Precision Agriculture in Nigeria: Geospatial Decision Support System Development for Smallholder Rice and Maize Production Optimisation in the North Central Zone

Smallholder agriculture in Nigeria's North Central zone accounts for the majority of national rice and maize production, yet crop yield gaps are substantial and input use efficiency is low, suggesting that spatially targeted precision agriculture interventions could significantly improve productivity and reduce environmental footprint. This dissertation develops an original geospatial decision support system for smallholder rice and maize production optimisation in the North Central zone, covering Benue, Kogi, Niger, Nassarawa, Kwara, and Plateau States, and makes original contributions to precision agriculture systems design for resource-constrained, fragmented farming landscape contexts. An integrated Earth Observation data architecture combining Sentinel-2 for crop mapping and growth monitoring, Sentinel-1 for soil moisture estimation, WorldView-3 for sub-field variability characterisation, and GPS field scouting data is developed and operationalised across 840 trial plots in collaboration with six state Agricultural Development Programmes. Original machine learning models for in-season yield prediction are developed and validated, achieving RMSE of 0.43 and 0.38 tonnes per hectare for rice and maize respectively at forty days before harvest. A smartphone-accessible GDSS application is designed through participatory prototype testing with 120 smallholder farmers, generating high usability ratings and measurable agronomic decision quality improvements in a randomised controlled trial. The dissertation introduces the Smallholder Precision Agriculture Readiness Framework as an original conceptual tool for assessing the institutional and infrastructure preconditions for GDSS adoption in developing country contexts. Keywords: precision agriculture, geospatial decision support, smallholder farming, North Central Nigeria, yield prediction.