This dataset contains MPI-ESM-1.2-LR output from the work package 1 (WP1) FRST scenario of idealized constant global re-/afforestation. The simulation branches from the 2014 CMIP6 historical concentration-driven simulation and spans 160 years (2015-01-01 to 2174-12-31), with anthropogenic (trace gases, aerosols, population density) and natural forcings (solar radiation, wildfire, lightning, natural aerosols) held constant at 2014 levels.
In the FRST scenario, land cover changes to forest PFTs on the entire fraction of hospitable land of each land grid cell. This is applied in a checkerboard-like pattern to half of all land grid cells within the first to fifth year of the simulation. Basically, all PFTs that are neither forest nor bare soil (unhospitable land) were reduced to the minimum fraction value (fract_small = 1e-10). The remaining forest fractions are increased such that fractions within a grid cell add up to 100 %. As the bare soil fraction is preserved, the resulting land cover map only contains forest PFTs and bare soil.
Generally, we ran the land model JSBACH3.2 with the following options:
- use_dynveg = false
- use_disturbance = true
- lcc_forcing_type = transitions
- lcc_scheme = 2
To be able to perform a land cover change to 100% forest in specific gird cells and using land use transitions at the same time we had to make changes to specific source files of the JSBACH3.2 code. Please contact the authors for details.
The approach mimics forest expansion across all vegetated, cropland and urban areas but avoids trees being planted in e.g. desert, high-altitude and tundra regions (unhospitable land). This approach is only possible for grid cells containing at least some forest PFTs. For grid cells without any forest PFTs, we calculated the latitudinal averages of the relative forest PFTs distributions consisting of different forest PFT species. These values are then considered to be representative for this latitudinal band and are used to replace all other vegetation in the grid cell.
The checkerboard-like pattern, with its homogeneous distribution of changed and unchanged grid cells, allows the application of an established method to separate local and nonlocal biogeophyiscal as well as biogoechemical effects of this land cover change (see Winckler et al., 2017 (doi: 10.1175/JCLI-D-16-0067.1), De Hertog et al., 2023 (doi: 10.5194/esd-14-629-2023), and Guo et al., 2025 (doi: 10.5194/esd-16-631-2025)).