Modeling Wetland Habitat Quality in the Rarh Tract of Eastern India

Abstract

Along with wetland loss, wetland habitat quality degradation is a growing concern that requires immediate attention. The current study aimed to assess the Wetland Habitat Quality State (WHQS) of Rarh region, Murshidabad, West Bengal. WHQS used a total of seventeen metrics, including water quality, hydrology, and landscape composition. Machine learning techniques such as ANN, SVM, RF, BAGGING, and REP-TREE were used to model WHQS. The effectiveness of the models was evaluated using statistical techniques such as the Receiver operating characteristics (ROC) curve. According to machine learning models, 6% of the area fall under very weak habitat quality zones in 1990 which increased by 15%, 26%, 41% in 2000, 2010 and 2020, respectively. Very strong portions of wetland area have been decreased from 32.74% in 1990 to 20.72% in 2020. The current study's findings could provide comprehensive research on the monitoring of habitat quality in wetlands, which will serve as the foundation for developing water resource management plans for the conservation, management, and restoration of wetlands.