Data-Driven Approach to Assess Spatial-Temporal Interactions of Groundwater and Precipitation in Choushui River Groundwater Basin, Taiwan

Abstract

The scarcity of groundwater and precipitation stations has limited accurate assessments of basin-scale groundwater systems. This study proposes a workflow that integrates satellite and on-site observations to improve the spatial and temporal resolution of the groundwater level and enable recharge estimations for the Choushui River groundwater basin (CRGB) in Western Taiwan. The workflow involves multiple data processing steps, including analysis of correlation, evaluation of residuals, and geostatistical interpolation based on kriging methods. The observed groundwater levels and recharge are then the basis to assess spatial-temporal interactions between groundwater and recharge in the CRGB from 2006 to 2015. Results of correlation analyses show the high correlation between the groundwater level and the land surface elevation in the study area. However, the multicollinearity problem exists for the additional precipitation data added in the correlation analyses. The correlation coefficient, root mean square error, and normalized root mean square parameters indicate that the Regression Kriging (RK) performs better the groundwater variations than the Ordinary Kriging (OK) dose. The data-driven approach estimates an annual groundwater recharge of approximately 1.40 billion tons, representing 37% of the yearly precipitation. The correlation between groundwater levels and groundwater recharge exhibits low or negative correlation zones in the groundwater basin. These zones might have resulted from multipurpose pumping activities and the river and drainage networks in the area. The event-based precipitation and groundwater level have shown strong recharge behavior in the low-land area of the basin. Artificial weir operations at the high-land mountain pass might considerably influence the groundwater and surface water interactions.

Cost-Effective Groundwater Potential Mapping by Integrating Multiple Remote Sensing Data and the Index–Overlay Method

“…The minimum of the possibility is used as the overall score for each criterion. The groundwater potential index (GWPI) is a dimensionless and quantification index used to calculate potential groundwater scores in various areas by integrating thematic layers and the corresponding weightings and ratings [28]. The GWPI can be determined by integrating these thematic layers into a Geographic Information System (GIS) platform.…”