Long-range streamflow prediction using a distributed hydrological model in a snowfed watershed

Inadequate planning of spring snowmelt discharge can lead to wastage of water resources for various purposes such as hydropower and lead to reduced capacity of the dams to control floods during rainy season. In this research we analyze how much can the predictive skill of long-range forecasts be improved by using a distributed hydrological model. We used the Water and Energy Budget-based Distributed Hydrological Model with improved snow physics (WEB-DHM-S) for generating long-range forecasts with a lead time of up to 3 months for the case of Kurobe River Basin in Japan. The predictive skills of two sets of simulations were compared (i) climatology and (ii) ensemble stream flow prediction (ESP). In the case of ESP, the initial conditions of WEB-DHM-S are updated using real-time datasets from Radar-AMeDAS, AMeDAS and JRA55. We found that the model initial conditions are particularly important during the spring snowmelt season and can improve the forecast predictive skills quite significantly compared with the climatology.