Lyu et al. (2025) An indicator framework for assessing forest ecosystem productivity resilience and transition risks under climate change

Identification

• Journal: Ecological Indicators
• Year: 2025
• Date: 2025-11-17
• Authors: Yingshuo Lyu, Xi Zheng, Han Wang, Teng Liu, Chutong Chao, Xiaoyang Ou
• DOI: 10.1016/j.ecolind.2025.114388

Research Groups

• School of Landscape Architecture, Beijing Forestry University, Beijing, China
• Department of Earth System Science, Ministry of Education Key Laboratory for Earth System Modeling, Institute for Global Change Studies, Tsinghua University, Beijing, China
• School of Systems Science and Institute of Nonequilibrium Systems, Beijing Normal University, Beijing, China
• Earth System Modelling, School of Engineering and Design, Technical University of Munich, Munich, Germany

Short Summary

This study develops a composite indicator framework using critical slowing down (CSD) metrics from gross primary production (GPP) to quantify forest ecosystem productivity resilience (EPR) and identify state transitions under climate change across China. It reveals that over half of China's forests experienced EPR decline, primarily due to climatic water availability, with a significant portion transitioning to unstable multistability.

Objective

• Quantify EPR spatiotemporal variation and its drivers.
• Validate the applicability of CSD metrics for EPR estimation.
• Characterize EPR states and identify potential transitions.
• Assess future risk of EPR across climate change scenarios.

Study Configuration

• Spatial Scale: China, at a 0.05° × 0.05° grid resolution (derived from 30 m to 500 m land cover data).
• Temporal Scale: Historical analysis from 2000 to 2018; future projections for 2060–2080.

Methodology and Data

• Models used:
  • Boosted Regression Tree (BRT) models for attribution analysis and future projections.
  • Geographically Weighted Regression Kriging (GWRK) for SPEI interpolation.
  • Gaussian Mixture Models for characterizing EPR states.
  • Seasonal-trend decomposition by Loess (STL) for deseasoning and detrending GPP and vegetation data.
  • GPP models: Statistical methods (e.g., NIRv), Light Use Efficiency (LUE) models (MODIS MOD17, revised EC-LUE, two-leaf LUE (TL-LUE)), and machine learning (FluxSat v2.0).
• Data sources:
  • Satellite: MODIS MCD12Q1, ESA-CCI, China Land Cover Dataset (CLCD) for land cover; MODIS MCD15A2H (Leaf Area Index - LAI), MODIS MOD13Q1 (Normalized Difference Vegetation Index - NDVI, Enhanced Vegetation Index - EVI) for vegetation dynamics.
  • Observation: China Meteorological Administration for monthly climate variables (maximum/minimum evaporation (EVPmax/EVPmin), precipitation (PRE), air temperature (TEM), air pressure (PRS), relative humidity (RHU), sunshine duration (SSD), wind speed (WIN), ground surface temperature (GST)); FLUXNET eddy covariance data for GPP validation at four flux tower sites (Xishuangbanna, Dinghushan, Qianyanzhou, Changbaishan).
  • Reanalysis/Model output: Ensemble of five GPP datasets; ten General Circulation Models (GCMs) from CMIP6 (ACCESSCM2, BCCCSM2MR, ECEarth3Veg, FIOESM20, GISSE21G, INMCM50, IPSLCM6ALR, MPIESM12HR, MRIESM20, UKESM10LL) for future climate projections under SSP2-4.5, SSP3-7.0, and SSP5-8.5 scenarios.
  • Index: 3-month Standardized Precipitation Evapotranspiration Index (SPEI).
  • Ancillary: Vegetation Map of the People’s Republic of China (1:1,000,000) for forest types.

Main Results

• 57.47% of forests in China experienced an EPR decline between 2000 and 2018, with 55.64% showing declining EPR despite increasing GPP trends, indicating expanding yet more vulnerable carbon sinks.
• Climatic water availability (SPEI, maximum temperature, precipitation) was identified as the dominant driver, accounting for over 90% of the variation in annual temporal autocorrelation (TAC) and its trend (δTAC).
• Coniferous-broadleaf forests in the temperate zone experienced the most severe abrupt declines under both water surplus and deficit conditions.
• EPR reductions and productivity declines are statistically linked and emerge as ecosystems approach critical thresholds.
• 24.58% of forests transitioned from a state of uniform stability (State I) to unstable multistability (State II), characterized by narrower and shallower attraction basins and higher amplitudes and rates of CSD change. This transition was most frequent in coniferous-broadleaf (37.32%) and broadleaf forests (30.95%), primarily occurring in climatic and geographic boundary regions.
• Future projections (2060–2080) indicate a spatially persistent and intensifying pattern of EPR risk, with SSP2-4.5 projected as the worst-case scenario, leading to the highest transition occurrence probabilities (0.18) and significant EPR loss. Up to 67.15% of China's forests are projected to be at risk of critical transition under this scenario.
• Under SSP2-4.5, 59.22% of forests are expected to continuously decline in EPR, with vulnerable carbon sinks shifting towards the productive southern regions. Conversely, the high-emission SSP5-8.5 scenario projects average EPR improvement in temperate forests due to rising temperatures.

Contributions

• Developed a composite indicator framework that quantifies Ecosystem Productivity Resilience (EPR) using Critical Slowing Down (CSD) metrics derived from Gross Primary Production (GPP) time series.
• Integrated structure and process characteristics of resilience to identify spatially explicit EPR states and potential critical transitions, addressing a gap in previous studies focused on temporal trends.
• Validated the applicability of CSD metrics for EPR estimation through alignment with abrupt productivity declines (ADs) and observed events.
• Provided a pixel-level diagnostic tool for resilience states, enabling risk mapping, early warning, and adaptive forest management strategies under climate change.
• Elucidated mechanisms driving EPR trajectories by linking carbon sink variability with functional resilience.

Funding

• National Natural Science Foundation of China (Grant No. 32371643).

Citation

@article{Lyu2025indicator,
  author = {Lyu, Yingshuo and Zheng, Xi and Wang, Han and Liu, Teng and Chao, Chutong and Ou, Xiaoyang},
  title = {An indicator framework for assessing forest ecosystem productivity resilience and transition risks under climate change},
  journal = {Ecological Indicators},
  year = {2025},
  doi = {10.1016/j.ecolind.2025.114388},
  url = {https://doi.org/10.1016/j.ecolind.2025.114388}
}