Predicted USDA soil great groups at 250 m (probabilities)

Distribution of the USDA soil great groups based on machine learning predictions from global compilation of soil profiles (>350,000 training points). To learn more about soil great groups please refer to the Illustrated Guide to Soil Taxonomy - NRCS - USDA. Processing steps are described in detail here. Antarctica is not included.

To access and visualize maps use: OpenLandMap.org

A back-up copy of all predictions (>65GB) can be downloaded from: http://gofile.me/6J25n/mQ3cHOOMr

If you discover a bug, artifact or inconsistency in the maps, or if you have a question please use some of the following channels:

• Technical issues and questions about the code: https://gitlab.com/openlandmap/global-layers/issues 
• General questions and comments: https://disqus.com/home/forums/landgis/

All files internally compressed using "COMPRESS=DEFLATE" creation option in GDAL. File naming convention:

• sol = theme: soil,
• grtgroup = variable: USDA great group,
• usda.argiustolls = determination method: USDA soil taxonomy class Argiustolls,
• p = probability,
• 250m = spatial resolution / block support: 250 m,
• s0..0cm = vertical reference: soil surface,
• 1950..2017 = time reference: period 1950-2017,
• v0.2 = version number: 0.2,