PART IV: CONCLUSIONS AND RECOMMENDATIONS 

 Conclusions 



41. To determine tsunami elevations, a record of local tectonic dis- 

 placements of the seabed was synthesized. In previous studies, Houston 

 and Garcia (1978) and Houston (1980) considered only distantly generated 

 tsunamis. These authors concluded that only the gross shape of the ground 

 deformation was necessary to determine tsunami elevations at distant loca- 

 tions. Hence, they synthesized a record of tectonically reasonable displace- 

 ments of the seabed knowing that the exact shape of each deformation was un- 

 important. The situation was not as simple in this study since all tsunamis 

 were locally generated. 



42. It is obvious that, in the near field, tsunami elevations will de- 

 pend on the shape of the seabed deformations. Still, the method used required 

 that a model deformation be defined. Hence, it was necessary to assume that 

 the standard deformation employed would result in the same 100- and 500-year 

 combined tsunami and tide elevations as would have resulted if actual histo- 

 rical sources had been employed. In light of the fact that tectonic displace- 

 ments are known for only two submarine earthquakes — the 1960 Chilean earth- 

 quake and the 1964 Alaskan earthquake — the assumption is not only reasonable 

 (as discussed in Part II) but also necessary. 



43. The numerical model used in this study accurately simulated tsunami 

 propagation in the open ocean of the Gulf of Alaska, on the narrow shelf of 

 the eastern Gulf of Alaska, and in Prince William Sound. In Cook Inlet the 

 water is sufficiently shallow such that the adequacy of the linear nondisper- 

 sive model equations may be questioned. Tsunami heights in the inlet, how- 

 ever, are fairly small since Kodiak Island, the Barren Islands, and Kenai Pen- 

 insula shelter the inlet from the major tsunami generating region. Hence, the 

 model is considered to be adequate in Cook Inlet also. 



Recommendat ions 



44. The elevations predicted in this report are at the shoreline but 

 can be assumed to equal runup elevations for most of the study area. There 

 are locations where time-dependent effects (e.g., lack of sufficient time to 



20 



