Kui et al. (2026) Detection of nine plateau lakes water level changes in Yunnan, China from ICESat-2 data

Identification

• Journal: Journal of Hydrology
• Year: 2026
• Date: 2026-03-29
• Authors: Mengyun Kui, Jinliang Wang, Jieying Lao, Jiya Pan, Jianpeng Zhang, Yuncheng Deng
• DOI: 10.1016/j.jhydrol.2026.135407

Research Groups

• Faculty of Geography, Yunnan Normal University, Kunming, Yunnan, China
• Key Laboratory of Remote Sensing of Resources and Environment in Yunnan Universities, Kunming, Yunnan, China
• Yunnan Geospatial Information Engineering Technology Research Center, Kunming, Yunnan, China
• School of Earth Sciences, Yunnan University, Kunming, Yunnan, China
• School of Economics and Management, Yunnan Normal University, Kunming, Yunnan, China

Short Summary

This study developed an outlier-detection framework for ICESat-2 ATL13 data to monitor water level changes in nine Yunnan plateau lakes from 2018 to 2024, revealing consistent declining trends for most lakes driven by seasonal natural factors modulated by anthropogenic influences.

Objective

• To retrieve and analyze the temporal variations of water levels for nine major plateau lakes in Yunnan Province using ICESat-2 ATL13 data with an enhanced outlier-detection framework, and to explore the primary drivers of these lake-level changes.

Study Configuration

• Spatial Scale: Nine major plateau lakes in Yunnan Province, China (Dianchi, Yangzonghai, Qilu Lake, Fuxian Lake, Yilong Lake, Xingyun Lake, Chenghai, Erhai, and Lugu Lake).
• Temporal Scale: October 2018 to December 2024.

Methodology and Data

• Models used: An outlier-detection framework integrating geospatial clustering and a dynamic-threshold robust Z-score approach; SHAP (SHapley Additive exPlanations) method for driver analysis.
• Data sources: ICESat-2 ATL13 global inland water dataset; observed annual water levels; Big Earth Data monthly water levels; DAHITI (Database for Hydrological Time Series of Inland Waters) trends.

Main Results

• ICESat-2 derived water levels showed high consistency with observed annual levels, Big Earth Data monthly water levels, and DAHITI trends, with R² values exceeding 0.8 and RMSE ranging from 0.04 to 0.13 m.
• Most lakes experienced declining water levels from 2018 to 2024, with Lake Fuxian, Lake Qilu, Lake Dianchi, and Yangzonghai showing the most pronounced decreases at −0.41 m/a, −0.23 m/a, −0.06 m/a, and −0.12 m/a, respectively.
• Lake-level drivers exhibited a seasonal pattern: "natural factors dominating and anthropogenic factors modulating." Precipitation–runoff inputs controlled water-level increases during the wet season, while temperature-induced evapotranspiration losses dominated declines during the dry season. Cropland area exerted opposite seasonal effects via reservoir irrigation return flows in the dry season and agricultural water consumption/discharge in the wet season.

Contributions

• Proposed and validated an enhanced outlier-detection framework for ICESat-2 ATL13 data, significantly improving the accuracy of water level retrieval for plateau lakes compared to previous studies.
• Provided a robust foundation for long-term monitoring of plateau lake water-level dynamics, offering reliable data support and scientific guidance for watershed water-resource regulation and lake ecosystem management.
• Identified the seasonal interplay of natural and anthropogenic factors driving water level changes in Yunnan's plateau lakes.

Funding

• Not specified in the provided text.

Citation

@article{Kui2026Detection,
  author = {Kui, Mengyun and Wang, Jinliang and Lao, Jieying and Pan, Jiya and Zhang, Jianpeng and Deng, Yuncheng},
  title = {Detection of nine plateau lakes water level changes in Yunnan, China from ICESat-2 data},
  journal = {Journal of Hydrology},
  year = {2026},
  doi = {10.1016/j.jhydrol.2026.135407},
  url = {https://doi.org/10.1016/j.jhydrol.2026.135407}
}