The detection of a multiple tsunami phenomenon

Abstract

A variable period band pass filter with a specified width is develop and applied to the long wave data recorded at Wake Island during the time period of March 9-12, 1957. This filter attenuates waves of periods other than the single period about which it is centered. The filter band is successively centered at the periods of March 9, 1957 which was generated in the Aleutian Islands is observed in all filtered records as well as in the original record. Several hours after the initial tsunami, additional bursts of wave energy begin to appear in the filtered wave records. These later wave arrivals are well correlated with the expected arrival times of long surface gravity waves generated by the aftershocks of the initial tsunamigenic earthquake. These secondary tsunamis are predominantly long period oscillations and are most distinctive in the 55 and 100 minute period filtered wave records. While these later tsunamis are not clearly in the original data, the 100 minute period components of some of the secondary tsunamis have greater amplitudes for the initial tsunami. Since permanent ground displacements are not always expected form this aftershock activity, these low energy waves are theorized to have been generated by non-permanent or transient ocean floor displacements. Some preliminary investigations involved the determination of the effect of rotation on tsunami propagation. A dispersive long wave train traveling on a rotating globe has a maximum group velocity associated with an intermediate wave period of the order of 60 minutes instead of the longest period wave as in the nonrotating case. The wave front is therefore found to be influenced by rotation.