Weather-Driven Exploratory Model of Powdery Mildew Development in Winter Wheat Based on Airborne Spore Trap Data

Powdery mildew, caused by Blumeria graminis, is a major foliar disease of winter wheat, and its development depends closely on weather and spore count. This exploratory study combines spore counts measured with a portable impactor-type spore trap with hourly meteorological data. Observations were made on five winter wheat varieties over three growing seasons in South Russia. Air and soil temperature, relative humidity, vapour pressure deficit ($VPD$) and precipitation were aggregated over sliding windows of 1–21 days and lagged periods, and seasonal epidemic metrics (infection pressure, stress index, rainfall index and phenology-weather index) were calculated. Predictors were selected by Spearman correlation corrected for pseudo-replication and a stability-selection method based on Elastic Net. Logarithmic conidial concentration was modelled by lagged regression, and disease severity by a two-part hurdle model. The combined model is able to replicate the contrasting epidemics of the three seasons in the fitted data, and its occurrence part was inverted to obtain provisional, weather-dependent spore concentration thresholds for given infection probabilities. However, these model-derived thresholds require validation using data from different seasons and plots before they can be used to inform decisions regarding the timing of fungicide applications.