Modeling vertical deformation along the San Andreas Fault System using geodetic, geologic, groundwater, and tide gauge data

Vertical motions along the San Andreas Fault System (SAFS) are recorded by several datasets. Geodetic (EarthScope PBO GPS & InSAR) data offer ideal spatial coverage, but often include anthropogenic effects and span short time periods (~5-20 years). Geologic data (from the SCEC Vertical Motion Database) primarily capture tectonic motions over long time periods (10 Ka to 7 Ma), but offer sparse spatial coverage. Tide gauge data (from the Permanent Service for Mean Sea Level) provide a temporal record of sea level change over intermediate time periods (~30-100+ years) and reflect variations in crustal uplift and subsidence at a few isolated locations along the California coastline. Using a 3-D viscoelastic earthquake cycle deformation model spanning the last 1000 years, this study aims to explore the first-order vertical tectonic motions reflected in each of these data sets. Previous work has shown that vertical GPS and geologic data in southern California do not correlate well, possibly related to groundwater contamination and the different timescales of the datasets. To isolate groundwater deformation recorded by the GPS data, a simple groundwater correction was applied to these data using regional well log data. To isolate a tectonic signal in the tide gauge data, global sea level rise and isostatic adjustment were removed from each station and additional processing was applied to eliminate major ocean-climate signals. A suite of vertical deformation models was then computed, reflecting variations in elastic plate thickness and mantle viscosity, to search the model parameter space for the optimal parameters that minimize residual data-model misfit. Preliminary results suggest that GPS, geologic, and tide gauge datasets are best fit by thick elastic plates ranging from 70 to 90 km and viscosities ranging from 1.5e18 to 1.5e19 Pa s. These values are consistent with previous results, however further work is needed to investigate sources of misfit and unmodeled phenomena. Additional work will include evaluation of deformation rates and a resulting optimized model using an integrated GPS - ALOS InSAR velocity field. A particular emphasis of this study will be placed on comparison of tide gauge time series with model-generated vertical motion from interseismic strain accumulation and major earthquake offsets over the last 100 years.