Comparison of Visual Rating Techniques and Spectral Reflectance Characteristics during Turf Variety Trial Evaluations.

Visual rating methods are often used to determine turf health and quality; however, these assessments are prone to bias and inconsistency. Vegetative indexes (VI), derived from spectral reflectance characteristics, and color analysis techniques are less subjective and may supplement visual methods. Research was conducted to compare visual rating techniques with VI’s derived from standard spectral analysis equipment used by turfgrass professionals.

Visual ratings and spectral reflectance measurements were taken during spring green-up of the 2013 National Turfgrass Evaluation Protocol (NTEP) bermudagrass test with the objective of comparing evaluation methods and gauging bermudagrass emergence from winter dormancy.  The experiment was planted June 28, 2013. Plots were maintained at a ½ inch mowing height with supplemental irrigation and fertility. Plots had reached total cover prior to winter dormancy. A -13°C low temperature was sufficient to cause significant winter injury during spring of 2014.

Visual ratings of percent green cover and turf color (1-9 scale) were assessed every two weeks by four experienced evaluators beginning on April 1, 2014. Spectral reflectance at 670 nm (red) and 780 nm (near infrared) were recorded by a Holland Scientific Crop Circle ACS430 from a height 60 cm above canopy level. Additionally, reflectance at 660 nm (red) and 850 nm (near infrared) were recorded using the FieldScout 500 NDVI Turf Color Meter at turf canopy level. Dark green color index (DGCI) and a derived visual rating (dVR) were computed for each plot using the Spectrum Technologies GreenIndex+ Turf App for iPhone. This device captures and processes image color relative to a green and yellow reference standard provided with the software.

Data were subject to analysis of variance. All methods detected significant cultivar differences that generally corresponded to visual rating methods (percent green cover and plot color). Data were also analyzed for correlation with visual rating methods. All methods were positively correlated (<0.0001). Some correlations showed stronger relationships with visual rating methods than others.