Fuzzy measure-based multicriteria land assessment for rainfed maize in West Africa for the current and a range of plausible future climates

Abstract

Crop production in West Africa is largely under rainfed conditions, making the system vulnerable to the impacts of climate change. However, the impact of future climate on the geographic range of many crops in West Africa is still uncertain. This is exacerbated by considerable uncertainty in projecting future West African climate by global circulation models (GCMs). In this paper, the current land suitability for rainfed maize (Zea mays L.) production in West Africa and how it would change in response to a range of future climate changes were assessed. A novel non-additive approach involving the integration of a fuzzy measure modelling procedure and the Choquet integral to deal with non-linear relationships among criteria was introduced. The supervised and unsupervised variants of the approach and the Ordered Weighted Averaging (OWA) aggregators were evaluated and compared using data on the current climate, soil characteristics, and actual maize yield estimates at some locations across the region. The supervised fuzzy measure modelling outperformed the unsupervised variant and the OWA operators, demonstrating the importance of considering interaction among criteria and incorporating human reasoning in multicriteria analysis. Consequently, the supervised modelling approach was used to estimate suitability scores for rainfed maize for the current and a range of plausible future climates which were generated from unique combinations of temperature and rainfall percentile values (5th, 50th and 95th) from 25 and 32 GCMs projections under the RCPs 2.6 and 8.5 scenarios, respectively. Under the current climate in West Africa, the combined suitable/highly suitable areas for rainfed maize production are in the humid savanna zones. In contrast, the forest region to the south exhibited moderate suitability with the semi-arid savanna zones being marginally suitable. Relative to the current climate in which 91.3% of West Africa is at least moderately suitable for maize production, suitability for maize will change in at least 43% of West Africa in response to future climate change. Across all scenarios and years, future suitability in the semi-arid region of the Sudano-Sahelian savannas shows strong sensitivity to uncertainty in rainfall projections. However, regardless of the trajectory of future rainfall projection, the 95th percentile temperature projection will decrease suitability in about 15% of the West African region by 2080 under the RCP8.5 scenario. Because of the incorporation of uncertainties of future climate projections in this analysis, the results from this study can be used in the development of flexible climate adaptation strategies.