Yang et al. (2025) Spatial heterogeneity and environmental drivers of drought vulnerability in the Yangtze River Basin

Identification

• Journal: Ecological Indicators
• Year: 2025
• Date: 2025-09-25
• Authors: Yuting Yang, Yunfei Feng, Xie He, Meng Li
• DOI: 10.1016/j.ecolind.2025.114246

Research Groups

• School of Geographic Sciences, Nantong University, Nantong, China
• Tangshan Key Laboratory of Simulation of Urban Ecosystem, Department of Resource Management, Tangshan Normal University, Tangshan, China

Short Summary

This study assessed the spatial heterogeneity and environmental drivers of drought vulnerability in the Yangtze River Basin (YRB) from 2001 to 2023, integrating exposure, sensitivity, and resilience. It found that resilience is the dominant determinant of vulnerability, with the central-northern YRB exhibiting the highest vulnerability due to concurrent high exposure and sensitivity with low resilience.

Objective

• To characterize the spatiotemporal patterns of drought vulnerability across the Yangtze River Basin from 2001 to 2023.
• To identify the dominant biophysical and climatic factors driving variability in ecological vulnerability.

Study Configuration

• Spatial Scale: Yangtze River Basin, China (approximately 1.8 million square kilometers, spanning 28°N–35°N, 90°E–122°E).
• Temporal Scale: 2001 to 2023.

Methodology and Data

• Models used:
  • Drought Vulnerability Index (DVI) based on the IPCC framework, integrating Exposure Index (EI), Sensitivity Index (SI), and Resilience Index (RI).
  • Entropy weight method for combining drought severity and duration into EI.
  • Autoregressive (AR) modeling (NDVIt = α × PDSIt + β × Tt + δ × NDVIt−1 + ε) to derive SI (α) and RI (1-δ).
  • eXtreme Gradient Boosting (XGBoost) with SHAP (Shapley Additive Explanations) analysis to identify environmental drivers.
  • Piecewise linear regression for validating threshold patterns.
• Data sources:
  • Palmer Drought Severity Index (PDSI) from TerraClimate dataset (4 km spatial resolution, monthly).
  • Normalized Difference Vegetation Index (NDVI) from MODIS MOD13A3 product (1 km spatial resolution, monthly).
  • Land cover data from MODIS MCD12Q1 product (500 m spatial resolution, IGBP classification).
  • Climate regionalization dataset for China (Resource and Environmental Science Data Center).
  • Predictor variables: altitude, soil organic carbon (SOC), soil texture (clay, sand, silt fractions), maximum rooting depth (MRD), aboveground biomass (AGB), belowground biomass (BGB), species richness (SR), tree cover (TC), aridity index (AI), precipitation (PREC), temperature (TEMP), solar radiation (RAD), human footprint index (HFP).

Main Results

• The central-northern YRB exhibits the highest drought vulnerability due to concurrent high exposure and sensitivity with low resilience.
• Western highlands show comparatively low vulnerability despite high exposure, supported by stronger ecosystem resilience.
• Resilience is the dominant determinant of vulnerability, with over 65 % of vegetated pixels showing a strong and significant negative correlation (R < –0.8, p < 0.05) between resilience and vulnerability.
• Altitude, precipitation, and species richness are the most influential environmental drivers of drought vulnerability.
• Environmental drivers often exhibit non-linear or threshold effects: altitude displays a U-shaped relationship (both lowlands and highlands are more vulnerable), and precipitation and species richness show thresholds beyond which vulnerability increases sharply.
• Shrub ecosystems, despite high drought exposure, exhibit relatively low overall vulnerability due to strong resilience, while forest ecosystems show significantly higher vulnerability due to lower resilience.
• From 2001 to 2023, the YRB showed a slight, non-significant increasing trend in PDSI (0.018 per year) and a non-significant upward trend in drought duration across the entire basin.

Contributions

• Developed a comprehensive composite drought vulnerability index for the Yangtze River Basin, integrating all three IPCC-defined components (exposure, sensitivity, and resilience) using multi-source remote sensing and reanalysis data.
• Quantified ecosystem response intensity and recovery under drought conditions using autoregressive models.
• Systematically identified and quantified the relative importance and non-linear contributions of biophysical and climatic factors to drought vulnerability using XGBoost and SHAP analysis.
• Highlighted the central role of resilience as the dominant modulator of drought vulnerability, providing a robust, indicator-based framework for ecological risk assessment and guiding resilience-oriented adaptation strategies.

Funding

• National Natural Science Foundation of China (42207129, 42301111)
• Strategic Pilot Science and Technology Project of the Chinese Academy of Sciences (XDA26050501)
• Central Guiding Local Fund Project of Tibet Autonomous Region (XZ202301YD0012C)

Citation

@article{Yang2025Spatial,
  author = {Yang, Yuting and Feng, Yunfei and He, Xie and Li, Meng},
  title = {Spatial heterogeneity and environmental drivers of drought vulnerability in the Yangtze River Basin},
  journal = {Ecological Indicators},
  year = {2025},
  doi = {10.1016/j.ecolind.2025.114246},
  url = {https://doi.org/10.1016/j.ecolind.2025.114246}
}