On the 22nd October, iSDAsoil was launched — a high spatial resolution soil information service for Africa mapped at 30 m spatial resolution for two standard depth intervals (0–20 cm and 20–50 cm). EnvirometriX, jointly with GiLAB and MultiOne, has helped produced predictions and build back-end and front-end solution for the system. As the main method to generate predictions, we used a 2-scale Ensemble Machine Learning, a combination of Sentinel-2, Landsat, DEM derivatives and coarse resolution covariates (MODIS, PROBA-V), and over 100,000 soil sampling training points. Interested in Ensemble Machine Learning? Please check our landmap package for automated interpolation / spatial prediction using Ensemble techniques. Most of the processing steps are explained in:
Hengl, T., MacMillan, R.A., (2019). Predictive Soil Mapping with R. OpenGeoHub foundation, Wageningen, the Netherlands, 370 pages, www.soilmapper.org, ISBN: 978-0-359-30635-0.
The iSDAsoil maps are available under an Open Data license CC-BY 4.0 via zenodo.org. iSDAsoil data can also be accessed through a REST API. To learn more about iSDAsoil, visit the Technical Information and FAQs pages.