Review of some tsunamis in canada

T.S. MURTY

1. December 6^th , 1917: Large tsunami in Halifax Harbor due to an explosion.

2. November 18^th, 1929: Large tsunami in Burin Inlet due to an earthquake off the coast of Newfoundland.

3. June 23, 1946: Small tsunami in the Strait of Georgia due to an earthquake on the Vancouver Island.

4. March 28^th 1964: Large tsunami on the coast of British Columbia due to an earthquake in Alaska.

5. April 27, 1975: Large tsunami in Kitimat Inlet due to a submarine landslide.

All these tsunamis have quite different characteristics as can be seen below. During the First World War, a munitions ship caught fire and exploded in Halifax harbor on December 6^th, 1917. In the harbour narrows the amplitude of the tsunami that was generated was about ten meters in amplitude.

The Grand Banks earthquake of November 18^th, 1929 created a tsunami of at least 12.2 metres in amplitude in the Burin Inlet. Some estimates put the maximum amplitude at about 35 meters, which could be an overestimate. The turbidity currents following the earthquake caused numerous cable breaks in the Atlantic Ocean. On June 23, 1946, a small tsunami occurred in the northern part of the Strait of Georgia (between Vancouver Island and the main land) following an earthquake on Vancouver Island close to the western shore of the Strait.

The May 1960 Chilean earthquake tsunami was observed at several locations of the Pacific Coast of Canada.

Following a large earthquake in Alaska on March 28^th, 1964, a major Pacific wide tsunami was generated. Outside of Alaska, the largest tsunami amplitude anywhere on the west coast of North America occurred not exactly on the open coast, but inland at Port Alberni at the head of the Alberni Inlet. This can be explained as due to the amplification of the tsunami through quarter wave resonance as the tsunami traveled from the mouth to the head of the inlet.

On April 27^th, 1975, following a major submarine landslide, a tsunami was generated in the Kitimat Inlet in the Douglas Channel system of the northern part of the coast of British Columbia.

2. The Halifax Harbour Explosion Tsunami
In the midst of World War I, shortly before 9 a.m. on December 6, 1917 the munitions ship Mont Blanc collided with the relief ship Imo in the Narrows of the Canadian port Halifax. The Mont Blanc was carrying the equivalent of about 2900 short tons (2000lb/ton) of TNT. A fire broke out onboard the Mont Blanc and the crew abandoned ship, which drifted into the Halifax side and grounded near one of the piers. Shortly about 9:00 a.m. the cargo exploded devastating a large section of the city. Estimates of the casualties are as high as 2000 dead and 9000 injured. Of these, many are thought to have died from the affects of the tsunami that followed. The documented damage of this, the greatest man made explosion to that date, would later be used in the Manhattan project to estimate the devastation of the first nuclear bombs. [1, 2].

There was no operational tide gauge at the time recording the changes in sea level, but there are several narrative reports of extreme high and low water and useful information might be obtained from them [3]. The explosion occurred very close to the time of low tide, about 0.5 m below mean water level.

To hindcast the tsunami, [1] ran a linearized finite element model covering the full area of Halifax Harbour from the head of the Bedford Basin to the seaward entrance to the harbour approach. The model was initialized with a setup derived using an empirical formula for waves resulting from explosions in water. Figures 1 to 2 respectively show the tsunami travel times and the tsunami amplitudes.


Figure 1 Figure 2


Figure 2. The maximum elevation (m) achieved during the model run and locations (estimated) of the anecdotal observations. The detail of the Narrows is shown in the lower left. From Greenberg et al, (1994).

Figure 1. The time of arrival in minutes of the initial wave taken to be the time the sea level first exceeded 0.2 m. From Greenberg et al, (1994)

The model results when put together with anecdotal reports give a consistent overview of the tsunami amplitude and progression. The study by [1, 2] indicates that in the Narrows at the explosion site, the wave was over 10 m high, but that the amplitude diminished greatly further away. In the Bedford Basin the tsunami would not have been damaging and in the outer reaches of the harbour, it would have been noticeable only to those looking for it.

This earthquake generated a tsunami with amplitudes of at least 12.2 m [4] in Burin Inlet on the south coast of Newfoundland and killed 26 people. Gregory [5] gave 30.5 m as the maximum amplitude in Burin Inlet. McIntosh [6] gave a value of 4.6 m for the tsunami amplitude at Lamaline. According to Johnstone, no tsunami was observed at Sable Island,

Although turbidity currents and tsunamis are not directly related, nevertheless both could be caused by earthquakes such as the 1929 Grand Banks earthquake. Turbidity currents and tsunamis could also be generated by landslides whether or not an earthquake occurred.

Most submarine telegraph cables from North America to Europe pass south of Newfoundland. At about 2032 h GCT, Nov. 18, 1929, an earthquake of magnitude 7.2 occurred on the continental slope southeast of the Cabot Trench. This generated a tsunami that caused considerable property damage and loss of life along the shores of Placentia Bay, which was discussed above.

Another important consequence occurred during the 13 h and 17 min following the earthquake. An orderly sequence of breaks occurred in the telegraph cables of 483 km south of the epicenter. According to Heezen and Ewing [10], although all cables along the continental slope and on the floor of the ocean south of the epicenter were broken, none on the continental shelf were disturbed. The exact times and locations of the cable breaks were known, respectively, from the telegraph records and resistance measurements.

Bucher [9] hypothesized that erosion in the submarine canyons caused by the tsunami left the cables unsupported, thus leading to breakage. Heezen and Ewing [10] criticized Bucher’s explanation on the grounds that the cable breaks were too regular in time and the cables were of different ages and breaking strengths.

Heezen and Ewing [10] showed rather convincingly that the successive series of breaks in the telegraph cables following the Grand Banks earthquake of 1929 were caused by the turbidity current generated.

Kitimat Inlet on the West Coast of British Columbia has a history of landslides: several slides occurred during the period 1952 to 1968, and also in 1971. On October 17, 1974, following a submarine slide, a water wave of 2.8 m amplitude was generated. On April 27, 1975, following a major slide, water waves with ranges up to 8.2 m were generated. In other parts of British Columbia, also major slides occur – for example, in Howe Sound, in 1995.

Depending on the nature of movement of the sediment, a slide will have two components: horizontal and vertical. Usually, one component dominates the other. For theoretical and laboratory studies, it is convenient to distinguish between vertical and horizontal slides, depending upon which component dominates. Another type of classification of slides is: (a)

Figure 4. Geography of the Pacific Coast of Canada

land-slides, (b) submarine or under-water slides. Actually, in several instances, there may be visual manifestations in the surrounding land, even in the case of a submarine slide.

Here, we concern ourselves specifically with the submarine slide of April 27, 1975 in the Kitimat Inlet. There are visual manifestations of the slide in the Moon Bay area even now, and it is estimated that about 3 x 10⁶m³ of material was involved in this bay alone. Casagrande [12] estimated the total volume of the material involved in the slide that gave rise to the water waves to be about 10⁷m³. Based on an examination of the hydrographic charts prepared before and after the slide, and some rough measurements of shore features, Murty [11] suggested that an upper limit for the total amount of material involved in the whole slide is about 26 x 10⁶m . Actually, it is not the volume of the material involved in the slide that directly enters into the calculation of the amplitude of the water waves generated: various factors associated with the slide together determine the amplitude of the waves.

At the time of occurrence of the slide of April 27, 1975, there was no major seismic event reported in that area. Also, there was no meteorological event at that time which could have caused large water-level oscillations.

Hence, there is no doubt that water waves were generated by the submarine slide. At least two water waves (and possibly three) were generated, and propagated into the connecting bays and channels. The largest wave was estimated to be 8.2 m in range (crest to trough).

Brown [13] mentions that the first wave was observed at 10:05 a.m. (Pacific Daylight Savings Time) and, at 10:15 a.m., the bottom was visible. The whole water-level disturbance lasted about an hour. In Bish Creek and Clio Bay, which are about 8 km from the site of the slide, at least one wave was observed. Some damage occurred in Bish Creek and the range in Clio Bay was estimated to be about 6.7 m. No wave was noted in Sue Bay, probably because of the complicated path needed for the wave to travel into that bay. Although Minette Bay is at the head of the Kitimat Arm, it is an extremely shallow bay (most of the time there is no water in it) and, further, the bay bottom is very rough because of vegetation and tree trunks and it is quite conceivable that the wave could not have penetrated and traveled in Minette Bay.

The Kitimat Submarine Slide of April 27, 1975 occurred approximately 53 minutes after the occurrence of low tide. There appears to be sufficient observational data to suggest that submarine slides appear in association with low tide. Murty [11] modeled this following a technique described in Miloh and Streim [14].

5. Results of Numerical Models for B.C. Tsunami Warning System
Maximum tsunami water levels and currents along the British Columbia outer coast have been computed for waves originating from Alaska, Chile, the Aleutian Islands (Shumagin Gap), and Kamchatka [15, 16]. Three computer models have been developed to generate and propagate a tsunami from each of these source regions in the Pacific Ocean to the continental shelf off Canada’s west coast, and into twenty separate inlet systems. The model predictions have been verified against water level measurements made at tide gauges after the March 28, 1964 Alaska earthquake. Simulated seabed motions giving rise to the Alaskan and Chilean tsunamis have been based on surveys of the vertical displacements made after the great earthquakes of 1964 (Alaska) and 1960 (Chile). Hypothetical bottom motions have been used for the Shumagin Gap and Kamachatka simulation. These simulations represent the largest tsuanamigenic events to be expected from these areas.

Maximum wave and current amplitudes have been tabulated for each simulated tsunami at 185 key locations along the British Columbia coast. On the north coast of British Columbia, the Alaska tsunami generated the largest amplitudes. In all other regions of the west coast, the largest amplitudes were generated by the Shumagin Gap simulations. Wave amplitudes in excess of 9 m were predicted at several locations along the coast and current speeds of 3 to 4 m/s were produced. The most vulnerable regions are the outer coast of Vancouver Island, the west coast of Graham Island, and the central coast of the mainland. Some areas, such as the north central coast, are sheltered enough to limit expected maximum water levels to less than 3 m.

We note that the effects of dry land flooding have not been included owing to the large amount of additional topographic data that is required and the special demands created for high-resolution, local-area models in many locations. This does not pose serious problems inmost areas. However, near the heads of inlets (where these is often an extensive area of flatlands associated with a river mouth) modeled wave heights may be appreciably overestimated. The four source areas shown in Figure 5 have been identified as likely sites for generation of tsunamis that could threaten Canada’s coastline. This is based on the occurrence of previous tsunamigenic earthquakes and on estimates of the likelihood of future great earthquakes in each area.

Figure 5. Epicenters of earthquakes used in tsunami simulations. The bold line is the boundary of the deep ocean model (DOM). (1: Alaska, 2: Chile, 3: Shumagin Gap, 4: Kamchatka). [16]

Tsunamis arriving at British Columbia’s outer coast propagate into all exposed inlet systems. Twenty of the more exposed systems have been identified and incorporated into this study. Each has a corresponding numerical model that uses time-varying water elevations at one or more connections to the continental shelf to calculate the water surface response and currents within the system. Construction of each model required detailed extraction of bathymetry and dimensional data (cross-sectional and surface areas) and calibration.

Seismic activity in the Pacific Ocean is confined primarily to zones adjacent to the continental margins where subduction of the oceanic plates under the continental landmasses episodically releases bursts of energy in the form of an earthquake. Other areas, such as the Hawaiian Islands, are also sites of large earthquakes but these do not pose a threat of producing a destructive tsunami in British Columbia waters.

If an earthquake results in vertical (dip-slip) motion of the oceanic crust, then it is tsunamigenic, that is, it will result in the deformation of the water surface and subsequent propagation of the disturbance outward form the source as a seismic sea wave (tsunami). Tsunamis from even moderately small earthquakes may result in significant damage within a short distance of the source. If the earthquake is sufficiently large (with a Richter scale magnitude greater than 7.5 to 8.0), however, then a large tsunami may be generated and damage will result at great distances from the source. This latter situation is most relevant to the west coast of Canada.

Three distinct numerical models were used to simulate tsunami generation and propagation to the inlets of British Columbia. A deep ocean model (DOM) with 0.5 resolution (Figure 5) has been used to simulate bottom motions that give rise to tsunamis and to propagate the resulting waves to the continental shelf (Figure 6) off Canada’s west coast. There, a 5-km

Figure 6. Shelf model grid outline (A) and outline of region used for field plots of northern British Columbia (B). From Dunbar et al, 1989.

Figure 7. Map showing water level and current locations in some inlets. From Dunbar et al, 1989

After confirming the correctness of all model components using historical tsunami data, the three models were run for a set of simulated tsunamigenic earthquakes. These included simulations for measured bottom displacements at the source regions of the 1960 Chilean and 1964 Alaskan tsuanmigenic earthquakes, and hypothetical earthquakes at the Shumagin Gap and Kamchatka Peninsula sites. In addition, the Alaskan simulation was repeated for a case where bottom motions were amplified by 25%. The simulations provided wave amplitudes and currents.

2. Greenberg, D.A., Murty, T.S. and Ruffman, A. (1994). Modeling the Tsunami From the 1917 Halifax Harbor Explosion, Science of Tsunami Hazards, Vol. 11, No. 2, 67-80.

3. Ruffman, A, Greenberg, D.A. and Murty, T.S. (1995). The Tsunami from the Explosion in Halifax Harbour, Ground Zero, pp. 327-344, Edited by A. Ruffman And C.D. Howell, Nimbus Publishing Ltd. Halifax.

4. Johnstone, J.H. L. 1930. The Acadian-Newfoundland earthquake of November 18, 1929. Proc. Trans. N.S. Inst. Sci. Halifax, N.S. 17:223-237.

5. Gregory, J.W. 1929. The earthquake south of Newfoundland and submarine canyons. Nature 124: 945-946.

6. McIntosh, D.S. 1930. The Acadian-Newfoundland earthquake. Proc. Trans. N.S. Inst. Sci. 17:213:222.

7. Murty, T.S. (1977). Seismic Sea Waves-Tsunamis, Bulletin No. 198, Fisheries Research Board of Canada, Ottawa, 337p.

8. Murty, T.S., and Wigen, S.O. 1976. Tsunami behavior on the Atlantic coast of Canada and some similarities to the Peru coast. Proc. IUGG Symp. tsunamis and tsunami Res. Jan. 29-Feb. 1, 1974. Wellington, N.Z., R. Soc. N.Z. Bull. 15:51-60.

9. Bucher, W.H. 1940. Submarine Valleys and Related Geologic Problems of the North Atlantic. Geol. Soc. Am. Bull. 51: 489-512.

10. Heezen, B.C., and Ewing, M. 1952. Turbidity currents and submarine slumps and the 1929 Grand Banks earthquake. Am. J. Sci. 250: 849-873.

11. Murty, T.S. (1979). Submarine Slide-Generated Water Waves in Kitimat Inlet, J. of British Columbia, Phys. Res., Vol. 84, No. C12, 7777-7779.

12. Casagrande, L. (1977). Kitimat Arm B.C. Slide of April 27, 1975. Unpubl. manuscript. Casagrande Consultants, 40 Massachusetts Avenue, Arlington, Mass. 30 pp. and Appendices.

13. Brown, W.E. (1975). Underwater Subsidence at Kitimat: Sunday 27 April, 1975 (Unpubl. man., Institute of Ocean Sciences, Patricia Bay, Fisheries and Environment Canada, Sidney, B.C. 8 pp.

14. Miloh, T and H.L. Strein (1978). Tsunami effects at coastal sites due to offshore faulting, tectonophysics, vol. 46, 347-356.

15. Dunbar, D.S., LeBlond, P.H. and Murty, T.S. (1989a). Evaluation of Tsunami Amplitudes for Pacific Coast of Canada, Progress in Oceanography, Vol. 26, 115-177.

16. Dunbar, D.S., LeBlond, P.H. and Murty, T.S. (1989b). Maximum Tsunami Amplitudes and Associated Currents on the Coast of British Columbia, Science of Tsunami Hazards, Vol. 7, No. 1, 3-44.