More info for the terms: cover, density, hardwood, presence, selection, shrubs, succession
Fishers are associated with areas of high cover and structural complexity (see Cover Requirements) in large tracts (see Landscape/scale effects) of mature and old-growth forests. Other site characteristics that can be important include presence of nearby water, slope, elevation, and snow characteristics.
Much of the information regarding fisher habitat is based on occurrence of fishers compared to availability. It is important to note that preference does not necessarily equate to superior habitat [30]. As a measure of preference, time spent in available habitat assumes that individuals have equal access to available habitats, which is unlikely (review by [35]). In addition, Powell [95] describes the case where foraging techniques for various prey have different efficiencies, resulting in relatively less time spent in the habitat of the prey associated with the more efficient foraging technique.
Fishers generally avoid early and/or prefer late successional stages, but in some cases they use fairly young forests extensively. For example, in predominantly grand fir (Abies grandis) and subalpine fir (Abies lasiocarpa) forests of north-central Idaho, stands subjectively categorized as pole-sapling age or younger were rarely used in summer or winter. In summer, mature forest and old-growth were preferred, but in winter young grand fir forests were preferred [65]. In Douglas-fir, lodgepole pine, and hybrid white spruce forests of south-central British Columbia, early seral stands (<10 years) were used by transient fishers significantly (P<0.05) less than expected, and young (41-80 years) forests were used more than expected based on availability in the landscape [128]. In northwestern California, fishers were associated with old-growth Douglas-fir forest [106]. However, use of varying succession classes by fishers in conifer stands of northwestern Montana was in accordance with availability [108].
In some locations, fishers prefer areas near water. In a relatively hot and dry California study area comprised primarily of Sierra mixed conifer, ponderosa pine (Pinus ponderosa), red fir (Abies magnifica), and montane hardwood habitats, there were significantly (P<0.05) more resting sites (51.9%) than random sites (27.8%) within 330 feet (100 m) of water [137]. In coniferous forests of north-central Idaho, several factors demonstrated the fisher's preference for riparian areas. Summer fisher locations were significantly (P<0.0001) closer (223 feet (68.0 m)) to water than random sites (400 feet (121.9 m)). In addition, fishers preferred grand fir/arrowleaf ragwort (Abies grandis/Senecio triangularis) riparian habitat in both summer and winter [65]. Similarly, fisher rest sites in northwestern Connecticut occurred in forest wetland habitat significantly (P<0.05) more often than expected (17%) in both winter (39%) and summer (29%) [68].
In some areas, fishers prefer comparatively steep slopes. In a mixed-wood forest of central Alberta, the average slope of fisher locations (5.8°) was significantly (P=0.002) greater than the average slope of random locations (1.9°) [17]. The average slope at resting sites (49.8%) in 2 California study areas, one predominantly comprised of Douglas-fir, white fir (Abies concolor), Oregon white oak (Quercus garryana), and tanoak (Lithocarpus densiflorus) stands and the other primarily Sierran mixed conifer, ponderosa pine, red fir, and montane hardwoods habitats, was significantly (P<0.05) steeper than random sites (42.6%) [137]. In coniferous forests of north-central Idaho, level slopes and benches were used (38%) much less than expected based on availability (71%) [65].
Fishers occur in a wide range of elevations, but generally prefer relatively low-elevation sites. In Maine, fishers were studied in an area that ranged from 0 to 1,210 feet (0-370 m) in elevation [10]. In the southern Sierra Nevada, fishers have been reported as high as 8,000 feet (2,438 m) [26]. In 75-year-old, even-age hardwood, mixed, and coniferous forests of White Mountain National Forest and surrounding areas of New Hampshire, use of varying elevations differed significantly (P<0.01) from expected. Based on availability fishers were expected to occur on sites from 1,100 to 2,000 feet (329-607 m) in elevation 264.7 times, and were expected on sites above 2,000 feet (≥608 m) 197.3 times. However, they preferred the low sites, occurring 363 and 99 times on low and high sites, respectively [66]. West of the Cascade Range crest in Washington, 87% of sighting and trapping records from 1894 to 1991 were from <3,300 feet (1,000 m), and none were from elevations >5,900 feet (1,800 m). However, east of the crest 70% of records were from sites >3,300 feet (1,000 m) and 18% were from sites 5,900 to 7,200 feet (1,800-2,200 m) in elevation. [13].
Fishers are likely affected by snow characteristics such as depth and consistency. Ninety-nine percent of the area in California where fishers were detected but American martens (Martes americana) were not was in the <5-inch (<13 cm) average snowfall zone; 1% was in the 5- to 9-inch (13-23 cm) average snowfall zone; and none was in the >9-inch (>23 cm) average snowfall zone [71]. In the low boreal region of southeastern Manitoba, significantly (P<0.005) fewer fisher tracks were observed during midwinter, when deep, soft snow conditions prevailed. Method of travel was also affected by snow conditions, with fishers walking in midwinter instead of bounding or galloping, and traveling on snowshoe hare (Lepus americanus) and their own trails more than they did during the thin snow cover of early winter and the crust conditions of late winter [102]. In addition, the stronger selection for low elevations west of the Cascade crest compared to the east (see previous paragraph) could have been due to the rain shadow, resulting in less snow on the east side compared to similar elevations on the west side [13]. However, in predominantly grand fir and subalpine fir cover types of north-central Idaho, fishers did not appear influenced by snow characteristics. Use of various elevations did not change over the seasons, and fishers preferred young forests in winter: a rather open habitat with large amounts of deciduous shrubs [65]. In addition, in an area of northwestern Connecticut with maximum snow depths less than 10 inches (<25 cm), there was no evidence that snow affected the fisher's habitat preferences [68].
Home range/density: Fisher home range size and density exhibit substantial variation, although male home ranges are larger than those of females. Small average female home range sizes (1,300 acres (527.5 ha), n=8) occurred in the southern Sierra Nevada [138]. Home ranges in the hardwoods and mixed-woods of southern Quebec were also comparatively small, with female (n=7) home ranges averaging 5.4 km² and male (n=3) home ranges averaging 9.2 km². Based on these estimates fishers occurred at a density of 2.7 fishers/10 km², while estimates from tracking and radio-collaring resulted in density estimates of 3.0 fishers/10 km² [53]. In conifer and mixed forests of south-central Maine, the median home range was 12.2 km² for females (n=5) and 25.5 km² for males (n=6). Density estimates in this area were 1 fisher/2.8-10.5 km² in the summer and 1 fisher/8.3-20.0 km² in winter [11]. Male home ranges in 2 California study areas, one predominantly comprised of Douglas-fir, white fir, Oregon white oak, and tanoak stands and the other primarily Sierran mixed conifer, ponderosa pine, red fir, and montane hardwoods habitats, were significantly (P<0.0001) larger (9,700 acres (3,934.5 ha), n=6) than female home ranges (2,400 acres (980.5 ha), n=15) [138]. In conifer forests of Idaho, fishers had very large home ranges. Average "year-long" estimates of home range size were 40.8 km² and 82.6 km² for females and males, respectively [65]. Powell and Zielinski [97] provide a thorough review of home ranges throughout the fisher's distribution, including notes on the methods used for calculation.
Landscape/scale effects: Fishers require large, well-connected habitat patches. In addition to their large home ranges, fishers can apparently be excluded from nearby habitat by relatively small expanses, possibly as short as 6 to 12 miles (10-20 km), of unsuitable environment [12,16]. In Douglas-fir forests of northwestern California, fishers were sensitive to habitat fragmentation. Fishers were detected in 70% of stands where <10% of the perimeter was clearcut edge and in less than 20% of stands where >75% of the stand's perimeter was clearcut edge. Fisher detections decreased markedly in stands <250 acres (100 ha) in area [106]. In mixed-wood forest of central Alberta, fishers used continuous (> 2 km²) forest blocks significantly (P<0.01) more than expected based on availability, and the woodlots (independent forested blocks ≤2 km²) used were significantly (P<0.01) larger (x=0.4 km²) than random woodlots (x=0.18 km²) [17].
Although edges between fisher habitats and open areas are apparently avoided [106], edges between various types of fisher habitat are often used. In 75-year-old, even-aged hardwood, mixed, and coniferous forests of the White Mountain National Forest, fishers occurred within 98 feet (30 m) of an edge (change in species composition, height or density class) significantly (P<0.01) more often than expected (observed=60, expected=42.5) in winter. The trend was the same in summer (observed=130, expected=114.6), but was not significant (P≥0.05) [66]. In a study in Douglas-fir, lodgepole pine, and hybrid white spruce forests of south-central British Columbia, Weir and Harestad [128] concluded that the fine-grained nature of the early and later seral stages provided sustainable fisher habitat. Likewise, fisher habitat use in central Maine suggested that areas with high interspersion of many forest types provided optimal habitat [10]. In addition, a review states that diverse forest communities would better conserve mustelids than a homogenous mature forest due to the variability of mustelid responses to conditions such as snow accumulations, microclimate conditions, prey availability, and predator densities [100].
Fishers select habitat on several scales. In predominantly upland hardwood habitat of upper peninsular Michigan, fishers used mixed pine (P. banksiana, P. resinosa, and P. strobus) habitat in accordance with availability at a fine scale (the area around fisher tracks), but used it more than expected based on its availability at a coarser scale (the general study area). Fishers selected (P<0.001) dense, lowland forests as rest sites at both these scales [95]. In forests dominated by Douglas-fir, lodgepole pine, and hybrid white spruce in south-central British Columbia, fishers displayed selectivity at the stand, patch, and element (resting, maternal den, and natal den habitat) scales [129]. In Douglas-fir forests of northwestern California, fishers were sensitive to fragmentation at the plot, stand, and 2,500-acre (1,000-ha) block scales [106].