Improving glacio-hydrological model calibration and model performance in cold regions using satellite snow cover data

Hydrological modeling realism is a central research question in hydrological studies. However, it is still a common practice to calibrate hydrological models using streamflow as a single hydrological variable, which can lead to large parameter uncertainty in hydrological simulations. To address this issue, this study employed a multi-variable calibration framework to reduce parameter uncertainty in a glacierized catchment. The current study employed multi-variable calibration using three different calibration schemes to calibrate a glacio-hydrological model (namely the FLEX^G) in northern Sweden. The schemes included using only gauged streamflow data (scheme 1), using satellite snow cover area (SCA) derived from MODIS data (scheme 2), and using both gauged streamflow data and satellite SCA data as references for calibration (scheme 3) of the FLEX^G model. This study integrated the objective functions of satellite-derived SCA and gauged streamflow into one criterion for the FLEX^G model calibration using a weight-based approach. Our results showed that calibrating the FLEX^G model based on solely satellite SCA data (from MODIS) produced an accurate simulation of SCA but poor simulation of streamflow. In contrast, calibrating the FLEX^G model based on the measured streamflow data resulted in minimum error for streamflow simulation but high error for SCA simulation. The promising results were achieved for glacio-hydrological simulation with acceptable accuracy for simulation of both streamflow and SCA, when both streamflow and SCA data were used for calibration of FLEX^G. Therefore, multi-variable calibration in a glacierized basin could provide more realistic hydrological modeling in terms of multiple glacio-hydrological variables.