Xue et al. (2025) Identifying Forest Drought Sensitivity Drivers in China Under Lagged and Accumulative Effects via XGBoost-SHAP

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Remote Sensing
• Year: 2025
• Date: 2025-08-20
• Authors: Ziqiu Xue, Simeng Diao, Fei Yang, Fei Long, Wenjuan Wang, Yan Liu
• DOI: 10.3390/rs17162903

Research Groups

Not specified in the provided text.

Short Summary

This study develops a drought sensitivity model for forests in China that incorporates lagged and accumulative effects, using machine learning and spatial analysis to identify the climatic and structural drivers of forest GPP response to drought.

Objective

• To systematically and quantitatively analyze the multi-factor drivers of forest drought sensitivity by accounting for lagged and accumulative effects.

Study Configuration

• Spatial Scale: National (China)
• Temporal Scale: Decadal (long-term)

Methodology and Data

• Models used: XGBoost–SHAP framework (for nonlinear associations and threshold effects) and Geodetector model (for spatially explicit interactions and coupling effects).
• Data sources: Long-term remote sensing datasets.

Main Results

• Lagged and Accumulative Effects: Lagged effects were observed in 99.52% and accumulative effects in 95.55% of forest Gross Primary Productivity (GPP); evergreen broadleaf forests showed the strongest effects, while deciduous needleleaf forests showed the weakest.
• Sensitivity Distribution: Evergreen needleleaf forests had the highest proportion of extreme drought sensitivity (16.94%), whereas deciduous needleleaf forests had the lowest (1.02%).
• Temporal Trend: The drought sensitivity index decreased in 67.12% of forests over the studied decades.
• Driving Factors: Temperature and precipitation are the primary drivers (exhibiting threshold effects). Climatic factors dominate evergreen forests, while forest age is a key driver for deciduous needleleaf forests.
• Interactive Effects: Spatial variations are driven by water–heat coupling in evergreen forests and structure–climate interactions in deciduous forests.

Contributions

The research fills a gap in the systematic quantitative analysis of drought sensitivity by integrating lagged and accumulative effects and employing a hybrid modeling approach (XGBoost-SHAP and Geodetector) to uncover nonlinear threshold effects and spatial coupling of drivers.

Funding

Not specified in the provided text.

Citation

@article{Xue2025Identifying,
  author = {Xue, Ziqiu and Diao, Simeng and Yang, Fei and Long, Fei and Wang, Wenjuan and Fang, Lantong and Liu, Yan},
  title = {Identifying Forest Drought Sensitivity Drivers in China Under Lagged and Accumulative Effects via XGBoost-SHAP},
  journal = {Remote Sensing},
  year = {2025},
  doi = {10.3390/rs17162903},
  url = {https://doi.org/10.3390/rs17162903}
}