Spatial correlation of groundwater level with natural factors using geographically weighted regression model in the Choushui River Alluvial Fan, Taiwan

Abstract

The groundwater of the Choushui River alluvial fan in Central Taiwan has been overexploited for a long time. It is essential to understand the factors governing changes in groundwater level (GWL) for the use of water resources. In this study, we first conducted a Mann–Kendall test to identify significant trends in the regional GWL and obtained its spatial characteristics using the Moran’s I index in the Choushui River alluvial fan. Furthermore, we established a geographically weighted regression (GWR) model to explore the spatial correlation between natural factors and GWL in dry and wet seasons from 1999 to 2019. The long-term trend analysis shows that the GWL of the Choushui River alluvial fan decline significantly. The Moran’s I index shows that the spatial distribution of GWL had a positive correlation in both dry and wet seasons. GWR model indicate that the GWL are affected by drainage density (Dd), slope (S), normalized difference vegetation index (NDVI), and precipitation (P) during the dry season, while Dd, S, NDVI, and wetness index (WI) have an effect on the GWL during the wet season. These results can not only describe the model applicability for exploring the relationship between natural factors and GWL but also be used as references for future regional water resource utilization and management.