Using GRACE Terrestrial Water Storage Data for Groundwater Flow Model Calibration of Alaşehir-Sarıgöl Sub-Basin, Turkiye

Water resources are under immense stress due to the continuous increase of anthropogenic and natural pressures. Therefore, effective water management backed by advanced tools and methods is essential for the sustainability of water resources. One of these tools is groundwater flow modeling, which can be used to estimate changes in groundwater storage. In this study, we propose an approach to improve groundwater flow modeling by supporting model calibration with remote sensing data. The approach is demonstrated on the Alaşehir-Sarıgöl sub-basin in the east of the Gediz River Basin, a water-stressed basin in western Turkiye. A MODFLOW-2005 based flow model is constructed to determine time series hydraulic head changes and aquifer storage. The model simulation period is from 2013 to 2021. The groundwater recharge input of the model is obtained by a remote sensing-supported water balance method (Batkan et al., 2022). Except for precipitation data measured at meteorological stations, other model parameters are remote sensing products. Evapotranspiration is obtained from the MODIS Global Evapotranspiration product (MOD16A2), and soil water content and runoff are obtained from the ERA-5 Land Model reanalysis dataset. Hydraulic parameters such as hydraulic conductivity and storage coefficient are determined as a result of the calibration of the groundwater flow model. Model performance is improved by using terrestrial water storage (TWS) data from NASA's GRACE mission in the calibration of the storage coefficient. TWS represents the total water content above and below ground in the unconfined aquifer, therefore data needs to be adjusted to obtain an estimate of groundwater storage. Streams in the region can be ignored as a contributor to the TWS as they are intermittent and have typically low discharges. The soil water content in the unconfined aquifer is determined using ERA-5 data. The calibrated model RMSE value is 7.4 m, which was subsequently improved to lower values after the conjunctive use of the GRACE-derived TWS data.

Keywords: groundwater flow modeling, model calibration, remote sensing, GRACE, ERA-5

Acknowledgment: This study is funded by the PRIMA program supported by the European Union under grant agreement No: 1924, project RESERVOIR (sustainable groundwater RESources managEment by integrating eaRth observation deriVed monitoring and flOw modelIng Results).