Groundwater is a massive portion of the total freshwater available, representing a primary source of water for domestic and agricultural uses, often threatened by climate change and human pressure. Its monitoring is thus a priority challenge for which satellite sensors can help in overcoming common issues related to in situ networks. This study proposes a novel approach for estimating groundwater recharge through satellite soil moisture, consisting in calibrating soil parameters commonly used in analytical formulations of the downward water flux due to gravitational forces through the inversion of the soil water balance and using rainfall rates as a benchmark. To do this, a soil moisture-based inversion approach is implemented over the Umbria region, central Italy, disposing of a dense network of piezometric stations for validation purposes. After a first step aimed at identifying the best performing combination of input data, 5 years (2017–2021) of groundwater recharge rates are estimated by exploiting 1 km Sentinel-1 soil moisture, 1 km potential evaporation, and in situ rainfall. Without any information on groundwater dynamics for calibration purposes, a good agreement between anomalies of estimated and ground-based monthly recharge is found. Over half of the pilot piezometric stations, monthly Pearson correlation is higher than 0.5 and the root mean square error is lower than 1.72. Performances increase the more the reference groundwater level records are correlated with rainfall, i.e., for undisturbed aquifers. Further analyses partially exploiting groundwater depth records for scaling estimated recharge rates to the observed ones show that yearly measured amounts can be quantified from satellite with a median percentage error of –6%. The proposed approach is a promising tool for remotely mapping and monitoring groundwater recharge, which is essential in the assessment of freshwater availability trends.
A novel approach for estimating groundwater recharge leveraging high-resolution satellite soil moisture
Dari, Jacopo
;Filippucci, Paolo;Morbidelli, Renato;Saltalippi, Carla;Flammini, Alessia
2025
Abstract
Groundwater is a massive portion of the total freshwater available, representing a primary source of water for domestic and agricultural uses, often threatened by climate change and human pressure. Its monitoring is thus a priority challenge for which satellite sensors can help in overcoming common issues related to in situ networks. This study proposes a novel approach for estimating groundwater recharge through satellite soil moisture, consisting in calibrating soil parameters commonly used in analytical formulations of the downward water flux due to gravitational forces through the inversion of the soil water balance and using rainfall rates as a benchmark. To do this, a soil moisture-based inversion approach is implemented over the Umbria region, central Italy, disposing of a dense network of piezometric stations for validation purposes. After a first step aimed at identifying the best performing combination of input data, 5 years (2017–2021) of groundwater recharge rates are estimated by exploiting 1 km Sentinel-1 soil moisture, 1 km potential evaporation, and in situ rainfall. Without any information on groundwater dynamics for calibration purposes, a good agreement between anomalies of estimated and ground-based monthly recharge is found. Over half of the pilot piezometric stations, monthly Pearson correlation is higher than 0.5 and the root mean square error is lower than 1.72. Performances increase the more the reference groundwater level records are correlated with rainfall, i.e., for undisturbed aquifers. Further analyses partially exploiting groundwater depth records for scaling estimated recharge rates to the observed ones show that yearly measured amounts can be quantified from satellite with a median percentage error of –6%. The proposed approach is a promising tool for remotely mapping and monitoring groundwater recharge, which is essential in the assessment of freshwater availability trends.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.