A multi-model approach for analyzing water balance dynamics in Kathmandu Valley, Nepal

Study region: Kathmandu Valley, Capital city of Nepal. Study focus: This study applied three hydrological models (i.e., SWAT, HBV, and BTOPMC) to analyze the water balance components and their temporal and seasonal variations in the Kathmandu Valley, Nepal. The water balance components were investigated using the same precipitation, climatic data, and potential evapotranspiration (PET) as input variables for each model. The yearly and seasonal variations in each component and the interactions among them were analyzed. There was a close agreement between the monthly observed and calibrated runoff at the watershed scale, and all the three models captured well the flow patterns for most of the seasons. New hydrological insights for the region: The average annual runoff in the study watershed calculated by the SWAT, HBV, and BTOPMC models was 887, 834, and 865 mm, corresponding to 59%, 55%, and 57% of the annual precipitation, respectively. The average annual evapotranspiration (ET) was 625, 623, and 718 mm, and the estimated yearly average total water storage (TWS) was 5, -35, and 29 mm, respectively. The long-term average TWS component was similar in all three models. ET had the lowest inter-annual variation and runoff had the greatest inter-annual variation in all models. Predictive analysis using the three models suggested a reasonable range for estimates of runoff, ET, and TWS. Although there was variation in the estimates among the different models, our results indicate a possible range of variation for those values, which is a useful finding for the short- and long-term planning of water resource development projects in the study area. The effects of historical water use, water stress, and climatic projections using multi-model water balance approaches offer a useful direction for future studies to enhance our understanding of anthropogenic effects in the Kathmandu Valley