General information The data is related to the manuscript "Decipher soil organic carbon dynamics and driving forces across China using machine learning" by Huiwen Li, Yiping Wu, Shuguang Liu, Jingfeng Xiao, Wenzhi Zhao, Ji Chen, Georgii Alexandrov, and Yue Cao. Corresponding author: Yiping Wu (yipingwu@xjtu.edu.cn) ================================ Methods are in detail described in the paper. ================================ This dataset contains all primary data that support the findings of this study, which are listed as below: 1, SOC and uncertainty maps estimated by the random forest models and the 9-fold cross validation models, respectively. 2, The relative contribution maps of the three time-variant driver categories (climate, vegetation, and environment). 3, The dominant driver maps for SOC dynamics in the two soil layers. 4, The changes in cropland SOC20 during the 1980s and 2000-2014. 5, The forest canopy height map in China. 6, The changes in SOC between different periods. 7, Other SOC density maps used to evaluate the spatial pattern of SOC. ================================ Data description 1, The SOC and uncertainty maps were synthesized in the file entitled "SOC_density_maps_in_China.nc4". This NC file contains two major variables: SOC and UC. The variable SOC contains the Chinese SOC density maps in the top 20 cm and top 100 cm soil layers during the 1980s, 2000s, 2010s, and 2000—2014 (eight maps in total). The variable UC contains eight uncertainty maps associated with each SOC map. This file can be easily visualized and analyzed by free NC file reading platforms such as Panoply, R, and Python (a netCDF4 module is recommend). 2, The relative contribution maps of the three time-variant driver categories for SOC20 and SOC100 are "RelCon_SOC20_Cli_Veg_Env_2000_2014.tif" and "RelCon_SOC100_Cli_Veg_Env_2000_2014.tif". There are three bands in each file: band1, band2, and band3 that represent the relative contributions of climate, vegetation, and environment driver category, respectively. 3, "Dominant_Driver_for_SOC20.tif" and "Dominant_Driver_for_SOC100.tif" present the dominant driver maps for SOC20 and SOC100, respectively. A single driver means that its relative contribution is larger than 0.6. 4, "Crop_SOC20_changes_between_1980s_2000-2014.tif" shows the changes in cropland SOC20 between the 1980s and 2000-2014 with a unit of kg C m^2. 5, "CHN_Forest_Canopy_Height.tif" is the forest canopy height map in China with a unit of m. 6, "SOC20_100_changes_between_1980s_2000s.tif" presents the change map of SOC20 and SOC100 between the 1980s and 2000s, see the Fig. 2 in the paper for the color legend. "SOC20_100_changes_between_1980s_2010s.tif" is the change map of SOC20 and SOC100 between the 1980s and 2010s. "SOC20_100_changes_between_1980s_2000-2014.tif" shows the change map of SOC20 and SOC100 between the 1980s and 2000-2014. "SOC20_100_changes_between_2000s_2010s.tif" presents the change map of SOC20 and SOC100 between the 2000s and 2010s. 7, "GSOCmap_CHN_SOC30_kgC_m2.tif" is the GSOCmap SOC density map in the top 30 cm soil layer with a unit of kg C m^2. "SoilGrid_CHN_SOC15_kg_m2.tif" is the SoilGrid SOC density map in the top 15 cm soil layer with a unit of kg C m^2. "SoilGrid_CHN_SOC100_kgC_m2.tif" is the SoilGrid SOC density map in the first meter soil layer with a unit of kg C m^2.