Snow Tracking and Trapping Harvest as Reliable Sources
for Inferring Abundance: A 9-year Comparison
Toshinori Kawaguchi, André Desrochers, and Héloïse Bastien
Northeastern Naturalist, Volume 22, Issue 4 (2015): 798–811
Full-text pdf (Accessible only to subscribers. To subscribe click here.)
Access Journal Content
Open access browsing of table of contents and abstract pages. Full text pdfs available for download for subscribers.
Current Issue: Vol. 30 (3)
Check out NENA's latest Monograph:
Monograph 22
Northeastern Naturalist
798
T. Kawaguchi, A. Desrochers, and H. Bastien
22001155 NORTHEASTERN NATURALIST 2V2(o4l). :2729,8 N–8o1. 14
Snow Tracking and Trapping Harvest as Reliable Sources
for Inferring Abundance: A 9-year Comparison
Toshinori Kawaguchi1,*, André Desrochers1, and Héloïse Bastien2
Abstract - Trapping harvest and snow tracking are frequently used to infer population
dynamics, yet there have been few evaluations of these indices. We developed population
indices for Martes americana (American Marten), Mustela spp. (weasels), and Tamiasciurus
hudsonicus (American Red Squirrel) from 9 years of snow-tracking data in eastern
Canada. We employed mean track counts per unit effort as population indices derived from
a generalized linear model (GLM) of track counts as a function of year and covariates including
forest age. Mean track counts were significantly correlated with American Marten
and weasel pelt sales and year effects in GLM were correlated with American Red Squirrel
and weasel pelt sales. The results of both methods are in agreement; therefore they are likely
valid sources to infer population dynamics for these species.
Introduction
Monitoring animal populations is key to understanding their ecosystems, functions,
and responses to anthropogenic and natural disturbances (Lindenmayer
et al. 2012). Making direct estimates of population densities by live-trapping or
mark–recapture methods is very labor intensive and expensive (Gese 2001). Thus,
ecologists often resort to population indices, such as those derived from snow
tracking (Pellikka et al. 2005) or trapping harvest (Roberts and Crimmins 2010)
data, observation reports by hunters (Simard et al. 2012), and scat surveys (Krebs
et al. 2001, Mowat and Slough 2003).
Snow tracking is non-invasive (Halfpenny et al. 1995) and is frequently used
to estimate relative abundance of mammals wintering in North America (Mowat
and Slough 2003) and Europe (Pellikka et al. 2005). Some studies assumed that
the number of tracks is proportional to population size and used tracking counts
per unit effort to infer relative abundances of mammals such as Lepus americanus
Erxleben (Snowshoe Hare; Jensen et al. 2012), Tamiasciurus hudsonicus Erxleben
(American Red Squirrel; Jensen et al. 2012), Martes americana Turton (American
Marten; Krebs 2011), and Mustela frenata Lichtenstein (Long-tailed Weasel;
Fitzgerald 1977). Snow tracking is relatively easy to conduct (Halfpenny et al.
1995) and has higher detection rates than other non-invasive techniques such as
camera traps and track plates (Gompper et al. 2006). Also, snow tracking does not
1Centre d’Étude de la Forêt and Département des Sciences du Bois et de la Forêt, Université
Laval, 2405, Rue de la Terrasse, Québec, PQ, G1V 0A6, Canada. 2Direction de la Gestion
de la Faune de la Capitale-Nationale et de la Chaudière-Appalaches, Ministère des Forêts,
de la Faune et des Parcs, 1300, Rue du Blizzard, local 100, Québec, PQ G2K 0G9, Canada.
*Corresponding author - toshinori.kawaguchi.1@ulaval.ca.
Manuscript Editor: Tom French
Northeastern Naturalist Vol. 22, No. 4
T. Kawaguchi, A. Desrochers, and H. Bastien
2015
799
require the baits or attractants that are employed in other population-monitoring
techniques (Raphael 1994).
However, there may be significant noise and bias in population indices obtained
from snow-tracking data. Track counts can be affected by weather conditions such
as strong winds and recent snow falls (Gompper et al. 2006, Raphael 1994), and
can vary with the activity level of animals. For example, activity of American Marten
(Thompson and Colgan 1994) and Mustela erminea L. (Short-tailed Weasel;
Robitaille and Baron 1987) decreases at extremely low temperatures . Presence of
prey can attract the predator, resulting in an increased track count for the predator
as shown for Lynx canadensis Kerr (Canada Lynx; Keim et al. 2011). Observer bias
and misidentification can also affect results.
In many jurisdictions, trapping harvest data in the form of total catch and catch
per unit effort is commonly used to infer relative population sizes and trends of
fur-bearing species (Douglas and Strickland 1987), particularly American Marten
and Pekania pennanti (formaly Martes pennanti) Erxleben (Fisher; Jensen et al.
2012). In the US, harvest surveys are frequently used to monitor Lynx rufus Schreber
(Bobcat) population status (Roberts and Crimmins 2010). In Quebec, Canada,
the number of pelts that are sold is the primary index used to track population size
for weasels (composed mostly of Short-tailed and Long-tailed Weasel), Mephitis
mephitis Schreber (Striped Skunk), and Ondatra zibethicus L. (Muskrat). For
Canada Lynx, American Marten, and Fisher, abundances can be inferred from the
logbooks that trappers are required to maintain. Those logbooks record the numbers
of trapped animals for each species, together with trapping effort. Numbers of
pelts sold can be obtained from fur-transaction reports. Trapping harvest is mainly
affected by trapping effort (Fortin and Cantin 2004), which in turn is influenced
by several ecological and socioeconomic factors, such as food abundance (Ryan et
al. 2004), market prices (Weinstein 1977), government quotas (Smith et al. 1984),
landscape changes (Raphael 1994), other disturbances over time (Raphael 1994),
and temperature (Kapfer and Potts 2012).
In addition to problems that are specific to each of these indirect methods, temporal
changes in sampling locations may introduce additional bias to the index.
Moreover, even if field surveys are conducted in fixed locations each year, successional
patterns in vegetation will affect the response of different species to the
sampled plots (Anderson 2001). As is the case with any model, methods that are
used to develop population indices should optimize the combination of simplicity
and accuracy.
Few studies have evaluated the reliability of population indices that have been
obtained through those indirect methods. In a 5-year study in Ontario, Thompson
and Colgan (1987) reported different population trends from trapping harvest and
snow tracking. Thompson et al. (1989) found no correlation between track counts
and trapping harvest, but they did find a significant correlation between track
counts and live-trapping data. The authors of those studies suggested that
track counts correctly described population changes over 5 years.
Northeastern Naturalist
800
T. Kawaguchi, A. Desrochers, and H. Bastien
2015 Vol. 22, No. 4
To evaluate population indices that were based on snow tracking and pelt sales,
we examined the correlation between the 2 techniques with data from 3 mammal
species: American Marten, weasels, and American Red Squirrel. Although the 2
indices did not directly measure population size, we hypothesized that if, after
accounting for sources of error described above, 2 indices describe population
changes accurately over time, then they would be highly correlated. We predicted
that if one or both of the indices failed to describe annual population dynamics, the
correlation coefficient between the 2 indices should not be different from zero for
the species concerned.
Field-Site Description
We conducted snow-tracking surveys in the Montmorency Forest, a 66-km2 area,
~80 km north of Quebec City (47°20'N, 71°10'W), Canada (Fig. 1). Trapping and
hunting for the 3 species were not allowed in this area. Most of the study area was
originally clear cut between 1941 and 1945 and is now managed with a combination
Figure 1. Spatial
location of snowtracking
site (the
M o n t m o r e n c y
Forest; black shading)
and trapping
area of Furbearer
Management Unit
(UGAF) 39 (gray
shading), southern
Quebec, Canada,
2004–2012.
Northeastern Naturalist Vol. 22, No. 4
T. Kawaguchi, A. Desrochers, and H. Bastien
2015
801
of clear cuts and selective cuts. The resulting forest is composed of mature (more
than 40-y-old), mid-succession (21- to 40-y-old), and regenerating (less than 20-yold)
forest stands, which comprised 55%, 25%, and 20% of the study area at the
time of the study, respectively. Locations of different-aged stands shift with time
due to continuing timber harvest and forest stand succession, but mean stand age
remained stable throughout the study period (42.97 ± 1.67 years, range 0–113 y).
A dense road network is present, with about 150 km of dirt roads. During winter,
several roads were machine-groomed as cross-country ski trails. Elevation ranges
from 650 m to 1000 m. From 1999 to 2011, annual mean temperature was 0.3 °C.
Annual precipitation was 1417 mm (33% as snow). Annual maximum snow depth
at the weather station in the Montmorency Forest ranged from 62 cm to 146 cm during
the period from 1999 to 2011 (Vigeant-Langlois and Desrochers 2011). Abies
balsamea (L.) Miller (Balsam Fir) dominated second-growth mature forest stands.
Picea mariana (Miller) Britton, Sterns, and Poggenburg (Black Spruce), Betula papyrifera
Marshall (White Birch), Populus tremuloides Michaux (Trembling Aspen),
and P. glauca (Moench) Voss (White Spruce) were also common. Recent (<5-y-old)
clear cuts were generally colonized by Rubus idaeus L. (Red Raspberry), Balsam
Fir, and White Birch (de Bellefeuille et al. 2001).
We obtained pelt-sales data from fur-transaction reports of Furbearer Management
Unit (French acronym, UGAF) 39, a 7934-km2 area (Fig. 1). In this area, Balsam
Fir and Black Spruce were the dominant plant species. White Birch, Trembling
Aspen, Betula alleghaniensis Britt. (Yellow Birch), and Acer spp. (maples) were
also common (Dussault et al. 2006). Similar to the Montmorency Forest, timber
harvest was conducted in UGAF 39, resulting in a heterogeneous stand mosaic
(Dussault et al. 2006). From 2004 to 2012, mean daily winter temperature ranged
from -16 oC to -9 oC and total annual winter precipitation ranged from 165 mm to
265 mm (Environment Canada 2014; Fig. 2).
In summary, the main difference between the snow-tracking and trapping areas
we studied was the presence of trapping in the latter. Thus, furbearers were possibly
more abundant in the snow-tracking area than in the trapping area, potentially resulting
in different movement patterns due to higher density. However, we assumed
that those 2 locations exhibited similar population trends because they overlapped
geographically and had similar forest composition and forest management. In addition,
wildlife populations within 100 km have been shown to vary synchronously
in various taxa, including mammals (Liebhold et al. 2004). Therefore, we expected
that few individuals changed their locations (Montmorency or UGAF 39) depending
on study year.
Methods
Snow tracking
We conducted snow tracking each winter (20 December–31 March) from 2004
to 2012. We counted tracks along a portion of a network of ~150 km of roads, 40
km of trails, and 60 km of off-trail straight-line transects (Fig. 3). None of the
roads that we surveyed had been snowplowed. Transect length depended upon snow
Northeastern Naturalist
802
T. Kawaguchi, A. Desrochers, and H. Bastien
2015 Vol. 22, No. 4
condition, current weather, time of day, and personnel availability (Table 1). Offtrail
transects were randomly selected from a systematic grid covering the entire
study area at the beginning of each year. However, in the selection process, we
removed transects that had been surveyed in the previous 2 years. For each year,
we surveyed selected transects only once to cover as large an area as possible. As a
result, we surveyed a mean ± SD of 91.3 ± 28.9 km of transects each year (Table 1).
We conducted our transect surveys within 24–72 hours after any snowfall >3 cm.
We recorded the locations of all tracks found within 2 m of either side of the transect
lines using a GPS receiver and identified the species based on track pattern
and size. We ignored conspecific tracks that were within 3 m of a recorded track.
We utilized temperature and snow depth as measured at an Environment Canada
weather station at the Montmorency Forest.
Figure 2. Inter-annual dynamics
of winter precipitation and winter
temperature in the study sites: the
Montmorency Forest and the Laurentides
Wildlife Reserves (UGAF
39), southern Quebec, Canada,
2004–2012. (a) Winter temperature
(oC), (b) winter precipitation (mm).
The data for 2006 was not available.
Northeastern Naturalist Vol. 22, No. 4
T. Kawaguchi, A. Desrochers, and H. Bastien
2015
803
We georeferenced track and transect data with ArcGIS (Version 10.1, ESRI,
Redlands, CA) and then split transects into 200-m fixed-length segments, resulting
in 4123 segments for the entire study (Table 1). We selected this fixed length as a
compromise that limited the number of zeros in the data while retaining a sufficiently
large number of sampling units. We counted tracks on each transect segment.
Within a 100-m buffer around each segment, we determined the mean forest-stand
age (weighted by area), slope (the difference between minimum and maximum elevation),
variance of age, and mean elevation. We calculated mean forest-stand age
with age in years and area (w) in hectares for each forest stand in a buffer:
Mean age = ik =
(w1)(age1) + (w2)(age2) + (w3)(age3) + … + (wi)(agei)
,
(w1) + (w2) + (w3) + … + (wi)
where i = the number of forest stands in a buffer. Buffers occasionally included
roads, rivers, and lakes; thus, we also calculated the percentage of vegetated area
inside each buffer.
Figure 3. Spatial distribution
of sampling transects in the
Montmorency Forest, southern
Quebec, Canada, 2004–
2012. Black lines indicate
off-trail transects and gray
lines indicate either roads or
trails.
Northeastern Naturalist
804
T. Kawaguchi, A. Desrochers, and H. Bastien
2015 Vol. 22, No. 4
Population indices from snow-tracking data
We assumed that track count is a function of population size, animal-activity
level, and exposure time of transect since the last disturbance. We developed 2
population indices. The first was mean track count per 200-m-transect segment per
hours of exposure since the last snowfall. This index has the advantage of simplicity,
but ratios raise statistical issues including spurious correlation (Atchley et al.
1976). Thus, we used a second index based on a generalized linear model (GLM)
with a negative binomial distribution and log link by using package MASS in the
software R (Venables and Ripley 2002). We used negative binomial distribution
instead of Poisson due to the large number of zero counts. In the GLMs, track
count was a function of year as a categorical variable, and combinations of the following
covariates: mean stand age, variance of stand age, mean temperature since
last disturbance, exposure time since last disturbance (snow or wind), slope, mean
elevation, proportion of vegetated area, and transect type (road or off-trail). We
included transect type in the model to account for differences in length of off-trail
transect surveyed among study years. Year effect represents a population index
(or anomaly) by comparing mean counts in a given year with those in the reference
year (2004). Thus, year-effect estimates reflected differences in mean track
counts between 2004 and other years, after accounting for the effect of covariates
on track counts described above. We were aware that GLMs do not account for
spatial autocorrelation, which likely occurred in track counts. However, average
Moran’s I values for American Marten, American Red Squirrel, and weasels over
all study years were 0.044 (range = 0.015–0.080), 0.062 (range = 0.007–0.119)
and 0.024 (range = 0.006–0.040), respectively. These values suggest weak spatial
autocorrelation in track counts after accounting for covariates. More importantly,
we considered years, not transects, as sampling units for statistical inference on the
comparison of time series. Thus we do not consider spatial autocorrelation an issue.
Table 1. Sampling effort for snow tracking and track counts for American Marten, American Red
Squirrel, and weasels at the Montmorency Forest, southern Quebec, Canada, 2004–2012.
Sampling 200-m American American
On road Off-trail events segments Marten Red Squirrel Weasels
Year /trails (km) (km) (days) (n) (count) (count) (count)
2004 33 17 14 252 70 133 36
2005 60 5 14 325 66 187 11
2006 51 23 12 370 88 63 58
2007 71 16 14 435 140 2061 130
2008 97 23 12 602 329 441 575
2009 58 12 11 346 98 78 76
2010 91 12 8 518 106 251 35
2011 110 13 8 619 163 175 89
2012 91 39 19 656 194 1497 71
Total 662 160 112 4123 1254 4886 1081
Mean 74 18 12.4 458 139 543 120
Standard deviation 25 10 3.4 146 83 723 174
Northeastern Naturalist Vol. 22, No. 4
T. Kawaguchi, A. Desrochers, and H. Bastien
2015
805
Pelt sales
We used pelt-sales data as a proxy for trapping harvest, based on fur-transaction
reports, which were obtained from Quebec’s Ministère des Forêts, de la Faune et
des Parcs. Pelt-sales data were composed mainly of trapping harvest totals and
number of animals hunted. There were 112 trappers in the area who used bodygripping
traps (Model 120, 160, or 220). Trapping season extended from 18 October
to 1 March of the following year, and there was no harvest limit for the 3 species
that we studied. However, the ministry asked trappers to stop trapping American
Marten when they began to catch more females than males late in the trapping
season. We obtained data of trapping effort (number of traps x number of nights)
for American Marten from trappers’ mandatory reports and daily logbooks from
the government. Data for American Red Squirrel and weasels were not available.
Average trapping effort for American Marten from 2004 to 2012 was 70,187 (range
= 52,046–87,756) trap-nights. Average number of trapping logbooks submitted was
72 (range = 58–83).
Population indices from pelt-sales data
Various confounding factors such as temperature and trapping effort can influence
both pelt sales and track counts. We calculated Pearson product-moment
correlations (r) between raw pelt sales and 4 factors: trapping effort, daily mean
temperature (January–March) (oC), winter total precipitation (mm), and pelt price
in the previous year adjusted for 2012 (Canadian dollars) in the UGAF 39. Only
factors which had significant correlation with pelt sales were used to test correlation
between 2 indices.
Testing correlation between two indices
In order to test if the 2 indices agreed, we calculated Pearson product-moment
correlation (r) between pelt sales and each of the annual snow-tracking indices. If
pelt sales had significant correlation with any potential external factors such as trapping
effort, winter temperature, and winter precipitation, we also modeled pelt sales
as a function of the significant external factors and one of the population indices derived
from snow-tracking data. When the estimated parameter of an annual tracking
index was significantly different from zero, we deemed that the 2 indices agreed.
We replicated the procedure for each population index based on snow tracking. We
conducted all statistical analyses in the software program R (R Development Core
Team 2015).
Results
Annual track counts were 543 ± 723 (range = 63–2061) for American Red Squirrel,
120 ± 174 (range = 11–575) for weasels, and 139 ± 83 (mean ± SD) (range =
66–329) for American Marten.
Mean annual pelt sales for American Red Squirrel, weasels, and American
Marten were 184 (range = 74 –319), 361 (242–636), and 921 (629–1334) pelts
respectively. None of the factors among trapping effort, winter temperature, winter
Northeastern Naturalist
806
T. Kawaguchi, A. Desrochers, and H. Bastien
2015 Vol. 22, No. 4
precipitation, and pelt price in the previous year were significantly correlated with
pelt sales (-0.64 < r < 0.44, P > 0.05).
Mean number of track counts hr-1 200 m-1 for American Red Squirrel, weasels,
and American Marten were 0.029 (0.004–0.131), 0.006 (0.007–0.027), and 0.008
(0.006–0.016), respectively. There was strong agreement between mean track
counts per unit effort and pelt sales in American Marten and weasels (Table 2,
Fig. 4), but the agreement was not statistically significant in the case of American
Red Squirrel (Table 2, Fig. 4).
Year-effect estimates of counts 200 m-1 (compared with 2004) in the GLM
ranged from -1.76 to 2.34 for American Red Squirrel, -1.54 to 1.91 for weasels, and
-0.61 to 0.81 for American Marten. Year-effect estimates agreed strongly with pelt
sales of weasels and American Red Squirrel (Table 2, Fig. 4), but the agreement
was not statistically significant in the case of American Marten (Table 2). Agreement
between GLM-based snow-tracking indices and pelt sales was the highest in
weasels, followed by American Red Squirrel and American Marten (Table 2).
Discussion
Track counts data were generally in strong agreement with pelt sales over the 9
years of this study, which indicated that both methods captured a real signal in the 3
taxa that we studied. Our results differed from those of Thompson et al. (1989), who
found no correlation between track counts and trapping harvest in Short-tailed Weasel,
Vulpes vulpes L. (Red Fox), or Canada Lynx. Agreement between track counts
and pelt sales is remarkable especially because they were derived from completely
Table 2. Pearson correlations between pelt sales and different population indices that were based on
snow tracking for the 3 species (n = 9): American Marten, American Red Squirrel, and weasels in
southern Quebec, Canada, 2004–2012. Year-effect GLM indicates estimates of year effect from a
generalized linear model.
Population index r P
American Marten
Mean of count/exposure hours 0.71 0.032
Year effect GLM 0.55 0.120
American Red Squirrel
Mean of count/exposure hours 0.57 0.100
Year effect GLM 0.77 0.020
Weasels
Mean of count/exposure hours 0.87 0.002
Year effect GLM 0.85 0.004
Figure 4 (following page). Comparison of population trends between snow tracking and pelt
sales across three taxa: (a) American Marten, (b) American Red Squirrel, and (c) weasels,
southern Quebec, Canada, 2002–2012. Two population indices are presented: left = year effect
of a Generalized Linear Model (Year effect GLM), right = tracks/exposure time. Black
lines represent pelt sales and gray lines represent population indices of year effect GLM
(right) or tracks/exposure time (left). Vertical bars represent standard errors.
Northeastern Naturalist Vol. 22, No. 4
T. Kawaguchi, A. Desrochers, and H. Bastien
2015
807
Northeastern Naturalist
808
T. Kawaguchi, A. Desrochers, and H. Bastien
2015 Vol. 22, No. 4
different methods, from different sets of locations, with contrasting trapping effort.
Thus, both data sets likely captured population phenomena occurring over the entire
study area.
Agreement between indices may arise from common biases, such as weather and
food-availability effects on exploratory behavior and movements. Thus, it could be
argued that the variation in the population indices had little to do with actual population
size. However, there was little confounding effect of trapping effort, mean
daily temperatures. or precipitation (snow) on indices, because none of these factors
were strongly correlated with indices; food-related biases may have occurred.
Jensen et al. (2012) reported that success rate of American Marten harvest was
lower in mast years of Fagus grandifolia Ehrhart (American Beech) and Sorbus
aucuparia L. (Mountain Ash) than in year of mast failure. When food is abundant,
American Marten could be less likely to be attracted to bait associated with trapping
devices, possibly leading to underestimates of the relative abundance. High
food-abundance might decrease track counts. For example, high levels of food
abundance was reported to decrease daily movement length of Mustela nivalis L.
(Least Weasel; Klemola et al. 1999), potentially leading to fewer track counts.
American Marten track counts showed less agreement with pelt sales than did
weasels. Lower agreement could have arisen due to enactment of a Quebec government
policy that recommended that trappers stop trapping adults (H. Bastien,
Ministère des Forêts, de la Faune et des Parcs, Quebec, Canada, pers. comm.).
However, we were unable to determine year of the implementation. In Michigan,
trapping harvest limits greatly impacted the number of Fishers harvested (Hiller et
al. 2011).
American Red Squirrel track counts also exhibited a lower degree of agreement
with pelt sales than did weasel track counts. Sales of American Red Squirrel
pelts might reflect population sizes poorly because not all pelts are sold, given
that American Red Squirrel is often used as bait for other furbearers (H. Bastien,
pers. comm.). Furthermore, American Red Squirrel pelts are considered of
no significant commercial value by local trappers. The price of a American Red
Squirrel pelts varied between 0.65 and 1.44 Canadian dollars (CAD) from 2004
to 2012 (where 1 CAD @ 1 USD), which was much less than the price of a weasel
pelt (range = 2.24 CAD–8.87 CAD) or an American Marten pelt (range = 44.88
CAD–121.71 CAD). Thus, the motivation for capture and pelt harvest contrasted
strongly between species.
Population indices are not substitutes for true estimates of abundance. We did
not compare the indices used in this study with population trends obtained from
direct measures of abundance, but given the high correlations obtained, at least in
the case of American Marten and weasels, our study adds to existing support for the
use of either snow tracking or pelt sales. With those indices, one should be able to
infer relative inter-annual population-changes of the furbearer species studied, and
possibly others such as Red Fox and Canada Lynx. Therefore, use of these indices
would be useful to investigate impact of local forestry and wildlife management on
population dynamics of furbearer species.
Northeastern Naturalist Vol. 22, No. 4
T. Kawaguchi, A. Desrochers, and H. Bastien
2015
809
Acknowledgments
Financial support for this project was provided by a scholarship to T. Kawaguchi from
the Leadership and Sustainable development Scholarship Program of Laval University, by
a scholarship to T. Kawaguchi from the Foundation F.-K.-Morrow, and by a Natural Sciences
and Engineering Research Council of Canada (NSERC) grant to A. Desrochers. We
are grateful to the 29 skilled field-workers who helped collect snow-tracking data, and to the
Ministère des Forêts, de la Faune et des Parcs for providing us with pelt-sales data, trappingeffort
data, and geographical information. We thank D. Fortin, L. Bélanger, and C. Samson
for their assistance in the design of the study and W.F.J. Parsons of Centre d’Ètude de Forêt
(CEF) for his assistance with linguistic corrections. Lastly, we thank the anonymous reviewers
and the editor for their thoughtful comments and suggestions to improve our manuscript.
Literature Cited
Anderson, D.R. 2001. The need to get the basics right in wildlife field studies. Wildlife
Society Bulletin 29:1294–1297.
Atchley, W.R., C.T. Gaskins, and D. Anderson. 1976. Statistical properties of ratios. I. Empirical
results. Systematic Zoology 25:137–148.
De Bellefeuille, S., L. Bélanger, J. Huot, and A. Cimon. 2001. Clear-cutting and regeneration
practices in Quebec boreal Balsam Fir forest: Effects on snowshoe hare. Canadian
Journal of Forest Research 31:41–51.
Douglas, C.W., and M.A. Strickland. 1987. Fisher. Pp. 511–529, In M.J. Novak, J.A. Baker,
M.E. Obbard, and B. Malloch (Eds.) Wildlife Furbearer Management and Conservation
in North America. Ontario Trappers Association, North Bay, ON, Canada. 1150 pp.
Dussault, C., M. Poulin, R. Courtois, and J.-P. Ouellet. 2006. Temporal and spatial distribution
of moose–vehicle accidents in the Laurentides Wildlife Reserve, Quebec, Canada.
Wildlife Biology 12:415–425.
Fitzgerald, B.M. 1977. Weasel predation on a cyclic population of the Montane Vole (Microtus
montanus) in California. Journal of Animal Ecology 46:367–397.
Fortin, C., and M. Cantin. 2004. Harvest status, reproduction, and mortality in a population
of American Martens in Quebec, Canada. Pp. 221–234, In D.J. Harrison, A.K. Fuller,
and G. Proulx (Eds.). Martens and Fishers (Martes) in Human-altered Environments: An
International Perspective. Springer, New York, NY. 280 pp.
Gese, E.M. 2001. Monitoring of terrestrial carnivore populations. Pp. 372–396, In J.L.
Gittleman, S.M. Funk, D.W. MacDonald, and R.K. Wayne (Eds.). Carnivore Conservation.
Cambridge University Press and The Zoological Society of London, Cambridge,
UK. 692 pp.
Gompper, M.E., R.W. Kays, J.C. Ray, S.D. Lapoint, D.A. Bogan, and J.R. Cryan. 2006. A
comparison of noninvasive techniques to survey carnivore communities in northeastern
North America. Wildlife Society Bulletin 34:1142–1151.
Halfpenny, J.C., R.W. Thompson, S.C. Morse, T. Holden, and P. Rezendes. 1995. Snow
tracking. Pp. 91–124, In W.J. Zielinski and T.E. Kucera (Eds.). American Marten, Fisher,
Lynx, and Wolverine: Survey methods for their detection. USDA Forest Service, Pacific
Southwest Research Station, Albany, CA.
Hiller, T.L., D.R. Etter, J.L. Belant, and A.J. Tyre 2011. Factors affecting harvests of
Fishers and American Martens in northern Michigan. Journal of Wildlife Management
75:1399–1405
Northeastern Naturalist
810
T. Kawaguchi, A. Desrochers, and H. Bastien
2015 Vol. 22, No. 4
Jensen, P.G., C.L. Demers, S.A. Mcnulty, W.J. Jakubas, and M.M. Humphries. 2012.
Marten and Fisher responses to fluctuations in prey populations and mast crops in the
northern hardwood forest. Journal of Wildlife Management 76:489–502.
Kapfer, P.M., and K.B. Potts. 2012. Socioeconomic and ecological correlates of Bobcat
harvest in Minnesota. Journal of Wildlife Management 76:237–242.
Keim, J.L., P.D. DeWitt, and S.R. Lele. 2011. Predators choose prey over prey habitats:
Evidence from a Lynx–hare system. Ecological Applications 21:1011–1016.
Klemola, T., E. Korpimäki, K. Norrdahl, M. Tanhuanpaa, and M. Koivula. 1999. Mobility
and habitat utilization of small mustelids in relation to cyclically fluctuating prey abundance.
Annales Zoologici Fennici 36:75–82.
Krebs, C. J. 2011. Of lemmings and Snowshoe Hares: The ecology of northern Canada.
Proceedings of the Royal Society B: Biological Sciences 278:481–489.
Krebs, C.J., R. Boonstra, S. Boutin, and A.R.E. Sinclair. 2001. What drives the 10-year
cycle of Snowshoe Hares? BioScience 51:25–35.
Liebhold, A., W.D. Koenig, and O.N. Bjørnstad. 2004. Spatial synchrony in population
dynamics. Annual Review of Ecology, Evolution, and Systematics 35:467–490.
Lindenmayer, D.B., G.E. Likens, A. Andersen, D. Bowman, C.M. Bull, E. Burns, C.R.
Dickman, A.A. Hoffmann, D.A. Keith, M.J. Liddell, A.J. Lowe, D.J. Metcalfe, S.R.
Phinn, J. Russell-Smith, N. Thurgate, and G.M. Wardle. 2012. Value of long-term ecological
studies. Austral Ecology 37:745–757.
Mowat, G., and B. Slough. 2003. Habitat preference of Canada Lynx through a cycle in
Snowshoe Hare abundance. Canadian Journal of Zoology 81:1736–1745.
Pellikka, J., H. Rita, and H. Lindén. 2005. Monitoring wildlife richness: Finnish applications
based on wildlife-triangle censuses. Annales Zoologici Fennici 42:123–134.
Raphael, M.G. 1994. Techniques for monitoring populations of Fishers and American
Martens. Pp. 224–240, In S.W. Buskirk, A.S. Harestad, M.G. Raphael, and R.A. Powell
(Eds.). Martens, Sables, and Fishers. Comstock Press, Ithaca, NY. 484 pp.
R Development Core Team. 2015. R: A language and environment for statistical computing.
R Foundation for Statistical Computing, Vienna, Austria.
Roberts, N.M., and S.M. Crimmins. 2010. Bobcat population status and management in
North America: Evidence of large-scale population increase. Journal of Fish and Wildlife
Management 1:169–174.
Robitaille, J.-F., and G. Baron.1987. Seasonal changes in the energy budget of captive Ermine,
Mustela erminea L. Canadian Journal of Zoology 65:2864-2871.
Ryan, C. W., J. C. Pack, W. K. Igo, J. C. Rieffenberger, and A. B. Billings. 2004. Relationship
of mast production to big-game harvests in West Virginia. Wildlife Society Bulletin
32:786–794.
Simard, M.A., S.D. Côté, A. Gingras, and T. Coulson. 2012. Tests of density dependence
using indices of relative abundance in a deer population. Oikos 121:1351–1363.
Smith, L.M., I.L. Brisbin, and G.C. White. 1984. An evaluation of total trapline captures
as estimates of furbearer abundance. Journal of Wildlife Management 48:1452–1455.
Thompson, I.D., and P.W. Colgan. 1987. Numerical responses of martens to a food shortage
in north-central Ontario. Journal of Wildlife Management 51:824–835.
Thompson, I.D., and P.W. Colgan. 1994. Marten activity in uncut and logged boreal forests
in Ontario. Journal of Wildlife Management 58:280–288.
Thompson, I.D., I.J. Davidson, S. O’Donnell, and F. Brazeau. 1989. Use of track transects
to measure the relative occurrence of some boreal mammals in uncut forest and regeneration
stands. Canadian Journal of Zoology 67:1816–1823.
Northeastern Naturalist Vol. 22, No. 4
T. Kawaguchi, A. Desrochers, and H. Bastien
2015
811
Venables, W.N., and B.D. Ripley. 2002. Modern Applied Statistics with S. 4th Edition.
Springer, New York, NY.
Vigeant-Langlois, C., and A. Desrochers. 2011. Movements of wintering American Marten
(Martes americana): Relative influences of prey activity and forest-stand age. Canadian
Journal of Forest Research 41:2202–2208.
Weinstein, M.S. 1977. Hares, Lynx, and trappers. The American Naturalist 111:806–808.