Si et al. (2026) Evolution trends and drivers of glacier and snowmelt induced floods in Upper Yarkant River Basin, Karakoram (1954–2020)

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2026
• Date: 2026-03-12
• Authors: Weian Si, Yonggang Ma, Tie Liu, Yijie Cao, X. L. Pan, Yunan Ling, Yue Huang
• DOI: 10.1016/j.ejrh.2026.103333

Research Groups

• State Key Laboratory of Ecological Safety and Sustainable Development in Arid Lands, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi, China
• College of Geography and Remote Sensing Sciences, Xinjiang University, Urumqi, China
• Xinjiang Key Laboratory of RS & GIS Application, Urumqi, China
• School of Geographic Information, Zhejiang University of Technology, Hangzhou, China
• University of Chinese Academy of Sciences, Beijing, China
• Tashkent Institute of Irrigation and Agricultural Mechanization Engineers, Tashkent, Uzbekistan

Short Summary

This study developed a multi-temporal framework to classify flood events in the Upper Yarkant River Basin (UYRB) from 1954 to 2020, focusing on the hydrological evolution and driving mechanisms of Glacier and Snowmelt Floods (GSMFs). It found that GSMFs are the predominant flood type, exhibiting increased peak discharge, reduced duration, earlier peak timing, and greater variability, primarily driven by extreme nocturnal warming and the 0 °C isotherm height with a 3-day lag.

Objective

• To develop a multi-temporal framework for the extraction and classification of multi-source flood events in alpine regions.
• To analyze the hydrological evolution of Glacier and Snowmelt Floods (GSMFs) in the Upper Yarkant River Basin (UYRB).
• To investigate the dominant factors, triggering mechanisms, and spatiotemporal variability of GSMFs in the UYRB.

Study Configuration

• Spatial Scale: Upper Yarkant River Basin (UYRB), Karakoram Mountains, Northwest China (approximately 46,000 km²), including the Tashkurgan River Basin (TRB) and Kuluklangan Upstream Watershed (KUW).
• Temporal Scale: 1954–2020 (70 years).

Methodology and Data

• Models used:
  • SFE_IFC software (for independent flood event extraction using Peak Over Threshold (POT) sampling)
  • K-Means clustering (for long-term flood classification)
  • Mann-Kendall (MK) test (for trend analysis)
  • Pettitt test (for abrupt change-point detection)
  • Generalized Pareto Distribution (for flood frequency analysis)
  • Random Forest (RF) regressor (for feature importance analysis)
  • Cross-Correlation Analysis (for temporal lag quantification)
• Data sources:
  • Hydrological Yearbook of China (River discharge: daily & hourly, 1954–2020, from KQ and KLK stations)
  • China Meteorological Administration (Air temperature & precipitation: daily, 1954–2020, from TSKG; Precipitation: daily, 1954–2020, from KQ & KLK; 0 °C isotherm height: daily, 1990–2020, from Kashgar & Hotian Stations)
  • Copernicus Climate Data Store (ERA5-Land temperature & precipitation: monthly, 1950–2020)
  • National Tibetan Plateau Data Center (Snow Depth (SD) dataset: daily, 1979–2020)
  • NASA Earthdata (MODIS snow cover product (MOD10A1): daily, 2000–2020)
  • Google Earth Engine (Sentinel-1 SAR imagery: 12-day, 2014–2020, for GLOF monitoring)

Main Results

• Glacier and Snowmelt Floods (GSMFs) are the primary flood type in the UYRB, constituting 93% of the total flood volume, with an average event duration of 17 days and a volume of approximately 1.5 × 10⁹ m³ per event.
• Over the 70-year study period, GSMFs have evolved to exhibit higher peak flows (KLK station: 26 m³/s per decade increase), shorter durations, earlier peak dates (KLK peak date: 1.6 days per decade earlier; KQ peak date: 6 days earlier post-1990s), and increased variability (KQ station's coefficient of variation (CV) increased from 0.23 to 0.28 post-1986).
• The frequency of GSMFs with return periods exceeding 10 years has increased since the 1990s, with the KLK station experiencing a higher occurrence rate of >10-year floods compared to the downstream KQ station.
• Sub-daily dynamics of GSMFs have intensified, marked by a delay in peak time, particularly at the Kaqun Station where a significant shift occurred in 1983. Since the 1980s, daily flood peaks have predominantly shifted to occur during the daytime, with daytime peaks in the early-to-mid flood stages and nighttime peaks in the later stages.
• Driving mechanisms of GSMF characteristics are highly scale-dependent:
  • Seasonal scale: Higher summer temperatures primarily induce an earlier onset of GSMFs rather than significantly increasing total flood volume. Warmer springs accelerate early snowmelt, reducing summer flood volume and duration.
  • Daily scale: GSMF discharge shows a significant 3-day lagged response to the 0 °C isotherm height (mean Pearson correlation coefficient (r) = 0.7 at Kashgar station), with a 100-meter rise in the isotherm corresponding to a runoff increase of 56 m³/s at the mountain outlet on the third day. Air temperature also exhibits a 3-day lag (mean r = 0.7 at TSKG station), where a 1 °C rise corresponds to a 77.4 m³/s discharge increase at the KQ station on the third day.
  • Diurnal scale: Elevated nighttime temperatures (TN90P) play a critical role in generating flood runoff by reducing refreezing and maintaining continuous melt within the snowpack. Extreme nocturnal warming is a dominant driver for peak discharge.
• The UYRB experienced a pronounced "warming-wetting" trend from 1957 to 2020, with annual precipitation increasing by 4.5 mm/decade (p < 0.05) and mean annual temperature rising by 0.32 °C/decade (p < 0.01). Extreme high temperatures (TX90P and TN90P) also showed highly significant upward trends.

Contributions

• Developed a novel multi-temporal scale framework for the extraction and classification of multi-source floods in alpine regions, providing a more accurate and event-specific approach than traditional methods.
• Provided comprehensive hydrological insights into the long-term evolution (1954–2020) and scale-dependent driving mechanisms of Glacier and Snowmelt Floods (GSMFs) in the Upper Yarkant River Basin, a crucial water source in an arid inland basin.
• Quantified the time-lagged response of GSMFs to meteorological drivers, such as the 3-day lag with the 0 °C isotherm height, which offers a critical predictive window for enhancing watershed-scale flood forecasting and early warning systems under a warming climate.
• Highlighted the significant role of nocturnal warming in generating meltwater runoff at the diurnal scale, contributing to a more nuanced understanding of cryospheric hydrological responses and their sensitivity to atmospheric thermal drivers.

Funding

• Outstanding Youth Fund Project of the Natural Science Foundation of Xinjiang Uygur Autonomous Region (2025D01E59)
• Xinjiang Talent Development Fund (XJRC-2025-KJ-PY-KJLJ-095)
• Key Research and Development Program of Xinjiang Uygur Autonomous Region (2022B03021-2)

Citation

@article{Si2026Evolution,
  author = {Si, Weian and Ma, Yonggang and Liu, Tie and Cao, Yijie and Pan, X. L. and Ling, Yunan and Huang, Yue},
  title = {Evolution trends and drivers of glacier and snowmelt induced floods in Upper Yarkant River Basin, Karakoram (1954–2020)},
  journal = {Journal of Hydrology Regional Studies},
  year = {2026},
  doi = {10.1016/j.ejrh.2026.103333},
  url = {https://doi.org/10.1016/j.ejrh.2026.103333}
}