water glaciers tends to depress it. Annual variations in sea level may cause further slight upward or downward shifts of the organisms’ upper growth limits. TEC’I‘ONICS 4.—0haracteristic parallel bands formed by zoned intertidal marine organisms long unlifted west shore of Knight Island, Pn‘nce William Sound. Encrusting lichens form the upper dark band and brown laminarians form the lower one; light—colored bamacle zone is between the two. Photograph taken at about 2-foot tide stage, Ma‘s7 20, 1964. 5.—Measuring height of the bamacle line at Port Bainbridge in an area of the Kenai Peninsula uplifted 5.7 feet. Protograph taken at 4.4-t‘oot tide stage, June 21, 1964. 113 I14 ALASKA EARTHQUAKE, MARCH 27, 191644 6.——Barnac1e line clearly marked by upper limit of light-gray barnacles and desiccated rockweed in dark patches and by the lower growth limit of dark-gray encrusting lichens of the splash zone. Uplift of 3.2 feet in Whale Bay, western Prince William Sound. Photograph taken at 30-foot tide stage, June 22, 1964. iv- 5% ; " fill, % w‘k 7.—Barnacle line on hull of 8.8. Goldbrook, Middleton Island. The upper growth limit of {the dark band of barnacles on this vertical surface is clearly defined and at a uniform level, even though it is in a locality that was exposed to open-coast surf conditions prior to the earthquake. Indicated uplift is 11.7 feet. Photograph taken at 65-foot tide stage, July 26, 1965. 156" 154° 152° 150" TE CTONICS 148° 146° 144” 142” 0?: y. i l i Homer Seldovia 15.4 - - ,-..e Trinity lsla/n‘d’s/ {\fi fl .1; 36° « , . ~. L 1’ X 0 50 100 150 MILES o so 100 150 KILOMETERS Q—l—J 1 l Epicenter Anchorage ~93 7.7“ ’ii’ Kayak Island/ ”Yakataga 9 ~ x _ '; x . . , . T*--.' ,. 7.8 ,,....600 FEE it”, ’Middleton island 1 s i 8.—Yariations of mean tide range in areas of south-central Alaska where tide levels were used as a datum in measuring vertical displacement of the coast. Triangles show locations of permanent tide gages prior to the earthquake. Data from U.S. Coast and Geodetic Survey (1964). Specific data on the normal up- per growth limit of barnacles and Fucus relative to tide levels were unavailable for the part of coastal south—central Alaska that was af- fected by vertical tectonic dis- placements associated with the earthquake at. the time this study was made. However, elsewhere on the Pacific Coast of North Amer- ica and in other areas of the world the height of the barnacle line roughly approximates annual mean high water along shores with a small or moderate tidal range. In the region where measurements of vertical displacements were made during this study, tidal range, and hence the height of the barnacle line, differs from place to place along the coast. For exam- ple, the mean tide range (that is the difference in height between mean high and mean 10w water) varies from a minimum of 6.4 feet along the ocean coast of the K0— diak group of islands to a maxi- mum of 15.9 feet at Homer near the entrance to Cook Inlet (fig. 8). Even higher tides—~as high as 30.3 feet—prevail within Cook Inlet. However, the exceptionally large tidal range, in combination with a general lack of stable rocky shores and an increasingly impoverished marine fauna. toward the head of the inlet precluded use of marine organisms for measuring vertical displacement. On figure 9 are shown the var— iations in the 196-1 positions of six of the important sea—level datum I15 planes corresponding to mean tide ranges of 6.4 feet to 15.9 feet in south-central Alaska.1 Kaye (1964) has reviewed the definition of these tidal planes, the way in which they are derived from tide- gage records, and the control they exert on the zonation of intertidal organisms. Field procedure for determining land—level changes by the barnacle- line method was to measure the height of the upper limit of bar- nacle or Fucus growth above or be- low water level at any stage of tide. On smooth steep rocky slopes sheltered from heavy surf, this line is sharply defined and can be readily determined to within 0.2 foot or less. On sloping shores or shores exposed to heavy surf it tends to be less regular, although even under such conditions it gen— erally can be determined with con- fidence to within 0.5 foot. At most places where the barnacle line was above water, its height was measured with a hand level or sur- veyor’s level and stadia rod. Where the barnacle line was visible under water, its depth below the surface was measured directly with the stadia rod. Stage of tide at the time of measurement was then determined from the U .8. Coast and Geodetic Survey table of predicted tides for 1 Extreme high and low water a_re_the high est and lowest predicted annual tides; the extreme high-water line is also the approxi- mate lower limit for growth of terrestrlal vegetation along shores sheltered from wave splash. Because tides in south-central Alaska are of the semidiurnal type. the levels of the two daily high and low tides differ, varying between a maximum (spring tides) and a minimum (neap tides), depending on. the interactions of the tide-producmg forces. Thus. mean high. water (MHW) and: mean low water (MLW); are the average heights of all high and low tides within the period of measurement. Mean lower low_ water (MLLW), the mean. of the spring indies, is the tidal datum plane commonly used as the zero datum by the U.S. Coast and Geodetic Survey. Mean tide level (MTL) is the exact midpoint of the range MHW—MLW. It daf- fers by a few tenths of a foot from mean sea level, the average of all hourly tide read- ings over a given period, because the tidal curve is not a simple sine curve, but rather is compounded of a number of simple sine curves, some of which have fixed phase rela- tionships with respect to each other (Mar- mer, 1 5.1, p. 70). I 16 ALASKA EARTHQUAKE, MARCH 27, 1964 30 1— 14 I I I I I I I I I I I I | I | I I | I | I | | 28 ——12 — | I I I _ I I | | ' I I I 26 ——1o — I I | I — | | I I I I I I 24 --8 — I I I _ I I | I I | I I I | I 22 -—6 — I I I I _ I I | I | I I I I I I 20 ——4 — ' ' I I I I I — I I I I I I I ' I I I __ _ | a E 18 2 I I I I Lu | I | I u. l I I . I z . I I I Mean tIde level I — 15 ——o — f - - I q I.- | I I I I | | | I w I I I I II: 14 “ ‘2 " I I I I — In I I I | o I I I I ' I I‘ 12 —— —4 — I | I I _ I I I I I I I I I I 10 -— -6 — I I I i ' I | I I I I ' 8 " ’8 ‘ I I I _ I ' . I I I I 6—— —10 — I I I I _ I I I I I I _L e g .9 4-- —12 — g a S g _ E 3 a a :f til? 0 m ___ _ | I _ 2 14 I I I I | I I I I I I I I I I O __ —16 1 I I I I I I I I I I I 6 7 8 9 10 11 12 13 14 15 16 MEAN TIDE RANGE, IN FEET 9.—Tidral parameters in areas of south-central Alaska where tide levels were used as a datum in measuring vertical displacement of the coast. Data from US. Coast and Geodetic Survey (1964). the closest reference station, and, finally, the position of the barnacle line relative to the mean lower low water (MLLW) datum was cal— culated. For measurements made close to the 16 US. Coast and Geodetic Survey tide gages that were installed in the area imme- diately following the earthquake, we later made corrections to the ‘ actual, rather than the predicted, tides. During the period of field- work, it was found that few tides deviated by as much as 1.5 feet from predictions; most were with— in a few tenths of a foot of 'pre— dicted values. Differences in the height of the barnacle line in local areas of simi- lar tide range provided a powerful tool for determining relative changes in land level along the coast even where the absolute change was not known. Absolute uplift or subsidence relative to sea level at any given locality was taken as the difference in height be- tween the measured elevation of the barnacle line and the “normal” preearthquake upper growth limit for the barnacles and Fucus as in- dicated in figure 10. The “normal” preearthquake upper growth limit of barnacles and F ucus relative to MLLW was determined empirically at the 17 localities listed on figure 10 where the amount of vertical displace- ment was known from Coast and Geodetic Survey tide gage read- ings. Its position at these localities was taken as the measured height relative to MLLW corrected for the amount of tectonic uplift or subsidence indicated by the tide- gage readings. On figure 10 these heights are plotted against the mean tide range at the station. The least-square curve through these points represents the average pre- earthquake height of the barnacle line for mean tides ranging from 6.4 to 15.9 feet in the study area. The curve suggests that the pre- earthquake barnacle line was close to mean high water for the lowest mean tides and that it lowered pro- gressively relative to MHW with an increase in mean tide range. For the lower tides of (SA—10.0 feet which prevail along most of the Gulf of Alaska and in Prince Wil- liam Sound, it ranged from MHW to 0.6 foot below MHW level; for the higher mean tides of as much as 15.4 feet at Seldovia in lower Cook Inlet and Shelikof Strait, the barnacle line dropped to 1.7 feet below MHW. The derived curve for the ap- proximate height of the barnacle line is only a crude approximation to the actual position of the bar- nacle line. 111 addition to uncer- tainties in measurement of the po- sition of the postearthquake barnacle line relative to sea level, it incorporates inherent errors in the determination of land-level change at the temporary tide-gage stations as discussed previously (p. I 11). GEOLOGICAL SURVEY EFFECTS OF THE EARTHQUAKE ON INTERTIDAL ORGANISMS A (right).—Sharply zoned intertidal marine organisms along west shore of Port Bain- bridge. The barnacle line, uplifted 6.1 feet, is at the contact between the upper black band of encrusting lichens and the light-gray band of barnacles. Greenish material in the lower part of the rock face is marine algae; its upper limit marks the approximate position of the postearthquake barnacle line. Photo- graph taken at 31-foot tide stage, June 18, 1964. 0 (right).—Postearthquake yearling barna- cles (light gray) among preearthquake barnacles (yellow). Within 4 months after the earthquake the new crop of barnacles had base diameters of as much as 0.3 inch. Large divisions in upper scale are inches. Photograph taken August 2, 1964. 307—634 0—69—-—4 PROFESSIONAL PAPER 543-1 PLATE 3 B (left).—Barnacle line defined by upper growth limit of olive-brown rockweed and lower limit of dark-gray encrusting lichens. Shoreline shown (Malina Bay, Afognak Island) subsided about 3 feet during the earthquake. Photograph taken at 65-foot tide stage, July 20, 1964. GEOLOGICAL SURVEY PROFESSIONAL PAPER 543-1: EFFECTS OF THE EARTHQUAKE 0N SHORELINE FEATURES A.—Living (olive-green) and desiccated (dark-brown) Fucus along the shore of Glacier Island, Prince William Sound. The top of the band of desiccated algae was near the preearthquake barnacle line and the top of the band of living algae was near the postearthquake barnacle line. The 3.0-f00t difference between their eleva- tions was a measure of the tectonic uplift in this area. Photograph taken at 88-foot tide stage, June 13, 1964. B.—-Extensive area of brown terrestrial vegetation at Kiliuda Bay, Kodiak Island, killed by salt-water immersion after about 4 feet of tectonic subsidence and an unknown amount of surfici'al subsidence. Dikelike gray ridge of beach gravel was built up in adjustment to the new higher base level. The area behind this beach ridge may eventually become a shallow lagoon. Photograph taken July 17, 1964. PLATE 4 18- 16- / ,_ LU LU LL E m" u '2 a 14_ / g ' / -‘ Upper growth a: limit of In 014 _ barnacles g 12 6.0){/ .10 _J 5 5/39 In gl 2 10 - y.” “>‘ Z '15 O /1. 3g . / 16 '4 ,_ s- 1 I 0 (D I l: 17 I 5 I I I I 6 8 10 12 I4 MEAN TIDE RANGE, IN FEET TECTONIC-S Data Points 1. Womens Bay, Kodiak Island 2. Seward, Kenai Peninsula ( average of 2 measurements, range indicated by vertical bar) . Chenega Island, Prince William Sound . Lazy Bay, Kodiak Island . Port Chalmers, Montague Island, Prince William Sound 6. Sawmill Bay, Evans Island, Prince William Sound 7. Port Gravina, Prince William Sound 8. Whittier, Prince William Sound 9. Cordova, Prince William Sound (average of 28 measurements; range indicated by ver- tical bar) 10. Larsen Bay, Kodiak Island 11. Port O’Brien, Kodiak Island 12. Seldovia, Kenai Peninsula 13. Port Dick, Kenai Peninsula 14. Port Chatham, Kenai Peninsula 15. Two Arm Bay, Kenai Peninsula 16. Tonki Bay, Afognak Island 17. Ugak Bay, Kodiak Island 01%“: 10.—Height of upper growth limit of barnacles above mean lower low water for mean tide ranges of (SA—15.9 feet in south-central Alaska. The least-square curve for the upper growth limit of barnacles is based on measurements made at 17 stations (dots) where the vertical displacement was determined independently by the US Coast and Geodetic Survey (Small, 1966) from tidegage observations. Standard deviation is 0.627 foot. However, the validity of using the empirically determined barnacle line was generally confirmed by our observations in late 1964 and in 1965 that the upper growth limits of new postearthquake barnacles and Fucus were generally within about :1 foot of this line. The precision of the land-level changes determined by the bar- nacle-line method varies within wide limits because of the numer- ous variables involved in making the measurements and the assump- tions inherent in the presumed preearthquake position of the bar- nacle line. The measurements are generally within 1 foot of changes estimated by« local residents or found by means of other tech- niques. Under the least favorable combination of circumstances, such as along segments of the coast exposed to heavy surf or swells or in areas of high and erratic tides in Cook Inlet and Shelikof Strait, measurements may locally be in error by as much as 21/2 feet. In such areas, however, more reliance was placed on changes indicated by these techniques that do not re- quire use of tide level as a datum. DIFFERENCES IN THE EXTREME PRE- AND POSTEARTHQUAKE GROWTH LIM- ITS OF SESSILE INTERTIDAL ORGA- NISMS AND TERRESTRIAL VEGETATION Throughout the region affected by tectonic land-level movements, postearthquake changes in the upper growth limit of barnacles and Fucus or in the lower growth limit of terrestrial vegetation pro- vided direct indications of the direction and approximate amount of movement. Thus, along uplifted shores a band of dead barnacles, F new, and other sessile organisms developed within 2 months after the earthquake. The height of this band reflected the amount of up- lift (pl. 4A). By July of 1964 a new postearthquake line of young barnacles, and in some places Fucus, was well established on most shores. The height of this line above or below the preearth- quake line furnished a direct meas- ure of the amount of uplift or subsidence (pl. 30 ; fig. 11). Simi- larly, at many places, the amount I17 of subsidence could be clearly de- termined within 2 months after the earthquake from certain ephem- eral features such as the eleva- tion to which the highest spring tides inundated terrestrial vegeta- tion (fig. 12). By the 1965 field season, land plants had become sufficiently well established over much of the uplifted shore that the approximate amount of uplift could be determined from differ- ences in the pre- and postearth- quake lower growth limits (fig. 13). This method was particularly useful in areas of uplift such as Middleton Island where, for some unknown reason, barnacles and F was had not become established in the intertidal zone even a year after the earthquake. Land-level changes determined from differences in pre- and post- earthquake positions of the barn- acle line 0r 0f the lower limits of terrestrial vegetation provide rea— sonably precise values for the tee— tonic movements where the post- earthquake growth limits have had time to reach a position in equi- librium with the local tides and where the earthquake-related dis- placements have not caused signifi- cant changes in the tidal charac- teristics. Such measurements are thought to represent the actual vertical change at least to within 1.0 foot, and probably to within 0.5 foot in most places. DIFFERENCES IN THE HEIGHTS OF PRE- AND POSTEARTHQUAKE STORM BEACHES Along uplifted sandy shores on Montague Island in Prince Wil- liam Sound and along the linear stretch of coast east of Kayak Island, the amount of uplift could be approximated in 1965 from the relative positions of pre- and post- earthquake storm beaches (fig. 14). The accuracy of such meas- urements is difficult to evaluate, although where they could be 118 ALASKA EARTHQUAKE, MARCH 27, 1964 11.——-Conspicuous white band of postearthquake barnacles along the shore of Kizhuyak Bay, Afognak Island. The difierence in elevation between the upper growth limit of the yearling barnacles in the photograph and the preearthquake barnacles, which were at water level, indicates at least 3 feet of tectonic subsidence. Photograph taken July 20, 1964. 12.—Drowned brush and trees along shore of Harriman Fiord, Prince William Sound. The color change (arrow) between the dead brown foliage (light gray) below and the green foliage (darker gray) reflects the position of the postearthquake extreme high-tide line. The difference in elevation between the lower growth limit of terrestrial vegetation and the new extreme high-tide line provided a measure of tectonic subsidence, which was 7.2 feet at this locality. Photograph taken at 0.9- foot tide stage, June 10, 1964. TE CTONICS 13.—Wi1d flowers and grass growing among dead barnacles (white) on shore of Middleton Island uplifted about 11 feet. The differences in the lower growth limits of pre- and postearthquake terrestrial vegetation provided a direct indication of the approximate amount of uplift. Photograph taken July 26, 1965. 14.——Coast at Cape Suckling uplifted about 13 feet during the earthquake. The diflerence in elevation between the postearthquake storm beach (marked by band of light-colored driftwood) and the preearthquake storm beach, which was above the base of the sea clifi, provided a crude measure of the uplift. The smooth area between the upper limits of driftwood land the sea cliff is now a marine terrace, and the fomer island in the foreground is a stack on its surface. The flat surface on the stack is probably an older marine terrace. Photograph taken at about zero tide stage, July 24, 1965. 307—634 O~a70——4 119 compared with changes at nearby rocky shores, they appear to give results consistent, within about 3 feet, with those obtained from barnacle lines. The measurements between Kayak Island and Yaka- taga (pl. 1) are particularly uncer- tain because the shore there con- sists of active sand dunes that had partially concealed the old storm- beach line by the time measure- ments were made in 1965. This method was used only where no other means was available for measuring vertical displacement. ESTIMATES OR MEASUREMENTS BY LOCAL RESIDENTS Where possible, data on local land-level changes were obtained from local residents in interviews and on form questionnaires. The amount of these changes and the confidence limits expressed by 0b- servers are shown on plates 1 and 2. .Most .of the estimates were made by fishermen, mariners, log- gers, and other coastal residents who had long experience in ob- serving the levels of local tides relative to familiar shoreline fea- tures. Consequently, most of their estimates or measurements of the vertical displacements are prob- ably correct to within a foot or less. HEIGHT 0F TIDAL BENCH MARKS RELATIVE TO SEA LEVEL At a few localities in Prince William Sound, changes in land level were determined by leveling from the water surface to US. Coast and Geodetic Survey tidal bench marks of known preearth- quake elevation. The accuracy of these determinations, which de- pends mainly upon the precision of leveling and the degree to which the actual tides at the time of measurement correspond with predicted tides, is believed to be within 0.5 foot. 120 COMPARISON OF PRE- AND POSTEARTHQUAKE DEPTH SOUNDINGS Submarine control for the off- shore uplift indicated southwest of Montague Island is provided by comparisons of detailed bot— tom soundings taken by the US. Coast and Geodetic Survey in 1927 and after the earthquake in 1964 (Malloy, 1964, p. 1048—1049; 1965, p. 22-26). Because of the technical problems involved in carrying out such surveys, how- ever, the inferred submarine dis— placements could locally be in er- ror by 10 feet or more. COMPARISON or PRE- AND POSTEARTHQUAKE LEVELINGS The amount and distribution of the vertical tectonic movements inland from the coast were de- fined by the U.S. Coast and Geo- detic Survey’s prompt releveling of 722 miles of previously sur- veyed first-order level lines con- necting the cities of Seward, An— chorage, Valdez and Fairbanks (pl. 1; fig. 3). Details of the methods used in these surveys were presented by Small (1966) ; changes in elevation between suc— cessive levelings at bench marks along these routes have been tab— ulated by Wood (1966, p. 124— 130). The changes indicated by this method are probably ac- curate to within a few tenths of a foot. Small (1966, p. 2) has cau- tioned that observed divergences in the results of the original lev- eling dating from 1923 and the postearthquake leveling in 1964 and 1965 may not be due entirely to movement associated with the earthquake because there is evi- dence for minor gradual regional movement along some lines where repeated levelings were made prior to 1964. At a few bench marks along these routes, which were located on unconsolidated deposits, the indicated subsidence ALASM EARTHQUAKE, MARCH 27, 19164 was anomalously large. Such anomalous measurements obvi— ously represented changes caused by accidental displacement of the bench marks or by local phenom- ena such as frost heaving, surfi- cial settling, or thawing of per— mafrost rather than to tectonic movements; they are not shown on plate 1. DISTRIBUTION OF LAND-LEVEL CHANGES Figure 3 summarizes the known and inferred areal distribution of land-level changes. The deforma— tion extends for almost 600 miles along the Gulf of Alaska coast from the southwest tip of the K0- diak group of islands through the Prince William Sound region and eastward to the vicinity of Yaka- taka where it seems to die out. The deformed region consists essen- tially of (1) a broad zone of subsi- dence centered along the axis of the Kodiak-Kenai—‘Chugach Moun- tains, (2) a major zone of uplift that borders it on the seaward side and extends from the coast onto the sea floor, and (3) a zone of slight uplift that borders it on the landward side and extends north— ward into the Alaska and Aleutian Ranges. The distribution of subsi— dence and uplift in these three zones is described below. THE ZONE OF SUBSIDENCE The zone of tectonic subsidence includes almost all of the Kodiak group of islands, most of the Kenai Peninsula, the northern and west- ern parts of Prince William Sound, and probably the western segment of the Ohugach Moun- tains (pls. 1, 2; fig. 3). Areas of subsidence in most rocky embayed coastal areas are clearly defined by the various criteria outlined in the preceding section or by qualitative indicators of shoreline submerg- ence. In sheltered embayments the changes may be noticeable where the subsidence is as little as 1 foot. Such effects were most pronounced in areas of lowest mean annual tide range (fig. 8), inasmuch as both the frequency and duration of shoreline immersion for a given amount of subsidence vary in— versely with the tidal range. Along coasts with large tidal ranges and nonrocky shores, such as the part of Cook Inlet north of Homer, the subsidence is known only from tide gage readings near Kenai and at Anchorage, from a few observations by local residents, and from leveling along the coast near the head of the inlet. In this area, and along Shelikof Strait, the northwestern limit of the zone of subsidence is poorly defined. It seems to be close to the west side of the inlet from Redoubt Bay south— westward to Kamishak Bay and probably extends inland between the south side of that bay and the general area of Katmai Bay. Control on the distribution and absolute amount of subsidence in- land from the coast is provided by the Coast and Geodetic Survey’s releveling of the first-order net shown on figure 3. Subsidence was indicated on all these lines south of the approximate southern mar- gin of the Alaska Range except in the immediate vicinity of Valdez where a few bench marks were up- lifted less than 0.‘2 foot (pl. 2). The leveling clearly demonstrates that the subsidence extends as a M! ad warp without abrupt changes of level across the Kenai Mountains northward from Seward and across the Chugach Mountains north of Valdez (pl. 1) . Within the Ohugach Mountains, subsidence of about half a foot extends eastward at least to ‘Chitina. The northern limits of the zone are approximate— ly defined 'by the leveling along The Alaska Railroad and Richard- son Highway. Because of the small measured land-level changes on those lines, errors of as little as 0.5 foot could cause shifts of as much as 30 miles in the position of the northern boundary of the zone. MAJOR ZONE OF UPLIFT The main zone of uplift on land, as determined from shoreline changes, includes (1) a narrow fringe of points, capes, and small islands along the seaward side of the Kodiak group of islands, (2) all but the extreme northwestern and northern parts of the Prince William Sound region, and (3) the coastal belt extending about 120 miles east of the sound. Direct indications of uplift of parts of the contiguous Continental Shelf are afforded by emergence of all the offshore islands and reefs, in- cluding Middleton Island near the edge of the Continental Shelf (pl. 1, 2; fig. 1). The extreme southwestern limit of the zone is believed to lie be- tween Sitkinak Island, which was uplifted about 11/2 feet according to a local resident (Mr. Hall Nel- son), and Chirikof Island, where there apparently was no change in level (Neal Hoisington, written commun., 1965). Its eastern limit is probably at, or just west of, Yakataga. The trend of the isobase con- tours in the northeastern part of the zone of uplift (fig. 3) and the distribution of aftershocks (fig. 2) seem to justify the inference that uplift also occurred over much of the submarine part of the con- tinental margin in a broad zone extending southwestward at least to the latitude of southern Kodiak Island. Seaward projection of the trend of the isobase contours from the area between Yaikataga and Middleton Island also suggests that uplift occurred over much of the continental slope and could have extended to the toe of the TE CTONICS continental slope, as is shown in figure 3 and on plate 1. Independent evidence for up- lift over a large segment of the Continental Shelf and slope comes from the seismic sea waves (tsun- ami) generated by the submarine movements. Because seismic sea waves are gravity waves set up in the ocean mainly by vertical dis- turbances of the sea bottom, the sense of displacement in the gen— erative area can ‘be determined under favorable conditions from the initial water motion at suit- ably situated tide stations. Tide- gage records outside the immedi- ate area afl'ected by the earthquake show an initial rise that indicates a positive wave resulting from up- ward motion of the sea bottom (Van Dorn, 1964, p. 166). The initial direction of water move- ment along the coast of the Gulf of Alaska within the area affected by the earthquake is less clear, how- ever, because there were no op- erative tide gages, and in many 10- calities the water movements reported by eyewitnesses were complicated by (1) changes of land level along the coast, (2) lo- cal waves generated mainly by submarine landslides, and (3) seiches, or other water disturb- ances related to horizontal tecton- ic displacement. The shape of the source area within which the train of seismic sea waves was generated can be approximated from an envelope of imaginary wave fronts project- ed back toward the wave source from observation stations along the shore at which arrival times are known. Distances traveled by the waves can be determined if both the wave velocity along the propagation path and the travel time are known. Because of their long wavelengths, seismic sea waves move as shallow water waves even in the deepest ocean, I21 and their approximate velocity is given by La Grange’s equation (in Lamb, 1932, p. 257) : V=\/9—h were 9 is the gravitational con- stant, and h is the water depth along the travel path (as deter- mined from nautical charts). Travel time is taken as the elapsed time between the main shock and the arrival of the first wave crest at shore stations. The distribution of tide gage stations outside the seismic sea-wave generative area precludes precise delineation of the source by this method. How- ever, its general position as de- rived by Van Dorn (1964, fig. 8), Pararas-Carayannis (1967), and MG. Spaeth (oral commun., Sept. 1964) is consistent with uplift in the broad zone that lies roughly between the Aleutian Trench axis and the coast and extends from the general area offshore from Yaka- taga southwestward to about the latitude of Kodiak. The position of the axis of the wave source shown on figure 15 was inferred from the arrival times of the initial wave crest along the adjacent coast, the gen- eral distribution of wave damage, and the reported movement direc- tions of the initial wave. Approxi- mate travel times to shore stations, sense of initial water motion, and the data sources are given in table 2. These data suggest that the wave crest was generated along one or more line sources within an elon- gate belt that extends about 350 miles from the vicinity of Monta- gue Island in Prince William Sound to the area offshore from Sitkalidak Island in the. Kodiak group. This inference is support- ed by the fact that at the north- east end of this axis on Montague Island warping and faulting have resulted in uplift of 38 feet in a. I22 156‘ b?“ 58"L l 4 4 ° w._.~.._<.‘__._.___———r \ \ i ALASKA EARTHQUAKE, MARCH 27, 1964 154° 15?." 150‘1 148° _ 146" _ l/ l , ~ I \ i I I I I’ ‘ ' . I xv”. I, i , /// I ' . / a .Anchomge Epicenter Valdez ‘ , , I a _ .9 ’4'" »__...,._..-.... 9 I . — \ X) / ’ , a / % PIEINCEO o ‘ o’WILLIAM . Cordeva ’ ’19 oo “cg I K I x3“ / ' G V? / _ $0 . ’ I SOUND % I I Montague Island Middleton I [I ‘ l ,I . e7 , Homer 3:? I Seldovia / . Island a ’l’ I [I I, ’ I & I’ ,’ 9 / I « ' [I / ‘\ ll ‘5 8‘ ’I A v I Afognak ,’ V33 ' Island , \Y I’ Y' I, ’ Q _ I0 O / I, ‘ I I ,/ G I , ’w , , Q I [I II ‘- ’ I +0 ‘ I I l l 4, l \ - , z ‘ "a Kdguyak ’,/ ‘\ I /(No wave damage) METERS/FEET FEET/METERS \ <31) z AI / 16 A 10 ,Trinity islands ,” 30 fl 30 l I,” 8 I 8 ‘\ I” 6 20 I 20 6 56° — ‘ \’ ’-/ 4 I \\ Aleutian Trench ””“s‘t‘ -—” 1° / \\\ 10 “~‘ ---- ‘flfl 2 \ 2 .— / \ O O “ \ o 50 lOOMILES —2—L "~§_/ _|>—2 _ 10 — 10 1 Kl o E ERS Profile along line A-A’ showing inferred initial wave form 00 L M T caused by tectonic displacements of the sea floor l l l o 50 L__l__l | EXPLANATION ___+_.__ Axis of subsidence up Wave travel direction inferred from shoreline damage or eyewitness accounts ______ ___. Calculated maximum distance trav- elled by initial wave __--__0 ______ _ Zero isobase contour o Epicenter of major aftershock (M_>_6.0) ® Dashed where approximate Station listed in table 2 Axis of uplift Dashed where inferred 15.—Submarine extension of the zone of maximum uplift and faulting on Montague Island as inferred from movement directions and calculated travel distances of seismic sea waves generated by the tectonic displacements. TECTONICS TABLE 2.—Travel times of seismic sea-wave crest to near-source observation stations [Locations shown on figure 15. Travel time from start of the earthquake (5:36 p.m., A.s.t.)] Station Travel Reported _ time sense of (minutes) first Data source No. Name motion (fig. 15) 1 Kaguyak _____ 38 j: 5 ? Larry Matfay (radioed message received at Old Harbor). 2 Old Harbor- _ _ 48 Up ______ Larry Matfay. 3 Cape Chiniak- 38 Up ______ Fleet Weather Central, Kodiak Naval Station. 4 Kalsin Bay- _ _ 70 ? US. Geological Survey stream gage. 5 Kodiak Naval 63 Up ______ Fleet Weather Central, Station. Kokiak Naval Station. 6 Kodiak _______ 45 :l: 3 Down_ _ _ _ Jerry Tilley. 7 Rocky Bay- _ _ 1 30 Down_ _ _ _ Guy Branson. 8 Seward _______ 30 :I: 5 Up ______ Scottie McRae. 9 Whidbey Bay- 19% :96 Up ______ Bill Sweeney. 10 Puget Bay- _ _ _ 20 :5; 2 Up ______ Sam Hatfield. 11 Middleton 1 20 Down_ _ _ _ Dwight Meeks Island 12 Cape 60 :1: 1 Down_- _ _ Charlie Bilderback. Yakataga. 13 Saltery Cove_ _ 1 30 ? Ron Hurst. 1 Approximate. belt about 6 miles wide (Plafker, 1968), and comparable displace- ments are known to have occurred on the adjacent sea floor (Malloy, 1964). Similarities in maximum wave-runup heights along physio- graphically comparable segments of coast, both on the Kenai Peninsula opposite Montague Is- land and on the ocean coast of Kodiak Island, suggest that the vertical sea-floor displacements that generated the waves in these two areas could be of the same order of magnitude. If so, the ini- tial wave form resulting from sea- floor displacement had the ap- proximate shape shown by profile A—A’, figure 15. Other factors, however, such as rate of uplift, ini- tial slope at the wave source, and energy loss along the propagation path preclude direct correlation of runup heights with displacement at the source. PROBABLE ZONE 0F SLIGHT UPLIFT Minor uplift, probably asso- ciated with the earthquake, has been detected in three areas ad- jacent to, and inland from, the zone of subsidence (fig. 3). The distribution of uplift in these three areas strongly suggests the possi- bility that they may be part of a continuous zone, roughly 100 miles wide, that parallels the major zones of subsidence and uplift. Slight uplift in the Alaska Range is indicated by U.S. Coast and Geodetic Survey releveling along both the Richardson High- way and The Alaska Railroad (Small, 1966, fig. 9). Comparison of the 1964 leveling with a line run in 1952 shows general uplift along the Richardson Highway in a zone from about 25 miles north of Glennallen to within 50 miles of Fairbanks (fig. 3). Maximum up- lift recorded on this line was 0.89 foot near the center of the zone, with irregular but generally pro- gressive decreases toward the north and south. Land—level changes indicated by comparison of 1922 and 1964 levelings along The Alaska Railroad are less con- I23 sistent but they are predominantly positive and are as much as 0.36 foot in a zone where the line crosses the Alaska Range. The relatively large amount of change (0.89 ft) in only 12 years between the successive surveys on the Rich- ardson Highway strongly suggests that at least the major part of the measured changes were probalbly associated with the earthquake. Furthermore, the fairly systematic rise and fall in the amounts of up- lift across the Alaska Range sug— gest that uplift is not due to sur- veying errors or errors inherent in tying the level lines to tidal datum planes. Thus, it is tentatively con- cluded that the uplift along these two lines represents earthquake- related uplift over a broad zone centered in the general area of the Alaska Range. Residents along the Iliamna, Chinitna, and Tuxedni Bays on the northwest shore of Cook Inlet (pl. 1; fig. 3) report a decrease in the height of tides after the earth— quake that suggests shoreline up- lift of 1—2 feet. There is little doubt about the validity of these estimates, particularly in Iliamna and Tuxedni Bays, where refer- ence marks existed whose pre- earthquake relationship to tide levels were precisely known. Be- tween 1963 and 1965 a slight up— lift, presumably related to the earthquake, occurred also on the west side of Augustine Island (R. L. Detterman, oral commun, 1965) where beach berms have been low- ered ‘1/3—11/2 feet. The observed changes along the northwest side of Cook Inlet strongly suggest slight tectonic uplift of the shore— line during the earthquake. They cannot be_ explained by changes in the tidal characteristics due to re- gional subsidence of the entrance to Cook Inlet, inasmuch as deepen- ing of the entrance would be ex- pected to increase, rather than 124 decrease, the height of the tides by facilitating diurnal movement of the tidal prism. GEOMETRY OF THE DEFORMATION The pattern of absolute vertical deformation associated with the earthquake is indicated by the isobase contours and profiles of plates 1 and 2 and is shown at a smaller scale in figure 3. The iso— base contours may be pictured as the amounts of vertical displace— ment of an imaginary surface that was horizontal before the earth- quake. The resulting map, there- fore, is a special form of structure- contour map showing the configu- ration of the deformed surface. The maps and profiles indicate that the deformation occurred in three broad elongate warps, each of which is from 100 to 130 miles wide and has axes that roughly parallel the trend of the continental mar- gin. Subsidence occurred in the middle warp and uplift in the ad- jacent warps on the seaward and landward sides. The zero isobases between the zone of subsidence and the adjacent zones of uplift are axes of tilt across which the sense of vertical displacement relative to the preearthquake position changes gradually. No abrupt changes of level have been found between the adjacent zones that would indicate vertical fault dis- placement between them. ZONE OF SUBSIDENCE The zone of subsidence is a syn- clinal downwarp whose axis is situated roughly along the crest of the coastal mountain ranges. The axis of subsidence plunges gently northeastward from the Kodiak Mountains and southwest- ward from the Chugach Moun- tains to a low of 71/2 feet on the south coast of the Kenai Penin— sula. In cross section the down- ALASKA EARTHQUAKE, MARCH 27 , 19'64 warp is strongly asymmetrical with an average tilt in the middle part of the deformed region of about 1 foot per 14 miles from the landward side towards the axis and a much steeper average tilt of 1 foot per 2—31/2 miles on the sea- ward side. The prevailing simple synclinal form of the downwarp is broken only by the slight warp- ing of the tilted surface near the axis of subsidence immediately north of Seward in an area where both triangulation and geologic data suggest the possibility of minor earthquake—related move- ment on a conspicuous north- south-trending lineament. Apparent reduction in crustal volume within the zone of subsi- dence, as calculated by summing the average volumes included be— tween successive isobase contours on plate 1, is about 29 cubic miles. Total area of the zone affected by subsidence is about 48,000 square miles, and the average amount of subsidence within it is roughly 21/2 feet. MAJOR ZONE 0F UPLIFT ‘ The major Zone of uplift along the continental margin is a broad upwarp with a maximum ampli- tude of 15 feet, upon which is superimposed a narrow belt, less than 10 miles wide, in which there has been strong. uplift asso- ciated with displacement on re- verse faults. The axis of the uplift- ed zone trends southwestward from Montague Island, presum- ably to the area offshore from Sit- kalidak Island in the Kodiak group. Maximum uplift along this axis on Montague Island is 38 feet and it may be as much as 50 feet on the sea floor. The position of the axis of up- lift to the northeast of Montague Island is uncertain; it may con- tinue offshore from Hinchinbrook Island and the Copper River Delta to intersect the coast at Cape Suck- ling where 13 feet of uplift was measured. Because of the scarcity of data points in the Cape Suck- jling area, the shape of the de- formed surface there cannot be closely defined and the possibility cannot be ruled out that the large amount of uplift there may reflect local warping or faulting. The part of the upwarp avail- able for observation in Prince William Sound has an irregular shape that suggests combined tilt- ing and warping. As indicated by profiles A—A’ and B—B’ on plate 1, the landward slope outside the narrow belt of extreme uplift av- erages 1 foot per 2.1 miles north- west of Montague Island and only 1 foot per 7.4 miles north of Hin- chinbrook Island. Local tilts as high as 1 foot in 185 feet occur within the belt of extreme uplift and surface faulting in southern Montague Island. The isobases in the central and southeastern part of Prince William Sound re— ‘flect a broad undulating platform 4—8 feet above its preearthquake ‘ position. In at least two areas of western Prince William Sound, local flattening or even reversals of slope are indicated. Data on the configuration of the upwarped surface in the Kodiak Island area, although less conclu- sive, suggest northwestward tilt— ing that is as steep as 1 foot per mile at Narrow Cape. Little is known about the shape of that part of the upwarped zone that is seaward from the axis of uplift because only a few points are available for observation. The slope between the most southeast- erly capes of Montague Island and Middleton Island 50 miles to the southeast averages 1 foot per 11 miles, but the shape of the surface in the water—covered area between these points is conjectural. Nor is it known whether the uplift sea- ward from Middleton Island dies out gradually toward the toe of the continental slope, as inferred on the profiles on plate 1, or whether it terminates abruptly in one or more faults or flexures on the slope. The apparent increase in crustal volume within the major zone of uplift is much less certain than that involved in subsidence, be— cause the distribution of uplift in extensive submarine areas must be inferred from the trend of isobase contours in the northeastern part of the zone and a few ofi‘shore con- trol points. If the deformation has the general form shown by the profiles on plate 1, the volume in- crease would be approximately 89 cubic miles, or roughly three times the decrease in the zone of subsi- dence. Total area of the uplifted zone is inferred to be roughly 60,000 square miles. Average amount of uplift is about 6 feet, except in the narrow axial belt of uplift and faulting extending southwestward from Montague Island, where it is probably 30 feet or more. PROBABLE ZONE 0!" SLIGHT UPLIFT Where the broad slight upwarp landward from the zone of subsi- dence is crossed by level lines, its axis seems to be centered along the crest of the Alaska Range and its maximum indicated uplift is 0.89 foot on the Richardson Highway line and 0.35 foot on The Alaska Railroad line. The upwarp crossed by these two lines of leveling may be part of a continuous zone that extends into the Aleutian Range of the Alaska Peninsula where up- lift of as much as 1.5 feet has been reported at several places in the Kamishak Bay-Tuxedni Bay area. A rough estimate of the apparent increase in crustal volume in the zone of slight uplift, based on an average uplift of 0.3 foot and an area of 24,000 square miles is about 1.0 cubic miles. TE CTONICS EARTHQUAKE FAULTS Faults on land associated with the 1964 earthquake were found only at two localities on south— western Montague Island in Prince William Sound and on the subsea continuation of one of these faults southwest of the island. Comparable faults entirely on the sea floor may have gone unde- tected. As far as could be deter- mined no definite movement occurred along any other faults on land, although faulting at depth is suspected in some areas of uncon- solidated surficial deposits charac— terized by linear zones of land- slides or surficial cracks. MONTAGUE ISLAND FAULTS The location of, and displace— ment across, the earthquake faults on and near Montague Island are shown on plate 1. Their surface characteristics and tectonic signi- cance are briefly summarized in the following paragraphs. In a separate volume of the Geological Survey’s series of papers on the Alaska earthquake, they are de- scribed in more detail (Plafker, 1967b). The longer of the two faults, the Patton Bay fault, is represented by a complex system, 22 miles long, of an echelon reverse faults and as- sociated flexures with an average N. 37° E. strike. Surface dip of the fault is northwest at about 85° near its southern end and 50°—75° elsewhere along the scarp. Dis- placement on the fault is almost entirely dip slip—the northwest side upthrown relative to the southeast side. The maximum measured vertical component of slip is 20—23 feet, and maximum indicated dip slip is about 26 feet. A left-lateral displacement com- ponent of less than 2 feet near the southern end of the fault is prob- ably a local phenomenon related to I25 a change in strike of the fault that causes it to trend at an oblique angle to the N. 53° W. principal horizontal stress direction. The Patton Bay fault system was traced by the U.S. Coast and Geodetic Survey (Malloy, 1964) for at least 17 miles on the sea floor southwest of Montague Is- land. Indirect evidence, from the distribution of large aftershocks associated with the earthquake and from the distribution of sub- marine scarps, suggests that the faulting on and near Montague Island occurred at the northeastern end of a reactivated submarine fault system. This system approxi- mately coincides with the axis of uplift inferred from seismic sea waves between the southeast coast of Kodiak Island and Montague Island (fig. 15). The fault ap— parently dies out on its north- western end, although the possi- bility cannot be ruled out that it is offset en echelon towards the southeast (in a righthanded sense) and continues northeastward off- shore from Montague Island at least as far as Hinchinbrook Island. The shorter of the two faults, the Hanning Bay fault, is a vir- tually continuous reverse fault with an average strike of N. 47° E. and a total length of about 4 miles. Dip of the fault is 52°—75° NW. at the surface. Displacement is dip slip except for a left-lateral strike- slip except for a left-lateral strike- of a foot near the southern limit of the exposure. The maximum measured vertical component of slip is 161/3 feet near the middle of the fault, the indicated dip slip at that locality being about 20 feet. The two reverse faults on Mon— tague Island and the postulated submarine extension of the Patton Bay fault constitute a zone within which crustal attenuation and maximum known uplift occurred I26 during the earthquake. Neverthe— less, there are no significant litho- logic differences in the rock se- quences across them to sug- gest that these faults form major tectonic boundaries. Furthermore, their spatial distribution relative to the regional zone of tectonic. up- lift associated with the earthquake, to the earthquake focal region, and to the epicenter of the main shock suggests that they are probably subsidiary features, rather than the primary faults along which the earthquake originated. OTHER POSSIBLE EARTHQUAKE FAULTS ON LAND As far as could be determined, there are no other surface faults on land along which movement occurred during the earthquake. A careful search for renewed movement on known preexisting faults did not reveal any detect— able surface displacements. Nor were any anomalous abrupt changes found in amounts of ver- tical movement along the coast or along level lines inland from the coast that would suggest signifi— cant displacement on concealed faults. All reports of suspected faulting that were checked in the field proved to be landslides or surficial cracks in unconsolidated deposits. It is reasonably certain that if additional faulting did in— deed occur, its surface expression is far more subtle than that on Montague Island. Some of the linear belts of con- centrated surficial cracking and landsliding may reflect displace- inents on concealed faults. Foster and Karlstrom (1967, p. F24) sug- gested that movement on a con- cealed fault may have produced a northeast-trending linear belt of conspicuous surface fissures on the Kenai Lowland in the western part of the Kenai Peninsula. However, no evidence has been found for ver- ALASKA EARTHQUAKE, MARCH 27, 19164 tical displacements where the belt crosses the US. Coast and Geo- detic Survey level line south of Anchorage, and there is a notable absence of aftershock activity along the postulated fault. A second possible line of fault movement lies along the broad north-south—trending topographic depression, referred to here as the “Kenai lineament,” that extends northward from Resurrection Bay through the valley containing the eastern arm of Kenai Lake (pl. 1). Faulting is suggested (1) by local concentrations of fissures seemingly unrelated to seismic shaking along The Alaska Rail- road (D. S. McCulloch, oral com- mun., October 1967), (2) by re— ported angular changes between points on either side of the linea- ment as indicated by comparison of ‘pre- and postearthquake trian- gulation surveys, and (3) by a distinct change in trend of isobase contours across the lineament (pl. 1). The geodetic data have been interpreted as suggesting left-lat- eral displacement of as much as 5 feet between stations about 4 miles apart on either side of the linea— ment (Wood, 1966, p. 122). These data, if correct, could indicate either slight movement on a north- south—trending concealed fault or crustal warping localized along the lineament. HORIZONTAL DISPLACEMENTS Although the vertical displace- ments that occurred during the earthquake are unusually large, they appear to be secondary to the horizontal displacements indicated by retriangulation over much of the deformed region. During 1964— 65, the US. Coast and Geodetic Survey carried out revisional tri- angulations in the area shown in figure 16. The resurvey includes an area of about 25,000 square miles bounded on the west by the Sew— ard-Anchorage highway, on the north by the Glenn Highway, and on the east by the Richardson Highway and the east coast of Prince William Sound. To the south, the resurvey extends to sta- tions on the Gulf of Alaska coast and on Middleton Island 50 miles offshore from the coast. A tellu- rometer traverse was also run around the south coast of the Ke- nai Peninsula from Seward to Ho— mer and from Homer to Moose Pass (at Kenai Lake) via Kenai. Because the precision of station 10- cations obtained by the tellurome- ter traverse is probably too low to yield meaningful data on earth— quake—related horizontal displace- ments, the stations are not shown in figure 16 and the indicated shifts of these stations are not considered here. METHODS OF MEASUREMENT Parkin (1966, p. 2—5) has de- scribed the procedures used in ad— justing the pre- and postearth- quake surveys. The preearthquake net consisted of: (1) a primary are extending along the highway route from Anchorage northeastward to Valdez via Glennallen, surveyed in 1941 and 1944, (2) a second-order arc across the north shore of Prince William Sound from Valdez to Perry Island, surveyed in 1947—48, (3) a third-order are surveyed from Perry Island to Anchorage between 1910 and 1914, (4) third- order triangulation between 1900 and 1961 for chart control across Prince William Sound and extend- ing south to Middleton Island and westward along the southern Ke- nai Peninsula to Seward, and (5) a double are from Seward north to connections at Turnagain Arm, surveyed by the U.‘S. Army Corps of Engineers in 1941—42. All these observations were combined into a single composite network and a TEC’I‘ONICS I 27 151° 149° 148° 147° 146° 145° 144° 9 I l 1 C5] 1 1...“..- lb 9 J! EEO/Station Klawasi ‘K OVOH1IVd VXSV'IV 3H.L_ a O o , ~ ‘~ O “\‘ \ Glennalien "\ ‘~ Em mm "_r I I 5‘ ., "‘ . o' A ‘0 ‘5 2 Whitney" “ 10I ‘13 I" I I .v“ ~... .- 12/ [1:3] 60° 100 MILES l O 50 ICIVO KlLOMETERS 590 l l | l EXPLANATION .- 30------ ................. Triangulation station Isothismic contours Number refers to table 4’ Showing approximate southeastward component _.8 a . . . . Direction of displacement relative tosw- gcaizgzzaizzigizl”sewage-(“3022523252 tions Fishhook and Klawasx (adjust— inferred ment 2, table 3) I Number is approximate change in feet . B . _________ - Lme of profile shown 1n figure 18 Direction of displacement relative to Sta- 6 tion Fishhook (adjustment 1, table 3) 0_—_ _ _'—o 10 Zero isobase between major zones of tee- Relative northwest-southeast component tonic uplift and subsidence of shortening indicated by resurvey of 0 60 FEET isolated segment of triangulation net #44 Number is measured change, in feet Vector Scale (Displacement, between 0—5fieef shown as 5 feet on map) 16.—Map showing horizontal tectonic displacements in the Prince William Sound region and nearby areas. Horizontal displace- ments based on triangulation surveys by US Coast and Geodetic Survey (Parkin, 1966, table 1). 3017—634 0—70—45 I28 free adjustment (an adjustment with no external constraints) was made in which one position—Sta- tion Fishhook—was held fixed. In- ternal scale and orientation for the net were furnished from 5 Laplace azimuths, 15 short taped base lines, and 1 tellurometer length, which were included in the adjustment as observation equations. The post- earthquake triangulation survey, . which was all first-order work, was adjusted in the same way as the earlier work. Probable errors in the geo— graphic positions of stations in southern Prince William Sound relative to the fixed station, as conservatively estimated by Par- kin from the residuals, are 15—20 feet for the preearthquake survey and 6-8 feet for the postearth- quake survey. These probable errors decrease progressively for stations closer to the fixed station. The horizontal shift of re- covered stations relative to Station Fishhook between the pre- and postearthquake surveys, as com- puted by Parkin, are listed in table 3 as adjustment 1 and are shown graphically as displacement vec- tors (dashed) in figure 16. Be- cause the postearthquake net was not carried northward to an area of stability, changes shown are relative rather than absolute. However, small angular shifts in the northern part of the net, as compared with those farther south, suggest that the northern part of the resurveyed net prob- ably approaches an area that was not strongly affected by horizontal distortion during the earthquake. Anomalous aspects of the adjust- ment are (1) a gradual increase in displacement along the Glenn Highway arc east of Station Fish- hook to almost 13 feet at Station Klawasi, and (2) an apparent 32- foot shift of Middleton Island southwestward in a direction al- ALASKA EARTHQUAKE, MARCH 27, 19'6-4 TABLE 3.—Pre- and postearthquake dt'fi'erences in triangulation station plane coordinates [Data after US. Coast and Geodetic Survey (Perkin; 1966, and B. K. Meade, written commun., June 24, 1966). Adjustment 1: station 139 (Fishhook) held fixed; orientation and scale from azimuths and baselines of preearth- quake net (Parkin, 1966, table 1). Adjustment 2: stations 139 (Fishhook) and 55 (Klawasi) held fixed for addi- tional orientation and scale of preearthquake net. Station locations are shown in figure 16 except for stations 180, 184, 186, 188, and 218. Azimuth: north=0°.] Adjustment 1 Adjustment 2 Position shifts Resultant vector Position shifts Resultant vector Station (feet) (feet) AX (east AY AX (east AY (+)-west (north Length Azimuth (+)-West (north Length Azimuth (—)) (—)-south (feet) (—)) (—)-south (feet) (+)) (+)) —13. 73 23. 23 125° +15. 51 —4. 65 16. 18 117° ~11. 50 24. 52 120° +18. 60 ~2. 32 18. 72 97° ~15. 39 23. 17 130° +12. 37 ~5. 23 13.43 113° ~14. 32 18. 16 140° +4. 17 ~4. 73 6. 30 139° ~11. 84 14.55 145° +1.96 ~0. 54 2.03 105° ~7. 82 12. 74 130° 0 0 0 __________ ~10. 47 14. 53 135° +2. 23 ~3. 29 3. 97 146° -6. 82 12. 72 120° +2.25 ~l. 23 2. 56 119° —9. 03 13. 51 130° +3. 34 —3. 60 4. 91 137° -7. 73 11. 47 130° +2. 99 —4. 16 5. 12 144° ~10. 23 12. 99 140° +3. 14 ~6. 13 7. 05 154° ~8. 88 10. 32 150° +1. 57 -6. 20 6. 40 166° ~12. 48 13.31 160° +1.62 ~9. 30 9. 44 170° —7. 09 8.06 150° +1.10 ~5. 41 5. 52 169° —9. 23 10. 01 155° +1. 54 ~7. 29 7. 45 168° —3. 91 4. 17 160° +0. 17 —3. 12 3. 12 177° ~6. 61 6. 77 170° +0.23 —5. 40 5. 40 178° -2. 37 2. 51 160° +0.81 —1. 81 1. 97 157° —4. 74 4. 99 160° +1.42 -3. 75 4. 01 159° 0 0 __________ 0 0 0 __________ —4. 71 4. 75 175° +1.93 —3. 76 4. 22 153° ~3. 68 3. 70 175° +1.82 —2 90 3. 42 148° —3. 41 3. 45 190° +1.35 -2. 70 3.02 153° +3. 15 3. 65 330° +0.37 ~2. 61 2. 64 172° —2. 2.10 195° +1.88 -1. 53 2. 42 129° +2.06 3. 21 310° +0.42 +1.68 1. 73 76° ~4. 38 4. 38 180° +2. 70 -3. 27 4. 24 140° —5. 56 5. 59 175° +3. 70 ~4. 18 5. 25 143° —2.43 2.88 210° +2.19 -l.95 2.93 132° ~58. 20 58. 42 175° +15. 34 ~49. 67 51. 98 163° —4. 83 5. 57 150° +6. 99 -2. 80 7. 55 112° —5. 50 6.41 150° +7. 79 —3. 40 8. 50 114° -10. 69 ll. 09 165° +6. 80 ~7. 51 10. 11 138° ~15. 58 15. 92 170° +6. 66 --12. 07 15. 81 140° ~14. 22 14. 76 165° +7. 58 ~10. 51 12. 96 144° ~21. 89 23. 93 155° +12. 96 ~16. 78 21. 20 142° ~24. 23 26. 50 155° +14. 12 —19. 00 23. 67 143° -14. 85 15.08 170° +8. 26 -10. 89 13.67 143° ~18. 23 18. 42 170° +8. 46 -13. 89 16. 26 149° ~17. 25 17. 28 185° +5. 88 ~12. 91 14.19 156° ~21. 59 21.67 175° +8. 35 ~16. 88 18. 83 154° ~41. 20 41. 54 175° +12. 88 ~34. 86 37. 16 160° -46. 34 47. 18 170° +16. 26 -39. 65 42. 85 158° ~11. 98 12. 66 160° +8. 41 —8. 89 12.24 137° ~22. 78 31. 76 135° +21. 22 ~13. 76 25. 29 123° ~28. 40 34. 43 145° +18. 18 -20. 46 27. 37 138° ~37. 17 41.30 155° +18. 70 ~29. 84 35. 22 148° ~45. 85 49. 33 160° +19. 76 ~38. 18 42.99 153° ~44. 77 47. 78 160° +18. 67 -37. 72 42. 09 154° ~22. 61 31. 24 225° ~17. ~6. 32 18. 16 250° —28. 77 28. 94 185° +2.34 ~17. 05 17.31 172° ~39. 50 43. 67 155° +20. 81 ~28. 26 35. 09 144° —41. 30 46. 26 155° +19. 35 ~28. 32 34. 29 146° ~23. 87 35.25 135° +2518 ~13. 30 29. 45 118° ~45. 30 50. 46 155° +23. 51 ~36. 48 43. 40 147° —48. 63 54. 05 155° +25. 25 ~39. 43 47. 82 147° ’67. 40 69. 92 165° +23. 88 ~57. 59 62. 34 157° —66. 56 70. 19 160° +28. 91 ~57. 49 64. 34 153° —66. 35 67. 30 170° +17. 12 ~55. 23 57. 82 163° most normal to that of stations The assumption that the base llne along the coast. The first of these anomalies is eliminated, and the second con- siderably reduced, by an alterna- tive preferred adjustment of the data in which two stations, Fish- hook and Klawasi spaced 140 miles apart, are held fixed to provide additional orientation and scale of the postearthquake net (fig. 17). remained relatively stable in length and azimuth is justified on the basis of its position in the seismically inactive part of the net where there was only slight verti— cal displacement and by its orien- tation roughly parallel to the trend of isobase contours and normal to the trend of the horizontal shifts. The revised adjustment involves a 1 40 (““65 D C Station FishhookA A /B\% station KIawaSi TECTONICS 12.74 feet 0 ./ 17.—Schematic diagram illustrating the method of deriving triangulation adjustment 2 from adjustment 1. According'to adjustment 1, the postearthquake position of line AB is given by .40; point B shifted 12.74 feet S. 50° E. to point 0. For adjustment 2, which assumes no change in distance or azimuth between A and B, the postearth- quake net was rotated counterclockwise 00000124277 radians (angle BAG) and reduced in scale by the amount D0, or a factor of 09999881256 (the ratio of AB/AO’). counterclockwise rotation of the earthquake net of 0.000012427 7 ra- dians and a decrease in scale by a factor of 0.9999881256——changes probably well within the limits of error of these surveys. The result— ing horizontal shifts, which ap- pear to be more consistent with the vertical displacements, are given in table 3 as adjustment 2 and are plotted vectorially in figure 16 as solid lines. Unless otherwise speci- fied, horizontal displacements re- ferred to in the following sections are those of adjustment 2. AMOUNT AND DISTRIBUTION OF THE DISPLACEMENTS Absolute magnitudes and pre- cise directions of the horizontal changes cannot be determined be- cause the preearthquake triangu- lation net consisted mainly of third—order surveys and because the postquake survey, all of which was precise first-order work, was not carried northward to an area unaffected by the earthquake. Nevertheless, most of the changes are so large and sys— tematic that there can be little doubt that they are in the gen- eral direction and are of the order of magnitude indicated by comparison of the two surveys. The true orientation and amount of displacement of the stations on the southeast shore of Montague Island (553) and 011 Middleton Island (552) are especially un- certain. This uncertainty exists because (1) both stations are in a part of the net where large dif— ferential earthquake-related ver- tical movements may have caused significant horizontal shifts in their positions, (2) the preearth- quake triangulation involving these stations was only third order and the stations were tied to the net in Prince William Sound through several figures that are geometrically weak, and (3) the stations are situated near the extremity of the net where errors in displacementrelative to the fixed stations are likely to be at a maximum. As a consequence, errors inherent in the adjust— ments could equal or exceed the observed displacements of these two stations in either of the two alternative adjustments. The pattern of horizontal dis- placements relative to stations Fishhook and Klawasi during the time between the surveys is brought out by the displacement vectors (solid) in figure 16. Ex- cept for Middleton Island (552), they show relative seaward move— ments that are predominantly toward the south-southeast in the western part of the area, almost due southeast in the central part, and east-southeast in the eastern part. Over the central part of the net, the magnitude of the dis- placements relative to the base I29 line increases progressively from the base line to a maximum of 64 feet at station 609 on the main- land immediately west of Prince William Sound, after which it decreases towards the southeast. In the western and eastern parts of the net, displacements show a progressive increase in magnitude to the most seaward stations amounting to as much as 52 feet south-southwest from Seward (218) and 31 feet near Cordova (555). In addition, resurveys of small isolated triangulation nets spanning the straits from La- touche and Knight Islands to Montague Island indicate rela- tive shortening of 10—13 feet in a northwest-southeast direction (fig. 16). The overall pattern of move- ment relative to the fixed sta— tions is emphasized by the iso- thismic contours (lines of equal horizontal movement) in figure 16 which show the approximate component of horizontal displace- ment in a S. 45" E. direction, or nearly parallel to the average trend of the vectors in the same area. Contours are based on the displacement vectors (adjustment 2, table 3) and on relative hori- zontal movements within an iso- lated segment of the triangula- tion net between Montague, La— touche, and Knight Islands (Parkin, 1966, p. 9; C. A. Whit- ten, written commun., 1965). Isothismic contours in figure 20 indicate that the entire area from the northern arc of the net to southwestern P r i n c e William Sound showed a relative extension in a seaward direction, whereas the part of the net southeast of the Knight Island-Latouche Is- land area showed a relative short— ening northwest-southeast direc- tion. In other words, the position of Middleton Island, and perhaps the area southeast of the island, FEET METERS 88 E. N ALASKA EARTHQUAKE, MARCH 27, 1964 ...... PATTON BAY FAULT SE METERS FEET ._ Z S. 45° E. component of horizontal displace- . - ' , 12 40 E 10 ment relative to Station Fishhook . - ' i 10 1...: 30 8 .' /\; 8 30 2 20 6 COOk Inlet _ . - ' / Middleton Island 6 20 —' ' I Vertical displacement l (proiected) _ . ' __ Al ut an % 10 4 Station Fishhook . . relative to sea level / 5 ----- '— .. 7.7—?— — N Tfenizh 4 10 a 2 (projected)\ ........ A, /’ ' ' - - .2 ----- 7 \7\?\ \2 SEA OJ— —_______ ..—- _2 {LEVEL -10 —2 —1o PATTON BAY FAULT o 1_._.—.— ._ ,__ ‘__ .‘_ ‘___._ _,__,__. _. ———=I .4 44L;f_1_4_¢_¢_s_.fifi 0 0 75 FEET -__._._n VECTOR SCALE 50 l O 150 KILOMETERS |__L._i__.|_i_l_..____—_l HORIZONTAL SCALE 100 MILES l 18.—Profi1e showing measured and. inferred tectonic displacements (above) and vectorial sum of the horizontal and vertical movements (below) along section B—B’ of plate 1 and figure 16. remained essentially fixed relative to the base line, whereas the inter— vening area was displaced in a relative seaward direction, the amount of displacement attaining a maximum in the Latouche- Knight Island area. Triangulation data indicating shortening across M o n t a g u e Island agree well with the observed imbrication on reverse faults in this area (Plafker, 1967b, p. G40—G41). However, the exact amount of shortening is uncertain because the geographic positions of the stations on the seaward side of Montague Island (553) and on Middleton Island (552) are sub- ject to large errors that may equal or exceed the indicated amounts of displacement at these stations. If the S. 45° W. component of dis- placement dies out at Middleton Island, as inferred in figure 16, the average contraction between that point and the 60-foot isothismic contour is 60 feet. That this amount of contraction may not be unreasonable is indicated by (1) the 10- to 13-foot shortening in a northwest-southeast d i r e c t i o 11 indicated by reobservation of the small isolated triangulation net spanning M 0 n t a g u e Strait between Montague Island and Latouche and Knight Islands, (2) the horizontal shortening of at least 9.3 feet, and possibly as much as 19 feet, across the Patton Bay fault that is indicated by sur- face mapping, and (3) the pro- nounced crustal warping that occurred on and near Montague Island. RELATIONSHIP TO REGIONAL VERTICAL DISPLACEMENTS AND SURFACE FAULTS A genetic relationship between the horizontal and vertical region— a1 displacements is strongly sug- gested by the orientation of the horizontal displacement vectors in a direction roughly normal to the trend of the isobases and by ap- proximate coincidence of the maxi- mum vertical displacements with areas of maximum transverse ex- tension or contraction. This rela- tionship is brought out by the pro— files in figure 18 which show mag- nitudes of the horizontal displace- ments in a relative S. 45° E. direc- tion and the vertical displacements relative to sea level along line B—B’ of figure 16. Also shown in figure 18 are the vectorial sums of the horizontal and vertical dis- placements along the line of pro— file, that is, the direction and rela- tive amount of movement of points on the ground surface along this line. The horizontal displacement data indicate that the zone of sub- sidence extended tranversely by an average of 1.1 X 10“, Or 1.1 parts in 10,000, and reached a maximum of about 3>< 10‘4 slightly seaward from the axis cf the subsided zone. By contrast, at least part of the zone of uplift seems to be one of. net transverse shortening result- ing from crustal warping and re- verse faulting. Average contrac- tion across the uplifted zone as far seaward as Middleton Island is about 10‘4 and it averages as much as 8><10—4 across the narrow belt of maximum uplift on Montague Island. Presumably, a comparable relationship exists between hori- zontal displacements and the earthquake-related vertical move- ments that ocurred outside the re- triangulated area. Extension of the retriangulation net over this area could provide a definitive test of this assumption. In a general way, the displace- ment vectors on either side of the Patton Bay fault (stations 610 and 553) are consistent with the field observations that the fault has undergone reverse movement with resultant crustal shortening by imbrication in the dip direc— tion. In detail, however, there is an unresolved discrepancy be- tween the observed dip-slip move- ment on the Patton Bay fault and the apparent left-lateral strike- slip shift of triangulation stations on either side of it (Plafker, 1967b, fig. 35). The discrepancy-was re- duced by the readjustment (ad- justment 2, table 3) used here, but not altogether eliminated. Ab— sence of an observable component of lateral slip on the fault sug- gests either that the displacement was taken up largely by horizontal distortion between the fault and the two triangulation stations or, more probably, that an error has been introduced into this part of the triangulation adjustment through a slight clockwise rota- tion of displacement vectors. It is significant that, regardless of the details of the horizontal displacements, the triangulation data suggest rebound of a broad segment of the continental mar— gin that had been elastically com- pressed and shortened by at least 64 feet prior to the earthquake. The vectors in figure 16 show the general sense and amount of the rebound within the retriangulazted area. This indicated rebound im- plies preearthquake regional com- pression oriented parallel to the trend of the vectors, or roughly normal to the continental margin and trend of the eastern end of the Aleutian Arc. TECTONICS TIME. AND RATE OF THE DEF‘ORMATION Instrumental records of the time and rate of tectonic move— ments in the deformed area are nonexistent. Three standard tide gages at Seward, Kodiak, and Homer were located where they might have been able to record the vertical land movements relative to sea level had they been opera— tive during the earthquake. How- ever, the Seward gage was de- stroyed in a submarine landslide at. the time of the earthquake, the K0- diak gage with the marigram for the month of March was lost when it was washed away by seismic sea waves half an hour after the earth- quake, and the Homer gage was made inoperative by the shaking. There were no accelerographs in the affected region to record the horizontal movements. As a conse- quence, the time and rate of the movements can only be inferred from the reports of eyewitnesses, from photographs taken after the earthquake, and from the water and atmospheric disturbances generated by the movements. EARTHQUAKE-RELATED MOVEMENTS Numerous eyewitness reports of immediate withdrawals of water from uplifted coastal areas indi— cate that much, if not all, of the deformation occurred during the 11/2—5 minutes of violent tremors. In most places, however, immedi- ate water disturbances resulting from submarine slides or other causes precluded estimates of rela— tive changes in level for several hours or days after the earth— quake. In an area uplifted 6.3 feet, one eyewitness (Gordon Mc- Mahan, oral commun., 1964), thought that the displacements were perceptible as a series of dis— tinct upward accelerations during the earthquake. Another eyewit- I31 ness (Guy Branson, oral commun., 1964), from an area that subsided 5 feet, described a definite drop- ping or sinking sensation toward the end of the strong ground mo- tion “as when a plane hits an air pocket.” No other observers re- ported perceptible accelerations in the direction of the tectonic dis— placements. All of the uplift and surface faulting at the southwest tip of Montague Island occurred prior to March 30th. On this date the up— lifted platform at Cape Cle'are and a part of the Patton Bay fault were phOtographed during a re- connaissance flight (fig. 20). Jim Osborne, a bush pilot who knows the Prince William Sound area intimately and is an exceptionally perceptive observer, informed me that all of the shoreline displace— ments took place prior to the morn- ing of March 28—the day he first flew over the area after the earth- quake. According to Osborne, there were no noticeable shoreline changes after the 28th. His evalu- ation is corroborated by residents along the coast in all areas affected by the tectonic displacements. Movement along strongly uplifted shores occurred at least fast enough to trap many mobile marine animals such as small fish, starfish, and snails above the tide level (fig. 24; p. I 36). That a substantial fraction of the net vertical displacement oc- curred very rapidly is also sug- gested by the pattern of seismic air and sea waves. The peaks be- tween compression and rarefaction on the La Jolla microbarograph record (Van Dorn, 1964, fig. 5) were 7 minutes apart, a difference which suggests a peak-to-peak separation at the origin of about 83 miles; this figure is in close agreement with the observed spac- ing between the axes of uplift and subsidence. As noted “by Van Dorn, I32 the recorded disturbance could only have been produced by verti- cal motions over a very large area, and in a time interval of the order of that required for an acoustic wave to propagate across the di— mensions of the generator. The elapsed time between the earth- quake and the arrival of the initial wave crest along the ocean coast of the Kenai Peninsula further sug- gests that the initiating disturb— ance along the submarine exten— sion of the axis of maximum uplift southwest of Montague Island (fig. 15) occurred during, or within a few minutes after, the earthquake. There is no direct evidence as to when the horizontal displacements, which in some inhabited localities were as much as 60 feet, occurred. No observers reported strong sys- tematic horizontal movements at any time during the main shock, nor could such movements be in- ferred with confidence from the incomplete data on the directions in which objects or structures fell. Nevertheless, as suggested on page I39, horizontal displacements probably occurred during the earthquake, and at a rate fast enough to cause waves in some bodies of surface water. Accelera- tions due to the permanent dis- placements probably were unde- tected by observers because they were masked by the strong ground motions resulting from the transi- ent. elastic seismic waves. PREEARTHQUAKE MOVEMENTS Vertical changes in the position of the shore relative to sea level have been noted within a period of hours prior to some major earth- quakes in Japan (Imamura, 1930, p. 141) . These changes, which have been termed “acute” tiltings or deformations by Japanese scien- tists, have been a subject of special interest because of their obvious ALASKA EARTHQUAKE, MARCH 27, potential importance in earth- quake prediction. During the field investigation of the 1964 Alaska earthquake, an effort was made to ascertain whether any premonitory changes of level were noted by residents in coastal areas or were recorded on operative tide gages. The only suggestion of preseismic changes was an observation made by an officer of Fleet Weather Central at the Kodiak Naval Station to the effect that tides in the area were at least 11/2, and possibly 21/2, feet lower than normal a few days be- fore the earthquake and that the low tides were apparently unre- lated to atmospheric conditions (Lt. C. R. Barney, oral commun., 1964). However, the loss of the March marigram prevented docu- mentation of the reported low tides. The Seward and Homer marigrams for the time preceding the earthquake do not show evi- dence of preseismic changes, nor have such changes been reported elsewhere by coastal residents. POSTEARTHQUAKE MOVEMENTS Relevelings, tidal observations, and gravity readings suggest either no postearthquake vertical changes or, perhaps, slight changes in the earthquake-affected region. The most convincing indication of continued postcarthquake move- ment comes from releveling in May—June and in Octolber 1964 of a line 22 miles long extending northwestward from Portage on Turnagain Arm and a third re— leveling from Portage to Anchor- age in the summer of 1965. Be- tween the preearthquake leveling and the initial postearthquake leveling, Portage subsided 5.6 feet, the area 22 miles to the northwest subsided about 4.9 feet, and An- chorage subsided about 2.3 feet (pl. 1). Comparison of the two 1964 1964 relevelings shows a progres- sive increase in divergence from northwest to southeast, which sug- gests additional relative sub— sidence of about 0.16 foot at Por- tage in the period between surveys (Small, 1966, p. 13). Comparison of the May—June 1964 and the 1965 leveling suggests relative post- earthquake subsidence of 0.36 foot at Anchorage and 0.52 foot at Portage during this interval (Small, 1966, p. 17). Unfortu- nately, neither the October 1964 line nor the 1965 line was tied to tidal bench marks, so the absolute postearthquake displacements are uncertain. Furthermore, because both Anchorage and Portage are situated in areas of extensive thick unconsolidated deposits, the possi- bility cannot be ruled out with the data available that some or all of the indicated subsidence may be due to continued consolidation of soft sediments. Small (1966, p. 18) also reports a gravity increase of about 0.18 mgal on Middleton Island relative to an Anchorage base station. This increase occurred between the time of a postearthquvake 1964 measure- ment and one made in 1965 which would indicate about 2 feet of ad- ditional uplift between surveys. However, the possibility that this large difference in the successive gravity readings may be due to meter drift in one or both surveys is suggested by the fact that resi— dents of the island did not notice changes in relative tide levels dur- ing this same interval. Two feet of uplift at Middleton Island should have been readily detectable along the shore. A comparison of tidal observa- tions made in 1964 and 1965 pro- vides data on the postearthquake land-level changes at 14 of the sta- tions listed in table 1. However, it is difficult to separate purely tec- tonic movements from meteor- ological effects and the effects of surficial compaction at gages sit- uated on soft sediments. Tidal ob- servations in the zone of uplift at Cordova, Port Chalmers, and Saw- mill Bay showed no detectable change suggestive of continued tectonic movements, but one sta- tion, Port Gravina, apparently subsided 0.3 foot between 1964 and 1965. In the zone of subsidence, gages at- Seward and Port O’Brien had no detectable change in mean sea level; five gages at bedrock sites showed slight rises ranging from 0.1 to 0.5 foot, possibly sug- gestive of postearthqurake tectonic uplift. Comparisons of 1964 and 1965 tidal observations at Valdez, Whittier, and Homer, in the zone of subsidence, indicated apparent continued subsidence ranging from 2.6 feet at Valdez to 0.3 foot at Whittier. The postearthquake subsidence at Valdez is definitely related to seaward extension and subsidence of the thick prism of deltaic deposits on which the tide gage is situated; much or all of the subsidence at the other two sites, both of which are on thick de- posits of unconsolidated sediment, could also have resulted from sur- ficial effects. The available data on postearth- quake changes outlined above are internally inconsistent and incon- clusive with reference to postearth— quake vertical movements. Disre- garding the Valdez, Whittier, and Homer stations, where superficial subsidence of unconsolidated de— posits is known or suspected to be large, the repeated postearthquake tidal observations indicate either recovery (by uplift) of as much as 0.5 foot or no change in the subsided zone. However, the re— peated levelings on the Portage— Anchorage line and repeated grav- ity readings at Anchorage have been interpreted as indicating ei— ther continued subsidence or sta- TECTONICS bility in that part of the zone of subsidence. On the other hand, tidal observations in the zone that was uplifted during the earth- quake suggests either postearth- quake subsidence of as much as 0.3 foot or stability, whereas the pair of gravity measurements at Mid- dleton Island in this zone suggest additional uplift of about 2 feet. Repeated tidal observations, level- ings, and gravity readings over a longer time period will be re— quired before definite conclusions may be drawn concerning the post- earthquake pattern of adjustments in the deformed region. It is abun— dantly clear, from available data, however, that there was no large rapid postearthquake recovery of vertical displacement comparable to the recoveries reported after some major earthquakes along the coasts of Japan and South Amer— 10a. EFFECTS OF THE TECTONIC DISPLACEMENTS Regional vertical tectonic dis— placements, bot-h upward and downward, have caused profound modifications in shoreline mor- phology and attendant widespread effects on the biota. Changes in the position of the shorelines relative to sea level directly affected nu— merous coastal installations, ship— ping, and the fishing and shell— fish industries. A major indirect effect of the vertical movements was the generation of a train of destructive seismic sea waves that were responsible for 35 of the 115 fatalities and for much of the property damage attributable to the earthquake. The movements also appear to have generated at- mospheric and ionospheric dis- turbances that were detectable at several places in the conterminous United States. I33 The systematic regional horizon— tal displacements may have caused waves in certain confined and semi- confined bodies of surface water, and related porosity changes may have caused temporary water losses from surface streams and lakes as well as drops in water levels of some wells that tap con- fined aquifers. Because the displacements were along faults that are under water and in uninhabited places on land they did not damage any works of man. Had they occurred in inhab- ited areas, however, these displace— ments surely would have caused extensive damage to structures built across them. It is also rea- sonabl y certain that phenomena re— lated to the reverse faulting, such as the landsliding, extension crack- ing, and severe warping that oc- curred in a belt as much as 3,000 feet wide adjacent to the fault traces (Plafker, 1967b), would have been a definite hazard for engineering works. Most of the effects resulting from vertical movement of the shoreline have been known from other earthquakes in coastal areas throughout the world. Especially detailed descriptions have been given by Tarr and Martin (1912) of the various physiographic and biologic effects of uplift and sub- sidence associated with the great earthquakes of 1899, centered near Yakutat Bay along the Gulf of Alaska coast. Effects of such move- ments on the works of man have also been amply documented for numerous major earthquakes along the coasts of South America, New Zealand, India, Japan, and elsewhere, most of which have been summarized by Richter (1958). Although submarine tectonic movements have long been sus- pected as the most probable gener- ative mechanism for seismic sea waves, the 1964 Alaska earthquake 134 provides what is probably the clearest evidence for a cause-and- efl'ect relationship between these two phenomena. Atmospheric dis- turbances of the type associated with the 1964 earthquake have been recorded previously after large volcanic explosions and nu— clear detonations, but they have never before been observed in :as- sociation with tectonic earth- quakes. To the writer’s knowledge, there are no published reports re— lating surface—water disturbances or ground—water changes to hori- zontal tectonic displacements dur- ing previous earthquakes. PHYSIOGRAPHIC CHANGES Tectonic subsidence, augmented locally by surficial subsidence of unconsolidated deposits, resulted in narrowing or, in extreme cases, complete submergence of beaches. Sea water inundated the lower reaches of some streams in subsided areas as much as 4,500 feet inland from the former mouths, and salt water encroached upon former beach-barred lakes at stream mouths or bay heads (Plafker and Kachadoorian, 1966, p. D27). Beach berms and deltas in subsided areas rapidly shifted landward and built up into equilibrium with the new, relatively higher sea lev— els ( pl. 43) . Former reefs and low- lying islands along the coast were submerged, and some tombolo-‘tied points or capes became islands. Wave action at the higher sea lev- els caused rapid erosion of shore- lines—especially those composed of poorly consolidated deposits that were brought within reach of the tides (fig. 19). An irreplace- able loss resulting from such ac- celerated erosion of these deposits was destruction of coastal archae- ological sites at several places in the Kodiak Island group and on the southern Kenai Peninsula. The major effect of tectonic up- lift was to shift the extreme high— uplifted ALASKA EARTHQUAKE, MARCH 27, 1964 19.—Spruce trees on a spit near the mouth of Resurrection Bay killed by salt-water immersion and undermined by erosion after the land subsided about 3 feet. Photo- graph taken at a 9-foot tide stage, July 10, 1964. tide line seaward and thereby ex- pose parts of the littoral and, at some places, the sublittoral zones (frontispiece; figs. 14, 20). In the areas of maximum uplift on south- western Montague Island, the emergent sea floor is as much as 1,800 feet wide (Plafker, 1967b, pl. 1, 2). As a consequence, former beaches and sublittoral marine de- posits were rapidly incised by streams that cut down through them to new, relatively lower base levels (fig. 21). In many places, beach-barred lakes were drained in varying degrees by incision of their outlet streams. About 8 or 9 feet of uplift at the outlet of shal— low Bering Lake, which formerly was reached by high tides, caused the lake to be suddenly reduced in area by about 4 square miles to a third its pieeart-hquake size. Beaches and deltas developed be- low, and seaward from, their pre— vious positions (fig. 14). Along the shores, preeart‘hquake beaches, sea cliffs, driftwood lines, sea caves, notches, stacks, and benches were elevated above their normal position relative to sea level. Similarly, in offshore areas, uplift created new islands and ex- posed reefs at stages of tide when they formerly were under water. TILTING OF LAKE BASINS Regional tilting or warping of the land surface seems to have caused permanent shoreline changes at Kenai and Tustumena lakes on the Kenai Peninsula. It may have had comparable effects on other lakes for which observa- tional data are unavailable. Tilting of Kenai Lake, which is about 25 miles long, is indicated by changes in the relative position of the bench marks that had been established near its ends prior to the earthquake. Although the ac- curacy of some of the recovered bench mark positions is open to question, the postearthquake sur- vey suggests that the western end of the lake sank 3.0 feet with re- spect to the east end, and that the dip of the tilted surface is N. 72° W. at 1 foot per 5.4 miles (McCul- loc‘h, 1966, p. A29). These data are corroborated by the fact that the west end of the lake is close to the axis of subsidence (pl. 1) and that residents report a relative lower- TECTONICS 20.-—Rocky surf-cut platform a quarter of a mile wide at Cape Cleare, Montague Island, exposed by 26 feet of tectonic uplift. The white band on the upper part of the platform consists mainly of barnacles and calcareous worm tubes; brown algae, or “kelp,” cover much of the surface below the barnacle zone. Photograph taken at about zero tide stage, March 31, 1964. Compare with frontispiece, taken 2 months later in same general area. 21.—Bay-head deposits in MacLeod Harbor, Montague Island, deeply incised by stream erosion following (about 33 feet of uplift. Arrows indicate the positions of pre- and postearthquake high-tide shorelines. Photograph taken August 6, 1965. I35 ing of the lake level at the eastern end after the earthquake (McCul- loch, 1966). The long axis of the 20-mile- long Tustumena Lake and its out- let stream, the Kasilof River, are oriented northwest-southeast, or almost normal to the projected trend of isobase contours in the area (pl. 1). After the earthquake, water levels at the inlet end of the lake reportedly rose above the banks; about 2 feet of southeast- ward tilt of the lake basin is thus suggested (J. D. Reardon, oral commun., 1965). The amount of tilt across the basin, as indicated by reported relative changes in lake levels, is in good agreement with that suggested by the spacing of isobase contours projected from the coast into the Kenai Lowland area (pl. 1). TILTING 0F RIVER DRAINAGES Regional tilting may also have temporarily reduced the flow of certain rivers, such as the Cop- per, Kenai, and Kasilof Rivers, whose flow directions were oppo- site to the regional tilt (pl. 1). The K'asilof River was reduced to a trickle the day after the earthquake (Alaska Dept. Fish and Game, 1965, p. 23) and‘ the Copper River reportedly ceased flowing at its mouth for several days. Immediately after the earthquake the Kenai River for almost a mile below the Kenai Lake outlet temporarily reversed its direction and flowed back towards Kenai Lake (McCulloch, 1966, p. A28), but it is not clear to what extent this reversal was due to tilting and to what extent it was related to the seiching of the lake. Because rivers and lakes were approximately at their lowest an- nual levels when the tilting oc- curred, slight changes in gradi- ent caused disproportionately I 36 large changes in discharge. The changes probably were largely re— lated to upstream tilting of the larger lake basins in the drainage systems with consequent reduc- tions or reversals of discharge un- til the basins onoe again filled to the spillover point. To some ex- tent, however, the reduced flow in the river channels may have resulted from the lowered gradi- ent of the beds. The regional tilting averaged 1 foot per 4.8 miles in the lower Copper River drainage and 1 foot per 10 miles or less in the Kenai Lowland. Other causes, such as channel blockage by river ice or landslides, may also have contributed to the reported temporary declines in discharge. BIOLOGIC CHANGES Vertical displacements of the shoreline strongly affected both the fauna and the flora over a vast segment of coastal south- central Alaska. Some of these ef- fects were apparent within days after the earthquake; others, which depend upon the complex interrelations of one organism to another and to their habitat, will not be known for a long time. Gr Dallas Hanna, who studied the biologic effects of the earthquake in the littoral zone, has given a graphic summary of these earth- quake-related changes (Hanna, 1964). The results of detailed governmental and private studies of the effects of the earthquake on intertidal organisms, land plants, and fish are to be reported in the Biology Volume of the planned series of publications of the Committee on the Alaska Earthquake of the National Academy of Sciences (W. L. Pe— trie, oral commun., 1968). The most conspicuous effect of subsidence was the fringe of ter- restrial vegetation killed by salt- water inundation at periods of high tides (pl. 43; figs. 12, 19). ALASKA EARTHQUAKE , MARCH 27, 19164 22.—Spruce trees in Nuk‘a Passage on the southern Kenai Peninsula killed by re- peated inundation with salt water in an area of 6.3 feet of tectonic subsidence. Algae and animals of the upper littoral have encroached upward into the former terrestrial environment. Inset shows a barmacle (white) and numerous Littom‘na or “periwinkles” (gray) on the roots of a tree. Photograph taken July 22, 1965. Virtually all nonclifl’ed shorelines that subsided more than 3 feet clearly showed fringes of dead vegetation within 2 months after the earthquake. In some sheltered localities at which vegetation ex- tended down to the extreme high- tide line, dead vegetation was noticeable even where subsidence was as little as 1 foot. Trees, bushes, beach grass, and muskeg along many former beaches were killed and partially buried in gravel or sand. Extensive areas of coastal marshland and for— est that formerly had provided winter forage for grazing animals or nesting grounds for migratory birds were inundated. In such places, marine organisms en- croached upward into the new lit— toral zone and it was not uncom- mon to find barnacles, limpets, and algae living on or among the re— mains of land plants (fig. 22). The effects of subsidence on ses- sile intertidal marine organisms submerged below their normal growth positions were not readily apparent. Undoubtedly, individu- als near the lower depth range of the species were adversely affected by the changed conditions and were gradually replaced by other organisms better adjusted to the deeper water environment. Effects of uplift on the biota of the littoral zone were more strik- ing than those resulting from sub- sidence, because the uplift caused complete extermination of orga- nisms that were permanently ele- vated above their normal ranges. The width of the resultant band of dead organisms depended, of course, on both the amount of uplift and the slope of the uplifted shore. In areas where uplift ex- ceeded the local tide range, as on islands in southern Prince William Sound, on parts of the mainland coast to the east of the Sound, and on several offshore islands on the Continental Shelf, destruction of the sessile organisms was almost absolute. Even many of the mobile forms—including starfish, gastro— pods, and small fish—did not sur- vive. Some of the effects of uplift TE CTONICS 23.—~Closeup View of surf-cut surface at Cape Cleare, Montague Island, shown in frontispliece. The white coating on the rocks consists primarily of desiccated cal- careous algae and bryoztoans; the dark ropelike objects are stipes of laminarian‘s (“kelp”). Photograph taken June 1, 1964. on organisms of the littoral zone are illustrated by plates 3A, 4A; figures 5, 23, and 24. The dramatic change with time in the appear— ance of the shore and sea floor after about 26 feet of uplift at the south- west end of Montague Island may best be appreciated by comparing the aerial photograph taken on March 30th, 3 days after the earth- quake (fig. 20), with one taken 2 months later on May 30th (fron- tispiece). By August 1964 a few land plants had encroached onto the fringe of shore reclaimed from the sea, and in the summer of 1965 scattered clumps of grasses and wildflowers grew everywhere, on raised beaches and deltas and in favorable localities on rock bench- es amid the dead and dried re- mains of marine organisms (fig. 13). In a few years the bleak as- pect of these fringes of uplifted shore should become subdued by a luxuriant cover of brush and timber comparable to that grow- ing on older uplifted marine ter- races in the area. By July 1965, land plants had covered much of the raised platform on Middleton Island and sea birds had already begun nesting in the former inter- tidal zone. Throughout the uplifted areas in and near Prince William Sound, the mortality of all types 137 of shellfish—including commer- cially important razor clams—has been estimated to be as high as 90 percent by G Dallas Hanna (oral commun, 1965). At many places where uplift exceeds the normal tide range, the clam population was literally wiped out. In such areas, the populations of birds, fish, and other animals that normally feed on shellfish must eventually readjust down- ward to the reduced food supply. The potential effect of the land- level changes on the important salmon runs in the affected areas cannot be fully evaluated until the matured 1964 hatch returns from the sea to spawn. Spawning areas for pink and chum salmon, which are intertidal spawners, received major damage due to changes in land level and seismic sea waves (Alaska Dept. Fish and Game, 1965, p. 3; Thorsteinsson, 1965). Spawning areas of upstream mi- grants, including the red and sil- ver salmon, where relatively un- affected by the earthquake. Many low-lying coastal lakes that were important habitats for 24.—Mass of dead starfish in a depression on the uplifted platform shown on figure 23. Photograph taken May 31, 1964. I38 waterfowl and for fresh-water fish were damaged mainly by salt- water pollution related to subsi- dence or to draining in varying degrees resulting from uplift and incision of their outlet streams. A few of the uplifted lakes that for- merly received salt water through their outlets during high tides be- come entirely fresh with conse- quent changes in the number and species of fish they can support. GENERATION 0F SEISMIC SEA WAVES Most major earthquakes in coastal areas that involve vertical tectonic displacements beneath the sea are accompanied by seismic sea waves, and the 1964 earthquake generated one of the larger seismic sea waves of recent times (Grantz and others, 1964, p. 11—12; Van Dorn, 1964; Plafker and Mayo, 1965; Plalfker and K'achadoorian, 1966; Pararas—Carayannis, 1967). Between the southern tip of Kodiak Island and Kayak Island, these Waves took 20 lives and caused destruction all along the ocean coast. The waves were espe— cially destructive along the ocean coast of the Kodiak group of islands and the Kenai Peninsula areas that had been lowered rela— tive to sea level by tectonic sub- sidence or by the combination of tectonic subsidence and compac- tion of unconsolidated deposits during the earthquake. In addi- tion, the waves, which were re- corded 0n tide gages throughout the Pacific Ocean, caused 15 deaths and major damage in British Co- lumbia, Oregon, and California. The wave—source mechanism was initially investigated by Van Born (1964), who concluded that the waves were generated by a di— polar displacement of water re- sulting from regional tectonic warping. He inferred that the positive pole of this disturbance included much of the shallow Con- ALASKA EARTHQUAKE, MARCH 27, 1964 tinental Shelf bordering the Gulf of Alaska Within the major zone of‘uplift, and that the negative pole lay largely under land or be- neath Cook Inlet and Shelikof Strait in the major zone of subsid- ence. From preliminary data on the amount and distribution of vertical displacements along the shore, Van Dorn (1964, p. 17) cal— culated that the total potential en- ergy imparted to the positive part of the seismic sea. wave by subma- rine uplift (assuming (1) vertical displacement of 6 feet that in- creases progressively from zero at the southwest end to 6 feet at the northwest end and (2) source di- mensions of 240 miles by 100 miles), was 1.7>< 1014 ft-‘lbs (2.3x 1021 ergs) , or only about 0.01—0.05 percent of the approximately 1024 to 2X 10“"1 ergs of seismic—wave energy released by the main shock. Pararas—Carayannis (1967), using source dimensions of 93 miles (150 km) by 435 miles (700 km) and the same average uplift as inferred by Van Dorn, arrived at a total water-Wave energy of 5.88><1021 ergs. The Geological Survey’s subse- quent studies of the vertical dis- placements on land and their prob- able extension beneath the sea provide additional data relevant to the probable configuration of the initial positive wave and its energy content. These data sug— gest that the initial wave form, due to vertical displacement of the sea floor on the Continental Shelf, probably had the general cross- sectional shape indicated by the profile in figure 15 and that the offshore areas involved in the up- lift and the amounts of sea-floor displacement are considerably greater than was indicated by pre- liminary reconnaissance surveys. Thus, the general shape of the de- formed surface on the Continental Shelf may be roughly approxi- mated by a broad low-amplitude upwarp with minimum dimensions of 400 by about 75 miles, superim- posed upon which is a narrow belt of maximum uplift about 6 miles wide that is inferred to extend some 350 miles southwestward from Montague Island. As indi- cated on profiles A—A’, B—B’ 0—0’ plate 1, average uplift acrOSS the broad upwarp is roughly about 12 feet and that across the narrow zone is probably at least 30 feet. Because this highly simplified model does not consider the addi- tional wave energy at the ends of the deformed region, where uplift gradually falls off to zero, or in that part lying seaward from the edge of the Continental Shelf where the deformation field is un- known, the calculated energy should be considered as a mini- mum. If the initial wave form approx- imates that of the uplift, total potential energy transferred to the water, Et was the sum of the energy in both the broad low-am- plitude part of the wave (E1) and in the narrow superimposed high- amplitude part (E2). Total poten- tial energy transferred to the water, E), derived by using Iida’s equation (1963, p. 65), was Et=E1+E2= P9<1014 ft-lbs (6.2)(102I ergS); E: (1.1) (32) (6)2(75) (400) (5,280)2 =1.06><1015 ft-lbs (1.4><1022 ergs) ; and their sum, E',=1.5>< 1015 ft—lbs (2X1022 ergs). These figures suggest that the total potential energy in the posi- tive part of the wave may be about an order of magnitude larger than that derived by Van Dorn, or 0.1— 0.5 percent of the seismic wave energy release. According to the model used, roughly one-third of the energy was concentrated in the narrow high—amplitude part of the wave along the axis of maximum uplift and two-thirds was dist-rib- uted over the low-amplitude part of the wave which has an area roughly 15 times larger. Thus, the relatively greater damage and higher wave runups along the outer coast of the Kodiak group of islands and the Kenai Peninsula, as compared to the ocean coast of Prince William Sound and the mainland east of the sound, ap- pears to be a function of prox- imity to the narrow zone of high wave-energy concentration along the axis of maximum uplift. ATMOSPHERIC EFFECTS An atmospheric pressure wave that was the atmospheric counter- part ofthe seismic sea waves was recorded on microbarographs at the University of California at Berkeley and at the Scripps Insti— tute of Oceanography at La Jolla, Calif. The wave traveled at the speed of sound in air (roughly 1,050 ft per sec in the lower atmos— phere), reaching Berkeley, 1,950 miles from the epicenter, 2 hours and 40 minutes after start of the earthquake (Bolt, 1964, p. 1095) and La J olla 39 minutes later (Van Dorn, 1964, fig. 5). Travel times to these stations correspond to an initiating disturbance in the epicentral region during the earth- quake. The pressure wave’s signa- ture further suggests that it was caused by the vertical tectonic dis- placements of the land and sea sur- faces that accompanied the earth- quake. TECTONICS The atmospheric pressure wave also seems to have caused a travel- ing ionospheric disturbance that was observed in Hawaii, Alaska, and the conterminous United States on high-frequency radio sounders (Row, 1966). The dis- turbance at Boulder, 0010., was characterized by an abrupt onset, speeds appropriate to sound waves above 100 km in altitude, an oscil- latory long-period tail, and an ini- tial negative doppler. Computa— tions by Row indicate that the es- sential features of the observations may be reproduced by sudden ver- tical ground displacement of the type observed in the epicentral region below a plane isothermal gravitating atmosphere. WATER DISTURBANCES POSSIBLY RELATED TO HORIZONTAL DISPLACEMENTS Water disturbances that accom— panied the earthquake in some lakes, fiords, and rivers may have been generated by inertial effects of the water bodies as the land mass was displaced horizontally beneath them. Horizontal move- ment of a deep steep-sided basin or fiord, if it occurred fast enough, would be expected to im- part potential energy to a con- tained water mass by changing its surface configuration as illus- trated diagrammatically by fig- ure 25. Thus, because of its in- ertia, water would tend to pile up above its original level along shores on the side of the basin opposite to the direction of dis- placement, and it would simul- taneously be lowered along shores in the direction of displacement. For a given amount and rate of displacement, the eifect of hori- zontal movement on the water . mass would be proportionally greatest where orientation of shores is normal to the direction of horizontal movement and rel- atively steep basin sides permit- ted the maximum energy to be 139 25.——Schematic diagram illustrating the postulated effect of a sudden hori- zontal displacement on water in an en‘ closed basin, the amount of displace ment assumed to be small relative to the dimensions of the basin. Dashed lines indicate the original position of the basin, solid lines the position af- ter displacement. Symbols along the basin margin in the plan view (above) indicate shores along which an initial rise (+) or drop (—) in water level would occur; profile A—A’ shows a possible configuration of the water surface immediately after the dis- placement. transferred from the basin to the contained water mass. McCulloch (1966, p. A39) has reported uninodal and multinodal seiche waves in Kenai Lake with half—wave amplitudes of 5—6 feet and initial runup heights that were locally as much as 30 feet. He inferred' that they were gen- erated by a tectonic tilting of the lake basin that amounted to no more than 3 feet. A possible al- ternative explanation, however, is that the waves and seiche in Kenai Lake—a lake which lies in a long narrow steep-sided glacial valley—resulted mainly from the 15—25 feet of south—southeast hor- izontal translation of the lake basin that accompanied the earth- quake in that area (fig. 16). Be- cause of the irregular shape of the basin and uncertainties re- garding the rate at which the horizontal displacements occurred, it is not possible to determine I40 quantitatively whether the hori- zontal displacements alone or in combination with tectonic tilting could generate the waves recorded at Kenai Lake. Sudden rises of water level dur- ing or immediately after the earthquake, observed at numerous coastal localities where there was no evidence for submarine sliding, strong tilting, or faulting could also have been caused by the hori- zontal displacements. Within Prince William Sound, where horizontal displacements in a south-to-southeast d i r e c t i o n ranged from about 20 to 62 feet (fig. 16), local waves of unknown origin were responsible for the loss of at least 28 lives and caused ex— tensive property damage at Che- nega, Port Ashton, Port San Juan, Port Oceanic, Perry Island, and probably at Port Nellie Juan and Point Nowell (Plafker and others, 1969). Similar waves that did not cause damage also were reported at Port Wells, Unakwik Inlet, Tatitlek, Naked Island, and sev- eral other localities in Prince Wil- liam Sound. Much of the damage from local waves was concen- trated along east-west- to north- east-southwest-trending shores in semiconfined bays or along deep steep-sided fiords and straits. These waves, which appeared at widely separated localities in the sound within minutes after the eanthquake was first felt, must have been generated locally and almost simultaneously. Most eye- witnesses observed a single large wave with runup as high as 70 feet (as at Chenega), preceded or fol— lowed by much smaller waves at intervals of a few minutes. The sudden onset, short period, and local distribution of the waves dis- tinguish them from the train of long-period seismic sea waves generated in the Gulf of Alaska that did not reach the outer coast ALASKA EARTHQUAKE, MARCH 27, 1964 of the Kenai Peninsula until about 20 minutes after the start of the earthquake. That the waves may have been generated by relative seaward movement of the land mass in Prince William Sound is suggested by (1) their appearance during the earthquake, (2) their occurrence in an area where there were large horizontal displace- ments, and by (3) the orientation and configuration of the affected shorelines. Unexplained waves were also observed in widely scattered coas- tal areas of the Kenai Peninsula and Kodiak Islands, where retri- angulation data are unavailable but where significant horizontal displacements probably occurred. For example, waves as high as 9 feet were reported by eyewitnesses in the Homer area during and im- mediately after the earthquake. Such waves could not be attrib- uted to sliding, slumping, or other causes (Waller, 1966a, p. D3—D4). The curious breaking and surging of the waves on the tidal flats sug- gested to one observer that “the land was being shoved under the bay” (Waller, 1966a, p. D4). Rapid, calm rises in water level of 9 feet at Kodiak (Plafker and Kachadoorian, 1966, p. D30) and of about 26 feet at Whittier (Plaf— ker and Mayo, 1965, p. 15) that cannot be readily ascribed to any other cause may also have been re- lated to horizontal displacement of the land. In summary, horizontal dis- placements of the magnitude indi- cated by retriangulation data, if they occurred fast enough. should theoretically generate waves in water bodies of suitable size and configuration. This movement may have been the cause, or a contrib- uting cause, of some waves ob- served in certain localities during or immediately after the earth- quake that cannot be directly re- lated to vertical tectonic displace- menits, regional tilting, seismic shaking, or submarine landslides. CHANGES IN ARTESIAN-WELL LEVELS Systematic long-term drops in water levels of wells tapping con~ fined aquifers in Pleistocene and late Tertiary strata were recorded at various widely spaced localities within the zone of tectonic sub- sidence (Waller, 1966a, p. D16- D18; 1966b, p. A18—A26) . Records of seven representative artesian wells from Anchorage, Chugiak, and four communities on the Kenai Peninsula are shown in figure 26. The residual drops in well levels at the time of the earth- quake range from about 7 to 25 feet, and none of the wells showed full recovery within a year after the earthquake. Observed long-term changes in well levels suggest changes in the physical structure of the aquifers and a net increase in aquifer—pore space. Such changes could be caused by rearrangement of grains or fractures as a result of the hori- zontal extension (on the order of 2>< 10-4) and (or) the elastic dila- tation that is known or inferred to have affected the areas in which these wells are located. A similar effect was looked for, but not found, in the oil wells of the Swanson River oil field located in the zone of tectonic subsidence near Kenai (R. I. Levorsen, writ- ten commun., 1966). Any small strain change that may have oc- curred probably was masked by changes in volume of the relatively compressible oil-water-gas mix- ture filling the pore space of the field reservoir. The only artesian water wells in the zone of tectonic uplift are in a thick deposit of glacial drift at Cordova (Waller, 1966b, p. A20—A21). Because these wells did not have recorders installed in them, their response to the earth- TECTONICS FEET 1 0 -_- Anchorage we“1,depth540';ane§an \ aquifer in Tertiary deposit —10 \\ \ ,..’~\_,,/-—__._ _20 I I l I l | I | I I I I l I L , _ 2 O —‘ Anchorage well 505, depth 453’: arte- \ sian aquifer in Pleistocene deposits -10 \ —2o \ // I_/ _30 | I I L | I I 3. 0 ‘—\ Chugiak wel|120,depth124’;artesian aquifer in PIeistocene deposits —5 \A __————-'"'”\ —10 _15 | I I | | I I I l l | I I | I 4. 0 _ —'_‘\ Kenai well 17, depth 163’; artesian aquifer _. _ _,———- ‘- 5 \ // \/ ‘10 \ \__./ __15 I I I 1 I I l I I I 1 I I 1 I 5 0 § \_. Anchor Point-Kasilof welI9; depth 100’, \ anesian aquifer /.\ _5 // \\ / \ __/f \l —10 \ _15 , I I | I I I I l I I | I | I I 6- O _\ Soldatna well 63, depth 100'; artesian \ aqufier —5 __ __ __ __ _’ / —— ‘- _10 I I I 1 1 (I 1 | 1 I 1 1 I 1 7 0 __‘ Homer well 15. depth 123’; artesian ' aquflerin Ternary deposns 5 \ ' ____ 'T' ” ’.’ _10 I I I \' I | l I I l | I I I I J F M A M J J A S O N D J F M 1964 1965 DATE 26.—Artesian well records from the zone of tectonic subsidence showing systematic drop in water levels at the time of the earthquake. After Waller (1966a, fig. 13 ; 1966b, figs. 14, 20, 22). I411 quake cannot be correlated with that of the wells in the zone that subsided. However, comparison of measurements in three wells made in July 1962 with measurements made 4: months after the earth- quake showed about a 1—foot rise in water level, rather than the re- sidual drop that characterized wells in the subsided zone. STREAMFLOW, LAKE LEVELS, AND SHALLOW WELLS Changes in the levels of many lakes, streams, and shallow wells in unconfined aquifers were ob- served at numerous localities within the zone of tectonic subsid- ence. In general, the reported changes involved temporary water losses (Waller, 1966b, p. A8—A11; Plafker and Kachadolorian, 1966, p. D23—D24). One of the more probable causes for such changes is an increase of intergranular or fracture porosity in the surround- ing materials consequent upon horizontal extension and elastic dilatation across the subsided zone during the earthquake. COASTAL FACILITIES AND SHIPPING Regional land-level changes— including both subsidence and uplift—caused direct and costly damage to homes, canneries, trans- portation routes, airfields, docks, harbors, and other facilities throughout the affected areas (figs. 27, 28). Many facilities that had otherwise been unaffected either by the earthquake or the destructive water waves associated with it were damaged by land-level changes. Such changes had rela- tively few short-term beneficial ef- fects on the works of man. Because the various forms of damage re- sulting from vertical tectonic movements have already been de- scribed in detail in the various re- ports of this series on effects to communities (US. Geol. Survey Professional Paper 542) and were I42 ALASKA EARTHQUAKE, MARCH 27, 1964 27.—Road along Womens Bay, Kodiak Island, in an area of about 5.5 feet of tectonic subsidence and an unknown, but probably substantial, amount of local settling of unconsolidated deposits. Since subsidence, the road has been flooded at high tide and subject to erosion by waves. Photograph taken at 40-foot tide stage, July 20, 1964. 28.—Canneries and fishermen’s homes along Orca Inlet in Prince William Sound placed above the reach of most tides due to about 6 feet of uplift. Photograph was taken on July 27, 1964, at a 9-foot tide stage, which would have reached beneath the docks prior to the earthquake. summarized by Hansen and Eckel (1966) and Eckel (1967), they need not be described here. GRAVITY CHANGES Vertical displacements were ac- companied by measurable changes in gravity at several stations where comparative pre- and postearth- quake gravity readings were made (D. F. Barnes, 1966; oral commun. 1966). The stations were distrib- uted in both the zones of subsi- dence and the zones of uplift where changes in elevation ranged from —5.8 feet at Portage to about + 11 feet at Middleton Island. Cor— responding gravity changes were between +0.5 milligals to —0.67 milligals. Barnes (1966, p. 455) noted that the gravity changes, at least in the uplifted area, tend to approximate the Bouger, rather than the free-air, gradients. Al- though uncertainties in relocating some of the station positions pre- clude firm conclusions, the data suggest that there has been a re distribution of mass in at least those parts of the deformed region where the changes correspond to Bouger gradients. COMPARISON WITH OTHER EARTHQUAKES In terms of areal extent of de- formation and amount of residual horizontal and vertical displace- ment, the 1964 Alaska earthquake is one of the most impressive tec- tonic evenrts ever recorded. This fact is brought out by table 4, which compares the deformation associated with the 1964 event with that of selected great earthquakes for which quantitative data are available. The area of observable crustal deformation, or probable deforma— tion, that accompanied the 1964 earthquake is larger than any such area known to have been associated with a single earthquake in his- toric times. Comparable tectonic deformations have probably oc- curred during other great historic earthquakes, but if so they were beneath the sea, along linear coast lines, or inland, where it generally is not possible to determine the areal extent of such features with any degree of confidence. For ex- ample, the area afl'ected by vertical displacements during the great series of Chilean earthquakes in May and June of 1960 extended TE CTONICS I43 TABLE 4.—C’omparative deformations of the 1964 Alaska earthquake and some other great earthquakes Approximate . Maximum Maximum relative Magnitude Probable area of Maxnnum vertical relative fault displacement Earthquake Date (Richter surface warping warping relative horizontal (feet) Data source scale) (square miles) to sea level (feet) distortion (H, horizontal; (feet) V, vertical) 1964 Alaska _____________ March 27, 1964.. 8. 4+ 108,300) (major zones +38, —7)é ........ 64 ___________ 211?; 20—23V _________ This paper. 0 y . Niigata, Japan _________ June 16, 1964..-. 7. 7 =20,000 ............... +4.9, —1. 5 ________ No data ..... Submarine only, Earthquake Research displacement Institute (1964). unknown. 1960 Chile ______________ May 22, 1960..-. 8.3 17,000+ ............... +10-13, —7. 2 ..... _...do. Np kpown surface Saint‘AInand (1961). an ts. Lituya Bay, Alaska... July 10, 1958 ..... 8. 0 No data; fault rup- None reported“ _ . ..... do ....... 21.5H; 6V ___________ Tocher (1960). fure 115+ miles ong. Gobi-Altai, Mongolia... December 4, 8.6 =32,000 (area of fault- No data ________________ do ....... 29H; 30V ____________ Florensov and Solo- 1957. (G.) ing . nenko (1963). Nankaido, Japan ....... Disamber 20. 8.5 20,000+ ............... +2. 9, —2. 0 ________ 6. 2 .......... N? Epown surface Inoue (1960, p. 84-85). 46. a ts. Hawkles gay, New February3, 1931. 7.9 2,000 (minimum) ...... +9, —2 to 3 _______ No data . 6-7H; 6—8V?_ _ ....... Henderson (1933). Zea an . Kwanto, Japan ......... Sepgt263mber l, 8.3 10+ __________________ +6. 6, —5. 3 ________ 15 ___________ 6.6V ................. Magi; 0(1932); Inamura San Francisco, Calif... . April 18, 1906...- 8. 3 No data; fault rup- Minor _____________ =20 _________ 21H; 3V _____________ Richter .(1958, p. 476— ture 270 miles long). . _ Yakutat Bay, Alaska... September 1899. 8. 5—8. 6 1,300 (minimum) ______ +47, —7(?) ........ No data ..... 8+H?; 8V (on sub- Tart and Martin sidiary faults). (1912). Assam, India ........... June 12, 1897. . .. 8. 7 2,000 (minim' ' um) ______ No data ________________ do _______ 011?; 35 ............ lighter (1958, p. 49— north-south some 420 miles along the Pacific Ocean coast, and at least 40 miles in an east-west direc- tion at the Gulf of Ancud near the southern end of the affected area (Saint—Amand, 1961). If the in- land extent of deformation to the north is comparable to that in the Gulf of Amend, the minimum area of surface warping on land would be 17,000 square miles. The distri- bution of aftershocks and the source area of the destructive train of seismic sea waves associated with the earthquake suggest fur— ther that significant movements oc- curred over an extensive area of the sea floor adjacent to the de- formed Pacific coast of Chile. Other major earthquakes in which long sections of the west coast of South America report— edly changed elevation ocurred in 1751, 1822, 1835, 1837, and 1906 (Richter, 1958) ; data on the areal distribution of the changes, how- ever, are scarce. Warping as- sociated with one of these, the violent Chilean earthquake of 1822, gave rise to speculation that an area of some 100,000 square miles of coastal Chile had been up- lifted (Lyell, 1874, p. 94), but there are few data to support this asser- tion. Imamura (1930) lists 26 Japanese earthquakes that result- ed in vertical displacements. Re- gional warping has accompanied some of the more recent Japanese earthquakes, notably the 1946 Nan- kaido earthquake (Inoue, 1960, p. 85) and the 1964 Niigata earth- quake (Hatori, 1965, p. 133—136). Elsewhere in the circum—Pacific region, additional large-scale ver- tical tectonic deformation was re- ported following the major earth- quakes of 1848, 1855, and 1931 in New Zealand, and 1762 in India and Burma. Known areas of de— formation associated with all of these earthquakes, however, are but small fractions of that in- volved in the 1964 Alaska event. The 38 feet of uplift on Monta- gue Island during the 1964 earth- quake is known to have been ex- ceeded only by the 47.3 feet of up- lift that occurred during the earth- quakes of 1899 that were centered at Yakutat Bay, 185 miles to the east (Tarr and Martin, 1912, pl. 14). Reported submarine vertical displacement offshore from Mon- tague Island, however, may equal or perhaps exceed the 1899 move- ments (Malloy, 1964, p. 1048). Subsidences roughly equal to, or slightly larger than, the 71/2 feet that accompanied the 1964 earth- quake have been recorded during previous seismic events, although at many places determination of the absolute amount has been com- plicated by surficial slumping or compaction effects in unconsoli- dated deposits. For example, the largest reported coastal subsidence, 17 feet, which accompanied the Great Cutch earthquake (Lyell, 1874, p. 98) was at the mouth of the Indus River in an area where sig- nificant surficial compaction of deltaic deposits was to \be ex- pected. Following the 1923 Kwanto, Japan, earthquake, unusually large submarine displacements, involving as much as 825 feet of uplift and 1,320 feet of subsidence on the sea floor in Sagami Bay, were inferred from a comparison of pre— and postearthquake sound- ings (Richter, 1958, p. 570—571). However, because of (1) possible errors in the positioning of the preearthquake survey and (2) the effects of submarine sliding, these data are of doubtful value for in- ferring tectonic movement; con- sequently they are not included in table 4. I44 The regional horizontal dis- placements associated with the 1964 earthquake, which were about 64 feet, are significantly larger than any previously re‘ corded. By contrast, regional dis- placements associated with the great 1923 Kwanto earthquake, as indicated by retriangulation, were less than 15 feet (Muto, 1932, fig. 6), and the maximum relative In the following sections the 1964 earthquake is viewed from the perspective of its broad rela- tionship to the Aleutian Arc and to other major structural ele- ments of south-central Alaska. Emphasis is placed on the geolog- ical and geophysical evidence for late Cenozoic tectonic movements in the immediate region affected by the earthquake. Earlier tec- tonic features and events, and contemporaneous features and events outside the region affected by the earthquake, are not of primary interest here, although they are mentioned where neces— sary to provide the setting of the principal features and events that are treated. Most of Alaska has been stud- ied geologically in a reconnais- sance manner, but detailed map- ping is still relatively uncommon and is largely confined to a few mining districts and to outcrops of potentially petroliferous or coal—bearing rocks around the margins of sedimentary basins. Gross aspects of the lithology and structure in outcrop areas are reasonably well known, but pres- ent knowledge of the geologic record in most places is inade- quate for a detailed interpreta- tion of the tectonic history. Sub- surface investigations in the vast ALASKA EARTHQUAKE, MARCH 27, 1964 displacement resulting from com- bined distortion and fault oflsets associated with the 1906 Cali- fornia earthquake was roughly 21 feet (Reid, 1911). Measured vertical displace- ments across the two subsidiary earthquake faults on Montague Island—the Patton Bay fault (20—23 ft) and Hanning Bay fault (161/3 ft)—are considerably TECTONIC SE'I'I‘ING terrestrial areas of sparse out- crops and in the submarine parts of the Continental Shelf and Aleutian Arc are in a state so rudimentary that geological in- terpretations based upon them must be considered as very ten- tative. The spatial distribution of earthquakes and the rapidly in- creasing number of available fault-plane solutions provide val- uable indirect evidence for the orientation and nature of dis- placements that have given rise to the earthquakes. Recent im- provements in locating hypocen- ters of Alaskan earthquakes should contribute substantially to resolving the precise relationship of the earthquakes to known structural features. THE ALE'UTIAN ARC The 1964 earthquake occurred in the tectonically complex region where northeastatrending struc- tural elements associated with the Aleutian Arc overlap and merge With arcuate structures of south- central Alaska. The belt of seis- mic activity and the major zones of tectonic deformation associated with the 1964 earthquake lie large- ly between and parallel to the chain of active volcanoes and the oceanic trench that constitute the arc and are presumably related larger than any others of definite reverse type previously described (Plafker, 1967b). A few normal or oblique-slip faults, however, have undergone larger vertical dis- placements. The largest of these was associated with the 1897 In- dian earthquake and reportedly was as much as 35 feet (Richter, 1958,p. 51). genetically to it (fig. 1). Conse— quently, it is pertinent to review briefly the data and current hy- potheses relevant to tectonic proc- esses within the arc. MAJOR FEATURES The Aleutian Arc, which sweeps 1,800 miles (2,800 km) across the North Pacific Ocean from Kam- chatka to southern Alaska, ex— hibits all of the striking features characteristic of the festoons of arcs that ring the Pacific basin. These are ( 1) an arcuate deep oceanic trench — the Aleutian Trench—Which is convex toward the ocean basin except near its eastern end, (2) a subparallel vol- canic chain on the concave, or in- ner, side, away from the ocean ba- sin——the Aleutian volcanic arc, (3) an associated belt of active seismicity between the trench and volcanic chain in which the lower limit of hypocen-ters tends to deep- en from the vicinity of the trench toward the arc, and (4) parallel zones of isostatic gravity lows over the trench bottom and gravity highs between the trench bottom and volcanic arc. The Aleutian Arc difl'eIs from most other arcs in that it partly traverses and partly follows along, the margin of the oceanic basin (fig. 1). In its western part the arc consists of the Aleutian Ridge, which is surmounted by a chain of volcanoes that comprise the Aleu- tian Islands, and the Aleutian Trench, which lies at a distance of less than 155 miles (250 km) on the convex (south) side of the ridge. As it approaches south-cen- tral Alaska, the distance between the volcanic arc and the trench gradually increases to more than 250 miles (400 km), the belt of seismic activity becomes broader and less well defined, and the trench gradually shallows to be- come indistinguishable from the floor of the North Pacific Ocean. It is in this eastern portion of the are that the line of volcanoes and the belt of seismicity overlap and in part merge with preexisting structural elements that roughly parallel the margin of the Gulf of Alaska (fig. 29). The group of vol- canoes in the Wrangell Moun- tains, which are separated from Mount Spurr, the most northerly volcano of the Aleutian Range, by a gap of 420 miles, aresimilar in age and composition to the Aleu- tian Arc volcanoes. These simi- larities, and their position near the eastern end of the Aleutian Trench and its associated belt of intermediate-depth earthquakes, suggest that the Wrangell Moun- tains volcanoes may also be geneti- cally related to the arc. The geology of the Aleutian Islands segment of the arc was summarized by Coats (1962), that of the Aleutian Range and nearby areas by Burk (1965). Both of these writers presented thorough reviews of available bathymetric, seismologic, and marine geophysi- cal data as well as hypotheses con- cerning origin of the arc. Results of detailed marine geophysical sur- veys along the eastern end of the Aleutian Arc, carried out during 1964 and 1965 by the Scripps In— stitute of Oceanography and the US. Coast and Geodetic Survey, TECTONICS had not been published at the time this report was completed. SEISMICITY Figures 1 and 30 indicate that, in plan, most of the belt of con- centrated seismic activity is be- tween the trench and the volcanic chain along the length of the arc; a small number of earthquakes oc- curred on the south wall of the Aleutian Trench and north of the volcanic chain. Shallow-depth (>70 km) earthquakes have oc- curred throughout the area in— cluded between the Aleutian Trench and the Aleutian volcanic are, but most of the intermediate- depth earthquakes (70—170 km) were located in the northern part of this area or north of the volcanic chain. The most easterly interme- diate-depth earthquakes have been along the coast of the North Paci- fic Ocean at long 145° W. in the vicinity of the Wrangell Moun- tains. In south-central Alaska the belt of seismicity associated with the Aleutian Arc bifurcates into a broad zone of shallow and inter— mediate depth shocks that sweeps northward into central Alaska and a belt. of shallow-focus earthquakes that extends eastward along the continental margin. Comparison of figures 2 and 31 shows that the spatial distribution of the seismi— city associated with the 1964 Alaska earthquake closely follows the pattern of previously recorded earthquakes along the arc and thereby implies a genetic relation~ ship between them, as suggested by Arthur Grantz, shortly after the event (Grantz and others, 1964, p. 2). From studies of earthquake dis- tribution along the Aleutian Arc, Beniofl (1954, p. 391) concluded that the hypocenters lie in a broad zone that dips northward from the vicinity of the Aleutian Trench at an average angle of 28°. Saint- I45 Amand (1957, p. 1348) interprets the same data as indicative of a zone dipping at an angle of about 45° between long 172° and 179° W. and at a “somewhat lower angle” in the eastern part of the arc. Detailed studies of seismicity associated with the 1964 earth— quake suggost that some planes within the earthquake focal region dip beneath the are at angles of less than 15°. Benioif postulated that the dip- ping seismic zone marked the loca- tion of a complex “reverse” fault (termed a “megathrust” by Coats, 1962, p. 103) along which the are relatively overrides the ocean basin. According to the sea floor spreading hypothesis advanced by Hess (1962, p. '617) and Dietz (1961), are structures are sites of down-welling mantle-convection currents, and the planar seismic zones dipping beneath them mark the zone of shearing produced by downward-moving material thrust against a less mobile block of the crust and upper mantle. This hy- pothesis is in general consistent with other data suggestive of ac- tive spreading of the sea floor in a northwest-southeast direction away from the East Pacific Rise (Vine, 1966, p. 1412), and with active regional shoreline submer- gence suggestive of a downward— directed component of deforma- tion near the eastern end of the are (p. I 60). Because of the scarcity of standard seismograph stations in Alaska as well as incomplete travel-time data for Alaska earth- quakes prior to 1964, the horizon- tal and vertical distribution of earthquakes could not be defined precisely enough to resolve details of structure within the broad seis- mic zone either at depth or later- ally along it. As a result, large un- certainties exist regarding the ALASKA EARTHQUAKE, MARCH 27, 19'64 I46 wmmhmz z_ mmbohzoo sz<2m3w _ _. A $352de com com 03 _ _ _ mm:___>_ com com OOH 1%ij my», 147 TE CTONICS $835093 53%: 2a ncflowm map 5 9593 9585? 95 was 3th Hauwio «o $25.35 23 "32. 525w Eamwosow .m.D 3:33:95 Scum can MEN .m .m E SE53 no gas 033an ”.333an a :85 co Econ“ 3386 .55 unwfiogfi £8250 Q23 3: «5 mm 333 do 53°33 ”:58 -033me 63535 dag? finaneofifism no ”#83me 3.555% and 33: Moon 382% 9:53am 530% 183.5; 63:83 93 mg 8:803 vwfifigwgmvludfi .893 2:8 Saga d .83 38m :25 5:2 ahme £5588 EyewiaEkBqu :H .muamfi d 6me 813m Am aomnm .355 $520 Kama d .mmad hzmvfifiptmb d .NSC Sax 93 :3th $me Nucahwxwwmfivmhng ER :oEEwm ”€35 “ENS—Aw .am 53 hmfiflmfiac 5%: 23 558mm: 3:89 GawEgoE «mg 236on aawfiaqu $28M .m xflwofi-a5=om *w $me $53k hwfimiaaum vua hum :caudm vino €30 £3-52 2:5 3. Gnmfigofi 2820mm peas Lofifimxmv d Kmmmc $850 «Eu .8sz 238—23 mama am‘fimwsso .N axon; copnfinewtwwnb D $32 3:85? ofiEEU 52 Emmam : \ \ \. mam 53m .m nsmtx 93.5% 333.33 98:3 53:3 SEEN :ESQ *m 932 wovvwn 3883a 3pr v2.» 3302mm sfimwaim *2 850m 3mm— :Sam 02 aQwEE/ofi 013mm: 333?: mmimamd 23:3. ”$05305 wsmooBm 550ch agoEgoE 9502on 5595 HE? miss.“ 6:“ BBS 5.32 unefifl? RENEE 333. :..»:aa .5334 mwuwomvi xmioumlw 338 @0352 33832 BRA 2:2 and SummoumEUm @5353 «o ‘wvmbm MEBonw was: 6:95. /)/ //) uEm SSosfifiSefi 33.339. Se :3 BE .50 “Ewggefimww N323 2.339. 33%5 35.2% .3th5 9.33. vuxmsQ £53 wsmnamu bnwwam I I II II Inluhl Al :33 3.3:: 23qu $52 :vaO SSE SENS fie Swagfisw 635:: 32%: 3:qu £35 wm3>wu .5 $3.59 Ia. dlqlalal r1 dldlqlql 333:3 S Ethfim 93:8 @mfimaQ £32.33 SEQ wwcmmman wwvgufi~ 33:8 maafimxoanaw $33.“? SENSE 22%; Ssgsa SEEN mxoe vmvvwn 33280 3ka 223%. EEKSR 323. 253:3 .33an mxuoh vovvwn 33280 BRA 80:33», 255 J86 .8 353 Mo mxoe wimgaxw oEwwcfiw ZO_._.200 Plafker and Rubin (1967). Rooted stomp on submerged beach. oun . 35 LJ-0033 Copper River Delta area __________ 860d:50 Relmm' 'tz (1966, p. 85) ..... Raoteditstump on deltaic and marsh e s s. 36 LJ—0034 ..... do _____________________________ l, 700:}:100 ~22. 0 ______________________ do ..................... B?)- 1 Approximate. 1960). Although surface breakage may have occurred along faults during other earthquakes in Alas- ka, such features could easily have gone undetected had they occurred in the vast uninhabited parts of the State. The sense of late Cenozoic dis- placement on the faults shown in figure 29 is given where known. In general, faults in south-central Alaska that trend northeast or cast are predominantly overthrusts or oblique overthrusts that dip northward at moderate to steep angles, with north sides relatively upthrown. Horizontal movements on major longitudinal faults are largely restricted to the northwest- trending F airweather fault and ’00 the northwest and east-west trend— ing parts of the Denali fault sys- tem, both of which are predomi- nantly right-lateral. Geologic rela- tionships across many of the faults suggest that they have undergone recurrent movement during Ceno- zoic and much of Mesozoic time and that the sense of displacement along some of them has changed with time. The orientation and the Quaternary displacement on these faults reflects tangential compres- sion oriented north-south to north- west-southeast along the margin of the Gulf of Alaska, perhaps rotat- ing to a more nearly cast—west direction in central Alaska. The broad fault pattern implies that the Pacific Ocean basin is moving north to northwest relative to the mainland. Thus, the ocean basin appears to be shearing relatively past the mainland of British Co— lumbia and southeastern Alaska along the system of northwest- trending right-lateral strike-slip faults, whereas it shears relatively beneath the continental margin and Aleutian Arc in central and western Alaska along a system of imbricate thrust (underthrust ? ) faults. THE PREEARTHQUAKE HOLOCENE (RECENT) RECORD OF VERTICAL SHORELINE MOVEMENTS Numerous records of preearth- quake differential shoreline move- ments relative to sea level provide data on the history of vertical tec- tonic movements during Holocene time. Reconnaissance studies of the displaced shorelines have brought out (1) a general similarity be- tween the pattern of earthquake displacements and the long-term trend of Holocene coastal emer- gence or submergence, and (2) a remarkable recent widespread su'b- mergence over much of the zone that was uplifted during the 1964 ’ earthquake and over at least part of the zone that subsided. In addi- tion, radiocarbon dating of mate- rial from coastal sites has provided quantitative data on the duration and rates of these Holocene move- ments. Data and interpretations pre- sented in this section are largely taken from a preliminary paper by Pla’fker and Rubin (1967). All radiocarbon-dated samples re- ferred to, sources, and other perti— nent data are listed in table 5; sample locations are shown on figure 33. As used herein, the terms “up- lift” and “subsidence” describe a tectonic rise or fall of the land; “emergence” and “submergence” indicate relative movements that may be the sum of both tectonic TE CTONICS movements and eustatic sea-level changes. Tectonic movements in— clude those that result from both diastrophic and isostatic processes. “Long—term” refers to shoreline changes relative to sea level having durations measured in thousands of years ; “short-term” refers to the general submergence that occurred during the 1,350 years or less prior to the earthquake. Some of the observed recent sub- mergence in areas shown on figure 33 is undoubtedly exaggerated by the local vibration-induced com- paction or slumping of uncon- solidated deposits during the earthquakes that frequently rock this seismically active region. As was demonstrated during the 1964 earthquake, surficial submergence may be substantial in areas of thick unconsolidated saturated de- posits—-especially in those deposits that are not constrained on one or more sides, such as deltas, spits, and barrier beaches. Consequently, rooted tree stumps, on bedrock or on thin beach deposits overlying bedrock, were used to determine amounts and rates of pre-1964 vertical displacement wherever possible. In the absence of a reliable local eustatic sea-level record for south- ern Alaska, it is assumed that Holocene sea levels were probably comparable to those in more tec- tonically stable parts of the world where the sea-level record has been worked out in some detail. Three recent interpretations of eustatic sea levels are shown on figure 34. Most of these studies suggest either (1) a rather rapid rise in sea level at an average rate of about 0.08 inch per year until it reached approximately its present level between 2,000 and 6,000 years ago (Coleman and Smith, 1964), or (2) that sea level has been ris— ing slowly and continuously from about —33 feet to its present level I55 during the past 7,000 years with a generally rising, but fluctuating, sea level between about 15,000 and 7,000 years B.P. (Curray, 1961). Some authorities believe that sea level reached its present position from 3,000 to 5,000 years ago and has been fluctuating above and be— low its present position by about 3—6 feet ever since (F airbridge, 1961). For present purposes, it is significant to note that none of these studies suggest that eustatic sea-level fluctuations during the last 10,000 years were large enough alone to explain the relative shore- line displacements found along the coast of south-central Alaska. LONG-TERM HOLOCENE EMERGENCE AND SUBMERGENCE The record of Holocene dis- placements in the area affected by the earthquake, as deduced from shoreline morphology and from radiocarbon dates, is outlined in the following section. Places re- ferred to and the spatial distribu- tion of radiocarbon-dated shore- line samples in relation to the 1964 deformation are shown in figure 33. The age and pre-1964 position of these samples relative to sea level is shown in figure 34. Much of the shore in the moun- tainous southern part of the area that subsided during the earth- quake exhibits the characteristic features of a deeply drowned coast. Submerged shorelines occur along most of Kodiak and the ad- jacent islands, the southern Kenai Peninsula, the south shore of Kachemak Bay, and part of Turn- again Arm. The northern and east- ern parts of the Prince William Sound region also appear to be submergent, although the evidence for relative shoreline movements is somewhat obscured by recent glaciation in most of the fiords. In these submerged mountainous areas, former rocky ridges are now peninsulas or islands, and I56 ALASKA EARTHQUAKE, MARCH 27, 1964 156° 154° 152° 150° 148° 146° 144° 142° 62° ’ r I ALKEETNA MOUNTAINS I l \ é” ,\\ f a x 1 Q‘ I l {3 ’ a: l r «é V II Cl v ' I Q TuaBedni ,3) “‘1 ~ . $3 . a1; 1 ' , lit 9. ‘ - ' u x 60° - . »/ Illv f4‘?‘ :4 .- 2200‘,» i2 " ,: Yakataga . '130~200 [OD 1&‘300 hinbrook Del ‘\‘ 4% Chzngxa / C.) Qdfifié 19 island, Kat Sf. ‘ ‘ e Suckling i; / 1" Mont’ague I ‘0 Q IaO—6O ‘9‘ Island :22: 28,29, 0 0\/’2/ $0 0 Yum" 7'18 Middleton ,r’ 2 Q”? ugustizleand 0 Island 0’ 100 ZOleland ’1 ‘ .32 gW Kamishak Boy I , P» ‘ 1 09 Q ’I b e I l 0K 8° 0 5 \(1‘0 0° 9 EXPLANATION 3 \\\\\\\ > Minimum area of short-term submer- gence . 5 Location of radiocarbon-dated sample used in determining the relative ages of shoreline features. Pertinent data relating to the samples are given in table 5 X 550 _ Emergent shoreline of Pleistocene or 100 150 WLES early Holocene age _ a I Number indicates approximate amount of 0 50 100 150 KILOMETERS uplift in feet _________ o—_-______ SUBMARINE CONTOURS IN FEET Zero isobase of vertical tectonic dis- placements associated With the 1964 earthquake 1 5 1 c l 1 33.—Areal distribution of vertical displacements associated with the 1964 earthquake, areas of preearthquake submergence, and locations of radiocarbon-dated samples used to determine relative ages of shoreline features. TECTONICS I 57 200 I I | | I I I — 60 180 — 26 160 — — 50 140 — 33 — 40 120 r- U) E E Lu I— u. In Z E —- E 8 100 — u; z o ‘5 -* 30 Z c: O ‘53 E Sample from zone of 1964 uplift E ”J E 80 — Cl I.I.I Sample from zone of 1964 subsidence A Sample from area of no change in 1964 _ 20 60 _ . 32 Vertical bars indicate approximate range of uncertainty in elevation relative to sea level. Number refers to sample listed in table 5 40 — 27 - 10 20 - 31 020 030 E < <1» -‘ In E 8 > >< -—I E o u} 0 D: Q: E _ gum ‘TFairbridge (1961) < LII ~\' LII (z) \.._ ~~~ O a) m E —20 \\ 5 ‘1 3 Lu ‘~~‘/Coleman and Smith (1964) <3 ti I.I.I LL ' \, ‘~~ 0: Lu 2 E a... ~~‘~ _ 10 I; E g Curray(1961)>‘~.,_ ~‘~~‘ m E a) .. s D _40 l l I | i | | J I (I) 0 1 2 3 4 5 6 7 8 9 10 TIME, IN MILLENNIA BEFORE 1964 34.—Pre-1964 positions of radiocarbondated samples relative to the Dre-1964 sea level and proposed eustatic sea levels of Curray (1961), Fail-bridge (1961), and Coleman and Smith (1964). The duration of the inferred period of short-term subsidence is indicated by the stippled area. Sample locations at Yukon Island (A), Girdwood (B), Middleton Island (0), Katalla (D), and KukIak Bay (E) are connected by solid lines. I58 ALASKA EARTHQUAKE, MARCH 27, 19'64 35.—Deeply submerged coast along the southern Kenai Peninsula. Embayments, which give the shore a distinctive scalloped appearance, are cirques whose floors are drowned to depths of as much as 300 feet below sea level. Shoreline shown was submerged an additional 2—5 feet in 1964. Photograph by T. L. Péwé. drowned river valleys or glacial cirques have become embayments (fig. 35). The general scarcity of well—developed sea cliffs, beaches, and similar shore features attests to the recency of the submergence. Elsewhere in the subsided zone, coastal bogs of terrestrial peat and some aboriginal dwelling sites that are now inundated by high tides also indicate long-continued sub- mergence relative to sea level. The most pronounced submer- gence appears to be in the vicinity of the axis of maximum subsidence during the 1964 earthquake—a region which roughly coincides with the crest of the Kenai and Kodiak Mountains (pl. 1). Al- though the absolute maximum amount of postglacial submer- gence cannot be determined, an indication of it is provided by dif— ferences in the altitudes of cirque floors of probable Wisconsin age along this part of the coast which presumably were formed at a fairly uniform level. The lowest cirque floors along the outer coast of Prince William Sound and in most of the Kodiak Island group— areas away from the region of maximum submergence—lie at al- titudes ranging from 800 to 1,000 feet above sea level, but cirque floors along the south coast of the Kenai Peninsula range in altitude- from 300 feet below sea level to 800 feet above sea level. This dif- ference in cirque levels suggests at least 300, and perhaps as much as several hundred feet, of submer- gence in the Kenai Peninsula area. In contrast to the zone of sub- sidence, the coast in those areas where the land has risen relative to the sea is generally smoother in outline and commonly exhibits, among other features, one or more wave-cut terraces or uplifted beaches rising to elevations of at least 200 feet (figs. 36, 37). In the major zone of uplift such features are characteristic of the points and capes on the seaward side of Ko- diak Island, the islands of the southern and eastern Prince Wil- liam Sound region, much of the mainland coast east of Cape Suck- ling, and Middleton Island on the Continental Shelf. Comparable emergent shores with postglacial terraces as much as 1,700 feet high occur all along the mainland coast to the east of the area that was affected by the 1964 earthquake. Relatively stable or emergent shores occur along parts of the Cook Inlet and Shelikof Strait coasts in areas that either subsided slightly, remained unchanged, or were slightly uplifted during the 1964 earthquake (fig. 33). Many of these emergent shorelines are probably pre-Holocene features related to high eustatic sea levels rather than to tectonic movements (Hopkins and others, 1965, p. 1113; Karlstrom, 1964, p. 34—37). The record of long-term Holo- cene deformation within the ma- jor zone of uplift in the area be- tween the Copper River Delta and Cape Suckling is seemingly anom- alous in that uplifted surfaces de- scribed as marine terraces, and drowned forests or Sphagnum-peat horizons, occur in close association with one another. Dated marine terraces at Ka-talla (nos. 26, 27, fig. 34) and Cape Suckling (no. 30, fig. 34) record net Holocene emergence; recent net submer- gence is indicated by (1) a dated wood sample from the Copper River Delta (no. 36, fig. 34) that was submerged 22 feet in the 1,700 years prior to the earthquake but was uplifted only about 6 feet at the time of the earthquake, (2) a widespread horizon of terrestrial peat of unknown age that was penetrated in borings at depths as much as 40 feet below sea level in the lower Copper River Delta and to a depth of 30 feet at Bering Lake (Tarr and Martin, 1914, p. 462—463), and (3) forest horizons in the Katalla and Cape Suckling areas submerged prior to the earth- quake by amounts that were con- siderably larger than the earth- quake-related uplift at these same TE CTONICS 36.——Muskeg-covered preearthquake marine terrace on Middleton Island at an alti- tude of 110—125 feet. It is one of five uplifted terraces on the island, and surf-cut rock platform exposed between the base of the sea cliiT and the new high-tide level is a sixth terrace formed by uplift of about 11 feet in 1964. White specks are seagulls. Photograph taken near 7-foot tide stage, April 4, 1964. 37.—The linear tree-covered beach ridge in this View is one of nine elevated beach ridges near Katalla, east of the Copper River Delta. Uplift of about 9 feet in 1964 shifted the shoreline several hundred feet seaward where another beach ridge is in process of formation. Photograph taken near zero tide stage, July 28, 1964. localities. The relative ages of dated samples in these areas sug- gest that the long-term displace- ment may have reversed direction from uplift to submergence during the time interval between about 3,770 and 1,700 years B.P. Radiocarbon dating of organic material from terraces, peat bogs, and archaeological sites in coastal areas afiected by the earthquake has provided some preliminary data on average long-term dis- placement trends relative to sea level at a number of localities. Dis- placement-time curves at the five I59 localities (A—E) in the area for which multiple samples are avail- able are shown in figure 34. Points on the graph are the preearthquake position relative to mean lower low water plotted against age in mil— lenia of radiocarbon-dated ma- terial from these five sites. The available dated samples were col- lected by several different workers over a period of 30 years and were analyzed in three different labora- tories. In spite of the small num- ber of samples at each site, uncer- tainties in their exact positions relative to sea level, and the ever- present problem of analytical or sampling errors, it is noteworthy that the results appear to be re- markably consistent with one an- other and with the displacements that occurred in 1964. These data indicate that the two localities at Yukon Island (A) and Girdwood (B) have subsided rela- tive to present sea level at an aver~ age rate of about 0.7—1.0 foot per century during the time interval from 2,800 to 700 years B.P., whereas those at Middleton Island (0) and Katalla (D) in the up- lifted area have risen at the much greater average rate of at least 3.3 feet per century between 7,650 and 1,350 years B.P. The rate of uplift would be increased somewhat if any of the deduced eustatic sea levels are used in place of the 1964 sea level. The shore at Kukak Bay (E), which was not afi'ected by the earthquake, has apparently under- gone no detectable net change in its present position relative to sea level since at least 1,450 B.P. Detailed study of the emergent shorelines shows that the long- term vertical movements occurred, at least in part, as a series of up- ward pulses separated by inter- vals of stability or even gradual submergence. Evidence for peri- odic uplift is exceptionally well displayed on Middleton Island I60 where the surface consists of a flight of five gently sloping and well-developed marine terraces separated by wave-cut cliffs or rises ranging from 20 to 30 feet each (Miller, 1953; Plafker, un- pub. data, 1963, 1964, 1965). The highest terrace, forming the flat central part of the island, has a general altitude of 130—165 feet. Sudden uplift of 10—13 feet dur- ing the 1964 earthquake, which ex— posed a sixth surf-cut terrace (shown on fig. 36), suggests that perhaps the earlier terraces may also have been exposed by a series of upward pulses during previous earthquakes. Radiocarbon-dated driftwood from the highest and lowest of these preearthquake plat— forms.on Middleton Island dates initial uplift of the island at 4,4‘70i250 years ago and uplift of the lowest preearthquake terrace at approximately 1,350:200 years B.P. (nos. 33, 31, table 5). Thus, the extrapolated average time in- terval between successive uplifts of the three intervening terraces would be close to 800 years. Because most of the dated samples were taken from deposits laid down when the postglacial eustatic sea level was rising grad- ually or was at about its present stand, the emergent shorelines probably result almost entirely from tectonic movements. Although some slightly submerged shorelines may be attributed to the effects of a late Holocene rise of sea level and to local surficial sub- sidence of unconsolidated deposits, in most instances the submergence is so large when compared to deduced eustatic curves that it must also be at least partly due to tectonic subsidence. Regional submergence of the Kenai and Kodiak Mountains is clearly anomalous in that the movements are in a direction that tends to reduce the elevation of ALASKA EARTHQUAKE, MARCH 27, 19'6-4 these youthful mountain ranges. Furthermore, because the net load of glacial ice on the mountains has been steadily diminished since the Pleistocene, isostatic adjustments, if any, should be upward. Thus, the submergence is a counter- isostatic process that can only be attributed to regional diastrophic movements. Less certain is the extent to which the regional postglacial up- lift reflects isostatic compensation resulting from unloading of ice and the extent to which it is caused by diastrophic movements. Both processes are undoubtedly in- volved. However, the known seis- mic activity of the region, the evidence for pulsating rather than continuous emergence, and the ap- parent late Holocene reversals in the sense of the net displacements over part of the region suggest that the movements were also, to a large extent, diastrophic in nature. SHORT-TERM TECI‘ONIC SUBMERGENCE Much of the coast in the region affected by tectonic movements during the 1964 earthquake had experienced a pronounced submer- gence for several centuries prior to that event. This phenomenon was probably first recorded by the great English navigator, George Vancouver, who explored Prince William Sound between May 25 and June 17, 1794 (Vancouver, 1801, v. 5, p. 335—336). He noted that on northern Montague Island “The shores are in general low, and as had been already observed, very swampy in many places, on which the sea appears to be making more rapid encroachments than I ever before saw or heard of * * * [trees along the shore] were re- duced to naked, dead white stumps, by the encroachment of the sea water to their roots; and some stumps of trees, with their roots still fast in the ground, were also found in no very advanced state of decay nearly as low down as the low water of spring tides.” The fact that low water of spring tides is about 13 feet below the normal lower growth limit of trees in this area indicates almost that amount of submergence prior to 1794. The characteristic appearance of these drowned forests in Prince William Sound and along the coast at Cape Suckling east of the Sound is illus- trated by figures 38 and 39. Evidence for active submergence along the coast of the type first described by Vancouver was sub- sequently noted by geologists, archaeologists, and botanists in the following places: (1) the Con— troller Bay area east of Prince William Sound (Tarr and Martin, 1914), (2) the Copper River re- gion (Sc‘hrader, 1900, p. 404; Reimnitz, 19616, p. 112—125), (3) around Prince William Sound (Grant and Higgins, 1910; Moffitt, 1954; Dachnowski-Stokes, 1941; De Laguna, 1956, (4) around Kachemak Bay on the southern Kenai Peninsula (De Laguna, 1934) , (5) in the upper Cook Inlet region ('Karlstrom, 1964, p. 48), and (6) in at least two localities on Kodiak Island (Clark, 1966). Reconnaissance studies by the Geological Survey of all these shorelines after the 1964 earth- quake suggest that the submer- gence observed is a regional phe— nomenon that has affected much, if not all, of the Prince William Sound region, the mainland coast and islands east of Prince William Sound, the south coast of the Kenai Peninsula at least as far as Kachemak Bay, parts of Turn- again Arm, and segments of the southeastern coast of the Kodiak group of islands. Undoubtedly, many more such localities could be found by detailed examination of the shorelines. Areas within TEC’I‘ONICS 38.——Bleached trunks of spruce trees on Latouche Island, Prince William Sound, killed by salt—water immersion and partially buried in beach gravel as a result of about 8 feet of submergence below preearthquake extreme high-tide leveL The locality was exposed by 8 feet of uplift in 1964. Photograph taken May 28, 1964. 39,—Spruce tree stumps (foreground) rooted in a thin layer of peat on a surf-cut bedrock surface about 14 feet below preearthquake extreme high water (indicated by the top of the line of driftwood below the present forest edge in the background). Radiomrbon age of a stump near the base of the stadira rod was 710:200 years (no. 29, fig. 34). These stumps were exposed by about 16 feet of uplift in 1964. Photograph taken July 24, 196/}. I61 which there is evidence of pre- earthquake submergence are indi- cated on figure 33 and the ages and positions relative to sea level of the dated samples are shown on figure 34. In a few scattered locali— ties within Prince William Sound, evidence of stable shorelines or of possible recent slight uplift (no more than 4 ft) was found. This evidence suggests that small areas may have acted as independent tectonic units that did not take part in the general submergence. The available data is insufficient for determining whether the sub- mergence, which affected the off- shore islands from Kayak Island to the Copper River Delta, ex- tended as far out on the Conti- nental Shelf as Middleton Island. Preearthquake sea levels that reached to the base of the prom- inent sea cliff encircling much of the island indicate either a long period of relative stability and erosion or some submergence since the last uplift of the island roughly 1,350 years ago. The record of preearthquake submergence in the zone that was lowered in relation to sea level is much less complete than that for the zone that was raised, mainly because much of the evidence is now below lowest low tide. Avail— able data suggest that submer- gence occurred along parts of the coast of the Kenai Peninsula and the Kodiak group of islands but that submergence probably was much less than in the area from Prince William Sound to Cape Suckling. No change in mean sea level attributable to tectonic move— ments was detected in a 21-year tidal record at Seward and a 15- year record at Kodiak. An appar- ently large preearthquake sub~ mergence on the delta 0f Spruce Creek near Seward of about 6.9 feet in less than 200 years, as indi- cated by drowned rooted tree I62 stumps (no. 12, fig. 34), may be partly due to slumping and (or) compaction of deltaic deposits. Radiocarbon dates f r o m drowned terrestrial plants that were probably killed by sea-water encroachment provide data that permit a crude estimate of the du- ration and average rate of the short-term preearthquake sub- mergence. Ages in radiocarbon years versus amount of submer- gence of samples taken from the most deeply submerged terrestrial vegetation at each site are plotted in the stippled area on figure 34. The fairly regular linear distribu- tion of 14- out of 16 samples, all from the area between Seward and Cape Suckling, suggests a general submergence at a rate that aver- aged roughly 1.7 feet per century for about 930 years. On the other hand, positions of two samples (nos. 15, 21, fig. 34) from this same general area, both of which are older than 930 years, reflect lower rates of submergence, as do three samples from the Cook Inlet re- gion (nos. 8, 11, 15, fig. 34). The maximum indicated average rate of submergence is 8.6 feet in 230 years or about 3.7 feet per century on Latouche Island in Prince William Sound. Absence of historic records of sudden earthquake-related rela- tive sea-level changes and of geo- morphic evidence for such move- ments suggest that the submer- gence was probably gradual or that it occurred in numerous small in— crements over a long period of time. This time interval, as in— ferred from dated submerged shorelines, was at least 930 years in the Prince William Sound ‘ region. The upper limit for the du- ration of the submergence is far less certain; if it corresponds to the oldest dated submerged-forest sample along the coast (no. 21, fig. ‘ 34) and the time of uplift of the ALASKA EARTHQUAKE, MARCH 27, 19614 lowest preearthquake marine ter— race on Middleton Island (no. 31, fig. 34), it could be as much as 1,350 years. Short—term submergence oc- curred at a time when sea level was at or near its present stand and in a region where overall isostatic displacements in response to gla- cial unloading should have been upward. Isostatic adjustments be- tween major earthquakes such as the 1964 event could conceivably be responsible for some of the sub— mergence in areas of tectonic up- lift. However, such adjustments are inadequate to explain either short-term submergence in the Copper River Delta-Cape Suck- ling area that exceeds earthquake- related uplift or submergence in areas of the Kenai Peninsula and Kodiak Island that subsided both before and during the earthquake. By implication, this fact suggests that the submergence was caused at least partly by di‘astrop‘hism that involved a significant down— wardadirected component of re— gional strain of variable amount and rate over much of the coastal belt affected by vertical displace- ments during the 1964 earthquake. TECTONIC IMPLICATIONS OF THE RECORD OF HlOLOCElNE. VERTICAL MOVEMENTS The history of Holocene verti- cal movements in the region af— fected by the 1964 earthquake, and in nearby areas, is a fragmentary one based largely on a rapid re— connaissance after the earthquake and on incidental investigations by others. Additional work—in- volving far more radiocarbon dait- ing—is required to obtain detailed data on vertical movements at se- lected localities in areas of net Holocene emergence and submer— gence. However, the following four tentative conclusions regard- ing the prehistoric tectonic move— ments seem to be indicated by the available data: 1. Areas of net Holocene emer- gence or submergence broadly correspond with those areas in which significant amounts of uplift and subsidence occurred during the 1964 earthquake. Thus, the tectonic movements that accompanied the earth- quake were apparently but one pulse in a long-continuing trend of deformation. This trend has resulted in regional emergence of parts of the continental mar- gin, simultaneous submergence of the Kenai-Kodiak Mountains belt, and either relative stability or emergence along the shores of Cook Inlet and parts of She- likof Strait. 2. The amounts of net long-term Holocene emergence and sub- mergence of the coast, which are locally considerably larger than the postulated eustatic sea-level changes for the same time inter- val, indicate that difl’erential displacement of the shoreline results largely from tectonic movements. Progressive sub— mergence of youthful moun- tains recently unloaded of ice and pulsating emergence of the Continental Shelf and Gulf of Alaska coast are suggestive of a dominant regional diastrophic deformation. 3. ‘Stepl‘ike flights of Holocene ma- rine surf—cut terraces at a num— ber of localities along the coast suggest that the long-term ver- tical movements occurred as a series of upward pulses that were separated by intervals of stability or even gradual sub- sidence that locally average 800 years or more in duration. These upward pulses probably repre- sent earthquake-related move- ments comparable in origin to those which affected parts of the same coast in 1964. 4. Gradual tectonic submergence prevailed during at least the past 900 years, and perhaps as long as 1,360 years, over much of the zone that was uplifted and over at least part of the zone TE CTONICS that subsided during the 1964 earthquake. This widespread submergence is tentatively in- terpreted as direct evidence for a significant downward- directed component of regional 163 strain preceding the earthquake and its duration as the approxi- mate time interval since the last major tectonic earthquake in this same region. MECHANISM OF THE EARTHQUAKE GENERAL CONSIDERATIONS According to the classic elastic rebound theory of earthquake gen- eration (Reid, 1911), which is gen— erally accepted by western geolo- gists and geophysicists, shallow earthquakes are generated by sud- den fracture or faulting follow— ing a period of slow deforma- tion during which energy is stored in the form of elastic strain within rock adjacent to the fault. When the strength of the rock is ex- ceeded, failure in the form of faulting occurs, the material on op- posite sides of the fault tends to rebound into a configuration of elastic equilibrium, and elastic strain potential is released in the form of heat, crushing of rock, and seismic-wave radiation. The drop in elastic strain potential is possibly augmented or partially absorbed by net changes in gravi- tational potential associated with vertical tectonic displacements. Field investigations demon- strate that there is little likelihood that the primary fault along which the 1964 earthquake occurred is ex- posed at the surface on land, nor is there evidence for movement on any of the known major continen- tal faults. If the earthquake origi- nated by rupture along one or more faults, the two most plausible models for the orientation and sense of movement on the primary fault consistent with the available fault-plane solutions and disloca- tion-theory analyses of the re- Aleutian Volcanic Arc (projected) h=20—50 krn V [Slight uplif‘tj Major zone of subsidence | l MILES °l 50- \ —/l\ l/ - j _ \ 100 —— 150 “ \ 7‘74“ P l l \ 9 Main shock (projected) Aleutian Trench axis Subsid— Major zone of up_|i_ft____ ence? KlLOMETERS MONTAGUE ISLAND FAULTS /’ o 7? , _// 100 150 200 O 50 O 50 \ 100 MILES 100 KlLOMETFRS |-’|__J 40,—Diagrammatic section showing two alternative interpretations of the 1964 Alaska earthquake based on focal mechanism studies (S. T. Algermissen, written mmmun., March 1965) and analyses of residual vertical displacements by dislocation theory (Savage and Hastie, 1966). Section is oriented through the focus of the main shock and roughly normal to the regional structural trend. P and T axes are the approxi- mate orientations of the principal pressure and tension axes, respectively, at the hypoeenter of the main shock indicated by the focal mechanism studies. sidual vertical displacements are (1) relative seaward thrusting along a fault that dips northwest- ward beneath the continental mar— gin at a low angle, and (2) dip- slip movement on a near-vertical fault, the ocean side being rela- tively upthrown, that strikes ap- proximately along the zero isobase between the major zones of uplift and subsidence (fig. 40) . These two models, which are discussed in the following sections, are referred to as the “thrust-fault model” and the “steep-fault model.” Whether the postulated shear- ing resulted from the overcoming of frictional resistance to sliding, as set forth in the elastic rebound theory of Reid (1911), or from some other process such as brittle fracture, creep instability, or prop~ agation of flaws cannot be ascer— tained from available data. Sev- eral writers (0 r o W a n, 1960; Griggs and Hamlin, 1960; Evison, 1963, p. 863—884) have pointed out I64 deficiencies in the elastic rebound mechanism for earthquakes at depths of more than a few miles, where frictional stress on dry fault planes must vastly exceed the rock strength. However, such consid- erations generally do not take into account the effect of pore fluids which must exist in the real earth down to the depths of the deepest hydrated mineral phases. Recent laboratory studies (Raleigh and Patterson, 1965, p. 3977—3978; Griggs and Blacic, 1965; Raleigh, 1967) suggest the possibility that the Reid mechanism may extend to depths at least as great as the lower crust and upper mantle in regions where frictional resistance to slid- ing on faults may be reduced by local anomalously high pore pres— sures or where the strength of the rock is lowered sufficiently in the presence of pore fluids to permit brittle fracture. An alternative mechanism to be considered is one that explains tec— tonic earthquakes as originating from sudden expansion and (or) contraction of large volumes of rock due to rapid phase changes (Evison, 1963). According to this concept, faulting is the result rather than the cause of earth- quakes. No evidence has yet been found from either natural or ex- perimental petrologic systems to support the idea that reconstruc- tive solid-solid or solid-liquid re- actions can occur fast enough throughout sufficiently large vol- umes of rock to generate major earthquakes or to produce regional vertical displacements (Ghent, 1965). Aside from the theoretical objections, it is difficult to conceive of a reasonable combination of equidimensional volume changes that could cause the observed pat- tern of vertical and horizontal dis- placements that accompanied this earthquake. For these reasons, there is little likelihood that phase ALASKA EARTHQUAKE, MARCH 27, 1964 changes were a primary mecha- nism for the 1964 earthquake, and this possibility will not be pursued further. THRUST-FAULT MODEL Most of the preliminary data available by the end of the 19.64 field season suggested that the earthquake and the associated tec— tonic deformation resulted from a relative seaward thrusting along the continental margin that was accompanied by elastic horizontal extension behind the thrust block (Plafker, 1965, p. 1686). The thrust—fault model has been strongly reinforced by data subse- quently obtained from (1) more detailed fieldwork on the surface displacements and preearthquake movements by the Geological Sur- vey during 1965, (2) comparison of pre- and post-earthquake trian- gulation surveys for a large seg— ment of the deformed region (Parkin, 1966, fig. 4), and (3) de- tailed focal-mechanism studies of the main shock and larger after- shocks (Stauder and Bollinger, 1966) . The essential features of the suggested model, which are illus- trated diagrammatically by figure 41, are presented below. OUTLINE OF THE MODEL According to the thrust-fault model, the earthquake resulted from shear fai‘lu-re along an in- ferred major zone of movement, or megathrust, which dips northwest- ward beneath the continental mar- gin from the vicinity of the Aleu- tian Trench, and on subsidiary faults within the overthrust plate. The zone within which movement occurred, as delineated by the belt of major aftershocks, parallels the Aleutian Trench on the northwest for a distance of some 600 miles and is from 110 to 180 miles wide (fig. 2). It lies almost entirely within the major zone of uplift and the portion of the adjacent zone of subsidence that is on the seaward side of the axis of subsidence. The pattern of earthquake-re- lated horizontal displacements (fig. 16) and the evidence for re- gional preearthquake shoreline submergence along the continental margin (fig. 33) suggest that shear failure followed a long period of elastic strain accumulation (930— 1,360 years) during which the up— per plate was horizontally com- pressed roughly normal to the trend of the Aleutian Arc and probably simultaneously depressed relative to sea level. A hypotheti- cal deformation cycle is illustrated diagrammatically in figure 42. The data suggest that the dynamic drive was probably provided by underthrust of the lower plate, with a downward-directed compo- nent of movement. R. K. Hose has pointed out that distortional drag due to an underthrust would have generated a shear couple in which the relative strains and orienta— tion of potential shear planes coin- cide reasonably well with those deduced from fault-plane solu- tions for the main shock and many of the larger aftershocks (oral commun., Nov. 1965). Alternative mechanisms such as movement of the continental margin over the oceanic basin, or almost horizontal movements, could not be respon- sible for the combination of both horizontal shortening and re- gional submergence indicated by the available data. Previous seismicity in the re- gion during the strain—accumula- tion phase of the 1964 earthquake may have resulted from adjust- ments within the strained volume of rock involving predominantly lateral offsets or relatively local dip-slip displacement. This pos— sibility is suggested by the fact that, although prior historic earthquakes with magnitudes of 7 or more have been recorded -7. Fault, showing sense and relative amount of displacement Queried where inferred ”Wk—”w Inferred megathrust zone I65 TECTONICS A /,r_- f “‘7' ‘3"V Z / Xi /‘uu r / AL> ‘/ ““my—“nmmiu \‘1 X‘ ' 1 . /' M-discontinuity (after Shor, 1962, fig. 4) .. .— 7... Relative direction and amount of tectonic displacement Queried where inferred * Epicenter of main shock I Epicenter of larger aftershock (M >5.0) and projected hypocenter of after- shock within 25 miles of vertical section gm Active or dormant volcano 41.—Schematic block diagram showing the postulated relationship of the seismic activity and tectonic displacement associated with the 1964 Alaska earthquake to major structural elements of south-central Alaska. Horizontal and vertical scales equal below sea level; vertical scale above sea level exaggerated x 4. Drawn by H. T. Koelbl. (Wood, 1966, p. 24) , none of these were accompanied by known re- gional tectonic deformation—and certainly none with deformation on the scale of the 1964 earthquake. Faulting at the time of the earthquake presumably was ini- tiated at the hypocenter of the main shock in northern Prince William Sound, from which point it propagated simultaneously up— dip towards the Aleutian Trench and along strike both towards the southwest and east within the area encompassed by the aftershocks. Shear failure was accompanied by elastic rebound in the upper plate above the thrust, which resulted in (1) relative seaward displacement and uplift of a part of the con- tinental margin by movement along the inferred megathrust and the subsidiary reverse faults that break through the upper plate to the surface, and (2) simultaneous elastic horizontal extension and vertical attenuation (subsidence) of the crustal slab behind the up- per plate. These movements, pos- sibly in combination with uni- dentified submarine faulting and (or) underthrusting of the lower plate in the opposite direction, re- sulted in the observed and inferred tectonic displacements at the surface. Indicated stress drops at the surface across the zone of subsid— ence (on the order of a few hun- dred bars) are comparable in mag— nitude to those reported for other tectonic earthquakes.- For the idealized case of homogeneous strain and purely elastic distortion of the crust, the stress drop (P) is a function of Young’s modulus (E) of the slab and the reduction of horizontal strain (6) across it: AP=E€ For an average E of 3X105 bars I 66 ALASKA EARTHQUAKE, MARCH 27, 1964 “Aleutian Volcanic Arc PATTON BAY FAULT \ ¢Aleutian Trench O m... -~vnv4b'£u um. “... mm"... \Inferred subcrustal flow dire tion Oceanic crust and mantle xxx: __.i_ Hypothetical strain ellipse showing postulated orientation of principal pressure axes and potential shear planes resulting from a shear couple XL\X:L\A __ Horizontal shortening, 1* Horizontal extension Slight uplift ?———?———’->'(-I. ——-S b' ? I u sldence Subsidence J. Uplift ' 0 100 200 MILES I, | 0 Too 200 KILOMETERS | Jr HORIZONTAL AND VEl‘QTICAL SCALE 42.—Diagrammatic time-sequential cross sections through the crust and upper mantle in the northern part of the region affected by the 1964 earthquake. A, Relatively unstralined condition after the last major earthquake. B, Strain buildup stage during which the continental margin is shortened and d‘ownwarped. 0, Observed and inferred displacements at time of the earthquake during which a segment of the continental margin is thrust seaward relative to the continent. Datum is the upper surface of the crust beneath the cover of water and low-velocity sediments. Vertical displacements at the surface, which are indicated by the profiles and by arrows showing sense and relative amount of movement, are about X 1,000 scale of the figure. (7>< 1011 dynes cm‘z) , and average and maximum strain across the subsided zone of 1.1X10" and 3X10‘4, respectively, the average stress drop was only about 77 bars (77><106 dynes cm”). The maxi- mum drop, which occurs near the axis of subsidence, was about 210 bars (210x 106 dynes cm”). The megathrust in the earth- quake focal region dips north to northwest at an angle of 9°—15°, according to fault-plane solutions of the main shock and several aftershocks, but its precise depth and configuration are uncertain. The limited available data on focal depths of the earthquake (p. I 4) and on deep crustal structure (Shor, 1962, fig. 4; Hales and As- ada, 1966, p. 420) suggest that the megathrust may coincide with the unstable interface at the base of the continental crust beneath the continental slope and shelf. It probably extends into the upper mantle toward the north to some unknown depth (no deeper than the deepest hypocenters in the re- gion ) where stress relaxation is sufliciently rapid to absorb the ap— plied force by plastic, rather than elastic, deformation. MAJOR UNRESOLVED PROBLEMS The cause of the slight uplift that occurred to the north of the major zone of subsidence is uncer- tain. Presumably, this uplift in- volves an elastic distortion related to strain changes resulting from seaward thrusting along the con- tinental margin at the time of the earthquake. It could represent an increase in vertical strain (Poisson bulge) resulting from a sudden in- crement in the regional horizontal compressive strain 0 r i e n t e d roughly normal to the zone (R. O. Burford, written commun, 1966). If this explanation for the uplift is correct, retri‘an'gu'lation within TECTONICS the areas of slight uplift should show transverse horizontal short- ening. A test of the suggestion could be made if the postearth- quake triangulation net were ex- tended northward into this zone. A second major problem is that there exists an apparent asymme- try in the volumes of uplift (89 cu mi, 372 cu km) and subsidence (29 cu mi, 122 cu km) in the two major zones that implies a substantial increase in net gravitational po- tential. By making the assump- tions that the displacements de- crease linearly from a maximum at the surface to zero at the base of the continental crust, that there were no density changes within the affected volume of crust, and that the crust has an average den- sity of 2.7, the approximate upper limit for this change may be cal- culated. An average lowering of 2% feet (75 cm) in a Crustal block having the dimensions of the zone of subsidence would result in a potential energy loss of 2.7 X1019 ft-lbs (3.6><1026 ergs). The in- crease in gravitational potential in the uplifted area would be about 4.1><1019 ft-lbs (5.6><1026 ergs), assuming an average uplift against gravity of roughly 6 feet (1.8 In) for a wedge-shaped block with a length of 475 miles (100 km), a width of 115 miles (185 km), and an average thickness of about 11 miles (18 km). These data suggest an apparent net increase in gravitational po- tential of about 1.5><1019 ft-lbs (2X1026 ergs). Even with the as- sumption that vertical displace-- ments e x t e n (1 only halfway through the slab, the indicated gravitational potential increase is still roughly two orders of magni— tude larger than the total released seismic-wave energy of about 1—2>< 1024 ergs (calculated from the empirically derived Guten- berg-Richter relationship between I67 energy and earthquake magnitude log E = 114+ 1.5 M ; Richter, 1958, p. 366). In fact, however, the net change in gravitational potential due to elastic-rebound mass redistribu- tion in the crust must be con- sidered as indeterminate because data are unavailable on (1) the seaward and continentward limits of the displacement field, (2) pos- sible elastic density changes at— tendant upon stress drops within and behind the upper plate, and (3) the extent to which the move- ments may have been compensated by elastic depression of the denser mantle beneath zones of uplift and a corresponding rise beneath the zone of subsidence. REPRESENTATION BY DISLOCATION THEORY Savage and Hastie (1966) and Stauder and Bollinger (1966) have used dislocation theory to compare theoretical profiles of vertical sur- face displacement for various gently dipping and horizontal overthrust models to the observed and inferred profile. Their as- sumed models, with the corre- sponding profiles along a vertical plane oriented normal to the fault strike, are plotted to a common scale on figure 43. These models have, of necessity, been highly generalized to permit a mathe— matical analysis by dislocation theory, and, in part, the studies are based upon preliminary and incomplete observational data. The basic assumption in this procedure is that the observed dis- placements can be modeled by the corresponding fields of a planar dislocation sheet in a homogeneous, isotropic, elastic half-space with displacement discontinuity match- ing the observed or inferred fault slip. Inasmuch as the depth, con— figuration, and net slip on the postulated faults are to a large de- gree speculative, considerable lati- I68 tude exists in the parameters used for those calculations. Further- more, because faults that break to the surface are not dislocations in a semi-infinite medium, their con- tribution to the deformation can only be examined qualitatively. Figure 43 shows that, although each of the assumed models can approximately account for the ob- served subsidence, none of them gives a close fit between the theo- retical and actual profiles in the uplifted zone. Stauder and B01- linger (1966, p. 5293) suggest that the observed tectonic surface dis- placements can be approximated to any desired degree by assuming combined differential-slip motion and a shallowly dipping thrust plane. Such a model would be more nearly in accord with the data which suggest a dipping master fault having the approximate con- figuration shown in figure 4311 with differential slip as indicated in figure 438. To be realistic, how— ever, it would also have to include the effects of (1) imbrication along known and suspected thrust or re- verse faults that break to the sur- face, (2) possible breaking to the surface along the shallowing lead- ing edge of the megathrust, and, perhaps (3) possible simultaneous underthrusting of the lower plate along an inclined fault plane. STEEP-FAULT MODEL OUTLINE OF THE MODEL According to the steep-fault model, the earthquake resulted from elastic rebound on a near- vertical fault that strikes approxi- mately along the line of zero change in land level, with the southeast block up and the north- west block down relative to sea level. A steeply dipping fault was suggested by a preliminary fault- plane solution based on surface waves (Press and Jackson, 1965, men EARTHQUAKE, MARCH 27, 1964 METERS/FEET 10 100 MILES Cook Inlet 100 MILES 0 200 KILOM’ETERS 3:100 Slip: 10m PATTON BAY F AU LT\ mil-g 7m um 18m > :3 Observed and inferred displacement. 1964 earthquake \. 100 200 KILOMETERS Middleton Island t 100 Aleutian Trench 0 A) KILOM ETERS 0 MILES D" O 50 / 11m 5m lip: 3—18m (as shown) 3mI 4m 7m , 11m Slip: 3-18m horizontal. ‘ 1-2.7m vertical (as shown) 43.—Comparison of observed and inferred profile of surface displacement along line 343’ of plate 1 (shaded line) with computed profiles for movement on faults of slip and cross-sectional configuration shown in diagram below the profiles. Model A from Savage and Hastie (1966, fig. 3) ; models B and 0 from Stauder and Bollinger (1966. fig. 7). p. 867; Press, 1965, p. 2404). How— ever, Savage and Hastie (1966, p. 4899—4900) have pointed out that a unique surface—wave solution for the fault orientation cannot be in- ferred from the surface waves, in— asmuch as the direction of rupture propagation is along the null axis and therefore the radiation of seismic waves will be essentially the same for either of the two pos- sible fault planes. The main appeal of the steep- fault model is that it can readily account for the gross distribution of uplift and subsidence in two major zones, as well as the occur- rence of the earthquake epicenter close to the zero isobase between these zones. However, as noted elsewhere (Plafker, 1965, p. 1686), there are no compelling geologic, seismolog- ic, or geodetic data in support of the steep-fault model. Aftershocks are not grouped along its postu— lated trace, but instead lie mainly in a broad belt along the conti— nental margin mainly to the south of the zero isobase (fig. 2). Fur- thermore no field evidence exists for new surface breakage in the vicinity of the zero isobase—de- spite the fact that it intersects the coast in more than 15 localities (fig. 3). And finally there is no evi- dence that the line corresponds to a major geologic boundary with the seaward side relatively up- thrown, as might be expected if it marked the trace of a major fault TECTONICS METERS/FEET 10 I 69 Observed and inferred displacement. 1964 earthquake \ 100 MILES \fi\ \ 1 ' _ r—’— \ I10 I ' 200 N\:\\0 . 0 100 200 KlLOMETERS \ // \ B>\/ ‘10 Middleton Aleutian Cook Inlet _5 PATTON BAY Island Twitch FA T 100 MILES UL | i200 KlLOMETERS 100 O——O J 150 ll '5”me \\ B. Slip: 10m KILOMETERS u.» 0 0 MILES 150 200 44.—Comparison of observed and inferred profile of surface displacement along line B—B’ of plate 1 (shaded line) with computed profiles for movement on steeply dip- ping faults of slip and cross-sectional configuration shown in the diagram below the profiles. Model A from Press and Jackson (1965, fig. 1) ; model B from Savage and Hastie (1966, fig. 3). along which vertical movement has occurred in the past. On the contrary the overwhelming ma- jority of surface faults that paral- lel the coast in this part of Alaska have exactly the opposite sense of displacement (fig. 29). The apparent absence of a sur- face dislocation along the zero iso- base between the major zones prompted the suggestion that the displacement represents flexure above a near—vertical fault at depth (Press and Jackson, 1965, p. 868; Press, 1965, p. 2405). Such a fault would have to extend to the considerable depth of 62—124 miles (100—200 km) below the free sur- face to account for the areal dis- tribution of residual vertical dis- placements normal to its inferred strike. REPRESENTATION BY DISLOCATION THEORY Observed residual tectonic dis— placements have been compared to the theoretical surface displace- ments that would occur on faults of varying inclination, slip, and dimension by application of dislocation theory (Press and Jackson, 1965; Press, 1965; Savage and Hastie, 1966). These analyses are subject to the same basic as— sumptions as were previously out- lined. Assumed models, with corresponding profiles along a vertical plane oriented normal to the fault strike, are plotted to a common scale on figure 44. The resultant profiles show the same general spatial distribution of up- lift and subsidence as the profile of observed and inferred vertical displacements along a northwest- southeast line through the south west tip of Montague Island. They differ fundamentally, however, in that the axes of uplift and sub- sidence are too close together by at least 50 miles (80 km), or a factor of one-half, and the indi- cated changes Within the two major zones are notably more equal in amplitude than in the observed profile. The distance between the two axes would tend to decrease even further if the faulting is shal- low—as is suggested by the spatial distribution of the aftershocks. Even if the vertical displacements could be explained by these mod- els, no reasonable combination of fault dimensions, slip, and dip could also duplicate the systematic horizontal displacements observed in the field (fig. 18). I70 ALASKA EARTHQUAKE, MARCH 27, 1964 SUMMARY AND CONCLUSIONS The earthquake of March 27, 1964, was accompanied by regional vertical and horizontal displace- ments over an area probably in excess of 110,000 square miles in south—central Alaska. Major defor- mation and related seismic activity were largely limited to an elongate segment of the continental margin lying between the Aleutian Trench and the chain of late Cenozoic volcanoes that comprise the Aleu- tian Volcanic Arc. Geologic evidence in the earth- quake—affected region suggests that the earthquake was but the most recent pulse in an episode of de- formation that probably began in late Pliocene time and has con- tinued intermittently to the pres- ent. The net effect of these movements has been uplift along the Gulf of Alaska coast and con- tinental margin by warping and imbricate faulting; subsidence or relative stability characterized most of the adjacent landward zone extending inland approx- imately to the Aleutian Volcanic Arc. The length of time since the last major tectonic earthquake that involved regional warping, as in- ferred from radiocarbon-dated displaced shorelines in the epicen— tral region, appears to have been , about 930—1,360 years. Because the primary fault or zone of faulting along which the earthquake is presumed to have occurred is not exposed at the sur— face on land, a unique solution for its orientation and sense of slip cannot be made. Nevertheless, the vertical displacements, when con- sidered in relationship to the focal- mechanism studies and spatial dis- tribution of seismicity, strongly favor the hypothesis that the pri- mary fault was a major thrust fault, or megathrust. The data suggest that the segment of the megathrust along which slippage occurred was 550—600 miles long and 110—180 miles wide and that it dips from the vicinity of the Aleutian Trench at a gentle angle beneath the continental margin. According to the thrust-fault hypothesis, the observed regional uplift and transverse shortening along the continental margin dur- ing the earthquake resulted from (1) relative seaward displacement of the upper plate along the dip- ping primary thrust, (2) imbrica- tion on the known subsidiary re- verse faults that break through the upper plate to the surface, and (3) crustal warping. Movement on other subsidiary submarine re- verse faults, as well as rebound of the lower plate toward the conti- nent, may have contributed to the uplift. Simultaneous subsidence in the zone behind the fault block presumably reflects an elastic hori- zontal extension and vertical at- tenuation of previously com- pressed crustal material. Stored elastic—strain energy within the thrust block and the segment of crust behind the block that was af- fected by subsidence was the pri- mary source of energy dissipated during the earthquake in the form of seismic waves, heat, crushing of rock, and, perhaps, net changes in gravitational potential. Major un- resolved problems are the cause of the slight uplift in the area north of the two major zones of defor- mation and the apparent large in- crease in potential energy over the displacement field. The implication of a genetic relationship between the 1964 Alaska earthquake and the Aleu- tian Arc appears inescapable in view of the nature of the surface deformation and seismic activity associated with the earthquake. Available geologic, geodetic, and geophysical data from the region affected by tectonic deformation during the earthquake are com- patible with the concept that are structures are sites of down-well- ing mantle convection currents and that planar seismic zones dip- ping beneath them mark the zone of shearing produced by down- ward-moving material thrust against a less mobile block of the crust and upper mantle. Conver- sely, the data provide severe con- straints for alternative hypotheses that relate deformation in the eastern Aleutian Arc, at least, primarily to transverse regional extension or to movement on major longitudinal strike-slip faults or shallow steep dip-slip faults. Preearthquake strain directed at a gentle angle downward beneath the arc in the epicentral region is suggested by geodetic evidence for horizontal shortening of as much as 64 feet roughly normal to the Gulf of Alaska coast and by geo- logic evidence for progressive coastal submergence in the same region in excess of 16 feet. If the duration of preearthquake strain accumulation, as inferred from radiocarbon-dated downed shore— lines, was about 930—1,360 years, average annual horizontal dis— placement was-about 0.59—0.83 inch per year (1.5 to 2.1 cm per year). The indicated shortening, which is probably a minimum value inas- much as there probably also was some permanent. shortening, is reasonably compatible with sea- floor spreading rates of about 2.9 cm per year in a northwesterly direction from the Juan de Fuca ridge. Such rates and spreading directions have been deduced from paleomagnetic studies of the north- eastern Pacific ocean floor (Vine, 1966, p. 1407). Alternative driving mechanisms, in which the upper Alaska Department of Fish and Game, 1965, Post-earthquake fisheries evaluation; an interim report on the March 1964 earthquake efiects on Alaska’s fishery resources: Juneau, Alaska, 72 p. Algermissen, S. T., 1965, Prince William Sound, Alaska earthquake of March 28, 1964, and aftershock se- quence [abs.]:Geol. Soc. America Spec. Paper 82, p. 2. Balakina, L. M., Shirokova, H. 1., and Vvedenskaya, A. V., 1961, Study of stresses and ruptures in earth- quake foci with the help of disloca- tion theory: Ottawa Dominion Ob- servatory Pub, v. 24, no. 10, p. 321- 327. Barnes, D. F., 1966, Gravity changes during the Alaska earthquake: J our. Geophys. Research, v. 71, no. 2, p. 451—456. Barnes, F. F., 1962, Geologic map of the lower Matanuska Valley, Alaska: US. Geol. Survey Misc. Geol. Inv. Map I—359, scale 1 : 63,360. Beniofi, Hugo, 1951, Earthquakes and rock creep; pt. 1, Creep characteris- tics of rocks and the origin of after- shocks: Seismol. Soc. America Bull., v. 41, no. 1, p. 31—62. 1954, Orogenesis and deep crustal istructure~additional evidence from seism‘ology: Geol. Soc. America Bull., v. 65, no. 5, p. 385—400. Bollinger, G. A., and Stauder, William, 1965, Geometry of earthquake foci stresses [abs] : Seismol. Soc. Amer- ica, Eastern Sec, Earthquake Notes, v. 36, nos. 3—4, p. 22. Bolt, B. A., 1964, Seismic air waves from the great 1964 Alaska earth- quakezNature, v. 202, no. 4937, p. 1095—1096. Brazee, R. J ., 1965, A study of T phases in the Aleutian earthquake series of March and April 1957 : Seismol. Soc. America, Eastern Sec. Earth- quake Notes, v. 36, nos. 1~2, p. 9—14. TE CTONICS plate overrides the ocean basin or in which the regional strain is di— dected horizontally, do not readily account either for the widespread preearthquake subsidence relative to sea level that accompanied hori— REFERENCES CITED Burk, C. A., 1965, Geology of the Alaska Peninsula—island arc and continen- tal margin: Geol. Soc. America Mem. 99, pt. 1, 250p. Carey, S. W., 1958, The tectonic ap- proach to continental drift, in Con- tinental drift—a symposium: Ho- bart, Australia, Tasmania Univ. Geol. Dept. [1956], p. 177—355. Clark, D. W., 1966, Perspectives in the prehistory of Kodiak Island, Alas- ka: Am. Antiquity, v. 31, no. 3, p. 358—371. Coats, R. R., 1962, Magma type and crustal structure in the Aleutian Arc, in The crust of the Pacific Basin: Am. Geophys. Union. Geo- phys. Mon. 6, p. 92—109. Coleman, J. M., and Smith, W. G., 1964, Late Recent rise of sea level: Geol. Soc. Am. Bull., v. 75, no. 9, p. 833— 840. Curray, J. R., 1961, Late Quaternary sea level—a discussion: Geol. Soc. Am. Bull., v. 72, no. 11, p. 1707— 1712. Dachnowski-Stokes, A. P., 1941, Peat resources in Alaska: US. Dept. Agriculture Tech. Bull. 769, 84 p. De Laguna, Frederica, 1934, The ar- chaeology of Cook Inlet, Alaska, with a chapter on Skeletal material, by Bruno Oetteking: Philadelphia Univ., Pa., Pennsylvania Press, 263 p. 1956, Chugach prehistory; the archaeology of Prince William Sound, Alaska: Seattle, Washing- ton Univ. Press, 289 p. Detterman, R. L., and Reed, B. L., 1964, Preliminary map of the geology of the Iliamna quadrangle, Alaska: US. Geol. Survey Misc. Geol. Inv. Map I—407, scale 1: 250,000. Dietz, R. S., 1961, Continent and ocean basin evolution by spreading of the sea floor: Nature, v. 190, no. 4779, p. 854—857. I71 zontal shortening in coastal areas affected by earthquake-related tectonic deformation or for the orientation of principal stress axes at the hypocenter of the main shock and many of its aftershocks. Earthquake Research Institute, 1964, Report of the Niigata earthquake of June 16, 1964: Spec. Bull. Tokyo Univ., Earthquake Research Inst., v. 8, 133 p. [In Japanese]. Eckel, E. B., 1967, Effects of the earth- quake of March 27, 1964, on air and water transport, communications, and utilities systems in south-cen- tral Alaska: US. Geol. Survey - Prof. Paper 545-B, p. B1—B27. Evison, F. F., 1963, Earthquakes and faults: Seismol. Soc. America Bull., v. 53, no. 5, p. 873—891. Ewing, Maurice, Ludwig, W. J ., and Ewing, John, 1965, Oceanic struc- tural history of the Bering Sea: Jour. Geophys. Research, v. 70, no. 18, p. 4593—4600. Fairbridge, R. W., 1961, Eustatic changes in sea level, in Physics and chemistry of the earth: New York, Pergamon Press, v. 4, p. 99-185. Florensov, N. A., and Solonenko, V. P., eds, 1963, The Gobi-Altai earth- quake: Izdatelo Akad. Nauk SSSR, 391 p. [in Russian; English trans- lation, U.S. Dept. of Commerce, 1965.] Foster, H. F., and Karlstrom, T. N. V., 1967, Ground breakage and associ- ated efiects in the Cook Inlet area, Alaska, resulting from the March 27, 1964, earthquake: U.S. Geol. Survey Prof. Paper 543—F, p. F1— F28. Furumoto, A. S., 1967, Analysis of Ray- leigh waves, pt. 2 of Source mecha— nism of the Alaska earthquake and tsunami of 27 March 1964: Pacific Sci, v. 21, no. 3, p. 311—316. Gates, G. 0., and Gryc, George, 1963, Structure and tectonic history of Alaska: Am. Assoc. Petroleum Geologists Mem. 2, p. 264—277. Gates, Olcott, and Gibson, W. M., 1956, Interpretation of the configuration of the Aleutian Ridge [Alaska]: Geol. Soc. America Bull., v. 67, no. 2, p. 127-146. I72 Ghent, E. D., 1965, Comments on phase changes and shallow focus earth- quakes: New Zealand Jour. Geol- ogy and Geophysics, v. 8, no. 5, p. 820-832. Grant, U. S., and Higgins, D. F., 1910, Reconnaissance of the geology and mineral resources of Prince William Sound, Alaska: U.S. Geol. Survey Bull. 443, 89 p. Grantz, Arthur, 1965, Geologic map and cross sections of the Nelchina area, south-central Alaska: U.S. Geol. Survey open-file map, 4 sheets, scale 1:63,360. 1966, Strike-slip faults in Alas- ka: U.S. Geol. Survey open-file re- port, 82 p. Grantz, Arthur, Plafker, George, and Kachadoorian, Reuben, 1964, Alas- ka’s Good Friday earthquake, March 27, 1964, a preliminary geo- logic evaluation: U.S. Geol. Survey Circ. 491, 35 p. Griggs, D. T., and Blacic, J. D., 1965, Quartz—anomalous weakness of synthetic crystals: Science, v. 147, no. 3655, p. 292—295. Griggs, D. T., and Handin, J. W., 1960, Observations on fracture and a hy- pothesis of earthquakes: Geol. Soc. America Mem. 79, p. 347—363. Gutenberg, Beno, and Richter, C. F., 1954, Seismicity of the earth and associated phenomena, 2d ed.: Princeton, N .J., Princeton Univ. Press, 310 p. Hales, A. L., and Alsaada, T., 1966, Crustal structure in coastal Alaska, m The earth beneath the continents: Am. Geophys. Union Geophys. Mon. 10, p. 420—432. Hamilton, Warren, and Myers, W. B., 1966, Cenozoic tectonics of the western United States; Reviews of Geophysics, v. 4, no. 4, p. 509—549. Hanna, G D., 1964, Biological effects of an earthquake: Pacific Discovery, v. 17, no. 6, p. 24—26. Hansen, W. R., and Eckel, E. B., 1966, A summary description of the Alaska earthquake—its setting and effects, in Hansen, Eckel, and others, The Alaska earthquake, March 27, 1964: Field investiga- tions and reconstruction effort: U.S. Geol. Survey Prof. Paper 541, p. 1—37. Hatori, Tokuta-ro, 1965, on the tsunami which accompanied the Niigata earthquake of June 16, 1964— source deformation, propagation, and tsunami run-up: Tokyo Univ., Earthquake Research Inst. Bull., v. 43, no. 1, p. 129—148. ALASKA EARTHQUAKE, MARCH 27, 1964 Henderson, J ., 1933, The geological as- pects of the Hawke’s Bay earth- quakes [New Zealand]: New Zea- land Jour, Sci. Technology, v. 15, no. 1, p. 38—75. Hess, H. B., 1955, The oceanic crust: Jour. Marine Research, v. 14, no. 4, p. 423439. 1957, The Vening Meinesz nega- tive gravity anomaly belt of is- land arcs 1926—56: Koninkl. Nederl. Geol.-Mijnbouka. Genoot., Verh., Geol. Ser., v. 18, 183-188. 1962, History of ocean basins, in Petrologic studios (Buddington volume): Geol. Soc. America, p. 599—620. Heusser, C. J., 1960, Late-Pleistocene environments of North Pacific North America—an elaboration of late-glacial and postglacial cli- matic, physiographic, |and biotic changes: Am. Geog. Soc. Spec. Pub. 35, 308 p. Hoare, J. M., 1961, Geology and tectonic setting of lower Kuskokwim-Bristol Bay region, Alaska: Am. Assoc. Petroleum Geologists Bull., v. 45, no. 5, p. 594—611. Hodgson, J. H., 1957, Nature of faulting in large earthquakes: Geol. Soc. America Bull, v. 68, no. 5, p. 611- 643. Hopkins, D. M., MacNeil, F. S., Merklin, R. L., and Petrov, 0. M., 1965, Quaternary correlations across Bering Strait: Science, v. 147, no 3662, p. 1107—1114. Iida, K., 1963, A relation of earthquake energy to tsunami energy and the estimation of the vertical displace- ment in a tsumami source: Nagoya [Japan] Univ. Inst. Earth Sci., Earth Sci. Jour., v. 11, no. 1, p. 49—67. Imamura, Akitsune, 1930, Topographi- cal changes accompanying earth- quakes or volcanic eruptions: Ja- pan Imperial Earthquake Inv. Comm., Foreign Languages Pubs, no. 25, 14.3 p. Inoue, E., 1960, Land deformation in Japan: Japan Geog. Survey Inst. Bull, v. 6, pt. 2—3, p. 73—134. Jordan, J. N., Lander, J. F. and Black, R. A., 1965, Aftershocks of the 4 February 1965 Rat Island Earth- quake: Science, v. 148, no. 3675, p. 1323—1325. Karlstrom, T. N. V., 1964, Quaternary geology of the Kenai Lowland and glacial history of the Cook Inlet region, Alaska: U.S. Geol. Survey Prof. Paper 443, 69 p. Kaye, C. A., 1964, The upper limit of barnacles as an index of sea level change on the New England Coast during the past 100 years: Jour. Geology, v. 72, no. 5, p. 580—600. Kelly, T. E., 1963, Geology and hydro- carbons in Cook Inlet basin, Alas- ka:Am. Assoc. Petroleum Geolo- gists Mem. 2, p. 278—296. Lamb, Horace, 1932, Hydrodynamics, 6th ed.: England, Cambridge Univ. Press, 738 p. Lanphere, M. A., 1966, Potassium-argon ages of Tertiary plutrons in the Prince William Sound region, Alaska, in Geological Survey re- search 1966: U.S. Geol. Survey Prof. Paper 550—D, p. D195—D198. Lensen, G. J ., 1961, Principal horizontal stress directions as an aid to the study of crustal deformation, in A symposium on earthquake mecha- nism: Ottawa Dominion Observa- tory Pub., v. 24, no. 10, p. 389—397. Levin, Betsy, Ives, P. 0., Oman, C. L., and Rubin, Meyer, 1965, U.S. Geo- logical Survey radiocarbon dates VIII : Radiocarbon, v. 7, p. 372—398. Lyell, Sir Charles, 1874, Principles of geology: New York, D. Appleton and 00., v. 2, 652 p. McCulloch D. S., 1966, Slide-induced waves, seiching, and ground frac- turing caused by the earthquake of March 27, 1964, at Kenai Lake, Alaska: U.S. Geol. Survey Prof. Paper 543—A, p. A1—A41. MacKevett, E. M., Jr., Berg, H. C., Plafker, George, and Jones, D. L., 1964, Preliminary geologic map of the McCarthy C-4 quadrangle, Alaska: U.S. Geol. Survey Misc. Geol. Inv. Map I—423, scale 1:63,360. Malloy R. J., 1964, Crustal uplift southwest of Montague Island, Alaska: Science, v. 146, no. 3647, p. 1048—1049. 1965, Gulf of Alaska—seafloor upheaval: Geo-Marine Technology, v. 1, no. 5, p. 22—26. Marmer, H. A., 1951, Tidal datum planes, revised ed.: U.S. Coast and Geodetic Survey Spec. Pub. 135, 142 p. Martin, G. C., Johnson, B. L., and Grant, U.S., 1915, Geology and mineral resources of the Kenai Peninsula, Alaska: U.S. Geol. Sur- vey Bull. 587, 243 p. Martin, G. C., and Katz, F. J., 1912, Geology and coal fields of the lower Matanuska Valley, Alaska: U.S. Geol. Survey Bull. 500, 98 p. Miller, D. J ., 1953, Late Cenozoic marine glacial sediments and marine ter— races of Middleton Island, Alaska: Jour. Geology, v. 61, no. 1, p. 17-40. 1961, Geology of the Katalla district, Gulf of Alaska Tertiary province, Alaska: US. Geol. Survey open-file map, scale 1: 96,000, 2 sheets. Miller, D. J ., and MacColl, R. S., 1964, Geologic map and sections of the northern part of the McCarthy A—4 quadrangle, Alaska: U.S. Geol. Survey Misc. Geol. Inv. Map I—410, scale 1 : 63,360. Miller, D. J., Payne, T. G., and Gryc, George, 1959, Geology of possible petroleum provinces in Alaska, with an annotated bibliography by E. H. Cobb: U.S. Geol. Survey Bull. 1094, 131 p. Mofiit, F. H., 1954, Geology of the Prince William Sound region, Alaska: U.S. Geol. Survey Bull. 989-E, p. 225— 310. Moore, G. W., 1967,_Preliminary geologic map of Kodiak Island and vicinity, Alaska: U.S. Geol. Survey open-file map, scale 1 : 250,000. Muto, K., 1932, A study of displacements of triangulation points: Tokyo Univ., Earthquake Research Inst. Bull., v. 10, p. 384—392. Orowan, E., 1960, Mechanism of seismic faulting: Geol. Soc. America Mem. 79, p. 323—345. Page, R. A., Jr., 1967, Aftershocks as a near-surface phenomenon [abs] : Am. Geophys. Union Trans, v. 48, no. 1, p. 205. Pararas-Carayannis, G e o r g e , 1967, Water waves, pt. 1 of Source me- chanism study of the Alaska earth- quake and tsunami of 27 March 1964: Pacific Science, v. 21, no. 3, p. 301—310. Parkin, E. J ., 1966, Horizontal displace- ments, pt. 2 of Alaskan surveys to determine crustal movement: U.S. Coast and Geodetic Survey, 11 p. Payne, T. G., 1955, Mesozoic and Ce- nozoic tectonic elements of Alaska : U.S. Geol. Survey Misc. Geol. Inv. Map 1—84, scale 1: 5,000,000. Peter, G., Elvers, D., Yellin, M., 1965, Geological structure of Aleutian Trench southwest of Kodiak Island: J our. Geophys. Research, v. 70, no. 2, p. 353—66. TE CTONICS Plafker, George, 1965, Tectonic defor- mation associated with the 1964 Alaskan earthquake: Science, v. 148, no. 3678, p. 1675—1687. ‘ 1967a, Geologic map of the Gulf of Alaska Tertiary Province, Alaska: US. Geol. Survey Misc. Geol. Inv. Map I—484, scale 1: 500,000. ———1967b, Surface faults on Mon- tague Island associated with the 1964 Alaska earthquake: U.S. Geol. Survey Prof. Paper 543—G, p. G1—G42. Plafker, George, and Kachadooriau, Reuben, 1966, Geologic effects of the March 1964 earthquake and associated seismic sea waves on Kodiak and nearby islands, Alaska : U.S. Geol. Survey Prof. Paper 543—D, p. D1—D46. Plafker, George, Kachadoorian, Reuben, Eckel, Edwin B., and Mayo, L. R., 1969, Effects of the earthquake of March 27, 1964, on various com- munities: U.S. Geol. Survey Prof. Paper 542—G, G1—G50. Plafker, George, and MacNeil, F. S., 1966, Stratigraphic significance of Tertiary fossils from the Orca Group in the Prince William Sound region, Alaska, in Geological Survey research 1966: U.S. Geol. Survey Prof. Paper 550—B, p. B62—B68. Plafker, George, and Mayo, L. R., 1965, Tectonic deformation, subaqueous slides, and destructive waves associ- ated with the Alaskan March 27, 1964, earthquake—an interim geo~ logic evaluation: U.S. Geol. Survey open-file report, 21 p. Plafker, George, and Rubin, Meyer, 1967, Vertical tectonic displace- ments in south-central Alaska dur- ing and prior to the great 1964 earthquake: Jour. Geosciences (Osaka City Univ.), v. 10, art. 1—7, p. 1—14. Press, Frank, 1965, Displacements, strains, and tilts at teleseismic distances: J our. Geophys. Research, v. 70, no. 10, p. 2395—2412. Press, Frank, and Jackson, David, 1965, Alaskan earthquake, 27 March 1964—vertical extent of faulting and elastic strain energy release: Science, v. 147, no. 3660, p. 867—868. Rainey, F. G., and Ralph, Elizabeth, 1959, Radiocarbon dating in the Arctic: Am. Antiquity, v. 24, no. 4, p. 365-374. I73 Raleigh, C. B., 1967, Tectonic implica- tions of serpentinite weakening: Royal Astron. Soc. Geophys. Jour., v. 14, p. 113—118. Raleigh, C. B., and Patterson, M. S'., 1965, Experimental deformation of serpentinite and its tectonic im- plications: Jour. Geophys. Re- search, v. 70, no. 16, p. 3965—3985. Reid, H. F., 1911, The California earth- quake of April 18, 1906. The me- chanics of the earthquake: Car- negie Inst. Washington Pub. 87, v. 2, 192 p. Reimnitz, Erk, 1966, Late Quaternary history and sedimentation of the Copper River Delta and vicinity: California Univ., San Diego, un- published Ph. D. thesis, 160 p. Richter, C. F., 1958, Elementary seismol- ogy: San Francisco, Calif, W. H. Freeman and 00., 768 p. Ricketts, E. F., and Calvin, Jack, 1962, Between Pacific tides; an account of the habit and habitats of some 500 of the common, conspicuous sea- shore invertebrates of the Pacific Coast between Sitka, Alaska, and northern Mexico, 3d ed.: Stanford, Calif, Stanford Univ. Press, 516 p. Rigg, G. B., and Miller, R. C., 1949, Intertidal plant and animal zon- ation in the vicinity of Neah Bay, Washington: California Acad. Sci. Proc., 4th ser., v. 26, no. 10, p. 323-351. Row, R. V., 1966, On the ionospheric long-period acoustic-gravity wave pulse launched by the great Alaskan earthquake of March 1964 [abs]: Am. Geophys. Union Trans, v. 47. no. 1, p. 51. Rubin, Meyer, and Alexander, Corrinne, 1958, U.S. Geological Survey radio- carbon dates IV : Science, v. 127, no. 3313, p. 1476-1487. Saint-Amand, Pierre, 1957, Geological and geophysical synthesis of the tectonics of portions of British Columbia, the Yukon Territory, and Alaska: Geol Soc. America Bull., v. 68, no. 10, p. 1343-1370. 1961, Les terremotos de mayo — Chile 1960 * * * an eyewitness account of the greatest catastrophe in recent history: China Lake, Calif, Michelson Lab., U.S. Naval Ordnance Test Sta. Tech. Art. 14, 39p. I74 Savage, J. G., and Hastie, L. M., 1966, Surface deformation associated with dip-slip faulting: Jour. Geo- phys. Research, v. 71, no. 20, p. 4897—4904. Schrader, F. C.. 1900, A reconnaissance of a part of Prince William Sound and the Copper River district, Alaska, in 1898: US. Geol. Survey 20th Ann. Rept., pt. 7, p. 341—423. Shor, G. G., J r., 1962, Seismic refraction studies off the coast of Alaska 1956- 1957: Seismol. Soc. America Bull., v. 52. no. 1, p. 37—57. 1966, Continental margins and island arcs of Western North America, in Continental margins and island arcs: Canada Geol. Survey Paper 66—15, p. 126—222. Small, J. B., 1966, Vertical bench mark displacement, pt. 1 of Alaskan surveys to determine crustal move- ment: US. Coast and Geod. Survey, 24 p. Stauder, William, 1967, Seismic evi- dence of present deformation in [island arc structures [abs]: Am. Geophys. Union Trans, v. 48, no. 1, p. 218—219. Stauder, William, and Bollinger, G. A., 1964, the S wave project for focal mechanism studies—Earthquakes of 1962: Seismol. Soc. America Bull., v. 54, no. 6, pt. B, p. 2199—2208. 1966, The focal mechanism of the Alaska earthquake of March 28, 1964, and of its aftershock se- quence: J our. Geophys. Research, v. 71, no. 22, p. 5283—5296. Stephenson, T. A., and Stephenson, Anne, 1949, The universal features of zonation between tide-marks on rocky coasts: Jour. Ecology, v. 37, p. 289—305. ALASKA EARTHQUAKE, MARCH 27, 19164 Tarr, R. S., and Martin, Lawrence, 1912, The earthquake at Yakutat Bay, Alaska, in September, 1899, with a preface by G. K. Gilbert: U.S. Geol. Survey Prof. Paper 69, 135 p. 1914, Alaskan glacier studies of the National Geographic Society in the Yakutat Bay, Prince William Sound and lower Copper River Regions: Washington, D.C., Natl. Geog. Soc., 498 p. Thorsteinsson, F. V., 1965, Effects of the Alaska earthquake on pink and. chum salmon runs in Prince William Sound: Auke Bay, Alaska, US. Bur. Commercial Fisheries Biol. Lab., 16 p. Tobin, D. G., and Sykes, L. R., 1966, Relationship of hypocenters of earthquakes to the geology of Alaska: Jour. Geophys. Research, v. 71, no. 6, p. 1659—1667. Tocher, Don, 1960, The Alaska earth- quake of July 10, 1958—M0vement on the Fairweather fault and field investigation of southern epicen-tral region: Seismol. Soc. America Bull., v. 50, no. 2, p. 267—292. Toksoz, M. N., Ben-Menahem, Ari, and Harkrider, D. G., 1965, Source mechanism of Alaska earthquake from long-period seismic surface waves [abs]: Am. Geophys. Union Trans, v. 46, no. 1, p. 154. US. Coast and Geodetic Survey, 1964, Tide tables, high and low water predictions, 1964, West Coast North and South America including the Hawaiian Islands: US. Coast and Geod. Survey, 224 p. Vancouver, John, ed., 1801, A voyage of discovery to the North Pacific Ocean and around the world * * * in the years 1790—1795 * * * under the command of Captain George Vancouver * ‘ *, new ed.:London, printed for J. Stockda'le, 6 v. Van Dorn, W. G., 1964, Source mech- anism of the tsunami of March 28, 1964, in Alaska: Coastal Eng. Conf., 9th, Lisbon 1964, Proc., p. 166—190. Vine, F. J ., 1966, Spreading of the ocean floor—new evidence: Science, v. 154, no. 3755, p. 1405—1415. Von Huene, Roland, Shor, G. R., Jr., and Saint-Amand, Pierre, 1966, Active faults and structure of the continental margin in the 1964 Alaskan aftershock sequence area [abs] : Am. Geophys. Union Trans, v. 47, no. 1, p. 176. Wahrhaf-tig, Clyde, 1965, The physio- graph/ic provinces of Alaska: U.vS. Geol. Survey Prof. Paper 482, 52 p. Waller, R. M., 1966a, Effects of the earthquake of March 27, 1964 in the Homer area, Alaska with a section, on Beach changes on Homer Spit, by K. W. Stanley: U.S. Geol. Sur- vey Prof. Paper 542—D, p. D1—D28. Waller, R. M., 1966b, Effects of the March 1964 Alaska earthquake on the hydrology of south-central Alaska: US. Geol. Survey Prof. Paper 544—A, p. A1—A28. Wolfe, J. A., Hopkins, D. M., and Leo- pold, E. B., 1966, Tertiary stra- tigraphy and paleobotany of the Cook Inlet region, Alaska: US. Geol. Survey Prof. Paper 398—A, p. A1—A29. Wood, F. J ., ed., 1966, Operational phases of the Coast and Geodetic Survey program in Alaska for the period March 27 to December 31, 1964, v. 1 of The Prince William Sound, Alaska, earthquake of 1964 [and aftershocks: US. Coast and Geodetic Survey, 263 p. Wyss, Max, and Brune, J. N., 1967, The Alaska earthquake of 28 March 1964—A complex multiple rupture: Bull. Seis. Soc. America, v. 57, no. 5, p. 1017—1023. ".5. GOVERNMENT PRINTING OFFICE : 1910 0—307-6“ UNITED STATES DEPARTMENT OF THE INTERIOR PROFESSIONAL PAPER 543—1 GEOLOGICAL SURVEY PLATE 1 EXPLANATION 2.0B* 2.DF* 2.05 2.0F 2.0TBM . —2.0B , —2.or . —2.0TBM Upper growth Upper growth Height of tidal limit of barna- limit of marine b e n c h m a r k cles (Balanus) algae (Fucus) Vertical tectonic displacement, in feet, measured along shore with tide level as a datum. Asterisk indicates 1965 measurement; all others are 1961, 2 OBB* 2.0FF 5 OVV* 2.038 2.0FF 5.0VV . —2.0EB . —2.0FF . —5.0VV . 10.0DD Upper growth Upper growth Lower growth Upper limit of limit of barna- limit of marine limit of terres- s t o r m t i d e cles (Balanus) algae ( F ucus ) trial vegetation driftwood. (In uplifted area Vertical tectonic displacement, in feet, from mea— only) sured difference between pre- and postearth— quake shoreline features Asterisk indicates 1965 measurement; all others are 1961, 51/211/2L* 5mm . —5‘Ai‘/2L Vertical tectonic displacement, in feet, estimated by local resident Asterisk indicates 1965 estimate; all others are 1961, —3,orr* . —3.on= ‘28“ 5’23” 5%? f Vertical tectonic displacement, in feet, measured at $2 w standard US. Coast and Geodetic Survey tide- gage station Asterisk indicates 1965 measurement; all others are 1961, —3.01* A —3.01 Vertical tectonic displacement, in feet, measured at temporary US. Coast and Geodetic Survey tide— gage station Preearthquake observations were made between 1906 and 1957; postearthquake measurements were made in 1961, and 1965 (asterisk). Most of the displacement occurred during the earthquake 223* 2.25 x _2.2s Vertical tectonic displacement, in feet, measured at U.S. Coast and Geodetic Survey bench mark Preearthquake observations were made between 1922 and 1952; postearthquake measurements were made in 1961, and 1.965 (asterisk). Most of the displacement occurred during the earthquake —2 Isobase of tectonic land—level change Contour interval is 2 feet from —8 to +10feet; 5feet from +10 to +35 feet. Solid where estimated precision is i % contour interval; dashed where i 1 contour interval; dotted where inferred ___.I______?____ Axis of tectonic uplift Dashed where approximately located; queried where inferred ————.- ..... *2 '5‘— ;fmzaw snow 5 l 1‘ Axis of tectonic subsidence Dashed where approximately located 5BB" 1.... - B 4 _ heading of 5.4’ less ,afact'pn of unconsolidated i—fi—L—‘I—fi-++-fi-IIA-Ib Reverse fault Dashed where approximately located; dotted where inferred. Barbs on upthrown side .4 U fi- Kenai lineament Zone of possible earthquake-related surface faulting or con- centrated warping and distortion. U, upthrown side. Arrows indicate sense of horizontal displacement sug— gested by retriangulation across the zone View» —4:E1L* 14 Major drainage affected by tectonic warping ‘ Dashed where uncertain. Arrow points in direction of tilt .: C" ‘ k—— Gulf of Alaska —> g +40: A Eff.) I f“ / T/FAULT7 gi +30 .fi ’\. o ”3 q |\ 7\/Postearthquake land surface § +20, “0' l<-—Shelikof Strait—fl T /,/'l/ 1” - ~~;L\\7‘ :3 +10, 4 0, . \\_3‘ I O’ “5 f ’ . —10’ J \Preearthquake land surface L 4 / 4,. ' 3.2m _10 , .M—znxl/ZL c A s B' "l . o 8 § , 'k‘ ,, , +3i1l / g l<——-—Gulf of Alaska 15E +5 3 — +40 “+3I/z—4v. (lakes) +30' 09 a a 2 5. _ , ‘ avian <5 1:6 o -"’ +30 r“ ' / , 4,! PATTON BAY FAULT -— s : § +20 <3 24 9 3‘44 h +204 ,1 ld .E {3 4_ +30. \l 2": x +20' — 3% 54; in“? 4;“ l<—-Gulf of Alaska—> Ea — +20' <( :5: v +10, T Eu: (3%” 8E: Postearthquake land surface T //’——?——?_—?——?\ 7 ~ +10' 0: —2— L l l I Preearthquake land surface ‘\\7\ \9 o, l. INTERIOR-GEOLOGICAL SURVEY. WASHINGTON. D.C.—1969—G68226 :1 —10' V —10’ PROFILES OF LAN D-LEVEL CHANGE ALONG A—A’, B—B’, AND C-C’ . 4 4 ‘2 ; a , l Base from U. 3. Geological Survey, 1956 ' ”"‘ ‘ ‘ ‘ Data from George Plafker, L. R. Mayo, J. B. Case, D. s. McCulloch, “ 2%. L. V . M. G. Bonilla, and U.S. Coast and Geodetic Survey, 1964—65 SCALE 1: 2 000 000 50 0 50 100 150 MlLES l—l | ' l———l l - l 50 O 50 100 150 KILOMEIERS l—-—l l———-l [_[ g l CONTOUR INTERVAL 200 FEET DATUM IS MEAN SEA LEVEL SUBMARINE CONTOURS lN FEET MAP SHOWING GROUND DEFORMATION RESULTING FROM THE 1964 ALASKA EARTHQUAKE IN SOUTH-CENTRAL ALASKA UNITED STATES DEPARTMENT OF‘ THE INTERIOR PROFESSIONAL PAPER 545—1 GEOLOGICAL SURVEY PLATE 2 -sec 146°00’ F“ W.- "i", " 61"<3C>’ / me got-53H l N 3 u l N e a 0 OJ" 108* Z.0F* 2.03 ZIOF 2.0TBM ‘ MM .—2.0B .—2.DF .—2.0TBM Upper growth Upper growth Height of tidal \ limit of barna— limit of marine b e n c h m a r k , cles (Balanus) algae (Fucus) l Vertical tectonic displacement, in feet, measured along shore with tide level as a datum Asterisk indicates 1965 measurement; all others are 1961, $1333 ._§;8§E .3333 . 10.00:) 590501 Upper growth Upper growth Lower growth Upper limit of , , , , . limit of barna- limit of marine limit of terres— s t o r m t i d e , . , , 37%;”, ' ' q _, f ~ 3 - cles (Balanus) algae (Fucus) trial vegetation driftwood. (In , 13%on \,®/. { ‘ _ U ‘ uplifted area I 0W $97 ?4* 23- I ‘ \ ’ only) ' ‘ . ’ ' , Vertical tectonic displacement, in feet, from mea— sured difference between pre- and post-earth— quake shoreline features Asterisk indicates 1965 measurement; all others are 1964 60°00 5Vz'tI/zL* SI/zinl. . ISMM Vertical tectonic displacement, in feet, estimated by Uplift, in feet, from pre- and postearthquake depth local resident soundings near Montague Island by the US. Asterisk indicates 1965 estimate; all others are 1961, Coast and Geodetic Survey Preearthquake soundings were made in 1927; postearth- ~ quake soundings in 1961, and 1965. Most of the change Ajigi" occurred during the earthquake 16333“ X.“ , I as r WM,“ .M “WM-” ,wr 5904?? 2;: X (P ____...... _.2 Subsidence, in feet, measured at temporary U.S. ‘2 ‘ is” Coast and Geodetic Survey tide-gage station Isobase 0f tectomc land-level change 1’ Preearthquake observations were made between 1906‘ and COMO?” interval is 2f“?t from ‘8 10 +10feet; 5feet from 45* 1957, postearthquake measurements were made in 1961, +10 to +35 feet. SOIId where estimated precision is :I: 1% and 1 .965 (asterisk). M ost of the displacement occurred contour interval; dashed where =I= 1 007L107“ interval; dotted I? during the earthquake where inferred 31' L 3‘ _A_A__A_.A._.A._A..4.. Reverse fault Dashed where approximately located; dotted where inferred. Barbs on upthrown side , ~l 553°3Q’ its so L ’ A I I 46°0t' =1 14:» Base from U. 8. Geological Survey, 1964 SCALE 1:500 000 Data from George Plafker, L. R. Mayo, J. B. Case, D. S. McCuIIoch, M. G. Bonilla, and U.S. Coast and Geodetic Survey, 1964—65 10 o 10 20 30 MILES I——-l l—I 3—1 T———-l *1 10 O 10 20 3O KILOMETERS l—l l—l l-—I l——l I CONTOUR INTERVAL 200 FEET A DATUM IS MEAN SEA LEVEL A’ I<—— MONTAGUE STRAIT ———>I ye GULF OF ALASKA ~> HANNING BAY FAULT ____—————-"—‘\ PATTON BAY FAULT 30' / __._——-———-—-—— 30’ 40' 40' 20’ __// /Postearthquake land surface 20/ _—————'————— ———— -—————————\__——__— 10' 1 I I 10, I Preearthquake land surface INTERIOR—GEOLOGICAL SURVEY. WASHINGTON. D.c.—1969—eeazzs O 1 2 3 4 5 MILES | I l l | | O 1 2 3 4 5 KILOMETERS PROFILE OF LAND—LEVEL CHANGE ALONG A-A’ MAP SHOWING GROUND DEFORMATION RESULTING FROM THE 1964 ALASKA EARTHQUAKE IN THE PRINCE WILLIAM SOUND REGION The Alaska Earthquake March 27, 1964 Regional Effects e X x: j A. m9; 1...... w ’%© Valdez ‘ “f '4 Lax; / Shore Processes .‘ and Beach Morphology GEOLOGICAL SURVEY PROFESSIONAL PAPER 543-J THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS Effects of the Alaska Earthquake of March 27, 1964 On Shore Processes and Beach Morphology By KIRK W. STANLEY The efiects of tectonic uplift and subsidence along 10,000 miles of shoreline, and the practical meaning of those efiects GEOLOGICAL SURVEY PROFESSIONAL PAPER 543—J UNITED STATES DEPARTMENT OF THE INTERIOR STEWART L. UDALL, Secretary GEOLOGICAL SURVEY William T. Pecora, Director UNITED STATES GOVERNMENT PRINTING OFFICE, WASHINGTON : 1968 For sale by the Superintendent of Documents, U.S. Government Printing Oflice Washington, DC. 20402 — Price 45 cents (paper cover) THE ALASKA EARTHQUAKE SERIES The US. Geological Survey is publishing the re- sults of its investigations of the Alaska earthquake of March 27, 1964, in a series of six Professional Papers. Professional Paper 543 describes the re- gional effects of the earthquake. Other Professional Papers in the series describe field investigations and reconstruction and the effects of the earthquake on communities, on the hydrologic regimen, and on transportation, utilities, and communications. CONTENTS Page Abstract _______________________ J1 Introduction ____________________ 1 Shorelines of the earthquake-af- fected region _________________ 2 Coastal features and earthquake effects _______________________ 3 Beaches ____________________ 3 Changes in profile and gradient ______________ 3 Minor beach features _____ 5 Low-water features ______ 6 Modern beach ridges _____ 7 Ancient beach ridges _____ 10 Stream-mouth changes _______ 10 Submergent areas _______ 10 Uplifted areas ___________ 11 Coastal erosion and movement of material _______________ 12 Erosion ________________ 12 Page Coastal features and earthquake effects —Continued Coastal erosion and movement of material—Continued Longshore material move- ment ________________ J 15 Upslope material move- ment ________________ 1 7 Biologic efl’ects of shoreline changes ______________________ 17 Fish _______________________ 17 Shellfish ____________________ 18 Wildfowl ___________________ 1 8 Effects on property and manmade structures _____________________ 19 Submergent areas ___________ 19 Emergent areas _____________ 19 Legal problems ______________ 19 Need for further studies __________ 20 References cited _________________ 20 ILLUSTRATIONS FIGURES Page . Index map of south-central Alaska showing coast- 9. Photograph of gullying and sloughing caused by lines affected by the earthquake ______________ VI headward stream erosion, Copper River Delta- . Diagram illustrating beach feature _____________ J3 10. Photograph of eastern shore of Cook Inlet showing . Photograph of sharp break-in-slope on shingle undercutting and sloughing of bluffs caused by beach _____________________________________ 4 subsidence _________________________________ . Sketch of changes of beach configuration in sub— 11. Photograph of bluff erosion at Kenai ____________ mergent areas ______________________________ 5 12~14. Photographs of timber seawall on Homer Spit: . Photograph of deposition of material on backslope 12. As constructed _________________________ of frontal ridge ____________________________ 8 13. Damaged, 6 months after construction“ _ _ . Aerial photograph of barrier beach at Breving 14. Almost completely destroyed by waves, 16 Lagoon, Seward Peninsula ___________________ 9 months after construction _____________ . Photograph showing undercutting and sloughing 15. Photograph of cobble-filled wire-mesh fence, of bluffs, Cook Inlet ________________________ 9 Larsen Bay, Kodiak Island __________________ . Oblique photograph of new land created by tec- 16. Profiles of updrift and downdrift fill of grain at tonic uplift, Copper River Delta _____________ 11 Homer Spit ________________________________ V Page .112 13 13 14 14 15 15 16 I . #1 < ,Tgfiele I.“ . 1219 l a a» a” t 01/4; [4/4 I) 4! (1'. ". * k _ %.Ch?“ega, {ll Hinchinbrook I. . ‘ a I i . , T fixelhn' v9 ,u qPort‘Oceanic ’7 ; Bay 6 flawmill Bay ‘ // I 3 " ontague I. A K S P» , , X) , P~ EXPLANATION ’ ,4 4 // Approximate line of zeroland level ' / _ § changelafter Plafker, 1965) «‘5 AFégvé‘K // 0 ‘5 “SD-Sap ‘ ’ / Po ,. Wakgmognak // Q fir- ~ 5’3 fiQuzinkie / I diak / V odiak kaval Station * / ’6 C5 0 50 L L nob ' L 150 MILES 56°15; ' I _, 1 56°15" 156° 144° 1.—-South-central Alaska showing coastlines afiected by the earthquake. Land to left of zero land-level change was generally lowered; land to right was raised. VI THE ALASKA EARTHQUAKE, MARCH 27, 1964: REGIONAL EFFECTS EFFECTS OF THE ALASKA EARTHQUAKE OF MARCH 27, 1964, ON SHORE PROCESSES AND BEACH MORPHOLOGY Some 10,000 miles of shoreline in south-central Alaska was affected by the subsidence or uplift associated with the great Alaska earthquake of March 27, 1964. The changes in shoreline processes and beach morphology that were sud- denly initiated by the earthquake were similar to those ordinarily caused by gradual changes in sea level operating over hundreds of years, while other more readily visible changes were sim- ilar to some of the effects of great but short-lived storms. Phenomena became available for observation within a few hours which would otherwise not have been available for many years. In the subsided areas—including the shorelines of the Kenai Peninsula, Kodiak Island, and Cook Inletflbeaches tended to flatten in gradient and to re- cede shoreward. Minor beach features One of the strongest earthquakes ever reported occurred in Alaska on March 27, 1964, at 5:36 pm. Alaska standard time. The epi- center was at Unakwik Inlet in Prince William Sound (fig. 1). The magnitude of the main shock was 8.4—8.6 on the Richter scale (Wood, 1966). The area of land and sea bottom affected by the earthquake is at least 70,000 square miles and may exceed 110,000. Forty thousand square miles was 1Geologist, Anchorage, Alaska: formerly Tidelands Supervisor, Alaska Department of Natural Resources, Division of Lands. By Kirk W. Stanley1 ABSTRACT were altered or destroyed on submerg- ence but began to reappear and to stabilize in their normal shapes within a few months after the earthquake. Frontal beach ridges migrated shore- ward and grew higher and wider than they were before. Along narrow beaches backed by bluffs, the relatively higher sea level led to Vigorous erosion of the bluff toes. Stream mouths were drowned and some were altered by seismic sea waves, but they adjusted within a few months to the new conditions. In the uplifted areas, generally around Prince William Sound, virtually all beaches were stranded out of reach of the sea. New beaches are gradually de- veloping to fit new sea levels, but the processes are slow, in part because the material on the lower parts of the old beaches is predominantly fine grained. Streams were lengthened in the emer- gent areas, and down cutting and bank erosion have increased. Except at Homer and a few small vil- lages, where groins, bulkheads, and [cobble-filled baskets were installed, there has been little attempt to protect the postearthquake shorelines. The few structures that were built have been only partially successful because there was too little time to study the habits of the new shore features and to design appropriate protection measures. Emer- gence of large areas that were once below water and permanent submer— gence of once-useful land areas have led to many problems of land use and ownership in addition to the destruction or relocation of Wildfowl, shellfish, and salmon habitats. INTRODUCTION lowered as much as 71/2 feet and 25,000 square miles was raised as much as 33 feet (Plafker, 1965, 1967). The coastlines affected by the earthquake are shown by figure 1. Definite limits of the coastal area of Alaska affected by the earth- quake have not been determined. Most authorities, however, agree that it is bounded by Yakataga (just southeast of the area shown in figure 1) on the east and the Kodiak group of islands on the southwest (Plafker, 1965). If all the shoreline irregularities of the mainland and the islands affected by the earthquake are included, the shoreline within the area ex— ceeds 10,000 nautical miles. Certain changes in beach forms occurred as a result of relative changes in sea level caused by up- lift or subsidence of the land dur- ing the earthquake. These changes were abrupt and thus cannot be unconditionally compared to a gradual change in sea level, but they did provide much informa- tion regarding normal shore proc- esses. Because sea—level changes were not only abrupt but also permanent, months were afforded for observations, which other- J1 J2 wise—as during storms would have been limited to hours or at the most several days. No attempt is made in this paper to present a detailed description of changes in beaches throughout either the submergent or emergent areas. A study of this type might have provided some interesting and descriptive material, but, con- sidering the great distances and the general remoteness of much of the shoreline involved, such a study would have been impractical. This paper is therefore restricted to descriptions of specific areas and ALASKA EARTHQUAKE, MARCH 27, problems. These limited observa— tions do not reflect all conditions and processes that occurred within the area affected by the earth- quake, but they do provide a basis for understanding the general processes that occurred. The chief objective of this report is to compare pre- and post-earth- quake processes. For this reason, Homer Spit is used as an example for much of the descriptive mat- ter; it happens to be one of the few places for which consider- able preearthquake information is 1964 available. Descriptions of coastal erosion primarily involve exam- ples along Cook Inlet because (1) much of the shoreline there is erod- ing and (2) the area is more densely populated than elsewhere. The earthquake not only caused physical changes which have left their marks upon the beach; it also had certain significant economic and legal consequences. These s0— ciological effects have an intercon— nection with the physical effects and are considered worthy of mention. SHORELINES The line of zero land movement, (fig. 1) trends southwest from the epicenter on Unakwik Inlet at the head of Prince William Sound, along the east shore of Kenai Pe- ninsula, to Kodiak Island. West of this line the land subsided; to the east it was uplifted. Anoma- lous areas of submergence, particu- larly along deltas and at the heads of bays, occur in both regions and were caused by compaction and settlement of sediments. Compac— tion of sediments in the uplifted area reduced the overall upward change in some localities; compac- tion in the subsided areas accen- tuated submergence. The landmass bordering the coastal area affected by the earth- quake is mountainous, and both it and the shoreline are characterized by glacial or periglacial features. The shoreline from Yakataga northwestward to the Copper River Delta is characterized by long sweeping beaches broken by wide-mouthed rivers and resem— bles a ria coast (Lobeck, 1939). The shoreline features were formed in part of outwash de- posits from the Bering and other large glaciers. Westward from the Copper River Delta the shoreline is glacially carved and is highly irregular, deeply incised, and fringed by numerous offshore rocks and reefs. The offshore area, particularly between the mouth of the Copper River and Cordova, is shallow and has numerous shoals, barrier islands, and spits. Active glaciers occupy the heads of many bays in the region, particularly in Prince William Sound. Along Kodiak Island the shore- line is characterized by fiords in classic forms. The bays, particu- larly along the Shelikof Strait side of Kodiak Island, have deep de- pressions at their heads and sub- marine threshholds of either rock or unconsolidated material at their mouths—a feature that, ac— cording to Guilcher (1958, p. 160), denotes a true fiord. The walls of the bays are usually steep, and many headlands are characterized by cliffs. Beaches are generally poorly developed and are com- OF THE EARTHQUAKE-AFFECTED REGION posed of medium to coarse shingle. Sandy beaches and constructional costal forms occur along the heads of bays, however, and well- developed sand-shingle beaches tens of miles in length are present along the southwesternmost part of the island. Along the southern Kenai Pe- ninsula and the western part of Prince William Sound, the shore- line resembles that of Kodiak Is— land in its irregularity, although bayhead depressions and subma— rine rock ramparts at the bay mouths are less well developed. However, the presence of active glaciers along the shoreline indi- cates that many of the indenta- tions are true fiord's. Here also the walls of many of the bays are steep, and cliffs are more numerous than along Kodiak Island. Nar- row, relatively steep beaches con- sist predominantly of shingle. Sandy beaches occur along the sides and heads of inlets and bays but are less well developed than along the eastern shoreline of Prince William Sound. EFFECTS ON SHORE PROCESSES AND BEACH MORPHOLOGY J3 Contrasting sharply with the ir- regular shorelines of Kodiak Is- land and Prince William Sound is the more uniform one of Cook In- let. Cook Inlet extends into the mainland more than 175 miles and narrows and shallows towards its head. The backshore, particularly east and north of the inlet, con- sists of low gently rolling glacial outwash plains. Wave-cut bluffs as high as several hundred feet occur in that area, and the adjacent beaches are generally well devel- oped. At the head of Cook Inlet the waters are shallow, and broad silty tidal flats are common. South of Tuxedni Bay, on the west shore of Cook Inlet, the low coastal plain pinches out, and mountains rise abruptly from the sea. Much of the shoreline affected by the earthquake presents seem- ingly contradictory evidence of both uplift and subsidence—a con- tradiction that perhaps is not sur- prising when viewed in the light of the complex tectonic history of the region. The high Chugach Mountains, Saint Elias Mountains, and Fairweather Range are ob- vious manifestations of strong tec— tonic uplift. Raised beaches along some coasts indicate more recent uplift, also, but the drowned fiord- like character of much of the shore- line, combined with numerous off- shore islands, skerries, and reefs, suggest coastal subsidence and submergence. Reconnaissance studies of the displaced shorelines, paced by nu— merous radiocarbon dates, have brought out a general similarity between the pattern of earthquake displacements and the long-term trend of Holocene coastal emer- gence or submergence, as well as a remarkable widespread submer- gence during the past several cen- turies over much of the zone that was uplifted during the earth- quake, and at least part of the zone that subsided (Plafker and Rubin, 1967; Plafker, 1968, in press). Thus, according to Plafker, the tectonic movements that accompa- nied the earthquake were but one pulse in a long—continuing trend of diastrophic deformation that has resulted in regional emergence of parts of the continental margin, simultaneous submergence of the Kenai-Kodiak Mountains belt, and either relative stability or emer- gence along the shores of Cook In- let and parts of Shelikof Strait. COASTAL FEATURES AND EARTHQUAKE EFFECTS BEACHES The beach face is the area be- tween high and low water (fig. 2). It is an ever-changing feature, but normally the changes are subtle and become noticeable only during severe storms. During the earthquake of 1964, rapid changes in land elevation caused obvious changes in shore processes and beach-face morphol- ogy. These changes were compara- ble in magnitude to changes that normally are caused by centuries- lo-ng fluctuations in sea level—or, paradoxically, to sudden changes caused by severe storms. Thus, changes in the beach face follow— ing the earthquake are important both to the study of fluctuating sea level as it affects a beach and to the engineering problems that might be met along the beach face as a result of changes caused by storms. 298—580 0—68—2 Frontal ridge or berm —‘\\ High water |¢——-———Upper beach face FEET 10 l——————i O 30 FEET — Break in slope 20-200' \ Low water (“W P———————- Beach face ——-——————>{ 2.—Diagram illustrating beach features. CHANGES IN PROFILE AND GRADIENT Obvious changes in profile and gradient occurred along shingle beaches within the submergent areas. In those areas, changes were noticed within 1 week after the earthquake. The most noticeable change was a flattening of the gradient and a recession of the beach face. Many shingle beaches within the submergent areas range in gradi- ent from 1: 8 to 1:30 and are characterized by a break-in-slope 50—200 feet seaward of the high- water line. The break-in-slope marks the location where waves act longest at high tide (King, 1959). The beach face (fig. 2) shoreward of the break (termed the “upper beach face”) has a steeper gradi- ent and coarser material than does the lower beach face. The break- in-slope shown in figure 3 is typical of those on many shingle beaches in Alaska. J4 ALASKA EARTHQUAKE, MARCH 27, 1964 3.—Sharp break-in-slope between upper and lower beach faces extends from mid- foreground to midcenter of photograph, just to right of man. These features, here seen in Kachemak Bay at low tide, are common to many shingle beaches. After earthquake-caused sub- sidence, waves reached higher on ' the beach face and caused severe scour erosion of the upper face; the break—in—slope g r a d u a l l y shifted shoreward. The upper limit of the swash was also ex- tended at some places, and the frontal ridge, or berm, was over— flowed. Part of the material eroded from the beach face was thus car— ried by the swash over the frontal ridge and was deposited along the landward side. A good example of profile and gradient changes along a shingle beach within the submergent areas was afforded by beaches along Homer Spit. Adjustment of the Homer Spit beaches to the new high-water level was gradual but can be described as occurring in three stages (fig. ‘4). The first stage was characterized by severe ero- sion and planing off of the beach face and crest by wave and swash overwash. Part of the eroded material was carried across the beach crest and onto the spit. The second stage was characterized by active erosion along the break-in- slope and the development of a new postearthquake frontal ridge. Dur— ing this stage the beach became noticeably convex upward because material from near the break-in— slope moved toward the upper beach face. The third stage was characterized by development of a well-defined break-in-slope that had gradually shifted landward. Erosion and recession of the beach face continued until the slope ap- proached the preeart'hquake gra— dient. During the third stage the new frontal ridge increased n0- ticeably in height and eventually retarded overflow. During this stage also, some of the material carried by the longshore drift be- gan to accumulate along the lower beach face near the break-in—slope. The observed erosional and deposi- tional effects of a relatively raised sea level on the Homer Spit beaches are in agreement with sug- gestions made originallyby Bruun (1962) and partly tested by Schwartz (1965) with small-scale laboratory experiments. Bruun concluded in part that a raised sea level is followed by shoreward dis- placement of the beach profile as the upper beach is eroded and that the amount of material eroded from the upper beach is equal in volume to that deposited on the nearshore bottom. Thus the rise of the nearshore bottom that results from this deposition will ulti- mately equal the original rise in relative sea level. Along beaches protected from severe storm waves, adjustment to subsidence difl'ered somewhat from that on exposed shingle beaches such as Homer Spit. Within the protected areas, for example those along the south side of Kachemak Bay and certain shores of Kodiak Island, recession of the beach face was more uniform and the convex- upward profiles were less notice- able because wave action was less turbulent. Erosion along the new break-in-slope and swash action along the beach crest were also less severe. Thus, the rate of beach-face recession along the beaches not acted on by large waves was more uniform and le% rapid than along exposed beaches. Even less noticeable changes oc- curred along sandy beaches, prob- ably because their gradient is usu- ally flatter than that of shingle beaches and because sand is less readily moved by wave action than is shingle. One area of subsidence where sandy silty beaches predominate is upper Cook Inlet. That region is characterized by an estuarine en- vironment where the beach mate- rial consists largely of silt-sized particles derived from glacier-fed streams (Karlstrom, 1964). Storm waves seldom exceed 4 feet in height. The beach gradient is as low as 1 2500 and, because the beach is acted upon by a tidal range of 35 feet, silty mud flats several EFFECTS ON SHORE PROCESSES AND BEACH MORPHOLOGY J5 lst stage Postearthquake HWL swash overfl0w Preearthquake HWL 2d stage Face convex upward 3d stage Postearthquake HWL Postea rthquake break-in-slope frontal ridge //‘~\ / \ / \\ Planing-off of face and crest Frontal ridge Fully developed postearthouake EXPLANATION Dashed line represents preearthquake frontal ridge Solid line represents beach profile at various postearthquake stages HWL High water line 4.—Sketch of three stages that characterized the postearthquake changes of beach configuration in the submergent areas. Sketch is based on behavior of the beach on the Cook Inlet side of Homer Spit, but applies to many other beaches. miles wide are exposed during low water. Subsidence of the region was not uniform but was as much as 2 feet. Some beaches examined shortly after the quake showed no noticeable change in profile or gradient, but by mid-1967, 3 years after the earthquake, subtle changes in profile had occurred. The action of shore processes following subsidence of the coastal region and the resultant effects on the beaches were similar to the effects of a severe storm. However the changes resulting from subsi- dence must be measured in years, whereas the maximum destructive action of a storm is usually meas- ured in hours. Within the emergent areas, as along the Copper River Delta, the profile and gradient of the up- lifted beaches remain unaffected with respect to wave action. The changes that have taken place are related to abandonment and stranding 0f the former beach faces above high water, and to ex- posure to the normal processes of subaerial erosion. In time, of course, new beaches will develop below the abandoned ones to fit the postearthquake high-water lines. MINOR BEACH FEATURES Minor beach forms including cusps, small beach ridges, and steps began to reappear on all submer- gent beaches within a few weeks after the earthquake. The new beach cusps were generally poorly developed and irregularly spaced upon the beach. Some of the hol— lows were overly large—being 30— 4:0 feet, wide as compared with pre— earthquake forms 5—15 feet wide. The outlines of the forms were vague, and at some places one horn was two to three times longer than the other. In all known examples, however, when the beach face itself began to revert to the preearth- quake form, the cusps also began to develop gradually into pre- earthquake forms. On beaches where the gradient was lowered by initial postearth- quake processes, cusps did not re- appear until the gradient had steepened, possibly because lack of mobility of the materials retarded the formation of the beach cusps. On beaches where the gradient was initially steepened after the earth- quake, p o o r 1 y defined cusps formed early, that is, within 3 months after submergence. These cusps were alternately destroyed and rebuilt by large storm waves, however—a process that suggests that beach cusps will not form, or at least will not persist, if the gradient of the beach flattens be- low a critical gradient. Along Homer Spit, this gradient appar- ently is about 1 :20 for shingle beaches. Prior to the earthquake, small beach ridges and steps occurred along the upper beach face of most shingle beaches. Submergence de- stroyed or greatly altered both features. However, the ridges and steps began to develop in approxi- mately the same location on the postearthquake beach face about 3 . months after subsidence. Unlike J6 beach cusps, the ridges and steps became stable within 1 year after submergence. LOW-WATER FEATURES The most prominent low-water features on beaches are ridges, run- nels, and submarine sandbars. All three forms were modified by the earthquake, and the changes ob- served give some insight into the movement of material and the de- velopment of such forms. RIDGE AND RUNNEL Ridges and runnels occur along most beaches of low gradient where sand is available and where the tide range is large enough to expose several hundred feet of beach face at low water. Good ex— amples are found along many beaches of Alaska but they are especially conspicuous at Cook Inlet. The ridges are composed chiefly of sandy material; the runnels, or troughs, are floored with gravel. The runnels provide channels which drain the beach on the ebb— ing tide. King (1959) suggests that a correlation can be made be— tween the most persistent ridges and the position at which the tide will stand for the longest period during the tidal cycle. The ridges along the Alaska coast are usually fairly stable, particularly those that are alined parallel to the coast and perpendicular to the direction of the dominant wave approach. In profile the normal sand ridge is asymmetrical, not unlike a rip— ple mark, but on a much larger scale; seaward—facing slopes are steeper than shoreward-facing ones. In general, individual ridges near the low-water line are higher than those farther up the beach face. In the areas of land subsidence, a noticeable scouring of sand ALASKA EARTHQUAKE, MARCH 27, occurred in the runnels within a few weeks after the earthquake, and at numerous places exposed a much coarser bed material. During the same period the ridges became rounded and more symmetrical in profile. The heights of many ridges increased a foot or more. At many places where the preearthquake ridges were hard enough to sup- port the weight of a man without appreciable indentation, the post- subsidence ridges were soft. Ap— proximately 30 days after the earthquake, many of the ridges had widened, some to as much as several hundred feet, from previ- ous widths of a few scores of feet. During the widening process the adjacent runnels were partly filled, and low basins were left in some areas to serve as drainage chan- nels. Ridges and runnels that were particularly conspicuous prior to the earthquake were modified to broad, somewhat undulating sand flats without definitely recogniz- able features. The landward shifting of the ridges was not a simple process of individual ridge migration. Instead, the ridges were first rounded then widened, and finally were coalesced—a process which often obscured and obliterated the intervening runnels. No appreci- able seaward migration of mate— rial occurred. The predominant landward migration seems to bear out King’s statements (1959) that seaward of the plunge zone, or break—in-slope of the beach face, there is a definite landward migra- tion of material. One year after the earthquake, most of the ridges and runnels had stabilized in approximately the same configurations as those be- fore the earthquake. The landward migration of the ridges apparently is not a simple process of removal from the sea- ward side of the ridges and deposi— 19‘64 tion on the landward side, such as that by which a new frontal ridge forms along the crest of a beach. Instead, the process seems to be one of flattening and spreading of the material followed by a land- ward movement of material en masse. As the landward migration of sand is slowed by its deposition at a higher elevation on the beach, runnels begin to form. ‘Vater draining from the beach contrib— utes to the process by eroding sand from the channels. The eroded sand is thereafter transported to the lower beach face where part of it is eventually carried down- beach by littoral currents. For this reason there was a persistent me- andering and relocation of the runnel courses during the first year following subsidence, some- thing that had not been observed before the earthquake. The mean- dering process is a result of the larger quantity of available sand and its increased movement along the low—water areas; part of the sand was deposited by the runnels themselves. 011 new beaches within the up- lifted areas, particularly along the Copper River Delta, incipient ridges and runnels became notice- able about a year after the earth- quake. They were poorly devel- oped, perhaps because the newly formed lower foreshore had not yet stabilized with respect to gra— dient and profile. In part, material in the ridges was transported land- ward and deposited on the upper beach face or on frontal ridges along the new high-water line. A few preearthquake sand ridges and runnels were partly stranded above the high-water line and so were altered or destroyed by storm waves. SUBMARINE SAND BARS Limited observations indicate some relocation of submarine bars. EFFECTS This relocation is suggested by previously reported hard bottoms in anchorages that, after the earth— quake, were composed of soft sand. This condition was particularly noticeable along Homer Spit where scuba diving showed that preearthquake areas of hard- packed sand were characterized by soft loose sand within a year after the earthquake. In some areas, particularly in Prince William Sound, submarine sand bars apparently were altered or destroyed by seismic sea waves or by local waves. G D. Hanna (written commun., 1965) states that soft sediments were scoured from many shallow bottom areas. Elsewhere submarine bars were al- tered even though tsunami effects were not evident along the shore- line. MODERN BEACH RIDGES Beach ridges (also referred to as “beach storm berms”) occur along many of the beaches of Alaska; in the submergent areas many of these ridges were altered during the earthquake. Observations since the earthquake have provided in- formation on the formation of beach ridges as well as their adj ust— ment to relatively higher sea level. Beach ridges of south-central Alaska are more perfectly formed along shingle beaches than along sand beaches where most are small— er and less well developed. Most ‘ constructional coastal forms, such as cuspate forelands, spits, and tombolos, are formed and enlarged by the development of successive beach ridge-s along the shoreline. The usual beach ridge consists of a mound or windrowlike deposit along the beach immediately above the high-water line. King (1959, p. 353) considers the ridges to be the product of steep storm waves which throw debris above the reach of normal waves. The larger ridges along the Alaska coast are composed of shingle, often with minor amounts of sand. A beach ridge along the present shoreline is generally referred to as a modern or frontal ridge, whereas the ridges farther inshore are referred to as ancient or old ridges. Most. individual ridges are 5—6 feet above high water and 8—10 feet wide, but some are as high as 20 feet and their bases may be as wide as 200 feet. Most ridges even— tually become stabilized by vege— tation that establishes itself on the landward slope. Frontal ridges grow gradually in both height and width by the deposition of material carried onto and across the ridge crest by waves and swashes during storms, As the ridges increase in height, the abil- ity of the overwash to carry debris across them decreases. Thereafter material will accumulate along the seaward side of the ridge; as the accumulation of debris progresses, a second ridge, seaward of the earlier one, will begin to form. The entire process may then be repeated. The time span between develop— ment of successive ridges is vari- able, being about a year or less along certain narrow stable beaches ( a “stable” beach being de— fined as one not enlarging seaward or retreating landward) but sev- eral hundred years or more along such large forms as cuspate fore- lands and spits. It is generally accepted that coastal forms, such as cuspate forelands and spits, widen sea— ward by the development of suc- cessive parallel and subparallel beach ridges. The landward parts of such coastal forms are usually characterized by a series of old vegetatirm-covered beach ridges that represent former shorelines that trend parallel or subparallel to the present shoreline. 0N SHORE PROCESSES AND BEACH MORPHOLOGY J7 RESPONSE TO SUBSIDENCE AND UPLIFT Changes in beach ridges occurred in many submergent areas. The magnitude of the changes depend- ed on (1) the extent to which the land was submerged, (2) the geo- graphic lo'cation of the beach with respect to storm waves, (3) the slope of the beach face, and (4) the type of beach material. In shel- tered areas where submergence was only about 1—2 feet, changes were minor, but in areas of greater sub- sidence and exposed sea conditions the changes were often major. In some areas, such as along Homer Split, subsidence caused waves to reach as much as 6 feet higher on the beach face than be- fore the earthquake, and swashes overflowed the crest of the frontal ridge. Material along the upper beach face, which was formerly above all but the highest waves and swashes, was scoured and erod- ed. Part of the eroded material was carried by the swash across the crest of the frontal ridge and de— posited along the landward slope (fig. 5). As overflow of the frontal ridges continued, material eroded from the beach face, including the face of the frontal ridge, was carried onto the backshore and deposited there. Continued overflow and erosion of the preearthquake fron— tal ridges reduced the crest height and caused the eroded material to spread out along the backshore. The beaches along Homer Spit (Stanley, in Waller, 1966a) are good examples of postearthquake beach—ridge development. lVithin approximately 30 days after the earthquake, the beach face in some areas had receded as much as 15 feet, but the frontal ridges had i11- creased as much as 30 feet in width and 2—3 feet in height. As the fron- tal ridges widened, coarse debris ALASKA EARTHQUAKE, MARCH 27, b.—Overflow of the crests of some beaches caused material to be carried and spread out along the backshore, left half of picture. Gradually the material accumulated into a new and higher frontal ridge. View along Homer Spit, looking southeast. was deposited near the crest and the finer material was carried to- ward the backsh'ore. Along the crest and the seaward side of the ridge, the material was poorly sorted because of the severe scour. On the landward slope of the fron- tal ridges, the sorting was gen- erally better and the grain size decreased. Along the extreme landward edge of the frontal ridge, the overflow material was delicately layered in fanlike forms and consisted of silt, sand, and pea- sized gravel. Continued overflow and the transfer of material from one side of the frontal ridges to the other caused the ridges to migrate to- ward the backshore. The transfer of material from one location to the other led to a net loss along the seaward side and a net gain along the landward side. Thus the process of landward ridge migra- tion was accomplished ‘by the transfer of material within the ridge proper. During overflow of the frontal ridges a greater volume of coarse material was deposited along the crest than along the landward side; the height of the ridges thus grad- ually increased. When the ridges had increased sufficiently in height to prevent overflow, migration stopped and the ridges stabilized. Observations of the migration of beach ridges after the earth- quake suggest an explanation as to why some ridges appear to migrate even under conditions not necessarily related to a rise in sea level. As stated above, beach ridges along Homer Spit migrated land— ward under conditions of rela- tively higher sea level, the migra- tion being brought about by the process of (overflow. Migration of the ridges ceased, however, when the width of the ridge was sulfi- cient to further the heightening by deposition of material along the crest. It is concluded, therefore, that a beach ridge will stabilize if there is sufficient backshore area upon which the overflow material 'an accumulate. If a barrier beach fronts a lagoon, then there is no platform 1964 for buildup of material, and land- ward migration of the beach ridge may continue indefinitely. This condition is exemplified by the barrier beach at Breving Lagoon, on the Seward Peninsula, far from the area where shorelines were af- fected by the earthquake (fig. 6). This barrier beach has a maximum width of 500 feet and is separated from the mainland by a shallow lagoon as much as a mile across. The beach is migrating landward, apparently because of storm-wave action and the resultant process of overflow. Material eroded by storm waves from the seaward side and transferred by overflow to the landward side of the barrier beach is carried into, and spread out on the floor of, the lagoon. Because the lagoon affords no plat-form on which overflow material can ac— cumulate there can ‘be no apprecia- ble widening of the barrier beach, and therefore, the height of the beach cannot be increased. Hence under normal storm conditions the beach will continue to migrate landward; the landward migration is not dependent on a. rise in sea level. Unlike the large ridges along shingle beaches at such places as Homer Spit, beach ridges along sandy beaches on the east shore of Cook Inlet were destroyed in many areas, particularly along the toes of low bluffs. In such areas the widths of beaches between the high—water line and the bluff toes prior to the earthquake ranged from 25 to 150 feet. After sub- sidence the beach widths were de- creased and the small beach ridges—those from 1 to -l feet high—were destroyed. The rela— tively narrow width of such beaches prevented overflow ma- terial from accumulating, and thus the beach ridges could not re— treat landvard when acted 011 by a sea level as much as 2 feet higher EFFECTS ON SHORE PROCESSES AND BEACH MORPHOLOGY J9 6.—Part of the barrier beach at Breving Lagoon, Seward Peninsula. Lagoon is north of (above) beach; Bering Sea is south of (below) beach. The lagoonward migration of the barrier beach is shown by superposition of the overflow upon the cuspate spit. Most beaches in the earthquake-affected part of Alaska migrated landward only short distances because they were backed by land platforms that allowed overwash material to accumulate. Photograph by Alaska Highway Department. (fig, 7), Thus, the material, i11— 7.—Submergence caused the high-water line to shift to a higher elevation on this beach on the eastern shore of Cook Inlet. The establishment of a new high—water stead of accumulating higher on . ‘ line caused undercutting and sloughmg of bluffs. Note absence of beach ridges. the beach (as it did along Homer Spit) was dissipated by wave action and was transported away from the site by the longshore drift. As the bluifs in such areas erode and recede, the beaches will eventually regain their preearth— quake widths, and new beach ridges similar to those that existed before the earthquake should develop. In theuplitted areas the Old frontal ridges are abandoned and stranded above high \‘ater. New ridges have formed in some places, and rather rapidly. In most areas, however, the growth of new ridges has been slow, probably because the predominantly fine material J10 along the preearthquake lower beach face does not form stable beach ridges as readily as does shingle. The old ridges in the up- lifted areas will gradually be covered by vegetation and will then assume the character of the typical old beach ridges of many coastal areas. New ridges along the lower high-water line will begin to develop as material continues to be worked by waves. However, the new ridges probably will require several years to stabilize, mainly because of a lack of coarse source material. ANCIENT BEACH RIDGES Postcarthquake studies of sub— sided coastal forms afford an eX— planation for the fact that the crests of some ancient beach ridges are uniformly lower than the mod— ern or frontal ridge. In some areas this condition is sufficiently pro- nounced to result in a basin—shaped area between the frontal ridge and the mainland. This habit of Alaska beach ridges has also been noted elsewhere (Bird, 1964; Johnson, 1919). Johnson (1919), Fisher (1955), and Zenkovitch (1959) attribute the relatively lower crest elevation of the older ridges to a continuous rise in sea level whereby each suc- cessive frontal ridge builds higher. This is well illustrated along Homer Spit; as a result of the earthquake-caused rise in relative sea level, the frontal ridge here has been built higher than the older ridges. If a rise in sea level is the only factor at work, however, the profile between the frontal ridge and mainland should have a gentle landward slope rather than the commonly observed basin-shaped profile. Studies by the author after the earthquake suggest that c0mpac~ tion and settlement of sediments ALASKA EARTHQUAKE, MARCH 27, by seismic shaking may also con- tribute to a change of coastal forms. The 1964 earthquake caused compaction and settlement of sedi- ments in many parts of south-cen- tral Alaska (Kachadoorian, 1965; Coulter and Migliaccio, 1966; Wal- ler, 1966a). Coastal forms must have been subjected to similar com- paction processes many times: During the past 50 years that offi— cial records have been kept, hun- dreds of earthquakes have occurred along the Alaska coast (US. Coast and Geodetic Survey, 1964, p. 23). As new beach ridges develop, the coastal form widens seaward with a corresponding increase in thickness of the column of sedi- ment. The coarsest material is usu- ally near the mainland and the smaller particles along the seaward side. If the mass of this sediment is repeatedly subjected to seismic and microseismic shocks, the sedi- ments—particularly the finer par- ticles—will lose bearing capacity and settlement will follow (Ter- zaghi and Peck, 1948). In some kinds of sediments, vibration can lead to spontaneous liquefaction which would cause additional set- tlement. Even though the coastal form nearest the mainland is oldest and hence has been subjected longest to seismic action, the sediment lay— er there is thinnest and the size of the particles is coarsest. Compac- tion by seismic shaking, therefore, would be less in such areas than in the seaward part of the land- form where the sediment is thicker and the particle size is smaller. Although the modern frontal ridge has been subjected to fewer earthquakes than have the ancient ridges, sediment compaction does occur there also. However, any de- crease in height caused by compac— tion would perhaps be offset by the addition of new material brought in by waves. Maximum decrease in 1964 height from compaction and set- tlement, therefore, probably would occur between the modern frontal ridge and the mainland and would lead to the typical basinlike profile that characterizes so many coastal constructional forms along the south-central coast of Alaska. STREAM-MOUTH CHANGES Stream mouths were changed throughout the areas of uplift and subsidence. In uplifted areas, streams were lengthened and in- cised into the elevated beach face. In the submergent areas, streams were shortened and drowned. SUBMERGENT AREAS Stream mouths were drowned throughout the submergent areas. Maximum drowning occurred mostly at the mouths of low-gra- dient streams that flowed across low-flying backshores composed of water—laid sediments. “chin such areas, subsidence is attributed both to tectonic movement and compac— tion of the sediments (Kachadoori— an and Plafker, 1967, p. F27; Ka- chadoorian, 1965, p. B2), and the extent of stream drowning is a function of both. The most notable examples of stream-mouth drowning are along the shores of Kodiak Island and the southern Kenai Peninsula where subsidence of the land was 5 feet or more in some locations (Alaska Dept. Fish and Game, 1965). Several excellent. photo- graphs of drowned streams on Kodiak Island are shown in a re- port on that area by Plafker and Kachadoorian (1966). During the earthquake many bays along Kodiak Island were hit by seismic sea waves. Spits, bay—mouth bars, and barrier beaches at or near stream mouths were altered or destroyed. Some of the lower of these landforms were EFFECTS overwashed, eroded, and reduced in height. New outlets formed where the tsunam‘is tore channels through ridges and spits. In all examples known to the writer, stream mouths that were acted upon by tsunamis were widened by scour and erosion. Since the earthquake, erosion by normal wave and swash overflow has gradually caused many coastal features such as spits and barrier beaches to flatten in profile and recede landward. Thus, the effect of the tsunami and Of normal wave action has reduced many once—con— spicuous forms to flattened, poorly defined deltalike features. In some areas of active shore drifting, material eroded by wave action from the stream mouths has re— plenished the beach along the downdrift side. In areas of weak or inactive shore drifting the eroded material is merely spread out on the lower beach face. In certain streams of low gradi- ent and velocity, the shores were subjected to strong wave action when the land subsided, and mate- rial eroded from the unconsoli— dated deposits along their banks, such as deltas and outwash fans, was transported by waves into the stream mouths. This process is par- ticularly noticeable along Turn— again Arm in upper Cook Inlet, where mud is accun’iulating in drowned stream mouths. In some areas the drowning of streams has decreased the supply of material for natural beach nourishment. By reducing stream gradients, submergence has led to prograding rather than degrading of channels. Stream erosion there fore is less effective than it was before the earthquake, and the re- duced quantity of debris *arried by the streams diminishes the quantity of material a *ailable to the longshore drift. The extent to which the beaches will be affected SHORE PROCESSES AND BEACH MORPHOLOGY J11, 8.—Tectonic uplift created new lands permanently above high water, such as this area along the Copper River Delta. View at high tide, several months after the earth- quake. The preearthquake high-water line is shown approximately by the dashed line. Photograph by US. Forest Service. by the decreased supply of mate— rial is not yet clear; however, be- cause of a decrease in stream—borne material, because wider stream mouths will retard longshore drifting and by-passing of mate— rial, and because of continuing wave action, beach erosion prob— ably will characterize many drowned stream—mouth areas for many years to come. UPLIFTED AREAS Changes at the mouths of streams in the uplifted areas are related to stream-course lengthen- ing and downcutting (fig. 8). These changes range from minor alterations along shorelines of little uplift to major changes along gently easily eroded beaches in areas that were appre— ciably uplifted. The effects of tsunamis on stream mouths in the uplifted areas varied widely. In many places, sand and silt were scoured sloping, from the mouths and carried with other debris into the upper reaches of the streams. Such deposit-ion caused considerable silting in stream mouths and channels, and in some streams caused a tempo- rary damming or blocking. Spits and barrier ridges at stream months were generally altered by the tsunamis. Many of the lower beach forms were appreciably changed, and material was redis- tributed along the beach into the stream mouth. Such redistribution of the material even obliterated some stream courses. Along appreciably uplifted shorelines composed of sand and silt, rapid gullying occurred at stream mouths (fig. 9). Along the Copper River Delta, an area of readily eroded sand and silt that was uplifted several feet, gullying commenced within hours after up- lift (Reimnitz and Marshall, 1965). Rapid headward erosion produced small waterfalls 1—2 feet high. J12 ALAsKA EARTHQUAKE, MARCH 27, 9.—Rapid gullying and sloughing caused by headward stream erosion along the uplifted coast on the Copper River Delta. Photograph by US. Forest Service. Along beaches of mixed sand and shingle, stream adjustments were slower. Small riflles formed at stream mouths along such beaches. Kirkby and Kirkby (1968) de— scribed in detail stream-mouth ad- justment along elevated intertidal zones composed of sand and shin- gle at Montague Island, where up— lift was as much as 33 feet. As on the Copper River Delta, stream ad- justment began immediately after uplift had occurred. Along some elevated shingle beaches, streams disappeared into the gravel of the uplifted inter- tidal zone. Generally this condition was temporary, and after several months the streams developed new courses across the intertidal zone. Most uplifted stream courses showed evidence of degradation and bank erosion within a few days after the earthquake. Slumping of bank material and debris deposited by seismic sea waves also contrib- uted material. In addition, after spring breakup, the normal in- crease of streamflow made deposi— tion of material all the greater. The overall result was an increase in material carried by the streams and deposited at their mouths. Along some stream mouths the newly deposited material has re— tarded channel development and has led to meandering. Along the larger streams the process has been mainly silting and dissection of newly formed deltas and spits. Although evidence indicates an increase in sediment carried by the streams, the length of time the increased load was carried prob- ably was short. By the fall of 1964, 6 months after the earthquake, the stream—carried sediment load was normal (Waller, 1966b). COASTAL EROSION AND MOVEMENT OF MATERIAL EROSION Coastal erosion is active along all shorelines within the area af- fected by the earthquake except on the rocky platforms. The shore- lines most affected are those com- posed of unconsolidated material within the submergent areas. Of the shorelines modified by 1964 wave erosion, none were affected more than the east shore of Cook Inlet from Kaehemak Bay to Turnagain Arm—a distance of more than 170 miles (fig. 1). There the entire shoreline is eroding. As already mentioned, the shoreline is a relatively uniform sandy beach with slopes as low as 1:300. The tidal range is about 22 feet along the southern section but increases to more than 30 feet near Turn- again Arm. Because of the low beach gradient and high tidal range, several thousand feet of tidelands are exposed at low water. The beach is backed by a line of bluffs about 200 feet high in most places, but locally as high as 600 feet in the Homer area. Most of the bluffs are of unsorted glacial material that is easily eroded. Prior to the earthquake, wave action was undercutting the bluffs in many areas. Most of the sandy fraction of the sloughed material drifted away, but part of the coarser material remained along the high-water line in the form of shingle on beach ridges. The ridges, which were as high as 8 feet, protected the toes of the bluffs from all but the larger storm waves. During the earthquake the shoreline subsided as much as 31/2 feet along the southern section near Homer and as much as 1 foot near Point Possession at the mouth of Turnagain Arm. Prior to sub- mergence the width of beach be- tween the high-water line and the toe of the bluff was generally less than 100 feet, and in some areas less than 20 feet. Submergcnce de- creased these widths and, after the earthquake, storm waves, particu- larly along the southern section, acted directly upon the toes of the bluffs. Many beach ridges that had previously protected the toes of the bluffs were exposed to waves of even, moderate height (fig. 10). EFFECTS ON SHORE 10.-—Undercutting and sloughing Of bluffs caused by tectonic subsidence, eastern shore of Cook Inlet. Driftwood along toe Of bluff indicates the position Of post- earthquake high-water line. ”3LL a: I 11.—Serious bluff erosion caused by subsidence along the waterfront of the town Of Kenai. View looking northwestward, about 30 days after the earthquake. By 1967, 3 years after the earth- quake, the bluff line in some areas had receded sufficiently to provide a width Of beach that afforded pro- tection from all but the larger storm waves. Elsewhere inter- mediate-size storm waves still (1967) erode the toes Of the bluffs. l'ntil the bluff line recedes enough to provide a sufficient beach to pre- vent wave runup to the toe, erosion and recession will continue. PROCESSES AND BEACH MORPHOLOGY J13 The two communities along this shoreline most seriously affected by blufl' erosion were Kenai and Homer. Kenai is at the mouth Of the Kenai River, a tidal estuary. During the earthquake the area subsided 12—18 inches. South Of the town, bluffs as much as 150 feet high consist Of readily eroded glacial material and alluvium. After regional subsidence, the preearthquake accumulation of sloughed debris along the toe Of the bluffs was quickly removed. Undercutting by waves and by the river began a few days after the earthquake, and within 3 months the bluffs had receded as much as 20 feet (fig. 11). In the Homer area the regional subsidence was about 31/2 feet. Par- ticularly damaging wave erosion occurred in the Millers Landing area, toward the east end of Homer on Kachemak Bay, along a low line Of bluffs composed Of peat, sand, and silt. After the earth- quake, undercutting caused seri- ous sloughing; within 6 months, the bluff line had receded as much as 8 feet. SHORE PROTECTION MEASURES Except at Homer and a few small villages, no serious effort has been made to construct erosion— preventive works anywhere in the earthquake-affected area. Along the Cook Inlet side Of Homer Spit, erosion on the lee side Of the groin farthest downdrift became critical. Retreat Of the beach was jeopard- izing the spit highway. In July 1964, a seawall and two additional groins were constructed. Two hun— dred feet Of the wall was con- structed to a height of 1—2 feet above the high-water line, and one section was built 5 feet higher above the high-water line. Figure 12 shows the timber bulkhead and the additional groins a short time after construction. Figures 13 and 14 Show the same area 6 and 16 Jl4 ALASKA EARTHQUAKE, MARCH 27, 1964 12.—Timber seawall and new groms along west side of Homer Spit as constructed 3% months after earthquake-caused subsidence. 13.—Seawall shown in figure 12, approximately 6 months after construction. EFFECTS 14.—Seawa11 shown in figure 12 some 16 months after construction. Note that the bulkhead and one of the groins were completely destroyed during this period owing to their excessive height. 15.-—Cobble-filled wire-mesh fence under construction at Larsen Bay, Kodiak Island, shortly after the earthquake and resultant subsidence 0f 2 feet. Most of the fence was destroyed by a storm before it could be completed. months later, respectively. Sixteen months after construction, the high timber bulkhead between the fifth and sixth groin had been de- stroyed, as had the last groin (fig. 14). Similar high bulkheads were constructed in other areas but most were destroyed within several months, or at most within a year, after the earthquake. Failure is attributed to the fact that the bulkheads were too high and were 0N SHORE PROCESSES AND BEACH MORPHOLOGY J15 anchored in unstable beach gravel. Elsewhere wire-mesh baskets filled with cobbles were placed along the beach to serve as bulk— heads. Such baskets appeared to function well along Homer Spit, even though they were pounded by waves as high as 10 feet. Erosion, moveover, occurred on neither the updrift nor the downdrift side. A similar cobble-filled wire-mesh fence was started at Larsen Bay on Kodiak Island (fig. 15), but most of it was destroyed by storm waves before it could be com- pleted. Wave erosion ‘ Of landslides caused by the earthquake was rapid, particularly along Cook In- let where the frontal edges Of the Slides were greatly modified within 2 years. Many slides consisted Of sandy-silty material. Silty-mud beaches have developed gradually, particularly on the downdrift sides. Erosion within the uplifted areas is evident along the post- earthquake lower high-water line. Wave erosion along most of the uplifted area is not serious so far as loss Of land is concerned; most streams across elevated tidal flats are incising channels. LONGSHORE MATERIAL MOVEMENT No significant changes resulting from the earthquake are known or reported for the directional habit of the longshore drift. The changes that have occurred are only in the quantity of material carried. Along most shorelines affected ‘by the earthquake, more material has entered the longshore drift than before. The additional material began to enter the sea immediately after the earthquake. This fact was made Obvious along many shore- lines by a broad band of muddy water that was several miles wide J16 ALASKA EARTHQUAKE, MARCH 27, 1964 SW 2 Elev. 28.6 ft m FEET EXPLANATION E 10 ___._. Groin Downdrift, southeast side Updrift, northwest side 0 0 30 FEET I l | l l I l l I I l 33 52 67 80 96 125 150 166 184 196 216 MEASUREMENT POINTS 16.—Profiles of updrift and. downdrift fill of groin aIt Homer Spit, measured (A) on March 2, 1964, a few weeks prior to the earthquake; (B) on August 28, 1964, 4 months after submergence; and (C) on September 30, 1965, 18 months after submer- gence. The change in configuration of the groin fill was caused by submergence and the resultant higher level of water on the fill. Note, however, that the gradient at 18 months approximates the preearthquake gradient. EFFECTS ON SHORE in some areas, but no direct effect, such as the formation Of new berms by the deposition Of the ma- terial, was observed along beaches until several weeks after the earth- quake. A good example Of the effect of a delayed longshore drift is de- scribed by Stanley (in Waller, 1966a, p. D24). At Homer Spit the additional quantity of material that had entered the longshore drift in the source area, 2—6 miles west Of the spit, did not reach the spit until about 30 days after the earthquake. Accelerated shoreline erosion in the submergent areas caused more material to enter the drift than had been entering before the earth- quake. However, in the same areas, material contributed by rivers and streams decreased. The converse was true in the uplifted areas. What effect the change in balance of the source and supply of mate- rial will have on the character of the longshore drift is not yet known. In planning future coastal projects the possible changes in source and supply that have oc- curred since the earthquake must be taken into consideration. UPSLOPE MATERIAL MOVEMENT Only one limited study of ma— terial movement was made within the first year following the earth- quake (Stanley, in Waller, 1966a). Significant changes in material movement may therefore have passed unnoticed. The study was made by the Alaska Division of Lands along the west side of Homer Spit at a system of filled groins that had been studied prior to the earth— quake. After submergence the high-water line rose 3.5 feet. W’ith- in 5 days after submergence, 1-2 feet of material had accumulated along the landward end of the groins and had spread inland to a width of 30—50 feet. The newly de- posited material was first thought to have been carried in by the long— shore drift. However, the first pro- files made 7 days after submer- gence indicated that the intergroin fills had actually receded and flattened and all had undergone a net loss of material (fig. 16A). In— asmuch as the material lost from the intergroin fill area about equaled the newly deposited ma- terial along the landward end of the groins, the material probably was derived from the adjacent in— terfill area rather than from some distant source. Additional profiles were made at about 15-day intervals for the next 60 days (fig. 163, 0). Within the first 30 days the intergroin fllls receded as material continued to accumulate along the landward PROCESSES AND BEACH MORPHOLOGY J17 ends of the groins (fig. 163), but about 30 days after the earthquake the intergroin fills began to en- large, and thereafter the enlarge- ment increased noticeably. The material causing the increase prob- ably was brought there by the longshore drift. Observations indicate that dur- ing the 30-day period following the earthquake, the material de- posited along the beach crest came from the lower-upper beach face of the groin fills and was carried upslope. Erosion along the lower- upper beach face continued until a new profie of equilibrium had been established. The dominant move— ment of material during the 30— day period following subsidence was upslope because the waves were reaching higher up the beach face. “raves running higher up the beach face shorten the travel distance of the swash and thus in- crease its carrying capacity. The increased shoreline erosion throughout the area unquestion— ably increased the supply of ma— terial to the longshore drift; after about 30 days a noticeably in- creased quantity of material was drifting alongshore, but during this first 30—day period following submergence the dominant motion of material was up the beach rather than along it. BIOLOGIC EFFECTS OF SHORELINE CHANGES FISH At the time this report was written (1967), the most compre- hensive discussion of the effects of the earthquake on the Alaska fish: eries was the one compiled by the Alaska Department of Fish and Game (1965). Although a. com— plete analysis and assessment of damage to this resource has not yet been made, the environment and habitat of the salmon is known to have been drastically changed in some areas. Part of Prince “’illiam Sound was uplifted as much as 33 feet, and great changes in the environ- ment and habitat of pink and chum salmon resulted. Similarly, in the submergent areas, such as Kodiak Island, important inter— tidal spawning grounds were inundated. G. Y. Harry, Jr. (written com- mun., 1964), reported, 5 months after the earthquake, that the greatest damage to the salmon in Alaska was probably in Prince \Villiam Sound; here 75 percent of the pink and chum salmon pro- duction comes from the intertidal J18 spawning areas. Thorsteinson (1964) states that within Prince William Sound the runs of pink salmon range from 3.2 to 8.7 mil- lion fish and the chum salmon from 0.4 to 0.6 million fish. Compounding the permanent damage caused by uplift and sub- mergence of the intertidal spawn- ing grounds was the temporary effect of tsunamis and local waves. In parts of Prince William Sound, the waves caused considerable scour, and debris and silt were car- ried upstream for several hundred feet. Some biologists believe that many salmon eggs were scattered by the movement of debris. When, all contributing factors in Prince William Sound are taken into ac- count, the salmon loss caused by the earthquake and its subsidiary effects is thought to be about one- quarter of a million salmon ( Noeren‘berg and Ossiander, 1964). In the Kodiak Island area, where submergence was as much as 6 feet, widespread flooding of the intertidal spawning areas oc- curred. In some places flooding of the intertidal area was helpful to the salmon in that waterfalls which formerly obstructed their upstream migration were elimi- nated. Along other streams, sub- mergence and flooding increased the size of the intertidal area. Re- moval of waterfalls in some areas and enlargement of other areas, thus increased the size and avail- ability of certain intertidal salmon—spawning grounds. W. L. Sheridan (written commun., 1965) stated that the changes in the intertidal spawn- ing habitat and environment have not been fully assessed, but that future production will decrease. Although some intertidal spawn- ing areas have increased in size, Sheridan indicates that many of l ALASKA EARTHQUAKE, MARCH 27, 1964 the preearthquake lower areas are now characterized by excessive quantities of sediments that de— crease the survival rate of salmon eggs and alevins. SHELLFISH One of the more important habitats of shellfish, particularly the razor clam, is along the Copper River Delta. The delta was up- lifted as much as 8 feet, and, be- cause of the low offshore gradient, extensive intertidal areas \" ere permanently elevated above high water. This area supported a large population of razor clams. The widespread death of these and other bivalves throughout Prince William Sound W a S evident within a few weeks after the earth— quake (G. B. Haven, written commun., 1965). Many of the clam beds are now above high water and are lost, but clam beds formerly below lower low water we I‘B elevated and are now accessible to clam diggers. In some areas of Prince \Villiam Sound, particu- larly along the western part, sub— sidence at the ‘bayheads was caused by local compaction of the sedi~ ments. Consequently, clam beds which were once readily accessible are now from 2 to 4 feet below low water and out of reach to conven— tional methods of harvesting. There is no general consensus as to what long-term ecological effect. either submergence or emergence has had upon the general clam or bivalve population within the af— fected areas. Certainly the initial mortality rate was high——es- pecially in the uplifted areas—as was proved by the skeleton. In the submergent areas the greater depth of water now present over the outermost clam beds has also resulted in bivalve mortality. WILD F'OWL Ordinarily sub-mergence or emergence would not be considered to affect waterfowl adversely, but, according to P. E. K. Shepherd (written commun., 1966), there has been an indirect effect. In the feeding ground of the dusky Can- ada goose in the Copper River Delta, the preferred food of the goose is a type of vegetation refer- red to as “forb-grass.” This grass—actually a combination of herbs and grasses—apparently re- quires oecasional inundation by marine waters and usually grows along the sides of tidal sloughs and other areas regularly inundated by the highest tides. Because of changed tide levels, the forb-grass is dying, and before it is reestab- lished at a higher elevation, the dusky Canada goose is likely to have changed its habitat. Shep- herd indicates that the habitat of the goose may be further changed because its former nesting grounds will become overgrown with coni- fers or other plants when the salt is leached out of the soil. To some extent these same changes are af- fecting the habitats of the dab- bling and diving ducks and the trumpeter swans. Brakish—water lagoons and sloughs are particularly favored as nesting or feeding grounds by some wildfowl. U plift. drained some of these areas or made them less brackish because of the inflow of fresh water. Thus, changed en- vironments, whether caused by up— lift or submergence, may ulti- mately contribute to a decrease in the Wildfowl population in specific areas. Whether new wildfowl com- munities will spring up in areas made more favorable by the earth— quake will not be known for some years. EFFECTS ON SHORE PROCESSES AND BEACH MORPHOLOGY J19 EFFECTS ON PROPERTY VALUES AND MANMADE STRUCTURES Earthquake-caused uplift and subsidence—and the consequent changes in coastal processes—af- fected the legalities of land owner- ship, and also led to many changes in property values and necessita- ted reconstruction or relocation of many coastal installations. A few of the resultant problems are touched on here. SUBMERGENT AREAS As already mentioned, coastal erosion is occurring in many of the submergent areas, and in some areas—such as on the Kenai Pe- ninsula—the rate of bluff recession is several feet per year. In other areas, erosion may still become a problem although the process is not yet noticeable. One such area is along upper Cook Inlet where submergence was less than 2 feet. The dramatic effects seen in areas of greater submergence are not present in this area, but there is unquestionably a gradual retreat of the shoreline. This retreat is oc- curring in some of the more densely populated areas where land values are high. At Turnagain Arm, wave ero— sion has damaged road and rail- road embankments that were raised after the earthquake. “’ave erosion in that area will probably continue, and the embankment will require constant surveillance and maintenance. In Homer, wave action has already damaged sec— tions of the newly constructed Homer Spit highway. Some shore installations that escaped severe damage at the time of the earthquake were, because of the higher stand of the sea, endan- gered by flooding a few weeks after the earthquake when tides were high. Several cannery wharves and loading platforms had to be raised. In some places, such as at the Wakefield Cannery at Port. ‘Vakefield on Kodiak Is- land, complete relocation of the cannery was necessary. In some submergent areas, wave scour has moved large quantities of gravel under or away from can— neries and other facilities built over the high-water line. For ex— ample, a cannery at Larsen Bay on Kodiak Island, which was built, in a seemingly ideal location and was adequately protected before the earthquake has incurred consider- able storm damage since the earth— quake. Erosion of gravel from the adjoining beach has endangered some of the cannery buildings. EMERGENT AREAS Some of the immediate effects of emergence that necessitated economic and engineering consid- eration were related to small-boat harbors. In the Cordova area the land was elevated about 8 feet. There emergence necessitated dredging and enlarging the small- boat harbor. Some of the channels of the Copper River Delta that had formerly been accessible to fishing boats required considerable redredging in order to afford ac- cess after the earthquake. In some uplifted areas, individuals who formerly had access to the sea by shallow channels or tidal sloughs found it necessary to dredge their access routes. In other areas, land- ing facilities, such as docks and wharves, could no longer accom— modate deep—draft vessels and had to be enlarged or relocated. In Cordova, because of the 8 feet of emergence, it has also been necessary to relocate the installa- tions upon which barges, scows, and fishing boats are stored in off- season months. At the beginning of the 1964 fishing season, follow— ing the earthquake, some craft had to be pulled from their storage areas onto the tidal flats by tractors before they could be refloated. LEGAL PROBLEMS Because the State of Alaska has jurisdiction over the bottoms of all navigable waters, legal prob— lems have arisen wherever land raised or lowered by the earth- quake abuts navigable waters. In Alaska the boundary of tidal waters is the line of mean high tide. A person whose Alaskan prop- erty abuts the line of mean high tide is called an “upland owner.” By common—law principles the up— land owner enjoys all littoral rights and privileges afforded him by the location of his land. Such rights and privileges include, among others, free and unob- structed ingress and egress to the sea. The line of mean high tide thus is the legal boundary that sep- arates' the upland owner’s private property from the State-owned tidelands. The line of mean high tide remains fixed unless the shore- line is altered by accretion or ero- sion. If these processes result from natural causes and the shoreline changes are gradual and imper- ceptible, the legal boundary will change location to match the line of mean high tide. However, if the changes are natural but sudden in character and wholly perceptible, J20 the process is termed “avulsion,” and the legal boundary remains fixed even though the mean high tide line has been displaced. Movement of the ‘land during the earthquake was abrupt, certainly occurring within a matter of hours if not of minutes, and flooding or withdrawal of waters from the land was equally abrupt. On Sep— tember 14, 1964, the attorney gen- eral for the State of Alaska de- scribed this process as one of avulsion and therefore the legal boundary line—the line of mean. high tide—remains fixed as it was the instant before the earthquake ALASKA EARTHQUAKE, MARCH 27, 1964 (State of Alaska Attorney Gen. Opinion 6, Sept. 14, 1964). Along the uplifted coasts many private parcels of land abutted the line of mean \high tide before the earthquake; afterwards, however, the natural line of mean high tide was shifted seaward, in some places several hundred feet. Because withdrawal of the water has been defined as avulsi‘on, the private ownership cannot follow the re- treating line of mean high tide but must instead remain fixed to the line existing the instant before the earthquake. Thus those persons who enjoyed all the rights and privileges afforded them as upland owners before the earthquake now find themselves unable to exercise their littoral rights and privileges. The opposite has occurred in the submergent areas. There, owners of land that albutted the line of mean high tide prior to the earth- quake now find their boundary line (that is, the preearthquake line of mean high tide) some distance sea- ward of the natural mean high tide line. These persons in fact own tidelands but the lands are legally described as uplands. NEED FOR FURTHER STUDIES One danger in attempting to rectify coastal-erosion problems is that protective measures are often taken without a clear understand- ing of littoral conditions. This problem has already manifested itself at various localities where installations that were constructed to protect the land have in fact destroyed it. Some of the later damage to roads and emb-ankments may per- haps be explained by the rather hurried reconstruction efl‘ort im— mediately following the earth— quake. The necessity for haste probably did not allow for full consideration of design methods with respect to unnatural slopes; such slopes can promote wave ero- sion. Time did not always allow for adequate investigation of near- shore conditions, such as wave runup and current direction; con— sequently, adequate safeguards to insure maximum protection against coastal erosion were not designed. Certainly, in the submer— gent areas the sea has encroached upon the land establishing a new profile of equilibrium. When arti- ficial impediments are placed in Alaska Department of Fish and Game, 1965, P0st~earthquake fisheries evaluation; an interim report on the March 1964 earthquake effects on Alaska’s fishery resources: Juneau, Alaska, 72 p. Bird, E. C. F., 1964, Coastal landforms; an introduction to coastal geomor- phology with Australian examples: Canberra, Australian Natl. Univ, 193 p. REFERENCES CITED Bruun, Per, 1962, Sea—level rise as a cause of shore erosion: Am. Soc. Civil Eng. Proc, v. .88, paper 3065, Jour. Waterways and Harbors Div., no. WW 1, p. 117—130. Coulter, H. W., and Migliaccio, R. R., 1966, Effects of the earthquake of March 27, 1964, at Valdez, Alaska: US. G901. Survey Prof. Paper 542— C, p. Cl—C36. such areas, the result more often than not promotes rather than re- tards wave erosion. The installation and construc— tion of shore protective works will undoubtedly become necessary long before a profile of equilibrium develops and stabilizes—a profile which will itself eventually retard shoreline erosion. Therefore, the economic ramifications of coastal erosion not now readily noticeable may, within the next few years, be— come a problem which must be dealt with if valuable areas of land are to be saved. Fisher, R. L., 1955, Cuspate spits of St. Lawrence Island, Alaska: Jour. Geology, v. 63, no. 2, p. 133—142. Guilcher. Andre, 1958, Coastal and sub- marine morphology [translation by B. \V. Sparks and R. H. W. Kneese]: New York, John \Viley and Sons, Inc.. 274 1). Johnson, D. W., 1919, Shore processes and shoreline deveolpment: New EFFECTS York, John Wiley and Sons, Inc., 584 p. Kachadoorian, Reuben, 1965, Eflects of the earthquake of March 27, 1964, at Whittier, Alaska; US. Geol. Sur- vey Prof. Paper 542—B, p. B1—B21. Kachadoorian, Reuben, and Plafker, George, 1967,'Efiects of the earth- quake of March 27, 1964, on the communities of Kodiak and nearby islands: U.S. Geol. Survey Prof. Paper 542—F, p. F1—F41. Karlstrom, T. N. V., 1964, Quaternary geology of the Kenai Lowland and glacial history of the Cook Inlet region, Alaska: US Geol. Survey Prof. Paper 443, 69 p. King, C. A. M., 1959, Beaches and coasts : London, Edward Arnold and 00.,403 p. Kirkby, M. J., and Kirkiby, A. V., 1968, Erosion and deposition on a beach raised by the 1964 earthquake, Mon- tague Island, Alaska: US Geol. Survey Prof. Paper 543—H. (In press.) Lobeck, A. K., 1939, Geomorphology; an introduction to the study of landscapes: New York, McGraw~ Hill Book Co., 731 p. Noerenberg, W. H., and Ossiander, F. J ., 1964, Effects of the March 27, 1964, earthquake on pink salmon alevin survival in Prince William Sound spawning streams: Alaska Dept. Fish and Game Inf. Leaflet 43, 10 p. Plafker, George, 1965, Tectonic defor- mation associated with the 1964 Alaska earthquake: Science, v. 148, no. 3678, p. 1675—1687. Plafker,, George, 1967, Surface faults on Montague Island associated with the 1964 Alaska earthquake: U.S. Geol. Survey Prof. Paper 543—G, p. G1—G42. 1968, Tectonics of the March 27, 1964, Alaska earthquake: U.S. Geol. Survey Prof. Paper 543—1. (In press.) Plafker, George, and Kachadoorian, Reuben, 1966, Geologic effects of the March 1964 earthquake and as- sociated seismic sea waves on Ko- diak and nearby islands, Alaska: US. Geol. Survey Prof. Paper 543— D, p. D1—D46. Plafker, George, and Rubin, Meyer, 1967, Vertical tectonic displace- ments in south-central Alaska dur- ing and prior to the great 1964 earthquake: Jour. Geoscience, Osaka City Univ., v. 10, art. 1—7, p. 53—66. Reimnitz, Elk, and Marshall, N. F., 1965, Effects of the Alaska earth- quake and tsunami on recent deltaic sediments: Jour. Geophys. Re- search, v. 70, no. 10, p. 2363—2376. Schwartz, Maurice, 1965, Laboratory study of sea-level rise as a cause of shore erosion: Jour. Geology, v. 73, no. 3, p. 528—534. Stanley, K. W., and Grey, H. J., 1966, Spray-on paint stripes to determine the direction of beach drifting: Jour. Geology, v. 74, no. 3, p. 357— 361. ON SHORE PROCESSES AND BEACH MORPHOLOGY J21 Terzaghi, Karl, and Peck, R. B., 1948, Soil mechanics in engineering prac- tice: New York, John Wiley and Sons, Inc., 566 p. Thorsteinson. F. V., 1964, Effects of the Alaska earthquake on pink and chum salmon runs in Prince Wil— liam Sound: U.S. Bur. Commercial Fisheries Biol. La.b., Auke Bay, Alaska, 16 p. Twenhofel, XV. S., 1952, Recent shore- line changes along the Pacific coast of Alaska: A111. Jour. Sci, v. 250, no. 7, p. 523—548. l.'.S. Coast and Geodetic Survey, 1964, Prince William Sound, Alaskan earthquakes, March—April 1964: US. Coast and Ge‘od. Survey, Seis- mology I')iv., Prelim. Rept, 83 p. Waller, R. M., 1966a, effects of the earthquake of March 27, 1964, in the Homer area, Alaska, with a sec- tion on Beach changes on Homer Spit, by K. W. Stanley: U.S. Geol. Survey Prof. Paper 542—D, 1). D1— D28. 1966b, Effects of the March 1964 Alaska earthquake on the hydrol- ogy of south-central Alaska: U.S. Geol. Survey Prof. Paper 544—A, p. A1—A28. \Vood, F. J., ed., 1966—67, The Prince William Sound, Alaska, earthquake of 1964 and aftershocks : US. Coast and Geod. Survey Pub. 10.3, v. 1, 1966, 236 p. ; v. 2, pt. A, 1967, 391 p. Zenkovitch, V. l’., 1959, On the genesis of cuspate spits along lagoon shores: Jour. Geology, V. 67, no. 3, p. 269— 277.