Research data supporting “How well can we forecast carbon credits generated by REDD+ projects?”

The supporting data contain, for all 26 ongoing REDD+ projects and for all sampled pixel sets within each project: 1. project_var.csv: project ID, country, year of start (t0), and project-level socio-economic and environmental variables (mean and standard error of reference carbon density for each land class, area, initial carbon density, average elevation/slope/remoteness of the project area and the surrounding region, average and growth rate of GDP per capita, corruption index) 1. boot_closs_obs_p.csv: Bootstrapped mean and 95% CI of observed annual carbon loss rates (MgC/ha/yr) of the project pixels, for each time interval from [t0, t1] to [t0, 2022] 2. boot_closs_obs_cf.csv: Bootstrapped mean and 95% CI of observed annual carbon loss rates (MgC/ha/yr) of the counterfactual pixels, for each time interval from [t0, t1] to [t0, 2022] 3. boot_closs_project.csv: Bootstrapped historical carbon loss rates (MgC/ha/yr) of the project pixels, for each time interval from [t-10, t0] to [t-1, t0] 4. boot_closs_regional.csv: Bootstrapped historical carbon loss rates (MgC/ha/yr) of the regional pixels, for each time interval from [t-10, t0] to [t-1, t0] 5. closs_annual.csv: Bootstrapped mean carbon loss rates (MgC/ha/yr) of the project, counterfactual and regional pixels, for each annual time interval from [t-10, t-9], [t-9, t-8]... to [2021, 2022]