| Sumario: | Mozambique faces challenges in staple food crop production, which makes crop yield prediction vital for effective policy-making on food security. The analytic framework that integrates satellite data and crop growth simulations to forecast regional crop yield can aid policy makers. The objectives of this study were to apply the analytic framework to three major crop production regions in Mozambique including Gaza, Manica, and Nampula provinces for maize, soybean, and rice. The gridded crop growth simulations were performed using Decision Support System for Agrotechnology Transfer (DSSAT). A set of crop management scenarios were applied to the crop growth simulations. One of these simulations were identified to obtain crop yield hindcasts by cell comparing leaf area index data derived from the simulations and the Moderate Resolution Imaging Spectroradiometer (MODIS) satellite data. Crop yield hindcasts were obtained using a percentile of crop yield distribution using three preceding growing seasons. It was found that the percentile used for crop yield hindcasts differed by crop and province. The accuracy of maize and soybean yield hindcasts was within an acceptable range, e.g., < 20% of crop yield in growing seasons, whereas that of rice yield hindcasts was considerably high. Crop yield predictions were limited by the use of crop management scenarios such as cultivars and fertilizer application. Despite biases and limitations in representing real farming conditions, the framework provided insights into improving staple food crop production. It was also highlighted that detailed knowledge on crop management practices such as cultivar and fertilizer applications would improve the reliability of the analytic framework to predict crop yield in the major production regions in Mozambique.
|
|---|