Durum wheat quality prediction in Mediterranean environments: From local to regional scale

•Delphi system, based on AFRCWHEAT2 model, for durum wheat quality forecast.•A scenario approach was applied to simulation of durum wheat quality.•Climatic variables potentially controlling grain protein content.•Operational mode for nine years in rainfed and water limiting conditions.•Accurate forecast as an useful planning tool.

Durum wheat is one of the most important agricultural crops in the Mediterranean area. In addition to yield, grain quality is very important in wheat markets because of the demand for high-quality end products such as pasta, couscous and bulgur wheat. Grain quality is directly affected by several agronomic and environmental factors. Our objective is to determine the general principles underlying how, in Mediterranean environments, grain protein content (GPC) is affected by these factors and provide a system model with high predictive ability. We initially evaluated the capability of the Delphi system to simulate GPC in the major Italian supply basins (Basilicata, Capitanata, Marche, Tuscany) for 9 years (1999–2007) a month ahead of harvesting and we then analyzed relations between Delphi system errors and selected environmental variables during flowering and grain filling stages. The results were evaluated on the basis of regression with observed GPC, while errors were calculated performing a linear correlation analysis with environmental variables. The model showed a high capability to reproduce the inter-annual variability, with important year to year differences, with better performance in the southern study areas (Basilicata and Capitanata). In this study the highest overestimation occurred in conjunction with the year (2004) characterized by the lowest quality in terms of GPC, lowest average temperature in May and highest yield production for the whole study period.