Modelling Crop Yield Response to Soil Carbon Inputs Across Nigerian Agro-Ecological Zones: A Bayesian Meta-Analysis

This study develops Bayesian meta-analysis models to quantify the yield response of major Nigerian crops to soil organic carbon inputs across agro-ecological zones, providing probabilistic estimates of carbon addition effects that incorporate parameter uncertainty. The relationship between soil organic carbon and crop yield is fundamental to agronomic management decisions, yet existing estimates from individual studies show high variability that makes them unreliable guides for extension recommendations. Bayesian meta-analysis offers a principled framework for synthesising evidence across studies while appropriately propagating uncertainty. This study systematically reviews Nigerian agronomic literature from 1990 to 2023, extracting yield response data from studies that simultaneously report soil organic carbon content and crop grain yield under varied organic matter management. A hierarchical Bayesian model is developed that estimates population-level yield-SOC relationships while accounting for between-study heterogeneity due to soil type, crop type, and management context. Zone-specific yield response functions are derived from the posterior distributions. Findings reveal a significant positive but non-linear yield response to SOC across all zones, with diminishing returns above 2.5 percent SOC in northern zones and above 3.5 percent in southern forest zones. The benefit of SOC increase is largest in the Sudan Savanna zone, where each 0.1 percent SOC increase is associated with a 4.2 percent yield gain. Zone-specific 95 percent credible intervals are substantially narrower than individual study uncertainty ranges. The study contributes original Bayesian SOC-yield models for Nigeria and recommends their use in precision soil fertility advisory services.

Keywords: Bayesian meta-analysis, soil organic carbon, crop yield, agro-ecological zones, Nigeria.