Factors Influencing the Postrelease Movements of Translocated Fishers: Implications for Translocation Success

Long-distance, postrelease dispersal of released animals has been identified as a primary factor limiting successful population reestablishment following species translocations (Armstrong and Seddon 2008; Moehrenschlager and Lloyd 2016; Swaisgood and Ruiz-Miranda 2018; Berger-tal et al. 2020). The uncertainty that dispersal imposes on translocation success is based on the unpredictable behavior of stressed individuals being released into a foreign environment and forced to move about safely to find food and cover, potential mates, and a suitable home range (Armstrong and Seddon 2008; Parker et al. 2012). This exploration can increase an individual's risk to predation and other sources of mortality, and the level of risk is likely to increase with postrelease movements of greater distance or duration. Individuals that disperse beyond selected release or recovery areas will likely encounter landscapes with less high-quality habitat, fewer potential mates, or a greater number of predators and hazards (Miller et al. 1999; Stamps and Swaisgood 2007; Spinola et al. 2008; Le Gouar et al. 2012; Harrington et al. 2013). Consequently, long-distance dispersers may be less likely to contribute to translocation success (Van Heezik et al. 2009; Ryckman et al. 2010; Swaisgood and Ruiz-Miranda 2018; Berger-Tal et al. 2020).

In the contiguous United States (U.S.), translocations have played a key role in restoring populations of the fisher Pekania pennanti, a midsized carnivore of the family Mustelidae (weasels and allies). Historically, fishers occurred throughout the northern temperate and boreal forests of North America (Powell 1993; Lewis et al. 2012). Extremely high prices paid for fisher pelts in the late 1800s and early 1900s (up to ∼$350 U.S./pelt; Seton 1926; Bailey 1936; Grinnell et al. 1937; Dalquest 1948) caused the fisher to be overexploited and extirpated from much of the southern portion of its range in the northern United States and southern Canada (Strickland et al. 1982; Powell 1993), including all of Washington State, where it is listed as an endangered species (Hayes and Lewis 2006; Lewis et al. 2018). Since the mid-1900s, ≥ 35 fisher translocations have been reported within the fisher's historical range in in North America, and ≥ 25 of these (∼71%) resulted in the successful establishment of a population (Lewis et al. 2012).

Many factors may influence the movements of translocated individuals following release, including sex and age (Moehrenschlager and MacDonald 2003; Van Heezik et al. 2009; Ryckman et al. 2010; Nussear et al. 2012), the timing of releases (Proulx et al. 1994), number and location of release sites (Powell et al. 2012), the size and quality of the release or recovery areas (Linklater and Swaisgood 2008; Le Gouar et al. 2012; Osborne and Seddon 2012), and the presence of conspecifics in a release area (Sjoasen 1997; Van Heezik et al. 2009; Mihoub et al. 2011). One or more of these factors could be correlated with reduced movements (in distance or duration), greater fidelity to a reintroduction area, or more rapid home range establishment. Managers could then use this information to shape translocation processes and founder populations to minimize postrelease movements and presumably enhance the odds of successfully restoring populations (Linklater and Swaisgood 2008; Canessa et al. 2014; Batson et al. 2015; Moehrenschlager and Lloyd 2016).

Despite the abundance of fisher translocation programs (Lewis et al. 2012), only one published study reported information about the postrelease movements of radiocollared individuals (Proulx et al. 1994). Information about postrelease movements from previous fisher translocations is largely anecdotal and indicates that these movements varied greatly in distance, duration, and outcome (Weckwerth and Wright 1968; Roy 1991; Heinemeyer 1993; Proulx et al. 1994; Weir 1995; Fontana et al. 1999; Serfass et al. 2001). These previous studies were instrumental in helping us frame a priori models of factors potentially affecting postrelease movements of fishers in Washington.

To reestablish a self-sustaining population in Washington, partners from State, Provincial, Federal, and private organizations translocated 90 fishers from central British Columbia to Olympic National Park from 2008 to 2010 (Lewis et al. 2018). We evaluated factors that could influence postrelease movements and translocation success to inform future translocation efforts in the State. Our specific objectives were to 1) describe the distance and duration of fisher movements from release sites to established home ranges; 2) evaluate the effects of sex, age, release date, and release-year cohort on the timing of home range establishment, the distance fishers move from a release site to an established home range, and the duration of these movements; 3) estimate home range size and locations; and 4) characterize exploratory forays made by fishers outside established home ranges. Data used for our analyses are available on U.S. Geological Survey ScienceBase catalog (Happe et al. 2022) and in Supplemental Materials (Reference Data S1, S2).

The study area (14,412 km² [5,564 mi²], centered at 47.76°N, −123.53°E) encompassed most of the Olympic Peninsula in western Washington, which is bordered by the Pacific Ocean to the west, the Strait of Juan de Fuca to the north, and Puget Sound to the east (Figure 1). The center of the Peninsula is dominated by the Olympic Mountains, which rise from sea level to 2,415 m atop Mount Olympus. The central mountainous core of the Peninsula lies mostly within Olympic National Park and Olympic National Forest, where management favors wilderness preservation and retention of late-seral forest characteristics. Lower elevations are mosaics of Federal, State, Tribal, and private lands that are managed with greater emphasis on commercial timber production (apart from low-elevation parts of Olympic National Park; Figure 1). We identified the combined Olympic National Park and Olympic National Forest (6,275 km²) as the Olympic Fisher Recovery Area (OFRA) because, on account of its abundance of mid- and late-seral forest cover, we considered it to be the optimal area in Washington State for supporting a self-sustaining population of fishers (Lewis and Hayes 2004; Hayes and Lewis 2006).

The Olympic Peninsula has a temperate, maritime climate characterized by warm summers and cool winters (Peel et al. 2007). Annual precipitation ranges from 315 to 500 cm on the western slope of the Olympic Mountains; whereas annual precipitation is typically < 40 cm in the rain-shadow of the Olympic Mountains (the northeastern corner of the Olympic Peninsula; Gavin and Brubaker 2015). Eighty percent of annual precipitation on the Olympic Peninsula falls from October through March. Most winter precipitation falls as rain at elevations < 305 m (1,000 ft) and as snow above 760 m (2,500 ft).

The moist climate and broad range of elevations support a diversity of conifer-forest vegetation types across the Peninsula (Franklin and Dyrness 1988; Houston and Schreiner 1994; National Park Service 2005). Low- to mid-elevation forests are dominated by Sitka spruce Picea sitchensis (primarily in the western valleys and Pacific coastal area), whereas Douglas-fir Pseudotsuga menziesii, western hemlock Tsuga heterophylla, western redcedar Thuja plicata, Pacific silver fir Abies amabilis zone, mountain hemlock Tsuga mertensiana, and subalpine fir Abies lasiocarpa assume greater dominance at mid- and higher elevations (Franklin and Dyrness 1988; Houston and Schreiner 1994). Forests in the Sitka spruce zone on the western Olympic Peninsula are often referred to as temperate rainforests, and are characterized by large-diameter trees, epiphytic plants, and dense understory vegetation (Franklin and Dyrness 1988). Hardwoods (red alder Alnus rubra, bigleaf maple Acer macrophyllum, and black cottonwood Populus trichocarpa) are common in riparian forests along major rivers (Houston and Schreiner 1994; National Park Service 2005).

In addition to translocated fishers, the Olympic Peninsula study area was occupied by a diverse community of mammalian carnivores. This community included black bears Ursus americanus, coyotes Canis latrans, mountain lions Puma concolor, bobcats Lynx rufus, river otters Lontra canadensis, Pacific martens Martes caurina, western spotted skunks Spilogale gracilis, long-tailed weasels Mustela frenata, and ermines Mustela erminea (Happe et al. 2020). Bobcats, mountain lions, and coyotes are known predators of fishers (Powell 1993; Lewis 2014; Wengert et al. 2014).

From January 2008 to February 2010, project partners captured 90 fishers (50 F, 40 M) in central British Columbia and temporarily held them in captivity for a health assessment and approval by a licensed veterinarian. We transported approved fishers to the Olympic Peninsula generally within 10–30 d following capture (with exception of five individuals that were held longer for additional veterinary care; Lewis 2014; Table 1). To facilitate postrelease monitoring, we equipped 82 (50 F, 32 M) fishers with a 40-g VHF radiocollar (model MI-2 with mortality sensor, Holohil Systems Ltd, Carp, Ontario, Canada; http://www.holohil.com). Of the remaining eight translocated fishers, we equipped five males (each weighing > 4.5 kg) with a 120-g Argos satellite collar (Kiwisat 202; Sirtrack Ltd., Havelock North, New Zealand; http://www.sirtrack.com) and equipped three males with a 41-g VHF transmitter (model IMP/310/L with mortality sensor; Telonics Inc., Mesa, AZ; http://www.telonics.com) surgically implanted in the abdominal cavity.

We transported fishers by truck from the captive facility in central British Columbia to Port Angeles, Washington, in 1 d. Upon arrival in Washington, we housed fishers overnight in their transport boxes and released them the following morning at predetermined release sites in the OFRA. From January 2008 to February 2010, we released 90 fishers at nine general areas within the Park to facilitate fisher occupancy of large landscapes dominated by late-seral conifer forest (Figure 1; Lewis 2014). Animal handling procedures met or exceeded guidelines of the American Society of Mammalogists for the use of wild mammals for research (Sikes and the Animal Care and Use Committee of the American Society of Mammalogists 2016) and were approved by the British Columbia Ministry of Environment (Permit WL07-40389).

We monitored postrelease movements by radiotracking fishers from a fixed-wing aircraft, although we also occasionally located them from the ground when close to roads. We attempted to locate each fisher at least once per week; however, inclement weather prevented flights during some weeks, and the large size and rugged terrain of the study area, the short transmission distance of the radio collars, and the extensive movements of many fishers, made it impossible to locate each fisher every week. During each flight, we surveyed a portion of the study area by flying parallel flight-lines (approximately 5 km apart) over forested habitats. We located fishers by scanning the frequencies of radio collars until a signal was detected. After a transmitter signal was detected, the pilot isolated a fisher to a specific location, which we recorded as coordinates of the Universal Transverse Mercator grid system using a global positioning system. We gave each location an accuracy rating based on the flight biologist's judgment of signal strength, altitude of the aircraft, and time spent locating the fisher. We gave an accuracy rating of 1.5 to locations with an estimated error < 500 m, a rating of 2 for locations with an estimated error < 1 km, a rating of 3 for locations with an estimated error < 5 km, and a rating of 4 for locations with an estimated error > 5 km. We gave a rating of 1 to ground locations where the location was known (e.g., carcass found, visual observation, fisher occupying a known den), or considered to have an error ≤ 100 m.

We evaluated the accuracy of aerial locations by placing test collars at 31 known locations within the study area that were representative of the range of topographic and vegetation conditions occupied by fishers. The locations of test collars were unknown to our pilots, so we obtained unbiased estimates of location errors by having pilots estimate test collar locations with the same intensity of search effort they used when locating a fisher. Test collar locations that were given an accuracy rating of 1.5 had a mean location error of 281 ± 47 m (SE), and locations given an accuracy rating of 2 had a mean location error of 420 ± 117 m. Thus, the accuracy of test-collar locations was generally greater than that indicated by the assigned accuracy rating.

For analyses of postrelease movements and home range estimation, we used only locations with accuracy ratings ≤ 2. To avoid temporal autocorrelation, we used locations collected ≥ 24 h apart. We used no more than 1 telemetry location/wk derived from ground radiotracking procedures to minimize any bias associated with a greater number of ground locations for fishers that occupied areas with greater vehicle access. We made an exception to this rule for male M011 because this male inhabited an area that was entirely accessible by vehicle and its movements were consistently monitored from the ground at routine intervals by the staff of the Makah Tribal Forestry Department.

After release, fishers generally explore their new environment prior to establishing a home range (Heinemeyer 1993; Proulx et al. 1994; Weir 1995). The nature of this exploration may vary considerably among individuals, but once a translocated fisher finds a suitable location, its movements become more localized and shorter in distance as it establishes a home range (Weir 1995). To identify a change in movement distances associated with home range establishment, we calculated the mean-squared distance (MSD) between consecutive locations for each fisher, following the methods of Weir (1995) and Weir et al. (2009). We calculated MSD by squaring the six individual distances between the first seven locations (e.g., distance from location 1 to 2, 2 to 3,…6 to 7) and then averaging those six distances. We repeated this calculation for all locations in each fisher's data set in increments of seven locations (i.e., distances between locations 1–7, 2–8, 3–9, etc.) until an MSD value was calculated for all groupings of six consecutive movement distances. To estimate the date when home range establishment began, we plotted MSD (y-axis) against date (x-axis) and examined these data to identify a date when MSD decreased noticeably and remained at a lower level; we used a minimum of 10 locations collected for ≥ 2 consecutive months to indicate home range establishment.

We estimated the distance from each fisher's release site to its newly established home range by measuring the distance from the release site to the geographic center of the first 10 home range locations (dependent variable: Distance; Data S1, Supplemental Material). We characterized the duration of a postrelease movement as the number of days between each fisher's release date and the estimated date of home range establishment (i.e., date of the first home range location; dependent variable: Duration; Data S1, Supplemental Material). We established several a priori models of factors potentially influencing either distance or duration of postrelease movements. Based on previous studies of dispersal behavior of mammals (Greenwood 1980; Clobert et al. 2012) and the limited dispersal information for fishers (Leonard 1980; Arthur et al. 1993; York 1996; Aubry et al. 2004), we hypothesized that postrelease movements (either distance or duration) may differ between sexes and age-classes of fishers (independent variables: Sex and Age; including Sex × Age interaction). We classed all released fishers as juveniles (i.e., age < 1 y old at time of release) or adults (≥ 1 y old at time of release [yearlings were approaching their second birthday at time of release]). The initiation of the breeding season has biological importance for fisher movements and reproductive behavior (Douglas and Strickland 1987; Powell 1993; Facka et al. 2016); therefore, we also hypothesized that the number of days from a release date until the start of the breeding season (1 March; Powell 1993; independent variable: Release Date) may influence postrelease movements. Consequently, we evaluated both the number of days from a release date to home range establishment and the calendar dates of home range establishment. Lastly, the presence of conspecifics within a release area can influence the movements of subsequently released individuals; therefore, we evaluated the influence of cohort membership on the postrelease movements of individuals released in year 1, 2, or 3 (independent variable: Release-Year Cohort). We modeled potential linear effects of one- and two-variable models (as well as sex × age-class interaction) on Distance and Duration using PROC GENMOD in SAS (SAS Institute, Inc. 2009). We evaluated relative parsimony among models by ranking candidate models with Akaike's Information Criteria for small samples (AIC_c; Burnham and Anderson 2002). We collectively examined the maximized log-likelihood and ΔAIC_c values and significance levels of individual model parameters to rule out uninformative model parameters (Burnham and Anderson 2002:131; Arnold 2010).

We estimated home range sizes for fishers with ≥ 20 locations collected over ≥ 6 consecutive months following the estimated date of home range establishment (Data S2, Supplemental Material). To allow comparisons among fisher populations, we used two common methods to estimate the configuration and size of fisher home ranges (Lofroth et al. 2010). We used the fixed-kernel method with smoothing parameters selected by least-squares cross-validation (Worton 1989) to estimate 95% (home range) and 50% (core area) utilization distribution contours. We also estimated home ranges using the 100% minimum convex polygon method. We used Home Range Tools for Arc GIS (Rodgers et al. 2007) to delineate home ranges and Program Animal Space Use (Horne and Garton 2009) to calculate smoothing parameters.

When estimating home ranges, we identified exploratory forays outside established home ranges and removed telemetry locations associated with those forays. We defined exploratory forays as movements associated with telemetry locations > 2 interquartile ranges from the median x or y location coordinates of all locations obtained during the period of home range occupancy, where those locations existed as isolated kernels (containing ≤ 2 locations) separate from the primary home range distribution (i.e., 95% fixed kernel utilization distribution). These forays occurred outside home range boundaries, so we eliminated all locations associated with exploratory forays and re-estimated home ranges and core areas from the remaining locations. We also measured the distances from outlier points to the perimeter of the home range to evaluate the seasonal patterns and magnitude of exploratory forays among radio-collared fishers.

Among the 90 fishers released from 2008 to 2010, 48 (27 F, 21 M) met our criteria for home range establishment (Table 2). Movement distances to home ranges, the number of days to home range establishment, and the timing of home range establishment, varied substantially among individuals (Figures 2–4). Fishers established home ranges at distances ranging from 2.6 to 108.0 km from their release sites and the number of days from release to home range establishment ranged from 2 to 383 d after release (Figures 2 and 3; Table 3). The best-supported model of factors associated with distance to a home range included effects of sex, age, and sex × age interaction (Table 4). This top model garnered 62% of the total model weight. The beta coefficient for sex × age interaction was highly significant (β = −28,694, SE = 12,417, P = 0.02), indicating that postrelease dispersal distances of adult males were approximately 28.7 km longer on average than the those of females or juvenile males (Table 3). Although there was moderate model support for individual effects of sex and age (without interaction), age effects clearly influenced movements of males. As a result of these age- and sex-related differences in postrelease movements, a greater proportion of females (18 of 27; 67%) than males (8 of 21; 38%) established home ranges nearer to their release sites and within or partially within the Olympic Recovery Area (z = 1.97, P = 0.04; Figure 2).

We found weak support for a slight negative effect of release date on postrelease movements of fishers. Models containing release date garnered 10% of the total model weights. The β coefficient for release date in the univariate model was −244 ± 150 (SE, t = 1.65, P = 0.10), indicating a small but marginally significant negative effect on movement distances for fishers released closer to the onset of the breeding season (1 March). There was essentially no model support for the effect of release-year cohort on postrelease movement distances of translocated fishers.

We found negligible support for the effects of any model variables on the duration of postrelease movements of fishers prior to home range establishment (Table 5). The null model, indicating no covariate effects, was the highest-supported model, garnering 40% of the total model weight. Univariate models containing potential effects of release date, sex, and age were moderately supported based on AIC_c values near 2, but log-likelihood values of these models were essentially equal, indicating model fit was not improved by the addition of these variables (Burnham and Anderson 2002; Arnold 2010).

Despite no discernable difference in the mean duration of postrelease movements among sex and age classes of fishers, the variance in movement duration was greater for females than for males (F = 3.02, P = 0.007). Twenty-six of 27 females established home ranges over an 11-mo period (from December to October) in their first year, while 19 of 21 males established home ranges within a 4-mo period (April to July; Figure 4). Males had a clear peak of home range establishment in June, whereas the distribution of establishment dates of females was bimodal with lesser peaks in April and August (Figure 4).

We estimated home range sizes for 29 (19 F, 10 M) of the 90 fishers released from 2008 to 2010 (Table 6; Figure 5). Female home ranges were estimated with an average of 40.1 ± 3.6 (SE) telemetry locations (range: 20–70) obtained over an average of 508 ± 45 d (range: 197–912), whereas the home ranges of males had an average of 30.1 ± 3.3 locations (range: 23–57) obtained over an average of 476 ± 55 d (range: 226–838). Mean home range sizes were larger for males than for females based on both the 95% utilization distribution contours (128.3 km² vs. 63.5 km², respectively; t-test for unequal variances = −2.82, df = 12, P = 0.015) and the 100% minimum convex polygon method (79.6 km² vs. 50.8 km², respectively; t-test for unequal variances = −2.55, df = 26, P = 0.017; Table 6; Figure 5).

The disproportionately greater number of females than males that established home ranges within the OFRA was greater still when considering only the 29 fishers for which we had sufficient location data to estimate home range size (Figure 5): 15 of 19 females (79%) as compared with 3 of 10 males (30%; z = 2.77, P = 0.005). There was limited overlap of estimated home ranges of males and females as indicated by only 3 of 10 male home ranges overlapping by > 5% with home ranges of 5 of the 19 females. The extent of home range overlap we observed may have been underestimated because of the possible presence and home range establishment of fishers with failed transmitters. Female home ranges overlapped spatially in 3 locations in the study area; male home ranges did not overlap (Figure 5).

Both male and female fishers made exploratory forays beyond the boundaries of their home ranges, and several did this on > 1 occasion (Table S1, Supplemental Material). We obtained 10 locations for males > 10 km from their home ranges, and 14 such locations for females (Figure 6). Examination of these locations indicated eight forays made by six females (ages 2–6 y) and five forays made by four males (ages 2–4 y; Table S1, Supplemental Material). Fishers were located > 10 km beyond their home range boundaries only during the breeding season (1 March to 30 June; Figure 6). Male forays were concentrated during the breeding season, whereas females also made shorter forays (< 10 km) outside the breeding season (Figure 6). The duration of these forays ranged from 26 to 130 d for males and from 12 to 100 d for females (Table S1, Supplemental Material). In addition to breeding season forays, we observed three females shift core-use-areas within established home ranges, with two of these spatial shifts initiated during the breeding season. We also observed three fishers (2 F, 1 M) abandon established home ranges during the breeding season and establish new home ranges > 20 km away after the breeding season.

We released 90 fishers mostly along the periphery of a large (6,275 km²) recovery area (the OFRA) because it comprised the largest block of continuous fisher habitat and because of the limited road access to its central interior. The OFRA was dominated by unmanaged mid- and late-seral forests, which have been considered high-quality habitat for fishers (Lewis and Hayes 2004; Raley et al. 2012; Lewis et al. 2016). We estimated that this area would support > 50 fishers because of its size and habitat composition (Lewis and Hayes 2004); however, we could not predict how individuals would move once released in the OFRA or if enough would remain in the recovery area to initiate a population. The movement distances and durations we observed varied broadly for individual fishers, resulting in many fishers dispersing beyond the boundary of the recovery area. Despite these broad movements, our follow-up monitoring of released fishers on the Olympic Peninsula study area from 2013 to 2016 indicated persistence and a widespread distribution of the founding individuals and their progeny for ≥ 6 y following the last releases (Happe et al. 2020). Hence, the movement patterns we observed contributed to successful population establishment, at least over the short term (Happe et al. 2020).

Our strategy of releasing the 90 founders across nine release areas over 3 y was based on two main objectives: 1) reducing the incidence of injuries or mortalities that could result from antagonistic interactions when releasing many, likely stressed, individuals simultaneously at a single release site; and 2) establishing fishers broadly across the recovery area to facilitate widespread occupancy and population growth. Based on the number of exploratory forays made by females during the breeding season, however, we suspect that broadly distributing fishers may have hindered their ability to quickly find potential mates. We hypothesize that the prevalence of extraterritorial movements of females we observed during the breeding season may have been associated with a scarcity of breeding males. Moreover, the wide spacing of release sites may have reduced the frequency of intraspecific interactions among released individuals and extended the exploratory movement phase as individuals sought mates, other conspecifics, and optimal habitats. Exploratory movements outside of established home ranges had not been reported before for female fishers, but have been reported for other mustelids (e.g., wolverines Gulo gulo; Aronsson and Persson 2018).

Variability in movements between the sexes and age classes of fishers resulted in an initial spatial segregation in the establishment of home ranges of fishers throughout the study area. As documented previously (Lewis et al. 2016), most females remained in the recovery area, which we see as an indication of habitat suitability and a positive indication for the reestablishment of a reproductive population. The shorter postrelease movement distances of females (compared with adult males) may also indicate greater preference by females for extensive blocks of old-growth forest. In addition to moving longer distances, most males established home ranges in landscapes dominated by younger, managed forests outside the OFRA boundary (Lewis et al. 2016).

Variability in both postrelease movement distances and durations highlights the extent to which individual behaviors may also influence restoration outcomes. In addition to sex-specific differences in habitat affinities of fishers, postrelease movements of fishers may also reflect individual differences in innate characteristics such as boldness, responses to stress, and responses to being released in a foreign environment (Dickens et al. 2010; Moehrenschlager and Lloyd 2016). Stamps and Swaisgood (2007) hypothesized that individuals from a source population in one area of a species' range may not readily identify important habitat qualities in another part of its range. Poor recognition of habitat quality could prompt released-individuals to search extensively for familiar-looking habitats and place them at greater risk of mortality. Environments of the source population and the release areas varied considerably in latitude (52.63° vs. 47.76°; ∼540 km), climate (continental dry cold vs. coastal maritime), vegetation associations, and predator assemblages. The source population of fishers was in the Central Plateau physiographic region of British Columbia, where vegetation is dominated by subboreal spruce and pine forests, and interior Douglas-fir forests (Pojar and Meidinger 1991) that are actively managed for timber production (British Columbia Ministry of Forests, Lands and Mines 2010). In contrast, the OFRA is dominated by temperate coniferous rainforests and mesic Douglas-fir, western hemlock, and Pacific silver-fir forests (Franklin and Dyrness 1988). In the OFRA, these forests are largely protected as wilderness or managed to promote the development of late-seral coniferous forests under the Northwest Forest Plan (USDA Forest Service, and USDI Bureau of Land Management 1994). Moreover, the source area supports populations of four carnivores that do not occur in the OFRA: grizzly bears Ursus arctos, gray wolves Canis lupus, wolverines, and Canada lynx Lynx canadensis (Cowan and Guiget 1965). We can only speculate as to the effects that these differences in the source and release environments may have had on released individuals; however, the substantial differences in these areas could influence how fishers perceive the release environment and how they move following release.

Postrelease dispersal distances of translocated fishers in Washington were typically greater than juvenile dispersal distances reported in resident native populations. Rather, postrelease dispersal distances in the OFRA were comparable to the longest postnatal dispersal distances previously reported. Arthur et al. (1993) evaluated the dispersal of 13 juvenile fishers in a resident population in Maine and found that dispersal distances ranged from 4.5 to 19.5 km and did not differ between males and females. In a resident population in Massachusetts, York (1996) reported greater dispersal distances for males (10–60 km; = 25; n = 10) and females (12–107 km; = 37; n = 19), which did not differ statistically by sex. Dispersal distances > 50 km were also reported for individual fishers in Manitoba (Leonard 1980) and Oregon (Aubry et al. 2004).

The mean duration of postrelease movements of fishers did not differ appreciably among sex- and age-classes, although the duration of movements of females was highly variable compared with males. Similarly, duration of movements was not demonstrably related to other process variables such as duration of captivity or release dates relative to the onset of breeding. The apparent lack of pattern amidst high variability in movement durations suggests the potential for other unmeasured factors (e.g., prey abundance or availability, reproductive status, habitat selection behavior) to influence movements and home range establishment, particularly for females. We did not determine the reproductive status of all females that established home ranges, but the effects of reproductive status on postrelease movements may have been an important unmeasured variable and warrants additional study. The narrow range of home range establishments by males near the end of breeding (in June) suggests that the timing of the breeding season had a more uniform influence on males than females.

The initial home ranges of fishers translocated from central British Columbia to the OFRA (this study) were among the largest reported for the species in western North America (Table 7). Habitat selection by translocated individuals is likely to be influenced by the natal environment (Stamps and Swaisgood 2007); therefore, translocated individuals may be inefficient at exploiting foreign environments and this inefficiency may require translocated individuals to use relatively large home ranges to meet their energetic needs (Weilenmann et al. 2010; Russell et al. 2010). However, this inefficiency (and a corresponding home range size) may diminish as individuals gain familiarity with a new environment (Moehrenschlager and Macdonald 2003; Nussear et al. 2012) and would be less likely to affect offspring born in the recovery area. Additional research will help us understand how natal environments and postrelease movements of founder individuals relate to home range establishment and habitat selection by subsequent generations of fishers inside and outside the OFRA. It will also enable us to further evaluate the suitability of the landscapes within and outside the OFRA for supporting a restored fisher population on the Olympic Peninsula (Happe et al. 2020).

Poor survival and long-distance dispersal of released individuals can have substantial effects on translocation success, so Armstrong and Seddon (2008) proposed several research questions to provide information for increasing the likelihood of success. These questions included “How are postrelease survival and dispersal affected by pre- and postrelease management?” Fishers may be less likely to leave a recovery area if they are released within the central interior of large landscapes (e.g., ≥1,000 km²) that are dominated by optimal habitats. We also see benefit in releasing fishers at a small number of release sites (2–3) that are not widely dispersed (e.g., within 1–2 male-home-range diameters from each other) to facilitate conspecific interactions (i.e., mate acquisition, responding to positive cues [e.g., scent, scat] that indicate habitat quality). This approach may be especially useful when few individuals have been released (e.g., the first year of a project; translocations with small founder populations; Le Gouar et al. 2012). Consolidating release sites could also reduce the number of risky breeding-season forays made by females.

The movements made by released females (and males) varied considerably by individual, release area, or the release year. Based on these findings, we have a better understanding of the proportion of the female population that could disperse beyond the bounds of a release area. An estimate of the loss of founder individuals to dispersal would be valuable information when planning the acquisition of an optimal founder population.

These implementation considerations were used in planning, conducting, and adaptively managing an ongoing fisher translocation in the Cascade Fisher Recovery Area in Washington (2008 to present; Lewis 2013, Lewis et al. 2020), where 81 fishers were released at 3 centrally located sites within the southern portion of the recovery area. These approaches may also have useful applications for planning proposed translocations of fishers in Oregon (Hiller 2015) or Montana (Coltrane and Inman 2021), for translocations being considered for wolverines in Colorado and California (Young et al. 2012; Wolverine Translocation Techniques Working Group 2013; McKelvey et al. 2014), or for translocations of other wide-ranging carnivores. Given the importance of postrelease movements and survival of released individuals for translocation success, an expansion of research and monitoring efforts to track factors affecting the relationships between postrelease movements and survival could provide key support for fisher recovery programs and those for other endangered carnivore populations. We also recognize several other fruitful avenues for future studies of translocated carnivores pertaining to relationships between postrelease movements of fishers, innate personality characteristics of founders, conspecific attraction, and habitat selection between natal and recovery areas.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Data S1. Movement distance and duration data for 48 fishers Pekania pennanti that established home ranges following release in Olympic National Park, Washington, 2008–2010. Data include covariates evaluated as influences of postrelease movement distance and duration, which include the sex, age, release date, and release-year cohort membership of each fisher.

Available: https://doi.org/10.3996/JFWM-21-023.S1 (15 KB XLSX)

Data S2. Home range estimate data and associated attribute data for 29 fishers Pekania pennanti (19 F, 10 M) released in Olympic National Park, Washington, 2008–2010. Home range estimate data (i.e., area in km²) include fixed-kernel estimates for 95% (home range) and 50% (core area) contours, and 100% minimum convex polygon estimates of home ranges. Associated fisher attribute data include the following variables: sex, release-year cohort membership, age at home range establishment, duration and timing of home range occupancy.

Available: https://doi.org/10.3996/JFWM-21-023.S2 (44 KB XLS)

Table S1. Breeding season forays > 10 km from a home range boundary for 10 of 29 fishers Pekania pennanti with estimated home ranges on the Olympic Peninsula, Washington, 2008–2011. The breeding season for fishers is 1 March to 30 June.

Available: https://doi.org/10.3996/JFWM-21-023.S3 (26 KB DOCX)

Reference S1. Aubry KB, Raley CM. 2006. Ecological characteristics of fishers (Martes pennanti) in the southern Oregon Cascade Range, update: July 2006. Olympia, Washington: U.S. Forest Service, Pacific Northwest Research Station.

Available: https://doi.org/10.3996/JFWM-21-023.S4 (1.028 MB PDF)

Reference S2. Buck S, Mullis C, Mossman A. 1983. Final Report: Corral Bottom–Hayfork Bally fisher study. Arcata, California: Humboldt State University.

Available: https://doi.org/10.3996/JFWM-21-023.S5 (3.585 MB PDF)

Reference S3. Davis LR. 2009. Denning ecology and habitat use by fisher (Martes pennanti) in pine dominated ecosystems of the Chilcotin Plateau. Master's thesis. Burnaby, British Columbia: Simon Fraser University.

Available: https://doi.org/10.3996/JFWM-21-023.S6 (555 KB PDF)

Reference S4. Fontana AJ, Teske IE, Pritchard K, Evans M. 1999. East Kootenay fisher reintroduction program, 1996. Cranbrook: British Columbia Ministry of Environment, Lands, and Parks.

Available: https://doi.org/10.3996/JFWM-21-023.S7 (2.157 MB PDF)

Reference S5. Heinemeyer KS. 1993. Temporal dynamics in the movements, habitat use, activity, and spacing of reintroduced fishers in northwestern Montana. Master's thesis. Missoula: University of Montana.

Available: https://doi.org/10.3996/JFWM-21-023.S8 (7.486 MB PDF)

Reference S6. Hiller TL. 2015. Feasibility assessment for the reintroduction of fishers in western Oregon, USA. Portland, Oregon: U.S. Fish and Wildlife Service.

Available: https://doi.org/10.3996/JFWM-21-023.S9 (3.335 MB PDF)

Reference S7. Jones JL. 1991. Habitat use of fisher in northcentral Idaho. Master's thesis. Moscow: University of Idaho.

Available: https://doi.org/10.3996/JFWM-21-023.S10 (11.273 MB PDF)

Reference S8. Leonard RD. 1980. The winter activity and movements, winter diet and breeding biology of the fisher (Martes pennanti) in southeastern Manitoba. Master's thesis. Winnipeg: University of Manitoba.

Available: https://doi.org/10.3996/JFWM-21-023.S11 (7.509 MB PDF)

Reference S9. Lewis JC, Ransom JI, Chestnut T, Werntz DO, Black S, Postigo JL, Moehrenschlager A. 2020. Cascades fisher reintroduction project: progress report for April 2019 to June 2020. Natural Resource Report NPS/PWR/NRR—2020/2211. Fort Collins, Colorado: National Park Service.

Available: https://doi.org/10.3996/JFWM-21-023.S12 (5.033 MB PDF) and http://www.npshistory.com/publications/mora/nrr-2020-2211.pdf

Reference S10. Mazzoni AK. 2002. Habitat use by fishers (Martes pennanti) in the southern Sierra Nevada, California. Master's thesis. Fresno: California State University.

Available: https://doi.org/10.3996/JFWM-21-023.S13 (1.465 MB PDF)

Reference S11.National Park Service. 2005. Fire management plan and environmental assessment. Port Angeles, Washington: Olympic National Park.

Available: https://doi.org/10.3996/JFWM-21-023.S14 (1.404 MB PDF)

Reference S12. Roy KD. 1991. Ecology of reintroduced fishers in the Cabinet Mountains of northwest Montana. Master's thesis. Missoula: University of Montana.

Available: https://doi.org/10.3996/JFWM-21-023.S15 (4.261 MB PDF)

Reference S13. Serfass TL, Brooks RP, Tzilkowski WM. 2001. Fisher reintroduction in Pennsylvania: final report of the Pennsylvania fisher reintroduction project. Frostburg, Maryland: Frostburg State University.

Available: https://doi.org/10.3996/JFWM-21-023.S16 (7.380 MB PDF)

Reference S14.USDA Forest Service, and USDI Bureau of Land Management. 1994. Record of decision for amendments to Forest Service and Bureau of Land Management planning documents within the range of the northern spotted owl. Portland, Oregon: U.S. Department of Agriculture Forest Service and U.S. Department of the Interior Bureau of Land Management.

Available: https://doi.org/10.3996/JFWM-21-023.S17 (650 KB PDF)

Reference S15. Weir RD. 1995. Diet, spatial organization, and habitat relationships of fishers in south-central British Columbia. Master's thesis. Burnaby, British Columbia: Simon Fraser University.

Available: https://doi.org/10.3996/JFWM-21-023.S18 (8.041 MB PDF)

Reference S16. Yaeger JS. 2005. Habitat at fisher resting sites in the Klamath Province of northern California. Master's thesis. Arcata, California: Humboldt State University.

Available: https://doi.org/10.3996/JFWM-21-023.S19 (625 KB PDF)

Reference S17. York EC. 1996. Fisher population dynamics in north-central Massachusetts. Master's thesis. Amherst: University of Massachusetts.

Available: https://doi.org/10.3996/JFWM-21-023.S20 (4.055 MB PDF)

This study was funded by the U.S. Fish and Wildlife Service, U.S. Geological Survey, National Park Service, Washington Department of Fish and Wildlife, Doris Duke Foundation, and Washington's National Parks Fund. Our project partner D. Werntz from Conservation Northwest was instrumental in getting funding for the project, serving as the administrator of project operations in British Columbia and helping with fisher preparations, transportation, and releases. With the British Columbia Ministry of the Environment, we are grateful to H. Schwantje (project veterinarian), E. Lofroth, and R. Wright for their efforts to provide healthy fishers from British Columbia. M. and D. Evans worked tirelessly with trappers to obtain and transport fishers, and provided housing and excellent care of each captured fisher. We thank H. Allen with the Washington Department of Fish and Wildlife, who provided essential support, guidance, and logistical assistance in all aspects of this project. We also thank P. Becker, G. Olsen, W. Michaelis, A. McMillan, G. Shirato, E. Gardner, M. Cope, I. Keren, K. Mansfield, A. Duff, and S. Pearson with the Washington Department of Fish and Wildlife; and R. Hoffman, C. Copass, K. Beirne, S. Gremel, and C. Hoffman with Olympic National Park for their administrative, logistical, analytical, and field support. Biologists K. Maurice (U.S. Fish and Wildlife Service), D. Houston (U.S. Geological Survey, retired), B. Howell (Olympic National Forest), K. Aubry (U.S. Forest Service Pacific Northwest Research Station, retired), and L. Davis (Davis Environmental Ltd.) provided valuable assistance with fisher handling, preparation, transporting and release activities. Biologists R. McCoy, S. Murphie, and staff from the Makah Tribal Forestry Department provided valuable assistance in tracking male fisher M011. We thank S. West, J. Lawler, T. Bradshaw, and A. Wirsing of the University of Washington and K. Aubry for sharing their insights and guidance as we designed and implemented this study. We greatly appreciate the skilled and dedicated pilots that enabled us to fly and locate these 90 fishers, including J. Well and R. Mowbray (Rite Brothers Aviation, Port Angeles), C. Cousins (Olympic Air, Shelton), and M. Kimbrel (Washington Department of Fish and Wildlife, Olympia). We thank S. Nelson, J. Hagar, and three anonymous reviewers for providing helpful comments on earlier versions of this manuscript.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.