Wang et al. (2025) Decoupling anthropogenic and climate impacts on vegetation dynamics in China’s Huaihe River Basin using geodetector

Identification

• Journal: Scientific Reports
• Year: 2025
• Date: 2025-11-18
• Authors: Xinyu Wang, Yan Ling Li, Shibo Fang, Chao Dong, Li Sun
• DOI: 10.1038/s41598-025-24198-y

Research Groups

• College of Information Science and Engineering, Shandong Agricultural University, Taian 271018, China.
• State Key Laboratory of Severe Weather, Chinese Academy of Meteorological Sciences, Beijing 100081, China.

Short Summary

This study developed a novel hybrid framework combining AR1 modeling and spatial autocorrelation analysis with Geodetector to decouple anthropogenic and climate impacts on vegetation dynamics in China's Huaihe River Basin (HRB) from 2000 to 2022. It found a significant basin-wide greening trend (0.00152 yr⁻¹ NDVI increase), with land use type being the dominant spatial driver and extreme climatic events governing temporal anomalies, highlighting complex nonlinear interactions between drivers.

Objective

• To understand the spatiotemporal dynamics of vegetation under the coupled influence of climatic and anthropogenic drivers in China’s Huaihe River Basin (HRB).
• To develop a novel hybrid framework combining Univariate Stationary First-order Gaussian Autoregressive (AR1) modeling with spatial autocorrelation analysis (Moran’s I index) for robust detection of vegetation trends, stability evaluation, and clustering characteristics, addressing inherent temporal and spatial autocorrelation challenges.
• To quantify the impacts of climatic extremes, land use change, and urbanization on Normalized Difference Vegetation Index (NDVI) patterns in the HRB from 2000 to 2022 using the Geodetector model.

Study Configuration

• Spatial Scale: Huaihe River Basin (HRB), China, covering approximately 270,000 km². Analysis performed at a 1 km × 1 km pixel resolution, with initial MODIS NDVI data at 500 m resolution.
• Temporal Scale: 2000–2022 (23 years) for NDVI and driving factors. Monthly composites for NDVI. Land-use data for 2000, 2005, 2010, 2015, 2020, and 2022.

Methodology and Data

• Models used:
  • Univariate Stationary First-order Gaussian Autoregressive (AR1) model (PARTS method) for robust trend detection and temporal autocorrelation correction.
  • Spatial autocorrelation analysis (Global and Local Moran’s I index) for quantifying basin-wide clustering patterns and identifying hotspots/cold spots.
  • Geodetector model (Divergence, Factor, and Interaction Detector) for quantifying the explanatory power (q-statistic) of individual driving factors and their nonlinear interaction effects.
  • Univariate linear regression model (least squares method) for overall and pixel-level NDVI trend analysis.
  • Coefficient of Variation (COV) for analyzing NDVI stability.
• Data sources:
  • Satellite/Remote Sensing: MODIS NDVI (Collection 6, 500 m spatial resolution, 16-day composites, 2000–2022) from NASA Earthdata Search. Land use type data (2000, 2005, 2010, 2015, 2020, 2022) derived from Landsat imagery, resampled to 1 km. Elevation (DEM) data from Shuttle Radar Topography Mission (SRTM) Version 4.1, resampled to 1 km.
  • Observation/Reanalysis/Derived:
    • Land use, soil type, vegetation condition, population, GDP, and climatic background data (dryness, humidity index, > 10 °C cumulative temperature) from the Resource and Environment Data Center of the Chinese Academy of Sciences.
    • Slope and aspect data generated from DEM.
    • Meteorological station data (annual mean temperature, annual precipitation, annual maximum/minimum temperature) from the National Climatic Data Center (NCDC), NOAA, interpolated to 1 km grids using inverse-distance weighting or thin-plate splines.
    • Standardized Precipitation Evapotapotranspiration Index (SPEI) for monthly drought and flood conditions.

Main Results

• The Huaihe River Basin (HRB) experienced a significant greening trend from 2000 to 2022, with a mean annual NDVI increase of 0.00152 yr⁻¹ (p < 0.05) when temporal autocorrelation was considered.
• Spatially, 42.89% of the basin showed significant NDVI growth, while 9.54% exhibited degradation. Notably, 47.57% showed non-significant increases, highlighting the importance of temporal autocorrelation correction in trend detection (without correction, 90.02% showed significant increase).
• Geodetector analysis identified land use type as the consistently dominant spatial heterogeneity driver of NDVI (q-values ranging from 0.35 to 0.42 across years).
• Extreme climatic events, particularly the 2000–2001 mega-drought, were found to govern temporal anomalies, with precipitation showing strong explanatory power (q=0.35) during that period.
• Urban expansion generally led to reduced vegetation cover, except in areas near water bodies where NDVI remained stable.
• Interaction analysis revealed nonlinear synergistic effects between anthropogenic activities (e.g., GDP, population) and climatic factors, with interaction q-values often exceeding the sum of individual factors (e.g., "precipitation × population" interaction yielded a q-value of 0.18 in 2022, higher than individual q-values of 0.022 for precipitation and 0.13 for population).
• Monthly NDVI and SPEI showed a strong positive correlation, indicating high vegetation responsiveness to hydroclimatic conditions, with a typical lag of 1–2 months. Both severe drought (SPEI < -1) and excessive moisture (SPEI > 1.5) were found to hinder vegetation growth.

Contributions

• Developed and validated a novel hybrid framework that integrates Univariate Stationary First-order Gaussian Autoregressive (AR1) modeling with spatial autocorrelation analysis and Geodetector, providing a robust method for analyzing vegetation dynamics in the presence of spatiotemporal autocorrelation.
• Provided a more accurate and decoupled quantification of the relative contributions of anthropogenic activities (land use change, urbanization, GDP, population) and climatic factors (temperature, precipitation, extreme events) to NDVI variations in the Huaihe River Basin.
• Demonstrated the critical importance of accounting for temporal and spatial autocorrelation in remote sensing time series analysis to prevent overestimation of significant trends and improve statistical reliability.
• Offered a replicable analytical framework for regional-scale ecological assessments, particularly valuable for climate-sensitive transitional ecosystems.
• Proposed specific, tiered management strategies for the HRB, including climate-adaptive zoning, riparian corridor conservation, and dynamic land-use optimization, to balance ecological restoration with urban development.

Funding

• National Natural Science Foundation of China, grant number 11701337.
• Natural Science Foundation of Shandong Province, grant number ZR2021MD096.
• Shandong Independent Innovation Project, grant number 2012CX90202.
• Horizontal Scientific Research Project, grant number YG2021010-03.

Citation

@article{Wang2025Decoupling,
  author = {Wang, Xinyu and Li, Yan Ling and Fang, Shibo and Dong, Chao and Sun, Li},
  title = {Decoupling anthropogenic and climate impacts on vegetation dynamics in China’s Huaihe River Basin using geodetector},
  journal = {Scientific Reports},
  year = {2025},
  doi = {10.1038/s41598-025-24198-y},
  url = {https://doi.org/10.1038/s41598-025-24198-y}
}