SEASONAL ABUNDANCE AND HABITAT ASSOCIATIONS OF WANDERING
CATS AND BIRDS IN A TEMPERATE ZONE BIODIVERSITY HOTSPOT
by
Olivia Wilson
B.Sc. (Hons), McMaster University, 2021

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
April 2025

© Olivia Wilson, 2025

ii
ABSTRACT
The interaction between birds and wandering domestic cats is an ongoing challenge for both
wildlife conservation and cat welfare, particularly in regions where high avian diversity
overlaps with dense human development and wandering cats. I examined the abundance,
richness and community structure of birds and the abundance of wandering domestic cats
(Felis catus), in the temperate biodiversity hotspot of the south Okanagan Valley, British
Columbia, Canada, between Okanagan Falls and Osoyoos, across an entire annual period. I
did this by pairing point counts and photos from trail cameras from 123 locations across five
seasonal periods between March 2022 and March 2023, assessing the habitat associations of
birds and cats across a variety of land use types, including urban, peri-urban, agricultural, and
natural. I conducted a total of 2380 point counts and used hierarchical modelling and
unconstrained ordination to examine bird abundance and species richness, and community
composition, respectively. My results revealed distinct seasonal patterns of bird abundance
and richness with these metrics being the highest during spring migration and the breeding
season. Urbanization and human development impacted the distribution of birds year-round,
especially in the non-breeding seasons when a large diversity of species used urban areas.
Using the same locations as the point counts, but shifting cameras every 28 days, I examined
local abundance of wandering cats. I showed that wandering cats were found in high
abundances in urban habitats year-round but overall had the highest abundances during the
early winter, spring, and summer. Wandering cats were detected at 100% of peri-urban sites,
97% of urban sites, 65% of agricultural sites and 42% of natural sites. I estimated an annual
average of 6,557 wandering cats within the study area with up to 82% of them being
unowned cats, equating to one cat for every two to three people. Overall, I demonstrate the
importance of identifying where birds are and what habitats they are using across the entire

iii
year and not only during distinct periods within the annual cycle (e.g. breeding). The high
numbers of wandering cats, combined with the diversity of birds and other wildlife, suggests
that cats likely have significant impacts on birds and other wildlife year-round in the study
region and likely elsewhere. In the south Okanagan Valley, management actions such as
outreach initiatives should take a seasonal and habitat-based approach. Outreach should focus
on encouraging urban residents to keep their cats indoors during the winter because many
species move into urban areas at this time. Resident in peri-urban and agricultural habitats
should be encouraged or incentivised to spay and neuter cats on their property and keep cats
inside during spring, summer, and fall when high cat numbers overlap with high bird
abundance and richness. Given the high abundance and richness of birds along with the
highest currently reported abundances of wandering cats per capita, these results stress the
urgent need for collaborative efforts among municipalities, stakeholders, and residents to
mitigate the ecological impact of wandering cats and help preserve the biodiversity of this
unique region.

iv
PREFACE
Funding for this Masters was provided by Mitacs Accelerate Grant in partnership with the
Stewardship Centre for British Columbia. Additional funding was supplied by Environment
and Climate Change Canada (Wildlife Research Division and Canadian Wildlife Services),
the Liber Ero Foundation, and a Graduate Entrance Research Scholarship from UNBC.
Permits for access to conduct research on private and provincial land were granted by
Nature Trust of British Columbia, BC Parks, the BC Ministry of Forests, and the Osoyoos
Desert Centre. Access to Vineyards, Wineries, and private residential property was granted by
the occupying resident. I went door-to-door asking residents for access to their property and
provided all of the information upfront, verbally and in a letter, working closely with
residents to ensure willing participation. Residents were always informed when I would be
accessing their property and were given the option to discontinue their involvement at any
time.
Although I use first-person, singular, throughout this thesis in describing the work
conducted, I would like to acknowledge that these studies represent, and will be published, as
collaborative works, as I could not have conducted this research otherwise. While I was the
lead researcher in the design, execution, analysis, and writing of the work presented here,
several others contributed significantly to be included in authorship. As such, the anticipated
authorship of the two main data chapters of the thesis (Chapters 2 and 3) will be, for Chapter
2, Tanya Luszcz and Elizabeth A. Gow, and for Chapter 3, Tanya Luszcz, DG Blair, Anna
Skurikhina, and Elizabeth A. Gow.

v
TABLE OF CONTENTS
ABSTRACT............................................................................................................................. II
PREFACE...............................................................................................................................IV
TABLE OF CONTENTS........................................................................................................ V
LIST OF TABLES ............................................................................................................. VIII
LIST OF FIGURES ...............................................................................................................IX
ACKNOWLEDGEMENTS .............................................................................................. XIII
CHAPTER 1: GENERAL INTRODUCTION ...................................................................... 1
1.1 SEASONALITY ................................................................................................................... 1
1.2 DOMESTIC CATS ................................................................................................................ 2
1.3 STUDY REGION ................................................................................................................. 6
1.4 BIRDS AND HABITAT ......................................................................................................... 7
1.5 THESIS OVERVIEW ............................................................................................................ 9
1.6 LITERATURE CITED ......................................................................................................... 12
CHAPTER 2: YEAR-ROUND DIFFERENCES IN BIRD ABUNDANCE, RICHNESS,
AND COMMUNITY COMPOSITION ACROSS AN URBAN TO RURAL GRADIENT
................................................................................................................................................. 19
2.1 INTRODUCTION ............................................................................................................... 19
2.2 METHODS ....................................................................................................................... 23
2.2.1 Study Area ............................................................................................................... 23
2.2.2 Study Design ........................................................................................................... 25
2.2.3 GIS Analysis ............................................................................................................ 28
2.2.4 Unmarked Analysis Design ..................................................................................... 30
2.2.5 Abundance Estimates .............................................................................................. 31
2.2.6 Species Richness ..................................................................................................... 32
2.2.7 Seasonal Analyses ................................................................................................... 33
2.2.8 NMDS Analysis ....................................................................................................... 33
2.3 RESULTS.......................................................................................................................... 34
2.3.1 Mean Abundance and Richness Across Seasons .................................................... 35
2.3.2 Seasonal Abundance and Richness Across Land Classes....................................... 35
2.3.3 Species Communities and Habitat Types ................................................................ 40
2.4 DISCUSSION .................................................................................................................... 42
2.5 LITERATURE CITED ......................................................................................................... 49

vi
CHAPTER 3: SEASONAL VARIATION IN DOMESTIC CAT (FELIS CATUS)
ABUNDANCE, POPULATION SIZE, AND HABITAT ASSOCIATIONS ACROSS A
MOSAIC OF HABITATS ..................................................................................................... 59
3.1 INTRODUCTION ............................................................................................................... 59
3.2 METHODS ....................................................................................................................... 62
3.2.1 Study Area ............................................................................................................... 62
3.2.2 Study Design ........................................................................................................... 65
3.2.3 GIS Analysis ............................................................................................................ 70
3.2.4 Statistical analysis .................................................................................................. 71
3.2.5 Model Selection and Seasonal Abundance Estimates ............................................ 73
3.2.6 Seasonal Analyses ................................................................................................... 74
3.2.7 Predicting Total Cat Abundance ............................................................................. 74
3.3 RESULTS.......................................................................................................................... 76
3.3.1 Detection probability .............................................................................................. 79
3.3.2 Local Cat Abundance ............................................................................................. 79
3.3.3 Seasonal and land class analyses ........................................................................... 80
3.3.4 Estimating wandering domestic cat population size .............................................. 82
3.4 DISCUSSION .................................................................................................................... 90
3.4.1 Management implications ....................................................................................... 95
4.5 LITERATURE CITED ......................................................................................................... 97
CHAPTER 4: CONCLUSIONS, FUTURE DIRECTIONS, AND MANAGEMENT AND
CONSERVATION IMPLICATIONS................................................................................. 104
4.1 DISCUSSION AND FUTURE DIRECTIONS ......................................................................... 104
4.2 MANAGEMENT AND CONSERVATION IMPLICATIONS ...................................................... 106
4.3 CONCLUSION................................................................................................................. 109
4.4 LITERATURE CITED ....................................................................................................... 110
APPENDIX 1........................................................................................................................ 114
APPENDIX 2........................................................................................................................ 114
APPENDIX 3........................................................................................................................ 115
APPENDIX 4........................................................................................................................ 116
APPENDIX 5........................................................................................................................ 116
APPENDIX 6........................................................................................................................ 117
APPENDIX 7........................................................................................................................ 123
APPENDIX 8........................................................................................................................ 128
APPENDIX 9........................................................................................................................ 129

vii
APPENDIX 10...................................................................................................................... 130
APPENDIX 11 ...................................................................................................................... 131
APPENDIX 12...................................................................................................................... 132
APPENDIX 13...................................................................................................................... 133
APPENDIX 14...................................................................................................................... 133
APPENDIX 15...................................................................................................................... 134
APPENDIX 16...................................................................................................................... 135
APPENDIX LITERATURE CITED .................................................................................. 137

viii
LIST OF TABLES
Table 2.1. Hypotheses and associated predictor variables used to explain local bird relative
abundance and species richness in the south Okanagan Valley, B.C. Source(s) indicate a study
or studies that provide evidence to support the hypothesis......................................................22
Table 2.2. Description of detection and site covariates used to estimate relative abundance
and species richness of birds in the south Okanagan Valley, B.C. Detection covariates
influenced the detection of species during point counts and site covariates describe the site
where point counts occurred....................................................................................................29
Table 2.3. Gdistsamp models used to estimate bird abundance for each season.....................32
Table 3.1. Description of detection and site covariates used to estimate relative abundance of
cats in the south Okanagan Valley, B.C. Detection covariates influenced the detection of
wandering cats and site covariates describe the area around the trail camera.........................64
Table 3.2. Seasonal summary statistics, including the number of sites where wandering cats
were detected, organized by land class, and the total number of individually identified
wandering cats identified using the trail camera during each season......................................77
Table 3.3. Top models for detection probability (p) when assessing the number of domestic
cats photographed at trail camera sites during all five seasons in south Okanagan Valley, B.C.
Global model included the co-variates: temp, date and precip (precipitation). Abundance
estimates (λ) are held constant (λ (.)).......................................................................................78
Table 3.4. Top models for abundance estimates (λ) when assessing the number of domestic
cats photographed at trail camera sites during all five seasons in south Okanagan Valley, B.C.
Global model included the co-variates: natural, vineyard, orchard, field, water, roads, and
buildings. Detection probability (p) is held constant (p(.))......................................................81
Table 3.5. Variables for each season used for predicting cat abundance. Natural is wetlands,
desert, and forests combined. Density was calculated using the study area of 141.2km2.......86

ix
LIST OF FIGURES
Figure 1.1. A) Wandering cat in the south Okanagan Valley. B) The same wandering cat seen
on trail camera image with a bird in its mouth. C) Trail cameras used in this study to capture
images of wandering cats. D) Trial camera strapped to a tree and secured with a lock............5
Figure 1.2. Representation of habitat types evaluated during this study, including urban (A
and B; e.g., residential areas of Osoyoos and Oliver), natural (C; e.g., Haynes’ Lease
Ecological Reserve, Osoyoos), peri-urban (D and E; e.g., low-density residential areas
around Okanagan Falls and Oliver), and agricultural (F; e.g., Montakarn Estate Winery).......7
Figure 2.1. Map of the study area in the south Okanagan Valley showing trail camera
locations in Okanagan Falls, Oliver, and Osoyoos. Location colours are classified by the
following land classes: Agriculture (AG), Natural (NA), Peri-urban (PU), and Urban (UR).
The red boarder represents the study area that was used to study wandering domestic cats and
the study area boundaries (Chapter 3). The inset map in the top right corner shows a red
circle around the south Okanagan Valley, located in the south-central area of British
Columbia, Canada. Map is in EPSG:3005 coordinate system (NAD83, BC Albers
projection)............................................................................................................................... 24
Figure 2.2. A) Mean seasonal abundance displaying non-significance between fall migration
(green) and early and late non-breeding season (blue and purple). Mean abundance was
highest during the breeding season (gold; 120 ± SE 29.4) and spring migration (pink; 104 ±
SE 31.8), and lowest throughout the fall migration and early and late non-breeding seasons
(fall migration 68.5 ± 28.4, early non-breeding 61.8 ± 21.4, late non-breeding 58.6 ± 31.3;
mean bird abundance ± SE). B) Mean seasonal richness, showing non-significance between
fall migration (green) and early non-breeding season (blue). Species richness was highest
during the breeding season (gold; 7.98 ± SE 2.64) and spring migration (pink; 6.76 ± SE
1.99) and lowest in the late non-breeding season (purple). The horizontal black line within
the box represents the median value, and the box represents 50% of the data, from the 25th to
75th percentile of each groups’ distribution of values, which is the interquartile range (IQR).
The whiskers extend to the highest and lowest points within 1.5 times the IQR and individual
dots are outliers beyond the whiskers......................................................................................37
Figure 2.3. Box plots showing the estimated relative abundance during the early nonbreeding seasons (A and C) and late non-breeding season (B and D). A) Estimated abundance
was highest in urban areas (75.4 ± SE 3.96) though not significantly different from peri-urban
(62.3 ± SE 4.37), and agricultural (70.1 ± SE 3.98) areas. B) Estimated abundance was
highest in peri-urban (80.8 ± SE 4.42) and urban (74.3 ± SE 3.99) areas. C) Species richness
was highest in urban (5.50 ± SE 0.378) and peri-urban (5.46 ± SE 0.420) areas, though not
significantly different from natural habitats (4.25 ± SE 0.372). D) Species richness was

x
highest in peri-urban (5.03 ± SE 0.427) and urban (4.79 ± 0.383) areas. See Table A6.3 and
Table A6.4 for P-values. The horizontal black line within the box represents the median
value, and the box represents 50% of the data, from the 25th to 75th percentile of each groups’
distribution of values, which is the interquartile range (IQR). The whiskers extend to the
highest and lowest points within 1.5 times the IQR and individual dots are outliers beyond
the whiskers..............................................................................................................................38
Figure 2.4. Box plots showing estimated abundance and species richness during the spring
migration (A and D), breeding season (B and E) and fall migration (C and F). A) Estimated
abundance was highest in urban (132.5 ± SE 4.30) and peri-urban (127.9 ± SE 3.95) areas
during the spring migration. B) Estimated abundance was highest in peri-urban (142.2 ± SE
4.32) and urban (125.1 ± SE 3.96) areas during the breeding season, though similar to natural
habitats (117.9 ± SE 3.90). C) During the fall migration, estimated abundance was highest in
peri-urban (107.8 ± SE 4.32) and urban (71.9 ± SE 3.96) areas. D) Species richness was not
significantly different between any land classes during the spring migration. E) Species
richness was highest in natural (9.47 ± SE 0.372) and peri-urban (9.10 ± SE 0.413) areas
during the breeding season. F) During the fall migration, species richness was lowest in
urban areas (4.65 ± SE 0.378) and not significantly different between any other land classes.
See Table A6.3 and Table A6.4 for P-values. The horizontal black line within the box
represents the median value, and the box represents 50% of the data, from the 25th to 75th
percentile of each groups’ distribution of values, which is the interquartile range (IQR). The
whiskers extend to the highest and lowest points within 1.5 times the IQR and individual dots
are outliers beyond the whisker................................................................................................39
Figure 2.5. NMDS comparing species habitat associations across A) the spring migration B)
breeding season, C) fall migration, D) early non-breeding season, and E) late non-breeding
season. NMDS1 is on the X axis and NMDS2 is on the Y axis. Additional information can be
found in Appendix 7 and remaining NMDS comparisons (1 vs. 3 and 2 vs. 3) can be found in
Figure A7.1. Birds are represented by four-letter banding codes and the corresponding
scientific and common names can be found in Table A6.1......................................................41
Figure 2.6. NMDS comparing species compositions across urban (UR), peri-urban (PU),
agricultural (AG), and natural (NA) areas. I also compared four seasons: A) the spring
migration B) breeding season, C) fall migration, D) early non-breeding season, and E) late
non-breeding season. Each point in the figure represents a point count and its location on the
NMDS is based on the birds that were observed during that point count. Points that are closer
together are more similar in species composition and points that are further apart are
dissimilar in their species composition. The circles around the points show a 95% confidence
interval. I can see a distinct separation between the natural and urban areas, which suggests

xi
that urban and natural habitats are comprised of different species. Remaining NMDS
comparisons (1 vs. 3 and 2 vs. 3) can be found in Figure A7.2...............................................42
Figure 3.1. Map of the study area in the south Okanagan Valley showing trail camera
locations in Okanagan Falls, Oliver, and Osoyoos. Location colours are classified by the
following land classes: Agriculture (AG), Natural (NA), Peri-urban (PU), and Urban (UR).
The red boarder represents the study area that was used to study wandering cats based on the
perimeter of the grid. The inset map in the top right corner shows a red circle around the
south Okanagan Valley, located in the south-central area of British Columbia, Canada. Map is
in EPSG:3005 coordinate system (NAD83, BC Albers projection)........................................66
Figure 3.2. Domestic cat coat patterns, as seen on the trail cameras. You can see that each cat
has a different coat pattern, making them individual identifiable............................................70
Figure 3.3. Example of 200-meter buffer divided into habitat variables. Variables within each
200 m better were used in the analyses....................................................................................71
Figure 3.4. Box plots showing the estimated abundance A) in different land classes
throughout the year and B) during each season. All land classes were significantly different
(p < 0.05 for all seasonal comparisons). Median abundance of cats per site was highest in
urban habitats (4.1), followed by peri-urban habitats (1.8), agricultural habitats (1.2), and
natural habitats (0.4). For seasonal comparisons, all land classes were not significantly
different (p > 0.05), except between early and late winter, which were significantly different
(p < 0.05). The median abundance of domestic cats per site was highest in the early winter
(2.1) and lowest in the late winter (1.3). The horizontal black line within the box represents
the median value, and the box represents 50% of the data, from the 25th to 75th percentile of
each groups’ distribution of values, which is the interquartile range (IQR). The whiskers
extend to the highest and lowest points within 1.5 times the IQR and individual dots are
outliers beyond the whiskers....................................................................................................84
Figure 3.5. Box plots showing the estimated abundance for each land class separated by
season. Mean abundance was significantly highest in urban habitats (UR) consistently across
each season (p < 0.05). Mean abundance in peri-urban (PU) and agricultural (AG) sites are
not significantly different in any season (p > 0.05). Mean abundance is significantly lowest in
natural (NA) habitats during each season (p < 0.05) except during the fall, where agricultural
and natural sites do not significantly differ (p > 0.05). The horizontal black line within the
box represents the median value, and the box represents 50% of the data, from the 25th to 75th
percentile of each groups’ distribution of values, which is the interquartile range (IQR). The
whiskers extend to the highest and lowest points within 1.5 times the IQR and individual dots
are outliers beyond the whiskers..............................................................................................85

xii

Figure 3.6. Habitat variables used in top models to estimate wandering cat populations in the
south Okanagan Valley, B.C. Maps show each grid square and the proportion of natural
habitats (A), vineyards (B), roads (C), buildings (D), fields (E) and water (F). Darker colours
represent higher values, while lighter colours represent lower values....................................87
Figure 3.7. Predicted abundance of wandering cats based on the proportion of natural habitat
in each cell during each season: A) spring, B) summer, C) Fall, D) Early winter, and E) Late
winter. Darker areas have higher densities of wandering cats. While maps look the same, the
local abundance per grid cell differs across seasons................................................................88
Figure 3.8. Predicted abundance of wandering cats based on the proportion of vineyards in
each cell during four seasons: A) spring, B) summer, C) Fall, and D) Early winter. Darker
areas have higher densities of wandering cats.........................................................................89

xiii
ACKNOWLEDGEMENTS
This thesis would not have been possible without the support and contributions of many
individuals and organizations, and I am deeply grateful to each of them. First and foremost, I
would like to express my heartfelt gratitude to my co-supervisor, Dr. Elizabeth Gow. Your
guidance, expertise, and unwavering support throughout this project have been invaluable.
You’ve taught me everything I know about birds and cats, and I cannot thank you enough for
your constant encouragement and advice. I would not have been able to complete this thesis
without you! I would also like to thank my co-supervisor Dr. Ken Otter and committee
member Dr. Nicola Koper, for their insightful feedback and support during this process. A
special thanks to Dr. Erin Baerwald and Kate Logan from UNBC for their help and
encouragement during my Masters.
I am incredibly grateful to everyone with the Stewardship Centre for British Columbia for
playing a pivotal role in launching this project and providing funding. Thank you, DG Blair,
for your vision and for making this research possible! A special thanks to Anna Skurikhina
for your behind-the-scenes work, your passion in outreach, and for being a friend when I was
new to the area. Additionally, I would like to give a huge thank you to Tanya Luszcz with the
Canadian Wildlife Services for funding my fifth season, which was crucial in completing a
year-round study. Your guidance and support throughout the project have been invaluable. I
also extend my appreciation to everyone who helped with fieldwork, with a special shout-out
to David Bell who did an awesome job carrying on the project during the winter. Finally, I
would also like to thank the Mitacs Accelerate Program for their financial support of this
project.
I am also deeply thankful for the organizations and government agencies that facilitated the
permits for conducting research in protected areas. Thank you to The Nature Trust of BC,
especially the Okanagan Field Crew, The Osoyoos Desert Centre, BC Parks, and the BC
Ministry of Forests. My sincere thanks also go to vineyards that hosted trail cameras: Arterra
Wines Canada, Montakarn Estate Winery, Hidden Chapel Winery, River Stone Estate Winery,
Red Horses Vineyard, and Burrowing Owl Estate Winery. Additionally, I am immensely
grateful to the many residents who allowed us to place trail cameras on their properties –
your generosity made this research possible, and it was a pleasure getting to know so many of
you!
Finally, I would like to express my deepest thanks to my friends and family. To my lab mates,
thank you for your constant support and encouragement. To my parents and sister, I
appreciate you guys so much and I am so grateful to have you as my family. Your love,
support, and encouragement have been a source of strength. To my partner, Andrew, thank
you for your patience, understanding, and for helping me survive the stressful times. I
couldn’t have done this without all of you.

1
Chapter 1: General Introduction
1.1 Seasonality
Most organisms, including animals, closely follow the annual cycle of environmental
fluctuations or changes, known as seasonality (Lisovski et al., 2021; Williams et al., 2017).
These environmental changes may include day length, precipitation, and temperature, which
ultimately influences resource availability (Lisovski et al., 2021). For animals, these seasonal
changes promote changes in hormones that can help individuals prepare for, cope with, and
adapt to seasonal changes (Bronson, 2009; Chen et al., 2020). Examples of some adaptations
driven by hormone changes are mating or breeding, migration, and moulting (Chen et al.,
2020). Additionally, as resource availability changes, animals will move to where resources
are abundant, which can drive their seasonal behaviour and movements (Huston &
Wolverton, 2009). These seasonal movements mean that the abundance and composition of
species or individuals within a given area change throughout the year (Briese & Smith, 1974;
Cameron et al., 2018; Scully et al., 2018). Seasonal changes in abundance and habitat
selection have been studied in a range of animals, such as small mammals (e.g., shrews,
Briese & Smith, 1974), larger mammals (e.g., seals, Cameron et al., 2018), domestic cats
(Cove et al., 2023; Horn et al., 2011), and birds (Blair, 1996; Ortega-Álvarez & MacGregorFors, 2009; Tu et al., 2020). Every year, over 2.6 billion birds fly from the Southern USA,
and Central and South America, to Canada during the spring and summer migration to follow
seasonal changes in resource abundance (Dokter et al., 2018; Huston & Wolverton, 2009).
Birds that can synchronize their movements with environmental patterns in their
surroundings, such as those related to food and weather, tend to experience greater success in
terms of survival and reproduction (Both et al., 2006). Both long- and short-distance migrants
aim to arrive when resources are at their peak (Huston & Wolverton, 2009), or they could

2
face a mismatch in resource abundance versus resource needs (Jones & Cresswell, 2010).
Birds face many dangers during migration such as bad weather, collisions with glass or
vehicles, pathogen transmission, and higher risk of predation (Jourdain et al., 2007; Lao et
al., 2023; Lind & Cresswell, 2006; Van Den Broeke & Gunkel, 2021). Mortality during
migration may be as much as fifteen times higher than during stationary periods for many
species, such as the black-throated blue warbler (Dendrioica caerulescens, Lind & Cresswell,
2006).
One predator that may influence mortality risk of birds differently throughout the year
are domestic cats. Globally, predation on birds by cats is well documented (Baker et al.,
2005, 2008; Beckerman et al., 2007). Predation rates are typically highest in the spring and
summer months when there are a high number of nestlings and fledglings that have less
experience evading predators (Baker et al., 2005). But predation risk may be proportionately
higher to individual birds in the late fall or winter if cat abundances do not change, but bird
abundances are lower (Blancher, 2013). Cat presence can also indirectly affect nest
productivity through a decrease in nest provisioning rates, in addition to predation rates by
predators (Bonnington et al., 2013).
1.2 Domestic cats
Domestic cats (Felis catus) descended from a wild ancestor, the African wildcat (Felis
silvestris lybica; Driscoll et al., 2007), a well-adapted and widespread predator, and have
retrained many of the physiological and behavioural traits that make them effective hunters
(Bradshaw, 2006). Cats have a short history of domestication and were domesticated to serve
as pest controllers only ~10,000 years ago and have more recently been termed “wild
companions” due to their dual roles in society as companion animals and pest controllers

3
(Crowley et al., 2020). Despite domestication, cats have retained their ancestral hunting
behaviours, enabling them to adapt to a variety of environments into which they have been
introduced (Bradshaw, 2006; Canadian Federation of Humane Societies [CFHS], 2017).
These innate hunting abilities allow cats to capture a variety of wild animals, posing serious
global threats to small vertebrate populations (Loss et al., 2013; Trouwborst et al., 2020).
Domestic cats are a common companion animal in Canada and the USA (CFHS,
2017; The Humane Society of the United States [HSUS], 2019), but unlike other companion
animals, their life-styles vary depending on their access to the outdoors. The classification of
domestic cats varies across studies, but in this thesis I will follow the definitions outlined by
(Crowley et al., 2020). Owned cats that are fully confined indoors are considered to have no
impact on birds or other wildlife, whereas those with indoor-outdoor access can wander
unsupervised and have minimal ecological impact. In contrast, unowned cats pose a greater
threat to wildlife, as they may rely on hunting for survival (van Heezik et al., 2010). Feral
cats fall within the category of unowned cats and are often self-sustaining, surviving
independently of human involvement. For this study, the term “wandering cats” refers to any
cat that is outdoors and unsupervised or unconstrained, encompassing both owned indooroutdoor and unowned cats. This term was found to be the most accepted by multiple groups
of people and stakeholders based on focus groups run by Nature Canada’s Cats and Birds
Campaign (Gow et al., 2024). Outdoor access provides cats with mental stimulation, physical
activity, and opportunities to engage in natural behaviours, but it also exposes them to
welfare risks (Rochlitz, 2005), highlighting the need for effective management strategies.
When cats are allowed to wander unconstrained outside, they may experience increased risks
of injury and illness through predation by other animals, vehicle collisions, and disease

4
transmission of illnesses such as toxoplasmosis (reviewed in Tan et al., 2020). For these
reasons, it is often recommended that cats are kept indoors, but many owners still allow their
cats outside.
Various groups (e.g., veterinarians, shelters, humane societies, conservation
organizations, policy makers, and members of the public) are concerned about issues
surrounding wandering cats (detailed in Saunders et al., 2021; Wald & Peterson, 2020) and
find common ground in the desire to curtail the number of wandering cats. Unowned cats
may be fed by people, but their movements are not constrained. Unlike many other
companion or domesticated animals, cats can persist, survive, and reproduce without any
human intervention, and thus feral and unmanaged outdoor populations of cats are common
(Legge et al., 2017; Tan et al., 2020). Quantifying the number of wandering domestic cats
allows us to understand anthropogenic threats facing birds and other species of concern
(Ferreira et al., 2011; Flockhart et al., 2016). It has also been a metric that is frequently
requested by municipal government officials, and from non-profit organizations to identify
where to direct limited resources to programs such as educational campaigns. Using this
information, I can also create a baseline for wildlife and cat populations that helps to inform
future management decisions and to assess the effectiveness of implemented management
strategies. This thesis arose by a need and desire from key interested parties (i.e., the
Stewardship Centre for British Columbia and Environment and Climate Change Canada) for
a detailed study of the abundance of birds and cats in the southern Okanagan Valley (see
below) to help direct educational campaigns and future funding and initiatives for other cat
management actions in the region.

5

Figure 1.1. A) Wandering cat in the south Okanagan Valley. B) The same wandering cat seen
on trail camera image with a bird in its mouth. C) Trail cameras used in this study to capture
images of wandering cats. D) Trial camera strapped to a tree and secured with a lock.
Some methods for estimating local cat abundance are walking line transects and
resident surveys, but they are sensitive to bias because they can limit the number of cats
detected (Flockhart et al., 2016; Hand, 2019). In contrast, the use of trail cameras (Elizondo
& Loss, 2016) can reduce bias seen in other methods and are increasingly being used as a
common tool to assess population sizes and abundance of various animals around the world
(Jhala et al., 2011; Sollmann et al., 2011; Tanwar et al., 2021), including wandering cats
(Clyde et al., 2022; Coe et al., 2021; Cove et al., 2023; Gow et al., 2024). Trail cameras offer
several advantages over other methods as they can detect both owned and unowned cats on
private property (e.g., in people’s yards) and during their peak activity periods (i.e., at night),
and are particularly effective in rural areas. Recently, trail cameras were used to: assess the
effectiveness of a trap-neuter-release management program for feral cats (Coe et al., 2021);
estimate local abundance and habitat associations of wandering cats (Clyde et al., 2022; Gow
et al., 2024); and, estimate occupancy and density of wandering cats in Washington, D.C

6
(Cove et al., 2023). For these reasons, I used trail cameras in this study to estimate the
abundance of wandering cats in the south Okanagan Valley study (Figure 1.1).

1.3 Study Region
Located in the south-central region of British Columbia, the south Okanagan Valley
(between Okanagan Falls and Osoyoos) is one of the most endangered ecosystems in Canada
and is part of the bunchgrass biogeoclimatic zone that is characterized by grassland, shrubsteppe, wetlands, and dry open forests consisting of ponderosa pines, cottonwoods, and
Douglas firs (Alldritt-McDowell et al., 1998). Its unique topography features a valley flanked
by large mountains, making the south Okanagan Valley part of an important avian migratory
route. These distinct habitats and topography also make it a significant area, supporting some
of the highest diversity of birds in Canada, with over 330 recorded bird species, including 24
provincially listed species at risk, 22 federally listed species at risk under the Species at Risk
Act (eBird, 2022; Okanagan Similkameen Stewardship [OSS], 2023). Many birds funnel
through the valley during migration, including long-distance, short-distance, and altitudinal
migrants. The temperate climate also ensures an abundance of food and other resources yearround, supporting many resident bird species that inhabit the south Okanagan Valley
throughout the entire year. Given its ecological significance and vulnerability, there is an
ecological need for research in this area to understand the challenges faced by bird
populations and where birds may be at greatest risk of domestic cats to help target
management actions.
I assessed four land class types during this study: urban, peri-urban, agricultural, and
natural (Figure 1.2). While overlap exists between these four land classes, the following
definitions outline the main attributes associated with each and classify them based on the

7
amount of anthropogenic disturbance. Urban habitats have the highest level of anthropogenic
disturbance and include a matrix of landscapes, such as pavement, buildings (residential and
commercial), green spaces, and habitat patches (e.g., transportation corridors, riverbanks, and
parks; Swanwick et al., 2003). Peri-urban habitats consisted of clusters of houses surrounded
by agricultural (e.g., orchards, vineyards, and crops) and/or natural (e.g., desert, forest,
riparian) habitat with houses often spaced apart (approximately 25-50 meters). Agricultural
land was used for agriculture (such as vineyards and orchards) or other farming practices,
targeting smaller vineyards/orchards that were <50 hectares. Natural habitats had the lowest
level of anthropogenic disturbance and included any protected habitats conserved
provincially, federally, or privately.

Figure 1.2. Representation of habitat types evaluated during this study, including urban (A
and B; e.g., residential areas of Osoyoos and Oliver), natural (C; e.g., Haynes’ Lease
Ecological Reserve, Osoyoos), peri-urban (D and E; e.g., low-density residential areas
around Okanagan Falls and Oliver), and agricultural (F; e.g., Montakarn Estate Winery).
1.4 Birds and Habitat
The abundance and diversity of birds in the south Okanagan Valley are shaped by
their ability to adapt to different levels of urbanization and resource availability across
habitats. Urbanization plays a large role in the abundance and richness of bird species, as

8
urban areas provide more human-provided resources such as food, water, perches, and
nesting places (Emlen, 1974; Mills et al., 1989). Some bird species take advantage of these
resources provided in urban environments and can be labeled as “urban exploiters” (Blair,
1996). Across bird species, few are urban exploiters, which results in an overall reduction in
the diversity of species in urban areas (Blair, 1996; Isaksson, 2018). Those who can exploit
the urban environment often thrive, leading to high abundances of these species (Blair, 1996).
Other species are “suburban adaptable” and can take advantage of human-provided resources
that come with urbanization, while exploiting the additional natural resources provided in
peri-urban areas. Lastly, many species are sensitive to urbanization and these “urban
avoiders” are at their highest densities in natural areas.
Areas with lower levels of urbanization, specifically peri-urban and natural areas,
tend to increase diversity and abundance of resources available to birds, and therefore
support a wider range of niches (Blair, 1996). Urban areas lack resource diversity, thereby
supporting few urban exploiters (Blair, 1996; Emlen, 1974). Other areas with lower levels of
urbanization include peri-urban, agricultural, and natural areas, each being able to support a
range of bird species. Agricultural areas in the south Okanagan Valley typically provide
limited resources to birds due to simplification of vegetation (e.g., monoculture) and reduced
insect populations due to pesticide use (Bain et al., 2020). This limits the number of birds that
can use this habitat type throughout the year, but there are many species that can take
advantage of the abundance of fresh fruit during the spring, summer, and fall. Natural and
peri-urban areas have higher resource availability and diversification, with the highest
abundance of resources available during the spring and summer (Blair, 1996; Jokimäki,
1999). Because of this, species richness and diversity are often highest in natural and peri-

9
urban areas (because they often border natural areas), and lowest in urban habitats
(Beissinger & Osborne, 1982; Emlen, 1974; Lancaster & Rees, 1979).
Abundance and diversity of birds in different habitats will vary between spring and
winter due to changes in resource availability. During the spring and summer, resources are
highest in peri-urban and natural areas leading to higher densities of species. While urban
areas still provide abundant food resources, the quality of this food, compared to that which
is available to birds naturally, enhances only a few feeding niches (Lancaster & Rees, 1979).
When winter arrives, the increasing snow will reduce vegetation and prey availability, which
will impact species distribution as snow levels change in the mountains and in the valley
(Johnson & Sherry, 2001; Leech & Crick, 2007). In the winter, most species in the Okanagan
Valley are residents, along with a few altitudinal migratory species. Additionally, during the
winter, urban and peri-urban environments with higher human population densities provide
food for birds, increasing abundance throughout the winter (Emlen, 1974; Lancaster & Rees,
1979). Many species will take advantage of the food and shelter provided by humans during
the winter when options are limited.
1.5 Thesis Overview
Globally, predation on birds by cats is well documented (Baker et al., 2005, 2008; Blancher,
2013; Krauze-Gryz et al., 2019; Li et al., 2021; Sedano-Cruz, 2022; van Heezik et al., 2010;
Woinarski et al., 2017). By quantifying the local abundance and numbers of wandering
domestic cats, we can assess the potential risk that birds face from cats and identify where
and when these threats may be highest. This study, developed in partnership with the
Stewardship Centre for British Columbia and the Canadian Wildlife Service (Environment
and Climate Change Canada), provides science-based information to guide targeted

10
education campaigns, other potential cat management options, and bird-focused conservation
and management, that overall aim to benefit people, birds, and cats. By examining both birds
(Chapter 2) and wandering cats (Chapter 3) in this thesis, I am able to identify where there is
overlap in local abundances of wandering cats, and high bird abundance and richness, and
how these patterns vary across the habitats and seasons. While more study and analysis
beyond the scope of this thesis would be needed to understand the true risks and impacts that
birds face from cats in this region, I hope this thesis creates an initial, and valuable, starting
point to identify potential predation pressure and risk to birds within the study region that can
be applied more broadly. For example, if peri-urban areas have a high cat abundance and
high bird abundance and species richness, this may indicate that focusing cat management or
education in these habitats may provide the greatest benefit compared to urban areas, which I
expect to have high relative cat abundance, but low relative bird abundance and species
richness. Thus, through this type of approach that highlights areas and seasons of overlap
between birds and cats, this thesis will contribute to helping to better inform conservation and
management efforts that can then be used by a wide diversity of interested parties and
identify where to direct limited resources.
I examined the year-round trends in abundance, species richness and community
structure of birds and the abundance of wandering domestic cats in the south Okanagan
valley, British Columbia, Canada, to understand where wandering domestic cats pose the
greatest risks to birds, which has never been looked at in this type of landscape. My research
was conducted between Okanagan Falls and Osoyoos, across a variety of habitats, including
urban, peri-urban, agricultural, and natural areas across five seasonal periods relevant to birds
(spring migration, breeding, fall migration, early non-breeding and late non-breeding). This
study was developed in partnership with the Stewardship Centre for British Columbia and

11
Environmental and Climate Change Canada (Wildlife Research Division and Canadian
Wildlife Services) to test hypotheses that will provide science-based information for targeted
educational campaigns. Chapter 2 will focus on birds, with the overall objective of
identifying how bird species richness and abundance varies across seasons and determine
patterns of habitat associations among birds throughout the entire year. Chapter 3 will focus
on wandering domestic cats, with the overall objective of identifying how cat abundance and
population size varies seasonally and across habitats. Chapter 4 discusses the connections
between birds (Chapter 2) and cats (Chapter 3).

12
1.6 Literature Cited
Alldritt-McDowell, J., Coupé, R., & Meidinger, M. (1998, July). The Ecology of the
Bunchgrass Zone. Ministry of Forests, British Columbia.
https://www.for.gov.bc.ca/hfd/pubs/docs/Bro/bro54.pdf
Bain, G. C., MacDonald, M. A., Hamer, R., Gardiner, R., Johnson, C. N., & Jones, M. E.
(2020). Changing bird communities of an agricultural landscape: Declines in arboreal
foragers, increases in large species. Royal Society Open Science, 7(3), 200076.
https://doi.org/10.1098/rsos.200076
Baker, P. J., Bentley, A. J., Ansell, R. J., & Harris, S. (2005). Impact of predation by domestic
cats Felis catus in an urban area. Mammal Review, 35(3–4), 302–312.
https://doi.org/10.1111/j.1365-2907.2005.00071.x
Baker, P. J., Molony, S. E., Stone, E., Cuthill, I. C., & Harris, S. (2008). Cats about town: Is
predation by free-ranging pet cats Felis catus likely to affect urban bird populations? Ibis,
150(s1), 86–99. https://doi.org/10.1111/j.1474-919X.2008.00836.x
Beckerman, A. P., Boots, M., & Gaston, K. J. (2007). Urban bird declines and the fear of cats.
Animal Conservation, 10(3), 320–325. https://doi.org/10.1111/j.1469-1795.2007.00115.x
Beissinger, S. R., & Osborne, D. R. (1982). Effects of Urbanization on Avian Community
Organization. The Condor, 84(1), 75–83. https://doi.org/10.2307/1367825
Blair, R. B. (1996). Land Use and Avian Species Diversity Along an Urban Gradient.
Ecological Applications, 6(2), 506–519. https://doi.org/10.2307/2269387
Blancher, P. (2013). Estimated Number of Birds Killed by House Cats (Felis catus) in
Canada. Avian Conservation and Ecology, 8(2), art3. https://doi.org/10.5751/ACE-00557080203
Bonnington, C., Gaston, K. J., & Evans, K. L. (2013). Fearing the feline: Domestic cats
reduce avian fecundity through trait‐mediated indirect effects that increase nest predation
by other species. Journal of Applied Ecology, 50(1), 15–24. https://doi.org/10.1111/13652664.12025
Both, C., Bouwhuis, S., Lessells, C. M., & Visser, M. E. (2006). Climate change and
population declines in a long-distance migratory bird. Nature, 441(7089), 81–83.
https://doi.org/10.1038/nature04539

13
Bradshaw, J. W. S. (2006). The evolutionary basis for the feeding behavior of domestic dogs
(Canis familiaris) and cats (Felis catus). The Journal of Nutrition, 136(7 Suppl).
https://doi.org/10.1093/jn/136.7.1927S
Briese, L. A., & Smith, M. H. (1974). Seasonal Abundance and Movement of Nine Species of
Small Mammals. Journal of Mammalogy, 55(3), 615–629.
https://doi.org/10.2307/1379550
Bronson, F. H. (2009). Climate change and seasonal reproduction in mammals. Philosophical
Transactions of the Royal Society B: Biological Sciences, 364(1534), 3331–3340.
https://doi.org/10.1098/rstb.2009.0140
Cameron, M. F., Frost, K. J., Hoef, J. M. V., Breed, G. A., Whiting, A. V., Goodwin, J., &
Boveng, P. L. (2018). Habitat selection and seasonal movements of young bearded seals
(Erignathus barbatus) in the Bering Sea. PLOS ONE, 13(2), e0192743.
https://doi.org/10.1371/journal.pone.0192743
Canadian Federation of Humane Societies [CFHS]. (2017). Cats in Canada 2017: A five-year
review of cat overpopulation. https://www.g2z.org.au/pdf/CFHS-Cats_In_Canada_2017FINAL_LR.pdf
Chen, J., Okimura, K., & Yoshimura, T. (2020). Light and Hormones in Seasonal Regulation
of Reproduction and Mood. Endocrinology, 161(9), bqaa130.
https://doi.org/10.1210/endocr/bqaa130
Clyde, H. E., Norris, D. R., Lupton, E., & Gow, E. A. (2022). Estimating habitat
characteristics associated with the abundance of free-roaming domestic cats across the
annual cycle. Wildlife Research, 49(7), 583–595. https://doi.org/10.1071/WR20205
Coe, S. T., Elmore, J. A., Elizondo, E. C., & Loss, S. R. (2021). Free-ranging domestic cat
abundance and sterilization percentage following five years of a trap–neuter–return
program. Wildlife Biology, 2021(1). https://doi.org/10.2981/wlb.00799
Cove, M. V., Herrmann, V., Herrera, D. J., Augustine, B. C., Flockhart, D. T. T., & McShea,
W. J. (2023). Counting the Capital’s cats: Estimating drivers of abundance of free-roaming
cats with a novel hierarchical model. Ecological Applications, 33(2), e2790.
https://doi.org/10.1002/eap.2790

14
Crowley, S. L., Cecchetti, M., & McDonald, R. A. (2020). Our Wild Companions: Domestic
cats in the Anthropocene. Trends in Ecology & Evolution, 35(6), 477–483.
https://doi.org/10.1016/j.tree.2020.01.008
Dokter, A. M., Farnsworth, A., Fink, D., Ruiz-Gutierrez, V., Hochachka, W. M., La Sorte, F.
A., Robinson, O. J., Rosenberg, K. V., & Kelling, S. (2018). Seasonal abundance and
survival of North America’s migratory avifauna determined by weather radar. Nature
Ecology & Evolution, 2(10), 1603–1609. https://doi.org/10.1038/s41559-018-0666-4
Driscoll, C. A., Menotti-Raymond, M., Roca, A. L., Hupe, K., Johnson, W. E., Geffen, E.,
Harley, E. H., Delibes, M., Pontier, D., Kitchener, A. C., Yamaguchi, N., O’brien, S. J., &
Macdonald, D. W. (2007). The Near Eastern origin of cat domestication. Science (New
York, N.Y.), 317(5837), 519–523. https://doi.org/10.1126/science.1139518
eBird. (2022). eBird [eBird, Cornell Lab of Ornithology, Ithaca, New York.]. eBird: An
Online Database of Bird Distribution and Abundance. https://ebird.org/region/CA-BCOS?yr=all
Elizondo, E., & Loss, S. (2016). Using trail cameras to estimate free-ranging domestic cat
abundance in urban areas. Wildlife Biology, 22, 246–252.
https://doi.org/10.2981/wlb.00237
Emlen, J. T. (1974). An Urban Bird Community in Tucson, Arizona: Derivation, Structure,
Regulation. The Condor, 76(2), 184–197. https://doi.org/10.2307/1366729
Ferreira, J. P., Leitão, I., Santos-Reis, M., & Revilla, E. (2011). Human-Related Factors
Regulate the Spatial Ecology of Domestic Cats in Sensitive Areas for Conservation. PLOS
ONE, 6(10), e25970. https://doi.org/10.1371/journal.pone.0025970
Flockhart, D. T. T., Norris, D. R., & Coe, J. B. (2016). Predicting free‐roaming cat population
densities in urban areas. Animal Conservation, 19(5), 472–483.
https://doi.org/10.1111/acv.12264
Gow, E. A., Aubrey, B., Cossar, L., Mazerolle, C., & Cheskey, E. (2024). Conducting
community-led research using trail cameras to develop baseline wandering domestic cat
local abundance estimates. FACETS, 9, 1–14. https://doi.org/10.1139/facets-2023-0033
Hand, A. (2019). Estimating feral cat densities using distance sampling in an urban
environment. Ecology and Evolution, 9(5), 2699–2705. https://doi.org/10.1002/ece3.4938

15
Horn, J. A., Mateus-Pinilla, N., Warner, R. E., & Heske, E. J. (2011). Home range, habitat
use, and activity patterns of free-roaming domestic cats. The Journal of Wildlife
Management, 75(5), 1177–1185. https://doi.org/10.1002/jwmg.145
Huston, M. A., & Wolverton, S. (2009). The global distribution of net primary production:
Resolving the paradox. Ecological Monographs, 79(3), 343–377.
https://doi.org/10.1890/08-0588.1
Isaksson, C. (2018). Impact of Urbanization on Birds. In D. T. Tietze (Ed.), Bird Species:
How They Arise, Modify and Vanish (pp. 235–257). Springer International Publishing.
https://doi.org/10.1007/978-3-319-91689-7_13
Jhala, Y., Qureshi, Q., & Gopal, R. (2011). Can the abundance of tigers be assessed from
their signs? Journal of Applied Ecology, 48(1), 14–24. https://doi.org/10.1111/j.13652664.2010.01901.x
Johnson, M. D., & Sherry, T. W. (2001). Effects of food availability on the distribution of
migratory warblers among habitats in Jamaica. Journal of Animal Ecology, 70(4), 546–
560. https://doi.org/10.1046/j.1365-2656.2001.00522.x
Jokimäki, J. (1999). Occurrence of breeding bird species in urban parks: Effects of park
structure and broad-scale variables. Urban Ecosystems, 3(1), 21–34.
https://doi.org/10.1023/A:1009505418327
Jones, T., & Cresswell, W. (2010). The phenology mismatch hypothesis: Are declines of
migrant birds linked to uneven global climate change? The Journal of Animal Ecology,
79(1), 98–108. https://doi.org/10.1111/j.1365-2656.2009.01610.x
Jourdain, E., Gauthier-Clerc, M., Bicout, D., & Sabatier, P. (2007). Bird Migration Routes
and Risk for Pathogen Dispersion into Western Mediterranean Wetlands. Emerging
Infectious Diseases, 13(3), 365–372. https://doi.org/10.3201/eid1303.060301
Lancaster, R. K., & Rees, W. E. (1979). Bird communities and the structure of urban habitats.
Canadian Journal of Zoology, 57(12), 2358–2368. https://doi.org/10.1139/z79-307
Lao, S., Anderson, A. W., Blair, R. B., Eckles, J. W., Turner, R. J., & Loss, S. R. (2023).
Bird–building collisions increase with weather conditions that favor nocturnal migration
and with inclement and changing weather. Ornithological Applications, 125(1), duac045.
https://doi.org/10.1093/ornithapp/duac045

16
Leech, D. I., & Crick, H. Q. P. (2007). Influence of climate change on the abundance,
distribution and phenology of woodland bird species in temperate regions. Ibis, 149(s2),
128–145. https://doi.org/10.1111/j.1474-919X.2007.00729.x
Legge, S., Murphy, B. P., McGregor, H., Woinarski, J. C. Z., Augusteyn, J., Ballard, G.,
Baseler, M., Buckmaster, T., Dickman, C. R., Doherty, T., Edwards, G., Eyre, T., Fancourt,
B. A., Ferguson, D., Forsyth, D. M., Geary, W. L., Gentle, M., Gillespie, G., Greenwood,
L., … Zewe, F. (2017). Enumerating a continental-scale threat: How many feral cats are in
Australia? Biological Conservation, 206, 293–303.
https://doi.org/10.1016/j.biocon.2016.11.032
Lind, J., & Cresswell, W. (2006). Anti-predation behaviour during bird migration; the benefit
of studying multiple behavioural dimensions. Journal of Ornithology, 147(2), 310–316.
https://doi.org/10.1007/s10336-005-0051-3
Lisovski, S., Gosbell, K., Minton, C., & Klaassen, M. (2021). Migration strategy as an
indicator of resilience to change in two shorebird species with contrasting population
trajectories. Journal of Animal Ecology, 90(9), 2005–2014. https://doi.org/10.1111/13652656.13393
Loss, S. R., Will, T., & Marra, P. P. (2013). The impact of free-ranging domestic cats on
wildlife of the United States. Nature Communications, 4(1), 1396.
https://doi.org/10.1038/ncomms2380
Mills, G. S., Dunning, J. B., Jr., & Bates, J. M. (1989). Effects of Urbanization on Breeding
Bird Community Structure in Southwestern Desert Habitats. The Condor, 91(2), 416–428.
https://doi.org/10.2307/1368320
Okanagan Similkameen Stewardship [OSS]. (2023). Birds. Osstewardship.
https://www.osstewardship.ca/birds
Ortega-Álvarez, R., & MacGregor-Fors, I. (2009). Living in the big city: Effects of urban
land-use on bird community structure, diversity, and composition. Landscape and Urban
Planning, 90(3), 189–195. https://doi.org/10.1016/j.landurbplan.2008.11.003
Rochlitz, I. (2005). A review of the housing requirements of domestic cats (Felis silvestris
catus) kept in the home. Applied Animal Behaviour Science, 93(1), 97–109.
https://doi.org/10.1016/j.applanim.2005.01.002

17
Saunders, S. P., Wu, J. X., Gow, E. A., Adams, E., Bateman, B. L., Bayard, T., Beilke, S.,
Dayer, A. A., Fournier, A. M. V., Fox, K., Heglund, P., Lerman, S. B., Michel, N. L.,
Paxton, E. H., Şekercioğlu, Ç. H., Smith, M. A., Thogmartin, W., Woodrey, M. S., & van
Riper, C., III. (2021). Bridging the research-implementation gap in avian conservation
with translational ecology. Ornithological Applications, 123(3), duab018.
https://doi.org/10.1093/ornithapp/duab018
Scully, A. E., Fisher, S., Miller, D. A. W., & Thornton, D. H. (2018). Influence of biotic
interactions on the distribution of Canada lynx (Lynx canadensis) at the southern edge of
their range. Journal of Mammalogy, 99(4), 760–772.
https://doi.org/10.1093/jmammal/gyy053
Sollmann, R., Furtado, M. M., Gardner, B., Hofer, H., Jácomo, A. T. A., Tôrres, N. M., &
Silveira, L. (2011). Improving density estimates for elusive carnivores: Accounting for
sex-specific detection and movements using spatial capture–recapture models for jaguars
in central Brazil. Biological Conservation, 144(3), 1017–1024.
https://doi.org/10.1016/j.biocon.2010.12.011
Swanwick, C., Dunnett, N., & Woolley, H. (2003). Nature, Role and Value of Green Space in
Towns and Cities: An Overview. Built Environment (1978-), 29(2), 94–106.
Tan, S. M. L., Stellato, A. C., & Niel, L. (2020). Uncontrolled Outdoor Access for Cats: An
Assessment of Risks and Benefits. Animals, 10(2), 258.
https://doi.org/10.3390/ani10020258
Tanwar, K. S., Sadhu, A., & Jhala, Y. V. (2021). Camera trap placement for evaluating species
richness, abundance, and activity. Scientific Reports, 11(1), 23050.
https://doi.org/10.1038/s41598-021-02459-w
The Humane Society of the United States [HSUS]. (2019). Pets by the numbers. Washington:
HSUS. https://humanepro.org/page/pets-by-the-numbers
Trouwborst, A., McCormack, P. C., & Martínez Camacho, E. (2020). Domestic cats and their
impacts on biodiversity: A blind spot in the application of nature conservation law. People
and Nature, 2(1), 235–250. https://doi.org/10.1002/pan3.10073
Tu, H.-M., Fan, M.-W., & Ko, J. C.-J. (2020). Different Habitat Types Affect Bird Richness
and Evenness. Scientific Reports, 10(1), 1221. https://doi.org/10.1038/s41598-020-582024

18
Van Den Broeke, M. S., & Gunkel, T. J. (2021). The influence of isolated thunderstorms and
the low-level wind field on nocturnally migrating birds in central North America. Remote
Sensing in Ecology and Conservation, 7(2), 187–197. https://doi.org/10.1002/rse2.179
van Heezik, Y., Smyth, A., Adams, A., & Gordon, J. (2010). Do domestic cats impose an
unsustainable harvest on urban bird populations? Biological Conservation, 143(1), 121–
130. https://doi.org/10.1016/j.biocon.2009.09.013
Wald, D. M., & Peterson, A. L. (2020). Cats and Conservationists: The Debate Over Who
Owns the Outdoors. Purdue University Press. https://doi.org/10.2307/j.ctvs1g9fn
Williams, C. M., Ragland, G. J., Betini, G., Buckley, L. B., Cheviron, Z. A., Donohue, K.,
Hereford, J., Humphries, M. M., Lisovski, S., Marshall, K. E., Schmidt, P. S., Sheldon, K.
S., Varpe, Ø., & Visser, M. E. (2017). Understanding Evolutionary Impacts of Seasonality:
An Introduction to the Symposium. Integrative and Comparative Biology, 57(5), 921–933.
https://doi.org/10.1093/icb/icx122

19
Chapter 2: Year-round Differences in Bird Abundance, Richness, and Community
Composition Across an Urban to Rural Gradient
2.1 Introduction
Environmental change, including anthropogenic disturbance, creates altered
environments that introduce new ecosystems and niches, thereby altering the spatial
distribution of species (Adhikari & Hansen, 2018; Gaston et al., 2003; Semenchuk et al.,
2022; Stein et al., 2014). Understanding how species respond to these alterations is crucial
and can be achieved by examining species abundance, richness, and community composition
across various land uses and seasons. Birds are suitable for studying the effects of land use
and seasonality on species distribution patterns and composition because of their mobility,
detectability, and diverse resource needs. Urbanization is a major form of land conversion
that significantly alters ecological processes and resource availability for birds, within and
around urban areas (Grimm et al., 2008; Isaksson, 2018; Rigal et al., 2023). Urban areas are
designed to accommodate human needs, leading to similar habitats among cities and within
cities, while still supporting diverse habitats within them through the diverse components and
structure of the urban matrix (Swanwick et al., 2003). These urban environments can provide
abundant resources such as food, water, perches, and nesting places to many species,
favouring birds with flexible and broad niche requirements (Emlen, 1974; Isaksson, 2018;
Mills et al., 1989). This uniformity in urban landscapes plays a significant role in influencing
the patterns of bird communities and structure both within urban areas and in areas that have
urban-like qualities (e.g. in peri-urban areas or urban boundaries; Callaghan et al., 2023;
Emlen, 1974; Isaksson, 2018; Mills et al., 1989; Stagoll et al., 2010; Tu et al., 2020).
Birds can be classified based on their ability to inhabit a range of habitats, which can
be categorized by varying degrees of urbanization. Certain bird species, known as “urban

20
exploiters”, thrive in urbanized environments due to their ability to capitalize on the available
resources (Blair, 1996). In contrast, “suburban adaptable” species, which can use both
human-provided and natural resources, are commonly found in peri-urban areas, while
“urban avoiders”, which are often native species, prefer natural habitats where the influences
of humans are minimal (Blair, 1996; Mckinney, 2006; Mills et al., 1989). Areas with less
homogenization and anthropogenic disturbance, such as peri-urban and natural habitats, offer
greater resource diversity, supporting a wider range of ecological niches and higher species
richness (Blair, 1996; Ortega-Álvarez & MacGregor-Fors, 2009; Stagoll et al., 2010).
Conversely, homogenization among urban habitats is expected to result in low species
richness but a high abundance of adaptable species that exploit available resources and
niches, resulting in environments dominated by species that prefer to live near humans (i.e.,
synanthropic species; Urban Homogenization Hypothesis; Blair, 1996; Emlen, 1974; Mills et
al., 1989; Morelli et al., 2016; Ortega-Álvarez & MacGregor-Fors, 2009; Tu et al., 2020).
Urbanization and land conversion significantly influence bird communities, but our
understanding of how these patterns change seasonally is limited due to a lack of full-annual
cycle studies. Seasonal or migratory movements are critical for birds to align their resource
needs with fluctuations in environmental productivity throughout the year (Fudickar et al.,
2021; Huston & Wolverton, 2009). However, most studies focus on spring migration and
breeding season, when abundance and richness are expected to be highest as birds move into
habitats with peak plant and insect productivity (Bauer & Hoye, 2014; Huston & Wolverton,
2009; Mac Arthur, 1959; Somveille et al., 2015). This emphasis on the breeding season limits
our understanding of bird habitat use during other seasonal periods. Studies examining
seasonal bird movements have relied on citizen science (e.g., eBird) and complex modelling
to predict migratory movements and factors that drive these movements (Fuentes et al., 2023;

21
Somveille et al., 2021) or have focused on individual species (Deshwal et al., 2021; RuedaUribe et al., 2024), rather than assessing community- or habitat-level differences. Therefore,
year-round studies assessing birds abundance, richness, and community composition across
the entire annual cycle and across habitats with varying degrees of urbanization are important
to understand how anthropogenic disturbance impacts birds year-round (Marra et al., 2015).
Anthropogenic effects (such as land conversion) may affect resources, vegetative
growth, and microclimates in human dominated landscapes (Clement et al., 2019; Rigal et
al., 2023), yet their seasonal impacts on bird communities remains poorly understood. In
North America, at least 25% of avian species are considered to be synanthropic (Johnston,
2001), yet, during the migration or non-breeding season, birds not typically considered
synanthropic may rely on urban areas for extended periods (Morales et al., 2022; Poirier et
al., 2024). During the non-breeding months (winter in the Northern Hemisphere), lower
temperatures and increased snowfall reduce natural vegetation and prey availability, which
may influence bird community composition and distribution, and patterns of abundance and
species richness (Johnson & Sherry, 2001; Leech & Crick, 2007). However, urban
environments can offer milder temperatures that may reduce metabolic costs for resident
species or facultative migrants in the northern temperate zone winter/non-breeding season
(Watson et al., 2024). As a result, urban areas may produce favourable conditions during
certain periods of the year that may cause some species to avoid migration altogether
(Bonnet-Lebrun et al., 2020), or short or altitudinal migrants may stay in habitats for longer
periods to take advantage of the available resources (Atwell et al., 2011). Habitats with
higher anthropogenic disturbance, such as urban or peri-urban habitats, may provide more
resources through human provisioning, while areas that have less anthropogenic disturbance,

22
such as agricultural and natural areas, may be less productive (Emlen, 1974; Lancaster &
Rees, 1979).
Here, I assess how bird abundance, species richness, and community composition
vary across habitats varying in degrees of anthropogenic disturbance and an entire annual
cycle in the south Okanagan Valley, British Columbia, which is one of the most biodiverse
regions in Canada. Using point counts conducted across five seasonal periods relevant for
birds and habitats ranging from urban, peri-urban, agricultural, and natural habitats, I had
three objectives. I aimed to: 1) identify how bird species richness and abundance varies
across seasons; 2) test the Winter Urban Association, and Urban Homogenization Hypotheses
(Table 2.1); and 3) determine patterns of habitat associations among birds throughout the
entire year.
Table 2.1. Hypotheses and associated predictor variables used to explain local bird relative
abundance and species richness in the south Okanagan Valley, B.C. Source(s) indicate a
study or studies that provide evidence to support the hypothesis.
Predictor
variable

Hypothesis

Hypothesized mechanism(s)

Source(s)

Urban
Homogenization
Hypothesis

Bird communities in urban areas
are very similar because urban
areas are all built the same to
accommodate human needs.
Therefore, the species that can
exploit these areas thrive in high
abundances, but this leads to low
species richness, as these
habitats are often dominated by
urban exploiter species.

Landclass*season

(Blair, 1996; Emlen,
1974; Isaksson,
2018; Mckinney,
2006; Morelli et al.,
2016; OrtegaÁlvarez &
MacGregor-Fors,
2009; Tu et al.,
2020)

Winter Urban
Association
Hypothesis

During the non-breeding
seasons, a lack of resources in
natural and agricultural areas will
draw birds into urban areas, as
humans provide resources for
birds. This will increase
abundance and species richness
in urban and peri-urban areas
during the non-breeding seasons.

Landclass*season

Emlen, 1974;
Jokimäki et al.,
1996; Lancaster &
Rees, 1979;
Tilghman, 1987;
Yaukey, 1996

23
2.2 Methods
2.2.1 Study Area
I studied seasonal patterns of bird abundance, richness, and habitat associations in the south
Okanagan Valley (within and surrounding the communities of Okanagan Falls, Oliver, and
Osoyoos) of British Columbia, Canada, which is the traditional, unceded, and ancestral
territory of the Syilx Okanagan Nation (Figure 2.1). Located in the south-central region of
British Columbia, the south Okanagan Valley is part of the northern extension of the Great
Basin Bird Conservation Region and is one of the most endangered ecosystems in Canada. It
is part of the bunchgrass biogeoclimatic zone, characterized by grassland, shrub-steppe,
wetlands, and dry open forests consisting of ponderosa pines, cottonwoods, and Douglas firs
(Alldritt-McDowell et al., 1998). The south Okanagan valley supports some of the highest
diversity of birds in Canada, with over 330 recorded bird species, including 24 provincially
listed species at risk and 22 federally listed species at risk under the Species at Risk Act
(eBird, 2022; Okanagan Similkameen Stewardship [OSS], 2023). Many birds funnel through
the valley during migration, including long-distance, short-distance, and altitudinal migrants.
This area consists of a mosaic of natural habitats (grasslands, deserts, riparian areas, and
forests), agriculture (e.g. vineyards and orchards), small urban centers, and peri-urban and
exurban areas. The diverse bird species, seasonal variation, and habitat variation of the region
emphasize the value of this region as a study area to test hypotheses about variation in
species abundance, richness, and community composition across the entire year.

24

Figure 2.1. Map of the study area in the south Okanagan Valley showing trail camera
locations in Okanagan Falls, Oliver, and Osoyoos. Location colours are classified by the
following land classes: Agriculture (AG), Natural (NA), Peri-urban (PU), and Urban (UR).
The red boarder represents the study area that was used to study wandering domestic cats and
the study area boundaries (Chapter 3). The inset map in the top right corner shows a red
circle around the south Okanagan Valley, located in the south-central area of British
Columbia, Canada. Map is in EPSG:3005 coordinate system (NAD83, BC Albers projection).

25
2.2.2 Study Design
I assessed seasonal patterns of bird abundance, species richness, and habitat associations
using a repeated-measures design stratified by season, conducting point count surveys from
March 2022 to March 2023 across five distinct seasonal periods relevant to bird ecology and
life history in this region. The five seasonal periods were spring migration (March 13 – May
18, 2022), breeding season (May 29 – July 29, 2022), fall migration (August 10 – October
12, 2022), early non-breeding season (October 20 – December 15, 2022), and late nonbreeding season (January 13 – March 23, 2023). The non-breeding season was divided into
early and late seasons because the resources available to birds are different during each of
these two time periods. For logistical reasons, some sites were added, discarded, or moved
throughout the year, leading to a total of 123 point count locations (see Appendix 1 for
details). The number of locations with completed point counts were as follows: 120 locations
for spring migration, breeding season, and fall migration; 119 locations for early nonbreeding season; and 114 locations for late non-breeding season.
I conducted point counts to examine bird communities and species abundance and
richness across different land use types. This research was part of a larger study examining
the abundance of domestic cats (Chapter 3) across seasons and land use types and therefore,
the point count locations were aligned with trail camera sites. This involved using a random
stratified design with point count locations at least 100m apart stratified by four land use
categories: urban, peri-urban, agricultural, and natural habitats (~30 point counts per land use
type; Figure 2.1). While the trail cameras involved in the broader study were on private
property (e.g. yards) or in natural habitats, to minimize disturbance to landowners, all point
counts occurred either along roadways or on paths within natural areas or agricultural fields
(see Appendix 2 for details of land permissions and site establishments).

26
I used satellite maps of the south Okanagan to categorize areas as urban, peri-urban,
agricultural, and natural before deciding which permits I would need and which
neighbourhoods to target for door-to-door recruitment. Definitions for each land use type
often vary (Sahana et al., 2023; Wolff et al., 2021) and are context dependent and so I
combined established definitions from the literature, with ground truthing, to ensure these
definitions, and thus categories, were appropriate within the context of my study area.
Therefore, I categorized urban areas as having extensive anthropogenic development that
included a matrix of landscapes such as pavement, buildings (residential and commercial),
and habitat patches (e.g., transportation corridors, riverbanks, and parks; Swanwick et al.,
2003). Peri-urban areas consisted of clusters of houses surrounded by agricultural (e.g.,
orchards, vineyards, and crops) and/or natural (e.g., desert, forest, riparian) land with
approximately 25-50 m between houses. Agricultural land was used for agriculture (such as
vineyards and orchards) or other farming practices, targeting smaller vineyards/orchards that
were <50 hectares. This included targeting houses residing next to agricultural land or
buildings/barns within agricultural areas. Natural habitats included any protected areas
conserved provincially, federally, or privately. Since these descriptors are continuous, I later
classified sites using QGIS to create discrete variables and refined categories based on
surrounding habitat variables for statistical analysis (see 2.2.3 GIS Analysis for more details).
When choosing sites, true randomization was impossible because locations were mainly on
private property and therefore dependent on the willingness of landowners and land
managers (except natural sites). This resulted in some sites (two sets of sites; n = 4) being
closer than 100 m apart (70 m and 40 m).
I incorporated a repeated measures component into each point-count session by
conducting two back-to-back 5-minute point counts at each location. For clarity throughout, I

27
will refer to a point count as each 5-minute survey, while a point-count session consists of
two consecutive point counts at a single site. During each seasonal period, I visited each site
twice, resulting in a total of 20 point counts per location (2 point counts per session x 2 visits
x 5 seasons = 20 point counts per location), unless if a site was added or discarded, resulting
in less than 5 seasons completed for a location (see Appendix 1 for more information).
During point counts, observers watched, listened, and recorded all birds that were seen or
heard, noting their distance from the surveyor. Distances were categorized into bins of 0–50
m and 50–100 m to align with the requirements of generalized distance sampling models (see
below) we used to estimate bird abundance and richness. In some cases, the 100 m radius
overlapped between point count sites, which meant that some point counts were not
statistically independent, and it was possible, yet unlikely, that some birds were counted more
than once. These methods have been used in other habitat-based point counts (Huff et al.,
2000) and the Ontario Breeding Bird Atlas (Ontario Breeding Bird Atlas, 2021). While
double counting can be a concern when estimating population sizes, for examining patterns
of abundance across habitats and seasons, statistical independence is not required (Martijn et
al., 2023). Thus, we did not remove point count locations that might have overlapped (i.e.,
<200m apart). At the start of each point-count session, the observer recorded the date, time,
cloud cover, wind, and precipitation (defined in Table 2.2). Point count routes, consisting of
10–15 point count locations, were completed by highly skilled birders (3 different birders
total). These routes began with morning civil twilight and were completed within 5 hours to
maximize detection rates during peak bird activity periods. The order of point count routes
was changed each time to account for temporal detection bias and activity variance between
species.

28
2.2.3 GIS Analysis
I used QGIS (Version 3.32 Lima; QGIS.org, 2024) to obtain habitat variables. I assessed
habitat variables that represent buildings, roads, water, vineyards, orchards, deserts, fields,
forest, and altitude. I accessed detailed OpenStreetMap layers for roads, lakes, rivers, and
buildings and imported them into QGIS. To obtain missing habitat information, I created
polygons around forests, lakes, rivers, deserts, fields, orchards, and vineyards, and updated
the buildings using the point creation function in QGIS. I used a 2023 Esri World Imagery
Wayback satellite image as a base layer to trace the polygons (Esri, 2023). I created 200m
buffers around each site and extracted the habitat variable information using the intersect
tool. Within each buffer, I categorized habitat variables into roads, buildings, desert, field,
forest, orchard, vineyard, and water. I then quantified the total length of roads, the total
number of buildings, and the proportion of desert, field, forest, orchard, vineyard, and water
within each 200 m buffer (defined in Table 2.2). I conducted these analyses in QGIS using
the EPSG:3005 coordinate system (NAD83, BC Albers projection). The habitat variables
extracted from the 200 m buffers were used to classify each point count location as either
urban, peri-urban, natural, or agricultural (defined in Table 2.2).

29
Table 2.2. Description of detection and site covariates used to estimate relative abundance
and species richness of birds in the south Okanagan Valley, B.C. Detection covariates
influenced the detection of species during point counts and site covariates describe the site
where point counts occurred.
Variables

Description

Detection covariates
date

Julian date when the point count was conducted.

time

Time at the start of the point count observational period.

wind

Strength of wind during point count using the Beaufort scale (Scale 0-4)

cloud

Cloud cover % overhead during the point count.

precip

Binary observation during point count, where 0 = no precipitation and 1 =
precipitation.

Site covariates
desert

Percentage of desert within a 200 m buffer around the point count
location.

field

Percentage of field within a 200 m buffer around the point count location.
Fields include public spaces such as parks and school yards, land that is
used as pasture or farming, and empty, overgrown lots.

orchard

Percentage of orchard within a 200 m buffer around the point count
location.

vineyard

Percentage of vineyard within a 200 m buffer around the point count
location.

forest

Percentage of forest within a 200 m buffer around the point count location.

water

Percentage of water within a 200 m buffer around the point count location.

roads

Privateroads and publicroads combined, to equal the total length of roads
(m) within a 200 m buffer around the point count location.

privateroad

Total length of private roads, in meters, within a 200 m buffer around the
point count location. Private roads include driveways (>40 m in length),
vineyard/orchard roads, and other impervious surfaces that are not
accessible to the public.

publicroad

Total length of public roads, in meters, within a 200 m buffer around the
point count location. Public roads are maintained by the city and are easily
accessible to the public.

buildings

All buildings within a 200 m buffer around the point count locations. This
includes high-density housing and all other buildings.

landclass

The land class type of the point count locations: urban, peri-urban, natural,
or agricultural

urban

Site buffer (200 m) containing >55 buildings

peri-urban

Site buffer (200 m) with 17 to 50 buildings and <49% vineyard and/or
orchard.

agricultural

Site buffer (200 m) characterized by >28% agriculture and <20 buildings.

natural

Site buffer (200 m) characterized by >25% natural habitat, <16 buildings,
and <29% agriculture.

altitude

Altitude (m) at the point count location.

30
2.2.4 Unmarked Analysis Design
I used distance sampling techniques to analyze data for unmarked animals (Fiske &
Chandler, 2011). Distance sampling involves measuring the distance from a point to an
observed organism and can be used to estimate abundance (Buckland et al., 2001). Distance
can be recorded by measuring the exact distance to the observed organisms from the
observer, however, since birds are often detected by their vocalizations, it can be difficult to
detect their exact location. As a result, distances are frequently recorded by grouping
observations into discrete distance intervals (Royle et al., 2004). I estimated bird abundance
using generalized distance sampling models using ‘gdistsamp’ within the unmarked package
(Fiske & Chandler, 2011) in Program R, V2.1.1. (R Core Team, 2021). The ‘unmarked’
models incorporate distance sampling and detection probability to estimate abundance, while
accounting for species that may go undetected during sampling (Fiske & Chandler, 2011).
Detection probability is the relationship between distance and the likelihood of detecting an
organism (Buckland et al., 2001). Within the unmarked package, gdistsamp organizes data
into three components: a distance matrix with observed birds per distance bin, the site
covariates detailing environmental and site influences, and the detection covariates, which
accounts for factors that would influence detection. Repeated measures are incorporated into
these matrices. This model also accounted for the mobility of birds by incorporating
temporary emigration, where individuals move in and out of the survey area throughout the
sampling period (Chandler et al., 2011). This represents a superpopulation meaning the
abundance estimates represent the number of individuals that could potentially use the point
count area (Chandler et al., 2011). See Appendix 3 for distance sampling equation and
assumptions.

31
I performed repeated counts (see above for how repeated measures are incorporated)
and integrated environmental covariates into my models to reduce measurement errors (e.g.,
false detection), improve detection probability, account for variation, and provide more
accurate estimates of abundance (Kellner et al., 2023). Site covariates, such as elevation and
habitat variables, were used to account for site-specific differences. I also included several
detection co-variates such as wind, time, and date. Prior to analysis, I standardized all
detection and site covariates using a z-transformation (mean of 0 and SD of 1).
2.2.5 Abundance Estimates
I followed the unmarked analysis workflow outlined in Kellner et al. (2023) to estimate
abundance. After completing the study design and data collection steps, I completed the data
input and organization step by creating an unmarkedFrame to store an observation matrix,
site covariates, and detection covariates. Given that most research on avian abundance has
focused on a single season, current statistical approaches are unable to handle multiple
seasons of point count data for sites, as well as multiple visits within each season. Thus, I
evaluated each season separately. I identified the key function of ‘halfnormal’ to create the
most parsimonious null model based on it having the lowest Akaike’s Information Criterion
corrected for a small sample size (AICc) score (out of models with key functions of: normal,
hazard, exponential, and uniform; (Bozdogan, 1987), and used it in subsequent gdistsamp
models to estimate abundance.
I then fit the models using the gdistsamp function where each model includes 3
formulas estimating: abundance (λ), availability (ϕ), and detection (ρ). Local abundance is
the superpopulation of birds that could potentially occupy the plot, availability is the portion
of the superpopulation that could be detected. The observation matrix influences abundance

32
estimates, while the site and detection covariates influence availability and detection,
respectively. I then used a correlation matrix and constructed models with non-correlated
combinations of covariates to assess potential covariation. In the model selection step, I
compared all gdistsamp models using AICc scores to determine the most parsimonious model
for each season. Finally, I used the most parsimonious gdistsamp model to estimate the
abundance of birds at each site using the GetP function, with the top model for each season
outlined in Table 2.3.
Table 2.3. Gdistsamp models used to estimate bird abundance for each season.
Availability
(φ)

Season

Abundance (λ)

Detection (ρ)

Spring
Migration

altitude + desert + field + agriculture +
forest + water + privateroad + landclass

date + time

wind + cloud
+ date + time

Breeding
Season

altitude + desert + field + agriculture +
forest + water + buildings + landclass

date + time

wind + cloud
+ date + time

Fall
Migration

altitude + desert + field + agriculture +
forest + water + roads + landclass

date + time

wind + cloud
+ date + time

Early
Nonbreeding
Season

altitude + desert + field + agriculture +
forest + water + roads + landclass

date + time

wind + cloud
+ date + time

Late Nonbreeding
Season

altitude + desert + field + vineyard +
forest + water + privateroad + buildings
+ landclass

date + time

wind + cloud
+ date + time

2.2.6 Species Richness
I calculated the average species richness at each site using the vegan package in R V2.1.1.
(Oksanen et al., 2024; R Core Team, 2021). I analyzed each season separately by calculating

33
the average species richness per site per season, including only birds that were observed
within 100 meters from the observer. To determine average species richness, I summed the
number of species observed during each point count and divided this sum by the number of
repeated point counts at the site for each season. These average richness outputs were then
used in subsequent analyses to understand which factors influence changes in species
richness across seasons and land classes.
2.2.7 Seasonal Analyses
I used Linear Mixed Effects Models (LMEs) to examine seasonal variation in species
richness and abundance. I used the average abundance and average species richness for each
site during each season. The abundance and richness were normally distributed, and
therefore, I used the ‘lme4’ and ‘lmerTest’ packages in program R to fit LMEs using the
‘lmer’ function (Bates et al., 2015; Kuznetsova et al., 2017) to test the Winter Urban
Association and Urban Homogenization Hypotheses. Predictor variables included land class,
season, and their interaction. The response variable was either abundance or richness, with
site as a random effect to account for repeated measures between seasons. I created a
pairwise comparison model, using lsmeans, to assess the patterns and interaction across
season and land class, with a Tukey adjustment to control for multiple comparisons.
2.2.8 NMDS Analysis
I tested the Urban Homogenization Hypothesis (Table 2.1) using non-metric
multidimensional scaling (NMDS), an unconstrained ordination, in R using the vegan
package to assess how species composition varies seasonally (Oksanen et al., 2024). I
analyzed each season separately and created an encounter history of the birds that were
observed during each point count. I excluded rare species, and thus only included species that

34
were observed at more than 10 point counts during a season (Spring, N = 35 species;
Breeding, N = 48 species; Fall, N = 37 species; Early non-breeding, N = 31 species; Late nonbreeding, N = 27 species). This approach simplified the model, lowered the stress scores, and
focused on species that were important for understanding community composition. I used the
Bray-Curtis distance metric because the ordination plots most accurately represented the bird
community composition and structure (see Appendix 4 for further rationale). I chose three
dimensions (k = 3) for each season to visualize these data because it resulted in the lowest
acceptable stress score (0.2; see Appendix 4 for more details). For the fall migration and nonbreeding seasons, I added a constant (0.05) to each observation because there were many
sites where no birds were seen leading to too many zero values in the encounter histories and
by adding the constant, I was able to improve robustness of the analysis (McCune & Grace,
2002). All stress values were below 0.2 (Appendix 5).
2.3 Results
I conducted 2380 point counts between March 2022 to March 2023, totalling 11,900 minutes
(~200 hours) of observation. Throughout this year-long study, I identified 146 species (N =
51,682 individual birds; Table A6.1). The most common bird species were European starlings
(Sturnus vulgaris; N = 10,432 observations), House sparrows (Passer domesticus; N = 9,463
observations), and California Quails (Callipepla californica; N = 5000 observations). The
mean species richness per site was 5.74 ± 0.064 se (median = 6, range: 0–17) and the mean
bird abundance per site was 82.8 ± 1.55 se (median = 80.62, range: 4.99–231.72). Overall,
the number of species observed varied between seasons and land uses (totals for species
observed during point counts are listed in Table A6.2), with natural habitats having the

35
highest total number of observed species across each season, while urban areas had the
lowest number of total species for every season except the late non-breeding season.
2.3.1 Mean Abundance and Richness Across Seasons
In my assessment of seasonal patterns using Linear Mixed Effects Models, I found that both
abundance and richness were significantly higher during breeding season (120.1 ± 2.64 SE,
95% CI: 114.8–125.3), followed by spring migration (104.8 ± 2.64 SE, 95% CI: 99.6–110.0;
spring migration vs breeding: t = -6.363, P < 0.0001; Figure 2.2). Additionally, mean bird
abundance was not significantly different between early and late non-breeding season (LME,
P = 0.7346; Figure 2.2), indicating that the abundance of birds did not change between
October and March. For richness, all seasons were significantly different (LME, P < 0.0001;
Figure 2.2). Richness was highest during the breeding season (7.99 ± 0.21 SE, 95% CI: 7.58–
8.39), followed by spring migration (6.77 ± 0.21 SE, 95% CI: 6.36–7.17; spring migration vs
breeding: t = -4.974, P < 0.0001).
2.3.2 Seasonal Abundance and Richness Across Land Classes
I showed support for the Winter Urban Association Hypotheses (Table 2.1) because
abundance and species richness were elevated in urban and peri-urban areas during the early
and late non-breeding seasons (Figure 3.4; Abundance: Table A6.3; Richness: Table A6.4).
During the early non-breeding season, the mean abundance was highest in urban, peri-urban,
and agricultural habitats, while mean species richness was highest in peri-urban, urban, and
natural habitats. During the late non-breeding season, the mean abundance and mean species
richness was highest in peri-urban and urban habitats, indicating that birds are using
urbanized habitats (urban and peri-urban) more than other habitats during the non-breeding
seasons.

36
There was support for the Urban Homogenization Hypothesis (Table 2.1) by
comparing abundance and richness in different land classes across the most productive
seasons (spring migration, breeding season, and fall migration). Abundance and richness
displayed different seasonal results during these three productive seasons (Figure 2.4;
Abundance: Table A6.3; Richness: Table A6.4). Mean abundance was highest in urban and
peri-urban areas across all three seasons. During the breeding season, abundance in urban
areas was very similar to abundance in natural habitats. Richness showed a different pattern
where urban areas have low species richness across all three seasons. During spring
migration, there is no significant difference in species richness between any of the land
classes. During both the breeding season and fall migration, richness was lowest in urban
areas, while abundance was high.

37

Figure 2.2. A) Mean seasonal abundance displaying non-significance between fall migration
(green) and early and late non-breeding season (blue and purple). Mean abundance was
highest during the breeding season (gold; 120 ± SE 29.4) and spring migration (pink; 104 ±
SE 31.8), and lowest throughout the fall migration and early and late non-breeding seasons
(fall migration 68.5 ± 28.4, early non-breeding 61.8 ± 21.4, late non-breeding 58.6 ± 31.3;
mean bird abundance ± SE). B) Mean seasonal richness, showing non-significance between
fall migration (green) and early non-breeding season (blue). Species richness was highest
during the breeding season (gold; 7.98 ± SE 2.64) and spring migration (pink; 6.76 ± SE
1.99) and lowest in the late non-breeding season (purple). The horizontal black line within
the box represents the median value, and the box represents 50% of the data, from the 25th to
75th percentile of each groups’ distribution of values, which is the interquartile range (IQR).
The whiskers extend to the highest and lowest points within 1.5 times the IQR and individual
dots are outliers beyond the whiskers.

38

Figure 2.3. Box plots showing the estimated relative abundance during the early nonbreeding seasons (A and C) and late non-breeding season (B and D). A) Estimated abundance
was highest in urban areas (75.4 ± SE 3.96) though not significantly different from peri-urban
(62.3 ± SE 4.37), and agricultural (70.1 ± SE 3.98) areas. B) Estimated abundance was
highest in peri-urban (80.8 ± SE 4.42) and urban (74.3 ± SE 3.99) areas. C) Species richness
was highest in urban (5.50 ± SE 0.378) and peri-urban (5.46 ± SE 0.420) areas, though not
significantly different from natural habitats (4.25 ± SE 0.372). D) Species richness was
highest in peri-urban (5.03 ± SE 0.427) and urban (4.79 ± 0.383) areas. See Table A6.3 and
Table A6.4 for P-values. The horizontal black line within the box represents the median
value, and the box represents 50% of the data, from the 25th to 75th percentile of each groups’
distribution of values, which is the interquartile range (IQR). The whiskers extend to the
highest and lowest points within 1.5 times the IQR and individual dots are outliers beyond
the whiskers.

39

Figure 2.4. Box plots showing estimated abundance and species richness during the spring
migration (A and D), breeding season (B and E) and fall migration (C and F). A) Estimated
abundance was highest in urban (132.5 ± SE 4.30) and peri-urban (127.9 ± SE 3.95) areas
during the spring migration. B) Estimated abundance was highest in peri-urban (142.2 ± SE
4.32) and urban (125.1 ± SE 3.96) areas during the breeding season, though similar to natural
habitats (117.9 ± SE 3.90). C) During the fall migration, estimated abundance was highest in
peri-urban (107.8 ± SE 4.32) and urban (71.9 ± SE 3.96) areas. D) Species richness was not
significantly different between any land classes during the spring migration. E) Species
richness was highest in natural (9.47 ± SE 0.372) and peri-urban (9.10 ± SE 0.413) areas
during the breeding season. F) During the fall migration, species richness was lowest in
urban areas (4.65 ± SE 0.378) and not significantly different between any other land classes.
See Table A6.3 and Table A6.4 for P-values. The horizontal black line within the box
represents the median value, and the box represents 50% of the data, from the 25th to 75th
percentile of each groups’ distribution of values, which is the interquartile range (IQR). The
whiskers extend to the highest and lowest points within 1.5 times the IQR and individual dots
are outliers beyond the whisker.

40
2.3.3 Species Communities and Habitat Types
I examined the dynamics of bird communities and their habitat associations (objective 2) by
visualizing patterns of similarity or dissimilarity between seasonal bird communities.
Synanthropic species, which were associated with buildings, included House Sparrows
(Passer domesticus), Eurasian Collared-doves (Streptopelia decaocto), and California Quail
(Callipepla californica). Species associated with vineyards included Vesper sparrows
(Pooecetes gramineus), American Robins (Turdus migratorius), and European Starlings
(Sturnus vulgaris). Species associated with deserts included Canyon Wrens (Catherpes
mexicanus), Western Meadowlarks (Sturnella neglecta), Black-billed Magpie (Pica
hudsonia), and Lazuli Bunting (Passerina amoena), and Western Bluebirds (Sialia
mexicana). Forest habitat was strongly associated with the Western Wood-peewee (Contopus
sordidulus). Natural habitat around water was associated with Song Sparrows (Melospiza
melodia) and Red-winged Blackbirds (Agelaius phoeniceus). During the non-breeding
seasons, water was associated with waterfowl, such as Mallards (Anas platyrhynchos),
Buffleheads (Bucephala albeola), and Common Goldeneye (Bucephala clangula; figures and
more details are in Appendix 7 and Figure A7.1). The late non-breeding season, the spring
migration, and the breeding season show similar patterns, with urban and natural habitats
have dissimilar species compositions (Figure 2.6; Figure A7.2).
In NMDS1 and NMDS2 there is a distinct separation between the natural and urban
areas, which suggests that urban and natural habitats are composed of different species.
Natural habitats also have greater variation in species (more spread out over NMDS1 and
NMDS2) compared to urban sites where points are more clustered together, indicating that
urban areas have similar community composition, which may be indicative of lower
diversity. In contrast, in peri-urban and agricultural areas there is considerable overlap

41
suggesting similarities in species composition. However, in agricultural areas there is more
diversity of species across sites. The fall migration and early non-breeding seasons show
different results from the other three seasons in terms of their species composition (Figure
2.6). During the fall migration, I observed an absence of discernible patterns across land
classes, as there is a lot of overlap between the land use circles. This observation implies that
there is a more uniform distribution of bird species across various sites. During the early nonbreeding season, there is some division among land class circles, implying increased
variation between the sites when compared to the fall migration.

Figure 2.5. NMDS comparing species habitat associations across A) the spring migration B)
breeding season, C) fall migration, D) early non-breeding season, and E) late non-breeding
season. NMDS1 is on the X axis and NMDS2 is on the Y axis. Additional information can be
found in Appendix 7 and remaining NMDS comparisons (1 vs. 3 and 2 vs. 3) can be found in
Figure A7.1. Birds are represented by four-letter banding codes and the corresponding
scientific and common names can be found in Table A6.1.

42

Figure 2.6. NMDS comparing species compositions across urban (UR), peri-urban (PU),
agricultural (AG), and natural (NA) areas. I also compared four seasons: A) the spring
migration B) breeding season, C) fall migration, D) early non-breeding season, and E) late
non-breeding season. Each point in the figure represents a point count and its location on the
NMDS is based on the birds that were observed during that point count. Points that are closer
together are more similar in species composition and points that are further apart are
dissimilar in their species composition. The circles around the points show a 95% confidence
interval. I can see a distinct separation between the natural and urban areas, which suggests
that urban and natural habitats are comprised of different species. Remaining NMDS
comparisons (1 vs. 3 and 2 vs. 3) can be found in Figure A7.2.
2.4 Discussion
This study provides a comprehensive year-round analysis of bird abundance, richness, and
community composition across a diversity of habitats and human modified landscapes. I
demonstrate how current analyses focusing on a single season (e.g. the breeding season) can
limit understanding of species habitat use and needs, and thus, ecological processes that may
be influencing these patterns of use. For instance, my results showed that abundance and
richness was highest during spring migration and breeding and that urbanization impacted the
distribution and community composition of species year-round. In areas where there are

43
mosaics of habitats, such as in my study region, the diversity of habitats may serve different
ecological purposes and provide resources or other benefits for numerous diverse species
communities, which can vary across the year. For example, urban, peri-urban, and
agricultural habitats may provide very important habitat during non-breeding for numerous
resident and migratory species, as these land class types had the highest abundances and
species richness of birds during these seasons. The findings highlight the importance of yearround studies in assessing avian community composition and habitat selection, challenging
long-held assumptions that these patterns remain consistent across regions.
Our findings do not align with past research on seasonal bird abundance in North
America, which has predominantly focused on Eastern regions during the breeding and
sometimes migratory periods, and this has led to a biased understanding, and often
generalization, of seasonal patterns of avian abundance. My observation that local bird
abundance was higher during spring migration and breeding compared to fall migration and
non-breeding differs from results found in other regions of North America. Another study in
western North America (southwestern Arizona) also observed higher bird abundance during
spring migration compared to the fall migration (van Riper III et al., 2008). Yet, a common
assumption in North America suggests that bird abundance peaks during the fall migration,
which is opposite of what I observed in my study. These assumptions are based on numerous
studies using mainly eastern North American sites, such as Dokter et al. (2018), which used
weather radar data from across the USA and found that the total biomass of migrating birds
was higher during the fall than the spring, and Horton et al. (2023), which used wood warbler
banding data across North America to show that there were over three times as many wood
warblers banded during the fall than the spring. One big reason for the differing patterns
observed in western North America may be related to its diverse topography and abruptly

44
changing habitats (Carlisle et al., 2009), such as numerous mountain ranges that create
elevational gradients and additional obstacles for migrating birds (La Sorte, Fink,
Hochachka, DeLong, et al., 2014; La Sorte, Fink, Hochachka, Farnsworth, et al., 2014). It is
unclear if total abundances of birds is actually lower in western Canada, or whether birds
take advantage of elevational gradients and higher elevations for food during the fall
migratory period (Wilson & Martin, 2005). Based on my findings, it is evident that more
research is needed on bird abundance and richness patterns throughout the year, especially in
western North America. This will help with avoiding generalized assumptions on bird
abundances and richness patterns based on studies from geographically different areas. Such
generalization could have detrimental consequences if conservation or restoration initiatives
apply standardized patterns of abundance across broad geographically different areas,
emphasizing the need for more independent year-round assessments of bird abundance,
richness, and associated community structures in western North America.
Urbanization may decrease the propensity of some birds to migrate and many species
that remain year-round choose to stay within urban habitats (reviewed by Bonnet-Lebrun et
al., 2020). While many birds migrate south during fall migration and the non-breeding
season, I found that abundance remained stable throughout early and late non-breeding. This
is unsurprising as many altitudinal and short-distance migrants remain in Canada during the
winter (Bonnet-Lebrun et al., 2020; Boyle, 2017), with western North America having the
highest number of altitudinal migrant species (Boyle, 2017). I also found that the birds that
remained during non-breeding were primarily using urban and peri-urban habitats, as
evidenced by the higher abundance and species richness compared to less urbanized
agricultural and natural habitats. These patterns have also been seen in other North American
studies (Clergeau et al., 1998; Murthy et al., 2016; Yaukey, 1996). In British Columbia, many

45
species move out of high elevation habitats during the non-breeding season, selecting lower
elevation ecosystems instead (Herbers et al., 2004), which may provide resources found in
abundance in urban habitats. Supplemental food in the form of feeders, garbage, and
cultivated plants can increase the abundance of birds in urban areas (McKinney, 2002), which
can be beneficial for birds during the sparse non-breeding season (Grubb & Cimprich, 1990;
Yaukey, 1996; Zuckerberg et al., 2011). Urban areas also act as heat islands, increasing the
average temperature (McKinney, 2002), and thereby making urban habitats more hospitable
for birds during the winter. These factors may all work together in the south Okanagan to
drive birds into urban habitats, which is why urbanization is important in dictating the
movement and community composition of birds during non-breeding (Bonnet-Lebrun et al.,
2020; Leveau et al., 2021; Saunders et al., 2022; Zuckerberg et al., 2011), and why it is
important to assess abundance, richness, and community composition of birds over a variety
of habitats including urban areas during non-breeding seasons for proper management and
conservation of overwintering species.
During early non-breeding, agricultural habitats had similar species composition and
richness relative to natural habitats, and similarly high abundance compared to urban and
peri-urban habitats. These findings differ from other studies and highlight the significant role
of human altered landscapes in driving community structure during the non-breeding period.
Recent studies have assessed the dramatic decline of farmland and grassland breeding birds
in North America (Murphy, 2003; Stanton et al., 2018; Valiela & Martinetto, 2007; van Vliet
et al., 2020), leading to the general assumption that agricultural habitat is bad for songbirds
and allies and very few species can benefit from access to agricultural food sources (Murphy,
2003; Stanton et al., 2018; Valiela & Martinetto, 2007; van Vliet et al., 2020). However, all
these studies were focused only on a single time period, the breeding season. While my

46
results also showed that agricultural habitats had the lowest abundance and richness of birds
during the breeding season, this was not the case year-round. Thus, my study shows that
conducting research in only a single-season and making broad scale inferences of habitat
quality or use for birds may be short-sighted, and these single season studies limit the full
understanding of the benefits (or detriments) of agricultural land to a wide diversity of
songbirds year-round. The high abundance of birds in agricultural habitat during early nonbreeding may be linked to the availability of remaining grapes and seeds, with grapes being a
major crop in the south Okanagan valley. Other studies using eastern North American eBird
data have also shown that some birds transition to agricultural areas during fall migration
(Zuckerberg et al., 2016), as changes in resource availability drive the movement and habitat
selection of birds. The potential benefits of agricultural (and urban and peri-urban) areas for
birds at various times of the year has typically been overlooked (or not even included) in past
studies, but I have shown agricultural land plays a critically important role in support of bird
communities outside of the breeding season.
Many species use urban habitats throughout the year, but species composition within
each season is often similar across urban sites, typically dominated by urban exploiters such
as Eurasian Collared-doves (Streptopelia decaocto) and House Sparrows (Passer
domesticus), resulting in homogenization characterized by low species richness and high
abundance (Beissinger & Osborne, 1982; Butler, 2003; Fontana et al., 2011; Ortega-Álvarez
& MacGregor-Fors, 2009). This pattern was evident during spring migration, breeding, and
fall migration, where most species were urban exploiters, which were in high abundance. But
during non-breeding, a wider diversity of species used urban habitats, including species not
typically common in urban areas during other times of the year. This led to increased
abundance and species richness, while maintaining homogenization as the same species were

47
observed across all urban sites. Non-breeding homogenization appeared to result from broad,
equal use of urbanized habitats by many species, including altitudinal migrants like Varied
Thrushes (Ixoreus naevius) and Dark-eyed Juncos (Junco hyemalis). However, there may be
elevated urban-associated risks (such as window collisions and predation by domestic cats)
for some species that are not adapted to urban environments and may be drawn to urban areas
for the abundant resources. For example, in Vancouver, British Columbia, Varied Thrushes (a
common urban species during non-breeding in this study) are 76.9 times more likely to
collide with glass compared to other species (De Groot et al., 2021). Homogenization can
also have a negative, destabilizing effect on bird communities (Devictor et al., 2007), but
targeted urban habitat management is important for proper bird conservation, not just during
non-breeding, but also year-round. Community composition is unique in urban habitats and
birds occupying this space face unique challenges that must be examined and considered
within the unique and complex environments that urban areas create.
In summary, I demonstrate the importance of identifying where birds are and what
habitats they are using across the entire year and not only during distinct periods within the
annual cycle (e.g. breeding). In addition, I emphasize caution in applying broad
generalization about patterns of bird abundance, richness, and communities at the continental
scale without adequate geographic and context specific knowledge and understanding. Urban
habitats provide important habitat for birds year-round yet are often only considered to
support a limited number of species, but my study suggests that urban areas provide
important resources for a large variety of bird species throughout the year and especially
during migration and non-breeding seasons. Effective conservation of birds and habitat
management requires the consideration of the entire annual cycle, the unique challenges that
birds face in western North America, and their seasonal changes in habitat use. Thus, more

48
research is needed that addresses and studies birds across the entire year and among a variety
of ecozones and habitats.

49
2.5 Literature Cited
Adhikari, A., & Hansen, A. J. (2018). Land use change and habitat fragmentation of wildland
ecosystems of the North Central United States. Landscape and Urban Planning, 177,
196–216. https://doi.org/10.1016/j.landurbplan.2018.04.014
Alldritt-McDowell, J., Coupé, R., & Meidinger, M. (1998, July). The Ecology of the
Bunchgrass Zone. Ministry of Forests, British Columbia.
https://www.for.gov.bc.ca/hfd/pubs/docs/Bro/bro54.pdf
Atwell, J. W., O’Neal, D. M., & Ketterson, Ellen. D. (2011). Animal migration as a moving
target for conservation: Intraspecies variation and responses to environmental change, as
illustrated in a sometimes migratory songbird. Environmental Law (Northwestern School
of Law), 41(2), 289–316.
Bates, D., Mächler, M., Bolker, B., & Walker, S. (2015). Fitting Linear Mixed-Effects
Models Using lme4. Journal of Statistical Software, 67, 1–48.
https://doi.org/10.18637/jss.v067.i01
Bauer, S., & Hoye, B. J. (2014). Migratory Animals Couple Biodiversity and Ecosystem
Functioning Worldwide. Science, 344(6179), 1242552.
https://doi.org/10.1126/science.1242552
Beissinger, S. R., & Osborne, D. R. (1982). Effects of Urbanization on Avian Community
Organization. The Condor, 84(1), 75–83. https://doi.org/10.2307/1367825
Blair, R. B. (1996). Land Use and Avian Species Diversity Along an Urban Gradient.
Ecological Applications, 6(2), 506–519. https://doi.org/10.2307/2269387
Bonnet-Lebrun, A.-S., Manica, A., & Rodrigues, A. S. L. (2020). Effects of urbanization on
bird migration. Biological Conservation, 244, 108423.
https://doi.org/10.1016/j.biocon.2020.108423
Boyle, W. A. (2017). Altitudinal bird migration in North America. The Auk, 134(2), 443–465.
https://doi.org/10.1642/AUK-16-228.1
Bozdogan, H. (1987). Model selection and Akaike’s Information Criterion (AIC): The
general theory and its analytical extensions. Psychometrika, 52(3), 345–370.
https://doi.org/10.1007/BF02294361

50
Buckland, S., Anderson, D., Burnham, K., Laake, J., Borchers, D., & Thomas, L. (2001).
Introduction to Distance Sampling: Estimating Abundance of Biological Populations. In
Oxford University Press (Vol. xv). https://doi.org/10.1093/oso/9780198506492.001.0001
Butler, C. J. (2003). The disproportionate effect of global warming on the arrival dates of
short-distance migratory birds in North America. Ibis, 145(3), 484–495.
https://doi.org/10.1046/j.1474-919X.2003.00193.x
Callaghan, C. T., Palacio, F. X., Benedetti, Y., Morelli, F., & Bowler, D. E. (2023). Largescale spatial variability in urban tolerance of birds. Journal of Animal Ecology, 92(2),
403–416. https://doi.org/10.1111/1365-2656.13862
Carlisle, J. D., Skagen, S. K., Kus, B. E., Riper, C. V., III, Paxtons, K. L., & Kelly, J. F.
(2009). Landbird Migration in the American West: Recent Progress and Future Research
Directions. The Condor, 111(2), 211–225. https://doi.org/10.1525/cond.2009.080096
Chandler, R. B., Royle, J. A., & King, D. I. (2011). Inference about density and temporary
emigration in unmarked populations. Ecology, 92(7), 1429–1435.
https://doi.org/10.1890/10-2433.1
Clement, M. J., Nichols, J. D., Collazo, J. A., Terando, A. J., Hines, J. E., & Williams, S. G.
(2019). Partitioning global change: Assessing the relative importance of changes in
climate and land cover for changes in avian distribution. Ecology and Evolution, 9(4),
1985–2003. https://doi.org/10.1002/ece3.4890
Clergeau, P., Savard, J.-P. L., Mennechez, G., & Falardeau, G. (1998). Bird Abundance and
Diversity along an Urban-Rural Gradient: A Comparative Study between Two Cities on
Different Continents. The Condor, 100(3), 413–425. https://doi.org/10.2307/1369707
De Groot, K. L., Porter, A. N., Norris, A. R., Huang, A. C., & Joy, R. (2021). Year-round
monitoring at a Pacific coastal campus reveals similar winter and spring collision
mortality and high vulnerability of the Varied Thrush. Ornithological Applications,
123(3), duab027. https://doi.org/10.1093/ornithapp/duab027
Deshwal, A., Panwar, P., Neal, J. C., & Young, M. A. (2021). Using long-term citizen science
data to understand distribution and habitat use of an irruptive species. Ecological
Informatics, 64, 101377. https://doi.org/10.1016/j.ecoinf.2021.101377
Devictor, V., Julliard, R., Couvet, D., Lee, A., & Jiguet, F. (2007). Functional homogenization
effect of urbanization on bird communities. Conservation Biology: The Journal of the

51
Society for Conservation Biology, 21(3), 741–751. https://doi.org/10.1111/j.15231739.2007.00671.x
Dokter, A. M., Farnsworth, A., Fink, D., Ruiz-Gutierrez, V., Hochachka, W. M., La Sorte, F.
A., Robinson, O. J., Rosenberg, K. V., & Kelling, S. (2018). Seasonal abundance and
survival of North America’s migratory avifauna determined by weather radar. Nature
Ecology & Evolution, 2(10), 1603–1609. https://doi.org/10.1038/s41559-018-0666-4
eBird. (2022). eBird [eBird, Cornell Lab of Ornithology, Ithaca, New York.]. eBird: An
Online Database of Bird Distribution and Abundance. https://ebird.org/region/CA-BCOS?yr=all
Emlen, J. T. (1974). An Urban Bird Community in Tucson, Arizona: Derivation, Structure,
Regulation. The Condor, 76(2), 184–197. https://doi.org/10.2307/1366729
Esri. (2023, February 23). Esri. “World Imagery (Wayback 2023-02-23)”.
https://www.arcgis.com/home/item.html?id=fb0f4638a6894bddb3292499a0aeb7b3
Fiske, I., & Chandler, R. (2011). unmarked: An R Package for Fitting Hierarchical Models of
Wildlife Occurrence and Abundance. Journal of Statistical Software, 43, 1–23.
https://doi.org/10.18637/jss.v043.i10
Fontana, S., Sattler, T., Bontadina, F., & Moretti, M. (2011). How to manage the urban green
to improve bird diversity and community structure. Landscape and Urban Planning,
101(3), 278–285. https://doi.org/10.1016/j.landurbplan.2011.02.033
Fudickar, A. M., Jahn, A. E., & Ketterson, E. D. (2021). Animal Migration: An Overview of
One of Nature’s Great Spectacles. Annual Review of Ecology, Evolution, and Systematics,
52(Volume 52, 2021), 479–497. https://doi.org/10.1146/annurev-ecolsys-012021-031035
Fuentes, M., Van Doren, B. M., Fink, D., & Sheldon, D. (2023). BirdFlow: Learning seasonal
bird movements from eBird data. Methods in Ecology and Evolution, 14(3), 923–938.
https://doi.org/10.1111/2041-210X.14052
Gaston, K. J., Blackburn, T. M., & Klein Goldewijk, K. (2003). Habitat conversion and
global avian biodiversity loss. Proceedings of the Royal Society B: Biological Sciences,
270(1521), 1293–1300. https://doi.org/10.1098/rspb.2002.2303

52
Grimm, N. B., Faeth, S. H., Golubiewski, N. E., Redman, C. L., Wu, J., Bai, X., & Briggs, J.
M. (2008). Global Change and the Ecology of Cities. Science, 319(5864), 756–760.
https://doi.org/10.1126/science.1150195
Grubb, T. C., & Cimprich, D. A. (1990). Supplementary Food Improves the Nutritional
Condition of Wintering Woodland Birds: Evidence from Ptilochronology. Ornis
Scandinavica (Scandinavian Journal of Ornithology), 21(4), 277–281.
https://doi.org/10.2307/3676392
Herbers, J. R., Serrouya, R., & Maxcy, K. A. (2004). Effects of elevation and forest cover on
winter birds in mature forest ecosystems of southern British Columbia. Canadian Journal
of Zoology, 82(11), 1720–1730. https://doi.org/10.1139/z04-151
Horton, K. G., Morris, S. R., Van Doren, B. M., & Covino, K. M. (2023). Six decades of
North American bird banding records reveal plasticity in migration phenology. Journal of
Animal Ecology, 92(3), 738–750. https://doi.org/10.1111/1365-2656.13887
Huff, M. H., Bettinger, K. A., Ferguson, H. L., Brown, M. J., & Altman, Bob. (2000). A
habitat-based point-count protocol for terrestrial birds, emphasizing Washington and
Oregon. (No. PNW-GTR-501; p. PNW-GTR-501). U.S. Department of Agriculture, Forest
Service, Pacific Northwest Research Station. https://doi.org/10.2737/PNW-GTR-501
Huston, M. A., & Wolverton, S. (2009). The global distribution of net primary production:
Resolving the paradox. Ecological Monographs, 79(3), 343–377.
https://doi.org/10.1890/08-0588.1
Isaksson, C. (2018). Impact of Urbanization on Birds. In D. T. Tietze (Ed.), Bird Species:
How They Arise, Modify and Vanish (pp. 235–257). Springer International Publishing.
https://doi.org/10.1007/978-3-319-91689-7_13
Johnson, M. D., & Sherry, T. W. (2001). Effects of food availability on the distribution of
migratory warblers among habitats in Jamaica. Journal of Animal Ecology, 70(4), 546–
560. https://doi.org/10.1046/j.1365-2656.2001.00522.x
Johnston, R. F. (2001). Synanthropic birds of North America. In J. M. Marzluff, R. Bowman,
& R. Donnelly (Eds.), Avian Ecology and Conservation in an Urbanizing World (pp. 49–
67). Springer US. https://doi.org/10.1007/978-1-4615-1531-9_3

53
Jokimäki, J., Suhonen, J., Inki, K., & Jokinen, S. (1996). Biogeographical comparison of
winter bird assemblages in urban environments in Finland. Journal of Biogeography,
23(3), 379–386. https://doi.org/10.1046/j.1365-2699.1996.00033.x
Kellner, K. F., Smith, A. D., Royle, J. A., Kéry, M., Belant, J. L., & Chandler, R. B. (2023).
The unmarked R package: Twelve years of advances in occurrence and abundance
modelling in ecology. Methods in Ecology and Evolution, 14(6), 1408–1415.
https://doi.org/10.1111/2041-210X.14123
Kuznetsova, A., Brockhoff, P. B., & Christensen, R. H. B. (2017). lmerTest Package: Tests in
Linear Mixed Effects Models. Journal of Statistical Software, 82, 1–26.
https://doi.org/10.18637/jss.v082.i13
La Sorte, F. A., Fink, D., Hochachka, W. M., DeLong, J. P., & Kelling, S. (2014). Spring
phenology of ecological productivity contributes to the use of looped migration strategies
by birds. Proceedings of the Royal Society B: Biological Sciences, 281(1793), 20140984.
https://doi.org/10.1098/rspb.2014.0984
La Sorte, F. A., Fink, D., Hochachka, W. M., Farnsworth, A., Rodewald, A. D., Rosenberg, K.
V., Sullivan, B. L., Winkler, D. W., Wood, C., & Kelling, S. (2014). The role of
atmospheric conditions in the seasonal dynamics of North American migration flyways.
Journal of Biogeography, 41(9), 1685–1696. https://doi.org/10.1111/jbi.12328
Lancaster, R. K., & Rees, W. E. (1979). Bird communities and the structure of urban habitats.
Canadian Journal of Zoology, 57(12), 2358–2368. https://doi.org/10.1139/z79-307
Leech, D. I., & Crick, H. Q. P. (2007). Influence of climate change on the abundance,
distribution and phenology of woodland bird species in temperate regions. Ibis, 149(s2),
128–145. https://doi.org/10.1111/j.1474-919X.2007.00729.x
Leveau, L. M., Jokimäki, J., & Kaisanlahti-Jokimäki, M.-L. (2021). Urbanization buffers
seasonal change in composition of bird communities: A multi-continental meta-analysis.
Journal of Biogeography, 48(10), 2391–2401. https://doi.org/10.1111/jbi.14236
Mac Arthur, R. H. (1959). On the Breeding Distribution Pattern of North American Migrant
Birds. The Auk, 76(3), 318–325. https://doi.org/10.2307/4081809
Marra, P. P., Cohen, E. B., Loss, S. R., Rutter, J. E., & Tonra, C. M. (2015). A call for full
annual cycle research in animal ecology. Biology Letters, 11(8), 20150552.
https://doi.org/10.1098/rsbl.2015.0552

54
Martijn, B., Jim, C., Natalie, B., & Thomas, N. (2023). Simulation-based assessment of the
performance of hierarchical abundance estimators for camera trap surveys of unmarked
species. Scientific Reports, 13(1), 16169. https://doi.org/10.1038/s41598-023-43184-w
McCune, B., & Grace, J. B. (2002). Analysis of ecological communities.
https://www.waterbouwkundiglaboratorium.be/nl/catalogus
Mckinney, M. (2006). Urbanization as a major cause of biotic homogenization. Biological
Conservation, 127, 247–260. https://doi.org/10.1016/j.biocon.2005.09.005
McKinney, M. L. (2002). Urbanization, Biodiversity, and Conservation: The impacts of
urbanization on native species are poorly studied, but educating a highly urbanized human
population about these impacts can greatly improve species conservation in all
ecosystems. BioScience, 52(10), 883–890. https://doi.org/10.1641/00063568(2002)052[0883:UBAC]2.0.CO;2
Mills, G. S., Dunning, J. B., Jr., & Bates, J. M. (1989). Effects of Urbanization on Breeding
Bird Community Structure in Southwestern Desert Habitats. The Condor, 91(2), 416–428.
https://doi.org/10.2307/1368320
Morales, A., Frei, B., Mitchell, G. W., Bégin-Marchand, C., & Elliott, K. H. (2022). Reduced
diurnal activity and increased stopover duration by molting Swainson’s Thrushes.
Ornithology, 139(2), ukab083. https://doi.org/10.1093/ornithology/ukab083
Morelli, F., Benedetti, Y., Ibáñez-Álamo, J. D., Jokimäki, J., Mänd, R., Tryjanowski, P., &
Møller, A. P. (2016). Evidence of evolutionary homogenization of bird communities in
urban environments across Europe. Global Ecology and Biogeography, 25(11), 1284–
1293. https://doi.org/10.1111/geb.12486
Murphy, M. T. (2003). Avian Population Trends Within the Evolving Agricultural Landscape
of Eastern and Central United States. The Auk, 120(1), 20–34.
https://doi.org/10.1093/auk/120.1.20
Murthy, A. C., Fristoe, T. S., & Burger, J. R. (2016). Homogenizing effects of cities on North
American winter bird diversity. Ecosphere, 7(1), e01216.
https://doi.org/10.1002/ecs2.1216
Okanagan Similkameen Stewardship [OSS]. (2023). Birds. Osstewardship.
https://www.osstewardship.ca/birds

55
Oksanen, J., Simpson, G. L., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., O’Hara,
R. B., Solymos, P., Stevens, M. H. H., Szoecs, E., Wagner, H., Barbour, M., Bedward, M.,
Bolker, B., Borcard, D., Carvalho, G., Chirico, M., Caceres, M. D., Durand, S., …
Weedon, J. (2024). vegan: Community Ecology Package (Version 2.6-6.1) [Computer
software]. https://cran.r-project.org/web/packages/vegan/index.html
Ontario Breeding Bird Atlas. (2021, June). Ontario Breeding Bird Atlas: Instructions for
Point Counts. https://www.birdsontario.org/wp-content/uploads/Instructions-for-PointCounts-June-2021.pdf
Ortega-Álvarez, R., & MacGregor-Fors, I. (2009). Living in the big city: Effects of urban
land-use on bird community structure, diversity, and composition. Landscape and Urban
Planning, 90(3), 189–195. https://doi.org/10.1016/j.landurbplan.2008.11.003
Poirier, V., Frei, B., Lefvert, M., Morales, A., & Elliott, K. H. (2024). Moult migrant
Tennessee Warblers undergo extensive stopover in peri-urban forests of southern Quebec.
Canadian Journal of Zoology, 102(3), 272–285. https://doi.org/10.1139/cjz-2023-0109
QGIS.org. (2024). QGIS Geographic Information System. QGIS Association.
http://www.qgis.org
R Core Team. (2021). R: A language and environment for statistical computing. [R
Foundation for Statistical Computing]. https://www.R-project.org/
Rakhimov, I. I., & Ibragimova, K. K. (2018). Morphological and ecological preadaptations as
the basis of bird synanthropization under transformed environment conditions. IOP
Conference Series: Earth and Environmental Science, 107(1), 012089.
https://doi.org/10.1088/1755-1315/107/1/012089
Rigal, S., Dakos, V., Alonso, H., Auniņš, A., Benkő, Z., Brotons, L., Chodkiewicz, T.,
Chylarecki, P., de Carli, E., del Moral, J. C., Domşa, C., Escandell, V., Fontaine, B.,
Foppen, R., Gregory, R., Harris, S., Herrando, S., Husby, M., Ieronymidou, C., …
Devictor, V. (2023). Farmland practices are driving bird population decline across Europe.
Proceedings of the National Academy of Sciences, 120(21), e2216573120.
https://doi.org/10.1073/pnas.2216573120
Royle, J. A., Dawson, D. K., & Bates, S. (2004). Modeling Abundance Effects in Distance
Sampling. Ecology, 85(6), 1591–1597. https://doi.org/10.1890/03-3127

56
Rueda-Uribe, C., Herrera-Alsina, L., Lancaster, L. T., Capellini, I., Layton, K. K. S., &
Travis, J. M. J. (2024). Citizen science data reveal altitudinal movement and seasonal
ecosystem use by hummingbirds in the Andes Mountains. Ecography, 2024(3), e06735.
https://doi.org/10.1111/ecog.06735
Sahana, M., Ravetz, J., Patel, P. P., Dadashpoor, H., & Follmann, A. (2023). Where Is the
Peri-Urban? A Systematic Review of Peri-Urban Research and Approaches for Its
Identification and Demarcation Worldwide. Remote Sensing, 15(5), Article 5.
https://doi.org/10.3390/rs15051316
Saunders, S. P., Meehan, T. D., Michel, N. L., Bateman, B. L., DeLuca, W., Deppe, J. L.,
Grand, J., LeBaron, G. S., Taylor, L., Westerkam, H., Wu, J. X., & Wilsey, C. B. (2022).
Unraveling a century of global change impacts on winter bird distributions in the eastern
United States. Global Change Biology, 28(7), 2221–2235.
https://doi.org/10.1111/gcb.16063
Semenchuk, P., Plutzar, C., Kastner, T., Matej, S., Bidoglio, G., Erb, K.-H., Essl, F., Haberl,
H., Wessely, J., Krausmann, F., & Dullinger, S. (2022). Relative effects of land conversion
and land-use intensity on terrestrial vertebrate diversity. Nature Communications, 13(1),
615. https://doi.org/10.1038/s41467-022-28245-4
Somveille, M., Bay, R. A., Smith, T. B., Marra, P. P., & Ruegg, K. C. (2021). A general
theory of avian migratory connectivity. Ecology Letters, 24(9), 1848–1858.
https://doi.org/10.1111/ele.13817
Somveille, M., Rodrigues, A. S. L., & Manica, A. (2015). Why do birds migrate? A
macroecological perspective. Global Ecology and Biogeography, 24(5/6), 664–674.
Stagoll, K., Manning, A. D., Knight, E., Fischer, J., & Lindenmayer, D. B. (2010). Using
bird–habitat relationships to inform urban planning. Landscape and Urban Planning,
98(1), 13–25. https://doi.org/10.1016/j.landurbplan.2010.07.006
Stanton, R. L., Morrissey, C. A., & Clark, R. G. (2018). Analysis of trends and agricultural
drivers of farmland bird declines in North America: A review. Agriculture, Ecosystems &
Environment, 254, 244–254. https://doi.org/10.1016/j.agee.2017.11.028
Stein, A., Gerstner, K., & Kreft, H. (2014). Environmental heterogeneity as a universal driver
of species richness across taxa, biomes and spatial scales. Ecology Letters, 17(7), 866–
880. https://doi.org/10.1111/ele.12277

57
Swanwick, C., Dunnett, N., & Woolley, H. (2003). Nature, Role and Value of Green Space in
Towns and Cities: An Overview. Built Environment (1978-), 29(2), 94–106.
Tilghman, N. G. (1987). Characteristics of urban woodlands affecting breeding bird diversity
and abundance. Landscape and Urban Planning, 14, 481–495.
https://doi.org/10.1016/0169-2046(87)90061-2
Tu, H.-M., Fan, M.-W., & Ko, J. C.-J. (2020). Different Habitat Types Affect Bird Richness
and Evenness. Scientific Reports, 10(1), 1221. https://doi.org/10.1038/s41598-020-582024
Valiela, I., & Martinetto, P. (2007). Changes in Bird Abundance in Eastern North America:
Urban Sprawl and Global Footprint? BioScience, 57(4), 360–370.
https://doi.org/10.1641/B570410
van Riper III, C., Paxton, K. L., van Riper, C. J., van Riper, K. A., & McGrath, L. (2008).
The Role of Protected Areas as Bird Stop-over Habitat: Ecology and Habitat Utilization
by Migrating Land Birds within Colorado River Riparian Forests of Southwestern North
America. Nebraska Cooperative Fish and Wildlife Research Unit - Staff Publications., 78.
van Vliet, H. E. J., Stutchbury, B. J. M., Newman, A. E. M., & Norris, D. R. (2020). The
impacts of agriculture on an obligate grassland bird of North America. Agriculture,
Ecosystems & Environment, 287, 106696. https://doi.org/10.1016/j.agee.2019.106696
Watson, H., Nilsson, J.-Å., Smith, E., Ottosson, F., Melander, O., Hegemann, A., Urhan, U.,
& Isaksson, C. (2024). Urbanisation-associated shifts in the avian metabolome within the
annual cycle. Science of The Total Environment, 944, 173624.
https://doi.org/10.1016/j.scitotenv.2024.173624
Wilson, S., & Martin, K. (2005). Songbird use of high-elevation habitat during the fall postbreeding and migratory periods. Écoscience, 12(4), 561–568.
Wolff, S., Mdemu, M. V., & Lakes, T. (2021). Defining the Peri-Urban: A Multidimensional
Characterization of Spatio-Temporal Land Use along an Urban–Rural Gradient in Dar es
Salaam, Tanzania. Land, 10(2), Article 2. https://doi.org/10.3390/land10020177
Yaukey, P. H. (1996). Patterns of Avian Population Density, Habitat Use, and Flocking
Behavior in Urban and Rural Habitats During Winter. The Professional Geographer,
48(1), 70–81. https://doi.org/10.1111/j.0033-0124.1996.00070.x

58
Zuckerberg, B., Bonter, D. N., Hochachka, W. M., Koenig, W. D., DeGaetano, A. T., &
Dickinson, J. L. (2011). Climatic constraints on wintering bird distributions are modified
by urbanization and weather. Journal of Animal Ecology, 80(2), 403–413.
https://doi.org/10.1111/j.1365-2656.2010.01780.x
Zuckerberg, B., Fink, D., La Sorte, F. A., Hochachka, W. M., & Kelling, S. (2016). Novel
seasonal land cover associations for eastern North American forest birds identified
through dynamic species distribution modelling. Diversity and Distributions, 22(6), 717–
730. https://doi.org/10.1111/ddi.12428

59
Chapter 3: Seasonal variation in domestic cat (Felis catus) abundance, population size,
and habitat associations across a mosaic of habitats
3.1 Introduction
Domestic cats (Felis catus) have retained their ancestral hunting behaviours, enabling them
to adapt to a variety of environments, both with and without human intervention (Bradshaw,
2006; Canadian Federation of Humane Societies [CFHS], 2017). These innate hunting
abilities allow cats to capture a variety of wild animals, posing serious global threats to small
vertebrate populations when they are allowed to wander unsupervised outdoors (i.e.,
wandering cats; Loss et al., 2013; Trouwborst et al., 2020). When cats are allowed to wander
freely outside, they may experience increased risks of injury and illness through predation by
other animals, vehicle collisions, and disease transmission of illnesses such as toxoplasmosis
(reviewed in Tan et al., 2020). Consequently, various groups (e.g., veterinarians, shelters,
humane societies, conservation organizations, policy makers, and members of the public) are
concerned about issues surrounding wandering cats (detailed in Saunders et al., 2021; Wald
& Peterson, 2020) and find common ground in the desire to curtail the number of wandering
cats.
Data on local cat abundance and population sizes is needed for municipal
governments to make informed policy changes and for organizations, including nongovernmental organizations such as the Stewardship Centre for British Columbia, to develop
educational campaigns and incentive programs promoting responsible cat ownership. The
absence of such data on wandering cat populations poses a significant challenge to
implementing effective management strategies, including municipal bylaws and policies.
Furthermore, abundance and population estimates are valuable to establish a baseline for

60
wandering cat populations, which is crucial for informing future management decisions and
evaluating the effectiveness of implemented strategies. Yet, many studies on wandering cats
have focused on only the spring and summer seasons (Bennett et al., 2021; Coleman &
Temple, 1993; Flockhart et al., 2016; Hand, 2019; Hanmer et al., 2017; Kays & DeWan,
2004). Therefore, there is limited knowledge on the seasonal dynamics of wandering cat
abundance and the factors that influence seasonal variation.
Wandering cats may inhabit and use a variety of habitats, which may be influenced by
both human activities and the ecological needs of cats (reviewed by Tan et al., 2020), but
only one study has directly explored seasonal variation in wandering cat abundance in North
America (Clyde et al., 2022). They found that wandering cat local abundance estimates were
higher in spring and summer compared to the fall and winter (Clyde et al., 2022), suggesting
that human behaviour (e.g. owners not putting cats out as much in the winter) may influence
cat abundances (Human Behaviour Hypothesis). Another study, conducted in Australia
examined broad patterns of seasonal variation, showing that cat abundance decreased during
dry periods and increased following periods of rainfall, suggesting that prey availability may
drive these patterns (Prey Availability Hypothesis; Legge et al., 2017). Based on these two
hypotheses, if cat abundances were highest during periods with most favourable temperature
(i.e., not below 0°C nor above 25°C), then human behaviour might be the main factor
influencing seasonal patterns. In contrast, if cat abundances are related to peak prey
abundances, such as in the spring and fall, then prey availability may be influencing seasonal
patterns.
The distribution and associated abundances of wandering cats are strongly influenced
by human presence, cat ownership, and human and cat behaviour (Bennett et al., 2021; Clyde

61
et al., 2022; Cove et al., 2023; Gow et al., 2024). In Canada, approximately 37% of
households own at least one cat, with an average of 1.64 cats per household, and 28% of
households allow their cat(s) unsupervised outdoor access (CFHS, 2017). When outdoors,
cats typically have small home ranges, often remaining within 100 meters of their homes
(Dunford et al., 2024; Kays et al., 2020), suggesting the highest densities of cats will be in
areas with high human populations, such as urban areas where there is a high density of
people and buildings (Human Presence and Cat Behaviour Hypothesis). Less urbanized
areas, such as peri-urban areas, may also have high abundances of cats. Peri-urban areas tend
to have fewer buildings than urban areas and are typically surrounded by agricultural lands or
natural habitat. Cats living in peri-urban areas often have access to a wider variety of
bordering habitats and have larger home ranges compared to those in urban habitats (Hall et
al., 2016; Hanmer et al., 2017). Peri-urban households also frequently have more cats (Dozier
et al., 2023) and are more likely to allow their cats outdoors (Clancy et al., 2003). Therefore,
relative abundance estimates may be similar between urban and peri-urban habitats
depending on human and cat behaviour (Hall et al., 2016; Hanmer et al., 2017; Kays et al.,
2020).
Areas with lower levels of urbanization, such as agricultural and natural areas, are
expected to have lower abundances of wandering cats. Agricultural habitats, such as orchards
and vineyards, typically have few buildings, yet have similar numbers of cats per household
to peri-urban areas (Coleman & Temple, 1993; Dozier et al., 2023). In addition to owned
cats, unowned barn cats are common in agricultural areas (i.e., cats that may be fed but have
no restrictions on movement). These cats are often unrestricted in their movement and used
for pest control, leading to some cats displaying home ranges that are 1.6 times larger than
urban wandering cats (Hall et al., 2016; Hanmer et al., 2017; Kays et al., 2020). Despite large

62
home ranges and high densities on farm, the low human population densities and Human
Behaviour Hypothesis lead me to expect that the abundance of wandering cats in agricultural
habitats will be lower than urban and peri-urban habitats. Finally, natural habitats, which are
characterized, in this study, by forests, deserts, wetlands, and few buildings, would have the
lowest abundances of cats, if any. Cats within these areas likely have the largest home ranges
(~ 70% larger than urban cats; Pirie et al., 2022) and may live in nearby households or be
unowned and survive without (or with minimal) human intervention.
Here, I use trail cameras to quantify cat local abundance across the entire year and
four land classes (urban, peri-urban, agricultural, and natural) in the south Okanagan Valley,
British Columbia (from Okanagan Falls to Osoyoos). I have three objectives: 1) Identify how
cat abundance and population size vary within the study region and across habitat types and
seasons; 2) Examine the Prey Availability and Human Behaviour Hypotheses by assessing
how local abundance of wandering cats varies seasonally; and 3) Test the Human Presence
and Cat Behaviour Hypothesis by examining how local cat abundance varies across habitat
types.
3.2 Methods
3.2.1 Study Area
I studied seasonal patterns of wandering domestic cat abundance and habitat associations in
the south Okanagan Valley of British Columbia, Canada (within and surrounding the
communities of Okanagan Falls, 49.34490N, -119.57328E; Oliver, 49.18276N, -119.54959E;
and Osoyoos, 49.031778N, -119.465111E). My study area extends laterally to the hills that
border the valley, covering a total area of 161.2 km2 (Figure 3.1), with a total human
population of ~22,217 (see Appendix 8 for more details of the study site boundaries and how

63
population was determined). Located in the south-central region of British Columbia, the
south Okanagan Valley is the traditional, unceded, and ancestral territory of the Syilx
Okanagan Nation, and one of the most endangered ecosystems in Canada. The Okanagan
Valley, including the study region, is heavily settled and has some of the highest numbers of
species at risk in Canada (Coristine et al., 2018). Given this, there are numerous conservation
plans, initiatives, and programs focused on species at risk recovery and habitat conservation
in the region (e.g., Restoration of Okanagan Sockeye Salmon, Alexander et al., 2024 and
Correia et al., 2024; Burrowing Owl recovery, Environment and Climate Change Canada,
2012; and Yellow-breasted Chat recovery, McKibbin & Bishop, 2008). It is part of the
bunchgrass biogeoclimatic zone that is characterized by grassland, shrub-steppe, wetlands,
and dry open forests consisting of ponderosa pines, cottonwoods, and Douglas firs (AlldrittMcDowell et al., 1998). The fertile lands, abundant lakes and rivers, long growing season,
and warm climate also make this area highly attractive for agriculture (e.g. vineyards, and
orchards), long-term residents, and seasonal recreation. This area consists of a mosaic of
natural habitats (grasslands, deserts, riparian areas, and forests), agriculture (e.g. vineyards
and orchards), small urban centers, and peri-urban areas (defined in Table 3.1).

64
Table 3.1. Description of detection and site covariates used to estimate relative abundance
of cats in the south Okanagan Valley, B.C. Detection covariates influenced the detection of
wandering cats and site covariates describe the area around the trail camera.
Variables

Description

Detection covariates
date

Julian date when the point count was conducted.

time

Time at the start of the point count observational period.

wind

Strength of wind during point count using the Beaufort scale (Scale 0-4)

cloud

Cloud cover % overhead during the point count.

precip

Binary observation during point count, where 0 = no precipitation and 1 =
precipitation.

Site covariates
desert

Percentage of desert within a 200 m buffer around the point count
location.

field

Percentage of field within a 200 m buffer around the point count location.
Fields include public spaces such as parks and school yards, land that is
used as pasture or farming, and empty, overgrown lots.

orchard

Percentage of orchard within a 200 m buffer around the point count
location.

vineyard

Percentage of vineyard within a 200 m buffer around the point count
location.

forest

Percentage of forest within a 200 m buffer around the point count location.

water

Percentage of water within a 200 m buffer around the point count location.

roads

Privateroads and publicroads combined, to equal the total length of roads
(m) within a 200 m buffer around the point count location.

privateroad

Total length of private roads, in meters, within a 200 m buffer around the
point count location. Private roads include driveways (>40 m in length),
vineyard/orchard roads, and other impervious surfaces that are not
accessible to the public.

publicroad

Total length of public roads, in meters, within a 200 m buffer around the
point count location. Public roads are maintained by the city and are easily
accessible to the public.

buildings

All buildings within a 200 m buffer around the point count locations. This
includes high-density housing and all other buildings.

landclass

The land class type of the point count locations: urban, peri-urban, natural,
or agricultural

urban

Site buffer (200 m) containing >55 buildings

peri-urban

Site buffer (200 m) with 17 to 50 buildings and <49% vineyard and/or
orchard.

agricultural

Site buffer (200 m) characterized by >28% agriculture and <20 buildings.

natural

Site buffer (200 m) characterized by >25% natural habitat, <16 buildings,
and <29% agriculture.

altitude

Altitude (m) at the point count location.

65
3.2.2 Study Design
I used 60 trail cameras to assess seasonal patterns of cat abundance, population size, and
habitat associations using a repeated measures design stratified by season, collecting data
across five distinct seasonal periods relevant to birds (a main prey sources of cats) from
March 2022 to March 2023. The five seasonal periods assessed were spring (March 8 – May
27, 2022), summer (May 27 – July 30, 2022), fall (August 8 – October 13, 2022), early
winter (October 13 – December 16, 2022), and late winter (January 13 – March 23, 2023).
This research was part of a larger project also looking at bird abundance and habitat
associations, and the resources available to birds are different throughout the winter; thus,
winter was divided into early and late seasons. Throughout the study, some sites were added,
discarded, or moved, leading to a total of 123 sites (see Appendix 1 for details). Total number
of sites per season were as follows: 120 sites for spring migration, breeding season, and fall
migration; 119 sites for early non-breeding season; and 114 sites for late non-breeding
season. Trail cameras were deployed for at least 4 weeks at each site during each season, then
moved to a second site within the same season for four weeks. I established these trail camera
sites using a random stratified design, following a framework for completing ethical camera
trapping fieldwork in urban environments (i.e., five P’s from Herrera et al., 2021:
partnership, planning, placements, public participation, and processing).

66

Figure 3.1. Map of the study area in the south Okanagan Valley showing trail camera
locations in Okanagan Falls, Oliver, and Osoyoos. Location colours are classified by the
following land classes: Agriculture (AG), Natural (NA), Peri-urban (PU), and Urban (UR).
The red boarder represents the study area that was used to study wandering cats based on the
perimeter of the grid. The inset map in the top right corner shows a red circle around the
south Okanagan Valley, located in the south-central area of British Columbia, Canada. Map is
in EPSG:3005 coordinate system (NAD83, BC Albers projection).

67
To determine wandering cat local abundance and habitat associations across an entire
year, I used satellite maps of the south Okanagan Valley to identify habitats as urban, periurban, agricultural, and natural based on broad definitions of each land class type. At the
stage of camera deployment, these broad categories helped me to identify which permits I
would need and the neighbourhoods to target for door-to-door recruitment. For the purposes
of camera deployment, I categorized urban areas as having extensive anthropogenic
development that included a matrix of landscapes such as pavement, buildings (residential
and commercial), and habitat patches (e.g., transportation corridors, riverbanks, and parks;
Swanwick et al., 2003). Peri-urban areas consisted of clusters of houses surrounded by
agricultural (e.g., orchards, vineyards, and crops) and/or natural (e.g., desert, forest, riparian)
land with approximately 25-50 m between houses. Agricultural land was used for agriculture
(such as vineyards and orchards) or other farming practices, targeting smaller
vineyards/orchards that were <50 hectares or houses residing next to agricultural land.
Natural habitats included any protected areas conserved provincially, federally, or privately.
Since these descriptors are continuous, I later classified sites using QGIS to create discrete
variables and refined categories based on surrounding habitat variables for statistical analysis
(see 3.2.3 GIS Analysis for more details).
When choosing sites, true randomization and equal distribution of cameras across
habitats during each 4-week deployment period was impossible because locations were
mainly on private property and therefore dependent on the willingness of landowners and
land managers (except natural sites). However, we aimed to have as equal distribution of
cameras across habitats as possible during each deployment period. I balanced the
distribution of cameras in each of the four land class types: urban (N = 32), peri-urban (N =
27), agricultural (N = 31), and natural (N = 33). This research was co-developed and

68
designed with the Stewardship Centre for British Columbia and the Canadian Wildlife
Service and Wildlife Research Division of Environment and Climate Change Canada (ECCC;
key contributors and co-authors on this paper). I used several methods to recruit participants
including word-of-mouth, networking, and media (see Appendix 9 for further details).
Overall, I found that the most successful way of forming partnerships was by going door-todoor to speak with the residents directly and ask if they would be willing to participate in the
study and providing them a letter to explain the study and provide contact information
(Additional details in Appendix 10; Example letter in Appendix 11).
Once I established trail camera locations, I considered the placement of cameras to
maximize cat detection, while ensuring the privacy of residents and the public. I used
Browning Strike Force HD Pro infrared trail cameras (model BTC-5HDP) with a 1/3s picture
trigger and a five-photo burst, which took enough pictures to identify individual cats and fastmoving animals (Clyde et al., 2022; see Appendix 12 for other technical considerations when
using trail cameras). When possible, cameras were pointed towards animal corridors, such as
fence lines and accessible paths to increase chances of detecting cats while avoiding
capturing pictures of humans (see Appendix 13 for additional ethical considerations I took
when placing cameras). Cameras were randomly placed as best as possible (given land use
permissions, available posts for attaching cameras, and resident concerns or limitations) with
a minimum of ~100 meters between cameras to maintain statistical independence, which is
the maximum distance I expect cats to wander based on another study suggesting cats stay
with ~100 m of their homes (Kays et al., 2020), Only two sets of sites (n = 4) were closer
than the desired 100 m apart (70 m and 40 m; see section 3.2.8 Statistical analysis to see how
I maintained statistical independence when one cat was identified at multiple sites).

69
Once pictures were collected, I processed ~7.1 million trail camera images to identify
individual cats and create encounter histories of wandering cats over the 123 trail camera
sites. Once the images were collected, images of human faces or other identifying features
were removed and deleted. Images were processed by individually identifying cats based on
their colour, coat pattern, and body size (Figure 3.2 highlights the diversity in coat pattern).
Using these markers, along with other features such as the presence and colour of collars, I
could identify 98% of cats to an individual level, which was consistent with other studies
(Clyde et al., 2022; Elizondo & Loss, 2016). It was only difficult to identify numerous black
cats at a trail camera site to an individual level, unless at night where pelage coats differ
under infrared photography. In cases where cats appeared identical, I followed the same
guidelines as Clyde et al. (2022) and Elizondo & Loss (2016), by assuming that there are as
many cats as seen at the same time in one photo. I used the identity of individual cats to
create encounter histories consisting of fourteen 24-hour sampling occasions (starting at
midnight), counting the total number of cats/camera site during the 24-hour sampling
occasion (see 3.2.8 Statistical analysis).

70

Figure 3.2. Domestic cat coat patterns, as seen on the trail cameras. Each cat has a different
coat pattern, making them individually identifiable.
3.2.3 GIS Analysis
Camera locations were stratified based on four broad habitat categories, but to extract
specific continuous habitat variables for statistical analyses, I used QGIS (Version 3.32 Lima;
QGIS.org, 2024; EPSG: 3005 coordinate system; NAD83, BC Albers projection). I assessed
habitat variables over the entire study area (Figure 3.1) that represented buildings, roads,
water, vineyards, orchards, deserts, fields, forest, and altitude (defined in Table 3.1). I
accessed detailed OpenStreetMap layers for roads, lakes, rivers, and buildings and imported
them into QGIS. To obtain missing habitat variables, I created polygons around forests,
rivers, deserts, fields, orchards, and vineyards, and updated the buildings using the point
creation function in QGIS. I used a 2023 Esri World Imagery Wayback satellite image as a
base layer to trace the polygons (Esri, 2023). I then created 200m buffers around each site to
extract habitat variable information using the intersect tool (see figure 3.3). Within each 200
m buffer, I quantified the total length of roads, the total number of buildings, and the

71
proportion of desert, field, forest, orchard, vineyard, and water (defined in Table 3.1). The
habitat variables extracted from the 200 m buffers were used to re-classify each trail camera
location as either urban, peri-urban, natural, or agricultural based on specific quantifiable
definitions (defined in Table 3.1; N = 100 correctly classified at start).

Figure 3.3. Example of 200-meter buffer divided into habitat variables. Variables within each
200 m buffer were used in the analyses.
3.2.4 Statistical analysis
To estimate local cat abundance, I used Hierarchical N-mixture models within the unmarked
package (Fiske & Chandler, 2011) in Program R, V2.1.1. (R Core Team, 2021), specifically
using the pcount model. I followed the methods in Clyde et al. (2022) and Gow et al. (2024),
which, in short, involves using hierarchical N-mixture models that incorporate habitat and
site covariates to estimate abundance and detection of domestic cats. Detection is the
probability (p) of detecting an individual, given that there are N individuals available (Kéry
& Royle, 2016). I used the number of individuals detected at a site during a 24-hour sampling
occasion, while accounting for overdispersion and non-perfect detection (Kéry & Royle,
2016). This also keeps sampling periods consistent, reducing over- or under-estimation of

72
local abundance estimates. The ‘unmarked’ model incorporates repeated measures and
detection probability to estimate local abundance, while accounting for species that may go
undetected during sampling (Fiske & Chandler, 2011). This hierarchical model is made up of
two parts: a detection model and an abundance model (Kéry & Royle, 2016; equations in
Appendix 14).
I modelled detection using three variables: average daily temperature (temp), Julian
date (date), and total daily precipitation (precip; Table 3.1). Temperature can influence the
likelihood of owners allowing their cats outdoors, as they may be less likely to allow them
outdoors in extreme heat or cold, or cats may move around less during extreme temperatures
(Forrest et al., 2023). Date accounts for any temporal patterns in either cat owner behaviour
(e.g. people vacationing) or broader seasonal patterns of cats themselves. Cat owners may be
less likely to put their cats outdoors when it is raining or snowing, or cats may be less active
(and thus less likely to be detected) during periods of precipitation (Forrest et al., 2023;
Goszczyński et al., 2009). Temperature and precipitation data was from ECCC from the
Osoyoos CS weather station (49°01'41.850" N, 119°26'27.570" W; ECCC 2022). The
average seasonal temperatures were 9.6 °C during the spring, 21.2 °C during the summer,
20.6 °C during the fall, 0.7 °C during the early winter, and 1.4 °C during the late winter. The
total precipitation was 38.1 mm during the spring, 68.7 mm during the summer, 4.2 mm
during the fall, 69.3 mm during the early winter, and 32.9 mm during the late winter.
Assumptions of this model include that the population remains closed during the
sampling periods (i.e., each of the five seasons), all individuals have an equal probability of
detection, and no animals are counted twice within a season. I assumed there was little to no
death or birth within the population during each two-month sampling period and that cats
could be equally detected across my study site by ensuring consistency in camera set-up and

73
locations, meeting the first and second assumptions. To meet the third assumption, I took
measures to avoid counting cats more than once by spacing cameras 100m apart as was
shown to be effective for limiting capturing individual cats at more than one camera.
However, there were 38 cats that were identified at more than one site in my study. When this
happened within the same season, if possible, I shifted the encounter histories to have 14-day
periods where the cat was only present at one site, and therefore only counted in one
encounter history, meeting the third assumption. When this was impossible (N = 26) I flipped
a coin to decide which site would be used in the encounter history.
3.2.5 Model Selection and Seasonal Abundance Estimates
I followed the unmarked analysis workflow outlined in Kellner et al., (2023) to estimate the
abundance of wandering cats. Prior to analysis, I standardized all detection and site
covariates using a z-transformation (mean of 0 and standard deviation of 1). Currently, the
unmarked package is unable to handle multiple seasons across multiple sites, as well as
multiple visits within each season, and therefore I examined each season separately. I used
the Variance Inflation Model (VIF) to assess multicollinearity amongst covariates in the
global model (included all detection and abundance covariates) and used the chi-squared
goodness-of-fit tests for N-mixture models, using the Nmix.gof.test function from the
AICcmodavg package (Mazerolle, 2023) in Program R, V2.1.1. (R Core Team, 2021). All
variables had a VIF value between 1 and 4.4, suggesting little to no multicollinearity among
the covariates (Murray & Sandercock, 2020).
I created separate models for each season and used the Quasi Akaike Information
Criterion (QAIC; Burnham et al., 2011; Burnham & Anderson, 2002) to help account for
some overdispersions and the dredge function to determine the top detection and abundance

74
models (QAIC = 0) for each season separately. I first determined the top detection model by
creating a detection-only model with all three detection covariates (date, precipitation, and
temperature), holding abundance (state) at a constant of 1. For determining the top abundance
model, I created a full model that included habitat variables (water, field, vineyard, orchard,
natural, roads, and buildings) and carried forward the top detection covariate(s) from the
detection-only model. Finally, I used the seasonal local abundance model to estimate the
local abundance of wandering cats within each buffer using the predict function.
3.2.6 Seasonal Analyses
I used a Kruskal-Wallace test followed by a post-hoc Dunn’s test to examine the seasonal
variation of wandering cat abundance between seasons and land class types. The abundance
data were not normally distributed, favouring non-parametric methods to examine the
significance of median differences in seasonal abundance. I used a Dunn’s post-hoc test to
compare abundance across seasons (total number of wandering cats estimated within the
study area during each season), and local abundance across land class types (the number of
wandering cats within a specific land class during each season). I generated maps of cat local
abundance by first obtaining the habitat variables in 200m x 200m grid cells over the entire
study area, then estimated the local abundance of wandering cats within each grid cell using
the top seasonal abundance model(s) (without the detection covariate) and set the remaining
habitat variables in the model to their mean (additional mapping methods in Appendix 15).
3.2.7 Predicting Total Cat Abundance
The local abundance estimates for each 200 m x 200 m grid cell were combined to estimate
the total abundance of wandering cats in the south Okanagan Valley, B.C. I assumed that each

75
grid cell accurately estimated the local abundance of urban wandering cats because cats
typically have a home range of 100 m from their homes (Kays et al., 2020). Since water isn’t
usable habitat for cats, I removed any grid cell that had a proportion of water greater than
50%, which avoided overestimating the population size. Assessing each season separately, I
combined the grid cell local abundances for each map created (as described above) and
calculated the average total abundance across all habitat covariates being assessed for that
season. Therefore, the population of wandering cats is described as the seasonal estimate of
wandering cats within the boundaries of the south Okanagan Valley. While the total study
area was 161.2km2, after removing grid cells with more than 50% water, the remaining area
was 141.2km2, which I used to calculate density of wandering cats.
I verified and compared my estimated population size to the estimated number of
owned wandering cats based on surveys conducted in Canada and the total number of
dwellings in the south Okanagan Valley. When combining the number of dwellings from
Okanagan Falls, Oliver and Osoyoos, I get a minimum number of 7,002 and maximum
number of 11,587 dwellings within the towns and surrounding areas that surpass my study
area (Government of Canada, 2023). Across Canada, an average of 36.9% of households
have at least one domestic cat, with an average of 1.64 cats per household and 28% of these
residents allow their cats to roam freely outdoors (CFHS, 2017). Surveys conducted in
British Columbia found that 36% of households had a domestic cat with an average of 1.69
cats per household, and 45% of cat owners allowed their cat unsupervised access to the
outdoors (Stewardship Centre for British Columbia, 2019). I subtracted the estimated number
of owned cats, that was calculated based on survey data, from the estimated population of
wandering cats in the south Okanagan Valley (based on my models), to approximately
estimate the number of unowned cats.

76
3.3 Results
I identified a total of 374 different cats across the year, with a range of 0 – 12 uniquely
identified cats per site and an average of 3.5 cats per camera site, or 4.8 cats per camera site
where I observed at least one cat. Wandering cats were detected by trail cameras at a total of
92 of 123 sites (74.8%). I detected the highest number of unique wandering cats during
spring (N = 220), followed by fall (N = 186), summer (N = 172), early winter (N = 156), and
late winter (N=153; Table 3.2). The percentage of sites where cats were detected was highest
during the spring (61.7%), followed by the early winter (57.6%), fall (57.5%), summer
(54.2%), and late winter (50%; Table 3.2). Wandering cats were detected at 100% of periurban sites, 97% of urban sites, 65% of agricultural sites and 42% of natural sites (Table 3.2).
Based on local abundance estimates, wandering cats were most prevalent in urban and periurban habitats during each season, followed by agricultural and natural habitats (Table 3.2).
Of the 374 wandering cats detected, 38 were observed to have traveled between two or more
camera locations; 22 wandering cats were detected at two camera locations, 7 cats were
detected at three sites, and 9 cats were observed at four sites. This was most prevalent at
agricultural (N = 12) and peri-urban (N = 9) sites, compared to urban and natural (N = 3 sites
each). Many of these cats travelled far distances, with 23 cats travelling more than 200 m
between sites and 8 cats that travelled more than 400m, including two of which travelled 567
meters between two peri-urban sites.

77
Table 3.2. Seasonal summary statistics, including the number of sites where wandering cats
were detected, organized by land class, and the total number of individually identified
wandering cats identified using the trail camera during each season.
Early
Late
Land Class
Spring
Summer
Fall
Full Year
Winter
Winter
Urban

27/31
= 87.1%

27/31
= 87.1%

28/31
= 90.3%

26/31
= 83.9%

24/29
= 82.8%

31/32
= 97%

Peri-urban

19/26
= 73 %

19/26
= 73%

16/26
= 61.5%

21/25
= 84%

18/24
= 75%

27/27
= 100%

Agricultural

19/31
= 61.3%

13/31
= 41.9%

17/31
= 54.8%

17/30
= 56.6%

11/29
= 37.9%

20/31
= 65%

Natural

9/32
= 28.1%

6/32
= 18.75%

8/32
= 25%

4/32
= 12.5%

4/32
= 12.5%

14/33
= 42%

TOTAL

74/120
= 61.7%

65/120
= 54.2%

69/120
= 57.5%

68/118
= 57.6%

57/114
= 50%

220

172

186

156

153

Number of
Individually
Identified
Cats

92/123
= 74.8%

374

78
Table 3.3. Top models for detection probability (p) when assessing the number of domestic
cats photographed at trail camera sites during all five seasons in south Okanagan Valley,
B.C. Global model included the co-variates: temp, date and precip (precipitation).
Abundance estimates (λ) are held constant (λ (.)).
SPRING

SUMMER

FALL

EARLY
WINTER

LATE
WINTER

Model
p (date) λ (.)
p (.) λ (.)
p (temp+ date) λ (.)
p (precip + date) λ (.)
p (precip) λ (.)
p (temp + date + precip) λ (.)
p (temp) λ (.)
p (temp + precip) λ (.)
p (.) λ (.)
p (precip) λ (.)
p (temp) λ (.)
p (date) λ (.)
p (date + precip) λ (.)
p (temp + precip) λ (.)
p (temp+ date) λ (.)
p (temp + date + precip) λ (.)
p (temp + precip) λ (.)
p (temp + date + precip) λ (.)
p (date + precip) λ (.)
p (temp) λ (.)
p (date) λ (.)
p (temp+ date) λ (.)
p (precip) λ (.)
p (.) λ (.)
p (temp + date + precip) λ (.)
p (date + precip) λ (.)
p (temp + precip) λ (.)
p (temp+ date) λ (.)
p (temp) λ (.)
p (date) λ (.)
p (precip) λ (.)
p (.) λ (.)
p (date + precip) λ (.)
p (temp + date + precip) λ (.)
p (date) λ (.)
p (temp+ date) λ (.)
p (temp + precip) λ (.)
p (temp) λ (.)
p (precip) λ (.)
p (.) λ (.)

df
4
3
5
5
4
6
4
5
3
4
4
4
5
5
5
6
5
6
5
4
4
5
4
3
6
5
5
5
4
4
4
3
5
6
4
5
5
4
4
3

logLik
-937.002
-938.191
-936.253
-936.255
-937.358
-935.516
-937.958
-937.117
-771.359
-770.765
-771.323
-771.327
-770.740
-770.751
-771.301
-770.719
-898.048
-897.698
-898.846
-900.150
-900.687
-899.892
-901.697
-902.822
-737.016
-738.546
-738.728
-739.065
-740.906
-741.155
-745.000
-748.977
-707.597
-706.591
-709.703
-708.993
-711.461
-714.058
-714.449
-716.518

QAICc
1882.4
1882.6
1883.1
1883.1
1883.1
1883.9
1884.3
1884.8
1549.0
1550.0
1551.1
1551.1
1552.2
1552.2
1553.3
1554.4
1806.7
1808.3
1808.3
1808.7
1809.8
1810.4
1811.8
1811.9
1486.9
1487.7
1488.0
1488.7
1490.2
1490.7
1498.4
1504.2
1425.8
1426.0
1427.8
1428.6
1433.5
1436.5
1437.3
1439.3

ΔQAICc
0
0.22
0.70
0.70
0.71
1.47
1.91
2.43
0
0.99
2.11
2.12
3.17
3.20
4.29
5.41
0
1.56
1.60
1.99
3.07
3.69
5.09
5.17
0
0.82
1.18
1.86
3.34
3.84
11.53
17.33
0
0.22
2.02
2.79
7.73
10.73
11.51
13.50

weight
0.194
0.173
0.136
0.136
0.136
0.093
0.074
0.057
0.345
0.210
0.120
0.120
0.071
0.070
0.040
0.023
0.356
0.163
0.160
0.132
0.077
0.056
0.028
0.027
0.339
0.225
0.188
0.134
0.064
0.050
0.001
0.000
0.394
0.353
0.144
0.098
0.008
0.002
0.001
0.000

79
3.3.1 Detection probability
The global model for detection probability included average daily temperature (temp), Julian
date (date), and total daily precipitation (precip; Table 3.1). The Variance Inflation Model
(VIF) showed that none of the variables were collinear. During the spring, the top model
contained one covariate, date (Table 3.3), with detection probabilities positively related to
date. The mean detection probability during the spring was 0.2261 ± 0.0323 (lower CI:
0.1366, upper CI: 0.3594). During the summer, the top model contained no covariates (Table
3.3), and detection held constant at 0.1822 ± 0.0293 SE (lower CI: 0.1315, upper CI: 2468).
During the fall, the top model contained temperature and precipitation, where detection was
positively related to precipitation and negatively influenced by temperature. The mean
detection probability was 0.237 ± 0.0312 (lower CI: 0.117, upper CI: 0.527). During the early
winter, the top model contained all three co-variates with detection positively related to
temperature but negatively influenced by date and precipitation. The mean detection
probability during the early winter was 0.1045 ± 0.028 (lower CI: 0.0114, upper CI: 0.3611).
During the late winter, date and precipitation were the top variables for estimating
abundance, with date positively and precipitation negatively influencing detection, and a
mean detection probability was 0.1753 ± 0.0316 (lower CI: 0.0433, upper CI: 0.3794).
3.3.2 Local Cat Abundance
The global model for local cat abundance during each season included the proportion of
vineyards, orchards, water, fields, and natural habitat (desert, forest, and wetlands combined),
as well as the number of buildings and the length of roads. None of the covariates showed
multicollinearity. Natural habitat, such as wetlands, forests, and deserts, were included in the
top models for each season indicating cats were the least abundant in natural areas compared

80
to other habitats (Table 3.4). Vineyards were also included in four top models; cats were
rarely found in the middle of vineyards, away from humans and buildings.
The habitat variables influencing local wandering cat abundance differed between
seasons, with a more consistent set of predictors during spring, summer, and early winter
(natural habitats and vineyards), and additional habitat variables becoming important in fall
and late winter (fall: natural habitats, vineyards, water and roads; late winter: natural habitats,
fields, and buildings). The mean estimated abundance of wandering cats per site during the
spring was 2.21 ± 0.498 (lower CI: 0.006, upper CI: 8.21), increasing during the summer to
2.33 ± 0.606 (lower CI: 0.004, upper CI: 8.114), and increasing again during the early winter
to 3.25 ± 1.084 (lower CI: 0.003, upper CI: 17.159). The mean estimated number of
wandering cats per site during the fall was 1.87 ± 0.478 (lower CI: 0.022, upper CI: 10.992)
and during the late winter it was similar 1.86 ± 0.476 (lower CI: 0.019, upper CI: 10.34;
Figure 3.4b).
3.3.3 Seasonal and land class analyses
The median seasonal estimated abundances were very similar, with no significant differences
between seasons except the early and late winter (Figure 3.4). Urban sites, characterized by a
higher number of buildings within their buffer areas, consistently exhibited the highest
estimated local abundance of wandering cats across all seasons (For all five seasons:
Kruskall-Wallace, P < 0.0001; post hoc Dunn Test, urban versus all other land classes, P <
0.01; Figure 3.5), supporting the Human Presence and Cat Behaviour Hypothesis. Over the
12-month sampling period, urban habitats had the highest median estimated abundance of 4.1
wandering cats per site (Figure 3.4).

81
Table 3.4. Top models for abundance estimates (λ) when assessing the number of domestic
cats photographed at trail camera sites during all five seasons in south Okanagan Valley,
B.C. Global model included the co-variates: natural, vineyard, orchard, field, water, roads,
and buildings. Detection probability (p) is held constant (p(.)).
Model

df

logLik

QAICc

ΔQAICc

weight

SPRING

p (.) λ (natural + vineyard)

5

-912.063

1834.7

0

0.154

6

-911.685

1836.3

1.49

0.073

6

-911.732

1836.4

1.58

0.070

6

-911.738

1836.3

1.59

0.069

6

-911.889

1836.6

1.89

0.060

SUMMER

p (.) λ (natural + vineyard +
roads)
p (.) λ (natural + vineyard +
field)
p (.) λ (natural + vineyard +
orchard)
p (.) λ (natural + vineyard +
water)
p (.) λ (natural + vineyard)

5

-751.523

1513.7

0

0.105

p (.) λ (natural + vineyard +
buildings)
p (.) λ (natural + vineyard +
roads)
p (.) λ (natural + vineyard +
field)
p (.) λ (natural + vineyard +
orchard)
p (.) λ (natural + vineyard +
roads + water)
p (.) λ (natural + vineyard)

6

-750.768

1514.5

0.77

0.072

6

-750.826

1514.6

0.89

0.067

6

-750.850

1514.6

0.94

0.066

6

-751.189

1515.3

1.63

0.047

7

-881.102

1777.4

0

0.124

5

-883.700

1778.0

0.63

0.091

p (.) λ (natural + vineyard +
water)
p (.) λ (natural + vineyard +
roads)
p (.) λ (natural + vineyard
+ orchard + roads + water)
p (.) λ (natural + vineyard)

6

-882.859

1778.6

1.20

0.068

6

-882.990

1778.9

1.47

0.060

8

-880.721

1779.0

1.59

0.056

5

-726.030

1462.6

0

0.128

p (.) λ (natural + vineyard +
buildings)
p (.) λ (natural + vineyard +
field)
p (.) λ (natural + vineyard +
orchard)
p (.) λ (natural + vineyard +
orchard + field)
p (.) λ (natural + buildings +
field)
p (.) λ (natural + roads +
buildings + field)
p (.) λ (natural + vineyard +
roads)
p (.) λ (natural + buildings +
water + field)
p (.) λ (natural + orchard +
buildings + field)

6

-725.051

1462.9

0.28

0.111

6

-725.296

1463.4

0.77

0.087

6

-725.656

1464.1

1.49

0.061

7

-724.666

1464.4

1.80

0.052

6

-695.826

1404.4

0

0.129

7

-695.348

1405.8

1.32

0.067

6

-696.709

1406.2

1.77

0.053

7

-695.632

1406.3

1.89

0.050

7

-695.698

1406.5

2.02

0.047

FALL

EARLY
WINTER

LATE
WINTER

82
The median estimated local abundance of wandering cats across the year was
significantly higher at peri-urban sites compared to agricultural sites, while agricultural sites
had significantly higher median local abundance than natural sites (Figure 3.4). Peri-urban
habitats generally supported more wandering cats than agricultural habitats, while natural
habitats had lowest median estimated local abundance except during fall, where the local
abundance was not significantly different from agricultural habitats (For all seasons except
fall: Kruskall-Wallace, P < 0.0001; post hoc Dunn Test, natural versus all other land classes,
P < 0.01; Figure 3.5). Across seasons there were no significant differences in median
estimated local abundance between peri-urban and agricultural sites (Kruskall-Wallace, P <
0.0001; post hoc Dunn Test, peri-urban versus agricultural, P > 0.05; Figure 3.5).
3.3.4 Estimating wandering domestic cat population size
I estimated the population size of wandering cats for the spring, summer, and early winter,
abundance using predicted values from top models that contained the natural and vineyard
covariates. For fall, I used the natural, vineyard, water, and roads covariates, and for late
winter I used natural, buildings, and field covariates (All covariates mapped in Figure 3.6).
The estimated average population size of wandering cats was highest during the early winter
(8,964; average density of 63.47 wandering cats/km2; Table 3.5), followed by the spring
(7,048; average density of 46.29 wandering cats/km2; Table 3.5) and summer (7,031; average
density of 49.78 wandering cats/km2; Table 3.5). The population size was lowest during the
fall (5,245; average density of 37.14 wandering cats/km2; Table 3.5) and late winter (4,496;
average density of 31.84 wandering cats/km2; Table 3.5). On average, the population across
the year and study area was 6,557 (range: 4,496 – 8,964) wandering cats, with an average
density of 46.44 (range: 37.14 – 63.47) wandering cats/km2. Seasonal predicted cat

83
abundance based on natural habitat showed low cat abundance in natural areas (Figure 3.7),
which was similar to the seasonal maps based on vineyards (Figure 3.8), where cats were not
associated with vineyards. All other habitat covariate maps for the fall and late winter can be
found in Appendix 16 (Figure A16.1 for the fall and Figure A16.2 for the late winter).
I used provincial and national cat owner surveys to estimate the number of owned and
unowned wandering cats in the south Okanagan Valley. Based on the provincial surveys and
the minimum number of dwellings, I estimated 1,917 owned wandering cats, with the
number increasing to 3,172 wandering cats with the maximum number of dwellings. Using
national owner surveys, the number of owned wandering cats was estimated to be 1,186
using the minimum number of dwellings, and 1,963 wandering cats with the maximum
number of dwellings. Using the average population size of 6,557 wandering cats estimated in
this study and subtracting the estimated owned cats from owner surveys, I estimated the
number of unowned cats to range from 4,485 to 4,640 based on provincial surveys, and from
4,594 to 5,371 based on national surveys, indicating the 68-83% of wandering cats were
unowned.

84

Figure 3.4. Box plots showing the estimated abundance A) in different land classes
throughout the year and B) during each season. All land classes were significantly different
(p < 0.05 for all seasonal comparisons). Median abundance of cats per site was highest in
urban habitats (4.1), followed by peri-urban habitats (1.8), agricultural habitats (1.2), and
natural habitats (0.4). For seasonal comparisons, all land classes were not significantly
different (p > 0.05), except between early and late winter, which were significantly different
(p < 0.05). The median abundance of domestic cats per site was highest in the early winter
(2.1) and lowest in the late winter (1.3). The horizontal black line within the box represents
the median value, and the box represents 50% of the data, from the 25th to 75th percentile of
each groups’ distribution of values, which is the interquartile range (IQR). The whiskers
extend to the highest and lowest points within 1.5 times the IQR and individual dots are
outliers beyond the whiskers.

85

Figure 3.5. Box plots showing the estimated abundance for each land class separated by
season. Mean abundance was significantly highest in urban habitats (UR) consistently across
each season (p < 0.05). Mean abundance in peri-urban (PU) and agricultural (AG) sites are
not significantly different in any season (p > 0.05). Mean abundance is significantly lowest in
natural (NA) habitats during each season (p < 0.05) except during the fall, where agricultural
and natural sites do not significantly differ (p > 0.05). The horizontal black line within the
box represents the median value, and the box represents 50% of the data, from the 25th to 75th
percentile of each groups’ distribution of values, which is the interquartile range (IQR). The
whiskers extend to the highest and lowest points within 1.5 times the IQR and individual dots
are outliers beyond the whiskers.

86
Table 3.5. Variables for each season used for predicting cat abundance. Natural is wetlands,
desert, and forests combined. Density was calculated using the study area of 141.2km2.
Model
Season
Estimated Abundance
Density (cats per km2)
Co-variate
Spring
Natural
6537 (range: 4264 to 10,098)
46.29 (range 30.19 to 71.50)
Vineyard
7559 (range: 4754 to 12,297)
53.52 (range 33.66 to 87.07)
Summer
Natural
6231 (range: 3901 to 10,069)
44.11 (range: 27.62 to 71.29)
Vineyard
7832 (range: 4473 to 14,419)
55.45 (range: 31.67 to 102.09)
Fall
Natural
5331 (range: 3633 to 7941)
37.74 (range: 25.72 to 56.22)
Vineyard
6423 (range: 4197 to 10,096)
45.49 (range: 29.72 to 71.50)
Water
7328 (range: 4912 to 10,959)
51.91 (range: 34.79 to 77.64)
Roads
1894 (range: 451 to 11,013)
13.41 (range: 3.19 to 77.97)
Early
Natural
7914 (range: 4208 to 14,964)
56.03 (range: 29.79 to 105.95)
Winter
Vineyard
10,015 (range: 4538 to 22,693) 70.91 (range: 32.13 to 160.67)
Late
Natural
4744 (range: 2898 to 7854)
33.59 (range: 20.52 to 55.61)
Winter
Buildings
2401 (range: 1467 to 3934)
17.00 (range: 10.38 to 27.85)
Field
6345 (range: 3123 to 21,129)
44.93 (range: 22.11 to 149.60)

87

Figure 3.6. Habitat variables used in top models to estimate wandering cat populations in the
south Okanagan Valley, B.C. Maps show each grid square and the proportion of natural
habitats (A), vineyards (B), roads (C), buildings (D), fields (E) and water (F). Darker colours
represent higher values, while lighter colours represent lower values.

88

Figure 3.7. Predicted abundance of wandering cats based on the proportion of natural habitat
in each cell during each season: A) spring, B) summer, C) Fall, D) Early winter, and E) Late
winter. Darker areas have higher densities of wandering cats. While maps look the same, the
local abundance per grid cell differs across seasons.

89

Figure 3.8. Predicted abundance of wandering cats based on the proportion of vineyards in
each cell during four seasons: A) spring, B) summer, C) Fall, and D) Early winter. Darker
areas have higher densities of wandering cats.

90
3.4 Discussion
In this multi-season and -habitat focused study, I demonstrated that wandering cats are most
abundant in urban habitats but also occur widely, and often in high local abundances, across
land classes. This suggests that regions like the south Okanagan, which feature a mosaic of
urban, agricultural, peri-urban and natural areas, harbour a proportionately large number of
cats in relation to the area and human population. Based on my estimated cat population size,
the south Okanagan Valley had at least 1 cat for every 3 people, which represents the highest
per capita density of wandering cats recorded in a North American based study thus far (Cove
et al., 2023; Flockhart et al., 2016; Gow et al., 2024). With an average predicted density of
46.44 cats/km2 across all habitat types, the density in the south Okanagan Valley is higher
than the only other multi-habitat assessment (i.e., Australia, 0.13-0.73 cats/km2; Legge et al.,
2017), and at least one urban area in North America (41.45 cats/km2, 1 cat for every 93
people, Washington, DC, USA, Cove et al., 2023). However, the density is lower than two
Canadian cities but higher based on wandering cats/person; Guelph, ON (88.36 cats/km2, one
cat for every 16 people, Flockhart et al., 2016) and Gatineau, PQ (62.24 cats/km2, one cat for
every 14 people, Gow et al., 2024). Cat population estimates were highest during the early
winter, spring, and summer, with fewer cats in fall and late winter. Population estimates of
wandering cats using trail cameras were nearly double those extrapolated from owner
surveys, meaning that the vast majority of wandering cats are likely unowned or households
were more likely to put their cats outdoors than national or provisional averages.
While my estimates provide valuable insights into the seasonal abundance of
wandering cats, there is inherent uncertainty in my results. My abundance models account for
factors influencing both cat abundance and detection, but other unmeasured variables may
contribute to fluctuations that I cannot fully capture. These limitations may lead to under- or

91
overestimation of wandering cat abundance across seasons. These uncertainties apply to both
owned and unowned wandering cat estimates; however, my results remain valuable in
identifying the seasons and habitats where wandering cat abundance may be highest and have
the greatest impact on wildlife. Despite limitations, my findings establish an important
foundation for understanding the seasonal patterns in wandering cat abundance and highlight
the need for further research to refine estimates and improve confidence in population
assessments. Wandering cats were most abundant in urban areas, while natural habitats
supported the lowest, but not zero, abundance, reflecting the strong influence of human
presence and habitat type on their local abundance, and supporting the Human Presence and
Cat Behaviour Hypothesis. This association between cats and urban areas aligns with other
studies in Ontario (Clyde et al., 2022) and Washington, DC (Cove et al., 2023), further
highlighting the influence of human presence in shaping wandering cat abundances across
different landscapes. While other studies have focused on binary classifications of habitat
types, such as rural vs urban (Kauhala et al., 2015), this approach often overlooks the
nuanced land-use patterns in wandering cat abundance. In peri-urban and agricultural
habitats, human influence also shaped the distribution of wandering cats, with lower
wandering cat abundances likely reflecting the reduced number of buildings and residents in
these areas. As predicted, natural habitats supported the lowest abundance of wandering cats,
reflecting the low levels of urbanization, greater distance from human settlements, and
limited human influence. However, the high overall abundance of wandering cats in the south
Okanagan Valley may be attributed to the unique integration of peri-urban and urban areas,
where urbanization intensity differs from that of a larger metropolitan center. The distinct
landscape, combined with the region’s temperate climate and abundant resources, may create
more favourable conditions for wandering cats. However, research on wandering cats in

92
small urban centers is limited, making it difficult to determine the mechanisms driving these
patterns.
While there were fewer wandering cats in natural habitats, they were still present and
can pose a threat to local wildlife, particularly birds, by exerting predation pressure,
especially if unowned or partially owned (e.g., barn cats). A study tracking cat movement
across four countries (USA, Australia, UK, and New Zealand) found that 11% of tracked cats
spent most of their outdoor time in natural habitats (Kays et al., 2020), while another study in
Guelph, ON, Canada, showed that 4% of GPS locations occurred in natural habitats or
greenspaces (Pyott et al., 2024). While I did not use tracking collars, these findings may
explain the use of natural habitats by wandering cats observed in my study. Cats that wander
into natural areas from nearby houses raise concerns for local wildlife, as studies in Great
Britain found that cats living adjacent to natural habitats killed more prey than inner
suburban cats, with an average of 7.91 prey/cat/year, 25% of which were birds (Pirie et al.,
2022). While predation pressure on birds and other animals may be highest in urban areas
due to greater cat densities (Legge et al., 2020), the impact of a single cat in natural habitats
is concerning, particularly due to the higher numbers of species at risk, including numerous
bird species found in the south Okanagan Valley (Chapter 2). Further research that tracks cat
movement and estimates predation rates of owned and unowned cats across different habitat
types in this region would provide more precise insights into the magnitude of their impact
on bird (and other animal) populations in this region.
Understanding how seasonal weather patterns influence the activity level and
detection of wandering cats is critical for accurately assessing changes in local abundances
and population sizes throughout the year. During fall and winter, the detection of wandering
cats was largely influenced by changes in weather, such as temperature and rain. During fall,

93
I saw a decrease in the abundance and detection of wandering cats, which may have been
influenced by the dry, hot weather. Similar dry hot weather has been associated with reduced
wandering cat density in Australia due to lower prey availability, with this pattern driven
predominantly across the dry regions of the country (Legge et al., 2017). Seasonal patterns of
feral cat abundance in Australia may be driven by changes in prey availability (i.e., more
prey during wet periods), thus it is possible that the seasonal variability that I observed in
local wandering cat abundance and population estimates may similarly be explained by prey
availability (i.e., Prey Availability Hypothesis), especially if the high estimated proportion of
unowned cats in my study area relies heavily on hunting animals to survive (see Chapter 2
for details about birds in the region). In addition, weather conditions can influence human
behaviour, such as reducing the likelihood of residents allowing their cats outdoors during
hot or cold temperatures (or rain or snow), thus lowering the abundance of wandering cats
(Forrest et al., 2023; Goszczyński et al., 2009). My study region is a popular tourist
destination, and the departure of vacationers in fall and the potential decrease in activity of
unowned cats due to hot, dry conditions could have further contributed to the observed
decline in wandering cat abundance estimates during the late winter, showing some support
for the Human Behaviour Hypothesis. While I was unable to examine the mechanism
explaining the seasonal abundance differences directly, it is possible that, given the large
estimated number and local abundance of unowned cats, ecological factors (e.g., prey
availability) and human factors (e.g., supplemental feeding, lack of spay-neuter programs, pet
dumping, etc.) may be more likely to drive wandering cat populations in this region, in
contrast to potential dynamics in large urban centres that may be influenced by a dynamic
between shelters, owned cats, and unowned cats (Flockhart et al., 2024; Flockhart & Coe,
2018).

94
Seasonal weather patterns, particularly snowfall, may influence the activity levels of
wandering cats, leading to the highest local population and local abundance estimates during
the early winter and decreasing during the late winter. Early winter local abundance and
population estimates were higher than those of all other seasons, possibly due to increased cat
activity levels as they searched for food before snow accumulation, and residents potentially
granting their owned cats more outdoor access as temperatures cooled and precipitation
increased following the fall. Additionally, during early winter, many birds moved into urban
areas where food is more readily available (Chapter 2). As cats are primarily found in urban
areas, there may be greater potential for hunting birds during this period of overlap between
high wandering cat abundance and high bird abundance, thereby increasing cat activity levels
and detection. Increased activity of wandering cats during the winter also aligns with findings
from GPS-tracked cats in Belfast, Ireland, which showed larger home ranges in winter
compared to summer (Dunford et al., 2024). While cats may have been less active when
snow accumulated, many were still outdoors, as evidenced by numerous observations of cats
walking through snow. Overall, early winter may provide the most realistic estimate of the
wandering cat population, as owned cats are likely outdoors, and unowned cats are more
detectable as they may be actively searching for food. During late winter, I observed a
decrease in the wandering cat abundance, potentially due to cats not surviving the winter and
higher rainfall, which has been shown to decrease cat activity level (Goszczyński et al.,
2009). During this period, I also saw a decrease in the number of prey species in the south
Okanagan Valley (Chapter 2), which may have contributed to decreased activity levels, such
as reduced exploratory and hunting behaviours. My findings suggest that wandering cats
have the potential to pose a significant year-round threat to birds and other small animals,
relative to seasonal prey availability (e.g. bats, insects, reptiles, amphibians, mammals etc.).

95
3.4.1 Management implications
Unfortunately, the high number of unowned wandering cats in the study region creates
wandering cat population management challenges and thus subsequent economic investments
for municipalities and other governments if they chose to take action. Without significant
multi-stakeholder actions and funding, welfare concerns (e.g., cats being hit by cars,
predation of cats, disease transmission, etc.), risks cats pose to wildlife, and subsequent calls
for initiatives to manage cat populations are likely to escalate and magnify. For example,
media reports indicate that the Okanagan Human Society has seen a substantial increase in
low-income spay neuter requests, which are estimated to cost up to $50,000 per month
(Black Press Media, 2024). This places significant strain on shelters and puts an economic
burden on local communities tasked with managing these unowned cats. While this
recognizes the growing public interest in controlling wandering cat populations, the
associated costs often prohibit effective action. Additionally, the economic burden of catdependent diseases may extend beyond the local level, with human healthcare costs in
Australia potentially reaching billions of dollars due to medical costs and lost productivity
(Legge et al., 2020). Effectively managing the high number of unowned wandering cats and
their ecological, welfare, and economic impacts requires a unified approach with
considerable collaboration and education across multiple sectors, including local
municipalities, animal welfare organizations, and the public.
In conclusion, my study highlights the widespread presence of wandering cats across
diverse land classes, raising significant concerns for both cat welfare and wildlife
conservation locally and more broadly. Here, I showed that cats were not confined to urban
areas, and their activity across the landscape would require landscape-wide management
strategies to mitigate their potential negative impacts on wildlife. By quantifying seasonal

96
variation in abundance and habitat use, I identified periods of increased cat activity and
potential greater risk to wildlife, such as birds that are also heavily driven by seasonal
patterns. By also identifying seasonal drivers of wandering cat detections, such as
temperature and precipitation, I was able to highlight periods of increased wandering cat
abundance, which could help in the development of strategies that account for these seasonal
dynamics. Effective solutions require a multifaceted approach that balances the welfare of
cats with the protection of vulnerable wildlife populations. Such approaches could involve a
combination of educational campaigns to reduce outdoor access for owned cats and the
implementation of new regulations by local governments, such as licensing or requiring spayneutering (Cecchetti et al., 2021). Collaborative efforts among municipalities, stakeholders,
and residents will be critical to mitigate the ecological impact of wandering cats and help
preserve the biodiversity of this unique region. Given the ecological sensitivity of the area,
public outreach and strong enforcement measures are needed, which ultimately requires
municipal government to legislate and enforce by-laws that curtail wandering cats and
address the environmental and economic impacts.

97
4.5 Literature Cited
Alexander, C. A. D., Alameddine, I., Machin, D., & Alex, K. (2024). A Weight-of-Evidence
Approach for Understanding the Recovery of Okanagan Sockeye Salmon. Environmental
Management, 74(6), 1063–1085. https://doi.org/10.1007/s00267-024-02031-y
Alldritt-McDowell, J., Coupé, R., & Meidinger, M. (1998, July). The Ecology of the
Bunchgrass Zone. Ministry of Forests, British Columbia.
https://www.for.gov.bc.ca/hfd/pubs/docs/Bro/bro54.pdf
Bennett, K. F. P., Evans, B. S., Clark, J. A., & Marra, P. P. (2021). Domestic Cat Abundance
and Activity Across a Residential Land Use Gradient. Frontiers in Ecology and Evolution,
9. https://doi.org/10.3389/fevo.2021.643845
Black Press Media. (2024, July 17). Humane society overrun as Okanagan SPCA forced to
limit animal intakes [News]. The Chilliwack Progress.
https://blackpress.tv/embed/55624/blackpress.tv/embed/55624/Humane_society_overrun_
as_Okanagan_SPCA_forced_to_limit_animal_intakes
Bradshaw, J. W. S. (2006). The evolutionary basis for the feeding behavior of domestic dogs
(Canis familiaris) and cats (Felis catus). The Journal of Nutrition, 136(7 Suppl).
https://doi.org/10.1093/jn/136.7.1927S
Burnham, K. P., & Anderson, D. R. (2002). Model Selection and Inference: A Practical
Information-Theoretic Approach.
Burnham, K. P., Anderson, D. R., & Huyvaert, K. P. (2011). AIC model selection and
multimodel inference in behavioral ecology: Some background, observations, and
comparisons. Behavioral Ecology and Sociobiology, 65(1), 23–35.
https://doi.org/10.1007/s00265-010-1029-6
Canadian Federation of Humane Societies [CFHS]. (2017). Cats in Canada 2017: A five-year
review of cat overpopulation. https://www.g2z.org.au/pdf/CFHS-Cats_In_Canada_2017FINAL_LR.pdf
Cecchetti, M., Crowley, S. L., & McDonald, R. A. (2021). Drivers and facilitators of hunting
behaviour in domestic cats and options for management. Mammal Review, 51(3), 307–
322. https://doi.org/10.1111/mam.12230

98
Clancy, E. A., Moore, A. S., & Bertone, E. R. (2003). Evaluation of cat and owner
characteristics and their relationships to outdoor access of owned cats.
https://doi.org/10.2460/javma.2003.222.1541
Clyde, H. E., Norris, D. R., Lupton, E., & Gow, E. A. (2022). Estimating habitat
characteristics associated with the abundance of free-roaming domestic cats across the
annual cycle. Wildlife Research, 49(7), 583–595. https://doi.org/10.1071/WR20205
Coleman, J. S., & Temple, S. A. (1993). Rural Residents’ Free-Ranging Domestic Cats: A
Survey. Wildlife Society Bulletin (1973-2006), 21(4), 381–390.
Coristine, L. E., Jacob, A. L., Schuster, R., Otto, S. P., Baron, N. E., Bennett, N. J., Bittick, S.
J., Dey, C., Favaro, B., Ford, A., Nowlan, L., Orihel, D., Palen, W. J., Polfus, J. L.,
Shiffman, D. S., Venter, O., & Woodley, S. (2018). Informing Canada’s commitment to
biodiversity conservation: A science-based framework to help guide protected areas
designation through Target 1 and beyond. FACETS, 3(1), 531–562.
https://doi.org/10.1139/facets-2017-0102
Correia, M., Alexis, S., & Dulic, A. (2024). Bringing the salmon home: A study of crosscultural collaboration in the Syilx Okanagan Territory of British Columbia. Ecology and
Society, 29(1). https://doi.org/10.5751/ES-14831-290115
Cove, M. V., Herrmann, V., Herrera, D. J., Augustine, B. C., Flockhart, D. T. T., & McShea,
W. J. (2023). Counting the Capital’s cats: Estimating drivers of abundance of free-roaming
cats with a novel hierarchical model. Ecological Applications, 33(2), e2790.
https://doi.org/10.1002/eap.2790
Dozier, M. E., Porter, B., Shivley, J. M., & Telle, M. R. (2023). Increased animal ownership
in the home is correlated with worse health outcomes based on health-care indicators
investigated in canines and felines in rural Mississippi.
https://doi.org/10.2460/javma.22.12.0589
Dunford, C. E., Loca, S., Marks, N. J., & Scantlebury, M. (2024). Seasonal habitat selection
and ranging of domestic cats (Felis catus) in rural and urban environments. Animal
Biotelemetry, 12(1), 13. https://doi.org/10.1186/s40317-024-00367-0
Elizondo, E., & Loss, S. (2016). Using trail cameras to estimate free-ranging domestic cat
abundance in urban areas. Wildlife Biology, 22, 246–252.
https://doi.org/10.2981/wlb.00237

99
Environment and Climate Change Canada. (2012, May 24). Burrowing Owl (Athene
cunicularia): Recovery strategy 2012 [Program results;research].
https://www.canada.ca/en/environment-climate-change/services/species-risk-publicregistry/recovery-strategies/burrowing-owl-2012.html
Environment and Climate Change Canada (ECCC). (2022). Osoyoos CS, British Columbia,
Weather Station.
https://climate.weather.gc.ca/climate_data/daily_data_e.html?hlyRange=1994-0201%7C2024-09-09&dlyRange=1990-08-01%7C2024-09-09&mlyRange=1990-0101%7C2006-1101&StationID=1041&Prov=BC&urlExtension=_e.html&searchType=stnName&optLimit
=yearRange&StartYear=1840&EndYear=2024&selRowPerPage=25&Line=1&searchMet
hod=contains&txtStationName=Osoyoos&timeframe=2&Day=9&Year=2022&Month=9#
Esri. (2023, February 23). Esri. “World Imagery” [Basemap]. Scale Not Given. “World
Imagery (Wayback 2023-02-23)”.
https://www.arcgis.com/home/item.html?id=fb0f4638a6894bddb3292499a0aeb7b3
Fiske, I., & Chandler, R. (2011). unmarked: An R Package for Fitting Hierarchical Models of
Wildlife Occurrence and Abundance. Journal of Statistical Software, 43, 1–23.
https://doi.org/10.18637/jss.v043.i10
Flockhart, D. T. T., & Coe, J. B. (2018). Multistate matrix population model to assess the
contributions and impacts on population abundance of domestic cats in urban areas
including owned cats, unowned cats, and cats in shelters. PLOS ONE, 13(2), e0192139.
https://doi.org/10.1371/journal.pone.0192139
Flockhart, D. T. T., Moore, S. M., Decker, S. T., & Kilgour, R. J. (2024). Geographic
Transitions of Domestic Cats in Urban Areas through Animal Adoption Centers and the
Implications for Population Dynamics. Journal of Applied Animal Welfare Science:
JAAWS, 1–19. https://doi.org/10.1080/10888705.2024.2438645
Flockhart, D. T. T., Norris, D. R., & Coe, J. B. (2016). Predicting free‐roaming cat population
densities in urban areas. Animal Conservation, 19(5), 472–483.
https://doi.org/10.1111/acv.12264
Forrest, R., Awawdeh, L., Pearson, M., & Waran, N. (2023). Pet Ownership in Aotearoa New
Zealand: A National Survey of Cat and Dog Owner Practices. Animals, 13(4), Article 4.
https://doi.org/10.3390/ani13040631

100
Forrester, T. R., Green, D. J., McKibbin, R., Morgan, T. C., Bezener, A. M., & Bishop, C. A.
(2023). Long-term decline in Brown-headed Cowbird (Molothrus ater) abundance has not
led to less brood parasitism of four riparian songbird species. The Wilson Journal of
Ornithology, 135(2), 154–171. https://doi.org/10.1676/22-00074
Goszczyński, J., Krauze, D., & Gryz, J. (2009). Activity and exploration range of house cats
in rural areas of central Poland. Folia Zoologica, 58(4), 363–371.
Government of Canada, S. C. (2023). Profile table, Census Profile, 2021 Census of
Population. https://www12.statcan.gc.ca/census-recensement/2021/dppd/prof/index.cfm?Lang=E
Gow, E. A., Aubrey, B., Cossar, L., Mazerolle, C., & Cheskey, E. (2024). Conducting
community-led research using trail cameras to develop baseline wandering domestic cat
local abundance estimates. FACETS, 9, 1–14. https://doi.org/10.1139/facets-2023-0033
Hall, C. M., Bryant, K. A., Haskard, K., Major, T., Bruce, S., & Calver, M. C. (2016). Factors
determining the home ranges of pet cats: A meta-analysis. Biological Conservation, 203,
313–320. https://doi.org/10.1016/j.biocon.2016.09.029
Hand, A. (2019). Estimating feral cat densities using distance sampling in an urban
environment. Ecology and Evolution, 9(5), 2699–2705. https://doi.org/10.1002/ece3.4938
Hanmer, H. J., Thomas, R. L., & Fellowes, M. D. E. (2017). Urbanisation influences range
size of the domestic cat (Felis catus): Consequences for conservation. Journal of Urban
Ecology, 3(1), jux014. https://doi.org/10.1093/jue/jux014
Herrera, D. J., Moore, S. M., Flockhart, D. T. T., McShea, W. J., & Cove, M. V. (2021).
Thinking outside the park: Recommendations for camera trapping mammal communities
in the urban matrix. Journal of Urban Ecology, 7(1), juaa036.
https://doi.org/10.1093/jue/juaa036
Kauhala, K., Talvitie, K., & Vuorisalo, T. (2015). Free-ranging house cats in urban and rural
areas in the north: Useful rodent killers or harmful bird predators? Folia Zoologica, 64(1),
45–55. https://doi.org/10.25225/fozo.v64.i1.a6.2015
Kays, R., Dunn, R. R., Parsons, A. W., Mcdonald, B., Perkins, T., Powers, S. A., Shell, L.,
McDonald, J. L., Cole, H., Kikillus, H., Woods, L., Tindle, H., & Roetman, P. (2020). The
small home ranges and large local ecological impacts of pet cats. Animal Conservation,
23(5), 516–523. https://doi.org/10.1111/acv.12563

101

Kays, R. W., & DeWan, A. A. (2004). Ecological impact of inside/outside house cats around
a suburban nature preserve. Animal Conservation, 7(3), 273–283.
https://doi.org/10.1017/S1367943004001489
Kellner, K. F., Smith, A. D., Royle, J. A., Kéry, M., Belant, J. L., & Chandler, R. B. (2023).
The unmarked R package: Twelve years of advances in occurrence and abundance
modelling in ecology. Methods in Ecology and Evolution, 14(6), 1408–1415.
https://doi.org/10.1111/2041-210X.14123
Kéry, M., & Royle, J. A. (Eds.). (2016). Applied Hierarchical Modeling in Ecology.
Academic Press. https://doi.org/10.1016/B978-0-12-801378-6.04001-7
Klenner, W., & Sullivan, T. P. (2025). Effect of type and extent of vegetative cover on
terrestrial small mammals in the semi-arid Okanagan Valley, British Columbia. Journal of
Arid Environments, 227, 105325. https://doi.org/10.1016/j.jaridenv.2025.105325
Legge, S., Murphy, B. P., McGregor, H., Woinarski, J. C. Z., Augusteyn, J., Ballard, G.,
Baseler, M., Buckmaster, T., Dickman, C. R., Doherty, T., Edwards, G., Eyre, T., Fancourt,
B. A., Ferguson, D., Forsyth, D. M., Geary, W. L., Gentle, M., Gillespie, G., Greenwood,
L., … Zewe, F. (2017). Enumerating a continental-scale threat: How many feral cats are in
Australia? Biological Conservation, 206, 293–303.
https://doi.org/10.1016/j.biocon.2016.11.032
Legge, S., Taggart, P. L., Dickman, C. R., Read, J. L., & Woinarski, J. C. Z. (2020). Catdependent diseases cost Australia AU$6 billion per year through impacts on human health
and livestock production. Wildlife Research, 47(8), 731–746.
https://doi.org/10.1071/WR20089
Legge, S., Woinarski, J. C. Z., Dickman, C. R., Murphy, B. P., Woolley, L.-A., & Calver, M.
C. (2020). We need to worry about Bella and Charlie: The impacts of pet cats on
Australian wildlife. Wildlife Research, 47(8), 523–539. https://doi.org/10.1071/WR19174
Loss, S. R., Will, T., & Marra, P. P. (2013). The impact of free-ranging domestic cats on
wildlife of the United States. Nature Communications, 4(1), 1396.
https://doi.org/10.1038/ncomms2380
Mazerolle, M. J. (2023). AICcmodavg: Model Selection and Multimodel Inference Based on
(Q)AIC(c) (Version 2.3-3) [Computer software]. https://cran.rproject.org/web/packages/AICcmodavg/index.html

102

McKibbin, R., & Bishop, C. A. (2008). Capture Success Rates of the Western Yellowbreasted Chat in the South Okanagan Valley, British Columbia, from 2005 through 2007.
Digital Commons: University of South Florida, 33(4).
https://digitalcommons.usf.edu/cgi/viewcontent.cgi?article=2801&context=nabb
McKinney, M. L. (2002). Urbanization, Biodiversity, and Conservation: The impacts of
urbanization on native species are poorly studied, but educating a highly urbanized human
population about these impacts can greatly improve species conservation in all
ecosystems. BioScience, 52(10), 883–890. https://doi.org/10.1641/00063568(2002)052[0883:UBAC]2.0.CO;2
Murray, D. L., & Sandercock, B. K. (2020). Population Ecology in Practice. John Wiley &
Sons.
Pirie, T. J., Thomas, R. L., & Fellowes, M. D. E. (2022). Pet cats (Felis catus) from urban
boundaries use different habitats, have larger home ranges and kill more prey than cats
from the suburbs. Landscape and Urban Planning, 220, 104338.
https://doi.org/10.1016/j.landurbplan.2021.104338
Pyott, M. L., Norris, D. R., Mitchell, G. W., Custode, L., & Gow, E. A. (2024). Home range
size and habitat selection of owned outdoor domestic cats (Felis catus) in urban
southwestern Ontario. PeerJ, 12, e17159. https://doi.org/10.7717/peerj.17159
QGIS.org. (2024). QGIS Geographic Information System. QGIS Association.
http://www.qgis.org
R Core Team. (2021). R: A language and environment for statistical computing. [R
Foundation for Statistical Computing]. https://www.R-project.org/
Saunders, S. P., Wu, J. X., Gow, E. A., Adams, E., Bateman, B. L., Bayard, T., Beilke, S.,
Dayer, A. A., Fournier, A. M. V., Fox, K., Heglund, P., Lerman, S. B., Michel, N. L.,
Paxton, E. H., Şekercioğlu, Ç. H., Smith, M. A., Thogmartin, W., Woodrey, M. S., & van
Riper, C., III. (2021). Bridging the research-implementation gap in avian conservation
with translational ecology. Ornithological Applications, 123(3), duab018.
https://doi.org/10.1093/ornithapp/duab018
Stewardship Centre for British Columbia. (2019). Reducing the impacts of cats on birds and
wildlife, 2017-2019.

103
https://stewardshipcentrebc.ca/PDF_docs/CatsBirds/Cat_BirdSurveyResults20172019.pdf
Swanwick, C., Dunnett, N., & Woolley, H. (2003). Nature, Role and Value of Green Space in
Towns and Cities: An Overview. Built Environment (1978-), 29(2), 94–106.
Tan, S. M. L., Stellato, A. C., & Niel, L. (2020). Uncontrolled Outdoor Access for Cats: An
Assessment of Risks and Benefits. Animals, 10(2), 258.
https://doi.org/10.3390/ani10020258
Trouwborst, A., McCormack, P. C., & Martínez Camacho, E. (2020). Domestic cats and their
impacts on biodiversity: A blind spot in the application of nature conservation law. People
and Nature, 2(1), 235–250. https://doi.org/10.1002/pan3.10073
Wald, D. M., & Peterson, A. L. (2020). Cats and Conservationists: The Debate Over Who
Owns the Outdoors.

104
Chapter 4: Conclusions, Future Directions, and Management and Conservation
Implications
4.1 Discussion and Future Directions
The high numbers of wandering cats in the south Okanagan Valley throughout the year,
combined with the diversity of birds and other wildlife, suggests that cats could have
significant impacts on birds and other wildlife year-round. Wandering cats pose a significant
threat to bird populations globally, contributing to the extinction of 22 species on islands
(Medina et al., 2011), and killing billions (if not trillions) of birds worldwide each year
(Blancher, 2013; Krauze-Gryz et al., 2019; Loss et al., 2013; Sedano-Cruz, 2022; Woinarski
et al., 2017; Woods et al., 2003). Cats may also affect mainland bird populations (Loss et al.,
2013). The impacts of wandering cats on wildlife are not evenly distributed, creating distinct
local ecological risks. The south Okanagan Valley acts as a funnel for migratory birds in
central B.C., which was evident from my findings (Chapter 2), where I identified 146 bird
species during point counts, though over 330 bird species inhabit this area across the year,
including 24 provincially listed species at risk and 22 federally listed species at risk under the
Species at Risk Act (eBird, 2022; Okanagan Similkameen Stewardship [OSS], 2023). Many
of these species at risk are found in natural habitats or border peri-urban or agricultural
habitats (e.g., Yellow-breasted Chat [Icteria virens auricollis], Lewis’s Woodpecker
[Melanerpes lewis], and Bobolink [Dolichonyx oryzivorus]), but are still at risk of predation
by cats that wander into natural areas. There are other habitats and seasons where the
abundance patterns of cats and birds overlap considerably, such as during the non-breeding
seasons when there is a high abundance and species richness of birds in urban areas along
with extremely high local abundances of wandering cats. I demonstrated the importance of
natural areas, especially during breeding for species of birds that are uncommon or were not

105
identified in other habitats, including numerous species at risk that may face predation by
domestic cats, such as Yellow-breasted Chat (Icteria virens auricollis), Bank Swallow
(Riparia riparia), and Bobolink (Dolichonyx oryzivorus). The fact that cats were observed in
these natural areas is concerning, especially given the potential impact, either directly or
indirectly, of a single cat on birds.
Although about half of owned cats kill wild animals (Loyd et al., 2013), up to 82%
percent of wandering cats in my study area are potentially unowned, with potentially up to
91% preying on wildlife (McGregor et al., 2015), collectively intensifying predation pressure
on migratory and non-migratory birds in the study region, as well as other animals including
native mammals, amphibians, reptiles and insects. This is exacerbated by the
disproportionately high abundance of wandering cats in urban and peri-urban areas in
relation to people, their presence in natural areas, and the presumably larger home ranges of
rural cats (Hanmer et al., 2017; Pirie et al., 2022). Although the local abundance estimates in
natural habitats were low, which may have also been attributed to the presence of predators,
cats were still detected at 42% of natural sites during the year. At six natural sites, cats
appeared more than 100 m from the nearest building, including one observation in a forest
that was 217 m from the closest structure. With this extensive movement into natural areas,
birds are at greater risk of encountering a wandering cat, and therefore risk of predation.
Blancher (2013) estimated the number of birds killed by cats every year in Canada using an
average predation rate of 2.8 birds killed per owned cat per year and 24 birds per unowned
cat per year. Although the exact predation rates in the south Okanagan Valley are currently
unknown, extrapolating these rates suggests that upwards of 100,000 birds are killed by
wandering cats in this region, and this would likely include many species at risk. The
potential impact on native small mammal species (including many endangered bat species) is

106
also exceptionally high. If using predation rate estimates based on those used in Loss et al.
(2013), ranging from 177.3 to 299.5 small mammals killed per cat annually, I estimate at
least 1 million small mammals are killed every year by domestic cats within the study region.
While these estimates are based on average and not median predation rates from studies in
other regions of the world, researching predation rates in the south Okanagan Valley would
provide a much-needed next step to better estimate the true impact that wandering cats have
on bird populations and other small animals in this region. Such research, when combined
with the cat population and abundance estimates (Chapter 3) along with bird abundance
estimates (Chapter 2), would provide important context to understand what the true impacts
of wandering cats are on birds and other animals that could be applied to other ecosystems.
4.2 Management and Conservation Implications
Despite the high per capita density of wandering cats in my study, many residents and the
general public overall are unaware of their prevalence due to the crepuscular and elusive
behaviour of cats, which makes them rarely seen (Merčnik et al., 2023). Numerous residents
were surprised to learn, through the trail camera images, that cats regularly frequented their
yards. This lack of awareness may be attributed to differing perspective on wandering cats, as
non-cat owners tend to express greater concern regarding their presence than cat owners
(Booth & Otter, 2024). The public’s unawareness and division regarding cat regulations
complicates efforts to build support for effective management strategies, such as enforceable
no-roam bylaws (Trouwborst et al., 2020). While their presence often goes unnoticed,
wandering cats have ecological impacts comparable to or exceeding those of wild carnivores,
with their high densities amplifying predation pressure on local wildlife (Kays et al., 2020).
Allowing cats to wander also imposes several welfare risks to the cats themselves, including

107
predation from other animals (e.g. coyotes, cougars, or bears), vehicle collisions, and disease
transmission (Tan et al., 2020). Cats are well known to acquire or transmit numerous
zoonoses through interactions with one cat to another (e.g., feline leukemia virus and rabies;
Fehlner-Gardiner et al., 2024; Powers et al., 2018), from wildlife (e.g., avian influenza and
toxoplasmosis; Burrough et al., 2024; Rahimi et al., 2015), and even transmit diseases to
other wildlife and humans (e.g., toxoplasmosis; Hollings et al., 2013; Rahimi et al., 2015).
While stakeholder groups such as Trap-Neuter-Release (TNR) groups, rescues, shelters, and
veterinarians (Canadian Veterinary Medical Association [CVMA], 2020) may be aware of the
challenges posed by wandering cats, municipalities – who have the power to implement
change – often underestimate these challenges or chose to ignore them (Canadian Federation
of Humane Societies [CFHS], 2017). This was evidenced by a 2017 survey in which the
majority of municipalities stated that cat overpopulation was not a concern (CFHS, 2017).
Removing all or most wandering cats from the landscape in the study region and
elsewhere is likely impossible, but management strategies may help mitigate their
environmental impact, benefiting wildlife, people, and the cats themselves. One proposed
approach for curtailing the number of owned cats in natural habitats involves buffer zones
around natural areas. This approach focuses on protecting birds and keeping cats out of
natural areas, which we’ve identified as important habitat for birds during spring migration,
breeding, and fall migration. Residents within the buffer would be prohibited from allowing
their cats outdoors and restricted from providing resources to unowned cats (e.g. can’t feed or
house barn cats that can leave the barn; Dunford et al., 2024; Pyott et al., 2024; R. L. Thomas
et al., 2014), and cats could be actively trapped in these buffers. In my study, only two cats
travelled more than 500 m between trail camera sites, suggesting that buffers of 750 meters
(Dunford et al., 2024) or 840 m (Pyott et al., 2024) may effectively protect wildlife in natural

108
habitats by restricting cat access and thereby reducing the risk to wildlife. This would be
critical for sensitive habitat in the south Okanagan Valley such as areas neighbouring
Important Bird Areas (IBA) and Migratory Bird Sanctuaries. For example, Vaseux Lake in
Okanagan Falls is a Migratory Bird Sanctuary, and the surrounding area is part of an IBA, yet
11 of 19 sites in this area detected at least one cat with a total of 14 wandering cats.
Other options to curtail the number of wandering cats includes by-law enforcement,
public outreach, and targeted funding, all of which would effectively reduce predation
pressure on birds throughout the south Okanagan Valley (also see Cecchetti et al., 2021). For
example, by-laws prohibiting owners from allowing their cats to wander outdoors would
effectively reduce the number of owned wandering cats. This has been introduced in
Osoyoos, but is not enforced, as one third (30.7%; N = 115) of the wandering cats I identified
were wandering in Osoyoos. Enforcing such by-laws requires resources, which could be
supported through licensing fees and penalties for wandering cats. Public education is also
crucial to raise awareness of the risks posed to wandering cats and threats cats pose to
wildlife (CFHS, 2017; Proulx, 1988). Outreach is often the first step in informing the public
and municipal governments about the environmental impacts and economic implication of
wandering cats, an important step in the south Okanagan Valley. Funding initiatives, such as
those spearheaded by the Stewardship Centre for British Columbia, are key to addressing the
high number of wandering cats in the south Okanagan Valley. In the south Okanagan Valley,
outreach initiatives should focus on peri-urban and agricultural habitats, where high cat
numbers overlap with high bird abundance and richness during the spring, summer, and fall.
During winter, outreach should target urban habitats to encourage residents to keep their cats
indoors, as many species move into urban areas at this time. Through collaboration, outreach,

109
and targeted funding, these efforts aim to reduce the impact of wandering cats on this
ecologically diverse and sensitive region.
4.3 Conclusion
In conclusion, the high population of owned and unowned wandering cats in the south
Okanagan Valley places significant predation pressure on birds and other wildlife, yet
controlling their numbers remains a controversial topic that requires input and collaboration
from various stakeholders. Targeted management strategies, such as establishing buffer zones
around natural habitats, strengthening by-law enforcement, enhancing public outreach, and
fostering community-based initiatives, are essential to mitigate the risks that wandering cats
pose to wildlife, themselves, and people. Using these data, I have established baselines that
can be used if repeated in future years to compare the success of cat management strategies. I
have also provided key seasons where birds and other wildlife may be most vulnerable to
predation. This provides a foundation for outreach efforts in the south Okanagan Valley, with
the goal of curtailing the number of wandering domestic cats, and in the process, reducing the
risk to birds and other wildlife.

110
4.4 Literature Cited
Blancher, P. (2013). Estimated Number of Birds Killed by House Cats (Felis catus) in
Canada. Avian Conservation and Ecology, 8(2), art3. https://doi.org/10.5751/ACE-00557080203
Booth, A. L., & Otter, K. (2024). The Law and the Pussycat: Public Perceptions of the Use of
Municipal Bylaws to Control Free-Roaming Domestic Cats in Canada. Journal of Applied
Animal Welfare Science, 27(3), 530–545. https://doi.org/10.1080/10888705.2022.2142059
Burrough, E. R., Magstadt, D. R., Petersen, B., Timmermans, S. J., Gauger, P. C., Zhang, J.,
Siepker, C., Mainenti, M., Li, G., Thompson, A. C., Gorden, P. J., Plummer, P. J., &
Main, R. (2024). Highly Pathogenic Avian Influenza A(H5N1) Clade 2.3.4.4b Virus
Infection in Domestic Dairy Cattle and Cats, United States, 2024. Emerging Infectious
Diseases, 30(7), 1335–1343. https://doi.org/10.3201/eid3007.240508
Canadian Federation of Humane Societies [CFHS]. (2017). Cats in Canada 2017: A five-year
review of cat overpopulation. https://www.g2z.org.au/pdf/CFHS-Cats_In_Canada_2017FINAL_LR.pdf
Canadian Veterinary Medical Association [CVMA]. (2020, April 29). Free-roaming owned,
Abandoned, and Feral Cats. Canadian Veterinary Medical Association: Position
Statements. https://www.canadianveterinarians.net/policy-and-outreach/positionstatements/statements/free-roaming-owned-abandoned-and-feral-cats/
Cecchetti, M., Crowley, S. L., & McDonald, R. A. (2021). Drivers and facilitators of hunting
behaviour in domestic cats and options for management. Mammal Review, 51(3), 307–
322. https://doi.org/10.1111/mam.12230
Dunford, C. E., Loca, S., Marks, N. J., & Scantlebury, M. (2024). Seasonal habitat selection
and ranging of domestic cats (Felis catus) in rural and urban environments. Animal
Biotelemetry, 12(1), 13. https://doi.org/10.1186/s40317-024-00367-0
eBird. (2022). eBird [eBird, Cornell Lab of Ornithology, Ithaca, New York.]. eBird: An
Online Database of Bird Distribution and Abundance. https://ebird.org/region/CA-BCOS?yr=all
Fehlner-Gardiner, C., Gongal, G., Tenzin, T., Sabeta, C., De Benedictis, P., Rocha, S. M.,
Vargas, A., Cediel-Becerra, N., Gomez, L. C., Maki, J., & Rupprecht, C. E. (2024).
Rabies in Cats—An Emerging Public Health Issue. Viruses, 16(10), Article 10.
https://doi.org/10.3390/v16101635

111
Hanmer, H. J., Thomas, R. L., & Fellowes, M. D. E. (2017). Urbanisation influences range
size of the domestic cat (Felis catus): Consequences for conservation. Journal of Urban
Ecology, 3(1), jux014. https://doi.org/10.1093/jue/jux014
Hollings, T., Jones, M., Mooney, N., & McCallum, H. (2013). Wildlife disease ecology in
changing landscapes: Mesopredator release and toxoplasmosis. International Journal for
Parasitology: Parasites and Wildlife, 2, 110–118.
https://doi.org/10.1016/j.ijppaw.2013.02.002
Kays, R., Dunn, R. R., Parsons, A. W., Mcdonald, B., Perkins, T., Powers, S. A., Shell, L.,
McDonald, J. L., Cole, H., Kikillus, H., Woods, L., Tindle, H., & Roetman, P. (2020). The
small home ranges and large local ecological impacts of pet cats. Animal Conservation,
23(5), 516–523. https://doi.org/10.1111/acv.12563
Krauze-Gryz, D., Gryz, J., & Żmihorski, M. (2019). Cats kill millions of vertebrates in Polish
farmland annually. Global Ecology and Conservation, 17, e00516.
https://doi.org/10.1016/j.gecco.2018.e00516
Loss, S. R., Will, T., & Marra, P. P. (2013). The impact of free-ranging domestic cats on
wildlife of the United States. Nature Communications, 4(1), 1396.
https://doi.org/10.1038/ncomms2380
Loyd, K. A. T., Hernandez, S. M., Carroll, J. P., Abernathy, K. J., & Marshall, G. J. (2013).
Quantifying free-roaming domestic cat predation using animal-borne video cameras.
Biological Conservation, 160, 183–189. https://doi.org/10.1016/j.biocon.2013.01.008
McGregor, H., Legge, S., Jones, M. E., & Johnson, C. N. (2015). Feral Cats Are Better
Killers in Open Habitats, Revealed by Animal-Borne Video. PLOS ONE, 10(8),
e0133915. https://doi.org/10.1371/journal.pone.0133915
Medina, F. M., Bonnaud, E., Vidal, E., Tershy, B. R., Zavaleta, E. S., Josh Donlan, C., Keitt,
B. S., Le Corre, M., Horwath, S. V., & Nogales, M. (2011). A global review of the
impacts of invasive cats on island endangered vertebrates. Global Change Biology,
17(11), 3503–3510. https://doi.org/10.1111/j.1365-2486.2011.02464.x
Merčnik, N., Prevolnik Povše, M., Škorjanc, D., & Skok, J. (2023). Chronobiology of freeranging domestic cats: Circadian, lunar and seasonal activity rhythms in a wildlife
corridor. Applied Animal Behaviour Science, 269, 106094.
https://doi.org/10.1016/j.applanim.2023.106094

112
Okanagan Similkameen Stewardship [OSS]. (2023). Birds. Osstewardship.
https://www.osstewardship.ca/birds
Pirie, T. J., Thomas, R. L., & Fellowes, M. D. E. (2022). Pet cats (Felis catus) from urban
boundaries use different habitats, have larger home ranges and kill more prey than cats
from the suburbs. Landscape and Urban Planning, 220, 104338.
https://doi.org/10.1016/j.landurbplan.2021.104338
Powers, J. A., Chiu, E. S., Kraberger, S. J., Roelke-Parker, M., Lowery, I., Erbeck, K.,
Troyer, R., Carver, S., & VandeWoude, S. (2018). Feline Leukemia Virus (FeLV) Disease
Outcomes in a Domestic Cat Breeding Colony: Relationship to Endogenous FeLV and
Other Chronic Viral Infections. Journal of Virology, 92(18), e00649-18.
https://doi.org/10.1128/JVI.00649-18
Proulx, G. (1988). Control of Urban Wildlife Predation by Cats Through Public Education.
Environmental Conservation, 15(4), 358–359.
https://doi.org/10.1017/S0376892900029878
Pyott, M. L., Norris, D. R., Mitchell, G. W., Custode, L., & Gow, E. A. (2024). Home range
size and habitat selection of owned outdoor domestic cats (Felis catus) in urban
southwestern Ontario. PeerJ, 12, e17159. https://doi.org/10.7717/peerj.17159
Rahimi, M. T., Daryani, A., Sarvi, S., Shokri, A., Ahmadpour, E., Teshnizi, S. H., Mizani,
A., & Sharif, M. (2015). Cats and Toxoplasma gondii: A systematic review and metaanalysis in Iran. The Onderstepoort Journal of Veterinary Research, 82(1), 823.
https://doi.org/10.4102/ojvr.v82i1.823
Sedano-Cruz, R. E. (2022). Estimated number of birds killed by domestic cats in Colombia.
Avian Conservation and Ecology, 17(2). https://doi.org/10.5751/ACE-02200-170216
Tan, S. M. L., Stellato, A. C., & Niel, L. (2020). Uncontrolled Outdoor Access for Cats: An
Assessment of Risks and Benefits. Animals, 10(2), 258.
https://doi.org/10.3390/ani10020258
Thomas, R. L., Baker, P. J., & Fellowes, M. D. E. (2014). Ranging characteristics of the
domestic cat (Felis catus) in an urban environment. Urban Ecosystems, 17(4), 911–921.
https://doi.org/10.1007/s11252-014-0360-5

113
Trouwborst, A., McCormack, P. C., & Martínez Camacho, E. (2020). Domestic cats and their
impacts on biodiversity: A blind spot in the application of nature conservation law. People
and Nature, 2(1), 235–250. https://doi.org/10.1002/pan3.10073
Woinarski, J. C. Z., Murphy, B. P., Legge, S. M., Garnett, S. T., Lawes, M. J., Comer, S.,
Dickman, C. R., Doherty, T. S., Edwards, G., Nankivell, A., Paton, D., Palmer, R., &
Woolley, L. A. (2017). How many birds are killed by cats in Australia? Biological
Conservation, 214, 76–87. https://doi.org/10.1016/j.biocon.2017.08.006
Woods, M., Mcdonald, R. A., & Harris, S. (2003). Predation of wildlife by domestic cats
Felis catus in Great Britain. Mammal Review, 33(2), 174–188.
https://doi.org/10.1046/j.1365-2907.2003.00017.x

114
Appendix 1
Addition, movement, or loss of trail camera sites.
Trail camera sites were moved, discarded, and added throughout the year. Trail camera sites
were added when homeowners reached out after the completion of the first season (N = 1).
Trail camera sites were lost due to the following reasons: homeowners no longer agreed to
participate in the study after one season (N = 1), the homeowners moved and I therefore lost
access to the site (N = 2), the homeowners didn’t reply when attempting to expand the project
for a 5th season (N = 2), or the trail camera was accidentally destroyed during the last season,
along with the fence it was attached to (N = 1). No theft occurred during this study and any
other loss of data was a result of technical errors. Trail camera sites were moved when the
vegetation grew too tall but couldn’t be removed and the camera couldn’t be repositioned in
the same place (N = 1), therefore I found the next closest spot to place the camera.

Appendix 2
Additional Study Design information regarding site selection.
Sites in natural habitats were obtained by collaborating with the Ministry of Forests, B.C.
Parks, and Nature Trust of British Columbia to acquire permits. In agricultural areas, I
obtained permission by emailing vineyards and speaking with vineyard owners in-person.
For urban and peri-urban sites, I collaborated with local organizations to help advertise for
volunteers and went door-to-door to seek landowner participation in areas with low numbers
of volunteers. Local organizations (e.g., Castanet and Penticton Naturalist Club) helped
advertise the project, reach out to homeowners, and share contact information with those
interested in participating.

115
Appendix 3
Distance Sampling equation and assumptions.
Distance sampling broadly belongs to the category of multinomial outcomes and estimates
abundance using the following equation:
𝑦𝑖𝑡 ~ 𝑀𝑢𝑙𝑡𝑖𝑛𝑜𝑚𝑖𝑎𝑙 (𝑁𝑖𝑡 , 𝜋𝑖𝑡 )
Where 𝑦𝑖𝑡 represents counts per plot 𝑖 on occasion 𝑡, 𝑁𝑖𝑡 denotes the total number of
individuals available for detection in plot 𝑖 on occasion 𝑡, and 𝜋𝑖𝑡 represents multinomial
probabilities derived from detection probabilities, 𝑝 (Chandler et al., 2011).
There are two other main assumptions for the gdistsamp model; first, that objects at the point
are detected with certainty and second, measurements of these detections are exact (Thomas
et al., 2010). The model addresses the first assumption by assigning higher detection
probabilities to birds closer to the observer, where they naturally have a greater likelihood of
being detected. To manage distance-dependent detection probability, my surveys categorized
observations into distance bins: 0-50 meters and 50-100 meters. This categorization allowed
me to estimate observation distances more accurately, thereby addressing the second
assumption of exact measurements.

116
Appendix 4
Bray-Curtis distance metric rationale.
I compared multiple distance metrics (Bray-Curtis, Canberra, Euclidean, and Gower) and
multiple axes (k = 1 to k = 4) for each season. The model with the lowest acceptable stress
(below 0.2) was chosen for each season and a stress plot was generated to assess the
adequacy of the model. Despite the Euclidean distance metric yielding the lowest stress
scores, the ordination plots produced did not accurately represent these data because
Euclidean distance does not take species identity into account (Legendre & Legendre, 2012).

Appendix 5
NMDS Stress Values.
Spring migration = 0.18, breeding season = 0.18, fall migration = 0.18, early non-breeding
season = 0.17, and late non-breeding season = 0.16.

117
Appendix 6
Table A6.1. All species observed during point counts during the spring migration (SM; N
= 480 point counts), breeding season (BS; N = 480 point counts), fall migration (FM; N =
480 point counts), early non-breeding season (EN; N = 476 point counts), and late nonbreeding season (LN; N = 464 point counts). Birds that are at higher risk of predation by a
domestic cat are denoted with an asterisk. The total number of individuals for each species
and season are calculated, as well as the total number of species seen during each season.
Common name, scientific name, and four-letter banding codes are provided. Only birds
seen within 100 meters from the observer are provided.
4-letter
banding
code
ALFL
AMCO
AMCR*

0
15
118

2
6
114

0
44
70

0
6
36

0
0
36

2
71
374

AMDI
AMGO*

0
155

0
215

0
276

4
124

0
120

4
890

Falco sparverius
Anthus rubescens
Turdus migratorius
Mareca americana

AMKE
AMPI
AMRO*
AMWI

2
0
848
6

3
0
511
0

4
4
335
0

1
1
1146
0

0
0
373
2

10
5
3213
8

Spizelloides arborea

ATSP*

2

0

0

3

1

6

Haliaeetus
leucocephalus
Bucephala islandica

BAEA

3

0

2

5

12

22

BAGO

0

0

0

9

0

9

Riparia riparia
Hirundo rustica
Pica hudsonia

BANS
BARS*
BBMA

0
5
19

33
45
38

0
66
10

0
0
27

0
0
16

33
116
110

Poecile atricapillus

BCCH*

105

118

199

130

170

722

Archilochus
alexandri
Megaceryle alcyon
Thryomanes
bewickii
Molothrus ater

BCHU

0

1

0

0

0

1

BEKI
BEWR*

0
8

1
4

21
12

9
17

0
15

31
56

BHCO*

20

58

3

0

0

81

Pheucticus
melanocephalus
Dolichonyx
oryzivorus
Bombycilla garrulus

BHGR

0

15

6

0

0

21

BOBO*

0

1

0

0

0

1

BOWA

12

0

0

281

309

602

Euphagus
cyanocephalus
Bucephala albeola
Icterus bullockii
Haemorhous
cassinii
Selasphorus calliope

BRBL*

224

322

425

0

0

971

BUFF
BUOR
CAFI

31
5
3

0
30
7

0
0
5

39
0
0

116
0
1

186
35
16

CAHU

0

6

0

0

0

6

Branta canadensis
Catherpes
mexicanus
Callipepla californica
Aphelocoma
californica

CANG
CANW*

19
18

118
30

88
33

81
14

126
25

432
120

CAQU*
CASJ

1358
0

914
0

1004
0

803
1

921
0

5000
1

Common Name

Scientific Name

Alder flycatcher
American coot
American crow

Empidonax alnorum
Fulica americana
Corvus
brachyrhynchos
Cinclus mexicanus
Spinus tristis

American dipper
American
goldfinch
American kestrel
American pipit
American robin
American
wigeon
American tree
sparrow
Bald eagle
Barrow’s
goldeneye
Bank swallow
Barn swallow
Black-billed
magpie
Black-capped
chickadee
Black-chinned
hummingbird
Belted kingfisher
Bewick’s wren
Brown-headed
cowbird
Black-headed
grosbeak
Bobolink
Bohemian
waxwing
Brewer’s
blackbird
Bufflehead
Bullock’s oriole
Cassin’s finch
Calliope
hummingbird
Canada goose
Canyon wren
California quail
California scrub
jay

SM

BS

FM

EN

LN

Total

118
Cedar waxwing
Chipping
sparrow
Chukar
Clark’s
nutcracker
Cliff swallow
Common
goldeneye
Cooper’s hawk
Common
merganser
Common raven
Common
yellowthroat
Dark-eyed junco
Downy
woodpecker
Dusky flycatcher
Eastern kingbird
Eurasian
collared-dove
European
starling
Evening
grosbeak
Fox sparrow
Gadwall
Great blue heron
Golden-crowned
kinglet
Great-horned
owl
Gray catbird
Gray flycatcher
Greater scaup
Greater whitefronted goose
Green-winged
teal
Harry
woodpecker
Herring gull
Hermit thrush
House finch
Hooded
merganser
House sparrow
House wren
Killdeer
Lesser goldfinch
Lark sparrow
Lazuli bunting
Least flycatcher
Lesser scaup
Lincoln’s
sparrow
Mallard
Marsh wren
Merlin
MacGillivray’s
warbler
Mountain
bluebird

Bombycilla
cedrorum
Spizella passerina

CEDW*

1

586

525

216

48

1376

CHSP

42

84

31

0

0

157

Alectoris chukar
Nucifraga
columbiana
Petrochelidon
pyrrhonota
Bucephala clangula

CHUK*
CLNU

0
0

0
7

0
6

7
0

0
0

7
13

CLSW

0

5

0

0

0

5

COGO

2

2

0

115

58

177

Accipiter cooperii
Mergus merganser

COHA
COME

1
10

0
0

1
0

4
32

5
6

11
48

Corvus corax
Geothlypis trichas

CORA
COYE

3
3

0
41

13
2

30
0

8
0

54
46

Junco hyemalis
Picoides pubescens

DEJU*
DOWO*

135
18

0
18

15
28

1028
8

726
16

1904
88

Empidonax
oberholseri
Tyrannus tyrannus
Streptopelia
decaocto
Sturnus vulgaris

DUFL

14

0

0

0

0

14

EAKI
EUCD*

0
425

30
408

70
377

0
202

0
273

100
1685

EUST*

1034

679

2492

5450

777

10432

Coccothraustes
vespertinus
Passerella iliaca
Mareca strepera
Ardea herodias
Regulus satrapa

EVGR*

2

2

16

0

0

20

FOSP
GADW
GBHE
GCKI

0
3
5
0

0
0
3
0

0
0
2
0

2
0
2
1

0
0
1
0

2
3
13
1

Bubo virginianus

GHOW

9

0

2

0

0

11

Dumetella
carolinensis
Empidonax wrightii
Aythya marila
Anser albifrons

GRCA*

0

46

51

2

0

99

GRFL
GRSC
GWFG

0
0
0

1
2
0

0
0
0

0
0
2

0
0
0

1
2
2

Anas carolinensis

GWTE

0

0

0

1

4

5

Leuconotopicus
villosus
Larus argentatus
Catharus guttatus
Haemorhous
mexicanus
Lophodytes
cucullatus
Passer domesticus
Troglodytes aedon
Charadrius vociferus
Spinus psaltria
Chondestes
grammacus
Passerina amoena
Empidonax minimus
Aythya affinis
Melospiza lincolnii

HAWO*

0

0

1

2

0

3

HERG
HETH
HOFI*

0
0
493

0
0
272

0
1
480

3
1
901

0
0
528

3
2
2674

HOME

10

1

0

10

0

21

HOSP*
HOWR*
KILL*
LEGO
LASP*

2528
2
102
0
0

1787
25
81
0
13

1336
7
48
0
9

1556
0
0
1
0

2256
0
4
0
0

9463
34
235
1
22

LAZB*
LEFL
LESC
LISP

0
0
7
1

87
1
0
0

28
0
0
9

0
0
12
2

0
0
28
0

115
1
47
12

Anas platyrhynchos
Cistothorus palustris
Falco columbarius
Geothlypis tolmiei

MALL
MAWR
MERL
MGWA*

93
3
0
0

37
5
1
0

63
0
0
1

629
4
2
0

317
0
2
0

1139
12
5
1

Sialia currucoides

MOBL*

3

0

0

1

0

4

119
Mountain
chickadee
Mourning dove
Nashville
warbler
Northern flicker
Northern
goshawk
Northern harrier
Northern
waterthrush
Northern shrike
Northern roughwinged swallow
Orange-crowned
warbler
Olive-sided
flycatcher
Osprey
Pacific wren
Pied-billed grebe
Peregrine falcon
Pine siskin
Pileated
woodpecker
Pygmy nuthatch
Ring-billed gull
Red-breasted
merganser
Red-breasted
nuthatch
Ruby-crowned
kinglet
Red crossbill
Red-eyed vireo
Ring-necked
duck
Ring-necked
pheasant
Rock pigeon
Rock wren
Red-tailed hawk
Ruddy duck
Ruffed grouse
Rufous
hummingbird
Red-winged
blackbird
Say’s phoebe
Savannah
sparrow
Snow bunting
Sora
Song sparrow
Spotted
sandpiper
Spotted towhee
Sharp-shinned
hawk
Steller’s jay
Townsend’s
solitaire
Tree swallow
Trumpeter swan
Turkey vulture
Varied thrush

Poecile gambeli

MOCH*

0

6

1

0

8

15

Zenaida macroura
Leiothlypis ruficapilla

MODO*
NAWA

29
11

55
2

53
0

4
0

0
0

141
13

Colaptes auratus
Accipiter gentilis

NOFL*
NOGO

340
0

200
0

224
0

284
3

159
0

1207
3

Circus cyaneus
Parkesia
noveboracensis
Lanius borealis
Stelgidopteryx
serripennis
Vermivora celata

NOHA
NOWA

4
0

1
0

0
5

2
0

0
0

7
5

NSHR
NRWS

0
39

0
58

0
2

4
0

7
0

11
99

OCWA

5

0

0

0

0

5

Contopus cooperi

OSFL

0

0

6

0

0

6

Pandion haliaetus
Troglodytes
pacificus
Podilymbus
podiceps
Falco peregrinus
Spinus pinus
Dryocopus pileatus

OSPR
PAWR

10
0

6
0

45
0

0
2

0
0

61
2

PBGR

7

4

9

0

0

20

PEFA
PISI*
PIWO

4
8
0

9
39
0

2
52
1

0
0
0

3
0
0

18
99
1

Sitta pygmaea
Larus delawarensis
Mergus serrator

PYNU
RBGU
RBME

12
0
0

32
9
0

33
3
0

30
0
2

123
0
0

230
12
2

Sitta canadensis

RBNU*

20

19

5

4

7

55

Regulus calendula

RCKI

61

0

1

3

0

65

Loxia curvirostra
Vireo olivaceus
Aythya collaris

RECR
REVI
RNDU

19
0
0

58
11
0

18
4
0

12
0
0

17
0
14

124
15
14

Phasianus colchicus

RNPH*

48

41

4

1

3

97

Columba livia
Salpinctes obsoletus
Buteo jamaicensis
Oxyura jamaicensis
Bonasa umbellus
Selasphorus rufus

ROPI*
ROWR*
RTHA
RUDU
RUGR
RUHU

43
2
6
12
1
1

12
0
1
9
0
0

65
0
3
0
0
0

328
0
12
6
0
0

120
2
4
0
0
0

568
4
26
27
1
1

Agelaius phoeniceus

RWBL*

464

432

81

171

69

1217

Sayornis saya
Passerculus
sandwichensis
Plectrophenax
nivalis
Porzana carolina
Melospiza melodia
Actitis macularius

SAPH
SAVS*

109
0

38
0

23
8

2
0

7
0

179
8

SNBU

0

0

0

4

0

4

SORA
SOSP*
SPSA

0
114
5

8
101
21

0
29
5

2
154
0

1
137
0

11
535
31

Pipilo maculatus
Accipiter striatus

SPTO*
SSHA

60
0

44
0

22
5

37
6

27
1

190
12

Cyanocitta stelleri
Myadestes
townsendi
Tachycineta bicolor
Cygnus buccinator
Cathartes aura
Ixoreus naevius

STJA*
TOSO

0
7

0
5

7
0

3
21

0
23

10
56

TRES
TRUS
TUVU
VATH*

21
0
38
0

37
0
0
0

6
0
0
0

0
3
0
35

0
8
0
9

64
11
38
44

120
Veery
Vesper sparrow
Violet-green
swallow
Warbling vireo
White-breasted
nuthatch
White-crowned
sparrow
Western
bluebird
Western kingbird
Western
meadowlark
Western tanager
Western woodpeewee
Willow flycatcher
Wilson’s warbler
Wood duck
White-throated
sparrow
White-throated
swift
Yellow-breasted
chat
Yellow rail
Yellow warbler
Yellow-headed
blackbird
Yellow-rumped
warbler
Total individuals
Total Species

Catharus
fuscescens
Pooecetes
gramineus
Tachycineta
thalassina
Vireo gilvus
Sitta carolinensis

VEER

0

8

0

0

0

8

VESP*

38

52

31

0

0

121

VGSW

213

142

6

0

0

361

WAVI
WBNU*

2
1

1
1

1
0

0
2

0
4

4
8

Zonotrichia
leucophrys
Sialia mexicana

WCSP*

952

0

38

111

22

1123

WEBL

57

9

24

159

38

287

Tyrannus verticalis
Sturnella neglecta

WEKI
WEME*

2
39

6
48

0
10

0
0

0
1

8
98

Piranga ludoviciana
Contopus sordidulus

WETA
WEWP

3
2

0
201

19
113

0
0

0
0

22
316

Empidonax traillii
Cardellina pusilla
Aix sponsa
Zonotrichia albicollis

WIFL*
WIWA
WODU
WTSP*

0
1
13
0

95
5
6
0

0
1
22
0

0
0
39
1

0
0
14
0

95
7
94
1

Aeronautes
saxatalis
Icteria virens

WTSW

2

30

0

0

0

32

YBCH*

6

38

2

0

0

46

Coturnicops
noveboracensis
Setophaga petechia
Xanthocephalus
xanthocephalus
Setophaga coronata

YERA

0

0

3

0

0

3

YEWA
YHBL

7
6

164
8

67
0

2
0

0
0

240
14

YRWA

107

0

78

19

1

205

10876
93

8888
92

9358
88

14431
82

8129
59

51682
146

Table A6.2. Total number of observed species in each land class during each season.
Urban
Peri-urban Agricultural
Natural
Spring migration

41

50

56

74

Breeding Season

45

53

59

79

Fall Migration

38

56

56

66

Early Non-breeding
Season
Late Non-Breeding
Season

37

51

40

56

30

35

27

52

121
Table A6.3. LMER pairwise abundance comparison across land classes for each season. I
assessed the four land classes, Agriculture (AG), Natural (NA), Urban (UR), and Periurban (PU), across 5 seasons.
Spring
Estimate
SE
DF
P-value
Migration
AG – NA
13.540
5.56
270
0.0729
AG – PU
-43.765
5.86
271
<.0001
AG - UR
-39.184
5.60
270
<.0001
NA - PU
-57.305
5.80
274
<.0001
NA - UR
-52.724
5.54
273
<.0001
PU - UR
4.581
5.84
274
0.8615
Breeding Season
AG – NA
-19.765
5.57
270
0.0025
AG – PU
-44.113
5.87
271
<.0001
AG - UR
-27.009
5.61
270
<.0001
NA - PU
-24.349
5.81
274
0.0002
NA - UR
-7.245
5.55
273
0.5607
PU - UR
17.104
5.86
274
0.0197
Fall Migration
AG – NA
1.315
5.57
270
0.9953
AG – PU
-55.807
5.87
271
<.0001
AG - UR
-19.964
5.61
270
0.0025
NA - PU
-57.123
5.81
274
<.0001
NA - UR
-21.279
5.55
273
0.0009
PU - UR
35.843
5.86
274
<.0001
Early Non-breeding Season
AG – NA
29.193
5.57
270
<.0001
AG – PU
7.832
5.91
276
0.5470
AG - UR
-5.250
5.61
270
0.7855
NA - PU
-21.362
5.85
279
0.0018
NA - UR
-34.443
5.55
273
<.0001
PU - UR
-13.081
5.89
274
0.1204
Late Non-breeding Season
AG – NA
0.472
5.59
273
0.9998
AG – PU
-37.576
5.97
284
<.0001
AG - UR
-31.096
5.66
277
<.0001
NA - PU
-38.003
5.89
284
<.0001
NA - UR
-31.523
5.58
277
<.0001
PU - UR
6.480
5.96
287
0.6971

122
Table A6.4. LMER pairwise species richness comparison across land classes for each
season. I assessed the four land classes, Agriculture (AG), Natural (NA), Urban (UR), and
Peri-urban (PU), across 5 seasons.
Spring
Estimate
Migration
AG – NA
-0.1621
AG – PU
-1.1387
AG - UR
-0.3598
NA - PU
-0.9766
NA - UR
-0.1977
PU - UR
0.7789
Breeding Season
AG – NA
-2.1248
AG – PU
-1.7552
AG - UR
1.1972
NA - PU
0.3696
NA - UR
3.3221
PU - UR
2.9525
Fall Migration
AG – NA
-0.3234
AG – PU
-1.0235
AG - UR
0.5924
NA - PU
-0.7001
NA - UR
0.9158
PU - UR
1.6159
Early Non-breeding Season
AG – NA
-0.4386
AG – PU
-1.6541
AG - UR
-1.6899
NA - PU
-1.2155
NA - UR
-1.2513
PU - UR
0.0358
Late Non-breeding Season
AG – NA
-0.7745
AG – PU
-2.5829
AG - UR
-2.3457
NA - PU
-1.8084
NA - UR
-1.5712
PU - UR
0.2372

SE

DF

P-value

0.531
0.560
0.535
0.555
0.530
0.559

418
419
418
423
422
423

0.9901
0.1774
0.9075
0.2948
0.9823
0.5045

0.531
0.561
0.536
0.556
0.531
0.560

416
417
416
420
419
420

0.0004
0.0100
0.1155
0.9104
<.0001
<.0001

0.531
0.561
0.536
0.556
0.531
0.560

416
417
416
420
419
420

0.9293
0.2627
0.6860
0.5898
0.3121
0.0214

0.531
0.566
0.536
0.561
0.531
0.565

416
423
416
425
419
425

0.8424
0.0190
0.0093
0.1346
0.0873
0.9999

0.535
0.574
0.543
0.567
0.535
0.574

421
433
425
432
424
436

0.4704
0.0001
0.0001
0.0082
0.0182
0.9762

123
Appendix 7
NMDS Seasonal Bird Communities
During spring migration, positive NMDS1 and negative NMDS2 were linked with desert
habitat and species such as Canyon wrens (Catherpes mexicanus), Western Meadowlarks
(Sturnella neglecta), and Western Bluebirds (Sialia mexicana). Conversely, negative NMDS1
values were associated with buildings and species like House Sparrows (Passer domesticus),
Eurasian Collared-doves (Streptopelia decaocto), and California Quail (Callipepla
californica). NMDS2 displayed positive associations with natural habitat variables including
water, forests, and fields, which corresponded to species such as European Starlings (Sturnus
vulgaris), Red-winged Blackbirds (Agelaius phoeniceus), Northern Flickers (Colaptes
auratus), and Song Sparrows (Melospiza melodia). For NMDS3, positive values were
associated with White-crowned Sparrows (Zonotrichia leucophrys), while negative values
associated with vineyards, altitude, and species such as American Robin (Turdus
migratorius), Say’s Phoebe (Sayornis saya), Dark-eyed Junco (Junco hyemalis), and Vesper
Sparrow (Pooecetes gramineus; Figure 2.5a and Figure A7.1).
During the breeding season, negative NDMS1 values corresponded to buildings and
orchards, as well as species like House Sparrow, Eurasian Collared-dove, and Brewer’s
Blackbirds (Euphagus cyanocephalus). The forest habitat variable was strongly associated
with both positive NMDS1 and negative NMDS2, which highlighted the strong association
between forests and species such as the Western Wood-peewee (Contopus sordidulus). Other
natural habitat variables, fields and water, were associated with both negative NMDS2 and
NMDS3 values, and with species such as European Starling, Song Sparrow, Red-winged
Blackbird, and Common Yellowthroat (Geothlypis trichas). Some species had a strong
association with negative NMDS3, such as Willow Flycatcher (Empidonax traillii) and Cedar

124
Waxwing (Bombycilla cedrorum). NMDS2 and NMDS1 revealed positive correlations with
desert habitats and species such as Canyon Wren, Western Meadowlark, Black-billed Magpie
(Pica hudsonia), and Lazuli Bunting (Passerina amoena). Positive NMDS3 values were
associated with vineyards and altitude, and species including Vesper Sparrow, California
Quail, American Robin, and Spotted Towhee (Pipilo maculatus; Figure 2.5b and Figure
A7.1).
During the fall migration, NMDS1 values were positively associated with deserts and
species such as Lazuli Bunting and Chipping Sparrows (Spizella passerina). Negative
NMDS1 values related to fields, buildings, and species including European Starling,
California Quail, Brewer’s Blackbird, House Sparrow, and Eurasian Collared-dove. NMDS2
showed positive associations with altitude, American Robin, and Yellow-rumped Warblers
(Setophaga coronata). For NMDS3, positive values were tied to natural habitat variables
such as water, forests, and fields, and with species including Cedar Waxwing, Northern
Flicker, Western Wood-peewee, and Yellow Warbler (Setophaga petechia). Negative NMDS3
values were associated with orchards, buildings, and species like American Crows (Corvus
brachyrhynchos) and Eurasian-collared Dove (Figure 2.5c and Figure A7.1).
During the early non-breeding season, NMDS1 values were positively associated with
the natural habitat variables of desert, forest, water, altitude, and only the Canyon Wren.
Negative NMDS1 values are associated with species like American robin, European starling,
California quail, Dark-eyed junco, and House finch (Haemorhous mexicanus). Positive
NMDS2 values exhibited associations with fields, forests and vineyards, and species such as
Northern Flicker and European Starling, while negative NMDS2 values related to buildings,
Eurasian Collared-dove, House Sparrow, and American Goldfinch (Spinus tristis). For
NMDS3, positive values are associated with fields and water, as well as species like House

125
Finch, Song Sparrow, Mallards (Anas platyrhynchos), Buffleheads (Bucephala albeola), and
Common Goldeneye (Bucephala clangula). Negative NMDS3 values correlated with altitude
and species such as Black-billed Magpie, Dark-eyed Junco, American Robin, and Western
Bluebird (Figure 2.5d and Figure A 7.1).
In the late non-breeding season, positive NMDS1 values were associated with
buildings and species like House Sparrow, Eurasian Collared-dove, House Finch, and
California Quail. Negative NMDS1 values are associated with habitat variables such as
vineyards, altitude, and forests, with no specific bird species associations. NMDS2 showed
positive values related to altitude, deserts, and buildings, and species such as Pygmy
Nuthatch (Sitta pygmaea), House Sparrow, and Canyon Wren. Conversely, negative NMDS2
values are associated with natural habitat variables like forests, water, and fields, and species
including Mallard, Song Sparrow, Bufflehead, Common Goldeneye, Black-capped Chickadee
(Poecile atricapillus), and Bewick’s Wren (Thryomanes bewickii). For NMDS3, positive
values are linked to species such as American Robin, European Starling, and Black-billed
Magpie, while negative values are associated with Dark-eyed Junco and Canada Goose
(Branta canadensis; Figure 2.5e and Figure A7.1).

126

Figure A7.1. NMDS comparing species habitat associations across A and F) the spring migration B and G) breeding season, C and
H) fall migration, D and I) early non-breeding season, and E and J) late non-breeding season. The top row is all NMDS1 compared
to NMDS3, while the bottom row is NMDS2 compared to NMDS3. Birds are represented by four-letter banding codes (See Table
A1 for 4-letter banding code and associated common and scientific names). Each row shows the three-dimensional comparisons
for that season

127

Figure A7.2. NMDS comparing species compositions across urban (UR), peri-urban (PU), agricultural (AG), and natural (NA)
areas. I also compared four seasons: A and F) the spring migration B and G) breeding season, C and H) fall migration, D and I)
early non-breeding season, and E and J) late non-breeding season. The top row shows NMDS1 vs NMDS3 and the bottom row
shows NMDS2 vs NMDS3. Each point in the figure represents a point count and its location on the NMDS is based on the birds
that were observed during that point count. Points that are closer together are more similar in species composition and points that
are further apart are dissimilar in their species composition. The circles around the points show a 95% confidence interval.

128
Appendix 8
South Okanagan Valley, B.C., study area details and human population estimates.
The study area consists of three main sections, each encompassing a town and its surrounding
area. Okanagan Falls is the smallest and most northern of the three towns, with a population
of 2,266 (Government of Canada, 2023). Okanagan Falls is part of Electoral District D,
which is 583.93 km2 with a population of 4016 (Government of Canada 2023). Oliver, which
is south of Okanagan Falls, has a population of 5,094 and is roughly 5.49 km2 (Government
of Canada, 2023). The study area surrounding Oliver is part of Electoral District C, which is
444.75 km2 in size and has a population of 3,986 (Government of Canada, 2023). The most
southern town is Osoyoos, which covers a total of 8.41 km2 and has a population of 5,556
(Government of Canada, 2023). Most of the study area surrounding Osoyoos is part of
Electoral District A, which has a population of 2,139 and is 258.04 km2 in size (Government
of Canada, 2023). The area northeast of Osoyoos is part of the census area of Osoyoos 1
Indian Reserve, belonging to the Osoyoos Indian Band, with a population of 1,426 and a size
of 130.34 km2 (Government of Canada, 2023). To estimate the total human population size, I
combined all census data, which covers a total of 1,430.96 km2 and has a total population of
22,217. This is an overestimation of the population size within the study area, but due to a
lack of data, I can only say the population is between 12,916 (total of all three town
populations) and 22,217 (total population estimates of towns and surrounding areas that
surpass my study area).

129
Appendix 9
Additional methods regarding partnership in using trail cameras.
The Stewardship Centre for British Columbia’s mission is to strengthen ecological
stewardship through programs and resources for organizations, governments, private sectors,
and the general public. This research aims to create the foundation of these resources and
programs in the south Okanagan Valley to curtail the number of wandering cats, reducing
welfare risks for both cats and wildlife. Partnership was also critical throughout the planning
and deployment phases of this research, especially in finding camera sites for this study. I
worked with local organizations and government groups to attain permits (e.g., The Nature
Trust of British Columbia and BC Parks), and I connected with local groups (e.g., Penticton
Naturalist Club) that would help me find interested participants to host trail cameras on
private property. Permits were obtained from the Ministry of Forests, BC Parks, and The
Nature Trust of British Columbia to put up cameras in natural habitats. There was also an
article in the local news (Castanet) which discussed my research and search for household
residents to let me put up cameras (Richardson, 2022). Additionally, the Stewardship Centre
for British Columbia helped in the creation and distribution of blogs that updated the public
on the research and any interesting findings throughout fieldwork (Wilson & Skurikhina,
2022a, 2022b).

130
Appendix 10
Additional information regarding the acquisition of trail camera sites.
Approximately 50% of the landowners I spoke with expressed interest in participating,
emphasizing the importance of door-to-door engagement in strategically placing cameras
within urban and peri-urban habitats. Speaking with residents face-to-face allowed me to
explain the project in detail, showcase the cameras, and address their questions. Going doorto-door was also essential for finding agricultural sites; while many vineyards were initially
contacted by email, responses were rare until follow-ups were conducted in person or over
the phone. This initial meetings and subsequent interactions during camera switches ensured
that property owners or workers became active participants in this research. I encouraged
them to talk to me about what they observed in their yards or properties and I in turn shared
highlights of interesting animals or other observations from the trail camera photos.

131
Appendix 11

Figure A11 Letter to landowners to assist with hosting trail cameras. Original letters only
included four periods between March and December, because the late winter was added later
in the project with the help of additional funding from Canadian Wildlife Services.

132
Appendix 12
Additional technical considerations when using trail cameras.
Multiple steps were taken to ensure the continuous capture and longevity of the trail
cameras. I used lithium batteries to keep the Browning Strike Force HD Pro infrared trail
cameras (model BTC-5HDP) functioning for as long as possible and checked the cameras biweekly to replace batteries and SD cards when necessary. Past experience with these cameras
has indicated alkaline batteries last between 4–8 months, and lithium batteries ~8 months.
Given the short deployment time of each camera, I did not expect cameras to run out of
batteries. Despite this, there were some technical difficulties (e.g., some cameras stopped
working when temperatures were too hot), which caused the camera to stop taking pictures
for extended periods. The Okanagan had multiple heat waves during July and August, which
held temperatures above 30°C, even at night, for weeklong periods, some days reaching over
40°C. In addition to malfunctioning, this seemed to have caused cameras to die quickly,
making it necessary to replace batteries every 1-2 weeks during the spring and summer
seasons. I mainly used SD cards that were 32 GB in capacity, with an additional 5 SD cards
with 128 GB of storage for sites where high volumes of photos were common. Many SD
cards filled up with pictures during windy periods or periods of vegetation growth, as the
grass grew to the height of the camera and triggered the sensor.

133
Appendix 13
Additional ethical considerations when using trail cameras
Many of the cameras were placed along public walking paths in natural habitats, but because
cameras were positions low to the ground to capture cats (30-60 cm above ground, i.e., cat
height), I was often able to avoid capturing human faces, avoiding privacy concerns. In
urban, peri-urban, and agricultural habitats, camera position was discussed with landowners
to ensure compliance with any of their conditions, such as avoiding windows, high traffic
areas, or spaces frequented by winery guests. This discussion with landowners was critical to
ensure they were completely comfortable with the study, camera position, and how images
would be stored or removed (i.e. images of human faces or identifying features were
removed).

Appendix 14
Unmarked model abundance and detection equations.
Abundance is calculated using the following equation:
𝑁𝑖 ~ 𝑁𝑒𝑔𝑎𝑡𝑖𝑣𝑒 𝐵𝑖𝑛𝑜𝑚𝑖𝑎𝑙 (𝜆𝑖 , 𝛼)
Where 𝑁𝑖 is the abundance, 𝜆𝑖 is the mean local abundance of cats at site 𝑖, and 𝛼 is the
dispersion parameter. The second part of the equation models detection using the following
equation:
𝑦𝑖𝑗 |𝑁𝑖 ~ 𝐵𝑖𝑛𝑜𝑚𝑖𝑎𝑙 (𝑁𝑖 , 𝑝𝑗 )
Where 𝑦𝑖𝑗 is the detection or non-detection of an individual cat at site 𝑖 during the 𝑗th
occasion and 𝑝𝑗 is the detection probability of an individual cat during the 𝑗th occasion (Kéry
& Royle, 2016).

134
Appendix 15
Mapping Predicted Cat Local Abundance in the South Okanagan Valley
I created predictive maps of cat abundance across Okanagan Falls, Oliver, and Osoyoos using
a predicted local cat abundance grid. I mapped a grid over the south Okanagan Valley in
QGIS, with each grid being 200 x 200 metres. The grid was placed to focus on covering the
towns of Okanagan Falls, Oliver, Osoyoos, and the space between and around the towns
(Figure 2.1 outlines the gridded study area in red). The grid extended laterally to the
mountains on either side, maintaining the focus in the valley, and longitudinally from the top
of Okanagan Falls to the Canada/USA border. I extracted the habitat variables for each grid
cell using the intersect tool, the same as used for extracting the buffer habitat variables (see
GIS Analysis). After the habitat variables were obtained for each grid cell, I estimated the
local abundance of wandering cats within each grid cell using the seasonal abundance model
(without the detection covariate). I mapped each habitat variable within the seasonal
abundance model separately and set the remaining habitat variables in the model to their
mean. Maps showing predicted local cat abundance were created using QGIS by joining the
predicted local abundance value with its corresponding grid cell and creating a graduated
colour scheme, breaking the abundance estimates into 4 classes of equal intervals.

135
Appendix 16

Figure A16.1 Predicted abundance of wandering cats during the fall based on A) the
proportion of water in each cell and B) the length of roads in each cell.

136

Figure A16.2 Predicted abundance of wandering cats during the late winter based on A) the
proportion of fields in each cell and B) the number of buildings in each cell.

137
Appendix Literature Cited

Chandler, R. B., Royle, J. A., & King, D. I. (2011). Inference about density and temporary
emigration in unmarked populations. Ecology, 92(7), 1429–1435.
https://doi.org/10.1890/10-2433.1
Government of Canada, S. C. (2023). Profile table, Census Profile, 2021 Census of
Population. https://www12.statcan.gc.ca/census-recensement/2021/dppd/prof/index.cfm?Lang=E
Kéry, M., & Royle, J. A. (Eds.). (2016). Applied Hierarchical Modeling in Ecology.
Academic Press. https://doi.org/10.1016/B978-0-12-801378-6.04001-7
Legendre, P., & Legendre, L. (2012). Chapter 11—Canonical analysis. In P. Legendre & L.
Legendre (Eds.), Developments in Environmental Modelling (Vol. 24, pp. 625–710).
Elsevier. https://doi.org/10.1016/B978-0-444-53868-0.50011-3
Richardson, C. (2022, March 21). Researcher looking for South Okanagan properties to host
trail cameras on their properties to monitor cats and other wildlife. Castanet.
https://www.castanet.net/news/Penticton/363436/Researcher-looking-for-SouthOkanagan-properties-to-host-trail-cameras-on-their-properties-to-monitor-cats-and-otherwildlife
Thomas, L., Buckland, S. T., Rexstad, E. A., Laake, J. L., Strindberg, S., Hedley, S. L.,
Bishop, J. R., Marques, T. A., & Burnham, K. P. (2010). Distance software: Design and
analysis of distance sampling surveys for estimating population size. The Journal of
Applied Ecology, 47(1), 5–14. https://doi.org/10.1111/j.1365-2664.2009.01737.x
Wilson, O., & Skurikhina, A. (2022a, February 22). Cats and Birds Research in the South
Okanagan. Stewardship Centre for BC. https://stewardshipcentrebc.ca/cb-researchokanagan/
Wilson, O., & Skurikhina, A. (2022b, August 13). August research update from the
Okanagan Cat Count. Stewardship Centre for BC.
https://stewardshipcentrebc.ca/okcatcount_august/