ire more remote. These results tend to agree with theoretical indications 

 E'er one-dimensional propagation. 



Data for the remaining stations of the west and central Pacific Ocean 

 ippear to accord best with a height decay law, H - r~^'^, which theoreti- 

 cally is appropriate to the leadixig waves of a tsunsmi in two-dimensional 

 [radial) expansion. 



The uniformity of the heat structure of the fundamental tsunami as 

 registered along the seaboard of the Americas suggests that the earth 

 lovement along the fault length must have been fairly uniform. The nijmber 

 )f waves in the leading envelope tends to be five, and the shape of the 

 mvelope is probably related to the shape of the initial vertical earth 

 lOvement. With distance from the origin the number of waves in the beat 

 .ncreases to 8, 9, or 10, but this is ascribed to an interference or 

 overlapping effect of competing wave trains. 



The period of the dominant tsunami waves (T - 1.8 hours) is found to 

 Lccord well with statistical trends relating tsunami period with earth- 

 quake magnitude M (Figure 71 )• Thus, an empirical relationship between 

 ' and M can be specified (Equation (l9), p. 10^* ) which suggests that 

 .arge magnitude earthquakes will always yield long-period tsunamis. 



Empirical relationships connecting tsunami period T and tsunami 

 source diameter S with earthquake magnitude M are found to yield closely 

 L direct proportionality between T and S (Equation (20), p. lOU ) . Such 

 I, proportionality, it turns out, can be established theoretically (Equa- 

 ;ion (25), p. 106). When wave period T is calculated in terms of tsunami 

 ;ource size S, the result (T - 1.99 hours) confirms quite well the wave 

 )eriod identified in the tide gage records. 



The subjective analyses made of selected marigrams (Figures U3 to 

 )6) appear to show low waves that are still longer in period than 2 hours, 

 ^hese may have derived from the horizontal thrust of the Alaskan landmass 

 )r may be nonlinear subharmonics of the main tsunami waves developed at 

 ,he confining boundary. 



When tsunami waves traverse a Continental Shelf, a degree of reso- 

 lance is possible even if the coastline is straight, and the shelf forms 

 i uniformly inclined plane out to a straight shelf edge parallel with 

 ;he coastline. The criterion for the development of this resonance is 

 ;he ratio of the effective relative depth (d/T^) at the shelf edge to the 

 square of the shelf slope (s^) (see Figure T6). The theoretical indica- 

 iions here require confirmation from experiments. In the presence of 

 leavy damping, as is likely to exist, the implication is that runup 

 leights of tsunami waves as a ratio of wave height H at the shelf edge 

 ire seldom likely to exceed a value of about h from this phenomenon alone. 

 :t has been concluded in general that the wave height Hj, at the coastal 

 joundary (not necessarily the runup height) for the Alaskan tsunami was 

 ibout 1.5 H. 



353 



