Uncertainties in river discharge simulations of the upper Indus basin in the Western Himalayas

Abstract

This study focuses on understanding and quantifying uncertainties in simulating river discharge in the Upper Indus Basin (UIB) of the Western Himalayas using a macro-scale semi-distributed hydrology model forced with multiple observed precipitation datasets and reanalysis products of near-surface wind-speed, maximum and minimum temperature during 2010–2012. We performed a suite of numerical simulations using a high-resolution setup of the variable infiltration capacity (VIC) hydrology model for the UIB. This model takes into account the balance of both water and surface energy budgets within each grid cell and incorporates sub-grid variability of topography to represent the effects of orographic precipitation and temperature lapse rate essential for hydrological modelling in the complex Himalayan terrain. While river discharges over non-mountainous basins are known to be generally sensitive to precipitation variations, it is noted that both precipitation and snowmelt processes critically influence seasonal river discharge in the UIB during the northern summer through surface temperature and wind-speed variations. Our study found that a marginal difference in temperature forcing can create large difference in snowmelt over UIB during summer season, which in turn increases the uncertainty in the summer monsoon river discharge. This analysis highlights the equally important need for the incorporation of realistic temperature data as that of precipitation product for the better simulation of land surface processes during various seasons, especially during summer, over snow covered UIB. Further analysis of daily simulations of the VIC model during 2010–2012 indicates that low and medium intensity river discharges tend to be associated with relatively lower spread among the ensemble members, as compared to the high intensity discharges which exhibit large ensemble spread. In particular, we noted a large increase in the spread of high flow simulations over the UIB during the flood episodes in the summer of 2010, arising from uncertainties in the precipitation forcing across multiple datasets. Our results emphasize the need for improved representation of precipitation and hydrological processes over the Himalayan region in weather and climate models for better management of water resources and flood forecasting in the UIB region under a changing climate.