World soil information developing from global, continental and national initiatives

Resulting from the GlobalSoilMap initiative and the Globally-integrated Africa Soil Information Service (AfSIS) project, soil property maps of the world were produced in 2014, following the maps of Sub-Saharan Africa produced in 2013. The two maps were fully compliant with the GlobalSoilMap specifications except for the spatial resolution of 1km. The primary soil data used as input for mapping were from the Africa Soil Profiles (AfSP) database and an early version of the World Soil Information Service (WoSIS) database, both compiled mainly from existing sources in collaboration with various partners in Africa and the world. The maps have more recently been updated to a resolution of 250m upon availability of additional data. These initiatives, using a top-down approach due to lack of budget to actively involve soil institutes worldwide, showed that state of the art baseline products can be developed cost-efficiently based on legacy soil data and can be updated when additional, possibly newly collected, data become available. As a result, several bottom-up initiatives in Africa have started similar work aiming to add value to national soil data holdings of mainly analogue and fragmented nature. In collaboration with ISRIC and facilitated by the global soil information facilities (GSIF), the Soil Research Institute in Ghana compiled and georeferenced data, collected during decades of soil studies applying standard procedures, for over 1000 soil profiles adding to the 400 in the AfSP. A similar successful initiative with the University of Dschang and the International Institute of Tropical Agriculture (IITA) in Cameroon compiled data for 1300 soil profiles, including 500 from the AfSP of which many could be georeferenced more accurately. Each of the countries developed a baseline version of the national SoilGrids, with and without using the global SoilGrids maps as covariate. These collaborative studies are good examples how national and global initiatives can strengthen each other as a feasible way forward to compile sufficient data, at low costs, to produce maps that are nationally and locally accurate, detailed and relevant as well as globally coherent, harmonised and complete. Such hybrid approach requires training of national experts and sufficient computational capacity to locally produce maps at high spatial resolution, i.e. 100m, and also requires global standards and mechanisms for sharing and merging of data and results.