Research News
Tarim Runoff to Remain Stable in 21st Century Even After Glacier Melt Peaks: Study
Editor: CAS_Editor | Jul 13, 2026
Print

A new study has revealed the hydrological response of river runoff to glacier peak water in the Tarim River Basin—China's largest inland river system—in response to glacier peak water, forecasting that the basin will sustain robust and stable runoff levels for the remainder of this century.

The study, led by Prof. CHEN Yaning from the Xinjiang Institute of Ecology and Geography (XIIEG) of the Chinese Academy of Sciences (CAS), was recently published in the Journal of Hydrology.

Glaciers are critical solid water reservoirs in arid regions and serve as a key source of river runoff in mountainous areas. Changes in glacier meltwater have significant implications for regional water security, ecological protection, and socioeconomic development.

As global temperatures rise, accelerated glacier melting is reshaping the timing and volume of meltwater entering rivers. Yet a critical question persists: will runoff drop sharply as soon as meltwater output hits its peak?

To answer this question, researchers used a Long Short-Term Memory (LSTM) model and integrated climate projections from Coupled Model Intercomparison Project Phase 6 (CMIP6) scenarios with glacier runoff datasets to assess future changes in seven headwaters catchments in the Tarim River Basin.

Model validation demonstrated that the LSTM model performed well in the Tarim River Basin. For monthly runoff simulation in the seven typical sub-basins, the median Nash-Sutcliffe Efficiency (NSE) coefficient, Kling-Gupta Efficiency (KGE), and Percent Bias (PBIAS) values were 0.77, 0.82, and −2.08%, respectively. The model also showed good performance for daily runoff simulation, with NSE, KGE, and PBIAS values of 0.53, 0.69, and −2.09%, respectively.

Under the SSP245, SSP370, and SSP585 scenarios, runoff in the basin's major rivers is expected to keep rising through the century, whereas the low-emission SSP126 scenario shows little overall change. The projected increases, however, vary markedly across different headwater systems.

By the end of the 21st century, runoff in the Aksu River on the southern slope of the Tianshan Mountains is projected to increase by 10% under SSP245 and 21% under SSP585, while runoff in the Kaidu River is projected to increase by 11% and 18%, respectively. Even sharper increases are forecast for rivers draining the northern Kunlun slopes: the Yarkand and Hotan rivers show the largest jumps relative to the 2000–2014 baseline under both scenarios.

Critically, the researchers found that reaching glacier peak water does not inevitably send river runoff into an immediate nosedive. To dissect this lag effect, they defined three 30-year phases around the peak year: a pre-peak window (45–16 years prior), a peak-transition window (15 years prior to 15 years after), and a post-peak window (16–45 years after).

Results from seven typical sub-basins in the Tarim River Basin showed that annual runoff remained relatively high during both the peak-transition and post-peak stages compared with the pre-peak stage. This points to a sustained buffer effect: even after the meltwater maximum is passed, river flows can hold steady or remain elevated for decades.

The findings, the researchers note, offer a scientific backbone for long-term water management in southern Xinjiang—balancing agricultural, ecological, and developmental needs. Beyond its regional impact, the study delivers a replicable modeling framework for glacier-fed alpine basins worldwide, supporting the sustainable stewardship of the "Central Asian Water Tower" and guiding climate-resilient water strategies in arid zones globally.