Development of site-specific fertilizer nutrient requirements for enhanced soil health management in West Africa and the Sahel

Agricultural productivity in West Africa continues to fall short of its potential, largely because fertilizer use is often insufficient, poorly targeted, or based on blanket and regional based recommendations that do not account for local soil and crop conditions. This has led to persistent yield gaps, inefficient nutrient use, and ongoing degradation of soils. In response, the IPI 1.4 deliverables were designed to generate more appropriate, site-specific fertilizer requirements for key crop–soil systems in priority areas, while also building a framework that can be scaled over time.

To achieve this, this initiative integrated curated legacy crop response to nutrient application data with newly generated nutrient omission trials, along with geospatial soil and climate information, using machine-learning approaches within the AgWise decision-support framework. This approach made it possible to produce spatially explicit V1 nutrient requirement maps for priority crops across the target countries. The work was carried out in close collaboration with national ministries, research institutions, and technical teams, ensuring both scientific quality and strong national ownership.

During the reporting period, major outputs included the development of Version 1 (V1) nutrient recommendations for key cereal systems, particularly maize in Nigeria and Ghana (covering about 5.1 million hectares of target watersheds across these countries belonging to Regional Hub’s wave-1 countries) and rice in Nigeria, Ghana, and Liberia (covering approximately 5.2 million hectares of target watersheds across these countries belonging to the same Regional Hub’s wave-1 countries). These outputs met the minimum accuracy threshold, with a model coefficient of determination (R²) of at least 0.7. However, the V1 outputs for Togo did not meet the R² threshold of 0.7 and were therefore not achieved or reported. In addition, the initiative piloted Virtual Agronomist as a scalable digital delivery tool and developed Version Zero of the agronomic gain assessment toolkit to standardize the measurement of productivity, resource-use efficiency, soil health, and resilience. These efforts are significant as they link recommendation development with both delivery mechanisms and evidence generation.

Going forward, the project will shift its focus to validation and refinement. Field trials are planned in Liberia, Ghana, and Nigeria to test and improve the V1 recommendations, leading to finalized V2 nutrient packages for operational use and eventual scaling across wider areas in West Africa and the Sahel. In Sierra Leone and Togo, the priority will be to conduct additional nutrient omission trials to strengthen the underlying data before moving to large-scale validation, given the current limitations in available data. Over time, the validated recommendations will increasingly be delivered through digital platforms such AKILIMO integrated through Soil Information System (SIS), alongside broader extension and scaling efforts through bundling of services and farmer-led experimentations.