Impacts of glacier melt on stream temperature in a cold region watershed: coupling the glacier melt model with hydrological model

Abstract:

The impacts of glacier melt processes on stream temperature, such as the cooling effect, have not been explicitly considered in the process-based model, which limits their capability in simulating stream temperature in the glacierized cold region. To overcome this limitation, we coupled the temperature index based glacier melt model with the SWAT equilibrium temperature model. The outputs of melt model provide inputs to the SWAT-ETM model and glacier melt runoff was used as another runoff component in addition to surface runoff, lateral flow, groundwater, and snowmelt when calculating the temperature of the local subbasin based on the mixing model. The coupled model was tested and applied in the upper Athabasca river basin in western Canada to simulate daily streamflow and stream temperature. It showed a good performance for stream temperature simulation with percent bias (PBIAS) less than ±5% and Nash-Sutcliffe efficiency (NSE) and coefficient of determination (R²) values greater than 0.85. The results showed coupled model was able to reflect the impacts from glacier melt by simulating higher streamflow and lower stream temperature during the melt season. The cooling effect of glacier melt runoff on stream temperature with decreasing trend from headwater to downstream was well represented. Moreover, the role of glacier melt runoff affecting stream temperature under climate change was analyzed by the scenario analysis. It showed that an increase in glacier melt runoff in near future would alleviate the average and peak stream temperature increase during melt season, while a decrease in glacier melt runoff in distant future would aggravate the stream temperature increase. Therefore, the impact of glacier melt runoff should be explicitly considered to accurately model stream temperature and predict future climate change impact on thermal regime in the glacierized watersheds. The model coupling framework developed in this study could be a useful tool for simulating stream temperature in the glacierized watersheds.