Evaluating Multiple Rating Methods Utilized in Turfgrass Weed Science

Abstract

Turfgrass weed scientists commonly use visual ratings (VR) to assign a numerical value to a turfgrass or weed response. These ratings lack quantifiable numerical values and are considered subjective. Alternatives to VR, including line intersect analysis (LIA) and digital image analysis (DIA), have been used to varying extents in turfgrass research. Alternatives can be expensive, labor intensive, and can require extensive calibration and increased time for data acquisition. Minimal research has been conducted evaluating rating methods used in turfgrass weed science. Trials were conducted in 2007 and 2008 to evaluate ratings methods used to quantify large crabgrass populations as influenced by tall fescue mowing height (2.5, 5.1, 7.6, and 10.2 cm). Percent large crabgrass cover was assessed utilizing VR, LIA, and DIA to determine if differences existed among evaluation methods. Pairwise comparisons, Pearson's correlation, and linear regression were performed to compare evaluations. All rating methods were significantly correlated to one another. Differences of large crabgrass cover estimates existed between LIA and DIA data at all mowing heights and between VR and DIA data at the 7.6 and 10.2 cm mowing heights in 2007. Authors believe that shadows produced by the turf canopy at higher (≥ 7.6 cm) mowing heights increased DIA estimates of large crabgrass cover. At trial initiation in 2007, researchers did not capture calibration images because the methodology to eliminate a shadow influence using a standard digital image had not been published. Additional DIA calibration in 2008 corrected for canopy shadows, and no differences were observed in large crabgrass cover between all evaluation methods indicated by nonsignificance pairwise comparisons and estimated regression parameters. These data indicate VR are no different than LIA or DIA in estimating large crabgrass cover as affected by tall fescue mowing height.

Herbicide Applications and Incorporation Methods Affect Dazomet Efficacy on Bermudagrass

“…Pearson correlation coefficients showed a strong positive relationship between cover determined via visual estimates and lineintersect analysis in North Carolina at 15 (r = 0.97, P < 0.0001) and 46 (r = 0.92, P < 0.0001) WAIT, as well as in Texas at 16 (r = 0.91, P < 0.0001) and 42 (r = 0.99, P < 0.0001) WAIT, indicating visual bermudagrass cover estimates increased proportionally with cover determined from line-intersect analysis (Table 2). These findings agree with Hoyle et al (2013), who reported correlations between visual cover estimates of large crabgrass [Digitaria sanguinalis (L.) Scop. ], a common weed in turfgrass systems, and lineintersect analysis were r = 0.93 to 0.98 (P < 0.001) across two experimental runs.…”

“…Bermudagrass cover was also determined by line-intersect analysis at 15 and 46 WAIT in North Carolina (0.1 • 0.1 m, 322 intersections), and 16 and 42 WAIT in Texas (0.1 • 0.1 m, 792 intersections). Intersections with bermudagrass beneath them were summed and divided by the total intersects to estimate percent cover (Hoyle et al, 2013).…”

Zoysiagrass Sod Establishment along Guardrails: Evaluation of Cultivars, Soil Preparation Techniques, and Planting Timings

“…Pearson correlation coefficients showed a strong positive relationship between visual cover estimates and zoysiagrass intersection counts at 40 ( r = 0.86, P < 0.001), 90 ( r = 0.87, P < 0.001), and 125 ( r = 0.91, P < 0.001) WAIE, indicating that visual zoysiagrass cover estimates increased with increased intersection counts. These findings agree with Hoyle et al (2013), as reported correlations between visual turfgrass cover estimates and line intersect analysis were r = 0.93 and 0.98 ( P < 0.001) in years 1 and 2, respectively. Therefore, visual zoysiagrass cover estimate data are presented to progressively identify variables that resulted in successful sod establishment along guardrails, arbitrarily set at ≥60% cover, and implications for in situ implementation.…”

“…The grid was partitioned into an inner grid (0.53 by 2.4 m) to quantify the original sod strip cover, and outer regions (two, 0.25 by 2.4 m) on both sides of the strip to quantify sod spread. Sod spread was calculated within a plot using the following equation: where Z intersections and T intersections equaled the number of intersections with zoysiagrass present outside of the original sod strip and the total number of intersections (92) outside of the original sod strip, respectively (Hoyle et al, 2013).…”

The Preservation and Protection of Native Biodiversity in the Guadalupe Nipomo Dunes Complex

“…Thompson in accordance with turfgrass standards for effectively measuring coverage (Hoyle, Yelverton, & Gannon, 2013;Thompson, personal communication, 2015). The From this experiment, we can conclude that the seedbank is not the variable controlling the successful invasion of E. calycina.…”