The product is generated in 2025. LPJ-GUESS (version 4.0, revision 6562) is forced with hourly ERA5 climate datasets to simulate global terrestrial CO2 flux at a resolution of 0.5 degree. LPJ-GUESS is a process-based dynamic global vegetation model, it uses time series data (e.g. climate forcing and atmospheric carbon dioxide concentrations with WMO CO2 X2019 scale) as input to simulate the effects of environmental change on vegetation structure and composition in terms of global plant functional types (PFTs), soil hydrology and biogeochemistry (Smith et al., 2001, https://web.nateko.lu.se/lpj-guess/). This simulation models natural fluxes, such as those from vegetation and soil under climate change, via natural processes such as biomass allocation, growth, reproduction, establishment, mortality, and disturbance. This means it does not include anthropogenic activities, such as forest management or agricultural management. The positive value means CO2 uptake from the atmosphere, and the negative value means CO2 release to the atmosphere.