Qiu et al. (2025) Large contribution of antecedent climate to ecosystem productivity anomalies during extreme events

Identification

• Journal: Nature Geoscience
• Year: 2025
• Date: 2025-11-21
• Authors: Jinghao Qiu, Yao Zhang, Mengyang Cai, Trevor F. Keenan, Hongying Zhang, Pierre Gentine, Mitra Asadollahi, Sha Zhou, Shilong Piao
• DOI: 10.1038/s41561-025-01856-4

Research Groups

• Institute of Carbon Neutrality, Sino-French Institute for Earth System Science, College of Urban and Environmental Sciences, Peking University, Beijing, China
• Climate and Ecosystem Sciences Division, Lawrence Berkeley National Laboratory, Berkeley, CA, USA
• Department of Environmental Science, Policy and Management, University of California, Berkeley, Berkeley, CA, USA
• Department of Earth and Environmental Engineering, Columbia University, New York, NY, USA
• Department of Geography, National University of Singapore, Singapore, Singapore
• State Key Laboratory of Earth Surface Processes and Disaster Risk Reduction, Faculty of Geographical Science, Beijing Normal University, Beijing, China

Short Summary

This study quantifies the significant contribution of antecedent climate conditions to ecosystem productivity anomalies during extreme events using an interpretable machine-learning framework, revealing that memory effects, particularly from precipitation, temperature, and vapour pressure deficit, substantially influence ecosystem responses, especially in semi-arid regions.

Objective

• To quantify the direction, strength, and influential duration of memory effects from antecedent climate on ecosystem gross primary productivity (GPP) anomalies during extreme events.

Study Configuration

• Spatial Scale: Global
• Temporal Scale: 1995–2020 (26 years)

Methodology and Data

• Models used: Interpretable machine-learning framework, Entity-Aware Long Short-Term Memory (EA-LSTM) model, Integrated Gradients (IG) explainer.
• Data sources:
  • FLUXNET 2015 dataset (eddy covariance data)
  • ERA5-Land dataset (reanalysis, meteorological variables)
  • Plant functional type dataset (MCD12Q1v061)
  • Canopy-height dataset
  • Soil properties (SoilGrids, Hihydrosoil)
  • Nitrogen deposition dataset
  • Topography data (GMTED2010)
  • TROPOMI SIF dataset (Solar-Induced Fluorescence)
  • FLUXCOM-X-BASE dataset
  • BESS v2.0 dataset
  • SMAP L4 carbon product
  • Global Forest Change dataset
  • GPP dataset generated in this study (via figshare)

Main Results

• Antecedent climate conditions contribute 38.2% to ecosystem productivity during extreme events.
• Precipitation accounts for 42.2% of these memory effects, followed by temperature (22.1%) and vapour pressure deficit (20.8%).
• Extreme events conditioned by long-term climatic variations often lead to higher productivity losses compared to short-term extremes.
• Semi-arid ecosystems exhibit the largest productivity anomalies and prolonged memory effects.

Contributions

• Quantifies the direction, strength, and duration of memory effects on ecosystem productivity anomalies during extreme events using an interpretable machine-learning framework.
• Provides an observation-constrained benchmark for memory effects, highlighting their role in regulating carbon flux variations.
• Reveals that long-term climatic variations exacerbate productivity losses, especially in semi-arid ecosystems.

Funding

• National Natural Science Foundation of China (42371096, 42141005)
• National Key R & D Program of China (2023YFF0805702)

Citation

@article{Qiu2025Large,
  author = {Qiu, Jinghao and Zhang, Yao and Cai, Mengyang and Keenan, Trevor F. and Zhang, Hongying and Gentine, Pierre and Luo, Xiangzhong and Asadollahi, Mitra and Zhou, Sha and Piao, Shilong},
  title = {Large contribution of antecedent climate to ecosystem productivity anomalies during extreme events},
  journal = {Nature Geoscience},
  year = {2025},
  doi = {10.1038/s41561-025-01856-4},
  url = {https://doi.org/10.1038/s41561-025-01856-4}
}