Application of the CERES-Wheat simulation model to durum wheat in two diverse mediterranean environments

Abstract

The CERES-Wheat model was applied to simulate growth and yield of durum wheat in two diverse Mediterranean environments, Breda (northern Syria) and Libertinia (eastern Sicily), using historical daily weather data. Two cultivars of known, good performance under Mediterranean conditions were examined. Three sowing dates were simulated: ‘normal’ (26 November), ‘early’ (1 November) and ‘late’ (20 December), for 12 seasons at Breda and six at Libertinia. Simulated grain yield and simulated date of anthesis were considered. Absolute differences between simulated and observed yield were independent of the recorded yield level. In a very harsh season at Breda, the model precision (in absolute terms) was of the order of magnitude of the low yield observed, thus making the simulation very inadequate. At both locations, ‘early’ sowing had the highest simulated yield, and ‘laterssowing the lowest. The water stress index CSD2, affecting leaf expansion and growth, was estimated for various growth stages. At ‘normal’ sowing, simulated yield was mainly affected by simulated stress around anthesis at Breda, and by stress during grain filling at Libertinia. The simulation suggested that anthesis date may be influenced by interactions of sowing date or cultivar with the environment. Alteration of the phenology of the simulated cultivar by modification of the genetic coefficients highlighted the positive effect of earliness on yield in these environments. Prospects for the application of the model to durum wheat in the Mediterranean area are discussed.