Use of Remote Sensing Data in Assessing the Impact of Climate and Land Use Land Cover Change on Groundwater Dynamics in Semi-Arid Regions

Abstract

This study examined how Land Use Land Cover (LULC) and climate affect groundwater dynamics in Southern India's semi-arid region (Chitravathi River basin) and identified the suitable sites for constructing the artificial recharge structures using geographical information systems (GIS). Coupled Model Intercomparison Project Phase6-Global Circulation Models (CMIP6-GCMs) climatic data is used to generate climate projections for the future. The GCMs are ranked for precipitation and temperatures using the Taylor Skill Score (TSS). Rating Metric (RM) was preferred to establish the final rank of the GCMs. Ensemble of projections from the top four ranked GCMs (MPI-ESM1-2-LR, EC-Earth3, MPI-ESM1-2-HR, and INM-CM5-0) were used as they estimated the most reliable forecasts for all the three considered parameters. MPI-ESM1-2-LR was the top-ranked GCM with an RM of 0.92. By using the ensemble GCMs, the six extreme precipitation indices, namely, consecutive dry days (CDD), consecutive wet days (CWD), total wet-day precipitation (PRCPTOT), R10mm (days), R95p (very wet days in mm), RX1day (maximum 1-day precipitation in mm) were calculated as per Expert Team on Climate Change Detection and Indices (ETCCDI) recommendations. Trend analysis of all the above parameters was calculated using Mann-Kendall (MK) and Spearman's rho tests. The future LULC map was produced using cellular automata and artificial neural networks (CA-ANN). Using standard modelling techniques, the SWAT model was used to evaluate the individual impact of climate change on groundwater recharge and the combined impacts of LULC and climate change. The SWAT model was calibrated for discharge data on a monthly basis at a gauging station. The overall accuracy of the SWAT was R2 = 0.83 and NSE = 0.81. The SWAT groundwater module estimates recharge for baseline (1985 - 2014), near-future (2015 - 2030), mid-future (2031 - 2060), and far-future (2061 - 2100) under the moderate SSP2-4.5 and extreme SSP5-8.5 emission scenarios. MODFLOW steady-state groundwater flow model was employed to predict future groundwater levels. Calibration of the model was performed based on seasonal groundwater level data spanning the years 2014 - 2022, and validation was carried out using data from 2020 - 2022. MODFLOW model exhibited good overall accuracy, with R2 values of 0.96 during calibration and 0.94 during validation. Based on the projected groundwater recharge and levels, a resiliency map of the basin was developed. Results revealed that recharge during constant LULC conditions ranged from 135 to 215 mm/year under SSP2-4.5 and 149 to 316 mm/year under the SSP5-8.5 scenario. Compared to baseline recharge (116.4 mm), the future groundwater recharge under both SSPs increased. The results also indicated that by the year 2060, under the SSP2-4.5 scenario, groundwater levels in the basin would decrease by 54 m, while under the SSP5-8.5 scenario, the decrease would be 62 m. By 2060, both SSPs indicate poor groundwater resiliency. Observations from the study highlight the non-resilient state of groundwater. In response, a study was conducted to identify locations conducive to artificial recharge of the groundwater system. Using remotely sensed data and GIS tools, thematic maps incorporating soil type, geology, topography, and groundwater information were overlaid. The Weighted Overlay Analysis (WOA) with assigned weighted scores and the Analytical Hierarchy Process (AHP) identified favourable sites for groundwater recharge. The resultant map guides the spatial distribution of these sites, offering valuable insights to safeguard and sustain the basin's groundwater resources.