Spatial aspects of light leaf spot (Pyrenopeziza brassicae) epidemic development on winter oilseed rape (Brassica napus) in the United Kingdom

In microplot experiments in 1998-99 and 1999-2000, the start of light leaf spot epidemics could be predicted from weather data, using empirical equations for Pyrenopeziza brassicae apothecial (ascospore) development. ascospore infection criteria, and the latent period of P. brassicae. The dates when P. brassicae sporulation was first observed fitted predictions and initial spread of light leaf spot from an inoculum source was mostly in the prevailing wind direction, with differences between the two growing seasons attributable to differences in wind patterns. Subsequent secondary spread of disease could be predicted using temperature and rainfall data, and observations fitted predicted dates. In both 1998-99 and 1999-2000, initial spatial patterns of observed disease in January were random, because data were not significantly different from a binomial distribution (P = 0.18). Analysis of spatial data from samples in February and March indicated aggregation, because data fit was significantly different from a binomial distribution (P less than or equal to 0.026). These data were described by a beta-binomial distribution, suggesting that the spatial distribution of light leaf spot becomes aggregated as secondary spread occurs. The importance of wind-dispersed ascospores in initiating epidemics and rain-splashed conidia in secondary localized spread in relation to strategies for sampling winter oilseed rape crops in the United Kingdom to assess light leaf spot is discussed.