Models for Predicting Potential Yield Loss of Wheat Caused by Stripe Rust in the U.S. Pacific Northwest

Abstract

Climatic variation in the U.S. Pacific Northwest (PNW) affects epidemics of wheat stripe rust caused by Puccinia striiformis f. sp. tritici. Previous models only estimated disease severity at the flowering stage, which may not predict the actual yield loss. To identify weather factors correlated to stripe rust epidemics and develop models for predicting potential yield loss, correlation and regression analyses were conducted using weather parameters and historical yield loss data from 1993 to 2007 for winter wheat and 1995 to 2007 for spring wheat. Among 1,376 weather variables, 54 were correlated to yield loss of winter wheat and 18 to yield loss of spring wheat. Among the seasons, winter temperature variables were more highly correlated to wheat yield loss than the other seasons. The sum of daily temperatures and accumulated negative degree days of February were more highly correlated to winter wheat yield loss than the other monthly winter variables. In addition, the number of winter rainfall days was found correlated with yield loss. Six yield loss models were selected for each of winter and spring wheats based on their better correlation coefficients, time of weather data availability during the crop season, and better performance in validation tests. Compared with previous models, the new system of using a series of the selected models has advantages that should make it more suitable for forecasting and managing stripe rust in the major wheat growing areas in the U.S. PNW, where the weather conditions have become more favorable to stripe rust.