Modelling water use in Nepal’s highlands: a multidisciplinary and probabilistic framework

Mountain communities in Nepal are increasingly exposed to climate-induced shifts in water availability, driven by glacial retreat, altered precipitation/snowmelt regimes, and declining groundwater sources. This study presents an integrated framework combining hydrological source analysis with socio-demographic survey data to evaluate seasonal water contributions and communitylevel water use patterns in the Upper Marsyangdi catchment, Manang District, Nepal. Isotopic (δ18O) and geochemical (silica) tracers were used in a Bayesian mixing model to quantify the seasonal contributions of glacial melt, snow, rain, and groundwater to river flow. Findings indicate that groundwater dominates pre-monsoon flow (60%–70%) while post-monsoon discharge reflects more balanced inputs from all sources. In parallel, 120 household surveys were analysed using Latent Class Analysis to characterise water use across domestic, agricultural, energy, and tourism sectors. Results reveal spatial and demographic gradients in water source dependency, including gender and occupation as important predictors of water use. Respondents reported perceived increases in spring flow, alongside reductions in the availability of snow for household and tourism use and deteriorating river water quality and quantity, particularly affecting hydropower operations. Adaptation strategies include increased reliance on water storage infrastructure and source switching. The study highlights the value of applying probabilistic methods to hydrological and sociocultural data to identify vulnerable populations and inform targeted, context-sensitive adaptation strategies. The proposed framework is transferable to other high-altitude regions, offering a robust approach for assessing climate resilience through the synthesis of scientific and local knowledge systems.