Characterizing Meteorological Scenarios Favorable for Septoria tritici Infections in Wheat and Estimation of Latent Periods

Abstract

Septoria tritici epidemics were monitored on winter wheat cv. Ritmo between 1995 and 2003 at 8 to 11 locations per year in Northern Germany (area between 53.70 and 54.38°N latitude and 8.83 and 10.88°E longitude) by counting the number of pycnidia on the leaves of plants obtained from plots under natural infection. Furthermore, meteorological data (leaf wetness, temperature, and precipitation) were recorded within the same period at the same locations. Groups of similar meteorological events were identified by hierarchical cluster analysis. The temporal distance of those clusters from the point of time when an increase of more than 70 S. tritici pycnidia was observed per leaf within 1 week was calculated for all epidemiological case studies and meteorological clusters. A cluster with average temperature = 13.62 ± 2.30°C, leaf wetness = 92.39 ± 4.15%, and precipitation = 0.04 ± 0.10 mm per day was consistently observed at 20.35 ± 4.15 days before epidemic outbreaks. This estimate of a latent period was significantly affected by geographic latitude, average temperature during infection, average temperature during the latent period, year, and precipitation, but not by leaf layer and longitude. On average, an increase in temperature during the infection period by 1°C decreased latent periods by 0.95 day. Latent periods were decreased by 0.2 day upon an increase of the average temperature by 1°C during the latent period. Average latent periods decreased by 1.7 days per degree of north latitude. Latent period estimates had lower coefficients of variation than temperature sums accumulated over latent periods. The usefulness of the approach described above for general epidemiology and for increasing fungicide efficacy by improving the timing of applications is discussed.