Estimation of the response of rainfed crops to heat stress and water stress must adequately account for the uncertainty in climatic and non-climatic factors that affect impact assessments. The objective of this research was to narrow the range of values characterizing the limits within which estimates are expected to fall in the diagnostics of agroclimatic risks. Assessments were made by analyzing historical observations and evaluating the influence of heat stress and rainfall variability on crop water demand, biomass and grain yields of short-cycle cultivars of pearl millet and maize. We used a wide range of consistent and practical sets of crop model ensemble analyses (based on crop management practices such as plant density, fertilization levels, early/late sowing dates and soil types) and climate model ensembles from 2 climate change hypothesis (A1b and RCP8.5) over the West African Sudan-Sahel. Recent rainfall developments show that hazardous intra-seasonal rainfall distribution affects crop productivity, with increased frequency and intensity of daily rainfall, false start and early cessation of the rainy season and decreasing diurnal temperature range. In 2011−2050 perspectives, relative to the 1981−2010 baseline, a slight increase in temperature (i.e. +0.6 to +0.8°C) combined with a stationary to moderate decrease in precipitation leads to a 10 to 15% (8 to 15%) decrease in aboveground biomass production (grain yield). When the warming is moderate (i.e. +1.4 to 1.8°C), the decline in grain yield worsens (10 to 20%), despite a slight increase in rainfall projections. At these rates of loss in crop production, resilience can be reinforced. However, it will require that climate-smart crop management practices be embedded in sub-seasonal and interannual monitoring and early warning systems.
KEY WORDS: Agroclimatic risks · Ensemble scenarios · Uncertainty management · Sensitivity
analysis · Pearl millet · Maize · West African Sudan-Sahel