Multi-time-window inversion for sea surface displacement of the 2011 Tohoku tsunami

Many studies have modeled the tsunami source of the 2011 Tohoku earthquake using different types of data. These studies show that the similar locations of large slip, but the details and the maximum slip amounts are different. Satake et al. (2013, BSSA) performed tsunami waveform inversion using offshore and coastal tsunami data to estimate spatial slip distribution and temporal process. Their result shows a very large slip occurred near the trench about 3 minutes after the rupture, then it propagated northward along the trench. This delayed slip is considered as the reason of large runup on the coast of Iwate, about 100 km of the maximum slip on the fault. On the other hand, Tappin et al. (2014, Marine Geology) claimed that a submarine landslide occurred offshore Iwate around 2 minutes after the earthquake, and it caused high runup on the coast. In this study, we followed and expanded the method of Satake et al. (2013 BSSA) to estimate the sea surface displacement rather than the slip on the fault plane. Sea surface displacement can be caused either by fault slip or other sources such as submarine landslide. In this approach, all sea surface units are 34 km-square. Total duration is set to 240 sec, which is divided into 4 windows, each 60 sec long. In order to cover the secondary sources which are not related to slip on the fault, we used a faster rupture speed of 4 km/sec, and assumed that every unit can move in the entire process. Our preliminary result shows that a 2 m water elevation first occurred between the epicenter and the trench, and then it propagated along the trench. About 3 minutes after the origin time, a large elevation occurred between 38°N and 39°N along the trench. This result is similar to Satake et al. (2013, BSSA). However, our result shows longer duration of sea surface elevation in the vicinity of the epicenter. Fig. Sea surface elevation at each minute after the earthquake origin time.