This dataset generated global long-term GPP data with a latitude of 0.05∘, a longitude of 0.05∘, and a time interval of 8 days by revising the light utilization efficiency model (i.e. EC-LUE model). In the revised EC-LUE model, several major environmental variables were integrated: atmospheric carbon dioxide concentration, radiation composition, and atmospheric water vapor pressure difference (VPD). The long-term changes in these environmental variables may have a significant impact on global vegetation productivity. The vorticity covariance (EC) measurement data from 95 tall towers from the FLUXNET2015 dataset were used to calibrate and validate the model, covering nine major ecosystem types worldwide. The revised EC-LUE model can explain 71% of the annual GPP spatial changes among 95 observation points. At over 95% of observation points, the correlation coefficient (R2) between tower estimation and model simulation of GPP seasonal variation is greater than 0.5. Among all 55 stations with observation years exceeding 5 years, the average annual GPP simulated by the tower estimation and model was 0.44, significantly higher than the original EC-LUE model (R2=0.36) and other LUE models (R2ranging from 0.06 to 0.30, with an average value of 0.16). On a global scale, there are significant differences in the size and interannual variations of GPP obtained from light utilization efficiency models, machine learning models, and process based biophysical models. The revised EC-LUE model quantifies the global average GPP from 1982 to 2017 as 106.2 ± 2.9 Pg C yr-1, with a trend of 0.15 Pg C yr-1. Sensitivity analysis shows that the GPP simulated by the EC-LUE model is sensitive to atmospheric CO2concentration, VPD, and radiation. During the period from 1982 to 2017, the increase in VPD (Pg C yr-1) partially offset the impact of CO2fertilization on global GPP (0.22 ± 0.07 Pg C yr-1). The long-term changes in environmental variables can be well reflected in the global GPP. Overall, the revised EC-LUE model can provide reliable long-term estimates of global GPP
1. Vorticity covariance tower data: FLUXNET2015 dataset（ http://www.fluxdata.org ）The variables including over 200 carbon flux, energy flux, and meteorological variables were collected and processed by the FLUXNET community on site
2. Global data
(1) Air temperature, Dew point temperature, Direct PAR, Diffuse PAR, sourced from https://gmao.gsfc.nasa.gov/reanalysis/MERRA-2/
(2) LAI data sourced from http://www.glass.umd.edu/Download.html
(3) The Land cover map data is sourced from https://lpdaac. usgs. gov/products/mcd12q1v006/
(4) C4The data for the crop percentage is sourced from https/doi. org/10.3334/ORNLDAAC/932
(5) CO concentration data is sourced from https//www.esrl.noaa. gov/gmd/ccgg/trends/
1. Using 95 EC sites in the FLUXNET2015 dataset to optimize parameters and evaluate the performance of the revised EC-LUE model for 95 EC sites in the dataset
2. The MCD12Q1 product using the IGBP classification scheme is used as the input land cover map
3. The ISLSCP II C4 vegetation percentage map is used to separate C3 and C4 crops
4. The National Oceanic and Atmospheric Administration (NOAA) Earth Systems Research Laboratory (ESRL) carbon monoxide concentration dataset is used to represent the effectiveness of carbon monoxide fertilization
The data quality is good
This work is licensed under
CC BY 4.0 (Creative Commons Attribution 4.0 International License).
CSTR:11738.11.NCDC.FIGSHARE.DB4095.2023
10.6084/m9.figshare.8942336.v3
