Examining Variations in Sensitivity of Cereal Crop Yield to Climate Change Variables across the Regions in Northern Ghana using Multilevel and Bayesian Multilevel Modeling

Abstract

The study applied both the multilevel and the Bayesian multilevel model approaches to investigate variations in the effects of climate variables on cereal crop yield in Northern Ghana with respect to the region of cultivation and year (time), and to compare the performance of the two models. Thirty-one years of data points on some climate variables and the annual yield of some selected cereal crops from the Meteorological Agency and the Ministry of Food and Agriculture of Ghana, respectively, were used. Results indicated significant variations in climate change impact across the regions and years of cultivation. Also, the study showed that the significant effect of humidity and sunshine (  and  respectively) on crop yield varies from one region to another, with humidity having the most variation. The study further revealed that the Bayesian Multilevel model performed better in its model scores and predictive ability. It concluded that there are variations in the impact of climate change on cereal crop yield in the regions in Northern Ghana and recommends that the climate characteristics of these regions should be taken into account in predicting future yield and adopting mitigation strategies.