Enhancing groundwater potential evaluation: Integrating borehole log data with hybrid-MCDM approach

Santanu Mallik,Avishek Chaudhuri,Tushar Das,Umesh Mishra

DOI: https://doi.org/10.1007/s12040-024-02462-5

2024-12-01

Journal of Earth System Science

Abstract:Groundwater plays a crucial role in the hydrologic cycle and serves as a vital natural resource for our ecosystem. With the increasing challenges associated with groundwater depletion and the necessity to develop effective mitigation strategies, it is crucial to identify areas with significant groundwater potential. The combination of geospatial tools and multi-criteria decision-making (MCDM) techniques has proven to be a highly effective approach for delineating groundwater potential zones. However, there is ambiguity in selecting the appropriate MCDM method and determining the relevant factors to delineate groundwater potential zones at a local scale. To overcome this issue in this study, a comprehensive framework has been developed that amalgamates different MCDM methods with geophysical and borehole log data. Four distinct models were developed: Analytic Hierarchy Process-Catastrophe Theory (AHP-CT), Analytic Hierarchy Process-Shannon Entropy (AHP-E), modified AHP-CT, and modified AHP-E method with the inclusion of borehole log data. The outcome of the groundwater potential maps indicates that a significant portion (37%) of the northwestern and southwestern regions of the study area Agartala city falls within the classification of high groundwater potential zones. The model outputs were validated using statistical performance indicators. The modified AHP-CT method exhibited the highest correlation coefficient of 0.68 with average groundwater depth, followed by the modified AHP-E method with a correlation coefficient of 0.63. The groundwater potential map was obtained using the Modified AHP-CT technique, identified 43.75, 19.71, and 13.04 km 2 of the study area as good, moderate, and poor groundwater potential zones, respectively. These results demonstrate that the hybrid methods employed in this study, particularly when incorporating borehole log data, yielded improved outcomes at a local scale. Therefore, in situations where data limitations hinder the use of machine learning models, the proposed methodology serves as a viable alternative for predicting and assessing groundwater potential.

geosciences, multidisciplinary