Detecting and Reducing the Effect of Ongoing Postseismic Deformation in Modern GPS Time Series Using a Hybrid Forward/Inverse Method in Southern California

High precision geodetic measurements of plate boundary zone deformation have provided an invaluable opportunity to analyze all aspects of the earthquake cycle over the last three decades. Yet, we still struggle to piece apart long- and short-term seismic and aseismic signals present in modern time series, in order to gain insight into long-term plate boundary motion and to shed light on long-term fault slip rates and the seismic hazards they help inform. Here, we apply a new hybrid forward/inverse modeling technique to identify and reduce the short- and long-term effects of viscoelastic postseismic deformation as observed by modern GPS instruments following large magnitude earthquakes in the southern California region. We assess ongoing postseismic deformation in the southwestern U.S. by analyzing all magnitude ≥M6.0 earthquakes that have occurred there and in Baja California and Sonora, M.X. since year 1800, finding that twelve events have measurable postseismic displacements in the modern day. With a forward modeling step, we produce modeled displacement time series of these twelve events using a reference model consisting of a laterally homogenous viscoelastic earth structure, which are then subtracted from modern GPS time series to produce a postseismic reduced dataset. In order to quantify the success of this forward model in reducing the postseismic signal, we use time series trend estimation to compare estimated logarithmic terms between our reduced dataset and our observed dataset for the 2010 M7.2 El Mayor-Cucapah earthquake. Using a calculated variance reduction number, our forward modeling reduced the postseismic signal for the El Mayor-Cucapah event across our southern California GPS network by 58% when calculated using TSFIT-estimated values and by 80% when calculated using Hector-estimated values. Anomaly maps produced using this hybrid forward/inverse method will guide our ongoing analysis as they highlight hotspots in which secondary processes may be occurring or where our laterally homogenous viscoelastic reference model may need adjusting.