State of the Greater Fundy Ecosystem

GREATER FUNDY ECOSYSTEM RESEARCH PROJECT

UNB Faculty of Forestry and Environmental Management

State of the Greater Fundy Ecosystem

American Marten Response to
Forest Characteristics

Maryse C. Bourgeois
Department of Biology
Acadia University, Wolfville, N.S.
B0P 1X0

The American Marten (Martes americana), once widely distributed throughout North America, has been extirpated, or greatly reduced in number in many areas, due to heavy trapping pressure and habitat destruction (Banfield, 1977). In New Brunswick, the species had been virtually eliminated from the southern part of the province by 1930. Recently, the Marten has become a species of interest to conservation managers and wildlife researchers because of its potential as an indicator of mature forests.

American Marten (Photo: G. Forbes)

In 1982, a recommendation was made to re-introduce Marten to Fundy National Park (Fundy NP) (Bateman, 1982). From 1984 to 1989, and then again in 1991, Marten were captured in northern New Brunswick and released in the Park. In total, 50 individuals (26 males, 19 females and five young born in captivity (one male, one female and three of unknown sex)) were released. Although evidence suggests that a breeding population has established itself in the area, it is not known if Fundy NP, due to its isolated nature, can support a viable population. In addition, the forest both inside and outside of the Park consist of habitats that differ in quality and suitability for Marten.

Most researchers who have examined Marten habitat requirements have found that Marten choose dense stands of conifer dominated forests (e.g., Koehler and Hornocker, 1977; Steventon and Major, 1982; Spencer et al., 1983; Slough, 1989). Marten, however, seem to be flexible in their use of habitat as they are also found in habitat types other than mature conifer, such as mixed forests (Soutiere, 1979; Steventon and Major, 1982; Katnik, 1992). Structural characteristics within the various forest types may influence use of these habitats. Overhead cover and stand maturity have been found to be important (Koehler and Hornocker, 1977; Hargis and McCullough, 1984; Snyder and Bissonette, 1987). Also, mixed-aged stands provide a larger variety of prey, and effective protection due to overhead cover during winter. Different levels of the vertical structure progressively become important as snow depth increases (Hargis and McCullough, 1984).

Direct quantitative observations of an animal's behaviour should provide information relating to the relative importance of habitats. As Marten travel seems to be influenced by a search for prey and a requirement for protective overhead cover (Hargis and McCullough, 1984), track patterns in the snow may serve as valuable indicators of such activities. Marshall (1951) suggested that Marten will spend more time in their selected habitat, as indicated by the more convoluted travel paths. The degree of tortuosity may therefore be a good measure of the relative importance of a habitat. A straight path suggests that the individual is crossing an area, but not choosing it for purposes other than traveling, whereas a tortuous path (characterized by twists and turns) is a sign of increased use.

Fractal analysis was used to compare the relative values of different habitats. The fractal dimension of a pattern is a measure of its degree of irregularity. Its value, which ranges between 1 and 2, increases as the complexity of the pattern increases. A fractal dimension which varies with scale may indicate the existence of distinct scale-dependent processes (Krummel et al., 1987; Sugihara and May, 1990).

Fractal analysis may provide a simple means of obtaining data on how an animal uses different habitats. Although habitat selection cannot be assessed by analyzing track patterns, a comparison of path tortuosities in different habitats may serve as a predictor of habitat selection. Detailed analyses of movement patterns may provide added information when used in conjunction with the traditional use/availability technique.

METHODS

Marten tracks were followed in 28 stand types during the winters of 1993-1994 and 1994-1995. Five male Marten were equipped with radio-transmitters during the first winter, and two were refitted with transmitters the second year. Tracks were located while searching along roads and trails, by radio-telemetry, or during track transect surveys conducted by park staff. In order to plot Marten paths, track patterns were measured. Track direction and distance traveled between each change in direction were recorded. Additional information recorded with the track data included: i) structural forest components present at the end of a track segment (portion of the trail between two noticeable changes in direction) and at sites that Marten investigated; ii) logs encountered; iii) presence of subnivean access holes; and iv) presence of Snowshoe Hare tracks. Forest composition surrounding the tracks was documented for each stand in which the Marten traveled.

Fractal dimensions were estimated for each Marten trail in each habitat, using a fractal dimension computer program. (Nams, 1996). They were estimated from the scale of 1 m to 50 m following the divider method (Mandelbrot, 1967). As track tortuosities differed between the scale of 1 to 4.5 m and the scale of 4.5 to 50 m, fractal dimensions were calculated at these two scales, in addition to the 1 to 50 m scale. Multiple regression analyses were used to assess the influence of the following variables on track tortuosity: presence of conifer and deciduous in the overstory and understory, stand maturity, and canopy closure.

RESULTS

A total of 3270 m of Marten tracks were followed. The fractal dimensions of 55 tracks were used in the analyses. The degree of tortuosity in Marten tracks varied with scale. Tracks were straighter at the scale of 1 to 4.5 m than at the scale of 4.5 to 50 m (P < 0.05; Wilcoxon test).

At P <= 0.10, both canopy closure and presence of conifer in the understory were found to influence track tortuosity at the 1 to 4.5 m scale (R2 = 0.239) (Table 1). At the other scales of observation, patterns were not influenced by species composition in the overstory or the understory, stand maturity, or canopy closure.

Out of 846 trail segments followed, 104 segments ended at structural forest components where Marten changed direction. This indicates that they responded to structures in some way once every 30.9 m of track. These structures consisted of i) tree bases (N (number) = 46; of which 38 were coniferous), ii) woody debris (N = 24), iii) snags (N = 11), iv) low hanging spruce branches (N = 4), v) subnivean access holes (N = 4), vi) animal tracks (N = 4; Squirrel = 1, Hare = 2, Weasel = 1), vii) frozen creek (N = 1), viii) rock (N = 1), and ix) unknown mound-like structure (N = 2). Subnivean access holes were encountered 10 times. One subnivean access way was found to be used, as indicated by the concentration of Marten tracks adjacent to it.

The median densities of trees along the Marten tracks in both conifer and mixed forest stands were 7 per 20 m2, while that in deciduous was 5 per 20 m2. Conifer and deciduous understories had tree densities of 8 per 20 m2 and mixed understories had 7 trees per 20 m2.

The relatively straight Marten paths observed at the 1 to 4.5 m scale suggests that Marten may be attracted to a specific habitat feature interspaced at approximately 4.5 m. The density of trees in the overstory and the understory were the only forest characteristics measured along the tracks which coincided with the 1 to 4.5 m scale. These linear patterns may therefore represent the animal inspecting trees. Also, Marten were often found to direct their travel course toward trees. This behaviour may be related to the animal searching for prey (Jedrzejewski et al., 1993) or attempting to maximize cover (Hargis and McCullough, 1984).

Marten may also respond to woody debris, subnivean access holes, and to the tracks of prey when these are encountered. Marten were, in fact, found to direct their course toward these structures on several occasions, and most frequently toward woody debris. A host of other factors may also influence Marten travel patterns. For instance, the spacing of subnivean prey may be important. Further investigation is needed to properly identify the forest components causing Marten to respond differently at these two spatial scales.

Track tortuosity was fairly constant from the 4.5 m to 50 m scale, suggesting that no additional processes influenced movement within this range of scales.

Track patterns at the 1 to 4.5 m scale were influenced by the proportion of canopy closure and by the presence of conifer in the understory. Tracks tended to be more complex in stands with high and low canopy closure than in stands providing partial canopy closure. The small sample of stands having a low canopy closure combined with this group's relatively large confidence interval puts in question the significance of this result. The response observed in stands having a high canopy closure may reflect an increase in the search for prey and/or a positive response to protective overhead cover.

Tracks also tended to be more tortuous in habitats with a notable presence of conifer in the understory. Conifer saplings may provide Marten with protective cover, thermal benefits, and/or good foraging habitat. The presence of young conifers may also favour prey availability. Crevices may form around low conifer branches touching the snow, thus providing greater accessibility to the subnivean layer (Hargis and McCullough, 1984; Corn and Raphael, 1992). Marten trails often passed under low conifer branches.

Habitat variables, which were not measured during the present investigation, may also influence Marten travel patterns. Again, the presence of subnivean prey may be important to Marten search patterns. Marten may detect subnivean prey from the snow surface, or they may have previous knowledge of the location of prey populations (Thompson, 1986; Corn and Raphael, 1992; Sherburne and Bissonette, 1994). Height of the first cover above the snow surface (Hargis and McCullough, 1984), hardness and thickness of snow (Raine, 1983), and density of trees in the overstory and the understory may also have an influence on Marten movement patterns.

There are certain limitations to using the fractal approach with respect to habitat evaluation. The examination of track patterns does not provide information on the habitat components which the animal perceives, nor does it disclose the reasons behind the observed response to specific habitat features (With, 1994). Background knowledge on the species under investigation will help to interpret the results, which is the case for all studies regardless of method used to examine habitat use.

This study was an exploratory examination of the value of using track patterns to identify habitat characteristics important to Marten during winter. Fractal analyses provided insight into Marten reactions to their surroundings by quantifying movement patterns. Findings were comparable to others obtained by investigators using the resource selection approach at a micro-habitat level. For instance, Hargis and McCullough (1984) found Marten to select forested areas with good canopy closure, and overhead cover less than 3 m above snow level for hunting, feeding, and resting.

IMPLICATIONS FOR MANAGEMENT

Use of fractal analysis may provide an innovative method of evaluating habitat from the perspective of the animal, which has been lacking in most ecological studies to date. Examining animal interactions with their micro-habitat may provide insight on the animal's reason for selecting stand types. Habitat selection at a stand level may remain undetected if animals are selecting micro-habitat structures rather than forest types. Specific forest components rather than stand types may be the basic influence on marten use of habitats.

Mixed aged stands, which provide suitable cover both at the canopy and the understory level, appear to be valuable to marten during periods of complete snow cover. However, it is important to note that these variables may not be the only ones meaningful to marten. Forest harvesting followed by planting of even-aged mono-cultures, does not retain the desired vertical structure. In order to preserve stands with these characteristics on the landscape, natural conifer dominated stands should be maintained within areas managed for fiber production.

Further Reading:

Bourgeois, M.C. 1995. Preliminary report on American marten habitat associations and winter foraging activity in Fundy National Park. Research Notes. FUN/95-14. Alma, N.B.

Bourgeois, M.C. 1997. Fractal analysis of movement patterns used to compare American Marten response to different habitat characteristics. Masters thesis. Department of Biology. Acadia University. Wolfeville, N.S.

- Table of Contents

- View other case studies

- Bibliography