Unveiling spatio-temporal shifts in the eco-hydrological state of an alpine lake using integrated GIS, geo-statistics, and machine learning approaches: insights into wetland sustainability

Mountain wetlands in the Himalayas are under increasing anthropogenic and climatic stress. In this study, a combined approach using GIS, geostatistical, and machine learning was adopted to analyse water quality and spatio-temporal changes in the eco-hydrological state of Khecheopalri Lake, a RAMSAR wetland situated in Sikkim, India. Observed water parameters and calculated indices (pH: 5.3; nutrient index: 4.73; Organic Pollution Index: 1.49) indicated a Class II level of contamination. The results suggested that the lake was eutrophic, with organic pollution spreading throughout the system, prompting immediate intervention to protect the lake's water resources and the communities depend on them. The lake surface area decreased by approximately 3.8% (from 0.079 km2 in 2013 to 0.076 km2 in 2023), and forecasting suggests a potential ∼4% additional decline by 2033 under the same conditions. These findings revealed distinct transformations and highlighted areas within the lake vulnerable to external influences. The study provides valuable insights into the interplay of human interventions and natural processes in shaping the lake's water quality. Moreover, it contributes to the sustainable management of freshwater resources in alignment with SDG6, aiming to mitigate anthropogenic threats and ensure the long-term viability of the wetland as a vital ecological and economic resource.