Water budget on the Dudh Koshi River (Nepal): Uncertainties on precipitation

- A 1000 mm/y deficit to balance the water budget in the Dudh Koshi River basin.
- Use of a hydrological snow model calibrated with remote sensing and discharge.
- Precipitation underestimation compensated by an overestimation of icemelt simulated.
- Observed precipitation underestimated by around 80%.

SummaryAlthough vital for millions of inhabitants, Himalayan water resources remain currently poorly known, mainly because of uncertainties on hydro-meteorological measurements. In this study, the authors propose a new assessment of the water budget components of the Dudh Koshi River basin (3720 km2, Eastern Nepal), taking into account the associated uncertainties. The water budget is studied through a cross analysis of field observations with the result of a daily hydrological conceptual distributed snow model. Both observed datasets of spatialized precipitations, interpolated with a co-kriging method, and of discharge, provided by the hydrological agency of Nepal, are completed by reanalysis data (NCEP/NCAR) for air temperature and potential evapotranspiration, as well as satellite snow products (MOD10A2) giving the dynamics of the snow cover area. According to the observation, the water budget on the basin is significantly unbalanced; it is attributed to a large underestimation of precipitation, typical of high mountain areas. By contrast, the water budget simulated by the modeling approach is well balanced; it is due to an unrealistic overestimation of the glacier melt volume. A reversing method led to assess the precipitation underestimation at around 80% of the annual amount. After the correction of the daily precipitation by this ratio, the simulated fluxes of rainfall, icemelt, and snowmelt gave 63%, 29%, and 8% of the annual discharge, respectively. This basin-wide precipitation correction is likely to change in respect to topographic or geographic parameters, or in respect to seasons, but due to an insufficient knowledge of the precipitation spatial variability, this could not be investigated here, although this may significantly change the respective proportions for rain, ice or snow melt.