Prévision de l’importance d’une épidémie de septoriose du blé à Septoria nodorum

Nous proposons un système de prévision de l’importance d’une épidémie de septoriose du blé à Septoria nodorum destiné à fonder une pratique de traitements fongicides raisonnés. La méthode repose sur un modèle mathématique de l’évolution de la maladie et sur la comparaison, à l’aide de méthodes statistiques descriptives, des sorties du modèle pour différentes séries de données climatiques.
A system of forecast of the severity of a wheat giume blotch epidemic due to Septoria nodorum is given. It is aimed to give indications on the necessity of fungicide treatments. The method is based on the use of a mathematical model describing the disease evolution. The input of the model are climatic datas and the output are different curves describing, for instance, the diseased leaf area with respect to time. It is possible to compare different curves coming from different climatic series : in particular the comparison of the curves corresponding to the « unknown » epidemic with these corresponding to various « known » epidemics can give indications on the severity of the unknown epidemic, depending on the distances of its curves with respect to those of « low » or « high » glume blotch epidemics. These comparisons arc made by means of definition of distances between curves, principal coordinates analysis of the distance tables so obtained and finally discriminant analysis, taking as populations the low and the high epidemics.