Global maps at 1 km spatial resolution of the predicted soil types (0–100% probabilities) at 1 km resolution based on the WRB 2022 (World Reference Base the international standard for soil classification) classification system. The training data comes from the following 3 main sources: WOSIS points available via: https://www.isric.org/explore/wosis; HWSD v2 (random draw of cca 20,000 points): https://iiasa.ac.at/models-tools-data/hwsd; Other national datasets / data from publications and projects. Predictions are based on using Rando Forest algorithm as implemented in the randomForestSRC package with cca 190 covariate layers representing soil forming factors (CHELSA Climate, Global Lithological DB GLiM, MODIS EVI and LST long-term derivatives, Digital Terrain model parameters and similar). All TIF files are provided as COGs, which means that you can open them directly in QGIS or similar. Publication explaining all modeling steps is pending. Update of the predictions takes about 4–5 hrs and will be regularly run provided that new training points are available. Disclaimer: These are initial results with limited accuracy and possible issues with quality of training points, location errors and harmonization issues. Use at own risk. Note: original list of soil types have been subset to classes that appear at least 10 times and at least in 2 countries. If you notice an error or artifact please report via the Github repository. Help us improve this dataset by contributing training points.

{"references": ["Batjes, N. H., Ribeiro, E., & Van Oostrum, A. (2020). Standardised soil profile data to support global mapping and modelling (WoSIS snapshot 2019). Earth System Science Data, 12(1), 299-320. https://doi.org/10.5194/essd-12-299-2020", "FAO & IIASA. 2023. Harmonized World Soil Database version 2.0. Rome and Laxenburg. https://doi.org/10.4060/cc3823en", "Hengl, T., Mendes de Jesus, J., Heuvelink, G. B., Ruiperez Gonzalez, M., Kilibarda, M., Blagoti\u0107, A., ... & Kempen, B. (2017). SoilGrids250m: Global gridded soil information based on machine learning. PLoS one, 12(2), e0169748.", "Krasilnikov, P., J. Ib\u00e1\u00f1ez Mart\u00ed, R.W. Arnold, and S. Shoba. (2009). A Handbook of Soil Terminology, Correlation and Classification. Edited by P. Krasilnikov, J. Ib\u00e1\u00f1ez Mart\u00ed, R.W. Arnold, and S. Shoba. London: Earthscan. 25 https://www.researchgate.net/profile/Juan_Ibanez3/publication/285586468_Handbook_of_Soil_Terminology_Corre lation_and_Classificati/links/5660452908ae4988a7bf10e4.pdf"]}