Resting site characteristics of American marten in the northern lower Peninsula of Michigan.
American marten (Martes americana) are found across a large geographic range throughout North America (Williams et al., 2007) and are often associated with late successional forest types that offer closed canopy cover, large mature trees, complex vertical and horizontal structure, and abundant coarse woody debris (CWD) (Buskirk and Ruggiero, 1994; Bowman and Robitaille, 1997; Chapin et al., 1997; Potvin et al., 2000). However, due to compositional differences in forest types and forest management practices across North America, these types of forest attributes can often be limited, ultimately forcing marten to utilize or adapt to a variety of habitat types that may not encompass all of the forest characteristics that constitute prime marten habitat (Buskirk and Powell, 1994; Potvin et al., 2000). The availability of forests in the Northern Lower Peninsula (NLP) retaining these structural requirements is often limited because of the relatively young age of forest stands, forcing marten to utilize habitat considered suboptimal.
Marten were extirpated in the early 1900s and later re-introduced in both the Upper and Lower Peninsulas of Michigan (Williams et al., 2007). This extirpation occurred due to overharvest of the species and habitat loss from logging and wildfires, which drastically changed the landscape throughout the state and ultimately changed forest species composition and habitat types (Earle et al., 2001; Williams et al., 2007). The NLP population has been slow to expand due to habitat fragmentation and relatively small founder numbers (Earle et al., 2001).
In addition to fragmentation across the landscape, forests in Michigan's NLP, although diverse in species composition, consist of entirely different age class configurations and physical structure than pre-European settlement forests. Therefore, forest stands that marten used prior to European settlement (i.e., forest stand 200 y ago) are very different from current habitat available to marten in the NLP. However, remnant legacy trees, defined as trees which have "achieved near-maximum size and age" and are "significantly larger and older than the average trees on the landscape" (Mazurek and Zielinski, 2004) do exist within the forest matrix and may be a key component of marten habitat.
For resting sites marten utilize specific habitat structures in their environment, such as large branches and tree cavities. Resting site structures have been identified as being a key component of high quality marten habitat and are essential for their survival (Bull and Heater, 2000). They provide refuge from terrestrial and avian predators, as well as thermal protection from inclement weather (Zalewski, 1997; Martin and Barrett, 1991; Buskirk, 1984). Marten are not physiologically adapted to cold weather, which requires them to use their environment to reduce thermoregulatory costs (Spencer, 1987; Buskirk et al., 1988; Taylor and Buskirk, 1994). Certain types of resting sites (e.g., exposed branches, stump platforms) provide refugia from predators but offer little protection from the elements, while others (e.g., cavities, subnivean areas, hollow logs) provide both (Martin and Barrett, 1991; Bull and Heater, 2000). During winter months when temperatures can drop below a marten's lower critical temperature of 16 C (Buskirk and Harlow, 1988), marten rely on resting sites such as cavities, hollow logs, and subnivean sites to minimize heat loss. Their body temperature, together with the insulating qualities of snow and the material of the resting site, can raise the temperature within the resting site above the ambient air temperature (Buskirk, 1984).
Research has been conducted describing resting sites in other parts of North America (Buskirk, 1984; Spencer, 1987; Buskirk et al., 1989; Zalewski, 1997); however, no studies have specifically focused on resting sites in the NLP. Because of the unique nature of the habitat marten appear to be using in the NLP (i.e., large tracts of mature red pine stands (Pinus resinosa) intermixed with secondary growth forests, set in a largely fragmented landscape) and the nature of the population (small, re-introduced population with lower densities), it is possible that marten are utilizing different microhabitats in the NLP than elsewhere in North America. Therefore, it is important to identify what specific habitat characteristics are being used in the NLP in order for forest management agencies can make informed decisions to maintain a viable marten population. While denning, foraging, breeding, and dispersal are all important activities which may require their own microhabitat characteristics, this research focused on identification of resting site structures occupied by marten and sought to identify characteristics of those resting sites and their associated habitat.
We conducted our study in Michigan's NLP (44[degrees]42'N, 85[degrees]40'W) in the Manistee National Forest (MNF) (Fig. 1). The MNF is managed for timber production while also maintaining a multiple-use policy (USDA Forest Service, 1996). Historically, Michigan had approximately 37 million acres of forested land (Kapp, 1999). However, post-European settlement, timber harvest and large wildfires fed by logging slash decimated a large proportion of these forests, which ultimately altered wildlife habitat in Michigan. In 1930 the Civilian Conservation Corps (CCC) reforested a large portion of the degraded landscape with conifer species, in particular red pine. As a result, habitat in the MNF varies widely throughout the forest and the landscape is highly fragmented. Additionally, fragmentation is caused by differing forest management practices of the many private inholdings.
Currently, the study area supports a variety of upland forest types comprised of mixed hardwood, aspen (Populus spp.), and second-growth conifer stands. Additionally, large tracts of nearly monoculture red pine plantations occur throughout the MNF, although these stands can also contain remnant deciduous trees. Predominant deciduous species include red oak (Quercus rubra), white oak (Quercus alba), black oak (Quercus velutina), black cherry (Prunus serotina), red maple (Acer rubrum), sugar maple (Acer saccharum), aspen, American beech (Fagus grandifolia), American basswood (Tilia americana), white ash (Fraxinus americana), iron wood (Carpinus caroliniana), yellow birch (Betula alleghaniensis), and witch hazel (Hammamelis virginiana). Coniferous species include red pine, white pine (Pinus strobus), jack pine (Pinus banksiana), and eastern hemlock (Tsuga canadensis).
The climate in Michigan's NLP varies seasonally with mild summers and cold winters, which can persist up to 6 mo. Average annual temperatures range from 16 C in spring/ summer (April-September) to 1 C in fall/winter (October-March; National Oceanic and Atmospheric Administration [NOAA] 2002). The mean annual snowfall averages 225 cm, while average annual rainfall is 76 cm (NOAA 2002). Topography is relativity flat with elevations ranging from 174-520 m in the NLP.
We live-trapped marten in May and July during 2011; January, May, July, and December during 2012; and May, June, July and August during 2013 using live traps (models 103 & 105, Tomahawk Live Trap Company, Tomahawk, Wisconsin, U.S.A.). Traps were baited with smoked pork, beaver, chicken, or venison and an olfactory long distance call lure was applied within 3 m of the trap location ("Gusto," F &T Fur Harvest Trading Post, Alpena, Michigan). Traps were placed on the ground and covered with a half of a 55 gal barrel or natural debris, depending on the season. Marten would then be transferred into a modified open-ended restraining cone (Desmarchelier et al., 2007) and anesthetized with isoflurane delivered in oxygen via a facemask. Marten temperature, pulse, respiratory rates, and oxygen saturation levels were monitored throughout field immobilization. Each marten was implanted subcutaneously with a passive integrated transponder tag (AVID Identification Systems Inc., Norco, California) for permanent unique identification. Marten were fitted with collar-mounted, mortality-sensing VHF radio transmitters that weighed [less than or equal to] 20 grams (Modified model RI-2D Holohil Systems Ltd., Ontario, Canada or Advanced Telemetry Systems, Isanti, Minnesota, U.S.A. model Ml 555). After completion of handling and anesthesia, marten were placed in a recovery box and transported back to the trap site where they were monitored until the effects of the anesthesia wore off ([bar.x]= 7.5 [+ or -] 3.5 min). Marten were released when they had fully recovered from the effects of anesthesia. Animal capture and handling activities were conducted using protocols established by Little River Band of Ottawa Indians (LRBOI) and Grand Valley State University Institutional Animal Care and Use Committee (protocol #12-05A) and under the guidance of a licensed veterinarian (Maria Spriggs, DVM).
RADIO TELEMETRY/RESTING SITES
Marten were located to identify resting site structures using radio telemetry at least once per week between 0800 and 1700 h from May 2011 through December 2013. Once resting sites were located, attempts were made to visually observe the marten in the structure. If the marten could not be seen in the structure, where possible, its presence was confirmed based on one or more of the following criteria: telemetry signal, tracks leading up to the structure, scat and/or prey remains found at the base of the structure, chew marks around cavity openings, or vocalizations from within the structure. If the precise structure could not be identified, we would not record any data other than the location and the instance was not included in our rest site data. Structures were categorized as hollow log, branch, cavity, nest (i.e., squirrel, corvid, or raptor nests), or subnivean site. Living trees or dead snags with resting sites were identified to species and the diameters at breast height (DBH in cm) were recorded. If a marten was found in a cavity, branch, or nest, the structure's height was measured using a hypsometer (Forestry Pro laser rangefinder/hypsometer, Forestry Suppliers, Jackson, Mississippi, U.S.A.). Canopy closure at each site was estimated using a densiometer (spherical densiometer model C, Forestry Suppliers, Jackson, Mississippi, U.S.A.) and classified as [less than or equal to] 33%, 34-66%, or [greater than or equal to] 67%. At each site we described the habitat type (e.g., red pine stands, red pine mixed with oaks, hardwoods), which was based on tree composition located within a 15 m radius of the site. Resting sites were grouped by season, summer (April-September) and winter (October-March). Locations of all resting sites were determined using Global Position System (GPS), and resting sites were flagged and marked. To minimize disturbance to marten while in their resting sites, vegetation data were collected after marten had left the site.
We randomly selected five resting site locations from each radio collared marten for vegetation sampling, unless a marten had less than five locations in total (n = 3 marten), in which case all resting sites were sampled. For these sites (n = 147) a circular plot was measured around the resting site structure. We chose a radius of 15 m for these plots so that microhabitat features and stand-level characteristics could be identified (Porter et al., 2005). Tree species and snags were identified within each plot and the DBH of all trees and snags [greater than or equal to] 10 cm was measured. Unlike resting sites, vegetation plot snags were not identified down to species due to the inability to distinguish between species for smaller snags (the vast majority of cases). Therefore, we classified them simply as "snags" and included those as a category of trees alongside our species analysis, because we felt they represent an important component of the habitat composition associated with vegetation plots. The average basal area ([m.sup.2]/ha) was calculated for each site from the DBH measurements. We measured the length and diameter of all downed coarse woody debris [greater than or equal to] 1.0 m long and [greater than or equal to] 10 cm in diameter, and the length and width of all brush piles [greater than or equal to] 1.0 [m.sup.2]. One quadrant of the plot was selected randomly and all saplings [greater than or equal to] 1.0 m tall were counted and identified. For each resting site surveyed, we established an associated control plot where we measured the same attributes as the resting site plots (Porter et al., 2005). Control plots were sampled to look for fine-scale differences between habitat used for resting sites and habitat that was available within forest stands. Control plots were selected in a random cardinal direction (0-360[degrees]) 60 m from the resting site structure. Sixty meters was selected to allow 30 m between the perimeters of the rest site plot and its associated control plot. This permitted a localized comparison but eliminated any overlap between plots when measuring forest vegetation.
Mann-Whitney U median tests were conducted to compare mean basal area, average volume of CWD, average area of brush piles, and average number of saplings between resting site plots and control plots. We chose nonparametric tests because our data failed normality tests. Additionally, Mann-Whitney tests were conducted to detect differences between species occurrence and average DBH size between resting and control plots. Species diversity of trees and saplings counted in resting and control plots were calculated using Shannon-Weiner index,
[H'.sub.Density] = [p.sub.i]ln[p.sub.i]
where [p.sub.i] = the proportion of the total density of species i (Elliott and Hewitt, 1997). We used a chi-squared test to look at resting site selection differences between summer and winter seasons, between sexes, and among deciduous, coniferous and snag species. For our analysis of average, we will report the variance as standard deviation.
We outfitted 25 marten (15 males and 10 females) with radio collars from which 522 individual resting site structures were identified. On average marten were located 27 [+ or -] 21 times (range 1-72). Based on our sampling, the average number of resting sites per marten we detected was 20.9 [+ or -] 16.2 (range 1-60). We visually confirmed structures 71% of the time, and for the remaining structures we relied on other clues to confirm the precise structure being utilized (e.g., tracks, radio frequency strength). We identified five unique structure types being utilized as resting sites: cavities, branches, nests (e.g., squirrel dreys, corvid, or raptor constructions), hollow logs, and subnivean sites. The three most commonly observed structures were cavities (n = 255, 48.9%), branches (n = 162, 31.0%), and nests (n = 90, 17.2%, Table 1).
We found marten use of resting sites differed between summer and winter ([chi square] = 52.3, df= 2, P < 0.001). During the summer season (April-September) branches (n = 137, 41.8%) and cavities (n = 129, 39.3%) were selected most often, while cavities (n = 126, 64.9%) and nests (n = 38, 19.6%) were selected most often in the winter season (October-March, Table 1).
Overall, male marten chose more nests and branches and females chose more cavities than expected ([chi square] = 32.3, df=2, P < 0.001).
STRUCTURE TYPE BY TREE SPECIES
The type of arboreal resting site structures we observed varied by tree species. Cavity (n = 132, 53%) and branch (n = 61, 38%) resting sites were located most often in oak species (i.e., black, red, white oak), whereas nest structures were found more often in red pines (n = 50, 56%, Fig. 2). The average heights of rest sites located in cavities (n = 249), branches (n = 159), and nests (n = 90) were 4.3 [+ or -] 2.1, 8.6 [+ or -] 3.1, and 12.7 [+ or -] 2.5 m, respectively.
Marten in our study area used forested stands that had a high percent canopy closure. We found 77.9% (n = 395) of sites were located in forest stands that had [greater than or equal to] 67% canopy closures, 19.7% (n = 100) with closures of between 34-66%, and only 2.4% (n = 12) with canopy closures of [less than or equal to] 33%.
TREE AND SNAG SPECIES
Of the 522 unique resting sites observed, 429 were found in live trees (86%) and 69 in snags (14%). We found 11 different live tree species utilized by marten for resting sites. The three most frequently used tree types were oak (40.6%), maple (18.4%), and red pine (18.2%). We found resting sites located in oak species (n = 174) had an average DBH of 53.2 [+ or -] 15.4 cm, maple species (n = 79) 45.0 [+ or -] 13.2 cm, and red pine (n = 78) 29.3 [+ or -] 5.9 cm (Table 2). The average DBH of all resting site live trees (n = 429, [bar.x] = 42.9 [+ or -] 10.3 cm) was significandy larger (U' = 268721, P = < 0.001) than the average DBH (n = 7764, [bar.x] = 21.1 [+ or -] 9.0 cm) of nonresting site trees found within the 15 m established vegetation plot around resting site structures. Nine different species of snags were used as resting sites. Snags were typically not decomposed to the point they could not be identified to species (e.g., bark present, hard/soft wood, tree species present, Table 3). The three most commonly observed snags species being utilized by marten were oak (43.5%, [bar.x] = 45.1 [+ or -] 11.6 cm DBH, n = 30), aspen (21.7%, [bar.x] = 34.8 [+ or -] 7.7 cm DBH, n = 15), and maple (15.9%, [bar.x] = 32.9 [+ or -] 8.5 cm DBH, n = 11, Table 3). The average DBH of all resting sites snags (n = 69, [bar.x] = 34.5 [+ or -] 8.6 cm) was significantly larger (U' = 3425, P = <0.001) then the average DBH (n = 767, [bar.x] = 17.6 [+ or -] 7.8 cm) of nonresting site snags found within the 15 m established vegetation plot around used snag structures. Resting site structures, including snags, were found in deciduous trees 63.3% (n = 325), in coniferous trees 21.6% (n = 111), and in snags 15.0% (n = 77) of the time. From the randomly-selected subset of resting site plots (n = 147), we found marten were selecting deciduous trees (n = 98) as resting sites significantly more and coniferous trees (n = 25) significantly less = 110.3, df = 2, P < 0.001) than would be expected based on availability (defined as the total number of trees of each type pooled from the 147 used resting site plots), while no differences were observed when comparing snags.
RESTING SITE REUSE
We observed marten re-using resting sites on 159 separate occasions throughout the course of our study. Of the 522 unique resting sites, we observed marten reusing 72 (13.7%) of those sites on more than one occasion. We found 450 of the resting sites were used only once, 48 were used twice, 13 were used three times, five were used four times, and two sites were used five times. Additionally, four resting sites were used six, seven, nine and ten times, respectively. We also observed 10 marten re-using 20 individual resting sites that had been previously occupied by different marten. The number of times resting sites were re-used varied by structure type. Cavities had the highest percentage of resting site re-use among the five structure types identified (i.e., cavities, nests, branches, hollow logs, and subnivean, Fig. 3). Resting site structures that were classified as reused sites were not included more than once in the total resting sites observed (n = 522).
RESTING VS. CONTROL PLOTS
We identified 15 different overstory tree species from 294 vegetation plots (n = 147 resting site plots, n = 147 control plots). Tree species richness per plot ranged from three to nine species ([bar.x] = 4.95 [+ or -] 1.16) in resting site plots and 1 to eight species ([bar.x] = 4.71 [+ or -] 1.17) in control plots and no significant difference was observed (U' = 9838, P = 0.17). The three overstory tree species found most frequently in resting site plots were red pine (n = 3377, 41.7%), white pine (n = 1134, 14.0%), and snags (n = 767, 9.5%). The average number of trees [greater than or equal to] 10 cm found in resting site plots (n = 147) was 55.2 [+ or -] 23.3. When looking at the differences between conifer and deciduous tree species across used resting site plots we observed approximately twice the amount of coniferous trees ([bar.x] = 33.1 [+ or -] 26.2) compared to deciduous trees ([bar.x] = 16.9 [+ or -] 10.1). The mean number of snags per resting site and control plots was 5.7 [+ or -] 4.2 and 6.2 [+ or -] 4.7, respectively.
We found that average basal area for resting site plots (n = 147, [bar.x] = 33.92 [+ or -] 9.04 [m.sup.2] ha) was significantly higher than control plots (n = 147, [bar.x] = 31.10 [+ or -] 8.69 [m.sup.2] ha, P = 0.007). When comparing the total volume of CWD between resting sites and control plots we found no significant difference (U' = 9567, P = 0.090). Additionally, we did not find any significant difference in the total area of brush piles between resting site and control plots (U' = 16,318, P = 0.525).
American marten in the NLP of Michigan used resting sites that were associated with large diameter trees, high basal area and high percent canopy closure. Frequently, these resting sites were located in old, legacy trees which typically are higher than surrounding trees, often adding more to localized canopy closure, and generally have diameters of over 70 cm. Marten likely selected for larger trees because they contain preferred resting site features (e.g., cavities and large branches). The use of larger diameter trees has also been documented in other studies. In pine marten, for example, Zalewski (1997) found that three tree species, oak (Quercus robur), lime (Tilia cordata), and spruce (Picea abies), with an average DBH of 59 cm, were used most often for resting sites. Older trees have had more time to develop suitable resting site structures, such as cavities and larger branches. Additionally, larger trees often will provide an ample amount of overhead structure and abundant canopy cover. High percent canopy cover provided by legacy trees also seems to be an important habitat component consistent with other studies (Bowman et al., 1997; Chapin et al., 1997; Bull et al., 2005).
During cooler months, marten must rely on environmental structures to combat extreme weather changes. Cavities provide insulating properties that provide secure, microclimatically stable resting sites that keep marten warmer during cooler temperatures (Buskirk et al., 1989; McComb and Noble, 1981b). We found that cavities were used frequently during the winter season, which suggests these types of structures meet certain requirements needed during cooler temperatures. To a lesser extent, preconstructed nests (e.g., squirrel dreys, corvid or raptor constructions) may also possess some of these insulating properties (Pulliainen, 1973), as we observed nests being used during the winter season as well. Cavities and nests, however, were used fairly uniformly throughout the year (Fig. 2) and might offer properties that may also be used as a cooling mechanism during bouts of high humidity during warmer temperatures (Zalewski, 1997). Despite the potential cooling properties of cavities and nests, we did find that NLP marten selected exposed branches more in the summer. During summer, the leaves on deciduous trees provide increased canopy closure, which may decrease the risk of avian predation and allow marten to use cooler, exposed structures more readily. Selection for deciduous-dominated stands has been detected elsewhere during summer (Fuller and Harrison, 2005; Siren et al., 2016), yet our findings increase the resolution of habitat utilization and also link behavior with selection.
Numerous studies have shown the use of resting sites between the underside of the snowpack and the ground is prevalent across a large portion of the American marten range (Chapin et al., 1997; Zalewski, 1997; Bull and Heater, 2000). We expected subnivean sites would also be widely used within our study area as well. However, we observed marten using subnivean structures only four times during the two winter seasons. This could be, in part, because of the limited amount of CWD found throughout our study area. Frequently subnivean sites are created in areas where logs and other structures create open pockets under the snow, where marten have been known to create extensive tunnel systems (Bull and Heater, 2000). We found marginally higher levels of CWD in resting site plots which may reflect a preference for areas with increased woody litter, though the non-significant result was likely due to: (a) the unavailability of ground structures in a relatively young forest, and (b) current forest management prescriptions which discourage high ambient CWD levels. In addition to resting sites, adequate CWD can promote healthy small mammal populations (Harmon et al., 1986; Carey and Johnson, 1995) as a prey base. Forest management which retains or establishes CWD (e.g., brush piles, logs) would be beneficial to marten by increasing the accessibility of subnivean sites and, likely, food availability, by providing more habitat for small mammals.
American marten habitat use varies throughout their range in North America and are generally thought of as being a late successional stage forest specialist, which has led to their presence being considered as an indicator of forest health (Ruggiero et al., 1994; Buskirk and Powell, 1994; Bisonette and Broekhuizen, 1995). They use boreal forests in the north part of their range and mixed conifer/deciduous forest to the east (Jensen, 2012; Fuller and Harrison, 2005; Cheveau et al., 2013; Siren et al., 2016). Notwithstanding their specialization in mature forests, they seem to be able to adapt to their surroundings and use other habitats, providing those habitats possess features such as tree cavities, CWD, and/or closed canopy cover. We found marten were selecting primarily deciduous trees for resting sites, but surrounding tree composition around a large portion of sites consisted largely of coniferous tree species, often red pines. These conifer stands, although essentially monocultures, did have larger DBH deciduous trees scattered throughout. This indicates that marten were selecting resting sites at a microhabitat scale (i.e., deciduous trees for resting sites) influenced by more stand-level characteristics (i.e., coniferous forest types surrounding resting sites). This is important because areas once were thought to be low quality marten habitat (pine plantations) may, in fact, be supporting the population to some degree, as long as a modest deciduous component is available. The diurnal activity patterns of marten vary depending on season and region (Buskirk, 1983; Zielinski et al., 1983; Thompson and Colgan, 1994), which includes resting activity. We only documented day time resting sites of marten, sampling resting sites at other times was not logistically feasible, but may have revealed additional resting site preferences.
Our comparison of microsite habitats between resting and control plots indicated marten in the MNF prefer higher basal areas in the surrounding area when selecting resting sites. However, species composition and vertical structure (i.e., layers within a forest stand) between resting and control plots were not significantly different. The higher basal area associated with resting site plots could be, in part, be due to the large DBH of resting site trees. Payer and Harrison (2003) found marten used areas with higher live tree basal areas than unused areas and speculated that unused areas lacked the complex vertical structure marten need in their habitat. Additionally, they found marten avoided using forest stands with basal areas below a threshold of 18 [m.sup.2]/ha. Although our average basal area per vegetation plot (34 [m.sup.2]/ha) supports this threshold, we did find six resting site plots below the 18 [m.sup.2]/ha threshold (range 8.9-16.9 [m.sup.2]/ha). This suggests that marten may utilize a different lower threshold in the NLP to some degree, though optimal habitat may require extents with larger basal areas.
We found marten reused some resting sites, which is similar to other marten studies (Buskirk, 1984; Martin and Barrett, 1991). Fager (1991) found that during the spring in western Montana, marten resting site structures were reused from one to six times. The majority of our sites were reused only once; however, this could be due to sampling effort, as some of our observed resting sites appeared to be used more frequently than documented based on anecdotal evidence (e.g., scat remains, chew marks around cavities, prey remains). Marten reuse of sites could indicate that resting sites are limited in those areas, or that the reused sites simply offer superior benefits than other sites/structures found in their home range. Additionally, searching for new resting sites daily may come at a cost to marten in terms of lost opportunities to forage or mate. Therefore, marten may employ some type of learned behavior to locate where resting sites are located within their home ranges, which would help decrease search time in locating new resting sites (Martin and Barrett, 1991).
Habitat loss is one of the major factors affecting the survival of American marten populations (Stone, 2010). Specifically, habitat loss from timber harvest has been reported to negatively affect marten (Fuller and Harrison, 2005; Thompson, 1994; Cheveau et al, 2013; Siren et al, 2016; Moriarty et al, 2015, 2016; Poole et al, 2004; Potvin et al, 2000). Removal of overhead cover, reduction in large diameter coarse woody debris, simplification of structural diversity (both vertical and horizontal), and alteration of tree composition (e.g., conversion of forest types to suboptimal species or age classes) are all factors that can negatively affect marten population density, home range size, and mortality rates (Andruskiw et al., 2008; Stone, 2010). Marten are secondary cavity users that rely on other species to create cavities for their use. Although they are not obligate cavity users, in Michigan these types of structures are critical to their survival because other resting sites opportunities are often scarce. Therefore, marten would benefit from forest management that provides habitat for primary cavity excavators, such as woodpeckers. Additionally, providing management agencies with specific clues on how to identify potential marten resting sites should allow tree markers or harvesters to avoid active or potential resting sites, thereby mitigating some of the negative impacts associated with timber removal and the implications logging may have on maintaining a sustainable marten population.
Maintaining complex forest structure, both vertical and horizontal, abundant coarse woody debris, high percent canopy closure, and high basal areas should be considered when managing habitat for American marten. Managing forests to maintain and improve habitat will require the adoption of silvicultural techniques that retain large-diameter trees, encourage the retention and accumulation of large snags and CWD, and retain high basal areas associated with resting sites. Additionally, forest stands that are predominantly composed of coniferous species and are infiltrated with hardwood species should be retained or promoted, particularly deciduous trees that have a DBH of [greater than or equal to] 32.8 cm (the mean diameter of resting site trees). This should promote both forest species diversity and resting site availability. More research is needed in the NLP to examine the effects of tree removal, especially in red pine stands. Marten cannot tolerate >30-40% open areas inside their home ranges (Chapin et al., 1997; Fuller and Harrison, 2005; Payer and Harrison, 2003; Dumyahn et al., 2007). Therefore, experimentation with timber removal in marten home ranges presents opportunities to explore how timber management techniques and strategies might affect this species.
Acknowledgments.--We acknowledge the Little River Band of Ottawa Indians, Grand Valley State University, and Mesker Park Zoo for collaborating and making this research project a reality. Without their collaboration this project would have not been feasible. We thank C. Schumacher who volunteered countless hours of his time. We would also like to thank the graduate and undergraduate students who helped collect data, in particular, T. Hillman, M. Cannon, D. Bradke, R. Huggart, and A. Spenski. A special thanks to J. Witt for contributing to the early phases of this project.
ANDRUSKIW, M., J. M. FRYXELL, I. D. THOMPSON, AND J. A. BAKER. 2008. Habitat mediated variation in predation risk by the American marten. Ecology. 89(8):2273-2280.
BISSONETTE, J. A. AND S. BROEKHUIZEN. 1995. Martes populations as indicators of habitat spatial patterns: the need for a multiscale approach, p. 95-121. In: W. Z. Lidicker (ed.) Landscape approaches in mammalian ecology and conservation. University of Minnesota Press, Minneapolis, Minnesota.
BOWMAN, J. C. AND J. F. ROBITAILLE. 1997. Winter habitat use of American marten Martes americana within second-growth forest in Ontario, Canada. Wildl. Biol., 3:97-105.
BULL, E. L. AND T. W. HEATER. 2000. Resting and denning sites of American marten in northeastern Oregon. Northwest Sci., 74:179-185.
--, T. W. HEATER, AND J. F. SHEPHERD. 2005. Habitat selection by the American marten in Northeastern Oregon. Northwest Sci., 79(1):37-43.
Buskirk, S. W. 1983. The ecology of marten in southcentral Alaska. Ph.D. Dissertation. University of Alaska, Fairbanks, Alaska.
--. 1984. Seasonal use of resting sites by marten in south central Alaska. J. Wildl. Manage., 48:950-953.
--, S. C. FORREST, AND R. G. MARTIN. 1989. Winter resting site ecology of marten in the central Rocky Mountains. J Wildl Manage., 53(1):191-196.
--, And H. J. HARLOW. 1988. Temperature regulation in American marten (Martes-americana) in winter. Nat. Geogr. Res., 4:208-218.
--And R. A. POWELL. 1994. Habitat ecology of fishers and American marten, p. 283-296. In: S. W. Burkirk, A. S. Harestad, M. G. Raphael, and R. A. Powell (eds.). Marten, sables, and fishers: biology and conservation. Cornell University Press, Ithaca, New York.
--And L. F. RUGGIERO. 1994. American marten, p. 7-37. In: L. F. Ruggiero, K. B. Aubrey, S. W. Buskirk, L. J. Lyon, and W. J. Zielinski (eds.). The scientific basis for preserving forest carnivores: American marten, fisher, lynx, and wolverine in the western United Sates. U.S. Forest Service General Technical Report RM254.
CAREY, A. B. AND M. L. JOHNSON. 1995. Small mammals in managed, naturally young, and old-growth forests. Ecol. Appl. 5:336-352.
CHAPIN, T. G., D. J. HARRISON, AND D. M. PHILLIPS. 1997. Seasonal habitat selection by marten in an untrapped forest preserve. J. Wildl. Manage., 61:707-717.
CHEVEAU, M., L. IMBEAU, P. DRAPEAU, AND L. BELANGER. 2013. Marten space use and habitat selection in managed coniferous boreal forests of eastern Canada. J. Wildl. Manage., 77:749-760.
DESMARCHELIER, M., M. CHEVEU, L. IMBEAU, AND S. LAIR. 2007. Field use of isoflurane as an inhalant anesthetic in the American Marten (Martes americana). J. Wildl. Dis., 43(4):719-725.
EARLE, R. D., L. H. MASTENBROOK, AND T. F. REIS. 2001. Distribution and abundance of marten in northern Michigan. Michigan Department of Natural Resources Wildlife Report No. 3321.
ELLIOTT, K. J. AND D. HEWEITT. 1997. Forest species in upper elevation hardwood forests in the southern Appalachian Mountains. Castanea, 62(1):32-42.
FAGER, C. W. 1991. Harvest dynamics and winter habitat use of the pine marten in southwestern Montana. Master Thesis. Montana State University. Bozeman, MT. p. 73.
FULLER, A. K. AND D. J. HARRISON. 2005. Influence of partial timber harvesting on American marten in North-central Maine. J. wildl. Manage., 69(2):710-722.
KAPP, R. O. 1999. Michigan late Pleistocene, Holocene, and presettlement vegetation and climate, p. 312-316. In. J.R. Halsey (eds.), Retrieving Michigan's buried past: The archaeology of the great lakes state. Cranbrook Institute of Science, bulletin 64, Bloomfield Hills, Michigan.
MARTIN, S. K. AND R. H. BARRETT. 1991. Resting site selection by marten at Sagehen Creek, California. Northwestern Nat., 72:37-42.
MAZUREK, M.J. AND W.J. ZIELINSKI. 2004. Individual legacy trees influence vertebrate wildlife diversity in commercial forests. For. Ecol. Manage., 193:321-334.
MCCOMB, W.C. AND R.E. NOBLE. 1981b. Microclimates of nest boxes and natural cavities in bottomland hardwoods. J. Wildl. Manage. 45:284-289
MORIARTY, K. M.,J. D. BAILEY, S. E. SMYTHE, ANDJ. VERSCHUYL. 2016. Distribution of Pacific marten in coastal Oregon. Northwestern Nat., 97:71-81.
--, C. W. EPPS, M. G. BEITS, D.J. HANCE, J. D. BAILEY, AND W.J. ZIELINSKI. 2015. Experimental evidence that simplified forest structure interacts with snow cover to influence functional connectivity for Pacific marten. Landscape Ecol. 30:1865-1877.
PAYER, D. C. AND D.J. HARRISON. 2003. Influence of forest structure on habitat use by American marten in an industrial forest. For. Ecol. Manage., 179(1-3):145-156.
POOLE, K. C., A. D. PORTER, A. DE VRIFS, A. C. MAUNDRELL, S. D. GRINDAL, AND C. C. ST. CLAIR. 2004. Suitability of a young deciduous-dominated forest for American marten and the effects of forest removal. Can. J. Zool, 82:423-435.
PORTER, A. D., C. C. ST. CLAIR, AND A. DE VRIES. 2005. Fine-scale selection by marten during winter in a young deciduous forest. Can. J. For. Res., 35:901-909.
POTVIN, F., L. BELANGER, AND K. LOWELL. 2000. Marten habitat selection in a clear-cut boreal landscape. Consent. Biol., 14:844-857.
PULLIAINEN, E. 1973. Winter ecology of the red squirrel (Sciurus vulgaris L.) in northeastern Lapland. Ann. Zool. Fennici 10:487-494.
RUGGIERO, L. F., K. B. AUBRY, S. W. BUSKIRK, L. J. LYON, AND W. J. ZIELINSKI. 1994. The scientific basis for conserving forest carnivores: American marten, fisher, lynx, and wolverine in the western United States (General Technical Report No. RM-254). U.S. Department of Agriculture, Forest Service, Rocky Mountain Forest and Range Experiment Station, Fort Collins, Colorado.
SIREN, A. P. K., P.J. PEKINS, M.J. DUCEY, AND J. R. KILBORN. 2016. Spatial ecology and resource selection of a high elevation American marten population in the northeastern United States. Can. J. Zool, 94:169-180.
SPENCER, W. D. 1987. Seasonal rest-site preferences of pine marten in the northern sierra Nevada. J. Wildl. Mange., 51:616-621.
STONE, K. 2010. Martes American. In: Fire effect information system [online]. U.S. Department of Agriculture, forest service, rocky mountain research station, fire sciences laboratory (producer). Available: http://www.fs.fed.us/database/feis/.
TAYLOR, S. L. AND S. W. BUSKIRK. 1994. Forest microenvironments and resting energetics of the American marten Martes americana. Ecography 17(3):249-256.
THOMPSON, I. D. 1994. Marten populations in uncut and logged boreal forests in Ontario. J. Wildl. Manage., 58:272-280.
--AND P. W. COLGAN. 1994. Marten activity in uncut and logged boreal forests in Ontario. J. Wildl. Manage., 58:280-288.
USDA FOREST SERVICE. 1996. American marten conservation strategy for the Huron-Manistee National Forest. USDA Forest Service. Huron-Manistee National Forest District Office. Cadillac, Michigan.
WILLIAMS, B. W., J. H. GILBERT, AND P. A. ZOLLNER. 2007. Historical perspective on the reintroduction of the fisher and American marten in Wisconsin and Michigan. USDA Forest Service, General Technical Report GTR-NRS-5.
ZALEWSKI, A. 1997. Patterns of resting site use by pine marten Martes martes in Bialowieza National Park (Poland). Acta Theriol., 42:153-168.
ZIELINSKI, W. J., W. D. SPENCER, AND R. H. BARRETT. 1983. Relationship between food habits and activity patterns of pine marten. J. Mamm., 64:387-396.
SUBMITTED 17 MARCH 2017
ACCEPTED 9 DECEMBER 2016
ROBERT L. SANDERS (1) AND ARI CORNMAN (4)
Little River Band of Ottawa Indians, Manistee, Michigan 49660
PAUL KEENLANCE (2) AND JOSEPH J. JACQUOT (3)
Department of Biology, Grand Valley State University, Allendale, Michigan 49401
DAVID E. UNGER (5)
Division of Natural Sciences, Maryville College, Maryville, Tennessee 37804
MARIA SPRIGGS (6)
Busch Gardens, Tampa, Florida 33612
(1) Corresponding author present address: Little River Band of Ottawa Indians, Manistee, Michigan 49660; e-mail: firstname.lastname@example.org
(2) e-mail: email@example.com
(3) e-mail: firstname.lastname@example.org
(4) Present address: Little River Band of Ottawa Indians, Manistee, Michigan 49660; e-mail: email@example.com
(5) e-mail: firstname.lastname@example.org
(6) e-mail: Maria.Spriggs@BuschGardens.com
Caption: FIG. 1.--Location of the study area in Michigan's Northern Lower Peninsula. American marten were monitored in the Manistee National Forest (MNF) from May 2011 through December 2013
Caption: FIG. 2.--Distribution of American marten resting site structures by tree species found from May 2011 through December 2013 in the Manistee National Forest in Michigan's Lower Peninsula. Tree species < 2% of the total were excluded from the figure
Caption: FIG. 3.--Number of times American marten were documented to reuse particular resting sites from May 2011 through December 2013 in the Manistee National Forest in Michigan Northern Lower Peninsula
TABLE 1.--Resting site structures used by American marten (15 males and 10 females) in the summer (Apr.-Sep.) and winter (Oct.-Mar.) seasons in the Manistee National Forest from May 2011 through December 2013. Males and females were pooled for analysis. Nests included those of squirrels, corvids, or raptors Resting Summer season Winter season Total resting site site structures structure n % n % n % Cavity 129 39.3 126 64.9 255 48.9 Branch 137 41.8 25 12.9 162 31.0 Nests 52 15.9 38 19.6 90 17.2 Hollow log 9 2.7 2 1.0 11 2.1 Subnivean 1 0.3 3 1.5 4 0.8 Total 328 100 194 100 522 100 TABLE 2.--Live tree species used for resting sites and associated type of resting site used by American marten in the Manistee National Forest from May 2011 through December 2013. Diameter at breast height (DBH) is a total average per tree species and measured in centimeters (cm, SD = standard deviation). Nests included those of squirrels, corvids, or raptors. Percentage is the total for the column Live tree Resting site structures species Cavity Branch n DBH (SD) % n DBH (SD) % Oak (a) 107 55.7 (14.7) 54.6 57 49.9 (15.9) 39.0 Maple (b) 35 46.7 (14.4) 17.9 34 44.0 (12.4) 23.3 Red pine 1 39.4 (n/a) 0.5 27 27.8 (5.8) 18.5 Aspen (c) 31 42.4 (9.0) 15.8 1 33.5 (n/a) 0.7 White 1 40.9 (n/a) 0.5 9 36.9 (14.3) 6.2 pine Basswood 11 49.6 (8.4) 5.6 4 42.2 (11.1) 2.7 Jack pine 0 n/a 0.0 9 26.5 (4.6) 6.2 American 3 45.9 (8.4) 1.5 2 32.7 (5.6) 1.4 beech Black 5 33.9 (12.4) 2.41 3 35.6 (n/a) 2.1 cherry White 1 64.5 (n/a) 0.5 0 n/a 0.0 ash Yellow 1 58.7 (n/a) 0.5 0 n/a 0.0 birch Total 196 47.8 (11.2) 100 146 36.6 (10.0) 100 Live tree Resting site structures Overall total species Nest n DBH (SD) % n DBH (SD) % Oak (a) 10 47.7 (15.8) 11.5 174 53.2 (15.4) 40.6 Maple (b) 10 41.7 (11.0) 11.5 79 45.0 (13.2) 18.4 Red pine 50 29.8 (5.7) 57.5 78 29.3 (5.9) 18.2 Aspen (c) 1 53.3 (n/a) 1.1 33 42.5 (9.1) 7.7 White 6 25.3 (8.1) 6.9 16 33.1 (12.9) 3.7 pine Basswood 1 41.4 (n/a) 1.1 16 46.9 (9.3) 3.7 Jack pine 6 27.9 (3.9) 6.9 15 27.1 (4.2) 3.5 American 2 25.7 (16.2) 2.3 7 36.3 (12.7) 1.6 beech Black 1 40.4 (n/a) 1.1 9 35.1 (10.4) 2.1 cherry White 0 n/a 0.00 1 64.5 (n/a) 0.2 ash Yellow 0 n/a 0.00 1 58.7 (n/a) 0.2 birch Total 87 37.0 (10.1) 100 429 42.9 (10.3) 100 (a) Oak species include black, red, and white oak (b) Maple species include red and sugar maple species (c) Aspen species include big tooth and quaking species TABLE 3.--Snag tree species used for resting sites and associated structure type used by American marten in the Manistee National Forest from May 2011 through December 2013. Diameter at breast height (DBH) is a total average per tree species and measured in centimeters (cm, SD = standard deviation). Percentage is a total for the column Resting site structures in snags Cavity Branch Snag tree species n DBH (SD) % n DBH (SD) % Oak 25 46.6 (11.7) 47.2 4 38.7 (10.3) 30.8 species (a) Aspen 15 34.8 (7.7) 28.3 0 n/a 0.0 species (b) Maple 9 33.0 (8.2) 17.0 2 32.4 (13.9) 15.4 species (c) Jack pine 0 n/a 0.0 5 19.1 (5.3) 38.5 Basswood 1 41.7 (n/a) 1.9 1 31.8 (n/a) 7.7 White pine 1 51.8 (n/a) 1.9 0 n/a 0.0 Red pine 1 30.1 (n/a) 1.9 0 n/a 0.0 White ash 0 n/a 0.0 1 44.4 (n/a) 7.7 Black 1 37.3 (n/a) 1.9 0 n/a 0.0 cherry Total 53 39.3 (9.2) 100 13 33.3 (9.8) 100 Resting site structures in snags Nest Overall total Snag tree species n DBH (SD) % n DBH (SD) % Oak 1 34.8 (n/a) 33.3 30 45.1 (11.6) 43.5 species (a) Aspen 0 n/a 0.0 15 34.8 (7.7) 21.7 species (b) Maple 0 n/a 0.0 11 32.9 (8.5) 15.9 species (c) Jack pine 2 32.4 (7.8) 66.7 7 22.9 (8.4) 10.1 Basswood 0 n/a 0.0 2 36.7 (7.0) 2.9 White pine 0 n/a 0.0 1 n/a 1.4 Red pine 0 n/a 0.0 1 n/a 1.4 White ash 0 n/a 0.0 1 n/a 1.4 Black 0 n/a 0.0 1 n/a 1.4 cherry Total 3 33.6 (7.8) 100 69 34.5 (8.6) 100 (a) Oak species include black, red, and white oak trees (b) Aspen species include big tooth and quaking aspen trees (c) Maple species include red and sugar maple trees
|Printer friendly Cite/link Email Feedback|
|Author:||Sanders, Robert L.; Cornman, Ari; Keenlance, Paul; Jacquot, Joseph J.; Unger, David E.; Spriggs, Mar|
|Publication:||The American Midland Naturalist|
|Date:||Apr 1, 2017|
|Previous Article:||Diets of fishers (Pekania pennanti) and evidence of intraspecific consumption in Pennsylvania.|
|Next Article:||Tree species preferences of foraging songbirds during spring migration in floodplain forests of the upper Mississippi River.|