Pulse Brain · Growing Health Evidence Index
Tier 3 — Observational / field trialPeer-reviewed

African soil properties and nutrients mapped at 30 m spatial resolution using two-scale ensemble machine learning

Tomislav Hengl, Matt Miller, Josip Križan, Keith Shepherd, Andrew Sila, Milan Kilibarda, Ognjen Antonijević, Luka Glušica, Achim Dobermann, Stephan M. Haefele, S. P. McGrath, Gifty Acquah, Jamie Collinson, Leandro Parente, Mohammadreza Sheykhmousa, Kazuki Saito, Jean‐Martial Johnson, Jordan Chamberlin, Francis B.T. Silatsa, Martin Yemefack, John Wendt, R.A. MacMillan, Ichsani Wheeler, Jonathan H. Crouch

Scientific Reports · 2021

Read source ↗ All evidence

Summary

This study presents a 30 m resolution Soil Information System for the African continent, the most comprehensive continental-scale soil mapping effort to date, synthesising approximately 150,000 soil samples with satellite and climatic data through ensemble machine learning. The authors produced spatially explicit predictions for 19 soil properties at three depths, with varying accuracy levels across properties; soil pH was most predictable (CCC = 0.900) whilst extractable phosphorus and sulphur presented greater challenges. Climatic variables—particularly SM2RAIN precipitation, bioclimatic data and land surface temperature—emerged as the dominant predictors of soil chemical properties at continental scale, alongside Sentinel-2 and Landsat spectral bands.

UK applicability

The methodology and machine learning framework may be transferable to UK soil mapping initiatives, though the African focus and climatic drivers mean direct application of the African predictions is limited. The study's approach to integrating diverse soil sampling networks and Earth Observation data could inform UK soil monitoring strategy, particularly for areas where direct sampling remains sparse.

Key measures

Soil pH, organic carbon, total nitrogen, total carbon, effective Cation Exchange Capacity (eCEC), extractable phosphorus, potassium, calcium, magnesium, sulphur, sodium, iron, zinc, silt, clay, sand, stone content, bulk density, and depth to bedrock. Predictions evaluated using fivefold spatial cross-validation with concordance correlation coefficient (CCC).

Outcomes reported

The study produced 30 m resolution predictions of 19 soil properties and classes across the African continent at three soil depths (0, 20 and 50 cm), based on approximately 150,000 soil samples and Earth Observation data. Prediction accuracy varied across properties, with soil pH achieving high accuracy (CCC = 0.900) whilst extractable phosphorus and sulphur were less predictable.

Theme
Measurement & metrics
Subject
Soil health assessment & monitoring
Study type
Research
Study design
Field trial / Spatial modelling study
Source type
Peer-reviewed study
Status
Published
Geography
Africa
System type
Other
DOI
10.1038/s41598-021-85639-y
Catalogue ID
BFmovi1txm-5gatpd

Topic tags

Pulse AI · ask about this record

Dig deeper with Pulse AI.

Pulse AI has read the whole catalogue. Ask about this record, its theme, or how the findings apply to UK farming and policy — every answer cites the underlying studies.