Giant Waves in Lituya Bay, Alaska; USGS PP 354-C

Possible Causes of 1936 Waves

Steven Dutch, Natural and Applied Sciences, University of Wisconsin - Green Bay
First-time Visitors: Please visit Site Map and Disclaimer. Use "Back" to return here.



Nature And Cause Of The Waves

All eyewitness accounts agree that the 1936 waves were preceded by or accompanied by a loud noise, and the two most detailed accounts agree that there were three waves of increasing size in the vicinity of Cenotaph Island, with estimates of maximum height ranging from 100 to 250 feet. The account of one of the men on the boat, in a better position for observation than the men on the island, indicates that the roaring noise from the head of the bay was heard as much as half an hour before the first wave was sighted. This account also indicates that the waves were spaced about 2 minutes apart, and were followed by recession of the water below normal level. One observer on the island estimated the rate of water movement (not necessarily the speed of the waves) at about 23 miles per hour. The time required for the first wave to travel from near the head of the bay to the west side of Cenotaph Island, as estimated by one observer on the boat, gives a speed of about 22 miles per hour. Evidence preserved in the trimlines indicates that the waves were generated at or near the head of Crillon Inlet, where at least one of them dashed up on the valley wall to a height of 490 feet or more. The maximum height of the trimlines near Cenotaph Island is 24 feet, suggesting that the observers' estimates of the height of the waves at this position in the bay is too large.

Possibly significant in the consideration of the origin of the 1936 waves is the fact that they occurred "during a period of unusually heavy rainfall" (Williams, 1938, p. 18). Although two eyewitness accounts differ as to the weather on the day of the waves, the weather records at other places in southeastern Alaska do indicate that the occurrence of these waves was preceded by heavy rainfall (U.S. Weather Bureau, 1938). At the two coastal stations nearest Lituya Bay-Sitka (fig. 14), and Cape St. Elias, about 260 miles northwest of Lituya Bay-precipitation averaged 45 percent above normal for the entire month of October 1936, and 150 percent above normal for the 6-day period preceding October 27. At the three nearest inland stationsJuneau, Haines, and Skagway (fig. 14)-precipitation averaged 42 percent above normal for the month of October, and 111 percent above normal for the 6-day period preceding October 27. These departures are based on weather records through 1957 (U.S. Weather Bureau, 1958).

Allen, in both the published account (Alaska Daily Press, 1936) and in the account related to Nolde (Jensen, Caroline, written communication, Dec. 23, 1958), attributed the 1936 destructive "flood" and waves in Lituya Bay to the sudden draining of an ice-dammed lake in the basin of North Crillon Glacier. Williams (1938, p. 18) presented this hypothesis in detail, showing in a diagram the supposed course followed by the wall of water as it rushed down the surface of the glacier and into the head of Crillon Inlet. Williams (written communication, Mar. 3, 1954) stated that when he visited Lituya Bay after the "flood" (in 1937) he climbed along the sides of the Crillon Glacier and noticed the highwater marks there.

Floods due to the sudden draining of ice-dammed, lakes are a frequent and well-known phenomenon in southern Alaska, and it is understandable that this hypothesis was proposed and generally accepted as the cause of the destruction in Lituya Bay in 1936. In papers given orally in 1954 the writer presented evidence opposing the ice, dammed lake hypothesis as follows : North Crillon Glacier is an actively moving, much crevassed stream of ice that has an average gradient of about 500 feet per mile; its drainage basin, now mapped from vertical aerial photographs and well known from aerial and ground observations (D. L. Rossman, written communication, 1957), lacks any topographic configuration in which a large body of water could be ponded, unless it is a chamber concealed beneath the glacier. An aerial photograph taken by Bradford Washburn in June 1937, less than a year after the supposed "flood," shows no derangement of the surficial moraine patterns on the surface of the lower part of North Crillon Glacier, such as certainly should have occurred if water had flowed down over the surface of the ice as inferred by Williams. The high-water marks mentioned by Williams may have been the scars of fresh rockslides.

Crillon Lake, into which the South Crillon Glacier discharges (pl. 2), has been mentioned also as the lake that drained at the time of the 1936 waves. Seismic investigation by Goldthwait (1936, p. 508), indicates, however, that the bedrock sill on the divide beneath the drainage of North Crillon Glacier into Lituya Bay and South Crillon Glacier into Crillon Lake is about at the same level as the surface of the lake.

The writer, after reviewing the evidence available in 1954, concluded that serious objections could be raised against many of the possible causes of the 1936 waves that had been suggested until then, and that conclusive support could not be marshaled for any of them. Despite the additional evidence obtained since then about the 1936 waves and despite the wealth of information gained from the 1958 wave, this opinion is still held as this report is written. It is necessary therefore, as in 1954, to present several possible causes, some more convincing than others, but none definitely proven.

Sudden Draining Of An Ice-Dammed Body Of Water

The writer has already given convincing arguments opposing the hypothesis of surface drainage from an ice-dammed lake in the North Crillon Glacier basin, although this hypothesis perhaps best explains the roaring sound heard before the waves were seen.

Two other variations of this hypothesis warrant consideration: (a) The water could have been ponded in a chamber within or beneath the North Crillon Glacier, or on the divide separating the drainage of the North and South Crillon Glaciers, then suddenly released beneath the glacier or through an ice tunnel below sea level in the tidal front of North Crillon Glacier. This might account for the sudden upwelling immediately in front of the glacier. However it seems unlikely that a chamber of sufficient size could form in a glacier as active as North Crillon. Also, if the chamber were very high in the glacier, as would be required to obtain a substantial hydraulic head, it seems unlikely that the water could have jetted out rapidly enough to generate giant waves. (b) A partly subglacial lake is present now, and existed in 1936 in the trench tributary to Gilbert Inlet, just northwest of the sharp bend in the Lituya Glacier (pl. 2). Aside from the probability of relatively slow drainage from this lake, it is also unlikely that drainage from beneath the Lituya Glacier would set up waves that rose highest at the opposite end of the trough forming the head of Lituya Bay.

Fault Displacement

In 1954, displacement along the Fairweather fault was suggested as a possible cause of the 1936 "flood wave," although evidence of an earthquake was lacking (Miller, 1954). Through the eyewitness accounts that were obtained since then, the date and approximate hour of occurrence of the 1936 waves are now known and it is possible to state definitely that no earthquake was felt in Lituya Bay and that no earthquake with an epicenter near Lituya Bay was recorded at that time on seismographs at Sitka, Alaska, or more distant stations (Tocher, Don, written communication, Aug. 1, 1958). Perry Byerly (oral communication, Jan. 22, 1954) believes that fault displacement sufficiently large to cause the waves could not fail to have caused an earthquake that would have been felt in the bay and recorded at seismographic stations more distant than Sitka. Therefore, it seems that fault displacement can be ruled out as a cause for the 1936 waves.

Rockslide, Avalanche, Or Landslide

The roaring sound reported by three eyewitnesses to the 1936 waves and said by one of the observers to have come from the head of the bay and to have preceded the waves, suggests a rockslide or avalanche. Some of the observed differences between the 1936 and 1958 waves, particularly the occurrence of three waves of increasing size in 1936, and the much higher velocity of the 1958 wave, might be advanced as an argument against a common origin. On the other hand, these differences in the wave patterns might be due to differences in the location of the sliding or falling rock mass and the manner in which it entered the water. This was demonstrated in 1934 in Tafjord, Norway, where a rockfall generated waves of about the same height and velocity as the 1936 waves in Lituya Bay, and where three waves of increasing height were observed (Kaldhol and Kolderup, 1937; table 1, this report). Three waves reportedly were generated also by two other landslides into Norwegian fiords, Langfjord in 1746 and Norddalsfjord in 1938 (Jorstad, 1956, p. 326, 330-331).

By analogy with the 1958 wave, a falling mass that caused the 1936 waves in Lituya Bay should have come from the southwest wall of Crillon Inlet, opposite the high point on the trimline. None of the previously published eyewitness accounts mention any evidence that a large mass of rock or ice had fallen into Lituya Bay at the time of the 1936 waves, and Fredrickson (written communication, Sept. 1958) states that he did not notice any such evidence when he went to the head of the bay a short time after the waves had occurred. Comparison of the trilens photographs of Lituya Bay taken in 1929 by the U.S. Navy with the 1948 vertical photographs indicate that sliding had occurred on the valley wall above and just south of the front of North Crillon Glacier at some time between 1929 and 1948. This slide scar, however, is directly above the delta that formed in front of North Crillon Glacier before 1929, and some evidence of a large slide in 1936 should have been preserved on the delta and should be visible in the 1948 photographs. The 1929 and 1948 photographs show scattered large blocks of rock on the delta surface, also small talus cones along the base of the cliff, suggesting that sliding in this area has taken place frequently but in small increments.

Elsewhere on both walls of Crillon Inlet the correspondence between the 1929 and 1948 photographs is so close, even as to individual trees, gulleys, and other distinctive patterns, as to definitely eliminate the possibility that a large slide occurred during this interval. Small fields of permanent snow and ice are on the northeast wall of Crillon Inlet above 3,400 feet altitude, but these too show close correspondence in shape and size on photographs taken by Bradford Washburn in 1934 and 1937, eliminating the possibility of a large avalanche. There is a possibility that a rockslide or avalanche of ice fell on the North Crillon Glacier causing movement that was transmitted through the glacier to the tidal front. This possibility may be eliminated at least for the lower 2 miles of the glacier by inspection of the 1937 photographs; it seems unlikely that any movement higher on the glacier would be transmitted to the front.

The photographs indicate the occurrence of small rockslides into Gilbert Inlet from both the southwest and northeast sides, and into Lituya Bay between Mudslide Creek and Crillon Inlet, at some time between 1929 and 1948, but these locations are all incompatible with the trimline pattern of the 1936 waves. If the writer's interpretations of the photographs are correct, and falling or sliding of a mass of a size greater than a few thousand cubic yards into Crillon Inlet is required to generate the 1936 waves, then landsliding or avalanching may virtually be eliminated as a possible cause.

Submarine Sliding

Submarine slides (submarine "landslides") have long been included among two or more hypothetical causes of tsunamis in the oceans. For example, Gutenberg (1939), and Shepard, Macdonald and Cox (1950, p. 394-395), offer opposing viewpoints. The tsunami associated with the 1908 earthquake in the Straits of Messina has been attributed to a turbidity current originating in a submarine slump (Heezen, 1957). Recent laboratory experimental work indicates that submarine slides are capable of generating tsunamis (Wiegel, 1955).

Soundings in Crillon and Gilbert Inlets indicate slopes of as much as 28�, through vertical distances of nearly 500 feet. Unconsolidated material was available in 1936 in the deltas built out from the fronts of both North Crillon and Lituya Glaciers, and may have been present in substantial thickness at other places around the head of the bay. Submarine slides could also have occurred in bedrock. Perhaps one of the most attractive aspects of submarine sliding as a possible cause of the 1936 waves is that it cannot be definitely disproved because the evidence, if any, is hidden beneath the bay. Considering the magnitude of the slopes available and the probability that a large submarine slide would involve material at least partly above water, submarine sliding seems unlikely as the cause of the 1936 waves, however. Unless two or more slides occurred in close succession, or the waves were reflected at the head of the bay, it is difficult to account for the observed fact that the third wave, rather than the first, was the largest.

Movement Of A Tidal Glacier Front

The trimlines at the head of Lituya Bay show clearly that at least the largest of the 1936 waves was generated at or near the tidal front of North Crillon Glacier, and attained maximum height on the northeast wall of Crillon Inlet within 3,500 feet of the glacier front. This evidence, reinforced by the known generation of waves at the fronts of other glaciers that discharge into water, lends strong support to some kind of movement of the Crillon Glacier front as the cause of the 1936 waves. Three types of movement must be considered : (a) calving of ice from the subaerial part of a glacier front into the water; (b) calving and sudden surfacing of ice from a submarine projection of a glacier front; and (c) almost instantaneous forward movement of a glacier front. One aspect of the 1936 wave pattern, the occurrence of three waves of increasing height in the vicinity of Cenotaph Island, could be explained either by repeated movements of any of these three types, or by interference, refraction or reflection of waves near the point of generation at the head of the bay. Calving from an ice front could have caused the roaring sound reported by eyewitnesses, although it seems unlikely that calving could have occurred continuously for as much as half an hour before the first wave was sighted.

No photographs showing the North Crillon Glacier front shortly before or shortly after the occurrence of the 1936 waves are available, but oblique aerial photographs taken by Bradford Washburn in the summer of 1934 and in June 1937 show little change in the position and configuration of the northeast half of the front. The delta and southwest half of the front on Crillon Inlet are not shown on the 1937 photographs. Based on the photographs taken in 1934, and assuming little change in the following 2 years, the tidal front of North Crillon Glacier at a time just preceding the occurrence of the 1936 waves was a nearly vertical wall of ice about 2, 7 00 feet long and 200 to 300 feet above water level, extending across about half of the total width of Crillon Inlet (fig. 17). If the ice front extended to the bottom of the inlet, as seems likely, its maximum height below water level was about 290 feet.

Calving of subaerial ice into water has been observed at the fronts of many glaciers discharging into lakes, rivers, bays, and even into the open ocean in Alaska, as well as in many other parts of the world. From observation or indirect evidence such calving has formed waves capable of eroding as much as 5 feet above high tide a mile or more from the ice front (Tarr, 1909, p. 33-34), but according to available data no waves even approaching the magnitude of the 1936 waves in Lituya Bay have resulted from glacier calving in Alaska. If calving were the cause of the giant waves in 1936, such waves should occur with greater frequency, not only in Lituya Bay but also at the fronts of many other tidal glaciers in Alaska. This would be true unless, as suggested by C. C. Bates (written communication, Apr. 7, 1955), simultaneous calving from two or more glacier fronts is a further requirement.

In the course of the model study of Lituya Bay, R. L. Wiegel and Don Tocher found that rotational fall of a partly submerged weight with a flat face, simulating the Crillon Glacier tidal front, formed wave traces that compare closely in configuration to the trimlines on the walls of Crillon Inlet. The maximum height reached by the wave in the model, however, was about equal to the height of the face of the weight above water level. This gives some basis for doubting that ice falling from the Crillon Glacier front could have raised a wave to a height much greater than the height of the front. C. C. Bates (written communication, Apr. 7, 1955) suggested that although ice falling from the front of North Crillon Glacier might provide only about 10 percent of the necessary volume increment, the remainder of the rise indicated by the trimline on the northeast shore of Crillon Inlet might come from uprush or local refraction effects.

Submarine calving from the glacier front was suggested as a possible cause of the 1936 waves by W. O. Field, Jr. (written communication, Dec. 5, 1952). Evidence of ice projecting below water level as much as 1,000 feet beyond the subaerial part of glacier fronts has been reported for glaciers in the Yakutat Bay area (Russell, 1891, p. 101-102; Tarr, 1909, p. 31-32). Field states that waves 25 feet or more in height are formed by calving of projecting submarine ice masses at the front of Muir Glacier in Glacier Bay. The configuration of the submarine parts of the tidal ice fronts in Lituya Bay is not known. The possibility that the deltas in front of the Lituya Glacier may have been underlain by ice was mentioned on page 60. In the few hours that either or both the North Crillon and Lituya Glacier fronts were in sight during the 1952, 1953, and 1958 field investigations, the writer did not see any calving of submarine ice. The appearance of the delta in front of North Crillon Glacier in the 1948 vertical photographs does not give evidence of disturbance by the sudden rise of an ice mass beneath it, and the remaining tidal part of the glacial front seems to be too small to provide a mass of sufficient size to generate the 1936 waves.

Slippage of an ice mass over its floor is generally accepted by glaciologists as a major mechanism of movement for glaciers on slopes (Sharp, 1954, p. 826). However, an instantaneous advance of a glacier front of more than a few inches has not been proven. Prospectors in Disenchantment Bay reported that during the largest of the Yakutat Bay earthquake shocks, on September 10, 1899, the tidal front of the Hubbard Glacier advanced or was thrust forward from one-half to three-quarters of a mile, but Tarr and Martin (1912, p. 16) believed this to be an erroneous interpretation of the enormous calving of ice from the glacier front. It seems likely that forward movement of the Crillon Glacier front of a few feet or even a few tens of feet would be required to raise a wave to the height indicated by the 1936 trimline in Crillon Inlet. Such movement of the glacier front should have disrupted the surface of the glacier for some distance above the front to such an extent that the changes should be evident on photographs taken in 1937 and later. An oblique aerial photograph taken by Bradford Wash- burn in June 1937 shows no unusual crevassing or disruption of the surficial moraine patterns on North Crillon Glacier.

Tsunami In The Ocean

Perry Byerly and J. P. Eaton (oral communication, Jan. 22, 1954) offered the suggestion that wave motion from a tsunami generated at sea might be transmitted either through the narrow entrance or through the spit at the mouth of Lituya Bay, causing a seiche wave or some other type to form inside the bay. In further support of this suggestion Byerly (written communication, Feb. 1, 1954) called attention to the following statement by McNown (1952, p. 163) : "It has been amply proved that the motion produced in a port can have an amplitude not only equal to but even a number of times greater than the amplitude of the wave that produces it. Furthermore, from theoretical considerations, this amplitude can occur equally well with an entrance width that is extremely small." According to tide-gage records (Neuman, 1938, p. 26) no tsunami occurred in the northeastern Pacific Ocean in October 1936. It is difficult, also, to understand how wave motion introduced at the mouth of Lituya Bay could have been transmitted without any obvious surface effects to the head of the bay, there to be amplified into three giant waves that traveled out the bay at high velocity.

Other Possible Causes

For the sake of completeness several other agents capable of generating waves are mentioned, although there is little or no evidence to recommend them as possible causes of the 1936 waves in Lituya Bay. Submarine volcanic ,activity is known to have given rise to large tsunamis in the sea, as, for example, the destruction of Krakatoa in 1883 (Williams, 1941, p. 255-256, 261). Waves caused by tidal action and by wind are well known. As examples of water waves set in motion by air waves, Press (1956) cited a sea wave with a 2-foot amplitude that followed the Krakatoa explosion, and a 10-foot wave at Chicago in 1954. Finally, Olaf Holtedahl (oral communication, 1957) suggested that a falling meteorite be added to the list of possible causes.


Return to Professor Dutch's Home Page

Created 3 August 2004, Last Update 01 July 2012

Not an official UW Green Bay site