Printer Friendly

Home-range size and habitat selection by male Island foxes (Urocyon littoralis) in a low-density population.

Documenting home-range size and habitat use patterns are fundamental components of developing effective management strategies for endangered species (Morrison et al., 1998; Powell, 2000; Coonan et al., 2010; Goble et al., 2012). However, complications can arise if these attributes vary with population density. A negative correlation between home-range size and population density has been documented in a wide variety of mammals including rodents (Erlinge et al., 1990), ungulates (Kjellander et al., 2004), and carnivores (Benson et al. 2006). Population density can also affect habitat use patterns because, at low population density, individuals can select high-quality habitat and avoid low-quality habitat (Krohn, 1992; Biggins et al., 2006), but at high population density individuals may occupy virtually any available habitat regardless of its quality (Van Horne, 1983). Density-dependent factors become especially complicating with insular species that cannot use long-distance dispersal to escape high local population densities.

These issues have become particularly relevant to the conservation and management of the endangered island fox (Urocyon littoralis), which is endemic to six of the eight Channel Islands off the coast of southern California (Fig. 1). Field studies conducted prior to the mid-1990s, when population densities were relatively high, concluded that island foxes had small home ranges, with considerable overlap among adjacent foxes, and exhibited little if any selection among habitat types (Laughrin, 1977; Moore and Collins, 1995; Crooks and Van Vuren, 1996). The situation changed dramatically in the late 1990s when the fox populations on Santa Rosa, San Miguel, and Santa Cruz, suddenly plummeted (Roemer, 1999; Coonan et al., 2003; Coonan et al., 2010). The crash was most severe on Santa Rosa Island (SRI), where the population of approximately 1,780 individuals in 1994 fell to 15 individuals (>99% reduction) by 1999 (Coonan et al., 2010). These declines were caused by predation from golden eagles (Aquila chrysaetos) which had established following the extirpation of bald eagles (Haliaeetus leucocephalus) and the introduction of nonnative ungulate prey such as pigs (Sus scrofa) and elk (Cervus canadensis; Roemer et al., 2002; Coonan et al., 2010). Populations on these three islands were subsequently listed as federally endangered (U.S. Fish and Wildlife Service, 2004), and captive breeding programs were initiated on all three islands to protect the surviving foxes while the eagles and ungulates were removed. The conservation strategy was successful; the captive breeding programs ended by 2008, and the fox populations on all three islands are now recovering toward their former abundance. However, the SRI population has had the slowest rate of population increase, as determined by a standardized live-trapping program initiated to monitor the population's recovery (Coonan, 2010; Coonan et al., 2010). Based on predecline estimates of population density, SRI probably supported an adult population of approximately 1,700 foxes (4 adults/[km.sup.2]), but had recovered to <400 foxes (0.86 adult foxes/[km.sup.2]) in 2009 when our study began (Coonan, 2010). This unusually low population density on SRI provided the opportunity to document home-range size, home-range overlap, and habitat selection when intraspecific competition for resources was presumably minimal. Documenting density-dependent changes in these attributes could be important for the successful management of this potentially conservation-reliant species (Scott et al., 2005). Our objectives were to 1) quantify home-range area and overlap, 2) document any selection for vegetation types or topography within home ranges, and 3) determine whether habitat selection within home ranges varied by time of day. We predicted that the SRI foxes would have larger home ranges than those previously reported for higher-density island fox populations, with minimal overlap between adjacent home ranges.

[FIGURE 1 OMITTED]

MATERIALS AND METHODS--SRI is the second largest (217 [km.sup.2]) of the eight California Channel Islands (Fig. 1) and is managed by the National Park Service (NPS). The island's climate is Mediterranean with a marine influence and its topography is rugged with a maximum elevation of 484 m. Vegetation types include coastal grassland, island chaparral, scrub oak woodland, coastal sage scrub, coastal bluff scrub, riparian woodland, Torrey pine forest, and bishop pine forest (Schoenherr et al., 1999).

We used Quantum 4000E Mini Collar GPS units (Telemetry Solutions, Concord, California), which included both a GPS receiver and a very high frequency (VHF) transmitter. Collars weighed 65-70 g depending on the specific design. Basic units stored the locality data within the unit and had to be retrieved in order to download the data. Some units included a remote download function which allowed the data to be downloaded using a specialized base station connected to a laptop computer. All units included a 6-h mortality sensor.

Prior to deploying the GPS collars on foxes, we assessed their performance by placing randomly-selected collars in the island's three main structural habitats: grassland (n = 5 trials), scrub (7 trials), and woodland (4 trials). For each trial, the test collar collected data for 1-3 h with 5-15 min between attempts. We then quantified the proportion of GPS location attempts that were successful ("fix rate") and the precision of the acquired locations compared to the locations acquired by a Trimble GeoXT survey-grade GPS unit (Trimble, Sunnyvale, California) with sub-meter accuracy. We assessed all three vegetation types on the same days to minimize confounding due to satellite configuration.

We captured foxes for GPS deployment from July through December 2009 as part of the ongoing NPS program to monitor island fox survival, reproduction, and general health (see Coonan et al., 2010). We captured the foxes in single-door, wire-mesh box traps (66 x 23 x 23 cm; Tomahawk Live Trap Co., Tomahawk, Wisconsin) baited with cat food and loganberry lure (Knobb Mountain Fur Co., Berwick, Pennsylvania). Traps were checked each morning and captured foxes were physically restrained without immobilizing drugs. We recorded the weight, sex, reproductive status, and general body condition of captured foxes. Foxes were captured and handled in accordance with accepted guidelines (Sikes et al., 2011). We deployed the collars on adult male foxes weighing [greater than or equal to] 1.65 kg, and we limited our deployments to the eastern half of SRI to ensure that the collared animals had adjacent home ranges.

We programmed the GPS collars to attempt to acquire a fix every 7 h, which yielded 3-4 locations per day and a fix attempt for each hour of the 24-h diel cycle every 7 d. We set the time allotted to acquire a fix at 90 s for 75% of the fix attempts and at 60 s for the remaining 25%, followed by an additional 45 s if the fix was not acquired in the allotted time, giving a total possible time of 105-135 s to acquire a fix. These settings yielded an expected GPS battery life of 200 d and a maximum of 670 locations per collar. To maximize the temporal overlap among the datasets, collars were programmed to remain dormant until 22 September 2009; collars deployed after this date (n = 2) began collecting data immediately.

Location data were retrieved by recapturing the foxes or via remote download from a fixed-wing aircraft. We used the locations to create 95% and 50% minimum convex polygons (MCPs) for each fox using the fixed mean algorithm in the Home Range Tools extension (http://flash.lakeheadu.ca/ ~arodgers/hre/) for ArcGIS 9 (ESRI, Redlands, California). MCPs are analytically and conceptually simple and can be directly compared with previous studies (Harris et al., 1990; White and Garrott, 1990). We clipped the MCPs by the island shoreline to exclude ocean areas where the foxes never entered. We calculated the overlap area between adjacent MCPs using the Intersect tool in ArcGIS, then calculated the proportion of overlap relative to the area of each home range (HR1 and HR2) using the following equation (Minta, 1993):

Mean Overlap = [square root of ([overlap/HR1] x [overlap/HR2])]

To assess habitat utilization, we also created 95% and 50% KDIs for each fox using the Geospatial Modeling Environment ver. 0.5.3 (www.spatialecology.com) and ArcGIS 10. We used a Gaussian kernel, set the bandwidth using plugin optimization (Jones et al. 1996), and used an output cell size of 10 m. Ocean areas were excluded as noted above for MCPs. We calculated the overlap area between adjacent KDIs using the same methods as for MCPs, and we also calculated the minimum overlap volume for adjacent 95% KDIs (Kertson and Marzluff, 2009).

We portioned fix times into three diel categories based on their timing relative to daily sunrise and sunset times (U.S. Naval Observatory, http://aa.usno.navy.mil/data/docs/RS_OneYear. php). We classified fixes that were within 2 h before or after local sunrise or sunset as crepuscular, fixes that occurred between 2 h after sunrise and 2 h before sunset as day, and fixes that occurred between 2 h after sunset and 2 h before sunrise as night.

When our study began, the most-recent vegetation data layer for SRI was based on aerial photographs from 1988, which clearly did not accurately represent the current vegetation boundaries. We therefore created a new vegetation map using interactive supervised image classification in ArcGIS 10, based upon 1-m resolution color infrared aerial imagery taken in 2009 (U.S. Department of Agriculture, http://www.fsa.usda.gov/ FSA/apfoapp?area=home&subject=prog&topic=nai). We defined six vegetation types: bare ground, chaparral, coastal scrub, grassland, lupine scrub, and woodland. To ensure accuracy, NPS staff familiar with the vegetation of our study site reviewed and manually edited the resulting vegetation classifications and polygon boundaries.

We characterized topography using topographic position index analysis (Jenness Enterprises, www.jennessent.com/ arcview/tpi.htm) in ArcGIS 10. Topographic position index is determined by evaluating each point relative to a generalized or smoothed surrounding terrain across the study area. We used a 3-m resolution light detection and ranging digital elevation model and determined the local mean elevation using a 500-m window. Residuals from the local mean elevation were used to assign each pixel to one of five topographic position index classes: valley bottom, low slope, mid slope, upper slope, and ridge.

To investigate selection for vegetation types or topographic classes, we used Euclidean distance analysis (Conner and Plowman, 2001). Euclidean distance analysis compares the observed fox locations against random points representing the expected habitat use in the absence of selection. This technique is appropriate for telemetry locations that may contain random spatial errors along with potential classification errors in the available resources (Conner and Plowman, 2001). Because Euclidean distance analysis is based on proximity rather than assignment, each location can be used to quantify selection for each available resource, including resources that are not actually used by each fox (Conner et al., 2003). We measured the linear (Euclidean) distance from each fox location and each random point to the nearest polygon of each vegetation community (six categories) or topographic class (five categories). For each fox, we created a ratio for each category of vegetation or topography by dividing the mean of the distances from the observed locations by the mean of the distances from the random points. We used a multivariate analysis of variance to determine if the average ratio across animals was significantly different from 1.0 (Conner and Plowman, 2001); ratios significantly >1.0 indicated avoidance of the habitat type (observed distance is greater than expected by chance) and ratios significantly <1.0 indicated selection.

We investigated habitat selection within the 95% KDI home range and within the 50% KDI core area. We defined the boundary of our study area as a 100% MCP around the complete set of fox locations, with the ocean portions removed as when calculating the animals' home ranges. We then generated 10,000 random points within the study area. Using the points that fell within the 95% KDIs (~5,000 of the 10,000 random points), we assessed selection within the home range by calculating the ratio of the average distance to each habitat type from the GPS collar locations, divided by the average distance to each habitat type from the random points within the home range. Similarly, we assessed selection within the core areas by calculating the ratio of the average distance to each habitat type from the GPS locations within the 50% KDI, divided by the average distance to each habitat type from the random points within the home range. Ratios were created for each fox and were natural log transformed to calculate the population average. We analyzed the ratios in SAS (SAS Institute, Cary, North Carolina) using a general linear model-multivariate analysis of variance to test whether habitat use differed from random (Conner and Plowman, 2001).

To examine diurnal patterns of habitat selection within the 95% KDIs, we used a multivariate analysis of variance with time of day (crepuscular, day, night) as a fixed effect. If the fixed effect was significant, we examined the results for individual habitat types. We used an orthogonal contrast in SAS to compare selection between diel categories (day vs. night, crepuscular vs. day, crepuscular vs. night) within habitat types. We examined vegetation and topographic categories separately. Results were considered statistically significant if P < 0.05.

RESULTS--The GPS collars performed well in the predeployment tests, with an overall fix rate of 99.4%, average time to fix of 27 s, and average location error of 8.2 m. Signal precision (horizontal dilution of precision; HDOP) and location error differed slightly among the three vegetation types tested. All of the grassland fixes were high precision (3D, HDOP < 5.0), with an average ([+ or -] SE) location error of 3.5 [+ or -] 0.3 m. In woodland, 96% of the fixes were high precision with an average location error of 11.6 [+ or -] 3.0 m; the one remaining fix was medium precision (2D, HDOP < 5.0) with a location error of 17 m. In scrub, 80% of the fixes were high precision with an average location error of 7.2 [+ or -] 0.7 m; 10% of the fixes were medium and low (HDOP > 5.0) precision with an average location error of 6.5 [+ or -] 0.9 m and 24.5 [+ or -] 11.1 m, respectively.

We deployed 14 GPS collars (eight store-on-board and six remote download models) on foxes. Collared foxes weighed 1.75-2.60 kg, resulting in unit/body mass ratios of 2.5-3.7%. We recovered a total of 11 data sets: eight collars were recovered during annual trapping efforts in 2010 and three additional partial data sets were acquired through remote downloads for animals that were not recaptured. No data were recovered from two animals with store on-board collars that were not recaptured, and one collar from a fox mortality was not included in our analyses due to its small data set (~3 wk).

The deployed GPS collars performed well but failed sooner than expected. The GPS components lasted an average ([+ or -]SE) of 16.5 [+ or -] 1.7 wk, which was 59.1 [+ or -] 6.0% of the expected operation time. Fix rate averaged 82.3 [+ or -] 2.1% (range: 72.0-92.8%), yielding an average of 364.0 [+ or -] 34.9 locations per fox (range: 220-524; Table 1). On average, high-precision fixes accounted for 88.1% of the locations for each fox, with medium and low precision accounting for 6.9% and 5.0%, respectively. (For more details on collar performance, see Cypher et al., 2014a and Drake, 2013). Avery small number of locations (n = 5) for the coastal foxes plotted in the ocean rather than on land, so we removed these points prior to analyses. All the other locations were used to generate the home-range estimates.

The mean 95% MCP home-range area was 3.39 [+ or -] 0.59 [km.sup.2] (Table 1) and mean 50% MCP area was 1.26 [+ or -] 0.23 [km.sup.2]. There was no significant linear association between the number of locations used and the 95% MCP area (linear regression; [F.sub.1,9] = 1.15, P = 0.31) or 50% MCP area ([F.sub.1,9] = 0.09, P = 0.77), or between fox mass and 95% MCP area ([F.sub.1,9] = 2.13, P = 0.18) or 50% MCP area ([F.sub.1,9] = 3.18, P = 0.11). Overlap between the 95% MCPs of neighboring males ranged from 0.1-28.3% (median = 5.3%) with an overlap area of 0.003-0.18 [km.sup.2] (median = 0.15 [km.sup.2]). There was no overlap among any of the 50% MCPs.

The 95% KDI home ranges were similar in size to the 95% MCPs (correlation = 0.87), averaging 3.82 [+ or -] 0.68 [km.sup.2] (Table 1). Most KDI home ranges had 2-3 core areas with a total average area of 1.06 [+ or -] 0.23 [km.sup.2]. Overlaps between 95% KDIs ranged from 2.1-11.2% (median = 6.0%) with an overlap area of 0.12-0.77 [km.sup.2] (median = 0.26 [km.sup.2]). Overlap between 50% KDIs was minimal: only three dyads of neighboring males had overlaps, ranging from 0.003-0.10 [km.sup.2] (median = 0.02 [km.sup.2]).

Based on the recovered GPS datasets, our study area encompassed 72.7 [km.sup.2] with grassland being the dominant vegetation type (54.7%), followed by scrub (24.1%) and chaparral (8.8%; Table 2). On average, fox home ranges were composed primarily of grassland (55.2%) and scrub (19.7%), but there was considerable variation among individual foxes (Table 2). There was no significant linear association between the area of the 95% KDI and the proportion of the primary habitats in the home range (grassland: [F.sub.1,9] = 0.73, P = 0.73; scrub: [F.sub.1,9] = 0.04, P = 0.85). The foxes exhibited no significant selection for vegetation type within their home ranges ([F.sub.6,5] = 3.36, P = 0.10). However, time of day did affect selection of vegetation type within the home range ([F.sub.18,71] = 2.35, P < 0.01; Fig. 2), with both grass and bare areas selected at night ([F.sub.1,30] = 15.54, P < 0.01). In the 50% KDI core areas, use of vegetation types did not differ from random locations ([F.sub.6,5] = 1.09, P = 0.47).

[FIGURE 2 OMITTED]

The topographic composition of the study area was 10.2% valley, 14.7% low-slope, 47.6% mid-slope, 20.1% steep-slope, and 7.4% ridge. On average, the topographic composition of fox home ranges was 9.3 [+ or -] 0.8% valley, 12.8-0.8% low-slope, 51.2-3.2% mid-slope, 20.2 2.2% steep-slope, and 6.4 [+ or -] 0.9% ridge. There was significant topographic selection within 95% KDI home ranges ([F.sub.6,5] = 4.72, P = 0.04; Fig. 3), with foxes being closer to valleys than expected ([t.sub.10] = -2.96, P = 0.01) and farther than expected from steep slopes (i10 = 2.78, P = 0.02). Topographic selection within the 95% KDI home ranges did not vary by time of day ([F.sub.15,72] = 0.68, P = 0.79). No topographic selection was detected within the 50% KDI core areas ([F.sub.6,5] = 1.22, P = 0.40).

[FIGURE 3 OMITTED]

DISCUSSION--We conducted the first telemetry-based home-range analysis for SRI foxes and only the second use of GPS collars on any island fox population. We documented the largest home ranges (averaging 3.39 [km.sup.2] for 95% MCPs and 3.82 [km.sup.2] for 95% KDIs) ever reported for any island fox population, ranging from 2-10 times larger than other studies (Table 3), although several of these studies may not have acquired enough locations to accurately estimate the home range, which may require 100 or more points for the MCP area to asymptote (Bekoff and Mech, 1984; White and Garrott, 1990). Roemer et al. (2001) conducted the most extensive homes-range analysis of island foxes published to date, tracking a total of 25 foxes from November 1993 through December 1994 on Santa Cruz Island. They divided this time period into three biological seasons and found that the seasonal 95% MCP home ranges were 0.44 [km.sup.2] (n = 14 foxes), 0.46 [km.sup.2] (n = 15 foxes), and 0.61 [km.sup.2] (n = 15 foxes), with the average annual home range being 0.55 [km.sup.2] (n = 14 foxes). Roemer et al. (2001) found that their homerange estimates reached an asymptote with 86 locations, yet their average home-range size was only 13% of ours. Studies using GPS telemetry usually acquire far more locations per animal than does traditional VHF telemetry, making it far easier to acquire the necessary number of locations for robust estimations of home-range size (Girard et al., 2002; Tomkiewicz et al., 2010). Yet in the only previous deployment of GPS telemetry collars on island foxes, with an average of 118 locations per fox, home ranges on Santa Catalina Island were less than one third of what we observed on SRI during approximately equivalent seasons (Table 3; Cypher et al., 2014a; King et al., 2014).

Biological factors that affect home-range size within a species include food availability, body mass, and population density (Benson et al., 2006; Schradin et al., 2010). Although we cannot eliminate the possibility that other factors may be involved, we believe that the hypothesis of a density-dependent response in home-range size is the most consistent with theoretical predictions and the available evidence. The habitats and fox body size on Santa Rosa are not sufficiently different from Santa Cruz or Santa Catalina to explain a 2-10 fold increase in home-range size (Moore and Collins, 1995; Schoenherr et al., 1999; Coonan et al., 2010). No estimate of food availability was available during our study, but because island foxes were the top predator on SRI prior to the arrival of golden eagles, their population crash would probably cause food such as mice, invertebrates, and fruit (Moore and Collins, 1995; Cypher et al., 2014b) to become more abundant, as was observed on San Miguel Island (Coonan et al., 2005). However, this increase in food availability would be expected to cause smaller home ranges (Benson et al., 2006), not the dramatically larger ones that we observed.

Other evidence suggests a negative correlation between island fox home-range size and population density. As the Santa Cruz population declined due to golden eagle predation, foxes expanded their home ranges by 16-266% when their neighbors were killed (Roemer, 1999; Roemer et al., 2001). On San Clemente, habitats with the highest population density had the smallest average home-range size and vice versa (Sanchez, 2012). A similar relationship between home-range size and population density has been documented for red fox (Vulpes vulpes; Trewhella et al., 1988), bobcat (Lynx rufus; Benson et al., 2006), and several other mammal species including rodents, deer, and small marsupials (e.g., Erlinge et al., 1990; Nelson, 1995; Bond and Wolff, 1999; Kjellander et al., 2004; Sale and Arnould, 2009; Schradin et al., 2010).

In contrast with home-range size, overlap between adjacent home ranges does not appear to vary strongly with population density. Our values are comparable to the 11% overlap among nonmated individuals (both males and females) reported from Santa Cruz as that population began to decline due to golden eagle predation (Roemer et al., 2001). An earlier study on Santa Cruz reported 30% overlap among 95% KDI home ranges when the population density was much higher (Crooks and Van Vuren, 1996). On San Clemente, home-range overlap was not significantly associated with population density (Sanchez, 2012). Island species typically have greater home-range overlap than do mainland species due to an increased tolerance for conspecifics that results from higher population densities where dispersal is constrained (Stamps and Buechner, 1985). However, our study is consistent with others in finding very little overlap among core areas within the home ranges (Crooks and Van Vuren, 1996; Roemer et al., 2001). By definition, core areas have higher utilization intensity than do other areas of the home range, presumably due to higher quality or quantity of resources in these areas (Powell, 2000). The fact that neighboring foxes have little overlap in these high-use areas suggests that core areas are actively defended from conspecifics, and this is consistent with the view that island foxes are territorial (Roemer, et al. 2001).

Previous studies have concluded that island foxes are habitat generalists because they inhabit all available vegetation types (Laughrin, 1977; Moore and Collins, 1995). Standardized trap arrays have documented higher fox densities in habitats such as sand dune, woodland, and scrub as compared to grasslands (Roemer, 1999; Sanchez, 2012). Of course, higher population densities in these habitats do not mean that the foxes prefer these habitats, nor that there are more resources in these habitats (Van Horne, 1983), especially given the comparatively high population densities typical for island foxes (Coonan et al., 2010). At low population densities, however, animals should be able to express their resource preferences due to lower intraspecific competition (Krohn, 1992). Despite the low population density during our study, we found no evidence of habitat selection until we partitioned the diel period into biologically relevant periods. At night, foxes selected for bare ground and grasslands, which likely represents foraging for prey such as deer mice (Peromyscus maniculatus) and insects (Cypher et al., 2014a). The SRI foxes also selected for valleys and avoided upper slopes within their home ranges. These patterns were evident throughout the diel period and so likely incorporate habitats used for day rests and den sites. SRI foxes have been observed resting under scrub cover in canyons, and their dens have been found in valleys near perennial streams and small pools. Valleys may also provide shelter from the high winds that are common on the outer Channel Islands.

Unlike the estimation of home-range size, which depends on the number of locations per animal, analyses of habitat selection are more dependent upon the number of collared animals (Garton et al., 2001; Girard et al., 2006). In this regard our sample size of 11 recovered data sets was small, although it is comparable to previous home-range studies on island fox (Table 3). Small sample size is a common problem for studies using GPS telemetry collars, primarily due to their high cost compared to traditional VHF collars (Girard et al., 2006; Hebblewhite and Haydon, 2010). Our small sample size was compounded by an unexpected level of variation among individuals (Table 2). Beyond its obviously detrimental effect upon our power to detect population-level trends, this variation is interesting because it indicates that even with very low intraspecific competition for resources, individual foxes still had large differences in their resource utilization as indicated by the habitat composition of their home ranges.

Unlike previous studies, we chose to collar only male foxes for several reasons. Our study was a trial deployment of a limited number of GPS collars, which are heavier than the VHF collars that have previously been used on island foxes (Cypher et al., 2014a). Our permit from the U.S. Fish and Wildlife Service specified that collars not exceed 4% of the collared animal's weight, which required collared animals to weigh [greater than or equal to] 1.65 kg. Both NPS and the Fish and Wildlife Service preferred we collar only males because males tend to be slightly larger than females (Moore and Collins, 1995), and therefore better able to carry the additional weight, as well as to avoid any possibility, however small, of compromising the success of reproductive females in this recovering population. By omitting females, we avoided the redundancy that would have resulted from collaring both members of a mated pair, which is the basic unit of the island fox social system (Roemer et al., 2001). The pair maintains a shared territory, so home-range sizes of mated foxes are similar, and the overlap between mates is [greater than or equal to] 70% as opposed to [less than or equal to] 30% between nonmated dyads (Crooks and Van Vuren, 1996; Roemer et al., 2001; Powers, 2009). Prior studies have found no significant difference between male and female home ranges overall (Crooks and Van Vuren, 1996; Roemer et al., 2001), although there may be seasonal differences related to extra-pair copulations (Ralls et al., 2013). Therefore, our results may also extend to adult females in this population, but this should be verified in the future.

Our study provides baseline data on habitat use and home-range characteristics for an island fox population at historically low population density following a near-extinction event. There appears to be a density-dependent relationship between fox population density and home-range size, but the pattern is confounded by differences among islands (e.g., habitat composition and shifting resource abundance) and studies (e.g., duration, seasonality, and methodology). Minimizing these confounding factors should clarify the pattern. Because the SRI population is recovering toward its pre-1990s levels of abundance and density, and the NPS does plan to continue monitoring this population, we recommend that this study be repeated when the population returns to its predecline levels of approximately 4 adult foxes/[km.sup.2]. Based on the pattern seen among island fox studies, we predict that the average home range at this density will be approximately 1 [km.sup.2], only 30% of what we observed in this study. GPS collars can acquire more locations at higher precision and more-evenly distributed throughout the diel period relative to VHF telemetry, so we recommend that GPS collars be used in future studies of island fox spatial ecology. We also recommend that the collars incorporate accelerometers to better indicate habitat-specific behaviors and activity (Brown et al., 2013). Additional insight into density-dependent responses, habitat use, and activity could inform habitat restoration and other conservation efforts for island foxes throughout their range.

Funding for this project was provided by the California Department of Fish and Game through the Endangered Species Act (Section-6) Grant-in-Aid Program, the National Park Service, and the Smithsonian Institution through a restricted endowment grant. J. Benson and three anonymous reviewers gave comments that significantly improved the manuscript. We thank R. Rudolph, D. Rodriguez, A. Guglielmino, and D. Sphar for assistance with geographic information system, botany, and field efforts. We thank B. Hudgens and the U.S. Navy for providing data from San Clemente Island and J. King for data from Catalina Island. J. Maldonado translated our abstract into Spanish. This study was conducted under U.S. Fish and Wildlife Service Recovery Permit TE08627-0.

LITERATURE CITED

BEKOFF, M., AND L. D. MECH. 1984. Computer simulation: simulation analyses of space use: home range estimates, variability, and sample size. Behavior Research Methods, Instruments and Computers 16:32-37.

BENSON, J. F., M. J. CHAMBERLAIN, AND B. D. LEOPOLD. 2006. Regulation of space use in a solitary felid: population density or prey availability? Animal Behavior 71:685-693.

BIGGINS, D. E., J. L. GODBEY, M. R. MATCHETT, AND T. M. LIVIERI. 2006. Habitat preferences and intraspecific competition in black-footed ferrets. Pages 129-140 in Recovery of the black-footed ferret: progress and continuing challenges (J. E. Roelle, B. J. Miller, J. L. Godbey, and D. E. Biggins, editors). U.S. Geological Survey, Reston, Virginia.

BOND, M. L., AND J. O. WOLLF. 1999. Does access to females or competition among males limit home-range size in a promiscuous rodent? Journal of Mammalogy 80:1243-1250.

BROWN, D. D., R. KAYS, M. WIKELSKI, R. WILSON, AND A. P. KIMLEY. 2013. Observing the unwatchable though acceleration logging of animal behavior. Animal Biotelemetry 1:1-16.

CONNER, L. M., AND B. W. PLOWMAN. 2001. Using Euclidean distance to assess nonrandom habitat use. Pages 275-290 in Radio telemetry and animal populations (J. Millspaugh and J. M. Marzluff, editors). Academic Press, San Diego, California.

CONNER, L. M., M. D. SMITH, AND L. W. BURGER. 2003. A comparison of distance-based and classification-based analyses of habitat use. Ecology 84:526-531.

COONAN, T. J. 2010. Summary report: Twelfth annual meeting of the Island Fox Working Group. Channel Islands National Park, Ventura, California.

COONAN, T. J., K. RUTZ, D. K. GARCELON, B. C. LATTA, M. M. GRAY, AND E. T. ASCHEHOUG. 2003. Progress in island fox recovery efforts on the northern Channel Islands. Proceedings of the California Island Symposium 6:263-273.

COONAN, T. J., C. A. SCHWEMM, AND D. K. GARCELON. 2010. Decline and recovery of the island fox: a case study for population recovery. Cambridge University Press, New York.

COONAN, T. J., C. A. SCHWEMM, G.W. ROEMER, D. K. GARCELON, AND L. MUNSON. 2005. Decline of an island fox subspecies to near extinction. Southwestern Naturalist 50:32-41.

CROOKS, K. R., AND D. VAN VUREN. 1995. Resource utilization by two insular endemic mammalian carnivores, the island fox and island spotted skunk. Oecologia 104:301-307.

CROOKS, K. R., AND D. VAN VUREN. 1996. Spatial organization of the island fox (Urocyon littoralis) on Santa Cruz Island, California. Journal of Mammalogy 77:801-806.

CYPHER, B. L., E. M. DRAKE, J. SAVAGE, J. L. KING, K. RALLS, T. J. COONAN, J. D. PERRINE, AND C. L. DUNCAN. 2014a. Evaluation of novel telemetry technologies for research on island foxes. Western North American Naturalist. 7:357-372.

CYPHER, B. L., A. Y. MADRID, C. L. VAN HORN JOB, E. C. KELLY, S.W. R. HARRISON, AND T. L. WESTALL. 2014b. Multi-population comparison of resource exploitation by island foxes: implications for conservation. Global Ecology and Conservation 2:255-266.

DRAKE, E. M. 2013. Home range and habitat use of Santa Rosa Island Foxes (Urocyon littoralis santarosae). M.S. thesis, California Polytechnic State University, San Luis Obispo, California.

ERLINGE, S., I. HOOGENBOOM, J. AGRELL, J. NELSON, AND M. SANDELL. 1990. Density-related home range size and overlap in adult field voles (Microtus agrestis) in southern Sweden. Journal of Mammalogy 71:597-603.

GARTON, E. O., M. J. WISDOM, F. A. LEBAN, AND B. K. JOHNSON. 2001. Experimental design for radiotelemetry studies. Pages 15-42 in Radio tracking and animal populations (J. J. Millspaugh and J. M. Marzluff, editors). Academic Press, New York.

GIRARD, I., C. DUSSAULT, J.-P. OUELLET, R. COURTOIS, AND A. CARON. 2006. Balancing number of locations with number of individuals in telemetry studies. Journal of Wildlife Management 70:1249-1256.

GIRARD, I., J.-P. OUELLET, R. COURTOIS, C. DUSSAULT, AND L. BRETON. 2002. Effects of sampling effort based on GPS telemetry on home-range size estimations. Journal of Wildlife Management 66:1290-1300.

GOBLE, D. D., J. A. WIENS, J. M. SCOTT, T. D. MALE, AND J. A. HALL. 2012. Conservation-reliant species. BioScience 62:869-873.

HARRIS, S. W., J. CRESSWELL, P. G. FORDE, W. J. TREWHELLA, T. WOOLARD, AND S. WRAY. 1990. Home range analysis using radio-tracking data: a review of problems and techniques particularly as applied to the study of mammals. Mammal Review 20:97-123.

HEBBLEWHITE, M., AND D. T. HAYDON. 2010. Distinguishing technology from biology: a critical review of the use of GPS telemetry data in ecology. Philosophical Transactions of the Royal Society B 365:2303-2312.

JONES, M. C., J. S. MARRON, AND S. J. SHEATHER. 1996. A brief survey of bandwidth selection for density estimation. Journal of the American Statistical Association 91:401-407.

KERTSON, B., AND J. MARZLUFF. 2009. Improving studies of resource selection by understanding resource use. Environmental Conservation 38:18-27.

KING, J. L., C. L. DUNCAN, AND D. K. GARCELON. 2014. Status of the Santa Catalina Island fox 13 years after its decline. Western North American Naturalist. 7:382-396.

KJELLANDER, P., A. J. HEWLSON, O. LIBERG, J. M. ANGIBAULT, E. BIDEAU, AND B. CARGNELUTTI. 2004. Experimental evidence for density-dependence of home range size in roe deer (Capreolus capreolus): a comparison of two long-term studies. Oecologia 139:478-485.

KROHN, W. B. 1992. In my experience: sequence of habitat occupancy and abandonment: potential standards for testing habitat models. Wildlife Society Bulletin 20:441-444. LAUGHRIN, L. L. 1977. The island fox: a field study of its behavior and ecology. Ph.D. dissertation, University of California, Santa Barbara, California.

MINTA, S. C. 1993. Use of spatio-temporal interaction statistics on badgers to test predictions of territorial polygyny theory. Oecologia 96:402-409.

MOORE, C. M., AND P. W. COLLINS. 1995. Urocyon littoralis. Mammalian Species 489:1-7.

MORRISON, M. L., B. C. MARCOT, AND R. W. MANNAN. 1998. Wildlifehabitat relationships. University of Wisconsin Press, Madison. NELSON, J. 1995. Intrasexual competition and spacing behavior in male field voles, Microtus agrestis, under constant female density and spatial-distribution. Oikos 73:9-14.

POWELL, R. A. 2000. Animal home ranges and territories and home range estimators. Pages 65-110 in Research techniques in animal ecology: controversies and consequences (L. Boitani and T. K. Fuller, editors). Columbia University Press, New York.

POWERS, R. M. 2009. Social behavior measured by home range overlap in island foxes (Urocyon littoralis dickey). M.S. thesis, San Francisco State University, San Francisco, California.

RALLS, K., J. N. SANCHEZ, J. SAVAGE, T. J. COONAN, B. R. HUDGENS, AND B. L. CYPHER. 2013. Social relationships and reproductive behavior of island foxes inferred from proximity logger data. Journal of Mammalogy 94:1185-1195.

ROEMER, G. W. 1999. The ecology and conservation of the island fox (Urocyon littoralis). Ph.D. dissertation, University of California, Los Angeles.

ROEMER, G. W., C. J. DONLAN, AND F. COURCHAMP. 2002. Golden eagles, feral pigs, and insular carnivores: how exotic species turn native predators into prey. Proceedings of the National Academy of Sciences 99:791-796.

ROEMER, G. W., D. A. SMITH, D. K. GARCELON, R. K. WAYNE. 2001. The behavioural ecology of the island fox (Urocyon littoralis). Journal of Zoology London 255:1-14.

SALE, M. G., AND J. P. Y. ARNOULD. 2009. Spatial and temporal organization in the Swamp Antechinus: comparison between island and mainland populations. Journal of Mammalogy 90:347-355.

SANCHEZ, J. N. 2012. Spatial ecology of disease spread in the island fox. M.S. thesis, Humboldt State University, Arcata, California.

SCHOENHERR, A. A., C. R. FELDMETH, AND M. J. EMERSON. 1999. Natural history of the islands of California. University of California Press, Berkeley.

SCHRADIN, C., G. SCHMOHL, H. G. RODEL, I. SCHOEPF, S.M. TREFFLER, J. BRENNER, M. BLEEKER, M. SCHUBERT, B. KONIG, AND N. PILLAY. 2010. Female home range size is regulated by resource distribution and intraspecific competition: a long-term field study. Animal Behaviour 79:195-203.

SCOTT, J. M., D. D. GOBLE, J. A.WIENS, D. S.WILCOVE, M. BEAN, AND T. MALE. 2005. Recovery of imperiled species under the Endangered Species Act: the need for a new approach. Frontiers in Ecology and the Environment 3:383-389.

SIKES, R. S., W. L. GANNON, and the Animal Care and Use Committee of the American Society of Mammalogists. 2011. Guidelines of the American Society of Mammalogists for the use of wild mammals in research. Journal of Mammalogy 92:235-253.

STAMPS, J. A., AND M. BUECHNER. 1985. The territorial defense hypothesis and the ecology of insular vertebrates. Quarterly Review of Biology 60:155-181.

TOMKIEWICZ, S. M., M. R. FULLER, J. G. KIE, AND K. K. BATES. 2010. Global positioning system and associated technologies in animal behavior and ecological research. Philosophical Transactions of the Royal Society B 365:2163-2176.

TREWHELLA, W. J., S. HARRIS, AND F. E. MCALLISTER. 1988. Dispersal distance, home range size and population density in the red fox (Vulpes vulpes): a quantitative analysis. Journal of Applied Ecology 25:423-434.

U.S. FISH AND WILDLIFE SERVICE. 2004. Endangered and threatened wildlife and plants: listing the San Miguel Island Fox, Santa Rosa Island Fox, Santa Cruz Island Fox, and Santa Catalina Island Fox as endangered. Federal Register 69(44):10335-10353.

VAN HORNE, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893-901.

WHITE, G. C., AND R. A. GARROTT. 1990. Analysis of wildlife radio-tracking data. Academic Press, San Diego, California.

Submitted 27 August 2014. Acceptance recommended by Associate

Editor, Stephen Glen Mech, 5 May 2015.

ELIZABETH M. DRAKE, BRIAN L. CYPHER, * KATHERINE RALLS, JOHN D. PERRINE, RUSSELL WHITE, AND TIMOTHYJ. COONAN

California Polytechnic State University, Biological Sciences Department, San Luis Obispo, CA 93401 (EMD, JDP)

California State University-Stanislaus, Endangered Species Recovery Program, Turlock, CA 95372 (BLC)

Smithsonian Conservation Biology Institute, National Zoological Park, Washington, D.C. 20008 (KR)

California Polytechnic State University, Information Services, San Luis Obispo, CA93401 (RW)

National Park Service, Channel Islands National Park, Ventura, CA 93001 (TJC)

* Correspondent: bcypher@esrp.csustan.edu
TABLE 1--Home-range sizes for male island foxes, Santa Rosa Island,
California, 2009-2010. "Fix rate" is the percentage of global
positioning system attempts that successfully generated a location;
MCP = minimum convex polygon; KDI = kernel density isopleth.

                                                Fix rate
Fox ID                Data dates                  (%)      Locations

M15       22 September 2009-12 February 2010    89.6          438
M33         22 September 2009-4 March 2010      82.6          450
M48        22 September 2009-7 December 2009    88.4          220
M52          1 December 2009-30 June 2010       72.0          524
M64        22 September 2009-18 January 2010    76.6          301
M66       22 September 2009-12 February 2010    92.8          440
M69       22 September 2009-16 February 2010    87.2          430
M71         22 September 2009-4 March 2010      85.6          481
M72       22 September 2009-14 December 2009    81.3          234
M73       22 September 2009-19 December 2009    72.8          222
M75        3 November 2009-12 February 2010     75.9          264
Average                   --                    82.3         364.0
SE                        --                     2.1          34.9

            95% MCP        95% KDI
Fox ID    ([km.sup.2])   ([km.sup.2])

M15           3.26           3.54
M33           2.61           2.63
M48           3.02           3.85
M52           7.31           5.65
M64           0.77           0.79
M66           1.31           1.51
M69           4.53           4.53
M71           3.94           4.43
M72           1.27           1.57
M73           3.56           4.78
M75           5.73           8.71
Average       3.39           3.82
SE            0.59           0.68

TABLE 2--Vegetation composition within the study area and the 95%
kernel density isopleth home ranges of male island foxes, Santa Rosa
Island, California, 2009-2010.

             Bare (%)   Chaparral (%)   Grassland (%)

Study Area    5.2        8.8            54.7
M15           6.5        0.0            72.5
M33           9.0        0.0            37.0
M48           6.6        0.0            11.6
M52           4.2        7.6            58.7
M64          15.8        0.0            51.7
M66           4.5        0.4            87.7
M69           6.6        5.0            46.2
M71           3.6       40.4            39.4
M72           3.6        0.0            67.7
M73           7.3        0.0            76.9
M75           1.9       21.0            57.3
Average       6.3        6.8            55.2
SE            1.1        3.9             6.4

             Lupine (%)   Scrub (%)   Woodland (%)

Study Area    5.3         24.1            1.9
M15           6.1         14.3            0.5
M33          31.9         21.9            0.2
M48          79.2          2.7            0.0
M52           0.0         26.6            3.0
M64           0.0         29.5            3.0
M66           0.0          7.4            0.0
M69           0.0         41.7            0.4
M71           0.0         10.5            6.0
M72           0.0         28.7            0.0
M73           0.0         15.6            0.2
M75           0.0         17.3            2.5
Average      10.7         19.7            1.4
SE           7.4           3.4            0.6

TABLE 3--Comparison of island fox home-range studies. Studies are
listed chronologically. The studies on Santa Rosa Island (this study)
and Santa Catalina Island, California, 2009-2010, used global
positioning system telemetry collars whereas the remainder used very
high frequency collars. No. Foxes = total number of acquired datasets,
with the number of males and females given in parentheses (not given
in Sanchez, 2012). Home ranges were calculated as 95% minimum convex
polygons (MCPs) for all studies except Sanchez (2012) and Crooks and
Van Vuren (1995), which used 100% MCPs. Roemer et al. (2001)
calculated home ranges for three seasons between November 1993 and
December 1994; the values shown below are for the season that
most-closely corresponded with the dates of our study.

                                              No. foxes
Island                  Study dates           (male, female)

Santa Rosa       September 2009-June 2010     11 (11, 0)
San Clemente     July 2010-February 2011      40 (?, ?)
Santa Catalina   November 2007-June 2008      17 (9, 8)

San Nicholas     December 2005-July 2006      14 (8, 6)
Santa Cruz       November 1993-March 1994     14 (8, 6)
Santa Cruz       January 1992-November 1992   12 (6, 6)

                 Locations          Home
Island            per fox    range ([km.sup.2])

Santa Rosa          364             3.39
San Clemente         20             0.42
Santa Catalina      118             1.05

San Nicholas         47             1.81
Santa Cruz          125             0.44
Santa Cruz           27             0.34

Island           Source

Santa Rosa       This study
San Clemente     Sanchez, 2012
Santa Catalina   King et al., 2014;
                 Cypher et al., 2014a
San Nicholas     Powers, 2009
Santa Cruz       Roemer et al., 2001
Santa Cruz       Crooks and Van Vuren, 1995
COPYRIGHT 2015 Southwestern Association of Naturalists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2015 Gale, Cengage Learning. All rights reserved.

 
Article Details
Printer friendly Cite/link Email Feedback
Author:Drake, Elizabeth M.; Cypher, Brian L.; Ralls, Katherine; Perrine, John D.; White, Russell; Coonan, T
Publication:Southwestern Naturalist
Article Type:Report
Date:Jun 1, 2015
Words:7377
Previous Article:Landscape-level assessment of Abert's squirrel and red squirrel in mixed conifer forest.
Next Article:Microsatellite variation in the Owens Valley vole (Microtus californicus vallicola).
Topics:

Terms of use | Privacy policy | Copyright © 2018 Farlex, Inc. | Feedback | For webmasters