Forecasting Spore Episodes ofBotrytis squamosain Commercial Onion Fields in New York

Publisher Website
Google Scholar
Abstract
Previously published models for forecasting sporulation by Botrytis squamosa often failed to forecast major spore episodes of this pathogen when weather and Hirst spore trap data collected in a commercial onion [Allium Cepa L.], field in Orange County, NY were used to evaluate the models. The sporulation index model for B. squamosa forecasted only 51.2% of the minor spore episodes (daily mean 1-9.9 spores per m3 of air) and 58.7% of the major spore episodes (mean .gtoreq. 10 spores per m3) in a 7-yr data set. The DINOV submodel of the BOTCAST predictive system for B. squamosa forecasted 81.8% of the minor spore episodes but only 56.5% of the major spore episodes in 2-yr data set in which all appropriate weather data were available. An alternative model was developed by deriving a set of decision rules (modifications from BOTCAST) and regression equations that identified conditions associated with the occurrence or nonoccurrence of significant spore episodes. Temperature, relative humidity, and calendar date were used as variables in the model, which was designed to issue a daily forecast-the inoculum production index (IPI)-for the presence or absence of significant sporulation. The IPI model correctly forecasted 66.7% of the minor spore episodes and 94.2% of the major spore episodes in the 7-yr data set used to develop the model. Similar results were obtained with a 4-yr data set not used in model development (76.2 and 88.6% of minor and major spore episodes, respectively), indicating the suitability of the model for use in independent growing seasons. By incorporating crop age rather than calendar date into the model (IPI2), further improvement was made in forecasting significant spore episodes. The IPI2 model should be useful in a predictive system for timing fungicide applications to control Botrytis leaf blight of onion in New York.