The origin of tsunamis excited by local earthquakes. Broadband waveform observation of local earthquakes

Citation

Ma, Kuo-Fong (1993) The origin of tsunamis excited by local earthquakes. Broadband waveform observation of local earthquakes. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/dkgr-2k53. https://resolver.caltech.edu/CaltechTHESIS:01092013-102744893

Abstract

The origins of tsunamis excited by the 1989 Lorna Prieta, the 1906 San Francisco, and 1975 Kalapana, Hawaii, earthquakes were examined in part I. Since tsunamis are mainly caused by vertical deformation under the sea-floor, the tsunami data allows us to constrain the vertical motion of the sea-floor during the earthquake and to determine the excitation mechanism of the tsunami.

The first arrival of the observed tsunami from the 1989 Lorna Prieta, California, earthquake observed at Montery was about 10 min. after the origin time of the earthquake. The synthetic tsunami computed for the uniform dislocation model determined from seismic data can explain the arrival time, polarity, and amplitude of the beginning of the tsunami but the period is too long. We tested other fault models with more localized slip distribution. None of the models could explain the observed period. The residual waveform, the observed minus the synthetic waveform, begins as a downward motion at about 18 min. after the origin time of the earthquake, and could be interpreted as due to a secondary source near Moss Landing. The volume of sediments involved in the slumping is approximately 0.012 km^3. We conclude that the most likely cause of the observed tsunami is the combination of the vertical uplift of the sea floor due to the main faulting and a localized slumping near Moss Landing.

The observation of tsunami excited by the 1906 San Francisco earthquake is curious because this earthquake is generally believed to be a strike-slip earthquake for which tsunamis are not usually expected. We show that the tsunami was caused by a local subsidence associated with a bend of the San Andreas fault offshore from the Golden Gate; no vertical fault motion was involved during the 1906 San Francisco earthquake.

The 1975 Kalapana, Hawaii, earthquake was accompanied by large tsunamis which were well recorded at several tide-gauge stations around the Hawaii islands. To examine the source of the tsunamis associated with the earthquake, we computed synthetic tsunamis for three tide-gage stations, Hilo, Honolulu and Kahului, using various dislocation models, Hilina fault models and slump models. Crustal deformation data were used to constrain the dislocation models. We could find a combination of a dislocation model for the earthquake and a Hilina fault model which can explain the observed crustal deformation inland fairly well. However, the tsunamis computed for this combined model are too early in first arrivals and too small in amplitudes. The residual tsunamis, observed synthetic, are not very different from the observed tsunamis and can be interpreted with a slump model which involves an uplift of 100-110 cm over an area of about 2500-3000 km^2 offshore. The total volume of displaced water associated with the slumping is about 2.5~3 Km^3.

The recent deployment of TERRAscope, a broadband and wide dynamic range seismic network in Southern California, provided us with a capability of recording complete waveforms of nearby earthquakes. These waveform data allow us to determine the overall faulting mechanisms, seismic moments, depths, stress drops, and the attenuation characteristics of the crust In part II. I investigated the waveforms of local earthquakes recorded at TERRAscope stations to understand the characteristics of the earthquake sequences.

The Pasadena earthquake (M_L=4.9) of 3 December, 1988, occurred at a depth of 16 km, probably on the Santa Monica-Raymond fault. Prior to this event, no earthquake larger than magnitude 4 had been recorded in this area since 1930. We determined the focal mechanisms and seismic moment of 9 aftershocks by combining the first-motion data and the waveform data of P, SV, and SH waves recorded at Pasadena TERRAscope station, since the first-motion data for most of the aftershocks are too sparse to determine the mechanism. The average orientations of the P and T axes of the aftershocks are consistent with the strike of the Raymond fault. The ratio of cumulative seismic moment of the aftershocks to the seismic moment of the main shock is significantly smaller than commonly observed.

The mechanisms and seismic moments of the Sierra Madre earthquake (M_L=5.8) of 28 June, 1991, sequence were determined using the same techniques that was applied in the 1988 Pasadena earthquake sequence. Most events located within 5 km west of the mainshock are similar to the mainshock in waveform. The mechanisms thus determined are thrust. Some events have high stress drops between 100 to 1000 bars; the mainshock is one of them. For other larger events, including the two largest aftershocks, the stress drops are between 10 to 100 bars. A few events located east of the mainshock have waveforms different from the mainshock and have strike-slip mechanisms. The ratio of cumulative seismic moments of the aftershocks to seismic moment of the mainshock is smaller than that of most events in California. The average Q_β values along the paths from the hypocenters of the Sierra Madre and the Pasadena earthquake to PAS are about 130 and 80 respectively.

The 1992 Landers earthquake is the largest event to have occurred in Southern California since 1952. We examined the waveforms of the aftershocks recorded at PFO TERRAscope station to see the correlation of the waveform and mechanisms determined from surface wave inversion. Since the depths of the events are usually not determined very well, the amplitude ratio of surface wave to body wave was used to examine the accuracy of the depths determined with various methods. Most of the events which occurred to the south of the mainshock epicenter have similar waveforms and mechanisms. Only a few events occurred to the north of the mainshock epicenter where large slip occurred during the mainshock. These events have dissimilar waveforms and mechanisms. A near vertical distribution of the aftershock extending to a depth of 15 km, or even deeper is found at about 18 km to the south of the Landers earthquake epicenter. About 72% of the total energy of the aftershocks were released from the region to the south of the mainshock epicenter. The ratio of cumulative seismic moment of the aftershock to that of mainshock is less than 1/100.

Thesis Files