return too soon to low areas, such as at Kaguyak. Spaeth and Berkman 

 (1972) note that several people were killed at Crescent City, California, 

 because they returned to a low area before the arrival of the largest 

 wave. A large amount of damage at Crescent City resulted from the failure 

 to remove vehicles, including a gasoline tank truck, from the endangered 

 area. 



3. Ionospheric Waves . 



To evaluate the possibility of a tsunami being generated by an earth- 

 quake, it is desirable to have information about the source mechanism of 

 the earthquake, i.e., whether the earthquake is a dip-slip type or a 

 strike-slip type. If the earthquake is a strike-slip type, it may be 

 assumed that a large transoceanic tsunami will not be generated, and 

 tsunami alerts can be canceled at all locations except those near the 

 epicenter. If the earthquake is a dip-slip type, there is a high proba- 

 bility that a tsunami may have been generated, and additional information 

 must be obtained from tide stations. 



Van Dorn (1965) indicated that a dipolar barometric wave in the 

 atmosphere was associated with the dipolar ground motion of the 1964 

 Alaska earthquake, and that this raised a possibility for early tsunami 

 prediction. Row (1972) discusses the atmospheric waves associated with 

 the Alaska earthquake in greater detail. Row indicates that there is 

 both an early-arriving pressure disturbance, associated with seismic 

 waves in the Earth, and the late-arriving disturbance (propagating at 

 about 300 meters per second) previously mentioned by Van Dorn. The late- 

 arriving disturbance is associated with the tectonic deformation and, at 

 distances far from the source, would arrive well in advance of gravity 

 waves traveling through the ocean. For example, at a distance of 5,000 

 kilometers (3,100 miles) from the source, a tsunami traveling across the 

 ocean at a speed of 200 meters per second (447 miles per hour) will 

 arrive approximately 2 hours 20 minutes after the atmospheric wave. 



Pressure disturbances also propagate through the ionosphere. Row 

 (1972) discussed the possible association between ground motion and 

 ionospheric waves. Furumoto (1970) reports on the use of a 10-megahertz 

 Doppler recording of Rayleigh waves to estimate the initial phase of the 

 source of the 11 August 1969 Kuril Islands earthquake. He notes that 

 this provides a rapid approach to source mechanism estimation. The 

 Doppler shift associated with the ionospheric waves can be monitored at 

 relatively low cost (Furumoto, 1970). Murty (1977) provides further 

 discussion on ionospheric effects. 



The use of atmospheric waves to estimate the ground motion of a 

 tsunami source requires further investigation. However, this method may 

 be useful for the Tsunami Warning System. 



4. Deep-Ocean Tsunami Gages . 3 



In addition to supplying information for the Tsunami Warning System, 

 the tide stations provide records of tsunami heights and periods. Unfor- 

 tunately, the local topography distorts the tsunami recorded on tide gages 

 near the coastline, and these records do not provide information on the 



198 



