Using T-phase spectra to estimate rupture depth and static offset of the seafloor

Analysis of T-phases recorded from recent tsunamogenic earthquakes indicates that they contain significantly more high-frequency (>50 Hz) relative to low-frequency (>10 Hz) T-phase energy with longer durations (>30 s) than nontsunamogenic events. Modeling T-phase spectrum explains this observation, where the spectrum of the recorded T-phase is the product of the source spectrum, propagation loss in the solid earth, conversion from seismic to acoustic T-phase energy, propagation loss in the ocean, and the instrument response. The propagation loss in the ocean can be accurately modeled. Assuming the seismic-to-acoustic conversion is constant over frequency, the remaining unknown term is loss in the solid earth which is primarily due to anelastic attenuation. A refined depth estimate can be obtained by modeling the solid-earth attenuation. That, coupled with the earthquake source mechanism, allows for estimation of the static ocean floor deformation and thus tsunami height. This approach has been applied to two tsunamogenic earthquakes (26, Dec. 2004, 28, Mar. 2005), and one other event. Both tsunamogenic events had significant (>1 m) seafloor uplift. In addition, the peak static offset estimated for the 26 Dec. event is consistent with estimates both from other studies, and from observations of the seafloor deformation.