| Sumario: | Crop productivity remains perennially lower in West Africa, and this significantly attributed to stale blanket fertilizer recommendations. This is also exacerbated by limited national extension services and widespread soil degradation. In response IITA and other technical partners, with funding from the World Bank are generating validated crop- and soil-specific nutrient combinations in support of the development of locally relevant Integrated Soil fertility Management (ISFM) recommendations through IPI 1.4. This builds on the digital soil mapping information generated through IPI 1.3. Implementation of IPI 1.4 is undertaken in close collaboration with national agricultural research partners. This is being undertaken with focus on Nigeria, Ghana, Togo, Liberia, Sierra Leone, Mali and Senegal in geographies where strategic scaling projects such as FSRP and Soil values are working in. The focus crops were rice, maize and cassava, based on country prioritisation exercises. The machine learning component of the AgWise fertilizer decision support framework was trained using legacy agronomic data from public repositories. The model performance was deemed to be acceptable at R2 of >0.5 and highly reliable at >0.75. This generated nutrient predictions which have however not undergone field validation (V0). These were validated using crowd sourced peer reviewed literature data. Field data were then used to test against the validated V0 models with a focus on maize Nigeria as these are the only Nutrient Omission Trial (NOT) results available
|
|---|