Research Plant Pathologist, Agricultural Research Service, U.S. Department of Agriculture and Department of Plant Pathology
Professor, Division of Biology, Kansas State University, Manhattan 66506-5502
Severe leaf rust epidemics, which result in economic yield reductions in the Great Plains wheat-producing region of the United States, are usually initiated by Puccinia recondita f. sp. tritici inoculum that has survived in the local field from the previous wheat crop until early spring. Models were developed for an epidemic year beginning at physiological maturity of one wheat crop to maturity of the following wheat crop. Meteorological variables for periods prior to final tiller development of the wheat crop during 1980 to 1992 at several sites in the central Great Plains winter-wheat-production area were used to model inoculum survival from one wheat crop until early spring of the next crop. Stepwise multiple regression was used to identify weather variables that explained the most variation in inoculum survival at the final tiller development wheat growth stage. Inoculum survival was recorded on a 0 to 9 scale with 0 indicating no survival and 9 indicating inoculum on all wheat plants in the field. Independent variables used in development of models were daily deviations from the 10-year average of maximum and minimum temperature, fungal temperature equivalence function, cumulative fungal temperature function, precipitation, cumulative precipitation, and snow cover averaged for 10-day periods prior to dates inoculum forecasts were desired. Models were constructed to forecast inoculum survival from data collected prior to fall wheat planting, the beginning of winter dormancy of the wheat, and the final tiller development wheat growth stage. Of the observed occurrences of leaf rust overwintering, 70% were forecast by models constructed using weather data prior to wheat planting decision time. Overwintering could be forecast by models constructed with data prior to the wheat entering winter dormancy 80% of the time. Models constructed with data collected prior to final tiller development in the spring forecast overwintering of leaf rust inoculum 95% of the time. Results from these models will be used to develop forecasts of leaf rust epidemics and resulting yield reductions.