Combining remote sensing, in situ observations and a snow model to estimate climate sensitivity of snow water equivalent and snowmelt runoff in a Himalayan catchment

Abstract

Snow is an important storage of water in the Himalayas. Previous studies on snowmelt have predominantly relied on remotely sensed snow cover, however this gives no information on the snow water equivalent (SWE). Therefore, in this study remotely sensed snow cover was combined with in situ observations and a snow model to estimate the climate sensitivity of SWE and snowmelt runoff in the Langtang catchment. Optimal model parameter values were obtained through data assimilation of remotely sensed snow cover (MOD10A2 snow cover product) and snow depth measurements using an Ensemble Kalman filter. The classification accuracy of the modelled snow cover is 78.2 % against Landsat 8 snow maps and 85.9 % against MOD10A2 snow cover. Landsat 8 snow maps and the MOD10A2 snow cover product were validated with in situ snow cover observations derived from surface temperatures and snow depth measurements for the first time in the Himalayas. The classification accuracies are 85.7% and 83.1% respectively. Modelled and observed snow depth do not match perfectly due to processes not included in the model such as sublimation and wind redistribution. The approach for modelling snow depth allows to validate the amount of modelled snow rather than only snow cover which shows great potential for future studies in the Himalayas. Climate sensitivity tests show a strong decrease of SWE in the valley for increasing temperature. Though, at high elevation decrease in SWE can be compensated by an increase in precipitation. Snowmelt runoff increases in winter and early melt season (December to May) as result of increasing temperature. However, during late melt season (June to September) most snow has already melted and consequently there is less snowmelt resulting in less snowmelt runoff during the late melt season.