General information This dataset contains soybean yield projections in Europe under historical (1981-2010) and future climate with moderate (RCP 4.5) to intense (RCP 8.5) warming, up to the 2050s and 2090s time horizons. The data has been generated by Guilpart et al. (2022) Data-driven projections suggest large opportunities to improve Europe's soybean self-sufficiency under climate change, Nature Food. All details can be found in this paper. A brief summary is provided below. Summary of soybean yield projections methodology Yield projections have been performed using data-driven relationships between climate and soybean yield derived from machine-learning (Random Forest). The Random Forest model was trained using (i) the the global dataset of historical yields updated version (Iizumi et al. 2014a), which includes grid-wise soybean yields worldwide with the grid size of 1.125 degree over 1981-2010, and (ii) the global retrospective meteorological forcing dataset tailored for agricultural application (GRASP, Iizumi et al. 2014b), which covers the period 1961–2010 at the same spatial resolution as yield data, i.e. a grid size of 1.125 degree. Time-detrended soybean yield data was related (using Random Forest) to 35 climate variables defined at a monthly time step over the seven months of the soybean growing season, plus the fraction of irrigated area, i.e. a total of 36 variables. The 35 climate variables are monthly mean daily minimum and maximum temperatures (Tmin and Tmax, degree Celsius), monthly total precipitation (rain, mm month-1), monthly mean daily total solar radiation (solar, MJ m-2 day-1), monthly mean air vapor pressure (VP, hPa). The fitted model showed high R² (higher than 0.9) and low RMSE (0.35 t ha-1) between observed and predicted yields based on cross-validation. Then, soybean yield projections under historical over whole Europe have been performed using the GRASP climate data, and yield projections under future climate have been performed using 16 climate change scenarios consisting of bias-corrected data of eight Global Circulation Models (GCM; GFDL-ESM2M, HadGEM2-ES, IPSL-CM5A-LR, MIROC5, MIROC-ESM, MIROC-ESM-CHEM, MRI-CGCM3, and NorESM1-M, used in the Coupled Model Intercomparison phase 5 (CMIP5) and two Representative Concentration Pathways (RCPs; 4.5 and 8.5 W m-2). Soybean growing season used for projections is April to October. All projections assumed irrigated fraction equals to zero. Projections are shown only on agricultural area (cropland plus pasture), in the year 2000. Soybean yield is expressed in tons per hectare. Files description RF_soybean_historical_GRASP_median_1981_2010.nc : random forest projections of soybean yield in Europe for the historical (1981-2010) period using GRASP climate data. This file contains the median yield (in tons per hectare) over 1981-2010. RF_soybean_rcp45_median_2050_2059.nc : random forest projections of soybean yield in Europe for the 2050-2059 time period under RCP4.5. This file contains the median yield (in tons per hectare) over 2050-2059 and the 8 GCMs. RF_soybean_rcp45_median_2090_2099.nc : random forest projections of soybean yield in Europe for the 2090-2099 time period under RCP4.5. This file contains the median yield (in tons per hectare) over 2090-2099 and the 8 GCMs. RF_soybean_rcp85_median_2050_2059.nc : random forest projections of soybean yield in Europe for the 2050-2059 time period under RCP8.5. This file contains the median yield (in tons per hectare) over 2050-2059 and the 8 GCMs. RF_soybean_rcp85_median_2090_2099.nc : random forest projections of soybean yield in Europe for the 2090-2099 time period under RCP8.5. This file contains the median yield (in tons per hectare) over 2090-2099 and the 8 GCMs. References Guilpart N. et al. (2022) Data-driven projections suggest large opportunities to improve Europe's soybean self-sufficiency under climate change, Nature Food. Iizumi T. et al. (2014a) Historical changes in global yields: Major cereal and legume crops from 1982 to 2006. Glob. Ecol. Biogeogr. 23, 346–357. Iizumi T. et al. (2014b). A meteorological forcing data set for global crop modeling: Development, evaluation, and intercomparison. J. Geophys. Res. Atmos. Res. 119, 363–384.