SubSaharan Africa nutrient maps at 250 m resolution

These are results of spatial predictions of content of soil macro and micro-nutrients across Sub-Saharan Africa at 250 m spatial resolution for standard depth interval of 0–30 cm. Model training was run using soil samples from ca. 59,000 locations (a compilation of soil samples from the AfSIS, EthioSIS, One Acre Fund, VitalSigns and legacy soil data) and an extensive stack of remote sensing covariates, as well as landform, lithologic and land cover maps. For generating spatial predictions, an ensemble model from the machine learning algorithms random forest and gradient boosting, as implemented in R packages ranger and xgboost, was used. The results of cross-validation showed that apart from S, P and B, significant models can be produced for most targeted nutrients (R-square between 40–85 %). A limiting factor for mapping nutrients using the existing point data in Africa is high spatial clustering of sampling locations with many countries / land cover and land use groups completely unrepresented.

This work presents a systematic update of maps previously published in Hengl et al. (2015). R code used to generate predictions is available via the github repository.

Please cite as:

Training points used to generate these maps (see also acknowledgments):

  • AfSIS (Africa Soil Information Service) Sentinel Sites: 18,000 soil samples at 9,600 locations (60 sites of 10 by 10 km).
  • EthioSIS (Ethiopia Soil Information Service): 15,000 topsoil samples (0–20 cm) from Ethiopia analyzed by conventional wet chemistry including Mehlich-3;
  • The Africa Soil Profiles database compiled for AfSIS: over 60,000 samples of 18,500 soil profiles collected from on average four depth intervals to on average 125 cm depth;
  • International Fertilizer Development Center (IFDC) projects co-funded by the government of The Netherlands: 3,500 topsoil samples (0–20 cm) for Uganda, Rwanda and Burundi also analyzed using soil spectroscopy;
  • One Acre Fund: some 2,400 topsoil samples (0–20 cm) for Uganda and Kenya;
  • University of California, Davis: some 1,800 topsoil samples (0–20 cm) for Kenya;
  • VitalSigns: 1,374 soil samples from Ghana, Rwanda, Tanzania and Uganda also analyzed using mid-infrared spectroscopy;

All maps, as majority of the SoilGrids products, are available for download under the Open Database License (ODbl) v1.0 and can be downloaded from without restrictions.

Predicted soil macro-nutrients for SubSaharan Africa. All values in ppm.
Predicted soil micro-nutrients for SubSaharan Africa. All values in ppm.

This study has been conducted primarily upon request of the Netherlands Environmental Assessment Agency (PBL). Acknowledgments are due to the various projects and organizations who made soil data collected from various countries available for this study, including projects partially or completely funded by the Bill and Melinda Gates Foundation (BMGF), such as the AfSIS (Africa Soil Information Service) project, which was co-funded by the Alliance for a Green Revolution in Africa (AGRA) and collected and compiled soil test data across SSA, the Vital Signs project with interventions in Ghana, Rwanda, Tanzania and Uganda, and the EthioSIS project, funded primarily by the Ethiopian government and co-funded by the the Bill and Melinda Gates Foundation and the Netherlands government through the CASCAPE project. Also co-funded by the Netherlands government are projects of the International Fertilizer Development Center (IFDC) in collaboration with the governments of Burundi, Rwanda and Uganda. The One Acre Fund made the collection of soil samples possible in Rwanda and Kenya and the University of California, Davis, in Kenya. We are grateful to these organizations for providing soil sample data and for commenting on the first drafts of the manuscript. ISRIC — World Soil Information is a non-profit foundation primarily funded by the Netherlands government.

These are results of spatial predictions based on using Machine Learning algorithms attached to the above-listed paper and hence some errors and artifacts are still possible. We aim at updating these maps regularly i.e. as the new training / point data arrives. See also ISRIC's general disclaimer. These maps are also available via Web Coverage Service at