Comparison of the shallow groundwater storage change estimated by a distributed hydrological model and GRACE satellite gravimetry in a well-irrigated plain of the Haihe River basin, China

Abstract

Groundwater is a key component of the global hydrological cycle. Quantitative approaches to estimate the spatial and temporal changes in groundwater storage (GWS) could provide solid foundation for sustainable management. This study made a comprehensive comparison of GWS changes based on two independent approaches: the Soil and Water Assessment Tool (SWAT) hydrological model and Gravity Recovery and Climate Experiment (GRACE) satellite mission data. The feasibility of using these two estimation methods to quantify shallow groundwater variations in the Haihe River basin of China, a globally representative overexploited area, was assessed. The GRACE-derived GWS changes were calculated based on independent component analysis and a joint inversion method. The results showed that the GRACE-estimated shallow GWS anomalies during 2003–2012 agree well with the SWAT-simulated data. For the variations in the long-term GWS depletion rate, we found that heterogeneity at 22 subbasin (approximately 500 to 1500 km2) scales was detected by the GRACE data, and these results were spatially coherent with the SWAT simulations based on the distributed modeling units and multiple inputs. The GRACE estimates could adequately replicate the general trends of GWS depletion in the winter wheat (Triticum aestivum L.) growth period and recovery in the summer maize (Zea mays L.) growth period. Specifically, comparisons between GRACE-based and SWAT-simulated results indicated that the signals of pumping for winter wheat irrigation in several key water requirement periods in April, May and November were identified by GRACE, as were seasonal changes in shallow GWS (e.g., an obvious depletion in spring). Despite an overestimation of the scope of the fluctuations in GRACE estimates compared with those of SWAT simulations from the perspective of the water flux, the effective identification of the impact of anthropogenic exploitation on shallow GWS in GRACE suggests that this monitoring tool could be applied in the groundwater management of this plain. This study demonstrated a good agreement between SWAT-simulated and GRACE-based shallow GWS changes from the perspectives of long-term trends and variations in crop seasons. Additionally, satisfactory consistency was observed from the perspective of the water flux in this semiarid and semihumid study region considering the influence of human activities.