Wang et al. (2025) A global hourly gross primary production dataset from 2001 to 2020

Identification

• Journal: Scientific Data
• Year: 2025
• Date: 2025-12-04
• Authors: Yong Wang, Zhi He, Wei Zhao, Gaofei Yin, Xiaobin Guan, Xinyao Xie
• DOI: 10.1038/s41597-025-06371-0

Research Groups

• Institute of Mountain Hazards and Environment, Chinese Academy of Sciences, Chengdu, China
• University of Chinese Academy of Sciences, Beijing, China
• Faculty of Geosciences and Environmental Engineering, Southwest Jiaotong University, Chengdu, China
• School of Resource and Environmental Sciences, Wuhan University, Wuhan, China

Short Summary

This paper introduces a novel global hourly gross primary production (GPP) dataset for 2001–2020, generated at a 0.1° spatial resolution using a modified radiation scalar two-leaf LUE (RTL-LUE) model. This dataset significantly improves the capture of short-term GPP variations and extreme environmental stresses, offering enhanced insights into terrestrial carbon dynamics.

Objective

• To develop and validate a global hourly gross primary production (GPP) dataset for 2001–2020 at a 0.1° spatial resolution, based on a modified radiation scalar two-leaf LUE (RTL-LUE) model, to better capture short-term variations in ecosystem productivity and improve insights into carbon fluxes.

Study Configuration

• Spatial Scale: Global, 0.1° (approximately 11 kilometers at the equator).
• Temporal Scale: Hourly, 2001–2020.

Methodology and Data

• Models used: Modified radiation scalar two-leaf LUE (RTL-LUE) model. Comparisons were made with the big-leaf MODIS GPP (MOD17A2) and the two-leaf LUE (TL-LUE) datasets.
• Data sources:
  • ERA5-land (hourly climate data: downward shortwave radiation, dewpoint temperature, air temperature, derived vapor pressure deficit).
  • GLASS Leaf Area Index (LAI) (8-day, 0.05°, aggregated to 0.1°).
  • MODIS land cover (MCD12C1, annual, 0.05°, aggregated to 0.1°).
  • NOAA atmospheric CO₂ concentration (monthly).
  • FLUXNET2015 network (184 eddy covariance (EC) flux sites for calibration and validation, providing 1-hour EC GPP).

Main Results

• The RTL-LUE GPP dataset estimates a global total GPP of 124.77 PgC/yr for 2001–2020, which is slightly lower than the 8-day TL-LUE GPP (126.92 PgC/yr) but higher than MOD17A2 GPP (113.90 PgC/yr).
• The hourly resolution of the RTL-LUE model effectively captures short-term extreme stresses, resulting in slightly lower but more realistic GPP estimates compared to coarser temporal resolution models.
• The dataset reveals that annual GPP variability can reach up to approximately 0.10 g C/m²/h at hourly scales.
• RTL-LUE GPP demonstrates robust performance against EC GPP at 184 flux towers, showing an average coefficient of determination (R²) of 0.84 and a root-mean-square error (RMSE) of 0.03 gC/m²/h at the Plant Functional Type (PFT) scale, and an average R² of 0.61 and RMSE of 0.15 gC/m²/h at the site scale.
• A widespread greening pattern is observed from 2001 to 2020, with approximately 49.64% of global vegetated land showing a GPP trend of 0 to 5 gC/m²/yr², particularly in the Sahel, India, eastern China, Europe, and boreal regions.
• All three GPP datasets (RTL-LUE, TL-LUE, MOD17A2) exhibit increasing trends during 2001–2020, with RTL-LUE showing a rate of 0.48 PgC/yr².

Contributions

• Generation of the first global hourly GPP dataset at 0.1° spatial resolution for 2001–2020, addressing the critical limitation of coarse temporal resolution in existing GPP products.
• Development and application of a modified RTL-LUE model that effectively captures short-term variations and extreme environmental stresses on GPP, leading to improved accuracy in GPP estimation.
• Provision of a valuable, publicly available dataset (https://doi.org/10.57760/sciencedb.29500) that can advance the understanding of ecosystem carbon cycling and improve carbon cycle models, especially for analyzing rapid vegetation responses to environmental changes.
• Comprehensive validation against 184 eddy covariance flux towers and detailed comparisons with other LUE-based GPP datasets, demonstrating the robust performance and highlighting the benefits of higher temporal resolution.

Funding

• National Natural Science Foundation of China (42471429, 42222109, 42201418)
• National Key Research and Development Program of China (2024YFF1306503)
• Chinese Academy of Sciences Youth Innovation Promotion Association (2023390)
• Sichuan Science and Technology Program (2024NSFSC0794)
• Science and Technology Research Program of Institute of Mountain Hazards and Environment Chinese Academy of Sciences (IMHE-ZYTS-05)

Citation

@article{Wang2025global,
  author = {Wang, Yong and He, Zhi and Zhao, Wei and Yin, Gaofei and Guan, Xiaobin and Xie, Xinyao},
  title = {A global hourly gross primary production dataset from 2001 to 2020},
  journal = {Scientific Data},
  year = {2025},
  doi = {10.1038/s41597-025-06371-0},
  url = {https://doi.org/10.1038/s41597-025-06371-0}
}