Harnessing Hydrochemical Characterisation and ANN-Driven Water Quality Modelling for Wetland Sustainability in Sudurpaschim Province, Central Himalaya, Nepal

Abstract

Wetland ecosystems in the Himalayan region face growing threats from climate change, human activities and environmental degradation. This study introduces an integrated approach to assess and predict the water quality index (WQI) for effective wetland management, focusing on the Alital and Bandatal Lakes in Nepal's Sudurpaschim Province. These lakes were selected due to their distinct ecological and geographical characteristics, as well as differing levels of human impact. A total of 40 water samples (20 from each lake) were collected, and 16 physicochemical parameters, including turbidity (Tur.), total dissolved solids (TDS) and major ions were analysed. Hydrochemical properties were characterised using graphical methods, such as the Gibbs and Piper diagrams and the WQI was computed using the arithmetic average method. The hydrochemical facies analysis indicated that carbonate weathering was the dominant process in both wetlands, with Bandatal showing significant anthropogenic influence. The findings revealed that Alital maintained ‘Excellent’ to ‘Good’ water quality, with an average TDS of 64 mg/L and Tur. of 2.14 NTU, reflecting minimal human impact. In contrast, Bandatal exhibited ‘Poor’ to ‘Unsuitable’ WQI classifications, with TDS averaging 115 mg/L and Tur. reaching 63.6 NTU, highlighting substantial human influences. An artificial neural network (ANN) model was developed to predict the WQI, demonstrating outstanding accuracy with an R² of 0.99 for both the training and testing phases. These results underscore the potential of the ANN model for proactive wetland management, aligning with sustainable development goals (SDGs) related to clean water and ecosystem restoration and providing globally applicable insights for wetland conservation.