Downscaling WGHM-Based Groundwater Storage Using Random Forest Method: A Regional Study over Qazvin Plain, Iran

Climate change, urbanization, and a growing population have led to rapid increase in groundwater (GW) use. As result, monitoring changes is essential for water managers decision-makers. Due the lack of reliable insufficient situ information, remote sensing hydrological models may be counted as alternative sources assess GW storage on regional global scales. However, often, these low spatial resolution water-related applications small scale. Therefore, main purpose this study downscale anomaly (GWSA) WaterGAP Global Hydrology Model (WGHM) from coarse (0.5 degrees) finer (0.1 using fine auxiliary datasets degrees), such evaporation (E), surface (SRO), subsurface runoff (SSRO), snow depth (SD), volumetric soil (SWVL), ERA5-Land model, well precipitation (Pre) measurement (GPM-IMERG) product. The Qazvin Plain central Iran was selected case region, it faces severe decline resources. Different statistical regression were tested GWSA downscaling find most suitable method. Moreover, since different budget components (such or storage) are known temporal lead lag relative each other, approach also incorporates time shift factor. model with highest skill score during training-validation applied predict final 0.1-degree GWSA. downscaled results showed high agreement levels over both interannual monthly scales, correlation coefficient 0.989 0.62, respectively. product represents clear proof that developed technique able learn high-resolution data capture features at higher resolution. major benefit proposed method lies utilization only available coverage free charge, while not requiring records training prediction. can potentially scale aquifers other geographical regions.