Zhe et al. (2025) Area dynamics of alpine lakes in the Yamzhog Yumco Basin: Optimized water indices reveal spatiotemporal patterns and key drivers

Identification

• Journal: Ecological Indicators
• Year: 2025
• Date: 2025-11-16
• Authors: Meng Zhe, Shaobo Sun
• DOI: 10.1016/j.ecolind.2025.114415

Research Groups

• Faculty of Education, Tianjin Normal University, Tianjin 300387, China
• Institute of Surface-Earth System Science, School of Earth System Science, Tianjin University, Tianjin 300072, China

Short Summary

This study evaluated and optimized water indices for accurate lake extraction in the Yamzhog Yumco Basin, reconstructing five-decade spatiotemporal lake area dynamics and identifying key climatic and anthropogenic drivers. It revealed that intense anthropogenic activities, particularly pumped-storage operations, can override climate-driven hydrological regimes in connected lakes, while isolated lakes follow regional climate trends.

Objective

• To evaluate and optimize seven prevalent water indices for accurate lake body extraction in the Yamzhog Yumco Basin (Tibetan Plateau) through a joint validation approach combining pixel-level classification accuracy and areal consistency.
• To reconstruct and analyze the five-decade (1972–2023) spatiotemporal dynamics of five alpine lakes in the basin.
• To identify and quantify the key climatic and anthropogenic drivers influencing lake area variations across distinct hydrological phases, particularly focusing on how anthropogenic activities interact with climate change.

Study Configuration

• Spatial Scale: Yamzhog Yumco Basin (YYB) on the Tibetan Plateau, covering 9,070 km². The study focused on five lakes: Yamzhog Yum Co (YZYC), Puma Yum Co (PMYC), Chen Co (CC), Kongmu Co (KMC), and Bajiu Co (BJC).
• Temporal Scale:
  • Landsat satellite imagery: 1972–2023 (96 scenes).
  • Meteorological data (Langkazi Station): 1972–2023.
  • Hydrological data (Baidi Hydrological Station for YZYC): Water levels (1975–2015), monthly evaporation (1975–2019).
  • Lake area time series: YZYC (1972–2023, with reconstructed data for missing years), other four lakes (1988–2023).

Methodology and Data

• Models used:
  • Water Indices: Normalized Difference Water Index (NDWI), Modified Normalized Difference Water Index (MNDWI), Enhanced Water Index (EWI), Automated Water Extraction Indices (AWEIsh, AWEInsh), Multi-Band Water Index (MBWI), Modified Automatic Water Extraction Index (MAWEI).
  • Statistical Models: Linear regression, Mann–Kendall (M–K) tests, moving t-tests, cumulative anomaly analysis for abrupt change detection, and Multiple Linear Regression (MLR) with cumulative climate anomalies for driver analysis.
• Data sources:
  • Satellite Imagery: Landsat Level-1 products (MSS from Landsats 1–2, TM from 4 to 5, ETM+ from 7, OLI from 8) obtained from USGS Earth Explorer.
  • In-situ Observations: Monthly precipitation and temperature data from Langkazi Meteorological Station. Water levels and monthly evaporation data for YZYC from Baidi Hydrological Station.
  • Reference Data: Manual Visual Interpretation (MVI) reference areas from 25 systematically selected Landsat scenes and high-resolution Google Earth imagery for validation.

Main Results

• Optimal Water Indices: MBWI performed optimally for Puma Yum Co (PMYC), while MAWEI was best for Yamzhog Yum Co (YZYC), Kongmu Co (KMC), Chen Co (CC), and Bajiu Co (BJC). Optimal indices achieved overall accuracies > 97%, Kappa coefficients > 0.93, and R² values ranging from 0.665 to 0.991 against MVI references.
• Lake Area Dynamics:
  • YZYC: Experienced significant long-term shrinkage at a rate of −15.56 km²/10a (R² = 0.709, p < 0.01) from 1972 to 2023, with abrupt change points in 1982, 1997, and 2005. Spatial changes included shoreline retreat, separation from Kongmu Co, and interior island expansion.
  • PMYC: Showed significant expansion at a rate of +2.24 km²/10a (R² = 0.634, p < 0.01) from 1988 to 2023, primarily northwestward.
  • CC and KMC: Both exhibited significant shrinkage at rates of −0.85 km²/10a (R² = 0.479, p < 0.01) and −0.58 km²/10a (R² = 0.665, p < 0.01), respectively, from 1988 to 2023.
  • BJC: Displayed the most pronounced fluctuations, undergoing expansion followed by marked shrinkage and re-fragmentation into discrete water bodies by 2022.
  • Basin-wide: YZYC dominated the basin's dynamics, accounting for 75.14% of the total variance in lake area.
• Drivers of Lake Area Dynamics:
  • YZYC: Exhibited a three-phase evolution:
    1. Climate-dominated (1975–2004): Climatic factors explained > 86% of variability. Cumulative evaporation anomalies dominated during 1975–1996 (β = −0.755, p < 0.01), while precipitation and temperature showed comparable positive contributions during 1997–2004.
    2. Anthropogenic-dominated (2005–2015): Rapid shrinkage occurred despite climate conditions favoring expansion. MLR showed an anomalous, highly significant negative temperature influence (β = −1.005, p < 0.01), indicating the overriding impact of pumped-storage operations.
    3. Multi-factor adjustment (2016–2023): Precipitation re-emerged as the primary contributor (β = 0.519), but the model was not statistically significant (p = 0.259), suggesting a complex interplay of attenuated anthropogenic impacts and resurgent climatic drivers.
  • Other Lakes: PMYC (isolated) expanded primarily due to precipitation (β = 0.509, p < 0.05), consistent with regional warming-wetting. Connected lakes (CC, BJC, KMC) largely synchronized with YZYC, reflecting the propagation of anthropogenic influence, with BJC's fluctuations amplified by shallow morphology, seepage, and pastoral water use. KMC showed partial decoupling due to glacial recharge.

Contributions

• Developed and validated a robust joint evaluation framework for water indices, combining pixel-level accuracy and areal consistency, which is crucial for accurate lake extraction in complex alpine environments.
• Provided the first comprehensive five-decade (1972–2023) spatiotemporal analysis of lake area dynamics in the Yamzhog Yumco Basin, including reconstruction for data-sparse periods, offering a long-term perspective on regional hydrological changes.
• Demonstrated a paradigm shift in lake drivers on the Tibetan Plateau, showing that intense anthropogenic activities (pumped-storage operations) can override and disrupt climate-driven hydrological regimes, a critical finding for water resource management in sensitive regions.
• Identified a three-phase evolutionary sequence of driving mechanisms (climate-dominated, anthropogenic-dominated, multi-factor adjustment) for YZYC, providing a process-informed understanding of complex human-environment interactions.
• Highlighted the synergistic interactions between specific natural factors (e.g., shallow morphology, hydrological connectivity) and anthropogenic disturbances in amplifying lake area fluctuations.

Funding

• National Natural Science Foundation of China (42171462)

Citation

@article{Zhe2025Area,
  author = {Zhe, Meng and Sun, Shaobo},
  title = {Area dynamics of alpine lakes in the Yamzhog Yumco Basin: Optimized water indices reveal spatiotemporal patterns and key drivers},
  journal = {Ecological Indicators},
  year = {2025},
  doi = {10.1016/j.ecolind.2025.114415},
  url = {https://doi.org/10.1016/j.ecolind.2025.114415}
}