IMPACTS OF INDUSTRIAL DEVELOPMENTS ON THE DISTRIBUTION AND
MOVEMENT ECOLOGY OF WOLVES (Canis lupus) AND WOODLAND CARIBOU
(Rangifer tarandus caribou) IN THE SOUTH PEACE REGION OF BRITISH
COLUMBIA

by

ELIZABETH PARR WILLIAMSON-EHLERS
B.Sc., University o f Vermont, 2002

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
IN
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES
(BIOLOGY)

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
April 2012

© Elizabeth P.Williamson-Ehlers, 2012

1+1

Library and Archives
Canada

Bibliotheque et
Archives Canada

Published Heritage
Branch

Direction du
Patrimoine de I'edition

395 Wellington Street
Ottawa ON K1A0N4
Canada

395, rue Wellington
Ottawa ON K1A 0N4
Canada
Your file Votre reference
ISBN:

978-0-494-94130-0

Our file Notre reference
ISBN:

978-0-494-94130-0

NOTICE:

AVIS:

The author has granted a non­
exclusive license allowing Library and
Archives Canada to reproduce,
publish, archive, preserve, conserve,
communicate to the public by
telecommunication or on the Internet,
loan, distrbute and sell theses
worldwide, for commercial or non­
commercial purposes, in microform,
paper, electronic and/or any other
formats.

L'auteur a accorde une licence non exclusive
permettant a la Bibliotheque et Archives
Canada de reproduire, publier, archiver,
sauvegarder, conserver, transmettre au public
par telecommunication ou par I'lnternet, preter,
distribuer et vendre des theses partout dans le
monde, a des fins commerciales ou autres, sur
support microforme, papier, electronique et/ou
autres formats.

The author retains copyright
ownership and moral rights in this
thesis. Neither the thesis nor
substantial extracts from it may be
printed or otherwise reproduced
without the author's permission.

L'auteur conserve la propriete du droit d'auteur
et des droits moraux qui protege cette these. Ni
la these ni des extraits substantiels de celle-ci
ne doivent etre imprimes ou autrement
reproduits sans son autorisation.

In compliance with the Canadian
Privacy Act some supporting forms
may have been removed from this
thesis.

Conform em ent a la loi canadienne sur la
protection de la vie privee, quelques
formulaires secondaires ont ete enleves de
cette these.

W hile these forms may be included
in the document page count, their
removal does not represent any loss
of content from the thesis.

Bien que ces formulaires aient inclus dans
la pagination, il n'y aura aucun contenu
manquant.

Canada

Abstract
Habitat alterations from anthropogenic disturbances across northeastern British
Columbia have resulted in large-scale modifications to predator-prey dynamics. I used GPS
collar locations and field data to quantify the responses o f wolves (Canis lupus) and
woodland caribou (Rangifer tarandus caribou) to the cumulative effects o f industrial
disturbance. I developed seasonal resource selection functions for caribou and count models
of habitat occupancy for wolves. I also related wolf movements to caribou habitat and
industrial features. Caribou occupying the boreal forest likely are more at risk from
industrial developments. My results suggest that caribou occupying these ecosystems are
subject to disturbance by human activity and a greater risk of spatial interactions with
wolves. However, these relationships are complicated by the positive and negative responses
of wolves to landscape change and the distribution o f other prey and predator species.

Table of Contents
Abstract............................................................................................................................................. i
Table of Contents............................................................................................................................ ii
List o f Figures.................................................................................................................................iv
List o f Tables................................................................................................................................ vii
List o f Appendices......................................................................................................................... ix
Acknowledgements......................................................................................................................... x
Chapter 1: General Research Introduction.................................................................................... 1
Organization of Thesis................................................................................................................... 6
Study Area..................................................................................................................................... 6
Woodland Caribou and Wolf Location Data.................................................................................. 9
Anthropogenic Disturbances in the South Peace Region of British Columbia............................... 12
Chapter 2: Effects of Anthropogenic Landscape Change on W olf (Canis lupis) and
Woodland Caribou (Rangifer tarandus caribou) Distribution.................................................. 16
Introduction.................................................................................................................................... 17
Methods.......................................................................................................................................... 20
Study Animals............................................................................................................................. 20
Defining Seasons......................................................................................................................... 22
Distribution of Caribou: Resource Selection Functions.................................................................22
Distribution of Wolves: Count Models.........................................................................................24
Results............................................................................................................................................ 35
Resource Selection by Season for Caribou....................................................................................38
Count Models for Wolves.............................................................................................................51
Discussion......................................................................................................................................60
Habitat Selection by Caribou and Wolves.....................................................................................61
Behavioural Responses of Wolves and Caribou to Industrial Disturbances................................... 64
Cumulative Effects of Resource Extraction and Development on Wolves and Caribou................ 69
Chapter 3: Movement Ecology of Wolves inan Industrialized Landscape.............................. 72
Introduction....................................................................................................................................73
Methods.......................................................................................................................................... 76
Study Area and Wolf Telemetry................................................................................................... 76
Defining Seasons......................................................................................................................... 77
Movement Paths, Rates, and Sinuosity.........................................................................................78
Resource and Human Disturbance Variables................................................................................ 79

Results............................................................................................................................................ 84
Discussion......................................................................................................................................96
Chapter 4: General Research Summary....................................................................................105
Literature C ited............................................................................................................................118
Appendix A.................................................................................................................................. 129
Appendix B .................................................................................................................................. 137
Appendix C .................................................................................................................................. 142
Appendix D .................................................................................................................................. 147
Appendix E .................................................................................................................................. 158

List of Figures
Figure 1. Locations from GPS collared wolves (symbols) and minimum convex
polygons (95% MCPs) for woodland caribou representing their current distribution
across the South Peace region of northeastern British Columbia. Distribution of
caribou includes all locations from members of the Quintette (n = 22) and
Bearhole/Redwillow (BHRW; n = 5) herds collected between April 2003 and
August 2009. W olf distribution includes all locations collected from wolves in five
packs (n = 16) between December 2007 and March 2010.......................................................... 7
Figure 2. Seasonal distribution of Quintette caribou (2003 - 2009) across the
South Peace region o f northeastern British Columbia................................................................ 10
Figure 3. Seasonal distribution of Bearhole/Redwillow (BHRW) caribou
(2007 - 2009) across the South Peace region of northeastern British Columbia......................11
Figure 4. Distribution of Quintette caribou (2003 - 2009) and three packs of wolves
(Upper Sukunka, Upper Murray and Onion Creek; 2008 - 2009) during the spring
season (April 1 - May 14) across the South Peace region of northeastern
British Columbia............................................................................................................................13
Figure 5. Minimum convex polygons (100% MCP) representing the area o f use (AOU)
associated with each o f 10 kill sites for members of the Chain Lakes wolf pack
(2008 - 2010) in the South Peace region of northeastern British Columbia............................26
Figure 6. A grid map o f habitat selection units (HSUs) developed from the average
area o f use (AOU) for collared members of the Chain Lakes wolf pack in the South
Peace region of northeastern British Columbia. Sizes o f HSU cells were determined as
the average wolf(s) area of use (AOU) affiliated with kill sites identified throughout
the territory.....................................................................................................................................27
Figure 7. Individual habitat selection units (HSUs) for the Chain Lakes wolf pack.
Random points were systematically generated for extracting habitat variables,
selection value of caribou habitat and disturbance attributes across each territory
for wolf packs in the South Peace region of northeastern British Columbia............................ 33
Figure 8. The percentage (%) o f used and available locations occurring within
each class of forest cover during the spring and calving seasons for caribou in the
Bearhole/Redwillow (BHRW) and Quintette herds. Model covariates for forest
cover are described in Table 1. An asterisk (*) indicates a forest cover class with
greater than 5% use by caribou.................................................................................................... 39

Figure 9. The percentage (%) of used and available locations occurring with each
class of forest cover during the summer/fall and winter seasons for caribou in the
Bearhole/Redwillow (BHRW) and Quintette herds. Model covariates for forest
cover are described in Table 1. An asterisk (*) indicates a forest cover class with
greater than 5% use by caribou.................................................................................................... 40
Figure 10. Coefficients for the parameters in the most parsimonious resource-selection
models for Bearhole/Redwillow (A; n = 3,401) and Quintette (B; n = 9,791) caribou
herds during the spring season. An asterisk (*) indicates a Gaussian term and variable
descriptions are given in Table 1................................................................................................. 42
Figure 11. Likelihood of occurrence of monitored caribou in the Quintette herd during
the calving season relative to the density o f forestry features (cutblocks and roads)
found across the South Peace region of northeastern British Columbia
(2003 - 2009). Habitat covariates were held at their mean values, while caribou
occurrence was allowed to vary with density of disturbance features...................................... 44
Figure 12. Likelihood of occurrence of monitored caribou in the Quintette herd during
the winter season relative to forestry cutblocks found across the South Peace region
of northeastern British Columbia (2003 —2009). Habitat covariates were held at their
mean values, while caribou occurrence was allowed to vary with distance from
disturbance features......................................................................................................................44
Figure 13. Coefficients for the parameters in the most parsimonious resource-selection
models for Bearhole/Redwillow (A; n = 2,200) and Quintette (B; n = 5,868) caribou
herds during the calving season. An asterisk (*) indicates a Gaussian term and
variable descriptions are given in Table 1.................................................................................. 45
Figure 14. Coefficients for the parameters in the most parsimonious resource-selection
models for Bearhole/Redwillow (A; n = 8,669) and Quintette (B; n = 22,458) caribou
herds during the summer/fall season. An asterisk (*) indicates a Gaussian term and
variable descriptions are given in Table 1.................................................................................. 48
Figure 15. Coefficients for the parameters in the most parsimonious
resource-selection models for Bearhole/Redwillow (A; n = 11,625) and Quintette
(B; n = 28,368) caribou herds during the winter season. An asterisk (*) indicates a
Gaussian term and variable descriptions are given in Table 1................................................... 50
Figure 16. Prey selection (%) by GPS collared wolves as identified through the
investigation of location clusters (2008 - 2010; n = 73 kills) across the South Peace
region of northeastern British Columbia......................................................................................52

Figure 17. Differences in the observed (withheld data) and predicted probability of
counts of wolf locations within habitat selection units (HSUs) for the Upper
Sukunka (A), Upper Murray (B), Onion Creek (C) and Chain Lakes (D) packs residing
in the South Peace region of northeastern British Columbia. Predicted data were
generated from the most parsimonious zero-inflated (ZINB) or negative binomial
(NBRM) regression model (Table 5). A value of zero indicated perfect prediction,
whereas positive values indicated under-prediction and negative values indicated
over-prediction.............................................................................................................................. 54
Figure 18. Mean monthly (± SE) movement rates (km/day) and sinuosity for wolf
movement paths sampled daily across the South Peace region o f northeastern British
Columbia. Movement paths were pooled for wolves by year (A, B) as well as across
all years (C; 2 0 0 8 -2 0 1 0 ).............................................................................................................85
Figure 19. Mean (± SE) monthly (2008 - 2010) movement rates (A, B) and sinuosity
(C, D) for daily (km/day) and weekly (km/week) sampling periods as they relate to
densities of linear (ha/km) and non-linear features (ha/km2) across the South Peace
region of northeastern British Columbia......................................................................................86
Figure 20. Coefficients for the parameters in the most parsimonious mixed-effects
models for daily (A; n = 1,599) and weekly (B; n = 212) movement rates during the
non-winter season for wolves in the South Peace region o f northeastern British
Columbia. An asterisk (*) indicates a Gaussian term and model variables are given
in Table 8.......................................................................................................................................90
Figure 21. Coefficients for the parameters in the most parsimonious mixed-effects
models for daily (A; n = 1,599) and weekly (B; n = 212) sinuosity during the
non-winter season for wolves in the South Peace region o f northeastern British
Columbia. An asterisk (*) indicates a Gaussian term and model variables are
given in Table 8............................................................................................................................ 92
Figure 22. Coefficients for the parameters in the most parsimonious mixed-effects
models for daily (A; n = 1,403) and weekly (B; n = 186) movement rates during the
winter season for wolves in the South Peace region of northeastern British Columbia.
An asterisk (*) indicates a Gaussian term and model variables are given in Table
8...................................................................................................................................................... 93
Figure 23. Coefficients for the parameters in the most parsimonious mixed-effects
models for daily (A; n = 1,403) and weekly (B; n = 186) sinuosity during the winter
season for wolves in the South Peace region o f northeastern British Columbia.
An asterisk (*) indicates a Gaussian term and model variables are given in Table
8...................................................................................................................................................... 95

List of Tables
Table 1. Description of variables used to model habitat selection for both caribou and
wolves across the South Peace region of northeastern British Columbia....................................... 29
Table 2. Statistical models representing hypothesized resource selection strategies of
northern woodland caribou and wolves monitored from 2003 - 2009 in the South Peace
region of northeastern British Columbia. Variables for solar insolation, distance and
density of human disturbances were modeled as either a linear or Gaussian (squared)
term depending on best fit for each season........................................................................................36
Table 3. Number of parameters (k), Akaike’s Information Criterion values (AICc), AICc
weights (AICW), and A AICc values presented for two top-ranked seasonal resource selection
models for members o f both the Bearhole/Redwillow (BHRW) and Quintette caribou
herds monitored from 2003 - 2009 across the South Peace region o f northeastern British
Columbia. Sample size o f caribou locations is presented in parentheses. Model covariates
are given in Table 2..............................................................................................................................37
Table 4. Results of seasonal resource selection function models and the affiliated nonlinear
avoidance distances (Dist; km) and densities (Dens; ha/km2) calculated using Gaussian
covariates for caribou across the South Peace region o f northeastern British Columbia
(Appendix E)........................................................................................................................................43
Table 5. Number of parameters (A), Akaike’s Information Criterion values (AICc), AICc
weights (AICW), and A AICc values for competing seasonal count models for wolves.
Models were developed (using ZINB or NBRM) for each o f four wolf packs monitored
from 2008 - 2010 across the South Peace region o f northeastern British Columbia.
Sample size used to define habitat selection units (HSUs) is presented in parentheses
for each pack. Model covariates are given in Table 2 and the full set o f candidate models
can be found in Appendix E................................................................................................................53
Table 6. Seasonal selection (S) and avoidance (A) of habitat features by wolves across the
South Peace region of northeastern British Columbia. Presence or absence (binary) and
the frequency o f habitat use (count) were determined using p coefficients from count
models. Models were developed for the Upper Sukunka (US; n = 33,599), Upper Murray
(UM; n = 35,959), Onion Creek (OC; n = 10,493) and Chain Lakes (CL; n = 3,389) packs.
Model covariates are given in Table 1 and Table 2.......................................................................... 56

Table 7. Seasonal selection (S) and avoidance (A) of disturbance features by wolves
across the South Peace region of northeastern British Columbia. Presence or absence
(binary) and the frequency of habitat use (count) were determined using (3 coefficients
from count models. Models were developed for the Upper Sukunka (US; n = 33,599),
Upper Murray (UM; n = 35,959), Onion Creek (OC; n = 10,493) and Chain Lakes
(CL; n = 3,389) packs. Model covariates are given in Table 1 and Table 2 .................................58
Table 8. Description of variables used to model movement of wolves across the South
Peace region of northeastern British Colum bia............................................................................... 80
Table 9. Candidate models to examine the movement o f wolves monitored between
2008 - 2010 across the South Peace region o f northeastern British Columbia. Each
model (except Land cover) was fit as either a linear or Gaussian (*squared) term
depending on best fit for each movement parameter and season. Distance was measured
in kilometers (km) and density was measured in hectares/unit area (linear features =
ha/km and non-linear features = ha/km2)...........................................................................................83
Table 10. Number o f parameters (k), Akaike’s Information Criterion (AICc) and AICc
weights (AICW) for linear regression models describing seasonal daily and weekly movement
rates of w o lv e s . Models were developed for wolves monitored between 2008 and 2010
across the South Peace region of northeastern British Columbia. Model covariates are
given in Table 9 and sample size o f seasonal movement paths is indicated in parentheses.........87
Table 11. Number o f parameters (k), Akaike’s Information Criterion (AICc) and AICc
weights (AIC„,) for logistic regression models describing seasonal daily and weekly sinuosity
of wolf movements. Models were developed for wolves monitored between 2008 and
2010 across the South Peace region of northeastern British Columbia. Model covariates
are given in Table 9 and sample size of seasonal movement paths is indicated in
parentheses........................................................................................................................................... 89
Table 12. The predicted and observed variation (f = increased, j = decreased) in
movement using movement rate and path sinuosity as indices of wolf behaviour across
the South Peace region o f northeastern British Columbia. If observed movements were
scale- or season-dependent, results are indicated in parentheses
(seasonal: NW = non-winter, W = winter; scale: daily or weekly).................................................97
Table 13. Hypothetical risk of wolves encountering caribou across the South Peace
region of northeastern British Columbia. Level of risk (low, low-moderate, moderate
or high) is based on the results from the resource selection functions (RSFs) for caribou,
and count and movement models for wolves that quantified the distribution and
movement ecology of GPS-collared animals...................................................................................108

List of Appendices
Appendix A. Seasonal distributions of caribou and wolves across the South Peace
region of northeastern British Columbia....................................................................................129
Appendix B. Fix rate and location error for GPS collars: methods for cleaning data
sets o f erroneous locations for caribou and wolves across the South Peace region of
northeastern British Columbia....................................................................................................137
Appendix C. Field investigations o f kill sites and calculations for areas of use (AOU)
for wolves across the South Peace region of northeastern British Columbia......................... 142
Appendix D. Beta (3) coefficient graphs and use/availability tables for wolves
across the South Peace region of northeastern British Columbia............................................147
Appendix E. Akaike’s Information Criterion values (AICc) and AICc weights (w)
for seasonal resource selection models for caribou and count models for wolves
monitored from 2003 - 2009 across the South Peace region of northeastern
British Columbia..........................................................................................................................158

Acknowledgements
The completion of this project would not have been possible without the considerable
help and support from multiple organizations and people along the way. The Habitat
Conservation Trust Foundation (HCTF), Canadian Association of Petroleum Producers
(CAPP), the BC Ministry of Forests, Lands and Natural Resource Operations, UNBC, West
Fraser Timber Company Ltd., Peace River Coal Ltd., and Western Coal (now under Walter
Energy, Inc.) provided funding for this project.
I thank Dr. Chris Johnson first and foremost, for his confidence, friendship, guidance,
patience, and constant unselfishness that will continue to guide my professional and personal
life. Chris inspired, heightened, and broadened my appreciation and application of
conservation ecology, writing, statistics and academics, all while demonstrating a passion for
loving what you do and the importance of balancing everything in life. I am grateful for the
opportunity to have been a part of his research group and to have learned from such a
talented conservationist. In addition, my advisory committee members, Dr. Dale Seip and
Dr. Kathy Parker, provided me with constructive comments and advice throughout the
duration of my degree. Dale’s expertise on caribou ecology was supreme, as well as his
ability to search for and secure funding to make this project possible. I also thank Dale for
opportunities to go to Kennedy Siding and re-connect with caribou after long hours on the
computer. Kathy Parker’s office was always open; her warmth, expert advice, friendship and
remarkable teaching ability will always be remembered and appreciated (not to mention our
many discussions focused on our mutual love for Jasper and introducing me to the best
cinnamon rolls ever).
I am very fortunate to have experienced caribou and wolf capture and handling with a
truly special and highly respectful trio. Brad and Diane Culling’s passion, trust,
organization, courteous and meticulous capture and handling skills, wit, positive energy and
coaching were impressive during our times in the field. I am grateful for their friendship and
ability to treat me like family during visits to FSJ; their zealous commitment to conservation
is contagious and our winter field days together in the South Peace will never be forgotten.
Greg Altoft’s exceptional piloting skills are unsurpassed. Greg’s excitement for wildlife and
the places outside our backyard o f Prince George were infectious and I always felt safe flying
into the Rockies and into often challenging (in my mind at least) landing spots. Thank you
Greg for going out of your way to help with personal endeavours, including helping Nick
plan one of the most memorable experiences in our lives; we are happy to share so many
fond memories with you.
The UNBC chapter o f The Wildlife Society welcomed my background and provided
me with the opportunity to stay involved and excited about the field of wildlife outside my
thesis. Members of the executive board (2010 - 2011) were great to work with and made my
involvement in the chapter meaningful. Ping Bai, Scott Emmons, and Roger Wheate endured
my challenges and accommodated my struggles along the way pertaining to GIS. Doug

Heard (BC Ministry of Environment), Brian Pate (West Fraser Timber Co.) and Mark
Sharrington (Shell) were willing and eager to help answer any questions I threw their way. I
also want to thank Dr. Mike Gillingham and Elena Jones for their willingness to help me
with statistics and data management.
To the many wonderful people on campus that made my personal life at UNBC and
around Prince George fun, entertaining and rewarding, I will always be thankful. The
Committee of Life (COL) on campus kept life exciting beyond the office and provided hours
of laughter, adventure and knowledge o f Canadian culture. The memories o f sushi nights,
curling, coffee hours, backcountry cabins, potlucks, brewing beer, winter fires, floating the
Nechako, wasting time, playing music and living life to its fullest with great friends in PG
will last a lifetime! Leslie Witter, my only lab mate, made our windowless lab bright by
filling it with talk about the amazing-ness o f caribou and the north, a great cup (ok, lots of
pots) o f coffee, the love of the trails in the Forest for the World and keeping the Bread Guy in
business together; I’m excited for our friendship and adventures to continue. I also thank all
my fellow graduate friends who helped provide great care for my best friends during times I
was called into the field. Furthermore, I want to thank the Yellowstone W olf Project for
being such a dedicated, strong and respectable organization that provided me with the handson research and observational experience I needed to be competitive for this project. A
special thank you goes out to one YWP friend and colleague in particular, who was largely
responsible for saving my life, that fine July day back in 2007; cheers to all our miles,
memories and conversations in the backcountry o f one of the greatest places on Earth.
I thank my incredible posse o f fur balls for walking into my life; Koya, Tiger, and
Takla. Thank you for your continued ability to make me laugh, your constant reminders that
snuggle and play sessions, walks, jogs, hikes, swims and skis are a far superior alternative to
working and your lifelong dedication to teaching me about the unparalleled bonds shared
between humans and animals. You all make my world a happy place everyday.
For their extraordinary support, I share this success with my parents (Stan and
Catharine Williamson), my sisters (Sarah and Kate) and their husbands and families (Russ,
Sadie, James, and Dave, Elle and Sydney). I thank my parents for encouraging me that
anything and everything is possible and for instilling in me an enormous appreciation for
music, nature, travel and beautiful places rooted at an early age from a life full o f visits to our
beautiful cabin in northern Minnesota (in addition to many other places around the world).
My most heartfelt and sincere thank you goes out to my partner in life, Nick. I could
not have completed this project without his support, humour, understanding, patience, and
undivided love at the end of each day (not to mention his excitement to move to Canada).
Nick was the most paramount and knowledgeable field assistant I could have asked for and
made exploring the backcountry of the South Peace region and our time living in Prince
George, truly unique and unforgettable. I very much look forward to sharing our life and our
love of adventures, wild places, family, friends, education, good food, travelling, and
creatures big and small, for all our years to come. jLa salud y el amor a mi alma gemela!

Chapter 1: General Research Introduction

l

Woodland caribou {Rangifer tarandus caribou) populations across North America
have declined since European advancement and colonization (Bergerud 1974). In some
locations, caribou range has contracted northward by roughly 35 km each decade since the
late 1880s (Edmonds 1991, Schaefer 2003, Hummel and Ray 2008). Woodland caribou now
receive considerable conservation attention across the western provinces, and throughout
much of boreal Canada. Habitat alteration and disturbance resulting from human
developments and predation, as an indirect effect of development activities, are thought to
contribute to the cross-continent decline o f this Rangifer subspecies (Fuller and Keith 1981,
James et al. 2004, Johnson et al. 2004a, Weclaw and Hudson 2004, Wittmer et al. 2007, StLaurent et al. 2009, Vors and Boyce 2009, DeCesare et al. 2010, Festa-Bianchet et al. 2011,
Hebblewhite 2011, Latham et al. 201 la, b). Anthropogenic disturbances are widespread
across portions of eastern British Columbia (BC) and caribou herds in these regions are listed
as threatened under the federal Species at Risk Act (SARA; Festa-Bianchet et al. 2011).
In BC, biologists and resource managers recognize three ecotypes o f woodland
caribou: mountain, northern and boreal (Heard and Vagt 1998). Mountain caribou range
across forests in subalpine and alpine habitats in the central and southeastern portions o f the
province. During winter, these caribou forage on abundant arboreal lichens {Bryoria spp.
and Alectoria sarmentosa) as deep snow restricts access to terrestrial lichens or vascular
plants (Stevenson and Hatler 1985, Jones et al. 2007). For these caribou, moving to higher
elevations in winter is an effective strategy for accessing forage and avoiding predators (Seip
1991, Seip and Cichowski 1996).
Caribou of the northern ecotype are found in mountainous and valley habitats
throughout central and northern BC. Northern caribou have highly variable wintering

2

strategies between years, populations and individuals; some caribou winter on high, wind­
swept alpine ridges, while others winter in lower-elevation pine-lichen forests (Bergerud
1978, Terry and Wood 1999, Johnson et al. 2002b). During winter, these caribou forage on
terrestrial lichens (Cladina mitis, Cetraria spp. and Cladonia spp.) that are found in pine
forests or wind-swept alpine habitats (Heard and Vagt 1998, Johnson et al. 2004a, Jones et al.
2007). Depending on snow conditions, northern caribou also forage on arboreal lichens
(Bryoria spp.) during the winter months (Johnson et al. 2004a).
The boreal ecotype o f caribou is found in the northeastern portion of the province and
prefers black spruce (Picea mariana) fen/bog complexes, and tends to avoid well-drained
areas (Bradshaw et al. 1995, Stuart-Smith et al. 1997, Rettie and Messier 2000, Dzus 2001,
Culling et al. 2006). A lack of topographic relief prevents boreal caribou from making
elevational migrations as demonstrated by the mountain and northern ecotypes (Stuart-Smith
et al. 1997, Culling et al. 2006). Ground lichens (C. stellaris, C. mitis and C. rangiferina) are
the dominant food source in winter (Bradshaw et al. 1995). Boreal caribou now occupy less
than half o f their historical range across the continent (Schaefer 2003).
Gray wolves (Canis lupus) once ranged throughout the northern hemisphere at
latitudes north of 15° - 20°N (Young and Goldman 1944, Nowak 1983, Mech and Boitani
2003, Paquet and Carbyn 2003). An increasing human population and the expansion and
advancement of agriculture in the late 1800s served as the catalyst for the general decline of
the gray wolf in North America. During that time, increased harvest of ungulates also
contributed to reductions in the distribution o f wolf populations (Paquet and Carbyn 2003).
In addition, predator control was implemented in the early 1900s, which led to wolf
eradication and extirpation from the western United States and neighbouring locations in

3

Canada (Paquet and Carbyn 2003). In southwestern Canada, wolves increased in number
between 1930 and 1950 as they responded to relaxed predator control programs and more
restrictive regulations for big game hunting which led to an expansion o f ungulate
populations (Nowak 1983, Gunson 1995).
Recent studies in BC and Alberta have demonstrated that roads, trails, geophysical
exploration lines, pipelines, electrical right-of-ways, cutblocks and oil and gas wells can alter
the movements, distributions and population dynamics of both caribou and wolves. Timber
harvesting is one of the primary agents o f habitat change. Large-scale harvesting reduces the
amount of habitat for caribou and increases the area of early-succession forests favoured by
moose and other ungulate species (Fuller and Keith 1981, Rempel et al. 1997, Schaefer 2003,
Johnson et al 2004a, Nitschke 2008). Linear features have resulted in negative impacts for
caribou, including increased human hunting, vehicle collisions, habitat reduction and
predation from enhanced encounter opportunities (Thurber et al. 1994, James and StuartSmith 2000, Dyer et al. 2002, Latham et al. 201 lc). Linear features have the ability to
change predator-prey dynamics by creating efficient travel routes for wolves and increasing
access to habitats used by caribou (Dyer et al. 2002, McCutchen 2007, Rinaldi 2010).
Landscape change and an increase in the abundance of other ungulate species now
limit the ability of caribou to effectively space-away from predators such as the wolf
(Rempel et al. 1997, Wittmer 2004, Latham 2009). Since the early 1900s, moose {Alces
alces) have expanded their distribution throughout BC resulting in a numerical and
distributional response by wolves (Bergerud and Elliot 1986, Spalding 1990, Seip 1992).
Known as “apparent competition”, deer and moose do not compete directly with caribou for
forage or space, but support larger numbers of wolves that prey on caribou opportunistically

4

(Holt 1977, DeCesare et al. 2010). Apparent competition is an important limiting factor for
many populations of woodland caribou in BC (Seip 1992, Hatter et al. 2002, Wittmer et al.
2005).
To conserve declining populations and manage the predators that historically co­
existed with caribou, land-use planners, biologists, and resource managers require
information that reveals how landscape change influences predator-prey dynamics. Such
information is essential in the South Peace region where there are increasing rates of
development for timber and coal reserves, natural gas deposits and wind energy. In addition,
there have been few studies of woodland caribou or gray wolves across that region. My
study investigated both the spatial dynamics and movement ecology of wolves in relation to
caribou and the presence and density o f industrial developments. I focused my research on
two broad themes. First, I investigated the spatial co-occurrence o f collared wolves and
caribou relative to habitat and disturbance factors. Second, I explored how wolves used
industrial features and disturbances when moving across the South Peace landscape. In the
context of those themes, I addressed two specific study objectives:
1)

to quantify seasonal selection or avoidance o f habitat and disturbance features for two
herds of woodland caribou using resource selection functions (RSFs) and four packs
of wolves using a count model based on biological sampling units, and

2)

to quantify movement parameters for wolves as they relate to a) cumulative effects
from human-caused disturbances at two scales, and b) the inferred distribution of
caribou.

5

Organization o f Thesis

I organized the thesis as two separate chapters to be submitted for journal publication,
followed by a final chapter summarizing the implications of my study findings. The portion
of my research that addressed resource selection by caribou and spatial dynamics o f wolves
across landscapes modified by human-caused developments (Objective 1) is presented in
Chapter 2. In Chapter 3 ,1 present methods and results that relate the presence and density of
industrial features and caribou habitat to seasonal movement behaviours of wolves. For
those analyses, movement behaviour is represented by the rate and sinuosity of the
movement paths o f monitored wolves (Objective 2). In the final chapter (Chapter 4), I
summarize findings and present the implications of my research for the conservation of
woodland caribou in the context of wolf distribution, predation behaviour, and development
practices across the South Peace region o f northeastern BC.

Study Area
The study area is located on the eastern slopes o f the Rocky Mountains and
encompasses approximately 12,000 km2 (Figure 1). Tumbler Ridge is located near the center
of the study area, which then extends northwest towards the town o f Mackenzie, northeast
towards Dawson Creek and south along the Alberta border. Four Biogeoclimatic Ecosystem
Classification (BEC) zones occur within that area (Sopuck 1985, Meidinger and Pojar 1991).
The Boreal White and Black Spruce (BWBS) zone is found at elevations ranging between
230 - 1300 m, with the majority of the BWBS occurring above 600 m (DeLong et al. 1991).
Air masses from the Arctic occur in frequent bursts, accounting for long, cold winters.

6

Chetwyftd

□

Quintette

I

I Bearhole/Redwillow

Rivers

Upper Murray (green circle)
+

Onion Creek (purple plus)

*

Lower Sukunka (blue star)

Chain Lakes (orange triangle)

o

Upper Sukunka (red diamond)

C ontours

o 5 to

Lake/W ater Body

j.

N
ao
KMomf ft

30

«

A

Figure 1. Locations from GPS collared wolves (symbols) and minimum convex polygons (95%
MCPs) for woodland caribou representing their current distribution across the South Peace region of
northeastern British Columbia. Distribution of caribou includes all locations from members of the
Quintette (n = 22) and Bearhole/Redwillow (BHRW; n = 5) herds collected between April 2003 and
August 2009. Wolf distribution includes all locations collected from wolves in five packs (n = 16)
between December 2007 and March 2010.
7

In the southern range of the BWBS zone, annual precipitation averages between 330 - 570
mm; snowfall accounts for approximately 45% o f the annual total (DeLong et al. 1991).
Prominent tree species within the BWBS include lodgepole pine (Pinus contorta), black
spruce (Picea mariana), white spruce (Picea glauca), trembling aspen (Populus tremuloides)
and cottonwood (Populus balsamifera). Fire disturbance occurs frequently in the BWBS and
therefore, these forests are characterized by a range o f age classes (DeLong et al. 1991).
The Sub-Boreal Spruce (SBS) zone is found throughout the low-elevation valley
bottoms and extends upwards into areas 1200 m in elevation (Meidinger et al. 1991). The
SBS zone receives 600 - 1000 mm of annual precipitation, where 35% falls as snow
(Meidinger et al. 1991). Dominant tree species include subalpine fir (Abies lasiocarpa),
hybrid white spruce (P. engelmannii x glauca), lodgepole pine, trembling aspen, paper birch
(Betula papyrifera) and Douglas-fir (Pseudotsuga menziesii; Meidinger et al. 1991).
The Engelmann Spruce-Subalpine Fir (ESSF) zone occurs across mountainous areas
with elevations between 900 - 1700 m (Coupe et al. 1991). Prominent tree species include
Engelmann spruce (Picea engelmannii) and subalpine fir; as elevation increases, the ESSF
transitions into the Alpine Tundra (AT) zone. At this junction, the ESSF forests become
more open and are characterised by stunted subalpine fir intermixed with alpine meadow.
Because of a lower fire frequency at higher elevations within the ESSF zone, older age
classes o f forest prevail (Coupe et al. 1991).
The AT zone is usually treeless and occurs above the ESSF. Prominent vegetation
includes ground lichens, sedges, mosses, grasses, dwarf shrubs and forbs (Pojar and Stewart
1991). The AT generally occurs at elevations above 2250 m in the southeastern portion of
the province and receives 700 - 3000 mm o f precipitation annually, which mostly (70 - 80%)

8

falls in the form o f snow (Pojar and Stewart 1991). Despite the heavy amounts o f snowfall,
wind blows snow off alpine ridgelines leading to high variability in snow depth.

Woodland Caribou and WolfLocation Data
The analyses I developed for this study were dependent on location data collected
from individual woodland caribou and wolves. I collected GPS location data from caribou
located in the Quintette and Bearhole/Redwillow (BHRW) herds. The Quintette herd is
found at higher elevations to the west o f the boreal forest and winters primarily on
windswept ridgelines in the alpine (Figure 2). Seip and Jones (2011) classified the
population as ‘increasing’ due to low adult mortality (9%) and high calf recruitment (20%).
The Quintette herd of caribou is estimated at 1 7 3 -2 1 8 individuals (Seip and Jones 2011).
The BHRW herd remains in the low-elevation boreal forests during winter (Figure 3).
From this point forward, I use the term “boreal” to broadly refer to the portion of lowelevation habitat occupied by caribou in the BHRW herd (although located on the same
latitude, this landscape lies just beyond the western edge boundary classified as the boreal
zone and is therefore, classified as hemiboreal; Brandt 2009). Seasonal movements of
caribou in the BHRW herd into the western and southwestern mountains suggest that
seasonally, some caribou may use mountainous habitats typically associated with the
Quintette herd (Figure 3, Appendix A). Unlike the Quintette herd, the BHRW herd is
classified as ‘declining’ due to high adult mortality (25%) and a low level o f calf recruitment
(10%); in 2008, the BHRW population was estimated at a minimum of 49 individuals (Seip
and Jones 2011).

9

a

Spring G P S Collar Locations (Apr 1 - May 14)

^

Calving G P S Collar Locations (May 15 - Jun 14)

*

Summer/Fall G P S Collar Locations (Jun 15 - Oct 31) L U

*

W inter G P S Collar Locations (Nov 1 - Mar 31)

Contours

-

Rivers

Lakes
n

Figure 2. Seasonal distribution of Quintette caribou (2003 - 2009) across the South Peace
region o f northeastern British Columbia.

10

m

*

Spring G PS Collar Locations (Apr 1 - May 14)

Contours

a

Calving G PS Collar Locations (May 15 - Jun 14)

Rivers

*

Summer/Fall G PS Collar Locations (Jun 15 - Oct 31)

Lakes

±

Winter G PS Collar Locations (Nov 1 - Mar 31)

0 25 5

*0

IS

20

Kilom eter*

Figure 3. Seasonal distribution of Bearhole/Redwillow (BHRW) caribou (2007 - 2009)
across the South Peace region of northeastern British Columbia.

11

I collected location and kill-site data from individual wolves within five packs (Figure 1).
The Lower Sukunka pack does not have territory overlap with caribou in either study herd,
whereas the Upper Sukunka pack resides further southeast and has a territory that coincides
with the annual range of the Quintette caribou herd (e.g., Figure 4, Appendix A). The
distribution of the Upper Murray and Onion Creek packs overlaps with habitats used by both
Quintette and BHRW caribou (Appendix A). The range o f the Chain Lakes pack overlaps
completely with caribou in the BHRW herd (Appendix A). The BHRW and Quintette herds
of caribou, in addition to each pack of wolves across the study area, are exposed to various
levels of disturbance resulting from logging, mining, extensive oil and gas (and currently
wind) exploration and extraction.

Anthropogenic Disturbances in the South Peace Region o f British Columbia
Since the early 1990s, eastern BC and western Alberta have experienced rapid landuse change from resource extraction activities, such as the exploration and development of
oil and gas reserves, in addition to large-scale commercial forestry, agriculture, mining, and
most recently, wind power (Sopuck 1985, Schneider et al. 2003). The cumulative effects
from these activities have produced forested landscapes that are progressively younger and
increasingly fragmented (Schneider et al. 2003). Nitschke (2008) reported that resource
development accounted for an 89% increase in edge habitats, a 67% increase in areas
containing habitats of early serai stages, and a 47% increase in the amount o f open
landscapes.
The Tumbler Ridge area has served as the center of the expanding energy sector in
central BC.

12

//'

'

...

*«. \
S '.

TUmbfer'&dge

\

*

>■;

a

Quintette 2009

a

Quintette 2008

«

Upper Sukunka 2009

• Upper Sukunka 2008

Rivers

•

Upper Murray 2009

■ Upper Murray 2008

Lakes

Onion Creek 2009

* Onion Creek 2008

a

Quintette 2003 - 2007

Contours

A
Figure 4. Distribution of Quintette caribou (2003 - 2009) and three packs o f wolves (Upper
Sukunka, Upper Murray and Onion Creek; 2008 - 2009) during the spring season (April 1 May 14) across the South Peace region o f northeastern British Columbia.

13

In August 2009, an Oil and Gas stimulus package was established by the provincial
government to boost the economy, make the oil and gas industry more competitive, and
attract new investors (Proulx 2010). Included in the stimulus package was a $50 million
allotment to invest in the development of affiliated infrastructure (i.e., roads and pipelines).
During 2009, there were new applications granted for 71 well sites, 63 pipelines, and 5
seismic lines located within 100 km o f Tumbler Ridge (Proulx 2010).
Two open-pit coal mines that specialize in the extraction of metallurgical (coking)
coal are found within the core winter range of the Quintette caribou herd. The Wolverine
mine has extracted approximately 2 million tons of coal annually since beginning production
in 2005. A proposed expansion would include the EB and Herman mines (2013) and is in the
early stages o f development. The Trend mine began extracting coal in December o f 2005
and is estimated to produce up to 2 million tons of coal annually through 2015. The Roman
mine (2013), an expansion project adjacent to Trend, would be located just to the south of the
existing facility and would more than double annual production of coal between the two
facilities. In addition, Teck Resources began a feasibility study in the fall of 2010 for the re­
opening o f the Quintette coal mine, a mine that has been dormant since 2000. If approved,
the Quintette mine would be in full operation by 2013, producing up to 3 million tons o f coal
per year. As of January of 2011, excavation processes had already begun in alpine areas that
provide core winter habitat for populations of caribou in the Quintette herd (B. Culling,
personal communication).
The development of wind energy is increasing across the South Peace region. As of
February 2010, tenures in the form of Investigative Use Permits (IUPs) were granted for
most ridgelines and mountain tops within an 85-km radius of Tumbler Ridge. Seven wind

14

projects, totalling more than $3 billion worth o f investment, are currently approved or
undergoing approval processes for initial phases of construction that would begin as early as
2012 (Finevara Wind Energy Inc., Capital Power Corp.). Although facilities related to wind
extraction were not included in my analyses, the ridgeline locations and associated road
networks required for the construction of wind turbines can include habitat for caribou and
should be recognized when considering the cumulative effects of future developments.
Large-scale forest harvesting has occurred in this region since the early 1970s (B.
Pate, personal communication). In 2010, the provincial government awarded Tumbler Ridge
a Community Forest Agreement. This 25-year agreement allows for an annual harvest of
20,000 m3 of timber. Initial stages o f planning are also underway to construct a
manufacturing facility for wood pellets. Once built, 600,000 m3 of wood biomass will be
required annually to supply this facility (Proulx 2010). Across the South Peace region,
forestry companies operate at 100% capacity during the winter (Nov. 15 ~ April 7, West
Fraser Timber C o personal communication).
Cumulative impacts related to activities associated with human development result in
negative consequences for populations o f wildlife (Johnson et al. 2005, Wamback and
Hilding-Rydevik 2009, Johnson and St-Laurent 2011). Large-scale exploration and
development of natural resources can lead to compounding instabilities for populations of
caribou: displacement from portions o f their range, increased movement and vigilance, and
altered predator-prey dynamics (Bradshaw et al. 1997, Nellemann and Cameron 1998,
Cameron et al. 2005, Faille et al. 2010, Latham et al. 201 la). Furthermore, these
relationships are complex and may be confounded by ecological sinks, non-linear responses
to certain feature types and time-lag effects.

15

Chapter 2: Effects of Anthropogenic Landscape Change on W olf (Cattis lupis) and
Woodland Caribou (Rangifer tarandus caribou) Distribution

16

Introduction
Understanding the distribution of organisms is fundamental to conservation. Where
the dynamics between predators and their prey are altered by industrial development, a better
understanding of the changing distribution of populations through space and time can provide
guidance for recovery and conservation efforts (Johnson and St-Laurent 2011). In general,
predators orient themselves to areas where there are greater densities o f prey, whereas prey
avoid areas with increased predation risk while attempting to meet nutritional requirements
(Sih 1984, Lima and Dill 1990, Rettie and Messier 2000). Industrial development can
influence that relationship. Loss of contiguous habitat, disturbance, and the generation of
linear corridors can alter the density and distribution of both species. For example, the
alteration o f habitats resulting from human developments can change the quantity and quality
of vegetation and force prey to concentrate in space facilitating predation (Dzus 2001,
Schlaepfer et al. 2002). Also increasing rates of predation, linear corridors can seasonally
increase the movement potential of predators that interact with low-density prey populations
(Bergerud et al. 1984, Jalkotzy et al. 1997, James and Stuart-Smith 2000, McKenzie 2006,
Latham et al. 201 lc).
Changing patterns of land use over the past 100 years have altered the relationships
among wolves, woodland caribou and other prey species. Wolves, a generalist species, now
serve as a primary predator of caribou (Bergerud and Elliot 1986, Seip 1992). Regenerating
forests resulting from human developments favour higher densities o f moose, elk, and deer.
This alteration in landscape composition facilitates a broader distribution of wolves and
increasing opportunities to use caribou as an alternate prey species (Fuller and Keith 1981,
James et al. 2004, Johnson et al. 2004a, Wittmer et al. 2007, DeCesare et al. 2010).

17

Contributing to this dynamic, large-scale resource exploration and extraction can result in a
variety of linear features that occur as narrow paths o f early-successional vegetation
communities. These features can increase the vagility of wolves and provide greater access
into the habitats o f caribou otherwise isolated by topography or vegetation (James and StuartSmith 2000, McCutchen 2007, Rinaldi 2010, Latham et al. 201 lc).
Variation in resource selection by caribou within seasons and across spatial scale may
be a behavioural strategy to decrease predictability to predators (Gustine et al. 2006b). Rettie
and Messier (2000) discussed the behavioural and distributional implications of predation
and hypothesised that caribou should respond to the most important limiting factor at the
scale o f the landscape and to less important factors at progressively smaller spatial scales. In
the case o f caribou populations limited by predators (Bergerud and Elliot 1986, Seip 1991,
Wittmer et al. 2005, Latham et al. 201 lb), this would involve selecting large areas with a
relatively lower risk of predation. Past research across BC and western Alberta has shown
that woodland caribou demonstrate variable distribution strategies across seasons and years.
For example, caribou of the same population will winter on high, windswept alpine ridges,
while others will winter in lower-elevation pine-lichen forests (Cichowski 1993, Terry and
Wood 1999, Johnson et al. 2004a, Jones et al. 2007). Johnson et al. (2002b) suggested that at
the patch scale forage was a more important factor than predation risk. Gustine et al. (2006b)
also found that only at larger spatial scales did caribou significantly increase their distance
from wolves.
Similar to caribou, wolves have the ability to adapt to local conditions including
spatial and temporal variation in prey availability (Mladenoff et al. 1999, Paquet and Carbyn
2003, Latham et al. 201 la). W olf distribution does, however, depend on landscape

18

conditions characterized by low densities of humans and active roads. For example,
Mladenoff et al. (1995) reported that areas occupied by wolves in the Northern Great Lakes
region had a much lower density of roads when compared to areas not used by wolves.
Packs selected forested habitats dominated by conifers on public land relative to areas with a
higher density of agricultural development (Mladenoff et al. 1995). In Italy, Corsi et al.
(1999) found wolves to be absent in areas supporting higher human and road densities and
greater levels of cultivation. Consistent with previous studies, Whittington et al. (2005) also
found that wolves in Alberta avoided areas o f high road and trail density, but selected lowuse roads and trails as travel corridors.
Despite the vast number of studies linking wolf occurrence with road density and
level of human use, there is still a need to better understand the behaviour and distribution of
wolves in areas where cumulative anthropogenic disturbances might influence predator-prey
dynamics (Nitschke 2008, Houle et al. 2010). Many studies o f wolf movement have
occurred in landscapes exposed to higher densities of and longer-term use by humans.
Although a correlation exists between wolf occurrence and a low probability o f encountering
humans, this relationship may not hold true where industrial footprints are large and human
densities remain low. Also, past studies that have considered wolf interactions with
industrial development have considered only a few disturbance types, but not the cumulative
effects of multiple types (but see Lesmerises et al. 2012). Considering the co-occurrence o f
caribou and wolves, there is uncertainty about the ability of caribou to adapt to predation risk
in the context of landscape change that includes altered successional dynamics and an
increase in the prevalence of linear corridors. There are unexplored relationships between
cumulative industrial developments and the interacting responses of wolves and caribou.

19

My principal research goal was to better understand the distribution and interactions
of wolves and caribou across a landscape with high levels o f industrial development. I first
developed resource selection functions (RSFs) to determine the seasonal distribution and
habitat selection of two herds o f northern woodland caribou. Then, I used count models to
investigate not only selection, but also the frequency o f wolf occurrence relative to
disturbance features and caribou habitat. Detailed investigations o f both caribou and wolf
habitat ecology serve as a foundation for increasing our knowledge o f the spatial and
temporal relationships of these two species. Such insights may also apply to other species
influenced by increasing human disturbances and apparent competition (Robinson et al.
2002, Kristan and Boarman 2003, Baldi et al. 2004, Bryant and Page 2005, Gibson et al.
2006). Understanding the spatial complexities of co-occurring populations can aid in
conservation planning for the long-term persistence of threatened species.

Methods
Study Animals
Woodland Caribou
A total of 27 caribou within two herds (Bearhole/Redwillow (BHRW) = 5, Quintette
= 22) were captured between February 2003 and March 2009 by net-gunning from a
helicopter. Caribou were fitted with either Televilt (n = 4; Televilt, TVP Positioning AB,
Bandygatan 2, SE-71134 Lindesberg, Sweden, Model: GPS-VHF remote download) or ATS
(n = 21; Advanced Telemetry System, 470 First Ave. N., Box 398, Isanti, Minnesota, USA,
Model: GPS Remote-Release Collar) GPS collars equipped with VHF transmitters and
remote-release devices. Televilt GPS collars were programmed to take fixes every four

20

hours and locations were downloaded remotely. All four Televilt GPS collars failed to
function as programmed and, therefore, each dataset was incomplete; animals were re­
captured and refitted with either a VHF (n = 1) or ATS GPS collar (n = 3). ATS collars were
programmed to take location fixes every 20 hours up until 2005; collars programmed after
April 2005 acquired fixes between two and six times daily. In addition, two female caribou
were captured in the study area in 2007 and collared with Lotek ARGOS GPS collars (F900
and F901 of the BHRW herd; Lotek Inc., Newmarket, Ontario, Canada). Data acquired from
each GPS collar were examined and screened for erroneous locations using a combination of
methods (Appendix B; Moen et al. 1997, D’Eon et al. 2002, D’Eon and Delparte 2005).
Wolves
Between March 2007 and March 2010, a total of 31 wolves from five packs (Lower
Sukunka, Upper Sukunka, Onion Creek, Upper Murray, and Chain Lakes) were captured
using a tranquilizer dart (Pneu-Dart, Inc. 15223 Route 87 Highway, Williamsport,
Pennsylvania USA, Model: 196 Projector) or net gun deployed from a helicopter. Each wolf
was fitted with either a remotely downloadable GPS (n = 16, Lotek Inc., Newmarket,
Ontario, Canada, model: GPS 4400S) or VHF (n = 15, Lotek) collar. GPS collars were
equipped with VHF transmitters, as well as remote-release devices. Collars were
programmed to take a location fix every three hours (n = 14; two collars were programmed
for high-frequency intervals and collected a location every 20 min) and were remotely
downloaded from a fixed-wing aircraft approximately bimonthly during routine tracking
flights. O f the 31 collared wolves, data from 16 were specific to the study area and used for
analysis. Similar to caribou, wolf data were screened and examined for erroneous locations
(Appendix B).

21

Defining Seasons
Drawing on variation in biology, snow conditions and movement patterns, Sopuck
(1985) and Jones et al. (2007) identified biological seasons for four herds o f caribou found
adjacent to, or within my study area. I used this information to define four primary seasons
for my study o f habitat selection by caribou: spring (April 1 - May 14), calving (May 15 June 14), summer/fall (June 15 - October 31), and winter (November 1 - March 31). Also, I
used past research (Mech 1970, Fuller 1989, Ballard et al. 1991, Kreeger 2003, Mech and
Boitani 2003, Packard 2003) to develop three biological seasons to model the response of
wolves to their surroundings: non-winter (April 16 - October 14), early winter (October 15 January 31) and late winter (February 1 - April 15). Non-winter months include the time
when wolves become responsible for the raising and rearing of pups and therefore, centralize
around dens or homesites (Mech 1970, Ballard et al. 1991). By mid-October, pups are
approximately six-months old and have grown large enough to travel and keep up with the
nomadic pack as they transition towards the winter months (Packard 2003). In North
America, breeding season occurs between late January and early April, depending on
latitude; this marks the transition into late winter (Kreeger 2003). Late winter extends until
the wolves begin localizing around a den site between the months o f March and May (Fuller
1989, Mech and Boitani 2003).

Distribution o f Caribou: Resource Selection Functions
I used resource selection functions (RSFs) to quantify the spatial relationships
between GPS-collared caribou and a number o f variables that were hypothesized to influence
caribou distribution. An RSF is any mathematical function that provides an estimate of
resource use that is proportional to the true probability of use (Manly et al. 2002).
22

Coefficients from RSFs represent selection for or avoidance o f a resource (i.e., habitat or
industrial features). Selection is assumed when an animal uses a resource out o f proportion
to the availability of that resource across some defined area (e.g., home range), or the
distance to a disturbance feature is less for animal observations relative to a comparison set
of random locations. I used GIS to apply RSF coefficients from the top-ranked models to the
corresponding spatial data and produced maps representing the relative value (poor- to highquality) of habitat, by season, across the range o f the Quintette and BHRW caribou herds.
I used a conditional fixed-effects logistic regression to develop the RSFs (Compton et
al. 2002, Manly et al. 2002). Instead of pooling used and available locations, a fixed-effects
logistic regression considers the difference between each used location and the set of
associated random locations. Pairing of used and random locations in space and time
provides a more precise definition of resource availability relative to the seasonal and annual
differences in the distribution of a monitored animal (Johnson et al. 2004b). RSFs estimated
from this style of matched regression were appropriate for my study as caribou have large
home ranges compared to their relocation intervals (Arthur et al. 1996, Compton et al. 2002,
Duchesne et al. 2010). All regression analyses were conducted using STATA (version 9.2,
StataCorp. 2007).
RSFs constructed using conditional logistic regression were dependent on a restricted
spatial domain, representing a specific distance an animal could have travelled during a time
period, for identifying resource availability. I used the programming interval between GPS
locations to define that spatial domain. For this calculation, I centered a circular buffer on
the preceding collar location for each individual study animal (Johnson et al. 2005). This
circle had a radius equivalent to the 95th percentile movement distance for a period o f 24

23

hours. Five comparison locations were then randomly selected from within this spatial and
temporal buffer, defined as the availability radius.
Similar to Johnson et al. (2005), I assumed that caribou would not respond to a
disturbance feature at excessively large distances. Thus, I used the conditional regression to
statistically remove the responses o f individual caribou locations that exceeded a set distance
threshold to individual disturbance features. The threshold was exceeded when the nearest
disturbance feature of a specific type (e.g., coal mine) was found outside the availability
radius for that caribou location. This approach allowed me to model a matched sample of
caribou and random locations based on the effects of habitat, while statistically removing
effects of an ecologically implausible ‘disturbance’ (Johnson et al. 2005).
Caribou were monitored independently throughout the study, but I pooled GPS
locations by herd for each season. Pooling locations forfeited my ability to detect variation
in resource use among individuals. However, pooling locations allowed for a sufficient
sample of relocations to build sets of complex seasonal models.

Distribution o f Wolves: Count Models
I used a statistical model based on counts to relate the number of wolf locations
within a habitat selection unit (HSU) to covariates that represented environmental or
industrial features that might explain the seasonal distribution of wolves. Count models
contained two parts; similar to RSFs, the binary portion o f the count model represented the
probability of occurrence o f wolves, while the count portion represented the relative
frequency of use in areas occupied by wolves (Nielsen et al. 2005, Sawyer et al. 2006).
Therefore, this technique had greater power, relative to the RSFs for caribou, to describe the
differential use of resources by wolves (Nielsen et al. 2005). Where possible, I used zero24

inflated count models to quantify the binary and count portions of the wolf location data. I
used wolf behaviour (i.e., predation) to identify a square sampling unit, the HSU, to model
the relative frequency of wolf locations relative to vegetation, selection value of caribou
habitat as determined from the RSF analysis, and disturbance attributes. Each HSU was
large enough to capture variation in wolf occurrence, as recorded using GPS collars (Sawyer
et al. 2006).
I defined the spatial extent o f the HSU as the average area occupied by wolves after
killing and consuming what was assumed to be a large prey item (e.g., moose, deer, caribou;
Figures 5, 6; Appendix C). During three summers (2008 - 2010), we investigated wolf kill
sites identified from clusters of GPS collar locations distributed throughout each pack
territory. Each cluster represented a grouping of GPS collar locations defined as two or more
consecutive locations within 200 m of one another. To minimize search effort of non-kill
sites (e.g., bed sites, etc.), we investigated clusters containing > four location fixes (four fixes
= 12 hours of time) only. The area o f use (AOU; ha) by collared wolves at each identified
kill site was calculated as the minimum convex polygon (100% MCPs) o f locations that
occurred within a one-week time period surrounding the assumed date o f kill (Figure 5). For
each pack territory, the area of a HSU was calculated as the mean of all AOUs for collared
wolves of that pack (e.g., Figure 6; Appendix C). Kills were identified for each collared wolf
(> 3 per pack) and throughout each pack territory (Appendix C).
Depending on the distribution o f data, count models were premised on the Poisson or
negative binomial distribution (Pielou 1969). I used a likelihood ratio test to check for over­
dispersion and determine if a Poisson (PRM) or negative binomial (NBRM) model was most
appropriate.

25

/

09-438 /

>

10- 00*

Area of U se (AOU)
M oose Kill

N

Rivers

A

Lake/W ater Body 0037876 16

3

Figure 5. Minimum convex polygons (100% MCP) representing the area o f use (AOU)
associated with each o f 10 kill sites for members of the Chain Lakes wolf pack (2008 - 2010)
in the South Peace region of northeastern British Columbia.

26

‘/ I ,

TumMtf Rtdg*

rt r 1

,

/-;• '?
rfb /fe

A rea of U se (AOU) Grid
G P S C ollar L ocations
C o n to u rs
R ivers
L ake/W ater Body

0 15 3

6

9

12

Kilometer*

Figure 6. A grid map of habitat selection units (HSUs) developed from the average area of
use (AOU) for collared members of the Chain Lakes wolf pack in the South Peace region o f
northeastern British Columbia. Sizes of HSU cells were determined as the average wolf(s)
area of use (AOU) affiliated with kill sites identified throughout the territory.

27

Both the PRM and NBRM can under-estimate the occurrence of zero counts. Therefore, I
used a Vuong Test (Vuong 1989) to determine if zero-inflated versions of each model (ZIP
or ZINB) were appropriate. Because data collected from GPS collars were correlated in
space and time, I used the robust option in Stata to adjust standard errors (SE) for an auto­
correlated error structure.

Resource and Human Disturbance Variables
Drawing from past research on wildlife-development interactions and observations of
the study area, I identified a number of resource and human disturbance variables for
modeling the responses o f caribou and wolves to their environments (Table 1). For each
seasonal RSF for caribou, I examined two categorical and multiple continuous variables:
forest cover type (categorical), serai stage of forest (categorical), solar insolation, and
distance to and density o f disturbance features. Human disturbance variables were grouped
by industry type as well as their ability to influence caribou and wolf behaviour across the
landscape: roads, linear features (roads, seismic lines and pipelines combined), forestry
(roads and cutblocks), open-pit operations for coal mining, oil and natural gas exploration
and extraction (mine/oil/gas; non-linear open-pit coal mine footprints, well and facility pads
> 1 ha), and cumulative effects from development features (linear features, forestry, and
mine/oil/gas combined).
I identified six variables that may be important predictors of seasonal wolf
distribution. For each season, I analyzed count models that contained combinations o f forest
cover type (categorical), serai stage o f forest (categorical), selection value o f caribou habitat
in pixel cells determined from the RSF analysis, and distance to and density o f disturbance
features.
28

Table 1. Description of variables used to model habitat selection for both caribou and wolves
across the South Peace region o f northeastern British Columbia.
Variable
Alpine
Blk Spruce
Fir
HBS
Other
Pine
Spruce
Tamarack
Tree Broadleaf
Tree - Other
Upland Nveg
Water
No Age Data
Young
Growing
Mature
Old
RSFBHRW
RSFQ
Solar
Insolation
Road
Seismic Line
Pipeline
SeisPipln
Cutblock
Mine
Oil and
Natural Gas
Water

Description
high elevation with few or no trees with primary cover being rock, snow, herbs,
shrubs, bryoids and terrestrial lichens
black spruce (Picea mariana)
subalpine fir (Abies lasiocarpa)
herbs (forbs, graminoids), bryoids and shrubs
specific to herd and season; combination of variables listed with too few
occurrences to model
lodgepole pine (Pinus contorta) and whitebark pine (P. albicaulis)
other spruce varieties: Picea spp., Engelmann (P. engelmannii), white (P. glauca),
hybrid (P. engelmannii x glauca)
tamarack (Larix laricina)
other non-listed broadleaf trees: aspen (Populus tremuloides), cottonwood (P.
balsamifera) and birch (Betula papyrifera)
other non-listed conifers, Douglas-fir (Pseudotsuga menziesii)
upland areas dominated by talus, rock, snow, tailing ponds, or no additional data
for land cover
lake, reservoir, river, stream or a non-spruce or tamarack dominated wetland
(caribou only)
no data available to determine serai age of forest
forest age 0 < 40 yrs
forest age 41 < 80 yrs
forest age 81 < 120 yrs
forest age > 121 yrs
RSF values for caribou in the Bearhole/Redwillow (BHRW) herd
RSF values for caribou in the Quintette herd
measure of incoming solar radiation on a surface (W/m2)
distance to road (km)
distance to seismic line (km)
distance to pipeline (km)
distance to seismic line and/or pipeline combined (movement models only; km)
distance to forestry cutblock (km)
distance to coal mine footprint (km)
distance to non-linear oil and gas well pad or facility pad > 1 hectare in size (km)
distance to water (wolves only; km)

29

I also tested the importance o f water (proximity) as an additional predictor o f wolf
distribution.
Habitat variables - Forest cover type and serai stage were estimated using the
provincial Vegetation Resource Inventory (VRI; BC Ministry of Forests and Range 2007a,
b). I used existing knowledge o f caribou ecology to consolidate categories o f forest cover
from the VRI into 11 new classes, based on the leading commercial or brush species (Table
1). Similar to forest cover, I categorized serai stage into five age classes based on regimes of
fire disturbance for dominant species in each BEC zone and past research pertaining to
habitat selection and behaviour o f woodland caribou (Medinger and Pojar 1991, Table 1).
Across my study area, VRI data were incomplete for a portion o f alpine-type habitats.
Therefore, I classified age in these ‘no age data’ habitats as late-succession forests (i.e., old).
Categorical variables for forest cover and age class were modeled with deviation coding
(Menard 2002). This method of coding takes individual variables and compares their
deviations to the grand mean across all categories.
Solar insolation - Solar insolation (SI) represented the amount of radiation striking a
surface. I used solar insolation in this study as a proxy of slope and aspect and therefore, as a
potential indicator of forage availability and snow conditions for caribou. Snow melt and
growth of vegetation can occur more rapidly in areas with increased radiation. In addition,
alpine areas that experience higher levels of solar radiation could be indicative of wind­
blown ridgelines that are often ideal habitats for northern woodland caribou in winter. I used
a digital elevation model (DEM 25m x 25m; BC Land and Resource Data Warehouse 2007)
to calculate seasonal averages o f SI in watts per square meter (W/m2) across the South Peace
region for each year caribou locations were collected (2003 - 2009). When mapping RSFs, I

30

used SI values from the most recent year (2009). I chose not to include elevation as a
topographical variable; elevation can often correlate with habitats classified as alpine which
further complicates results and model interpretation.
Disturbance features - 1 used databases from government and industry to identify the
location of roads and forestry cutblocks (BC Land and Resource Data Warehouse 2007; West
Fraser Timber Company). I did not classify roads by use or status. During the period of
monitoring for caribou and wolves, the Wolverine and Trend coal mines were fully
operational and spatial data were acquired directly from their parent corporations (Western
Coal and Peace River Coal Ltd.). This variable representing mines was applied to caribou
(Quintette herd) and wolves (Upper Sukunka, Upper Murray, and Onion Creek) that occurred
within the vicinity of active coal mines. Lastly, I used the Oil and Gas Commission of BC’s
public database, complete through 2009, to identify the spatial locations of seismic lines,
pipelines, well sites and other developed areas related to the exploration and development of
oil and natural gas reserves across the South Peace (http://www.ogc.gov.bc.ca/GIS.asp,
2009).
I calculated the distance (km) from caribou and wolf locations to human disturbance
features as well as the density of disturbance features (total area of features/unit area; linear
features = ha/km, non-linear features = ha/km2) at each animal location using IDRISI (V
15.0, The Andes Edition; Eastman 2006). I used a standard moving-window algorithm to
calculate the density o f disturbance features. I fit RSF models to three sizes of moving
windows (0.56 ha, 1.56 ha and 3.06 ha) and used Akaike’s Information Criterion for small
sample sizes (AICc) and Akaike weights (AICW; see Model Selection and Validation below

31

for more information) to identify the best-fitting moving window size for the analysis of
habitat selection by caribou.
Modeling nonlinear responses - 1 used a Gaussian function to model the nonlinear
responses (if applicable) of caribou or wolves to disturbance features. For each seasonal
model, I used Akaike weights (w) to determine if a linear or Gaussian term was most
appropriate. Where I observed a nonlinear relationship, I determined a threshold value using
the point o f inflection for each disturbance type or class. Values indicating disturbance
thresholds for caribou need to be interpreted cautiously, as these thresholds may be unique to
the South Peace study area, study animals, and/or my chosen method of analyses (e.g.,
logistic regression, size o f availability radius, etc.). A variety of analytical tools are available
to researchers to aid in the definition o f an ecological threshold (e.g., Nielsen et al. 2009,
Leblond et al. 2011), but there remains uncertainty surrounding the ability to correctly
identify these points of change (Ficetola and Denoel 2009).
Distribution of caribou habitat - 1 multiplied coefficients from the most parsimonious
RSF models by the corresponding GIS data layer to generate seasonal maps illustrating the
most strongly selected habitats by collared caribou from the BHRW and Quintette herds. I
used these maps to model the response of wolves to habitats of different value to caribou
across the South Peace region.
Random point generation for count models - When constructing the count models for
wolf location data, I systematically generated random points for each pack territory (e.g.,
Figure 7). I then extracted values for each point and took the median value across each HSU
to quantify habitat class, RSF value of caribou habitat, and distance to or density of
disturbance feature.

32

Hwy 52

H abitat S election Unit (HSU)
R an d o m Points
V , 3117

R ivers
» ♦

L ake/W ater Body

OOtSQ3

08

0»

1

•

V#
3 W 3 \#

# 32Kk

KacntNm

Figure 7. Individual habitat selection units (HSUs) for the Chain Lakes wolf pack. Random
points were systematically generated for extracting habitat variables, selection value of
caribou habitat and disturbance attributes across each territory for wolf packs in the South
Peace region of northeastern British Columbia.

33

Model Selection and Interpretation
I used Akaike’s Information Criterion for small sample sizes (AICc) and Akaike
weights (AIC„.) to identify the most parsimonious model from a suite of ecologically
plausible candidate models for both caribou and wolves (Anderson et al. 2000).I also used
the delta (A) AICc as a measure to compare each candidate to the top-ranked model (i.e., the
model with the lowest AICc; Burnham and Anderson 2002). I reported coefficients (P) from
the most parsimonious model and used 95% confidence intervals to illustrate the precision of
each covariate. For covariates that fell close to or overlapped with 0, selection or avoidance
of habitat or disturbance features could not be determined. I used tolerance scores to assess
collinearity among variables (Menard 2002). Where tolerance scores were less than the
threshold value of 0 .2 ,1 used bivariate correlation and visual inspection o f standard errors to
determine if there was a large effect on model inference. Where collinearity occurred
between disturbance variables, I preferentially retained linear features to better understand
how these disturbances might influence the distribution of caribou and wolves.

Model Validation
I used £-fold cross validation to assess the capability of the most parsimonious RSF
model to predict resource selection by caribou (Boyce et al. 2002). Here, I determined if
there was a Spearman rank correlation (rs) between the predicted RSF values and the
frequency of occurrence of animal locations (Boyce et al. 2002). I also examined the
classification accuracy of top-ranked models by using the more conservative receiver
operating characteristic (ROC) curve. Models demonstrating an area under the ROC curve
(AUC) > 0.7 are thought to perform well, whereas a score o f 1 represents perfect
discrimination between used and available locations (Hosmer and Lemeshow 2000). I

34

generated independent k-fold and AUC scores by withholding approximately 20% of the
animal locations from the model-building process.
For the count models for wolves, I randomly partitioned wolf locations into training
(80%) and testing (20%) groups. Using the withheld data, I determined if there was a
relationship between the observed probabilities o f counts and the predicted probabilities of
counts (prcounts.ado: Long and Freese 2006). As a second measure o f model fit and
prediction, I calculated the unstandardized residuals. Perfect prediction occurred when the
mean residuals for a count class equaled zero, whereas positive values indicated under­
prediction and negative values indicated over-prediction.

Results
I used a total of 38,116 GPS collar locations from members of the
Bearhole/Redwillow (BHRW: 12,297 locations) and Quintette (25,819 locations) caribou
herds to develop 19 seasonal resource-selection models (Table 2). For all four seasons, the
most parsimonious models for both BHRW and Quintette caribou were also the most
complex in each candidate set and contained variables for all habitat and human-caused
disturbances (Table 3). The predictive ability o f the cumulative effects (CE) model for the
BHRW herd ranged from a mean rs= 0.820 in calving to r = 0.981 in winter (AUC = 0.737
and 0.725, respectively).
The most parsimonious model for BHRW in summer/fall demonstrated poor
predictive ability using &-fold cross validation, but the more conservative ROC (AUC =
0.726) implied an acceptable level of discrimination.

35

Table 2. Statistical models representing hypothesized resource selection strategies of
northern woodland caribou and wolves monitored from 2003 - 2009 in the South Peace
region o f northeastern British Columbia. Variables for solar insolation, distance and density
o f human disturbances were modeled as either a linear or Gaussian (squared) term depending
on best fit for each season.
Model name
Forest Cover

Forest Age
Solar Insolation
Landscape
Road Distance (Dist; km)
Road Density (Dens; ha/km2)
Road Dist and Dens
Linear Feature (LF) Dist
Linear Feature (LF) Dens
Linear Feature (LF) Dist and
Dens
Forestry (FOR) Dist
Forestry (FOR) Dens
Forestry (FOR) Dist and Dens
Mine, Oil, and/or Natural Gas
(MOG) Dist

Covariates included in model
Forest cover type (alpine, black spruce, fir, HBS, pine,
spruce, tamarack, broadleaf trees, other trees, upland non­
vegetated, and water)
Forest age class (0 - 4)
Solar insolation (W/m2)
Forest Cover + Forest Age + Solar Insolation
Landscape + Dist to Road
Landscape + Road Dens
Landscape + Dist to Road + Road Dens
Landscape + Dist to LF (road, seismic line and/or pipeline)
Landscape + LF Dens
Landscape + Dist to LF + LF Dens
Landscape + Dist to Cutblock + Dist to Roads
Landscape + Cutblock Dens + Road Dens
Landscape + Dist to Cutblock + Dist to Roads + Cutblock
Dens + Road Dens
Landscape + Dist to MOG + Dist to LF

Mine, Oil, and/or Natural Gas
(MOG) Dens

Landscape + MOG Dens + LF Dens

Mine, Oil, and/or Natural Gas
(MOG) Dist and Dens

Landscape + Dist to MOG + Dist to LF+ MOG Dens + LF
Dens

Cumulative Effects (CE) Dist

Landscape + Dist to LF + Dist to Cutblock + Dist to MOG

Cumulative Effects (CE) Dens
Cumulative Effects (CE) Dist
and Dens

Landscape + LF Dens + Cutblock Dens + MOG Dens
Landscape + Dist to LF + LF Dens + Dist to Cutblock + For
Dens + Dist to MOG + MOG Dens

36

Table 3. Number o f parameters (k), Akaike’s Information Criterion values (AICc), AICc weights (AIC*,), and A AICc values presented
for two top-ranked seasonal resource selection models for members o f the Bearhole/Redwillow (BHRW) and Quintette caribou herds
monitored from 2003 - 2009 across the South Peace region o f northeastern British Columbia. Sample size o f caribou locations is
presented in parentheses. Model covariates are given in Table 2.
BHRW

Spring (n = 3,401)

Sum/Fall (n = 8,669)

Calving (n = 2,200)

Winter (n = 11,625)

Model Covariates

k

AICc

AA1C

AIC„,

k

AICc

AAIC

AIC h

k

AICc

AAIC

AIC h

k

AICc

AAIC

AIC b

CE Dist*

26

3137.2

4.6

0.09

25

1854.6

3.5

0.15

27

7402.6

48.2

<0.001

32

10065.1

65.4

<0.001

CE Dist + CE

30

3132.5

0.0

0.91

27

1851.1

0.0

0.85

31

7354.4

0.0

1.00

34

9999.7

0.0

1.00

Dens*c

Spring (n = 9,791)

Quintette

Winter (n = 28,368)

Sum/Fall (n = 22,458)

Calving (n = 5,868)

Model Covariates

k

AICc

AAIC

AIC„

k

AICc

AAIC

AIC„

k

AICc

AAIC

AIC h

k

AICc

AAIC

AIC„

CE Dist*

34

5466.2

10.6

<0.001

35

4102.0

6.3

0.04

30

15933.3

30.1

<0.001

34

19949.3

275.8

<0.001

CE Dist* + CE

37

5455.5

0.0

1.00

38

4095.8

0.0

0.96

33

15903.2

0.0

1.00

39

19673.6

0.0

1.00

Dens
“Gaussian (squared) term was most parsimonious in at least one seasonal candidate model
*Linear term was most parsimonious in at least one seasonal candidate model

For Quintette caribou, models containing covariates for cumulative effects also predicted
well from a mean rs = 0.893 in summer/fall to a mean rs = 0.984 in spring (AUC = 0.815 and
0.866). Models excluding disturbance features failed to enter into the final top-model set for
each season. Collared caribou demonstrated clear differences in the use o f forest cover,
forest age, solar insolation, and disturbance features across the South Peace study area.

Resource Selection by Season fo r Caribou
Differences in selection for forest cover and age were evident between the two
caribou herds for each season (Figures 8, 9). In general, the BHRW herd selected for lowelevation habitats across the boreal forest dominated by black spruce (all four seasons) and
tamarack. Caribou that overwintered in the boreal forest selected older pine-leading stands in
addition to black spruce and tamarack. A proportion o f the BHRW herd was observed
migrating to higher elevations during the calving and summer/fall seasons and selected
subalpine fir and alpine habitats, in addition to habitats dominated by herbs, bryoids, and
shrubs (Figures 8, 9; Appendix A). Quintette caribou remained at high-elevations and
selected alpine, subalpine fir, spruce and pine-leading habitats of late-succession throughout
the year. Caribou in both the BHRW and Quintette herds selected for areas with greater
levels of solar insolation (SI) across each o f the four seasons. Seasonal co-occurrence o f
caribou and wolves was greatest near forestry cutblocks and in areas with increased densities
of disturbance features, as both species were observed selecting for these disturbance types.
However, these interactions varied for both the BHRW and Quintette herds o f caribou
according to differences in topography and industrial development found across the boreal or
mountainous landscape.

38

Bearhole/Redwillow - Spring

Quintette - Spring

0.80 i
0.70 0.60 -

, 0.60

++

i Available

g 0.40
£

0.30
0.20

■ U sed

i Used

#
— 0.50

"

0.40

a

0.30

mA vailable

-

0.10
0.00

&

^

O'

$

&
Forest Cover

Forest Cover

&

Quintette - Calving

Bearhole/Redwillow - Calving

0.80 -

V,&

^*

0.70 0.60 *

■ U sed

I Used

0.50 -

i Available

0.40 V

0.30 -

—

0.40

«

0.30 -

■ A vailable

0.20 0.10
0.00 -

<tT

•f
.-A-'

o^
Forest Cover

-

J «y£>

<,V

Forest Cover

Figure 8. The percentage (%) o f used and available locations occurring within each class o f forest cover during the spring and calving
seasons for caribou in the Bearhole/Redwillow (BHRW) and Quintette herds. Model covariates for forest cover are described in Table
1. An asterisk (*) indicates a forest cover class with greater than 5% use by caribou.

Quintette - Summer/Fall

Bearhole/Redwillow - Summer/Fall
0.70
0.60 *

0.60
I U sed ■ A vailable

0.40

° -5 0

I U sed ■ A vailable

'

— 0.40
§ 0.30 -

» 0.30
a

—

lJu

0.20

*

^

“

0.20
0.10

0.00
At-

rO y
J y' A

As

J 1

At

jgr

r

Forest Cover

Forest Cover

J?

x,V

4?

Quintette - Winter

Bearhole/Redwillow - Winter

0.70

&

0.60
0.50

I U sed • A vailable

I U sed * Available

St"
— 0.40
a o.3o

aV

0.20
0.10
0.00

JLl

LI

-

*

cf
.V*-'

^

O'

Forest Cover

4?

vV?

Forest Cover

Figure 9. The percentage (%) o f used and available locations occurring with each class o f forest cover during the summer/fall and
winter seasons for caribou in the Bearhole/Redwillow (BHRW) and Quintette herds. Model covariates for forest cover are described
in Table 1. An asterisk (*) indicates a forest cover class with greater than 5% use by caribou.

Spring. Some members o f the BHRW caribou herd occupied the boreal forest prior
to calving and selected habitats dominated by black spruce o f a moderate age during spring
( 4 1 - 1 2 0 years; Figure 10). These habitats contained 30% of all locations for the BHRW
herd during the spring season (Figure 8). Caribou in the BHRW herd avoided alpine habitats
as well as forests classified as old. In contrast, Quintette caribou remained at higher
elevations during spring, selected alpine and subalpine habitats and avoided forests
dominated by black spruce, tamarack, broadleaf and other mixed-conifer species. Alpine
habitats contained 74.2% of all locations for the Quintette caribou during the spring season
(Figure 8). The mean solar incidence was similar for caribou in both herds during this period
(BHRW: 1030.9 ± 331.4 W/m2; Quintette: 1029.7 ± 120.3).
The BHRW and Quintette herds demonstrated a nonlinear avoidance response to road
and pipeline features during spring (Figure 10, Table 4). BHRW caribou selected against
roads to an unknown distance, whereas Quintette caribou showed an avoidance response up
to 3.5 km (Table 4; e.g., Figures 11, 12). Pipelines were avoided up to 2.5 and 20 km by
BHRW and Quintette caribou, respectively. Caribou in the BHRW herd selected for areas
that were adjacent to cutblocks. Quintette caribou, in contrast, avoided individual cutblocks,
but demonstrated a higher relative probability of occurrence within areas with increased
densities o f forestry features (cutblocks and roads). During spring, Quintette caribou selected
for areas that were closer than random to coal mines.
Calving. Similar to spring, BHRW remained in the low-elevation boreal forest to
calve and continued their selection for habitats dominated by black spruce. Caribou in the
Quintette herd remained at high elevations to calve in alpine-dominated landscapes (Figure
13).

41

HBS

Up_Nveg

u
*tn 1
a*
c01 o
*c
£
u -1

Blk S pruce

G rowing

Spruce
Pine

^ M atu re

il

Tam

{ ■

Pipln
Road
SI - S e i s

Old

$
Pipln*
Seis*

Young

OG
^

♦
Ctblk

LF FOR

..
FOR*

OG*

TreeO

to.

-2

No Age D ata

Fir

Alpine

-4

B

No Age Data

Alpine

Fir

Road

*

*

Ctblk
Sprucf

20

SI
•TreeO

o4>
U -1

t

♦ ♦
W a te r

HBS

-3

Seis

T

?

i*

I
Old
Old
i\ atu re
M

SI*

♦

f

^ad*

Pipln

OG

MOG

FOR

♦ ♦♦
S * eiT P 7 p ln -C *

♦

*

.*

’

M |n e

Y oung]
Up_Nveg

G rowing

TreeBL

Figure 10. Coefficients for the parameters in the most parsimonious resource-selection models for
Bearhole/Redwillow (A; n = 3,401) and Quintette (B; n = 9,791) caribou herds during the spring
season. An asterisk (*) indicates a Gaussian term and variable descriptions are given in Table 1.
42

Table 4. Results of seasonal resource selection function models and the affiliated nonlinear
avoidance distances (Dist; km) and densities (Dens; ha/km ) calculated using Gaussian
covariates for caribou across the South Peace region o f northeastern British Columbia
(Appendix E).

Sum/Fall

linear

Calving
4.5

linear

Winter
11

Seismic Dist

0.6

linear

2.25

3

Pipeline Dist

2.5

linear

2

linear

Cutblock Dist

linear

6

linear

20

Oil/Gas Dist

21

10.5

7

4.5

Forestry Dens

24

linear

56

linear

Q uintette
Road Dist

Spring
3.5

Calving
3.5

Sum/Fall
4.5

Winter
3.5

Seismic Dist

6

3.5

2.5

1.5

Pipeline Dist

20

5.5

3.5

2

Cutblock Dist

3

20

3.5

4.5

Oil/Gas Dist

0.6

15

1

0.9

Mine Dist

linear

5

4.5

linear

Forestry Dens

linear

44

linear

28

BHRW
Road Dist

Spring

*Values may be unique to the South Peace study area, study animals, and/or my chosen method o f analyses
(e.g., logistic regression, size of availability radius, etc.; Ficetola and Denoel 2009).

43

0.98 0.97 0.96
0.95
0.94 0.93 0.92
0.91 0.90

100
Density of forestry cutblocks (ha/km2)

Figure 11. Likelihood of occurrence o f monitored caribou in the Quintette herd during the
calving season relative to the density o f forestry features (cutblocks and roads) found across
the South Peace region o f northeastern British Columbia (2003 - 2009). Habitat covariates
were held at their mean values, while caribou occurrence was allowed to vary with density of
disturbance features.

1
2
3
4
4.5
5
6
7
Distance from forestry cutblock feature (km)

Figure 12. Likelihood o f occurrence of monitored caribou in the Quintette herd during the
winter season relative to forestry cutblocks found across the South Peace region of
northeastern British Columbia (2003 - 2009). Habitat covariates were held at their mean
values, while caribou occurrence was allowed to vary with distance from disturbance
features.

44

Tam

No Age D ata

Fir
Blk S pruce
OG

H
tt

M atu re

Pine

u
*in 0
n
ch

Road
SI

* Road*

♦ ♦■■■

OG*

*

G rowing
Young

Alpine

»

,LF FOR

Seis

S pruce

l-i

♦

#
Old

*

Ctblk*

Ctblk
Pipln

-2
Up_Nveg

-3 TreeO

-4

5
4
3

No Age D ata

B

Alpine

Fir

Pine
S pruce

O2

*m
<n 1

TreeBL

-H X

&
•So

0 |d
T SI

Up_Nveg

x
"

se is
i

T * ♦— ♦—

c tb lk
P ip ln - £

M ine
F0R
O G **
MOG ♦

^ - ♦—

SI* Road* Seis* ^ b i k *
OG

Young

I-i

— i
M ine* LF FOR*

^ ^ M a tu re

-2
HBS

TreeO

Grow ing

-3
-4

W a te r

Figure 13. Coefficients for the parameters in the most parsimonious resource-selection models for
Bearhole/Redwillow (A; n = 2,200) and Quintette (B; n = 5,868) caribou herds during the calving
season. An asterisk (*) indicates a Gaussian term and variable descriptions are given in Table 1.
45

The use of cover types and patch age for both herds was more variable during the calving
season. Some members of the BHRW herd calved at higher elevations in subalpine fir
habitat, while others selected tamarack and black spruce-leading stands in the low-elevation
boreal forests. Subalpine fir, tamarack and black spruce habitats contained 51% o f all
calving locations for BHRW caribou (Figure 13). Caribou in the BHRW herd avoided
forests of early-successional ages. Quintette caribou selected older patches o f pine and
spruce in addition to alpine habitats during calving (containing 86.7% o f GPS collar
locations; Figure 8), but avoided areas dominated by water, herbs, bryoids, shrubs, and
forests classified up to 120 years of age. Solar incidence during calving ranged from a mean
of 1036.7 W/m2 (± 103.1) to 1068.0 W/m2 (± 133.1) for the BHRW and Quintette herds,
respectively.
The probability of finding caribou near roads remained low for herds residing in
boreal and alpine habitats during calving (Figure 13). Oil and gas facilities were avoided by
BHRW caribou up to a distance of 10.5 km and caribou in the Quintette herd demonstrated
nonlinear avoidance responses to seismic lines. Coefficients suggested that cutblocks and
coal mines had a negative influence on the distribution o f caribou during calving. Location
data also suggested that caribou from each herd had a larger relative probability of
occurrence in the vicinity of some disturbance features. During calving, Quintette caribou
were found in habitats closer than random to oil and gas features. The BHRW demonstrated
selection of habitats located near seismic lines, pipelines and forestry cutblocks. Caribou
from the Quintette herd also demonstrated a tolerance of forestry disturbances up to a density
of 44 ha/km2 (Figure 11, Table 4), after which an avoidance response was observed.

46

Summer/Fall. During the summer/fall season, individuals within the BHRW herd
remained divided between high- and low-elevation habitats while raising their calves (Figure
9). BHRW caribou inhabiting mountainous terrain selected herb, bryoid, and shrub
communities, although use and availability of these habitat types were extremely low across
their seasonal range (Figure 9, Figure 14). Caribou in the BHRW herd that remained in the
low-elevation boreal forest continued to select black spruce and tamarack-leading stands
during the summer and fall months. Alpine habitats and forests dominated by spruce (non-P.
mariana) were avoided by BHRW caribou (Figure 14). Similar to the calving season,
Quintette caribou demonstrated strong selection for alpine, pine and spruce-leading habitats,
where 79.9% of all summer/fall locations occurred (Figure 9). Forests dominated by
broadleaf and other mixed conifers were avoided by caribou in the Quintette herd. Caribou
in the BHRW herd were located in habitats with increased levels of solar insolation (x =
1038.3 ± 109.7 W/m2), whereas locations for Quintette caribou were exposed to relatively
less solar insolation during the summer/fall season (x = 1022.8 ± 143.7 W/m ).
Avoidance of anthropogenic disturbances was more prominent during the summer
and fall months (Figure 14). BHRW demonstrated an avoidance to roads up to an unknown
distance, whereas Quintette caribou avoided roads up to a distance of 4.5 km (Table 4). In
addition, coal mines were avoided throughout the summer and fall and both BHRW and
Quintette caribou avoided features related to oil and gas extraction. Quintette and BHRW
caribou avoided seismic lines to a distance of 2.5 and 2.25 km, respectively. Avoidance
responses towards pipelines were also comparable between herds during these post-calving
months. As the density of linear features increased, the relative probability of caribou
occurrence decreased.

47

HBS

Tam

o
*LA 1
o>

t

Blk S p ru c e |

*

1 0

Up_Nveg

Young

Old
I GrowingT

i’

Pine

+■

Seis
£

*“ <§

M atu re

i

Pipln

SI

Road

OG
C tblk:

♦ ----- *
Seis*

FOR*

----- ^

Pipln*

..

06

Fir
FOR

Alpine

-2

T reeO

W ater.
1

S pruce
No Age Data

-3

No Age Data

B
A lpine
Fir

S pruce
Pine

Road

u
*
m
ot

♦
SI

g

M ine
Seis p 'P ln

* ♦ ---*
Tfpeo------------------! ' ♦ ---- * -* .*♦ ----

0

0
£
u8 -1A
ca.

^

W aterY oung

I« .

°

-2
HBS

hi

^

R o ad *

OG

MOG FOR

*

Seis* Pipln*Ctblk*OG * M ine*

LF

Old

I M ea tu re
G rowing

Up_Nveg

-3

-4

TreeBL

Figure 14. Coefficients for the parameters in the most parsimonious resource-selection models for
Bearhole/Redwillow (A; n = 8,669) and Quintette (B; n = 22,458) caribou herds during the
summer/fall season. An asterisk (*) indicates a Gaussian term and variable descriptions are given in
Table 1.
48

Forestry cutblocks were avoided by caribou in the Quintette herd to a distance of 3.5 km.
For BHRW, negative coefficients suggested an avoidance o f habitats where the density of
forestry features exceeded 56 ha/km2. Quintette caribou also avoided habitats with increased
densities of mines and oil and gas features, but the top-ranked model produced coefficients
that suggested selection of habitats near forestry cutblocks.
Winter. BHRW and Quintette caribou demonstrated variation in resource selection
during winter (Figure 15). Animals in both herds demonstrated selection for alpine,
subalpine fir and pine-dominated landscapes o f older age classes, although habitats used by
caribou in the two herds differed considerably (Figures 8, 9). During the winter months,
66.3% o f Quintette locations occurred in the alpine while only 0.2% o f the BHRW locations
occurred in the alpine. Likewise, 58.7% of locations for caribou in the BHRW herd and
8.9% of locations for the Quintette herd were located in pine-leading forests (Figure 9).
Members of the BHRW herd overwintering in the boreal forest selected habitats dominated
by pine, black spruce, tamarack and other species of spruce. Both herds avoided mixed
conifer, broadleaf, upland areas without vegetation, and water-dominated habitats. During
winter, caribou in the BHRW herd were located in habitats with less solar incidence (x =
1029.7 ± 62.1 W/m2) as compared to caribou in the Quintette herd (x = 1082.4 ± 113.8
W/m2).
During winter, industrial activities influenced the habitat selection o f the Quintette
and BHRW herds (Figure 15). Quintette caribou avoided areas of their seasonal range where
cutblock (and associated road) densities exceeded 28 ha/km2 in addition to habitats located
within close proximity to roads and cutblocks (Figure 12, Table 4).

49

No Age Data

B|k S pruce p .n e

Tam

Al-inf I T T
o 2
*l/l 1
CD
1

11 II tTreeBL
Fir

I
"Spruce

tV -l
I

OG

♦

Road

i

SI

G row ing

♦

I***

-2

old

Young

+
Seis
Ctblk*
LFFOR
♦
♦ ..
"♦---Si* Road* S e is* 4 ♦
Ctblk
0G *
Pipln

M atu re

-3
-4
W atei

-5

HBS

Up_Nveg

-6

T reeO

-7

3

B

Alpine

2.5

No Age D ata

2
1.5

C 0.5
01

■0

Road
Fir

p *ne

*

ft

Ctblk
Seis Pipln

Spruce

+?

£

.TreeBL

5 °
CO.

T reeO w a t e r 4

-0.5

Road

Y° un|

-1
-1.5

B lkS pruo!
HBS

i_

M atu re

Old

*

*
OG
*

FOR MOG

^ ----- ♦ * ♦ ♦ ♦

*Seis* Pipln* C tb ll^

OG* FGR

LF

M ine

G rowing
Up_Nveg

Figure 15. Coefficients for the parameters in the most parsimonious resource-selection models for
Bearhole/Redwillow (A; n = 11,625) and Quintette (B; n = 28,368) caribou herds during the winter
season. An asterisk (*) indicates a Gaussian term and variable descriptions are given in Table 1.

50

Landscape features related to oil and gas exploration and extraction were avoided by caribou
wintering in high-elevation mountainous habitats. Caribou of the BHRW herd also
demonstrated avoidance o f roads, cutblocks, oil and natural gas features. Quintette caribou
however, were observed closer than their random locations to coal mines during winter.

Count Models fo r Wolves
I used 24,075 GPS collar locations for Upper Sukunka (US; n = 6,783), Upper
Murray (UM; n = 6,478), Onion Creek (OC; n = 4,624) and Chain Lakes (CL; n = 6,190) to
generate statistical count models. Seventy-three kill sites (US, n = 20; UM, n = 15; OC, n =
17; CL, n = 21; Figure 16) served as the foundation for determining the average wolf area of
use (AOU) and size o f the habitat selection unit (HSU) for each pack (Appendix C). Habitat
Selection Units ranged in size from 6.6 ha for the Upper Sukunka pack to 155.4 ha for the
Upper Murray pack o f wolves (Appendix C). The Vuong test suggested that count data were
fitted better to zero-inflated regression models (ZINB) for all packs across seasons (Vuong
4.31 < z < 9.08, p < 0.001). I was unable to maximize the log likelihood values for five
seasonal models using the preferred ZINB. For each o f these five seasons, I chose to use the
more simplistic NBRM. Tolerance scores for continuous variables were low for some packs
and variables. With a primary study focus on linear features, I excluded cutblocks and non­
linear features linked to oil and natural gas from candidate models for the Upper Sukunka
and Chain Lakes packs. For the Onion Creek pack, distance to mine was highly correlated
with distance to oil and gas features; I retained the distance to mine variable as the Onion
Creek territory was less influenced by features associated with oil and gas development.

51

70
60
50
I 40

|
<2 30

*

20

10

0
M oose

D eer

Elk

M t. G o at C aribou U nknow n

O th e r

Prey Species

Figure 16. Prey selection (%) by GPS collared wolves as identified through the investigation
o f location clusters (2008 - 2010; n = 73 kills) across the South Peace region o f northeastern
British Columbia.

Based on AICc, a combination o f landscape attributes, selection value o f caribou
habitat and human disturbances was best able to model the seasonal occurrence and relative
frequency of habitat use by wolves across the South Peace region (Table 5). For all three
seasons, the most parsimonious models for both the Onion Creek and Chain Lakes packs
were also the most complex in each candidate set and contained variables for landscape,
caribou habitat and anthropogenic disturbances. During late winter, the second-ranked
model for the Upper Sukunka pack was comparable to the top-ranked model (A AICc = 1.0).
I observed similar model selection uncertainty for the two top-ranked models for the Upper
Murray pack during winter (A AICc = 1.9). The predictive ability o f seasonal count models
was generally good for HSUs across the study area (Figure 17).

52

Table 5. Number o f parameters (k), Akaike’s Information Criterion values (AICc), AICc weights (AIC*), and A AICc values for
competing seasonal count models for wolves. Models were developed (using ZINB or NBRM) for each o f four w olf packs monitored
from 2008 - 2010 across the South Peace region o f northeastern British Columbia. Sample size used to define habitat selection units
(HSUs) is presented in parentheses for each pack. Model covariates are given in Table 2 and the full set o f candidate models can be
found in Appendix E.
Non-WinterNBRM

Upper Sukunka (n = 33,599)

Late WinterNBRM

Early Winter2™6

Model Covariates

k

AICc

AAIC

AICH

k

AICc

AAIC

A IC K

k

AICc

AAIC

AIC„

MOG Dist“

22

9636.6

44.4

<0.001

23

5087.1

37.1

<0.001

22

6116.9

0

0.45

MOG Dist + MOG Dens“

23

9629.1

36.9

<0.001

25

5084.8

34.9

<0.001

23

6118.3

1.4

0.22

CE Dist“

19

9598.8

6.6

0.04

26

5073.1

23.2

<0.001

23

6117.8

0.9

0.27

CE Dist + CE Dens“

21

9592.2

0

0.96

32

5049.9

0

1

25

6120.6

3.7

0.07

Early WinterNBRM

Non-WinterNBRM

Upper Murray (n = 35,959)

Model Covariates
MOG Dist
MOG Dist + MOG Dens
CE Dist“

k
16
18
20

AICc

AAIC

AIC„

11027.1
11029.8
11026.9

39.2
41.9
39

CE Dist + CE Dens“

27

10987.9

0

Model Covariates
CE Dist + CE Dens

k
20

Chain Lakes (n = 3,389)

Model Covariates
CE Dist + CE Dens

k
18

AICc

AAIC

6722.1
0
Non-Winter2™6

AICc

AAIC

AIC„

<0.001
0.72
<0.001

k
14
15
16

5816
5778.8
5789.3

132.3
95.1
105.6

<0.001
<0.001
<0.001

0.28

18

5683.6

0

1

AICc

AAIC

AIC„

<0.001
<0.001
<0.001

k
15
18
15

4417.2
4357.2
4417.2

59.9
0
59.9

1

17

4359.1

1.9

Early Winter2™6

Non-Winter2™6

Onion Creek (n = 10,493)

Late Winter2™6

a ic m

1

AICc

AAIC

AIC„

4014.2

0

1

k
25
k
22

AICc

AAIC

3079.1
0
Early WinterNBRM

Late Winter2™6
AIC„

1

AICc

AAIC

a ic m

7174.7

0

1

“Gaussian (squared) term was most parsimonious in at least one seasonal candidate model

k
23
k
18

AICc

AAIC

a ic m

1
3675.6
0
Late Winter2™6
AICc

AAIC

a ic m

3252.2

0

1

0.001

0.005
0.004

0.001

© N on-W inter

0.003

□ Early W in ter
“ ■

0.002

■o 0.000

A Late W in ter

30

£ 0.001
“--

•a

40

50

0.001
© N on-W inter
□ Early W in te r

-

0.001
A Late W in te r

-0.003

Wolf Count

-

0.002

0.008 -

0.03
0.03

0,006

0.004 - |- j

© N on-W inter

0.02

© N on-W in ter

□ Early W inter

0.02

□ Early W in te r

A L ate W in ter

p 0.01

A Late W in ter

C 0.002 |#

■p 0.01
•

Wolf Count

<P

0.00

0 .0 0 0

40

g -0.01
w

20

o -0.01

-0.004 J

Wolf Count

-

0.02

Figure 17. Differences in the observed (withheld data) and predicted probability o f counts of wolf locations within habitat selection
units (HSUs) for the Upper Sukunka (A), Upper Murray (B), Onion Creek (C) and Chain Lakes (D) packs residing in the South Peace
region o f northeastern British Columbia. Predicted data were generated from the most parsimonious zero-inflated (ZINB) or negative
binomial regression model (NBRM; Table 5). A value o f zero indicated perfect prediction, whereas positive values indicated under­
prediction and negative values indicated over-prediction.

Inspection of residuals indicated the probability of models to predict occurrences o f wolves
across the landscape was relatively poor for zero and low frequencies, but improved as wolf
locations increased within habitat selection units.

Occurrence and frequency o f w olf locations across the South Peace
Seasonal variation in the selection and use of forest cover, caribou habitat, and
disturbance features was observed for each study pack of wolves. Wolves residing in
mountainous regions occurred in pine-dominated forests throughout the year. Habitats
dominated by broadleaf or mixed-conifer trees, water, and non-linear features were also
important indicators of wolf occurrence. Linear features were avoided by boreal and
mountainous wolves during each of the three seasons. In general, wolves were infrequently
located in habitats valued as high quality for BHRW or Quintette caribou.
Non-winter. During the non-winter season, Onion Creek wolves occurred in pinedominated habitats and wolves in both the Onion Creek and Chain Lakes packs avoided
mature and late-successional forests (81 - 120, and > 120 years) dominated by broadleaf
species (Table 6; Appendix D). ZINB models showed that processes influencing the
presence or absence of wolves on a landscape were different from those affecting the
frequency of use. Three packs with territories in the mountainous regions (Upper Sukunka,
Upper Murray, and Onion Creek) showed higher frequencies o f locations in HSUs containing
upland or spruce habitats. Wolves in the Chain Lakes pack were commonly located in the
lower elevation boreal areas with aspen, cottonwood and birch o f unknown ages. Habitats
dominated by water bodies were frequently selected by all packs. Wolves rarely used latesuccessional forests containing pine and other mixed-conifer species.

55

Table 6. Seasonal selection (S) and avoidance (A) o f habitat features by wolves across the South Peace region o f northeastern British
Columbia. Presence or absence (binary) and the frequency o f habitat use (count) were determined using p coefficients from count
models. Models were developed for the Upper Sukunka (US; n = 33,599), Upper Murray (UM; n = 35,959), Onion Creek (OC; n =
10,493) and Chain Lakes (CL; n = 3,389) packs. Model covariates are given in Table 1 and Table 2.
Non-winter fS)
Variable

Binary

Count

Earlv winter (Si

Late winter fS)

Non-winter (A)

Binary

Binary

Binary

Count

Alpine

Count

Count

Earlv winter fAl

Late winter (Ai

Binary

Binary

Count

US
CL

Black spruce
No VRI

UM

UM

UM

Other

UM,US

CL

OC,UM

Pine

Count

OC

OC

Spruce

OC

US
UM

UM

CL

OC

UM

US

Tamarack
CL

Tree broadleaf
Tree other

CL

US

US

OC

US

UM,US

No age

us
OC

UM

UM,US

YG (0-80 yrs)

OC

YGM (0-120 yrs)

US

CL

OC

us

UM

Young
Growing

UM

Mature
Old
Water06
RSF BHRW

CL
CL
UM,CL

CL,OC,UM,US
UM

OC

R SFQ uintette

UM
CL

UM

CL,UM
US

OC,UM

US,OC

UM

CL
OC

UM

US

CL,OC
UM

" Covariate measuring distance (km) to a feature; selection is therefore represented by a -P coefficient and avoidance is represented by a +P coefficient
6 Either a Gaussian (squared) or linear term was used in the top model

UM
OC,US

Upper Murray, Onion Creek and Chain Lakes wolves avoided higher quality habitats for
Quintette caribou during the non-winter months. Only wolves from the Upper Murray pack
frequented habitats selected by the BHRW caribou herd (Table 6).
Members o f the boreal Chain Lakes pack were present in HSUs with few roads and
few locations occurred in areas with high densities o f linear features. Members were
observed in HSUs near forestry cutblocks, but the total number of locations was not strongly
related to such features. Wolves in the Onion Creek and Upper Sukunka packs avoided
seismic lines, pipelines, and coal mines (Onion Creek only; Table 7, Appendix D). Although
habitats near roads were selected by wolves in the Upper Murray pack, non-linear features
were more informative in describing wolf distribution between mid-April and mid-October.
Only Upper Sukunka and Upper Murray wolves frequented areas near coal mines and oil and
gas facilities; however, as the density of these features increased, the frequency of wolf
locations decreased.
Early winter. Similar to non-winter, wolves in the Onion Creek pack occurred in
HSUs where pine was the predominant species. Upper Sukunka wolves were present in
habitats o f primarily mixed conifer. Habitats dominated by broadleaf trees were avoided by
wolves in the Upper Sukunka pack, but were frequently used by members o f the Chain Lakes
pack in the boreal forest and the Upper Murray pack residing in the mountains (Appendix D).
Higher frequencies of wolf locations occurred in early-successional forests classified as pine
(Onion Creek), upland and habitats dominated by herbs, bryoids, and shrubs (Chain Lakes
and Upper Murray; Table 6).
The frequency of wolf locations was not related to habitats strongly selected by
caribou.

57

Table 7. Seasonal selection (S) and avoidance (A) o f disturbance features by wolves across the South Peace region o f northeastern
British Columbia. Presence or absence (binary) and the frequency of habitat use (count) were determined using P coefficients from
count models. Models were developed for the Upper Sukunka (US; n = 33,599), Upper Murray (UM; n = 35,959), Onion Creek (OC;
n = 10,493) and Chain Lakes (CL; n = 3,389) packs. Model covariates are given in Table 1 and Table 2.
Non-winter (S)
Variable

Binary

Road0,4

Count

Earlv winter (SI
Binary

Count

Late winter (SI
Binary

Count

UM

SeisPipln"4
Ctblk"4

US

US

OG"4

US

US

US

Mine"'4

UM

US,OC

UM

CL

Count

CL
OC

Earlv winter (A)

Late winter (A)

Binary

Count

Binary

CL

US

US

UM

OC,US

US

UM

OC,UM

CL

Count
OC
OC

US
OC

M O G D en s4

OC

OC,US

OC,US

CL,UM

F O R D en s4
L F D e n s4

Non-winter (A)
Binary

UM

CL,OC,UM
CL

UM

US
UM

UM,US

UM

UM,CL

OC

" Covariate measuring distance (km) to a feature; selection is therefore represented by a -0 coefficient and avoidance is represented by a +p coefficient
4 Either a Gaussian (squared) or linear term was used in the top model

CL, OC

Forests o f late succession (>121 years o f age) and HSUs classified as black spruce or
pine-leading contained few locations o f wolves across the territory o f the Chain Lakes pack.
In addition, quality habitat for caribou in the BHRW herd was avoided by most wolves in the
boreal forest. Only members of the Onion Creek pack occurred, but were not frequently
located in HSUs containing high-value habitat for caribou in the BHRW herd (Table 6).
Areas containing cutblock, oil, gas, and coal mine features supported high frequencies of
wolf locations during early winter for both the Upper Sukunka and Onion Creek packs.
Conversely, boreal wolves were uncommon in HSUs close to cutblock features or with a
high density o f linear features or cutblocks (Table 7).
Late Winter. Between February and mid-April, the presence o f wolves was best
described by a variety o f forest cover types. Upper Murray wolves used mountainous
habitats classified as upland and alpine, as well as communities dominated by herbs, bryoids,
shrubs, water (ice) or broadleaf trees. Wolves in the Onion Creek pack occurred in HSUs
where mixed conifers prevailed. As in other seasons, wolves in the Upper Sukunka and
Upper Murray packs frequented forests dominated by aspen, cottonwood, birch, and pine
between 0 and 120 years of age. Both Upper Murray and Chain Lakes wolves did occur,
although not frequently, in late-successional forests during late winter (> 120 years). Wolves
in the Upper Murray pack were absent from mature (8 1 -1 2 0 years) forests dominated by
white, Engelmann, or hybrid spruce. In addition, HSUs with communities of herbs, bryoids,
shrubs or upland areas, all contained low frequencies of wolf locations. Throughout late
winter, both Upper Murray and Onion Creek wolves also demonstrated an avoidance o f
habitats containing water features (Table 6).

59

Although Upper Sukunka wolves rarely had the opportunity to overlap populations of
woodland caribou, they demonstrated increased frequencies of use o f alpine habitats during
late winter. In contrast, and consistent with the early winter, Onion Creek and Chain Lakes
wolves did not frequently occur in habitats selected by caribou in the boreal forest. Upper
Murray wolves also demonstrated an avoidance o f habitats used by the Quintette caribou
herd during late winter (Table 6).
Anthropogenic disturbances continued to influence the distribution o f wolves across
the study area throughout the late-winter months (Table 7). Wolves in the Upper Sukunka
and Upper Murray packs frequented HSUs near oil and gas features as well as habitats with
greater densities o f cutblocks. Avoidance o f disturbance features was more apparent in late
winter for wolves in both the boreal forest and in the mountains. Packs occurring in
mountainous portions of the study area were absent or rarely occurred in areas near linear
features (Upper Murray and Onion Creek), coal mines (Upper Sukunka and Onion Creek),
and areas with high densities of cutblocks and roads (Upper Murray). In addition, locations
o f wolves from both the Chain Lakes and Onion Creek packs were uncommon or absent
from habitats with relatively high densities o f linear features.

Discussion
This study supports the general conclusions of others that the cumulative effects of
industrial development have strong influences on the patterns o f habitat selection and
distribution for both wolves and woodland caribou in mountainous and boreal ecosystems
(Dyer et al. 2001, James et al. 2004, Vors et al. 2007, Nitschke 2008, Houle et al. 2010).
However, my results suggest that regionally-specific information and knowledge of the

60

processes of predator-prey interactions are essential for understanding the ecological impacts
of those cumulate effects. This is especially the case for woodland caribou, a threatened
species that is influenced directly and indirectly by disturbance, habitat modification and
altered predator-prey dynamics.
I used an innovative combination of field and statistical methods to understand the
seasonal distribution of wolves relative to caribou habitat and industrial development. The
application o f count models to HSUs allowed me to develop statistical relationships that
represented the frequency of habitat use, not simply habitat selection. The number o f wolf
locations in an HSU may be associated with predatory behaviour, such as hunting and prey
handling, or the size of pack territories. In addition, the inclusion of the RSF variable that
quantified the selection value of habitats for monitored caribou herds, provided a more
holistic description o f habitats related to the distribution o f caribou. Resource selection
functions represented not only vegetation that would serve as forage for caribou, but also
human disturbances that influence the distribution o f each herd.

Habitat Selection by Caribou and Wolves
My study o f the BHRW and Quintette herds provided a unique opportunity to observe
differences in behaviours between populations o f caribou that winter in low-elevation boreal
and high-elevation alpine habitats, respectively. Few studies have looked at behaviourally
distinct populations of caribou as they respond to direct threats from industrial encroachment
and predation by wolves. Caribou o f the BHRW herd demonstrated selection for mature and
late-successional forests dominated by black spruce (all four seasons), tamarack, and to a
lesser extent, subalpine fir, alpine and communities of herbs, bryoids, and shrubs. Caribou

61

that overwintered in the boreal forest selected black spruce, tamarack, and older pine-leading
stands. Although a small proportion of the BHRW herd was observed selecting highelevation habitats during winter, GPS collar locations were rare in alpine habitats. Quintette
caribou selected alpine, subalpine fir, spruce and pine-leading habitats o f late-succession
during winter. Similar results were documented for caribou in the Quintette herd by Sopuck
(1985) and Jones et al. (2007).
Across all seasons, caribou in both herds were observed avoiding early-successional
habitats dominated by aspen, cottonwood, birch, and mixed conifers. These avoidance
behaviours may be a result o f the increased abundance of other ungulates and associated
predators typically found in these forest types. As documented for other populations of both
northern (Cichowski 1993, Johnson et al. 2002b) and boreal ecotypes (Saher and
Schmiegelow 2004, Culling et al. 2006, Neufeld 2006, Courbin et al. 2009), my results
suggest that high-risk habitats are avoided by caribou from the Quintette and BHRW herds.
Like caribou, wolves residing in mountainous regions demonstrated selection for
pine-dominated forests throughout the year. Unlike caribou, wolves frequented habitats of
early serai ages. During all three seasons, habitats dominated by broadleaf or mixed-conifer
trees and water were important indicators of wolf, but not caribou occurrence. In addition,
wolves favoured habitats dominated by herbs, bryoids, and shrubs during the winter, whereas
caribou in both the BHRW and Quintette herds avoided these habitats. Upper Murray and
Onion Creek wolves demonstrated some selection of habitats used by BHRW caribou.
Caribou in both the BHRW and Quintette herds were at a relatively low risk of
predation during late winter, even though the use of subalpine (e.g., subalpine fir, upland and
herbs, bryoids, and shrub-dominated habitats often in ‘other’ category; Appendix D) and

62

alpine habitats by wolves generally increased during this season. In contrast to Latham
(2009), who reported increased levels of overlap between caribou and wolves during winter
in the low-elevation forests of western Alberta, black spruce and tamarack forests were rarely
selected by either of the two packs of wolves I monitored. Furthermore, the Chain Lakes and
Onion Creek packs avoided habitats classified as high quality for BHRW caribou during
winter. These findings are supported by observations of prey remains at kill sites where
caribou accounted for 1.3% of identified wolf kills in the South Peace region (Figure 16;
Appendix C). I lack information delineating the habitats of other prey species, but my results
are comparable with past studies suggesting wolf populations are typically supported by prey
other than caribou (i.e., moose, deer, elk, beaver, other small mammals and birds; Figure 16;
Bergerud et al. 1984, James et al. 2004, Gustine et al. 2006b, DeCesare et al. 2010, Latham et
al. 201 la, Milakovic and Parker 2011, Steenweg 2011). Although my data suggest wolves
are not using habitat patches selected by caribou, the level of spatial separation remains
greater for the Quintette herd than the BHRW herd because wolves have increased
opportunities, with relatively low costs o f movement, to use caribou habitat across the boreal
forest.
Solar insolation correlated with the distribution o f caribou. However, interpreting the
mechanism by which this variable influenced caribou was challenging. Levels o f solar
insolation were generally less for BHRW under the cover of the boreal forest than for
Quintette caribou residing in exposed alpine and subalpine habitats. In future studies, a
topographic variable representing windblown ridgelines in alpine habitats, in addition to solar
incidence, could further our understanding of caribou distribution.

63

Habitat selection studies can suggest that certain resources or habitat types are
important simply as a product of the availability o f those types. For example, I observed
selection for alpine habitats by a relatively few caribou of the BHRW herd - this relationship
was the product o f the low availability o f the habitat not a high level o f use. To clarify such
relationships, I determined the seasonal availability and use of each forest cover type (e.g.,
Figures 8, 9; Appendix D). Special consideration should be given to habitat types that are
commonly used and selected.

Behavioural Responses o f Wolves and Caribou to Industrial Disturbances
Wolves in all four packs demonstrated avoidance o f linear features during each of the
three seasons. Also, I modeled a low frequency o f wolf occurrence in habitats with high
densities o f roads, seismic lines and/or pipelines. These findings parallel similar studies of
wolves in industrial landscapes (Thurber et al. 1994, Whittington et al. 2005, Houle et al.
2010). Such avoidance responses are likely due to direct and indirect risks associated with
exposure to areas used by humans (i.e., mortalities from increased access for hunting,
trapping, and vehicular collisions; Fuller 1989, Mladenoff et al. 1995, Callaghan 2002,
Hebblewhite and Merrill 2008, Person et al. 2008). Because only one pack (Upper Murray)
was observed frequently using habitats near roads during the non-winter season, my results
suggest that cumulative road densities across the majority of the study area may have
surpassed levels acceptable for travel by wolves (-0.25 km/km2- 0.6 km/km2; Mech et al.
1988, Fuller 1989, Merrill 2000, Person 2008).
Caribou demonstrated a strong avoidance o f linear features during all four seasons,
although roads were the only feature consistently avoided each season. Similar to Nellemann

64

et al. (2001), Dyer et al. (2002) and Latham (2009), roads were avoided (but up to a greater
distance of 3.5 km) by both BHRW and Quintette caribou most significantly during the
winter months. Winter is the busiest season for activities related to the exploration and
development of petroleum reserves and forestry operations and could explain the observed
avoidance by caribou. Although I used conditional regressions to statistically remove the
responses of individual caribou locations to disturbance features that occurred at large
distances, nonlinear avoidance thresholds still occurred at large distances not currently
reported in the literature (Table 4). Future studies could consider exploring alternative
statistical methods that are not constrained by a Gaussian function, such as I used, and should
also consider analysing caribou behaviour in the presence of individual disturbances at
multiple scales.
Unlike avoidance of roads, caribou in the BHRW herd demonstrated selection for
specific habitats near seismic lines and pipelines during calving and winter. These linear
features vary in intensity of disturbance and age, and although I did not measure such factors,
they may explain seasonal tolerance by BHRW caribou. Close proximity to these features
may also suggest that caribou have not yet reached a threshold level of intolerance. Also,
caribou may demonstrate long-term fidelity to seasonal habitats that become adjacent to
early-successional habitats or industrial developments. Persistent use of such sites may
increase the risk of predation for caribou and their calves, thus, serving as ecological traps
(Schlaepfer et al. 2002, Faille et al. 2010). While past studies suggest predation risk for
caribou increases near linear features, results from the count models for wolves suggest that
risk is less severe. Wolves generally avoided roads, seismic lines and pipelines, as well as
habitats that supported greater densities of linear features. Contrary to other studies (James

65

and Stuart-Smith 2000, Latham et al. 201 lc), my results suggest that in this area, the current
density of linear features may not result in a direct increase in predation risk for caribou.
High densities o f linear features also influenced the distribution o f caribou across a
larger regional area. During calving, summer and fall, caribou avoided areas o f their range
with high densities of linear features. Like Polfus et al. (2011) and Curatolo and Murphy’s
(1986) study, my results suggests that high levels of human use near roads indirectly results
in functional habitat loss for caribou across the South Peace region.
Wolves in boreal and mountainous habitats occurred in areas closer to, and with
higher densities o f cutblocks during the non-winter season. Wolves may be advantageously
selecting these habitats for increased hunting opportunities of moose and deer. Wolves may
also frequent these habitats because they are suitable for denning or homesites. The use of
cutblocks by wolves was less consistent during early and late winter. The Upper Murray
pack frequently selected habitats closer to and with greater densities of cutblocks in early
winter. Unlike early winter, all four packs avoided cutblocks during the late-winter months the time of year when forestry, oil, and gas industries are most active and when deep snow
begins to restrict moose from foraging in cutblocks (D. Heard, personal communication).
Similarly, Houle et al. (2009) found that wolf occurrence decreased as cutblock density
increased in Quebec. Wolves in the South Peace may be responding to the cumulative
influence of roads and cutblocks at both a home range and regional scale (inter-pack; Houle
et al. 2010). Also, there is often little browse for moose in newly harvested areas (Nielsen et
al. 2005). The infrequent occurrence of wolves could indicate a relatively large proportion of
recent cutblocks, and their associated roads, in some territories as opposed to those with older
regenerating cutblocks containing more suitable habitat for moose (Courtois et al. 1998).

66

Caribou in both boreal and mountainous habitats responded differently to cutblocks
than to linear features. BHRW caribou selected habitats within close proximity to individual
cutblocks during spring and calving, but avoided areas with high densities o f cutblocks
during the summer, fall, and winter seasons. Quintette caribou also avoided habitats with
higher densities of cutblocks during winter. Similar to wolves, caribou from both herds were
found within areas of their ranges containing high densities o f cutblocks during calving.
These results, though counterintuitive, further support my hypotheses that female caribou
may select for particular habitat characteristics regardless of human disturbance or predation
risk (i.e., fidelity; Rettie and Messier 2001, Wittmer et al. 2006, Faille et al. 2010).
There are a number of plausible explanations for the observed distribution of caribou
near cutblocks. Behaviours associated with the learned use o f distinct calving sites may take
precedence over the risks associated with spending increased amounts o f time in early serai
forests. Similar to Hins et al. (2009), caribou across the region might also select remnant
strips o f old-growth forest often found adjacent to cutblocks. Thus, I may have observed a
pattern of selection associated with juxtaposition, not composition o f habitats. Alternatively,
or in combination, caribou may be demonstrating seasonal tolerances towards regenerating
cutblocks as there can be a time lag o f 20 years between the initial phases of forestry
extraction and avoidance of those areas (Nielsen et al. 2005, Vors et al. 2007). In total, my
results suggest the possibility of maladaptive sinks for populations of caribou across the
South Peace region. These negative fitness outcomes may be subject to a lag effect, being
realised only after moose and associated predators adjust their distribution to emerging
habitats (Nielsen et al. 2005).

67

Features associated with the development of oil and gas deposits also influenced
behaviours o f wolves and caribou across the study area. The Chain Lakes and Upper Murray
packs avoided areas o f their range with a high density of oil or gas features during the non­
winter seasons. In contrast, wolves in the Sukunka Valley demonstrated a greater frequency
of use of habitats within close proximity to oil or gas features. These patterns of selection
suggest that levels of human activity associated with oil and gas development vary across the
territories of collared wolves, or some wolves have developed strategies to accommodate
disturbance stimuli (Hebblewhite and Merrill 2008). Caribou were located further than
random from well pads and other oil and gas sites during calving, summer/fall, and winter,
but demonstrated the greatest avoidance during calving (BHRW) and summer/fall. My
results are similar to studies on Arctic caribou herds where the indirect losses o f higherquality habitat were most apparent during post-calving seasons (Cameron et al. 2005,
Johnson et al. 2005). During the non-winter months, co-occurrence o f wolves and caribou
near non-linear features associated with oil and gas development was rare.
The Upper Murray and Onion Creek packs o f wolves occupied mountainous
territories adjacent to, but were infrequently located near coal mines. Two packs o f wolves
(Upper Sukunka and Onion Creek) usually avoided mines during the non-winter and latewinter seasons. As wolves focus on the rearing of pups, the high levels o f human activity
and vehicular traffic associated with mine sites might deter them from frequenting those
areas (Lesmerises et al. 2012). During winter, wolves may naturally avoid industrial
features, such as mines, if they continue to hunt primary prey in the valley bottoms. Coal
mines occurred only within the range o f Quintette caribou (Sopuck 1985). These caribou
avoided mines up to a distance of 5 km during calving and throughout the summer and fall

68

months, but selected habitats near mines during spring and winter. Elongated ridges in alpine
wintering habitats are of high value to Quintette caribou. Again, caribou may trade-off the
learned use of high-quality habitat (i.e., fidelity) with a tolerance of human activities and
disturbance.

Cumulative Effects o f Resource Extraction and Development on Wolves and Caribou
The cumulative effects of anthropogenic activities are now recognized as one of the
most pressing problems facing the conservation and management of wildlife (Vistnes and
Nellemann 2001, Johnson et al. 2005, Vors et al. 2007, Johnson and St-Laurent 2011,
Krausman and Harris 2011). Habitat alterations from large-scale forestry, oil, natural gas,
and mineral exploration, have resulted in dramatic transformations of the South Peace region
and continue to threaten the ecological integrity of the landscape (Nitschke 2008).
Avoidance of habitats with high densities of linear (i.e., roads, seismic lines and/or pipelines)
or non-linear disturbance features (i.e., cutblocks, coal mines, oil and gas facilities) strongly
suggests that industrial activities have reduced the quality and quantity of contiguous habitat
for caribou across this region. Human-caused disturbance in combination with altered
vegetation communities result in compounding instabilities for populations o f caribou:
increased movement and vigilance, displacement from portions o f the range and altered
predator-prey dynamics (Bradshaw et al. 1997, Nellemann and Cameron 1998, Cameron et
al. 2005, Faille et al. 2010, Latham et al. 201 la). Furthermore, these relationships are
complex and may be confounded by ecological sinks and lag effects.
Habitat and movement analyses, in addition to field investigations of wolf kill sites
from my study area suggest that co-occurrence between caribou and wolves is rare. In

69

general, wolf packs rarely selected habitats that were ranked as high quality for either herd of
caribou. Similar to caribou, wolves avoided habitats with high densities of linear and non­
linear features. Wolves also avoided roads, seismic lines, and/or pipelines, but selected
habitats within close proximity to non-linear features (i.e., cutblock, oil or gas footprints)
during some seasons, where a presence o f ungulates, other than caribou, was likely.
However, if caribou continue to demonstrate seasonal fidelity to developments that support
early-successional habitats or predator movement, risks o f encountering their primary
predators increase. Furthermore, although caribou kills from wolves were infrequently
identified during field investigations across the South Peace region, slight increases in the
rate o f adult mortality from predation can have significant impacts on the stability of small
herds o f caribou (Wittmer et al. 2005, Gustine et al. 2006a, Latham et al. 201 lb).
A challenge for resource managers is to balance the demand for expanding coal
mines, oil and gas reserves, and wind-farms with caribou conservation. New projects are
being proposed and constructed across caribou winter range throughout the South Peace
region. The continued rate of development and resulting loss of contiguous habitat across
this area will likely push already small populations of caribou further into decline (Seip and
Jones 2011). Caribou inhabiting the low-elevation boreal habitats may be demonstrating a
maladaptive strategy in the context of multiple disturbance regimes on the landscape.
Specifically, encounters between caribou and wolves are most likely to occur in areas closer
to and with higher densities of cutblocks, as both species were observed selecting these
features during the non-winter season. As my results suggest, however, interactions among
predators, caribou and land-use development are not easily predicted or temporally static.
Further monitoring o f caribou and wolves is necessary in the context of a changing and

70

interacting landscape to understand when distribution strategies o f these species begin to be
affected and to minimize changes that permanently alter the ability of landscapes to support
populations o f caribou.

71

Chapter 3: Movement Ecology of Wolves in an Industrialized Landscape

72

Introduction
Throughout Canada, agriculture and industrial activities provide economic
development, but are also responsible for habitat change, fragmentation, altered community
dynamics, and ultimately, a reduction in biodiversity (Bradshaw et al. 1997, Dyer et al. 2001,
Schneider et al. 2003, Festa-Bianchet et al. 2011). Since the early 1990s, the Peace River
and Moberly regions of northeastern British Columbia have undergone rapid land-use change
as a result o f large-scale commercial forestry, energy, and mineral development (Nitschke
2008). Woodland caribou are now o f considerable conservation concern across that region.
Throughout much of boreal Canada, habitat alteration and disturbance resulting from human
developments are responsible for declining herds, a loss of connectivity of contiguous habitat
and increasing predation through apparent competition (Vors and Boyce 2009, FestaBianchet et al. 2011).
Activities related to large-scale resource exploration and extraction serve as a catalyst
for creating efficient travel corridors for wolves, a primary predator of caribou in the boreal
forest. Roads, seismic lines, pipelines, and other linear features (e.g., power lines) can
provide greater mobility for wolves as well as access to habitats that would otherwise be
isolated by topography or snow. Following human developments, early serai forests become
more abundant and support regenerating habitats that favour higher densities o f ungulate
species, such as moose, elk, and deer. This change in landscape composition increases the
distribution of wolves and the likelihood of interactions with caribou (Fuller and Keith 1981,
James et al. 2004, Johnson et al. 2004a, Wittmer et al. 2007, DeCesare et al. 2010).
Movement parameters describing animal paths can provide an index o f animal
behaviour relative to variation in resource availability (e.g., Ferguson et al. 1998, Johnson et

73

al. 2002b, Nams and Bourgeois 2004, Whittington et al. 2005). Behaviours associated with
movement can increase our understanding o f how wolves hunt prey and use landscapes
altered by human developments. W olf movements can be categorized as dispersal,
movements within territories, and prey searching (Mech 1974). To minimize the energetic
costs of movement or maximize encounter rates, wolves travel roads, trails or other linear
features that have little human use (Mech 1970, Thurber et al. 1994, Paquet and Carbyn
2003, Wittington et al. 2005). In the valley bottoms of Jasper, Alberta, Whittington et al.
(2005) studied the spatial responses o f wolves to roads and trails. Using snow tracking to
identify movement paths, they found that wolves avoided areas with high densities of trails
and roads. Consistent with other studies, wolves selected areas near low-use trails and
roadways (Whittington et al. 2005). McCutchen (2007), also working in Alberta, looked at
wolf use of linear corridors and how these features may be contributing to declining caribou
populations. Based on simulation models, she found that the use o f linear corridors by
wolves did not contribute to increased rates of predation on caribou. Caribou predation was
most influenced by an increase in the total number of wolves on the landscape (McCutchen
2007).
Past research has suggested wolves move more efficiently through habitats within
close proximity to linear features with low human use (Thurber et al. 1994, James and StuartSmith 2000, Whittington et al. 2005, Rinaldi 2010), but researchers have not yet looked at
the variation in movement behaviour across multiple seasonal and temporal scales in direct
relation to populations o f caribou. Studying movement parameters at both fine and coarse
scales can increase our knowledge of factors that may influence seasonal predation rates on
caribou and how the movements o f wolves are influenced by human-caused changes on the

74

landscape. Furthermore, understanding the relationship between carnivore movements and
landscape composition may have applications to other predator-prey systems influenced by
human developments (Kinley and Apps 2001, Robinson et al. 2002, Bryant and Page 2005,
Gibson 2006, Cooley et al. 2008).
In this chapter, I quantified variation in wolf movement and used these measures as
an index of wolf behaviour in relation to the distribution of woodland caribou and industrial
features. I accounted for factors such as cover type and distance to water, that may also
influence seasonal movement rates and the sinuosity of movement paths by wolves. Based
on past research and results from Chapter 2 , 1 predicted that movements o f wolves would
differ seasonally and according to the condition of the landscape. As winter progressed, wolf
movements would be less sinuous and movement rates would decrease due to the additional
energy expenditure required to travel through deep snow as well as the increased availability
of vulnerable prey across the landscape. Alternatively, during the non-winter months wolf
travel would be more efficient and movement rates would increase as a variety o f prey
species and ages (i.e., neonates, rodents, birds, etc.) become available. Sinuosity o f paths
would vary depending on the seasonal availability of prey and increase in habitats across the
study area where wolves spend more time searching and hunting.
I expected wolves to travel at increased rates and in a more linear direction in alpine
habitats where fewer vegetative barriers, changes in topography, and increased snow
hardness reduce the energetic costs of movement. As tree cover thickens, wolves would
move more slowly and sinuously. Movement rates and time spent searching and hunting
throughout non-conifer habitats would increase due to the availability of browse preferred by

75

moose, deer and elk. Likewise, searching and hunting behaviours would increase for wolves
in seasonal areas supporting populations of caribou.
If wolves in the South Peace region behave similarly to other populations across
North America, I would expect less sinuous movements across areas o f the landscape
influenced by human developments. Non-linear features with low human use would aid
behaviours of hunting and prey searching, and linear features would facilitate linear travel
and movement across pack territories. As wolves travel close to, or across early-successional
forests and where habitat for primary prey is plentiful, I would expect greater sinuosity of
movement paths as searching and hunting behaviours increase. Finally, I expected wolf
movement to differ between daily (fme-scale use) and weekly (course-scale use) spatial
scales. At the daily scale, short-term movements by wolves would indicate behaviours
associated with hunting and searching. Alternatively, I expected weekly movements, which
facilitate patrol and defense o f territories, to result in greater use of caribou habitat as wolves
had increased opportunities to use features in mountainous and boreal habitats (e.g., alpine,
established game trails) during large-scale movements.

Methods
Study Area and Wolf Telemetry
Located on the eastern slopes of the Rocky Mountains in northern British Columbia,
the South Peace study area is approximately 12,000 km2 (Figure 1, Chapter 1). Tumbler
Ridge is located near the center of the area, which then extends northwest towards the town
of Mackenzie, northeast towards Dawson Creek and south along the Alberta border. Four
Biogeoclimatic Ecosystem Classification zones characterize the study area: Boreal White and
76

Black Spruce (BWBS), Sub-Boreal Spruce (SBS), Engelmann Spruce - Subalpine Fir
(ESSF), and Alpine Tundra (AT; Meidinger and Pojar 1991). Large-scale commercial
forestry, natural gas, oil, mineral, and most recently, wind developments exist throughout the
region (Sopuck 1985, Nitschke 2008). The cumulative effects resulting from these industrial
developments have produced forested landscapes that are progressively younger and
increasingly fragmented (see Chapter 2 for a more comprehensive description o f the study
area).
Between 2008 and 2010, 16 wolves were captured and fitted with GPS collars (Lotek
Inc., Newmarket, Ontario, Canada, model: GPS 4400S). Collars were programmed to take a
location fix every three hours (n = 14; two collars were programmed for high-frequency
intervals and collected locations every 20 minutes) and were remotely downloaded from a
fixed-wing aircraft approximately bimonthly during routine tracking flights. Data were
examined for erroneous locations using the number of satellites required to obtain locations
(2D or 3D) and visual inspection (Appendix B).

Defining Seasons
I used past research to develop two biological seasons to model the movement of
wolves: non-winter (April 16 - October 14) and winter (October 15 - April 15). Non-winter
months included the time when wolves are responsible for the rearing and raising of pups and
therefore, centralize around dens, rendezvous or homesites (Mech 1970, Ballard et al. 1991).
By mid-October, pups are approximately six months old and have grown large enough to
travel with the nomadic pack as they transition towards the winter months (Packard 2003).
Winter extends through the breeding season until the wolves begin localizing around den
sites between March and May (Mech and Boitani 2003).
77

Movement Paths, Rates, and Sinuosity
I created movement paths using consecutive GPS collar locations recorded over daily
and weekly intervals. These paths allowed me to compare the relationship between
movement rate or path sinuosity and land cover, caribou habitat and disturbance variables.
Paths generated from 24-hour relocation intervals allowed me to identify fine-scale
behaviours and provided results that were comparable to past studies of wolf movement
(Fritts and Mech 1981, J?drzejewski et al. 2001, Walton et al. 2001, Whittington et al. 2005).
Wolves patrol territories in cyclic patterns approximately every week (J?drzejewski et al.
2001); therefore, I analysed movement patterns over a longer 7-day period.
I assumed a straight-line distance between consecutive GPS locations when inferring
movement paths. I used Julian dates from the GPS collars to define the temporal extent of
each 24-hour (i.e., Julian calendar date = 1, 2, 3, etc.) and 7-day path segment (i.e., Julian
calendar dates 1 - 7, 8 - 14, 15 - 21, etc.). Movement paths were considered incomplete if
the number of acquired locations was less than 50% of the total number of expected GPS
fixes for each temporally constrained interval. I pooled movement paths across individual
wolves; pooled movement paths provided sufficient sample size for statistical analysis. I
calculated movement rate as the total distance travelled (km) by individual wolves for each
daily and weekly interval. I calculated the sinuosity of each path as the total distance of all
line segments divided by the net displacement (i.e., distance between the start and end
locations o f each path).
I used polygonal buffers around each movement path to quantify the characteristics of
the landscape traversed by collared wolves. I used high-frequency location data (relocation
interval = 20 minutes) to determine an appropriate buffer for each daily and weekly

78

movement path. I used the same buffer size for daily and weekly paths; a series of daily
movements served as the foundation for calculating weekly movement paths. I grouped
high-frequency locations into 24-hour intervals and applied 100% MCPs around each
temporally constrained group of locations to represent the total area (km2) available to each
o f the collared wolves. The width o f all buffers was determined and calculated as the median
distance (km) across each daily MCP polygon.

Resource and Human Disturbance Variables
I drew from past research on wildlife-development interactions and observations of
the study area to identify a number of variables that I hypothesized would influence the
movement behaviours of wolves (Table 1, Chapter 2). I examined five classes of variables
within each buffered polygon: forest cover, caribou habitat, distance to water, and distance to
and density of disturbance features.
Habitat Variables. - Forest cover was estimated using the provincial Vegetation
Resource Inventory (VRI; BC Ministry of Forests and Range, 2007a, b). I consolidated the
vegetation types into four super-classes: alpine, conifer, deciduous, and mixed-other forests
(Table 8). Each class was converted into a binary raster layer so the average value (%) could
be extracted for each daily and weekly movement polygon. I also tested the seasonal
importance of water (proximity) as an additional predictor of wolf movement across the
landscape. Water features included lakes, rivers, creeks/streams, and reservoirs.
Values from the spatial resource selection function (RSF) analyses (Chapter 2) for
Bearhole/Redwillow (BHRW) and Quintette caribou were extracted for each season. Non­
winter represented the median value for caribou habitat modeled during the spring, calving,
and summer/fall, whereas winter was used in its original context.
79

Table 8. Description of variables used to model movement o f wolves across the South Peace region o f northeastern British Columbia.
Variable
Alpine
Conifer

Deciduous
(Decid)
Mixed-Other
Water
BHRW
Quintette (Q)
Road
SeisPipln
Cutblock (Ctblks)
Mine
Oil and Gas (OG)
LF Dens
NLF Dens

Description
high elevation with few or no trees with primary cover being rock, snow, herbs, shrubs, bryoids and terrestrial
lichens
includes black spruce (Picea mariana), tamarack (Larix laricina), subalpine fir {Abies lasiocarpa), lodgepole pine
{Pinus contorta) and whitebark pine {P. albicaulis), other spruce varieties: Picea spp., Engelmann {P.
engelmannii), white {P. glauca),hybrid {P. engelmannii x glauca),
includes aspen {Populus tremuloides), cottonwood {P. balsamifera), birch (Betula papyrifera)
includes Douglas-fir {Pseudotsuga menziesii), upland areas dominated by talus, rock, snow, tailing ponds, herbs
(forbs, graminoids), bryoids and shrubs
distance to water (km)
RSF values for caribou in the Bearhole/Redwillow herd
RSF values for caribou in the Quintette herd
distance to road (km)
distance to seismic line and/or pipeline combined (km)
distance to forestry cutblock (km)
distance to coal mine footprint (km)
distance to non-linear oil and gas well pad or facility pad > 1 hectare in size (km)
density (ha/km) of linear features on the landscape (roads, seismic lines, and pipelines)
density (ha/km2) o f non-linear features on the landscape (cutblocks, mine, oil, and gas facilities)

There was no overlap in the range of the Lower and Upper Sukunka wolf packs and the
BHRW caribou herd; thus, I did not apply RSF values to those movement paths. Similarly,
because wolves in the Chain Lakes pack do not have opportunities to overlap with caribou in
the Quintette herd, the caribou habitat variable was excluded from those seasonal movement
models.
Disturbance Features. - 1 used databases from government and industry to identify
the location of disturbance features across the South Peace region (BC Land and Resource
Data Warehouse 2007, Oil and Gas Commission of BC 2009, West Fraser Timber Company
Ltd., Western Coal, Inc., Peace River Coal Ltd.). Following methods from Chapter 2 , 1 used
the most parsimonious moving window (1.56 hectares), identified during the RSF analysis to
calculate the density o f industrial features (linear: ha/km; non-linear: ha/km2). I combined
spatial data for forestry (cutblocks) and mine/oil/gas to create a variable representing the
density of non-linear features (ha/km2). GIS calculations for distance and density were
computed using IDRISI (The Andes Edition; Eastman 2006). I used Hawth’s Tools and
GME (Spatial Ecology LLC 2009) in ArcGIS 9.3 (2009; ESRI, Redlands, CA) to create and
develop daily and weekly movement paths for wolves, as well as to attribute habitat, caribou
RSF and disturbance values to movement paths.

Modeling Movement o f Wolves
I used mixed effects generalized linear models to statistically relate movement
distance and sinuosity to landscape variables recorded within the area (km2) buffered around
each 24-hour or 7-day movement interval. Pooling movement paths for wolves from all
packs resulted in a nested sampling design. Adding a random effect accounted for additional
variation that may have occurred among individuals or packs (Gillies et al. 2006,
81

Hebblewhite and Merrill 2008). I conducted a sensitivity analysis to determine if additional
variation was best described using a random effect for individual wolf, pack, or wolf and
pack. Each model contained a random effect for “pack”.
I used linear regression to model movement rate. I used a square root transformation
to normalize those data. Because of extremely non-normal data, I transformed the sinuosity
measures into binary categories and applied logistic regression. I used the median value
across each seasonal dataset to classify paths as high (1) or low (0) sinuosity.
I built a suite of 18 ecologically plausible candidate models to determine the
influence of habitat and disturbance variables on wolf movement (Table 9). Variables for
distance (km) and density (total area of features/unit area; linear features = ha/km, non-linear
features = ha/km2) were modeled as linear and as 2-term Gaussian functions (distance to road
+ distance to road squared) for each season. I used tolerance scores (> 0.2) and visual
inspection of bivariate correlation matrices to assess excessive multicollinearity. Where
collinearity occurred between disturbance variables, I preferentially removed non-linear
features to retain the oftentimes more abundant linear features.
I used the AICc (A) difference to select the most parsimonious fixed effects linear or
logistic regression model for each season (Burnham and Anderson 2002). If competing
models were present, I considered the model with the smallest A AICc to be the most
parsimonious. I applied the random effect to the most parsimonious model and reran the
analysis to generate model coefficients. I then used the coefficient of determination (R2) to
assess predictive fit for linear regression models. I partitioned w olf movement paths into
training (80%) and testing (20%) groups.

82

Table 9. Candidate models to examine the movement o f wolves monitored between 2008 - 2010 across the South Peace region of
northeastern British Columbia. Each model (except Land cover) was fit as either a linear or Gaussian (* squared) term depending on
best fit for each movement parameter and season. Distance was measured in kilometers (km) and density was measured in
hectares/unit area (linear features = ha/km and non-linear features = ha/km2).
M odel G roup
H abitat

L in ear F eatu res (LF)

C u m u lativ e E ffects (CE)

M odel N am e
L an d cover

M odel V ariables
% land co v er (alpine, conifer, decid u o u s, m ix ed-species)

C aribou/W ater (C arW at)

C arib o u R SF + w ater d istance (D ist)

C aribou/W ater (C arW at)*

C arib o u R SF + w ater (D ist)2

Landscape*

% land cover + C aribou R SF + w a ter (D ist)2

R oad D istance (D ist)

L andscape + R oad Dist

R oad D istance (D ist)*

L andscape + R oad D ist2

L F (roads, seism ic lines and/or pipelines) D istance (Dist)

L andscape + L F Dist

L F D istance (D ist)*

L an dscape + LF D ist2

LF D ensity (LF D ens)

L andscape + LF D ens

L F D ensity (L F D ens)*

L andscape + LF D ens2

L F Total (LF C E)

L an dscape + L F D ist + L F D ens

L F T otal (LF CE)*

L andscape + LF D ist2+ L F D e n s2

C E D istance (C E D ist)

L andscape + LF D ist + N o n -L in ear F eature (N L F ) D ist

C E D istance (C E D ist)*

L andscape + LF D ist2 + N L F D ist2

C E D ensity (C E D ens)

L an dscape + LF D ens + N L F D ens

C E D ensity (C E D ens)*

L andscape + LF D ens2 + N L F D ens2

C E T otal (C E )

L andscape + LF D ist + LF D ens+ N L F D ist + N L F D ens

C E T otal (CE)*

L andscape + L F D ist2+ L F D ens2 + N L F D ist2+ N L F D en s2

Using the withheld data, I assessed residuals to determine if there was a relationship between
the observed values and the predicted movement rates. I also evaluated fit for the top-ranked
logistic regression (sinuosity) model by calculating the area under the receiver operating
characteristic curve (ROC; Hosmer and Lemeshow 2000).

Results
I used a total of 25,254 GPS locations collected from wolves to develop 3,749 daily
and 493 weekly movement paths. Two wolves of the Chain Lakes pack provided an
additional 8,493 high-frequency locations (n = 168 daily MCPs). The daily area used by
these wolves had a median width of 4.44 km; I used these data to identify the area o f use
around each daily and weekly movement path. In general, I observed variation between
annual and seasonal movement rates and path sinuosity when movements were pooled for
collared wolves across the South Peace study area (Figure 18). As predicted, movement rates
of wolves were highest during the non-winter season. However, seasonal variation in
movement was greater than variation in the use or proximity to linear and non-linear features,
suggesting that other factors also influenced the movement dynamics of wolves (Figure 19).
For each season, the most parsimonious models for daily and weekly movement rates
were also the most complex and contained variables for all habitat cover types and humancaused disturbances (Table 10). Models with a random effect for pack performed best across
all seasonal movement rate and sinuosity models. One model (daily movement rate during
the winter season) was an exception and performed better with a random effect for individual
wolf. More than 30% o f the variation in movement rate was explained by the weekly (non­
winter R2 - 0.3429, winter R2 = 0.5012) regression models.

84

2008

I•of 12

Month

2008

B

2010

100

> 40

-20

M onth
M ovem en t Rate

»

Sinuosity

80

Jan

Feb

Mar April

M ay June

July

Aug

Sept

Oct

Nov

Dec

Month

Figure 18. Mean monthly (±SE) movement rates (km/day) and sinuosity for wolf movement
paths sampled daily across the South Peace region of northeastern British Columbia.
Movement paths were pooled for wolves by year (A, B) as well as across all years (C; 2008 2010 ).
85

— • — LF D ens

M o v em e n t R ate (km /day)

Daily Sinuosity

LF D ens

NLF D ens

A

M o v em e n t R ate (k m /w eek )

W eekly Sinuosity (Log)

’♦ “"NLF D ens

NLF D ens

LF D ens

B

NLF D ens

80
70

50
° 40
30

Jan

Feb

M ar April

M ay

Ju n e

July

Aug

S ept

O ct

Nov

Dec

Figure 19. Mean (±SE) monthly (2008 - 2010) movement rates (A, B) and sinuosity (C, D) for daily (km/day) and weekly (km/week)
sampling periods as they relate to densities o f linear (ha/km) and non-linear features (ha/km2) across the South Peace region o f
northeastern British Columbia.

Table 10. Number o f parameters (k), Akaike’s Information Criterion (AICc) and AICc weights (AIC*) for linear regression models
describing seasonal daily and weekly movement rates o f wolves. Models were developed for wolves monitored between 2008 and
2010 across the South Peace region of northeastern British Columbia. Model covariates are given in Table 9 and sample size o f
seasonal movement paths is indicated in parentheses.
Weekly

Daily
Non-Winter (n = 1599)i

Winter (n = 1403)

Non-Winter (n = 212)

Winter (n = 186)

Land cover

k
5

A IQ
5172.64

AAIC
97.49

<0.001

AICc
4333.39

AAIC
226.19

AIC„.
<0.001

AICc
327.47

AAIC
21.50

AIC m
<0.001

AICc
349.33

AAIC
81.57

AIC„
<0.001

CarWat

4

5220.72

145.57

<0.001

4334.25

227.05

<0.001

349.14

43.17

<0.001

349.41

81.65

<0.001

CarWat*

5

5209.19

134.04

<0.001

4305.40

198.19

<0.001

347.59

41.62

<0.001

350.59

82.83

<0.001

Landscape*

9

5166.58

91.43

<0.001

4302.65

195.45

<0.001

330.23

24.26

<0.001

342.88

75.11

<0.001

Road Dist

9

5177.73

102.58

<0.001

4329.89

222.68

<0.001

331.47

25.50

<0.001

346.86

79.10

<0.001

Road Dist*

11

5165.58

90.43

<0.001

4278.91

171.71

<0.001

331.67

25.70

<0.001

342.34

74.58

<0.001

LF Dist

10

5179.74

104.59

<0.001

4331.52

224.31

<0.001

333.55

27.58

<0.001

344.81

77.05

<0.001

LF Dist*

13

5162.24

87.09

<0.001

4274.96

167.75

<0.001

323.62

17.65

<0.001

335.45

67.69

<0.001

LF Dens

9

5180.51

105.36

<0.001

4329.65

222.45

<0.001

331.24

25.27

<0.001

345.45

77.69

<0.001

LF Dens*

11

5165.81

90.66

<0.001

4276.23

169.03

<0.001

328.12

22.15

<0.001

316.67

48.90

<0.001

LF CE

11

5181.76

106.61

<0.001

4333.38

226.18

<0.001

334.89

28.92

<0.001

346.91

79.15

<0.001

LF CE*

15

5157.87

82.72

<0.001

4204.44

97.24

<0.001

316.03

10.06

0.01

293.83

26.07

<0.001

CE Dist

12

5154.54

79.39

<0.001

4330.62

223.42

<0.001

329.24

23.27

<0.001

345.00

77.24

<0.001

CE Dist*

17

5131.77

56.62

<0.001

4249.40

142.20

<0.001

316.38

10.41

0.01

338.90

71.13

<0.001

CE Dens

10

5164.93

89.78

<0.001

4331.52

224.32

<0.001

332.10

26.13

<0.001

346.01

78.25

<0.001

CE Dens*

13

5130.10

54.95

<0.001

4260.68

153.48

<0.001

320.37

14.40

<0.001

320.36

52.60

<0.001

CE

14

5154.84

79.69

<0.001

4329.71

222.50

<0.001

333.65

27.68

<0.001

327.71

59.95

<0.001

CE*

21

5075.15

0.00

1.00

4107.20

0.00

1.00

305.97

0.00

0.99

267.76

0.00

1.00

Model

♦Gaussian (squared) term

a ic h

When modeling daily movement rates, less than 20% of the overall variation was explained
(non-winter R2 = 0.117, winter R2 = 0.1746) by variables representing habitat and disturbance
features. Inspection of residuals indicated the ability of models to predict daily and weekly
movement rates for wolves was generally good as values were evenly dispersed around zero.
Similar to movement rate, the most parsimonious daily logistic regression models for
sinuosity were also the most complex in each candidate set (Table 11) and received strong
support (AICW> 0.95 for both seasons). Area under the curve (AUC) scores indicated poor
predictive performance for the best-ranked seasonal models (non-winter AUC = 0.55, winter
AUC = 0.62). Results from the weekly sinuosity models indicated multiple candidate models
had reasonable support compared to the top-ranked model (Table 11). Models were not
averaged due to the complexity of the random effect. During the non-winter season, the most
parsimonious model for weekly sinuosity included covariates for forest cover, distance to
water, caribou habitat and distance to linear features. The top model for weekly sinuosity
during winter was similar to non-winter, but included covariates for linear feature density (in
addition to distance) across the landscape. Model fit was generally poor for weekly sinuosity
models during the non-winter season (AUC = 0.64), but improved during winter (AUC =
0.75).

Cumulative Effects o f Industrial Disturbances on Seasonal Wolf Movements
Non-Winter. - Daily movement rates for wolves decreased as they traveled closer to
water features and cutblocks (Figure 20). However, the large confidence interval
surrounding the coefficient for cutblocks suggested considerable variation in response by
wolves.

88

Table 11. Number o f parameters (k), Akaike’s Information Criterion (AICc) and AICc weights (AICW) for logistic regression models
describing seasonal daily and weekly sinuosity o f w olf movements. Models were developed for wolves monitored between 2008 and
2010 across the South Peace region of northeastern British Columbia. Model covariates are given in Table 9 and sample size o f
seasonal movement paths is indicated in parentheses.
Daily

Weekly

M odel
Land cover

k
5

Non-Winter (n = 1599)
AICc
AAIC
a ic m
2220.91
8.88
0.01

Winter (n = 1403)
AICc
AAIC
AIC„
1950.97
30.61
<0.001

Non-Winter (n = 212)
AAIC
AICc
AIC„
300.58
3.75
0.05

W inter (n = 186)
AICc
AAIC
AIC„
250.70
10.21
<0.001

CarWat

4

2222.94

10.91

<0.001

1951.25

30.89

<0.001

299.26

2.43

0.10

263.77

23.28

<0.001

CarWat*

5

2221.59

9.56

0.01

1951.54

31.18

<0.001

301.16

4.34

0.04

264.82

24.33

<0.001

Landscape*

9

2223.52

11.49

<0.001

1954.94

34.57

<0.001

305.04

8.22

0.01

240.68

0.19

0.24

Road Dist

9

2226.88

14.85

<0.001

1954.96

34.60

<0.001

304.95

8.12

0.01

243.46

2.97

0.06

Road Dist*

11

2217.98

5.96

0.04

1955.19

34.83

<0.001

309.25

12.43

<0.001

244.43

3.94

0.04

LF Dist

10

2227.13

15.11

<0.001

1954.94

34.58

<0.001

297.92

1.09

0.20

244.14

3.65

0.04

LF Dist*

13

2221.83

9.81

0.01

1956.63

36.26

<0.001

304.52

7.69

0.01

246.54

6.05

0.01

LF Dens

9

2226.70

14.68

<0.001

1957.07

36.71

<0.001

303.62

6.80

0.01

244.18

3.69

0.04

LF Dens*

11

2225.78

13.75

<0.001

1953.28

32.92

<0.001

307.80

10.98

<0.001

242.31

1.82

0.11

LF CE

11

2228.67

16.64

<0.001

1956.66

36.30

<0.001

300.08

3.25

0.07

244.36

3.87

0.04

LFC E*

15

2223.77

11.74

<0.001

1956.09

35.72

<0.001

305.96

9.14

<0.001

246.77

6.28

0.01

CE Dist

12

2225.98

13.95

<0.001

1951.49

31.13

<0.001

296.82

0.00

0.35

247.18

6.69

0.01

CE Dist*

17

2224.52

12.50

<0.001

1933.46

13.10

<0.001

304.09

7.27

0.01

250.94

10.45

<0.001

CE Dens

10

2223.79

11.76

<0.001

1957.02

36.66

<0.001

300.67

3.85

0.05

242.17

1.68

0.12

CE Dens*

13

2218.63

6.60

0.03

1951.75

31.39

<0.001

307.12

10.30

<0.001

240.49

0.00

0.27

CE

14

2227.79

15.76

<0.001

1955.33

34.97

<0.001

299.51

2.69

0.09

248.48

7.99

<0.001

CE*

21

2212.03

0.00

0.88

1920.36

0.00

1.00

310.40

13.57

<0.001

252.27

11.78

<0.001

‘ Gaussian (squared) term.

Conifer
M ixed O th e r

2 -

Ctblks
LF D ens
Alpine

o
*m
o>
+i
2
e
•o
£

NLF D ens

Road

W a te r

SeisPipIn

BHRW

♦
W a te r*

i

M ine

+- 1* ♦ ..
Q

Road* Se,sP'P ln * c t b |ks* M ine*

-1

♦

NLF Dens*

LF Dens*

Decid

-2

B
LF D ens

1.5
Conifer
Alpine
Ctblks
Road
u

*in

cn

NLF Dens

0.5
Decid T W a te r

c
a*

0
£
2

SeisPipIn

BHRW
W a te r*

SeisPipIn'

Q
Road*

M ine

Ctblks*

M ine*

jr
NLF Dens*

-0.5
M ixed O th e r

LF Dens*

-1

Figure 20. Coefficients for the parameters in the most parsimonious mixed-effects models
for daily (A; n = 1,599) and weekly (B; n = 212) movement rates during the non-winter
season for wolves in the South Peace region of northeastern British Columbia. An asterisk
(*) indicates a Gaussian term and model variables are given in Table 8.
90

Higher daily and weekly movement rates were associated with high densities o f non-linear
features (oil and gas well sites, facility stations, coal mines). Weekly movement rates
decreased slightly near coal mines, but increased in alpine habitats. Roads were the only
linear feature wolves responded to during the non-winter months; movement rates decreased
as wolves travelled within close proximity to roads at the weekly scale.
During the non-winter season, movement paths became increasingly linear as wolves
travelled near deciduous habitats and areas with higher densities o f non-linear features
(Figure 21). In addition, linear movements increased slightly near roads and coal mines as
wolves traversed across the landscape. Sinuosity o f weekly movement paths increased near
conifer forests, mixed-species forests and cutblocks. At the daily scale, I observed a slight
increase in sinuous movements in habitats of high quality for the BHRW and Quintette
caribou herds. Similarly, sinuosity of weekly paths increased slightly in BHRW, but not
Quintette habitat.
Winter. - Daily movement patterns o f wolves during the winter season were
influenced more by forest cover, caribou habitat and disturbance features than during non­
winter (Figures 21, 22). Daily movement rates decreased in habitats dominated by conifer
forests; weekly movement rates increased in mixed-species forests. Movement rates
decreased as wolves approached water features and cutblocks (Figure 22). Although to a
lesser extent, movement rates of wolves also decreased near roads, seismic lines, pipelines,
and mines during winter. However, as densities o f linear and non-linear disturbance features
increased across the landscape, wolves increased daily and weekly movement rates. Daily
movement rates also increased in habitats important to BHRW caribou.

91

1.5
M ixed O th e r
Decid

0.5

LF Dens*

u

*in

BHRW Q

♦ ...♦

*_

..

W ate r*

c
41

NLF Dens*

Road* S eisP ipIn*Ctblks* M ine*
M ine

SeisPipIn

\

0

£

8 -0.5
Road

NLF D ens

W ate r

-1

Ctblks

LF Dens

Alpine
Conifer

-1.5

B

Alpine
M ixed O th e r

C onifer

u
at

Road

ae
'5

Ctblks

BHRW

£

SeispipIn

3

i

M ine

r

—♦ —

OG

l -2
W a te r

-4

Decid

-6

Figure 21. Coefficients for the parameters in the most parsimonious mixed-effects models
for daily (A; n = 1,599) and weekly (B; n = 212) sinuosity during the non-winter season for
wolves in the South Peace region of northeastern British Columbia. An asterisk (*) indicates
a Gaussian term and model variables are given in Table 8.
92

W a te r
LF D ens

Mixed O th e r
Ctblks

Road

NLF Dens
BHRW
Decid

M ine

SeisPipIn

Ctblks*
CO. - 1

M ine

NLF Dens*
LF Dens

W a te r'

1 C onifer
A lpine

10

B

W ate r

8

6
4

u
2
a?
in
0i
a 0
S

LF Dens
C onifer
M ixed_O th
Ctblks

. 1

1 .

SeisPipIn

BHRW

♦ ...♦

I
♦

Road* SeisPipIn* Ctblks*

■o

NLF Dens

M ine

♦

,.a

♦ ------M ine*

^ NLF Dens*

-2
Decid

-4

LF Dens*
W ater*

Alpine

-6

-8
Figure 22. Coefficients for the parameters in the most parsimonious mixed-effects models
for daily (A; n = 1,403) and weekly (B; n = 186) movement rates during the winter season for
wolves in the South Peace region of northeastern British Columbia. An asterisk (*) indicates
a Gaussian term and model variables are given in Table 8.
93

Sinuosity of daily and weekly paths decreased in habitats dominated by deciduous
forest (Figure 23). Conifer forests facilitated increased daily sinuosity, but decreased
sinuosity of weekly movement paths. Wolves demonstrated linear travel in alpine habitats at
the scale of a week, but not the day. During winter, the sinuosity of wolf movements
increased slightly in habitats o f high quality for Quintette caribou. Wolves demonstrated
increased linear travel where the density of non-linear features was high and in habitats
valued as important to BHRW caribou. As wolves traveled close to coal mines, I observed a
slight relative increase in sinuous movements.

94

C onifer

Alpine

G 1
*LSI

W ater*

Ol

f

S o
0

1

Q

Road* S e is P ip In * ^

-----SeisPipIn

TLF Dens*
_ NLF Dens*

M ine
4►

BHRW

■ -i

M ine*

I
LF D ens

-INLF Dens

Road

Decid
W a te r

Ctblks

M ixed O th e r

-3

B

15
W ater*
M ixed_O ther

10

|_p Dens

j

5

u
*

0
c

a

-5

E

-15

+i
£ -io H
a>
*0

LF Dens*

I ?
I

T

pine

T

1
BHRW

NLF D en s*

M

♦

NLF D ens

Deci

C onifer

£ -20
-25
-30
-35
W a te r

-40

Figure 23. Coefficients for the parameters in the most parsimonious mixed-effects models
for daily (A; n = 1,403) and weekly (B; n = 186) sinuosity during the winter season for
wolves in the South Peace region o f northeastern British Columbia. An asterisk (*) indicates
a Gaussian term and model variables are given in Table 8.
95

Discussion
I used two parameters o f movement as an index of wolf behaviour across forested
boreal and mountainous environments occupied by woodland caribou. Considering the large
range of factors that influence animal movement and the broad spatiotemporal scales of
analysis I developed, the majority o f explanatory models had strong statistical relationships.
My results indicated that the cumulative effects from industrial disturbances had an influence
on the movement behaviour of wolves in both environments. Past studies of wolf movement
have not quantified compounding effects from multiple sources o f human disturbances (e.g.,
forestry and oil/gas extraction), determined how these behaviours change across
spatiotemporal scales, or examined how wolves move across areas supporting populations of
caribou (but see Kuzyk et al. 2004, Neufeld 2006, Houle et al. 2010, Latham et al. 201 lc).
Following my predictions, the influence of habitat and development features on movement
varied across season and scale (Table 12).
At the weekly scale, my results indicated that movement rates were generally higher
for wolves across the South Peace region during the non-winter months (Figure 19). If wolf
packs across the study area successfully reproduced throughout the duration of this study,
increased movement rates (up to 2 km/hr) could result from wolves rapidly travelling back to
dens or homesites after feeding bouts (Mech 1994). However, as responsibilities associated
with pup care are dependent on an individual’s pack status and because I pooled movement
rates, behavioural interpretation remains challenging without investigating the direct
ecological determinants of path characteristics (e.g., behaviour, activity type or association
with a den or homesite).

96

Table 12. The predicted and observed variation ( “T = increased, 4^ = decreased) in movement using movement rate and path sinuosity
as indices o f wolf behaviour across the South Peace region o f northeastern British Columbia. If observed movements were scale- or
season-dependent, results are indicated in parentheses (seasonal: NW = non-winter, W = winter; scale: daily or weekly).
M ovem ent Index
Factor
Season/Scale
N on-w inter

W inter

D aily M ovem ents

W eekly M ovem ents

H a b ita t C lass
A lpine
Forest cover type:
conifer
Forest cover type:
m ixed-species
Forest cover type:
deciduous

Hypothesized M ovem ent Response o f W olves
M ovem ent rates increase in response to reproduction
and greater availability o f prey. Sinuosity o f
movements decrease concurrent with less human
disturbance.
M ovem ent rates and sinuosity decrease in response to
greater snow accum ulation and availability o f
vulnerable prey.
M ovement rates decrease and sinuosity increase as
short-term m ovem ents are associated with hunting and
searching o f prey.
M ovem ent rates increase and sinuosity decrease as
long-term m ovem ents are associated with territory use
and patrol.

Predicted

M ovem ent Rate
Observed

Path Sinuositv
Predicted
Observed

T

T

i

N ot statistically
influential

i

1

i

N ot statistically
influential

i

i

T

t

T

T

1

1

M ovem ent rates increase and sinuosity decrease in
response to reduced travel resistance.

f

t (weekly)

I

l (weekly)

M ovem ent rates decrease and sinuosity increase in
response to greater prey availability and selection o f
habitats for den/hom esites.
M ovem ent rates decrease and sinuosity increase in
response to greater prey availability and selection o f
habitats for den/homesites.
M ovem ent rates decrease and sinuosity increase in
response to greater prey availability and selection o f
habitats for den/homesites.

|

NW: N ot statistically
influential, W: j (daily)

T

N W : t , W: f

I

NW : N ot statistically
influential, W: f
(weekly)
NW : N ot statistically
influential

f

N W : t, W: T
(weekly)

t

N W : I (weekly),
W: 1

I

Table 12. Continued.

Factor

Hypothesized M ovem ent Response o f W olves

M ovem ent Index
M ovem ent Rate
Path Sinuositv
Predicted
O bserved
Predicted
Observed

Caribou Habitat
RSF Caribou BH RW ( Ch.2)

M ovem ent rates decrease and sinuosity increase in
response to greater availability o f caribou as prey.

1

NW : N ot statistically
influential, W: f

T

NW : t , W: 4 (daily)

R SF Caribou Q uintette ( Ch.2)

M ovem ent rates decrease and sinuosity increase in
response to greater availability o f caribou as prey.

i

N ot statistically
influential

T

T (slight)

M ovem ent rates increase and sinuosity decrease in
response to reduced travel resistance. Linear features
facilitate rapid travel by wolves.

t

NW : J, (roads, weekly),
W: I (daily)

i

N on-linear feature - M ovem ent rates decrease and sinuosity increase in
Proxim ity
response to greater prey availability.

i

NW : f (cutblocks, daily;
mines, weekly), W: 4
(daily)

T

N W : 4 (roads,
weekly), W: N ot
statistically
influential
NW : | (cutblocks,
daily), 4 (mines,
weekly), W: 4
(m ines, daily)

L inear feature Density

t

NW : N ot statistically
influential, W: |

i

N ot statistically
influential

1

t

T

4 (daily)

Human
Disturbances
L inear feature Proxim ity

M ovem ent rates increase and sinuosity decrease in
response to reduced travel resistance and low human-

N on-linear feature - M ovem ent rates decrease and sinuosity increase in
Density
response to greater prey availability.

At the scale of weekly movements, wolves traveled through alpine habitats at greater
speeds, but only during the non-winter months (Table 12). Alpine habitats did not affect
movement rates in winter, but resulted in more linear travel as wolves traversed the
landscape. Past studies found that wolves avoided conifer forests during winter (Kunkel and
Pletscher 2001, Houle et al. 2010, Milakovic et al. 2011). In partial support to this, my
results suggest that conifer forests facilitated linear, or direct travel through these forests as
wolves traversed the landscape during winter. Linear travel has been shown to aid in the
maintenance of territories and facilitate the element o f surprise during hunting bouts (Mech
et al. 1998). Dissimilar to the abovementioned studies, however, slow and sinuous daily
movements in conifer habitats indicated that wolves also demonstrated behaviours associated
with brief hunting bouts.
Unlike conifer and mixed-species forests, deciduous habitats facilitated linear
movements during both seasons. Hebblewhite et al. (2009) and Laurian et al. (2008)
reported that populations of moose, deer, and wolves were better supported in post-harvested
forests where deciduous species prevailed. My results did not detect increases in sinuosity or
variations in the rate o f movement at either daily or weekly scales across such habitats.
Linear travel by wolves through deciduous forests during winter might suggest ungulates
seasonally shift habitat use, or that wolves, by remaining unpredictable, avoid creating
localized reductions in prey (Mech and Boitani 2003, J^drzejewski et al 2001, Fortin et al.
2005). The adoption o f linear movements by wolves in deciduous forests could also be in
response to environmental cues or interactions with the structure of the landscape which I did
not assess during this study. Furthermore, sinuous movements associated with hunting may
become detectable only at finer observational scales (Morales et al. 2010).

99

Water features influenced movement parameters at both spatiotemporal scales (Table
12). Throughout the year, wolves traveled more slowly near water, but this response was
statistically significant for daily models only. Lowland or riparian habitats provide wolves
with increased opportunities to hunt moose, deer, and beaver and are also important in the
selection o f natal dens and homesites (Mech 1970, Packard 2003, Latham 2009). As
important as lakes, rivers, or creeks may be for prey searching and hunting, water features
did not facilitate direct travel in winter, as weekly movement paths were increasingly
sinuous.
Habitats important to caribou did not influence the movement rates o f collared wolves
during the non-winter season, although sinuosity increased slightly at both scales during this
period. Spatial separation between BHRW caribou and wolves may occur in the boreal
forest as wolves were observed travelling more rapidly in habitats classified as black spruce,
tamarack, or other peatland-type complexes - habitats that may also become seasonally void
of non-caribou prey (i.e., moose; James et al. 2004, Chapter 2). However, if populations of
caribou remain small and isolated, even an opportunistic kill in these habitat types will
influence the decline of small remnant populations (Kinley and Apps 2001, Festa-Bianchet et
al. 2006, McLellan et al. 2010).
In contrast to movements in habitats used by caribou in the BHRW herd, the sinuosity
of movement paths for wolves increased slightly in habitats used by Quintette caribou (Table
12). A covariate representing density o f primary prey, which is unavailable across the South
Peace region, would further our understanding o f predation behaviour by wolves in caribou
habitat as well as how movements through particular habitats influence other inter-specific
interactions. My study provides supportive evidence (i.e., prey selection, count and

100

movement analyses) that encounters between caribou and wolves resulting from increased
use o f disturbance features is less significant to population declines than the potential number
and variety of alternate prey available to wolves across the landscape (McCutchen 2007,
Latham 201 la, b, Tremblay-Gendron 2012).
Industrial disturbances influenced the movement behaviour o f wolves at both
spatiotemporal scales throughout the year. Practices related to forestry, natural gas, oil, and
coal extraction all rely on an affiliated network o f roads for site access. During the non­
winter months, wolves decreased weekly travel rates and sinuous movements near roads.
Levels of human activity drop during the non-winter months, thus allowing low-risk
opportunities for wolves to travel along road corridors for short intervals of time.
During winter, close proximity o f wolves to roads may have been associated with
searching and hunting for prey (Table 12). Moose select habitats near roads to forage not
only on abundant vegetation and accumulating mineral deposits, but also to travel across
seasonal ranges (Fraser and Thomas 1984, Child et al. 1991, Rea 2003, Laurian et al. 2008).
If roads were simply used by wolves to increase travel efficiency, I would expect an increase
in movement rates and linear travel. My data do not support this hypothesis as increased
linear movements were observed only at larger spatial scales during non-winter, when human
activity related to industrial activities is less. Therefore, I suspect that road corridors are
more important for hunting than increased travel efficiency.
The cumulative effects of roads and other linear features had the greatest influence on
wolf movements at the daily scale during winter. Daily movement rates decreased near
seismic lines and pipelines (in addition to roads) until these features became abundant across
the landscape, at which point rates increased. In conjunction with decreasing rates, the

101

cumulative densities of roads, seismic lines and pipelines did not result in more linear travel
for wolves at either the daily or weekly scale. Slower movement rates during winter suggest
that wolves hunt along these corridors, but that behaviour is restricted to the linear corridor,
at least for the scales of movement that I observed.
Contrary to movements near individual linear features, wolves travelled quickly
through habitats with high densities of linear features (Table 12). This suggests that there
may be a landscape effect, where wolves can exploit individual features within their range
for hunting, but high densities of features, and associated human disturbance, result in the
avoidance o f such areas. Similar avoidance responses of high densities o f linear features
were noted for wolves studied in Alberta and Quebec (James 1999, Whittington et al. 2005,
Hebblewhite and Merrill 2008, Houle et al. 2010).
Non-linear industrial features affected movement parameters of wolves more
consistently than linear features at both scales. Movement rates decreased each season near
forestry cutblocks and coal mines, but increased at both scales once the density of features
increased (Table 12). The initial phases of exploration or construction associated with forest
harvesting and oil and natural gas extraction result in high levels o f disturbance to wildlife
(Bradshaw et al. 1997). Well sites and other oil and gas facilities differ from cutblocks in
that human presence occurs throughout the year once initial developments are complete.
Across the South Peace region, activity at such sites peaks between September and
December; increased rates and linear movements near oil and gas extraction facilities during
winter suggest wolves move rapidly through these areas in response to higher levels of
human activity (Houle et al. 2010, Energy and Resources Conservation Board 2011). For
this study, I did not classify the age o f disturbance features, or the level o f human activity

102

associated with each disturbance type. However, such data could contribute to our
understanding of the movement behaviours of wolves relative to prey availability and wolf
tolerance for human presence.
In conclusion, habitat features had a strong influence on the movement parameters of
wolves when considering broad spatiotemporal scales. Disturbance features facilitated
behaviours associated with hunting and searching more during winter. In addition, wolves
generally increased linear movements in winter when territory patrol intensifies during the
breeding season. As suggested by Mech et al. (1998), linear travel may also facilitate the
element o f surprise during hunting. Wolves decreased movement rates when close to
disturbance features, suggesting that hunting behaviours are associated with those sites.
Increased movement rates and linear travel through habitats containing high densities of
disturbance features suggest wolves avoid spending time in high-risk areas associated with
human presence. Due to the complexity of cumulative effects from activities associated with
resource exploration and extraction, I was unable to detect obvious correlations between wolf
movement and increased predation risk for caribou. However, patterns o f w olf movement
(i.e., increased sinuosity and decreased movement rates) indicate caribou are most vulnerable
to predation when in close proximity to disturbance features (Table 12).
Recorded movements provide researchers with an opportunity to better understand
population dynamics as they relate to finer-scale animal behaviours (Turchin 1998).
However, it is important to consider the scale o f data collection, sample size and accuracy of
locations when interpreting behavioural patterns from animal movement paths. My results
provide some novel insights on the responses of wolves to landscape heterogeneity, but only

103

across a small range of behavioural scales and with limited inference to the mechanisms
influencing movement.
Quantifying movements of wolves across two spatial scales furthers our
understanding of wolf distribution in habitats supporting populations of caribou within a
matrix of industrial developments. Daily movements o f collared animals provide evidence
that wolf behaviour is driven by a combination o f seasonal life cycle stages, environmental
factors, prey availability and human disturbances (Peters and Mech 1975, Bibikov et al.
1985, J?drzejewski et al. 2001). Weekly movements corroborate that wolf behaviour is
associated with territory maintenance and patrol (Mech et al. 1998, J^drzejewski et al. 2001).
In conjunction with past studies, my results demonstrate that movement behaviours of an
apex predator are seasonally influenced by complex relationships that occur at multiple
spatiotemporal scales (Johnson et al. 2002a, Morales and Ellner 2002).

104

Chapter 4: General Research Summary

105

Woodland caribou are a species of increasing conservation concern. Habitat
alteration, disturbance and predation have resulted in population declines across much of
their distribution (Vors and Boyce 2009, Festa-Bianchet et al. 2011). Activities related to
large-scale resource exploration and extraction displace caribou from areas of their range and
result in early serai habitats. This change in landscape structure increases the accessibility of
caribou habitat, influences movement efficiency of predators, and supports a broader
distribution and density o f predators that use caribou as a secondary or alternate prey species
(Fuller and Keith 1981, James et al. 2004, Johnson et al. 2004a, Wittmer et al. 2007,
DeCesare et al. 2010).
My study was conducted at a regional-scale and was designed to quantify the impacts
o f a variety of human-disturbance features on wolf and caribou interactions across a rapidly
developing landscape in northeastern British Columbia. Because caribou herds within the
study area winter in different habitats (Figures 2, 3; Appendix A), I quantified seasonal
variations in wolf distribution in the context of herd-specific distribution strategies. To
address these objectives, I used animal locations collected with global positioning system
(GPS) collars, field data and information theoretic model comparisons to develop seasonal
resource selection functions (RSFs) for caribou, count models of habitat occupancy by
wolves and movement parameters that served as indices o f wolf behaviour.
In Chapter 2 , 1 used RSFs to quantify the spatial relationships between two herds of
collared northern woodland caribou and a number of variables that were hypothesized to
influence their distribution. I then used a count model to relate the number of wolf locations
within a habitat selection unit (HSU) to covariates that represented environmental or
industrial features that might explain the seasonal distribution o f wolves. Count models

106

contained two parts; similar to RSFs, the binary portion of the count model represented the
probability of occurrence o f wolves, while the count portion represented the relative
frequency of use in areas occupied by wolves (Nielsen et al. 2005, Sawyer et al. 2006).
Therefore, this technique had greater power, relative to the RSFs for caribou, to describe the
differential use o f resources by wolves (Nielsen et al. 2005).
Habitat selection for both the Bearhole/Redwillow (BHRW) and Quintette herds of
caribou were best described using a combination o f forest cover, forest age, solar insolation,
and distance to and density of disturbance features for all four seasons (spring, calving,
summer/fall, and winter). BHRW caribou selected for mature and late-successional forests
dominated by black spruce, tamarack and subalpine fir; these caribou expanded their use of
the winter range to include pine forests o f late succession. Caribou in both herds were at
relatively low risk of increased predation by wolves during winter (Table 13). Black spruce
and tamarack forests were rarely selected by either of the two overlapping packs of wolves.
Furthermore, the Chain Lakes and Onion Creek packs avoided habitats classified as high
quality for BHRW caribou throughout the winter months. Quintette caribou consistently
selected alpine, subalpine, spruce and pine-leading habitats year round; wolves demonstrated
avoidance behaviours of habitats selected by caribou in the Quintette herd during all three
seasons (non-winter, early winter and late winter).
The frequency of occurrence of wolves across their seasonal range was best described
using a combination of forest class, forest age, caribou habitat, and cumulative effects from
anthropogenic disturbances.

107

Table 13. Hypothetical risk o f wolves encountering caribou across the South Peace region o f northeastern British Columbia. Level o f
risk (low, low-moderate, moderate or high) is based on the results from the resource selection functions (RSFs) for caribou, and count and
movement models for wolves that quantified the distribution and movement ecology o f GPS-collared animals.

108

Habitat/Disturbance Type

Supporting
Risk of Encounter Analysis

Alpine

Low

RSF, Count,
Movement

Wolves generally avoided and increased linear movements across alpine
habitats.

Conifer

Moderate

RSF, Movement

Wolves selected for early serai forests.

Pine

Moderate

RSF, Count

Wolves selected for early serai forests.

Deciduous

Low

RSF

Caribou avoided deciduous habitats.

Mixed-species

Low

RSF

Caribou avoided mixed-species habitats.

Black spruce/
tamarack/peatland

Low

Count, Movement

Wolves avoided these lowland habitats.

Water

Low

RSF, Count

Caribou avoided habitats near lakes, rivers or creeks.

Caribou habitat - RSF
values for BHRW

Low-Moderate

RSF, Count,
Movement

Wolves selected early serai, subalpine fir, pine, and conifer forests
(seasonal). Wolves avoided black spruce/tamarack/peatland in addition to
BHRW RSF habitats during winter.

Caribou habitat - RSF
values for Quintette

Low-Moderate

RSF, Count,
Movement

Wolves avoided Quintette RSF habitats throughout the year. Wolves
increased sinuous movements in subalpine fir, pine and conifer habitats.

Comments

Table 13. Continued.
Habitat/Disturbance Type

Supporting
Risk of Encounter Analysis

Roads

Low

RSF, Count,
Movement

Both caribou and wolves avoided roads. However, decreased movement
rates suggest habitats near roads can have some encounter risk to caribou.

Seismic lines

Low - Moderate

RSF, Count,
Movement

BHRW selected habitats near seismic lines during calving and winter.
Wolves avoided and were infrequently located near these habitats throughout
the year. However, slight decreases in movement rates suggest moderate
encounter risk to caribou during winter.

Pipelines

Low - Moderate

RSF, Count,
Movement

BHRW selected habitats near pipelines during calving and winter. Wolves
avoided and were infrequently located near these features throughout the
year. However, slight decreases in movement rates suggest moderate
encounter risk to caribou during winter.

Cutblocks

High (NW),
Moderate (W)

RSF, Count,
Movement

BHRW selected habitats near cutblocks during spring and calving. Wolves
reduced use of cutblocks during winter.

Mine/Oil/Gas features
(MOG)

Low - Moderate

RSF, Count,
Movement

Quintette selected habitats near MOG features during spring and winter.
Caribou and wolf co-occurrence is most likely during the late-winter months.

High densities of linear
features

Low - Moderate

RSF, Count,
Movement

BHRW selected habitats near linear features during calving and summer/fall
and are therefore, at moderate encounter risk as Chain Lakes wolves were
observed infrequently selecting these habitats during the non-winter months.
Wolves generally increased movement rates across these habitats.

Comments

Similar to caribou, wolves residing in mountainous regions selected for pine-dominated
forests throughout the year, but unlike caribou, wolves frequented habitats of early
succession. Broadleaf forests, mixed-species forests, water, and shrub habitats were
important indicators o f wolf occurrence, but not of caribou occurrence. Two packs of wolves
residing in mountainous portions o f the study area demonstrated selection o f habitats used by
boreal caribou. However, these wolves were often found distant to the known range o f the
BHRW caribou during winter (Table 13; Appendix A).
During all four seasons, caribou demonstrated a strong avoidance o f linear features
that can serve as travel corridors or habitat for predators and other ungulate species. Also,
linear features were the most likely places for human activity, possibly displacing caribou
from adjacent habitats. Roads were avoided most strongly during the winter months by both
herds and were the only features consistently avoided by caribou each season (up to distances
between 3.5 and 11 km). Similarly, high densities of linear features influenced the seasonal
distribution of caribou. Winter is the busiest season for activities related to petroleum and
forestry exploration and development and could explain the strong avoidance of roads,
seismic lines, and pipelines.
Caribou in the BHRW herd did select habitats within close proximity to and with
increased densities of linear features (seismic lines and pipelines) during calving,
summer/fall and winter. Selection of habitats near linear sites suggests that caribou can
tolerate some levels o f disturbance. Alternatively, the high level o f industrial activity across
the range of the BHRW may offer few intact habitats distant from linear features. Caribou
may also show long-term fidelity to habitats that are degraded, but now act as an ecological
sink relative to disturbance or occurrence of predators (Schlaepher et al. 2002, Faille et al.

no

2010). Similar to caribou, wolves in all four packs demonstrated avoidance of linear features
during each of the three seasons. Furthermore, I modeled a low frequency of wolf
occurrence in habitats with high densities of roads, seismic lines and/or pipelines. Avoidance
behaviours are likely due to the direct and indirect risks associated with exposure to areas
used by humans (i.e., mortalities from increased access for hunting, trapping, and vehicular
collisions) and my results suggest that cumulative densities of these features across the
majority of the study area high (Mech et al. 1988, Fuller 1989, Mladenoff et al. 1995,
Callaghan 2002, Person 2008).
Non-linear disturbances across the landscape also influenced the distribution and
occurrence o f wolves and caribou (Table 13). Caribou in the boreal forest responded
differently to cutblocks than did caribou in the mountains. BHRW caribou selected habitats
near individual cutblocks during spring, calving, and winter, but were more sensitive to
increased densities of cutblocks during the summer/fall and winter seasons. Quintette
caribou also avoided habitats with higher densities of cutblocks during winter, but selected
these habitats most strongly during calving. These results indicate that calving sites occur
adjacent to early-successional forests and that selection o f particular habitat characteristics
may take precedence over the risks associated with spending increased amounts o f time near
early serai forests (Faille et al. 2010). Also, both herds of caribou may be demonstrating
some seasonal tolerance to regenerating cutblocks; there can be a time lag o f 20 years
between the initial cut and the regeneration of high-quality habitats for moose that can result
in the eventual extirpation o f caribou from those areas (Nielsen et al. 2005, Vors et al. 2007).
Similar to caribou, wolves in the boreal and mountainous portions of the study area occurred
in habitats closer to, and with higher densities o f cutblocks (and roads) during the non-winter

ill

season (Table 13). Wolves presumably selected those habitats for increased hunting
opportunities of moose and deer (Laurian et al. 2008, Hebblewhite et al. 2009) or because
they were suitable for denning or homesites.
Caribou were located more distant than random from features associated with the
development of oil and gas deposits during calving, summer/fall, and winter, but
demonstrated the greatest avoidance during calving (BHRW) and summer/fall. Wolves in
the Chain Lakes pack did not frequent areas with relatively high densities o f oil or gas
features during the non-winter season. For wolves in the boreal forest, there was no strong
pattern o f selection or avoidance of these features. Caribou of the Quintette herd avoided
coal mines up to a distance o f 5 km during calving and throughout the summer and fall
months, but selected habitats near mines during spring and winter. Elongated ridges in the
alpine are of high value to Quintette caribou during winter, thus, selection o f areas near
mines may be explained by the use of these important winter habitats. Two packs of wolves
that occupied territories adjacent to mine sites were infrequently located in habitats close to
mine footprints.
My field investigations and statistical results suggest that co-occurrence between
caribou and wolves is rare (Tables 12, 13), but due to the small size and isolation of caribou
herds, any amount of adult or neonate mortality from predation could have severe impacts on
herd stability and recruitment (Wittmer et al. 2005, Courbin et al. 2009). Wolves residing in
mountainous and boreal habitats appear to be supported by other prey species (i.e., moose,
deer, elk, beaver, small mammals and birds; Figure 16; Bergerud et al. 1984, James et al.
2004, Gustine et al. 2006b, DeCesare et al. 2010, Gillingham et al. 2010, Milakovic and
Parker 2011, Steenweg 2011). Similar to McCutchen (2007) and Latham et al. (201 la) in

112

Alberta, my results suggest that encounters between caribou and wolves resulting from
increased use of disturbance features by wolves is less significant to population declines than
the potential number and variety of alternate prey to support high densities o f multiple
predators (McCutchen 2007, Latham 201 la, b). In summary, my results from the analysis of
caribou and wolf distribution revealed that:
•

During winter, caribou are at relatively low risk of encountering wolves (Table
13). Caribou selected black spruce, tamarack, alpine, subalpine and pine-leading
habitats of late succession. Wolves also selected subalpine and pine-leading
habitats, but of early succession. Wolves avoided high-quality habitats for
Quintette caribou throughout the year. Caribou are likely at greatest risk of
encountering wolves in forests dominated by subalpine species, spruce, and pine.

•

Linear features, as well as habitats with high densities o f linear features, were
avoided by both caribou and wolves across all seasons. Wolves also
demonstrated low frequencies of occurrence where densities of linear features
were high. BHRW caribou did select areas where seismic lines and pipelines
occurred during calving, summer/fall, and winter.

•

Cutblocks influenced the distribution of both caribou and wolves. Both species
seasonally selected habitats close to cutblocks, as well as habitats with higher
densities of cutblocks. However, during summer/fall and winter, BHRW caribou
avoided habitats with increased densities o f cutblocks. Similarly, Quintette
caribou avoided habitats with a high density o f cutblocks during winter.

•

Non-linear features associated with mine/oil/gas development were generally
avoided by both caribou and wolves. Caribou avoided these features most during

the calving and summer/fall season (coal mines, Quintette caribou only).
Alternatively, Quintette caribou were found within close proximity to mine
features during the spring and winter months. Wolves were infrequently located
in habitats near coal mines or where densities o f non-linear features were high.
In Chapter 3 , 1 quantified seasonal variation in wolf movement. I examined 1) how
human changes to the landscape affected the speed at which wolves moved and 2) the
sinuosity o f movement paths in the context of the inferred distribution o f caribou (Chapter 2).
For each season, the rate and sinuosity of wolf movements were best explained using the full
suite of habitat and human disturbance variables. This result was consistent across daily and
weekly periods, although the weekly period demonstrated better model fit.
Alpine habitats did not affect travel rates in winter, but resulted in more linear
movements for wolves. W olf travel was more sinuous in conifer and mixed-species forests
during non-winter, but linear through conifer forests during winter and deciduous habitats
during both seasons at the scale o f weekly movements. On the contrary, daily movement
paths were more sinuous throughout conifer habitats during winter. Water features did not
facilitate linear travel as weekly movement paths were sinuous. At the daily scale,
movement rates decreased near lakes, rivers, or creeks and suggested that habitats near water
features provided wolves with increased hunting opportunities.
The occurrence of habitats I assessed as important to caribou did not influence the
movement rates o f collared wolves during the non-winter season. Spatial separation between
BHRW caribou and wolves may occur in the boreal forest as wolves were observed
travelling more rapidly in habitats classified as black spruce, tamarack, or other peatland-type
complexes and where the presence o f other prey may be minimal (e.g., moose; James et al.

114

2004, Chapter 2). However, spatial separation between wolves and caribou may occur at
finer scales than analyzed here (i.e., patch scale) and will likely vary between boreal and
mountainous habitats. In contrast to caribou in the BHRW herd, the sinuosity o f movement
paths for wolves increased in habitats used by Quintette caribou.
Industrial disturbances influenced movement behaviours of wolves throughout the
year (Table 12). Paralleling the distribution patterns o f wolves (Chapter 2), non-linear
features affected movement parameters more than linear features did at both the fine and
coarse scale. Daily movement rates decreased near forestry cutblocks, coal mines, and oil
and gas facilities, but increased where those features were relatively dense across the study
area. In addition to decreases in daily travel rates, movement was sinuous and suggested
wolves spent time hunting and searching near these habitats. As densities o f non-linear
features increased across the study area, wolves avoided these areas associated with human
presence. In summary, my results from the analysis of wolf movement revealed that:

•

In general, movement rates of wolves were higher during the non-winter months.
However, seasonal variation in movement was greater than variation in the use or
proximity to linear and non-linear features, suggesting that other factors also
influenced the movement dynamics of wolves (Figure 19).

•

Habitat and disturbance features better explained wolf movements during the
weekly as compared to the daily temporal scale.

•

Linear movements generally increased during winter and paralleled past studies
that suggested linear travel was associated with the maintenance of territories.

115

•

Wolves decreased movement rates, but not sinuosity within close proximity to
disturbance features, thus implying behaviours near such features were more
closely associated with searching and hunting.

•

Wolves increased movement rates and linear travel through areas with higher
densities of linear and non-linear industrial features; this response suggested that
wolves avoided spending time in high-risk areas associated with human activities.

Due to the complex set of interacting habitat variables, range of prey types and
variety of activities associated with resource exploration and extraction, I was unable to
detect obvious correlations between wolf movement and increased opportunities to encounter
caribou (Table 13). However, patterns of wolf movement and distribution (Chapter 2)
indicated that caribou may be most vulnerable to wolf encounters when in close proximity to
cutblocks. Future studies o f the cumulative effects o f development on the distribution of
wolf and caribou populations should include interactions associated with the ecology of
moose, deer, elk and other predators including bears, wolverines and cougars. In addition, it
is unclear how caribou behaviour might be influenced by short- and long-range wolf
movements as well as wolf presence across overlapping habitats. Quantifying current and
future levels of direct and indirect habitat loss resulting from industrial developments would
also provide additional support to managers focusing on the long-term conservation of
woodland caribou.
Activities associated with forestry, oil, natural gas, and mineral exploration and
development have resulted in dramatic transformations of the South Peace region and
continue to threaten the ecological integrity o f these landscapes (Nitschke 2008). Reductions
in the quantity and quality o f contiguous habitats can result in compounding instabilities for

116

populations of caribou: a reduction in the availability of habitat, altered predator-prey
dynamics and increased movement rates that can lead to reductions in body mass and
reproductive success (Bradshaw et al. 1997, Nellemann and Cameron 1998, Cameron et al.
2005, Faille et al. 2010). Due to the complex interactions between the cumulative effects of
disturbance and the distribution of caribou, I may not have captured all the dynamics (e.g.,
ecological sinks, time lags, etc.) responsible for influencing selection or avoidance
behaviours. In a region where wolf territories overlap caribou range, I was unable to
corroborate (i.e., through the investigation o f kill sites, count or movement models) that
wolves select, or frequently use habitats of high value to caribou. However, it remains
unclear how distributions of caribou respond to variations in wolf movement or the increased
presence o f wolves across portions o f their home range. Furthermore, I did not assess vital
rates or population change across caribou herds, the ultimate measures o f cumulative
impacts. Recent (2008) population inventory data has shown, however, that the BHRW herd
is in decline while the Quintette population of caribou is increasing (Seip and Jones 2011).
Quantifying the distribution o f caribou and the frequency of habitat use and
movement by wolves increased our understanding of predator-prey dynamics across a
changing landscape. My study, based on habitat selection, movement ecology, and
behaviours linked to predation, indicates that disturbance effects from anthropogenic
developments occur at multiple scales (i.e., patch scale and valley scale) for both caribou and
wolves. My results indicate there is relatively little spatial overlap among the two species
with this overlap being greatest in the boreal forest, where wolves have increased
opportunities to adjust behaviours to increase their use of high-quality habitat for caribou.
Caribou inhabiting the low-elevation boreal habitats may be demonstrating a maladaptive

117

strategy in the context o f multiple disturbance regimes on the landscape. Specifically,
encounters between caribou and wolves are most likely to occur in areas closer to and with
higher densities of cutblocks, as both species were observed selecting these features during
the non-winter season. As my results suggest, however, interactions among predators,
caribou and land-use development are not easily predicted or temporally static. As the
density and types of industrial disturbances increase across the boreal forest, predators and
other ungulates will become more widespread and predation risk for caribou will increase
with the reduction of available refugia.

Literature Cited
Anderson, D.R., K.P. Burnham and W.L. Thompson. 2000. Null hypothesis testing:
problems, prevalence, and an alternative. Journal of Wildlife Management 64:912-923.
Arthur, S.M., F.J. Manly, L.L. McDonald and G.W. Gamer. 1996. Assessing habitat
selection when availability changes. Ecology 77:215-227.
Baldi, R., A. Pelliza-Sbriller, D. Elston and S. Albon. 2004. High potential for competition
between guanacos and sheep in Patagonia. Journal of Wildlife Management 68:924938.
Ballard, W.B., L.A. Ayres, C.L. Gardner and J.W. Foster. 1991. Den site activity patterns of
gray wolves, Canis Lupus, in south-central Alaska. Canadian Field Naturalist 105:497504.
Bergerud, A.T. 1974. Decline o f caribou in North America following settlement. Journal of
Wildlife Management 38:757-770.
Bergerud, A.T. 1978. The status and management of caribou in British Columbia. Report to
the Minister of Recreation and Conservation. BC Ministry of Recreation and
Conservation. Victoria, BC, Canada.
Bergerud, A.T., H.E. Butler and D.R. Miller. 1984. Antipredator tactics o f calving caribou:
dispersion in mountains. Canadian Journal of Zoology 62:1566-1575.
Bergerud, A.T., and J.P. Elliot. 1986. Dynamics o f caribou and wolves in northern British
Columbia. Canadian Journal of Zoology 64:1515-1529.
Bibikov, D.I., A.N. Kudaktin and A.N. Filimonov. 1985. Territoriality, movements. Pages
415-431 in The wolf: history, systematics, morphology, ecology. {In Russian}.
Bibikov, D.I. editor. Nauka, Moscow, Russia.
Boyce, M.S., P.R. Vernier, S.E. Nielson and F.K.A. Schmiegelow. 2002. Evaluating resource
selection functions. Ecological Modeling 157:281-300.

118

Bradshaw, C.J.A., D.M. Hebert, A.B. Rippin and S. Boutin. 1995. Winter peatland habitat
section by woodland caribou in northeastern Alberta. Canadian Journal of Zoology
73:1567-1574.
Bradshaw, C.J.A, S. Boutin and D.M. Hebert. 1997. Effects of petroleum exploration on
woodland caribou in northeastern Alberta. Journal o f Wildlife Management 61:11271133.
Brandt, J.P. 2009. The extent of the North American boreal zone. Environmental Reviews
17:101-161.
British Columbia Ministry o f Forests and Range. 2007a. Composite VRI relational data
dictionary. Draft version 1.0b. Available at:
http://www.for.gov.bc.ca/hts/vridata/standards/. Date accessed: July 17, 2009.
British Columbia Ministry o f Forests and Range. 2007b. Forest Analysis and Inventory
Branch. 2007a. Vegetation Resource Inventory. Available at:
http://www.for.gov.bc.ca/hts/vri/. Date accessed: March 10, 2009.
British Columbia Integrated Land Management Bureau. 2007. Land and resource data
warehouse. TRIM. Available at: http://www.lrdw.ca/. Date accessed: March 10,2009.
Bryant, A.A., and R.E. Page. 2005. Timing and causes of mortality in the endangered
Vancouver Island marmot (Marmota Vancouverensis). Canadian Journal of Zoology
83:674-682.
Burnham, K.P., and D.R. Anderson. 2002. Model selection and inference: a practical
information-theoretic approach. 2nd Edition. Springer-Verlag. New York, NY, USA.
Callaghan, C.J. 2002. The ecology of gray wolf ( Canis lupus) habitat use, survival, and
persistence of gray wolves in the central Rocky Mountains. Ph.D. Dissertation,
University of Guelph, Guelph, Ontario, Canada.
Cameron, R.D., W.T. Smith, R.G. White and B. Griffith. 2005. Central Arctic Caribou and
petroleum development: distributional, nutritional, and reproductive implications.
Arctic 58:1-9.
Child, K.N., S.P. Barry and D.A. Aitken. 1991. Moose mortality on highways and railways in
British Columbia. Alces 27:41-49.
Cichowski, D.B. 1993. Seasonal movements, habitat use, and winter feeding ecology of
woodland caribou in west-central British Columbia. Land Management Rep. No.79,
British Columbia Ministry of Forests, Victoria, British Columbia, Canada.
Compton, B.W., J.M. Rhymer and M. McCollough. 2002. Habitat selection by wood turtles
(Clemmys insculpta ): an application o f paired logistic regression. Ecology 83:833-843.
Cooley, H.S., H.S. Robinson, R.B. Wielgus and C.S. Lambert. 2008. Cougar prey selection
in a white-tailed deer and mule deer community. Journal of Wildlife Management
72:99-106.
Corsi, F., E. Dupre and L. Boitani. 1999. A large-scale model o f wolf distribution in Italy for
conservation planning. Conservation Biology 13:130-159.
Coupe, R., A.C. Stewart and B.M. Wikeem. 1991. Engelmann spruce - subalpine fir zone.
Pages 223-236 in The ecosystems of British Columbia. Meidinger, D.V and J. Pojar,
editors. Special Report Series No 6. BC Ministry of Forests. Victoria, BC, Canada.
Courbin, N., D. Fortin, C. Dussault and R. Courtois. 2009. Landscape management for
woodland caribou: the protection of forest blocks influences wolf-caribou co­
occurrence. Landscape Ecology 24: 1375-1388.

119

Courtois, R., J.P. Ouellet and B. Gagne. 1998. Characteristics of cutovers used by moose
(AIces alces) in early winter. Alces 34:201-211.
Culling, D.E., B.A. Culling, T.J. Raabis and A.C. Creagh. 2006. Ecology and seasonal
habitat selection of boreal caribou in the Snake-Sahtaneh watershed, British Columbia,
2000-2004. Report for Canadian Forest Products Ltd., Fort Nelson, British Columbia,
Canada.
Curatolo, J.A., and S.M. Murphy. 1986. The effects o f pipelines, roads and traffic on the
movement of caribou, Rangifer tarandus. Canadian Field Naturalist 100: 218:225.
D’Eon, R.G., R. Serrouya, G. Smith and C.O. Kochanny. 2002. GPS radiotelemetry error and
bias in mountainous terrain. Wildlife Society Bulletin 30:430-439.
D’Eon, R.G., and D. Delparte. 2005. Effects of radio-collar position and orientation on GPS
radio-collar performance, and the implications of PDOP in data screening. Journal of
Applied Ecology 42:383-388.
DeCesare, N.J., M. Hebblewhite, H.S. Robinson and M. Musiani. 2010. Endangered,
apparently: The role of apparent competition in endangered species conservation.
Animal Conservation 13:353-362.
DeLong, C., R.M. Annas and A.C. Stewart. 1991. Boreal white and black spruce zone. Pages
237-250 in The ecosystems o f British Columbia. Meidinger, D.V and J. Pojar, editors.
Special Report Series No 6. BC Ministry o f Forests. Victoria, BC, Canada.
Duchesne, T., D. Fortin and N. Courbin. 2010. Mixed conditional logistic regression for
habitat selection studies. Journal of Animal Ecology 79:548-555.
Dyer, S.J., J.P. O’Neill, S.M. Wasel and S. Boutin. 2001. Avoidance of industrial
development by woodland caribou. Journal of Wildlife Management 65:531-542.
Dyer, S.J., J.P. O'Neill, S.M. Wasel and S. Boutin. 2002. Quantifying barrier effects of roads
and seismic lines on movements of female woodland caribou in northeastern Alberta.
Canadian Journal o f Zoology 80:839-845.
Dzus, E. 2001. Status of the woodland caribou {Rangifer tarandus caribou) in Alberta.
Alberta Environment, Fisheries and Wildlife Management Division, and Alberta
Conservation Association, Wildlife Status Report No. 30, Edmonton, Canada.
Energy Resources Conservation Board (ECRB). 2011. Annual Catalog: production, maps,
and services. Energy resources conservation board, Calgary, Alberta, Canada.
Edmonds, E.J. 1991. The status of woodland caribou in western North America. Rangifer
7:91-107.
Faille, G., C. Dussault, J-P. Ouellet, D. Fortin, R. Courtois, M-H. St-Laurent and C. Dussault.
2010. Range fidelity: the missing link between caribou decline and habitat alteration?
Biological Conservation 143:2840-2850.
Ferguson, S.H., W.J. Rettie and F. Messier. 1998. Fractal measures o f female caribou
movements. Rangifer Special Issue No. 10:139-147.
Festa-Bianchet, M., T. Coulson, J-M. Gaillard, J.T. Hogg and F. Pelletier. 2006. Stochastic
predation events and population persistence in bighorn sheep. Proceedings of the Royal
Society, Biology 263:1537-1543.
Ficetola, G.F., and M. Denoel. 2009. Ecological thresholds: an assessment o f methods to
identify abrupt changes in species-habitat relationships. Ecography 32:1075-1084.
Festa-Bianchet, M., J.C. Ray, S. Boutin, S.D. Cote and A. Gunn. 2011. Conservation of
caribou {Rangifer tarandus) in Canada: an uncertain future. Canadian Journal of
Zoology 89:419-434.
120

Fortin, D., H.L. Byer, M.S. Boyce, D.W. Smith, T. Duchesne and J.S. Mao. 2005. Wolves
influence elk movements: behaviour shapes a trophic cascade in Yellowstone National
Park. Ecology 86:1320-1330.
Fraser, D., and E.R. Thomas. 1982. Moose-vehicle accidents in Ontario: relation to highway
salt. Wildlife Society Bulletin 10:261-265.
Fritts, S.H., and L.D. Mech. 1981. Dynamics, movements and feeding ecology of a newly
protected wolf population in northwestern Minnesota. Wildlife Monographs 80:1-79.
Fuller, T.K. 1989. Denning behaviour of wolves in north-central Minnesota. American Field
Naturalist 121:184-188.
Fuller, T.K., and L.B. Keith. 1981. Woodland caribou population dynamics in northeastern
Alberta. Journal of Wildlife Management 45:197-213.
Gibson, L. 2006. The role of lethal control in managing the effects of apparent competition
on endangered prey species. Wildlife Society Bulletin 34:1220-1224.
Gillies, C.S., M. Hebblewhite, S.E. Nielsen, M.A. Krawchuk, C.L. Aldridge, J.L. Frair, D.J.
Saher, S.E. Stevens and C.L. Jerde. 2006. Application o f random effects to the study of
resource selection by animals. Journal o f Animal Ecology 75:887-898.
Gillingham, M.P., R.W. Steenweg and D.C. Heard. 2010. Parsnip caribou recovery trial report on activities during 2009 - 2010. Peace/Williston Fish and Wildlife
Compensation Program Report No. 339.
Gunson, J.R. 1995. Wolves: Their characteristics, history, prey relationships and
management in Alberta. Alberta Environmental Protection, Alberta, Canada.
Gustine, D.D., K.L. Parker, R.J. Lay, M.P.Gillingham and D.C. Heard. 2006a. Calf survival
of woodland caribou in a multi-predator ecosystem. Wildlife Monographs 165:1-32.
Gustine, D.D., K.L. Parker, R.J. Lay, M.P.Gillingham and D.C. Heard. 2006b. Interpreting
resource selection at different scales for woodland caribou in winter. Journal of
Wildlife Management 70:1601-1614.
Hatter, I., D. Butler, A. Fontana, D. Hebert, T. Kinley and S. McNay. 2002. A strategy for
the recovery of mountain caribou in British Columbia. British Columbia Ministry of
Water, Land, and Air Protection. Victoria, British Columbia. Available at
http://www.env.gov.bc.ca/wld/documents/mtcaribou_rcvrystrat02.pdf
Heard, D.C., and K.L. Vagt. 1998. Caribou in British Columbia: a 1996 status report.
Rangifer Special Issue No. 10:159-172.
Hebblewhite, M., and E. Merrill. 2008. Modelling wildlife-human relationships for social
species with mixed-effects resource selection models. Journal o f Applied Ecology
45:834-844.
Hebblewhite, M., R.H. Munro and E.H. Merrill. 2009. Trophic consequences o f postfire
logging in a wolf-ungulate system. Forest Ecology Management 257:1053-1062.
Hebblewhite, M. 2011. Effects of energy development on ungulates. Pages 71-94 in Energy
development and wildlife conservation in western North America. D.E. Naugle editor.
Island press, Washington, USA.
Hins, C., J.-P. Ouellet, C. Dussault and M.-H. St. Laurent. 2009. Habitat selection by forestdwelling caribou in managed boreal forest of eastern Canada: evidence o f a landscape
configuration effect. Forest Ecology and Management 257:636-643.
Holt, R.D. 1977. Predation, apparent competition and the structure o f prey communities.
Theoretical Population Biology 12:197-229.

121

Hosmer, D.W., and S. Lemeshow. 2000. Applied logistic regression. Second edition. John
Wiley and Sons, New York, New York, USA.
Houle, M., D. Fortin, C. Dussault, R. Courtois and J.-P. Ouellet. 2010. Cumulative effects of
forestry on habitat use by gray wolf (Canis lupus) in the boreal forest. Landscape
Ecology 25:419-433.
Hummel, M., and J.C. Ray (Editors). 2008. Caribou and the North: a shared future. Dundum
Press Limited, Toronto, Ontario, Canada.
Jalkotzy, M.G., P.I. Ross and M.D. Nasserden. 1997. The effects of linear developments on
wild- life: a review of selected scientific literature. Prepared for the Canadian
Association of Petroleum Producers, Arc Wildlife Services, Calgary, Alberta, Canada.
James, A.R.C. 1999. Effects of industrial development on the predator-prey relationship
between wolves and caribou in northeastern Alberta. Dissertation. University of
Alberta, Edmonton, Alberta, Canada.
James, A.R.C., and A.K. Stuart-Smith. 2000. Distribution of caribou and wolves in relation
to linear corridors. Journal of Wildlife Management 64:154-159.
James, A.R.C., S. Boutin, D.M. Hebert and A.B. Rippin. 2004. Spatial separation o f caribou
from moose and its relation to predation by wolves. Journal o f Wildlife Management
68:799-809.
J?drzejewski, W., K. Schmidt, J. Theuerkauf, B. Jedrzejewska and H. Okarma. 2001. Daily
movements and territory use by radio-collared wolves {Canis lupus) in Bialowieza
Primeval Forest in Poland. Canadian Journal o f Zoology 79:1993-2004.
Johnson, C.J., K.L. Parker, D.C. Heard and M.P. Gillingham. 2002a. Movement parameters
of ungulates and scale-specific responses to the environment. Journal of Animal
Ecology, 71:225-235.
Johnson, C.J., K.L. Parker, D.C. Heard and M.P. Gillingham. 2002b. A multiscale
behavioural approach to understanding the movements o f woodland caribou. Ecological
Applications 12:1840-1860.
Johnson, C.J., K.L. Parker, D.C. Heard and D.R. Seip. 2004a. Movements, foraging habits,
and habitat use strategies of northern woodland caribou during winter: Implications for
forest practices in British Columbia. BC Journal of Ecosystems and Management 5:2235.
Johnson, C.J., D.R. Seip and M.S. Boyce. 2004b. A quantitative approach to conservation
planning: using resource selection functions to map the distribution of mountain
caribou at multiple spatial scales. Journal of Applied Ecology 41:238-251.
Johnson, C.J., M.S. Boyce, R.L. Case, H.D. ClufF, A.Gunn and R. Mulders. 2005.
Cumulative effects of human developments on Arctic wildlife. Wildlife Monographs
160:1-36.
Johnson, C.J., and M.-H. St-Laurent. 2011. Unifying framework for understanding impacts
of human developments on wildlife. Pages 27-54 in Energy development and wildlife
conservation in western North America. Naugle, D.E., editor.
Jones, E.S., M.P. Gillingham, D.R. Seip and D.C. Heard. 2007. Comparison o f seasonal
habitat selection between threatened woodland caribou ecotypes in central British
Columbia. Rangifer Special Issue No. 17:111-128.
Kinley, T.A., and C.D. Apps. 2001. Mortality patterns in a subpopulation of endangered
mountain caribou. Wildlife Society Bulletin 29:158-164.

122

Krausman, P.R., and L.K. Harris. 2011. Cumulative effects in wildlife management: impact
mitigation. CRC Press, Taylor and Francis Group, Boca Raton, FL, USA.
Kreeger, T.J. 2003. The internal wolf: physiology, pathology, and pharmacology. Pages 192217 in Mech, L.D and L. Boitani, editors. Wolves: behaviour, ecology, and
conservation. The University of Chicago Press, Chicago, II, USA.
Kristan, W.B. Ill, and W.I. Boarman. 2003. Spatial pattern of risk o f common raven
predation on desert tortoises. Ecology 84:2432-2443.
Kunkel K.E., and D.H. Pletscher. 2001. Winter hunting patterns of wolves in and near
Glacier National Park, Montana. Journal o f Wildlife Management 65:520-530.
Kuzyk, G.W., J. Kneteman and F.K.A. Schmiegelow. 2004. Winter habitat use by wolves,
Canis lupus, in relation to forest harvesting in west-central Alberta. Canadian FieldNaturalist 118:368-375.
Latham, A.D.M. 2009. W olf ecology and caribou-primary prey-wolf spatial relationships in
low productivity peatland complexes in northeastern Alberta. PhD Dissertation.
University o f Alberta, Edmonton, Alberta, Canada.
Latham, A.D.M, M.C. Latham, N.A. McCutchen and S. Boutin. 201 la. Invading white-tailed
deer change wolf-caribou dynamics in northeastern Alberta. Journal of Wildlife
Management 75:204-212.
Latham, A.D.M, M.C. Latham and M.S. Boyce. 201 lb. Habitat selection and spatial
relationships of black bears (Ursus americanus) with woodland caribou (Rangifer
tarandus caribou) in northeastern Alberta. Canadian Journal of Zoology 89:267-277.
Latham, A.D.M, M.C. Latham, M.S. Boyce and S. Boutin. 201 lc. Movement responses by
wolves to industrial linear features and its effect on woodland caribou in northeastern
Alberta. Ecological Applications 21: 2854 - 2865.
Laurian, C., C. Dussault, J.-P. Ouellet, R. Courtois, M. Poulin and L. Breton. 2008.
Behaviour of moose relative to a road network. Journal of Wildlife Management
72:1550-1557.
Leblond, M., J. Frair, D. Fortin, C. Dussault, J-P. Ouellet and R. Courtois. 2011. Assessing
the influence o f resource covariates at multiple spatial scales: an application to forestdwelling caribou faced with intensive human activity. Landscape Ecology 26: 14331446
Lesmerises, F., C. Dussault and M-H. St-Laurent. 2012. W olf habitat selection is shaped by
human activities in a highly managed boreal forest. Forest Ecology and Management
276:125-131.
Lima, S.L., and L.M. Dill 1990. Behavioural decisions made under the risk of predation: a
review and prospectus. Canadian Journal of Zoology 68:619-640.
Long, J.S., and J. Freese. 2006. Regression models for categorically dependant variables
using Stata. 2nd Ed. College Station, TX: Stata Press.
Manly, B.F.J., L.L. McDonald, D.L. Thomas, T.L. McDonald and E.P. Erickson. 2002.
Resource selection by animals: statistical design and analysis for field studies. Kluwer
Academic Publishers, Dordrecht, Netherlands.
McCutchen, N.A. 2007. Factors affecting caribou survival in northern Alberta: the role o f
wolves, moose, and linear features. Ph.D Dissertation. University of Alberta,
Edmonton, Alberta, Canada.
McKenzie, H.W. 2006. Linear features impact predator-prey encounters: analysis with first
passage time. MSc. Thesis. University o f Alberta, Edmonton, Alberta, Canada.
123

McLellan, B.N., R. Serrouya, H.U. Wittmer and S. Boutin. 2010. Predator-mediated Allee
effects in multi-prey systems. Ecology 91:286-292.
McLoughlin, P.D., E. Dzus, B. Wynes and S. Boutin. 2003. Declines in populations of
woodland caribou. Journal of Wildlife Management 67:755-761.
Mech, L.D. 1970. The wolf: the ecology and behaviour of an endangered species. Natural
History Press, Garden City, NY, USA.
Mech, L.D. 1974. Canis lupus. Mammalian Species 37:1-6.
Mech, L.D. 1994. Regular and homeward travel speeds o f wolves. Journal o f Mammalogy
74:741-742.
Mech, L.D., S.H. Fritts, G.L. Radde and W.J. Paul. 1988. W olf distribution and road density
in Minnesota. Wildlife Society Bulletin 16:85-87.
Mech, L.D., L.G. Adams, T.J. Meier, J.W. Burch and B.W. Dale. 1998. The wolves of
Denali. University o f Minnesota Press, Minneapolis, MN, USA.
Mech, L.D., and L. Boitani. 2003. Wolves: behaviour, ecology, and conservation. The
University o f Chicago Press, Chicago, II, USA.
Meidinger, D.V., and J. Pojar. 1991. The ecosystems of British Columbia. Special Report
Series No 6. BC Ministry of Forests. Victoria, British Columbia, Canada.
Meidinger, D.V., J. Pojar and W.L. Harper. 1991. Sub-Boreal Spruce zone. Pages 210-221
in The ecosystems of British Columbia. Meidinger, D.V and J. Pojar, editors. Special
Report Series No 6. BC Ministry of Forests. Victoria, British Columbia, Canada.
Menard, S. 2002. Applied logistic regression analysis. Sage University Paper Series on
Quantitative Applications in the Social Sciences, 07-106. Thousand Oaks, California.
USA.
Merrill, S.B. 2000. Road densities and gray wolf, Canis lupus, habitat suitability: an
exception. Canadian Field-Naturalist 114:312-313.
Milakovic, B., and K.L. Parker. 2011. Using stable isotopes to define diets of wolves in
northern British Columbia, Canada. Journal of Mammalogy 92:295-304.
Milakovic, B., K.L. Parker, D.D. Gustine, R.J. Lay, A.B.D. Walker and M.P. Gillingham.
2011. Habitat selection by a focal predator (Canis lupus) in a multi-prey ecosystem of
the northern Rockies. Journal of Mammalogy 92:568-582.
Mladenoff, D.J., T.A. Sickley, R.G. Haight and A.P. Wydeven. 1995. A regional landscape
analysis and prediction o f favourable gray wolf habitat in the northern Great Lakes
region. Conservation Biology 9:279-294.
Mladenoff, D.J., T.A. Sickley and A.P. Wydeven. 1999. Predicting gray wolf landscape
recolonization: logistic regression models vs. new field data. Ecological Applications
9:37-44.
Moen, R., J. Pastor and Y. Cohen. 1997. Accuracy of GPS telemetry collar locations with
differential correction. Journal of Wildlife Management 61:530-539.
Morales J.M. and S.P. Ellner. 2002. Scaling up movement in heterogeneous landscapes: the
importance of behavior. Ecology 83:2240-2247.
Morales, J.M., P.R. Moorcrofit, J. Matthiopoulos, J.L. Frair, J.G. Kie, R.A. Powell, E.H.
Merrill and D.T. Haydon. 2010. Building the bridge between animal movement and
population dynamics. Philosophical Transactions o f the Royal Society B 365:22892301.

124

Nams, V.O., and M. Bourgeois. 2004. Fractal analysis measures habitat use at different
spatial scales: an example with American marten. Canadian Journal of Zoology
82:1738-1747.
Nellemann, C., and R.D. Cameron. 1998. Cumulative impacts o f an evolving oil-field
complex on the distribution of calving caribou. Canadian Journal of Zoology 76:14251430.
Nellemann, C., l.Vistnes, P. Jordhoy and O. Strand. 2001. Winter distribution o f wild
reindeer in relation to power lines, roads and resorts. Biological Conservation 101:351360.
Neufeld, L. 2006. Spatial dynamics o f wolves and caribou in an industrial forest landscape in
west-central Alberta. MSc Thesis. University of Alberta, Edmonton, Alberta, Canada.
Nielsen, S.E., C.J. Johnson, D.C. Heard and M.S. Boyce. 2005. Can models o f presenceabsence be used to scale abundance? Two case studies considering extremes in life
history. Ecography 28:197-208.
Nielsen, S.E., J. Cranston and G.B. Stenhouse. 2009. Identification o f priority areas for
grizzly bear conservation and recovery in Alberta, Canada. Journal o f Conservation
Planning 5: 38-60.
Nitschke, C.R. 2008. The cumulative effects o f resource development on biodiversity and
ecological integrity in the Peace-Moberly region o f Northeast British Columbia,
Canada. Biodiversity Conservation 17:1715-1740.
Nowak, R.M. 1983. A perspective on the taxonomy of wolves in North America. Pages 1019 in L.N. Carbyn, editor. Wolves in Canada and Alaska: Their status, biology, and
management (Report Series 45). Canadian Wildlife Service, Ottawa, Canada.
Packard, J.M. 2003. W olf behaviour: reproductive, social, and intelligent. Pages 35-65 in
L.D. Mech and L. Boitani, editors. Wolves: behaviour, ecology and conservation. The
University o f Chicago Press, Chicago, II, USA.
Paquet, P.C., and L.N. Carbyn. 2003. Gray wolf: Canis lupus and allies. Pages 482-510 in
G.A. Feldhamer, B.C. Thompson, and J.A. Chapman, eds. Wild mammals of North
America: Biology, management, and conservation. John Hopkins University Press,
Baltimore, Maryland, USA.
Person, D.K., and A.L. Russell. 2008. Correlates of mortality in an exploited wolf
population. Journal of Wildlife Management 72:1540-1549.
Peters, R.P., and L.D. Mech. 1975. Scent-marking in wolves: A field study. American
Scientist 63:628-637.
Pielou, E.C. 1969. An introduction to mathematical ecology. Wiley-Interscience, New York,
New York, USA.
Pojar, J., and A.C. Stewart. 1991. Alpine tundra zone. Pages 263-274 in The ecosystems of
British Columbia. Meidinger, D.V and J. Pojar, editors. Special Report Series No 6.
BC Ministry of Forests. Victoria, British Columbia, Canada.
Polfus, J.L., M. Hebblewhite, and K. Heinemeyer. 2011. Identifying indirect habitat loss and
avoidance o f human infrastructure by northern mountain woodland caribou. Biological
Conservation 144:2637 - 2646.
Proulx, R. 2010. Economic development outlook and business opportunity summary. Report
for the District of Tumbler Ridge. Tumbler Ridge, British Columbia, Canada.
Rea, R.V. 2003. Modifying roadside vegetation management practices to reduce vehicular
collisions with moose Alces alces. Wildlife Biology 9:81-91.

125

Rempel, R.S., P.C. Elkie, A.R. Rodgers and M.J. Gluck. 1997. Timber-management and
natural-disturbance effects on moose habitat: landscape evaluation. Journal o f Wildlife
Management 61:517-524.
Rettie, W.J., and F. Messier. 2000. Hierarchical habitat selection by woodland caribou; its
relationship to limiting factors. Ecography 23:466-478.
Rettie, W.J., and F. Messier. 2001. Range use and movement rates of woodland caribou in
Saskatchewan. Canadian Journal of Zoology 79:1933-1940.
Rinaldi, T.A. 2010. Influence of linear features and snowmachine activity on resource
selection by wolves. MSc Thesis. University o f Northern British Columbia, Prince
George, British Columbia, Canada.
Robinson, H.S., R.B. Wielgus and J.C. Gwilliam. 2002. Cougar predation and population
growth of sympatric mule deer and white-tailed deer. Canadian Journal of Zoology
80:556-568.
Saher, D.J., and F.K.A. Schmiegelow. 2004. Movement pathways and habitat selection by
woodland caribou during spring migration. Rangifer Special Issue No. 16:143-154.
Sawyer, H., R.M. Nielson, F. Lindzey and L.L. McDonald. 2006. Winter habitat selection of
mule deer before and during development of a natural gas field. Journal of Wildlife
Management 70:396-403.
Schaefer, J.A. 2003. Long-term range recession and the persistence o f caribou in the taiga.
Conservation Biology 17:1435-1439.
Schlaepfer, M.A., M.C. Runge and P.W. Sherman. 2002. Ecological and evolutionary traps.
Trends in Ecology and Evolution 17:474-480.
Schneider, R.R., J.B. Stelfox, S. Boutin and S. Wasel. 2003. Managing the cumulative
impacts of land uses in the Western Canadian Sedimentary Basin: a modeling approach.
Conservation Ecology 7:8.
Seip, D.R. 1991. Predation and caribou populations. Rangifer Special Issue No. 7:46-52.
Seip, D.R. 1992. Factors limiting woodland caribou populations and their interrelationships
with wolves and moose in southeastern British Columbia. Canadian Journal of Zoology
70:1494-1503.
Seip, D.R., and D.B. Cichowski. 1996. Population ecology o f caribou in British Columbia.
Rangifer Special Issue No. 9:73-80.
Seip, D.R., and E.L. Jones. 2011. Population status of threatened caribou herds in the central
Rockies ecoregion of British Columbia. Unpublished report.
Sih, A. 1984. The behavioural response race between predator and prey. American Naturalist
123:143-150.
Sopuck, L.G. 1985. Movements and distribution of caribou in relation to the Quintette coal
development. Report prepared for Denison Mines Ltd.
Spalding, D.J. 1990. The early history of moose (.Alces alces): distribution and relative
abundance in British Columbia. Contributions to Natural Science Number 11. Royal
British Columbia Museum, Victoria, British Columbia, Canada.
St-Laurent M-H., C. Dussault, J. Ferron and R. Gagnon. 2009. Dissecting habitat loss and
fragmentation effects following logging in boreal forest: conservation perspectives
from landscape simulations. Biological Conservation 142:2240 - 2249.
StataCorp. 2005. Stata Statistical Software: Release 9. College Station, TX: StataCorp LP.
Steenweg, R.W. 2011. Interactions of wolves, mountain caribou and an increased moosehunting quota - primary-prey management as an approach to caribou recovery. MSc

126

Thesis. University of Northern British Columbia, Prince George, British Columbia,
Canada.
Stevenson, S.K., and D.F. Hatler. 1985. Woodland caribou and their habitat in southern and
central British Columbia. Volume 1. Land Management Report No. 23. BC Ministry of
Forests. Victoria, British Columbia, Canada.
Stuart-Smith, K.A., C.J.A. Bradshaw, S. Boutin, D.M. Hebert and A.B. Rippin. 1997.
Woodland caribou relative to landscape patterns in northeastern Alberta. Journal of
Wildlife Management 61:622-633.
Teny, E.L., and M.D. Wood. 1999. Seasonal movements and habitat selection by woodland
caribou in the Wolverine herd, north central B.C. Phase 2: 1994-1997. Rep. No. 204,
Peace/Williston Fish and Wildlife Compensation Program, Prince George, British
Columbia, Canada.
Thurber, J.M., R.O. Peterson, T.D. Drummer and S. A. Thomasma. 1994. Gray wolf response
to refuge boundaries and roads in Alaska. Wildlife Society Bulletin 22:61-68.
Tremblay-Gendron, S. 2012. Influence des proies sur le deplacement d'un predateur: etude
du systeme loup-orignal-caribou (Influence o f prey on predator's movements: a study
on the wolf-caribou-moose system). MSc Thesis. Universite du Quebec a Rimouski,
Rimouski, Quebec, Canada.
Turchin, P. 1998. Quantitative analysis of movement: measuring and modeling population
redistribution in animals and plants. Sinauer Associates Inc., Sunderland,
Massachusets,U SA.
Vistnes, I., and C. Nellemann. 2001. Avoidance o f cabins, roads and power lines by reindeer
during calving. Journal o f Wildlife Management 65:915-925.
Vors, L.S., J.A. Schaeffer, B.A. Pond, A.R. Rodgers and B.R. Patterson. 2007. Woodland
caribou extirpation and anthropogenic landscape disturbance in Ontario. Journal of
Wildlife Management 71:1249-1256.
Vors, L.S., and M.S. Boyce. 2009. Global declines of caribou and reindeer. Global Change
Biology 15:2626-2633.
Vuong, Q. 1989. Likelihood ratio tests for model selection and non-nested hypotheses.
Econometrica 57:307-334.
Walton, L.R., H.D. Cluff, P.C. Paquet and M.A. Ramsay. 2001. Movement patterns of
barren-ground wolves in the central Canadian arctic. Journal o f Mammology 82:867876.
Wamback, A., and T. Hilding-Rydevik. 2009. Cumulative effects in Swedish EIA practice:
Difficulties and obstacles. Environmental Impact Assessment Review 29:107-115.
Weclaw, P., and R.J. Hudson. 2004. Simulation of conservation and management of
woodland caribou. Ecological Modelling 177:75-94.
Whittington, J., C.C. St. Clair and G. Mercer. 2005. Spatial responses of wolves to roads and
trails in mountain valleys. Ecological Applications 15:543-553.
Wittmer, H.U. 2004. Mechanisms underlying the decline o f mountain caribou {Rangifer
tarandus caribou) in British Columbia. Ph.D Dissertation. University of British
Columbia, Vancouver, British Columbia, Canada.
Wittmer, H.U., A.R.E. Sinclair and B.N. McLellan. 2005. The role o f predation in the decline
and extirpation of woodland caribou. Oecologia 144:257-267.

127

Wittmer, H.U., B.N. McLellan and F.W. Hovey. 2006. Factors influencing variation in site
fidelity of woodland caribou (Rangifer tarandus caribou) in south eastern British
Columbia. Canadian Journal of Zoology 84:537-545.
Wittmer, H.U., B.N. McLellan, R. Serrouya and C.D. Apps. 2007. Changes in landscape
composition influence the decline of a threatened woodland caribou population. Journal
of Animal Ecology 76:568-579.
Young, S.P. and E.A. Goldman. 1944. The wolves of North America. American Wildlife
Institute, Washington, D.C.USA.

128

r

v.

-

/ — ' S . ■W tfH L rf* *

"V' L /fk - O
>v

BHRW 2009

a

BHRW 2008

Chain Lakes 2009

♦ Chain Lakes 2008

Onion Creek 2009

* Onion Creek 2008

Upper Murray 2009

■ Upper Murray 2008

BHRW 2007

a

Contours
Rivers
□

0

25

5

10

15

Lakes
N
20

Appendix A. Figure 1. Distribution of Bearhole/Redwillow caribou (BHRW; 2007 - 2009)
and three packs o f wolves (Chain lakes, Onion Creek, and Upper Murray; 2008 - 2009)
during the spring season (April 1 - May 14) across the South Peace region of northeastern
British Columbia.
129

a

BHRW 2009

a

BHRW 2008

*

Chain Lakes 2009

♦

Chain Lakes 2008

*

Onion Creek 2009

* Onion Creek 2008

*

Upper Murray 2009

■

a

BHRW 2007

Contours
Rivers
[ ~ ] Lakes

Upper Murray 2008
0

3

8

12
Kilometer*

18

24

,

m

Appendix A. Figure 2. Distribution of Bearhole/Redwillow caribou (BHRW; 2007 - 2009)
and three packs of wolves (Chain lakes, Onion Creek, and Upper Murray; 2008 - 2009)
during the calving season (May 15 - June 14) across the South Peace region o f northeastern
British Columbia.
130

a

BHRW 2009

*

Chain Lakes 2009

▲ BHRW 2008
♦ Chain Lakes 2008

*

Onion Creek 2009

*

Onion Creek 2008

■

Upper Murray 2009

■

Upper Murray 2008

*

BHRW 2007

Contours
Rivers
1

0 25 5

10

1 Lakes
15

2Q

k

Kilometer*

Appendix A. Figure 3. Distribution of Bearhole/Redwillow caribou (BHRW; 2007 - 2009)
and three packs of wolves (Chain lakes, Onion Creek, and Upper Murray; 2008 - 2009)
during the summer/fall season (June 15 - October 31) across the South Peace region of
northeastern British Columbia.
131

1

ft

a

BHRW 2009

a

♦ Chain Lakes 2009
* Onion Creek 2009
■ Upper Murray 2009

BHRW 2008

a

♦ Chain Lakes 2008
* Onion Creek 2008
■ Upper Murray 2008

_ _ _ _ _

BHRW 2006 & 2007

Contours

Rivers
Lakes
N

0

3.5

7

14

21

20

A

Kilometer*

Appendix A. Figure 4. Distribution of Bearhole/Redwillow caribou (BHRW; 2007 - 2009)
and three packs of wolves (Chain lakes, Onion Creek, and Upper Murray; 2008 - 2009)
during the winter season (November 1 - Mar 31) across the South Peace region of
northeastern British Columbia.
132

#

Tumb

a

Quintette 2009

a

Quintette 2008

*

Upper Sukunka 2009

•

Upper Sukunka 2008

*

Upper Murray 2009

■

Upper Murray 2008

FT""] Lake*

*

Onion Creek 2009

*

Onion Creek 2008

N

a

Quintette 2003 - 2007

Contour*
River*

0

25

5

10
Kiiormtors

15

20

A

Appendix A. Figure 5. Distribution of Quintette caribou (2003 - 2009) and three packs of
wolves (Upper Sukunka, Upper Murray and Onion Creek; 2008 - 2009) during the spring
season (April 1 - May 14) across the South Peace region o f northeastern British Columbia.

133

a

*
*
*

Quintette 2009
Upper Sukunka 2009
Upper Murray 2009
Onion Creek 2009

Quintette 2008
• Upper Sukunka 2008
■ Upper Murray 2008
* Onion Creek 2008

a

a

Quintette 2003 - 2007

C ontours

Rivers

(7 7 1 Lakes
N
Kriorrwtent

Appendix A. Figure 6. Distribution of Quintette caribou (2003 - 2009) and three packs of
wolves (Upper Sukunka, Upper Murray and Onion Creek; 2008 - 2009) during the calving
season (May 15 - June 14) across the South Peace region o f northeastern British Columbia.

134

JVTumtfleKv

\

2
Quintette 2009

Quintette 2008

Upper Sukunka 2009

Upper Sukunka 2008

Rivers

Upper Murray 2008

Lakes

Upper Murray 2009
Onion Creek 2009

*

Quintette 2003 - 2007

Contours

Onion Creek 2008
0

25

5

Appendix A. Figure 7. Distribution o f Quintette caribou (2003 - 2009) and three packs of
wolves (Upper Sukunka, Upper Murray and Onion Creek; 2008 - 2009) during the
summer/fall season (June 15 - October 31) across the South Peace region of northeastern
British Columbia.
135

&
V.
*

Quintette 2009

*

Quintette 2008

*

Upper Sukunka 2009

•

Upper Sukunka 2008

*

Upper Murray 2009

■

Upper Murray 2008

*

Onion Creek 2009

*

Onion Creek 2008

*

Quintette 2003 - 2007

C ontours

Rivers

I

~~] Lakes
N

0 2.5 5

10

15

20

Ktk>motors

Appendix A. Figure 8. Distribution of Quintette caribou (2003 - 2009) and three packs of
wolves (Upper Sukunka, Upper Murray and Onion Creek; 2008 - 2009) during the winter
season (November 1 - March 31) across the South Peace region of northeastern British
Columbia.
136

Appendix B. Table 1. GPS collar fix and success rate (based on 24-hour period) for woodland caribou in the South Peace region o f
northeastern British Columbia.

Caribou

Dates of Collar

Total
Days

Fix
Rate

Attempted
Fixes

Acquired
Fixes

Fix
Success
Rate

Rec
Altitude

2D
% 2D
Fixes'5 Fixes'5

3D
% 3D
Fixes'5 Fixes'5

Quintette

ATS
car012A3

12/11/2003

4/17/2005

493

1

575

474

0.8241

Yes

33

7.16

428

92.84

car0313

12/11/2003

4/17/2005

493

1

575

550

0.9562

Yes

48

8.92

490

91.08

car0403

4/4/2005

10/25/2006

569

1 ,2

569,1138

670

1.18, 0.59

Yes

66

10.28

576

89.72

car0413

4/4/2005

1/30/2006

303

1 ,2

303, 606

303

1, 0.50

Yes

61

21.63

221

78.37

car0423

4/4/2005

7/20/2006

472

1 ,2

472, 944

561

1.19, 0.59

Yes

22

3.99

529

96.01

car045

12/22/2005

4/4/2007

468

5

2340

1600

0.6838

Yes

44

2.77

1545

97.23

car046

12/22/2005

12/3/2006

346

3

1038

462

0.4451

Yes

56

12.33

398

87.67

car057

1/30/2007

1/16/2009

717

4

2868

2830

0.9868

Yes

157

5.61

2644

94.39

car059

1/30/2007

10/31/2008

640

4

2560

1989

0.777

Yes

706

37.57

1173

62.43

car068

4/4/2007

3/14/2009

710

4

2840

2530

0.8908

Yes

449

18.91

1925

81.09

car069

4/4/2007

1/3/2009

640

4

2560

2093

0.8176

Yes

673

34.32

1288

65.68

car070

4/4/2007

3/7/2009

703

4

2812

2688

0.9559

Yes

252

9.63

2365

90.37

car071

4/5/2007

1/2/2009

638

4

2552

2243

0.8789

Yes

253

11.72

1906

88.28

car072

4 /5/2007

5/29/2007

54

4

216

141

0.6528

Yes

56

45.53

67

54.47

car074

4 /5/2007

2/25/2009

692

4

2768

2442

0.8822

Yes

725

31.12

1605

68.88

car075

4/5/2007

3/14/2009

4

2836

2630

0.9274

Yes

328

12.81

2233

87.19

carl05

1/10/2009

7/31/2009

70S
202

4

808

680

0.8416

Yes

27

4.01

646

95.99

c a rll2

1/24/2009

12/21/2009

331

4

1324

662

0.5

Yes

20

5.05

376

94.95

3Quintette collars initially fixed at 20hr (1-2/day) intervals, and then at 12 hours intervals.
bpremised on data already screened for erroneous locations.

Appendix B. Table 1. Continued.
Fix
Caribou

Dates of Collar

Total

Fix

Attempted

Acquired

Success

Rec

2D

% 2D

3D

% 3D

Days

Rate

Fixes

Fixes

Rate

Altitude

Fixes'5

Fixes'5

Fixes6

Fixes'5

303

3

909

734

0.8075

Yes

68

9.26

666

90.74

Bearhole

Start

End

car054

12/20/2006

10/19/2007

car055

12/20/2006

8/3/2007

226

3

678

592

0.8732

Yes

24

4.05

568

95.95

car077

12/18/2007

12/17/2009

730

3

2190

1652

0.7543

Yes

352

21.32

1299

78.68

car078

1/8/2008

5/14/2009

492

3

1476

1130

0.7656

Yes

158

14.21

954

85.79

F900

12/21/2006

1/26/2009

767

6

4602

4325

0.9398

Yes

791

18.57

3468

81.43

F901

1/30/2007

1/26/2009

727

6

4362

4008

0.9188

Yes

665

16.79

3296

83.21

car012

2/12/2003

4/7/2003

54

6

324

321

0.9907

No

104

32.4

217

67.6

car013

2/12/2003

3/19/2003

35

6

210

98

0.4667

No

30

30.61

68

69.39

car014

2/12/2003

7/26/2003

164

6

984

886

0.9004

No

257

29.01

629

70.99

car015

2/12/2003

3/16/2003

32

6

192

174

0.9063

No

21

12.07

153

87.93

Redwillow

Quintette

TV

aQuintette collars initially fixed at 20hr (1-2/day) intervals, and then at 12 hours intervals.
bpremised on data already screened for erroneous locations.

Appendix B. Table 2. Category classes of GPS collar locations developed to determine
erroneous locations for caribou. All but 2.99% of 3D and 16.4% of 2D fixes were used in
statistical analyses.

GPS
Location
Category
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27

2D
Filter
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1

PDOP
Filter
N/A
1
1
1
2
2
2
3
3
3
4
4
5
5
4
4
4
5
5
5
3
3
2
2
1
1
1
2

Elevation
Filter
N/A
1
2
3
1
2
3
1
2
3
1
2
1
2
3
4
5
3
4
5
4
5
4
5
4
5
0
0

139

Frequency
40,506
4,152
1,302
629
185
50
46
121
38
28
83
22
96
34
9
5
21
24
12
37
19
30
23
48
371
894
775
7

% of
all
Locs
81.72
8.38
2.63
1.27
0.37
0.1
0.09
0.24
0.08
0.06
0.17
0.04
0.19
0.07
0.02
0.01
0.04
0.05
0.02
0.07
0.04
0.06
0.05
0.1
0.75
1.8
1.56
0.01

Loc Class
Removed?
(Y/N)
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
N
N
N
N
N
N
N
N
N
N
N
Y
N

Appendix B. Table 3. GPS collar fix and success rate (based on 24-hour period) for wolves in the South Peace region of northeastern British
Columbia. Due to the improved precision of GPS collars, wolf locations were checked using GIS and dates of collar deployment. Twodimensional fixes (2D) accounted for 0.04 - 17.19% of all downloaded locations. Recorded altitude was used to ensure elevation was recorded
from the GPS collar. The Chain Lakes pack had a number of locations that occurred beyond the BC/Alberta boarder and were thus discarded from
analyses.

Total
Days

Fix Rate
(per 24h)

Attempted
Fixes

Acquired
Fixes

Fix
Success
Rate

Recorded
Altitude

3D
Fixes

%
3D
Fixes

2D
Fixes

%
2D
Fixes

Pack
Lower Sukunka

Dates of Collar
Start
End

W008aai
W008b

1/7/08
2/1/08

1/31/08
1/14/09

25
348

1
8

25
2760

24
2384

0.96
0.86

1
1

23
2070

99.96
99.86

1
290

0.04
0.14

WO10a6
WO10b
W027
W029

1/22/08
2/1/08
2/17/09
3/14/09

1/31/08
10/27/08
4/11/10
3/20/10

9
269
418
371

1
8
8
8

9
2152
3344
2968

7
1712
2754
2310

0.78
0.80
0.82
0.78

1
1
1
1

5
1485
2356
1906

0.60
0.85
85.55
82.51

2
229
398
404

0.40
0.15
14.45
17.49

Upper Murray
W015
W019
W022
W023

2/3/08
2/4/08
2/10/09
2/10/09

1/14/09
3/8/08
2/5/10
9/27/09

346
33
360
229

8
8
8
8

2768
264
2880
1832

2384
232
2325
1537

0.86
0.88
0.81
0.84

1
1
1
1

2073
207
1999
1346

86.95
89.22
85.98
87.57

311
25
326
191

13.05
10.78
14.02
12.43

Upper Sukunka

a These data were excluded from analyses as they occurred beyond the range o f monitored caribou
6 GPS collar locations separated into two classes (a,b) due to a programming changes in their GPS collars

Appendix B. Table 3. Continued.

Total
Days

Fix Rate
(per 24h)

Attempted
Fixes

Acquired
Fixes

Fix
Success
Rate

Recorded
Altitude

3D
Fixes

% 3D
Fixes

2D
Fixes

%
2D
Fixes

286
54
354

8
8
8

2288
432
2832

1950
411
2263

0.85
0.95
0.80

1
1
1

1697
375
1874

87.03
91.24
82.81

253
36
389

12.97
8.76
17.19

Chain Lakes
W016
3496
2/3/08 4/15/09
437
8
3173
0.91
1
2824 91.30 269.00
W024
2/10/09 11/3/09
8
2128
1614
1
266
0.76
1269 87.22 186.00
W028
3/14/09 2/5/10
328
8
2624
1
1414 86.11 228.00
2521
0.96
12/3/09 3/19/10
W030
106
72
7632
7462
0.98
1
3609 95.88 155.00
7992
W031
2/11/10 6/2/10
111
72
7442
0.93
1
4516 95.50 213.00
3Analyses were based on wolf locations occurring in British Columbia only; locations for Chain Lakes that occurred in Alberta
were discarded.

8.70
12.78
13.89
4.12
4.50

Pack
Dates of Collar
Onion Creek
Start
End
W020
2/5/08 11/17/08
2/16/09 4/11/09
W025
W026
2/16/09 2/5/10

BC
Total3

3093
1455
1642
3764
4729

Appendix C. Table 1. Kill sites identified using clusters of GPS locations and site
investigations for wolves in the South Peace region o f northeastern British Columbia. Field
investigations were completed over three years (5/2008 - 9/2010); a total of 73 kill sites were
used to analyze wolf area of use (AOU).
Kill ID
1

Kill#
08-001

Species
Moose

Age Class
Adult

Sex
Unknown

Pack
Chain Lakes

Wolf
W016

Date
2/9/2008

2

08-002

Moose

Adult

Unknown

Chain Lakes

W016

2/18/2008

3

08-003

Unknown

Adult

Unknown

Chain Lakes

W016

4/20/2008

4

08-004

Elk

Unknown

Unknown

Onion Creek

W020

4/15/2008

5

08-005

Moose

Adult

Unknown

Upper Murray

W019

2/20/2008

6

08-006

Moose

Adult

Unknown

Upper Murray

W015

4/22/2008

7

08-007

Moose

Adult

Unknown

Upper Sukunka

W010

3/16/2008

8

08-008

Moose

Unknown

Unknown

Upper Murray

W015

5/8/2008

9

08-009

Moose

Adult

Unknown

Upper Murray

W015

5/17/2008

10

08-010

Moose

Adult

Female

Onion Creek

W020

7/31/2008

11

08-011

Moose

Adult

Female

Chain Lakes

W016

8/22/2008

12

08-012

Moose

Adult

Unknown

Chain Lakes

W016

8/13/2008

13

09-001

Deer

Adult

Unknown

Onion Creek

W025

2/23/2009

14

09-002

Elk

Adult

Unknown

Onion Creek

W025

3/21/2009

15

09-003

Moose

Unknown

Unknown

Onion Creek

W026

2/24/2009

16

09-004

Deer

Adult

Unknown

Onion Creek

W026

3/12/2009

17

09-005

Deer

Adult

Unknown

Onion Creek

W025

3/15/2009

18

09-006

Deer

Adult

Unknown

Onion Creek

W025

4/6/2009

19

09-007

Deer

Adult

Unknown

Onion Creek

W026

4/6/2009

20

09-008

Moose

Adult

Unknown

Onion Creek

W026

2/22/2009

21

09-009

Unknown

Unknown

Unknown

Upper Sukunka

W027

9/23/2009

22

09-010

Moose

Adult

Unknown

Onion Creek

W025

3/12/2009

23

09-011

Elk

Adult

Male

Onion Creek

W026

3/19/2009

24

09-012

Moose

Unknown

Unknown

Upper Sukunka

W029

3/24/2009

25

09-013

Unknown

Adult

Unknown

Upper Sukunka

W027

3/22/2009

142

Appendix C. Table 1. Continued.
Kill ID

Kill#

Species

Age Class

Sex

Pack

Wolf

Date

26

09-014

Deer

Adult

Unknown

Upper Sukunka

W029

4/12/2009

27

09-015

Deer

Adult

Unknown

Upper Sukunka

W029

4/12/2009

28

09-016

Unknown

Unknown

Unknown

Upper Sukunka

W027

4/6/2009

29

09-017

Moose

Adult

Unknown

Chain Lakes

W016

2/1/2009

30

09-018

Moose

Adult

Unknown

Chain Lakes

W028

3/15/2009

31

09-019

Moose

Calf

Unknown

Chain Lakes

W028

3/16/2009

32

09-020

Moose

Adult

Unknown

Chain Lakes

W016

2/3/2009

33

09-021

Moose

Adult

Unknown

Upper Murray

W022

5/21/2009

34

09-022

Moose

Adult

Unknown

Onion Creek

W026

6/3/2009

35

09-023

Unknown

Calf

Unknown

Onion Creek

W026

6/8/2009

36

09-024

Muskrat

Adult

Unknown

Upper Murray

W022

5/27/2009

37

09-025

Moose

Adult

Unknown

Upper Murray

W022

5/27/2009

38

09-026

Mt. Goat

Adult

Unknown

Upper Sukunka

W029

5/6/2009

39

09-027

Moose

Calf

Unknown

Onion Creek

W026

6/5/2009

40

09-028

Moose

Adult

Female

Onion Creek

W026

6/5/2009

41

09-029

Caribou

Adult

Unknown

Upper Murray

W022

5/19/2009

42

09-030

Moose

Adult

Unknown

Onion Creek

W026

5/27/2009

43

09-031

Moose

Yearling

Unknown

Upper Sukunka

W027

3/19/2009

44

09-032

Moose

Adult

Unknown

Upper Sukunka

W027

6/3/2009

45

09-033

Moose

Adult

Unknown

Chain Lakes

W024

5/13/2009

46

09-034

Moose

Adult

Male

Chain Lakes

W028

5/10/2009

47

09-035

Moose

Yearling

Unknown

Chain Lakes

W028

4/18/2009

48

09-036

Moose

Adult

Female

Chain Lakes

W028

3/29/2009

49

09-037

Moose

Adult

Female

Chain Lakes

W024

3/22/2009

50

09-038

Moose

Adult

Female

Chain Lakes

W028

7/11/2009

143

Appendix C. Table 1. Continued.
Kill ID

Kill#

Species

Age Class

Sex

Pack

Wolf

Date

51

09-039

Unknown

Calf

Unknown

Upper Murray

W023

6/30/2009

52

09-040

Unknown

Unknown

Unknown

Upper Murray

W023

7/5/2009

53

09-041

Mt. Goat

Unknown

Unknown

Upper Murray

W023

7/17/2009

54

09-042

Unknown

Unknown

Unknown

Upper Murray

W022

6/15/2009

55

09-043

Unknown

Unknown

Unknown

Upper Murray

W022

6/30/2009

56

09-044

Moose

Adult

Unknown

Upper Sukunka

W029

8/19/2009

57

09-045

Moose

Adult

Unknown

Upper Sukunka

W029

8/11/2009

58

09-046

Moose

Adult

Female

Chain Lakes

W024

8/9/2009

59

09-047

Moose

Adult

Unknown

Chain Lakes

W024

8/1/2009

60

09-048

Moose

Yearling

Unknown

Upper Sukunka

W027

9/25/2009

61

09-049

Moose

Unknown

Unknown

Upper Murray

W023

8/28/2009

62

09-050

Moose

Adult

Male

Upper Murray

W023

9/22/2009

63

10-001

Deer

Unknown

Unknown

Upper Sukunka

W029

3/15/2010

64

10-002

Deer

Adult

Unknown

Upper Sukunka

W029

2/26/2010

65

10-003

Deer

Adult

Unknown

Upper Sukunka

W029

2/26/2010

66

10-004

Deer

Adult

Male

Upper Sukunka

W029

3/4/2010

67

10-005

Deer

Unknown

Unknown

Upper Sukunka

W029

2/19/2010

68

10-006

Moose

Adult

Female

Chain Lakes

W031

3/17/2010

69

10-007

Deer

Adult

Unknown

Chain Lakes

W028

1/4/2010

70

10-008

Moose

Adult

Male

Chain Lakes

W028

12/31/2009

71

10-009

Deer

Adult

Unknown

Chain Lakes

W028

2/8/2010

72

10-010

Moose

Calf

Unknown

Upper Sukunka

W027

3/13/2010

73

10-011

Moose

Adult

Unknown

Upper Sukunka

W027

2/13/2010

144

Appendix C. Table 2. Area o f use (AOU) calculated using 100% MCPs that surrounded kill
sites for each wolf across the South Peace region o f northeastern British Columbia. Kill sites
were identified through the investigation of location clusters over three summers (2008 2010 ).

AOU (ha)

# Pts
in Cluster

W olf Pack

86.5

93.02

34

Chain Lakes

1.00

99.2

99.62

26

Chain Lakes

804663

0.80

64.7

80.47

18

Chain Lakes

Male

3500262

3.50

1225.2

350.03

5

Chain Lakes

W031a

Female

3500262

3.50

1225.2

350.03

5

Chain Lakes

W020

Male

64809

0.06

0.4

6.48

23

Onion Creek

W025

Male

95208

0.12

1.4

11.90

17

Onion Creek

W026

Male

69866

0.09

0.8

8.89

14

Onion Creek

W015

Male

4672523

4.67

2183.2

467.25

15

Upper Murray

W019

Female

205363

0.21

4.2

20.54

23

Upper Murray

W022

Female

276753

0.33

11.0

33.21

12.2

Upper Murray

W023

Female

1007388.8

1.0

101.5

100.7

14

Upper Murray

W010

Female

62194

0.06

0.4

6.22

22

Upper Sukunka

W027

Male

95526.4

0.1

0.9

9.6

13.9

Upper Sukunka

W olf ID

Sex

AOU
(m2)

AOU
AOU
(km2) (per 100 Ha)

W016

Male

930161

0.93

W024

Male

996167

W028

Female

W030°

W029
Female
0.2
4.1
14
41226.0
0.04
Upper Sukunka
a W030 and W031 were fitted with high-frequency collars programed to collect locations every 20 mins;
data were therefore, excluded from AOU calculations to maintain consistancy across all packs.

145

■AOU size (ha)

■ Avg. # of wolf locations per kill site
155.4

Q 100

is 40
9.1 18
Chain Lakes

Onion Creek

■

18

6.6 _ 17
_

™ 17
7.9,

Upper Murray Upper Sukunka Lower Sukunka

Pack

Appendix C. Figure 1. Average area o f use (AOU; ha) by pack identified through the
investigation o f wolf kill sites over three summers (2008 - 2010) across the South Peace
region of northeastern British Columbia.

146

1.5

T r e e B le a f

1

YG
W ater*

0.5

Road

Old
O th e r

C ID IK
T

1 —^ —

0

u_
X

T
1

C_BHRW

$ -0.5

Pine

e
*

-1

FOR
F
01

LF

j —♦

SeisPipIn

M atu re

OG
MOG

Blk S pruce

u -1.5
CO.

-2

-2.5
W a te r

-3

Appendix D. Figure 1. Coefficients from count model describing frequency of occurrence of wolf
locations within habitat selection units (HSUs; n = 3,389) relative to environmental and disturbance
covariates during the non-winter season (April 16 - Oct 14) from wolves collared in the Chain Lakes
pack. An asterick (*) indicates a Guassian (squared) term was used in the model.

7
W a te r

6

5
4

u

Old

X
3
IA

91
*13 2*
cOJ „
‘3
1
£
2 0

MOG

Blk S pruce

Road

Pine

i

4

co.

OG

C BHRW

-1

O th e r

-2

LF
SeisPipIn

T ree_B leaf
Ctblk

M atu re

-3

W ater*

-4

Appendix D. Figure 2. Coefficients from binary model describing the presense or absense of wolves
within HSUs relative to environmental and disturbance covariates during the non-winter season from
wolves collared in the Chain Lakes pack.
147

2
1.5

O th er

1

T ree_B leaf
W ate r*

YG
0.5 -|

u
*
S

0

-1

H ‘1

Blk S pruce

Pine

ii_i
t

„

g -0.5

0
1
.9

S pruce

t

M a tu re

i

ra m

0 ld

Road

t

4 ... i

Ctblk

OG*

*

LF
FOR

SeisPipIn

MOG

OG

-C BHRW

-1.5

-2
W a te r
-2.5

Appendix D. Figure 3. Coefficients from count model describing frequency of occurrence of wolf
locations within HSUs during the early winter season (Oct 15 - Jan 31) from wolves collared in the
Chain Lakes pack.

W a te r

0 2
*
0i/1i
■H

1

2c
01
I 0

1u

ca

MOG
O th e r

Pine
T

YG

\

Blk S pruce

t

T ree Bleaf

OG

Road

} u in
T

M a tu re

Ctblk

4T T

LF

-

i

Old

SeisPipIn

FOR

C BHRW

W ate r*

Appendix D. Figure 4. Coefficients from count model describing frequency of occurrence of wolf
locations within HSUs during the late winter season (Feb 1 - April 15) from wolves collared in the
Chain Lakes pack.
148

12
W a te r

10
8
6
4
2

Blk S pruce

Road

Old

Pine

YG

Ctblk

FOR

0
2

O th e r

7 ree Bleaf

M atu re

OG

SeisPipIn

C_BHRW

MOG

■4
6

8

W ate r*

10
A ppendix D. Figure 5. C oefficients from binary m odel describing the presense o r absense o f w olves
w ithin H SU s during the late w inter season from w olves collared in the C hain Lakes pack.

W ater*

4.5

2.5

u
£

m

O th e r

Spruce

0.5

*
Pine

•8 -1-5

Alpine

YG

C_BHRW

M a tu re

T ree O th e r
T ree Bleaf

♦
C_Q

Old

SeisPipIn
#

♦

♦

Road

M ine

Fo r

♦

LF

♦

-♦

MOG

-3.5

-5.5
W a te r

-7.5

Appendix D. Figure 6. Coefficients from count model describing frequency of occurrence of wolf
locations within HSUs (n = 10,493) during the non-winter season from wolves collared in the Onion
Creek pack.

149

Water

u
X
1f \
Ot
■H

M ■

A'P 'n e

C

4>
IV o

P'ne

I *Mi-H
T ree O th e r

t

S pruce

3

SeisPipIn

M a tu re

C_Q

+

Old

YG

♦-

i

.

‘
T
Road

4

FOR

M ine

LF
MOG

C- BHRW

O th er

T ree Bleaf

W ater*

Appendix D. Figure 7. Coefficients from binary model describing the presense or absense of wolves
within HSUs during the non-winter season from wolves collared in the Onion Creek pack.

W a te r’

Pine
Alpine

Road

T ree Bleaf

SeisPipIn

Old
C BHRW, Road*

M ine

S pruce
O th er

FOR
M a tu re

M ine

CO.

T ree O th e r
SeisPipIn

W a te r

Appendix D. Figure 8. Coefficients from count model describing frequency of occurrence of wolf
locations within HSUs during the early winter season from wolves collared in the Onion Creek pack.

150

6
W a te r

5
4
3

u
*ts» 2
O*

Pine
Road

44

tft 1
C
01
1 0
8
u
ca -1
-2

O th e r T S pruce

YG

SeisPipIn’

Old

C BHRW

:1 H}-*-A
Tree_Bleaf.

M ine*
Road*

M a tu re

i

LF

4 4-4
-

FOR

M ine

T ree O th e r

Alpine

-3

SeisPipIn
W ater*

-4

Appendix D. Figure 9. Coefficients from binary model describing the presense or absense of wolves
within HSUs during the early winter season from wolves collared in the Onion Creek pack.

W ater'

lil

O th e r

Pine
CO.

SeisPipIn

S pruce

Alpine

YG

T ree Bleaf

M atu re

C BHRW

C_Q

Road

FOR
FOR

Old

W a te r

Appendix D. Figure 10. Coefficients from count model describing frequency of occurrence of wolf
locations within HSUs during the late winter season from wolves collared in the Onion Creek pack.

151

W a te r

o
*in

an
•h 2

Alpine
O th e r

SeisPipIn

sp ru c e
M atu re

0
E

C_Q

Road

-1 .

3

i

M ine
Ctblk

♦

C BHRW
YG

LF*

FOR*
FOR

Old

T ree Bleaf

LF

Pine
W ate r*

-4

A ppendix D. Figure 11. C oefficients from binary m odel describing the presense or absense o f w olves
w ithin H SU s during the late w inter season from w olves collared in the O nion C reek pack.

1.5
NoVRI
G row ing
O th e r

0.5

M atu re

S pruce

i

u
m
at
•H

C_Q
C_BHRW

♦ ♦
Road

-0.5

-1
-1.5

FOR

♦ -* £

Young
T re e_ O th e r

Road*

SeisPipIn
M ine

MOG*

LF*

Ii

FOR"

LF

Old

MOG

No Age

Pine
W a te r

-2

Appendix D. Figure 12. Coefficients from count model describing frequency of occurrence of wolf
locations within HSUs (n = 35,959) during the non-winter season from wolves collared in the Upper
Murray pack.

152

2

NoVRI

1.5

No Age

_u

Road

0.5 -

C BHRW

Spruce

*th
Ol
+1

SeisPipIn

i-

YGM

0

i.

C Q

1

Road*

0 -0.5

» ■ ...........

M ine

FOR
LF

O th e r

-1

T ree O th e r

-1.5
W a te r

-2

Appendix D . Figure 13. Coefficients from count model describing frequency of occurrence of wolf
locations within HSUs during the early winter season from wolves collared in the Upper Murray
pack.
2.5
Pine

2

No Age
Old

1.5

o
X
m
a>
c

u
•a

£

1

NoVRI
Road
C_BHRW

0.5
YG

$

C_Q

SeisPipIn
Road*

♦

0

♦
FOR

8

^ -0.5

LF

i

O th er

-1
Spruce
W a te r

-1.5
T re e_ O th er

M atu re

-2

Appendix D. Figure 14. Coefficients from count model describing frequency of occurrence of wolf
locations within HSUs during the late winter season from wolves collared in the Upper Murray pack.

153

Water
O th e r
NoVRI
Road
Old
Pine
YG

u
*in
cr*

C_BHRW

T ree O th er

I

c

i
0

No Age

SeisPipIn

C Q

Road*

LF
^

FOR

M atu re

S pruce

-3

A ppendix D. Figure 15. C oefficients from binary m odel describing the presense or absense o f w olves
w ithin H SU s during the late w inter season from w olves collared in the U pper M urray pack.

W ater*
O th er

i

Road

L

u
*LA
eh
•H -1

e
V
O
£
2 -2

Ctblk

YGM

i

....

Old

C q

Alpine

SeisPipIn

M ine
Ctblk *

M ine*

FOR

♦

♦

LF

OG

No_Age
T ree O th er

-3
W a te r

Appendix D. Figure 16. Coefficients from count model describing frequency of occurrence of wolf
locations within HSUs (n = 33,599) during the non-winter season from wolves collared in the Upper
Sukunka pack.

154

4
3

A lpine

Road
W ater*

2

YGM
No_Age

O th e r

SeisPipIn

D 1
X
in
at
■
h 0
ttt

CQ

*■»

^

c

41

OG*
x

M ine*

FOR*

SeisPipIn*
FOR

■° -1
<E 1

I

i

Ctblk*

Road*

^

M ine

LF*

LF

^

T ree_B leaf

-2

T ree O th e r

OG
Old

-3
-4

Ctblk

W a te r

-5
Appendix D. Figure 17. Coefficients from count model describing frequency of occurrence of wolf
locations within HSUs during the early winter season from wolves collared in the Upper Sukunka
pack.

W a te r
Road

A lpine

T re e_ O th e r
SeisPipIn

in

Ctblk1

YGM

FOR*

SeisPipIn

O th e r

CO.

OG
M ine

No_Age

LF*

FOR

Road*

Old

M ine

W a te r’

OG

T ree Bleaf
Ctblk

Appendix D. Figure 18. Coefficients from binary model describing the presense or absense of wolves
within HSUs during the early winter season from wolves collared in the Upper Sukunka pack.

155

M ine
Alpine

*

Road
T ree_B leaf

W ater*

YGM

X
lA

SeisPipIn*

C_Q

01
■H 0
W
i/t
e

I -1

M ine*

—♦
^

j

O th e r

T

No_Age

I

T

Road*

T

SeisPipIn

#
OG

S pruce

3

Old

CO.

-2

-3
W a te r

-4

Appendix D. Figure 19. Coefficients from count model describing frequency of occurrence of wolf
locations within HSUs during the late winter season from wolves collared in the Upper Sukunka
pack.

156

Appendix D. Table 1. The percent (%) o f total habitat used, frequently used, or available across the range o f wolves based on the
occurrence o f Habitat Selection Units (HSUs) dominated by a land cover type. Use includes Habitat Selection Units (HSUs) with > 1
wolf location, whereas Frequent Use includes HSUs with > 10 wolf locations. Model variables are described in Table 1.

Variables
U pper Sukunka
Alpine
Other
Tree Broadleaf
Tree Other
Upper M urray
No VRI
Other
Spruce
Tree Other
Onion Creek
Alpine
Other
Pine
Spruce
Tree Broadleaf
Tree Other
Chain Lakes
Alpine
Black Spruce
Subalpine Fir
HBS
No VRI
Pine
Spruce
Tamarack
Tree Broadleaf
Tree Other
Upland NonVeg
Water

EARLY W INTER
Frequent
Use
Use

Availability

32.54
32.72
14.44
20.29

28.00
52.00
12.00
8.00

35.86
20.79
2.19
41.16

48.55
12.30
15.21
23.94

67.86
14.29
7.14
10.71

18.37
32.29
18.33
31.01

8.94
13.13
22.91
24.86
18.99
11.17

5.26
10.53
21.05
52.63
10.53
0.00

9.95
12.23
28.28
22.17
3.93
23.45

0.21
17.95
0.11
0.53
1.38
43.45
7.31
4.91
21.26
2.35
0.11

0.00
14.94
0.00
0.00
0.00
31.17
9.74
6.49
34.42
2.60
0.00

0.50
18.52
1.06
1.17
3.22
47.13
9.10
3.28
15.03
0.27
0.42

Variables
Upper Sukunka
Alpine
Other
Pine
Spruce
Tree Broadleaf
Upper M urray
No VRI
Other
Pine
Spruce
Tree Other
Onion Creek
Alpine
Other
Pine
Spruce
Tree Broadleaf
Chain Lakes
Alpine
Black Spruce
Subalpine Fir
HBS
NoVRI
Pine
Spruce
Tamarack
Tree Broadleaf
Tree Other

0.43

0.65

0.30

Water

LATE W INTER
Frequent
Use
Use

Availability

30.80
33.43
9.94
12.29
13.54

25.00
50.00
3.57
14.29
7.14

35.91
29.56
12.23
20.15
2.15

15.01
13.20
28.57
31.46
11.75

8.82
5.88
32.35
44.12
8.82

18.72
32.33
12.51
18.14
18.31

5.64
20.32
33.63
21.44
18.96

8.00
24.00
16.00
36.00
16.00

10.07
35.89
27.94
22.27
3.83

0.26
20.04
0.26
1.05
0.00
40.40
7.12
4.22
23.47
3.16

0.00
14.29
0.00
0.00
0.00
37.14
11.43
0.00
31.43
5.71

0.45
18.24
0.91
1.03
3.35
47.00
8.86
3.61
15.71
0.49

0.00

0.00

0.36

Variables
Upper Sukunka
Alpine
Other
Tree Other
U pper M urray
No VRI
Other
Pine
Spruce
Tree Other
Onion Creek
Alpine
Other
Pine
Spruce
Tree Broadleaf
Tree Other
Chain Lakes
Alpine
Black Spruce
HBS
No VRI
Pine
Spruce
Tamarack
Tree Broadleaf
Tree Other
Upland NonVeg
Water

NON- WINTER
Frequent
Use
Use

Availability

26.62
41.22
32.16

17.02
59.57
23.40

36.06
22.82
41.12

15.01
13.20
28.57
31.46
11.75

8.82
5.88
32.35
44.12
8.82

18.72
32.33
12.51
18.14
18.31

8.94
13.13
22.91
24.86
18.99
11.17

5.26
10.53
21.05
52.63
10.53
0.00

9.95
12.23
28.28
22.17
3.93
23.45

0.20
15.22
0.40
1.78
39.33
7.31
7.11
25.30
2.77
0.20
0.40

0.00
14.29
0.00
0.00
23.81
12.70
9.52
36.51
1.59
0.00
1.59

0.47
18.82
1.12
2.95
47.23
8.81
3.18
15.17
1.55
0.38
0.32

Appendix E. Table 1. Number of parameters (&), Akaike’s Information Criterion values (AICc), and AICc weights (w) for seasonal
resource selection models for the Bearhole/Redwillow (BHRW) caribou herd monitored from 2006 - 2009 across the South Peace
region o f northeastern British Columbia. Sample size o f caribou locations is represented in parentheses.
Spring (n = 3,401)

AICc
8063.0
9189.7

AAIC
708.6
1835.3

<0.001
<0.001

k
10
5
1
16

k
11
5

AICc
12018.2
12339.0

AAIC
2018.4
2339.3

AIC„
<0.001
<0.001

1653.9
446.7

AICh
<0.001
<0.001
<0.001
<0.001

9008.3
7801.1

3
19

12353.9
11856.8

2354.2
1857.1

<0.001
<0.001

252.8
272.7

<0.001
<0.001

19
17

7795.9
7743.8

441.5
389.4

<0.001
<0.001

22
20

11696.1
11739.0

1696.3
1739.3

<0.001
<0.001

2099.6

248.6

<0.001

1961.7
2128.8

110.6
277.7

<0.001
<0.001

20

7740.0

1645.7

<0.001

<0.001
<0.001

23
28
20

11645.4

7448.8
7724.0

385.6
94.5
369.6

<0.001

25
17

10817.6
11849.8

817.9
1850.0

<0.001
<0.001

1963.0
2079.1
2103.9

111.9

<0.001

228.1
252.9

26
22
18

7420.4
7789.4
7723.6

66.0
435.0
369.2

<0.001

29

10819.6

819.9

<0.001

<0.001
<0.001

<0.001
<0.001

25
21

11489.5
11851.8

1489.7
1852.1

<0.001
<0.001

2065.9
1921.3

214.8

<0.001

24

7717.0

362.6

<0.001

27

11421.4

1421.7

<0.001

70.3

<0.001

28
17

7429.3
7724.0

75.0
369.6

<0.001
<0.001

31

10247.2

247.5

<0.001

Model Covariates
Forest cover
Forest age
Solar insolation6,c
Landscape

k
9
5
1
15

AICc
3550.2
3596.8

AAIC
417.7
464.3

AICm
<0.001
<0.001

k
8
5

AICc
2205.4
2319.9

AAIC
354.4
468.8

AIC„
<0.001
<0.001

3583.9
3519.2

451.4
386.7

<0.001
<0.001

1
14

2253.1
2127.1

402.1
276.1

Road Dist6
Road Dens
Road Dist*c + Road
Dens

18
16

3514.3
3519.5

381.8
387.0

<0.001
<0.001

17
15

2103.9
2123.8

382.7
77.7
386.3

<0.001
<0.001
<0.001

18
23
15

LF Distfc
LF Dens
LF Dist6 c + LF Dens

19

3515.2

24
16

3210.2
3518.8

25
21
17

3211.9

79.4

<0.001

24

FOR Distic
FOR Dens
FOR Dist* c + FOR
Dens

3447.0
3521.2

314.5
388.7

<0.001
<0.001

16
16

23

3450.1

MOG Dist3,*

27

3210.4

317.6
77.9

<0.001
<0.001

18
26

MOG Dens3
MOG Dist3,6c +
MOG Dens3
CE Dist*
CE Dens
CE Dist6 c+ CE Dens

Winter (n = 11,625)

Summer/Fall (n = 8,669)

Calving (n = 2,200)

'

'

28

3212.2

79.7

<0.001

27

1922.9

71.9

<0.001

29

7401.9

47.5

<0.001

32

10249.1

249.4

<0.001

26
19

3137.2
3522.9

4.6
390.4

0.09
<0.001

25
16

1854.6
2107.6

3.5
256.6

0.15
<0.001

27
20

7402.6
7693.3

48.2
338.9

<0,001
<0.001

32
21

10065.1
11851.8

65.4
1852.1

<0.001
<0.001

30

3132.5

0.0

0.91

27

1851.1

0.0

0.85

31

7354.4

0.0

1.00

34

9999.7

0.0

1.00

“MOG may or may not have been included due to seasonal proximity (distance) from herd
*Gaussian (squared) term was most parsimonious in at least one seasonal candidate model
“L inear term was most parsimonious in at least one seasonal candidate model

Appendix E. Table 2. Number o f parameters (k), Akaike’s Information Criterion values (AICc), and AICc weights (w) for seasonal
resource selection models for the Quintette caribou herd monitored from 2003 —2009 across the South Peace region o f northeastern
British Columbia. Sample size o f caribou locations is in parentheses.
Calving (n - 5,868)

Spring (n = 9,791)
Model Covariates
Forest cover
Forest age
Solar insolation41
Landscape

k
8
5
3
16
19
17

AAIC
844.0
1063.4

AIC„
<0.001
<0.001

k
9
5

3
17

AICc
4939.8
5159.1
5101.2
4411.0

1005.4
315.2

<0.001
<0.001

1
15

19
18
20

4370.0
4369.2
4330.9

274.2
273.4

<0.001
<0.001

309.2

<0.001
<0.001
<0.001

235.2

<0.001

18
16
19

289.1
958.1

<0.001
<0.001

26
17

4288.3
4396.5

192.6
300.8

<0.001
<0.001

24
16

AICc
7490.9
8277.3
8274.9
6415.0

AAIC
2035.3
2821.8

AIC„
<0.001
<0.001

k
9
5

2819.3
959.5

<0.001
<0.001

308.3
895.7

Road Dist4
Road Dens
Road Dist6 + Road Dens

20

5763.9
6351.2
5764.8

LF Dist6
LF Dens

25
17

5744.6
6413.6

Winter (n = 28,368)

Summer/Fall (n = 22,458)
k
10
5

AICc
24357.5
26493.5

5539.2
2490.5

AIC„.
<0.001
<0.001
<0.001
<0.001

3
18

16966.6

1081.2
2242.4
1063.4

<0.001
<0.001
<0.001

16737.4
18345.4

834.2
2442.2

<0.001
<0.001

AICc
19032.4
20116.6
21442.4
18393.7

AAIC
3129.3
4213.4

16984.3
18145.5

AIC„.
<0.001
<0.001

25082.7
22322.4

AAIC
4684.0
6819.9
5409.2
2648.9

21
19
22

21227.9
22068.9

1554.4
2395.4

21197.5

1524.0

<0.001
<0.001
<0.001

27
19

20969.4
22281.9

1295.9
2608.3

<0.001
<0.001

<0.001
<0.001

LF Dist6 + LF Dens

26

5746.5

291.0

<0.001

27

4285.4

189.6

<0.001

25

16725.5

822.4

<0.001

28

20968.7

1295.1

<0.001

FOR Dist6
FOR Dens
FOR Dist6 + FOR Dens

22
18

5671.5
5663.5

216.0
207.9

<0.001
<0.001

24
19

4276.0
4359.6

180.2
263.8

<0.001
<0.001

21
17

16939.2
18140.1

1036.1
2237.0

<0.001
<0.001

24
20

21086.0
22064.9

1412.4
2391.3

<0.001
<0.001

24

5651.3

195.8

<0.001

26

4265.1

169.3

<0.001

23

16844.7

941.6

<0.001

26

20765.9

MOG Dist3,6
MOG Dens3
MOG Dist3,6 + MOG
Dens3

31
18

5520.7
6353.6

65.1
898.0

<0.001
<0.001

32
19

4175.2
4397.4

79.5
301.6

<0.001
<0.001

30
17

15979.6
18299.9

76.4
2396.7

<0.001
<0.001

31
20

20109.2
22239.8

1092.3
435.7
2566.2

<0.001
<0.001
<0.001

33

5522.0

66.4

<0.001

34

4172.5

76.8

15968.0

64.9

<0.001

429.4

5466.2
6322.1

10.6
866.6

<0.001
<0.001

35
20

4102.0
4386.9

6.3
291.2

30
18

15933.3
18301.0

30.1
2397.8

<0.001
<0.001

33
34
23

20103.0

34
19

<0.001
0.04
<0.001

32

CE Dist6
CE Dens

19949.3
22166.5

275.8
2492.9

<0.001
<0.001
<0.001

CE Dist6 + CE Dens

37

5455.5

0.0

1.00

38

4095.8

0.0

0.96

33

15903.2

0.0

1.00

39

19673.6

0.0

1.00

“MOG may or may not have been included due to seasonal proximity (distant) from herd
^Gaussian (squared) term was most parsimonious in at least one seasonal candidate model
T in ear term was most parsimonious in at least one seasonal candidate model

Appendix E. Table 3. Number o f parameters (k), Akaike’s Information Criterion values (AICc), and AICc weights (w) for seasonal
count models for wolf packs monitored between 2008 - 2010 across the South Peace region o f northeastern British Columbia.

Non-Winter“brm

Early Winter*"1’

Late Winter"1"™

160

Upper Sukunka (n = 33,599)
Model
#
Model Covariates

k

AICc

AAIC

AIC„

k

AICc

AAIC

AIC„

k

AICc

AAIC

AIC„

1
2

Forest cover
Forest age

3
3

10824.8
10882.1

1232.6
1289.8

<0.001
<0.001

4
3

6146.5
6316.9

1096.6
1267.0

<0.001
<0.001

5
3

7565.3
7556.3

1448.5
1439.4

<0.001
<0.001

3
4
5

Water Dist"
Caribou
Landscape

3
1
10

10810.1
10898.4
10642.1

1217.9
1306.1
1049.9

<0.001
<0.001
<0.001

3
1
11

6208.0
6291.4
5975.5

1158.1
1241.4
925.5

<0.001
<0.001
<0.001

3
1
12

7496.5
7552.9
7322.5

1379.7
1436.1
1205.7

<0.001
<0.001
<0.001

6

LF Dist0

7

LF Dist + LF Dens"

16
17

10529.5
10516.4

937.3
924.2

<0.001
<0.001

17
18

5867.9
5867.3

818.0
817.3

<0.001
<0.001

18
19

7170.0
7168.0

1053.2
1051.1

<0.001
<0.001

8

FOR Dist"

9

FOR Dist + FOR Dens"

16
15

10508.7
10529.1

916.5
936.9

<0.001
<0.001

17
18

5862.1
5860.3

812.2
810.4

<0.001
<0.001

16
17

7190.2
7192.2

1073.3
1075.3

<0.001
<0.001

10

MOG Dist"

22

9636.6

44.4

<0.001

23

5087.1

37.1

<0.001

22

6116.9

0.0

0.45

11

MOG Dist + MOG Dens"

23

9629.1

36.9

<0.001

25

5084.8

34.9

<0.001

23

6118.3

1.4

0.22

12

CE Dist"

13

CE Dist + CE Dens"

19
21

9598.8
9592.2

6.6
0.0

0.04
0.96

26
32

5073.1
5049.9

23.2
0.0

<0.001
1.00

23
25

6117.8
6120.6

0.9
3.7

0.27
0.07

‘’Gaussian (squared) term was most parsimonious in seasonal candidate model

Appendix E. Table 3. Continued.

Upper M urray (n = 35,959)
Model
#
1

Non-W inter"”1”

Late W inter”"’

Early W inter"’™

Model Covariates

k

AICc

AAIC

AIC„

k

AICc

AAIC

AIC„.

k

AICc

Forest cover

5

11682.2

694.3

<0.001

4

5130.2

773.0

<0.001

5
4

6224.2

540.6

<0.001

6364.4

680.8

<0.001

1

6262.5
6233.8

578.8

<0.001

550.1

<0.001

6066.4

382.8

<0.001

AAIC

AIC„

2

Forest age

5

11699.1

711.2

<0.001

3

5177.4

820.2

<0.001

3
4

Water Dist

1

562.7

<0.001

1

2

731.0

<0.001

2

5264.5
5252.1

907.3

13

11516.4

528.5

<0.001

10

5061.3

894.9
704.1

<0.001
<0.001

5

Caribou
Landscape

11550.6
11718.9

<0.001

2
12

6

LF Dist0

19

11312.2

324.4

<0.001

12

4826.5

469.3

<0.001

18

5789.1

105.4

<0.001

7

LF Dist + LF Dens°

22

11289.2

301.4

<0.001

13

4824.1

466.9

<0.001

19

5737.4

53.8

<0.001

8

FO RDist0

16

11521.4

533.5

<0.001

11

5043.9

686.7

<0.001

15

5982.2

298.6

<0.001

9

FOR Dist + FOR Dens0

19

11515.4

527.6

<0.001

14

5036.2

679.0

<0.001

18

5978.9

295.3

<0.001

10

MOG Dist

16

11027.1

39.2

<0.001

15

4417.2

59.9

<0.001

14

5816.0

132.3

<0.001

11

MOG Dist + MOG Dens

18

11029.8

41.9

<0.001

18

4357.2

0.0

0.72

15

5778.8

95.1

<0.001

12

CE Dist°

20

11026.9

39.0

<0.001

15

4417.2

59.9

<0.001

16

5789.3

105.6

<0.001

13

CE Dist + CE Dens°

27

10987.9

0.0

1.00

17

4359.1

1.9

0.28

18

5683.6

0.0

1.00

“Gaussian (squared) term was most parsimonious in seasonal candidate model

Appendix E. Table 3. Continued.

O nion Creek (n = 10,493)

Model
#

Early W inter anb

Non-W interaBb

Late W interaob

Model Covariates

k

AICc

AAIC

AICm

k

AICc

AAIC

AICv,

k

AICc

AAIC

AIC„

1
2

Forest cover
Forest age

6
3

7139.0
7136.2

416.9
414.1

<0.001
<0.001

6
3

3565.0
3659.9

485.9
580.9

<0.001
<0.001

5
3

4339.3
4423.6

663.7
748.0

<0.001
<0.001

3
4
5

Water Dist"
Caribou
Landscape

3
2
14

7030.8
7150.9
6942.3

308.7
428.8
220.2

<0.001
<0.001
<0.001

3
2
14

3563.9
3513.5
3364.5

484.9
434.4
285.4

<0.001
<0.001
<0.001

3
2
13

4321.1
4357.5
4153.7

645.5
681.9
478.1

<0.001
<0.001
<0.001

6

LF Dist"

20

6873.8

151.7

<0.001

16

3326.9

247.8

<0.001

19

3975.5

299.8

<0.001

7

LF Dist + LF Dens"

21

6872.8

150.7

<0.001

17

3314.6

235.5

<0.001

20

3968.8

293.2

<0.001

8

FOR Dist"

15

6930.0

6891.9

<0.001

17

3345.7

266.7

<0.001

19

4042.1

366.5

<0.001

9

FOR Dist + FOR Dens"

16

6925.8

6887.7

<0.001

18

3338.8

259.7

<0.001

20

4018.7

343.1

<0.001

10

MOG Dist

17

6731.6

147.0

<0.001

23

3092.9

13.8

<0.001

20

3694.7

19.1

<0.001

11

MOG Dist + MOG Dens

19

6727.0

142.5

<0.001

24

3093.4

14.3

<0.001

21

3692.9

17.3

<0.001

12

CE Dist

17

6731.6

9.4

<0.001

23

3092.9

13.8

<0.001

17

3711.1

35.4

<0.001

13

CE Dist + CE Dens

20

6722.1

0.0

1.00

25

3079.1

0.0

1.00

23

3675.6

0.0

1.00

"Gaussian (squared) term was most parsimonious in seasonal candidate model

Appendix E. Table 3. Continued.

Non-W inter“Qb

Chain Lakes (n = 3 3 8 9 )

Model
#

L ate W inter“"b

Early Winter"1’™

Model Covariates

k

AICc

AAIC

AIC„

k

AICc

AAIC

AIC„

k

AICc

AAIC

a ic m

1
2

Forest cover
Forest age

4
3

4422.9
4368.2

408.7
354.0

<0.001
<0.001

6
3

7414.5
7394.8

239.9
220.1

<0.001
<0.001

4
3

3412.9
3394.0

160.7
141.8

<0.001
<0.001

3
4
5

Water Dist“
Caribou
Landscape

3
1
11

4329.2
4365.9
4140.0

315.0
351.7
125.7

<0.001
<0.001
<0.001

3
1
13

7357.8
7390.5
7263.0

183.2
215.8
88.4

<0.001
<0.001
<0.001

3
1
11

3392.0
3373.9
3320.5

139.8
121.7
68.3

<0.001
<0.001
<0.001

6

LF Dist“

7

LF Dist + LF Dens

13
14

4095.2
4055.4

81.0
41.2

<0.001
<0.001

15
16

7212.5
7189.9

37.8
15.2

<0.001
<0.001

17
18

3295.7
3270.8

43.5
18.7

<0.001
<0.001

8

FOR Dist“

9

FOR Dist + FOR Dens“

17
18

4061.2
4044.7

46.9
30.5

<0.001
<0.001

19
20

7206.8
7203.9

32.1
29.2

<0.001
<0.001

13
14

3289.0
3280.2

36.8
28.1

<0.001
<0.001

10

MOG Dist

11

MOG Dist + MOG Dens

14
18

4083.7
4067.7

69.4
53.4

<0.001
<0.001

22
24

7212.9
7189.2

38.2
14.5

<0.001
<0.001

14
16

3297.6
3278.4

45.4
26.3

<0.001
<0.001

12
13

CE Dist
CE Dist + CE Dens

15
18

4041.9
4014.2

27.7
0.0

<0.001
1.00

19
22

7199.6
7174.7

24.9
0.0

<0.001
1.00

15
18

3278.5
3252.2

26.3
0.0

<0.001
1.00

“Gaussian (squared) term was most parsimonious in seasonal candidate model