Rupture process of the 1964 Prince William Sound, Alaska, earthquake from the combined inversion of seismic, tsunami, and geodetic data

Four great earthquakes (1952, 1960, 1964, and 2004) have occurred since seismic monitoring began and only two since the installation of a global seismic network. A reexamination of the 1964 (M 9.2) Prince William Sound (PWS), Alaska, earthquake is timely due to the 2004 Sumatra earthquake because it adds constraints to the potential range of source parameters for these types of infrequent events and aids in the ability to predict the ground motions for other subduction zone including Cascadia. We first measured the durations of high-frequency energy radiation from teleseismic P wave codas recorded by short-period World Wide Standardized Seismographic Network (WWSSN SP) instruments to put constraints on the earthquake rupture length. The durations ranged between 250 and 600 s and are shorter in time between azimuths of 190° and 260°. The fault length is estimated to range from 540 to 740 km, rupturing at average speeds of 1.4-2 km/s in a 220°-238° direction. We developed a multiple time window kinematic rupture model for the PWS earthquake from the least squares inversion of teleseismic P waves, tsunami tide gauge records, and geodetic leveling survey observations based on the Green's function technique. We assume three major fault segments for the subduction zone dipping 6-12° based on geologic data. The subfault size on the megathrust was set to 50 × 50 km. The Patton Bay fault (PBF) imbricate thrust was assigned a 60 di with a subfault size of 20 × 20 km. We estimated a seismic moment of 5.52 × 10²² N m (M_w9.12). We identified three areas of major moment release, on the PWS segment near Montague and Middleton Islands extending out to the trench, beneath the Portlock anticline on the Cook segment between 58°N fracture zone and the Kodiak-Bowie seamount chain and finally on the Katmai segment off the coast of Kodiak Island extending to the trench. Our preferred rupture model has a peak slip of 14.9 in along the megathrust and 17.4 in along the PBF. The addition of the PBF resulted in an 8% improvement in residuals and also had a significant effect on the overall slip distribution. Moment release occurred as deep as 50 km depth slab contour near the 350°C isotherm and mantle fore-arc wedge, which is about 10-20 km deeper than previous coupling depths.