Updating HTDP for two recent earthquakes in California

This paper describes the National Geodetic Survey (NGS) process for selecting and validating models for two recent earthquakes prior to their inclusion in the Horizontal Time Dependent Positioning software (HTDP v 2.9), the 28 September 2004 Parkfield, California earthquake with a magnitude of 6.0 and the 22 December 2003 San Simeon, California earthquake with a magnitude of 6.5 . The process of modeling and correcting for deformation can be well illustrated by utilizing the data from these earthquakes because they represent contrasting styles of deformation; Parkfield exhibited strike-slip motion and San Simeon was predominantly a blind-thrust earthquake. Also, while both earthquakes were relatively small, they produced significant displacements (though restricted to a relatively small area near the epicenter). The Parkfield earthquake in particular produced an unusually high proportion of post-seismic deformation. This deformation can accumulate during a period of a few months after the earthquake compared to co-seismic deformation that occurs suddenly during an earthquake. Quantifying and characterizing deformation at these disparate time scales represents a major challenge to developing the accurate models of deformation that are required to support modern high-accuracy surveying in tectonically active areas like California. We found that the best dislocation models for this purpose are determined from inverting a combination of GPS vectors and InSAR data as opposed to ones derived from seismic body waves. This is because including direct measurements of the deformation field provides a more faithful estimate of changes in relative position than would otherwise be the case. This study also shows the importance of including separate dislocation models of the co-seismic and post-seismic deformation, which is important because the post-seismic deformation can be comparable in magnitude to the co-seismic deformation, as it was for the Parkfield earthquake.