Investigating earthquake cycle vertical deformation recorded by GPS and regional tide gauge stations in California

Geodetic and tide gauge measurements of vertical deformation record localized zones of uplift and subsidence that may document critical components of both long and short-period earthquake cycle deformation. In this study, we compare vertical tide gauge data from the Permanent Service for Mean Sea Level (PSMSL) and vertical GPS data from the EarthScope Plate Boundary Observatory (PBO) for 10 approximately co-located station pairs along coastal California from Point Reyes, CA to Ensenada, Mexico. To compare these two datasets, we first truncate both datasets so that they span a common time frame for all stations (2007 - 2012). PSMSL data are treated for both average global sea level rise (~1.8 mm/yr) and global isostatic adjustment. We then calculate a 2-month running mean for tide gauge and a 1-month running mean for GPS datasets to smooth out daily oceanographic or anthropologic disturbances but maintain the overall trend of each dataset. As major ocean-climate signals, such as El Nino, are considered regional features of the Pacific Ocean and likely common to all California tide gauge stations, we subtract a reference sea level record (San Francisco station) from all other stations to eliminate this signal. The GPS and tide gauge data show varying degrees of correlation spanning both 3-month and 4-year time-scales. We infer that the slope of vertical displacements are largely controlled by interseismic motions, however displacements from major earthquakes are evident and are required to explain some of the unique signatures in the tide gauge and GPS data. Specifically, we find that stations from both datasets in Southern California show an anomalous trend since the 2010 Baja California earthquake. To further investigate this trend and others, we compare these data to vertical motions estimated by a suite of 3-D viscoelastic earthquake cycle deformation models. Long-term tide gauge time series are well simulated by the models, but short-term time series are not as well predicted; additional parameter adjustments are needed to improve these. Alternatively, both tide gauge and GPS data show a better short-term than long-term correlation; oceanographic and possibly groundwater effects could be responsible for these differences.