A simulation model for the development of brown rust epidemics in winter wheat

A model simulating the progress of Puccinia recondita severity, expressed as a percentage of rusted leaf area (both as average and its 95% confidence interval) on individual wheat leaves over the course of a growing season, with a time step of one day, was elaborated using laboratory and field data from literature. Data on the stages of each infection cycle (uredospore germination, penetration, latency, uredium eruption and infectiousness) were transformed into model parameters by curve fitting, Montecarlo stochastic procedures, corrections and empirical assumptions. Data on host growth, like the timing of all phenological stages, the dynamic of the green area of each leaf from appearance to complete senescence, and tillering were obtained from a specific sub-model. Model validation was performed on actual data not used in model building and representing a wide range of conditions (several winter wheat cultivars grown at eight locations in northern Italy between 1990 and 1994) by using subjective, non-parametric and parametric tests: it revealed a satisfactory agreement between the data simulated by the model and actual data.