2010 SOUTHEASTERN NATURALIST 9(3):573–586 Landscape Scale Correlates of Fox Squirrel Presence on Golf Courses in Coastal South Carolina Kristin Meehan1,2 and Patrick G.R. Jodice3,* Abstract - The coastal plain of the southeastern United States has experienced rapid and intense urbanization that has substantially changed the area’s landscape. Much of this landscape change has been associated with the development of golf courses and associated communities, and cumulatively, these landscape units occupy thousands of hectares in the region. The opportunity for these golf courses to support native wildlife requires examination as these habitats are becoming more common and often represent some of the largest tracts of remaining open space within an area. Although declining throughout the southeastern US, Sciurus niger (Fox Squirrel) populations are still found on golf courses in this region. We investigated the relationship between Fox Squirrel presence and landscape characteristics on 98 golf courses in coastal South Carolina. Visual and telephone surveys indicated Fox Squirrels were present on 68 of the courses surveyed. The best predictor of Fox Squirrel presence on a course was the presence of a Fox Squirrel population on the nearest neighboring course. The probability of Fox Squirrels being present on a course if they were also present on the nearest neighboring course was 87.3%. Course age was the best predictor of Fox Squirrel presence on golf courses without a Fox Squirrel population on their nearest neighboring course. Our results suggest that regional Fox Squirrel populations may be stabilized by multi-patch population dynamics. Introduction The role of golf courses in wildlife conservation has become increasingly important in areas where development is prevalent, where native habitats are declining and becoming fragmented, and where golf courses often represent some of the largest tracts of remaining open space (Angold et al. 2006, Yasuda and Koike 2006, Zipperer et al. 2000). For example, the coastal plain of the southeastern United States was historically characterized by a mix of forest, agricultural land, and rural communities. Since the mid-1960s, rapid and intense urbanization has substantially changed the landscape of the Southeast (Allen and Lu 2003). A common change to this landscape has been the development of golf courses and associated communities, which now occupy thousands of hectares in the region (Jones et al. 2005). Many of these golf courses were created from large parcels of undeveloped or rural land, and when originally constructed, were surrounded by coastal forest. As such, the landscape was often characterized by isolated golf courses 1Department of Forestry and Natural Resources and South Carolina Cooperative Fish and Wildlife Research Unit, Clemson University, Clemson, SC 29634. 2Current address - Landmælingar Íslands, Stillholti 16-18, 300 Akranes, Iceland. 3US Geological Survey, South Carolina Cooperative Fish and Wildlife Research Unit, Clemson University, Clemson, SC 29634. *Corresponding author - pjodice@clemson.edu. 574 Southeastern Naturalist Vol. 9, No. 3 surrounded by a matrix of undeveloped or lightly developed lands. However, as urban areas have expanded and as development has increased adjacent to golf courses, many golf courses have instead become surrounded by higherintensity development. In these instances, golf courses frequently become the largest parcels of open space remaining in a matrix of development and may represent important wildlife habitat. Currently, the maintenance or preservation of specific types of habitat for wildlife on golf courses occurs infrequently by design, and any planning that does occur appears to happen at a within-course scale. The potential for golf courses to act as wildlife habitat could be improved if species requirements were considered prior to development (Love 1999, Terman 1997) and if landscape-scale factors (e.g., multiple courses, adjacent undeveloped lands, corridors) were considered, the latter being especially critical for species that require habitat patches that are larger than those offered by most courses. Such an approach, however, requires species-specific studies that provide an understanding of the factors which affect species presence specifically in a setting comprised of urban golf courses (Hodgkison et al. 2007, Mason 2006). Often these data are lacking, or in cases where data do exist, much of it pertains to requirements of avian species (e.g., Cristol and Rodewald 2005) or to species that do not require extensive landscapes, such as small mammals or herpetofauna (Barthelmess 2004, Montieth and Paton 2006). Habitat-use data for meso- or large mammals that move among patches in an urban landscape (e.g., Adkins and Stott 1998, Ditgen et al. 2007) are rare due to the difficulty of observing these species and the complexity of examining landscape-level attributes. Our goal was to investigate the occurrence of Sciurus niger (L.) (Fox Squirrel) on golf courses along the coast of South Carolina. This species is declining throughout much of the southeastern US (Loeb and Moncrief 1993), but is often observed on golf courses in the region. Use of golf courses by Fox Squirrels may be due in part to similarities in habitat structure between this species’ commonly used native habitat (mature pinelands with open-understory and mast producing hardwoods) and that provided by many of the golf courses in the region, and to abundant food resources that often occur on courses (Jodice and Humphrey 1992, Lee et al. 2008). Population viability for this species may be affected, however, not only by within-course habitat attributes but also by the matrix and landscape-scale characteristics surrounding golf courses (Ditgen 1999, Jodice and Humphrey 1993). For example, Fox Squirrels inhabiting golf courses move between courses either as part of their naturally high rate of natal dispersal or simply to access other individuals or areas for breeding or feeding (Ditgen 1999). Thus, both within- course habitat structure and landscape characteristics require attention for the conservation of this species in these golf course settings. Although neither of these relationships have been examined extensively, the former has received more attention than the latter in the southeastern US (e.g., Ditgen et al. 2007; Jodice and Humphrey 1992, 1993; Lee et al. 2008). Therefore, 2010 K. Meehan and P.G.R. Jodice 575 our objective was to use a model-selection approach to assess the relationship between Fox Squirrel presence on golf courses along the South Carolina coast and landscape-scale features. Determining the importance of habitat features for Fox Squirrel presence on golf courses at the landscape-scale will provide insight into the potential long-term viability of this declining species in the southeastern coastal plain. Methods Study species and study area The southeastern Fox Squirrel is listed as a species of concern in South Carolina. Fox Squirrel populations appear to be declining throughout much of the southeastern US due in large part to the loss of native Pinus palustris Mill. (Longleaf Pine) habitat (Harrigal 1993, Loeb and Moncrief 1993, Weigl et al. 1989). Home ranges of southeastern Fox Squirrels vary by region and habitat conditions, but often are >35 ha (Ditgen 1999, Jodice 1993, Kantola and Humphrey 1990, Perkins and Connor 2004). Road mortality can be significant (McCleary et al. 2008), even in areas with relatively low traffic volumes (Lee et al. 2008), and dispersing subadult Fox Squirrels may move between habitat patches in urban regions, traveling up to 6 km (Ditgen 1999). Our study was conducted on 98 golf courses that were located within the Coastal Plain and Coastal Zone ecoregions of South Carolina (Fig. 1). Figure 1. Location of study area in coastal South Carolina, showing coastal ecoregions. 576 Southeastern Naturalist Vol. 9, No. 3 The Coastal Zone ecoregion is a subset of the Coastal Plain and is found seaward of the state inland marine waters boundary. The courses included in this study occurred along the entire length of the South Carolina coast, although there was a large unstudied area in the central part of the coast which included numerous public lands and few golf courses (Fig. 2). We only included golf courses in the study if the landscaping was complete as of 1999 to ensure that the landcover data, based on 1999 satellite imagery, matched the study courses. Golf courses which substantially differed in appearance between aerial photos and the national landcover dataset (NLCD) were not included in the study. Furthermore, any golf courses located on sea islands where Fox Squirrels were absent outside of golf course habitats were not included in analyses. Golf courses were counted as one entity for data analysis when they adjoined each other, shared the same owner, and were managed as one unit. Fox Squirrel surveys The goal of the surveys was to determine presence or absence of Fox Squirrels on southeastern golf courses. We conducted phone surveys with staff at 98 golf courses. Typically the contact was either a course Figure 2. Extents for which land-cover data were analyzed to determine models of Fox Squirrel presence on 98 golf courses along the coast of South Carolina. Study courses were not located along the central portion of the coast because that area is relatively undeveloped, with large parcels of land in public ownership. 2010 K. Meehan and P.G.R. Jodice 577 superintendent or golf professional, as they were generally the most familiar with landscaping and ecological issues on the course. Along with questions about design and characteristics of golf courses, we also asked whether or not Fox Squirrels had been observed on the golf course during the past year. Although we requested estimates of Fox Squirrel abundance based upon number of sightings per month, we only used presence or absence categories for data analysis because we were not certain that estimates from course personnel were accurate and we were not able to conduct on-site surveys that resulted in adequate or balanced coverage across all golf courses (see below). Because the Fox Squirrel differs substantially from Sciurus carolinensis (L.) (Gray Squirrel) in size, coloration, and behavior, and because we included questions in our interviews that focused on coloration and behavior, we were confident that respondents were correctly reporting presence of Fox Squirrels. We also conducted on-site surveys for Fox Squirrels at 51 of the 98 golf courses to determine the accuracy of the interviews with respect to Fox Squirrel presence or absence and to obtain relative estimates of abundance. These abundance estimates were not used in data analysis, but rather as a qualitative assessment of population status on our study courses. We did not attempt to obtain a more accurate measure of density or abundance because we were restricted to conducting surveys from golf carts and from cart paths by most course personnel and had limited time on each course. These restrictions made it difficult to survey all portions of a golf course. Nonetheless, the high visibility within golf courses allowed for a thorough search of the active areas such as greens, fairways, tee boxes, and other managed landscapes which appear to represent the majority of habitat used by Fox Squirrels on golf courses in the southeastern US (Ditgen et al. 2007, Jodice and Humphrey 1992). Therefore, we are confident that presence and absence could be determined accurately. All surveys were conducted from golf carts driven at ca. 3.5 km/h along cart paths. We surveyed for squirrels using a strip-transect design (50 m on either side of the cart path). Because golf courses maintain a relatively open understory, we are confident that we met the assumption of equal detectability of individuals located on or near the ground throughout the strip transect. Surveys were conducted twice on each course: once between May and June 2005, and once between November 2005 and February 2006. Spring and summer surveys were conducted between sunrise and 1200 h, while fall and winter surveys were conducted between sunrise + 1 h and 1400 h. Survey times coincided with higher incidences of foraging and traveling activity of Fox Squirrels in the southeastern US (Jodice and Humphrey 1992) and provided the highest probability of detection. Given that Fox Squirrels in the southeastern US allocate substantial time to foraging and traveling on the ground during our survey hours and that these activities often occur along the ground (Jodice and Humphrey 1992, Koprowski 2005), we suggest that our ability to determine presence or absence was not compromised by 578 Southeastern Naturalist Vol. 9, No. 3 limiting our surveys to the area observed from the cart (ground, lower tree boles, lower canopy). Remote sensing/GIS We used aerial photos taken in 1999 (most recent complete set available), mailing addresses from the United States Golf Association member listings, GPS data collected during field visits, or geocoding of street addresses using ArcGIS 9.1 (ESRI, copyright 1999–2005) to locate and identify golf courses in the geographic information system. Because cadastral data were not universally available, we were not able to spatially define all golf courses by property boundaries. Instead we created a minimum convex polygon that circumscribed the outermost fairway for each course. This method resulted in each golf course being represented by a polygon that included all of the actively managed areas of the course. This delineation also matched the spatial extent of our on-site survey data, which were collected from cart paths. On some courses, this delineation resulted in an area that excluded associated housing developments or unmanaged land. These excluded areas were, however, included in analyses as buffers clipped to each course (see caption for Fig. 2 and description below for landscape clipping at three different scales). Land cover, tree canopy closure, and paved surface cover were extracted from the 2001 National Land Cover Database (http://www.mrlc. gov/mrlc2k_nlcd.asp accessed 12/12/2006), the most current land-cover data available for the study area. The NLCD is recommended for analysis of wildlife habitat data at the landscape scale (Cunningham 2006). Wetland classes were combined because previous research demonstrated that Fox Squirrels rarely use these land classes (Jodice 1993, Kantola and Humphrey 1990, Koprowski 1994). Grassland classes also were combined because Fox Squirrels appear too use these habitats primarily as movement corridors (Nupp and Swihart 2000). The multiple classes for developed land also were merged because the developed land-cover subtypes had lower levels of accuracy compared to other classes in our study area (Homer et al. 2004). A roads layer was created from 2000 Census Tiger/Line data, which was converted to a raster dataset based on the average width of each road type in the data set (American Association of State Highway and Transportation Officials 2004). We used Hawth’s Analysis Tools (Beyer 2004) and ArcGIS (ESRI, Redlands, CA) to clip the land-cover and road grids to individual golf courses at three different scales: the golf course itself, the golf course with a 1-km buffer, and the golf course with a 5-km buffer (Fig. 2). We used Fragstats 3.3 (McGarigal et al. 2002) to calculate landscape-, patch-, and class-scale metrics for all datasets. Data analysis We developed 22 logistic regression models (Table 1) to assess the relationship between presence of Fox Squirrels on a golf course and various 2010 K. Meehan and P.G.R. Jodice 579 landscape and habitat attributes on and adjacent to the course. We considered various aspects of the ecology of Fox Squirrels, urban ecology, and fragmentation effects on wildlife when developing these a priori models. For example, models included factors such as the proximity of other populations of Fox Squirrels (models 1–8), the extent of development on or surrounding the golf course (models 9–14), vegetation structure on and adjacent to the golf course (models 15–17), and individual attributes of golf courses (models 18–22). Select models included habitat and landscape characteristics surrounding the golf course by including the area within 1 and 5 km of the golf course polygon. We also used 19 of the 22 models to analyze data only on golf courses (n = 35) that did not have Fox Squirrels on their nearest neighboring golf course (Table 1). Table 1. Logistic regression models used for analysis of Fox Squirrel presence on 98 golf courses in coastal South Carolina. ModelA Parameters included in model Models that consider proximity of Fox Squirrel populations 1 Distance to edge of nearest neighboring course with Fox Squirrels; total area of courses within 5 km with Fox Squirrels 2 Distance to edge of nearest neighboring course with Fox Squirrels; total road area within 1 km of course 3 Presence of Fox Squirrels on nearest neighboring course 4 Distance to nearest golf course with Fox Squirrels 5 Total area of all courses within 5 km with Fox Squirrels 6 Distance to edge of nearest neighboring course with Fox Squirrels; total road area within 1 km of course; Presence of Fox Squirrels on nearest neighboring course 7 Distance to edge of nearest neighboring course with Fox Squirrels; total area of courses with Fox Squirrel populations within 5 km; presence of Fox Squirrels on the nearest neighboring course 8 Distance to nearest golf course with Fox Squirrels Models that consider the amount of developed area on and nearby the course 9 Total developed landcover within course 10 Total developed landcover within 1km of course 11 Developed landcover within 5 km of course 12 Total area on course with >50% impervious cover 13 Area of all road types within the course boundary 14 Area of all roads within 1km of the course Models that consider the amount of various vegetation classes on and nearby the course 15 Area within course of coniferous, deciduous, mixed, and shrub/scrub forest; grass, wetland, and open water 16 Area of cover within 1 km of course of coniferous, deciduous, and scrub/shrub forest; grass, wetland, and open water 17 Area of cover within 5 km of course of coniferous, deciduous, and scrub/shrub forest; grass, wetland, and open water Models that consider individual course attributes 18 Total course area; patch fractal dimension 19 Simpson's diversity index 20 Gray Squirrel presence or absence 21 Latitude of course 22 Year course was built AModels 6 and 7 were not used in instances where Fox Squirrels were absent from the nearest neighboring course. 580 Southeastern Naturalist Vol. 9, No. 3 Once the models were created, we analyzed each using PROC LOGISTIC (SAS/STAT system version 9.1, 1999–2005 SAS Inc., Cary, NC) and then used a model-selection approach to rank and assess models (Burnham and Anderson 2002). Models were ranked using AIC values corrected for small sample size (AICc), and we used AICc weights (wi ) to define a set of models which included the best model in 90% of samples (i.e., 90% confi- dence set of models, sensu Burnham and Anderson [2002]). This group was comprised of all the models which, when ranked, had a cumulative weight of approximately 90%. When using the model-selection approach, it is common to calculate estimates for regression coefficients (and hence odds ratios) for each variable in the confidence set, and to do so using a weighting scheme based on the model averages (Burnham and Anderson 2002). We applied this strategy to results that included all courses and present the weighted odds ratios in the results. For results that included just the subset of courses for which there were not Fox Squirrels present on the nearest neighboring course, we only present the odds ratios for variables from the highest ranked model because that model weight was substantially higher than all other models. All means and coefficient values are reported ± 1 SE unless otherwise noted. Results Survey and landscape summary statistics Fox Squirrels were reported as “present” on 35 of 47 (74.5%) golf courses contacted by telephone only and were observed on 33 of 51 (64.7%) golf courses where onsite surveys were conducted. On these latter golf courses, the maximum number of Fox Squirrels counted was 55, although on only 8 of these courses did we count ≥10 Fox Squirrels. Of the 51 golf courses where surveys were conducted, 36 reported via telephone surveys that Fox Squirrels were present. Hence, there was a 91.7% agreement in determining presence of absence or Fox Squirrels between phone and on-the-ground surveys. The three golf courses with conflicting results each categorized Fox Squirrels as very rare with <1 observation estimated per month. The proportion of courses with Fox Squirrels present did not differ between those surveyed in person and those for which presence was determined via telephone interviews (χ²=1.1, P >0.2). Based on these results, we therefore considered the phone surveys to be accurate for classification of Fox Squirrels as present or absent and subsequently analyzed data from all 98 golf courses. The golf courses included in this study had a mean age of 24 ± 1.5 years and a mean area of 126.8 ± 7.2 ha (range 25.5 to 400.4 ha). The golf courses included in this study were moderately forested, with 44.0 ± 1.7% (range 9.3%–83.4%) of the total area of the course having greater than 20% canopy cover (n.b., these values refer only to the course polygon itself and hence primarily include the fairways and interior roughs). The mean distance from a course to the next nearest course was 1.4 ± 0.2 km. For golf courses with Fox Squirrels, the mean distance to the nearest 2010 K. Meehan and P.G.R. Jodice 581 neighboring course that also supported Fox Squirrels was 1.9 ± 0.5 km, while for golf courses without Fox Squirrels the mean distance to the nearest neighboring course that supported Fox Squirrels was significantly (t96 = 2.2, P = 0.03) greater at 2.5 ± 0.4 km. Modeling Fox Squirrel presence Two models best predicted the presence of Fox Squirrels on golf courses along the South Carolina coast, and the 90% confidence set of models included only these two models (Table 2). The difference in the AIC values between the first and second-ranked models, however, was only 1.04, and hence these two models are not easily differentiated. Nonetheless, both the first- and second-ranked models included a term for the presence or absence of a Fox Squirrel population on the nearest neighboring course (NNFP). The odds ratio for NNFP indicated there was an 87.3% probability that Fox Squirrels would be present on a course if they also were present on the nearest neighbor course, but only a 31.4% probability that Fox Squirrels would be present on a course if they were not present on the nearest neighboring course. The second-ranked model also included variables that reflected proximity to nearby populations of Fox Squirrels (Table 2). We also assessed the relationship between Fox Squirrel presence and landscape and habitat variables just on courses where Fox Squirrels were not present on the nearest neighboring golf course (n = 35). We did this to determine which factors may be important to Fox Squirrel presence in the absence of a nearest neighboring course population. Fox Squirrel presence on this subset of golf courses was best predicted by the age of the golf course (Model 22; Table 3). This model had a 57% relative likelihood of being the best model for predicting Fox Squirrel presence on this subset of golf courses and was approximately 6.3 times more likely to be the best model than the next-highest ranked model. The odds ratio from this model indicated that the probability of a course having Fox Squirrels improved by 1.74% with each 1-year increase in course age. Coefficient estimates for other parameters in the 90% confidence set of models were very close to zero; since these parameters did not contribute strongly to the predictive power of these models, they are not discussed. Table 2. Model selection statistics from logistic regression modeling of Fox Squirrel presence on 98 golf courses in coastal South Carolina. Only models included in the 90% confidence set are presented. Cumulative sum of Model AICc AICc # Variables KA ΔAICc weight weights 3 Presence of Fox Squirrels on nearest neighboring 2 0 0.59 0.590 course 7 Distance to edge of nearest neighboring course with 4 1.044 0.35 0.939 Fox Squirrels; total area of courses with Fox Squirrel populations within 5 km; presence of Fox Squirrels on the nearest neighbor course ANumber of parameters in the model. 582 Southeastern Naturalist Vol. 9, No. 3 Discussion The models that best predicted the presence of southeastern Fox Squirrels on golf courses in coastal South Carolina included terms that were related to the presence of conspecifics on the nearest neighboring golf-course and the distance to and prevalence of those golf courses. Furthermore, the mean distance between two neighboring courses when each had Fox Squirrels was less compared to the distance between two neighboring courses where only one had Fox Squirrels. Perhaps surprisingly, variables that accounted for the amount of forested land within 1 km and 5 km of the golf course did not have a strong relationship with Fox Squirrel presence on a golf course. These results demonstrate that the proximity of golf courses with conspecifics increases the probability of Fox Squirrel presence on a golf course. Although we did not specifically assess movement patterns, the importance of the nearest neighboring course variable suggests that golf courses may be acting as habitat patches and that movement between these patches may be critical for maintaining populations. For example, Fox Squirrels can range widely and exhibit high rates of dispersal in both natural and urban patches (Ditgen 1999, Jodice 1993). Most of our study courses, which averaged ca. 125 ha in area, were small relative to the reported home-range size for Fox Squirrels in similar landscapes in the southeast. For example, Ditgen (1999) reported that home-range size of Fox Squirrels on golf courses in southwestern Florida were as large 304 ha across genders and age groups. The courses in that study are best characterized as open spaces within a matrix of intense development and formed an overall landscape similar to many of those available in our study area (e.g., the Myrtle Beach region along the northern coast and the Hilton Head region along the southern coast). Ditgen Table 3. Model selection statistics from logistic regression modeling of Fox Squirrel presence on golf courses in coastal South Carolina, which did not have a Fox Squirrel population on their nearest neighbor course (n = 35). Only models included in the 90% confidence set are presented. Cumulative Model AICc sum of # Variables KA ΔAICc weight AICc weights 22 Course age 2 0 0.571 0.571 15 Area within course of coniferous, deciduous, 7 3.672 0.091 0.662 mixed and shrub/scrub forest; grass, wetland, and open water 13 Area of all road types within the course boundary 2 5.235 0.042 0.703 16 Area of cover within 1km of course of coniferous, 6 5.239 0.042 0.745 deciduous, and scrub/shrub forest; grass, wetland, and open water 4 Distance to nearest course with Fox Squirrels 2 5.649 0.034 0.779 19 Simpson’s diversity index for habitat 2 6.045 0.028 0.806 12 Total area on course with >50% impervious cover 2 6.155 0.026 0.833 21 Latitude of course 2 6.526 0.022 0.855 10 Total developed landcover within 1 km of course 2 6.84 0.019 0.873 11 Total developed landcover within 5 km of course 2 6.87 0.018 0.892 9 Total developed landcover on course 2 6.886 0.018 0.91 ANumber of parameters in the model. 2010 K. Meehan and P.G.R. Jodice 583 (1999) observed substantial movement among golf courses and, as in our study, found that landscape-scale factors were more important in determining population size on a course compared to course-scale factors. We did observe a positive effect of individual attributes of golf courses (i.e., factors operating at a spatial scale less than the landscape level) on Fox Squirrel presence on courses which did not have conspecifics on their nearest neighboring course. The best model of this subset included only a term for the age of the golf course. The high ranking of this variable may be due to a combination of factors. Older golf courses may provide a more stable habitat since landscaping is more mature and development less active. The age of the golf course also may have acted as a surrogate for the location of the golf course in this study. Golf courses in the southern part of the state were older on average and appeared from our survey data to have higher abundances of Fox Squirrels (Meehan 2007). A greater frequency of occurrence of Fox Squirrels on southern courses may ultimately reflect historic abundances, however, as the central and southern portions of the South Carolina coast have traditionally had the greatest density of Fox Squirrels in the state (Harrigal 1993). Research conducted on Fox Squirrels in urban and fragmented settings has found that a combination of landscape and patch attributes tends to affect Fox Squirrel presence. Deuser et al. (1988) and Ditgen (1999) reported that Fox Squirrel presence in fragmented woodlots and golf courses, respectively, was primarily related to patch isolation and between-patch variables, which is similar to the nearest neighboring course effect we documented. Jodice and Humphrey (1992) and Lee et al. (2008), however, both suggested that local habitat features, or within-patch variables such as food availability, were important factors affecting Fox Squirrel populations. This observation is similar to the course-level effects we observed for courses without Fox Squirrels on their nearest neighboring course. We suggest that these two sets of apparently conflicting results may rather be factors operating at different spatial scales, where the ability of individuals to move among courses has a strong effect on presence versus absence, but local habitat features have a larger role in determining population abundance. Therefore our data should not be taken to suggest that Fox Squirrels do not respond to habitat factors within the boundaries of a golf course. Our inability to detect these effects is likely due to the scale of our research, as well as our analysis of squirrel populations at a presence/absence level. Our data, specifically the nearest-neighboring-course effect, suggest that movement of Fox Squirrels among golf courses in coastal South Carolina may be important and that golf courses in this region may have a greater opportunity to support or retain Fox Squirrel populations when the land area surrounding or near the course also supports Fox Squirrels. If so, the population stability of Fox Squirrels on golf courses in this region may be enhanced by the existence of movement corridors between nearby golf courses. This connectivity may be especially important in areas comprised 584 Southeastern Naturalist Vol. 9, No. 3 of smaller courses with apparently lower densities of Fox Squirrels, such as those found along the northern coast of South Carolina (e.g., the Myrtle Beach area; Meehan 2007). In areas of the state with larger courses and a lower level of present development overall, those interested in conservation of Fox Squirrels should consider maintaining large contiguous patches of habitat within areas experiencing development and by clustering development of new golf courses. Future research should investigate the metapopulation structure of Fox Squirrels on golf courses in this region. Furthermore, the development of a long-term monitoring plan focused on golf course habitats should be considered. Acknowledgments Funding for this research was provided by the National Fish and Wildlife Foundation Wildlife Links Program through the United States Golf Association. The US Geological Survey, South Carolina Cooperative Fish and Wildlife Research Unit supplied logistical support. The staff of the Waddell Mariculture Center, Bears Bluff National Fish Hatchery, and the Belle W. Baruch Institute for Marine and Coastal Sciences provided field housing. Thanks go to all the golf courses that participated in the research, and especially those that allowed repeated access to their course for surveys. This manuscript benefited from suggestions by J. Allen, S. Loeb, B. Song, R. Applegate, and anonymous peer-reviewers for constructive comments on the manuscript. The South Carolina Cooperative Fish and Wildlife Research Unit is supported jointly by the US Geological Survey, the South Carolina Department of Natural Resources, and Clemson University. Literature Cited Adkins, C.A., and P. Stott. 1998. Home ranges, movements and habitat associations of Red Foxes, Vulpes vulpes, in suburban Toronto, Ontario, Canada. Journal of Zoology 244:335–346. Allen, J., and K. Lu. 2003. Modeling and prediction of future urban growth in the Charleston region of South Carolina: A GIS-based integrated approach. Conservation Ecology 8:2. Available online at http://www.consecol.org/vol8/iss2/art2. Accessed June 2006. American Association of State Highway and Transportation Officials (AASHTO). 2004. AASHTO green book: A policy on geometric design of highways and streets. American Association of State Highway and Transportation Officials, Washington, DC. Angold, P.G., J.P. Sadler, M.O. Hill, A. Pullin, S. Rushton, K. Austin, E. Small, B. Wood, R. Wadsworth, R. Sanderson, and K. Thompson. 2006. Biodiversity in urban habitat patches. Science of the Total Environment 360:196–204. Barthelmess, E.L. 2004. Managing golf courses for small-mammal diversity. USGA Turfgrass and Environmental Research Online 3:1–10. Available online at http:// usgatero.msu.edu/v03/n22.pdf. Accessed June 2006. Beyer, H.L. 2004. Hawth's analysis tools for ArcGIS. Available online at http://www. spatialecology.com/htools. Accessed June 2006. Burnham, K.P., and D. Anderson. 2002. Model Selection and Multi-model Inference: A Practical Information-theoretic Approach. Springer-Verlag, New York, NY. 2010 K. Meehan and P.G.R. Jodice 585 Cristol, D.A., and A.D. Rodewald. 2005. Introduction: Can golf courses play a role in avian conservation? Wildlife Society Bulletin 33:407–410. Cunningham, M.A. 2006. Accuracy assessment of digitized and classified landcover data for wildlife habitat. Landscape and Urban Planning 78:217–228. Ditgen, R. 1999. Population estimates, habitat requirements, and landscape design and management for urban populations of the endemic Big Cypress Fox Squirrel (Sciurus niger avicennia). Ph.D. Dissertation. University of Florida, Gainesville, fl. Ditgen, R., J.D. Shepherd, and S.R. Humphrey. 2007. Big Cypress Fox Squirrel (Sciurus niger avicennia) diet, activity, and habitat use on a golf course in southwest Florida. American Midland Naturalist 158:403–414. Deuser, R.D., J.L. Dooley, and G.J. Taylor. 1988. Habitat structure, forest composition, and landscape dimensions as components of habitat suitability for the Delmarva Fox Squirrel. Pp. 414–421, In R.C. Szaro, K.E. Severson, and D.R. Patton (Eds.). Management of Amphibians, Reptiles, and Small Mammals in North America. US Forest Service General Technical Report RM-166, Rocky Mountain Forest and Range Experiment Station, Fort Collins, CO. Harrigal, D. 1993. Fox Squirrel (Sciurus niger) distribution and habitat presence in South Carolina. P. 53, In N.D. Moncrief, J.W. Edwards, and P.A. Tappe (Eds). Proceedings of the Second Symposium on Southeastern Fox Squirrels (Sciurus niger). Virginia Museum of Natural History, Martinsville, VA. 84 pp. Hodgkison, S.C., J.M. Hero, and J. Warnken. 2007. The conservation value of suburban golf courses in a rapidly urbanising region of Australia. Landscape and Urban Planning 79:323–337. Homer, C., C. Huang, L. Yang, B. Wylie, and M. Coan. 2004. Development of a 2001 national landcover database for the United States. Photogrammetric Engineering and Remote Sensing 70:829–840. Jodice, P.G.R. 1993. Movement patterns of translocated Big Cypress Fox Squirrels (Sciurus niger avicennia). Florida Scientist 56:1–6 Jodice, P.G.R., and S.R. Humphrey. 1992. Activity and diet of an urban-population of Big Cypress Fox Squirrels. Journal of Wildlife Management 56:685–692. Jodice, P.G.R., and S.R. Humphrey. 1993. Activity and diet of an urban population of Big Cypress Fox Squirrels: A reply. Journal of Wildlife Management 57:930–933. Jones, S.G., D.H. Gordon, G.M. Phillips, and B.R.D. Richardson. 2005. Avian community response to a golf course landscape unit gradient. Wildlife Society Bulletin 33:422–434. Kantola, A.T., and S.R. Humphrey. 1990. Habitat use by Sherman’s Fox Squirrel (Sciurus niger shermani) in Florida. Journal of Mammalogy 71:411–419. Koprowski, J.L. 1994. Sciurus niger. Mammalian Species 479:1–9. Koprowski, J.L. 2005. The response of tree squirrels to fragmentation: A review and synthesis. Animal Conservation 8:369–376. Lee, J.C., D.A. Osborn, and K.V. Miller. 2008. Characteristics of a high-density population of southern Fox Squirrels. American Midland Naturalist 159:385–393. Loeb, S.C., and N.D. Moncrief. 1993. The biology of Fox Squirrels (Sciurus niger) in the southeast: A review. Pp. 1–20, In N.D. Moncrief, J.W. Edwards, and P.A. Tappe (Eds). Proceedings of the Second Symposium on Southeastern Fox Squirrels (Sciurus niger). Virginia Museum of Natural History, Martinsville, VA. 84 pp. 586 Southeastern Naturalist Vol. 9, No. 3 Love, B. 1999. An environmental approach to golf course development. American Society of Golf Course Architects, Chicago, IL. Mason, C.F. 2006. Avian species richness and numbers in the built environment: Can new housing developments be good for birds? Biodiversity and Conservation 15:2365–2378. McCleary, R.A., R.R. Lopez, N.J. Silvy, and D.L. Gallant. 2008. Fox Squirrel survival in urban and rural environments. Journal of Wildlife Management 72:133–137. McGarigal, K., S.A. Cushman, M.C. Neel, E. Ene. 2002. Fragstats: Spatial pattern analysis program for categorical maps. University of Massachusets, Amherst, MA. Meehan, K. 2007. Landscape-scale correlates of Fox Squirrel (Sciurus niger) presence on South Carolina golf courses. M.Sc. Thesis. Clemson University, Clemson, SC. Montieth, K.E., and P.W.C. Paton. 2006. Emigration behavior of Spotted Salamanders on golf courses in southern Rhode Island. Journal of Herpetology 40:195–205. Nupp, T.E., and R.K. Swihart. 2000. Landscape-level correlates of small-mammal assemblages in forest fragments of farmland. Journal of Mammalogy 81:512–526. Perkins, M.W., and L.M. Conner. 2004. Habitat use of Fox Squirrels in southwestern Georgia. Journal of Wildlife Management 68:509–513. Terman, M.R. 1997. Natural links: Naturalistic golf courses as wildlife habitat. Landscape and Urban Planning 38:183–197. Weigl, P.D., M.A. Steele, L.J. Sherman, J.C. Ha, and T.L. Sharpe. 1989. The ecology of the Fox Squirrel (Sciurus niger) in North Carolina: Implications for survival in the southeast. Tall Timbers Research Station, Tallahassee, fl. 93 pp. Yasuda, M., and F. Koike. 2006. Do golf courses provide a refuge for flora and fauna in Japanese urban landscapes? Landscape and Urban Planning 75:58–68. Zipperer, W.C., J. Wu, R.V. Pouyat, and S.T.A. Pickett. 2000. The application of ecological principles to urban and urbanizing landscapes. Ecological Applications 10:685–688.