Improving streamflow and flood simulations in three headwater catchments of the Tarim River based on a coupled glacier-hydrological model

Abstract

Understanding hydrological changes, especially those of snow and/or glacier melt-induced floods, is vital for early flood warnings in alpine regions. However, there is limited information available regarding warming-induced changes in hydrological and flooding processes in high-mountain regions owing to data scarcity and inadequate considerations within hydrological models. This study aims to improve streamflow and flood simulations in three high-mountain headwater catchments of the Tarim River in northwest China. An improved scheme for calculating glacier melt (GM2) was coupled in the Variable Infiltration Capacity model (named VIC-GM2). Driven by bias-corrected meteorological data, VIC-GM2 exhibited good performance, as evidenced by NS, KGE and PBIAS values between the simulated and observed daily streamflow ranging from 0.69 to 0.79, 0.70 to 0.81 and −2.68% to 19.6%, respectively. In addition, annual contributions of glacier melt and snow melt to total streamflow were greater than 80% and less than 16% among the three catchments, respectively. The flood characteristics were inconsistent among the three catchments, except that flooding mainly occurred in summer. The MAF (mean annual flood) in the Yarkand River was largest, followed by that in the Yurungkax River and the Qaraqash River. The FPOT (frequency of peak over threshold) and FQ99 (frequency of Q99) (times/year) fluctuated within 1 – 1.75 and 0.25 – 1 in the three catchments, respectively. This study provides references for understanding cryosphere hydrological processes and quantifying the contribution of different components (i.e., glacier melt, snow melt and rainfall) to streamflow under a changing climate in the upper Tarim River.