OUP user menu

If we build habitat, will they come? Woody debris structures and conservation of forest mammals

Thomas P. Sullivan, Druscilla S. Sullivan, Pontus M. F. Lindgren, Douglas B. Ransome
DOI: http://dx.doi.org/10.1644/11-MAMM-A-250.1 1456-1468 First published online: 17 December 2012

Abstract

Coarse woody debris (CWD) from forest harvesting and salvage wood from wildfire and insect outbreaks provide habitat for an array of forest-floor small mammal species and some of their mammalian predators. Because forest clear-cutting reduces abundance of many mammal species, strategic management of postharvest debris could help maintain abundance and diversity of forest mammals on harvested sites. We tested hypotheses that abundance and species diversity of forest-floor small mammals would be lower on conventional clear-cuts than in uncut forest; abundance and species diversity of forest-floor small mammals and relative activity and species richness of winter mammals would be higher on clear-cut sites with woody debris arranged in large piles or windrows than a dispersed treatment of debris. Small mammals were intensively livetrapped, and winter mammals snow-tracked, from 2007 to 2009 in replicated (n = 3) woody debris treatments of dispersed, piles, windrows, and uncut mature forest at 3 study areas in south-central British Columbia, Canada. We captured all 9 species of forest-floor small mammals. Compared with uncut forest, clear-cutting had no effect on mean total abundance of the small-mammal community, and species richness and diversity were either similar or higher. With respect to habitat preference, generalist species increased while specialist species declined. Habitat structures of large piles and windrows of woody debris on clear-cuts dramatically ameliorated these responses. On the basis of track counts, relative activity and species richness of winter mammals were enhanced by these structures, but the response was species specific. This is the 1st investigation showing significant increases in abundance and species diversity of forest-floor small mammals associated with constructed piles and windrows of postharvest woody debris on clear-cuts. Large-scale CWD structures as piles or windrows have clear conservation implications for mammals in commercial forest landscapes.

Key words
  • forest-floor small mammals
  • forest harvesting
  • habitat structures
  • mammalian carnivores
  • piles and windrows
  • species richness and diversity
  • woody debris

Down wood or coarse woody debris (CWD) on the floor of deciduous and coniferous forests provides many important functions that are essential to maintaining biodiversity and long-term ecosystem productivity (Bunnell and Houde 2010; Harmon et al. 1986; McComb and Lindenmayer 1999). CWD provides reserves of nutrients and water (Laiho and Prescott 2004), microsites and substrates for seedlings (Harmon and Franklin 1989) and various saprobic and mycorrhizal fungi (Smith et al. 2000), and contributes to habitat quality for a wide range of mammal species (Maser et al. 1979; McComb 2003). Woody debris is created as a result of stand development and by natural (e.g., wildfire, insect outbreaks, wind events) and anthropogenic disturbances such as harvesting (Agee 1999; Spies et al. 1988). Woody debris is the residue (slash) occurring after conventional and salvage harvesting of forests (Hesselink 2010; Lindenmayer et al. 2008). Excess woody debris from felling operations is typically burned to reduce a perceived fire hazard.

Vertebrate use of CWD as habitat is primarily by small mammals, at least in the Pacific Northwest (PNW) of North America (Bunnell and Houde 2010; McComb 2003). On the forest floor, CWD provides opportunities for foraging, resting, reproduction, and thermal and security cover for many mammal species (Amaranthus et al. 1994; Carey and Johnson 1995; Maser et al. 1978; McComb 2003). Terrestrial small mammals are widespread across temperate and boreal forest ecosystems and may serve as ecological indicators of significant change in forest structure and function (Carey and Harrington 2001; Pearce and Venier 2005). These functions include prey for many predators (Martin 1994), consumers of plants, plant products (Carey et al. 1999), and invertebrates (Gunther et al. 1983), and dispersal of fungal spores (Maser et al. 1978).

Maintaining mammal diversity, as a component of biodiversity, is a major conservation goal in commercial forest landscapes and an array of harvesting systems that leaves residual live trees after harvest (green-tree retention [GTR]) has evolved (Franklin et al. 1997; Rosenvald and Lõhmus 2008). However, clear-cutting of forests remains the dominant silvicultural system in much of North America and northern Europe. The responses of small mammals inhabiting the forest floor are species specific, with generalists, which occupy a variety of habitats (e.g., deer mouse [Peromyscus maniculatus], northwestern chipmunk [Neotamias amoenus], voles in the genus Microtus), persisting on clear-cuts, although some for variable periods, and specialists, which require a specific habitat such as closed-canopy forest (e.g., southern red-backed vole [Myodes gapperi]), tending to disappear on clear-cuts (Fisher and Wilkinson 2005; Lehmkuhl et al. 1999; Sullivan et al. 2008; Zwolak 2009).

Abundances of mammalian carnivores are also reduced by clear-cutting, with loss of preferred prey species, den sites, and other components of forest stand structure (Fisher and Wilkinson 2005). In terms of mammalian biodiversity, carnivores such as coyotes {Canis latrans), red foxes (Vulpes vulpes), lynx (Lynx canadensis), cougars (Puma concolor), weasels (Mustela spp.), and American martens (Martes americana) also use CWD, particularly logs, as habitat for denning, nesting, and foraging (Bull 2002; McComb 2003). Weasels and martens seem to use incidental debris piles in reports reviewed by Bunnell and Houde (2010). Thus, strategic management of postharvest CWD could help maintain abundance and diversity of forest mammals on clear-cuts. Large-scale salvage harvesting of those forests influenced by wildfire and insect outbreaks typically create large (>100 ha) openings where habitat creation is much needed (Lindenmayer et al. 2008).

We know of no experimental manipulations of woody debris at a real-world scale that includes extremes in amounts and configurations of debris on clear-cuts to examine the responses of forest-floor small mammals and some of their predators. Thus, we tested hypotheses that abundance and species diversity of forest-floor small mammals would be lower on conventional clear-cuts than in uncut forest; and abundance and species diversity of forest-floor small mammals and relative activity and species richness of winter mammals would be higher on clear-cut sites with woody debris arranged in large piles or windrows than a dispersed treatment of debris.

Methods

Study areas.—Three study areas were located in south-central British Columbia, Canada: China Valley (50°44′N, 119°28;'W) 25 km west of Salmon Arm in the upper Interior Douglas fir (IDFdk; d,k = dry precipitation regime, cool temperature regime), Montane Spruce (MSdm; d,m = dry precipitation regime, mild temperature regime), and lower Interior Cedar Hemlock (ICHdk) biogeoclimatic zones; the Aberdeen Plateau (50°09′N, 119°12′W) 22 km southeast of Vernon in the upper IDFdk and MSdm subzones; and Summerland (49°40′N, 119°53′W) in the Bald Range 25 km west of Summerland in the upper IDFdk and MSdm subzones. Topography in all areas was rolling hills at 1,125 to 1,520 m elevation.

The upper IDF and MS have a cool, continental climate with cold winters and moderately short, warm summers. The average temperature is below 0°C for 2–5 months, and above 10°C for 2–5 months, with mean annual precipitation ranging from 30 to 90 cm. Open to closed mature forests of Douglas-fir (Pseudotsuga menziesii) cover much of the IDF zone, with even-aged postfire lodgepole pine (Pinus contorta var. latifolia) stands at higher elevations. In the MS landscape, hybrid interior spruce (Picea glauca × P. engelmannii) and subalpine fir (Abies lasiocarpa) are the dominant shade-tolerant climax trees. Trembling aspen (Populus tremuloides) is a common serai species and black cottonwood (Populus trichocarpa) occurs on some moist sites (Meidinger and Pojar 1991).

The ICH has an interior, continental climate with cool wet winters and warm dry summers. Mean annual temperature ranges from 2° C to 8.7° C. The temperature averages below 0° C for 2–5 months and above 10° C for 3–5 months of the year. Mean annual precipitation is 50–120 cm, 25–50% of which falls as snow. Upland coniferous forests dominate the ICH landscape and comprise the highest diversity of tree species of any zone in British Columbia. Western red cedar (Thuja plicata) and western hemlock (Tsuga heterophylla) dominate mature climax forests, with lodgepole pine, white spruce, Engelmann spruce, their hybrids, and subalpine fir common in these stands (Meidinger and Pojar 1991).

Before harvesting, all study stands were composed of a mixture of lodgepole pine with variable amounts of Douglas-fir, interior spruce, and subalpine fir. Average ages of lodgepole pine ranged from 80 to 120 years and for Douglas-fir and other conifers ranged from 120 to 220 years. Average tree heights ranged from 10.5 to 19.5 m for lodgepole pine and from 16.7 to 27.5 m for Douglas-fir and other conifer species.

Experimental design.—Each of the 3 study areas had a randomized complete block design (based on topography and geographic location) with 3 replicate sites each of: CWD dispersed uniformly over each site (control); CWD distributed into several piles (average of 2–3 piles/ha); CWD distributed into windrows; and uncut old-growth/mature forest (Fig. 1). The 12 sites (4 treatments × 3 replicates) at each study area were selected on the basis of operational scale, harvest sites that were the size of typical forestry operations, and reasonable proximity of sites to one another within a study area (Hurlbert 1984). All sites within a study area were far enough apart to be statistically independent: China Valley an average (± SE) of 0.58 ± 0.18 km (range 0.2–2.3 km); Aberdeen an average (± SE) of 0.58 ± 0.23 km (range 0.2–2.7 km); and Summerland an average (± SE) of 0.77 ± 0.29 km (range 0.2–3.0 km).

Fig. 1

Photographs of treatment sites at the Summerland study area in south-central British Columbia, Canada: a) dispersed coarse woody debris (CWD), b) piles of CWD, c) windrows of CWD, and d) uncut mature/old growth forest.

Woody debris treatments.—All timber harvesting was targeted at lodgepole pine salvage after, or impending, mountain pine beetle (MPB) attack at all 3 study areas. The harvesting system was clear-cut with reserves of Douglas-fir and some spruce as GTR units. Harvesting and subsequent CWD treatments were installed in October 2006 at the Summerland area where debris piles and windrows were created during processing of cut timber, followed by some specific site preparation work with a Caterpillar tractor. Debris piles and windrows averaged 2–3 m in height and 5–8 m in diameter or width. China Valley sites were harvested in 2005 and 2006. Debris piles and windrows were installed at these sites using an excavator and backhoe as site preparation treatments during September–October 2006. Debris piles and windrows were smaller than those at Summerland, averaging 1–2 m in height and 2–3 m in diameter, and generally were installed in the 2nd year after harvesting. At Aberdeen, sites were harvested in January–February 2007 with installation of treatments in March 2007. Completion of site preparation work was done in August 2007. However, our sampling regime commenced in May–-June 2007, with accommodation of the site preparation work, and hence was concurrent with the other 2 study areas. Size of debris piles and windrows at Aberdeen was similar to the China Valley study area.

In visual estimates of the various piles and windrows, wood (butts, tops, and unmerchantable logs from 10- to 50-cm diameter) ranged from 70% to 80% of debris loads. Branches and fines made up 20% to 30% (<10-cm-diameter pieces). This range varied on those sites with much cull wood (e.g., MPB and other mortality agents) to 90–100% of debris loads. Volumes of downed wood in the dispersed treatments were measured using the line-intersect method of Van Wagner (1968) and in the piles and windrow treatments by the method of Hardy (1996).

Forest-floor small mammals.—Forest-floor small mammals were sampled at 4-week intervals from May to October 2007, 2008, and 2009 at Summerland, and at 4-week (2007) and 8-week (2008–2009) intervals at China Valley and Aberdeen. One livetrapping grid (1 ha) was located in each of the 12 sites, at each study area, and had 49 (7 × 7) trap stations at 14.3-m intervals with 1 or 2 Longworth live traps at each station. Number of traps at a station was increased when trap occupancy was greater than 60%. Traps were supplied with whole oats, a slice of carrot, and cotton as bedding. Each trap had a 30-cm ×30-cm plywood cover for protection from sunlight (heat) and precipitation. Traps were set on the afternoon of day 1, checked on the morning and afternoon of day 2 and morning of day 3, and then locked open between trapping periods. Forest-floor small mammal species sampled by this procedure included the southern red-backed vole, long-tailed vole (Microtus longicaudus), meadow vole (M. pennsylvanicus), heather vole (Phenacomys intermedius), deer mouse, northwestern chipmunk, montane shrew (Sorex monticolus), masked shrew (S. cinereus), and short-tailed weasel (Mustela erminea). All animals captured (except shrews and weasels) were ear-tagged with serially numbered tags and point of capture recorded. Animals were released on the grids immediately after processing (Krebs et al. 1969). The overnight trapping technique resulted in a high mortality rate for shrews. Therefore, shrews were collected, frozen, and later identified according to tooth patterns (Nagorsen 1996). All handling of animals followed guidelines approved by the American Society of Mammalogists (Sikes et al. 2011) and the Animal Care Committee, University of British Columbia.

Abundance estimates were derived from the Jolly–Seber (J-S) stochastic model for open populations (Jolly and Dickson 1983; Seber 1982), with small sample size corrections (Seber 1982). The J-S model assumes that marked and unmarked animals have the same capture probability in each sampling session (Krebs 1999). Because our traps were locked open between trap sessions, we expected that trap responses of animals dissipated between our monthly sampling sessions. Thus, the J-S model was used to estimate populations of the major species. The reliability of the J-S model declines when population sizes are very low and no marked animals are captured (Krebs 1999). Number of individual animals captured was used as the population estimate for the first and last sampling weeks, and for the enumeration of the relatively less common heather voles, shrews, and weasels. There were 3 summer (May–September) and 3 winter (October-April) periods.

Winter tracking of mammals.–Track counts assume that the importance or attractiveness (or both) of particular habitats or features can be inferred by the relative frequency of animal occurrence at, or near, them. Animal track crossings were counted along permanent transects that maximized the length of transect within a treatment site and were located in a manner that avoided edge effects. Thus, their location maximized distance to site edge so that sampled tracks best reflected animal activity within the treatment site rather than simply capturing activity associated with the edge. There was 1 continuous transect per site and the strategy for locating transects was the same for all sites at all 3 study areas. Variability in size (hectares) of treatment sites resulted in different lengths of transects. Thus, mean (± SE) transect lengths, averaged across all 12 sites at each study area, were 417 ± 37, 567 ± 40, and 756 ± 46 m at Aberdeen, China Valley, and Summerland, respectively. Data were standardized by dividing the number tallied by the product of transect distance× elapsed time since last snowfall (usually expressed as tracks/km-day), although the time unit varied from half-days (12-h periods) to a week in the studies cited. Snow-tracking was conducted within the 12 treatment sites at each study area following methodology for small- and medium-sized carnivores and ungulates (Bowman and Robitaille 1997; D'Eon et al. 2006; Thompson et al. 1989). Therefore, our snow-tracking methodology was used to document all mammal activity, including coyotes, red foxes, wolves (Canis lupus), weasels, American martens, lynx, bobcats (L. rufus), cougars, red squirrels (Tamiasciurus hudsonicus), snowshoe hares (Lepus americanus), mule deer (Odocoileus hemionus), and moose (Alces alces), and is best described as a mammalian biodiversity sample.

Timing of samples depended on snow conditions. We attempted to sample 3 days after a fresh snowfall as recommended. Two sampling sessions (early and late winter) were conducted at each study area in each of 3 winters (2007-2008, 2008–2009, and 2009–2010). All 12 sites at a given study area were sampled within the same week. However, different weather conditions experienced at our 3 widely separated study areas resulted in different timing of samples. All animal tracks were identified continuously along a given transect length. Time since the last snowfall was accounted for by dividing the number of tracks by the number of 12- or 24-h periods since the last snowfall. We summarized track count data by calculating the number of tracks per km per 12-h period for each of the 12 sites at each study area for each of the 3 winters. These data were used to compare relative activity (habitat use) among the 12 treatment sites at a given study area each year. There was insufficient snow for tracking in the second sampling session at China Valley in 2010.

Diversity measurements.—Species richness was the total number of species sampled for the mammal communities in each site (Krebs 1999). Species diversity was based on the Shannon–Wiener index (Burton et al. 1992; Magurran 2004). Diversity of small mammals was calculated using the estimated abundance of each species for a given sampling period and averaged over the number of sampling periods for each year. Diversity of mammals in the winter tracking was based on relative activity (abundance of tracks) of each species and calculated in an identical manner.

Statistical analysis.—A one-way analysis of variance (ANOVA; SPSS 17.0, SPSS Institute Inc., 2007) was conducted to detect differences in mean volumes of woody debris in treatments of dispersed, piles, and windrows at the 3 study areas. A repeated-measures analysis of variance (RM-ANOVA), on the basis of mean annual samples of small-mammal populations per grid or mean annual counts of snow tracks per treatment site, was used to determine the effect of woody debris treatments and time (years 2007–2009) on mean total abundance, mean species richness, and mean species diversity of the forest-floor small-mammal community, as well as the mean abundance of each major small-mammal species. This same analysis was conducted on mean relative abundance of winter track counts for each major carnivore species and for species richness of this winter community of mammals. RM-ANOVAs were conducted separately for each study area because an initial RM-ANOVA of the randomized block design (i.e., study areas as blocks) indicated that there was a significant block × treatment interaction for mean total abundance and mean abundance of 3 of 4 of the major small-mammal species. In addition, there were dramatic differences in the dimensions of CWD structures between Summerland and those at China Valley and Aberdeen. Count data were transformed using square-root or log (base 10) transformations and attributes based on percentage or proportion data with arcsinh (inverse hyperbolic sine) to better approximate homogeneity of variance as measured by the Levene statistic (Fowler et al. 1998). Mauchly's W test statistic was used to test for sphericity (independence of data among repeated measures; Kuehl 1994; Littel 1989). For data found to be correlated among years, the Huynh–Feldt correction was used to adjust the degrees of freedom of the within-subjects F-ratio (Huynh and Feldt 1976). Logarithmic (In based) regression analyses were conducted to determine the relationship and upper limits between mean total abundance, species richness, and species diversity of small mammals with volume of woody debris/ha (dispersed, piles, or windrows; Zar 1999). Overall means (n = 9; 3 replicates × 3 years) ± 95% confidence intervals were calculated for total abundance,

species richness, and species diversity of forest-floor small mammals. Duncan's multiple range test (DMRT) was used to compare mean values based on RM-ANOVA results. For all analyses, P = 0.05.

Results

Woody debris treatments.—The dispersed treatments were reasonably similar (F2,6 =2.13; P = 0.20) in mean volume (range of 135 to 213 m3/ha), as would be expected after clear-cut harvesting and dispersion of debris over the area of each cutblock. Mean volume of dispersed CWD in the forest treatments was also similar (F2,6 = 0.92; P = 0.45), with a range of 125 to 180 m3/ha. The mean volume of CWD/ha in piles (F2,6 = 10.98; P = 0.01) was significantly greater at Summerland than at the other 2 study areas (Fig. 2a). Mean number of piles/ha was also highest at Summerland (F2,6 = 5.50; P = 0.04). Mean volume of CWD/ha in windrows (F2,6 = 6.57; P = 0.03) was significantly greater at Summerland than at the other 2 areas, and mean number of windrows/ha followed this pattern but was not formally significant (F2,6 = 4.21; P = 0.07; Fig. 2b).

Fig. 2

Mean volume (m3) of CWD/ha in a) piles and b) windrows at the China Valley, Aberdeen, and Summerland study areas. Histograms with different letters are significantly different by Duncan's multiple range test.

Forest-floor small mammals.—There were significant block (study area)× treatment interactions for mean total abundance (F6,24 = 4.09; P < 0.01) of forest-floor small mammals and mean abundance of 3 major species: deer mice (F6,24 = 5.28; P < 0.01), red-backed voles (F6,24 = 11.44; P < 0.01), and long-tailed voles (F6,24 = 7.55; P < 0.01), but not for northwestern chipmunks (F6,24 = 0.74; P = 0.62). Thus, on the basis of the results of this analysis and the significant differences in amounts of woody debris among study areas, analyses were conducted separately for each study area.

A total of 9 species of forest-floor small mammals, composed of 5,209 individuals, was captured. The proportions of individuals were 1,663 (31.9%) at China Valley, 1,217 (23.4%) at Aberdeen, and 2,329 (44.7%) at Summerland (Table 1). Mean total abundance of small mammals was similar among sites at each of the China Valley (F3,6 = 3.14; P =0.11) and Aberdeen (F3,6 = 0.95; P = 0.47) study areas (Figs. 3a and 3b). However, mean total abundance of small mammals was significantly (F3,6 = 34.16; P < 0.01) different among sites (DMRT; P = 0.05) at Summerland, with overall numbers 1.8 to 2.3 times higher in the piles and windrows than the dispersed CWD and forest treatments (Fig. 3c). Mean species richness and species diversity of small mammals also followed this pattern with similar (P > 0.05) values among treatment sites at China Valley and Aberdeen (Figs. 4 and 5). Again, mean species richness (F3,6 = 30.91; P < 0.01) and diversity (F3,6 = 44.31; P < 0.01) at Summerland were highest in the piles and windrows, followed by the dispersed and then forest treatments (Figs. 4 and 5). There were no significant effects of time or treatment × time interactions in these analyses.

View this table:
Table 1

Total number of individuals captured for the 9 species of forest-floor small mammals at each of the 3 study areas.

Study area
SpeciesChina ValleyAberdeenSummerlandTotals
Myodes gapperi3585445921,494
Peromyscus maniculatus5761207071,403
Neotamias amoenus3943744341,202
Microtus longicaudus4980454583
Sorex monticolus1636178302
Microtus pennsylvanicus74208102
Sorex cinereus2041438
Phenacomys intermedius172836
Mustela erminea2871449
Totals1,6631,2172,3295,209
Fig. 3

Overall mean (n = 9; 3 replicate sites × 3 years) ± 95% confidence interval (CI) total abundance of forest-floor small mammals in the dispersed, piles, windrows, and forest treatments at the China Valley, Aberdeen, and Summerland study areas in 2007–2009. Mean values with different letters are significantly different by Duncan's multiple range test (adjusted for multiple contrasts).

Fig. 4

Overall mean (n = 9; 3 replicate sites × 3 years) ± 95% confidence interval (CI) species richness of forest-floor small mammals in the dispersed, piles, windrows, and forest treatments at the China Valley, Aberdeen, and Summerland study areas in 2007–2009. Mean values with different letters are significantly different by Duncan's multiple range test (adjusted for multiple contrasts).

Fig. 5

Overall mean (n = 9; 3 replicate sites × 3 years) ± 95% confidence interval (CI) species diversity of forest-floor small mammals in the dispersed, piles, windrows, and forest treatments at the China Valley, Aberdeen, and Summerland study areas in 2007–2009. Mean values with different letters are significantly different by Duncan's multiple range test (adjusted for multiple contrasts).

Mean abundance of deer mice was significantly (F3, 6 = 5.73; P = 0.03) different among sites at China Valley, with the dispersed, piles, and windrow treatments 4.0 to 6.3 times higher (DMRT; P = 0.05) than the forest (Table 2). Mean abundance of deer mice was similar (P > 0.05) among sites at Aberdeen and Summerland. Mean abundance of red-backed voles was significantly (F3,6 = 138.21; P < 0.01) different among sites at China Valley, with the highest (DMRT; P = 0.05) numbers of this microtine in the forest and lowest in the dispersed treatment (Table 2). Conversely, mean abundance of this closed-canopy specialist was also significantly different among sites at Summerland (F3,6 = 142.28; P < 0.01), but with the highest (DMRT; P = 0.05) numbers in the windrows, followed by piles and forest sites, and the lowest in the dispersed treatment (Table 2). Red-backed voles at Aberdeen also tended to follow this pattern but it was not statistically significant.

View this table:
Table 2

Overall mean (n = 3 replicate sites) ± SE total abundance of each species of forest-floor small mammals in each of the 4 treatment sites at the 3 study areas and repeated-measures analysis of variance (RM-ANOVA) results. Columns with different letters are significantly different by Duncan's multiple range test (adjusted for multiple contrasts). Bold P-values indicate statistical significance.

RM-ANOVA
SiteTreatment
Species and study areaDispersedPilesWindrowsForestF3,6P
Peromyscus maniculatus
China Valley9.62a ± 1.497.49a ± 1.0611.83a ± 1.601.87b ± 0.665.730.03
Aberdeen0.88 ± 0.321.21 ± 0.391.33 ± 0.321.14 ± 0.251.390.33
Summerland9.39 ± 1.247.07 ± 0.795.36 ± 0.977.35 ± 1.252.180.19
Myodes gapperi
China Valley0.79b ± 0.450.61b ± 0.310.24b ± 0.2014.17a ± 1.59138.21<0.01
Aberdeen3.23 ± 1.016.20 ± 1.237.20 ± 1.3410.76 ± 2.493.540.09
Summerland1.07c ± 0.287.23b ± 1.5312.41a ± 1.676.80b ±1.16142.28<0.01
Neotamias amoenus
China Valley10.17a ± 1.689.92a ± 1.2512.15a ± 1.102.58b ± 0.625.380.04
Aberdeen9.22a ± 0.688.32a ± 1.1010.08a ± 0.923.70b ± 0.787.600.02
Summerland8.51a ± 0.9614.90a ± 1.3713.17a ± 1.512.36b ± 0.4634.73<0.01
Microtus longicaudus
China Valley0.38 ± 0.140.42 ± 0.280.26 ± 0.100.09 ± 0.090.730.57
Aberdeen0.44 ± 0.161.07 ± 0.371.70 ± 0.420.07 ± 0.073.900.07
Summerland1.07b ± 0.328.64a ± 1.696.94a ± 1.210.00c ± 0.0046.12<0.01
Sorex spp.
China Valley1.51 ± 0.501.68 ± 0.381.35 ± 0.390.22 ± 0.102.160.19
Aberdeen0.49 ± 0.170.65 ± 0.280.65 ± 0.300.39 ± 0.170.780.55
Summerland0.24ab ± 0.070.19ab ± 0.050.54a ±0.180.00b ± 0.005.390.04
Mustela erminea
China Valley0.14 ± 0.050.27 ± 0.120.41 ± 0.150.00 ± 0.003.900.07
Aberdeen0.00 ± 0.000.02 ± 0.020.02 ± 0.020.14 ± 0.092.200.19
Summerland0.00b ± 0.000.08ab ± 0.030.13a ± 0.050.06ab ± 0.035.000.05

Mean abundance of northwestern chipmunks was significantly (P ≤ 0.04) different among sites at all 3 study areas, with more chipmunks (DMRT; P = 0.05) in the CWD treatments than forest (Table 2). Across all comparisons, chipmunks were 2.2 to 6.3 times higher in the CWD treatments than forest. The long-tailed vole was common at Summerland, but not at the other 2 study areas, with significantly (F3,6 = 46.12; P < 0.01) higher (DMRT; P = 0.05) mean numbers in the piles and windrows than in the dispersed and forest sites (Table 2). Mean abundance of the 2 shrew species, pooled for this analysis, was similar (P > 0.05) among sites at China Valley and Aberdeen, but significantly (F3,6 = 5.39; P = 0.04) different at Summerland. Shrew numbers were higher (DMRT; P = 0.05) in the windrows than in forest, similar in all CWD treatments, and also similar in the dispersed, piles, and forest (Table 2). Mean number of weasel captures was similar (P > 0.05) across sites at China Valley and Aberdeen (Table 2), but again, at Summerland mean (± SE) abundance of weasels was significantly (F3,6 = 5.00; P - 0.05) different among sites, with 0.13 ± 0.05 in windrows, 0.08± 0.03 in piles, 0.06 ± 0.03 in forest, and no captures in the dispersed treatment. There were no significant time × treatment interactions in these analyses of individual species' responses.

There was a positive relationship (r = 0.65; P < 0.01) between mean abundance of small mammals and mean volume of CWD/ha across the dispersed, piles, and windrow treatments at the 3 study areas (Fig. 6a). We found similar relationships for mean species richness (r = 0.59; P < 0.01) and mean species diversity (r = 0.61; P < 0.01) with mean volume of CWD/ha (Fig. 6b and c). A range of 300–400 m3/ha of CWD in piles or windrows seemed required as habitat for small mammals on these sites.

Fig. 6

Regression relationship of mean a) total abundance of forest-floor small mammals/ha, b) species richness, and c) species diversity, with mean volume (m3) of woody debris/ha in dispersed, piles, and windrows at the three study areas. Data points for China Valley (▲), Aberdeen (■), and Summerland (●).

Winter tracking of mammals.—Mean species richness of winter mammals, as indicated by track counts, was similar (P > 0.05) among treatments at both China Valley and Aberdeen. However, there was a significant (F6,16 = 3.49; P = 0.02) treatment × time interaction, with a higher number of species in the forest than in other treatment sites at Aberdeen in 2007. Mean species richness of tracks at Summerland was significantly (F3,6 = 10.25; P < 0.01) different among sites, with highest (DMRT; P = 0.05) levels in the forest and piles, similar numbers of species in the piles and windrows, and lowest levels in the dispersed sites (Fig. 7a). In terms of a significant (P < 0.05) change through time, mean species richness increased from 2007 to 2009 at both Aberdeen and Summerland.

Fig. 7

Overall mean (n = 9; 3 replicate sites × 3 years) ± 95% confidence interval (CY). a) Species richness of track counts of winter mammals, and number of b) weasel and c) coyote tracks (per kilometer per 12-h period) in the dispersed, piles, windrows, and forest treatments at the Summerland study area in 2007–2009. Mean values with different letters are significantly different by Duncan's multiple range test (adjusted for multiple contrasts).

Weasels and coyotes were the 2 most common carnivores encountered during winter track counts. Mean track counts of weasels and coyotes were similar (P > 0.05) among sites at both China Valley and Aberdeen. At Summerland there was a significant difference among sites for weasels (F3,6 = 9.62; P = 0.01), with highest (DMRT; P = 0.05) numbers of tracks in the piles and windrows than in forest, in the piles than in dispersed, with similar levels in the dispersed and windrow sites (Fig. 7b). Mean track counts of coyotes also followed this pattern at Summerland, but the difference was not formally significant (F3,6 = 3.73; P = 0.08; Fig. 7c).

Discussion

This study is the 1st investigation of the responses of mammals to constructed piles and windrows of postharvest woody debris compared with the conventional dispersed debris on clear-cuts. Our first hypothesis, that abundance and diversity responses of forest-floor small mammals would be lower on conventional clear-cuts compared with uncut forest, was not supported, because the response varied among animal species. When comparing the forest and dispersed CWD on conventional clear-cut sites at China Valley and Aberdeen, complete removal of the forest cover by clear-cutting had no effect on mean total abundance, species richness, or species diversity of the small-mammal community. This pattern also occurred at Summerland for abundance, but not for richness and diversity, which were significantly higher in the dispersed CWD than forest sites. This latter result was also reported for 1–4 years postharvest by Sullivan and Sullivan (2001), but not 5–8 years after clear-cutting (Sullivan et al. 2008). As reported by several authors (Fisher and Wilkinson 2005; Lehmkuhl et al. 1999; Zwolak 2009), generalists such as the deer mouse, northwestern chipmunk, voles, and shrews persist on clear-cuts, sometimes for variable periods, whereas specialists such as the southern red-backed vole tend to disappear. Higher levels of species richness and diversity on clear-cuts are represented by these generalist species. Our results support this pattern of species-specific responses to conventional clear-cutting. However, the advent of constructed piles and windrows of woody debris on clear-cut sites, at an appropriate scale and timing since harvest at the Summerland study area, yielded novel results for maintenance of both generalist and specialist small mammals.

Our 2nd hypothesis, that, compared with a dispersed (conventional) treatment of woody debris on clear-cut sites, debris arranged in large piles or windrows at a real-world scale would have a higher abundance and species diversity of forest-floor small mammals, seemed to be supported. The strongest results were recorded at Summerland, where debris structures were generally 5 to 8 times larger than those at the China Valley and Aberdeen areas. This difference in scale of debris structures, and timing of their creation postharvest, was most clearly seen in the response of red-backed voles. There were few red-backed voles in the CWD treatments at China Valley that were constructed after this species had essentially disappeared from the clear-cut sites. Red-backed vole popultions decline dramatically on clear-cuts in coniferous and mixed coniferous–deciduous forests (Klenner and Sullivan 2003; Moses and Boutin 2001; Sullivan et al. 1999; Zwolak 2009). At Aberdeen and Summerland, mean abundance of red-backed vole populations was consistently higher in the windrows than in other sites, whereas red-backed voles in the piles and in forest were similar. Thus, this closed-canopy specialist was maintained on clear-cuts for at least 3 years postharvest and contributed to the higher overall abundance and species richness–diversity of the forest-floor small-mammal community in the piles and windrows.

The lack of response of deer mice to our piles and windrows was similar to other studies where there were no relationships of this ubiquitous species to dispersed CWD or logs on the forest floor (Craig et al. 2006; Hayes and Cross 1987; Smith and Maguire 2004). Alternative results in sites of natural CWD accumulations found more deer mice (Steel et al. 1999) and cotton mice (Peromyscus gossypinusLoeb 1999) in debris piles than at reference sites. The similar responses of northwest chipmunk and shrew populations across our CWD treatments followed other cited works where these species use CWD as travel paths but have only weak or no population response (Dueser and Shugart 1978; Loeb 1999; McCay and Komoroski 2004; Waldien et al. 2006).

Pauli et al. (2006), working in a range of sites of extensive natural blowdown of mature coniferous forest, and Loeb (1999), investigating tornado blowdown in pine forest, reported that abundance of small mammals was higher in those sites most affected by blowdown and consequent high amounts of CWD. A similar result in a smaller-scale blowdown study was reported by Powell and Brooks (1981). Loeb (1999) also reported that species composition of small mammals was similar in blowdown and control sites, whereas Pauli et al. (2006) found the lowest species diversity in the sites with extensive blowdown. Riverine debris piles in a riparian corridor had a greater abundance and species richness of small mammals than nearby reference sites without piles (Steel et al. 1999). Maguire (2002) reported a positive relationship between species richness of small mammals and volume of CWD. Again, these contradictory results suggest strongly that large-scale experimental manipulations of CWD, as done in our study, are needed to generate unequivocal results in terms of responses in abundance and diversity of forest-floor small mammals.

Our 3rd hypothesis, that relative activity of members of a winter community of mammals would also be increased by creation of piles and windrows of debris, was supported in part, but was also species specific. Although overall species richness was higher in piles and windrows than in dispersed CWD at Summerland, only relative abundance of coyotes and weasels followed this pattern. Other predators, including lynx, foxes, cougars, wolves, and martens, were present in the general area, but did not selectively choose among our CWD structures. The other species encountered were either herbivores (hares, deer, moose) or omnivores (red squirrels) and primarily were recorded in the forest with occasional forays out into the openings with their respective CWD structures.

The benefits of CWD to forest-floor small mammals also seemed positive for mustelids and other predators. Weasels were recorded in piles and windrows more than in dispersed treatments at the Summerland study area, and a tendency overall (1.9 times higher) at the 3 study areas. Predators, particularly mustelids (Mustela spp. and Martes americana), are known to forage and select paths near downed wood (Buskirk and Zielinski 2003; Simms 1979). These small carnivores use logs and debris piles as den and resting sites (Bunnell and Houde 2010; Buskirk et al. 1989). Piles with an “open” circumference, compared with the linear nature of windrows, may make small mammals more susceptible to predation, particularly by weasels (Mustela spp.—Lisgo et al. 2002). However, weasels and martens are prey species themselves and they may seek out windrows as “safe” corridors to cross openings.

The size of our experimental units were all typical forestry operations (range of means from 4.6 to 10.1 ha) and in reasonable, but independent, proximity to each other at a given study area. Ideally, all CWD treatments would have been done in the same manner and year of harvest. However, this was not possible because of logistical and methodological constraints of operating in 3 different forest company jurisdictions. Results from the 3 study areas suggest strongly that our inferences are applicable to a wide range of clear-cut openings among forest ecosystems in the south-central part of British Columbia.

It could be argued that the less frequent sampling regime at China Valley and Aberdeen in 2008–2009 may have affected our abundance estimates of each species and measurements of species richness and diversity, but we saw no evidence of bias. Our results persisted for 3 years postharvest, but they should still be viewed with caution as small-mammal populations and communities will likely fluctuate in these CWD habitats and ecosystems over time. Continued monitoring is required to determine long-term responses to these novel structures. In general, structures composed of debris made up of large wood (butts, tops, and unmerchantable logs) will very likely last decades compared with those with a large component of branches and fines. Sites with much cull wood (e.g., MPB, fires, and other mortality agents) would likely provide longer-lasting habitat structures than those where merchantable material is high and branches and fines make up >50% of debris loads. The winter track counts had just 2 sampling sessions per year and this was a minimum effort to determine relative abundances of these species. Although dictated by suitable snow-tracking conditions and access to study sites, additional samples each year would have been ideal.

As reviewed by Bunnell and Houde (2010), relationships between dispersed CWD volumes and abundance of small mammals have been weak and variable (Bowman et al. 2000; Bunnell et al. 1999; Carey and Johnson 1995; Corn and Bury 1991; West 1991). As evidenced by our significant differences in the relative sizes and dimensions of piles and windrows created at the 3 study areas, real-world tests of the relationship of CWD structures and mammals is very scale dependent, as are various patterns of managing for biodiversity (Bunnell and Huggard 1999). Thus, it is not surprising that few unambiguous results have been obtained, to date. Craig (2002) and Craig et al. (2006) did experimental manipulations of debris, but at a small scale, and consequently found no differences in small-mammal responses to a range of dispersed CWD. Other dispersed CWD studies have been incidental or anecdotal in evaluating the utility of such structures to forest-floor small mammals (Bunnell and Houde 2010).

CWD structures as piles or windrows have clear conservation implications for the native forest-floor small-mammal community. We captured all 9 common species and the persistence of the red-backed vole was particularly important, as this species disappears from clear-cuts and does not reappear in regenerated stands for decades (Ransome et al. 2009; St-Laurent et al. 2008; Sullivan et al. 2011). The red-backed vole is a principal prey for marten, which also is uncommon on recently logged sites, and is a species of concern in Canada and other parts of North America (Thompson 1991).

Piles or windrows at least 2 m in height and 5 m in width or diameter provide habitat for forest-floor small mammals on clear-cuts. These structures need to be created at the time of forest harvesting and log processing to provide sufficient volume of woody material for at least 300 m3/ha, and to minimize both operational costs and excessive movement of heavy machinery over the forest floor. Depending on operational conditions and availability of woody debris, at least 1 windrow or a series of piles should connect patches of mature forest and riparian areas to allow small mammals and other species to access and traverse clear-cut openings. This will be particularly important on conventional-sized openings (5-20 ha), but also on much larger (>100 ha) salvage harvesting operations in beetle-killed as well as burned forests, at least in the PNW of North America. The fate of down wood from harvesting operations remains unclear as excess woody debris is currently burned on-site for fire hazard reduction or removed for bioenergy production. However, our results suggest strongly that if we build large-scale CWD habitat structures as piles or windrows, they will have clear conservation implications for mammals in commercial forest landscapes.

Acknowledgments

We thank the Forest Science Program (Forest Investment Account), BC Ministry of Forests and Range, the Okanagan Innovative Forestry Society (Innovative Forest Practices Association), Gorman Bros. Lumber Ltd., Tolko Industries Ltd., and Federated Coop for financial and logistical support. We thank J. Konken, T. Seebacher, H. Sullivan, and B. Tilling for assistance with the fieldwork.

Footnotes

  • Associate Editor was Lisa A. Shipley.

Literature Cited

View Abstract