The seismic and tsunami hazard posed by great subduction zone earthquakes has long been recognized, but in light of the December 26, 2004 Mw 9+ Sumatra-Andaman Islands earthquake, the resulting tsunami, infrastructure damage and huge loss of life, it is clear that the danger posed by such events remains underestimated. The Sumatra great earthquake is one of the four largest events in the instrumental record, comparable in many respects to the 1960 Mw 9.5 Chilean and 1964 Mw 9.2 Alaska subduction megathrust events that generated significant damage and tsumani within the Pacific ocean basin. The Sumatra-Andaman Islands event triggered a second great earthquake, the March 28, 2005 Mw 8.7 Nias event, directly to the south, increasing concerns about the occurrence of a third great earthquake offshore southern Sumatra in the near-term. This study consists of a series of high-precision earthquake relocation of the 2004 and 2005 Sumatra aftershock sequences and prior regional seismicity and computation of compressional and shear seismic wave speeds in three-dimensions for the Sumatra region in order to better constrain how thermal, mechanical, compositional, hydrological, and rupture processes interact along subduction megathrusts during great earthquakes. The 2004 and 2005 main shock and aftershock sequences generated over 5000 earthquakes recorded globally by seismic networks along 1700 km of subduction system. In the Sumatra region, however, the seismogenic portion of the subduction megathrust lies primarily within an oceanic environment, which ultimately limits the amount and coverage of local and regional seismic data that will become available, and emphasizes the need for state-of-the-art analysis of the available global data. The increased precision earthquake catalog is used to refine subduction zone geometry and to explore spatio-temporal variability within and between the aftershock series. Comparisons between aftershocks and prior regional seismicity provide insight into processes within the subduction zone leading up to these great earthquakes, while comparisons between compressional and shear wave speed models and regional seismicity patterns can be related to variations in rupture behavior along the subduction megathrust. The study builds over three stages: 1) probabilistic phase re-determination, waveform cross-correlation of phases to reduce picking error, and single-event earthquake relocations using the Engdahl, van der Hilst, and Buland method, 2) high-precision relative earthquake locations using teleseismic double-difference methods, and 3) inversion for regional compressional and shear wave speed models in three dimensions using teleseismic DD tomography methods. Among the broader impacts are: 1) The methods developed significantly advance the use of global catalog and waveform data to study regional-scale problems where high spatial density local seismometer deployments are not feasible, and all techniques are shared with the public; 2) Data analysis and interpretation involves undergraduate, graduate, and young investigator collaboration with well-established experts in the fields of earthquake location and tomography; and 3) The project involves five institutions, including the Institut Teknologi Bandung, and strengthens ties with the scientific community in Indonesia.
