Quantifying the intrarater repeatability and interrater reliability of visual and remote-sensing disease-assessment methods in the alfalfa foliar pathosystem

Abstract

The precision of visual and remote-sensing assessment methods for the alfalfa foliar pathosystems was studied. Precision was defined as the intrarater repeatability and interrater reliability of a disease-assessment method. Visual disease assessments were performed on 10 alfalfa stems per plot by four raters for disease incidence, disease severity, and percent defoliation. Remote-sensing assessments were also performed by the same four raters, using a hand-held, multispectral radiometer to measure the percentage of sunlight reflected from alfalfa canopies. The F statistics, intercepts, slopes, coefficients of determination (R 2), standard errors of the estimate for y, and coefficients of variation were used to quantify and compare the repeatabilities and reliabilities of each assessment method. Among the three visual disease-assessment methods evaluated, percent defoliation had the highest intrarater repeatability and interrater reliability with R 2 ranging from 0.89 to 0.97, whereas R 2 ranged from 0.15 to 0.95 among the four raters for disease incidence and disease severity. For intrarater repeatabilities and interrater reliabilities with the remote-sensing assessment method, R 2 ranged from 0.87 to 0.99. The standard errors of the estimate for y and coefficients of variation values for intrarater repeatabilities and interrater reliabilities, when using the remote-sensing assessment method, were approximately one half of the values for the percent-defoliation assessment method. In summary, percent defoliation had the best precision among the three visual assessment methods, and the percent-reflectance assessment method had the highest precision compared with all three visual assessment methods.