Estimation of surface hydrological diffusivity and atmospheric flux bias using GRACE satellite data

Variations in terrestrial water storage (TWS), as observed by the GRACE/GRACE-FO  missions, provide unique insights into large-scale hydrological processes. However, translating these satellite observations into transport parameters such as surface diffusivity, lateral water fluxes, and groundwater recharge remains challenging. In this study, we propose using a surface diffusion-advection model coupled with a WGHM data assimilation framework of gridded GRACE solutions to estimate subsurface diffusivity and systematic precipitation–evapotranspiration biases simultaneously. The global kinematic hydrology model represents the lateral and vertical transport of water by diffusion, while GRACE observations represent the total water storage. In the steepest descent 4D Var-like procedure, the parameter gradients of the objective function are computed using the hydrological model's adjoint. Errors on derived diffusivities are also computed. The optimised parameters enable us to diagnose effective surface diffusivity and lateral water fluxes, as well as net groundwater recharge. This framework provides a physically consistent interpretation of GRACE-observed mass redistribution and offers new perspectives on large-scale hydrological transferts.