Effective conservation planning for endangered species depends on an understanding of space use patterns. Black-footed ferrets Mustela nigripes depend on prairie dogs Cynomys sp. as prey and use their burrow systems for shelter. The availability of areas with high densities of active prairie dog burrows is the major factor thought to affect their selection of sites and resources. However, we have little knowledge about how the spatial distribution of active prairie dog burrows might influence the spatial organization and home-range size of ferrets. We monitored the movements of black-footed ferrets on a black-tailed prairie dog C. ludovicianus colony in South Dakota to document ferret space use patterns. Home ranges of female ferrets were 22.9 - 95.6 ha in size (x̄ = 56.3 ha, SE = 19.7, N = 6), while male ferret home ranges were on average more than twice as large as those of females (x̄ = 128.3 ha, SE = 68.5, N = 3). The home-range size of female ferrets was correlated with mean active prairie dog burrow utilization distribution (UD) value within ferret home ranges, where home-range size decreased as active prairie dog burrow UD value increased (r2 = 0.974, P < 0.001, N = 6). Ferret space use overlapped more extensively than previously reported, with up to 43% UD overlap between a ferret and the nearest adjacent ferret of the same sex. Areas of overlap tended to have higher active prairie dog burrow UD values, suggesting that the spatial distribution of active prairie dog burrows influenced both home-range size and the amount of space use overlap between ferrets. These findings emphasize the potential influence of resource distribution on carnivore sociobiology and the importance of considering that distribution in assessing habitat for the reintroduction of specialized species.
Successful conservation of rare or declining carnivores requires management of habitat on a scale appropriate for the biological functions of those species (Woodroffe & Ginsberg 1998, Herfindal et al. 2005). The amount of space required for establishing self-sustaining populations of carnivores varies by species, yet within species the spatial distribution of resources may greatly influence carnivore space use patterns (Macdonald 1981, 1983, Norbury et al. 1998). For formerly extirpated species, such as the black-footed ferret Mustela nigripes (hereafter referred to as ferret), there is an urgent need to better understand how habitat conditions influence space use patterns, because such information might help improve reintroduction success (Griffith et al. 1989, Benson & Chamberlain 2007).
Spatial ecology of ferrets has been studied little due to their rarity and typically solitary, nocturnal and subterranean nature (Paunovich & Forrest 1987, Forrest et al. 1988). Ferrets occur exclusively on prairie dog Cynomys sp. colonies (Biggins et al. 2006b), rely on prairie dogs as their primary prey source (Sheets et al. 1972, Campbell et al. 1987), and spend most of their lives in underground burrow systems created by prairie dogs (Henderson et al. 1969, Richardson et al. 1987). Male ferrets, similar to other mustelids, do not assist in litter rearing, frequently have multiple female partners and have been hypothesized to select for areas frequented by female ferrets (Biggins et al. 1985, Richardson et al. 1987). Female ferrets are reproductively active immediately prior to one year of age and produce one litter each year throughout their life span (Miller et al. 1996). Female ferrets are solitary and are hypothesized to select and compete for areas with high densities of prairie dog burrows (Richardson et al. 1987, Biggins et al. 2006b, Jachowski 2007b). Yet how the spatial distribution of prairie dog burrows directly influences ferret home range and spacing patterns remains poorly understood (Biggins et al. 2006b).
Prey density is a major determinant of home-range size for predators (Sandell 1989, McLoughlin et al. 2000, Nilsen et al. 2005), and likely influences ferret space use patterns. Macdonald (1983) proposed the resource dispersion hypothesis for explaining variation in home-range size, where home-range size is primarily determined by the spatial distribution of resource patches. Although this hypothesis was developed for social carnivores, similar patterns have been observed for numerous solitary carnivore species (Carr & Macdonald 1986, Johnson et al. 2002, Eide 2004). Male ferret home ranges are typically much larger than those of females, though home-range size also varies greatly even among individuals of the same sex (Biggins et al. 1985, Livieri 2007). Under the resource dispersion hypothesis, such variation likely occurs due to variations in prey availability. The spatial distribution of prairie dogs and their burrows is heterogeneous (Jachowski et al. 2008), thus we hypothesized that ferrets occupying areas with higher densities of prairie dog burrows would have smaller home ranges.
Overlap of home ranges among ferrets of opposite sexes is believed to be common, with male home ranges overlapping the home ranges of multiple females, but overlap of home ranges among members of the same sex is thought to be less frequent (Richardson et al. 1987). Intra-sexual overlap in home ranges has been documented among males and among females; however, these observed overlaps were considered ‘boundary patrol’ movements (Richardson et al. 1987:235). Space use overlap for other solitary carnivore species has been reported to vary in response to fluctuations in prey abundance (Eide et al. 2004, Cochrane et al. 2006). For example, when habitat quality or prey availability drops below a certain threshold, there is a net benefit to defending quality areas of food resources (Carpenter & MacMillen 1976, McLoughlin et al. 2000). Female ferrets in particular would be expected to exhibit territoriality similar to other species that bear altricial young (Wolff 1997, Wolff & Peterson 1998). Given this, along with strong evidence that ferrets select for specific areas with high prairie dog abundance (Biggins et al. 2006b, Jachowski 2007b), we hypothesized that intra-sexual space use is uncommon between female ferrets and where it does occur it is likely a function of prey abundance.
We studied a successfully reintroduced ferret population to test hypotheses about ferret space use patterns and the relationship of prairie dog burrow distribution to ferret space use during the non-breeding or litter-rearing (May-October) season. Specifically, we quantified: 1) home-range size of individual ferrets, 2) spatial overlap between adjacent ferrets, and 3) spatial distribution of prairie dog burrows and how it was related to ferret home-range size and space use overlap.
Material and methods
Study area
Given the relatively recent reintroduction of ferrets to the Great Plains of North America (Lockhart et al. 2006, Jachowski & Lockhart 2009), we selected a study area with a long history of occupancy by ferrets. The Conata Basin is a portion of the Buffalo Gap National Grassland in southwestern South Dakota, which has extensive black-tailed prairie dog Cynomys ludovicianus colonies (approximately 11,803 ha; Fig. 1; Livieri 2006). From 1996 to 2000, 164 captive-born ferrets were released at Conata Basin (Livieri 2006). The population stabilized at an annual minimum population estimate of 200-290 individuals during 2000-2006 (Livieri 2006).
Within Conata Basin, we selected the distinct North Exclosure prairie dog colony as our study area (see Fig. 1). Prairie dog populations form distinct colonies that are typically defined by the maximum extent of their burrow systems (King 1959). The North Exclosure prairie dog colony was 202 ha in size and comprised 2% of the entire Conata Basin recovery area (see Fig. 1). We selected this colony for the following reasons: 1) the logistical benefits of its size for monitoring ferrets and prairie dog burrows, 2) its geographic isolation from other colonies by ravines of ∼ 15 m depth on the east, west and south sides, and 3) the continuous occupancy of the colony by ferrets for eight years prior to the initiation of our study. In 1997, six female and four male ferrets were released into the North Exclosure prairie dog colony. Subsequent annual surveys showed that the colony was occupied continuously by 3-11 adult ferrets since that time without subsequent releases (Livieri 2006). No captive-born animals were present during our study.
Field observations
We collected location data on all ferrets occurring in the study area during nearly consecutive nights from June through October in 2005 and 2006. We used systematic spotlight survey techniques to obtain ferret locations (Biggins et al. 2006a, Jachowski 2007a). We chose spotlight surveys instead of telemetry or other commonly used tracking techniques due to: 1) their logistical benefit for surveying a large flat area with short (< 10 cm) vegetation, 2) their widespread acceptance and use in locating and monitoring ferrets (Campbell et al. 1985, Grenier et al. 2009), and 3) our concern for the endangered status of ferrets and the need to minimize capturing or disturbing females during the litter-rearing period (May through October). We established a survey route that maximized coverage of the colony and minimized overlap. We traveled the route once each hour during the night to sample ferret locations. We drove the survey route with a 4-wheel-drive vehicle using a roof-mounted Lightforce® spotlight of > 1 million candlelight power. We traveled 8-16 km/hour with constant side-to-side scanning of the spotlight to detect ferret ‘eye shine’. Upon seeing ferret eye shine, we were able to identify an individual ferret > 94% of the time by visually identifying the animals or recording the animal's unique transponder number (Grenier et al. 2009). Every individual within our study area was marked with a PIT (passive integrated transponder) implant (Fagerstone & Johns 1987) prior to the start of the study. We captured male individuals upon first encounter to apply unique dye-markings on the front and back of the neck to allow for quick visual individual identification (Grenier et al. 2009). We waited until August to capture and dye-mark female ferrets to limit disturbance to females while they cared for young kits. In subsequent observations, if we were unable to discern clearly the dye mark on an individual, we placed a PIT reader at the burrow entrance for up to three days to record the unique PIT number. We obtained GPS coordinates for the location of each ferret observation using hand-held, Garmin GPS 12XL Personal Navigator units that have an accuracy error of < 7 m. When we observed a ferret more than once a night, we selected the first observation made ≥ 24 hours after the previous observation to allow for one activity-period cycle between observations (Goodrich & Buskirk 1998, Lair 1987).
We mapped the location of all prairie dog burrows in the colony in 2005 using Trimble CMT MC-V GPS data loggers (Trimble Navigation Limited, Sunnyvale, California, USA) mounted on all-terrain vehicles (see Jachowski et al. 2008). We downloaded burrow location data and differentially corrected locations using the nearest base station in GPS Pathfinder® Office 3.0 (Trimble Navigation Limited, Westminster, Colorado, USA). We obtained correction rates of 99-100% and therefore assumed burrow location error to be ≤ 1 m. While mapping, we classified burrows as active or inactive based on the presence of fresh scat or recent soil disturbance at burrow entrances (Biggins et al. 1993, Dullum 2001, Jachowski et al. 2008). We used active prairie dog burrows as a measure of the spatial distribution of resources for ferrets based on previous findings that ferrets spend a majority of their lives in prairie dog burrow systems and typically are sighted near active burrow entrances (Jachowski 2007b), that active prairie dog burrows are correlated with the presence of prairie dogs, and that burrow distribution commonly is used to examine patterns of prairie dog spatial distribution and abundance (Biggins et al. 1993, Powell et al. 1994, Johnson & Collinge 2004).
Data analysis
To assess space use by individual ferrets, we used locations to estimate fixed kernel utilization distributions (UDs; Kernohan et al. 2001). Our assumption that the population was effectively closed due to physical barriers and prairie dog colony boundaries was supported by our monitoring, because we regularly relocated all individuals in our study area. Therefore, we included all observed animals in our analyses. We acquired more than the minimum number of observations (N > 30; Seaman et al. 1999) on each individual within the 5-month sampling period each year and calculated a 95% fixed kernel UD for each ferret (Fig. 2). We used Matlab (Mathworks Incorporated, Natick, Massachusetts, USA) and ‘plug-in’ methods for bandwidth selection (Wand & Jones 1995, Gitzen et al. 2006) using the ‘Kde folder’ (Beardah & Baxter 1995) to create individual UDs. We utilized data from all ferrets observed during each sampling year to analyse space use overlap. To avoid pseudoreplication in our estimates of mean home-range size, we randomly selected one year for each of the three animals observed during both years (Hurlbert 1984).
We assessed space use overlap of ferrets by determining how much overlap existed in the 95% kernel UD of each ferret with all other ferrets observed during the same sampling year. We used the Volume of Intersection (VI) technique to quantify the degree of overlap in volume of two individual ferret UDs as
where f̂A is the estimated UD for ferret A and f̂B is the UD for ferret B (Seidel 1992, Millspaugh et al. 2004). VI scores range from 0 to 1 where a VI score of 1 indicates complete overlap of the UDs.
To assess the relationship between the distribution of active prairie dog burrows and ferret home-range size, we established the home-range size for each individual ferret by measuring the maximum extent of each ferret's 95% fixed kernel UD. We then quantified the spatial distribution of prairie dog burrows by using the comprehensive location data of active burrows to create a UD of all active prairie dog burrows in the study area using the fixed kernel approach described above (Jachowski et al. 2008). We used Spatial Analyst tools within ArcGIS 9.1 (Environmental Systems Research Institute, Redlands, California, USA) to clip the active prairie dog burrow UD by the maximum extent of each ferret's 95% fixed kernel home range. We then measured the mean active prairie dog burrow UD value within each ferret's home range. Using linear regression, we then assessed the relationship of home-range size to mean active burrow UD value within the home range for each ferret.
We also assessed the relationship between home-range overlap and active prairie dog burrow density. We paired individual ferrets of the same sex that had the highest degree of UD space use overlap (i.e. highest VI values). We randomly selected 200 locations from within the area where space use overlap occurred to measure active burrow UD values and compared those values to 200 randomly selected locations from the area within the selected ferret's home range where no overlap occurred. We then calculated the mean and standard errors for the UD values at those locations and compared the distribution of values by sex.
Results
We collected > 30 observations on each of nine individual ferrets in our study area (three males and six females) during our study period (x̄ = 47.6, SD = 11.1, range: 32-71; Table 1). We did not observe other adult ferrets at any point during our study, supporting our assumption that the population was closed and that we measured space use for all animals present in the study area. Of the nine individuals, three (one male and two females) were present in both years (Table 2).
Ferret home-range size was related to the spatial distribution of active prairie dog burrows (see Fig. 2). Home-range estimates for female ferrets ranged from 22.9 to 95.6 ha (x̄ = 56.3 ha, SE = 19.7, N = 6; see Table 1). Home-range estimates of males averaged more than twice the size of female home ranges (x̄ = 128.3 ha, SE = 64.5, N = 3). Because of few individuals, we omitted males when assessing the influence of resource distribution on home-range size. Female ferret home-range size showed a negative relationship to mean active burrow UD values within the home range (r2 = 0.974, P <0.001, N = 6; Fig. 3), indicating that female ferrets had smaller home ranges in areas with the highest relative burrow density.
We observed considerable overlap in space use between ferrets of the same and opposite sex. Although space use overlap between females varied widely (range of VI values: 0.01-0.43), six of the eight females monitored had at least 0.18 or 18% overlap in volume of their UD with another female (range: 0.18-0.42; see Table 2). Similarly, six of eight females had at least 0.13 or 13% overlap in the volume of their UD with another male (range: 0.13-0.40; see Table 2). The presence of only two males during each year of our study restricted our analyses to two measures of overlap (one for each pair each year).
Ferrets were more likely to exhibit space use overlap in areas of relatively high active prairie dog burrow density during the litter-rearing season. Active burrow UD values within areas of overlap between a female ferret and its nearest same-sex neighbour were higher (x̄ = 0.06, SE = 0.00, range: 0.00-0.10) than in areas without overlap (x̄ = 0.04, SE = 0.00, range: 0.00-0.11). In contrast, active burrow UD values of a male ferret and its nearest same-sex neighbour were similar within areas of overlap (x̄ = 0.04, SE = 0.00, range: 0.00-0.11) and in areas without overlap (x̄ = 0.04, SE = 0.00, range: 0.00-0.12).
Discussion
As predicted, ferret home-range size varied among individuals based on sex and the spatial distribution of resources, a pattern which exists for other solitary carnivore species (Sandell 1989, Eide et al. 2004). Sex-specific differences in home-range estimates during the litter-rearing season in our study were similar to annual home-range estimates of 132 ha for males and 65 ha for females previously reported by Livieri (2007) for ferrets in our study area. The strong inverse relationship between female ferret home-range size and the availability of active prairie dog burrows that we observed provides unique support for the hypothesis that the spatial distribution of this critical resource likely explains most of the variation in home-range size observed among individuals of the same sex.
In contrast to previous hypotheses regarding female ferret territoriality, we found that where resources were more abundant, home ranges of female ferrets tended to overlap. Generally, when habitat quality is high, individuals exhibit less territoriality (Carpenter & MacMillen 1976, McLoughlin et al. 2000). For example, Eide et al. (2004) observed weaker territoriality and increased sociality among arctic foxes Alopex lagopus where resources were clumped, highly abundant and predictable. Similarly, female ferrets in our study exhibited decreased territorial behaviour in areas of high quality habitat or where prey were more abundant. These findings contradict previous hypotheses that predict relatively exclusive territories among female ferrets (Clark 1989), and provide novel support for the hypothesized relationship between habitat quality and space use overlap that has been suggested to occur in ferrets (Biggins 2000, Biggins et al. 2006b, Livieri 2007). The relatively high degree of overlap (up to 42%) that we observed among female ferrets goes beyond previous observations of occasional or exploratory overlap among ferret territories (Paunovich & Forrest 1987, Richardson et al. 1987), and suggests that for female ferrets rearing young where prey and burrows are relatively abundant, the cost of defending exclusive territories might outweigh any potential benefit from defending exclusive territories.
Our findings support application of the resource dispersion hypothesis (Macdonald 1983) to solitary carnivore species. Sandell (1989) proposed that the resource dispersion hypothesis could be extended beyond social carnivores. Broadly applied to solitary carnivores, females would have exclusive ranges when food resources are ‘stable and evenly distributed’ and would have overlapping ranges when food distribution and timing varied. Support for this hypothesis has been mixed (Herfindal et al. 2005), where a direct relationship between habitat quality and home-range size or space use overlap has not been universally observed in social (Creel & Creel 2002) or solitary (Bixler & Gittleman 2000) carnivore species. Clear support of the importance of resource distribution on ferret spatial ecology was likely due to 1) the importance of active prairie dog burrows to highly specialized ferrets, 2) the heterogeneous distribution of resources, and 3) the fine scale measures of resource distribution that we employed. Similar to other mustelid species, ferrets are likely primarily concerned with locating prey, limiting vulnerability to predators and conserving energy (King & Powell 2007). For ferrets, these three factors are associated with active prairie dog burrows. Where Bixler & Gittleman (2000) found no relationship between home-range size and habitat conditions for omnivorous striped skunks Mephitis mephitis, which likely viewed their site as having a uniform distribution of resources, active prairie dog burrows are heterogeneously distributed within colonies (Jachowski et al. 2008) and are occupied by populations of prairie dogs which can fluctuate annually due to biotic and abiotic factors (Hoogland 1995). Previous failures to detect a close association between a critical resource and carnivore space use patterns could also be due to poor classification of habitat quality (Herfindal et al. 2005). By precisely mapping locations of active prairie dog burrows throughout a colony and taking a utilization distribution approach which measures the relative distribution of those burrows, we quantified the fine-scale distribution of resources available to ferrets within our study area.
Although the spatial distribution of prairie dog burrows likely influenced home-range size and space use overlap between individual ferrets, factors such as season, rank, kinship, physiological condition and age also have been identified as influencing space use patterns of other carnivores (Rolls et al. 2001, Boydston et al. 2003, Støen et al. 2005). While our small sample size and season of study did not allow for analysis of these factors, we suspect that seasonality, in particular, is likely to affect space use by ferrets. Our period of study occurred during the litter-rearing season, but ferret space use likely differs between breeding and non-breeding times in a pattern similar to that of other mustelids (Erlinge 1977, Johnson et al. 2000, King & Powell 2007). Female ferrets might have smaller home ranges during the litter-rearing period (Robitaille & Raymond 1995) and males might be less territorial during that period than during the winter breeding season (Richardson et al. 1987). Additionally, ferret populations typically experience high turnover within and between years because of their relatively short life span and high reproductive rate (Grenier et al. 2007). This, too, has the potential to affect ferret spatial organization through the frequency with which newly vacated habitat becomes available.
Conservation implications
Current coarse-scale techniques used to evaluate the distribution of critical resources might be insufficient to determine the suitability of reintroduction sites for extirpated species. With the advancement of computer mapping software, the use of spatially explicit techniques to assess resource distribution has become a valuable tool in assessing reintroduction site suitability for extirpated carnivore species (Mladenoff & Sickley 1999, Schadt et al. 2002). For carnivores in particular, field surveys of prey density are logistically demanding and rarely are made at a fine spatial scale (Nilsen et al. 2005). Currently, prairie dog populations at ferret reintroduction sites are assessed by measuring the maximum spatial extent of active prairie dog burrows as well as the average density of active burrows (based on counting the number of active burrows on approximately 10% of a colony) within each area on an annual or semi-annual basis (Biggins et al. 1993). These measurements are used by the U.S. Fish and Wildlife Service in assessing and ranking the biological suitability of reintroduction sites in a competitive process of allocating a limited number of captive and wild-born ferrets for release into the wild (Jachowski & Lockhart 2009). Using such an approach, the release or relocation of > 3,000 ferrets into 18 reintroduction sites over the last 18 years has resulted in establishing only four relative large (100-300 individuals), self-sustaining populations (Jachowski & Lockhart 2009). While our findings generally support the use of prairie dog burrow distribution and density as a coarse indicator of habitat suitability, the dynamic nature of prairie dog populations within colonies and the influence of active burrow density on ferret space use suggest that finer-scale and more spatially explicit measurements of the spatial distribution of prairie dogs might help to predict the suitability of sites for reintroduction of ferrets. By identifying areas of high prairie dog density that are likely to serve as high quality habitat for ferrets, managers could more effectively prioritize potential reintroduction sites and more accurately focus efforts to protect prairie dog populations from human influences or disease outbreaks.
Given the role of habitat quality in ferret space use requirements, managers might be able to increase the number of ferrets within a reintroduction site and improve reintroduction success by enhancing prairie dog population density. Reintroduced ferret populations have persisted without augmentation at four of the largest prairie dog complexes (Lockhart et al. 2006, Jachowski & Lockhart 2009), but few remaining prairie dog colonies or complexes of such a size exist (Forrest 2005). To achieve recovery goals, self-sustaining ferret populations will have to be established in smaller prairie dog colonies. It might be possible to maintain more ferrets in a relatively small area if that area is able to maintain a high population density of prairie dogs. This would minimize the total area required for each ferret and maximize the amount of tolerated space use overlap. Management by altering vegetative cover (Milne-Laux & Sweitzer 2006), reducing shooting pressure (Pauli & Buskirk 2007) and prairie dog translocation (Truett et al. 2001, Dullum et al. 2005) could enhance prairie dog abundance at reintroduction sites. Encouraging such management practices could improve the ability of sites to maintain higher densities of ferrets and increase the likelihood of self-sustaining ferret populations.
Acknowledgments
funding and logistical support for this project were provided by the U.S. Fish and Wildlife Service, U.S. Forest Service, Prairie Wildlife Research, National Fish and Wildlife Foundation (Grant number 2006-0058-0000), South Dakota Department of Game, Fish and Parks Small Grants Program, South Dakota Department of Game, Fish and Parks Competitive Grants Program, Disney Centennial Refuge Scholarship, Missouri Chapter of the Wildlife Society and the University of Missouri. This study was funded in part by federal funding through State Wildlife Grant T-35-R, Study #2435, administered through the U.S. Fish and Wildlife Service. We thank R. Jachowski and D. Hanson for help with fieldwork. We also thank K. Atchley and D. Sargent who provided valuable assistance in support of this project. We thank R. Jachowski, C. Bodinof, D. Eads, C. Rittenhouse, T. Bonnot, C. Hansen, J. Kolar, B. Keller and B. Dijak for comments that improved the manuscript. Any use of trade, product or firm names is for descriptive purposes only and does not imply endorsement by the U.S. Government.
References
Figure 1.
Maximum extent of prairie dog colonies (solid grey) in 2005 within the Conata Basin region of the Buffalo Gap National Grassland managed by the U.S. Forest Service (white), National Park Service (speckled) and private landowners (cross-hatched) in southwestern South Dakota. The North Exclosure prairie dog colony study area is highlighted in black.
Figure 2.
Comprehensive mapping of active prairie dog burrows was used to calculate 95% fixed kernel utilization distribution (UD) of active burrows at the Conata Basin study site in 2005 (A). Areas of high active prairie dog burrow distribution (dark areas) are similar to areas with large UD values for individual female ferrets observed during 2005: F02015 (B), F02047 (C), F03031 (D) and F04019 (E) as well as 2006: F03031 (F), F04019 (G), F05003 (H) and F05005 (I). The height and darkness of the ferret UD indicates areas with a higher probability of use by that individual ferret.
Figure 3.
Home ranges of individual female ferrets (N = 6) as a function of mean active prairie dog burrow utilization distribution (UD) value at the Conata Basin in 2005 and 2006. Standard error values for all ferrets were < 0.006. The mean prairie dog UD value for the entire study area was 0.05 (SE = 0.02).
Table 1.
Number of locations and home-range estimates (perimeter of 95% of UD by volume) for male and female ferrets monitored in 2005 and 2006 (year of observation) at the Conata Basin, South Dakota.
Table 2.
Amount of overlap between ferret utilization distributions, as measured through volume of intersection scores (Seidel 1992, Millspaugh et al. 2004) for all ferrets present on the study area during 2005 (A) and 2006 (B) at the Conata Basin, South Dakota. Volume of intersection scores range from 0 to 1, with 1 representing complete overlap in space use between two ferrets.
