Ecological Research

, Volume 20, Issue 1, pp 95–101

Home range and habitat use by the sable Martes zibellina brachyura in a Japanese cool-temperate mixed forest

Authors

    • Graduate School of Environmental Earth ScienceHokkaido University
  • Seigo Higashi
    • Graduate School of Environmental Earth ScienceHokkaido University
Original Article

DOI: 10.1007/s11284-004-0012-y

Cite this article as:
Miyoshi, K. & Higashi, S. Ecol Res (2005) 20: 95. doi:10.1007/s11284-004-0012-y

Abstract

Home range and habitat use of the sable Martes zibellina brachyura were studied in a cool-temperate mixed forest in northernmost Japan. In both sexes, some sables showed a wide range of migration without establishing home ranges and the others had home ranges of 0.50–1.78 km2 (mean: 1.12±SD 0.495 km2, n =6) which were not significantly correlated with body weight or age. The analysis of canine tooth annuli revealed that the maximum age was 5.5 years. The home ranges of some sables overlapped so extensively that the home ranges and even the core areas did not appear exclusive to other sables. We determined resting sites and foraging routes in snow in winter. Comparison of vegetation at the resting sites and foraging routes with habitat availability suggested that the sables preferred resting in dense-tree forests with many tree species and debris probably in order to avoid predators (red foxes) and strong wind and foraging in forests of climax succession which are usually rich in their prey such as voles and mice.

Keywords

CWDHabitat useHome range Martes zibellinaTerritoriality

Introduction

The genus Martes consists of seven species which are distributed in Eurasia and North America. Sable M. zibellina (Linnaeus), Japanese marten M. melampus (Wagner), American marten M. Americana (Turton) and pine marten M. martes (Linnaeus) are similar to each other in body size and habitat preference and are recognized belonging to a “superspecies” (Anderson 1970). The sable Martes zibellina is distributed in northeastern Asia including Siberia, Sakhalin, northern Mongolia, northern China and northernmost Japan (Bakeyev and Sinitsyn 1994). Ellerman and Morrison-Scott (1966) classified them into 14 subspecies. The Japanese sable ( M. zibellina brachyura) is restricted to cool-temperate mixed forests of the northernmost island of Hokkaido. Murakami (2003) analyzed the scat contents of this subspecies and reported that they feed on small mammals such as voles ( Clethrionomys spp.) throughout the year, with occasional ingestion of insects and fruit in summer and autumn. Kurose et al. (1999) and Hosoda et al. (1999) analyzed the genetic diversity of M. zibellina brachyura and found that this subspecies is genetically less diverse than Russian subspecies of sable and Japanese marten, evidently representing a relatively new immigration to Hokkaido Island from the continent. Apart from these studies, M. zibellina brachyura has rarely been investigated and in particular ecological data are so deficient that this subspecies is categorized DD (data deficient) on the IUCN Red List. In the present study, therefore, we conducted surveys of home range and habitat use of this subspecies to obtain basic ecological data.

The home range and habitat use have frequently been studied in American marten (e.g. Buskirk and Macdonald 1989; Chapin et al. 1997; Potvin et al. 2000; Smith and Schaefer 2002) and pine marten (e.g. Clevenger 1993; Zalewski et al. 1995; Zalewski 1997). These studies suggest intraspecific variation of the home range size and territoriality probably depending on habitat conditions. Only Buskirk et al. (1996) and Xu et al. (1997) have ever investigated the home range and habitat use of sables in the taiga of northeastern China where the habitat is very different from that of cool-temperate mixed forests dominated by dwarf bamboos and deciduous broad-leaved trees. This paper is the first report on the home range and habitat use of sables inhabiting mixed forests.

Materials and methods

Study area

In Nakagawa Experiment Forest of Hokkaido University (44°50’ N; 142°00’ E; 50–330 m asl), this study was conducted from 23 November 2000 to 24 November 2001 at station 1 and from 24 August 2002 to 27 December 2003 at station 2. In this forest, the mean monthly temperature ranged from 20.3°C (August) to –10.4°C (January) and mean annual precipitation was 1,650 mm, with up to 2 m of snow cover in winter. This forest is dominated by fir Abies sachalinensis Mast., spruce Picea jezoensis Carr., oak Quercus crispula Blume, maple Acer mono Maxim. and birch Betula ermanii Cham. with a floor of dwarf bamboos Sasa senanensis Rehder and S. kurilensis Makino and Shibata and inhabited by such mammals as brown bear Ursus arctos Linnaeus, sika deer Cervus nippon Temminck, red fox Vulpes vulpes Linnaeus, mountain hare Lepus timidus Linnaeus, chipmunk Tamias sibiricus (Laxmann), grey red-backed vole Clethrionomys rufocanus (Sundevall), and large Japanese field mouse Apodemus speciosus (Temminck). Sables prey upon voles and mice and are hunted by foxes. In addition, white-tailed eagle Haliaeetus albicilla (Linnaeus) and mountain hawk eagle Spizaetus nipalensis (Hodgson) appeared to be predators for sables.

Methods

Capture of sables and location of home range

Live trapping of non-juvenile sables was made using cage traps. Captured sables were immobilized with intramuscular injection of medetomidine hydrochloride (0.3 mg/kg) and ketamine hydrochloride (2.5 mg/kg). We determined sex, weighed each individual, and categorized wear degree of canine teeth into three classes: 1, almost intact with little wear; 2, worn less than half of canine length; and 3, worn more than half of canine length. Then we attached a 20 g radio collar to the sables, reversed them with atipamezole (1.5 mg/kg) and released them at the capture site.

Locations of the radio-collared sables were determined by triangulation techniques (Bailey 1974) in which the direction of each sable was confirmed from three points within 10 min, and the location was represented by a triangle on a topographical map. In the present study, we recorded locations only when the size of the triangle was smaller than 25 mm2 on a 1:25,000 map. We tried to locate the sable daily from the day after release. Finally, we determined the size of minimum convex polygon (MCP) where the outermost locations were connected, in order to estimate the home range size of the sable; and core areas in each home range were determined for sables which were successfully located more than 90 times, by using the kernel method (Worton 1989) equipped with Animal Movement extension (Hooge and Eichenlaub 1997) of ArcView 3.1 (ESRI, Redlands, California), where we used 50% for probability contours.

The ages of dead sables were determined by counting the number of canine tooth annuli under a microscope (Strickland et al. 1982).

Analyses of habitat use

When the home range of a sable was determined, it was pictured on a vegetation map digitized for ArcView 3.1 which automatically measured area of each stand type and counted the number of locations in it. Following Neu et al. (1974), the stand-scale habitat preference was analyzed by comparing the number of locations with the stand area in the home range (cf. Fig. 3).

By locating radio-collared sables in winter, we were able to determine their resting sites and foraging routes on the snow. In order to analyze the vegetational and topographical characteristics at the resting sites and foraging routes, we conducted vegetational and topographical surveys in summers of 2001 and 2003 at stations 1 and 2, respectively. At these stations, circular plots (radius=5 m) were set up on the intersections of 300 m×200 m interval grids (open circles in Fig. 4) and at resting sites (closed circles) and foraging routes (closed triangles). In each plot, we recorded DBH (diameter at breast height), height and species name of trees (≥3 cm in DBH). Canopy cover was estimated from hemispherical photographs taken by a fisheye lens camera (Kuusipalo 1985). The amount of dead trees (CWD: coarse woody debris) was estimated by ∑πri2 Li where Li and ri are length and mean radius of the i th debris collected in a plot, respectively. In this survey, the debris was defined as standing dead trees and fallen stems or twigs of 10 cm or more diameter. In addition, steepness and aspect of the plot were measured using a clinometer with compass and the elevation was estimated from a topographical map. From these data we lined up the following items for each plot: (1) maximum DBH: the DBH of the largest tree; (2) basal area of coniferous trees and (3) of broad-leaved trees; (4) number of trees; (5) number of tree species; (6) tree height: represented by a median of all trees; (7) canopy cover; (8) CWD; (9) steepness; (10) aspect; and (11) elevation. The vegetational or topographical characteristics at resting sites and foraging routes were analyzed by comparing the 11 items between resting sites or foraging routes and grid intersections representing the habitat availability in the forest. For this analysis, we used Kolmogorov-Smirnov (K-S) test and Jacobs’s preference index (Jacobs 1974).

Results

Migration and home range

A total of 19 males and 8 females were captured by cage traps, and the mean body weight of males was 893.0±SD 87.1 g which was significantly larger than 668.1±81.6 g of females ( t =–6.232, P <0.0001). These non-juvenile sables were radio-tagged and tracked, and 14 of them were successfully located more than five times (Table 1) until they were evidently hunted by red foxes (3 sables), died of unknown mortal factors (2 sables), detached a transmitter (1 sable) or were lost probably due to troubles of batteries or transmitters or to long-distance emigration (8 sables).
Table 1

Details of 14 radio-tagged sables and four carcass samples. M male; F female

Sable

Weight (g)

Number of locations

Tracking period

Fate

Canine teeth wear

Estimated age

Tracked individuals

 M1

860

127

Dec 2000–May 2001

Lost

1

-

 M2

770

31

Mar–May 2001

Lost

1

-

 M3

890

29

Mar–May 2001

Lost

2

-

 M4

1050

11

Oct 2001–Mar. 2003

Lost

2

-

 M5

1010

44

Aug–Dec 2002

Dead

3

-

 M6

1060

177

Aug 2002–Dec 2003

Preyed

2

4.5

 M7

890

5

Nov 2002

Lost

1

-

 M8

980

5

Dec 2002–Feb 2003

Transmitter-detached

3

-

 M9

990

38

Mar–July 2003

Lost

1

-

 M10

860

33

Apr–June 2003

Dead

3

4

 F1

730

7

Jan–Mar 2003

Preyed

1

-

 F2

740

15

Jan–Apr 2003

Lost

2

-

 F3

600

5

Mar–Apr 2003

Lost

1

-

 F4

600

94

Apr–Sep 2003

Preyed

1

2

Carcass samplesa

 Male

-

-

-

Carcass

3

5.5

 Male

-

-

-

Carcass

3

5.5

 Male

-

-

-

Carcass

1

0.5

 Female

-

-

-

Carcass

1

0.5

aDied of road traffic accident

Figure 1 shows the relationship between the number of locations and the accumulated area of minimum convex polygon (MCP). In 8 of the 14 sables, the home range was not determined due to the small number of locations and/or the absence of home range (Fig. 1A). The longest migration was about 10.5 km in M4 (Fig. 2). In the other 6 sables, the accumulated area of MCP reached asymptote, which probably indicates the home range size of each sable (Fig. 1B). The home range size estimated by MCP 100% method ranged from 0.50 km2 (M9) to 1.78 km2 (M1), with an average of 1.12±SD 0.495 km2. The home range size was not significantly correlated with body weight (Spearman’s rank correlation coefficient rs =–0.39, P =0.36) nor with canine teeth wear ( r s=0.30, P =0.58) which was significantly correlated with age ( r s =0.86, P =0.034) estimated from cementum annuli of canine teeth of 7 dead sables including 4 carcasses of sables which died of road traffic accidents (Table 1). The maximum age was 5.5 years. The home range was established not only by old sables (M5, M10) but also by young sables (M1, M9). The resident sables showed an extensive overlap of home ranges (Fig. 2), suggesting that they were not exclusive to each other. For M1, M6 and F4 in which the number of locations exceeded 90, we could determine the site of their core areas. As shown in Fig. 2, the core areas of M6 and F4 overlapped with home ranges of M9 and M5 and of M10, respectively, suggesting that they were not territorial. Such females as F1 and F2 appeared to be transients wandering in station 2 without establishing home ranges. M4 was found at station 1 in the 2001−2002 winter and recaptured by a trap at station 2 on 28 March 2003, though it was unknown whether or not he had a home range in 2003.
Fig. 1

Accumulation curves of minimum convex polygon area with increasing number of locations. A Curves did not reach asymptote; B curves reached asymptote

Fig. 2

Home ranges of six sables (M1, M5 , M6 , M9 , M10 and F4) and migration routes of eight sables (M2, M3 , M4 , M7 , M8 , F1 , F2 and F3), in which locations were connected by straight lines in the order of discovery. For M1, M6 and F4 in which the number of locations exceeded 90, behavioral core areas are shaded

Habitat use

Figure 3 shows the distribution of locations within six home ranges, which were divided into five stand types: coniferous stand, deciduous stand, mixed-forest stand, plantation and open area. The observed number of locations in each stand type was compared with expected number of locations which was calculated by L A i/ H where L is the observed number of locations, H is the home range size and Ai is the total area of stand type i. However, chi-square test did not detect any significant difference between observed and expected numbers of locations, suggesting that sables do not show stand-scale preference for habitats.
Fig. 3

Distribution of locations within six home ranges (cf. Fig. 2). The observed number of locations in each vegetation type was compared with expected number of locations which was calculated by L · Ai/ H where L is the observed number of locations, H is the home range size and Ai is the total area of vegetation type i

Figure 4 shows the distribution of three categories of circular plots where we surveyed vegetation and topography. The availability of habitats was estimated from the survey of 186 plots set up on intersections of 300m×200m grids (open circles in Fig. 4), and the average ±SE was 38.4±1.79 cm maximum DBH, 706.3±127.22 cm2 basal area of coniferous trees, 1,859.4±168.74 cm2 basal area of broad-leaved trees, 67.1±1.35% canopy cover, 8.8±0.43 m tree height, 8.0±0.66 the number of trees, 4.9±0.28 number of tree species, 0.8±0.23 m3 CWD, 23.4±1.14° steepness, 130.3±4.28 m elevation and 152.2±10.60° in aspect. The vegetational and topographical characteristics at resting sites and foraging routes were analyzed by comparing vegetation and topography in the plots of resting sites (57 closed circles in Fig. 4) and foraging routes (71 closed triangles) with the availability. While none of topographical factors, i.e. steepness, elevation and aspect, affected sables in choosing resting sites and foraging routes, some vegetational factors seemed to be involved in their habitat preference. Since sables often took a rest under fallen trees, CWD at resting sites was significantly larger than availability (Fig. 5A). In addition, the numbers of trees and tree species at resting sites were also significantly larger than availability (Fig. 5B,C). At foraging routes, CWD, maximum DBH, basal area of coniferous trees and canopy cover were significantly larger than availability (Fig. 5A,D,E,F).
Fig. 4

Distribution of three categories of circular plots (radius=5 m) where we surveyed vegetation and topography. The plots for the survey of microhabitat availability were established at 300 m×200 m intervals. The plots for the survey of foraging routes (solid curves) were chosen at 300 m intervals after a resting site was located

Fig. 5

Comparison of frequency distribution between availability and use of microhabitats. K-S test detected significant difference for 7 of 11 attributes. In CWD, the availability was significantly different with the use at resting sites (RS) and foraging routes (FR). Index of preference was calculated, using Jacobs’s (1974) index

Discussion

The coexistence of residents and transients has been reported in pine marten (Chapin et al. 1997) and American marten (Paragi et al. 1996; Chapin et al. 1998) as well as the sable in the present study. Although these two types of life style have been found in both sexes, females are more transient than males (Phillips et al. 1998), probably since females may be limited in habitat choice because of denning requirements (Wynne and Sherburne 1984), or more specific prey requirements than males (Holmes and Powell 1994), and thus may be forced to abandon a home range rather than shift or expand their range (Phillips et al. 1998). Although Powell (1994) and Fuller et al. (2001) considered that the home ranges are hardly established by low-ranking male martens which are too young or too aged to win the intraspecific competition for resource-rich habitats, the present study indicates that the home range is established not only by old sables but also by young sables which have almost intact canine teeth usually possessed by sables younger than 2 years, suggesting that even young males have a chance to establish a home range in the sables.

The mean size of home range is 3.5–45.0 km2 in American marten (Raine 1982; Taylor and Abrey 1982; Phillips et al. 1998; Smith and Schaefer 2002) and 1.8–58.0 km2 in pine marten (Pulliainen 1984; Overskaug et al. 1994; Zalewski et al. 1995), which are much larger than 1.12 km2 of the sable in the present study. The size of terrestrial animal’s home range may be affected by many factors such as latitude (Gompper and Gittleman 1991), habitat quality (Lindstedt et al. 1986; Buskirk and Macdonald 1989) and population density (Wolff 1993). Although more data are required to determine the effects of these factors, the home range size of sables may vary across latitude, because Xu et al. (1997) reported that sables have a mean home range of 7.18 km2 in females and 13.03 km2 in males in Daxinganling Mountains (52° N, 123° E), northeastern China, that is situated at latitudinally 8° north of the present study site. While Daxinganling Mountains is located in taiga dominated by larch Larix gmelini Ledeb., the present study site is situated in a cool-temperate mixed forest inhabited by a high density of mice and voles which are main prey of sables. In addition, the thick snow cover may offer a suitable habitat to mice and voles in winter.

The present study suggests that sables do not show territoriality. However, Powell (1994) mentioned that martens often show a local variation of intraspecific interactions. For instance, American martens showed an intrasexual territoriality in northwestern Maine (Wynne and Sherburne 1984) and north central Ontario (Thompson and Colgan 1987) but not in northeastern Minnesota where the home ranges overlapped with each other extensively (Mech and Rogers 1977). These observations suggest that more data are required to determine whether the sable is territorial or not.

Whereas Paragi (1996) did not find any stand-scale habitat preference in American marten, other authors have reported that American martens prefer coniferous forest to other habitats (Koehler and Hornocker 1977; Soutiere 1979; Buskirk and Powell 1994), because coniferous forests are in the final stage of ecological succession in Canada and Alaska, and the martens effectively hunt prey or avoid predators in old-growth forests (Buskirk and Powell 1994). Sables do not clearly show any stand-scale habitat preference because coniferous stands constitute only 3% of the present study site. Instead, sables show patch-scale habitat preference. In Daxinganling Mountains, northeastern China, Buskirk et al. (1996) found that sables preferred foraging in the habitats with larger values of CWD, maximum DBH, basal area of coniferous trees and canopy cover, as also shown in the present study. According to Hale (1999), all of these vegetational factors characterize the old-growth forests, suggesting the high similarity of habitat preference between sable and American marten. In the habitats with well-covered canopies, sables may be able to avoid attacks by raptors such as white-tailed eagle Haliaeetus albicilla and mountain hawk eagle Spizaetus nipalensis, just like American martens which prefer foraging in a forest with well-covered canopy to avoid the attacks by raptors (Buskirk et al. 1989; Buskirk and Powell 1994). In the analysis of resting sites, Buskirk et al. (1996) found that sables prefer resting in debris-rich microhabitat, as also suggested in the present study. The debris-rich and dense-tree forests with a large number of tree species may be effective for the sables to avoid strong wind and predators (red foxes) which usually forage in open lands or sparse-tree forests. Buskirk et al. (1989) reported that American martens take a rest in debris-rich habitats to avoid their terrestrial predator (e.g., coyote).

Acknowledgements

We wish to thank M. Asano, Y. Iinuma, Y. Ikegami, T. Kubo, T. Kurumada, T. Murai, T. Murakami, M. Noda and K. Waseda for their helpful assistance and critical advice throughout this study. Cordial thanks are also due to M. Saitoh, H. Sugiyama and other staff members of Nakagawa Experiment Forest of Hokkaido University for their assistance and kind encouragement throughout the fieldwork.

Copyright information

© The Ecological Society of Japan 2004