Analysing sensitivity of conditioning factors for wetland habitat quality modelling

Abstract

Wetland habitat quality degradation, in addition to wetland loss, is a growing concern that demands rapid action. The purpose of the current study is to assess the sensitivity analysis of conditioning factors on the Wetland Habitat Quality State (WHQS) in the Rarh tract of Murshidabad, West Bengal. Total seventeen metrics in total, including water quality, hydrology, and landscape composition were employed to mapping WHQS. WHQS was modelled using machine learning techniques. The efficiency of the models was evaluated using statistical methods such as the Receiver operating characteristics (ROC) curve. Very strong portions of wetland area have been decreased from 32.74% in 1990 to 20.72% in 2020. According to the study, wetland fragmentation, distance from agricultural land, and trophic status factor of the wetlands are key elements that affect wetland habitat quality to a large extent. The current study's findings could serve as the foundation for developing water resource management plans for the management, conservation, and restoration of wetlands, in addition to doing extensive research on monitoring habitat quality in wetlands.