Detailed soil information for Africa now available
Predicted exchangeable Mg, Ca and K (in cmol+/kg) and Al concentration (ppm) using random forests RK model (PLOS article)
Zoom in on an African village, and see whether the soil is clayey or sandy, or how much nutrients or organic carbon it contains. This is now possible since ISRIC and partners have launched detailed freely accessible soil property maps for Africa. The maps, that can be explored in Google Earth, contain predictions of more than 20 soil properties at six standard depths at 250 meter resolution. The academic article about the maps appeared in PLOS ONE, (25 June).
The new maps, commissioned by the Africa Soil Information Services project (AfSIS), are important for studies on agricultural development, environment and food security. In Africa, significant amounts of soil nutrients are lost every year due to land degradation and soil exhaustion. However, improving land management is impossible without local information on soil properties such as sand-silt-clay content, water-holding capacity, or nutrient content. Unfortunately accurate soil information has been difficult to obtain for governments or research institutes, because existing soil profiles records are scattered over many sources. The aim of AfSIS, an international project funded by the Bill and Melinda Gates Foundation, is to improve this situation, among others by creating up-to-date digital soil property maps at high spatial resolution.
More accurate version
ISRIC had already launched a first digital soil information system of Africa in 2013. Now it has launched a more accurate version of the maps, drawing on 18,000 geo-referenced soil profile records from 41 countries compiled from over 450 data sources, ranging from digital databases, books, reports and articles. Meanwhile, AfSIS teams had analysed newly collected topsoil samples from over 9,500 locations using a combination of soil spectroscopy and laboratory analysis. These old and new soil data records, together with a large number of remote-sensing based images of vegetation, climate and terrain relief, were used to estimate the soil properties for the non-desert parts of Africa.
One mouse click on these maps can tell you for instance that Cameroon, Ghana and West Congo have greater than normal levels of organic carbon in the topsoil. Or that the soil along the northern border of South Africa contains only low amounts of organic carbon. "These new maps are more reliable than the first ones, because we have used more soil profile records, more detailed remote sensing images and more advanced statistical methods”, says ISRIC researcher Tom Hengl.
But there is much room to improve reliability. The team has used ‘only’ 28,500 soil profiles for an area of 18.3 million square kilometers, which is in fact a low sampling intensity. And most of these profiles reflect the period from 1975 to 2005, while some soil properties, such as soil organic content, have probably changed over time. The international mapping team is excited about possibilities for increasing the number of soil profile data used and refining the prediction models. The team looks forward to additional involvement from experts at national soil agencies and regional offices, says Tom Hengl: "This is inherently a joint effort."
The digital soil property maps are one of the most important products of the AfSIS project, which is currently headed by the Columbia University Global Center in Nairobi. Other AfSIS products include up-to-date soil databases, methodology for land degradation surveillance and training of soil scientists. The ultimate goal of such information systems is to empower farmers so that they can use their smartphones to download real-time soil data, along with corresponding fertilization recommendations. A partner of AfSIS has already begun testing an app that can give answers to questions such as “How much maize can I expect in this village?” or “Which nutrients do I need to add to my soils to obtain best yields?”.