Liu et al. (2026) EGO: a global 0.05° hourly GPP dataset for monitoring diurnal photosynthesis dynamics

Identification

• Journal: Earth system science data
• Year: 2026
• Date: 2026-09-09
• Authors: Xi Liu, Xing Li, Dalei Hao, Jingfeng Xiao, Yanan ZHOU, Cenliang Zhao, Zikang Diao, Fuqiang Qu, Shangrong Lin, Xiangzhuo Liu, Zhaoying Zhang, Xinjie Liu, Helin Zhang
• DOI: 10.5194/essd-18-6613-2026

Research Groups

• School of Geography and Planning, Sun Yat-sen University, Guangzhou 510006, China
• Atmospheric, Climate, & Earth Sciences Division, Pacific Northwest National Laboratory, Richland, WA 99354, USA
• Earth Systems Research Center, Institute for the Study of Earth, Oceans, and Space, University of New Hampshire, Durham, NH 03824, USA
• INRAE, Bordeaux Sciences Agro, UMR 1391 ISPA, Villenave-d'Ornon, France
• International Institute for Earth System Sciences, Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application, Nanjing University, Nanjing, Jiangsu 210023, China
• Key Laboratory of Digital Earth Science, Aerospace Information Research Institute, Chinese Academy of Sciences, Beijing 100094, China
• Research Institute of Agriculture and Life Sciences, Seoul National University, Seoul 08826, Republic of Korea

Short Summary

This study presents a new global hourly GPP dataset (EGO) developed using a causal-constrained machine-learning framework based on eddy-covariance site observations. EGO achieves high accuracy in reproducing diurnal photosynthetic dynamics and captures vegetation responses to extreme climatic events.

Objective

• Develop a global 0.05° hourly GPP product upscaled from flux-tower observations.
• Evaluate the product's ability to reproduce key features of the diurnal photosynthetic cycle, including diurnal GPP centroid and relative midday depression.

Study Configuration

• Spatial Scale: Global coverage at 0.05° resolution.
• Temporal Scale: Hourly time step from 2000 to 2022.

Methodology and Data

• Models used: Causal Knowledge-driven Machine Learning (CKML-GPP) model, which integrates the PCMCI algorithm with XGBoost.
• Data sources:
  • Eddy covariance measurements from 190 flux sites.
  • Global gridded datasets of environmental variables (ERA5-Land reanalysis).
  • Satellite datasets (MODIS MCD43C4, MOD15A2H).

Main Results

• The CKML-GPP model achieves an R² of 0.76 and an RMSE of 4.17 μmol CO₂ m⁻² s⁻¹ on independent test sites.
• EGO outperforms two existing hourly upscaling products (FLUXCOM and X-BASE) in terms of overall accuracy, diurnal dynamics, and spatial patterns.

Contributions

• This study provides a reliable global 0.05° hourly GPP product that captures diurnal photosynthetic behavior and responses to extreme climatic events.
• The CKML-GPP model offers a promising approach for integrating causal inference with machine learning to improve the accuracy of flux upscaling products.

Funding

• Not specified in the provided text.

Citation

@article{Liu2026EGO,
  author = {Liu, Xi and Li, Xing and Hao, Dalei and Xiao, Jingfeng and ZHOU, Yanan and Zhao, Cenliang and Diao, Zikang and Qu, Fuqiang and Lin, Shangrong and Liu, Xiangzhuo and Zhang, Zhaoying and Liu, Xinjie and Zhang, Helin},
  title = {EGO: a global 0.05° hourly GPP dataset for monitoring diurnal photosynthesis dynamics},
  journal = {Earth system science data},
  year = {2026},
  doi = {10.5194/essd-18-6613-2026},
  url = {https://doi.org/10.5194/essd-18-6613-2026}
}