T21D-2857
The Nature of Co-seismic Rupture Zone of the 2010 Mentawai Tsunami Earthquake from Full Waveform Inversion of Long Offset Seismic Reflection Data
Tuesday, 15 December 2015
Poster Hall (Moscone South)
Gabriel Huot and Satish Chandra Singh, Institut de Physique du Globe de Paris, Paris, France
Abstract:
The Sumatra subduction zone is one of the most seismically active zone on Earth. In the last one decade alone, it has hosted three Mw>8.4 great earthquakes (2004, 2005, 2007) along with 2010 tsunami earthquake. Although the 2007 Mentawai earthquake had Mw=8.4, it did not produce tsunami whereas the 2010 earthquake had Mw=7.8 only in the same region, it produced a large tsunami with a run up height of up to 8 m on Pagai Island, taking 800 lives. Therefore, understanding why an earthquake produce tsunami is fundamental for risk assessment as well for subduction zone processes. Prior to the 2010 earthquake we had acquired ultra-long offsets seismic reflection data in 2009 in the co-seismic slip zone using a 15 km long streamer, the longest streamer ever used, and found that the earth ruptured the frontal section of the subduction zone, which is normally believed to be aseismic, and possibly produced the tsunami. In order to quantify the nature of the co-seismic rupture zone and its link with the tsunami generation, we performed full waveform inversion of seismic reflection data. In order to obtain the high-resolution velocity model for the full waveform inversion, we first downward continue the data to the seafloor, picked first arrivals, and performed tomography. We used the tomographic velocity model as an input to the full waveform inversion. This process also reduced the computation cost significantly as the water depth in this area is 5.5 km. The resulting models shows the presence of thrust faults extending up to the subducting oceanic plate, suggesting that the frontal section of the subduction in this region was indeed locked, capable of hosting great earthquakes. Our inverted model provides the resolution of tens of meters, allowing to characterize the nature of the megathrust and other faults, and hence estimate the effective porosity, permeability and stress along these faults, subsequently the pore pressure.