Diao et al. (2026) Impacts of Climate Change, Human Activities, and Their Interactions on China’s Gross Primary Productivity

Identification

• Journal: Remote Sensing
• Year: 2026
• Date: 2026-01-14
• Authors: Yiwei Diao, Jie Lai, Lijun Huang, Anzhi Wang, Jiabing Wu, Y. Liu, L. D. Shen, Yuan Zhang, Rongrong Cai, W. B. Fei, Hao Zhou
• DOI: 10.3390/rs18020275

Research Groups

• Key Laboratory of Ecosystem Carbon Source and Sink, China Meteorological Administration, Wuxi University, Wuxi 214105, China
• CAS Key Laboratory of Forest Ecology and Silviculture, Institute of Applied Ecology, Chinese Academy of Sciences, Shenyang 110016, China
• University of Chinese Academy of Sciences, Beijing 101408, China
• Zhejiang Institute of Meteorological Science, Hangzhou 310008, China
• College of Meteorology and Oceanography, National University of Defense Technology, Changsha 410073, China

Short Summary

This study quantifies the spatio-temporal dynamics and driving mechanisms of Gross Primary Productivity (GPP) across China from 2001-2020, revealing a significant nationwide GPP increase driven by complex, ecosystem-specific nonlinear interactions among vegetation, climate, topography, and human activities.

Objective

• To quantify the contributions and interactions of climate, vegetation, topography, and human factors on Gross Primary Productivity (GPP) across China using GPP data from 2001–2020.
• To determine how climate, vegetation, topography, and human activities individually and interactively influence GPP across China.
• To assess whether interaction effects differ systematically among major ecosystem types.

Study Configuration

• Spatial Scale: Mainland China (approximately 9.6 million km²), with data harmonized to a 0.1° (approximately 10 km) spatial resolution.
• Temporal Scale: 2001–2020 (20 years), analyzed annually, seasonally (spring, summer, autumn, winter), and for the growing season.

Methodology and Data

• Models used: Mann–Kendall (MK) trend test, Sen’s slope estimator, Extreme Gradient Boosting (XGBoost) algorithm, SHapley Additive exPlanations (SHAP) for factor attribution and interaction analysis.
• Data sources:
  • Penman−Monteith–Leuning Version 2 (PML−V2) (China) terrestrial GPP dataset (daily, 500 m).
  • MODIS Land Cover (MCD12Q1) product (yearly, 500 m).
  • Land Surface Water Index (LSWI) derived from MODIS MOD09A1 surface reflectance (8-day, 500 m).
  • China Meteorological Forcing Dataset (CMFD) meteorological variables (3-hourly, 0.1°) including precipitation (PRECIP), surface pressure (PSFC), downward shortwave radiation (SWDOWN), air temperature (T2D), and specific humidity (Q), from which relative humidity (RH) and vapor pressure deficit (VPD) were calculated.
  • Global Land Surface Satellite Leaf Area Index V60 (GLASS LAI V60) product (8-day, 1 km).
  • Shuttle Radar Topography Mission (SRTM) Digital Elevation Model (DEM) (30 m) for Elevation, Slope, and Aspect.
  • Human Influence Index (HII) from the Wildlife Conservation Society (WCS) (annual, 1 km).
  • Landscape pattern indices (patch density (PD), landscape shape index (LSI), aggregation index (AI)) calculated from land cover data.

Main Results

• Nationwide GPP Trends: China's GPP exhibited a significant upward trend from 2001–2020, with the most pronounced increases in the Northeast Plain, North China Plain, and Loess Plateau. Regional increases were most rapid in Northeast, East, and Central South China (Sen's slopes: 0.0467 g C m⁻² d⁻¹ yr⁻¹, 0.0460 g C m⁻² d⁻¹ yr⁻¹, and 0.0438 g C m⁻² d⁻¹ yr⁻¹, respectively).
• Seasonal GPP Trends: GPP increased nationwide in spring and winter. During summer, GPP significantly increased in the Northeast Plain, North China Plain, and Loess Plateau, but declined in southern China (Jiangnan Hills, Yunnan–Guizhou Plateau). GPP declined significantly across most regions in autumn.
• Vegetation-Specific GPP Trends: Deciduous broadleaf forests (DBF), croplands, grasslands, and savannas showed widespread GPP increases. Croplands exhibited the largest GPP increase (0.0656 g C m⁻² d⁻¹ yr⁻¹). Deciduous needleleaf forests (DNF) showed no significant change.
• Dominant Drivers: Leaf Area Index (LAI) was identified as the primary contributor to GPP variations (overall contribution 54.6%), followed by climate factors (24.9%), topography (12.4%), and human activities (8.1%).
• Ecosystem-Specific Sensitivities:
  • Forest GPP (e.g., ENF, EBF) is strongly associated with climate–vegetation interactions (e.g., SWDOWN × LSWI in DBF). In DNF, human activities (HII, 39.7%) and climate (29.7%) are primary drivers, with a notable HII–PSFC interaction.
  • Grassland GPP is primarily linked to topography and human disturbance (e.g., LAI × HII, LAI × Elevation).
  • Cropland GPP is mainly related to management practices and environmental conditions, with interactions among patch density (PD), temperature (T2D), and LSWI.
  • Savannas and shrublands showed relatively weaker interaction strengths.
• Factor Interactions: GPP variations arise from coupled, nonlinear interactions. The most pronounced pairwise interaction was between surface pressure (PSFC) and elevation (2.1%), followed by LAI and PSFC (1.8%), LAI and elevation (1.5%), and LAI and LSWI (1.5%). High vapor pressure deficit (VPD) consistently suppressed GPP.

Contributions

• Provides a comprehensive, nationwide synthesis of GPP dynamics and their drivers across major ecosystems in China, addressing a previous limitation of studies focusing on specific regions or vegetation types.
• Quantifies the extent of nonlinear interactions among climate, vegetation, topography, and human activities in regulating GPP, moving beyond additive effects and revealing ecosystem-specific interaction regimes.
• Employs a machine-learning-based interpretability framework (XGBoost with SHAP) to explicitly characterize these complex, nonlinear, and ecosystem-dependent interactions, offering a mechanistic, interaction-centered interpretation of GPP regulation.
• Highlights the necessity of ecosystem-specific management and restoration strategies for carbon sequestration and climate adaptation, providing a basis for improving carbon cycle modeling and climate change adaptation planning.

Funding

• Open Research Fund Project of Key Laboratory of Ecosystem Carbon Source and Sink, China Meteorological Administration (Grant Number: ECSS-CMA202305)
• National Natural Science Foundation of China (Grant Number: 32401667)
• Natural Science Foundation of Liaoning Province (2024-BSBA-62)

Citation

@article{Diao2026Impacts,
  author = {Diao, Yiwei and Lai, Jie and Huang, Lijun and Wang, Anzhi and Wu, Jiabing and Liu, Y. and Shen, L. D. and Zhang, Yuan and Cai, Rongrong and Fei, W. B. and Zhou, Hao},
  title = {Impacts of Climate Change, Human Activities, and Their Interactions on China’s Gross Primary Productivity},
  journal = {Remote Sensing},
  year = {2026},
  doi = {10.3390/rs18020275},
  url = {https://doi.org/10.3390/rs18020275}
}