Zhu et al. (2025) Exploring interactive effects of water stress and ecological restoration on vegetation eco-regimes using interpretable machine learning based on kernel NDVI

Identification

• Journal: Journal of Hydrology
• Year: 2025
• Date: 2025-12-14
• Authors: Yongwei Zhu, S. S. Jiang, Liliang Ren, Shuping Du, Hao Cui, Miao He, Chong‐Yu Xu
• DOI: 10.1016/j.jhydrol.2025.134789

Research Groups

• The National Key Laboratory of Water Disaster Prevention, Hohai University, Nanjing, China
• Cooperative Innovation Center for Water Safety and Hydro-Science, Hohai University, Nanjing, China
• College of Hydrology and Water Resources, Hohai University, Nanjing, China
• Key Laboratory of Soil and Water Processes in Watershed, Hohai University, Nanjing, China
• Department of Geosciences, University of Oslo, Oslo, Norway

Short Summary

This study developed a novel methodology using 18 vegetation indicators, an ecological restoration index, and interpretable machine learning to assess vegetation eco-regime changes across China. It found that precipitation, surface solar radiation, and ecological restoration are the dominant factors influencing these dynamics, with significant shifts in eco-regimes observed since 2002.

Objective

• To develop a novel methodology for assessing vegetation eco-regime changes by incorporating 18 comprehensive indicators and an ecological restoration index, and to systematically evaluate the interactive impacts of water stress and ecological restoration on vegetation eco-regimes across China using interpretable machine learning.

Study Configuration

• Spatial Scale: China
• Temporal Scale: 1982–2022 (Baseline Period: 1982–2001; Change Period: 2002–2022)

Methodology and Data

• Models used: Interpretable Machine Learning (IML) approaches, Range of Variability Approach (implied for eco-regime analysis).
• Data sources: Kernel Normalized Difference Vegetation Index (NDVI) (likely satellite-derived), 18 comprehensive vegetation indicators, an ecological restoration index, precipitation data, surface solar radiation data, and vapor pressure deficit data (sources not explicitly detailed but typically from reanalysis or observations).

Main Results

• The proportion of mutation years in China’s kernel NDVI peaked in 2002, leading to the definition of a Baseline Period (1982–2001) and a Change Period (2002–2022).
• During the Change Period, the degree of change in vegetation eco-regimes across China ranged from 21.5 % to 89.4 %, with a median value of 64.9 %.
• The proportions of low, medium, and high change categories were 40.7 %, 59.1 %, and 0.2 %, respectively.
• Interpretable Machine Learning identified precipitation, surface solar radiation, and ecological restoration as the three dominant factors governing vegetation eco-regime dynamics, contributing a combined mean of 74.8 % during both periods.
• The transition threshold of vegetation eco-regimes increased by 40 mm for precipitation and by 0.3 hPa for vapor pressure deficit.

Contributions

• Proposes a novel and comprehensive methodology for assessing vegetation eco-regime changes using 18 indicators and an ecological restoration index.
• Employs interpretable machine learning to systematically quantify and explain the interactive effects of water stress and ecological restoration on vegetation eco-regimes.
• Identifies and quantifies the dominant factors (precipitation, surface solar radiation, ecological restoration) influencing vegetation dynamics across China.
• Provides quantitative insights into the degree of change in vegetation eco-regimes and identifies critical transition thresholds for key climatic variables.
• Offers a transferable approach for global assessments of vegetation-climate-restoration interactions, informing targeted ecological strategies.

Funding

• Not explicitly stated in the provided text.

Citation

@article{Zhu2025Exploring,
  author = {Zhu, Yongwei and Jiang, S. S. and Ren, Liliang and Du, Shuping and Cui, Hao and He, Miao and Xu, Chong‐Yu},
  title = {Exploring interactive effects of water stress and ecological restoration on vegetation eco-regimes using interpretable machine learning based on kernel NDVI},
  journal = {Journal of Hydrology},
  year = {2025},
  doi = {10.1016/j.jhydrol.2025.134789},
  url = {https://doi.org/10.1016/j.jhydrol.2025.134789}
}