A Stochastic Programming Model for Agricultural Water Resource Allocation Under Climate Uncertainty in Kebbi State

Climate variability represents a growing threat to the sustainability of irrigated agriculture in Nigeria's Northwest zone, where water resource availability for dry-season farming is increasingly uncertain due to shifting precipitation patterns and rising temperatures. This study develops a two-stage stochastic programming model for agricultural water allocation under climate uncertainty in the Kebbi State irrigation command areas along the Rima River. The first-stage decisions involve infrastructure capacity allocation across four irrigation districts, while second-stage decisions optimize water distribution across crop types and districts conditional on realized rainfall and river inflow scenarios. Twelve climate scenarios are generated by coupling historical variability data with projections from the Coordinated Regional Downscaling Experiment for Africa database. Scenario probabilities are derived using a Bayesian updating procedure. The stochastic model is compared against a deterministic benchmark using expected value of perfect information and value of stochastic solution metrics. Results indicate a value of stochastic solution of approximately N1.2 billion annually, demonstrating the significant economic benefit of incorporating uncertainty explicitly into water allocation planning. The optimal first-stage infrastructure investments prioritize rehabilitation of the Dabai irrigation scheme, which yields the highest expected return across scenarios. Robustness analysis confirms solution stability across a wide range of scenario probability specifications. This study fills a methodological gap in water resource operations research applied to Nigerian agricultural irrigation, providing a scalable framework for climate-adaptive water governance. Keywords: stochastic programming, water resource allocation, climate uncertainty, Kebbi State, irrigation