Estimation of visual quality and canopy characteristics of turfgrass using spectral reflectance and digital imagery

Abstract

Radiometric methods may provide more objective and quantitative assessments of turf quality and density than visual ratings and may be useful in measuring green leaf area index (LAI), aboveground biomass, and chlorophyll concentration. This three-year study was conducted near Manhattan, KS to examine: 1) relationships between canopy reflectance and visual quality and density ratings in four cool-season grasses tall fescue (Festuca arundinacea Schreb.), Kentucky bluegrass (Poa pratensis L.) and two hybrid bluegrasses (HBG); 2) effects of species, mowing height, and irrigation deficit on relationships between visual quality and reflectance; 3) comparisons of visual quality with reflectance and digital images of individual plots; and 4) relationships of LAI, aboveground biomass, and chlorophyll concentration with canopy reflectance in the same four grasses and in perennial ryegrass (Lolium perenne), zoysiagrass (Zoysia japonica Stued.), and bermudagrass [Cynodon dactylon (L.) Pers.]. Reflectance was strongly correlated with visual ratings in the normalized difference vegetation index (NDVI, [935-661]/[935+661] nm, r = 0.88), the near infrared to red (NIR/R [935/661] nm, r = 0.83), Stress1 (706/760 nm, r= 0.84), and Stress2 (706/813 nm, r= 0.70) ratios and at wavelengths 613 (r= 0.74) and 661 nm (r = 0.80), but correlations varied among years at each wavelength and vegetation index. For density, highest correlations were in NDVI (r=0.86), R661 (r=-0.84), and Stress2 (r=-0.82). Regressions between reflectance and quality and density ratings indicated cultivar- and mowing height-specific models. Irrigation-deficit strongly affected reflectance in KBG and both HBG but not in TF, indicating greater sensitivity to drought of bluegrasses than TF. Digital images indicated strong correlations between percentage green cover and visual quality (r=0.89). However, wide ranges in visual quality were observed in plots with similar green cover or NDVI for reasons that are not apparent. Correlations of LAI, aboveground biomass, and chlorophyll concentration with reflectance were strong in some species at different wavelengths and ratios. Results indicated both potential and limitations in using spectral reflectance to estimate turfgrass canopy characteristics.