Résumé
Crop simulation models (CSM) dynamically estimate agricultural production as a function of weather and soil conditions, and crop management. They can be used for evaluating cultivars in actual and future tropical conditions. In Northern Cameroon, cotton (Gossypium hirsutum L.) is grown exclusively in rainfed conditions and its yield has been decreasing steadily since the 80s. Therefore our objectives were to evaluate the usefulness of CSM to identify current rainfed cotton ideotypes for two contrasted environments of northern Cameroon. Based on field observations constituting the minimum dataset, phenology, morphology, leaf area index and yield simulated by CROPGRO-Cotton were successfully calibrated and validated in our conditions. Results showed that ideotype should have earlier anthesis date, longer reproductive duration, thicker leaves with higher potential assimilation rate compare to the reference cultivar (L484). In the North region, it seemed that having bigger leaf than L484 should be favorable whereas in the Far North smaller ones are more suitable to local drought-prone conditions. We concluded that morpho-physiological traits could and should be imported into breeding programs in FS generation where high genetic diversity still exist and plant material start to be considered as a line rather than a single plant. Consequently, we invite breeders to target cultivars with low "emergence to anthesis" to reproductive ratio, thick leaf, high chlorophyll content, and smaller leaf for the conditions with the lowest water availability and bigger ones for the conditions with best water availability. (Résumé d'auteur)