Extracting and separating different sources of hydrology-induced deformation in geodetic datasets (Invited)

Abstract : Space-based geodesy offers a new, complementary way to observe hydrological processes that deform the solid Earth. Known sources of hydrogeodetic deformation include changes in localized hydrological loads such as lakes, larger-scale variations in terrestrial water storage linked to climate as well as fluctuations in groundwater levels which can activate a poroelastic or inelastic porous response. Discriminating between these distinct sources of deformation in geodetic datasets such as GNSS and InSAR is essential to accurately invert for regional fluctuations in water mass and constrain aquifer hydromechanical properties. Blind source separation techniques such as Independent Component Analysis (ICA) help in isolating hydrology-induced deformation from other sources of deformation and noise in geodetic datasets but cannot necessarily distinguish between statistically-correlated hydrological processes. Here, we propose a general framework to accomplish this task by relying on continental-scale gravimetric observations and local field measurements. We first account for deformation due to long-wavelength loads by considering the response of a spherical elastic PREM Earth to hydrological loads inferred from GRACE. We then project the residual geodetic time series onto temporal functions representative of local hydrology and hence extract the associated deformation. The temporal functions may be sampled at a single point (e.g., lake level time series) or may require statistical analysis to be extracted from a heterogeneous dataset (e.g. network of groundwater monitoring wells). The final step consists in validating the extracted signals with simple elastic, poroelastic and inelastic deformation models. We demonstrate the methodology through case studies in different hydrological settings.