Constraints on Postseismic Deformation Processes After the December 2004 and March 2005 Sumatra Earthquakes From Continuous GPS Observations

The December 26, 2004, and March 28, 2005, Sumatra earthquakes are the two largest earthquakes ever to be monitored by the Global Positioning System (GPS). GPS data have already provided unique constraints on the rupture process. Here we present GPS analysis results of the postseismic deformation that has occurred, and continues to occur, after both events. We will concentrate on the differences between the two earthquakes and evaluate whether the postseismic deformation after the December event could have contributed to, or triggered, the March event. Initial analysis of the time-series of sites SAMP (in northern Sumatra) and NTUS (Singapore) has indicated that the postseismic deformation after both events is significant and is consistent with an afterslip process, i.e., the time-series can be well modeled with a logarithmic function. We find from the analysis of these two stations that the expected post-seismic deformation after one year from the occurrence of the first event is expected to be 10-30% of the coseismic displacements, while it is likely >80% for the second event. We also find that the directions of the post-seismic offsets are not parallel to the coseismic offsets indicating that the afterslip is more complex than continued slip on the coseismic rupture plane. We extend our analysis in order to better understand the differences between the postseismic deformation for both events (and their relationship to the coseismic rupture) and to place better constraints on the location of the afterslip processes with respect to the rupture. We analyze data from Caltech's SuGAr network together with stations SAMP and NTUS and other regional stations of the IGS network. We use all available data to simultaneously solve for the coseismic and afterslip parameters of both events. We will present dislocation models using the inferred postseismic displacements in order to locate the afterslip and also attempt to identify the possible effects of viscoelastic relaxation that may now becoming evident in the postseismic time-series.