SECT. 5] 



LONG OCEAN WAVES 



659 



documented accounts is that of the tsunami of 1 April, 1946, associated with an 

 earthquake in the Aleutian trough (Shepard, MacDonald and Cox, 1950). An 

 interesting discussion of waves caused by the eruption of a submarine volcano 

 is due to Unoki and Nakano (1953). We have already referred to the findings of 

 Ewing and Press (1953) that some of the distant tide-gauge disturbances follow- 

 ing the eruption of Krakatoa were induced by atmospheric waves. This historic 

 eruption was the subject of a fascinating report by the Royal Society (Symons, 

 1888). A rock-slide associated with the Alaska earthquake of 10 July, 1958, 

 resulted in a long wave whose crest rose up to 1700 feet (!) above mean water 

 level (Miller, 1960). Nakano (1955) has given an account of tsunamis from 

 atomic explosions. 



The study of tsunami waves differs from that of the background in one 



Fig. 8. Tide record of 5 November, 1952, at Pago Pago, Samoa. 



important aspect : there is a definite (impulsive) beginning to the disturbance. 

 Fig. 8 shows the initial arrival of the Kamchatka tsunami at Pago Pago, 

 Samoa (Zerbe, 1953). The relatively small size of the initial crest is typical of 

 many tsunamis. In this particular case the arrival time can be measured to the 

 nearest few minutes ; in many instances the initial onset is not nearly as clear 

 cut. Arrival times have been computed assuming propagation at a speed of 

 ■\/{gh) along the great circle route between epicenter and station (see, for 

 example, Zetler, 1947). In an ocean basin of variable depth the wave fronts 

 propagate according to Fermat's principle, and the rays are not precisely great 

 circle routes. Nevertheless the great circle approximation ought to be a good 

 one, and in fact computed and observed travel times are generally in accord to 

 within a few per cent. 



