Modelling wetland eco-hydrological state and its role on ecosystem service potentiality

The application of the machine learning model for modelling environmental susceptibility is widespread. However, its application coupling with the Pressure-State-Response (PSR) framework for modelling wetland eco-hydrological state (EHS) is rare but it is necessary for wetland management and restoration. Dam-induced hydrological modification is a crucial driver of EHS change. How EHS conditions can control ecosystem services potentiality (ESP) including provisioning, regulating and cultural services is another vital unresolved issue that needs to be analysed for the well-being of the people. Considering this, the present study tried to model the EHS of wetlands in the Tangon River of India and Bangladesh blending both machine learning algorithms and PSR framework. It also aimed to assess the influence of EHS on ESP. Tangon River is a floodplain river, which is hydrologically intervened through the Boda dam in the Panchagarh district of Bangladesh, which causes hydrological modification in the river and riparian floodplain wetlands. All the applied models successfully explored that 28.90% to 47.89% of the wetland area under a very good EHS zone was reduced to 28.90 % to 31.41%, indicating the gradual degradation of EHS quality over a wider part of the wetland after hydrological intervention. The performance of the applied models is acceptable, but REPTree and Bagging models were found best-performing models for EHS modelling. This study also established that the EHS condition of the wetlands positively influences the ESP of the wetlands. So, improvement of EHS is a way forward for promoting the serviceability of the wetland.