Use of a multispectral radiometer for noninvasive assessments of foliar disease caused by ray blight in pyrethrum

Foliar disease due to ray blight (Phoma ligulicola) in pyrethrum was quantified at three locations over 2 years in Tasmania, Australia. To obtain a range of ray blight disease intensities, replicated plots were treated with fungicides that varied in efficacy to control ray blight. Visual disease assessments and measurement of canopy reflectance were made at least once during spring (September through December). Visual assessments involved removal of flowering stems at ground level from which measurements of defoliation severity and the incidence of stems with ray blight were obtained. Reflectance of sunlight from pyrethrum canopies was measured at 485, 560, 660, 830, and 1,650 nm using a handheld multispectral radiometer. Measurements from these wavelengths also were used to calculate all possible reflectance ratios, as well as four vegetative indices. Relationships between wavelength bands, reflectance ratios, vegetative indices, and disease intensity measures were described by linear regression analyses. Several wavelength bands, ratios, and vegetative indices were significantly related in a linear fashion to visual measures of disease intensity. The most consistent relationships, with high R² and low coefficients of variation values, varied with crop growth stage over time. The ratio 830/560 was identified as the best predictor of stem height, defoliation severity, and number of flowers produced on each stem in October. However, reflectance within the near-infrared range (830 nm) and the difference vegetative index was superior in November. The use of radiometric assessment of disease was noninvasive and provided savings in disease assessment time, which is critical where visual assessment is difficult and requires destructive sampling, as with pyrethrum.