TRAFFIC TOLERANCE AND DIVOT RECOVERY OF EIGHT BERMUDAGRASS CULTIVARS

Transcription

1 TRAFFIC TOLERANCE AND DIVOT RECOVERY OF EIGHT BERMUDAGRASS CULTIVARS By BRADLEY WILLIAMS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA

2 2010 Bradley Thomas Williams 2

3 To TDW 3

4 ACKNOWLEDGMENTS With sincere gratitude, I want to thank my major professor, Dr. Jason Kruse for his guidance and direction throughout my degree program. I greatly appreciate the time and effort that you invested in me during my time at the University of Florida. To my committee members, Dr. Jerry Sartain and Dr. J. Bryan Unruh, thank you for your support and assistance with my research. I appreciate all the contributions that you made in my research. I would also like to thank Jason Haugh for the countless hours of help in the field, Dr. Kevin Kenworthy for the use of equipment, and Mark Kann and the staff at the Plant Science Research and Education Unit for general maintenance of the research plots. To my family, thank you for all the support and encouragement you have given me over the years. I would also like to thank Katie for her support and the hours of listening to me talk about this research. I could not have made it to this point without all of you. 4

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS... 4 LIST OF TABLES... 7 LIST OF FIGURES... 9 LIST OF ABBREVIATIONS ABSTRACT CHAPTER 1 GENERAL INTRODUCTION Traffic Tolerance Divot Recovery Cultivar Selection Fertility TRAFFIC TOLERANCE OF EIGHT BERMUDAGRASS CULTIVARS UNDER VARYING FERTILITY RATES Introduction Material and Methods Digital Image Analysis Changes in Turfgrass due to Traffic Treatments Discussion AN ENHANCED METHOD OF TRACKING DIVOT RECOVERY IN TURFGRASS Introduction Materials and Methods Colored Sand Divot Recovery Assessment Using DIA Digital Image Analysis Results Discussion RECUPERATIVE ABILTY OF EIGHT BERMUDAGRASS CULTIVARS Introduction Materials and Methods Results

6 Discussion GENERAL CONCLUSIONS APPENDIX A WEATHER DATA CITRA, FL B EQUIPMENT LIST OF REFERENCES BIOGRAPHICAL SKETCH

7 LIST OF TABLES Table page 2-1 Analysis of variance for percent green cover, dark green color index (DGCI), normalized difference vegetation index (NDVI), visual color, visual quality, and visual density Analysis of variance for change in percent green cover, dark green color index (DGCI), visual color, visual quality, and visual density between traffic treatments Percent green cover of bermudagrass cultivars during Percent green cover of bermudagrass cultivars during Change in percent green cover as a result traffic and fertility treatments in 2009 and Change in percent green cover as a result of traffic treatments in Dark green color index (DGCI) values for bermudagrass cultivars on each collection date during Normalized Difference Vegetation Index (NDVI) values for bermudagrass cultivars during 2009 and Visual ratings of bermudagrass cultivars for color, quality, and density during 2009 and Change in dark green color index (DGCI) values for bermudagrass cultivars on each collection date during 2009 and Change in visual color (Δvisual color) for bermudagrass cultivars on each collection date during 2009 and Change in visual quality (Δvisual quality) for bermudagrass cultivars on each collection date during 2009 and Change in visual density (Δvisual density) for bermudagrass cultivars on each collection date during 2009 and Analysis of variance for divot recovery (%) among methods of analysis (SDA, SDAP, and PSA) Set 1 showing percent recovery by day with standard errors Set 2 showing percent recovery by day with standard errors

8 3-4 Set 3 showing percent recovery by day with standard errors Pearson correlation coefficients compare method of digital image analysis (DIA) with the visual analysis of the corresponding divot. n = Infrared (IR) temperature data for divot sand Analysis of variance for percent recovery of divots in Percent divot recovery over time during set Percent divot recovery over time during set Percent divot recovery over time during set A-1 Monthly averages were obtained from Florida Automated Weather Network (FAWN). Solar Radiation was measured in watts per sq. meter

9 LIST OF FIGURES Figure page 2-1 Percent green cover for eight cultivars of bermudagrass in Color spectrum from SigmaScan Pro software shows hue and saturation values Digital image analysis of one divot on the day of injury Divots analyzed using both methods of DIA Percent recovery for divots filled with pink sand Recovery curves using all methods of analysis Divots were analyzed using sesveral methods of analysis Recovery curves for each of the eight bermudagrass cultivars beginning May 18, 2010 (set 1) Recovery curves for each of the eight bermudagrass cultivars beginning June 24, 2010 (set 2) Recovery curves for each of the eight bermudagrass cultivars beginning July 29, B-1 Modified Cady traffic simulator B-2 Light box that was used to provide uniform light on each plot without shadows B-3 A specialized divot machine was used to produce uniform divots on each plot B-4 Light box that was used to take uniform photos of the divots under consistent light

10 LIST OF ABBREVIATIONS DGCI DIA K N NDVI NTEP P PSA SDA SDAP UF Digital Green Color Index Digital Image Analysis Potassium Nitrogen Normalized Difference Vegetation Index National Turfgrass Evaluation Program Phosphorous Pink Sand Analysis Standard Divot Analysis Standard Divot Analysis with Pink Sand University of Florida 10

11 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Science TRAFFIC TOLERANCE AND DIVOT RECOVERY OF EIGHT BERMUDAGRASS CULTIVARS Chair: Jason Kruse Major: Horticultural Science By Bradley Williams December 2010 Injury to turfgrass occurs in a variety of ways on a golf course or athletic field, whether it is mechanical injury from contact with athletes, compaction as a result of play, or complete removal of turf in the form of divots. Stress or injury as a result of play is termed traffic, and can be detrimental to a playing surface over time, causing turf to thin. In the southern United States, Tifway bermudagrass has been the industry standard on golf courses and athletic fields for more than forty years. In recent years more cultivars of bermudagrass have become available as breeders are constantly taking steps towards increasing the stress tolerances of bermudagrass. This research was designed to examine the traffic tolerance and recovery rate of several new lines of bermudagrass and compare these cultivars to long time industry standards. The objective of the first study was to determine bermudagrass cultivar response to traffic under varying fertility rates. Cultivars consisted of Celebration, Floratex, Hybrid1, Riviera, T11, TifGrand, TifSport, and Tifway. Sub plots received traffic or nontraffic treatments with sub-sub plots receiving 24.4, 36.6, and 48.8 kg N ha -1 per month. Traffic tolerance was measured by visual ratings of color, quality, and density along with 11

12 digital image analysis (DIA) to determine percent cover and dark green color index (DGCI) values. Normalized difference vegetation index (NDVI) values were also collected using multispectral radiometry to determine which cultivar and fertility rate was best suited to handle traffic. Results indicated that Celebration, Hybrid1, T11, and TifGrand were able to tolerate traffic better than industry standards Tifway and TifSport. Celebration consistently yielded the darkest green color while Hybrid1 and T11 maintained the highest density in both years of the study. No differences were perceived among fertility rates in either 2009 or The objective of the second study was to develop a new method of analysis that could accurately calculate the area of a divot regardless of the condition of the surrounding turf. Standard divot analysis (SDA) techniques found in the literature provided analysis with a high degree of variability. To combat this problem, topdressing sand used to backfill divots was colored pink and DIA software was altered to determine the percentage of pink (hue and saturation 0-100) pixels in the photo. Two divots were created per plot in Celebration bermudagrass using a specialized divot machine. One divot was backfilled with standard white topdressing sand and analyzed using the SDA method. The other divot was backfilled with pink colored sand and analyzed using both the standard divot analysis - pink divot (SDAP) method selecting for green pixels and the newly developed pink sand analysis method (PSA). Each divot was also visually analyzed for percent recovery. This study was repeated three times during When comparing to visual analysis of the divots, correlation coefficients were relatively low (avg. r = 0.725). It was found that backfilling divots with pink colored sand and slightly altering the DIA software to select for pink pixels, the percent divot 12

13 recovery could be calculated more accurately. The pink sand analysis method correlated well with visual analysis of the divot (avg. r = 0.955). This new method of analysis proved to be much more accurate and had the ability to precisely measure the area of the divot in less time. The objective of the third study was to examine the differences in divot recovery among eight bermudagrass cultivars under varying fertility and traffic treatments. Cultivars consisted of Celebration, Floratex, Hybrid1, Riviera, T11, TifGrand, TifSport, and Tifway. Sub plots received traffic or non-traffic treatments with sub-sub plots received fertility treatments of 24.4, 36.6, and 48.8 kg N ha -1 per month. This study was repeated three times during the summer of 2010 and analyzed using the pink sand analysis method developed in the previous study. Results indicated that Celebration, Floratex, and T11 performed better than industry standards TifSport and Tifway. TifGrand, a newly released cultivar performed similarly to TifSport and Tifway. No differences were observed between fertility rates or traffic treatments in any of three sets of divots. Celebration, Floratex, and T11 quickly recovered from divot injury, attaining 50% recovery in less than ten days and reaching 100% recovery two to four days sooner than Hybrid1, Riviera, TifGrand, TifSport, and Tifway. 13

14 CHAPTER 1 GENERAL INTRODUCTION There are several criteria that need to be considered when selecting a bermudagrass cultivar (Cynodon dactylon [L.] Pers.; C. dactylon x C. transvaalensis [Burtt-Davy]) for use on a golf course or athletic field. Bermudagrass is commonly established on sports fields in the southern United States because it provides a very dense, green turf cover that is able to tolerate drought and heat as well as a wide variety of soil types, ph, textures, fertilities, and temperatures (Hanna and Maw, 2007). Arguably the most important factor is its ability to withstand injury and re-grow quickly to maintain a high quality playing surface. As play increases, injury from traffic and divots can accumulate causing a stand of turf to thin, creating an unfavorable, unsafe, and unattractive playing surface. Historically, selection of traffic tolerant cultivars has not been a large field of study, but is becoming essential as more cultivars are developed each year (Trappe et al., 2008; 2010b; Williams et al., 2010). The National Turfgrass Evaluation Program (NTEP) is the best method of comparing cultivars because of its national reputation and the high number of cultivar entries into these trials each year. NTEP implemented a traffic tolerance trial in 2009 (Morris, 2009), showing the first step towards gathering information on cultivar differences in traffic tolerance. Selecting a bermudagrass that is genetically able to withstand high amounts of traffic, can give turfgrass managers a competitive advantage from the start. Traffic Tolerance Traffic creates two major stresses to turfgrass. First, the shoots of the plant experience physical injury or wear, and secondly, roots are adversely affected by soil compaction (Carrow and Petrovic, 1992; Carrow and Wiecko, 1989; Trenholm et al. 14

15 2000). The plant s ability to withstand traffic has been linked to nitrogen (N) fertilization (Ebdon et al. 2010; Trenholm et al. 2001a), potassium (K) fertilization (Trenholm et al. 2000), silica (Si) content (Trenholm et al. 2001b), height of cut (Youngner 1961), thatch depth (Dunn et al., 1994; Vavrek, 2002), quantity of shoot tissue (Beard et al., 1981), leaf texture (Williams et al., 2010), water content (Trenholm et al., 2000), and growth rate (Trenholm et al. 1999b; 2000; Youngner, 1962). With all these compounding factors it is difficult to pinpoint which cultural practices are most important to promote traffic tolerance. In past studies, traffic has been applied using a wide range of machinery from rolling drums to cleated aerifiers as a way to simulate traffic stress. Rolling drum traffic simulators include the differential slip wear machine (Canaway, 1976), Georgia soilcompaction wear simulator (Minner and Valverde 2005), and the Brinkman traffic simulator (Cockerham and Brinkman, 1989). These all have cleated rollers and differential skidding action to create shearing stress. The weight of the drums creates the downward force to replicate soil compaction. More recently developed, the Cady traffic simulator (Henderson et al., 2005) works very differently than the rolling drum simulator. The Cady traffic simulator is a modified aerifier that replaces the tines with four feet. These feet are composed of cleated tires that alternately strike the ground creating force in three directions. The Cady traffic simulator has shown to provide the most realistic traffic damage as it causes physical abrasion or wear to crowns, stolons, and leaves while the downward force causes soil compaction. This traffic simulator is able to provide more force and shearing action than other traffic machines. It has the ability to quickly and easily produce similar amounts of traffic as a National Football 15

16 League (NFL) game (Henderson et al. 2005). This amount of traffic makes it possible to distinguish which cultivars are more adapted to handle the high degree of traffic. Determining the traffic tolerance of newly developed bermudagrass cultivars is a necessary science to determine if they are able to outperform industry standards. The ability of turfgrass to tolerate traffic has been measured using several methods. The NTEP visual ratings have been the standard method to rate plots for color, quality, and density. These ratings score many different characteristics of the turf on a scale from 1 to 9 with 6 being the minimal acceptable level (Morris, 2001). Though this method has been used for some time, it has been scrutinized for potential subjectivity and inconsistencies in the ratings (Horst et al., 1984; Skogley and Sawyer, 1992). In direct response, a new method of rating turf plots using digital image analysis (DIA) software was developed with the goal of removing any bias in the rating system. DIA has become a commonly used method to rate turfgrass plots objectively. DIA is conducted using SigmaScan Pro software (v. 5.0, SPSS, Inc., Chicago, IL 60611) and a batch analysis macro (Karcher and Richardson, 2005) to analyze a photo of turfgrass. This method can analyze photos for percent green cover (Richardson et al., 2001) and dark green color index (DGCI) (Karcher and Richardson, 2003). Percent green cover is calculated by counting the number of green pixels in the photo. Work done by Richardson et al. (2001) describes healthy green bermudagrass and zoysiagrass as having color thresholds with hue ranging from 57 and 107 and saturation between 0 and 100. The number of pixels in this target range is then divided by the total number of pixels in the photo to determine percent turf cover. The DGCI value is determined by 16

17 digitally evaluating the hue, saturation, and brightness levels in a photo. The macro then calculates the relative dark green color using the following equation: ( ) ( ) (1-1) DIA allows plots to be quickly and objectively analyzed by a computer. The use of this technology not only allows researchers to more quickly evaluate plots but also provides consistent, unbiased analysis. A third method of analysis to determine turfgrass health uses multispectral radiometry and calculates a corresponding normalized difference vegetation index (NDVI) value. This process uses an optical sensor to measure red light reflectance at 660 and 850 nm. NDVI is then calculated by (R NIR R red )/(R NIR + R red ) where R NIR is the reflectance in the near infrared wavelength (850 nm) and R red is the reflectance in the red wavelength (660 nm) (Johnson et al., 2009). NDVI has been shown in bermudagrass to be closely correlated with irrigation, N fertilization, and turf quality (Xiong et al., 2007). NDVI has also been successfully used to determine traffic stress on both seashore paspalum (Paspalum vaginatum Swartz) and bermudagrass (Trenholm, 1999a). This is another method that works to objectively determine the relative health of the plant. Divot Recovery The ability of turfgrass to recovery from divot injury is another important field of study often related to traffic tolerance. Faster divot recovery can increase playability and appearance of the turf while reducing inputs. The NTEP trials have not traditionally incorporated a measure of recuperative potential (Morris, 2006), making it difficult to draw conclusions about differences among bermudagrass cultivars. Divot recovery has 17

18 historically been a time consuming and laborious process (Fry et al., 2008; Richardson et al., 2001; Ward and Thompson, 1971) making it an undesirable field of research. However, with technological advances, it has become a common topic of study (Austin et al., 1995; Fagerness and Yelverton, 1999; Fry et al., 2008; Karcher et al. 2005a; 2005b; Martin et al., 2001; Trappe et al., 2009; 2010). Divots can now be created much faster and digital images can be used to determine the area of the divot. Fry et al. (2008) describes several types of divot machines that have been produced, enabling researchers to quickly create a large number of divots. Karcher et al. (2005a) developed a technique of using DIA software to calculate the area of a divot, making it easier to analyze large numbers of divots with minimal increase in labor. Utilizing these new technologies, divot research can be conducted more quickly than using conventional methods of visual analysis (Austin et al., 1995) or line intersect analysis (Calhoun, 1996). The DIA method, however, seems to have a large degree of variability, as several researchers have found it difficult to determine differences between treatments (Kaminski et al., 2005; Schmitz et al., 2005). Schmitz et al. (2005) noted a great deal of variability as differences in divot recovery were difficult to find among cultivars. Kaminski et al. (2005) did not explicitly state that their work contained variability in divot recovery rates, yet their data shows some inconsistencies. They visually rated turf that had been subjected to several treatments of growth regulators, for aggressiveness on a 0 to 5 scale with 0 being little to no lateral growth and 5 being strong lateral growth with numerous stolons. Divot recovery was evaluated using DIA software to calculate percent recovery based on the number of green pixels in the photo. Although aggressiveness or lateral growth would 18

19 be directly correlated to divot recovery their data had significant discrepancies. Differences in aggressiveness were found on 12 of 13 analysis dates with a range of ratings from 1.2 to 4.6, showing a strong visual difference among treatments. However, differences in divot recovery were only found on 1 of 13 analysis days during this three month study. This shows a clear lack of correlation between aggressiveness (lateral growth) and divot recovery. Furthermore, percent divot recovery on days 3, 5, 6, 8, and 9 within a single treatment were 47%, 75%, 64%, 83%, and 62% respectively. These data points do not represent recovery well as recovery should never decrease unless additional injury occurs. The inconsistencies between aggressiveness and divot recovery rate as well as the deviation in percent divot recovery shows a high degree of variability in their work, and is likely due to a weakness in the method of analysis. Cultivar Selection Cultivar development is an ongoing science with the goal of finding a single genotype that contains desirable traits while minimizing undesirable traits. Plant breeders are constantly developing and testing new cultivars for increased stress tolerances. Cultivar selection is more important than ever for turfgrass managers since breeders are selecting for highly desirable characteristics displayed in cultivars from all over the world (Baltensperger et al., 1993; Taliaferro, 1992; 1995; Taliaferro et al., 2004;). The following studies will examine several characteristics among eight cultivars of bermudagrass. Two vegetatively propagated cultivars of common bermudagrass (C. dactylon [L.] Pers.) were selected for this study. Celebration bermudagrass was chosen because it s increasing popularity in the southern United States and T11, an experimental bermudagrass line from the University of Georgia (UGA). Two seeded cultivars of bermudagrass were selected. Riviera bermudagrass which has performed 19

20 well in Kentucky (Williams et al. 2010), Oklahoma (Wu et al., 2009) and Arkansas (Trappe et al., 2010a) and Hybrid1 an experimental from UGA. Four vegetatively propagated cultivars of hybrid bermudagrass (C. dactylon [L.] Pers. x C. transvaalensis [Burt Davy]) were also selected including industry standards Tifway, TifSport, and Floratex as well as newly released cultivar TifGrand (Hanna et al. 2010), from the UGA breeding program. Newly developed cultivars Celebration, and TifGrand are promoted to have good traffic tolerance and divot recovery (Sod Solutions, 2006; Vlack, 2010); however, little research has been done to confirm these claims. A comparison of these cultivars will give insight into which turfgrass has the best ability to withstand traffic and recover from divot injury. Fertility Turfgrass nutrition has been linked to traffic tolerance (Ebdon et al. 2010; Trenholm et al. 2001a), although it is unknown how new cultivars of bermudagrass respond to differing fertility levels. Fertility rates on bermudagrass vary greatly depending on desired color, quality, and budget. In the past, bermudagrasses were generally thought to require the same fertility and increased traffic equaled increased fertility need. Trenholm et al. (2000) has shown a relationship between increased potassium concentrations and increased wear tolerance in seashore paspalum and bermudagrass. However, newer cultivars of bermudagrass are being bred for reduced fertility needs. Cultivars have shown to respond differently to fertility treatments and are able to survive with smaller amounts of fertilizer while continually producing dense, dark green cover. This thesis describes the findings of several research projects conducted at the University of Florida during 2009 and 2010 aimed to determine the traffic tolerance and 20

21 divot recovery rate of eight bermudagrass cultivars in Florida. Selecting for improved traffic tolerance in bermudagrass cultivars has the potential to greatly impact turfgrass managers. Higher traffic tolerance has the ability to provide a better playing surface using fewer inputs. Celebration, Floratex, Hybrid1, Riviera, T11, TifGrand, TifSport, and Tifway were treated with three fertility rates to determine if nutrition plays a role in traffic or divot recovery. Additionally, the method of divot analysis commonly used in the literature was found to have a great deal of variability so a new method was created to reduce or eliminate this issue. Chapter three details a new method of analysis that was developed and tested during 2010 with the goal of increasing the accuracy of digital image analysis used in divot research. 21

22 CHAPTER 2 TRAFFIC TOLERANCE OF EIGHT BERMUDAGRASS CULTIVARS UNDER VARYING FERTILITY RATES Introduction Thin areas of turf on a golf course or athletic field are unsightly and have a detrimental effect on the playability of that surface. On a golf course, tee boxes, landing areas and other high traffic zones can quickly deteriorate as the amount of play increases. Similarly, turf coverage can decrease quickly due to greater traffic between the hash marks on a football field or around the goal mouths of a soccer or lacrosse field, resulting in an unsightly and unsafe playing surface. Traffic stress has a detrimental effect on the shoot system of the plant and causes a reduction in overall turf quality (Trenholm et al., 2000). To help combat this decline in the southern United States, superintendents and turfgrass managers are turning to new cultivars of bermudagrass that are marketed as having high wear tolerance and recuperative ability (Sod Solutions, 2006; Vlack, 2010). However, there is limited information comparing the performance of these new bermudagrass cultivars to existing industry standards. New bermudagrass cultivars are constantly being bred and selected for drought tolerance, insect resistance, traffic tolerance, disease resistance, shade tolerance, ideal mowing height, growth rate and other desirable factors. Several new experimental lines of bermudagrass developed at the University of Georgia have shown in preliminary testing to outperform industry standards in many management aspects. One characteristic yet to be thoroughly evaluated is the traffic tolerance and recuperative ability of these new lines under differing fertility rates. This study was set up to compare genetic color, density, and turf quality of eight bermudagrass cultivars. Genotypes included industry standards Tifway, TifSport, Floratex, and Riviera. These were 22

23 compared with two newly released cultivars Celebration and TifGrand (Hanna et al., 2010) as well as experimental cultivars T11, and Hybrid1. This was accomplished by applying simulated traffic on a weekly basis and tracking percent green cover, digital green color index (DGCI), normalized difference vegetation index (NDVI), visual color, visual quality, and density. Wear tolerance was measured with the use of digital image analysis (DIA) software (SigmaScan, v. 5.0, SPSS, Inc., Chicago, IL 60611) and the "Turf Analysis" macro (Karcher and Richardson, 2005). Trappe et al. (2008) previously studied the differences in wear tolerance of some bermudagrass cultivars in Arkansas. However, since that time, new cultivars have been released that need to be evaluated for their ability to withstand traffic in the Deep South. The objectives of this study were (i) to characterize a typical response of these bermudagrass genotypes to varying rates of traffic, (ii) to establish fertility recommendations for each of the genotypes studied under high- and low-traffic conditions, and (iii) to determine which cultivars keep the highest density, quality, and darkest green color when subjected to traffic. Material and Methods This study was conducted at the University of Florida Plant Science Research and Education Unit located near Citra, Florida. The study began on August 1, 2009 and continued through September 1, Plots were maintained similarly to a golf course fairway or athletic field and mowed three times a week at a 1.2 cm height of cut. Treatments were arranged in a split-split plot design with three replications. Whole plots were 18.2 m 2 (4.26 m x 4.26 m) and consisted of the following eight bermudagrass cultivars: Tifway, TifSport, Floratex, Riviera, Celebration, TifGrand, Hybrid1, and T11. Sub-plots were 9.1 m 2 (4.26 m x 2.13 m) and consisted of the following traffic treatments: 1) no traffic; and 2) simulated traffic applied weekly. Fertility treatments 23

24 were randomized within each sub-plot and were applied at 24.4, 36.6, or 48.8 kg N ha -1 per month. Sub-sub-plots measured 3.0 m 2 (2.13 m x 1.42 m) and fertilized with turf fertilizer blend containing 50% slow release nitrogen in a split application every two weeks. In addition to N-P-K the fertilizer contained 2% Mg, 1% Fe, and 1% Mn. Traffic was applied weekly throughout the growing season using a modified Cady traffic simulator (Henderson et al., 2005) that was constructed using a John Deere Aercore 800 aerifier (Appendix B-1). The 2009 traffic treatment began on August 1 and continued through November 14. In 2010 traffic treatments were applied during two separate periods, from May 18 to June 16, and from July 28 to August 25. Traffic was not applied continuously during the summer of 2010 due to unexpected equipment failure. Visual ratings of color, quality, and density were collected along with digital pictures that were processed using DIA to determine percent green cover (Richardson et al. 2001) and DGCI (Karcher and Richardson, 2003). Visual ratings were based on a 1 to 9 scale with 9 representing optimum healthy turf, 6 representing an acceptable color or quality turf for athletic fields and golf courses, and 1 representing brown or dead turf (Morris, 2001). Normalized difference vegetation index (NDVI) values were also collected biweekly with a Crop Circle ACS-210 (Holland Scientific Inc., Lincoln, NE 68516). NDVI was calculated as (R NIR R red )/(R NIR + R red ) where R NIR is the reflectance in the near infrared wavelength (850 nm) and R red is the reflectance in the red wavelength (660 nm) (Johnson et al., 2009). Digital Image Analysis Digital images of each plot were collected twice monthly using a Sony Cybershot camera, model DSC-H10 (Sony Corp., New York, NY 10022). Camera settings consisted of ISO100, 1/40 second shutter speed, and aperture of F3.5. The camera 24

25 was mounted on a large portable light box which provided an enclosed area with artificial light to ensure consistency and minimal shadowing. The light box measured 53 x 61 x 51 cm with four ten-watt compact florescent, 6500 kelvin daylight bulbs mounted to the ceiling to cast consistent light onto the target area (Appendix B-2). The light box ensured uniformity of both light and camera height so all pictures were identical. Photos were downloaded to a personal computer and resized to an area of 600x800 pixels using ACDSee Pro (v. 2.5., ACD Systems International Inc., Victoria, British Columbia). This was done in preparation for analysis as the macro cannot readily process images larger than one megapixel (Karcher and Richardson, 2005). Percent green cover was calculated by SigmaScan Pro software (v. 5.0, SPSS, Inc., Chicago, IL 60611) using the Turf Analysis macro (Karcher and Richardson, 2005). The program counted the number of pixels in the target range (hue ; sat 0-100) and divided by the total number of pixels in the photo. Change in percent green cover was determine by subtracting percent green cover of traffic plots from untrafficked plots. This made it possible to track the effect that traffic had on percent cover, independently of other factors. DGCI values were determined using the same software and macro. The DGCI was created by Karcher and Richardson (2003) to measure the relative dark green color of an image using the following equation: ( ) ( ) (2-1) Changes in turfgrass due to traffic treatments were determined by calculating the delta (Δ) value between traffic treatments within each fertility and cultivar treatment. Values of percent cover, DCGI, visual color, visual quality, and visual density values for a specific fertility and cultivar combination under no traffic were subtracted from the 25

26 same treatment receiving traffic. This allowed the effects of traffic to be pulled out, removing any compounding effects of cultivar or fertility. An analysis of variance (PROC ANOVA) was performed to test treatment effects on wear tolerance for percent green cover, DCGI, NDVI, visual color, visual quality, and visual density (Table 2-1) along with the changes in turf due to traffic (Δpercent green cover, ΔDCGI, Δvisual color, Δvisual quality, and Δvisual density) (Table 2-2). Additionally, data were analyzed with General Linear Model regression (PROC GLM) with the appropriate error terms to test relationships between wear treatments and turf color, quality, density, DGCI, percent green cover, and delta values (SAS, 2008). Mean separation was determined by date through Duncan s means separation at P = Results In both years of the study, the driving factor between turfgrass performance was genetic, as the greatest differences were seen among bermudagrass cultivars (P < ). No differences were observed in visual turf quality, color, density, or NDVI value as a result of fertility treatments. This is supported by digital image analysis which also indicated no relationship between fertility rate and percent green cover or DGCI values. Cultivar performance did not differ among fertility rates, therefore data was analyzed across fertility treatments. Analysis of variance indicated a significant interaction between days after initial treatment (DAIT), cultivar, and traffic in 2009 and between DAIT and cultivar in 2010 (Table 2-1). In 2009, under no traffic, TifGrand was in the top statistical category on all nine collection dates (Table 2-3). Hybrid1 and Celebration were in the top category on 26

27 7 and 6 dates, respectively. Floratex, TifSport, and Tifway had the lowest percent cover and were only in the top category for the first 3 collection dates (Table 2-8). When subjected to traffic, Hybrid1 and TifGrand kept the greatest percent green cover (Table 2-3). Celebration and Riviera were in the next statistical category while the other cultivars lost significantly more coverage throughout the season as a result of traffic stress (Table 2-3). Plots receiving simulated traffic treatments declined from 0 to 20% during the growing season (Figure 2-1). All cultivars declined in density beginning around day 80 and continuing through day 101. In 2010, Celebration and T11 were in the top statistical category at every collection date, whereas, Hybrid1, Riviera, and TifGrand were in the top category on all but three dates. In both 2009 and 2010, the Celebration, Hybrid1, T11, and TifGrand cultivars all performed better than the Tifway, TifSport, and Floratex cultivars (Table 2-4). Results of statistical tests of dark green color show DAIT by cultivar interaction in 2009 and 2010 (Table 2-1). In both years, Celebration and T11 had the darkest green color on nearly all collection dates (Table 2-7). Hybrid1 and T11 were in the second statistical category showing relatively dark green color on several collection dates. Floratex, Riviera, TifSport, and Tifway were among the lowest performing cultivars on almost all dates in both years (Table 2-7). Cultivar by traffic interactions were observed in NDVI values and means are displayed in Table 2-8. Under no traffic, Hybrid1; and TifGrand and T11 had the highest NDVI values during 2009 and 2010, respectively. Under traffic conditions Hybrid1 had the highest value in 2009 and T11 in Cultivars Floratex, TifSport and Tifway were 27

28 again in the lowest statistical category under both traffic treatments in both years of the study. Analysis of variance indicated an interaction between cultivar and traffic treatments (P < 0.05) for visual color, quality, and density during both years of the study (Table 2-1). No differences were observed between fertility treatments during either 2009 or Visual color, quality, and density ratings differed among cultivars within traffic treatments. TifGrand kept the best visual color, under both traffic treatments in Celebration, Hybrid1, T11 and Tifway provided the second best color under traffic treatments. All cultivars performed similarly under each traffic treatment in 2010 with the exception of Floratex which provided a lighter green turf, however still above minimum acceptable levels (Table 2-9). Evaluations in 2009 of turf quality showed, without traffic, Celebration, Hybrid1, T11, TifGrand, and TifSport maintained the highest quality turf. In 2010, there was less separation between cultivars with all cultivars performing better than Floratex and Riviera under no traffic, and better than Floratex under simulated traffic conditions. Again, all cultivars were above minimum acceptable levels for turf quality. In both years of the study, traffic reduced visual turf quality by 0.5 to 1.0 points across all cultivars (Table 2-9). Changes in Turfgrass due to Traffic Treatments To determine the effect that traffic had on the turf, each fertility and cultivar combination receiving traffic was compared with the same treatment under no traffic. Analysis of variance indicated DAIT by cultivar interaction for ΔDGCI, Δvisual color, Δvisual quality, and Δvisual density for both years of the study (Table 2-2). Interaction 28

29 between cultivar and fertility was observed for Δpercent green cover in both years and additionally in 2009, cultivar by DAIT interaction. Change in percent green cover in 2009 showed Floratex, Hybrid1, Riviera, T11, and Tifway performed better than Celebration, TifGrand, and TifSport (Table 2-6). Fertility by cultivar interaction was also observed in both years of the study. Celebration and TifSport were in the lowest statistical category in 2009 in both the high and low fertility treatments (Table 2-5). In 2010 there was no notable separation between cultivars or fertility treatments (Table 2-5). Analysis of ΔDGCI indicated Celebration and T11 performed the best in 2009, losing very little dark green color as a result of traffic stress. In 2010, Celebration, Floratex, Riviera, TifGrand, and TifSport all performed better than other cultivars (Table 2-10). Hybrid1, T11, and Tifway did not perform as well, becoming lighter green when subjected to traffic stress. Change in visual evaluations as a result of traffic showed cultivar by DAIT interaction for visual color, quality, and density in both years of the study (Table 2-2). Analysis of Δvisual color ratings in 2009 indicated Celebration, Floratex, Hybrid1, Riviera, T11, and Tifway performed better than TifGrand and TifSport (Table 2-11). In 2010, Celebration, Floratex, Riviera, TifGrand, and Tifway all performed better than Hybrid1, T11, and TifSport, losing less visual color as a result of traffic (Table 2-11). ΔVisual quality ratings showed Hybrid1, Riviera, T11, and TifGrand performed better than other cultivars in 2009 and all cultivars performed better than Floratex and TifSport in 2010 (Table 2-12). Ratings of visual density in 2009 indicated Hybrid1, Riviera, and TifGrand performed well on almost all collection dates (Table 2-13). In 2010, all 29

30 cultivars performed better than Floratex and TifSport as they had the most days in the lowest statistical category (Table 2-13). Discussion In this study there was not a single cultivar that performed the best in every aspect of evaluation. Instead, several cultivars performed well in multiple areas of assessment. Celebration, Hybrid1, T11, and TifGrand all performed well under both traffic and notraffic treatments. Celebration produced high visual ratings, consistently above the minimum level of acceptability, and yielded the darkest green color of any of the cultivars evaluated according to DGCI analysis. When evaluating cultivars for percent cover, all cultivars began the season in the same statistical category (data not shown). As traffic damage continued throughout the season, cultivars began to lose coverage, leaving only the most traffic-tolerant cultivars with high density stands. Hybrid1 and TifGrand performed the best in both years of the study while Celebration, Riviera, and T11 performed well in one of the two years. Tifway, TifSport, and Floratex consistently rated in the lowest statistical category for percent coverage. Once traffic damage began to accumulate, these cultivars could not recover fast enough and percent cover greatly declined. In 2009, plots receiving simulated traffic treatments declined by up to 20% (Figure 2-1). However, traffic did not cause percent green cover to decline as much in This may have been due to the increased growth rates during the warmer summer months keeping density high even when subjected to weekly traffic treatments (Appendix A). Another factor in 2010 was that traffic was not able to be applied during the month of July due to equipment failure. This allowed trafficked plots to fully recover before they started receiving traffic again the first week in August. 30

31 Differences among cultivars were also supported by NDVI data in both 2009 and NDVI is an optical sensory value that has been shown in bermudagrass to be closely correlated with irrigation, N fertilization, turf quality (Xiong et al., 2007) as well as traffic stress (Trenholm et al., 1999a). In this study NDVI did not correlate closely with fertility rate, however, it did seem to be related to overall turf quality. Hybrid1 and T11 performed well in 2009 and 2010 respectively, supporting the superior visual quality ratings for these cultivars, revealing that these grasses are less stressed under traffic than cultivars that are currently used by the industry. Changes in turf due to traffic were evaluated by calculating delta values between traffic treatments and comparing cultivar and fertility effects. Again it was found that not one cultivar performed best in all categories however, several cultivars performed well in multiple categories. It was found that in 2009, cultivars Hybrid1, Riviera, and T11 were affected the least by traffic over all evaluations. The change in percent cover, DGCI, and visual ratings were not greatly affected by traffic treatments in these cultivars. Celebration and TifSport were affected the most by traffic as shown by the reduction in color, quality, and density of these cultivars as a result of traffic stress. In 2010, there were fewer differences between cultivars with all performing similarly besides Floratex. Floratex was in a lower statistical category more often than any other cultivar in the study suggesting that traffic has a much more detrimental effect on this cultivar. There were no significant differences among fertility rates for any of the evaluation parameters. This result indicates that fertility rates may have been too narrow in range to establish differences in growth, color, or density response. Trenholm et al. (2000) 31

32 reports mechanisms for wear tolerance include higher shoot density, K concentration, and leaf moisture. However, in this study, no differences were seen even though K concentrations varied among treatments. This suggests that sufficient K concentrations may have been achieved at the lowest fertility level, allowing all rates to respond similarly to traffic treatments. Carroll and Petrovic (1991) similarly found no differences in traffic tolerance of Kentucky bluegrass despite variation in N and K fertility. Visual color differences were observed among fertility treatments during the spring of 2010 however no other significant differences were seen during the period of the study. Future research needs to be done with a wider range of fertility levels to determine the appropriate fertilization regime. However, this study has shown that although Tifway has been the industry standard in Florida for over forty years, there are now other, better options for high traffic areas. Celebration, Hybrid1, T11 and TifGrand all perform very well in this climate, providing a very dense, green stand of turf that can withstand a great deal of traffic. 32

33 Table 2-1. Analysis of variance for percent green cover, dark green color index (DGCI), normalized difference vegetation index (NDVI), visual color, visual quality, and visual density for 2009 and Traffic Tolerance ANOVA % Green Cover DGCI NDVI Effect df Pr > F Pr > F Pr > F Pr > F Pr > F Pr > F Days after initial 1 <0.001 < <0.001 <0.001 <0.001 treatment (DAIT) Cultivar 7 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 Fertility Traffic <0.001 DAIT*Cultivar 7 < <0.001 < DAIT*Traffic Cultivar*Traffic 7 < < DAIT*Cultivar*Traffic Visual Color Visual Quality Visual Density Effect df Pr > F Pr > F Pr > F Pr > F Pr > F Pr > F Days after initial < <0.001 <0.001 <0.001 treatment (DAIT) Cultivar 7 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 Fertility Traffic < <0.001 DAIT*Cultivar 7 <0.001 <0.001 < < DAIT*Traffic Cultivar*Traffic 7 < < < DAIT*Cultivar*Traffic Full ANOVA table is not shown; all interactions not displayed here had Pr > F values above

34 Table 2-2. Analysis of variance for change in percent green cover, dark green color index (DGCI), visual color, visual quality, and visual density between traffic treatments in 2009 and Traffic Tolerance ANOVA Δ % Green Cover Δ DGCI Δ Visual Color Δ Visual Quality Δ Visual Density Effect df Pr > F Pr > F Pr > F Pr > F Pr > F Pr > F Pr > F Pr > F Pr > F Pr > F Days after initial 1 < <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 <0.001 treatment (DAIT) Cultivar 7 < <0.001 <0.001 <0.001 <0.001 Fertility < < < <0.001 DAIT*Cultivar 7 < <0.001 <0.001 DAIT*Fertility Cultivar*Fertility < DAIT* Cultivar*Fertility

35 Table 2-3. Percent green cover of bermudagrass cultivars during Percent Green Cover 2009 No Traffic Days Cultivar Celebration 74.3 ab 77.1 ab 91.3 a 95.7 ab 89.2 a 84.8 ab 95.6 bc 75.0 b 73.5 bc Floratex 70.7 ab 81.2 ab 84.7 b 89.8 c 89.7 a 71.4 cd 95.0 bc 68.8 c 68.9 cd Hybrid b 71.4 ab 90.1 a 91.5 abc 93.0 a 82.9 b 96.7 ab 77.9 b 78.9 ab Riviera 80.0 a 82.8 ab 91.7 a 92.5 abc 90.3 a 70.5 cd 93.9 c 67.1 cd 67.7 cd T ab 69.3 b 89.9 a 93.0 abc 92.4 a 76.6 bc 94.2 c 63.5 de 63.3 de TifGrand 68.0 ab 79.4 ab 95.4 a 96.6 a 94.0 a 92.2 a 97.8 a 85.5 a 83.4 a TifSport 79.9 a 83.6 ab 91.7 a 89.9 bc 83.6 b 63.5 d 88.5 e 61.5 e 60.1 e Tifway 73.3 ab 85.0 a 90.9 a 90.7 bc 80.7 b 70.9 cd 91.3 d 60.2 e 59.6 e Traffic Days Cultivar Celebration 54.9 b 59.7 c 71.5 e 89.7 a 87.4 ab 87.6 a 94.4 ab 70.3 a 57.5 bc Floratex 68.0 a 80.9 a 78.4 d 84.3 b 84.5 bc 74.6 cd 94.1 ab 64.9 ab 55.1 bcd Hybrid a 85.4 a 89.6 abc 93.1 a 89.6 ab 79.3 bc 96.1 a 69.8 a 65.4 a Riviera 71.1 a 84.0 a 89.4 abc 89.9 a 88.9 ab 70.4 d 92.4 b 60.1 bc 49.8 cde T a 66.0 bc 86.6 bc 89.8 a 87.0 ab 79.2 bc 94.3 ab 61.3 bc 54.1 bcd TifGrand 71.2 a 88.2 a 94.0 a 90.7 a 91.1 a 84.5 ab 95.8 ab 67.2 a 60.5 ab TifSport 68.7 a 70.4 b 84.2 cd 76.8 c 75.5 d 62.8 e 85.1 d 49.5 d 43.6 e Tifway 80.6 a 89.2 a 91.1 ab 83.2 b 81.4 c 68.6 de 88.6 c 56.6 c 48.8 de Percent green cover was calculated using digital image analysis software selecting pixels with hue and saturation Means with the same letter in a given year and traffic treatment were not significantly different (P = 0.05) according to Duncan s means separation. 35

36 Table 2-4.Percent green cover of bermudagrass cultivars during Percent Green Cover 2010 Days Cultivar Celebration 55.1 ab 73.1 a 85.9 bc 86.0 abc 96.2 ab 97.4 a 95.6 a 95.2 ab 95.0 a 97.7 a Floratex 42.9 c 65.6 b 88.4 abc 88.4 abc 90.2 c 74.6 c 68.8 c 79.7 d 87.6 b 83.4 e Hybrid ab 74.9 a 90.4 ab 94.4 a 98.0 a 91.0 b 80.9 b 94.6 ab 89.0 b 91.5 bcd Riviera 54.9 ab 77.2 a 84.2 c 87.1 abc 93.9 abc 92.6 ab 90.2 a 95.3 ab 95.5 a 91.6 bcd T a 76.6 a 93.4 a 93.5 ab 98.3 a 97.5 a 96.1 a 97.8 a 96.2 a 93.2 abc TifGrand 47.2 bc 71.1 ab 89.3 abc 78.2 c 93.3 bc 96.0 ab 92.5 a 92.3 abc 94.6 a 96.2 ab TifSport 48.4 bc 71.8 ab 85.6 bc 82.0 bc 97.9 a 91.3 b 75.2 bc 87.6 c 88.3 b 90.0 cd Tifway 46.5 bc 68.0 ab 83.5 c 83.9 abc 95.4 ab 94.1 ab 76.8 b 90.7 bc 89.9 b 87.5 de Percent green cover was calculated using digital image analysis software selecting pixels with hue and saturation Means with the same letter in a given year and traffic treatment were not significantly different (P = 0.05) according to Duncan s means separation. 36

37 Table 2-5. Change in percent green cover as a result traffic and fertility treatments in 2009 and Relative Reduction in Percent Green Cover 2009 Fertility Rate Cultivar 24.4 Kg N ha Kg N ha Kg N ha -1 Celebration b -7.6 a -8.1 cd Floratex -3.0 a -4.3 a -4.8 abcd Hybrid1 0.0 a -2.8 a -0.2 ab Riviera -5.2 a -2.6 a -6.9 bcd T a -3.5 a -2.3 abc TifGrand -5.0 a -7.1 a -4.2 abcd TifSport b -6.7 a -9.5 d Tifway -1.0 a -3.8 a 0.4 a 2010 Fertility Rate Cultivar 24.4 Kg N ha Kg N ha Kg N ha -1 Celebration -0.9 a 1.7 a 2.0 a Floratex -2.2 a -1.5 a -1.6 a Hybrid1-0.1 a -2.2 a -0.2 a Riviera -0.4 a -3.5 a 1.9 a T a -1.9 a -1.2 a TifGrand -0.9 a -0.3 a -1.4 a TifSport -2.3 b -1.3 a 1.1 a Tifway -1.8 a -1.6 a 0.9 a Percent green cover was calculated using digital image analysis software selecting pixels with hue and saturation Means with the same letter in a given year and traffic treatment were not significantly different (P = 0.05) according to Duncan s means separation. 37

38 Table 2-6. Change in percent green cover as a result of traffic treatments in Relative Reduction in Percent Green Cover 2009 Days After Initial Treatment Cultivar Celebration c d b -6.1 b -1.8 ab 2.8 ab -1.1 a -4.7 a ab Floratex -2.7 ab -0.3 abc -6.3 a -5.5 ab -5.2 ab 4.1 a -0.9 a -3.9 a ab Hybrid1 5.4 a 14.0 a -0.4 a 1.5 a -3.3 ab -3.6 ab -0.6 a -8.1 ab ab Riviera -8.8 abc 1.2 abc -2.3 a -2.6 ab -1.4 ab -2.8 ab -1.5 a -7.0 ab ab T ab -3.3 bcd -3.4 a -3.2 ab -5.4 ab 2.6 ab 0.2 a -2.2 a -9.2 a TifGrand 3.1 ab 8.9 ab -1.4 a -5.9 b -2.9 ab -7.9 b -2.0 a c b TifSport bc cd -7.5 a c -8.1 b -2.6 ab -3.4 a b ab Tifway 7.3 a 4.1 ab 0.2 a -7.5 bc 0.8 a -2.4 ab -2.7 a -3.6 a a Percent green cover was calculated using digital image analysis software selecting pixels with hue and saturation Means with the same letter in a given year and traffic treatment were not significantly different (P = 0.05) according to Duncan s means separation. 38

39 Table 2-7. Dark green color index (DGCI) values for bermudagrass cultivars on each collection date during Dark Green Color Index Days Cultivar Celebration a a a a a a a a a Floratex c a d d b c b d b Hybrid ab a b b a b a b a Riviera bc a c d b c b cd b T bc a bc c a c b cd b TifGrand ab a a bc a a a ab a TifSport bc a bc d bc c b bc a Tifway bc a bc d c c b cd b Days Cultivar Celebration a a a a a a a a a Floratex d c c b c d e e e Hybrid b b ab a b c cd bc de Riviera c b c b c c bc bc cd T b bc ab a a a a ab ab TifGrand c b bc b b ab b ab bc TifSport cd c c b b c de d de Tifway d c c b b bc e cd cd Dark green color index (DGCI) value was calculated using: DGCI = [(Hue-60)/60 + (1-Saturation) + (1-Brightness)]/3. Means with the same letter in a given year and traffic treatment were not significantly different (P = 0.05) according to Duncan s means separation. 39

40 Table 2-8. Normalized Difference Vegetation Index (NDVI) values for bermudagrass cultivars during 2009 and Normalized Difference Vegetation Index No Traffic Traffic Cultivars Celebration b d e c Floratex e f e e Hybrid a b a b Riviera c c c b T d a d a TifGrand a d b c TifSport f e f e Tifway f f e d Normalized difference vegetation index (NDVI) value was calculated using: NDVI = (RNIR R red )/(R NIR + R red ). Means with the same letter in a given year and traffic treatment were not significantly different (P = 0.05) according to Duncan s means separation. 40

41 Table 2-9. Visual ratings of bermudagrass cultivars for color, quality, and density during 2009 and Visual Ratings No Traffic Visual Color Visual Quality Visual Density Cultivars Celebration 7.1 b 6.8 a 7.4 ab 6.9 ab 98.8 a 93.4 ab Floratex 6.7 d 6.4 b 7.0 d 6.5 c 97.5 a 91.3 b Hybrid1 7.2 b 7.0 a 7.3 abc 6.9 ab 97.3 a 93.5 ab Riviera 6.9 c 6.8 a 7.1 cd 6.8 bc 95.6 b 93.1 ab T b 7.2 a 7.5 a 7.2 a 98.1 a 94.7 a TifGrand 7.6 a 7.1 a 7.4 ab 7.0 ab 94.6 b 93.8 ab TifSport 7.1 b 6.9 a 7.4 ab 7.0 ab 97.5 a 94.7 a Tifway 7.1 b 6.8 a 7.2 bcd 6.9 ab 95.6 b 93.4 ab Traffic Visual Color Visual Quality Visual Density Celebration 6.8 bc 6.8 a 6.5 bc 6.7 a 89.1 abc 91.9 a Floratex 6.5 d 6.3 b 6.3 c 6.2 b 87.9 bc 86.9 b Hybrid1 7.0 b 7.0 a 6.8 a 6.8 a 91.2 a 93.0 a Riviera 6.6 cd 6.8 a 6.6 ab 6.6 ab 91.4 a 98.9 ab T b 7.1 a 6.8 a 7.0 a 91.0 ab 92.9 a TifGrand 7.3 a 7.0 a 6.7 ab 6.8 a 88.3 abc 94.5 a TifSport 6.6 cd 6.7 a 6.3 c 6.6 ab 84.1 d 90.5 a Tifway 6.8 bc 6.8 a 6.5 bc 6.8 a 87.2 c 92.0 a Average of biweekly rating in 2009 and Scale is from 1 to 9, 9=optimum turf color, 6=acceptable turf color. Average of biweekly rating in 2009 and Scale is from 1 to 9, 9=optimum turf quality, 6=acceptable turf quality. Means with the same letter in a given year and traffic treatment were not significantly different (P = 0.05) according to Duncan s means separation. 41

42 Table Change in dark green color index (DGCI) values for bermudagrass cultivars on each collection date during 2009 and Relative Reduction in Dark Green Color Index (DGCI) Days Cultivar Celebration a a b a a ab ab a ab Floratex a a a ab d abc abc ab a Hybrid a a a a bcd c abcd ab abc Riviera a a a a abc bc cd ab bc T a a a a cd a a ab ab TifGrand a a a ab abcd c bcd b c TifSport a a a b bcd abc d ab c Tifway a a a ab ab c cd ab abc Days Cultivar Celebration a a b a ab a ab ab ab Floratex ab a ab a ab a a ab ab Hybrid ab a b a b a abc b ab Riviera ab a ab a a a bc a a T b a ab a b a ab ab ab TifGrand b a ab a ab a ab ab ab TifSport ab a a a ab a abc ab ab Tifway ab a ab a ab a c ab b Dark green color index (DGCI) value was calculated using: DGCI = [(Hue-60)/60 + (1-Saturation) + (1-Brightness)]/3. Means with the same letter in a given year and traffic treatment were not significantly different (P = 0.05) according to Duncan s means separation. 42

43 Table Change in visual color (Δvisual color) for bermudagrass cultivars on each collection date during 2009 and Relative Reduction in Turf Color Days Cultivar Celebration -0.2 a -0.2 b -0.3 a -0.1 a -0.3 a 0.0 ab 0.0 ab -0.6 ab -1.2 ab Floratex 0.0 a -0.1 b -0.1 a 0.0 a -0.4 a 0.2 a 0.0 ab -0.2 a -1.4 ab Hybrid1 0.1 a 0.2 ab 0.1 a -0.2 a -0.6 ab 0.1 ab -0.1 abc -0.6 ab -1.0 a Riviera 0.0 a 0.0 ab -0.1 a -0.1 a -0.4 a -0.2 b -0.1 abc -0.6 ab -1.2 ab T a 0.1 ab 0.0 a -0.1 a -0.7 ab 0.1 ab 0.1 a -0.4 ab -1.0 a TifGrand 0.1 a 0.4 a 0.0 a -0.3 a -0.3 a -0.1 ab -0.4 c -0.8 bc -1.2 ab TifSport -0.1 a 0.0 ab -0.1 a 0.0 a -0.9 b -0.1ab -0.3 bc -1.2 c -1.5 b Tifway 0.2 a 0.2 ab -0.1 a -0.2 a -0.6 ab -0.1 ab -0.4 c -0.4 ab -1.1 ab Days Cultivar Celebration -0.2 a 0.1 a 0.0 a -0.1 ab 0.0 a -0.1 a 0.0 a -0.1 ab -0.1 a Floratex -0.2 a -0.1 a -0.2 a -0.2 ab -0.2 a -0.3 a -0.1 a -0.1 ab 0.1 a Hybrid1-0.1 a -0.1 a 0.1 a 0.2 a -0.1 a -0.3 a -0.1 a -0.2 b 0.1 a Riviera -0.1 a -0.1 a -0.2 a 0.1 ab -0.3 a -0.3 a -0.1 a -0.1 ab 0.1 a T a -0.2 a -0.1 a -0.3 b -0.2 a -0.1 a -0.1 a 0.0 ab 0.0 a TifGrand -0.1 a 0.1 a -0.1 a -0.1 ab -0.1 a -0.2 a -0.2 a 0.0 ab -0.1 a TifSport -0.7 b -0.4 a -0.2 a -0.2 ab -0.1 a -0.3 a -0.1 a -0.1 ab 0.2 a Tifway -0.2 a -0.3 a -0.1 a 0.1 ab 0.1 a -0.1 a -0.2 a 0.1 a 0.2 a Average of biweekly rating in 2009 and Scale is from 1 to 9, 9=optimum turf color, 6=acceptable turf color Means with the same letter in a given year and traffic treatment were not significantly different (P = 0.05) according to Duncan s means separation. 43

44 Table Change in visual quality (Δvisual quality) for bermudagrass cultivars on each collection date during 2009 and Relative Reduction in Turf Quality Days Cultivar Celebration -0.7 c -0.2 ab -0.8 b -0.9 c -0.9 a -0.9 ab -1.1 c -0.8 a -2.1 a Floratex -0.1 ab -0.3 b -0.2 a -0.6 bc -0.8 a -0.6 a -0.7 abc -0.9 a -2.4 a Hybrid1 0.3 a 0.3 ab 0.1 a -0.4 abc -0.8 a -0.6 a -0.4 ab -1.1 a -2.0 a Riviera 0.0 ab 0.0 ab 0.0 a 0.0 a -0.9 a -0.4 a -0.4 ab -1.0 a -1.8 a T ab 0.1 ab -0.4 ab -0.8 c -1.1 a -0.6 a -0.3 a -0.9 a -1.8 a TifGrand -0.1 ab 0.4 a -0.2 a -0.2 ab -1.0 a -0.7 a -1.0 bc -1.1 a -2.2 a TifSport -0.3 bc 0.0 ab -0.4 ab -0.8 c -1.3 a -1.3 b -1.3 c -1.7 b -2.3 a Tifway 0.1 ab 0.3 ab 0.1 a -0.6 bc -1.2 a -0.9 ab -1.2 c -1.1 a -1.9 a Days Cultivar Celebration -0.1 ab -0.3 ab -0.1 a -0.1 a -0.3 a -0.4 a 0.0 a 0.0 a 0.0 a Floratex -0.7 bc -0.2 ab -0.3 a -0.2 a -0.7 a -0.3 a -0.1 a -0.1 a 0.0 a Hybrid1 0.0 a -0.1 ab -0.1 a 0.1 a -0.2 a -0.3 a 0.0 a -0.1 a 0.0 a Riviera -0.4 abc -0.4 ab -0.2 a -0.1 a -0.3 a -0.7 a 0.0 a 0.0 a 0.1 a T abc -0.1 a -0.2 a -0.4 a -0.6 a -0.3 a 0.1 a 0.0 a -0.1 a TifGrand -0.2 ab 0.1 a -0.1 a 0.1 a -0.1 a -0.2 a -0.1 a 0.0 a 0.0 a TifSport -0.8 c -0.8 b -0.4 a -0.1 a -0.3 a -0.8 a 0.0 a -0.1 a -0.1 a Tifway -0.3 abc -0.6 ab -0.2 a 0.1 a -0.2 a -0.5 a -0.1 a 0.1 a 0.1 a Average of biweekly rating in 2009 and Scale is from 1 to 9, 9=optimum turf quality, 6=acceptable turf quality. Means with the same letter in a given year and traffic treatment were not significantly different (P = 0.05) according to Duncan s means separation. 44

45 Table Change in visual density (Δvisual density) for bermudagrass cultivars on each collection date during 2009 and Relative Reduction in Visual Density Days Cultivar Celebration -9.4 c -3.9 b -5.0 b -7.8 bc -5.6 a -7.8 ab c a ab Floratex -0.6 ab -3.9 b -3.3 b -7.8 bc -8.3 ab -7.2 ab bc a bc Hybrid1 2.8 a 1.1 ab -1.1 ab -7.2 bc -3.9 a -5.6 a -3.9 a a ab Riviera 1.1 ab 3.9 a 2.2 a -1.1 a -5.0 a -3.3 a -3.3 a a ab T b 0.0 ab -1.1 ab -6.7 bc -7.8 ab -3.9 a -4.4 ab a ab TifGrand 1.1 ab 5.6 a -2.2 ab -4.4 abc -6.7 a -5.6 a bc a a TifSport -3.3 b -1.1 ab -4.4 b -9.4 c c b c b c Tifway 0.0 ab 3.3 ab 1.7 a -2.8 ab bc ab c a ab Days Cultivar Celebration -2.2 ab -0.6 ab -0.6 a 0.0 a -1.7 ab -3.9 a 0.0 a 0.0 a 0.0 a Floratex bc -6.7 b -8.3 b -5.0 a -6.7 b -5.0 a 0.0 a 0.0 a 0.0 a Hybrid1-1.1 a -0.6 ab 0.0 a 0.0 a -1.7 ab -1.1 a 0.0 a 0.0 a 0.0 a Riviera -8.3 abc -5.0 ab -4.4 ab -1.7 a -5.0 ab -6.7 a 0.0 a 0.0 a 0.0 a T ab -1.7 ab -2.8 ab -2.8 a -1.7 ab 0.0 a 0.0 a 0.0 a 0.0 a TifGrand -5.0 ab 3.9 a -1.1 a -0.6 a 0.0 a 0.0 a 0.0 a 0.0 a 0.0 a TifSport c -7.2 b -5.6 ab -1.7 a -2.2 ab -3.9 a 1.7 a 0.0 a 0.0 a Tifway -5.0 ab -3.3 ab -3.3 ab 1.1 a -2.2 ab -4.4 a 0.0 a 0.0 a 0.0 a Average of biweekly rating in 2009 and Density was estimated by visual analysis of the whole plot. Means with the same letter in a given year and traffic treatment were not significantly different (P = 0.05) according to Duncan s means separation. 45

46 Figure 2-1. Percent green cover for eight cultivars of bermudagrass (Tifway, TifSport, Floratex, Riviera, Celebration, TifGrand, Hybrid1, and T11) in 2009 including traffic and nontrafficked plots. Solid trendline represents nontrafficked plots and dashed trendline represents plots receiving weekly traffic treatments. 46

47 CHAPTER 3 AN ENHANCED METHOD OF TRACKING DIVOT RECOVERY IN TURFGRASS Introduction Divot recovery research has traditionally been limited due to the time consuming, laborious nature of the research (Fry et al., 2008; Richardson et al., 2001). Uniform divots need to be created and regrowth needs to be accurately tracked over time. This field of study has progressed from creating divots with a 10.2 cm cup-cutter and pitching wedge (Ward and Thompson, 1971) to removing turf with a gasoline powered saw with blades stacked twelve wide (Fry et al., 2008). In 1971, the injury and recovery of the divot was assessed by using a grid and a plastic overlay tracing of individual divots. Today, digital image analysis (DIA) software can evaluate a photo and calculate the percent recovery of a divot (Karcher et al., 2005a). These new methods of creating and analyzing divots have immensely reduced the amount of time and labor involved in conducting divot research. The standard divot analysis (SDA) method described by Karcher et al. (2005a) involves creating a divot in the turf canopy and backfilling it with topdressing sand which is typically white in color. The divoted area is then photographed several times a week and subjected to DIA. The DIA software, SigmaScan Pro (v. 5.0, SPSS, Inc., Chicago, IL 60611) is used to measure the number of pixels within a target hue and saturation. Hue values range from 0 to 255 and directly relates to the colors in a photo. Saturation values range from 0 to 100 and describes the intensity of a specific hue (Figure 3-1). Richardson et al. (2001) described healthy bermudagrass and zoysiagrass as having a hue range from 57 to 107 and a saturation range from 0 to 100. Therefore, in a photo of a divot the number of target (green) pixels (hue and saturation 0-100) are 47

48 digitally counted and divided by the total number of pixels to give a percent green cover (Figure 3-2). This number is then compared to the percent green cover on the day the divot was removed and a percent recovery value can be calculated (Karcher et al, 2005a). While using this method of analysis, Schmitz et al. (2005) reported a great deal of variability in a divot recovery study involving seed/soil ratios and Kaminski et al. (2005) also had a large degree of variability in their results when studying the effects of growth regulators on divot recovery. During the first year of a study at the University of Florida examining divot recovery rates under traffic, we found it difficult to assess recovery results due to a large amount of variability in the turf surrounding the divot. Upon closer examination, turfgrass areas surrounding the divot in the picture that were not 100% dense or 100% green were not selected by the software as turf cover. Therefore, this percentage of discolored or missing turf, outside the divot area, was categorized as non-turf or divot area. As traffic treatments continued, the surrounding turf changed over time and continued to cause variability in the results. When conducting research on turf with minimal stress this problem may be negligible. However, when confounding factors that may impact canopy uniformity such as simulated traffic, low fertility, disease, or drought stress, are introduced, variability can increase, making analysis difficult. Therefore, it was necessary to identify steps to increase the precision of the divot analysis methodology while not losing any of its efficiency. The objectives of this study were (i) to develop a method of analysis that will more accurately determine the divot area regardless of turfgrass health while maintaining equal or better analytical efficiency and (ii) evaluate the validity of this process as a potential method of divot analysis. 48

49 Materials and Methods This study was conducted at the University of Florida Plant Science Research and Education Unit located near Citra, Florida. Plots of Celebration bermudagrass were maintained similar to a golf course fairway and mowed three times a week at a 1.2 cm height of cut. Celebration was selected because it has been shown to recover from divots faster (Karcher et al., 2005a) and withstand traffic injury better than other bermudagrass cultivars (Rutledge et al., 2005). This was part of a separate, larger study with treatments arranged in a split-split plot design with three replications. Whole plots were 18.2 m 2 (4.26 m x 4.26 m) and sub-plots were 9.1 m 2 (4.26 m x 2.13 m) and consisted of the following traffic treatments: 1) no traffic; and 2) simulated traffic applied weekly. Fertility treatments were randomized within each sub-plot and were applied at 24.4, 36.6, or 48.8 kg N ha -1 per month. Sub-sub-plots measured 3.0 m 2 (2.13 m x 1.42 m)and were fertilized with a turf fertilizer blend with 50% slow release nitrogen in a split application every two weeks. Colored Sand Colored sand was produced to backfill divots with the goal of allowing the software to more easily recognize the difference between the turf area and the divot area. Pink was chosen because it is at the farthest end of the spectrum (hue ) (Figure 3-1b) in the DIA software. This color was vastly different from green turf (hue ) (Figure 3-1a) exposed soil from traffic or other damage (hue 30-50), dry clippings (hue 35-65), stolons, (hue 35-65) or anything else naturally found on the turf. The pink sand was created by mixing 18.9 liters of dry white topdressing sand with 60 ml pink pigment, 240 ml polyurethane, and 480 ml paint thinner. The pink pigment was magenta mist (Valspar 211-6) obtained from Lowe s Paint Center. Minwax

50 super durability polyurethane varnish (Minwax Co., Upper Saddle River, NJ 07458) was added to adhere pigment to sand grains. Sand was put into a small concrete mixer and mixing was initiated. Polyurethane, paint thinner, and pigment were premixed in a beaker and added to the sand. Mixing continued for approximately three hours until the sand was dry to the touch. During initial testing the colored sand was found to be slightly hydrophobic compared with original white sand, likely due to the addition of polyurethane. To combat the hydrophobic property of the colored sand, 15 ml of Revolution (Aquatrols, Paulsboro, NJ 08066) wetting agent was mixed into 500 ml water and added to the dry pink sand. The sand was then continually mixed again in the concrete mixer until completely dry. The addition of wetting agent nullified any hydrophobic properties as colored sand accepted water similarly to white sand. White topdressing sand used for the SDA was also mixed with equal amounts of polyurethane and paint thinner to limit any chemical differences between the white and pink sand. The same procedure was used mixing all components, minus the pint pigment, and wetting agent was again added to reduce hybrophobicity. Divot Recovery Assessment Using DIA Two divots were taken from each sub-sub-plot to compare methods of analysis. Divots were created using a specialized divot machine that was designed and constructed at the University of Florida (Appendix B-3). One divot per plot was backfilled using white topdressing sand and the other was backfilled using pink colored sand. Digital pictures were collected two times per week throughout the period of the study. Photos were taken until divots were visually determined to be fully recovered. The study was repeated three times between May and August of Initial injury of divots occurred on May 18 (set 1), June 24 (set 2), and July 29 (set 3). Recovery was 50

51 tracked using four methods to determine percent divot recovery: standard divot analysis of the white divot (SDA), standard divot analysis of the pink divot (SDAP), pink sand analysis (PSA), and visual estimates. The SDA method is described by Karcher et al. (2005a), analyzing the white sand divot and using DIA to select for healthy green turf (hue and saturation 0-100) (Figure 3-2). Secondly, the SDAP method analyzed the pink sand divot using the same method where DIA selected for healthy green turf (hue and saturation 0-100) (Figure 3-3b). The PSA method analyzed the pink sand divot for pink pixels (hue and saturation 0-100), representing the area of the divot (Figure 3-3c). The PSA method measured the percentage of pink pixels or divot area in the photo, therefore this number was then subtracted from 100% to get percent turf cover making it comparable to SDA and SDAP methods. Percent recovery in both cases was then calculated by using this equation. % recovery [ %cover x -%cover %cover 0 ] 100 (3-1) In this equation, %cover x was the percent turf cover in the photo on the day it was collected and %cover 0 was the percent turf cover of the photo on the day of divot injury (day 0). Lastly, visual ratings of each divot were collected, estimating the percent divot recovery since the day of injury. This was done by comparing each divot photo to the photo of the same divot on the day of injury and estimating percent recovery. Recognizing that color differences between the sands could result in temperature differences, surface temperatures were taken one day after divot injury occurred at midday under full-sunlight using an infrared thermometer (model TN400L) (Metris Instruments, South Setauket, NY 11720). 51

52 Digital Image Analysis Digital images of each divot were collected using a portable light box and a Sony Cybershot DSC-H10 camera (Sony Corp., New York, NY 10022). Camera settings consisted of ISO100, 1/80 second shutter speed, and aperture of F3.5. The camera was mounted 30.5 cm above the ground on the light box, which provided an enclosed area with artificial light to ensure consistency along with minimal shadowing (Appendix B-4). The light box was 30.5 cm on each side with one ten-watt compact florescent, 6500 kelvin daylight, bulb mounted inside to cast light onto the target area. The light box provided uniformity of both light and camera height so all pictures were identical. The photo area was 28 x 21 cm. Digital images were then uploaded to a personal computer where they were cropped and resized using ACDSee Pro software (v. 2.5., ACD Systems International Inc., Victoria, British Columbia). Raw images were 3060x2448 pixels and final images were 800x800 pixels. This was done in preparation for DIA as the macro cannot readily process images larger than one megapixel (Karcher and Richardson, 2005). An analysis of variance (PROC ANOVA) was performed to test each method of analysis (Table 3-1). Pearson correlation coefficients (PROC CORR) were calculated to test the correlation between visual and digital analysis of white and pink sand divots. Additionally, data were analyzed with General Linear Model regression (PROC GLM) with the appropriate error terms to test relationships between percent recovery and methods of analysis (SAS, 2008). Mean separation was determined by date through Duncan s means separation at P =

53 Results Divots were photographed twice weekly until visual ratings reached 100% recovery. Divots analyzed using SDA and SDAP typically only showed 60-70% recovery at this time (100% visual recovery). When utilizing the PSA method, divots showed to be greater than 99% recovered in all cases (Table 3-2; 3-3; 3-4). Pearson Correlation Coefficients were calculated to determine the correlation between digital (SDA, SDAP, and PSA) and visual analysis. The PSA method was highly correlated (r > 0.90) with the visual analysis in all three sets (Table 3-5). SDA and SDAP had much lower correlation coefficients when compared to the visual analysis of the white and pink sand divots respectively (Table 3-5). Additionally, Figure 3-4 illustrates the comparison between DIA (PSA and SDAP) and visual analysis of the pink sand divot. Visual analysis was more closely related to PSA than SDAP as seen in Figure 3-4. Error bars were large when using the SDAP method and significant differences were found between SDAP and visual analysis after day fifteen. Furthermore, Figure 3-4b compares the PSA method with visual analysis, showing a definite correlation between the PSA and visual analysis methods. Visual analysis and PSA were similar on every collection day. Means comparisons were conducted to determine differences between methods of analysis. The SDA and SDAP method proved to be similar while PSA was statistically different on every collection date, excluding day zero, in sets one and three (Table 3-2; 3-4). During set two PSA was statistically different than SDA and SDAP on all but three collection days (Table 3-3). The PSA method allowed the accuracy of DIA to increase significantly in all three sets, as confirmed by the reduction in standard error (Table 3-2; 3-3; 3-4). Standard errors for SDA and SDAP ranged from 2% to 14% while the PSA 53

54 method had error <5% on all but two collection days and error <1% on more than half of all collection days (Table 3-2; 3-3; 3-4). To determine any differences in sand temperatures, surface temperatures of the two sands were compared with no differences noted (Table 3-6). The temperature of the colored sand, therefore, does not appear to influence divot recovery. It was also found that the PSA method analyzed the images more quickly than when using SDA and SDAP methods. SigmaScan (v. 5.0, SPSS, Inc., Chicago, IL 60611) took an average of 22.8 seconds to analyze one photo using the SDA and SDAP methods, while the PSA method only took an average of 2.8 seconds (Data not shown). Visual analysis of the images was much slower, averaging 2 minutes per divot. Discussion Results indicated the PSA method had less variability and was more precise in measuring divot recovery than SDA or SDAP. Pearson s correlation coefficients showed that PSA was more highly correlated with visual analysis than either SDA or SDAP. This suggests that the PSA method more accurately describes how the divots recover over time when compared with SDA and SDAP. Additionally this is supported by the reduction in standard error values. Standard errors were lower overall when using PSA compared with either of the standard methods of analysis (Table 3-2; 3-3; 3-4). This was illustrated by Figure 3-4b as error bars for SDAP are very large. The high degree of error made statistical differences more unlikely between SDAP and visual analysis. Despite the large degree of variability, significant differences were seen after day 15. This was not the case when using the PSA method. Figure 3-4a depicts the small amount of error observed and the close relationship that PSA had with visual analysis of divot recovery. The reduction in variability is reiterated in Figure 3-5 as the 54

55 R 2 value for PSA was much higher than SDA or SDAP methods of analysis. Traffic did not increase variability using standard methods of analysis as originally thought, however, the already high degree of variability in the non-traffic plots (Table 3-2; 3-3; 3-4) made it difficult for significant changes to occur. The PSA method provided precise analysis of the divot area regardless of any changes to the surrounding turfgrass. The SDA and SDAP methods both rely on the surrounding turf to calculate percent recovery whereas PSA focuses on the sand in the divot to calculate recovery. There was little variation between SDA and SDAP, as expected, because the only difference was sand color, which was not selected for by the software when using these methods of analysis. The small variation between these methods was due to the slight differences in divot size and the random variability in the surrounding turf. The PSA method has proven to have many benefits, mainly the improved accuracy, when conducting divot research; however there were a few issues that were discovered during this study. Clippings that would settle in the divot, on top of the sand, could alter analysis. This was similar to dry clippings on the turf that skewed the standard method of analysis. To combat this, a backpack blower was used at idle speed to move clippings out of the divot before photos were taken. Another issue that was discovered was that insects, mainly red imported fire ants (Solenopsis invicta Buren) and mole crickets (Scapteriscus spp.), would bring native soil to the surface of the divot covering some of the pink sand. This was not often a problem and was alleviated by either carefully removing the native soil or covering it up with additional colored sand. The final problem that was found when using colored sand was that the 55

56 traffic simulator would hit the divot directly and spread the sand, moving it out of the divot. To alleviate this problem additional sand was added to level divots before photos were taken. Although there were some minor issues with the PSA method, the benefits outweigh the problems. The elimination of variability in turf cover led to much more precise analysis of the divot area and therefore gave more accurate rate of recovery results. Damage from traffic or other external sources was not confused with divot damage because the pink sand was only located within the divot area. Obtaining precise measurements of divot area quickly and easily proves to be invaluable to obtaining more concrete results. Using the PSA method also led to faster analysis by DIA software. The images were analyzed in a fraction of the time as the software only had to count pink pixels which typically consisted of less than 20% of the image (128,000 pixels) whereas counting green pixels was typically greater than 60% of the image (384,000 pixels). SDA worked well for evaluating divot recovery when the turf surrounding the divot was one hundred percent dense and one hundred percent green in color. This is rarely the case as turf is frequently injured, diseased, scalped, soil is exposed, or leaves are sloughed off and turn brown. These external factors present a major problem when using SDA and SDAP, causing a high degree of variability and a lack of correlation with visual analysis. However, it was found that backfilling divots with pink colored sand and using DIA software to select for the pink color, these problems were eliminated. Not only was the accuracy increased when using the PSA method, but so was the speed of analysis. Divot research has incorporated several new technologies over the years and 56

57 implementing the PSA method will allow divot research to reach the next level, as a fast, simple, and precise way to measure divot recovery. 57

58 Table 3-1. Analysis of variance for divot recovery (%) among methods of analysis (SDA, SDAP, and PSA). Method of Analysis Effect df F value Pr > F Days after injury (DAI) <.0001 Method <.0001 Fertility Traffic Set <.0001 DAI*Method <.0001 Fertility*Method Traffic*Method Set*Method Set*Traffic*Method Methods of analysis consisted of standard divot analysis (SDA), standard divot analysis pink divot (SDAP), and pink sand analysis (PSA). Fertility treatments included 0.5, 0.75, and 1.0 lb N/1000ft 2 per month. Traffic treatments consisted of no traffic and traffic applied once weekly. This study was repeated three times between June and August, Initial injury of divots occurred on May 18 (set 1), June 24 (set 2), and July 29 (set 3). 58

59 Table 3-2. Set 1 showing percent recovery by day with standard errors. Variability in Analysis Methods No Traffic SDA SDAP PSA Day % Error % Error % Error a a a ab b a b b a b b a b b a b b a b b a b b a <0.1 Traffic SDA SDAP PSA Day % Error % Error % Error a a a b b a b b a b b a b b a b b a b b a b b a <0.1 SDA: Standard divot analysis analyzes white sand divots for green pixels (hue , sat 0-100). SDAP: Standard divot analysis - pink divot analyzes pink sand divots for green pixels (hue , sat 0-100). PSA: Pink sand analysis analyzes pink sand divots for pink pixels (hue , sat 0-100). Means with the same letter in each row are not statistically different (P < 0.05). 59

60 Table 3-3. Set 2 showing percent recovery by day with standard errors. No traffic was applied during this set of divots due to equipment failure. Variability in Analysis Methods SDA SDAP PSA Day % Error % Error % Error a a a a a a b ab a b b a c b a b b a b b a b b a b b a b b a <0.1 SDA: Standard divot analysis analyzes white sand divots for green pixels (hue , sat 0-100). SDAP: Standard divot analysis pink divot analyzes pink sand divots for green pixels (hue , sat 0-100). PSA: Pink sand analysis analyzes pink sand divots for pink pixels (hue , sat 0-100). Means with the same letter in each row are not statistically different (P < 0.05). 60

61 Table 3-4. Set 3 showing percent recovery by day with standard errors. Variability in Analysis Methods No Traffic SDA SDAP PSA Day % Error % Error % Error a a a c b a c b a c b a b b s b b a < c b a <0.1 Traffic SDA SDAP PSA Day % Error % Error % Error a a a b b a c b a b b a b b a < b b a < c b a <0.1 SDA: Standard divot analysis analyzes white sand divots for green pixels (hue , sat 0-100). SDAP: Standard divot analysis pink divot analyzes pink sand divots for green pixels (hue , sat 0-100). PSA: Pink sand analysis analyzes pink sand divots for pink pixels (hue , sat 0-100). Means with the same letter in each row are not statistically different (P < 0.05). 61

62 Table 3-5. Pearson correlation coefficients compare method of digital image analysis (DIA) with the visual analysis of the corresponding divot (n = 468). Correlation of Digital and Visual Analysis Set 1 Set 2 Set 3 DIA r P r P r P SDA < < <.0001 SDAP < < <.0001 PSA < < <.0001 r: Pearson correlation coefficient. SDA: Standard divot analysis analyzes white sand divots for green pixels (hue57-107, sat 0-100). SDAP: Standard divot analysis with pink sand analyzes pink sand divots for green pixels (hue57-107, sat 0-100). PSA: Pink sand analysis analyzes pink sand divots for pink pixels (hue , sat 0-100). Table 3-6. Infrared (IR) temperature data for divot sand. Divot Sand Temperature Effect df F value Pr > F Temperature Avg. Temps (C) White 44.2 Pink 44.6 Surface temperatures were collected using an infrared thermometer one day after divot injury. All measurements were taken at mid-day under full sunlight. 62

63 100 a b Figure 3-1. Color spectrum from SigmaScan Pro software (v. 5.0, SPSS, Inc., Chicago, IL 60611) shows hue and saturation values. Area (a) highlights the range of healthy green bermudagrass from hue 57 to 107. Area (b) highlights the hue range ( ) of the pink sand. a b Day % green cover Figure 3-2. Digital image of one divot on the day of injury. Standard divot analysis (SDA) finds that 54.3% of the image is green turf area. The red overlay (b) shows all pixels that were found to be in the target area (hue and saturation 0-100). 63

64 a b c Day % green cover 100% % pink pixels = 88.4% green cover Figure 3-3. Divots analyzed using both methods of DIA. Photo (a), (b), and (c) are the same images of a divot taken on the day of injury. Photo (a) is the image after being cropped and resized to 800x800 pixels. Photo (b) shows a red overlay of selected pixels using standard divot analysis (SDA) (hue and saturation 0-100). Photo (c) shows red overlay of selected pixels using the new methodology of pink sand analysis (PSA) (hue and saturation 0-100). 64

65 Methods of Divot Analysis 100 a Divot Recovery (%) b PSA Visual Analysis SDAP Visual Analysis Days after Injury Figure 3-4. Percent recovery for divots filled with pink sand. Methods of analysis include standard divot analysis of pink divot (SDAP), pink sand analysis (PSA), and visual ratings. Visual analysis data is identical in both (a) and (b) as they both describe the same divot. 65

66 Methods of Analysis Divot Recovery (%) PSA SDA SDAP Days after Injury Figure 3-5. Recovery curves using standard divot analysis method for divots filled with both white (SDA) and pink sand (SDAP) as well as the pink sand analysis (PSA) method. Solid trend line describes PSA, small dashed trend line describes SDA, and medium dashed trend line describes SDAP. PSA: Pink sand analysis y = -0.14x x R 2 = SDA: Standard divot analysis y = -0.06x x R 2 = SDAP: Standard divot analysis pink divot y = -0.03x x R 2 =

67 CHAPTER 4 RECUPERATIVE ABILTY OF EIGHT BERMUDAGRASS CULTIVARS Introduction Turfgrass can be damaged in various ways on a golf course but one of the most devastating injuries occurs when a golfer physically removes an area of turf in the form of a divot. This most often occurs when a golfer plays an iron shot from the fairway or tee. Similar damage can be observed on an athletic field as players pivot, digging their cleats into the turf and removing a small area of the turf canopy. As a result of this damage, one of the main concerns of turfgrass mangers is to encourage the turf to regrow and fill-in bare areas as quickly as possible. Bermudagrass (Cynodon dactylon [L.] Pers.; C. dactylon x C. transvaalensis [Burtt-Davy]) is considered one of the fastest growing warm-season grasses, making it an excellent choice for fairways, tees, and athletic fields that often experience large amounts of divot damage. Recently, many new cultivars of bermudagrass have been produced with improved color, quality, and stress tolerances (Taliaferro et al., 2004), but it is unknown how these cultivars compare in their ability to recover from divot injury. Cultivar selection is an ongoing field of study, as new cultivars are constantly being developed and tested to determine beneficial attributes. The National Turfgrass Evaluation Program (NTEP) trials, the most recognized program comparing bermudagrass cultivars at locations around the United States, does not currently include a measure of recuperative potential (Morris, 2006). Karcher et al. (2005a; 2005b) and Trappe et al. (2009) represent the main published research comparing divot recovery among cultivars in the South. Their work has given some preliminary information but more work is needed to determine how other factors such as fertility rate, irrigation, 67

68 traffic, or shade may affect regrowth of the damaged turf. This study focuses on how fertility rate, simulated traffic, and cultivar selection influence the recuperative ability of bermudagrass when subjected to divot damage. Newly released cultivars being investigated in this study will include commercially available cultivars Celebration and TifGrand along with experimental cultivars Hybrid1 and T11 from the University of Georgia breeding program. These will be compared with some of the industry standards in Tifway, TifSport, Riviera, and Floratex. Determining which cultivars of bermudagrass recover from divot injury the fastest will be valuable information to turfgrass managers at high use facilities. The objectives of this study were (i) to analyze the difference in recovery rate between bermudagrass cultivars, (ii) to determine the effect that fertility rate has on the speed of divot recovery, and (iii) to assess the effect that traffic plays on the time that it takes for divots to recover. Materials and Methods This study was conducted at the University of Florida Plant Science Research and Education Unit located near Citra, Florida. The study was repeated three times between May and August, Initial injury of divots occurred on May 18 (set 1), June 24 (set 2), and July 29 (set 3). Plots were maintained similar to a golf course fairway or athletic field and mowed three times a week at a 1.2 cm height of cut. Treatments were arranged in a split-split plot design with three replications. Whole plots were 18.2 m 2 (4.26 m x 4.26 m) and consisted of eight bermudagrass cultivars: Celebration, Floratex, Hybrid1, Riviera, T11, TifGrand, TifSport, and Tifway. Sub-plots were 9.1 m 2 (4.26 m x 2.13 m) and consisted of the following traffic treatments: 1) no traffic; and 2) simulated traffic applied weekly. Fertility treatments were randomized within each sub-plot and were applied at 24.4, 36.6, or 48.8 kg N ha -1 per month. Sub-sub-plots measured

69 m 2 (2.13 m x 1.42 m) and fertilized with turf fertilizer blend containing 50% slow release nitrogen in a split application every two weeks. In addition to N-P-K the fertilizer contained 2% Mg, 1% Fe, and 1% Mn. Divots were removed using a specialized divot machine constructed at the University of Florida (Appendix B-3) and then backfilled with pink colored sand. Divots measured 105 cm 2 on average. All divots were tracked until visual ratings reached100% recovery. Visual ratings of each divot were compared to the photo of that same divot on the day of injury (day 0) and percent recovery was estimated. Traffic was applied weekly throughout the growing season using a modified Cady traffic simulator (Henderson et al., 2005) that was constructed using a John Deere Aercore 800 aerifier (Appendix B-1). Traffic treatments were applied during two separate periods, from May 18 to June 16, and from July 28 to August 25. Traffic was not applied continuously during the study due to equipment failure. Digital pictures of each divot were collected twice per week and processed using digital image analysis (DIA) software to determine percent green cover (Richardson et al., 2001). Digital image analysis. Digital images of each divot were collected using a Sony Cybershot DSC-H10 camera (Sony Corp., New York, NY 10022). Camera settings consisted of ISO100, 1/80 second shutter speed, and aperture of F3.5. The camera was mounted 30.5 cm above the ground on a portable light box, which provided an enclosed area with artificial light to ensure consistency along with minimal shadowing (Appendix B-4). The light box was 30.5 cm on each side with one ten-watt compact florescent, 6500 kelvin daylight bulb mounted inside to cast light onto the target area. The light box provided uniformity of both light and camera height so all pictures were 69

70 identical. The photo area was 28 x 21 cm. Digital images were uploaded to a personal computer where they were cropped and resized using ACDSee Pro (v. 2.5., ACD Systems International Inc., Victoria, British Columbia). Raw images were 3060x2448 pixels and final images were 800x800 pixels. This was done in preparation for digital image analysis as the macro cannot readily handle images larger than one megapixel (Karcher and Richardson, 2005). Divots were backfilled with pink colored sand and analyzed using the pink sand analysis (PSA) method described in Chapter 3 (Figure 4-1). DIA software calculated the percentage of divot area in the photo by counting the number of pink pixels (hue and saturation 0-100) and dividing by the total number of pixels in the photo. This number was then subtracted from 100% to get a percent turf cover. The following equation was used to determine percent recovery for each divot on a particular day. % recovery [ ] 100 (4-1) In this equation, %cover x is the percent turf cover of the photo on the day it was collected and %cover 0 is the percent turf cover of the photo on the day of divot injury (day 0). An analysis of variance (PROC ANOVA) was performed to test cultivar, fertility, and traffic effects (Table 4-1). Additionally, data were analyzed with General Linear Model regression (PROC GLM) with the appropriate error terms to test relationships between percent recovery and cultivar, fertility and traffic treatments (SAS, 2008). Mean separation was determined by date through Duncan s means separation at P =

71 Results Results indicated that neither fertility nor traffic treatments were significant (P > 0.05) during any of the divot sets (Table 4-1). The primary factor in recovery was clearly genetic as recovery speed was directly associated with differences among cultivars. Each of the three sets of divots was analyzed separately as they were found to be significantly different. In set one Floratex and T11 recovered faster than other cultivars, achieving 50% recovery within 7 days and 95% recovery by day 17. Celebration achieved 50% recovery by day 7 and 95% recovery by day 20. Hybrid1, Riviera, TifGrand, TifSport, and Tifway all performed similarly, needing between 10 and 14 days to achieve 50% recovery and between 20 and 23 days to reach 95% recovery. (Table 4-2). Divots in set two were created on June 24, 2010 and tracked for 33 days. Cultivars in set two performed similarly to set one, as Celebration, Floratex, and T11 all recovered faster than other cultivars. TifGrand, TifSport and Tifway were in the lowest statistical category at nearly every collection date (Table 4-3). All cultivars recovered more quickly during set three, and again Celebration and T11 were in the top category on all collection dates under both traffic treatments. These cultivars reached 75% recovery in the first 8 days and 95% recovery by day 11. TifGrand, Tifway, and TifSport were in the lowest statistical category on all days prior to day 18 when 95% recovery was achieved (Table 4-4). Differences in recovery rate among cultivars are illustrated in Figures 4-2; 4-3; 4-4. A quadratic model described the data well as R 2 values for all cultivars were >0.97 (Figure 4-2; 4-3; 4-4). Cultivars fell into three distinct groupings. Fastest recovering cultivars were Celebration, Floratex, and T11 which averaged 15 days to reach 95% recovery. Hybrid1 and Riviera were in the next group, recovering slightly slower, 71

72 reaching 95% recovery in 17 days. Lastly, cultivars that took the longest to recover from divot injury were TifGrand, TifSport, and Tifway, needing 20 days to reach 95% recovery. Discussion In all three sets of divots, cultivar effects were significant while traffic and fertility treatments were not. Nitrogen rate did not play a factor in recovery rate at any point during this study. Fertility treatments may have been too narrow to see differences or minimal nutrition levels were available with the lowest fertility treatment. Nutritional deficiencies were not observed at any point in the study suggesting that all fertility levels were satisfactory. Additionally, results indicated that traffic did not affect lateral growth rate as no differences were observed between treatments. Wear stress on the turf and compaction caused by the traffic did not adversely affect the growth rate of the plant as initially expected. The key component in determining recovery rate was found to be differences in bermudagrass cultivars. Although sets were statistically different, similar trends were seen over all three recovery periods. Differences among sets were likely due to the variation in temperature and sunlight. All cultivars recovered faster during set three likely due to a seasonal response (Appendix A-1) causing grass to increase lateral growth. Recovery of divots in set one occurred in early summer (May), set two in mid-summer (June), and set three tracked divot recovery during late summer (August). Over the entire period of the study, T11 preformed the best of all bermudagrass cultivars, and was in the top statistical category on all but two collection dates. T11 shows to be a promising experimental cultivar, exhibiting a high lateral growth rate even when subjected to traffic stress. Celebration was the fastest recovering commercially 72

73 available cultivar in this study, typically reaching 50% recovery by day 7 and 95% recovery by day 15. Floratex did well in early- and mid-summer, but divot recovery slowed in late summer, suggesting a decrease in growth rate. Seeded cultivars Riviera and Hybrid1 performed similarly reaching 50% recovery in an average of 8 to 10 days and 95% recovery in 16 days; slower than Celebration, T11, and Floratex, but faster than vegetative cultivars Tifway, Tifsport, and Tifgrand. Karcher et al. (2005a) found that seeded cultivars of bermudagrass typically took longer to recover from divot injury. TifGrand preformed similar to Tifway and TifSport. These cultivars were significantly slower to recover from divot injury taking an average of 10 days to reach 50% recovery and 18 days to reach 95% recovery. TifGrand, TifSport, and Tifway were in the lowest statistical category at almost every collection date. This data agrees with Karcher et al. (2005a) who observed Tifway and TifSport to be among the slowest cultivars to recover from divot injury. Celebration, Floratex, and T11 performed well throughout the study. These cultivars outperformed the industry standards, reaching full recovery an average of 3 to 5 days sooner, than Tifway and TifSport. As turfgrass managers are faced with more play, it is critical that the turf can handle the increased stress. While industry standards, Tifway and TifSport, have performed well in the past,, there are new, improved cultivars available today. Choosing a cultivar that has a genetic advantage in recuperative ability will allow the playing surface to recover more rapidly, increasing both aesthetics and playability. 73

74 Table 4-1. Analysis of variance for percent recovery of divots in Divot Recovery (%) Effect df F value Pr > F Days after injury (DAI) < Cultivar Fertility Traffic Set DAI*Cultivar Fertility*Cultivar Traffic*Cultivar Set*Cultivar Eight bermudagrass cultivars were included in this study included, Celebration, Floratex, Hybrid1, Riviera, T11, TifGrand, TifSport, Tifway. Fertility treatments included 0.5, 0.75, and 1.0 lb N/1000ft 2 per month. Traffic treatments consisted of no traffic and traffic applied once weekly. This study was repeated three times between June and August, Initial injury of divots occurred on May 18 (set 1), June 24 (set 2), and July 29 (set 3). 74

75 Table 4-2. Percent divot recovery over time during set 1 for the eight bermudagrass cultivars: Celebration, Floratex, Hybrid1, Riviera, T11, TifGrand, TifSport, and Tifway. Divot Recovery Over Time Set 1 Days Cultivar Celebration 0.0 a 52.5 ab 69.7 abc 86.5 bc 92.8 b 95.9 ab 98.6 ab 99.9 a Floratex 0.0 a 56.2 ab 72.9 ab 92.0 ab 95.6 ab 96.9 ab 99.2 ab 99.8 a Hybrid1 0.0 a 45.8 bc 62.3 bcd 84.1 c 97.6 a 96.0 ab 99.3 ab 99.8 a Riviera 0.0 a 43.9 bc 59.5 cd 85.0 c 92.8 b 94.9 b 98.3 ab 99.7 ab T a 61.0 a 75.4 a 94.4 a 97.6 a 98.5 a 99.5 a 99.9 a TifGrand 0.0 a 36.0 c 53.2 d 82.2 c 92.0 b 94.5 b 97.6 b 99.3 b TifSport 0.0 a 36.2 c 53.2 d 81.3 c 93.2 b 94.5 b 98.4 ab 99.7 ab Tifway 0.0 a 42.7 bc 56.6 d 83.4 c 93.3 b 94.4 b 98.0 ab 99.6 ab Means with the same letter on a given day and traffic treatment were not significantly different (P 0.05) according to Duncan s means separation. 75

76 Table 4-3. Percent divot recovery over time during set 2 for the eight bermudagrass cultivars: Celebration, Floratex, Hybrid1, Riviera, T11, TifGrand, TifSport, and Tifway. Divot Recovery Over Time Set 2 Days Cultivar Celebration 0.0 a 34.8 ab 58.1 ab 85.2 b 95.2 abc 97.5 ab 99.6 a 99.6 a 99.8 a Floratex 0.0 a 38.9 a 61.0 a 85.7 b 97.2 ab 98.7 a 99.2 a 99.0 ab 99.3 b Hybrid1 0.0 a 32.5 abc 51.9 bc 78.6 c 92.0 bc 96.2 ab 99.2 a 99.3 a 99.5 ab Riviera 0.0 a 29.3 bcd 49.8 c 80.2 bc 89.1 c 93.4 b 98.5 a 98.8 ab 99.4 ab T a 33.9 ab 59.3 a 91.7 a 99.3 a 99.7 a 99.8 a 99.8 a 99.8 a TifGrand 0.0 a 25.8 cd 39.0 d 68.4 d 72.2 d 80.4 d 95.5 b 98.1 bc 99.2 b TifSport 0.0 a 23.1 d 39.2 d 71.8 d 78.5 d 85.7 c 96.5 b 96.9 d 97.8 d Tifway 0.0 a 22.2 d 33.1 d 69.2 d 75.4 d 83.2 cd 96.6 b 97.7 cd 98.5 c Means with the same letter on a given day and traffic treatment were not significantly different (P 0.05) according to Duncan s m eans separation. 76

77 Table 4-4. Percent divot recovery over time during set 3 for the eight bermudagrass cultivars: Celebration, Floratex, Hybrid1, Riviera, T11, TifGrand, TifSport, and Tifway. Divot Recovery Over Time Set 3 Days Cultivar Celebration 0.0 a 45.4 a 83.7 a 96.7 a 99.1 a 99.5 a 99.9 a Floratex 0.0 a 34.3 bc 77.2 ab 92.8 ab 97.7 bc 97.8 d 99.3 c Hybrid1 0.0 a 36.6 b 64.4 c 89.9 b 97.1 c 98.2 cd 99.6 b Riviera 0.0 a 36.0 b 74.5 b 89.7 b 98.2 abc 98.6 bc 99.7 ab T a 34.4 bc 75.9 b 93.8 ab 98.8 ab 99.1 ab 99.9 a TifGrand 0.0 a 32.8 bc 59.2 c 78.6 cd 94.7 d 98.3 cd 99.6 b TifSport 0.0 a 27.1 c 45.6 d 77.0 d 95.3 d 96.4 e 99.1 d Tifway 0.0 a 32.5 bc 58.4 c 82.9 c 94.9 d 97.0 e 99.2 cd Means with the same letter on a given day and traffic treatment were not significantly different (P 0.05) according to Duncan s means separation. 77

78 a b Day 0 100% % pink pixels = 88.4% green cover Figure 4-1. Divots were analyzed using SigmaScan Pro software (v. 5.0, SPSS, Inc., Chicago, IL 60611). Photo (a) and (b) are the same image of a divot taken on the day on injury. Photo (a) is the image after being cropped and resized to 800x800 pixels. Photo (b) shows red overlay of selected pixels with new methodology utilizing colored sand (hue and saturation 0-100). 78

79 Cultivar Recovery - Set 1 Divot Recovery (%) Celebration Floratex Hybrid1 Riviera T11 Tifgrand Tifsport Tifway Days after Injury Figure 4-2. Recovery curves for each of the eight bermudagrass cultivars (Celebration, Floratex, Hybrid1, Riviera, T11, TifGrand, TifSport, and Tifway) beginning May 18, 2010 (set 1). Line equations and R 2 values can be found below; cultivars are listed in order from fastest to slowest recovery. Recovery curves were truncated when recovery reached 100% as divots were fully recovered at this time. (n=6). T11 y = x x R² = 0.99 Floratex y = x x R² = 0.99 Celebration y = x x R² = 0.99 Hybrid1 y = x x R² = 0.99 Riviera y = x x R² = 0.99 Tifway y = x x R² = 0.99 TifSport y = x x R² = 0.98 TifGrand y = x x R² =

80 Cultivar Recovery - Set 2 Divot Recovery (%) Celebration Floratex Hybrid1 Riviera T11 TifGrand TifSport Tifway Days after Injury Figure 4-3. Recovery curves for each of the eight bermudagrass cultivars (Celebration, Floratex, Hybrid1, Riviera, T11, TifGrand, TifSport, and Tifway) beginning June 24, 2010 (set 2). Line equations and R 2 values can be found below; cultivars are listed in order from fastest to slowest recovery. Recovery curves were truncated when recovery reached 100% as divots were fully recovered at this time. Simulated traffic was not applied to any plots during this set of divots (n=6). T11 y = x x R 2 = 0.97 Floratex y = x x R 2 = 0.98 Celebration y = x x R 2 = 0.98 Hybrid1 y = x x R 2 = 0.99 Riviera y = x x R 2 = 0.99 TifSport y = x x R 2 = 0.99 Tifway y = x x R 2 = 0.99 TifGrand y = x x R 2 =

81 Cultivar Recovery - Set 3 Divot Recovery (%) Celebration Floratex Hybrid1 Riviera T11 TifGrand TifSport Tifway Days after Injury Figure 4-4. Recovery curves for each of the eight bermudagrass cultivars (Celebration, Floratex, Hybrid1, Riviera, T11, TifGrand, TifSport, and Tifway) beginning July 29, Line equations and R 2 values can be found below; cultivars are listed in order from fastest to slowest recovery. Recovery curves were truncated when recovery reached 100% as divots were fully recovered at this time. (n=6). Celebration y = x x R² = 0.97 T11 y = x x R² = 0.98 Floratex y = x x R² = 0.98 Riviera y = x x R² = 0.99 Hybrid1 y = x x R² = 0.99 Tifway y = x x R² = 0.99 TifGrand y = x x R² = 0.99 TifSport y = x x R² =

82 CHAPTER 5 GENERAL CONCLUSIONS Bermudagrass selection has become an important field of study, and as more cultivars are available, it is necessary to test the many aspects of performance. While there are many factors that need to be considered when selecting a cultivar, traffic tolerance and divot recovery are important characteristics to consider. The studies included in this thesis have shown that the industry standards may not be providing the best color, quality, or density when compared to some of the new improved cultivars. The first study found the most traffic tolerant bermudagrass cultivars to be Celebration, Hybrid1, T11 and TifGrand. These cultivars provided better cover, color, and quality than industry standards Tifway and TifSport as well as Floratex and Riviera. Fertility rates of 24.4, 36.6, and 48.8 kg N ha -1 per month did not make a difference in turfgrass performance at any point during the study period. Further research examining wider fertility ranges would be useful to locate any definitive points of change in turfgrass traffic tolerance or divot recovery. Experimental cultivars Hybrid1 and T11 have proven to handle traffic stress well, and more research will determine if these cultivars are ready for commercial release. The second study compared standard divot analysis (SDA) to pink sand analysis (PSA) as a method of tracking divot recovery. Results indicated that the SDA method had a great deal of variability and did not correlate well with visual analysis of the divot recovery. The PSA method was able to provide a more accurate divot analysis by eliminating the variability caused by the surrounding turfgrass and focusing analysis on the sand within the divot. Furthermore, the PSA method closely correlated to visual analysis in all three sets of divots. This method was used in study three to compare 82

83 cultivar responses to divot injury under several traffic and fertility treatments. This study found that divot recovery rate was a direct response of bermudagrass cultivar. Celebration, Floratex, and T11 consistently recovered from divot injury 3 to 4 days faster than Hybrid1, Riviera, TifGrand, Tifway, or TifSport. Turfgrass managers looking to increase cover in damaged areas could achieve this goal by regrassing with an improved cultivar of bermudagrass. Similar research will need to be completed in areas around the country to determine the limitations of these cultivars. These studies analyzed recovery of bermudagrass cultivars under traffic and divot stress and found some overlap between cultivars and their ability to recover. Celebration, Hybrid1, T11 and TifGrand were all found to be very traffic tolerant, however not all four of these cultivars displayed rapid recovery from divot injury. Celebration and T11 were the most tolerant cultivars to both traffic and divot injuries. Hybrid1 and TifGrand, provided dark green, dense turf when subjected to simulated traffic, however, they were not able to recover from divot injury as quickly as Celebration or T11. Conversely, Floratex showed to recover rapidly from divot injury, but declined in color, quality, and density when subjected to traffic. The combined results of these studies show that Celebration and T11 are more injury tolerant than current industry standard Tifway and TifSport. Tifway bermudagrass has long been the standard in the state of Florida and other areas in the South; however these studies have shown that turfgrass managers today have better, more efficient options. More research is needed to determine if T11 is ready for commercial release, however, Celebration is currently on the market. Turfgrass managers at high use facilities can increase the traffic tolerance and divot 83

84 recovery of their playing surface by installing Celebration bermudagrass. Whether it is building a new course, renovating an old one, or just converting a tee, there are new cultivars that are worth considering. 84

85 APPENDIX A WEATHER DATA CITRA, FL Weather Data Citra, FL Jul Aug Sep Oct Nov Avg Temp (C) Max Temp Min Temp Rainfall (cm) Solar Radiation (w/m^2) Apr May Jun Jul Aug Avg Temp (C) Max Temp Min Temp Rainfall (cm) Solar Radiation (w/m^2) A-1. Monthly averages were obtained from Florida Automated Weather Network (FAWN). Solar Radiation was measured in watts per sq. meter. 85

86 APPENDIX B EQUIPMENT B-1. Modified Cady traffic simulator constructed from a John Deere Aerocore 800 applied traffic once per week during the growing season. 86

87 B-2. Light box that was used to provide uniform light on each plot without shadows. The light box measured 53 x 61 x 51 cm with four ten-watt compact florescent, 6500 kelvin daylight bulbs mounted to the ceiling to cast consistent light onto the target area. B-3. A specialized divot machine was used to produce uniform divots on each plot. It was designed and constructed at the University of Florida from a modified clay pigeon thrower and mimics the swing action of a golf club. 87

88 B-4. Light box that was used to take uniform photos of the divots under consistent light. The light box was 30.5 cm on each side with one ten-watt compact florescent, 6500 kelvin daylight bulb mounted inside to cast light onto the target area. 88