The increasing rainfall variability in West Africa is a great challenge for crop productivity in small-scale farming systems, thus jeopardizing food security. Rainfed rice is particularly sensitive to inconsistent rainfall, especially during the reproductive stage. It is, therefore, necessary to develop management practices suited to the change of rainfall pattern over the growth seasons. In this study, the modeling technology with the rice model ORYZA (v3) was used to identify appropriate rainfed rice growing seasons for a better adaptation of farmers to climate variability. The potential yields, the favorable sowing periods, the optimum sowing dates, and the attainable yields of two contrasted cultivars were determined. After successfully calibrating and validating the model, it predicted potential yields of 5.5 to 6.5 tons/ha for the early maturing variety WAB56-104 (90-100 days), while potential yields of 4 to and 5.5 tons/ha was predicted for the late maturing variety CG14 (115-125 days). In rainfed conditions, two favorable sowing periods were identified from the model scenario analysis. The first period spans from late February to late April and the second from late July to early September. Farmers can double their actual yield of 1.5 tons/ha if they follow the recommended sowing dates and good agricultural practices. Indeed, the yield of 3.5 tons/ha was found with the variety WAB 56-104 sown on around 16 April in San-Pedro and around 2 April in Dimbokro. The yield of 3 tons/ha for the variety CG14 could be achieved if the sowing is done on around 18 March in San-Pedro and around 21 March in Dimbokro.

This study aims to identify the optimal time of planting upland rice in the Center-west of C