From Pixels to Piezometers: Bridging Satellite Gravimetry and Groundwater-Level Analysis to Estimate Aquifer Specific Yield

Groundwater is a critical resource in arid and semi-arid environments, where it supports agriculture and ensures drinking water. However, in many regions, including Morocco, this resource is experiencing significant stress, urging the need to understand the aquifers’ behavior. In this context, providing estimates of specific yield (Sy) is essential. This study combines terrestrial water storage (TWS) from GRACE-SeDA (Gravity Recovery and Climate Experiment Self-Supervised Data Assimilation), and the groundwater storage (GWS) residual obtained from it by subtracting modeled soil moisture, snow water equivalent, and reservoir storage, with groundwater levels. A detailed analysis of groundwater-level time series using Sen’s slope shows a general and systematic depletion in almost all of the piezometers, with level decline ranging from weak trends to strong declines above −0.14 m/month. Similarly, the cumulative monthly GWS from the median of the 25-member ensemble shows a decreasing trend with a Sen’s slope of −13.23 million cubic meters (MCM) per month. A robust Huber loss function is employed to optimize Sy by minimizing the residuals between derived GWS anomalies and changes inferred from groundwater-level observations. Sy estimates exhibit spatial variability across the aquifer, ranging from 0.59 to 3.49% for GWS-derived values. The specific yield estimates are drawn from a 15-member ensemble crossing three soil moisture and five snow water equivalent sources. The obtained results fall within the range of observed Sy values. Nevertheless, uncertainties in remote sensing retrievals, spatial resolution, and subsurface heterogeneity limit the robustness of the approach.