Quantifying glacier accumulation in the western Tianshan Mountains considering snow redistribution

Snow redistribution in alpine catchments plays an important role in hydrological modeling. However, snow redistribution has rarely been incorporated into semi-distributed hydrological models. This omission has introduced considerable uncertainty and potential bias in glacier mass balance simulations. This study addressed this challenge by developing an empirical snow redistribution module (SRD) within the SWAT–Glacier model (SWAT–Glacier–SRD). The snow redistribution amount was a function of slope and snow density under a certain threshold. This model was applied to the Kumarik catchment, the largest tributary in the Tarim River Basin. To constrain the hydrological components, the model was calibrated using multiple observations, including runoff, glacier area, mass balance, and snow balance indices. Compared to SWAT–Glacier, SWAT–Glacier–SRD was able to more effectively simulate spring streamflow. Kling–Gupta Efficiency (KGE) was 0.82 in the calibration period and ranged from 0.57 to 0.91 in the validation period. The model also achieved improved performance in reproducing glacier- and snow-related indices. From 1975 to 2019, the mean rates of glacier accumulation and ablation were estimated at 0.573 m w.e. and 0.850 m w.e., respectively, resulting in a glacier mass balance of −0.277 m w.e. Glacier accumulation increased substantially from 0.460 m w.e. during 1975–1998 to 0.697 m w.e. during 1999–2019. The rate of glacier mass loss slowed, with glacier mass balance increasing from −0.359 m w.e. (1975–1998) to −0.187 m w.e. (1999–2019).