Do riparian buffer strips mitigate the impacts of ...

Viewer
Transcript

Biological Conservation 113 (2003) 133–140 www.elsevier.com/locate/biocon

Do riparian buﬀer strips mitigate the impacts of clearcutting on small mammals? Kristina L. Cockle*, John S. Richardson Department of Forest Sciences, 3041-2424 Main Mall, The University of British Columbia, Vancouver, BC, Canada, V6T 1Z4 Received 1 May 2002; received in revised form 15 September 2002; accepted 20 October 2002

Abstract We assessed the impact of clearcutting on small mammals in riparian areas and evaluated riparian buﬀer strips as a tool for conserving small mammals in managed forests. Over two summers, we trapped small mammals of seven species in riparian areas in southwestern British Columbia, Canada. Communities of small mammals were compared across three diﬀerent habitat types: (1) clearcut to the stream bank, (2) clearcut with a 30 m riparian buﬀer strip, and (3) control (no logging). Species richness was signiﬁcantly lower in clearcuts than in controls and buﬀers. On clearcut sites, creeping voles were more abundant, but red-backed voles and dusky shrews were less abundant than at the control sites. At sites with riparian buﬀer strips, both voles were present in numbers similar to those found in controls, but dusky shrews were less common. Signiﬁcantly more deer mice and creeping voles were infested with bot ﬂies at clearcut sites than at buﬀer sites, and no animals were infested at any of the control sites. Riparian reserves appear to be useful in reducing the short-term impacts of clearcutting on small mammal communities, though they do not eliminate these impacts altogether. # 2003 Elsevier Science Ltd. All rights reserved. Keywords: Forestry; Insectivores; Riparian zones; Populations; Rodents

1. Introduction Clearcutting remains the predominant method of forest harvesting in many parts of the world, despite concerns about wildlife, biodiversity, and long-term eﬀects on forest ecosystems. Forest harvesting has negative impacts on some species, and positive or neutral eﬀects on others (e.g. Laurance, 1990; deMaynadier and Hunter, 1995; Duguay et al., 2000; Payer and Harrison, 2000; de Bellefeuille et al., 2001). As forested lands are rapidly being converted to managed forests, there is concern for those species negatively aﬀected by harvesting. Governments commonly respond to this concern by introducing requirements for small habitat reserves, designed to reduce the impacts of timber extraction on wildlife. Riparian areas within forests are generally cooler, wetter, more structurally complex, and more productive * Corresponding author. Present address: Department of Biology, Dalhousie University, Halifax, Nova Scotia, Canada, B3H 4J1. E-mail addresses: [email protected] (K.L. Cockle), jrichard@ interchg.ubc.ca (J.S. Richardson).

than upland areas, so they are often home to a distinct community of plants and animals (e.g. Bilby, 1988; McComb et al., 1993b; Kelsey and West, 1998). They provide important habitat for many aquatic and terrestrial species, some of which are riparian obligates. Riparian buﬀer strips are retained primarily for the purpose of protecting aquatic ecosystems, and particularly ﬁsh, but policy-makers claim that they also help to conserve terrestrial wildlife by providing habitat and corridors in fragmented landscapes (Forest Ecosystem Management Assessment Team, 1993). Despite these claims, very little information is actually available on the use of riparian buﬀer strips by terrestrial species (Hagar, 1999; Darveau et al., 2001; Vesely and McComb, 2002). Buﬀer strips are long and narrow, resulting in the absence of interior forest, and an abundance of edge habitat. Given the widespread use of riparian buﬀer strips to conserve biodiversity in managed forests, it is important to make sure that these buffer strips are indeed used by terrestrial wildlife. Small mammals, such as shrews, moles, voles, and mice, perform several important roles in the forest ecosystem. They are the primary prey for many carnivorous

0006-3207/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0006-3207(02)00357-9

134

K.L. Cockle, J.S. Richardson / Biological Conservation 113 (2003) 133–140

mammals, snakes, and birds (Cross, 1988; Kelsey and West, 1998), some help to disperse mycorrhizal fungi (Terwilliger and Pastor, 1999), and some aﬀect plant species composition and soil fertility through selective herbivory and seed dispersal (Sirotnak, 2000). Many small mammal species are associated with, and even dependent upon riparian areas (e.g. Bamﬁeld, 1974; Cross, 1985; Gomez and Anthony, 1998), so the protection of riparian habitat may be particularly important for conserving small mammals in managed forests. Studies in upland areas show that logging often leads to changes in the abundance of some species of small mammals (Tevis, 1956; Gashwiler, 1970; Hooven, 1973; Hooven and Black, 1976; Martell and Radvanyi, 1977; Kirkland, 1990; McComb et al., 1993a), but does not aﬀect overall species richness of small mammals (Kirkland, 1990). While many studies have looked at the eﬀects of forest harvesting on wildlife, the majority have been conducted in upland forests rather than riparian areas. Studies in western Oregon have demonstrated that riparian areas are important habitats for some small mammals, and may be critical to the conservation of these species (Cross, 1985; Anthony et al., 1987). Riparian reserve zones may function to provide for this wildlife conservation objective (Cross, 1985; Darveau et al., 2001). Despite the widespread acceptance of riparian reserve zones as a measure to protect streamsides, few studies have tested whether these riparian buﬀers are eﬀective in providing either habitat or corridors for small mammals. Our study was designed to test the hypothesis that ﬁxed-width riparian reserves can maintain population sizes of small mammals in managed forests. First, we wished to assess the impact of clearcutting of mature, second-growth forest on small mammals in riparian areas. Secondly, we wanted to determine whether retaining riparian buﬀer strips within clearcuts could mitigate the impact of logging. We measured the impact of clearcutting by considering the number of animals we captured, the species richness of the small mammal community, the survival rate, the average body mass, and the rate of parasite infestation among small mammals. We compared these variables between three ‘‘treatments’’: recently clearcut riparian areas, riparian areas with 30 m buﬀer strips, and riparian areas in intact second growth forest.

2. Materials and methods 2.1. Study area This study was conducted from July to September 1999, and from May to August 2000, in the Malcolm Knapp Research Forest, near Maple Ridge, British Columbia (49 160 N, 122 340 W). Eight study sites were

chosen, all in riparian areas. All of the sites had a common stand history. They were harvested in the early 1900s, and were all naturally regenerated following a stand-initiating wildﬁre in 1931. Thus, prior to harvesting for our study, all sites were 68-year-old second-growth forest, made up of Western hemlock (Tsuga heterophylla), Douglas-ﬁr (Pseudotsuga menziesii) and Western redcedar (Thuja plicata) in the canopy, and vine maple (Acer circinatum), salmonberry (Rubus spectabilis) and huckleberries (Vaccinium spp.) in the understory. All sites contained similar amounts of coarse woody debris. Treatments were randomly assigned to each of the sites. Sites I, E and B were clearcuts, with no riparian buﬀer-strips (hereafter referred to as ‘‘clearcuts’’). The South Creek site and H-buﬀer were both clearcuts with 30 m buﬀer strips (hereafter referred to as ‘‘buﬀer sites’’). Mike Creek, H-control, and Spring Creek were control sites, located in rotation-age second growth forests. All of the buﬀer and clearcut sites were logged between September 1998 and February 1999 (5–10 months prior to our study). 2.2. Field methods This study was based on a capture–mark–recapture system. We used a 77 grid of Longworth live traps, set near stakes 8 m apart, with rows parallel to the stream on each site. Rows of stakes were placed starting at 2 m from the stream, extending out to 50 m away, and traps were placed within 1 m of each stake, near the largest piece of coarse woody debris available. On the clearcut sites all traps were in the open, and on the control sites all traps were in the forest. At the buﬀer sites, the ﬁrst four rows were in the buﬀer strip (forest), while the last three rows were in the clearcut (open). Traps were baited with apple, oats, and sunﬂower seeds, and we provided raw cotton for bedding. We covered them with wood or moss for thermal protection. In total, there were nine trapping sessions: four in 1999 and ﬁve in 2000. Each trapping session consisted of three consecutive trapping nights, during which the traps were set in the evenings and checked every morning. The length of time between trapping sessions varied from 2 to 4 weeks. The trapping eﬀort was the same for each site (49 traps27 nights), but the actual number of trapnights included per site varied depending on disturbance by bears, slugs, and other animals that tripped the trap and made it impossible to catch a small mammal. To account for this, we standardized capture rates by numbers per 100 trapnights, counting only those trapnights in which the trap was available to capture a small mammal (i.e. the trap was found the next morning either empty and still set, or tripped and containing a small mammal). When rodents were caught for the ﬁrst time, we tagged them using numbered metal ear-tags, weighed

K.L. Cockle, J.S. Richardson / Biological Conservation 113 (2003) 133–140

them, determined their sex, inspected them for bot ﬂy (Cuterebra sp.) larvae, and released them. On subsequent captures they were simply recorded, inspected for bot ﬂy larvae and released. The insectivores were too small to mark using ear-tags, so they were weighed and released without marking. We were unable to identify live shrews, so these were simply recorded as Sorex spp. Dead shrews were taken back to the lab, where we later determined their species based on dentition (Woodward, 1994).

135

We determined the eﬀect of the streams themselves on small mammal distribution by counting the number of captures in each row. We used maximum likelihood estimations with distance as a covariate to determine the trends in capture rates with the distance from each stream, and to compare these trends between treatments. This analysis could only be performed for the three species that were captured on all eight sites. Species richness was calculated for each site in each year, using the rarefaction method described in Krebs (1999).

2.3. Data analysis We used maximum likelihood estimation through general linear models for our analyses of treatment eﬀects (PROC GENMOD, SAS version 8.0). Where we had data from both years we used a two-way design (treatments and year), including the interaction term. Percentage data were transformed (arc sine square-root). In all cases the algorithm for estimation converged, thus meeting assumptions required for the test. We used the planned contrasts statement with least square means to compare treatments. We used =0.05 as our critical signiﬁcance level, unless noted. It is important to note that our buﬀer site grids had approximately half of the traps in the riparian reserve (forested) and half in the harvest zone, but since we used the numbers across a grid we did not separate parts of the grid for analysis. We attempted to estimate population sizes from mark–recapture models (PROGRAM MARK), but our data were too sparse and insectivores were not marked. Thus, we used the capture rate (standardized as the number of individuals captured per hundred trap nights) as a measure for comparing population size between treatments. For the more common species we were able to use the number of individuals captured on each site for each trapping session. This should be a reasonable indicator of relative population sizes for comparison between treatments, assuming that the populations are closed (no immigration or emigration) for the 3 days of each trapping session. In seven cases, tagged animals were observed to move between two sites (one a clearcut, the other a 30 m buﬀer site) that were located less than 300 m apart. In each of these cases, the animal was counted as belonging to the site where it was captured most often. For less common species, there were not enough data to analyze each trapping session separately, so instead we looked at the number of individuals captured over the course of each summer. Unfortunately this method does not give a good estimate of how many animals are using the site at once, because we cannot assume that the population is closed for 4 months. Thus we do not know if a site with a high number of individuals is capable of supporting many animals, or if it simply has a high turnover rate of individuals.

3. Results 3.1. Presence and abundance of small mammals In 27 nights of trapping, we captured at least 994 individuals of seven diﬀerent species of small mammals. The deer mouse [Peromyscus maniculatus (Wagner); 465 individuals] was the most common species captured, followed by the creeping vole [Microtus oregoni (Bachman); 281 individuals] and then the vagrant shrew (Sorex vagrans Baird; 153 individuals). Dusky shrews (Sorex monticolus Merriam; 50 individuals) and Southern red-backed voles [Clethrionomys gapperi (Vigors); 30 individuals] were captured on some sites. On a few occasions, we caught American shrew moles [Neurotrichus gibbsii (Baird); 13 individuals], and at the Mike Creek control site we caught Paciﬁc jumping mice (Zapus trinotatus Rhoads; two individuals). Jumping mice were so rare that they were not included in any of the analyses except for the analysis of species richness. Most shrews and moles died in the traps. This is nearly always the case when traps are set overnight (Sullivan et al., 1998). Live insectivores could not be marked and the two shrew species we captured cannot be distinguished from one another in the ﬁeld, so we could not count live shrews. On average, for every shrew found alive, ﬁve were found dead. The trap mortality rate for shrews did not diﬀer between treatments (maximum likelihood, year 2000 data, w2=3.67, d.f.=2, 5, P > 0.15), so it is reasonable to assume that the sites where we collected the most dead shrews were also the sites where shrews were most abundant. Thus, for shrews and moles, only dead animals were counted in the ﬁnal analysis. 3.2. Number of individuals captured Deer mice did not diﬀer in abundance between treatments for any of the nine trapping sessions (all P > 0.05). The total number of deer mice captured (per 100 trap nights) did not diﬀer signiﬁcantly between treatments in either year (w2=2.66, d.f.=2, 10, P > 0.25), and there was no interaction between treatment and

136

K.L. Cockle, J.S. Richardson / Biological Conservation 113 (2003) 133–140

year. In 2000 deer mice were more than 2.9 times higher in density than in 1999 (w2 =21.82, d.f.=1, 10, P < 0.0001). Creeping vole abundance (per 100 trap nights) differed signiﬁcantly between treatments (w2=11.14, d.f.=2, 10, P < 0.004), but there was no signiﬁcant eﬀect of year, or treatment by year interaction (both with P > 0.5). Creeping voles were more than twice as abundant on the clearcuts compared to the other two treatments (P < 0.008), and there was no signiﬁcant diﬀerence between numbers on the controls versus the buﬀer sites (P > 0.4; Fig. 1). The red-backed vole was lower in abundance in the clearcuts, with none captured on those three grids in 1999 and only one in 2000 (Fig. 1). Variation amongst

sites resulted in no signiﬁcant diﬀerence overall (w2=4.75, d.f.=2, 10, P < 0.1); however, the least square mean contrasts showed a signiﬁcantly lower density on clearcuts (P < 0.05) compared with the other two treatments. There was no signiﬁcant eﬀect of year or year by treatment interaction. The vagrant shrew was the third most common species on our study sites (Fig. 1), but showed no signiﬁcant diﬀerences between treatments (P > 0.27) or years (P > 0.21), and no treatment by year interaction (P > 0.5). The dusky shrew showed a large treatment eﬀect (w2=21.53, d.f.=2, 10, P < 0.0001). Dusky shrews were signiﬁcantly more abundant on the controls in contrast to the other treatments (least square means, P < 0.0002), and the buﬀer site abundances were intermediate and signiﬁcantly higher than the clearcuts (P < 0.03). Dusky shrews were not found at clearcut sites in 1999. There was a signiﬁcant year eﬀect on the abundance of dusky shrews (w2=4.74, d.f.=2, 10, P < 0.03), with numbers in 2000 1.74 times higher than in 1999, but there was no year by treatment interaction (P > 0.7). Aside from jumping mice, the least often trapped species was the American shrew mole. The shrew mole had a small treatment eﬀect (P=0.067), but a signiﬁcant treatment by year eﬀect (w2=6.72, d.f.=2, 10, P < 0.035). The buﬀer sites had signiﬁcantly lower densities than the controls overall (P < 0.02). Shrew moles were most common on controls in 1999, but most common on clearcuts in 2000 (Fig. 1). 3.3. Eﬀect of distance from the stream

Fig. 1. Mean number of animals captured per 100 trap nights in each treatment in 1999 (black bars) and 2000 (open bars) for small mammals in second-growth coniferous forest, buﬀer strips, and clearcuts in southwestern British Columbia. Bars indicate 1 standard error. Means and standard errors calculated from sample sizes of n=3, 2, 3 for controls, buﬀers, and clearcuts, respectively.

For the two most common species, deer mice and creeping voles, we looked at changes in abundance across the transriparian gradient, using distance from the stream edge as a covariate. Deer mice showed no signiﬁcant eﬀect of distance from streamside in 1999, but a signiﬁcant decrease with increasing distance in 2000 (Fig. 2; year eﬀect, w2=134.9, d.f.=1, 98, P < 0.0001; distance eﬀect, w2=8.01, d.f.=1,98, P < 0.005). There was a signiﬁcant eﬀect of distance by treatment (w2=8.74, d.f.=2,98, P < 0.02) indicating that the pattern varied by treatment. Individual treatments in the year 2000 showed signiﬁcant decreases in abundance with distance from the stream for the controls (P=0.0003) and 30 m buﬀers (P=0.008), while there was no signiﬁcant slope to the relation for clearcut sites (P=0.61). Creeping voles showed the opposite trend with increasing abundance as distance increased (Fig. 2, w2=25.82, d.f.=1, 45, P < 0.0001). There was a signiﬁcant interaction of treatment and distance (w2=8.52, d.f.=2, 45, P< 0.02), indicating diﬀerent patterns by treatment. The primary diﬀerence was for the 30 m buﬀer sites, where there was a very large increase in abundance beyond the riparian reserve zone (Fig. 2). Vagrant shrew abundance was not related to distance

K.L. Cockle, J.S. Richardson / Biological Conservation 113 (2003) 133–140

137

Fig. 3. Species richness of small mammals at study sites in southwestern British Columbia for 1999 and 2000, based on rarefaction to account for abundance – richness relations. The sites include three control sites (68 year old conifer forest), two sites with 30 m riparian reserves remaining after harvesting, and three clearcut sites. Bars indicate means of the sites within treatment 1 standard error.

Fig. 2. Number of captures of deer mice (Peromyscus maniculatus) and creeping voles (Microtus oregoni) versus distance from the stream for each of the treatments in 2000. Numbers are the total number of captures at each distance. Error bars have been omitted for clarity, but there were no signiﬁcant site by distance eﬀects within treatments, so the patterns are general for sites within a treatment. Sample sizes at each distance are n=3, 2, 3 for controls, buﬀers, and clearcuts, respectively.

from the stream in either year (P > 0.60) and there was no interaction between distance and treatment (P > 0.95). 3.4. Species richness Species richness (Fig. 3) was signiﬁcantly diﬀerent by treatment (w2=14.87, d.f.=2, 10, P < 0.001), with clearcuts signiﬁcantly lower than the other two treatments (P < 0.001). There was no signiﬁcant diﬀerence between the controls and buﬀers in species richness (P=0.60). There was no signiﬁcant year or interaction eﬀect on species richness (P > 0.08), but the control and buﬀer sites did reverse rank orders between years. 3.5. Demographics and condition Sex ratios, average weight of adult males, and juvenile to adult ratios could only be analyzed for deer mice, as

the other species were too rare on certain sites to provide suﬃcient data. There were no signiﬁcant diﬀerences between treatments in either year for these population parameters (Table 1). In August 1999, a number of creeping voles and deer mice began to show abdominal swellings and exit holes characteristic of bot ﬂy larvae (Cuterebra sp.) infestation. The incidence of bot ﬂy parasitism among deer mice diﬀered signiﬁcantly between treatments (w2=12.86, d.f.=2, 5, P < 0.002). On average, 24% of deer mice in the clearcuts were infested, diﬀering signiﬁcantly from the control sites where no mice were infested with the bot ﬂy (P < 0.0001). Deer mice in buffer strips showed intermediate rates of infestation (5%), signiﬁcantly diﬀerent from either the controls or the clearcuts (both with P < 0.05; Table 1). Creeping voles also experienced high bot ﬂy infestation rates in clearcuts; in fact, all of the infested voles were living in the clearcuts (25 infestations among 80 creeping voles in the clearcuts), and thus the treatment eﬀect was signiﬁcant (w2=20.33, d.f.=2, 5, P < 0.0001). Bot ﬂy infestations were extremely rare in 2000 (only three animals were found to be infested), so no comparisons were made between treatments.

4. Discussion 4.1. Eﬀects of clearcutting Compared with nearby forested areas, clearcuts showed a higher abundance of creeping voles, lower abundance of dusky shrews and red-backed voles, lower species richness, and higher rates of bot ﬂy infestation. While studies in upland forested areas report that small mammal diversity either increases or remains the same

138

K.L. Cockle, J.S. Richardson / Biological Conservation 113 (2003) 133–140

Table 1 Population parameters for deer mice, and rates of infestation of bot ﬂy on deer mice and creeping voles (meanstandard error) for each of the three treatments Year Deer mice % Juvenile

Mean weight of adult males in grams % Infested with bot ﬂy

1999 2000 1999 2000 2000 1999

Creeping vole % Infested with bot ﬂy

1999

% Female

Clearcut (3 sites) 47.29.2 38.58.1 42.39.7 58.21.3 19.00.6 24.07.1

(12, 10, 14) (51, 30, 60) (12, 10, 14) (51, 30, 60) (9, 9, 18) (12, 10, 14)

30.07.99 (17, 45,18)

Buﬀer (2 sites) 57.4 7.4 49.0 15.7 44.4 5.6 62.4 14.0 18.0 1.1 5.0 5.0

(18, (21, (18, (21, (11, (18,

10) 31) 10) 31) 6) 10)

Control (3 sites)

Probability

43.31.1 52.214.6 42.48.2 49.92.8 18.30.3 0.0

0.72 0.58 0.96 0.80 0.57 <0.002

0.0 (15, 5)

(13, 21, 12) (53, 57, 52) (13, 21, 12) (53, 57, 52) (3, 15, 9) (13, 21, 12)

0.0 (3, 8, 16)

<0.0001

Numbers in parentheses show the number of animals used in the calculation of each parameter at each site. Statistical signiﬁcance of the comparison between treatments is based on maximum likelihood estimation.

after clearcutting (Cross, 1985; Kirkland, 1990), results from riparian areas suggest that clearcutting reduces species richness (Cross, 1985; our study). In our study, species richness was low in clearcuts because of the absence of the four species least often captured, so differences may have been only a product of our small sample size. However, other studies have found similar evidence that these species decline in numbers when forests are clearcut (shrew moles: Tevis, 1956; Cross, 1985; Cole et al., 1998; red-backed voles: Gashwiler, 1970; Martell and Radvanyi, 1977; Martell, 1983). Deer mice and creeping voles usually increase (Gashwiler, 1970; Hooven and Black, 1976; Martell and Radvanyi, 1977; Martell, 1983; Cross, 1985) or remain similar (Sullivan, 1979; Cole et al., 1998) in abundance after clearcutting, while shrews and red-backed voles usually decrease in abundance (Tevis, 1956; Gashwiler, 1970; Hooven and Black, 1976; Martell, 1983; Cross, 1985; Cole et al., 1998). Shrew moles are captured only rarely in most studies, but they appear to be associated with riparian areas and are present only at low abundance on cutblocks (Tevis, 1956; Cross, 1985; Doyle, 1990; Cole et al., 1998; Gomez and Anthony, 1998), consistent with our ﬁndings. Abundance is not always a good measure of habitat quality (Van Horne, 1983), so we also considered demographic and other condition measures. Sullivan (1979) hypothesizes that clearcuts might be a sink for subordinate animals that are excluded from the remnant forest by aggressive adults. Our results do not support this hypothesis, since age ratios and adult body mass of deer mice were similar in clearcuts and control sites. Deer mice are considered habitat generalists (Krebs and Wingate, 1976; Galindo and Krebs, 1984; Bayne and Hobson, 1998; Darveau et al., 2001), and our results suggest that they adapt well to living in clearcuts. It is worth noting, however, that the incidence of bot ﬂy infestation was signiﬁcantly higher at clearcut sites, among both deer mice and creeping voles. Bot ﬂy parasitism has been

found to reduce survival, reproduction and growth in Townsend’s voles (Microtus townsendii; Boonstra et al., 1980) and red-backed voles (Boonstra et al., 1980; Martell, 1983), and to reduce body mass by 5% in white-footed mice (Peromyscus leucopus; Munger and Karasov, 1991, 1993). 4.2. Eﬀectiveness of buﬀers In our study, buﬀer strips appeared to help lessen the short-term impact of logging on riparian communities. Where buﬀer strips were retained along streams, creeping vole abundance, red-backed vole abundance, and species richness were similar to the values for the control sites. Since only half of the traps on the buﬀer grids were located within the buﬀers, our results indicate that the eﬀects of forest harvesting on small mammals were reduced within buﬀer strips. Like us, Cross (1985) found that buﬀer strips were similar in species richness to unlogged controls, and recommended the use of buﬀers to reduce the impacts of clearcutting. However, on our buﬀer sites, the abundance of dusky shrews and the rate of bot ﬂy infestation in deer mice were intermediate between controls and clearcuts. Riparian corridors are probably not necessary for species like deer mice that adapt easily to clearcuts (e.g. Darveau et al., 2001), but may be critical for maintaining connectivity for other species. Small mammals operate on a spatial scale where riparian reserves of 30 m width might suﬃce to conserve important habitat or provide corridors through the landscape (Darveau et al., 2001; Vesely and McComb, 2002). These reserves may also be important sources for the recolonization of regenerating forest stands by small mammals that are otherwise reduced on clearcuts. The maintenance of small mammal populations may also have important implications for other wildlife dependent on small mammals. This study shows that small mammals of various species make use of riparian buﬀer strips, and the results

K.L. Cockle, J.S. Richardson / Biological Conservation 113 (2003) 133–140

suggest that buﬀer strips are useful in conserving small mammals. Results from similar studies in other regions will be necessary to complete this picture. Our results cannot be extrapolated beyond the immediate vicinity of our study site and it is unlikely that all small mammals and all forest types would respond similarly. In order to be more certain about cause and eﬀect, there is an urgent need for studies that compare sites both before and after clearcutting. Studies of demography and movements will be necessary to determine whether riparian buﬀer strips are wide enough, or have enough inﬂuence on adjacent clearcuts, to provide landscape connectivity and support viable populations of all small mammals during the many years needed for a forest stand to regenerate. Our results support the use of riparian buﬀer strips, but longer-term, larger-scale research will be needed before we can tell whether current buﬀer requirements are suﬃcient to conserve small mammals at a landscape level.

Acknowledgements We gratefully acknowledge our ﬁeld assistants— Pindy Saran, Nadia Baker, and Graham Rhodes. We would also like to thank the many volunteers who helped with ﬁeld work in the autumn of 1999. We are grateful to Walt Klenner for the loan of the Longworth traps and to David Huggard for instruction in identifying shrews. We thank Peter Arcese and two reviewers for suggestions on the manuscript. Financial support for this project was provided by NSERC (Canada) and Forest Renewal B.C.

References Anthony, R.G., Forsman, E.D., Green, G.A., Witmer, G., Nelson, S.K., 1987. Small mammal populations in riparian zones of diﬀerent-aged coniferous forests. The Murrelet 68, 94–102. Bamﬁeld, A.W.F., 1974. The Mammals of Canada. University of Toronto Press, Toronto. Bayne, E.M., Hobson, K.A., 1998. The eﬀects of habitat fragmentation by forestry and agriculture on the abundance of small mammals in the southern boreal mixedwood forest. Can. J. Zool. 76, 62– 69. Bilby, R.E., 1988. Interactions between aquatic and terrestrial systems. In: Raedeke, K.J. (Ed.), Streamside Management. Riparian Wildlife and Forestry Interactions, University of Washington, Institute of Forest Resources, Contribution No. 59, Seattle, Washington. Boonstra, R., Krebs, C.J., Beacham, T.D., 1980. Impact of bot ﬂy parasitism on Microtus townsendii populations. Can. J. Zool. 58, 1683–1692. Cole, E.C., McComb, W.C., Newton, M., Leeming, J.P., Chambers, C.L., 1998. Response of small mammals to clearcutting, burning, and glyphosate application in the oregon coast range. J. Wildl. Manag. 62, 1207–1216. Cross, S.P., 1985. Responses of small mammals to forest riparian perturbations. In: Riparian Ecosystems and their Management:

139

Reconciling Conﬂicting Uses. First North American Riparian Conference. (USDA Forest Service General Technical Report RM-120). Cross, S.P., 1988. Riparian systems and small mammals and bats. In: Raedeke, K.J. (Ed.), Streamside Management: Riparian Wildlife and Forestry Interactions. University of Washington, Institute of Forest Resources, Contribution No. 59, Seattle, Washington. Darveau, M., Labbe´, P., Beauchesne, P., Be´langer, L., Huot, J., 2001. The use of riparian forest strips by small mammals in a boreal balsam ﬁr forest. Forest Ecology and Management 143, 95–104. de Bellefeuille, S., Be´langer, L., Huot, J., Cimon, A., 2001. Clear-cutting and regeneration practices in Quebec boreal balsam forest: eﬀects on snowshoe hare. Can. J. For. Res. 31, 41–51. deMaynadier, P.G., Hunter Jr., M.L., 1995. The relationship between forest management and amphibian ecology: a review of the North American literature. Environmental Review 3, 230–261. Doyle, A.T., 1990. Use of riparian and upland habitats by small mammals. J. Mammal. 71, 14–23. Duguay, J.P., Wood, P.B., Miller, G.W., 2000. Eﬀects of timber harvests on invertebrate biomass and avian nest success. Wildlife Society Bulletin 28, 1123–1131. Forest Ecosystem Management Assessment Team, 1993. Forest ecosystem management: an ecological, economic, and social assessment. US Department of Agriculture, US Department of the Interior, US Department of Commerce, and Environmental Protection Agency, Portland, Oregon, USA. Galindo, C., Krebs, C.J., 1984. Habitat use and abundance of deer mice: interactions with meadow voles and red-backed voles. Can. J. Zool. 63, 1870–1879. Gashwiler, J.S., 1970. Plant and mammal changes on a clearcut in west-central Oregon. Ecology 51, 1018–1026. Gomez, D.M., Anthony, R.G., 1998. Small mammal abundance in riparian and upland areas of ﬁve seral stages in Western Oregon. Northwest Sci. 72, 293–302. Hagar, J.C., 1999. Inﬂuence of riparian buﬀer width on bird assemblages in western Oregon. J. Wildl. Manag. 63, 484–496. Hooven, E.F., 1973. Response of Oregon creeping vole to the clearcutting of a Douglas-ﬁr forest. Northwest Sci. 47, 256–264. Hooven, E.F., Black, H.C., 1976. Eﬀects of some clearcutting practices on small-mammal populations in western Oregon. Northwest Sci. 50, 189–208. Kelsey, K.A., West, S.D., 1998. Riparian Wildlife. In: Naiman, R.J., Bilby, R.E. (Eds.), River Ecology and Management: Lessons from the Paciﬁc Coastal Ecoregion. Springer-Verlag, New York, pp. 235– 258. (Chapter 10). Kirkland, G.L., 1990. Patterns of initial small mammal community change after clearcutting of temperate North American forests. Oikos 59, 313–320. Krebs, C.J., 1999. Ecological Methodology. Benjamin/Cummings Publishing Company, California. Krebs, C.J., Wingate, I., 1976. Small mammal communities of the Kluane region, Yukon Territory. Can. Field-Nat. 90, 379–389. Laurance, W.F., 1990. Comparative responses of ﬁve arboreal marsupials to tropical forest fragmentation. J. Mammal. 71, 641–653. Martell, A.M., 1983. Changes in small mammal communities after logging in north-central Ontario. Can. J. Zool. 61, 970–980. Martell, A.M., Radvanyi, A., 1977. Changes in small mammal populations after clearcutting of northern Ontario black spruce forest. Can. Field-Nat. 91, 41–46. McComb, W.C., Chambers, C.L., Newton, M., 1993a. Small mammal and amphibian communities and habitat associations in red alder stands, central Oregon Coast Range. Northwest Sci. 67, 181–188. McComb, W.C., McGarigal, K., Anthony, R.G., 1993b. Small mammal and amphibian abundance in streamside and upslope habitats of mature Douglas-ﬁr stands, western Oregon. Northwest Sci. 67, 7–15. Munger, J.C., Karasov, W.H., 1991. Sublethal parasites in whitefooted mice: impact on survival and reproduction. Can. J. Zool. 69, 398–404.

140

K.L. Cockle, J.S. Richardson / Biological Conservation 113 (2003) 133–140

Munger, J.C., Karasov, W.H., 1993. Costs of bot ﬂy infection in white-footed mice: energy and mass ﬂow. Can. J. Zool 72, 166– 173. Payer, D.C., Harrison, D.J., 2000. Structural diﬀerences between forests regenerating following spruce budworm defoliation and clear-cut harvesting: implications for marten. Can. J. For. Res. 30, 1965–1972. Sirotnak, J.M., 2000. Direct and indirect eﬀects of herbivores on nitrogen dynamics: voles in riparian areas. Ecology 81, 78–87. Sullivan, T.P., 1979. Demography of populations of deer mice in coastal forest and clear-cut (logged) habitats. Can. J. Zool. 57, 1636–1648. Sullivan, T.P., Nowotny, C., Lautenschlager, R.A., Wagner, R.G., 1998. Silvicultural use of herbicide in sub-boreal spruce forest:

implications for small mammal population dynamics. J. Wildl. Manag. 62, 1196–1206. Terwilliger, J., Pastor, J., 1999. Small mammals, ectomycorrhizae, and conifer succession in beaver meadows. Oikos 85, 83–94. Tevis, L., 1956. Responses of small mammal populations to logging of Douglas-ﬁr. J. Mammal. 37, 189–196. Van Horne, B., 1983. Density as a misleading indicator of habitat quality. J. Wildl. Manag. 47, 893–901. Vesely, D.G., McComb, W.C., 2002. Salamander abundance and amphibian species richness in riparian buﬀer strips in the Oregon Coast Range. Forest Sci. 48, 291–297. Woodward, S.M., 1994. Identiﬁcation of Sorex monticolus and Sorex vagrans from British Columbia. Northwest Sci. 68, 277–283.