Exploring the Spatio-Temporal Dynamics of the Association between Climate Variables and Post-Harvest Loss of Cassava across Benue State, Nigeria

Abstract

Nigeria is reckoned as one of the world’s largest producers of cassava and contributes around 60 million tons annually. Despite its significance, post-harvest losses remain high within communities. This study investigated the spatio-temporal dynamics and influencing factors of PHLs across Benue State, Nigeria. By the integration of classical spatial statistics with Bayesian hierarchical modeling via the Integrated Nested Laplace Approximation (INLA), the study analyzed how cassava PHL evolved over time across locations. Results from the study showed a consistent weak negative spatial autocorrelation (Moran’s I ≈ −0.31, p ≈ 0.055–0.063), indicating no strong clustering of PHL values but a mild dispersion pattern. The Local Moran’s I (LISA) analysis revealed no statistically significant hotspots or outliers. It was observed that mean PHL values increased modestly from 29,968 in 2021 to 31,917 in 2024, and LGAs such as Ado, Ukum, and Ogbadibo recorded the highest losses, while Katsina-Ala, Obi, and Ado exhibited the strongest increasing trends. Temperature was identified as a small but credible positive predictor of PHL. Strong spatial random effects as highlighted in Ado (0.629), Ukum (0.541), and Kwande (−1.278) showed areas with significantly higher and lower losses, respectively, after controlling for climatic variables. By providing a comprehensive spatio-temporal analysis of cassava PHLs across Benue State, our study fills a critical knowledge gap in agriculture and spatial statistics. The findings from this study provide location-specific intelligence and suggest that interventions should prioritize high-loss LGAs while adopting a mix of targeted and statewide strategies.