Xue et al. (2025) Climate–human interactions influence widespread peatland subsidence and soil carbon stock vulnerability in China

Identification

• Journal: Communications Earth & Environment
• Year: 2025
• Date: 2025-11-21
• Authors: Zhenshan Xue, Ruxu Li, Ming Jiang, Yuanchun Zou, Haitao Wu, Xianguo Lü, Yeqiao Wang, Enyong Tian, Rihong Zhang
• DOI: 10.1038/s43247-025-02896-9

Research Groups

• Northeast Institute of Geography and Agroecology, Chinese Academy of Sciences, Changchun, China
• University of Chinese Academy of Sciences, Beijing, China
• Department of Natural Resources Science, University of Rhode Island, Kingston, RI, USA

Short Summary

This study provides the first national-scale assessment of peatland subsidence across China by integrating satellite radar observations with advanced modeling. It reveals widespread subsidence, particularly in hotspots like the Zoigê Plateau, driven by a combination of climate variability, drought, peat depth, and human pressures, projecting that over 65% of China's peatland carbon stock will be vulnerable under future high-emission scenarios.

Objective

• Delineate the spatial distribution of peatlands and their key soil properties across China.
• Characterize spatial and temporal patterns of peatland surface deformation across eight representative regions using satellite-based SBAS-InSAR analysis.
• Predict national-scale subsidence risk under current and future scenarios, and identify the dominant climatic, ecological, edaphic, topographic, and anthropogenic drivers shaping peatland vulnerability.

Study Configuration

• Spatial Scale: National scale across mainland China, with detailed analysis in eight representative peatland regions and classification into 15 ecological subregions. Predictions generated at 1 km spatial resolution.
• Temporal Scale: InSAR observations from January 2017 to December 2023. Climate-related predictors for current conditions represent the 2010–2020 period. Future scenario simulations project conditions for the end of the 21st century (SSP1-2.6, SSP3-7.0, SSP5-8.5).

Methodology and Data

• Models used:
  • Small Baseline Subset Interferometric Synthetic Aperture Radar (SBAS-InSAR) technique for surface deformation.
  • Random Forest regression models for peat depth, bulk density, and total organic carbon.
  • Ensemble machine learning framework (Random Forest, GBM, XGBoost) for peatland distribution mapping.
  • XGBoost regression model for national-scale peatland subsidence prediction.
  • Minimum Cost Flow (MCF) algorithm for phase unwrapping.
  • Singular Value Decomposition (SVD) for time-series inversion.
• Data sources:
  • Satellite radar observations: Sentinel-1 SLC images (IW mode, VV polarization).
  • Topographic data: 30 m Shuttle Radar Topography Mission (SRTM) DEM.
  • Field data: 935 peat profile samples (depth, bulk density, total organic carbon) from the 2013–2018 national survey of marsh wetland resources and ecological benefits, supplemented by published studies.
  • Validation data: GNSS time series data from the Mohe GNSS station (HLMH).
  • Environmental predictors: A comprehensive set of 30 raster predictors covering climatic, topographic, vegetation, soil, and anthropogenic conditions.
  • Climate projections: Multi-model ensemble mean of five Coupled Model Intercomparison Project Phase 6 (CMIP6) global climate models (GFDL-ESM4, UKESM1-0-LL, MPI-ESM1-2-HR, IPSL-CM6A-LR, and MRI-ESM2-0) under Shared Socioeconomic Pathway (SSP) scenarios.

Main Results

• China's peatlands cover approximately 22,981 km², storing an estimated 3.672 PgC of soil organic carbon (SOC), with an average SOC density of 159.7 kg C m⁻².
• Peatland subsidence is widespread, with hotspots in the Zoigê Plateau (ZP), Yunnan–Guizhou Plateau (YGP), and Daxing’an Mountains (DXAM). ZP alone accounts for 1.63 PgC of the national stock.
• SBAS-InSAR analysis (2017–2023) showed significant subsidence, with a root mean square error (RMSE) of 4.32 mm compared to GNSS data. Mean annual subsidence rates reached −5.73 mm·yr⁻¹ in Zoigê and Honghe peatlands, and −5.48 mm·yr⁻¹ in Caohai Peatland.
• Machine learning identified precipitation (mean absolute SHAP = 0.719) as the most influential driver of subsidence, followed by net primary productivity (0.600), livestock density (0.427), aspect (0.374), soil sand content (0.294), and peat depth (0.293).
• Under current conditions, 61.4% of China's peatlands are subsiding. Projections indicate a decrease to 56.1% under SSP1-2.6, but a marked increase to 68.1% under the high-emission SSP5-8.5 scenario.
• The proportion of China's peatland carbon stock located in subsiding areas is 63.3% (2.27 PgC) currently, rising to 65.2% (2.34 PgC) under SSP5-8.5.
• Estimated hydrological function loss due to subsidence is approximately 1.15 × 10⁸ m³·yr⁻¹ currently, increasing to 1.43 × 10⁸ m³·yr⁻¹ under SSP5-8.5.

Contributions

• Provides the first national-scale assessment of peatland subsidence across China, integrating satellite radar observations and advanced machine learning.
• Identifies widespread peatland subsidence hotspots and quantifies their rates, revealing the combined influence of climatic (e.g., precipitation variability, drought) and anthropogenic (e.g., livestock density, drainage) pressures.
• Projects future peatland subsidence risks under different climate scenarios (SSP1-2.6, SSP3-7.0, SSP5-8.5), highlighting the increasing vulnerability of China's peatland carbon stock and hydrological functions.
• Quantifies the potential loss of hydrological function, emphasizing broader implications for regional water resource stability.
• Underscores the urgent need for region-specific and scenario-informed conservation strategies to mitigate compounded risks to peatland resilience.

Funding

• National Natural Science Foundation of China (grant nos. 42430511, 42494822, and U2243230)
• National Key Research and Development Program of China (grant nos. 2023YFF1304502 and 2022YFF1300900)

Citation

@article{Xue2025Climatehuman,
  author = {Xue, Zhenshan and Li, Ruxu and Jiang, Ming and Zou, Yuanchun and Wu, Haitao and Lü, Xianguo and Wang, Yeqiao and Tian, Enyong and Zhang, Rihong},
  title = {Climate–human interactions influence widespread peatland subsidence and soil carbon stock vulnerability in China},
  journal = {Communications Earth & Environment},
  year = {2025},
  doi = {10.1038/s43247-025-02896-9},
  url = {https://doi.org/10.1038/s43247-025-02896-9}
}