The 2004 Sumatra-Andaman Earthquake: Non-linear Joint Inversion of Geodetic and Tsunami Data

The December 26, 2004 an earthquake of M_w=9.2 initiated off the West coast of northern Sumatra and propagated northward up to the Andaman and Nicobar Islands: it was the largest event ever occurred during the last 40 years, since the beginning of the era of modern digital recording. This earthquake generated the most devastating tsunami of the last centuries, causing more than 250,000 fatalities and spreading massive damages along the coast of the whole Indian Ocean. The aim of this work is to infer the characteristic of the earthquake rupture, in terms of slip distribution, rupture speed and rise time along the causative fault, by using different datasets. In particular we use several tide- gauge records distributed along the coast of the Indian Ocean, satellite measured sea level anomalies collected on passing of the tsunami wave at the open sea and GPS measurements of the coseismic static displacement. We divide the fault into a set of subfaults both along strike and dip direction. Tsunami waves are computed by numerically solving the non-linear shallow water equations through a finite- difference technique in a staggered grid. The coseismic GPS offsets are calculated using the semi-analytic global deformation model, which is a spherical, layered, self-gravitating incompressible model which can account for both elastic and viscoelastic responses induced by a point dislocation. The global solution is then computed as the linear superposition of the solutions corresponding to the individual subfaults (Green's functions), for both the case of tsunami waves and GPS displacement. We search for the rupture parameters simultaneously, by means of a two-stages non-linear inversion procedure based on the simulated annealing technique that allows extracting the stable features of the earthquake rupture. We perform several tests, by inverting the data separately and jointly, to validate the method and to estimate the resolution.