extreme, in Womens Bay, the depth shelves to zero on the tidal flats 

 (see Figure 88). Hence any free oscillation of the basin would have 

 rather small amplitude in Womens Bay, maximum amplitude at City Dock, 

 and reduced amplitude at Kodiak City. We believe the enigma is thus 

 explained, and that the oscillation Tilley observed was the direct re- 

 sult of the jolt of land movement in the southwest direction occurring 

 during the earthquake. Figiore kO shows that many observations of ground 

 motion from NE to SW were reported throughout the region. 



In Figure 90 we have superimposed upon our hypothetical marigram 

 the large tsunami system that seemed apparent in Figure 38c. In Figure 

 97 we apply the subjective analysis technique to find the residual os- 

 cillations. Thus, Figure 9Tc suggests the nature of the wave system 

 riding on the main tsunami system. The accuracy of Figures 90 and 97 

 after about 11:00 p.m. is doubtful. Prior to that time, however, we 

 suspect that the oscillation shown is a combination of the Continental 

 Shelf oscillation (Figure 9Tc ) noted in Figure 38c and the oscillation 

 in St. Pauls Harbor (Figure 97d). 



2. Tsunami Damage at Kodiak City 



The picture we now have of the tsunami (Figure 90) suggests that 

 the "second" wave, after Tilley 's first wave, had an amplitude of some 20 

 feet or a height of about ^40 feet. This wave is presumed to be dominantly 

 the result of a shelf oscillation such as discussed in Section III-2, with 

 a second mode period of approximately 100 minutes. The interval of time 

 between the "second" wave, or first crest on Figure 97c, and the next 

 crest (Figure 97c) is about 2 hours, however. 



In Figure 97d we have assiomed the amplitude of the first oscillation 

 in St. Pauls Harbor, as it might have occurred at Kodiak City, to be some- 

 what less than Tilley 's observation. The second wave of Figure 90 is now 

 seen to comprise a 10-foot amplitude shelf oscillation apparently combined 

 (slightly out of phase) with a 15-foot amplitude oscillation of St. Pauls 

 Harbor. Most of the water for the antinodal resultant of these waves at 

 Kodiak City would have been drawn locally from St. Pauls Harbor accounting 

 for the fact that velocities of horizontal flow during the rise of the 

 wave at Kodiak were small (at the antinode of a seiche they are nominally 

 zero). However, after the "second" wave (Figure 90), the shelf oscilla- 

 tion went to work in draining water from the entire area. Figure 97c 

 suggests that it would have done this to the extent of a drop in water 

 level of about 20 feet over the entire area of St. Pauls Harbor. An 

 extra drop in level at Kodiak City of another 20 feet would have been 

 occasioned by the local oscillation (Figure 97d). 



This total drop of water level of nearly Uo feet in 35 minutes 

 (Figure 90) would have occurred over the entire area of the Inner 

 Anchorage and boat harbor which, as Figure 98 shows, are contained 

 between the island string (Gull, Uski and Near Islands) and the 

 mainland. 



Text resimes on page 156 

 147 



