WINTER ROOSTING ECOLOGY OF SILVER-HAIRED BATS (LASIONYCTERIS
NOCTIVAGANS) IN SOUTHERN BRITISH COLUMBIA
by
Emily de Freitas
B.Sc. University of Guelph, 2016

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

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
August 2023
© Emily de Freitas, 2023

ii

TABLE OF CONTENTS
Abstract .......................................................................................................................................... iv
List of Tables ................................................................................................................................. vi
List of Figures ................................................................................................................................ ix
Acknowledgements ....................................................................................................................... xii
1. Introduction ................................................................................................................................. 1
1.1

Background ...................................................................................................................... 1

1.2

Objectives ......................................................................................................................... 4

1.3 Organization of Thesis .......................................................................................................... 5
2. Seasonal (Summer and Winter) Roost Characterization and Use by Silver-Haired Bats
(Lasionycteris noctivagans) ............................................................................................................ 6
2.1 Introduction ........................................................................................................................... 6
2.1.1 Roost Selection by Silver-Haired Bats (Lasionycteris noctivagans)............................ 10
2.2 Methods ............................................................................................................................... 14
2.2.1 Study Site...................................................................................................................... 14
2.2.2 Data Collection ............................................................................................................. 16
2.2.3 Data Analysis ................................................................................................................ 20
2.3 Results ................................................................................................................................. 27
2.3.1 Winter Roosts ............................................................................................................... 27
2.3.2 Summer Roosts ............................................................................................................. 37
2.3.3 Winter Versus Summer Roost Comparison.................................................................. 43
2.3.4 Roost Re-Use ................................................................................................................ 49
2.4 Discussion ........................................................................................................................... 51
3. Microclimates and Torpor Patterns Inside Silver-Haired Bat (Lasionycteris noctivagans)
Hibernacula ................................................................................................................................... 57
3.1 Introduction ......................................................................................................................... 57
3.2 Methods ............................................................................................................................... 63
3.2.1 Study Site...................................................................................................................... 63
3.2.2 Data Collection ............................................................................................................. 64
3.2.3 Data Analysis ................................................................................................................ 66
3.3 Results ................................................................................................................................. 69
3.3.1 Roost Microclimates ..................................................................................................... 69

iii

3.3.2 Torpor and Arousal Patterns ......................................................................................... 78
3.4 Discussion ........................................................................................................................... 82
4. General Discussion ................................................................................................................... 88
Literature Cited ............................................................................................................................. 93
Appendix A: Skin Temperature Measurements – Problems and Observations.......................... 110
Appendix B: Acoustics Around Hibernacula and Singing Silver-Haired Bats (Lasionycteris
noctivagans) ................................................................................................................................ 114
Introduction ............................................................................................................................. 114
Methods ................................................................................................................................... 115
Results ..................................................................................................................................... 118
Discussion ............................................................................................................................... 121

iv

Abstract
Many animals spend a considerable amount of time within a shelter, and shelter availability can
influence the distribution, abundance, and diversity of animal populations. Shelters used during
winter, when ambient temperatures are low (< 0°C), must provide animals with adequate
protection to survive through months of adverse conditions. For bats, shelters (roosts) used
during winter must also provide microclimates that support energetic requirements for
hibernation. Silver-haired bats (Lasionycteris noctivagans) are a tree-roosting species that are
migratory throughout much of their range. In parts of northwestern North America, however,
they may hibernate locally. Hibernation sites (hibernacula) in regions where winter temperatures
are below freezing are typically caves, underground mines, or rock crevice features, as these
protect bats from cold temperatures and provide a humid environment that prevents dehydration.
In a forested area near Beasley in southern British Columbia, silver-haired bats use an abandoned
mine, rock crevices, and trees as roosts during winter. The use of trees as roosts during winter in
cold regions (average temperatures < 0°C) is poorly understood. I sought to investigate the
winter ecology of tree-roosting silver-haired bats in southern British Columbia. In Chapter 2, I
described the characteristics of trees used as hibernacula by silver-haired bats. I hypothesize that
bats select winter roost trees non-randomly, and differently across seasons. In Chapter 3, I
investigated microclimates and torpor patterns inside silver-haired bat mine, tree, and rock
crevice hibernacula. I hypothesize that microclimates vary among winter roost types, and
because of these differences, torpor, arousal and movement patterns will differ among tree, rock
crevice and mine winter roosts. A majority (66.7%, n = 22) of silver-haired bats used trees as
winter roosts during the study. Winter tree roosts were selected non-randomly, and compared to
summer roosts, bats used features that provided insulation, such as, cavities in large-diameter,

v

low-decay trees in areas of low canopy closure. Winter trees roosts were colder but more humid
compared to the mine and rock crevice roosts. Bats switched among roosts throughout winter
depending on ambient conditions but did not alter torpor patterns among roost types. Bats
typically used humid roosts (trees) on drier days, and less humid roosts (mine) on more humid
days. Further, bats used the well-insulated mine on colder days, and poorly insulated trees and
rock crevice roosts on warmer days. Despite this, bats showed more tolerance for colder roost
temperatures than expected, remaining in trees during periods of cold (-9°C) temperatures. I
conclude that there may be a trade-off between roost humidity and temperature, with bats using
multiple hibernacula to optimize energetic benefits in the context of ambient conditions. Trees
are an important part of silver-haired bat winter ecology. The use of trees as hibernacula needs to
be better understood to inform forest management decisions that better support silver-haired bat
conservation by protecting important winter roosts.

vi

List of Tables
Table 2.1. Tree decay-stage classification system. Modified from Vegetation Resources
Inventory (VRI) Ground Sampling Procedures Handbook 2018. ................................................ 19
Table 2.2. A priori models used in conditional logistic regression to compare winter roosts used
by silver-haired bats (Lasionycteris noctivagans) from December – March 2021, 2022, to
potentially available roosts in the Smallwood Creek area near Beasley, British Columbia. DBH
indicates diameter at breast height. Roost type is either exfoliating bark or cavity. Canopy
placement is the location of the roost tree within the canopy (above/within or below). The closest
potential roost is the distance to the nearest tree that has an available roosting feature (m). Snag
basal area is the cross-sectional area of all trees in decay class > 2 in a plot (m 2/ha), measured at
breast height (1.3 m). Δ Height indicates the distance between roost height, and tree-top (m).
Distance to nearest edge indicates the distance (m) to the nearest linear feature (i.e., road,
cutblock, vehicle trail). Three-meter (3m) canopy closure (%) is the percent of sky obscured by
vegetation at 3m from the roost. Bark cover (%) is the percentage of the tree (approximate) that
is covered in bark. ......................................................................................................................... 22
Table 2.3. A priori models used in conditional logistic regression to compare summer roosts
used by silver-haired bats (Lasionycteris noctivagans) from June – August 2021, to potentially
available roosts in the Smallwood Creek area near Beasley, British Columbia. DBH indicates
diameter at breast height. Roost type is either exfoliating bark or cavity. The closest potential
roost is the distance to the nearest tree that has an available roosting feature (m). Snag basal area
is the cross-sectional area of all trees in decay class > 2 in a plot (m 2/ha), measured at breast
height (1.3m). Δ Height indicates the distance between roost height, and tree-top (m). Distance to
nearest edge indicates the distance (m) to the nearest linear feature (i.e., road, cutblock, vehicle
trail). Three-meter (3m) canopy closure (%) is the percent of sky obscured by vegetation at 3m
from the roost. Bark cover (%) is the percentage of the tree (approximate) that is covered in bark.
....................................................................................................................................................... 23
Table 2.4. A priori models used in logistic regression to compare summer (June – August 2021)
and winter (December – March 2021, 2022) roosts used by silver-haired bats (Lasionycteris
noctivagans) in the Smallwood Creek area near Beasley, British Columbia. Roost type is either
exfoliating bark or cavity. DBH indicates diameter at breast height. Roost type is either
exfoliating bark or cavity. The closest potential roost is the distance to the nearest tree that has an
available roosting feature (m). Snag basal area is the cross-sectional area of all trees in decay
class > 2 in a plot (m2/ha), measured at breast height (1.3m). Δ Height indicates the distance
between roost height, and tree-top (m). Distance to nearest edge indicates the distance (m) to the
nearest linear feature (i.e., road, cutblock, vehicle trail). Three-meter (3m) canopy closure (%) is
the percent of sky obscured by vegetation at 3m from the roost. Bark cover (%) is the percentage
of the tree (approximate) that is covered in bark. ......................................................................... 25

vii

Table 2.5. Top conditional logistic regression models with ΔQAICc < 5 for comparing winter
tree roosts used by silver-haired bats (Lasionycteris noctivagans) to potentially available trees
across two winters (December 2020 – March 2021; January 2021 – March 2022) at the
Smallwood Creek area near Beasley, British Columbia. Possible explanatory variables the
candidate model set included: diameter at breast height (DBH; cm), distance to closest potential
roost (m), snag basal area (m2/ha), and Δ Height (m). The number of parameters in each model
(K), the quasi-Akaike’s Information Criterion corrected for small sample sizes (QAICc) score,
ΔQAICc score, model weight (Weight) and area under the receiver operating characteristic
(ROC) score for all models are included. ..................................................................................... 32
Table 2.6. Mean ± standard deviation (SD) and range (max – min) of variables included in the
top models for winter roost trees used by silver-haired bats (Lasionycteris noctivagans) and
potentially available trees (random) identified across two winters (December 2020 – March
2021; December 2021 – March 2022) in the Smallwood Creek area near Beasley, British
Columbia. Variables in each model include: DBH, which is diameter at breast height (cm), the
closest potential roost (m), Δ Height (m) which is the distance between tree top and roost height),
percent bark cover (%), distance to the nearest edge which is the distance to the nearest linear
features (i.e., road, cutblock, vehicle trail), the snag basal area (m 2/ha) in the plot, and percent
canopy closure 3m away from the roost. ...................................................................................... 35
Table 2.7. Conditional logistic regression models for characteristics of summer tree roost use by
silver-haired bats (Lasionycteris noctivagans) compared to potentially available trees identified
from 01 June – 31 August 2021 in the Smallwood Creek area near Beasley, British Columbia
with ΔQAICc < 5. The number of parameters (K), the Akaike’s information Criterion corrected
for small sample sizes (QAICc) score, ΔQAICc score, model weight (Weight) and area under the
receiver operating characteristic (ROC) score for all models are included. The variables in each
model are: snag basal area (m2/ha), distance to closest potential roost (m), 3m canopy closure
(%) and distance to nearest edge (m). ........................................................................................... 41
Table 2.8. Mean ± standard deviation of variables included in the top models to compare
potentially available roost trees with summer roost trees used by silver-haired bats (Lasionycteris
noctivagans) from 01 June – 31 August 2021 in the Smallwood Creek area near Beasley, British
Columbia. Variables in each model include: snag basal area (m 2/ha), the closest potential roost
(m), and canopy closure (%) at 3m from the roost. ...................................................................... 42
Table 2.9. Top logistic regression model with AICc < 5 and the null model for characteristics of
summer versus winter tree roost use by silver-haired bats (Lasionycteris noctivagans) identified
from June – August 2021 and December – March (2021, 2022) in the Smallwood Creek area
near Beasley, British Columbia. The number of parameters (K), the Akaike’s information
Criterion corrected for small sample sizes (AICc) score, ΔAICc score, model weight (Weight)

viii

and area under the receiver operating characteristic (ROC) score for all models are included. The
variables in the top model are: 3m canopy closure (%) and roost type (exfoliating bark or cavity).
....................................................................................................................................................... 47
Table 2.10. Parameter estimates, standard errors (SE), odds ratios, and 95% lower and upper
confidence intervals (CI) for odds ratios in the top logistic regression model for winter versus
summer tree roost use by silver-haired bats (Lasionycteris noctivagans) from December – March
(2021, 2022) and June – August 2021 in the Smallwood Creek area near Beasley, British
Columbia. ...................................................................................................................................... 48
Table 3.1. Mean ± standard deviation (SD) and range (minimum – maximum) of ambient
(outside) temperature (°C) and relative humidity (%) recorded during winter from 20 December
– 29 March across five years (2012, 2014, 2018, 2021, 2022) at the Smallwood Creek area near
Beasley, British Columbia. ........................................................................................................... 70
Table 3.2. Parameter estimates, standard errors (SE) and 95% lower and upper confidence
intervals (CI) in the logistic regression model for arousal likelihoods determined by changes in
weather variables for silver-haired bats (Lasionycteris noctivagans) using rock, tree and mine
roosts in Beasley, British Columbia from December – March over five years (2012, 2014, 2018,
2021, 2022). Odds ratios with confidence intervals not overlapping one are considered drivers of
the model. Variables are: ΔP, absolute value of 24-hour change in ambient pressure (hPa); ΔT,
absolute value of 24-hour change in ambient temperature (°C); ΔVPD absolute value of 24-hour
change in ambient vapour pressure deficit.................................................................................... 80
Table B-1. Summary of low frequency bat recordings from six bat detectors (SM4Bat, Swift or
SM2+) deployed at silver-haired bat (Lasionycteris noctivagans) hibernacula in the Smallwood
Creek area near Beasley, British Columbia from January – April 2021. An additional detector
was deployed but did not capture any acoustic data during the sampling period and was excluded
from analysis. The number of low frequency bat calls included files that were labelled as LANO
(silver-haired bat), EPFULANO (big drown bat; Eptesicus fuscus –silver-haired bat dyad), or
25K (low frequency bat). ............................................................................................................ 119

ix

List of Figures
Figure 2.1. Map view of all winter tree (blue circles), winter rock (yellow triangles), and
summer tree (orange diamonds) roosts used by silver-haired bats (Lasionycteris noctivagans)
located across six field seasons (winter: December – March 2012, 2014, 2018, 2020, 2021,
summer: June 2021 –August 2021) at the Smallwood Creek area outside of Beasley, British
Columbia. The white star indicates the location of the Queen Victoria Mine. ............................. 28
Figure 2.2. The proportion of tree species used as roosts (Roost) and potentially available tree
species (Available) used by silver-haired bats (Lasionycteris noctivagans) identified across five
winters (December – March; 2012, 2014, 2018, 2020, 2021) at the Smallwood Creek area
outside of Beasley, British Columbia. .......................................................................................... 29
Figure 2.3. The proportion of decay classes of trees used as roosts (Roost) and potentially
available tree species (Available) used by silver-haired bats (Lasionycteris noctivagans)
identified across five winters (December – March; 2012, 2014, 2018, 2021, 2022) at the
Smallwood Creek area outside of Beasley, British Columbia. ..................................................... 31
Figure 2.4. An image of a silver-haired bat (Lasionycteris noctivagans) using a drill hole (from
previous mine exploration activities) as a hibernaculum on 17 February 2022 in the Smallwood
Creek area near Beasley, British Columbia. This bat did not have a transmitter and was located
visually; however, this feature had been used the day before by a radiotagged silver-haired bat,
who was located the next day in a rock crevice nearby. The bat with no radiotransmitter
possessed an aluminium band on its left forearm, indicating that it had been previously captured,
and was likely a male, based on banding protocols, where males are banded on the left forearm,
and females are banded on the right forearm. ............................................................................... 36
Figure 2.5. The proportion of tree species used as summer roosts (Roost) and potentially
available tree species (Available) used by silver-haired bats (Lasionycteris noctivagans)
identified from June 2021 – August 2021 at the Smallwood Creek area outside of Beasley,
British Columbia. .......................................................................................................................... 38
Figure 2.6. The proportion of decay classes of trees used as summer roosts (Roost) and
potentially available trees (Available) used by silver-haired bats (Lasionycteris noctivagans)
identified from June 2021 – August 2021 at the Smallwood Creek area outside of Beasley,
British Columbia. .......................................................................................................................... 39
Figure 2.7. The proportion of tree species used as summer roosts (identified from June 2021 –
August 2021) and winter roosts (December – March 2012, 2014, 2018, 2021, 2022) by silver-

x

haired bats (Lasionycteris noctivagans) at the Smallwood Creek area outside of Beasley, British
Columbia. ...................................................................................................................................... 45
Figure 2.8. The proportion of decay classes used as summer roosts (identified from June 2021 –
August 2021) and winter roosts (December – March 2012, 2014, 2018, 2021, 2022) by silverhaired bats (Lasionycteris noctivagans) at the Smallwood Creek area outside of Beasley, British
Columbia. ...................................................................................................................................... 46
Figure 2.9. A still image of a trail camera video showing a silver-haired bat (Lasionycteris
noctivagans) visiting an identified winter tree roost in the Smallwood Creek area near Beasley,
British Columbia on 23 February 2022 (3 additional visits were documented on 24 January, 21
February, and 23 February 2022). This roost was located using radio telemetry on 26 January
2021. Identification of this as a silver-haired bat is possible because of its size and distinct
coloration. ..................................................................................................................................... 50
Figure 3.1. Mean daytime temperatures (°C; left) and nighttime temperatures (°C; right) in mine
(n = 1), rock (n = 3) and tree (n = 5) hibernacula used by silver-haired bats (Lasionycteris
noctivagans) from 20 December 2021– 29 March 2022 and 15 January 2014 – 05 March 2014 at
the Smallwood Creek area near Beasley, British Columbia. The dark bar in the center of the
boxplot indicates the median value of the data, the grey box limits represent the 75 th (upper) and
25th (lower) quartiles, and the whiskers represent minimum and maximum values..................... 72
Figure 3.2. Mean daily temperatures (°C) ± 1 standard deviation (shading) recorded inside mine
(n = 1), rock (n = 2), and tree (n = 5) hibernacula used by silver-haired bats (Lasionycteris
noctivagans), and ambient temperature (outdoor) recorded from 20 December 2021 – 29 March
2022 in the Smallwood Creek area near Beasley, British Columbia. ........................................... 73
Figure 3.3. Ambient (Ta), roost (Tr) and skin (Tsk) temperatures of two different silver-haired
bats (Lasionycteris noctivagans) roosting in two different tree roosts from January 2021–
February 2021, near Beasley, British Columbia. Sharp increases in Tsk indicate arousals.
Measurements on the two animals occur on different days. ......................................................... 74
Figure 3.4. Mean daytime vapour pressure deficit (VPD; hPa; left) and nighttime vapour
pressure deficit (hPa; right) in recorded hibernacula used by silver-haired bats (Lasionycteris
noctivagans; mine (n = 1), rock (n = 3) and tree (n = 5) recorded from 20 December 2021 – 29
March 2022 and 15 January 2014 – 05 March 2014 at the Smallwood Creek area near Beasley
British Columbia. The dark bar in the center of the boxplot indicates the median value of the
data, the grey box limits represent the 75th (upper) and 25th (lower) quartiles, and the whiskers
represent minimum and maximum values. ................................................................................... 76

xi

Figure 3.5. Mean daily internal roost humidity (RH, %) ± 1 standard deviation (shading)
recorded in mine (n = 1), rock (n = 2) and tree (n = 3) hibernacula used by silver-haired bats
(Lasionycteris noctivagans) and ambient (outdoor) humidity recorded from 20 December 2021–
29 March 2022 in the Smallwood Creek area near Beasley, British Columbia. .......................... 77
Figure 3.6. Ambient (Ta), roost (Tr) and skin (Tsk) temperatures of three different silver-haired
bats (Lasionycteris noctivagans) roosting in the mine (A), a rock crevice roost (B) and a tree
roost (C) from 31 January 2021– 08 March 2021 in the Smallwood Creek area near Beasley,
British Columbia. Sharp increases in T sk indicate arousals. Measurements of A, B, and C occur
on different days. .......................................................................................................................... 79
Figure 3.7. Frequency of arousals resulting in a roost switching event during either positive (+)
or negative (-) 24-hour changes in pressure (hPa), temperature (°C), or vapour pressure deficit
(VPD) occurring from the three winter roost types (mine, rock and tree) to another roost by
silver-haired bats (Lasionycteris noctivagans) in the Smallwood Creek area near Beasley, British
Columbia from December – March of five years (2012, 2014, 2018, 2021, 2022). .................... 81
Figure A-1. Skin (Tsk) temperature (°C) a silver-haired bat (Lasionycteris noctivagans)
hibernating in the Queen Victoria Mine from January 2022 – March 2022 in the Smallwood
Creek area near Beasley, British Columbia, recorded on a Lotek SRX–400 unit. Sharp increases
in Tsk indicate arousals. Skin temperature changed widely during torpor based on 1 beat per
minute increments, highlighted by the arrows. In this case, a transmitter rate of 22 was equivalent
to Tsk of 12.16°C (lower arrow), and a transmitter rate of 23 was equivalent to 17.43°C. ........ 113
Figure B-1. Map of acoustic detectors deployed from 08 December 2020 to 16 – 29 April 2021
at silver-haired bat (Lasionycteris noctivagans) hibernacula in the Smallwood Creek area near
Beasley, British Columbia. The location of detectors placed at tree hibernacula are marked with
yellow diamonds. The location of the Queen Victoria Mine detector is marked with a red star.
..................................................................................................................................................... 117
Figure B-2. Silver-haired bat (Lasionycteris noctivagans) acoustic recordings captured from the
Smallwood Creek area near Beasley, British Columbia from December 2020 – April 2021. A) A
typical repeated silver-haired bat ‘song’ phrase; B) and C) Previously undocumented and unique
silver-haired bat social calls. ....................................................................................................... 120

xii

Acknowledgements
I have a very long list of people to thank for getting me here.
First, I would like to thank my supervisor, Erin Baerwald for her support and patience
throughout the project. Erin brought insight, unwavering encouragement, and most importantly,
laughs through this process. Erin provided the best environment to learn in, and challenged me to
improve, and ask big questions. I am a better scientist because of her.
I would like to thank Cori Lausen, my committee member and unofficial co-supervisor.
She was instrumental in making this project happen, discovered bats in the Queen Victoria Mine,
and was the first to track silver-haired bats to trees in the area. Thank you for providing
invaluable support and expertise, and for convincing me that sitting outside all night in the dead
of winter would be fun (it was). I know what it means to be a good field scientist because of her
mentorship.
I am grateful to those who provided assistance and feedback on project design, including
my committee members, Russ Dawson and Phil Burton for their valuable insight on early drafts,
and Kristin Jonasson for helping me work through analysis problems, and providing much
needed expertise.
For assistance with field work, I thank Tory Rhoads, Brett Gandy, Caroline Lafond, and
many volunteers. I am eternally grateful you chose to wallow through snow, bushwhack, carry
heavy packs, chase bouncing telemetry signals, wade through stinky wetlands and work long
nights with me. Catching and tracking bats is not easy work, and I could not have done it without
you.
Thank you to the WSC Canada staff, Heather Gates, Dana Blouin and Jason Rae for their
field, technical and administrative expertise. This team helped me when I had no idea what I was
doing, and I am grateful for their support.
This project would not have been possible without a long list of funding partners,
including Wildlife Conservation Society Canada, Mitacs Accelerate Program, University of
Northern British Columbia, Fish and Wildlife Compensation Fund, Columbia Basin Trust, and to
Carley Doleman and Dave DeRosa with the Okanagan Nation Alliance, who greatly supported
this project.
I thank the rest of the Baerwaldians for being the most supportive lab out there. I will
miss seeing you and your dogs on my zoom screen every week.
To my family and friends who provided support every step of the way. I would like to
thank my dad for inspiring me to be a scientist through countless (insane) stories of his time as a
grad student. To my mom, for always encouraging me to pursue my goals, even if that involved
chasing bats- one day I will convince you to love them as much as I do.
Finally, to Luke, who has been my rock through everything. Thank you for your support
and willingness to take on this adventure with me. I truly could not have done this without you.

1

1. Introduction
1.1 Background
Conserving habitat requires understanding an animal’s needs and how these change
across seasons and years. In simple terms, habitat is where an animal lives and refers to the
environmental requirements that an individual, population, or species needs to survive (Block
and Brennan 1993). In Canada, federal conservation and management strategies are focused on
conserving ‘critical habitat’ that species require for the survival because it is of high quality, or it
is occupied during a key period of the lifecycle (Environment and Climate Change Canada
2002). For many species, conserving critical habitat involves conserving shelters (i.e., roosts,
burrows, or dens) used during sensitive periods, such as during the reproductive or hibernation
seasons (Grillet et al. 2010; Martin et al. 2016). The use of shelters is a common behaviour
among animals. Animals use shelters for protection from adverse weather and predators, and as a
place to rest, raise young, and interact socially (Kunz and Lumsden 2003; Clement and
Castleberry 2012; Finkbeiner et al. 2012; Tsunoda et al. 2018). Many animals spend a
considerable amount of time within a shelter, and as such, shelter availability can influence the
distribution and abundance of animal populations (Humphrey 1975; Moretto and Francis 2017).
Because of their importance, the selection of shelters occurs non-randomly, with animals
choosing shelters based on various factors influenced by individual needs, life-history strategies,
and seasonality.
Shelters used during winter in temperate regions, when ambient temperatures are low (<
0°C), must provide animals with adequate protection to survive through months of adverse
conditions. Winter in temperate regions also brings the challenge of low food abundance for

2

many animals, which, combined with the high energetic cost of staying warm in cold conditions,
creates a negative energy balance. To mitigate these challenges, animals can either migrate to
warmer locations where food is abundant, compensate by using energy from stored fat, or
hibernate to conserve energy. For animals that hibernate, shelters must protect them from subfreezing temperatures, while still meeting other physiological needs. It is, therefore, important to
understand the thermal properties and physiological implications of winter shelters (e.g.,
Walsberg 1986; Cooper 1999; Kalcounis-Rüppell et al. 2005).
During winter, when ambient temperatures are cold, endothermic animals must use
energy to maintain body temperature (Tb) and this is particularly costly for small animals that
have higher rates of heat loss compared to larger individuals (Bradley and Deavers 1980). To
offset this heat loss, some animals employ heterothermy. Through metabolic heat production,
endotherms can maintain an elevated and stable body temperature across a range of ambient
temperatures (Geiser 2004). Conversely, heterotherms can reduce body temperature and
metabolic rate to conform with environmental conditions (Geiser 2004). This reduction in energy
expenditures by lowering activity levels and metabolism is a physiological process known as
torpor (Geiser and Ruf 1995; Speakman 2008). This process is used by hibernators to offset the
energetic costs associated with winter (Geiser and Ruf 1995). Hibernation is commonly used by
small mammals, and is characterized by the prolonged use of torpor, lasting days or weeks
(Geiser and Ruf 1995; Geiser 2004). Energy expenditure during hibernation can be reduced by
choosing a shelter with internal conditions that promote prolonged bouts of torpor (Thomas and
Cloutier 1992; Webb et al. 1996). As such, shelter choice can help optimize energy budgets in
winter.

3

All bats in Canada are small, nocturnal insectivores that frequently employ heterothermy
throughout the year. As heterotherms, shelters (i.e., roosts) for bats are important for
thermoregulation and choice varies by species, reproductive status, location, food supply and
season. In summer, bats use caves, rock crevices, anthropogenic structures, and trees as roosts.
For bats that hibernate, winter roosts differ from summer roosts, and individuals choose winter
roosts that help conserve energy and optimize torpor expression. Torpor expression in bats is
strongly influenced by roost microclimate (i.e., temperature and humidity; Twente et al. 1985;
Thomas and Geiser 1997; Geiser 2004). In cold, temperate climates, bats may spend over half
the year in hibernation, and as such, considerable research has focused on understanding
hibernation ecology (Weller et al. 2009; Weller et al. 2018). Despite this, little is known about
the overwintering ecology of many species of bat (Weller et al. 2018). In British Columbia, for
example, 14 of the 15 resident bat species hibernate, but few winter roosts (hibernacula) have
been located (Lausen et al. 2022). The few known hibernacula in western Canada are typically in
caves or mine features that protect from extreme cold winter conditions but are cool enough to
promote torpor expression (Weller et al. 2018; Lausen et al. 2022).
Silver-haired bats (Lasionycteris noctivagans) are a tree-roosting species found
throughout forested regions in North America and are common in British Columbia. They belong
to the guild of cavity-dwelling bats and typically roost in tree-bole crevices, previously
excavated cavities, and under loose bark (Kunz 1982). The silver-haired bat is migratory
throughout its range, typically moving to southern portions of the United States during the winter
(Cryan 2003). Although migration is common, silver-haired bats have been found hibernating at
more northern latitudes, such as Minnesota and British Columbia (Nagorsen et al. 1993; Kurta et
al. 2018; Lausen et al. 2022) and have been recorded acoustically in winter in Southeast Alaska

4

(Blejwas et al. 2014). These winter records of silver-haired bats have begun to shift our
perspective on their migratory strategies. In British Columbia, silver-haired bats have been found
overwintering in underground mines, rock crevices, and trees (Nagorsen et al. 1993; Lausen et al.
2022). The use of trees as winter roosts in cold climates is poorly understood, and it has been
hypothesized that trees may not provide sufficient protection from cold ambient temperatures
during winter (Burles 2014).

Provincially, silver-haired bats are yellow-listed, meaning populations are considered
secure (BC CDC 2015). Although not currently listed federally on the Species at Risk Registry,
this species has been recommended for endangered by the Committee on the Status of
Endangered Wildlife in Canada as endangered (COSEWIC 2023). These individuals face threats
of mortality from wind energy facilities (Arnett et al. 2008), habitat loss through forest clearing
(Kunz and Lumsden 2003), and potential threats from disease (Bernard et al. 2015). For other
federally listed bat species, the little brown myotis (Myotis lucifugus) and the tri-colored bat
(Perimyotis subfalvus), winter roosts are considered critical habitat. Understanding the
overwintering ecology and tree use by silver-haired bats is important for conservation, especially
considering the anticipated change in their legal status.
1.2 Objectives
I investigated the winter ecology of silver-haired bats in southern British Columbia. In
Chapter 2, I (1) determine characteristics of trees used as hibernacula by silver-haired bats and if
they use trees non-randomly on the landscape and (2) compare winter tree-roost characteristics to
summer roosts used at the same site. I hypothesize that bats select roosts non-randomly, and that
winter roosts differ from summer roosts. In Chapter 3, I (3) investigate microclimates inside

5

silver-haired bat hibernacula, and hypothesize that microclimates vary among roost types but are
buffered against ambient temperatures. I also determine (4) how silver-haired bats use torpor in
the winter, and hypothesize that torpor, arousal, and movement patterns differ among roost types.
1.3 Organization of Thesis
This thesis is written manuscript-style, with two stand-alone data chapters and a general
introduction, conclusion, and appendices. This results in inevitable repetition of background
material, and methods throughout the text. Chapter 2 and Chapter 3 will be submitted for review
upon completion of this thesis. Upon publication, Chapter 2will be co-authored by Dr. Cori
Lausen, Thomas Hill and Dr. Erin Baerwald, and Chapter 3 will be co-authored by Dr. Cori
Lausen, Dr. Erin Baerwald and Dr. Kristin Jonasson. The methods, results and discussion that
address objectives 1 and 2 (listed above) are addressed in Chapter 2 (Seasonal (Summer and
Winter) Roost Characterization and Use by Silver-Haired Bats). Chapter 3 (Microclimates and
Torpor Patterns Inside Silver-Haired Bat Hibernacula) discusses the methods, results and
discussion addressing objectives 3 and 4.

6

2. Seasonal (Summer and Winter) Roost Characterization and Use by Silver-Haired Bats
(Lasionycteris noctivagans)
2.1 Introduction
Shelter sites (i.e., roosts, dens, or burrows) are essential for many animals, providing
protection from adverse weather, predators, and as a place to rest, raise young, and interact
socially (Kunz and Lumsden 2003; Clement and Castleberry 2012; Finkbeiner et al. 2012;
Tsunoda et al. 2018). Many species deliberately choose shelter locations based on a variety of
different factors that are influenced by individual needs and life-history strategies. Shelters used
by bats (i.e., roosts) must meet very specific criteria, and as such, we typically consider them a
limiting resource (Kunz and Lumsden 2003). In addition to selecting roosts based on the abovementioned criteria, bats select roosts based on microclimates that help meet energetic
requirements. Bats have high energy demands and are required to balance a tight energy budget.
Flight, the primary method of locomotion for bats is energetically intensive compared to running
in terrestrial animals (Thomas and Suthers 1972), and because of a high surface area to volume
ratio, bats have higher levels of heat loss compared to mammals of similar size (Speakman and
Thomas 2003). As insectivores (insect-eating), bats are unable to acquire energy when insect
prey is unavailable, which, in Canada, occurs during winter. To combat these avenues of energy
loss, bats make use of energy-saving strategies, including selecting roosts that provide
microclimates that reduce energy expenditures.
All temperate-zone bats can fluctuate between defending an elevated and stable body
temperature (Tb), or use heterothermy to save energy (Geiser 2004). Maintaining typical
mammalian Tb requires significant energy, but as heterothermic endotherms, bats do not have to
maintain a stable and elevated Tb constantly. When endotherms self-regulate, body temperature is

7

maintained at a consistent elevated level using metabolism, despite fluctuations in ambient
temperature (Geiser 2004). Conversely, during heterothermy, individuals can reduce or increase
metabolic rate in such a way that body temperature more closely conforms to ambient conditions
(Geiser 2004). Reducing Tb can be beneficial when energy supply is low, usually caused by low
prey abundance. Heterotherms can reduce Tb below the standard 30 – 40°C of euthermic
conditions (Barnes 1989; Geiser and Ruf 1995). This process, known as torpor, results in
reduced metabolic activity and, consequently, a reduction in physiological function, such as
growth and reproduction (Speakman 2008). The thermal properties of a roost may promote or
discourage torpor use, and microclimate is thus a crucial consideration in roost selection for bats.
Bats roost in a variety of structures, including caves, rock crevices, anthropogenic
structures (e.g., houses and bridges), and trees (Kunz and Lumsden 2003). Tree roosts are
particularly important for bats, as over half of all bat species use trees or other plants as roosts
(Kunz and Lumsden 2003). Forest loss is a growing concern globally (Hansen et al. 2013), and a
potential threat to tree-roosting bats that rely on forests. As such, considerable focus has been
directed to understanding the ecology of tree-roosting bats, and the physical characteristics
associated with tree roosts, to better support their conservation in the face of habitat loss
(Kalcounis-Rüppell et al. 2005). The importance of tree roost characteristics as they relate to
microclimates can vary among individuals when there are differences in energetic needs. For
instance, reproductive female bats have different thermoregulatory requirements than males and
non-reproductive females. Reproduction is energetically expensive, especially in females.
Pregnant and lactating female bats will select warmer roosts to facilitate higher Tb and metabolic
rates, which promotes fetal development, lactation, and offspring development (Racey and Swift
1981; Zahn 1999; Speakman 2008). For example, cavity-roosting bats select tree cavities with

8

thick walls, typically with south-facing orientations that provide thermal benefits through solar
exposure and insulation (Wiebe 2001; Parsons et al. 2003). In contrast, males and nonreproductive females might be more opportunistic in their summer roost selection because they
do not have the same reproductive expenditures as pregnant or lactating females. In many
temperate-zone bats, males use torpor more often and for more extended periods than
reproductive females (Hamilton and Barclay 1994). Males may be more likely to use thermally
labile roosts, such as space under exfoliating bark, that allow them to exploit daily torpor use
(Hamilton and Barclay 1994).
Many species of cavity- and bark-roosting bats select for specific structural
characteristics of trees or roosts beyond those that influence microclimate (Kalcounis-Rüppell et
al. 2005). Although features like tree diameter, species, height, canopy position, and decay status
directly influence microclimate, they can provide additional benefits beyond energetics. For
example, trees that are taller, larger, and more distant from neighbouring trees might help
facilitate orientation by bats (Campbell et al. 1996) and reduce predation risk (Betts 1998). Large
trees may also provide more roosting opportunities for bats to choose from than small trees
(Harper et al. 2005). Beyond tree and roost characteristics, bats may also select roosts at a stand
level. For example, many species of tree-roosting bat select roosts in open-canopy stands with
high snag (i.e., a standing dead tree) densities (Mattson et al. 1996; Cryan et al. 2001; KalcounisRüppell et al. 2005). High densities of snags may provide more access to alternative roosts to
support frequent roost-switching during summer (Mattson et al. 1996). The proximity of the
roost to a canopy gap can be an important factor in roost selection for some bats (KalcounisRüppell et al. 2005), as they provide bats with travel corridors and foraging opportunities for

9

open-space foragers (Blakey et al. 2017), resulting in the areas being used more intensely than
areas without gaps (Tena et al. 2020).
In the winter at northern latitudes, insect prey is limited, so bats hibernate to conserve
energy or else migrate to avoid harsh conditions. During hibernation torpor is periodically
interrupted by hibernal arousals, where bats warm back up to euthermic Tb and perform
necessary biological functions (Trachsel et al. 1991; Thomas and Cloutier 1992; Burton and
Reichman 1999; Park et al. 2003; Ben-Hamo et al. 2013). For hibernating bats, repeated arousal
from torpor can consume critical fat reserves and reduce survival (Speakman et al. 1991), so bats
commonly use extended torpor bouts, lasting days to weeks (Geiser 2004) and short arousals,
lasting hours, to ensure their fat stores last until the end of winter (Speakman et al. 1991; Geiser
2004). Extended torpor requires specific roosting conditions (microclimates), and bats in eastern
North America typically use cool, humid, and well-insulated winter roosts (hibernacula) to
protect them from sub-freezing ambient temperatures (Webb et al. 1996). Most of these
hibernacula are rock crevices or mines, especially at northern latitudes (e.g., Swanson and Evans
1936; Clark et al. 1997; Kunz et al. 1997), which are typically well-insulated and meet thermal
requirements. The hibernation strategies of bats in western North America are largely unknown
because discovering hibernacula is difficult (Weller et al. 2018), except in cases where
individuals form large groups while hibernating, such as Townsend's big-eared bats
(Corynorhinus townsendii) (Ingersoll et al. 2010; Weller et al. 2018; Whiting et al. 2018). In the
southern United States, many tree-roosting bat species may overwinter in trees (Boyles and
Robbins; Newman et al. 2021), which have minimal thermal insulation and promote short torpor
bouts and passive re-warming for daily arousal, allowing bats to feed throughout the winter.

10

Winter temperatures in these areas are mild (i.e., > 0°C daily average), and it is unknown how
common this behaviour is at northern latitudes.
2.1.1 Roost Selection by Silver-Haired Bats (Lasionycteris noctivagans)
Silver-haired bats are a tree-roosting species found throughout forested regions in North
America and are common in British Columbia. They belong to the guild of cavity-dwelling bats
and typically roost in tree-bole crevices, excavated cavities, and under loose bark (Kunz 1982).
Roost selection by silver-haired bats follows the above-described patterns, with individuals using
different roost types depending on reproductive status, sex, age, and season. In summer, silverhaired bats select roost trees that have recently died or are in early stages of decay (Crampton
and Barclay 1998), are larger in diameter than those that are available on the landscape, and
taller than the surrounding canopy (Campbell et al. 1996; Vonhof and Barclay 1996; Barclay et
al. 1998; Betts 1998). Trees that are taller, larger, and more distant from neighbouring trees
might help facilitate orientation by bats (Campbell et al. 1996), maximize solar exposure for
roost warmth, and reduce predation risk (Betts 1998). Large trees may also have more cavities
available for use (Harper et al. 2005). Newly dead trees have firmer wood that provides more
insulation than rotten wood (Hooge et al. 1999), and the increased presence of bark on newly
dead trees may further increase insulation (Bär and Mayr 2020) and provide roost space.
Female silver-haired bats often form small maternity colonies of under 30 individuals and
choose warm roosts to promote reproductive function (Vonhof and Barclay 1996; McAlpine et
al. 2021). These maternity colonies are found in tree crevices or holes excavated by woodpeckers
(Vonhof and Barclay 1996; McAlpine et al. 2021), and these roosts provide more insulation than
alternative roosts, such as exfoliating bark. The location of a roost within the canopy and the
amount of sunlight it receives will also influence roost temperature. Pregnant and lactating

11

silver-haired bats preferentially select trees with increased solar exposure, such as those in a
canopy gap or above the canopy (Campbell et al. 1996; Mattson et al. 1996; Betts 1998).
Because they do not have the same energetic constraints, non-reproductive silver-haired bats are
more flexible with their roost selection. They roost under loose bark, in crevices, and tree
hollows (Vonhof and Barclay 1996; Vonhof and Gwilliam 2007). Non-reproductive individuals
select trees of varying heights; some studies show selection for trees above the canopy
(Crampton and Barclay 1998; Vonhof and Gwilliam 2007), while others report selection for
shorter trees (Barclay et al. 1988). Difference in canopy closure may influence these choices,
with shorter trees in areas with open canopies receiving more insolation than taller trees within a
closed canopy.
The selection of specific tree species by silver-haired bats is poorly understood. In the
Pacific Northwest (i.e., Oregon, Washington, and British Columbia), silver-haired bats appear to
prefer western white pine (Pinus monticola) and ponderosa pine (P. ponderosa) (Campbell et al.
1996; Vonhof and Barclay 1996). Trembling aspen (Populus tremuloides) is preferentially
selected by silver-haired bats in Alberta (Crampton and Barclay 1998) and parts of British
Columbia (Vonhof and Gwilliam 2007). Conversely, other studies have suggested that tree
species did not appear to play a role in roost selection (Betts 1998).
Silver-haired bats are migratory throughout much of their range (Cryan 2003). In the
winter, they are commonly found in southern and eastern portions of the United States (Cryan
2003). In Arkansas, silver-haired bats have been found roosting under south-facing bark and in
tree cavities (Perry et al. 2010), similar to other tree-roosting bat species overwintering in the
southern United States (Boyles and Robbins 2006; Newman et al. 2021). Although silver-haired
bats are considered migratory, they have been found overwintering at more northerly latitudes

12

(Nagorsen et al. 1993; Kurta et al. 2018; Lausen et al. 2022). Silver-haired bats can overwinter as
far north as the -12.2°C minimum daily January isocline (Kurta et al. 2018), which includes
much of British Columbia. Where they are found overwintering in these cold regions, they use a
variety of well-insulated roost types commonly used by other hibernating bats, including
underground, abandoned mines (Bonewell et al. 2017; Kurta et al. 2018), buildings (Brigham
1995), and rock crevices (Weller et al. 2018; Lausen et al. 2022). Silver-haired bats have also
been found roosting in trees in southern British Columbia during the winter (Nagorsen et al.
1993; Lausen et al. 2022), but the characteristics of these roosts have not been described. The
large body of work on tree-roosting bats typically describes summer roost selection (reviewed in
Kalcounis-Rüppell et al. 2005). Furthermore, all previous research on the use of tree roosts in
winter by temperate-zone bats has been in areas where mean winter temperatures rarely drop
below 0°C. To fully understand the winter ecology of silver-haired bats, the use of trees as
hibernacula in regions where winter temperatures are below 0°C must be examined. Winter treeroost selection in colder climates is poorly studied, and in areas where bats occupy trees yearround, little research exists on seasonal preferences. Bats hibernating in trees may be susceptible
to roost loss and can die if unable to escape a falling tree while in deep torpor (Cryan and
Veilleux 2007). Since many of the recommendations on bat protection in managed forests
considers only summer use of trees, understanding how bats use forests in the winter, and if this
use differs from summer use or not, will better inform timber harvest timing windows to support
bat conservation.
To date, three silver-haired bat hibernation areas have been identified in British
Columbia: one is a mine near Nelway, where logistics have made studies challenging, another is
an open, ponderosa pine forest near Castlegar, where bats hibernate in a south-facing rocky bluff,

13

and anecdotally, in ponderosa pine trees (C. Lausen, personal communication, [July 2023]).).
The third is a hibernation area near Beasley, consisting of a shallow abandoned mine (the Queen
Victoria Mine), a forested area and a rock outcrop. Previous winter research at the Queen
Victoria Mine showed that silver-haired bats hibernate in trees and rock crevices adjacent to the
mine and alternate among these three types of roost throughout winter (Lausen et al. 2022).
Based on summer captures, it is known that silver-haired bats use the mine as a night roost, and
likely the forested area in the summer months, but this has not been investigated. My research
aimed to characterize winter roosts used by silver-haired bats and compare with summer roosts
in the same area. Because of differing thermoregulatory requirements across seasons, I
hypothesized that silver-haired bats select roosts non-randomly, and that winter roosts differ
from roosts selected in summer. Specifically, I predicted:
a) In winter, silver-haired bats select roosts with characteristics that protect against subfreezing ambient temperatures, such as roosts in large-diameter, low-decay trees and the
use of cavities instead of exfoliating bark. Because I expected these types of features to
be available at low densities on the landscape, winter roosts were significantly different
from what is available.
b) In summer, males and non-reproductive females select roosts with characteristics that
have low insulation capacity, such as exfoliating bark roosts in small-diameter trees. I did
not expect these types of features to be limiting on the landscape, and I predicted summer
roosts were not significantly different from what was potentially available. Conversely,
reproductive females (if captured) select well-insulated, warm roosts, such as cavities in
large-diameter trees, and in areas with high solar exposure. I expected these

14

characteristics will be significantly different from what is available because they are
similar to expected winter tree roost characteristics and will be limiting on the landscape.
c) Winter roosts are significantly different from summer roosts used by non-reproductive
individuals because of differing thermoregulatory requirements. Specifically, I predicted
winter roosts were more likely to have characteristics that increase thermal stability, and
thus prolonged torpor use compared to summer roosts used by males and nonreproductive females, which promote thermal instability and daily torpor cycles. I
predicted roosts used by reproductive females in the summer were similar in structure
and surrounding habitat features to the above-described winter roosts, and because of
this, would not differ significantly from winter roosts.
2.2 Methods
2.2.1 Study Site
The study area is in the Smallwood Creek drainage above Beasley, British Columbia.
(49.49350, -117.44960). This area is located within the Interior Cedar-Hemlock Biogeoclimatic
zone (Pojar et al. 1987). The dominant tree species are Douglas-fir (Pseudotsuga menziesii),
western redcedar (Thuja plicata), western hemlock (Tsuga heterophylla), and ponderosa pine.
The Queen Victoria Mine, an underground copper-silver-gold mine used until 1956, is a focal bat
roost feature in this area. In winter, it is a hibernaculum for the silver-haired bat, Townsend's bigeared bat and the California myotis (Myotis californicus). From 2012 – 2018, radio telemetry
work conducted here identified eight trees used as winter roosts for silver-haired bats, where
individuals used space under exfoliating bark and in cavities throughout the winter (Cori L.
Lausen, [Wildlife Conservation Society Canada, Kaslo, British Columbia], personal

15

communication, [December 2020]). In summer, the mine serves mainly as a night roost and
social gathering location for these same species in addition to at least five others (big brown bat,
Eptesicus fuscus; Yuma myotis, M. yumanensis; long-eared myotis, M. evotis; long-legged
myotis, M. volans; and little brown myotis, M. lucifugus). The mine consists of an underground
opening with several chambers, which typically remain above 0°C in the winter and provide a
suitable selection of hibernation locations. In the summer, bats frequent the mine and roost in the
adjacent forested area (Cori L. Lausen, [Wildlife Conservation Society Canada, Kaslo, British
Columbia], personal communication, [December 2020]). This site is one of only three identified
silver-haired bat mine hibernacula in western North America, all in the West Kootenay region of
British Columbia (Lausen et al. 2022). It is also the only silver-haired bat mine hibernaculum in
the province that is safe to access, and thus it is an ideal research location. Despite this site being
the only area in British Columbia where silver-haired bats have been verified to hibernate in
trees, timber harvesting occurs within the habitat surrounding the Queen Victoria Mine. Logging
practices can severely reduce roosting habitat through direct tree removal and subsequent
landscape modification (Russo et al. 2010; Borkin et al. 2011), and risk of direct mortality of
bats in winter is known.
I focused my efforts in a 3km radius around the mine because radiotagged silver-haired
bats to date had not been tracked more than 1km from the mine in the winter (Cori L. Lausen,
[Wildlife Conservation Society Canada, Kaslo, British Columbia], personal communication,
[December 2020]). This provided sufficient netting locations for summer capture while also
allowing for direct comparison to winter captures.

16

2.2.2 Data Collection
I captured free-flying silver-haired bats using mist nets (Avinet Research Supplies,
Maine, USA) for two winter seasons (from December 2020 – March 2021 and January – March
2022) and one summer season (June – August 2021). In winter, I placed one double-high 12m
net across the Queen Victoria Mine entrance to capture bats as they entered and exited the mine.
In summer, I similarly captured free-flying bats at the mine with one double-high net at the
entrance, but additionally in the surrounding forested area by placing multiple nets along flyways
and over water bodies within a 3km radius of Queen Victoria Mine. I used additional data from
bats captured by Wildlife Conservation Society Canada in the winters of 2012, 2014, and 2018
for five total years of winter data.
I recorded sex, age, reproductive condition, mass, and forearm length of each captured
bat. To identify recaptured individuals, I banded bats with a 2.9mm aluminum-lipped metal band
(Porzana Ltd., Sussex, United Kingdom). Female bats were banded on the right forearm and
male bats were banded on the left forearm, which allows for differentiation between sexes
without handling, if bats are observed within a roost. I affixed 0.42g temperature-sensitive radiotransmitters (LB-2NT; Holohil Systems Ltd., Ontario, Canada) to bats by trimming the fur in the
interscapular region and attaching transmitters using a nontoxic latex glue (Osto-bond; Montreal
Ostomy, Quebec, Canada). The combined mass of the transmitter and glue was less than or equal
to 5% of the individual's body mass, as per standard guidelines (Aldridge and Brigham 1988). To
ensure transmitter attachment, I held bats in cloth bags for 15 minutes and then released them at
the capture site. I handled all animals with the approval of the UNBC Animal Care and Use
Committee (Protocol 2020-17) and under provincial bat capture permit MRCB20-598305.

17

I used a hand-held radio telemetry receiver (R-1000 Telemetry System, Communication
Specialists Inc., California, USA) to locate individuals in their day roosts the morning after
capture. The exact position of the bat within the tree was determined by scanning the tree at close
range with the receiver or visually observing the transmitter antenna with binoculars. Once I
located a bat, I placed a data-logging receiver (SRX400/600/800/1200; Lotek Wireless Inc.,
Ontario, Canada) with a three- or five-element Yagi antenna at each identified roost. I set data
logging units to record transmitters every 10 minutes. I used data-logging receivers to record bat
movement, i.e., exiting and returning to a roost and the length of roost occupancy. I tracked bats
and data-logged transmitter signal daily for the duration of the life of the transmitter, which was
approximately 45 days, or until it was not detected for more than seven days. I also placed a
data-logging receiver programmed with all transmitter frequencies used in the study, inside the
mine for the duration of the study, because I expected frequent use by bats, especially during the
winter.
To monitor roost use across seasons and years, I installed remote-trigger infrared trail
cameras (Prime Low-Glow, Bushnell Outdoor Products, Kansas, USA) at trees identified as
roosts in the winter of 2020 – 2021. I installed cameras at two trees on 24 November 2021 and
ran them continuously until 27 September 2022 (614 camera nights). I installed cameras on an
adjacent tree 5 m from the roost tree, with the camera pointed directly at the roost entrance.
Cameras recorded 15-second-long videos when triggered by motion, from sunset until sunrise
during the monitoring period. I defined a bat visit as a video capturing a bat landing on the tree
or repeatedly approaching the tree.
Once I identified a tree roost, I marked it with flagging tape and recorded the location
using a hand-held GPS unit. Once the radiotagged bat left the roost (determined using

18

radiotelemetry), I assessed roosts by collecting information at three scales: roost, tree, and
surrounding habitat (plot) features. I recorded the roost type, delta height (Δ height; the distance
between tree top and roost height), tree species, diameter at breast height in cm (1.3 m above
ground; DBH), decay class, position within the canopy (above or at the height of the canopy, or
below), and percent remaining bark. Δ height was used instead of the height of the roost off the
ground because it provides information on what part of a tree is being used by a bat, regardless
of tree height. A bat using a roost that is higher up in a tree may be more exposed to wind and
sun compared to one that is using a roost lower down in the tree. I determined the decay class
following the Vegetation Resources Inventory Ground Sampling Procedures Handbook
(Ministry of Forests, Lands and Natural Resource Operations 2018; Table 2.1). I defined position
within the canopy as one of two categories: above (including dominant and co-dominant trees) or
below (including intermediate or suppressed trees). Using a digital camera, I took photos of the
canopy in all four cardinal directions at 3m and 15m from the base of the roost tree. I converted
canopy closure photos to binary images and used ImageJ version 1.8 (National Institute for
Health) to estimate the percent canopy closure. I averaged percent canopy closure from each
cardinal direction to estimate canopy closure at 3m and 15m.

19

Table 2.1. Tree decay-stage classification system. Modified from Vegetation Resources
Inventory (VRI) Ground Sampling Procedures Handbook 2018.
Code
1
2
3
4
5
6
7
8

Description
Live and healthy; no decay or obvious defects.
Live and unhealthy; internal decay or growth deformities.
Recently dead; needles or twigs may be present, roots sound.
Dead; no needles/twigs, 50% of branches lost, loose bark, top usually broken, roots
stable.
Dead; most branches/bark absent, some internal decay, roots of larger trees stable.
Dead; no branches/bark, sapwood/heartwood sloughing, lateral roots or larger trees
softening.
Dead; extensive internal decay, outer shell may be hard but hollow, lateral roosts
decomposed, 1/2 of original height.
Dead; extensive internal decay, outer shell may be hard but hollow, lateral roosts
decomposed, 1/3 of original height.

20

Within 0.1-ha circular plots, using roost trees as a focal point (Vonhof and Barclay 1996;
Limpert et al. 2007), I recorded DBH, tree species, decay class and roost availability of all trees
> 10cm DBH. I also recorded the distance from the roost tree to the nearest available roost, and
the distance to the nearest linear opening (e.g., road, powerline, forest edge). I then used the
DBH of measured trees in the plot to calculate tree bole cross-sectional areas, and used these
values to calculate snag basal area (m2/ha) for each plot. Snag basal area is the cross-sectional
area of all trees in decay class > 2 in a plot measured at breast height (1.3 m above ground). This
give a measure of how many available roosts are near the roost tree.
For each identified roost tree, I collected the same suite of data at a randomly-selected
tree within the vicinity of the roost tree. These trees, hereafter called “potentially available
roosts” were > 10cm DBH, had exfoliating bark, or a cavity that could be available as a roost for
bats, and were within 100m of the identified roost tree. Potentially available roosts were selected
using QGIS software (Version 3.16.1), where I buffered each roost tree by 100m and used a
randomly generated point within the buffer to survey for available roosts. If a suitable tree was
not located within 10m of the randomly generated point, a new randomly generated location was
surveyed, and the process repeated until a suitable candidate was discovered. I took this approach
instead of buffering by the average nightly distance travelled by a bat because of our limited
knowledge of the home range size of silver-haired bats.
2.2.3 Data Analysis
I used R Studio version 1.3.1093 (R Development Core Team 2020) for all statistical
analyses. R packages are noted for each analysis throughout. Given the low number of captured
individuals in winter (n = 33), winter trees identified (n = 30), and the assumption that males and
females do not have differing thermoregulatory needs in the winter, I pooled data for both sexes

21

for winter tree roosts. Although females and males have different roosting requirements in the
summer, none of the female bats I captured in the summer were reproductive, so I also pooled
data for both sexes for summer roosts. Mean values for variables are reported as mean ± standard
deviation (SD). I conducted two separate analyses to compare identified roosts to potentially
available roosts, one for winter roosts and one for summer roosts. Data collection methods did
not differ across seasons, and methodologies are the same for each analysis, so they are
described below once unless specific discrepancies occurred.
To test my hypothesis that silver-haired bats use roosts non-randomly relative to available
trees, I used conditional logistic regression with a 1:1 matched case design (Hosmer and
Lemeshow 2000; R package ‘survival’; Therneau 2023). Conditional logistic regression was
conducted on sets of biologically feasible a priori models developed using literature on roosting
ecology, containing combinations of nine variables (Table 2.2, Table 2.3). This method is
appropriate under these conditions because potentially available trees were not selected
independently from roost trees, and instead represent a matched set of roost tree and potentially
available tree. I did not assess a global model because of model over-saturation due to small
sample size.

22

Table 2.2. A priori models used in conditional logistic regression to compare winter roosts used
by silver-haired bats (Lasionycteris noctivagans) from December – March 2021, 2022, to
potentially available roosts in the Smallwood Creek area near Beasley, British Columbia. DBH
indicates diameter at breast height. Roost type is either exfoliating bark or cavity. Canopy
placement is the location of the roost tree within the canopy (above/within or below). The closest
potential roost is the distance to the nearest tree that has an available roosting feature (m). Snag
basal area is the cross-sectional area of all trees in decay class > 2 in a plot (m 2/ha), measured at
breast height (1.3 m). Δ Height indicates the distance between roost height, and tree-top (m).
Distance to nearest edge indicates the distance (m) to the nearest linear feature (i.e., road,
cutblock, vehicle trail). Three-meter (3m) canopy closure (%) is the percent of sky obscured by
vegetation at 3m from the roost. Bark cover (%) is the percentage of the tree (approximate) that
is covered in bark.
Model
1
2
3
4
5
6
7
8
9
10
11
12

Variables
DBH (cm)
Roost type
Canopy placement
Closest potential roost (m) + Snag basal area (m2/ha)
DBH (cm) + Δ Height (m)
Distance to nearest edge (m) + Closest potential roost (m)
Δ Height + Bark cover (%)
DBH (cm) + Bark cover (%)
Closest potential roost (m) + 3m canopy closure (%)
3m canopy closure (%) + Snag basal area (m2/ha)
Distance to nearest edge (m) + DBH (cm)
Bark cover (%) + 3m canopy closure (%)

23

Table 2.3. A priori models used in conditional logistic regression to compare summer roosts
used by silver-haired bats (Lasionycteris noctivagans) from June – August 2021, to potentially
available roosts in the Smallwood Creek area near Beasley, British Columbia. DBH indicates
diameter at breast height. Roost type is either exfoliating bark or cavity. The closest potential
roost is the distance to the nearest tree that has an available roosting feature (m). Snag basal area
is the cross-sectional area of all trees in decay class > 2 in a plot (m 2/ha), measured at breast
height (1.3m). Δ Height indicates the distance between roost height, and tree-top (m). Distance to
nearest edge indicates the distance (m) to the nearest linear feature (i.e., road, cutblock, vehicle
trail). Three-meter (3m) canopy closure (%) is the percent of sky obscured by vegetation at 3m
from the roost. Bark cover (%) is the percentage of the tree (approximate) that is covered in bark.
Model
1
2
3
4
5
6
7
8
9
10
11

Variables
DBH (cm)
Roost type
Closest potential roost (m) + Snag basal area (m2/ha)
DBH (cm) + Δ Height (m)
Distance to nearest edge (m) + Closest potential roost (m)
Δ Height + Bark cover (%)
DBH (cm) + Bark cover (%)
Closest potential roost (m) + 3m canopy closure (%)
3m canopy closure (%) + Snag basal area (m2/ha)
Distance to nearest edge (m) + DBH (cm)
Bark cover (%) + 3m canopy closure (%)

24

Before model selection, I tested for collinearity using Pearson’s correlations and excluded
variables if they were highly correlated (Pearson’s r ≥ 0.7; Dormann et al. 2012). For both data
sets, tree height and DBH were correlated (r = 0.7, r = 0.8, respectively), so I used DBH in the
models. Tree diameter is commonly reported when describing bat roosting preferences
(Kalcounis-Rüppell et al. 2005), allowing a better comparison to available literature. Three-meter
and 15m canopy closure averages were also correlated (r = 0.7) for both data sets. I used 3m
canopy closure averages instead of 15m canopy closure averages because this better describes
the immediate roost microclimate conditions with respect to sun exposure, which is of interest. I
ranked models using Akaike’s Information Criterion corrected for overdispersion in the data and
small sample sizes (QAICc; R package ‘MuMIn’, Bartoń 2023) and considered the models
plausible if they had a ΔAICc ≤ 4, and likely as a top model if it had a ΔAIC c ≤ 2 (Burnham and
Anderson 2002). (Burnham and Anderson 2002).
To determine if winter tree roosts used by silver-haired bats differed from summer tree
roosts, I compared them using binary logistic regression (R package ‘lme4’, Bates et al. 2015) on
a set of a priori-developed models based on biologically relevant hypotheses (Table 2.4). I did
not assess a global model, because of model over-saturation due to small sample size, but I did
assess an intercept-only (null) model. There was collinearity between DBH and tree height, so I
removed tree height from the models and used DBH instead. There was also collinearity between
3m and 15m canopy closure estimates, so I used 3m canopy closure estimates, following the
same logic described above. I ranked models using Akaike’s Information Criterion corrected for
small sample sizes (AICc; R package ‘MuMIn’, Bartoń 2023) and considered a model plausible
if it had a ΔAICc ≤ 4, and competitive as a top model if it had a ΔAIC c ≤ 2 (Burnham and
Anderson 2002).

25

Table 2.4. A priori models used in logistic regression to compare summer (June – August 2021)
and winter (December – March 2021, 2022) roosts used by silver-haired bats (Lasionycteris
noctivagans) in the Smallwood Creek area near Beasley, British Columbia. Roost type is either
exfoliating bark or cavity. DBH indicates diameter at breast height. Roost type is either
exfoliating bark or cavity. The closest potential roost is the distance to the nearest tree that has an
available roosting feature (m). Snag basal area is the cross-sectional area of all trees in decay
class > 2 in a plot (m2/ha), measured at breast height (1.3m). Δ Height indicates the distance
between roost height, and tree-top (m). Distance to nearest edge indicates the distance (m) to the
nearest linear feature (i.e., road, cutblock, vehicle trail). Three-meter (3m) canopy closure (%) is
the percent of sky obscured by vegetation at 3m from the roost. Bark cover (%) is the percentage
of the tree (approximate) that is covered in bark.
Model
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16

Variables
Roost type
DBH (cm)
DBH (cm) + 3m canopy closure (%) + closest potential roost (m)
Δ Height (cm) + Canopy placement
Closest potential roost (m) + Snag basal area (m2/ha)
3m canopy closure (%) + Roost type
Roost type + Bark cover (%)
Distance to water (m)
3m canopy closure (%)
Distance to water (m) + Distance to nearest edge (m)
Bark cover (%) + 3m canopy closure (%)
Snag basal area (m2/ha) + Δ height (m)
Closest potential roost (m) + Distance to nearest edge (m) + Distance to water
(m)
Roost type + Canopy placement
Bark cover (%) + DBH (cm)
DBH (cm) + Δ Height

26

For all analyses, I reported parameter estimates ± standard error (SE), and odds ratios of top
models, which describe the ratio of the probability of an outcome (e.g., roost use), and the
strength and direction of the outcome. I determined the importance of variables based on if 95%
confidence intervals of the odds ratios overlapped one. If a variable appeared in more than one of
the top models, I used multimodel inference (R package ‘AICcmodavg’, Mazerolle 2023) to
average parameter estimates across top models and reported unconditional SE (Burnham and
Anderson 2002). I used the area under the receiver operating characteristic curve (ROC; Hosmer
and Lemeshow 2000; R package ‘ROCR’, Sing et al. 2005) to examine the model accuracy of
top models. This method assesses the discrimination capacity of a given model (Pearce and
Ferrier 2000) and is commonly used with logistic regression (Fielding and Bell 1997). I
considered the performance of the models with ROC > 0.7 as good and ROC > 0.9 as excellent
(Hosmer and Lemeshow 2000).
I analyzed roost aspect, tree species and decay class using different methodologies,
because categorical variables saturate models quickly when sample sizes are small. I used
Rayleigh's (Z) circular statistic to determine if mean roost aspect differed from random for both
summer and winter tree roosts (Zar 1999; R package ‘circular’, Agostinelli and Lund 2022). I
used a Fisher's exact test to compare the proportion of tree species used as roosts with available
tree species calculated from plots, and to compare summer roost trees to winter roost tree
(Limpert et al. 2007; Monarchino et al. 2020). I repeated this procedure with decay class,
comparing decay class of roost trees with the decay classes of available trees, calculated from
plot transects, and to compare summer roost trees to winter roost trees. The significance level for
all Fisher’s tests was assessed at α < 0.05.

27

2.3 Results
2.3.1 Winter Roosts
Thirty-three adult silver-haired bats were captured and radio-tracked (16 female; 17
male) across five winter seasons (December – March; 2012, 2014, 2018, 2021 and 2022). Bats
were tracked for 9 – 61 days ( ̅ = 30.8 ± 13.4) and used 1 – 4 roosts ( ̅ = 1.8 ± 0.8). Of the
silver-haired bats tracked during winter, 22 (66.7%) used trees at least once, 25 (75.8%) used the
mine at least once, and 5 (15.1%) used rock crevices at least once as hibernacula. Bats roosted
longer in the mine ( ̅ = 16.7 ± 13.6 days, range = 3 – 47 days) compared to trees ( ̅ =11.3 ± 7.4,
range = 1 – 27 days) and rock crevices ( ̅ =10.8 ± 11.0 days, range = 1 – 31 days).
A total of 30 tree roosts were identified across five winters: eight from 2012 – 2018 and
22 from 2021 – 2022 (Figure 2.1). Roosts were primarily in cavities (43.3%) and under
exfoliating bark (36.7%), although, on six occasions (20%), the exact roost feature could not be
identified. The proportion of tree species used by bats differed significantly from potentially
available trees (Fisher’s exact test, P = 0.005). The most frequently used tree species were
ponderosa pine (n = 8) and trembling aspen (n = 13), and these were found at low densities
(1.0% and 6.8% respectively). Bats occasionally roosted in Douglas-fir, western larch (Larix
occidentalis), and paper birch (Betula papyrifera) (Figure 2).

28

Figure 2.1. Map view of all winter tree (blue circles), winter rock (yellow triangles), and
summer tree (orange diamonds) roosts used by silver-haired bats (Lasionycteris noctivagans)
located across six field seasons (winter: December – March 2012, 2014, 2018, 2020, 2021,
summer: June 2021 –August 2021) at the Smallwood Creek area outside of Beasley, British
Columbia. The white star indicates the location of the Queen Victoria Mine.

29

Figure 2.2. The proportion of tree species used as roosts (Roost) and potentially available tree
species (Available) used by silver-haired bats (Lasionycteris noctivagans) identified across five
winters (December – March; 2012, 2014, 2018, 2020, 2021) at the Smallwood Creek area
outside of Beasley, British Columbia.

30

The proportion of decay classes used by bats in winter differed significantly from
available decay classes (Fisher’s exact test, P = 0.0004; Figure 2.3). All roost trees (n = 30) were
in decay classes 2 – 6, with bats primarily using trees in lower stages of decay (2 – 4, 80.0%).
Bats frequently used trees in decay class two (43.3%), and trees in this decay class were found at
low densities (10.4%). Bats never used trees in decay class one, but trees in decay class one were
found in high densities (65.7%). Mean roost orientation (170.0° ± 120.9°) did not differ
significantly from random (Z = 0.10, P = 0.81).
For winter tree roosts identified in 2012, 2014, or 2018 (n = 8), habitat data were not
collected, so I did not use them in the conditional logistic regression analysis. Additionally, I
could not determine the exact roost feature at one of the identified winter tree roosts in 2021, so
my final sample size was 21 choice sets for conditional logistic regression and subsequent model
selection, with each set consisting of one used tree and one potentially available tree. Six of the
12 candidate models tested had ΔQAICc ≤ 4 and were considered in the top model set (Table
2.5). All models had good predictive model accuracy (ROC > 0.7; Table 2.5), suggesting weak
support for one best model to explain winter roost use by silver-haired bats.

31

Figure 2.3. The proportion of decay classes of trees used as roosts (Roost) and potentially
available tree species (Available) used by silver-haired bats (Lasionycteris noctivagans)
identified across five winters (December – March; 2012, 2014, 2018, 2021, 2022) at the
Smallwood Creek area outside of Beasley, British Columbia.

32

Table 2.5. Top conditional logistic regression models with ΔQAICc < 5 for comparing winter
tree roosts used by silver-haired bats (Lasionycteris noctivagans) to potentially available trees
across two winters (December 2020 – March 2021; January 2021 – March 2022) at the
Smallwood Creek area near Beasley, British Columbia. Possible explanatory variables the
candidate model set included: diameter at breast height (DBH; cm), distance to closest potential
roost (m), snag basal area (m2/ha), and Δ Height (m). The number of parameters in each model
(K), the quasi-Akaike’s Information Criterion corrected for small sample sizes (QAICc) score,
ΔQAICc score, model weight (Weight) and area under the receiver operating characteristic
(ROC) score for all models are included.
Model Structure

K

QAICc

ΔQAICc

Weight

ROC

DBH

3

24.62

--

0.25

0.79

DBH + Δ Height

4

24.98

0.36

0.21

0.81

DBH + % Bark cover

4

25.19

0.57

0.19

0.81

DBH + Distance to edge

4

26.32

1.70

0.11

0.78

Closest potential roost + Snag basal area

4

27.92

3.30

0.05

0.77

Closest potential roost + 3m canopy closure

4

28.23

3.61

0.04

0.76

33

The top four models contained the variable DBH (model-averaged estimate = 0.06 ± 0.03
[unconditional SE], odds ratio = 1.06, unconditional CI = 0.99 – 1.14). These models all had a
ΔQAICc < 2, meaning they were competitive as top models. Two of the top six models
contained the variable closest potential roost (model-averaged estimate = -0.1 ± 0.06
[unconditional SE], odds ratio = 0.90, unconditional CI = 0.80 – 1.02), however, they had
ΔQAICc > 2, meaning they may be less competitive as top models. For both DBH and distance
to closest roost, the 95% unconditional confidence intervals for the odds ratios overlapped one,
indicating they are weak predictors. Other variables in the top model set include Δ Height
(estimate = -0.07 ± 0.06, odds ratio = 0.93, CI = 0.83 – 1.05), percent bark cover (estimate = 0.05 ± 0.06, odds ratio = 0.95, CI = 0.84 – 1.07), distance to edge (estimate = -0.006 ± 0.009,
odds ratio = 1.01, CI = 0.99 – 1.02), snag basal area (estimate = -0.007 ± 0.007, odds ratio =
1.01, CI = 0.99 – 1.02) and 3m canopy closure (estimate = -0.02 ± 0.03, odds ratio = 0.99, CI =
0.92 – 1.04). All variables in the top model sets have odds ratios with confidence intervals
overlapping one and are weak predictors of winter roosts. Although none of the top models
contained strong predictors, the presence of the variables DBH and closest potential roost in
multiple top models suggests these characteristics are important for predicting winter roost use.
Based on the top models, increasing DBH or decreasing distance to the closest potential roost
increases the likelihood that a bat will occupy a tree in winter. Observed differences in variables
support this result; the mean DBH of roost trees was larger than the potentially available roost
tree DBH, and roost trees were closer to potential roosts compared to potentially available trees
(Table 2.6).
I located six roosts in rock crevices used by three radiotagged silver-haired bats. Due to
the small sample size, I did not conduct complete habitat assessments on these roosts. All rock

34

crevice roosts, except one, were in the same rocky outcrop ~200 m from the Queen Victoria
Mine. To my knowledge, this is the only known rock outcrop within a 1km radius of the mine.
Rock crevice roosts were in vertical fissures in rock cliff faces (n = 4), a drill hole in an alcove of
a rock cliff (n = 1; Figure 2.4), or a vertical fissure on the outside of the mine (n = 1). All rock
crevice roosts were southwest to southeast facing. On one occasion, a bat with a transmitter
roosted openly in the shallow drill hole (Figure 2.4), and I observed this roost being occupied by
silver-haired bats without transmitters periodically across multiple winters.

35

Table 2.6. Mean ± standard deviation (SD) and range (max – min) of variables included in the
top models for winter roost trees used by silver-haired bats (Lasionycteris noctivagans) and
potentially available trees (random) identified across two winters (December 2020 – March
2021; December 2021 – March 2022) in the Smallwood Creek area near Beasley, British
Columbia. Variables in each model include: DBH, which is diameter at breast height (cm), the
closest potential roost (m), Δ Height (m) which is the distance between tree top and roost height),
percent bark cover (%), distance to the nearest edge which is the distance to the nearest linear
features (i.e., road, cutblock, vehicle trail), the snag basal area (m 2/ha) in the plot, and percent
canopy closure 3m away from the roost.
Variable

Roost (n = 21)

Available (n = 21)

Mean ± SD

Range

Mean ± SD

Range

DBH (cm)

61.72 ± 36.18

26.43 – 172.61

39.66 ± 19.88

14.20 – 75.90

Closest potential roost
(m)

8.12 ± 5.54

1.52 – 22.76

12.53 ± 8.24

0.00 – 33.42

Δ Height (m)

17.64 ± 14.47

2.08 – 41.41

14.49 ± 10.64

0.25 – 31.56

Bark cover (%)

86.88 ± 24.73

5.00 – 100.00

94.05 ± 7.12

80.00 – 100.00

Distance to edge (m)

75.64 ± 85.77

3.30 – 332.49

85.03 ± 75.53

0.01 – 347.38

Snag basal area (m2/ha)

85.52 ± 69.21

4.20 – 215.00

66.26 ± 43.27

24.00 – 229.15

3m canopy closure (%)

54.80 ± 12.19

32.67 – 76.04

61.01 ± 11.59

34.81 – 74.75

36

Figure 2.4. An image of a silver-haired bat (Lasionycteris noctivagans) using a drill hole (from
previous mine exploration activities) as a hibernaculum on 17 February 2022 in the Smallwood
Creek area near Beasley, British Columbia. This bat did not have a transmitter and was located
visually; however, this feature had been used the day before by a radiotagged silver-haired bat,
who was located the next day in a rock crevice nearby. The bat with no radiotransmitter
possessed an aluminium band on its left forearm, indicating that it had been previously captured,
and was likely a male, based on banding protocols, where males are banded on the left forearm,
and females are banded on the right forearm.

37

2.3.2 Summer Roosts
I captured eight adult silver-haired bats (6 male, 2 female) from 01 June – 31 August
2021. The two females were both non-reproductive. I could not determine the roosting location
of one bat, so I excluded it from the analysis. I tracked bats for 2 – 12 days ( ̅ = 6.0 ± 3.9 days).
Bats used from 2 – 4 ( = 3.1 ± 1.2) tree roosts.
I located 21 tree roosts during the summer (Figure 2.1). Sixteen roosts were located under
exfoliating bark (76.2%) and in two in cavities (9.5%), and on three occasions (14.3%), the exact
roost feature within a tree could not be identified. The proportion of tree species used by bats
differed significantly from available trees (Fisher’s exact test, P = 0.0004). Bats mainly roosted
in Douglas-fir (33.3%) and trembling aspen (28.6%), and these species were found at relatively
low densities (22.5% and 2.5%, respectively; Figure 2.5). Western redcedar (Thuja plicata) was
found at relatively high densities, however summer roosts were not found in this species (Figure
2.5).
All roosts were in trees in decay classes two to five. The proportion of decay classes used
by bats differed significantly from available decay classes (Fisher’s exact test, P = 0.04; Figure
2.6). Bats used trees in decay class four (n = 12, 57.1%), and decay class five (n = 4, 19.0%)
most frequently. Trees in decay class one were never used, even though they were found at a
density of 70.7% (Figure 2.6). Bats did not use roosts with a particular orientation around trees,
as mean roost orientation (194.0° ± 92.5°) did not differ significantly from random (Z = 0.19, P =
0.24).

38

Figure 2.5. The proportion of tree species used as summer roosts (Roost) and potentially
available tree species (Available) used by silver-haired bats (Lasionycteris noctivagans)
identified from June 2021 – August 2021 at the Smallwood Creek area outside of Beasley,
British Columbia.

39

Figure 2.6. The proportion of decay classes of trees used as summer roosts (Roost) and
potentially available trees (Available) used by silver-haired bats (Lasionycteris noctivagans)
identified from June 2021 – August 2021 at the Smallwood Creek area outside of Beasley,
British Columbia.

40

The exact roost could not be identified on three occasions, so I used a sample size of 18
choice sets for conditional logistic regression. Two of the 11 a priori models had low ΔQAICc (≤
4), and good model accuracy (ROC > 0.7; Table 2.7). Both top models contained the variable
snag basal area (model-averaged estimate = -0.01 ± 0.01 [unconditional SE], odds ratio = 0.99
unconditional 95% CI = 0.97 – 1.00). The 95% unconditional confidence interval for the odds
ratio does not overlap one, but does include one, suggesting this variable is a weak predictor for
summer roosts. Other variables included in the top model set include the distance to the closest
potential roost (estimate = 0.17 ± 0.10 odds ratio = 1.19, CI = 0.97 – 1.45) and 3m canopy
closure (estimate = 0.05 ± 0.06, odds ratio = 1.05, CI = 0.92 – 1.19). Confidence intervals of
odds ratios for both variables overlapped one, indicating these variables were weak predictors of
summer roost use. Based on the top model, increasing snag basal area increases the likelihood
that a bat will occupy a roost in the summer. Observed differences in variables support this
result; the snag basal area of roost tree plots was higher compared to available tree plots.
Although distance to closest potential roost was a weak predictor, roost trees were closer to
potential roosts compared to potentially available trees (Table 2.8).

41

Table 2.7. Conditional logistic regression models for characteristics of summer tree roost use by
silver-haired bats (Lasionycteris noctivagans) compared to potentially available trees identified
from 01 June – 31 August 2021 in the Smallwood Creek area near Beasley, British Columbia
with ΔQAICc < 5. The number of parameters (K), the Akaike’s information Criterion corrected
for small sample sizes (QAICc) score, ΔQAICc score, model weight (Weight) and area under the
receiver operating characteristic (ROC) score for all models are included. The variables in each
model are: snag basal area (m2/ha), distance to closest potential roost (m), 3m canopy closure
(%) and distance to nearest edge (m).
Model Structure

K

QAICc

ΔQAICc

Weight

ROC

Snag basal area + Closest potential
roost

4

20.67

--

0.57

0.86

Snag basal area + 3m canopy closure

4

23.89

3.22

0.11

0.79

Closest potential roost + Distance to
edge

4

24.95

4.28

0.07

0.73

42

Table 2.8. Mean ± standard deviation of variables included in the top models to compare
potentially available roost trees with summer roost trees used by silver-haired bats (Lasionycteris
noctivagans) from 01 June – 31 August 2021 in the Smallwood Creek area near Beasley, British
Columbia. Variables in each model include: snag basal area (m 2/ha), the closest potential roost
(m), and canopy closure (%) at 3m from the roost.
Variable

Roost (n = 18)

Available (n = 18)

Mean ± SD

Range

Mean ± SD

Range

Snag basal area
(m2/ha)

120.10 ± 86.49

24.00 – 337.00

62.81 ± 48.97

22.00 – 141.00

Closest potential
roost (m)

7.20 ± 4.48

1.65 – 16.4

10.80 ± 9.19

1.40 – 17.90

3m canopy closure
(%)

70.79 ± 4.85

55.92 – 77.61

71.22 ± 6.59

59.13 – 82.42

43

2.3.3 Winter Versus Summer Roost Comparison
The species of trees used as roosts differed significantly across seasons (Fisher’s exact
test, P = 0.017; Figure 2.7). Eight (26.7%) winter roosts were in ponderosa pine, but silverhaired bats did not use ponderosa pines in summer. Conversely, four (19.1%) summer roosts
were in grand fir (Abies grandis), but bats did not use grand fir in the winter. Bats commonly
used trembling aspen and Douglas-fir in both summer and winter. Bats used trees in different
decay classes across seasons (Fisher’s exact test, P = 0.008; Figure 2.8). Decay class four was
most frequently used in the summer (n = 12, 54.5%), compared to class two in winter (n = 13,
43.3%).
Of the models tested to compare winter and summer roost trees, one carried 96% of
model weight and had excellent model accuracy (ROC > 0.9; Table 2.9). The top model included
the two variables: 3m canopy closure and roost type, both strong predictors with odds ratio
confidence intervals not overlapping one (Table 2.10). For every 1% increase in canopy closure
at 3m from the tree, the odds of that roost being used in the winter decrease by 24.0%. The use of
exfoliating bark instead of a cavity decreases the odds of a tree being occupied as a roost in
winter compared to the summer by 97.0%. Summer roost trees were in areas with higher canopy
closure ( ̅ = 70.3% ± 5.0%) compared to winter roost trees ( ̅ = 55.1% ± 12.3%). Additionally,
bats roosted more frequently in exfoliating bark in the summer (n = 16, 76%) but used both
cavities (n = 13, 43.3%) and exfoliating bark (n = 11, 36.0%) in winter.
Roosting locations of silver-haired bats also varied across seasons. Most summer roosts
(n = 16, 76.2%) were located within 1km from the bats’ capture location with the furthest a bat
was found roosting being 2km from the capture location. In winter, roosts were typically within

44

500 m of the mine (i.e., the capture location), except for one tree roost located approximately 750
m southwest of the mine (Figure 2.1).

45

Figure 2.7. The proportion of tree species used as summer roosts (identified from June 2021 –
August 2021) and winter roosts (December – March 2012, 2014, 2018, 2021, 2022) by silverhaired bats (Lasionycteris noctivagans) at the Smallwood Creek area outside of Beasley, British
Columbia.

46

Figure 2.8. The proportion of decay classes used as summer roosts (identified from June 2021 –
August 2021) and winter roosts (December – March 2012, 2014, 2018, 2021, 2022) by silverhaired bats (Lasionycteris noctivagans) at the Smallwood Creek area outside of Beasley, British
Columbia.

47

Table 2.9. Top logistic regression model with AICc < 5 and the null model for characteristics of
summer versus winter tree roost use by silver-haired bats (Lasionycteris noctivagans) identified
from June – August 2021 and December – March (2021, 2022) in the Smallwood Creek area
near Beasley, British Columbia. The number of parameters (K), the Akaike’s information
Criterion corrected for small sample sizes (AICc) score, ΔAICc score, model weight (Weight)
and area under the receiver operating characteristic (ROC) score for all models are included. The
variables in the top model are: 3m canopy closure (%) and roost type (exfoliating bark or cavity).
Model Structure

K

AICc

ΔAICc

Weight

ROC

3m canopy closure (%) + Roost Type

4

27.52

--

0.91

0.92

Intercept-only (null)

2

55.94

28.42

0.00

0.50

48

Table 2.10. Parameter estimates, standard errors (SE), odds ratios, and 95% lower and upper
confidence intervals (CI) for odds ratios in the top logistic regression model for winter versus
summer tree roost use by silver-haired bats (Lasionycteris noctivagans) from December – March
(2021, 2022) and June – August 2021 in the Smallwood Creek area near Beasley, British
Columbia.
Variable

Estimate SE

Odds ratio

95% Lower CI

95% Upper CI

3m canopy closure (%)

-0.27

0.10

0.76

0.62

0.93

Roost type: exfoliating
bark

-3.53

1.33

0.03

0.00

0.39

49

2.3.4 Roost Re-Use
Of the three tree roosts where cameras were deployed from November 2021 to March
2022, I observed use at one roost the subsequent winter. Imagery showed silver-haired bats using
this tree on four separate occasions in winter of 2022 (Figure 2.9). It was not possible to identify
if this was the same individual across years due to camera resolution; however, I could identify it
as a silver-haired bat because of distinguishable characteristics (i.e., pelage colour and size).
Additionally, in April 2022, another study conducted in partnership with the Ministry of
Environment and Wildlife Conservation Society Canada tracked two silver-haired bats to tree
roosts used by radiotagged bats the previous winter (BC Ministry of Environment and Wildlife
Conservation Society Canada, unpublished data). Of the three trail cameras I installed at winter
tree roosts, there was activity at two roosts during summer. Images of bats visiting known tree
roosts were captured 17 times from 01 May – 30 September 2022. I identified bats as silverhaired on six occasions. Bats in the remaining 11 videos were unidentifiable to species because a
lack of visible defining features visible, or poor video quality.
None of the bats captured and fit with radio-transmitters in the summer were found using
previously identified winter roosts in summer months. Two male silver-haired bats tracked in the
winter of 2020 – 2021 were recaptured and tracked the following summer, suggesting year-round
site fidelity among males of this species, although the sample size is small. These males did not
use the same roosts in summer as they did in winter.

50

Figure 2.9. A still image of a trail camera video showing a silver-haired bat (Lasionycteris
noctivagans) visiting an identified winter tree roost in the Smallwood Creek area near Beasley,
British Columbia on 23 February 2022 (3 additional visits were documented on 24 January, 21
February, and 23 February 2022). This roost was located using radio telemetry on 26 January
2021. Identification of this as a silver-haired bat is possible because of its size and distinct
coloration.

51

2.4 Discussion
I described winter roost tree characteristics used by silver-haired bats and compared them
to roost trees used in the summer. Trees used by silver-haired bats in the winter differed from
those used in summer. Consistent with my prediction, in winter, silver-haired bats use cavities in
large-diameter, low-decay trees. Elsewhere, these features have been demonstrated to promote
thermally stable, warm internal temperatures (Hooge et al. 1999; Wiebe 2001; Parsons et al.
2003; Bär and Mayr 2020). Specific microclimate characteristics of tree roosts are addressed in
Chapter 3, where I found tree roosts to be buffered against ambient temperatures, but not as
warm or thermally stable compared to the mine or rock crevice roosts.
Bats used trees in earlier stages of decay in the winter compared to potentially available
trees and those used in the summer. Trees in lower stages of decay (i.e., live, declining trees, or
newly dead trees) typically have firm wood and more bark cover that provides more insulation
than highly decayed wood (Hooge et al. 1999; Bär and Mayr 2020). Decay class may be
especially important when considering cavity roosts because advantages from insulation are
more evident than bark roosts (Turbill and Geiser 2008). As expected, bats roosted more
frequently in cavity roosts in the winter compared to exfoliating bark roosts in the summer, as
seen in other tree-roosting bats (Boyles and Robbins 2006; Newman et al. 2021). However,
silver-haired bats also used exfoliating bark occasionally at this site in winter. Although this has
been observed with other bat species in winter, it typically occurs where winter climates are
warmer, and ambient temperatures rarely drop below 0°C (Perry et al. 2010), as exfoliating bark
is likely not as thermally buffered as cavities. It is possible that silver-haired bats used bark
roosts only on warmer days, when freezing was not a risk at this site, however, small sample
sizes limited my ability to conduct this analysis.

52

Winter roosts were in larger diameter trees compared to potentially available roosts.
Roosts in large-diameter trees may provide better insulation from cold ambient temperatures than
smaller trees, and cavities in large-diameter trees have thick walls, which contribute to more
stable microclimates (Wiebe 2001; Parsons et al. 2003). The use of large diameter trees in winter
compared to summer has been seen elsewhere (Boyles and Robbins 2006). Although tree
diameter influences microclimate, other hypotheses may also explain the selection of larger trees
in winter. Tree diameter showed strong collinearity with tree height, and as such, I can assume
that the large-diameter trees used by silver-haired bats in winter were also tall. Tall trees act as
visual markers on the landscape and facilitate orientation by bats (Campbell et al. 1996). Bats
may preferentially use roosts that are easy to locate and may travel shorter distances between
roosts in winter. Because flight is energetically expensive (Thomas and Suthers 1972) and
limiting movements may help reduce energetic costs. Plot characteristics of winter roosts further
support the hypothesis that bats chose roosts that are easy to locate. Winter roosts were in areas
of high snag density, and closer to potential roosts compared to potentially available trees. High
densities of snags, or roosts near one another might allow more access to alternative roosts to
support roost-switching (Mattson et al. 1996), meaning less energy expended during roostswitching events.
Winter tree roosts were in areas with lower canopy closure compared to summer roosts.
Although canopy closure may be affected by the lack of deciduous leaves in the winter, resulting
in lower canopy closure in winter, the forest at the site is primarily coniferous trees. As such, I
expected this effect to be minimal on seasonal differences in canopy closure. In winter, tricolored
bats (Perimyotis subflavus) use trees in areas with high canopy closure and reduced solar
exposure may keep roosts cool in an area with mild winter temperatures to promote winter torpor

53

(Newman et al. 2021). Because winter temperatures at this site are cold (< 0°C), the use of areas
with low canopy closure, and thus increased solar exposure may help keep tree roosts warm
enough to prevent freezing.
The species of trees most often used as roosts at the site were those that were also
relatively rare. Bats commonly used ponderosa pine and trembling aspen in winter, however,
both species are found at low density (1% and 6.8%, respectively). Bats may have chosen these
species preferentially because they inherently have high-quality roost attributes. Trembling aspen
are prone to internal decay, resulting in more useable cavities (Chambers et al. 1997; Martin et
al. 2004). Ponderosa pine trees have thick bark that loosens in large sheets, providing space for
sizable roosts (Bull et al. 1997). The thick bark of ponderosa pine also decreases wood drying
and increases the likelihood that trees break along the stem instead of at the base, allowing the
tree to remain standing and provide habitat (Bull et al. 1997). Decreased wood drying in
ponderosa pine may help increase humidity within roosts, which is important for hibernating bats
(Thomas and Geiser 1997). Decreased humidity leads to higher levels of evaporative water loss
in bats, resulting in increased arousal rates (Thomas and Geiser 1997). Bats that experience
lower rate of evaporative water loss while roosting in ponderosa pine may have reduced arousal
costs, leading to increased energy savings during winter (see Chapter 3 for more details).
Furthermore, ponderosa pine have a high percentage of sapwood, which decays readily (Hallett
et al. 2001). The relative speed at which these tree species decay may explain why bats chose
ponderosa pine and trembling aspen more often, compared to the more commonly observed
Douglas-fir, which decays more slowly (Hallett et al. 2001). Additionally, ponderosa pine, while
rare, are large at this site ( ̅ = 97.8 ± 53.5 cm DBH), and those used as roost trees typically had
deep heartwood cavities. These deep cavities would have been more insulated than smaller

54

counterparts, and the large trees may have been easy to locate by bats, increasing the likelihood
that they were used as winter roosts.
Winter tree roosts at my site showed notable differences to those used by bats in areas
with warmer winter temperatures, and this suggests that winter roost use is highly dependent on
local climate. I observed bats primarily use cavities, whereas the use of exfoliating bark in the
winter is more common in warmer areas (Turbill and Geiser 2008; Perry et al. 2010). Exfoliating
bark is likely not as well-insulated as cavities, but freezing risk may not be a concern in these
milder climates. Bats also preferentially roost on the south side of trees elsewhere (Hein et al.
2008; Perry et al. 2010), whereas I did not notice a trend in selection for specific aspects. In
warmer areas, bats benefit from thermally unstable roosts (i.e., exfoliating bark, areas with
intense afternoon sun) by using passive re-warming to arouse from torpor when conditions may
be favourable for foraging (Turbill and Geiser 2008; Chenery et al. 2022). Foraging is likely not
possible in the winter at my study site, and this may help explain observed differences.
As I predicted, silver-haired bats did not use well-insulated roosts in the summer. I
captured adult males (n = 6) and non-reproductive adult females (n = 2) in the summer, and they
do not have the thermoregulatory requirements of pregnant females (Racey and Swift 1981;
Zahn 1999) or hibernating bats (Webb et al. 1996). Thus, they may be more flexible in their roost
choices. I observed bats commonly using bark roosts in areas of high canopy closure, both
features which would result in cooler internal roost temperatures. Elsewhere, males and nonreproductive female bats use cooler, thermally unstable roosts (Monarchino et al. 2020)
compared to reproductive females, which would allow them to use daily torpor, and benefit from
passive re-warming, thus saving energy. Because I did not capture any reproductive females, it
was impossible to determine preferred roost characteristics. However, the lack of reproductive

55

females suggests that roost availability or climate conditions at this site were not conducive to
reproduction by silver-haired bats.
The distribution of roost trees throughout the study site varied between winter and
summer. In winter, roost trees were found in close proximity (< 500m) to the capture location
(i.e., the Queen Victoria Mine), whereas in summer, bats were found within 2km of their capture
location. Little is known about movement patterns of tree bats in winter, but where they have
been recorded, tree bats are typically relocated within 1.5km of capture location (Hein et al.
2008). Bats do not have the same energetic constraints in summer as they do winter and may
travel farther to find a suitable roost. The intensive use of a relatively small area by silver-haired
bats in winter highlights the importance of identifying and managing a focal winter roosting area
to support the needs of tree-roosting bats in winter. Although the differences in travel distances
observed can be explained biologically, biases may have been introduced through differences in
capture methodology between seasons. In winter, I captured bats using a net placed outside of
the Queen Victoria Mine, which means only individuals who frequent the mine were captured. It
is possible that silver-haired bats overwinter elsewhere at the study site, without frequenting the
mine. Further research should focus on conducting winter capture efforts elsewhere in the
Smallwood Creek area, or in other areas devoid of underground mines where silver-haired bats
may hibernate.
My results indicate some male silver-haired bats show fidelity to an area across seasons
and years. Banding records from this study indicate that male silver-haired bats show year-round
site fidelity, suggesting that males in this area may not be migratory. Silver-haired bats are
considered migratory throughout most of their range (Cryan 2003), and this provides evidence
that a sedentary population of males exists in British Columbia. Therefore, I can conclude that

56

trees provide critical winter habitat for silver-haired bats and remain essential over time. Winter
tree roosts used by silver-haired bats display features that were not common in this forest system
(i.e., large trees, rare species), and bats exhibited roost and site fidelity at these trees. These
results suggest that these roosts provide unique benefits, potentially beyond thermoregulatory
requirements, highlighting the importance of protecting these trees. This work is the first
descriptive study of winter tree roost use by silver-haired bats in British Columbia and provides
data that improves our understanding of winter habitat requirements of silver-haired bats.
However, these data represent a small area of the province, and further work needs to be
conducted to determine if silver-haired bats hibernate in trees elsewhere in British Columbia.

57

3. Microclimates and Torpor Patterns Inside Silver-Haired Bat (Lasionycteris noctivagans)
Hibernacula
3.1 Introduction
During winter in regions with cold climates (where temperatures drop below 0°C),
animals face energetic stressors, including limited food abundance and cold temperatures. To
mitigate these challenges, animals may migrate to escape harsh conditions, or hibernate to
conserve energy. Hibernation is a common strategy for small mammals living in temperate
regions (Geiser and Ruf 1995). During hibernation, mammals employ multi-day periods of
torpor, a physiological process where metabolic rate and body temperature (Tb) are depressed
(Geiser 2004). During hibernation, Tb typically falls to 0°C to 10°C and can closely track
ambient temperature (Speakman and Thomas 2003; Geiser 2004). Hibernators employ
heterothermy to save energy. Maintaining typical mammalian Tb requires significant energy, but
as heterothermic endotherms, bats do not have to defend a stable and elevated Tb constantly.
When endotherms self-regulate, body temperature is maintained at an elevated level by using
metabolism, despite fluctuations in ambient temperature (Geiser 2004). Conversely, during
heterothermy, individuals can reduce or increase metabolic rate so that body temperature more
closely conforms to ambient conditions (Geiser 2004).

Throughout hibernation, torpor is interrupted periodically by euthermic bouts, known as
arousals. The primary purpose of hibernal arousal is still debated. Arousals may be used to clear
metabolic waste, rehydrate, or to combat sleep deprivation or decreased immunocompetence
(Trachsel et al. 1991; Thomas and Cloutier 1992; Burton and Reichman 1999; Park et al. 2003;
Ben-Hamo et al. 2013). For bats, arousals may also be used for drinking, foraging, copulating, or

58

switching roosts (Speakman and Racey 1991; Thomas and Geiser 1997; Falxa 2007; Newman et
al. 2021). Arousals account for most of the winter energy expenditure (Thomas et al. 1990;
Geiser and Ruf 1995), and to conserve energy, bats employ extended torpor bouts that last more
than 30 days, with short arousals that last hours (Park et al. 2000; Geiser 2004; Jonasson and
Willis 2012). Energetics of torpor and arousal in bats are influenced by roost microclimates,
primarily temperature and humidity (Twente et al. 1985; Thomas and Geiser 1997; Geiser 2004).

Torpor bout duration (TBD) during winter is correlated with roost temperatures, with
TBD increasing as temperature decreases down to a minimum physiological threshold (Geiser
and Broome 1993; Geiser 2004). The amount of energy expended during torpor, torpid metabolic
rate, reaches a minimum at a set roost temperature (Tmin) but can increase rapidly when subject to
small increases or decreases in roost temperature, resulting in an increase in metabolic rate, or an
arousal (Twente et al. 1985; Dunbar 2007; Boyles and McKechnie 2010). The Tmin at which
torpid metabolic rate reaches a minimum varies with species and along a latitudinal gradient
(Dunbar and Brigham 2010; McGuire et al. 2021) but is typically from 2°C to 10°C across North
America (e.g., Dunbar 2007; Dunbar and Tomasi 2006; McGuire et al. 2021). Bats at southern
latitudes typically hibernate at a higher Tmin than their northern counterparts (Dunbar and
Brigham 2010).
Evaporative water loss (EWL), essential in regulating torpor, is influenced by roost
microclimate. Evaporative water loss is the sum of water lost through respiration and cutaneous
water loss, the latter of which is related to the difference in water vapour pressure at the skin's
surface and the air (Thomas and Cloutier 1992). During torpor, respiration is depressed, and as
such EWL is largely a result of cutaneous water loss (Hosken and Withers 1997), resulting in

59

EWL being lower during torpor compared to euthermic conditions (Webb et al. 1995; MuñozGarcia et al. 2012). Because of their large surface area to volume ratio, and non-furred wings,
bats experience high rates of EWL compared to other mammals of a similar size (Hosken and
Withers 1997). Metabolically produced water is thought to be insufficient for long-term
hydration in bats (Thomas and Cloutier 1992), meaning deep torpor is a period of negative water
balance for bats. Significant water loss causes arousal, where bats must drink to avoid
dehydration (Thomas and Geiser 1997). Evaporative water loss during torpor decreases with
increasing roost humidity, with near saturation (100% relative humidity; RH) resulting in the
lowest rates of EWL (Thomas and Cloutier 1992). Low rates of EWL have been correlated with
prolonged TBD, reduced arousals during hibernation, and increased overwinter survival (Thomas
and Geiser 1997; Ben-Hamo et al. 2013). Interspecific and intraspecific variation in sensitivity to
EWL during hibernation exists, with certain species or populations more tolerant to dry
conditions (Klüg-Baerwald et al. 2017; McGuire et al. 2021).
Roost use is an important mechanism to optimize winter energy budgets and minimize
water loss. Using one or more appropriate winter roosts (hibernacula) that meet an individual’s
physiological needs during hibernation is essential. Bats may remain within a roost for the entire
winter, or periodically switch between roosts to exploit differences in microclimate depending on
physiological needs (Boyles et al. 2017; Newman et al. 2021). Energy expenditures can be
reduced by using hibernacula with cool, above-freezing internal temperatures (approaching Tmin)
and high humidity (~100%; Thomas and Cloutier 1992; Webb et al. 1996). These conditions are
ideal for promoting prolonged, deep torpor, which may be beneficial in northern climates, where
winters are long and energy conservation is essential (e.g., Clark et al. 1997; Humphries et al.
2003). There is, however, a trade-off between the physiological costs and the energetic benefits

60

gained from deep torpor (Humphries et al. 2003). Bats need to balance sufficient fat reserves to
survive the winter while minimizing the adverse effects of torpor, such as dehydration, reduced
immunocompetence and metabolic waste build up (Humphries et al. 2003). These trade-offs can
be managed by using surplus fat reserves to reduce the depth and duration of torpor or increasing
torpor use if energy reserves are limited or needed for reproduction in the spring (Humphries et
al. 2003). Bats may employ different roosting strategies throughout the winter depending on
energy reserves (Wermundsen and Siivonen 2010; Ryan et al. 2019).
Most described hibernacula in Canada are in caves or underground mines with stable
temperatures above freezing and near-saturation humidity (Webb et al. 1996; Weller et al. 2018).
Although these types of hibernacula are the most common, bats can hibernate in a variety of
other structures, including rock crevices, mudstone erosion holes, buildings, bridges, and trees
(Lausen and Barclay 2006; Turbill and Geiser 2008; Halsall et al. 2012; Newman et al. 2021).
Variability in hibernaculum preference exists among species, and along a latitudinal gradient.
Bats at more southerly latitudes (i.e., the southern United States) are not constrained by freezing
temperatures and may supplement fat reserves with occasional foraging (Stawski and Currie
2016). In these warm locations, bats commonly overwinter in trees, where they employ shorter,
shallower torpor bouts, and switch roosts more frequently than bats hibernating at northern
latitudes. Microclimates inside winter tree roosts are relatively unstable and track ambient
temperatures closely (Turbill and Geiser 2008; Newman et al. 2021), and because of this, their
use as hibernacula is uncommon in cold climates.
In eastern North America, some species of bats hibernate in large, conspicuous groups,
making the identification of hibernacula relatively easy (Weller et al. 2018). Few hibernacula
have been discovered in British Columbia, or other areas of western North America (Nagorsen et

61

al. 1993; Weller et al. 2018). Where hibernating bats have been found in the west, they are in
smaller, inconspicuous groups or individually, and this contributes to our lack of understanding
of hibernation strategies in the west (Weller et al. 2018). In British Columbia, hibernacula have
been found in caves, mines, and rock crevices (Nagorsen et al. 1993; Lausen et al. 2022). The
unstable internal microclimates of trees suggests that they may not be suitable as hibernacula in
cold regions; however, the use of trees in winter has been observed in British Columbia
(Nagorsen et al. 1993; Burles 2014; Lausen et al. 2022) and may be restricted to areas of the
province with mild winter seasons (Burles 2014).
One location where bats use trees in winter is the Smallwood Creek area in the southern
British Columbia interior. This area has an abandoned, underground mine, which is known to
host three hibernating bat species: the Townsend’s big-eared bat (Corynorhinus townsendii), the
California myotis (Myotis californicus), and the silver-haired bat (Lasionycteris noctivagans).
Previous research found that silver-haired bats hibernate in trees adjacent to the mine and may
alternate among the mine, trees, and rock crevices throughout winter (Lausen et al. 2022). Most
research on winter tree-roost use by bats has been conducted in the southern United States and
Australia (e.g., Dunbar and Tomasi 2006; Turbill and Geiser 2008; Newman et al. 2021; Jorge et
al. 2021), and hibernation in trees by bats at northern latitudes is understudied. Therefore, it is
unclear what types of microclimates exist in these features and what types of torpor patterns treeroosting bats exhibit in the winter in colder regions. This represents a significant gap in our
knowledge on the winter ecology of temperate-zone bats. Understanding torpor patterns and
microclimate conditions inside winter roosts used by silver-haired bats will better define winter
habitat needs of bats in the province and inform forestry practices to support bat conservation.

62

I investigated (1) microclimates inside silver-haired bat hibernacula and determined if
bats experience different microclimates across a range of hibernaculum types (mine, trees and
rock crevice roosts). I also aimed to determine (2) how silver-haired bats use torpor in the winter
and if torpor patterns differ among roost types. Because of differing insulative properties among
rocks, tree roosts, and mine features, I hypothesized that microclimates vary among roost types
but are buffered against ambient temperatures. Specifically, I predicted:
a) Hibernacula microclimates would be above 0°C and had low vapour pressure deficit
(VPD), a measure of moisture in the air.
b) The mine would be thermally stable and had low VPD. Rock crevices and tree cavities
would be less thermally stable and have higher VPD compared to the mine.
c) Bats would use the most thermally buffered roosts (i.e., the mine) on colder days, and use
less thermally buffered roosts (i.e., rock crevices and tree cavities) during warmer periods
and switch among roost types as temperatures change. Additionally, bats would primarily
use low VPD roosts during periods of low ambient humidity.
Because of the predicted differences in internal microclimates among roost types, I
hypothesized that torpor, arousal, and movement patterns would differ among roost types.
Specifically, I predicted:
a) If bats exhibit different torpor patterns among roost types, bats would longer torpor bouts
(and decreased arousals) in the mine compared to rock crevice and tree roosts, based on
microclimate predictions described above.
b) Furthermore, bats would switch roosts when ambient weather conditions change.
Specifically, when ambient temperatures cool, bats move to the more insulated mine to

63

prevent from freezing. Furthermore, when ambient humidity decreases, bats move from
less humid roosts (i.e., trees and rock crevices) to more humid roosts (i.e., mine).
3.2 Methods
3.2.1 Study Site
The study area is the Smallwood Creek drainage above Beasley, British Columbia.
(49.49350, -117.44960). This area is within the Interior Cedar-Hemlock Biogeoclimatic zone
(Pojar et al. 1987). The dominant tree species are Douglas-fir (Pseudotsuga menziesii), western
redcedar (Thuja plicata), western hemlock (Tsuga heterophylla), and ponderosa pine. The Queen
Victoria Mine, an underground copper-silver-gold mine used until 1956 is a focal bat roost
feature in this area. In winter, it is a hibernaculum for the silver-haired bat, Townsend's big-eared
bat and the California myotis. From 2012 – 2018, radio telemetry work conducted here identified
eight trees used as winter roosts by silver-haired bats, where individuals used space under
exfoliating bark and in cavities throughout the winter (Cori L. Lausen, [Wildlife Conservation
Society Canada, Kaslo, British Columbia], personal communication, [December 2020]). In
summer, the mine serves mainly as a night roost and social gathering location for these same
species in addition to at least five other species (big brown bat, Eptesicus fuscus; Yuma myotis,
M. yumanensis; long-eared myotis, M. evotis; long-legged myotis, M. volans; and little brown
myotis, M. lucifugus). The mine consists of an underground opening with several rooms that
typically remain above 0°C in the winter and provide a suitable selection of hibernation
locations. In the summer, bats frequent the mine and roost in the adjacent forested area (Cori L.
Lausen, [Wildlife Conservation Society Canada, Kaslo, British Columbia], personal
communication, [December 2020]). This site is one of only two identified silver-haired bat mine

64

hibernacula in western North America, both in the West Kootenay region of British Columbia
(Lausen et al. 2022). It is also the only silver-haired bat mine hibernaculum in the province that
is safe to access, and thus it is an ideal research location. Despite this site being one of three
known mine hibernacula in British Columbia, timber harvesting occurs within the habitat
surrounding the Queen Victoria Mine. Logging practices can severely reduce roosting habitat
through direct tree removal and subsequent landscape modification (Russo et al. 2010; Borkin et
al. 2011).
3.2.2 Data Collection
I captured free-flying silver-haired bats using mist nets (Avinet Research Supplies,
Maine, USA) for two winter seasons (2021, 2022), and used additional data from bats captured
in 2012, 2014, and 2018, for a total of five years of data collection. Staff members of the
Wildlife Conservation Society Canada collected data for 2012, 2014 and 2018. I defined winter
as the time period from 21 December (winter solstice) to 21 March (spring equinox). I placed
one 12m double-high net across the Queen Victoria Mine entrance to capture bats as they entered
and exited the mine. In the winter, bats can reliably be captured when exiting and entering the
mine during mild (> -7°C) winter nights, and no other nets were deployed.
I affixed bats with 0.42g temperature-sensitive radio-transmitters (LB–2NT; Holohil
Systems Ltd., Ontario, Canada) by trimming the fur in the interscapular region and attaching
transmitters using a nontoxic latex glue (Osto-bond; Montreal Ostomy, Quebec, Canada). The
combined mass of the transmitter and glue was less than or equal to 5% of the individual's body
mass, as per standards (Aldridge and Brigham 1988). To ensure transmitter attachment, I held
bats in cloth bags for 15 minutes and then released them at the capture site. I conducted all

65

animal handling with the approval of the UNBC Animal Care Committee (Protocol 2020–17)
and under provincial bat capture permit MRCB20-598305.
I used a hand-held radio telemetry receiver (R–1000 Telemetry System, Communication
Specialists Inc., California, USA) to locate individuals in their day roosts the morning after
capture. The exact position of the bat within the tree was determined by scanning the tree at close
range with the receiver or visually observing the transmitter antenna aerial, if it protruded from
the roost, with binoculars. I tracked bats daily for the duration of the life of the transmitter, which
is approximately 45 days, or until they were undetectable for more than seven days.
Once I located the exact location of a roost, I installed temperature and relative humidity
data loggers (HOBO U2302A; Onset Computer Corporation, Massachusetts, USA) into roosts to
record microclimates every 30 minutes. When possible, I installed data loggers next to the bat
while it was roosting, otherwise, I installed loggers after the bat had vacated the roost. I placed a
microstation (HOBO H21–USB; Onset Computer Corporation, Massachusetts, USA) in a central
location and used it to record ambient temperature, humidity, and atmospheric pressure every 30
minutes for the duration of the study. However, in 2014 and 2018, there was no microstation at
the study site, so I used publicly available weather data (temperature, humidity and atmospheric
pressure) from a weather station located 10km away from the study site to estimate ambient
conditions (Environment Canada 2023). To determine that data were comparable, I compared
Environment Canada data from 2022 with data collected at the study site from the same year.
Temperature, relative humidity, and atmospheric pressure were highly correlated between
stations (Pearson’s r = 0.9) and varied little.
At all active roosts (i.e., occupied by a radiotagged bat), I installed a data-logging
receiver (SRX400, SRX600, SRX800D or SRX1200; Lotek Wireless Inc., Ontario, Canada)

66

outside of the roost with a three- or five- element Yagi antenna. I used data-logging receivers to
record bat movement (i.e., time the bat exited roost, the length of roost occupancy), and torpor
information. Data-logging receivers recorded the transmitter pulse interval (later converted to
skin temperature) every 10 minutes for the duration of the bats’ occupancy, or until the
transmitter failed. Occasionally, data quality was poor, and pulse interval was recorded once per
hour. Length of roost occupancy was determined by calculating time of continuous recording of
bat transmitter, and a bat was assumed to have left a roost when the signal was undetectable. I
moved data-logging receivers when bats moved, and new roosts were identified. I installed a
data-logging receiver inside the mine for the duration of the study because I expected frequent
use by bats throughout the winter.
3.2.3 Data Analysis
I used R Studio version 1.3.1093 (R Development Core Team 2020) for all statistical
analyses, and reported specific R packages used throughout. Variables are reported as mean ±
standard deviation (SD). I compared internal microclimates (temperature and humidity) among
the three types of hibernacula used by silver-haired bats: mine, rock crevices and tree cavities.
Relative humidity has been identified as ineffective at determining evaporative water loss rates
in bats (Kurta 2014), so I used vapour pressure deficit (VPD), a measure of absolute moisture,
that has been used elsewhere when comparing roost types used by bats in the winter (Newman et
al. 2021). It is calculated using saturated water vapour pressure, which requires simultaneous
measurements of temperature and relative humidity (Brice and Hall 2016), using the following
equation: VPD = [(100-RH)/100]*SVP, where SVP (saturated vapour pressure) =
[610.7*107.5T(237.3+T)]/1000, RH = relative humidity (%), and T = temperature (°C). I used
microclimate measurements from dates only on which I had a complete winter of data recorded

67

from all roosts simultaneously (20 December 2021 – 29 March 2022 and 15 January 2014 – 05
March 2014).
Because daytime and nighttime temperatures vary, I split data into daytime and nighttime
periods, delineating the two periods using local civil sunset and sunrise times. Prior to modelling,
I confirmed that the data met assumptions of normality by examining quantile-quantile (Q – Q)
and residual versus fitted value plots, and when necessary, applied an appropriate transformation
to normalize the data. To determine if internal temperature or VPD differed among roost types, I
used analysis of variance (ANOVA, R Package: ‘lme4’, Bates et al. 2015) to compare daily
mean daytime and daily mean nighttime among roost types with a random effect of individual
roost identity to account for repeated measures. To determine the thermal buffering capacity of
each roost type, I used ANOVA to test the effect of roost type on mean roost-ambient difference
(RAD) in temperature. Roost ambient differences were calculated by subtracting the roost
temperature from the associated ambient (outdoor) temperature. To determine if bats used
different roost types depending on ambient (outdoor) weather, I used one-way ANOVA to test
for differences in ambient temperature on days bats used the three roost types with a random
effect of individual to account for repeated measures. I repeated all above analyses on VPD,
using inverse-transformed daily mean daytime VPD, square-root transformed daily mean
nighttime VPD, and square-root transformed RAD VPD. For post-hoc analysis, I used Tukey’s
Honest Significant Difference (HSD) test to assess the differences between pairs of group means.
To calculate the skin temperature (Tsk) of radiotagged bats, I used polynomial equations
(calibration curves), obtained from the transmitter manufacturers for each transmitter to convert
pulse interval to Tsk. I defined torpor entry as a decrease in Tsk below normothermic levels (Tsk <
25°C), and arousals as a sharp increase in Tsk to normothermic levels in under 30 minutes (Tsk >

68

25°C). I used Tsk > 25°C instead of the standard 28°C (i.e., Turbill and Geiser 2008; Chenerey et
al. 2022) because I recorded bats flying when Tsk was 25°C. I calculated torpor and arousal
duration only when complete data were available (i.e., no data gaps before or after torpor entry
and exit). To determine if time of arousals in a 24-hour period differed from random, I used
Rayleigh's (Z) circular statistics (Zar 1999, R package ‘circular’, Agostinelli and Lund 2022.).
The Rayleigh’s test is used to determine if the distribution of circular data, such as a 24-hour
time period, differs significantly from random (Zar 1999).
Because of the age of the receivers used and suspected problems with transmitter
calibration, I did not conduct any analysis on Tsk or body temperature and only used Tsk to
determine the start and end of torpor bouts (see Appendix A for more information). To determine
if torpor bout length was dependent on roost type, I used an analysis of covariance (ANCOVA)
to test the effect of roost type on torpor bout duration (square-root transformed to normalize the
data). I used day of winter (where 01 December = Day 1) that the torpor bout ended as a
covariate, because bats typically use longer torpor bouts later in the winter (French 1985). I
included a random effect of individual to account for repeated measurements, however, the
variance was indistinguishable from zero, so this effect was removed.
Changes in weather (i.e., incoming snowstorms or cold, high-pressure Arctic outflow
systems) are often associated with rapid changes in ambient temperature, humidity, and
barometric pressure. To determine if changes in weather predicted whether bats aroused or not, I
used logistic regression analysis on 24-hour changes in ambient temperature (ΔT), VPD (ΔVPD),
and pressure (ΔP) (R package: ‘lme4’, Bates et al. 2015). I used absolute values for all climate
variables to account for the possibility of both positive and negative changes in weather. I
included a random effect of individual to account for repeated measures. The variance of the

69

random effect was indistinguishable from zero, so it was removed from the final model. I tested
for collinearity among variables using a correlation matrix and did not include variables if they
were highly correlated (Pearson’s r ≥ 0.7). There was no significant correlation among variables,
so none were removed from the model. I used a likelihood ratio test (χ2) to assess the fit of the
model to the data and used the area under the receiver operating characteristic curve (ROC;
Hosmer and Lemeshow 2000, R package ‘ROCR’, Sing et al. 2015) to examine the model
accuracy. This method assesses the discrimination capacity of a given model (Pearce and Ferrier
2000) and is commonly used with logistic regression (Fielding and Bell 1997). I considered the
performance of the model with ROC > 0.7 as good and ROC > 0.9 as excellent.
3.3 Results
3.3.1 Roost Microclimates
I monitored the internal microclimate at nine roosts (three rock crevices, five trees, one
mine) in the winters of 2014 and 2022. Microclimates were not analyzed in 2012, 2018 or 2021
because of incomplete data collection. The mean ambient outside temperature and humidity
varied among years, but all monitoring years experienced periods of cold weather (< -7°C; Table
3.11).

70

Table 3.11. Mean ± standard deviation (SD) and range (minimum – maximum) of ambient
(outside) temperature (°C) and relative humidity (%) recorded during winter from 20 December
– 29 March across five years (2012, 2014, 2018, 2021, 2022) at the Smallwood Creek area near
Beasley, British Columbia.
Temperature (°C)

Humidity (%)

Year
2012

Mean ± SD
-1.2 ± 3.4

Range
-11.8 – 8.9

Mean ± SD
90.9 ± 9.9

Range
46.1 – 100

2014

0.1 ± 4.6

-14.8 – 13.2

79.4 ± 15.4

27 – 97

2018

-0.5 ± 3.6

-12.9 – 9.7

85.4 ± 13.5

31 – 100

2021

-1.4 ± 3.4

-15.1 – 6.9

88.9 ± 10.6

38.7 – 100

2022

-2.3 ± 3.7

-17.1 – 13.1

88.4 ± 11.2

37.2 – 100

71

Mean daytime roost temperatures differed significantly among the three roost types (F2,8
= 25.7, P < 0.001; Figure 3.1). Post hoc ANOVA analysis showed that mean daytime roost
temperatures in the mine were significantly warmer than tree roosts (P < 0.001) and rock
crevices (P < 0.001), and rock crevice roosts were significantly warmer than tree roosts (P <
0.01). Similarly, nighttime mean roost temperatures differed significantly among the three roost
types (F2,8 = 20.1, P < 0.001; Figure 3.1). Post hoc ANOVA analysis showed that mean nighttime
temperatures in the mine were significantly warmer than rock crevice (P < 0.001) and tree roosts
(P < 0.0001), and rock crevice roosts were significantly warmer than tree roosts (P < 0.01).
Both rock and tree roost temperatures closely tracked ambient temperatures, with no
relationship between mine and ambient temperatures (Figure 3.2), which is reflected in the
significant difference in RAD temperatures observed among the three roost types (F2,8 = 8.8, P <
0.01). The RAD temperatures were significantly lower in tree roosts compared to rock crevice
roosts (P < 0.05) and the mine (P < 0.001), meaning trees provide less stable environments.
There was no significant difference in RAD temperatures between rock crevice roosts and the
mine (P = 0.06).
Rock crevice and tree roosts had internal temperatures below freezing for extended
periods (Figure 3.2). When occupied, tree roost temperatures were below 0°C for 50.0% of the
time, whereas rock crevice roosts were below 0°C for 16.5% of the time. The lowest temperature
recorded inside a roost while occupied was -9°C in a tree roost (Figure 3.3), and this bat
remained in the tree roost with internal temperatures at -9°C for 12 hours. Additionally, I
observed torpor use when roost temperatures were below 0°C (Figure 3.3).

72

Figure 3.1. Mean daytime temperatures (°C; left) and nighttime temperatures (°C; right) in mine
(n = 1), rock (n = 3) and tree (n =5) hibernacula used by silver-haired bats (Lasionycteris
noctivagans) from 20 December 2021– 29 March 2022 and 15 January 2014 – 05 March 2014 at
the Smallwood Creek area near Beasley, British Columbia. The dark bar in the center of the
boxplot indicates the median value of the data, the grey box limits represent the 75 th (upper) and
25th (lower) quartiles, and the whiskers represent minimum and maximum values.

73

Figure 3.2. Mean daily temperatures (°C) ± 1 standard deviation (shading) recorded inside mine
(n = 1), rock (n = 2), and tree (n = 5) hibernacula used by silver-haired bats (Lasionycteris
noctivagans), and ambient temperature (outdoor) recorded from 20 December 2021 – 29 March
2022 in the Smallwood Creek area near Beasley, British Columbia.

74

Figure 3.3. Ambient (Ta), roost (Tr) and skin (Tsk) temperatures of two different silver-haired
bats (Lasionycteris noctivagans) roosting in two different tree roosts from January 2021–
February 2021, near Beasley, British Columbia. Sharp increases in Tsk indicate arousals.
Measurements on the two animals occur on different days.

75

Mean daytime VPD was significantly higher in the mine compared to tree roosts
(Tukey’s HSD P < 0.05), but not rock crevice roosts (Tukey’s HSD P = 0.09; Figure 3.4). No
difference was observed in mean daytime VPD in rock crevice roosts compared to tree roosts
(Tukey’s HSD P = 0.75; Figure 3.4). Mean nighttime VPD in the mine was higher compared to
tree roosts (Tukey’s HSD P < 0.05), and rock crevices (Tukey’s HSD P < 0.05; Fig 7). There
was no difference in mean nighttime VPD in tree roosts compared to the rock crevice roosts
(Tukey’s HSD P = 0.85). These VPD results are supported by the relative humidity differences
observed among the roost types. The mine had lower humidity on average compared to tree and
rock crevice roosts (Figure 3.5).
I used telemetry data from 31 bats and ambient weather conditions from five winters
(2012, 2014, 2018, 2021 and 2022; Table 3.11) to determine whether roost type use depended on
ambient temperature and VPD. Mean daily ambient temperatures differed among days bats used
the mine, trees and rock crevice roosts (F2,227 = 50.01, P < 0.0001). Mean ambient temperatures
were colder on days bats used the mine compared to rock roosts (P < 0.001) and tree roosts (P <
0.001). Mean ambient temperatures were not colder on days bats used rock roosts than tree
roosts (P = 0.87). Mean ambient VPD differed significantly among days bats used the mine, rock
and tree roosts (F2,654 = 7.89, P < 0.001). Specifically, ambient VPD was higher on days bats
used trees compared to the mine (P < 0.001), but not between the other roost types.

76

Figure 3.4. Mean daytime vapour pressure deficit (VPD; hPa; left) and nighttime vapour
pressure deficit (hPa; right) in recorded hibernacula used by silver-haired bats (Lasionycteris
noctivagans; mine (n = 1), rock (n = 3) and tree (n = 5) recorded from 20 December 2021 – 29
March 2022 and 15 January 2014 – 05 March 2014 at the Smallwood Creek area near Beasley
British Columbia. The dark bar in the center of the boxplot indicates the median value of the
data, the grey box limits represent the 75th (upper) and 25th (lower) quartiles, and the whiskers
represent minimum and maximum values.

77

Figure 3.5. Mean daily internal roost humidity (RH, %) ± 1 standard deviation (shading)
recorded in mine (n = 1), rock (n = 2) and tree (n = 3) hibernacula used by silver-haired bats
(Lasionycteris noctivagans) and ambient (outdoor) humidity recorded from 20 December 2021–
29 March 2022 in the Smallwood Creek area near Beasley, British Columbia.

78

3.3.2 Torpor and Arousal Patterns
Skin temperatures were measured for 29 silver-haired bats across five winters for a total
of 96 torpor bouts and 558.2 monitoring days. Bats used torpor every day of the study, but torpor
bout duration varied widely, with bats using short (< 24h) and multi-day torpor bouts. Mean
torpor bout duration was 5.8 ± 5.0 days, with no significant difference in duration among roost
types (F1,2 = 0.60, P = 0.55). The longest torpor bouts recorded were 23.0 days in the mine, 18.3
days in a tree, and 18.8 days in a rock crevice roost. Ninety-five arousal events were recorded
during the monitoring period. During arousal, bats remained normothermic for 0.89h ± 0.77h
(range: 0.02h – 4.2h).
Time of arousal differed significantly from random (Z = 0.58, P < 0.001), with most
arousals occurring 0h to 4h after sunset (67.4%, n = 64), and few occurring midday (6.3%, n =
6). Roost switching was associated with 43.2% of all arousals (n = 41). Forty (42.1%) arousals
did not involve any type of movement, and 14 (14.7%) involved flight, but not roost switching.
Bats roosting in trees were occasionally passively re-warmed (exogenous heating preceding
arousals) as roost temperatures rose, but this did not occur in the mine or rock crevices (Figure
3.6). Where roost temperatures and Tsk were recorded simultaneously, of all arousals that
occurred in tree roosts, passive re-warming was associated with three (37.5%) arousals.
In the logistic regression model used to predict if weather variables affect arousals, only
changes in 24-hour absolute (both positive and negative) ambient temperature (ΔT) affected
arousals. Specifically, bats were more likely to arouse with increasing values of ΔT, but not with
increasing values of ΔVPD or ΔP (Table 3.12). The data were a good fit for the model (P <
0.05), and the model had good predictive probability (ROC > 0.7).

79

A

B

C

Figure 3.6. Ambient (Ta), roost (Tr) and skin (Tsk) temperatures of three different silver-haired
bats (Lasionycteris noctivagans) roosting in the mine (A), a rock crevice roost (B) and a tree
roost (C) from 31 January 2021– 08 March 2021 in the Smallwood Creek area near Beasley,
British Columbia. Sharp increases in T sk indicate arousals. Measurements of A, B, and C occur
on different days.

80

Table 3.12. Parameter estimates, standard errors (SE) and 95% lower and upper confidence
intervals (CI) in the logistic regression model for arousal likelihoods determined by changes in
weather variables for silver-haired bats (Lasionycteris noctivagans) using rock, tree and mine
roosts in Beasley, British Columbia from December – March over five years (2012, 2014, 2018,
2021, 2022). Odds ratios with confidence intervals not overlapping one are considered drivers of
the model. Variables are: ΔP, absolute value of 24-hour change in ambient pressure (hPa); ΔT,
absolute value of 24-hour change in ambient temperature (°C); ΔVPD absolute value of 24-hour
change in ambient vapour pressure deficit.
Variable

Estimate

SE

Odds ratio

95% Lower CI

95% Upper CI

ΔP

-0.03

0.03

1.00

0.91

1.02

ΔT

0.25

0.08

1.28

1.10

1.49

ΔVPD

-0.17

0.24

0.84

0.51

1.32

81

Arousals resulting in roost switching in the mine typically occurred during positive
changes in temperature in the preceding 24-hour period (73.3%, n = 11), and arousals resulting in
roost switching from a tree to an alternate feature occurred during positive changes in VPD
(68.2%, n = 15; Figure 3.7).

Figure 3.7. Frequency of arousals resulting in a roost switching event during either positive (+)
or negative (-) 24-hour changes in pressure (hPa), temperature (°C), or vapour pressure deficit
(VPD) occurring from the three winter roost types (mine, rock and tree) to another roost by
silver-haired bats (Lasionycteris noctivagans) in the Smallwood Creek area near Beasley, British
Columbia from December – March of five years (2012, 2014, 2018, 2021, 2022).

82

3.4 Discussion
The use of tree and rock hibernacula by bats has not been well studied in Canada (Lausen
and Barclay 2006; Perry et al. 2010; Klüg-Baerwald et al. 2017), and this study is the first to
investigate microclimates and torpor patterns of tree-hibernating bats in this area. In addition to
hibernating in the mine and rock crevices, bats used multi-day bouts of torpor in trees, lasting up
to 18 days. Microclimates differed among roost types, however, torpor patterns varied less than
expected.
As hypothesized, tree roosts, and to a lesser extent, rock crevices were colder and had
more variable internal temperatures than inside the adjacent mine. Tree roosts had limited
buffering capacity relative to ambient temperature, suggesting they are poor insulators during
cold weather. This is consistent with what has been reported elsewhere, where rock crevices and
trees are less insulated than cave hibernacula (Webb et al. 1996; Brack 2007; Turbill and Geiser
2008; Klüg-Baerwald et al. 2017). Contrary to my prediction, the mine had higher VPD
compared to tree and rock crevice roosts. Higher rates of VPD are correlated with increased
EWL (Thomas and Cloutier 1992; Kurta 2014), meaning bats would experience higher rates of
EWL and dehydration while in the mine compared to alternative features. Where mine and rock
crevice hibernacula have been compared elsewhere, mines were found to be more humid (KlügBaerwald et al. 2017). The Queen Victoria Mine has large, above ground openings that allow for
increased air flow, and this may have resulted in the lower humidity here compared to mines or
caves elsewhere.
Silver-haired bats at this site used tree and rock crevices, which were thermally unstable,
and frequently had internal temperatures below 0°C. Occasionally, bats withstood cold (-9.0°C)
temperatures inside the roost, and one occasion, a bat remained in a tree roost with internal

83

temperatures at - 9°C for 12 hours. During this time, ambient (outdoor) temperatures were below
-15°C (Figure 3.6). Where bats are found overwintering in trees elsewhere, winter temperatures
are typically mild, and they use thermally unstable roosts (e.g., Boyles and Robbins 2006; Jorge
et al. 2021; Newman et al. 2021), because freezing is not a risk. Here, I have shown that this also
occurs in colder climates. Most hibernators avoid sub-freezing roost temperatures (Geiser 2004),
although the use of colder hibernacula can occur (Barnes 1989; Reimer 2014; Klüg-Baerwald et
al. 2017). I observed torpor use when roost temperatures were below 0°C (Figure 3.6),
suggesting that silver-haired bats may be more robust to cold roost temperatures during torpor or
may not be as energetically constrained during hibernation as expected. Because of difficulties in
accurately assessing Tsk at low ambient temperatures, I was unable to determine the depth of
torpor, and further research should be conducted to investigate physiological strategies employed
by these bats.
Additionally, silver-haired bats roosting in trees experienced large daily increases in roost
temperature, and I observed passive rewarming resulting in arousals in tree roosts. In treeroosting bats, passive rewarming, and subsequent arousal because of the daily warming cycle is
common (Twente et al. 1985; Currie et al. 2015). Elsewhere where bats overwinter in trees and
buildings, passive rewarming is used to save energy on arousals, when foraging may be possible
(Dunbar and Tomasi 2006; Turbill and Geiser 2008; Halsall et al. 2012). Passive rewarming can
benefit bats by an energy savings of 20% to 47% during arousal and reduced metabolic stress
(Halsall et al. 2012). Because caves and mines are typically well-buffered against ambient
temperatures, bats are unable to benefit from passive rewarming inside these roosts. The energy
savings obtained through passive rewarming in trees may reduce the energetic costs associated
with hibernation and contribute to overwinter survival (Halsall et al. 2012). This may be

84

especially important for silver-haired bats in my study because bats face long periods with low
insect availability compared to tree roosting bats in southern regions in winter.
Although silver-haired bats tolerate cold roost temperatures, they used warmer roost
types on colder days and humid roost types on drier days. This indicates there may be a trade-off
between roost humidity and temperature, with many, but not all bats using multiple hibernacula
to optimize energetic benefits in the context of ambient conditions. Some species of hibernating
bats require near-saturation roost humidity, typically avoiding dry conditions (Boyles et al.
2022), to prevent evaporative water loss (Thomas and Cloutier 1992). Stable temperatures and
high humidity increase torpor duration and reduce arousals (Geiser and Broome 1993; Geiser
2004; Ben-Hamo et al. 2013), however, bats are unable to satisfy both conditions in a single
location at this site. A lower temperature threshold may exist in tree roosts, below which bats are
unable to tolerate, resulting in a switch to a warmer, but drier roost. Conversely, higher vapour
pressure deficit in a warmer roost may result in arousals, and movement to a more humid roost.
Roost switching throughout winter based on ambient conditions is common in tree-roosting bats
to exploit microclimate conditions (Mormann and Robbins 2007; Newman et al. 2021). The use
of tree and rock roosts might increase the risk of freezing during extreme or extended cold
periods, and such an event was documented at this site when the entrance to an occupied rock
crevice roost iced over, and the radio-tracked bat succumbed to exposure (Lausen et al. 2022).
While the mine provides bats with a permanent and predictable refuge, where freezing is not a
risk, not all silver-haired bats used this roost during freezing conditions, suggesting that other
factors may be contributing to winter roost use beyond microclimates.
Roost-switching in response to weather conditions was weakly supported by torpor and
arousal patterns. Bats were more likely to arouse with greater changes in ambient temperature.

85

Furthermore, positive changes in ambient temperature were associated with roost-switching
events that resulted in bats leaving the mine, meaning bats may use warming outdoor
temperatures as an indicator to switch to a less thermally stable roost (i.e., rock or tree). Changes
in ambient VPD and pressure did not influence likelihood of arousal, however, most arousals
resulting in bats leaving tree roosts occurred during positive changes in ambient VPD. This result
was unexpected, as bats typically roosted in trees on the driest days, and a positive change in
VPD would indicate drier weather. In British Columbia, cold weather is typically a result of a
dry, arctic outflow systems. This roost switching response could be a result of extreme cold
temperatures causing bats to abandon trees in favour of a well insulated roost.
Contrary to my prediction, changes in barometric pressure did not influence the
likelihood of arousal, or the likelihood of roost switching, despite changes in barometric pressure
being associated with increased bat activity during summer (Paige 1995; Turbill 2008). Changes
in barometric activity can influence prey availability, and it is thought that bats use increasing
barometric pressure an indicator of suitable foraging conditions (Paige 1995). Because it is
unlikely that silver-haired bats are foraging in winter here, this could help explain the lack of
relationship observed between barometric pressure and arousal likelihood. Additionally, changes
in weather patterns can occur on a variety of time scales, and my analysis may not have reflected
the time scale that bats use to respond to weather. Roost microclimates are also likely to respond
to these changes differently depending on their insulation properties (Turbill and Geiser 2008;
Perry 2013). Because of small sample sizes, I was unable to analyze roost microclimates as
predictors of arousals, and it is possible that changing microclimate conditions would have been
better predictors of arousals and movement than ambient climate.

86

Contrary to my prediction, I did not observe differences in torpor bout duration (TBD)
among roost types, and bats exhibited different torpor patterns compared to other cave and treehibernating bats. The maximum TBD I recorded was lower than for other cave-hibernating bats
in Canada, where TBD can last up to 48 days (Jonasson and Willis 2012; Czenze and Willis
2015; Czenze et al. 2017). Additionally, maximum TBD in trees was longer than for other treehibernating bats, where torpor bouts infrequently exceed 10 days (Stawski et al. 2009; Clare and
Currie 2016; Chenery et al. 2022). Relatively short TBD and frequent roost switching suggests
microclimates in all three roost types may be sub-optimal for extended torpor, however, this may
not represent a physiological pitfall. Given the results presented here, silver-haired bats appear to
optimize, not maximize torpor expression throughout the winter (Humphries et al. 2003). It has
been suggested that metabolic physiology is not the only consideration in microclimate selection
by hibernating bats (Boyles et al. 2017). There may be unknown physiological or ecological
factors, such as predator avoidance, or social interaction that drive roost selection and influence
subsequent microclimate use. Although not well documented, rodents may prey on hibernating
bats (e.g., Haarsma and Kaal 2016), and may influence bat hibernation, as possible predators of
roosting bats (Lewis 1996; Lausen 2001). Bushy-tailed woodrats (Neotoma cinerea) do not
hibernate, and the Queen Victoria Mine housed a woodrat in all winters of study.
The roost microclimates, and torpor and movement patterns in relation to changes in
ambient climate suggest that silver-haired bats do not require one ‘ideal’ hibernaculum
microclimate, but rather a range of hibernaculum types to support physiological and ecological
needs throughout the winter. Winter temperatures in Southern British Columbia are mild
compared to many other parts of Canada, with Nelson experiencing an average of 100 days a
year where daily minimum temperatures drop below 0°C, compared to upwards of 130 days in

87

the prairie provinces (Environment and Climate Change Canada 2010). It may be that silverhaired bats are able to hibernate in trees here because this area has fewer days of harsh weather.
Excess fat storage facilitates roosting in cold roosts (e.g., trees), which are less limiting on the
landscape than thermally stable ones (e.g., mines). In areas of British Columbia, where bats use
trees to hibernate, a well-insulated roost, like a mine, or cave, may be essential as a refuge from
extreme cold temperatures. Where bats use multiple roosts throughout winter, the core roosting
range should be managed to ensure a sufficient supply of winter roosts are available to support
the population. Further work should focus on identifying hibernation sites and determine the
extent that hibernating in trees is common in cold climates, and how hibernation behaviours are
influenced by winter length, severity, and predation risks.

88

4. General Discussion
Roost loss is a major threat for forest-dwelling bats (Frick et al. 2019), and therefore,
understanding seasonal roosting ecology is essential for conservation. Winter is an energetic
bottleneck for bats, and hibernation is one strategy used to conserve energy (Geiser 2004).
Hibernaculum microclimate influences torpor bout duration, arousals and, consequently,
overwintering survival (French 1985; Boyles and Brack 2009). As such, choosing one or more
hibernacula that meet an individual’s energetic needs is an important component of winter
roosting ecology for bats.
I investigated hibernation ecology and described characteristics of winter roosts used by
silver-haired bats in southern British Columbia. This population of bats uses thermally unstable
trees as winter roosts, which is unusual for bats overwintering in a cold climate. Using trees as
hibernacula is common in warmer climates (e.g., Turbill and Geiser 2008; Hein et al. 2008;
Chenery et al. 2022), where bats can forage in the winter and do not have the same energetic
constraints as those that overwinter in colder climates. In Canada, bats face up to nine months
annually of low prey abundance. As such, during winter, bats typically hibernate in mines, caves,
or rock crevice features, because these roosts can protect bats from cold ambient temperatures,
and allow them to use long, deep bouts of torpor (Jonasson and Willis 2012; Klüg-Baerwald et
al. 2017; Czenze et al. 2017). Traditionally, ‘optimal’ hibernacula conditions are described as
promoting long deep, torpor bouts (Thomas and Cloutier 1992; Webb et al. 1996). However,
temperature and humidity are only two of many factors that influence selection of hibernation
sites, and bats successfully hibernate in a variety of microclimates (Boyles et al. 2017). My
research supports the hypothesis that a ‘one-size fits all’ approach to hibernacula selection is
unlikely, and that bats require multiple hibernation sites within a season and across years.

89

In Chapter 2, I characterized winter tree roosts used by silver-haired bats and compared
these to potentially available trees and summer tree roosts. I also described roost use and
movement patterns throughout the winter. Silver-haired bats used multiple winter roosts,
including the Queen Victoria Mine, trees and (to a lesser extent), rock crevices. I saw evidence
of roost fidelity within seasons, and roosts being used by different individuals across years.
Silver-haired bats used cavities in large-diameter trees in lower stages of decay in winter more
often compared to summer. I suggest that the characteristics of these roosts improve their
thermal insulation, which is important in preventing bats from freezing. Furthermore, bats used
winter tree roosts that are easy to locate when flying, and close to other roosts, potentially
providing energy savings. Features of winter tree roosts differed from summer tree roosts, where
bats primarily used exfoliating bark, in areas with higher canopy closure. The types of trees used
by silver-haired bats in the winter were not common in this forest ecosystem and had
characteristics of old-growth forest (i.e., large diameter, tall trees with evidence of decay),
suggesting that mature and older forests are important as winter roosting habitat.
In Chapter 3, I described roost microclimates and compared torpor patterns among roost
types. I also described movement, torpor and arousal patterns in relation to ambient weather
conditions. Although winter tree-roosts had insulating characteristics as described in Chapter 2,
they were significantly colder compared to the mine and rock crevices. Bats showed greater
tolerance for colder temperatures than expected inside trees, and this may be to benefit from the
high humidity, and thus lower rates of evaporative water loss (EWL) in these roosts relative to
the mine and rock crevices. Bats generally moved to more insulated roosts on colder days,
suggesting that despite the physical characteristics of tree roosts that improve insulation, they are
insufficient during periods of extreme cold. Given these results, I conclude that optimal

90

hibernation conditions are not found in one roost, and movement may be necessary to offset the
costs of either high rates of evaporative water loss or extreme cold temperatures. Although
microclimates differed significantly among roost types, torpor bout duration did not and was
relatively short compared to other cave-hibernating species (Jonasson and Willis 2012; Czenze et
al. 2017). I conclude that silver-haired bats optimize, not maximize, torpor expression throughout
the winter and may not require deep, prolonged torpor bouts to survive the winter.
Although microclimate is an important consideration for bat roost selection, other
hypotheses could help explain the use of trees in the winter. The use of trees in the winter may
provide a means of predator avoidance, or social benefits related to mating behaviour. For
hibernating bats in North America, mating occurs at hibernation sites during the fall (Thomas et
al. 1979; Glover and Altringham 2008) and may extend into winter. In fact, mating has been
observed in winter in the Smallwood Creek area through capture of silver-haired bats displaying
evidence of mating (i.e., females exuding sperm from their vagina and males with erect penises)
outside of the Queen Victoria Mine (Cori L. Lausen, [Wildlife Conservation Society Canada,
Kaslo, British Columbia], personal communication, [December 2020]).
Mating requires communication among conspecifics, and I recorded unique social
vocalizations produced by silver-haired bats called ‘songs’, and additional social calls at tree
roosts that may be related to mating and support this hypothesis (see Appendix B). Silver-haired
bat songs may be associated with mating and courtship behaviours (Lausen et al. 2023,
submitted). For migratory tree bats in North America, such as the silver-haired bat, hoary bat
(Lasiurus cinereus), and eastern red bat (Lasiurus borealis), mating may occur at trees along the
migration route during autumn (Cryan and Brown 2007; Cryan 2008; Cryan et al. 2012). The
mating behaviours of tree bats in North America have not been well studied, but in Europe, many

91

male tree-roosting bats exhibit resource defence polygyny at tree roosts (Sachteleben and von
Helversen 2006; Jahelková and Horáček 2011). Male bats emit unique social vocalizations (i.e.,
display calls) at roosts to attract females (Sachteleben and von Helversen 2006). North American
tree-roosting bats may exhibit similar behaviours and use tall, large trees as meeting points for
mating (Cryan and Brown 2007; Cryan 2008; Cryan et al 2012). The high percentage of songs,
and unique social vocalizations recorded at tree hibernacula suggests that they may play an
important role in courtship, and male silver-haired bats may defend winter roost trees as a mating
territory. Although songs were recorded at the Queen Victoria Mine, no other silver-haired bat
social vocalizations were recorded, suggesting mating may occur at the mine, but trees serve
some other social function, perhaps related to courtship, that is not present at the mine.
Silver-haired bats are considered migratory throughout most of their range (Cryan 2003);
however, I recorded year-round use of the Smallwood Creek area by male silver-haired bats, and
low capture success of females in summer, suggesting that a sedentary population of males exists
here. Variation in migratory behaviour among populations of the same species exists across a
variety of taxa (Chapman et al. 2011; Lehnert et al. 2018). Among eastern populations of silverhaired bats, females are thought to migrate longer distances than males (Fraser et al. 2017), and
this pattern appears to exist at my study site. The existence of a sedentary population of silverhaired bats at the Smallwood Creek area indicates that the mine, and surrounding forested area,
provide important habitat for this population throughout the year.

My research is the first descriptive study of winter tree roosts used by bats in British
Columbia and describes the hibernation ecology of a typically migratory species. I provide data
that improves our understanding of winter habitat requirements of silver-haired bats in the

92

province. I conclude that trees provide an important component of winter habitat for silver-haired
bats, and bats require a variety of different hibernacula to meet energetic needs throughout the
winter. Where bats use trees as hibernacula in the province, more stable roosts (i.e., mines,
caves), likely exist as a refuge in areas that experience extreme cold. These results have
important implications for the management of silver-haired bats. In British Columbia, the
legislation that provides critical bat habitat with protection (i.e., critical habitat as defined by
SARA recovery strategies for federally listed species, the Wildlife Habitat Features Order –
Kootenay-Boundary region only, the Forest and Range Practices Act, and the Mines Act) place
emphasis on mines, caves and rock features as hibernacula for bats. Although there is mention of
winter tree use in some parts of the province (Province of British Columbia 2018), little
empirical data exists to identify these features, and it is unknown how common overwintering in
trees is for other hibernating bat species, or if this behaviour is pervasive throughout the
province. Winters are mild in the west Kootenay region, where my study took place, compared to
other areas of the province (e.g., northern British Columbia), and trees may not be sufficiently
insulated to provide winter habitat in colder areas. Further work should aim to determine if treehibernation occurs elsewhere or if it is restricted to relatively mild areas of the province. These
data will aid resource managers in identifying winter tree roosts.

93

Literature Cited
Agostinelli C., Lund U. 2022. R package ‘circular’: Circular statistics. R version 0.4-95. https://rforge.r-project.org/projects/circular/
Aldridge H.D.J.N., Brigham R.M. 1988. Load carrying and maneuverability in an insectivorous
bat: A test of the 5% “rule” of radio-telemetry. Journal of Mammalogy 69:379–382.
Arnett E.B., Brown W.K., Erickson W.P., Fielder J.K., Hamilton B.L., Henry T.B., Jain A.,
Johnson G. D., Kerns J., et al. 2008. Patterns of bat fatalities at wind energy facilities in North
America. The Journal of Wildlife Management 72:61–78.
Audet D., Thomas D.W. 1996. Evaluation of the accuracy of body temperature measurement
using external radio transmitters. Canadian Journal of Zoology 74:1778–1781.
Bär A., Mayr S. 2020. Bark insulation: Ten central alpine tree species compared. Forest Ecology
and Management 474:118361.
Barclay R.M.R., Faure P.A., Farr D.R. 1988. Roosting behavior and roost selection by migrating
silver-haired bats (Lasionycteris noctivagans). Journal of Mammalogy 69:821–825.
Barclay R.M.R., Kalcounis M.C., Crampton L., Stefan C., Vonhof M. J., Wilkinson L., Brigham
R.M. 1996. Can external radiotransmitters be used to assess body temperature and torpor in bats?
Journal of Mammalogy 77:1102–1106.
Barlow K.E., Jones G. 1997. Function of pipistrelle social calls: field data and a playback
experiment. Animal Behaviour 53:991–999.
Bartoń K. 2023. MuMIn: Multi-model inference. R version 1.47.5. https://cran.rproject.org/web/packages/MuMIn/MuMIn.pdf
Barnes B.M. 1989. Freeze avoidance in a mammal: Body temperatures below 0°C in an arctic
hibernator. Science 244:1593–1595.
Bates D., Mächler M., Bolker B., Walker S. 2015. Fitting linear mixed-effects models using
lme4. Journal of Statistical Software 67:1-48.

94

Ben-Hamo M., Muñoz-Garcia A., Williams J.B., Korine C., Pinshow B. 2013. Waking to drink:
rates of evaporative water loss determine arousal frequency in hibernating bats. Journal of
Experimental Biology 216:573–577.
Bender M.J., Hartman G.D. 2015. Bat activity increases with barometric pressure and
temperature during autumn in Central Georgia. Southeastern Naturalist 14:231–242.
Bernard R.F., Foster J.T., Willcox E.V., Parise K.L., McCracken G.F. 2015. Molecular detection
of the causative agent of white-nose syndrome on Rafinesque’s big-eared bats (Corynorhinus
rafinesquii) and two species of migratory bats in the southeastern USA. Journal of Wildlife
Diseases 51:519–522.
Betts B.J. 1998. Roosts used by maternity colonies of silver-haired bats in Northeastern Oregon.
Journal of Mammalogy 79:643–650.
Blakey R.V., Law B.S., Kingsford R.T., Stoklosa J. 2017. Terrestrial laser scanning reveals
below-canopy bat trait relationships with forest structure. Remote Sensing of Environment
198:40–51.
Blejwas K.M., Lausen C.L., Rhea-Fournier D. 2014. Acoustic monitoring provides first records
of hoary bats (Lasiurus cinereus) and delineates the distribution of silver-haired bats
(Lasionycteris noctivagans) in Southeast Alaska. Northwestern Naturalist 95:236–250.
Block W.M., Brennan L.A. 1993. The habitat concept in ornithology. In: Power, D.M., editors.
Current Ornithology: Vol. 11. Springer US, Boston, USA; p.25-91.
Bohn S.J. 2017. Tall timber: Roost tree selection of reproductive female silver-haired bats
(Lasionycteris noctivagans). Master’s thesis, The University of Regina, Regina, Saskatchewan,
Canada.
Bonewell L.R., Hayes M.A., LaMantia-Olson N., Wostl E., Navo K.W. 2017. Silver-haired bats
associated with abandoned mines in Colorado provide insights into winter habitat and roost use.
Western North American Naturalist 77:404–407.

95

Borkin K.M., O’Donnell C., Parsons S. 2011. Bat colony size reduction coincides with clear-fell
harvest operations and high rates of roost loss in plantation forest. Biodiversity and Conservation
20:3537–3548.
Boyles J.G., Robbins L.W. 2006. Characteristics of summer and winter roost trees used by
evening bats (Nycticeius humeralis) in southwestern Missouri. The American Midland Naturalist
155:210–220.
Boyles J.G., Brack V. Jr. 2009. Modeling survival rates of hibernating mammals with individualbased models of energy expenditure. Journal of Mammalogy 90:9–16.
Boyles J.G., McKechnie A.E. 2010. Energy conservation in hibernating endotherms: Why
“suboptimal” temperatures are optimal. Ecological Modelling 221:1644–1647.
Boyles J.G., Boyles E, Dunlap R.K., Johnson S.A., Brack V. 2017. Long-term microclimate
measurements add further evidence that there is no “optimal” temperature for bat hibernation.
Mammalian Biology 86:9–16.
Boyles J.G., Johnson E.M., Fuller N.W., Silas K.A., Hou L., Frick W.F., McGuire L.P. 2022.
Behavioural microclimate selection and physiological responses to environmental conditions in a
hibernating bat. Canadian Journal of Zoology 100:233–238.
Brack V. 2007. Temperatures and locations used by hibernating bats, including Myotis sodalis
(Indiana bat), in a limestone mine: Implications for conservation and management.
Environmental Management 40:739–746.
Bradley S.R., Deavers D.R. 1980. A re-examination of the relationship between thermal
conductance and body weight in mammals. Comparative Biochemistry and Physiology Part A:
Physiology 65:465–476.
Brice T., Hall T. 2016. Heat index calculator. National Weather Service.
https://www.weather.gov/epz/wxcalc_heatindex. Accessed Feb 23 2023.
Brigham R.M. 1995. A winter record for the silver-haired bat in Saskatchewan. Blue Jay 53:168168.

96

Brigham R.M., Vonhof M.J., Barclay R.M.R, Gwilliam J.C. 1997. Roosting behavior and roostsite preferences of forest-dwelling California bats (Myotis californicus). Journal of Mammalogy
78:1231–1239.
B.C. Conservation Data Centre. 2015. Species summary: Lasionycteris noctivagans. B.C.
Ministry. of Environment. Available: https://a100.gov.bc.ca/pub/eswp/ (accessed Apr 24, 2023).
Britzke E.R., Gillam E.H., Murray K.L. 2013. Current state of understanding of ultrasonic
detectors for the study of bat ecology. Acta Theriologica 58:109–117.
Bull E.L., Parks C.G., Torgersen T.R. 1997. Trees and logs important to wildlife in the interior
Columbia River basin. General Technical Report PNW-GTR-391. Portland, OR. Pacific
Northwest Research Station, Forest Service, U.S. Department of Agriculture.
Burles D.W. 2014. Aspects of the winter ecology of bats on Haida Gwaii, British Columbia.
Northwestern Naturalist 95:289–299.
Burnham K.P. and Anderson D.R. 2002. Model selection and multimodel inference: a practical
information-theoretic approach. Springer US, Boston, USA; p.1-488.
Burton R.S., Reichman O.J. 1999. Does immune challenge affect torpor duration? Functional
Ecology 13:232–237.
Campbell L.A., Halleti J.G., O’Connell M.A. 1996. Conservation of bats in managed forests:
Use of roosts by Lasionycteris noctivagans. Journal of Mammology 77:976-984.
Chambers C.L., Carrigan T., Sabin T.E., Tappeiner J., McComb W.C. 1997. Use of artificially
created Douglas-fir snags by cavity-nesting birds. Western Journal of Applied Forestry 12:93–
97.
Chambers C.L., Mast J.N. 2005. Ponderosa pine snag dynamics and cavity excavation following
wildfire in northern Arizona. Forest Ecology and Management 216:227–240.
Chapman B.B., Brönmark C., Nilsson J.-Å., aHansson L.-A. 2011. The ecology and evolution of
partial migration. Oikos 120:1764–1775.

97

Chaverri G., Ancillotto L., Russo D. 2018. Social communication in bats. Biological Reviews
93:1938–1954.
Chenery M., Geiser F., Stawski C. 2022. Thermal biology and roost selection of free-ranging
male little forest bats, Vespadelus vulturnus, during winter. Journal of Mammalogy 103:1-9.
Clare S., Currie S.E. 2016. Effect of roost choice on winter torpor patterns of a free-ranging
insectivorous bat. Australian Journal of Zoology 64:132–137.
Clark B.K., Clark B.S., Leslie D.M.Jr. 1997. Seasonal variation in use of caves by the
endangered Ozark big-eared bat (Corynorhinus townsendii ingens) in Oklahoma. The American
Midland Naturalist 137:388–392.
Clement M.J., Castleberry S.B. 2013. Summer tree roost selection by Rafinesque’s big-eared bat.
The Journal of Wildlife Management 77:414–422.
Cooper S.J. 1999. The thermal and energetic significance of cavity roosting in mountain
chickadees and juniper titmice. The Condor 101:863–866.
Committee on the Status of Endangered Wildlife in Canada (COSEWIC). 2023. COSEWIC
status report in preparation with anticipated assessment dates: April 2023 wildlife species.
Available online at: https://www.cosewic.ca/index.php/en-ca/reports/status-reportspreparation.html
.
Crampton L.H., Barclay R.M.R. 1998. Selection of roosting and foraging habitat by bats in
different-aged aspen mixedwood stands. Conservation Biology 12:1347–1358.
Cryan P., Bogan M., Yanega G. 2001. Roosting habits of four bat species in the Black Hills of
South Dakota. Acta Chiropterologica 3:43–52.
Cryan P.M. 2003. Seasonal distribution of migratory tree bats (Lasiurus and Lasionycteris) in
North America. Journal of Mammology 84:579–593.

98

Cryan P.M., Veilleux J.P. 2007. Migration and use of autumn, winter, and spring roosts by tree
bats. Pages 153-175. In: Lacki M.J., Hayes J.P., Kurta A, editors. Bats in forests: conservation
and management. Johns Hopkins University Press, Baltimore, Maryland, USA.
Cryan P.M. 2008. Mating behavior as a possible cause of bat fatalities at wind turbines. The
Journal of Wildlife Management 72:845–849.
Cryan P.M., Brown A.C. 2007. Migration of bats past a remote island offers clues toward the
problem of bat fatalities at wind turbines. Biological Conservation 139:1–11.
Cryan P.M., Jameson J.W., Baerwald E.F., Willis C.K.R., Barclay R.M.R., Snider E.A., Crichton
E.G. 2012. Evidence of late-summer mating readiness and early sexual maturation in migratory
tree-roosting bats found dead at wind turbines. PLOS ONE 7:e47586.
Currie S.E., Noy K., Geiser F. 2015. Passive rewarming from torpor in hibernating bats:
minimizing metabolic costs and cardiac demands. American Journal of Physiology-Regulatory,
Integrative and Comparative Physiology 308:34-41.
Czenze Z.J., Willis C.K.R. 2015. Warming up and shipping out: arousal and emergence timing in
hibernating little brown bats (Myotis lucifugus). Journal of Comparative Physiology B 185:575–
586.
Czenze Z.J., Jonasson K.A., Willis C.K.R. 2017. Thrifty females, frisky males: Winter energetics
of hibernating bats from a cold climate. Physiological and Biochemical Zoology 90:502–511.
Dormann, C.F., Elith J., Bacher S., Buchmann C., Carl G., Carre G., Garcia Marquez J.R.,
Gruber B., Lafourcade B., Leitao P.J., et al. 2012. Collinearity: a review of methods to deal with
it and a simulation study evaluating their performance. Ecography 36:27–46.
Dunbar M.B. 2007. Thermal energetics of torpid silver-haired bats Lasionycteris noctivagans.
Acta Theriologica 52:65–68.
Dunbar M.B., Tomasi T.E. 2006. Arousal patterns, metabolic rate, and an energy budget of
eastern red bats (Lasiurus borealis) in winter. Journal of Mammalogy 87:1096–1102.

99

Dunbar M.B., Brigham R.M. 2010. Thermoregulatory variation among populations of bats along
a latitudinal gradient. Journal of Comparative Physiology B 180:885–893.
Environment Canada. 2023. Historical data.
https://climate.weather.gc.ca/historical_data/search_historic_data_e.html. Accessed 27 Feb 2023.
Environment and Climate Change Canada. 2002. Species at risk act (S.C. 2002, c. 29).
https://laws.justice.gc.ca/eng/acts/s-15.3/. Accessed 21 Feb 2023.
Falxa G. 2007. Winter foraging of silver-haired and California myotis bats in western
Washington. Northwestern Naturalist 88:98–100.
Fielding A.H., Bell J.F. 1997. A review of methods for the assessment of prediction errors in
conservation presence/absence models. Environmental Conservation 24:38–49.
Finkbeiner S.D., Briscoe A.D., Reed R.D. 2012. The benefit of being a social butterfly:
communal roosting deters predation. Proceedings of the Royal Society B: Biological Sciences
279:2769–2776.
Fraser E.E., Brooks D., Longstaffe F.J. 2017. Stable isotope investigation of the migratory
behavior of silver-haired bats (Lasionycteris noctivagans) in eastern North America. Journal of
Mammalogy 98:1225–1235.
French A.R. 1985. Allometries of the durations of torpid and euthermic intervals during
mammalian hibernation: A test of the theory of metabolic control of the timing of changes in
body temperature. Journal of Comparative Physiology B 156:13–19.
Frick W.F., Kingston T., Flanders J. 2020. A review of the major threats and challenges to global
bat conservation. Annals of the New York Academy of Sciences 1469:5–25.
Geiser F., Broome L.S. 1993. The effect of temperature on the pattern of torpor in a marsupial
hibernator. Journal of Comparative Physiology B 163:133–137.
Geiser F., Ruf T. 1995. Hibernation versus daily torpor in mammals and birds: Physiological
variables and classification of torpor patterns. Physiological Zoology 68:935–966.

100

Geiser F. 2004. Metabolic rate and body temperature reduction during hibernation and daily
torpor. Annual Review of Physiology 66:239–274.
Glover A.M., Altringham J.D. 2008. Cave selection and use by swarming bat species. Biological
Conservation 141:1493–1504.
Grillet P., Cheylan M., Thirion J-M., Dore F., Bonnet X., Dauge C., Chollet S., Antoine
Marchand M. 2010. Rabbit burrows or artificial refuges are a critical habitat component for the
threatened lizard, Timon lepidus (Sauria, Lacertidae). Biodiversity and Conservation 19:2039–
2051.
Haarsma A.-J., R. Kaal. 2016. Predation of wood mice (Apodemus sylvaticus) on hibernating
bats. Population Ecology 58:567–576.
Hamilton I.M., Barclay R.M.R. 1994. Patterns of daily torpor and day-roost selection by male
and female big brown bats (Eptesicus fuscus). Canadian Journal of Zoology 72:744–749.
Hallett J.G., Lopez T, O’Connell M.A., Borysewicz M. 2001. Decay dynamics and avian use of
artificially created snags. Northwest Science 75:378–386.
Halsall A.L., Boyles J.G., Whitaker Jr J.O. 2012. Body temperature patterns of big brown bats
during winter in a building hibernaculum. Journal of Mammalogy 93:497–503.
Hammesfahr A.M., Ohms R.E. 2018. Winter bat activity in a landscape without traditional
hibernacula. U.S. National Park Service Publication Papers 232: 1-41.
Hansen M.C., Potapov P.V., Moore R., Hancher M., Turubanova S.A., Tyukavina A., Thau D.,
Stehman S.V., Goetz S.J., Loveland T.R., et al. 2013. High-resolution global maps of 21stcentury forest cover change. Science 342:850–853.
Harper M.J., McCarthy M.A., van der Ree R. 2005. The abundance of hollow-bearing trees in
urban dry sclerophyll forest and the effect of wind on hollow development. Biological
Conservation 122:181–192.

101

Hein C.D., Castleberry S.B., Miller K.V. 2008. Male seminole bat winter roost-site selection in a
managed forest. The Journal of Wildlife Management 72:1756–1764.
Hooge P.N., Stanback M.T., Koenig W.D. 1999. Nest-site selection in the acorn woodpecker.
The Auk 116:45–54.
Hosken D.J., Withers P.C. 1997. Temperature regulation and metabolism of an Australian bat,
Chalinolobus gouldii (Chiroptera: Vespertilionidae) when euthermic and torpid. Journal of
Comparative Physiology B: Biochemical, Systemic, and Environmental Physiology 167:71–80.
Hosmer D.W., Lemeshow S. 2000. Applied logistic regression. John Wiley and Sons, Inc, New
York, USA.
Humphrey S.R. 1975. Nursery roosts and community diversity of nearctic bats. Journal of
Mammalogy 56:321–346.
Humphries M.M., Thomas D.W., Speakman J.R. 2002. Climate-mediated energetic constraints
on the distribution of hibernating mammals. Nature 418:313–316.
Humphries M.M., Thomas D.W., Kramer D.L. 2003. The role of energy availability in
mammalian hibernation: A cost‐benefit approach. Physiological and Biochemical Zoology:
Ecological and Evolutionary Approaches 76:165–179.
Ingersoll T.E., Navo K.W., de Valpine P. 2010. Microclimate preferences during swarming and
hibernation in the townsend’s big-eared bat, Corynorhinus townsendii. Journal of Mammalogy
91:1242–1250.
Jahelková H., Horáček I. 2011. Mating system of a migratory bat, Nathusius’ pipistrelle
(Pipistrellus nathusii): different male strategies. Acta Chiropterologica 13:123–137.
Jonasson K.A., Willis C.K.R. 2012. Hibernation energetics of free-ranging little brown bats.
Journal of Experimental Biology 215:2141–2149.
Jorge M.H., Ford W.M., Sweeten S.E., Freeze S.R., True M.C., St. Germain M.J., Taylor H.,
Gorman K.M., Garrison E.P., Cherry M.J. 2021. Winter roost selection of Lasiurine tree bats in a
pyric landscape. PLOS ONE 16:e0245695.

102

Kalcounis‐Rüppell M.C., Psyllakis J.M, Brigham R.M. 2005. Tree roost selection by bats: an
empirical synthesis using meta-analysis. Wildlife Society Bulletin 33:1123–1132.
Klüg-Baerwald B.J., Lausen C.L., Willis C.K.R, Brigham R.M. 2017. Home is where you hang
your bat: winter roost selection by prairie-living big brown bats. Journal of Mammalogy 98:752–
760.
Kunz T.H. 1982. Lasionycteris noctivagans. Mammalian Species. 172:1-5.
Kunz T.H., Wrazen J.A., Burnett C.D. 1997. Changes in body mass and fat reserves in prehibernating little brown bats (Myotis lucifugus). Écoscience 5:8–17.
Kunz T.H., Lumsden L.F. 2003. Ecology of cavity and foliage roosting bats. In: Kunz T.H.,
Fenton M.B., editors. Bat ecology. University of Chicago Press, Chicago, USA; p.3-89.
Kurta A. 2014. The misuse of relative humidity in ecological studies of hibernating bats. Acta
Chiropterologica 16:249–254.
Kurta A., Auteri G.G., Hofmann J.E., Mengelkoch J.M., White J.P., Whitaker Jr. J.O., Cooley T.
Melotti J. 2017. Influence of a large lake on the winter range of a small mammal: Lake michigan
and the silver-haired bat (Lasionycteris noctivagans). Diversity 10:1–24.
Lausen C.L. 2001. Thermoregulation and roost selection by reproductive big brown bats
(Eptesicus fuscus) roosting in the South Saskatchewan River Valley, Alberta: rock-roosting and
building-roosting colonies. M.Sc. Thesis, University of Calgary, Alberta.
Lausen C.L., Barclay R.M.R. 2006. Winter bat activity in the Canadian prairies. Canadian
Journal of Zoology 84:1079–1086.
Lausen C.L., Nagorsen D.W., Brigham R.M., Hobbs. J. 2022. Bats of British Columbia. 2 nd ed.
Royal BC Museum, Victoria, British Columbia, Canada.

103

Lemen C.A., Freeman P.W., White J.A. 2016. Acoustic evidence of bats using rock crevices in
winter: A call for more research on winter roosts in North America. Transactions of the Nebraska
Academy of Sciences and Affiliated Societies.
Lehnert L.S., Kramer-Schadt S., Teige T., Hoffmeiser U., Popa-Lisseanu A., Bontadina F.,
Ciechanowski M., Dechmann D.K.N., Kravchenko K., Presetnik P., et al. 2018. Variability and
repeatability of noctule bat migration in Central Europe: evidence for partial and differential
migration. Proceedings of the Royal Society B: Biological Sciences 285:20182174.
Lewis S.E. 1996. Low roost-site fidelity in pallid bats: associated factors and effect on group
stability. Behavioral Ecology and Sociobiology 39:335–344.
Limpert D.L., Birch D.L., Scott M.S., Andre M., Gillam E. 2007. Tree selection and landscape
analysis of eastern red bat day roosts. The Journal of Wildlife Management 71:478–486.
Loeb S., Rodhouse T.J., Ellison L.E., Lausen C.L., Reichard J.D., Irvine K.M., Ingersoll T.E.,
Coleman J.T.H., Thogmartin W.E., Sauer J.R., et al. 2015. A plan for the North American bat
bonitoring program (NABat). General Technical Repport SRS-208. Asheville, NC: U.S.
Department of Agriculture Forest Service, Southern Research Station. 100 p.
Martin, K., Aitken K.E.H., Wiebe K.L. 2004. Nest sites and nest webs for cavity-nesting
communities in interior British Columbia, Canada: Nest characteristics and niche partitioning.
The Condor 106:5–19.
Martin T.G., Camaclang A.E., Possingham H.P., Maguire L.A., Chadès I. 2016. Timing of
protection of critical habitat matters. Conservation Letters 10:308–316.
Mattson T.A., Buskirk S.W., Stanton N.L. 1996. Roost sites of the silver-haired bat
(Lasionycteris noctivagans) in the Black Hills, South Dakota. The Great Basin Naturalist
56:247–253.
Mazerolle M.J. 2023. AICcmodavg: Model selection and multimodel inference based on
(Q)AIC(c). R version 2.3.2. https://cran.r-project.org/package=AICcmodavg.
McAlpine D.F., Bullied J.L., Seymour P.D. 2021. A maternity roost of silver-haired bats
(Lasionycteris noctivagans) in New Brunswick: First evidence of parturition in atlantic Canada.
Northeastern Naturalist 28:1–6.

104

McGuire L.P., Fuller N.W., Dzal Y.A., Haase C.G., Klug-Baerwald B.J., Silas K.A., Plowright
R.K., Lausen C.L., Willis C.K.R., Olson S.H. 2021. Interspecific variation in evaporative water
loss and temperature response, but not metabolic rate, among hibernating bats. Scientific Reports
11:20759.
Michener G.R. 1992. Sexual differences in over-winter torpor patterns of Richardson’s ground
squirrels in natural hibernacula. Oecologia 89:397–406.
Milchram M., Suarez-Rubio M., Schröder A., Bruckner A. 2020. Estimating population density
of insectivorous bats based on stationary acoustic detectors: A case study. Ecology and Evolution
10:1135–1144.
Ministry of Forests, Land and Natural Resource Operations. 2018. Vegetation resources
inventory-British Columbia. Forest Analysis and Inventory Branch.
Monarchino M.N., Behan M.L., Johnson J.S. 2020. Summer day-roost selection by eastern red
bats varies between areas with different land-use histories. PLOS ONE 15:e0237103.
Moretto L., Francis C.M. 2017. What factors limit bat abundance and diversity in temperate,
North American urban environments? Journal of Urban Ecology 3:1-9.
Mormann B.M., Robbins L.W. 2007. Winter roosting ecology of eastern red bats in southwest
Missouri. The Journal of Wildlife Management 71:213–217.
Muñoz-Garcia A., Ben-Hamo M., Pinshow B., Williams J.B., Korine C. 2012. The relationship
between cutaneous water loss and thermoregulatory state in Kuhl’s pipistrelle Pipistrellus kuhlii,
a Vespertillionid bat. Physiological and biochemical zoology 85:516–525.
Murray K.L., Britzke E.R., Hadley B.M., Robbins L.W. 1999. Surveying bat communities: a
comparison between mist nets and the Anabat II bat detector system. Acta Chiropterologica
1:105-112.
Nagorsen D.W., Bryant A.A., Kerridge D., Roberts G., Roberts A., and Sarell M.J. 1993. Winter
bat records for British Columbia. Northwestern Naturalist 74:61-61.

105

Newman B.A., Loeb S.C., Jachowski D.S. 2021. Winter roosting ecology of tricolored bats
(Perimyotis subflavus) in trees and bridges. Journal of Mammalogy 102:1331–1341.
O’Keefe J.M., Loeb S.C., Lanham J.D., H. S. Hill H.S. 2009. Macrohabitat factors affect day
roost selection by eastern red bats and eastern pipistrelles in the southern Appalachian
Mountains, USA. Forest Ecology and Management 257:1757–1763.
Paige K.N. 1995. Bats and barometric pressure: Conserving limited energy and tracking insects
from the roost. Functional Ecology 9:463–467.
Park K.J., Jones G., Ransome R.D. 2000. Torpor, arousal and activity of hibernating greater
horseshoe bats (Rhinolophus ferrumequinum). Functional Ecology 14:580–588.
Parsons S., Lewis K.J., Psyllakis J.M. 2003. Relationships between roosting habitat of bats and
decay of aspen in the sub-boreal forests of British Columbia. Forest Ecology and Management
177:559–570.
Pearce J., Ferrier S. 2000. Evaluating the predictive performance of habitat models developed
using logistic regression. Ecological Modelling 133:225–245.
Perry R.W., Saugey D.A., Crump B.G. 2010. Winter roosting ecology of silver-haired bats in an
Arkansas Forest. Southeastern Naturalist 9:563–572.
Perry R.W. 2013. A review of factors affecting cave climates for hibernating bats in temperate
North America. Environmental Reviews 21:28–39.
Pojar J., Klinka K., and Meidinger D.V. 1987. Biogeoclimatic ecosystem classification in British
Columbia. Forest Ecology and Management 22:119–154.
Province of British Columbia. 2018. Kootenay Boundary Wildlife Habitat Features Order.
https://www2.gov.bc.ca/gov/content/environment/natural-resource-stewardship/laws-policiesstandards-guidance/legislation-regulation/forest-range-practices-act/government-actionsregulation/wildlife-habitat-features/kootenay-boundary-wildlife-habitat-features-order. Accessed
Feb 23 2023.
R Development Core Team. 2020. R: A language and environment for statistical computing. R
Foundation for Statistical Computing, Vienna, Austria.

106

Racey P.A., Swift S.M. 1981. Variations in gestation length in a colony of pipistrelle bats
(Pipistrellus pipistrellus) from year to year. Journal of Reproductive Fertility 61:123–129.
Reimer J.P. 2014. Bat activity and use of hibernacula in Wood Buffalo National Park, Alberta.
Northwestern Naturalist 95:277–288.
Russo D., Cistrone L., Garonna A.P., Jones G. 2010. Reconsidering the importance of harvested
forests for the conservation of tree-dwelling bats. Biodiversity and Conservation 19:2501–2515.
Ryan C.C., Burns L.E., Broders H.G. 2019. Changes in underground roosting patterns to
optimize energy conservation in hibernating bats. Canadian Journal of Zoology 97:1064–1071.
Sachteleben J., von Helversen O. 2006. Songflight behaviour and mating system of the
pipistrelle bat (Pipistrellus pipistrellus) in an urban habitat. Acta Chiropterologica 8:391–401.
Sedgeley J.A. 2001. Quality of cavity microclimate as a factor influencing selection of maternity
roosts by a tree-dwelling bat, Chalinolobus tuberculatus, in New Zealand. Journal of Applied
Ecology 38:425–438.
Sing T., Sander O., Beerwinkel N., Lengauer T. 2005. ROCR: visualizing classifier performance
in R. Bioinformatics 21: 7881. http://rocr.bioinf.mpi-sb.mpg.de.
Smotherman M., Knörnschild M., Smarsh G., Bohn K. 2016. The origins and diversity of bat
songs. Journal of Comparative Physiology A 202:535–554.
Speakman J.R., Webb P.I., Racey P.A. 1991. Effects of disturbance on the energy expenditure of
hibernating bats. Journal of Applied Ecology 28:1087–1104.
Speakman J.R., Thomas D.W. 2003 Physiological ecology and energetics of bats. In: Kunz T.H.,
Fenton M.B., editors. Bat ecology. University of Chicago Press, Chicago, USA; p. 430-750.
Speakman J.R. 2008. The physiological costs of reproduction in small mammals. Philosophical
Transactions of the Royal Society B: Biological Sciences 363:375–398.

107

Stawski C., Turbill C., Geiser F. 2009. Hibernation by a free-ranging subtropical bat
(Nyctophilus bifax). Journal of Comparative Physiology B 179:433–441.
Stawski C., Currie S.E. 2016. Effect of roost choice on winter torpor patterns of a free-ranging
insectivorous bat. Australian Journal of Zoology 64:132–137.
Swanson G., Evans C. 1936. The hibernation of certain bats in southern Minnesota. Journal of
Mammalogy 17:39–43.
Tena E., de Paz Ó., de la Peña R., Fandos G., Redondo M., and Tellería J.L. 2020. Mind the gap:
Effects of canopy clearings on temperate forest bat assemblages. Forest Ecology and
Management 474:118341.
Therneau T.M. 2023. A package for survival analysis in R. R version 3.5-5. https://CRAN.Rproject.org/package=survival.
Thomas S.P., Suthers R.A. 1972. The physiology and energetics of bat flight. Journal of
Experimental Biology 57:317–335.
Thomas D.W., Fenton M.B., Barclay R.M.R. 1979. Social behavior of the little brown bat,
Myotis lucifugus: I. Mating behavior. Behavioral Ecology and Sociobiology 6:129–136.
Thomas D.W., Dorais M., Bergeron J.-M. 1990. Winter energy budgets and cost of arousals for
hibernating little brown bats, Myotis lucifugus. Journal of Mammalogy 71:475–479.
Thomas D.W., Cloutier D. 1992. Evaporative water loss by hibernating little brown bats, Myotis
lucifugus. Physiological Zoology 65:443–456.
Thomas D.W., Geiser F. 1997. Periodic arousals in hibernating mammals: is evaporative water
loss involved? Functional Ecology 11:585–591.
Trachsel L., Edgar D.M., Heller H.C. 1991. Are ground squirrels sleep deprived during
hibernation? American Journal of Physiology-Regulatory, Integrative and Comparative
Physiology 260:1123–1129.

108

Tsunoda M., Newman C., Buesching C.D., Macdonald D.W., Kaneko Y. 2018. Badger setts
provide thermal refugia, buffering changeable surface weather conditions. Journal of Thermal
Biology 74:226–233.
Turbill C. 2008. Winter activity of Australian tree-roosting bats: influence of temperature and
climatic patterns. Journal of Zoology 276:285–290.
Turbill C., Geiser F. 2008. Hibernation by tree-roosting bats. Journal of Comparative Physiology
B 178:597–605.
Twente J.W., Twente J., Brack Jr V. 1985. The duration of the period of hibernation of three
species of vespertilionid bats. II. Laboratory studies. Canadian Journal of Zoology 63:2955–
2961.
Vonhof M.J., Barclay R.M.R. 1996. Roost-site selection and roosting ecology of forest-dwelling
bats in southern British Columbia. Canadian Journal of Zoology 74:1797–1805.
Vonhof M.J., Gwilliam J.C. 2007. Intra- and interspecific patterns of day roost selection by three
species of forest-dwelling bats in Southern British Columbia. Forest Ecology and Management
252:165–175.
Walsberg G.E. 1986. Thermal consequences of roost-site selection: The relative importance of
three modes of heat conservation. The Auk 103:1–7.
Webb P.I., Speakman J.R., Racey P.A. 1995. Evaporative water loss in two sympatric species of
vespertilionid bat, Plecotus auritus and Myotis daubentoni: relation to foraging mode and
implications for roost site selection. Journal of Zoology 235:269–278.
Webb P.I., Speakman J.R., Racey P.A. 1996. How hot is a hibernaculum? A review of the
temperatures at which bats hibernate. Canadian Journal of Zoology 74:761–765.
Weller T.J., Cryan P.M., O’Shea T.J. 2009. Broadening the focus of bat conservation and
research in the USA for the 21st century. Endangered Species Research 8:129–145.

109

Weller T.J., Rodhouse T.J., Neubaum D.J., Ormsbee P.C., Dixon R.D., Popp D.L., Williams
J.A., Osborn S.D., Rogers B.W., Beard L. O., et al. 2018. A review of bat hibernacula across the
western United States: Implications for white-nose syndrome surveillance and management.
PLOS ONE 13:e0205647.
Wermundsen T., Siivonen Y. 2010. Seasonal variation in use of winter roosts by five bat species
in south-east Finland. Open Life Sciences 5:262–273.
Whiting J.C., Doering B., Wright G., Englestead D.K., Frye J.A., Stefanic T. 2018. Bat
hibernacula in caves of southern Idaho: Implications for monitoring and management. Western
North American Naturalist 78:165–173.
Wiebe K.L. 2001. Microclimate of tree cavity nests: Is it important for reproductive success in
northern flickers? The Auk 118:412–421.
Willis C.K.R., Brigham R.M.R. 2003. Defining torpor in free-ranging bats: experimental
evaluation of external temperature-sensitive radiotransmitters and the concept of active
temperature. Journal of Comparative Physiology B 173:379–389.
Willis C.K.R., Brigham R.M. 2007. Social thermoregulation exerts more influence than
microclimate on forest roost preferences by a cavity-dwelling bat. Behavioral Ecology and
Sociobiology 62:97–108.
Zahn A. 1999. Reproductive success, colony size and roost temperature in attic-dwelling bat
Myotis myotis. Journal of Zoology 247:275–280.
Zar J.H. 1999. Biostatistical analysis. 4 th ed. Prentice Hall International, New Jersey, USA.

110

Appendix A: Skin Temperature Measurements – Problems and Observations
Assessing body temperature (Tb) in hibernating mammals can allow us to measure torpor
depth and provides insight into energy requirements over winter (Dunbar and Tomasi 2006;
Jonasson and Willis 2012). Understanding winter energy requirements can be useful for
identifying critical winter habitat, responses to environmental conditions, and impacts from
disease (Michener 1992; Boyles and Brack 2009; Frick et al. 2020). In my original study design,
I planned to assess differences in Tb of silver-haired bats (Lasionycteris noctivagans) hibernating
in a mine, rock crevices, and tree roosts, to determine if metabolic rates during torpor differed
given the differing microclimates expected across these roost types. Specifically, I aimed to
investigate the depth of torpor used in cold, unstable roosts (i.e., trees and rocks).
Direct measurement of body temperature requires implantation of transmitters, but
temperature-sensitive transmitters affixed to the skin are a reliable proxy commonly used to
study torpor in bats (Barclay et al. 1996; Willis and Brigham 2003). These record pulse interval,
which can be converted to skin temperature (Tsk), and skin temperature can accurately imply
body temperature (Audet and Thomas 1996; Willis and Brigham 2003). Throughout the course
of the study, I used multiple telemetry receiver types to record data from bats in roosts (SRX–
400, SRX–600, SRX–800, SRX–1200; Lotek Wireless Inc., Newmarket, Ontario). I deployed a
Lotek SRX–400 in the Queen Victoria Mine for the duration of the study and moved newer
models (SRX–600/800/1200) among active tree and rock crevice roosts. I attempted to
standardize data by setting all Lotek receivers, regardless of model, to record with the same
parameters. A notable exception to this was the number of recordings the unit took before it
created a data point, otherwise known as the group size. The newer models recorded two
transmitter pulses per recording time, and took an average of this recording, whereas the SRX–

111

400 model recorded six pulses before taking an average. Based on my observations, the biggest
difference in the change in group size between the units is the amount of noise, or unreliable
recordings, they produced. The newer models, with a group size of two, were more likely to
produce more data points attributed to noise than the older model. This noise was easily filtered
out during the analysis, and in my opinion, did not contribute significantly to the differences in
temperatures observed between unit types.
The SRX–400 model records transmitter rate with no decimal places (e.g., 25), whereas
the newer models record transmitter rate with two decimal places (e.g., 25.12). When I converted
pulse intervals to skin temperatures, small fluctuations in pulse interval resulted in large
fluctuations in skin temperature, up to 5°C for every one beat per minute (Figure A-1). The
newer receivers, placed at tree and rock crevice roosts, record pulse interval with higher
sensitivity, and I did not observe these large fluctuations in temperatures (Figure 3.3). Although
the SRX–400 receivers, when coupled with Holohil transmitters (used here), can record skin
temperatures that accurately reflect body temperature (Audet and Thomas 1996), the difference
in sensitivity between the unit placed in the mine and units placed at other roosts made direct
comparisons of skin temperatures between the unit types difficult.
For data collected in the winters of 2021 and 2022, I noticed unusually high Tsk compared
to earlier data collection (2012, 2014, 2018). During 2021 and 2022, bats were recorded with
torpid Tsk up to 20°C in all roost types (Figure 3.3; Figure 3.6), whereas torpid Tsk of bats
recorded before 2021 was typically < 15°C. Because calibration of transmitters was not
completed prior to field work, it is impossible to determine if this is a result of transmitter
malfunction, or a biological phenomenon. Because I observed high Tsk in tree, rock, and mine

112

roosts, it is likely that if it is a biological phenomenon, it is not linked to differences in
microclimate.

113

Figure A-1. Skin (Tsk) temperature (°C) a silver-haired bat (Lasionycteris noctivagans)
hibernating in the Queen Victoria Mine from January 2022 – March 2022 in the Smallwood
Creek area near Beasley, British Columbia, recorded on a Lotek SRX–400 unit. Sharp increases
in Tsk indicate arousals. Skin temperature changed widely during torpor based on 1 beat per
minute increments, highlighted by the arrows. In this case, a transmitter rate of 22 was equivalent
to Tsk of 12.16°C (lower arrow), and a transmitter rate of 23 was equivalent to 17.43°C.

114

Appendix B: Acoustics Around Hibernacula and Singing Silver-Haired Bats (Lasionycteris
noctivagans)
Information in this Appendix was completed as a collaborative effort with the Ministry of
Environment and Climate Change Strategy
Introduction
Identifying bat hibernacula can be difficult, because bats often roost in underground
structures or deep in rock crevices or tree cavities, making them hard to access and locate
visually. Capture and radio telemetry of bats during winter to identify hibernacula can also be
problematic, because bats are relatively inactive during hibernation. Because bat hibernacula are
difficult to find, and winter capture can be labour intensive, deploying ultrasonic detectors may
be a viable, cost-effective alternative to identify approximate locations of bat hibernacula and
species presence (Lemen et al. 2016; Hammesfahr and Ohms 2018).
The use of ultrasonic detectors is common practice in the study of bat ecology (Britzke et
al. 2013). When deployed passively and over multiple days, ultrasonic detectors can give insight
into species’ presence, activity patterns, and abundance (Loeb et al. 2015; Milchram et al. 2020).
Compared to capture, the deployment of detectors and subsequent analysis of data is costeffective, non-invasive and not labour-intensive (Murray et al. 1999), making them an ideal
research tool in many scenarios.
Bats navigate and detect prey using echolocation. During echolocation bats produce
sounds, or calls. Some bats produce sounds for communication instead of prey detection or
navigation (Chaverri et al. 2018). These sounds, hereafter referred to as social calls, are distinct,
and can be used to identify species, and give insight into social behaviour (Chaverri et al. 2018).
Silver-haired bats (Lasionycteris noctivagans; LANO) produce a distinct social vocalization

115

called a ‘song’, which can be useful for identifying the species when similarly echolocating
species co-occur (Lausen et al. 2022; Lausen et al., in review). Songs have been documented in
only a small percentage of bat species and are defined as a social call with a repeating pattern
(reviewed in Smotherman et al. 2016). The exact function of the silver-haired song is unknown,
but elsewhere bat songs are commonly associated with mating or resource defence (Barlow and
Jones 1997; Smotherman et al. 2016). Silver-haired bat songs are most often recorded in the
winter (Lausen et al. 2023, submitted), and thus, regardless of their function, may indicate the
presence of hibernacula when recorded during the hibernation season.
As part of a long-term joint monitoring with the Ministry of Environment and Climate
Change Strategy, I installed bat detectors in the Queen Victoria Mine, and at surrounding tree
roosts, in the Smallwood Creek area of British Columbia, to record activity patterns and acoustic
signatures. My objective was to determine if silver-haired bats hibernating in this area produced
patterns of unique acoustic signatures that could be used to identify hibernation sites.

Methods
From 08 December 2020 to 16 – 29 April 2021, I installed bat detectors (SM2+ Bat
Detector, Wildlife Acoustics, Inc. Maynard, Massachusetts) in the Queen Victoria Mine, and
different units (SM4Bat; Wildlife Acoustics Inc. Maynard, Massachusetts, and Swift; Titley
Scientific Columbia, Missouri) at tree roosts, as identified using radio telemetry (see Chapter 2
for methods; Figure B-1). In the Queen Victoria Mine, I placed the bat detector 3m high on a
pole, close to the mine entrance. At tree roosts, I deployed detectors 3m to 7m from the tree with
the microphone pointed towards the roost entrance. In one instance, a detector was placed in
between two known roost trees which were in close proximity. Once deployed, detectors

116

recorded continuously until 16 – 29 April 2021. All detectors were programmed to turn on 30
minutes before sunset and turn off 30 minutes after sunrise each day. Detectors recorded in zerocrossing (i.e., recording the loudest frequency at a given time, instead of the full range) in the
mine and full spectrum at tree roosts. Full spectrum recording uses more battery life, and to
reduce visitation to the mine detector, and thus disturbance inside the hibernaculum, zerocrossing was used on this unit.

Bat echolocation recordings were analyzed manually to identify species, or species
groups using Anabat Insight version 1.9.2 (Full-spectrum files; Titley Scientific, Columbia,
Missouri) or AnalookW version 4.4a (Zero-crossing; Titley Scientific, Columbia, Missouri).
When species identification was impossible due to overlapping characteristics of certain species,
or poor file quality, I placed potential silver-haired bat calls into one of two groups: an
EPFULANO dyad (silver-haired bat or big brown bat; Eptesicus fuscus), or a 25K (lowfrequency bat) group. All potential silver-haired bat calls were pooled by combining files
identified as LANO, EPFULANO and 25K together, and are hereafter referred to as lowfrequency bat calls. Variable mean values are presented as mean ± 1 standard deviation.

117

Figure B-1. Map of acoustic detectors deployed from 08 December 2020 to 16 – 29 April 2021
at silver-haired bat (Lasionycteris noctivagans) hibernacula in the Smallwood Creek area near
Beasley, British Columbia. The location of detectors placed at tree hibernacula are marked with
yellow diamonds. The location of the Queen Victoria Mine detector is marked with a red star.

118

Results
From December 2020 to April 2021, I deployed eight acoustic detectors (seven tree
roosts; one mine) and recorded 453 nights of activity. No data were recorded from one of the
detectors over the deployment period; thus, I excluded it from subsequent analysis. Bats were
active throughout the winter and recorded on 50.1% ± 25.5% of nights when detectors were
deployed and recording (detector-nights; Table B-1). The only bat species recorded throughout
the winter were those that hibernate in the Queen Victoria Mine (silver-haired bat; Townsend’s
big-eared bat, Corynorhinus townsendii; California myotis, Myotis californicus). I recorded 1830
low-frequency bat passes (Table B-1). Silver-haired bat passes were recorded on each detector.
Detectors placed at tree roost 3/5 recorded the highest proportion of low-frequency bat
passes/night (Table B-1).

Of all low-frequency calls recorded in the mine, 14.6% were confirmed silver-haired bat
songs (Figure B-2). For low-frequency calls pooled across detectors at tree roosts, 14.1% were
confirmed silver-haired bat songs, although the percentage of songs recorded was highly variable
among tree roost detectors (Table B-1).

119

Table B-1. Summary of low frequency bat recordings from six bat detectors (SM4Bat, Swift or
SM2+) deployed at silver-haired bat (Lasionycteris noctivagans) hibernacula in the Smallwood
Creek area near Beasley, British Columbia from January – April 2021. An additional detector
was deployed but did not capture any acoustic data during the sampling period and was excluded
from analysis. The number of low frequency bat calls included files that were labelled as LANO
(silver-haired bat), EPFULANO (big drown bat; Eptesicus fuscus –silver-haired bat dyad), or
25K (low frequency bat).
Detector ID

Detector Nights

Nights With
Bat Activity

Low-Frequency
Bat Passes

Percent Silverhaired Song

Tree Roost 1

68

29

90

5.6

Tree Roost 3/5

74

39

358

26.5

Tree Roost 4

59

25

83

3.6

59

47

746

10.5

37

11

25

12.0

Tree Roost 8

54

9

14

14.3

Mine

102

89

603

14.6

Tree Roost 6
Tree Roost 7

120

A

B

C

Figure B-2. Silver-haired bat (Lasionycteris noctivagans) acoustic recordings captured from the
Smallwood Creek area near Beasley, British Columbia from December 2020 – April 2021. A) A
typical repeated silver-haired bat ‘song’ phrase; B) and C) Previously undocumented and unique
silver-haired bat social calls.

121

Discussion
Bats known to hibernate in the area were recorded on over half of the detector-nights, and
only species known to hibernate in the immediate vicinity were recorded on the detector. Six
other species of bat have the potential to hibernate in the west Kootenay region (Lausen et al.
2022), however, none of these species were recorded on the detectors, indicating they do not
hibernate in the area. As such, bat passes recorded during the winter are likely to originate from
individuals that are hibernating near by, and locally hibernating bats can be recorded
acoustically.
Two of the tree roosts with the highest activity levels of low-frequency bats (Table B-1)
were close (< 80m apart) and connected by a linear flyway (de-activated forestry skid road),
suggesting an overlap in detections. Through radio telemetry, I identified nine silver-haired bat
tree hibernacula near the two detectors. This area has a relatively high density of large-diameter
ponderosa pine (Pinus ponderosa) and Douglas-fir (Pseudostuga menzeisii) trees, which were
identified as high-value roost trees (see Chapter 2). As such, this area may be high-quality
habitat for this species, which would explain the high number of low-frequency calls recorded.
I recorded silver-haired bat songs at all identified hibernacula. The songs consist of three
different pulses which form a repetitive pattern: the lead, droplet and a series of higher frequency
chirps (Figure B-2A). The songs have a highly repetitive pattern, with a short time between
phrases. The proportion of songs recorded here is higher than documented elsewhere across
western North America, where silver-haired bat songs comprise 7% – 10% of all low-frequency
passes in the winter (Lausen et al., in review). Silver-haired bat songs are most often recorded in

122

the winter (Lausen et al., in review), and their regularity at known hibernacula suggests that this
type of acoustic signature may be useful in identifying silver-haired bat hibernacula.
At three of the detectors placed at tree roosts, I recorded previously undocumented silverhaired bat acoustic calls (Figure B-2B, C) unlike a typical song phrase (Figure B-2A). These
calls may be for social purposes, because they were often recorded close in time to song phrases
or as a duet (i.e., coordinated singing with a conspecific during mating; Figure B-2C). I did not
detect these distinct acoustic calls in the Queen Victoria Mine. Thus, I hypothesize that winter
tree roosts have an important social function. Because these distinct acoustic calls were rare, I
cannot draw concrete conclusions, but suggest further study.

The high proportion of silver-haired bat song recorded here, relative to what has been
documented elsewhere in western North America indicates acoustic detectors could be a useful,
non-invasive and cost-effective tool for identifying potential silver-haired bat hibernacula. The
detection of low-frequency bats, particularly silver-haired bat songs may allow researchers to
refine their search area when locating hibernacula based on acoustic detections. Further data
collection should focus on recordings from areas devoid of winter roosting opportunities as a
control measure and compare acoustic activity from other hibernation roosts (rock crevices,
mines).