Sensitivity of CSM-CERES-Maize model to soil available water and impact on rainfed maize grown in the Brazilian Cerrado

The response of maize to variations in soil available water (AW) is a function of the interactions among plant, soil and weather conditions. We studied the sensitivity of a previously calibrated CSM-CERES-Maize model and the response of rainfed maize grain yield to soil AW. The study was conducted for conditions in southeast Brazil. The model was set for weekly sowings, from August to September, for a total of 52 sowing dates. At each sowing date, six scenarios of soil AW using field capacity estimated at -4 kPa, -6 kPa, -10 kPa, -20 kPa, -33 kPa and determined in situ, were used. For each sowing date, the model was also set for rainfall reductions of 10%, 20%, 30%, 40% and 50%. The simulated results showed the sensitivity of the model to soil AW, which in turn affected grain yield of maize among sowing dates. For the highest yielding sowing date, a reduction of 48.3% in average grain yield was simulated with soil AW using FC at -4 kPa and FC at -33 kPa. Additionally, our simulations indicated significant correlation between grain yield and total crop evapotranspiration and between grain yield and maximum leaf area index. Scenarios of low rainfall had little effect on yield with high soil AW. Therefore, our simulations indicate that accurate information on FC is needed for the simulation of maize grown under rainfed conditions. Our simulations also indicate that the best sowing window for maize in southeastern Brazil ranges from Oct 17 to Nov 28.