Li et al. (2026) Natural versus anthropogenic forces affecting spatial variations of Ebinur Lake in Northwest China over the past two decades 2001–2020

Identification

• Journal: Journal of Hydrology Regional Studies
• Year: 2026
• Date: 2026-01-06
• Authors: Zhihui Li, Fei Zhang, Zhuohan Jiang, Ngai Weng Chan, P. P. Anil Kumar, Gowhar Meraj, Wei Wang, Lifei Wei, Xu Ma
• DOI: 10.1016/j.ejrh.2025.103078

Research Groups

• College of Meteorology and Oceanography, National University of Defense Technology, Changsha, China
• College of Geography and Remote Sensing Sciences, Xinjiang University, Urumqi, China
• College of Geography and Environmental Sciences, Zhejiang Normal University, Jinhua, China
• GeoInformatic Unit, Geography Section, School of Humanities, Universiti Sains Malaysia, Penang, Malaysia
• Institute for Global Environmental Strategies, Hayama, Japan
• Department of Biology, Chemistry and Environmental Sciences, College of Arts and Sciences, American University of Sharjah, Sharjah, United Arab Emirates
• School of Geographical Sciences, China West Normal University, Nanchong, China
• Faculty of Resources and Environmental Science, Hubei University, Wuhan, China

Short Summary

This study quantifies the influence of natural and anthropogenic factors on Ebinur Lake's water area fluctuations in Northwest China from 2001 to 2020, revealing a significant decline in lake area primarily driven by inflow river run-off and population density.

Objective

• To accurately extract and validate the water area of Ebinur Lake from 2001 to 2020 using the Normalised Difference Water Index (NDWI) on Landsat imagery within the Google Earth Engine (GEE) platform.
• To monitor the monthly and annual changes in Ebinur Lake's area over the past two decades.
• To construct and apply an eXtreme Gradient Boosting (XGBoost) model coupled with SHapley Additive exPlanations (SHAP) analysis to quantify the contributions of various natural and anthropogenic factors to these changes.

Study Configuration

• Spatial Scale: Ebinur Lake Watershed, Northwest China (44°54′–45°08′N, 82°35′–83°10′E).
• Temporal Scale: 2001–2020.

Methodology and Data

• Models used: Google Earth Engine (GEE), Normalised Difference Water Index (NDWI), Otsu thresholding method, eXtreme Gradient Boosting (XGBoost) model, SHapley Additive exPlanations (SHAP) analysis, Monte Carlo framework, Time-Series Cross-Validation (TSCV).
• Data sources:
  • Remote Sensing: Landsat Collection 2 Level 2_T1 imagery (2001–2020), JRC global surface water product (for validation).
  • Hydrological: Daily average temperatures and precipitation from three meteorological stations (Ala Pass, Bole, Jing River stations, 2001–2020), runoff data for Bortala and Jing Rivers (2001–2020) from Xinjiang Uygur Autonomous Region Water Resources Department.
  • Socio-economic: Regional statistical data from the Xinjiang Statistical Yearbook (2001–2020), including primary industry added value, crop sowing area, livestock numbers, and registered population.

Main Results

• Ebinur Lake exhibited significant monthly variations, with its largest average extent in May (637.35 km²) and smallest in October (480.74 km²).
• Interannual fluctuations showed a peak area of 946.32 km² in 2003, decreasing to a minimum of 409.33 km² by 2015. The observed long-term decline trend of -9.97 km²/yr was not statistically significant (p > 0.05) due to high uncertainty from drought-induced accuracy degradation.
• SHAP-based analysis indicated that inflow river run-off was the most significant natural factor influencing lake area changes, showing a strong correlation with combined inflow from the Jing and Bortala Rivers (R² = 0.6718 for 2001–2017).
• Among human drivers, population growth (from 6.16 × 10⁵ to 7.51 × 10⁵ persons) and the expansion of sown cropland (from 1.22 × 10⁵ to 3.28 × 10⁵ hectares) were identified as dominant factors increasing irrigation demand and contributing to lake contraction.
• The XGBoost model demonstrated robust performance with a Validation RMSE of 43.43 and an average R² of 0.51, indicating good generalization.

Contributions

• Provides a comprehensive, long-term (2001-2020) quantitative analysis of Ebinur Lake's water area dynamics and their drivers, addressing a gap in previous short-term or qualitative studies.
• Quantifies the contributions of both natural (runoff, precipitation, temperature) and anthropogenic (population, primary industry, crop sowing, livestock) factors using an XGBoost model coupled with SHAP analysis.
• Offers a robust methodological framework (GEE, NDWI, XGBoost-SHAP, Monte Carlo uncertainty analysis) applicable to similar vulnerable lake systems worldwide.
• Highlights the importance of hydroclimate-stratified accuracy reporting and integrating Synthetic Aperture Radar (SAR) data for improved monitoring in arid regions.

Funding

• National Natural Science Foundation of China (42261006)
• State Key Laboratory of Lake Science and Environment (2022SKL007)
• Universiti Sains Malaysia Internationalisation Incentive Scheme (R502-KR-ARP004–00AUPRM003-K134)

Citation

@article{Li2026Natural,
  author = {Li, Zhihui and Zhang, Fei and Jiang, Zhuohan and Chan, Ngai Weng and Kumar, P. P. Anil and Meraj, Gowhar and Wang, Wei and Wei, Lifei and Ma, Xu},
  title = {Natural versus anthropogenic forces affecting spatial variations of Ebinur Lake in Northwest China over the past two decades 2001–2020},
  journal = {Journal of Hydrology Regional Studies},
  year = {2026},
  doi = {10.1016/j.ejrh.2025.103078},
  url = {https://doi.org/10.1016/j.ejrh.2025.103078}
}