Modeling Fusarium head blight and deosynivalenol content in barley in response to field temperature and wetness durations

Fusarium head blight (FHB), caused by the fungus Gibberella zeae (anamorph: Fusarium graminearum), continues to be a serious problem for barley producers in the U.S. Northern Great Plains and elsewhere. Field experiments were conducted during the 2005-9 growing seasons to evaluate the combined effects of temperature and wetness durations (relative humidity > 90%) prior to full head emergence on disease development and deoxynivalenol (DON) accumulation in malting barley. Disease incidence (number of diseased spikes/total), average severity (number of disease spikelets/total), and DON content in the grain (mg/kg) were collected from 51 location*years for three varieties, 'Conlon', 'Robust' and 'Tradition'. A binary DON response variable, eDON, was created based on the DON content for each variety at every location*year where '0' and '1' represent values below or above a threshold of 0.5 mg/kg. A Weibull function was calculated to predict the probability of infection using the average temperature and weighted wetness durations in the field during the 10-day prior and including the full head emergence day. Disease incidence, severity, DON content and eDON were significantly correlated to the Weibull variable calculated from the field weather data (p < 0.001). The sensitivity, specificity and total prediction accuracy obtained from the confusion matrix after dividing the observations at a cut-off of 0.5 of Weibull variable in comparison with eDON were greater than 80%.