1998 Iowa Turfgrass Research Report
Radiometric Assessment of Dollar Spot Severity on a Creeping Bentgrass Green
Mark L. Gleason, Forrest W. Nutter, Jr., and Nick E. Christians
Introduction
Few guidelines are available to help researchers, consultants, and agrichemical industry personnel to select a "best" method for assessing the severity of turfgrass diseases. The "best" method should be easy to use, offer reproducible results, and provide a fast and accurate measure of disease severity.
Rating systems for turfgrass diseases - usually based on visual assessment of test plots - are highly subjective and vary greatly from person to person. This crazy quilt of methods slows the development of new disease control products because assessment errors and variability can mask or distort differences between treatments.
Radiometers have shown promise for improving assessment of some turf diseases. Research in the early 1990's by our group (Nutter et al., Phytopathology 83:806-812, 1993) showed that assessments of dollar spot severity using a commercially available radiometer (Crop Scan Inc.) were not only more accurate than visual assessments but also faster to make.
This report summarizes a 1997 project to test radiometric assessment against visual assessment in the "real world" context of fungicide trials. The purpose of the trial was to evaluate radiometry's potential to improve the quality of research in new product evaluation.
Procedures
A field trial was conducted in September 1997 on green-height Penncross creeping bentgrass at the ISU Horticulture Research Farm. Plot design was a randomized complete block with 39 fungicide treatments (38 treatments selected by agrichemical companies plus one untreated check treatment) and four replications per treatment. The entire plot was inoculated with rye grain infested with the dollar spot pathogen, Sclerotinia homeocarpa, approximately one month before plot ratings were made, and fungicide treatments began five days after inoculation. Applications of fungicides were made with a bicycle sprayer using flat fan nozzles at 30 psi, using 5 gal water/1000 ft2.
By the first week of September, plots displayed a wide range of dollar spot severity, from disease-free to severely infected. On 9 September, all plots were rated visually and radiometrically by four individuals. Radiometric readings, using a CropScan radiometer at 610 and 810 nm wavelengths, were made in full sun between 11 am and 2 pm. Both visual and radiometric assessments were made on the central 1 m2 of each plot.
Results
Statistical analysis of the data showed that: 1. There was a significant interaction between the fungicide treatment and the person doing the visual assessment, but not the radiometric assessment.
2. The coefficient of variation (a measure of the precision of the measurements; the lower the number, the more precise the measurements) was 31.5 for visual ratings and 1.4 for radiometric ratings.
3. For visual ratings, the slope of the graphs comparing one rater's ratings to that of another rater - an indicator of systematic bias (varies with disease severity) among the raters - was significant for 10 of the 12 combinations of four raters; for radiometric ratings, none of the 12 combinations was significant.
4. The intercept of the graphs comparing one rater's ratings with another's - an indicator of a constant bias (does not vary with disease severity) was significant for 8 of the 12 comparisons for visual assessment but for none of the radiometric assessment comparisons.
5. Rank correlation of treatments between visual and radiometric assessments was relatively high (91%). However, preliminary analysis of the data suggested that the rank order of several treatments was significantly different between visual and radiometric methods. This means that the rating of the relative success of at least some of the fungicide treatments at controlling dollar spot would have been different with the two methods.
Interpretation
1. Radiometric assessment of dollar spot severity was more precise and less subject to bias by individual raters than visual assessment.
2. Although the rankings of treatments were fairly similar by the two methods, some treatments were ranked significantly different by these methods.
3. An earlier study (Nutter et al., 1993) showed that radiometric ratings of dollar spot severity were also more accurate (i.e., closer to the true level of disease) than visual ratings, and required only 2/3 the time to complete.
Bottom line
The study demonstrated that radiometric
assessment has excellent potential for improving the
accuracy, precision, and speed of dollar spot ratings. The
method may also have potential to improve ratings of other
turfgrass diseases.
 |
ISU Horticulture:Publications:1998 Turfgrass Report | College of Agriculture |