Analysis of systematic differences from GPS-measured and GRACE-modeled deformation in Central Valley, California

Abstract

Crustal seasonal displacement signals, which are commonly attributed to surface mass redistributions, can be measured by continuous GPS, modeled by GRACE and loading models. Previous studies have shown that the three methods generally agree with one another. However, the discrepancy among them in some regions has not yet been investigated comprehensively. In this paper, we compare the vertical annual displacement signals in the Central Valley, California derived from GPS, GRACE and loading models. The results show a general agreement from these three methods for most sites, which reach the maximum during the dry late summer and autumn. Irregular annual terms with peaks during the wet winter and spring are detected from GPS solutions for the sites located in places with extensive groundwater depletion. However, annual vertical variations for these same sites derived from GRACE and loading models reach the maximum in August and minimum in February. To explain such apparent discrepancy, we find that the vertical components of abnormal sites show a strong correlation with in situ groundwater data, which display peaks during cold months. In addition, with the assistance of water table depth data, we perform hydrological simulations based on Terzaghi’s Principle, Mogi’s Model and Green’s function method. The results suggest that the discrepancy from GPS-measured and GRACE-modeled deformation is induced by the seasonal variations of groundwater.