Description: The ability of a country or region to feed itself in the upcoming decades is a question of importance. The population in West Africa is expected to increase significantly in the next 30 years. The responses of food crops to short term climate change is therefore critical to the population at large and the decision makers tasked with providing food for their people. An ensemble of near term climate projections are used to simulate maize, millet and sorghum in West Africa in the recent historic and near term future. The mean yields are not expected to alter significantly, while there is an increase in inter annual variability. This increase in variability increases the likelihood of crop failures, which are defined as yield negative anomalies beyond one standard deviation during a period of 20 years. The increasing variability increases the frequency and intensity of crop failures across West Africa. The mean return frequency between mild maize crop failures from process based crop models increases from once every 6.8 years to once every 4.5 years. The mean return time frequency for severe crop failures (beyond 1.5 standard deviations) also almost doubles from once every 16.5 years to once every 8.5 years. Two adaptation responses to climate change, the adoption of heat-resistant cultivars and the use of captured rainwater have been investigated using one crop model in an idealised sensitivity test. The generalised adoption of a cultivar resistant to high temperature stress during flowering is shown to be more beneficial than using rainwater harvesting by both increasing yields and the return frequency of crop failures.