Characterization of multiple disease systems and cultivar susceptibilities for the analysis of yield losses in winter wheat

Abstract

The relationship between winter wheat ( Triticum aestivum L.) cultivar susceptibility to four main fungal diseases, multiple disease systems, and yield loss (YL) levels was studied. Data from 276 wheat cultivar trials carried out in major wheat-growing areas in France over 12 years (1991–2002) were analyzed. Two variables were defined for each disease: actual disease intensity, referring to individual cultivar and potential disease intensity for each trial, calculated from the disease intensities measured on susceptible cultivars. Cultivar was represented by its level of susceptibility to each disease, previously given by GEVES and based on a semi-quantitative scale. YL was estimated for a given cultivar as the difference between the mean yield of fungicide-protected and of untreated plots. Nonparametric multivariate analysis provided a characterization of relationships among variables. Eight cultivar susceptibility profiles (CP), five potential disease profiles (PDP), and six actual disease profiles (ADP) were determined by cluster analysis. Correspondence analyses led to a good description of variation in YL when potential or ADP were considered. General linear models were developed for each of the five PDP, which involved a significant interaction between CP and PDP on YL. These empirical models provided a means of quantifying the effects of CP on YL. These models represent a practical tool to support choices for multiple-disease resistances in a given area, depending on the prevailing disease profile. Results from this study can also be used to improve the disease module of an agronomic model for wheat aimed at designing “cultivar×crop management” combinations for a given environment and cost/price ratio.