SoilFER programme · Online training platform

Map the soils that feed the world, from your browser.

Learn digital soil mapping with the SoilFER Training Manual. Practise in your own SoilFER Learn lab with JupyterLab, RStudio and QGIS. Then make soil maps of your own country, with their uncertainty. All you need is a web browser.

Sign in to your lab Explore the course Try the workflow ↓
  • 6 modules
  • JupyterLab · RStudio · QGIS
  • Maps by depth, in 2D and 3D
  • Free & open source
From soil profile to soil map: observe a profile, sample many sites across the landscape, measure the samples in the laboratory, model with covariates, and predict a soil property map with its uncertainty.ABCRHorizon B · 30–60 cmsample KS-017 → labSampling site KS-003Sampling site KS-010Sampling site KS-017Sampling site KS-024Sampling site KS-031Sampling site KS-038Sampling site KS-045Sampling site KS-052Sampling site KS-059Sampling site KS-06610 sites · 3 depths · stratifiedLaboratory resultsSiteDepthSOC %pHP mg/kgKS-0010–301.826.412KS-00130–600.946.97KS-0170–302.316.115KS-0420–301.157.39KS-0580–302.675.918MIR spectrum · KS-017Covariates + soil data → modelElevationClimateLand coverQuantile RegressionForestPredicted SOC 3.1 %Predicted SOC 3.3 %Predicted SOC 3.5 %Predicted SOC 3.2 %Predicted SOC 3.2 %Predicted SOC 3.4 %Predicted SOC 3.3 %Predicted SOC 3.1 %Predicted SOC 3.1 %Predicted SOC 3.0 %Predicted SOC 2.9 %Predicted SOC 2.7 %Predicted SOC 3.1 %Predicted SOC 2.8 %Predicted SOC 2.5 %Predicted SOC 2.5 %Predicted SOC 3.0 %Predicted SOC 2.5 %Predicted SOC 2.6 %Predicted SOC 2.4 %Predicted SOC 2.7 %Predicted SOC 2.5 %Predicted SOC 2.5 %Predicted SOC 2.6 %Predicted SOC 2.3 %Predicted SOC 2.4 %Predicted SOC 2.2 %Predicted SOC 2.0 %Predicted SOC 2.4 %Predicted SOC 2.4 %Predicted SOC 2.3 %Predicted SOC 2.1 %Predicted SOC 2.5 %Predicted SOC 2.4 %Predicted SOC 2.3 %Predicted SOC 1.9 %Predicted SOC 1.8 %Predicted SOC 1.6 %Predicted SOC 1.1 %Predicted SOC 1.0 %Predicted SOC 1.6 %Predicted SOC 1.3 %Predicted SOC 1.2 %Predicted SOC 1.1 %Predicted SOC 1.4 %Predicted SOC 1.2 %Predicted SOC 1.1 %Predicted SOC 1.3 %Predicted SOC 2.9 %Predicted SOC 2.7 %Predicted SOC 2.8 %Predicted SOC 2.4 %Predicted SOC 2.5 %Predicted SOC 2.3 %Predicted SOC 2.3 %Predicted SOC 2.2 %Predicted SOC 1.9 %Predicted SOC 1.7 %Predicted SOC 1.6 %Predicted SOC 1.7 %Predicted SOC 2.3 %Predicted SOC 2.4 %Predicted SOC 2.4 %Predicted SOC 2.5 %Predicted SOC 2.1 %Predicted SOC 2.2 %Predicted SOC 2.2 %Predicted SOC 2.4 %Predicted SOC 1.7 %Predicted SOC 1.9 %Predicted SOC 2.1 %Predicted SOC 2.2 %Predicted SOC 2.4 %Predicted SOC 2.4 %Predicted SOC 2.1 %Predicted SOC 1.7 %Predicted SOC 2.3 %Predicted SOC 2.2 %Predicted SOC 1.9 %Predicted SOC 1.4 %Predicted SOC 1.9 %Predicted SOC 1.7 %Predicted SOC 1.5 %Predicted SOC 1.3 %Predicted SOC 1.5 %Predicted SOC 1.1 %Predicted SOC 1.2 %Predicted SOC 1.4 %Predicted SOC 1.2 %Predicted SOC 1.2 %Predicted SOC 1.1 %Predicted SOC 1.4 %Predicted SOC 1.2 %Predicted SOC 1.3 %Predicted SOC 1.3 %Predicted SOC 1.5 %Predicted SOC 1.4 %Predicted SOC 1.4 %Predicted SOC 1.4 %Predicted SOC 1.2 %Predicted SOC 1.1 %Predicted SOC 1.0 %Predicted SOC 1.0 %Predicted SOC 1.0 %Predicted SOC 0.8 %Predicted SOC 0.7 %Predicted SOC 0.8 %Predicted SOC 1.0 %Predicted SOC 1.4 %Predicted SOC 1.6 %Predicted SOC 1.8 %Predicted SOC 1.8 %Predicted SOC 1.5 %Predicted SOC 1.6 %Predicted SOC 1.7 %Predicted SOC 1.6 %Predicted SOC 1.2 %Predicted SOC 1.4 %Predicted SOC 1.3 %Predicted SOC 1.4 %Predicted SOC 1.7 %Predicted SOC 1.3 %Predicted SOC 1.0 %Predicted SOC 1.0 %Predicted SOC 1.3 %Predicted SOC 1.1 %Predicted SOC 0.9 %Predicted SOC 1.1 %Predicted SOC 1.2 %Predicted SOC 1.0 %Predicted SOC 1.1 %Predicted SOC 1.3 %Predicted SOC 1.1 %Predicted SOC 1.3 %Predicted SOC 1.6 %Predicted SOC 1.7 %Predicted SOC 1.1 %Predicted SOC 1.3 %Predicted SOC 1.6 %Predicted SOC 1.6 %Predicted SOC 1.5 %Predicted SOC 1.7 %Predicted SOC 2.0 %Predicted SOC 2.0 %Predicted SOC 0.8 %Predicted SOC 0.9 %Predicted SOC 0.9 %Predicted SOC 1.1 %Predicted SOC 1.0 %Predicted SOC 1.1 %Predicted SOC 1.1 %Predicted SOC 1.3 %Predicted SOC 1.0 %Predicted SOC 1.1 %Predicted SOC 1.2 %Predicted SOC 1.2 %Predicted SOC 1.3 %Predicted SOC 1.4 %Predicted SOC 1.4 %Predicted SOC 1.2 %Predicted SOC 1.3 %Predicted SOC 1.4 %Predicted SOC 1.3 %Predicted SOC 1.3 %Predicted SOC 1.2 %Predicted SOC 1.2 %Predicted SOC 1.1 %Predicted SOC 1.2 %Predicted SOC 1.0 %Predicted SOC 1.3 %Predicted SOC 1.4 %Predicted SOC 1.6 %Predicted SOC 1.2 %Predicted SOC 1.3 %Predicted SOC 1.8 %Predicted SOC 2.2 %Predicted SOC 1.4 %Predicted SOC 1.6 %Predicted SOC 2.1 %Predicted SOC 2.5 %Predicted SOC 2.0 %Predicted SOC 2.1 %Predicted SOC 2.3 %Predicted SOC 2.3 %Predicted SOC 2.5 %Predicted SOC 2.6 %Predicted SOC 2.6 %Predicted SOC 2.4 %Predicted SOC 2.7 %Predicted SOC 2.7 %Predicted SOC 2.5 %Predicted SOC 2.3 %Predicted SOC 1.1 %Predicted SOC 1.2 %Predicted SOC 1.3 %Predicted SOC 1.4 %Predicted SOC 1.2 %Predicted SOC 1.4 %Predicted SOC 1.4 %Predicted SOC 1.1 %Predicted SOC 1.3 %Predicted SOC 1.3 %Predicted SOC 1.0 %Predicted SOC 0.8 %Predicted SOC 1.2 %Predicted SOC 1.0 %Predicted SOC 1.0 %Predicted SOC 1.2 %Predicted SOC 0.9 %Predicted SOC 1.0 %Predicted SOC 0.9 %Predicted SOC 1.0 %Predicted SOC 0.9 %Predicted SOC 0.8 %Predicted SOC 0.9 %Predicted SOC 1.3 %Predicted SOC 1.3 %Predicted SOC 1.8 %Predicted SOC 2.3 %Predicted SOC 2.6 %Predicted SOC 1.4 %Predicted SOC 1.8 %Predicted SOC 2.4 %Predicted SOC 2.7 %Predicted SOC 1.6 %Predicted SOC 2.1 %Predicted SOC 2.4 %Predicted SOC 2.4 %Predicted SOC 2.8 %Predicted SOC 2.7 %Predicted SOC 2.5 %Predicted SOC 2.5 %Predicted SOC 2.6 %Predicted SOC 2.7 %Predicted SOC 2.6 %Predicted SOC 2.4 %Predicted SOC 2.5 %Predicted SOC 2.4 %Predicted SOC 2.3 %Predicted SOC 2.4 %SOC 0–30 cmlowhighuncertaintyObserveSampleMeasureModelMap
With the financial
support of U.S. Department of State sealUnited StatesDepartment of State From the People of JapanFrom the People of JapanMinistry of Foreign Affairs Implemented by FAO through the Global Soil Partnership
What you can do here

Learn it, practise it, map your own country

Everything opens in your web browser. Your trainer gives you a username and password for the training.

Practise in your lab

Your own page picks up where you left off: notebooks, RStudio, QGIS and the course data, with handouts from your trainer and a button to ask for help.

How to get access →

Bring your own data

Upload a file of soil profiles in almost any layout. DSM Studio reads it, finds the coordinates and depths, and shows you the points on a map.

About DSM Studio ↓

Make soil maps

Get a map for each depth with its uncertainty and the area where the model can be trusted. Covariates can come from Google Earth Engine.

About DSM Studio ↓

Look at any map

Drop a GeoTIFF or a results zip on the Map viewer. Read values, draw a profile, see it on the terrain in 3D. Your files stay on your computer.

Open the Map viewer →
Try it · digital soil mapping in five steps

From a soil pit to a soil map

Dig a profile, choose where to sample, read the laboratory data, train a model and draw the map with its uncertainty. Every step reacts to what you do: this is the workflow you learn in the course.

Curriculum

Digital Soil Mapping track

Six modules, from your first line of R to national map products.

00

Welcome & your workspace

Sign in, open JupyterLab, RStudio and QGIS, and upload your own data.

30 min · Beginner 01

R, spatial data & soil data preparation

Clean and harmonise KSSL soil data and build the covariate stack.

4 sessions · Beginner 02

Soil sampling design

Three-stage sampling with covariate space coverage and Newhall climate regimes.

3 sessions · Intermediate 03

Digital soil mapping

Quantile Regression Forest, Boruta, prediction with uncertainty and the area of applicability.

2 sessions · Intermediate 04

Soil spectroscopy

Predict soil properties from MIR spectra and use them for mapping.

4 sessions · Advanced 05

Data standards & sharing

Metadata, standards and services for national soil information systems.

1 session · Intermediate
DSM Studio

Soil maps of your country, without writing code

  • Your own data: an Excel or CSV file of soil profiles, in long or wide layout, in many languages and coordinate systems
  • Covariates ready: the course stack, your own rasters, or layers fetched from Google Earth Engine for your area
  • Maps you can trust: several models compared, a map for each depth, prediction intervals, cross-validation and the area of applicability
  • Inside your country: maps cut to the UN boundary or to your own boundary file, ready to view in 2D or 3D and to download

DSM Studio is opened to a training when the course reaches mapping. It runs the same methods as Module 3.

DSM Studio results: a soil pH map of Kansas with its model quality, and an option to grey out where the model extrapolates
For trainers

Run a training from one page

Trainers see every participant of their training at a glance and help them while they work.

Trainings

Create accounts and print credential slips, see who is in their lab, and set the memory for each lab while the training runs.

Live help

Participants raise a hand from their lab. Reply, join their notebook when they allow it, or send a notice and handouts to everyone.

The SoilFER Learn lab

A complete soil data science workstation in your browser

  • R 4.6 and Python with every package used in the manual (terra, sf, caret, ranger, prospectr, clhs, aqp…) and Google Earth Engine clients
  • Course data in place: KSSL soil data, covariate rasters and MIR spectra, exactly where the scripts expect them
  • Any extra package in seconds: install.packages() gets ready-made binaries, %pip install for Python
  • Upload your own data and keep your files between sessions; look at inputs and results in QGIS or the Map viewer
  • Help when you are stuck: ask a trainer from your notebook, in RStudio or on your lab page
How to get access
The SoilFER programme

Soil information for resilient agri-food systems

SoilFER builds national soil information systems, soil nutrient and soil property maps and decision support for soil fertility in Central America and Africa. National experts learn here to produce those maps with the methods of the Global Soil Partnership.

SoilFER · GCP/GLO/1127/USA

Soil mapping for resilient agri-food systems in Central America and sub-Saharan Africa

National soil information systems, soil nutrient and nutrient budget maps, decision support systems and sustainable soil management. Funded by the United States Department of State.

GuatemalaHondurasGhanaKenyaZambia
SoilFER-VACS · GCP/GLO/1242/JPN

Enhancing integrated soil-crop management for sustainable food systems in Africa

Part of the Vision for Adapted Crops and Soils (VACS): soil information, crop suitability and integrated soil-crop advisory. Funded by the People of Japan.

MozambiqueTunisia
7beneficiary countries, and regional trainings in sub-Saharan Africa and Central America
6modules, from soil data preparation to data sharing
100%open source: course material, code and this platform

Ready for your training event

Trainers set up a training in a few minutes: accounts and credential slips, a private lab for every participant with the course data and tools, and live help during the sessions. Only a web browser is needed.

About the platform Training resources on GitHub