A damage function for stem rust of perennial ryegrass seed crops

Observations of naturally occurring stem rust epidemics and seed yields in perennial ryegrass were taken in 19 field experiments conducted over the course of 9 years. Epidemic severity differed among years and also among experimental treatments (fungicide regimes) within years. In each experiment, attainable yield was represented by the nondiseased treatment, and yields of other treatments were expressed as relative yield (a proportion of the attainable yield). Yield loss (difference between attainable and actual yield) in the nonprotected treatments was 0 to 98% due to yearly differences in epidemic conditions. Fungicides were effective in reducing stem rust injury and damage when properly timed. Disease severity in the upper canopy was estimated at approximately weekly intervals and converted to proportion of the plant area diseased. The complementary value, proportion of area healthy, and its integral over time, healthy area duration (HAD), were calculated. Regression analyses were conducted using various phenological time intervals of HAD as the independent variable. The best intervals of HAD for predicting relative yield were centered on the midpoint time between anthesis and harvest. The regression equation (r2=0.89) for relative yield as a function of HAD during the 3-week interval was selected and rearranged to produce a quadratic damage function. This damage function estimates yield loss at 5, 22, and 42% for critical-interval diseased proportions of 1, 5, and 10%, respectively. Yield data collected from field experiments not used in model development correlated well (r2=0.9) with yields predicted by the damage function from their observed disease severity.