The University of Maine

DigitalCommons@UMaine

Electronic Theses and Dissertations Fogler Library

12-2021

Forest Disturbance and Occupancy Patterns of Carnivores:

Results of a Large-scale Field Study in Maine, USA
Bryn E. Evans
University of Maine, [email protected]

Follow this and additional works at: https://digitalcommons.library.umaine.edu/etd

Part of the Environmental Monitoring Commons, Forest Biology Commons, Natural Resources and
Conservation Commons, Other Animal Sciences Commons, Other Ecology and Evolutionary Biology
Commons, Other Forestry and Forest Sciences Commons, Population Biology Commons, Research
Methods in Life Sciences Commons, and the Zoology Commons

Recommended Citation
Evans, Bryn E., "Forest Disturbance and Occupancy Patterns of Carnivores: Results of a Large-scale Field
Study in Maine, USA" (2021). Electronic Theses and Dissertations. 3512.
https://digitalcommons.library.umaine.edu/etd/3512

This Open-Access Thesis is brought to you for free and open access by DigitalCommons@UMaine. It has been
accepted for inclusion in Electronic Theses and Dissertations by an authorized administrator of
DigitalCommons@UMaine. For more information, please contact [email protected].
 FOREST DISTURBANCE AND OCCUPANCY PATTERNS OF CARNIVORES: RESULTS OF A LARGE-SCALE

FIELD STUDY IN MAINE, USA

By

Bryn Elen Hutchins Evans

B.S. University of California-Davis, 2009

M.S. University of Wisconsin-Madison, 2017

A DISSERTATION

Submitted in Partial Fulfillment of the

Requirements for the Degree of

Doctor of Philosophy

(in Wildlife Ecology)

The Graduate School

The University of Maine

December 2021

Advisory Committee:

Dr. Alessio Mortelliti, Associate Professor of Wildlife Habitat Conservation, Advisor

Dr. Walter Jakubas, Maine Inland Fisheries and Wildlife

Dr. Cynthia Loftin, Associate Professor of Wildlife Ecology

Dr. Daniel Hayes, Associate Professor of Geospatial Analysis & Remote Sensing

Dr. J. Pascal Berrill, Professor of Silviculture and Restoration

© 2021 Bryn Evans

All Rights Reserved

ii
 FOREST DISTURBANCE AND OCCUPANCY PATTERNS OF CARNIVORES: RESULTS OF A LARGE-SCALE

FIELD STUDY IN MAINE, USA

By Bryn Elen Hutchins Evans

Dissertation Advisor: Dr. Alessio Mortelliti

An Abstract of the Dissertation Presented

in Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy
(in Wildlife Ecology)
December 2021

Understanding trends in the abundance and distribution of carnivores is important at global,

regional and local scales due to their ecological role, their aesthetic and economic value, and the

numerous threats to their populations. Carnivores in Maine range from the American black bear (Ursus

americanus), to numerous native mesocarnivore species, such as American marten (Martes americana),

fisher (Pekania pennanti), coyote (Canis latrans), red fox (Vulpes vulpes), bobcat (Lynx rufus), Canada

lynx (Lynx canadensis) and to two small weasel species (Mustela erminea and Neogale frenata). Though

smaller than their apex carnivore cousins, Mesocarnivores are essential components of ecosystems and

have complex impacts on prey species and intraguild dynamics. However, these species can vary in how

they respond to human disturbances, from direct declines due to unregulated harvest and habitat loss,

and their ability to adapt to land-use change.

Maine is a working landscape which provides habitat for diverse wildlife species coincident with

extensive forest harvest industries, as well as tourism and recreation. The intensity, timing, and

configuration of harvest activities all interact to modify the landscape, with cascading impacts on the

distribution of many animals. Forest management practices have changed through time (Maine Forest

Service 2003) with potentially unpredictable outcomes (e.g. Simons 2009). However, the extent to which

carnivore species adapt to land use change is a key knowledge gap that needs to be addressed to ensure
proper management and conservation going forward. I investigated these patterns by designing a

natural experiment across the forested landscape of Maine, and by collecting detection data on multiple

species at camera trapping survey stations deployed along a gradient of forest disturbance. My

dissertation aims to collect broad-scale, relevant information for carnivore management and

conservation, and assess the efficacy of motion-triggered trail cameras for long-term monitoring.

My work is divided into four sections, reflected by the four chapters included in the dissertation.

My first goal was to determine the optimal number and configuration of camera-trap transects, to

balance between reasonable effort expended and high-quality data collection. I used multi-method

occupancy analyses to compare between one, two or three camera units spaced either 100 m or 150 m

apart. We found that a design with three cameras spaced 100 m apart increased detection probabilities

up to five-fold over a single camera trap, and thus used this configuration for the duration of the

following research.

Once the survey unit was selected, I established a large-scale, multi-year camera trapping

regimen across the northern two-thirds of Maine. Survey sites were selected in compliance with a

natural experimental design, replicating across all combinations of a) forest disturbance intensity, b)

latitude, and c) fur trapping harvest reports for key furbearing species. In the second chapter I present

this study design in more detail, and use the resulting data to investigate the interspecies dynamics of

marten and fisher, two species of interest to the state of Maine that co-exist in several geographic areas

and partition habitat in distinct ways. Both species are sensitive to habitat change resulting from timber

harvest, which was a more important factor in occupancy patterns than intraguild dynamics.

In chapter three, I took advantage of the large data set I collected to provide a landscape scale

understanding of long-tailed and short-tailed weasel distribution patterns in the face of habitat change.

Both of these species are poorly studied, and may be in decline in North American. My results indicate

that short-tailed weasel are widespread in Maine and do not appear limited by forest harvest practices,
while long-tailed weasel are rarer and more apt to be present in southern Maine. Finally in chapter four

I ran models incorporating multiple states for species occupancy, beyond mere present or absent, to

understand the dynamics of black bears and of black bear reproduction across managed forests in

Maine. I found that generally disturbance at a small scale was positively associated with both occupancy

and probability of reproduction, while the availability of hardwood trees (an important food source for

bears) was also positively linked to the probability of female bears being with cubs.

In addition to meeting our stake holder needs for informed management guidelines, I hope that

many of my findings will be directly relevant to the broader research community—as camera trapping

equipment becomes more affordable, it will become feasible to both monitor and rigorously study

wildlife populations in remote locations and under many scenarios of human land-use.
 DEDICATION

For my parents.

vi
 ACKNOWLEDGEMENTS

I am forever honored and humbled to have been a part of this much bigger and nobler effort

undertaken by so many excellent people. This project was brought about by hard work, perseverance,

and the sincere desire to understand and protect the wild creatures of Maine far before I arrived on the

scene to lend my small contribution. I thank everyone at the Maine Department of Inland Fisheries and

Wildlife that put their trust in this project, and to Cory Mosby, I tip my hat: I know your enthusiastic

support from the get-go helped take an improbable plan to an actual possibility, and I personally am

grateful for that. And for the infinite wisdom of fuzzy slippers.

To the many landowners and managers that so graciously granted access for this project, you

truly made it happen. It was a genuine pleasure for me to get to work with you, and to walk in the

woods that you care for! And Brian Roth and the CFRU, I am very thankful for the assistance you lent all

along the way. Shevenell Webb, it’s been a joy to work with you, let’s try to get out in the field again

someday!

To my committee: the last few years have been turbulent in so many ways, and I know each of

you has faced changes and challenges even beyond a global pandemic. Yet you've been there when I

needed you, and somehow guided me through the ordeal that was comprehensive exams, and I did not

succumb to anxiety as I was sure I would. Dan, it’s been an honor to get my grubby non-tech-savvy field-

biologist hands on the beautiful data layers that your students have produced, thank you for helping me

bumble through actually using them! And Pascal, I actually *enjoyed* your readings for comps, what an

unexpected treat. Cyndy, it has been an honor, and I so much value your time and insight. Keep leading

the way! Wally, I have barely scratched the surface of your broad knowledge on the glorious state of

Maine. Perhaps you’ll let me get you a cup of coffee when times are a little easier, and tell me all about

it.

vii
 Speaking of broad knowledge and know-how, Joe Z, Molly, Katherine, and Rena, I don’t know

what I would have done without you!

I'm sure many acknowledgments sections of dissertations express that it was only made possible

by the support of a good advisor. And it may well be true. In this case I know for an absolute,

indisputable fact that I would not have made it to this point without Alessio Mortelliti. By rejoicing

together in successes, navigating me through tough spots (sorry Squeezebox) and supporting me

through my struggles, I was never alone on a journey that I found daunting, sometimes beyond my

capabilities. I owe my triumphs to your uncompromising commitment to your students, and though I

have fallen short many times, you set an amazing example of hard work, integrity, and intelligence.

Though it won’t be in an academic realm, I hope to draw upon my time under your guidance to—when

the time comes to really buckle down and do the work—catch the kangaroo.

I would also like to offer my sincere thanks to folks that have guided, encouraged, and cajoled

me on my path to this point. To Max, Tavis, and David: the Mendo days forever shaped my priorities in

life. That is, showers, WiFi and flush toilets are infinitely less important than animal tracks to follow

along dusty roads, cold mountain streams to dive into after chasing telemetry beeps all day, and friends

to gather with around a campfire. To Shawn and Tim: I still think you over-estimated my academic

ability, but your support and faith gave me the courage for this latest leap, and I am glad I took it.

I also owe a huge dept of gratitude to my lab mates, department-mates, housemates, and so on

and so forth into the wonderful web of community that is Orono, Maine. The friends I've made here are

too numerous and too precious to give you all proper homage in such a short space. I'll do my very best

to pay all your kindnesses and support forward, instead. Be it a quick cup of coffee to vent next week, a

workout buddy next summer, or a couch to crash on sometime next decade, I will have your back as you

have had mine.

viii
 There are, however, two honorable mentions: Alli Brehm, my sistah from another mistah: I

would credit you with keeping me sane over the last five years. But the truth is, you kept me company

by being insane together, which is immeasurably better. And Sarah Vogel, my SarBear, my rock. Thank

you for not ever letting me give up. Or letting me starve.

And obviously, I wouldn’t be here at all without the unwavering, unending, unconditional love

and support of my family. Joe, I did not imagine I would find the missing piece of my heart in between

field sessions covered in skunk stink, but life works in strange ways. And I’m so grateful it does. But I

promise to put my bait clothes straight outside whenever I come home to you.

My love and gratitude for my parents, to whom I dedicate this dissertation, goes beyond words.

You guys are THE BEST.

ix
 TABLE OF CONTENTS

DEDICATION ................................................................................................................................................. VI

ACKNOWLEDGEMENTS ............................................................................................................................... VII

LIST OF TABLES ........................................................................................................................................... XIV

LIST OF FIGURES ......................................................................................................................................... XVI

CO-AUTHORSHIP STATEMENT .................................................................................................................. XVII

CHAPTER 1: ASSESSING ARRAYS OF MULTIPLE TRAIL CAMERAS TO DETECT

NORTH AMERICAN MAMMALS .................................................................................................................... 1

1.1 Chapter abstract ........................................................................................................................ 1

1.2 Introduction ............................................................................................................................... 2

1.2.1 Study area ....................................................................................................................... 5

1.3 Methods ..................................................................................................................................... 6

1.3.1 Field methods ................................................................................................................. 6

1.3.2 Analytical methods ......................................................................................................... 9

1.4 Results ...................................................................................................................................... 14

1.5 Discussion................................................................................................................................. 20

1.5.1 Camera trap number and spacing ................................................................................ 21

1.5.2 Landscape and camera site features ............................................................................ 23

1.5.3 Camera trapping survey design .................................................................................... 24

1.6 Acknowledgments.................................................................................................................... 26

CHAPTER 2: MARTEN, FISHER, AND FOREST DISTURBANCE IN MAINE ...................................................... 27

2.1 Chapter abstract ...................................................................................................................... 27

2.2 Introduction ............................................................................................................................. 28

2.3 Methods ................................................................................................................................... 31

x
 2.3.1 Study area ..................................................................................................................... 31

2.3.2 Study design ................................................................................................................. 33

2.3.3 Occupancy models........................................................................................................ 35

2.4 Results ...................................................................................................................................... 37

2.5 Discussion................................................................................................................................. 45

2.5.1 Effects of forest disturbance on marten and fisher occupancy patterns ..................... 45

2.5.2 Effects of snow depth on marten and fisher occupancy patterns ............................... 46

2.5.3 Intraguild effects on occupancy patterns ..................................................................... 46

2.5.4 Effects of forest composition on occupancy patterns of marten and fisher................ 47

2.5.5 Conclusions ................................................................................................................... 48

2.6 Acknowledgements.................................................................................................................. 49

CHAPTER 3: FOREST DISTURBANCE AND OCCUPANCY PATTERNS OF SHORT-TAILED WEASEL

AND LONG-TAILED WEASEL IN MAINE ....................................................................................................... 50

3.1 Chapter abstract ...................................................................................................................... 50

3.2 Introduction ............................................................................................................................. 51

3.2.1 Species biology ............................................................................................................. 53

3.2.2 Study area ..................................................................................................................... 54

3.3 Methods ................................................................................................................................... 56

3.3.1 Data Collection ............................................................................................................. 56

3.3.2 Image Analysis .............................................................................................................. 57

3.3.3 Occupancy Models ....................................................................................................... 59

3.4 Results ...................................................................................................................................... 60

3.4.1 Image Data.................................................................................................................... 60

3.4.2 Occupancy Analyses ..................................................................................................... 61

xi
 3.5 Discussion................................................................................................................................. 66

3.5.1 Habitat relationships .................................................................................................... 66

3.5.2 Large-scale distribution and co-existence patterns ..................................................... 67

3.5.3 Detectability of weasels ............................................................................................... 68

3.5.4 Management Implications............................................................................................ 69

3.6 Acknowledgements.................................................................................................................. 70

CHAPTER 4: AMERICAN BLACK BEAR OCCUPANCY AND REPRODUCTION IN MANAGED

TIMBER FORESTS OF MAINE ....................................................................................................................... 71

4.1 Chapter abstract ...................................................................................................................... 71

4.2 Introduction ............................................................................................................................. 71

4.3 Methods ................................................................................................................................... 74

4.3.1 Study area ..................................................................................................................... 74

4.3.2 Study design and data collection.................................................................................. 77

4.3.3 Analyses ........................................................................................................................ 78

4.4 Results ...................................................................................................................................... 80

4.4.1 Occupancy models........................................................................................................ 80

4.5 Discussion................................................................................................................................. 85

4.5.1 Management implications ............................................................................................ 87

REFERENCES ................................................................................................................................................ 88

APPENDIX A: ASSOCIATED PUBLICATION ABSTRACTS .............................................................................. 106

APPENDIX B: SUPPORTING INFORMATION FOR CAMERA ARRAY SELECTION (CHAPTER 1) .................... 108

Appendix B.1: Comparison between 24-hour and 1-week detection windows .......................... 108

Appendix B.2: Results of phase 1 modeling................................................................................. 111

Appendix B.3: Model comparison between different possible pairs .......................................... 112

xii
APPENDIX C: FOREST DISTURBANCE INDEX .............................................................................................. 115

APPENDIX D: SUPPORTING INFORMATION FOR LARGE-SCALE CAMERA SURVEYS (CHAPTERS 2-4) ....... 115

Appendix D.1: Study design for large-scale camera trapping surveys......................................... 117

Appendix D.2: Two-species models for marten and fisher interaction ....................................... 122

Appendix D.3: Additional results of marten and fisher two-species models .............................. 125

Appendix D.4: Weasel image tagging protocol........................................................................ 12930

Appendix D.5: Full model ranking for weasel multi-season false-positive occupancy models ... 132

Appendix D.6: Correlograms indicating no spatial-autocorrelation for bear models ................. 136

BIOGRAPHY...............................................................................................................................................138

xiii
 LIST OF TABLES

Table 1.1. Definitions of terms used in study design and modeling to assess the influence of

different numbers and spacing of camera traps on the detection of mammals ............................ 13

Table 1.2. Top-ranked occupancy models for six species surveyed with camera trap transects

in Maine, USA, 2017 ........................................................................................................................ 15

Table 1.3. Parameter estimates for six mammal species in Maine ................................................ 18

Table 1.4. Influence of camera site features on detection probabilities for six mammal

species in Maine, USA..................................................................................................................... 20

Table 2.1. Ranking of single species, multi-season models for marten and fisher ........................ 39

Table 2.2. Ranking of two-species, multi-season models for marten and fisher ........................... 41

Table 3.1. The top ranking AICc models for weasel occupancy in Maine ...................................... 62

Table 4.1. Survey effort across four summers in Maine, USA. ....................................................... 78

Table 4.2. Parameters of dynamic multi-state occupancy models. ............................................... 79

Table 4.3. AIC model ranking for multi-state occupancy models of American black bears ........... 81

Table 4.4. Parameter estimates for multi-state occupancy model of American black bears ........ 82

Table 4.5. Detection probabilities for American black bears in Maine .......................................... 84

Table B.1. Comparisons by species for 24-hour temporal binning versus 1-week binning for

detection histories. ....................................................................................................................... 110

Table B.2. Model ranking comparing the inclusion of the “method” covariate for detection

probability versus a null detection model. ................................................................................... 111

Table B.3. Comparison of model ranking output between the “original” pairing and the

“alternative” pairing configurations possible from out T-array of five camera traps. ................. 113

Table B.4. Comparison of beta estimates between the “original” pairing and the

“alternative” pairing configurations possible from out T-array of five camera traps. ................. 114

xiv
Table D.1. Parameterization of American marten occupancy models used to compare the

relative importance of interspecific dynamics from fisher presence ........................................... 123

Table D.2. Parameterization of fisher occupancy models used to compare the relative

importance of interspecific dynamics from American marten presence ..................................... 124

Table D.3. Survey effort over seven seasons of camera trap transects ....................................... 125

Table D.4. Naive occupancy rates for marten and fisher ............................................................. 125

Table D.5. Parameter estimates and standard errors for single-species, multi-season

occupancy models of marten and fisher....................................................................................... 126

Table D.6. Short-tailed weasel (Mustela erminea) model ranking. .............................................. 132

Table D.7. Long-tailed weasel (Neogale frenata) model ranking. ................................................ 134

xv
 LIST OF FIGURES

Fig 1.1. Three study areas surveyed with multiple camera traps in Maine, USA ............................. 7

Fig 1.2. A survey site composed of five camera trap units in a T-configuration to detect

terrestrial mammals in Maine, USA.. ................................................................................................ 8

Fig 1.3. Model-averaged estimated detection probabilities and standard errors for surveys

conducted with differing combinations of camera traps ............................................................... 17

Fig 2.1. Map of survey stations (n=197) to detect marten and fisher in Maine, USA .................... 32

Fig 2.2. Predictions from top ranking multi-season models for marten and fisher........................ 43

Fig 3.1. Map of Maine and 197 camera trap survey site locations and design .............................. 55

Fig 3.3. Results of multi-season, false-positive occupancy modeling for short-tailed weasels ...... 63

Fig 3.4. Results of multi-season, false-positive occupancy modeling for long-tailed weasels ....... 65

Fig 4.1. Survey sites and number of summers deployed to detect black bears in Maine .............. 76

Fig 4.2. Occupancy probabilities from multi-state, multi-season models of black bears............... 83

Fig. 4.3. Detection probabilities for American black bears in Maine. ............................................ 84

Fig B.1. Detection probabilities comparing between 24-hour and 1-week temporal bins .......... 109

Fig D.1. Boxplots illustrating the independence of predictor variables ....................................... 117

Fig D.2. Forest disturbance categories at the township level across Maine, USA ........................ 118

Fig D.3. Reported hunter harvest success over the 2014 to 2018 seasons .................................. 119

Fig D.4. Correlation coefficients between predictor variables for occupancy models................. 120

Fig D.5. Example configuration of our survey transects ............................................................... 121

Fig D.6. Example images of American marten and fisher at the same camera sites.................... 127

Fig D.7. Map of detections of marten and fisher from camera trapping in Maine ...................... 128

Fig D.8. Example of the SmallMeasure tool to provide distances between image pixels ............ 130

Fig D.9. Example of weasel species idenfitication protocol.......................................................... 131

xvi
 CO-AUTHORSHIP STATEMENT

Bryn Evans has been the primary researcher for the design, collection, and analyses of the data included

in this dissertation. The scope and scale of this project would not have been possible without extensive

support and collaboration from both her academic advisor, Dr. Alessio Mortelliti, as well as innumerable

individuals acting in many capacities. While many of these people have been mentioned in the

acknowledgements, several people have contributed at the level of co-authorship on manuscripts

resulting from the work reported here. For each of the following chapters, the author of the dissertation

is the lead author, but we ask that readers please acknowledge the contributions of coauthors as

described below.

Chapter 1, Assessing arrays of multiple trail cameras to detect North American mammals, was

coauthored by Mr. Cory Mosby (Maine Department of Inland Fisheries and Wildlife, Bangor, Maine) and

Dr. Alessio Mortelliti (University of Maine, Department of Wildlife Fisheries and Conservation Biology)

and has been published: PLoS ONE 14(6):e0217543. https://doi.org/10.1371/journal.pone.0217543,

Received September 13, 2018, Accepted May 14, 2019, Published June 17, 2019.

Chapter 2, Effects of forest disturbance, snow depth and intraguild dynamics on American

marten and fisher occupancy in Maine, USA, was coauthored with Dr. Alessio Mortelliti, and has been

accepted for publication: Ecosphere, Submitted November 24, 2021, accepted December 7, 2021.

Chapter 3, which has been titled Forest disturbance and occupancy patterns of short-tailed

weasel (Mustela erminea) and long-tailed weasel (Neogale frenata): results from a large-scale natural

experiment in Maine, USA for publication, was coauthored by Dr. Alessio Mortelliti. This manuscript has

been submitted for publication and is currently seeking reviewers: Journal of Mammalogy, Submitted

December 3, 2021, seeking reviewers as of December 12, 2021.

Chapter 4, American black bear occupancy and reproduction in managed timber forests of

Maine, has been authored in collaboration with Dr. Alessio Mortelliti, Amay Bolinjkar (MWC 2021,

xvii
Wildlife Conservation, University of Maine), and Griffin Archambault (BSc 2020, Wildlife Fisheries and

Conservation Biology, University of Maine) and is being prepared for submission to The Journal of

Wildlife Management.

xviii
 CHAPTER 1: ASSESSING ARRAYS OF MULTIPLE TRAIL CAMERAS TO DETECT

NORTH AMERICAN MAMMALS

1.1 Chapter abstract 1

Motion triggered camera traps are an increasingly popular tool for wildlife research and can be used to

survey for multiple species simultaneously. As with all survey techniques, it is crucial to conduct camera

trapping research following study designs that include adequate spatial and temporal replication, and

sufficient probability of detecting species presence. The use and configuration of multiple camera traps

within a single survey site are understudied considerations that could have a substantial impact on

detection probability. Our objective was to test the role that camera number (one, two or three units),

and spacing along a linear transect (100 m or 150 m), have on the probability of detecting a species

given it is present. From January to March, 2017 we collected data on six mammal species in Maine,

USA: coyote (Canis latrans), fisher (Pekania pennanti), American marten (Martes americana), short-

tailed weasel (Mustela erminea), snowshoe hare (Lepus americanus), and American red squirrel

(Tamiasciurus hudsonicus). We used multi-scale occupancy modelling to compare pooled detection

histories of different configuration of five cameras deployed at the same survey site (n = 32), and how

the configuration would influence the probability of detecting a species given it was available at the site.

Across all six species, we found substantial increases in probability of detection as the number of

cameras increased from one to two (22 to 400 percent increase), regardless of the spacing between

cameras. For most species the magnitude of the increase was less substantial when adding a third

camera (4 to 85 percent increase), with coyote and snowshoe hare showing a pronounced effect. The

Note on citation:1
This chapter has been published in a peer reviewed journal, and any readers wishing to cite our findings
should do so as follows: Evans, B. E., C. E. Mosby, and A. Mortelliti. 2019. Assessing arrays of multiple
trail cameras to detect North American mammals. PLoS ONE 14(6):e0217543. To receive a copy of the
journal publication at no charge, please visit: https://www.researchgate.net/profile/Bryn-Evans-2 or
email: [email protected].

1
influence of survey station features also varied by species. We suggest that using pooled data from two

or three cameras at a survey site is a cost-effective approach to increase detection success over a single

camera.

1.2 Introduction

Motion-triggered camera traps are an increasingly popular tool for wildlife research (Rovero et al. 2013,

Burton et al. 2015). Early camera trapping studies initially built upon existing capture-mark-recapture

analysis-framework (Karanth 1995), but the approach has since expanded to study unmarked individuals

ranging from the species to the community level (Kays et al. 2011, O’Connell et al. 2011). Camera

trapping studies have been proposed as a tool for large-scale, long-term monitoring (Steenweg et al.

2016), and are an important development for terrestrial wildlife research (Rowcliffe and Carbone 2008,

Rowcliffe 2017). Data collected by camera trapping can be analyzed in an occupancy modeling

framework (MacKenzie et al. 2017), where consecutive days or weeks of data collection create a

detection history of animal visits but do not require the researchers to be present at all occasions.

Originally designed for distinct patches of habitat, occupancy modeling has been expanded to include

use in continuous habitat, where the activity patterns of animals that bring them into the detection

range of a remote sensor correspond to presence in the area (Efford and Dawson 2012, Neilson et al.

2018, Stewart et al. 2018). Occupancy modeling is a flexible, statistically robust approach which

accounts for biases in parameter estimates caused by false-absences or temporary unavailability

(MacKenzie et al. 2002). The approach offers further biological insights by modeling covariates

(MacKenzie et al. 2017) and provides ecologically relevant information to researchers and managers

(Royle and Dorazio 2008).

Study design is a fundamental step in any wildlife research project. Factors to consider include

selecting an appropriate scale and study sites that meet the goals of the study, adequate spacing

between sites for independence, and the trade-off between effort invested at every survey site (often

2
the number of visits to a site, or duration of camera deployment) and the total number of sites (Guillera-

Arroita et al. 2010, Guillera-Arroita and Lahoz-Monfort 2012). When considering the study objective,

there are also trade-offs between addressing specific questions regarding one, or a few, focal species

versus collecting broader data at a community level (Wintle et al. 2010, Noon et al. 2012). An approach

which may maximize efficacy for some species may not be optimal for all, and to implement an effective

study design information on the actual impact of these trade-offs is essential.

The specific aspect of study design we investigate below is the use of multiple camera traps,

placed at unique microsites but within in a single survey site, to increase the cumulative detection

probability of target species. While there has been research to address the trade-off between total

number of camera survey sites in a study area and deployment time at each (Shannon et al. 2014),

relatively fewer studies have assessed the impact of increasing the number of cameras contributing data

to a single survey site. Increasing the number of cameras, and the heterogeneity of the available habitat

surveyed by placing them in distinct microsites, should increase detection probability. Although it is not

the goal of most research studies itself, having a reasonable chance of successfully detecting species

when they are present is essential to avoid biased results (Guillera-Arroita et al. 2010, Guillera-Arroita

and Lahoz-Monfort 2012). This is especially important for species that occur at low density, such as

carnivores (Boitani and Powell 2012). Recently, O’Connor and colleagues (O’Connor et al. 2017) assessed

the benefits of increasing number of cameras in randomized clusters for four unmarked North American

mammal species, and did find a dramatic gain in detection probability when more than one camera was

used. Similarly, researchers in Australia found an increase in detection for feral cats when increasing the

number of cameras in a cluster up to five (Stokeld et al. 2015). These previous studies, however, used

grouped or random arrangements. A study design using a consistent arrangement and spacing of

cameras at each study site would reduce variation between sites but allow each site to include different

small-scale landscape features, and may increase the likelihood of detecting organisms with clustered

3
distributions (Elzinga et al. 1998) or activity patterns that are not perfectly uniform across the landscape.

This movement behavior is expected for animals seeking optimal foraging opportunities, avoiding

predation, and moving throughout a heterogeneous home range (Feldhamer et al. 2003, Elbroch and

Rinehart 2011, Turcotte and Desrochers 2003, King and Powell 2007, Moriarty et al. 2015). Our aim was

to help fill this knowledge gap by comparing the detection success (defined as a higher likelihood of

observing the species if it is available in the area, or conversely lower chances of a false absence)

between camera trap survey methods that differed in the number of cameras deployed at one, two or

three separate microsites within each survey site and the spacing between them. This differs from the

use of multiple cameras at one single location to assess missed detections (Pease et al. 2016)[29] or to

assist individual identification of animals with unique markings (such as [Karanth and Nichols 1998]).

Our study used multi-scale occupancy models to compare detection probabilities between

survey methods that varied in the number (one, two or three) and spacing (100 m or 150 m for transects

of two- and three-cameras) of cameras set in a “T” arrangement at a single survey site. This approach is

a subset of occupancy modeling designed to compare efficacy of different data-collection techniques

while accounting for non-independence of co-located devices or detection methods (Nichols et al.

2008). Though prior research has implemented a multi-scale approach using camera trap data, these

studies typically compare cameras to other survey techniques (such as track plates, snow tracking or

genetic capture-recapture [Fisher et al. 2016, Clare et al. 2017]). Our objectives were 1) to use a multi-

scale approach to determine the most effective arrangement of camera traps in terms of both spacing

and number of cameras deployed to increase detection probability and 2) to exploit data from multiple

species collected at the same sites to illustrate how the optimal study designs may vary between

species. Our primary goal is to investigate, on behalf of future research, how much is gained in increased

detection probability at a single survey site by using multiple cameras. By placing cameras in close

proximity but at unique microsites to attract or opportunistically detect wild mammals, and by pooling

4
the resulting detecting histories, we predict that the risk of false-absences is decreased and therefore

stronger inferences can be drawn.

We analyzed the detection histories of six terrestrial mammal species: coyote (Canis latrans),

fisher (Pekania pennanti), American marten (Martes americana), short-tailed weasel (Mustela erminea),

snowshoe hare (Lepus americanus), and American red squirrel (Tamiasciurus hudsonicus). These species

span across a range of trophic levels and body sizes, which impact movement patterns and space use

and therefore the detection probability during camera surveys. The carnivore species represent both the

cursorial hunting strategy of coyotes and the more opportunistic, exploratory mustelids (King and

Powell 2007, Elbroch and Rinehart 2011) which could influence the likelihood of detection. Even with a

baited site, a broader camera spacing might increase the chances that at least one camera fell along the

movement path of a coyote, where a denser spacing might increase the overall attractant and draw

mustelids closer as they investigate their surroundings. Similarly snowshoe hare are likely to cover

greater distances than red squirrels, and thus a longer spacing might increase detection success if this

were a target species. By incorporating site covariates that were likely to play a role in animal occupancy

and detection, we began to assess how the transect design may have differing impact on the ability of a

given survey to detect a target species if it is present.

1.2.1 Study area

Our study took place in central and northern Maine at the transition zone of the Eastern Deciduous

Forest and the Boreal Forest (Gawler et al. 1996, McWilliams et al. 2005) (Fig 1.1). We surveyed in three

study areas: Scraggly Lake (100 km2 centered at 46.2455° N, -68.7348° W), Telos Road and the

Nahmakanta Public Reserve Lands (430 km2, 45.9163° N, -69.1564° W) and northeastern Moosehead

Lake (340 km2, 45.7073° N, -69.6910° W). These three areas are spaced a minimum of 30 km apart, to

capture potential geographic variation in animal distribution patterns over north-central Maine. The

region receives precipitation year round (mean annual precipitation 112 cm) and experience mild

5
summers and cold winters (mean summer temperature 18° C, winter -8° C) (NOAA 2019). All three study

areas are heavily forested and are comprised of a mosaic of different age forest stands resulting from

recent timber harvest, renewing stands and preserved areas. Coniferous trees are predominant, and

include balsam fir (Abies balsamea), spruces (Picea spp.), and white pine (Pinus strobus), while mixed or

fully hardwood stands include white and yellow birches (Betula papyrifera and B. alleghaniensis), red

and sugar maples (Acer rubrum and A. saccharum), and American beech (Fagus grandifolia).

1.3 Methods

1.3.1 Field methods

We deployed camera trap survey sites across a range of forest stand ages and management regimes in

northern Maine to detect terrestrial mammal species during winter 2017. We used randomized

locations, generated prior to conducting field work, to avoid bias in habitat sampled. We traveled as

close as possible to these locations by either truck or snowmobile before hiking off of the motorized trail

to set cameras. Each site (sensu MacKenzie et al. 2002) was sampled with a detection array (sensu

Nichols et al. 2008) which consisted of five cameras arranged in a T-configuration (Fig 1.2). This

configuration allowed us to test how detection probability was impacted by different arrangements of

cameras, placed at nearby microsites and thus dependent in terms of habitat characteristics and

underlying species occupancy status (Karanth and Nichols 1998). We consider our sampling unit to be

each site with five camera and spaced these to be independent of one another in terms of overall

species use of the surrounding habitat (goal distance of 5 k between nearest sites, mean 5.1, range 1.6 –

12.7).

6
Fig 1.1. Three study areas surveyed with multiple camera traps in Maine, USA. 12 sites each composed
of five camera traps were deployed in the Scraggly Lake (SL) area during January 2017, 12 sites in the
Telos Road/Nahmakanta Public Reserve Lands (TN) from February to March 2017, and 8 sites in the
Moosehead Lake (ML) area from late February to March 2017.

7
Fig 1.2. A survey site composed of five camera trap units in a T-configuration to detect terrestrial
mammals in Maine, USA. Detection histories from individual cameras were pooled in five different
combinations to assess the impact of survey methods that varied by number of cameras (1-3) and
spacing (short, 100 m, or long, 150 m between cameras): 1 unit (camera A), 2 short (cameras B and C), 3
short (A, B, and C), 2 long (D and E) 3 long (B, D, and E).

Three cameras were spaced in a straight line 100 m apart, and two cameras were spaced 150 m

perpendicular from the central camera. Each camera microsite consisted of one Bushnell Trophy Cam

Essential 2 passive infrared (PIR) triggered camera (Overland Park, KS, USA) attached to a tree at

approximately 1 m above snow level. The site was baited with a portion of beaver (Castor canadensis)

meat enclosed in a suet cage wired to a tree 3-4 m in front of the trail camera and scent lure designed to

attract furbearers (skunk essence and Vaseline based, Kenduskeag, ME, USA). Cameras were programed

to take a single image for every PIR trigger event, and to record time-lapse images at 03:00 and 15:00 to

capture weather events impacting performance. The microsite and the placement of cameras were

chosen to optimize unimpeded access to, and visibility of, the bait. We sought out natural clearings and

8
likely movement corridors to increase the probability of animals visiting a site. We deployed camera trap

arrays for two to three weeks from January to March, 2017. We did not rebait the sites, instead using

the suet cage to prevent meat bait from being consumed, and revisited to retrieve the cameras and

reset in the next study site. The techniques implemented in this non-invasive wildlife research study

were presented to the University of Maine Institutional Animal Care and Use Committee and were

approved under Protocol A2018-05-06.

1.3.2 Analytical methods

Detection histories were created for six species of terrestrial mammals at each camera unit. All images

recorded in a single day were collapsed into a daily detection result of 0 (not detected) or 1 (detected at

that camera). Daily detections were used following similar, relatively short-term camera trapping studies

(Mortelliti and Boitani 2008, Shannon et al. 2014). Although many arbitrary cut-offs have been used in

the camera trapping literature (Burton et al. 2015) a 24-hour period captures natural movement

patterns of activity for wildlife species. Non-independence between time units may occur regardless of

binning, and our data did not indicate a strong trend at 24 hours. This approach also guaranteed enough

revisits over the two- to three-week study duration to avoid models without solutions (Guillera-Arroita

et al. 2010, Guillera-Arroita and Lahoz-Monfort 2012), and when we tested for an impact of 24-hours

versus 1-week bins, trends were the same for detection results but model fit was poorer in the collapsed

dataset (see Appendix B, Fig B.1 and Table B.1). Any days in which a camera was not active or was

blocked by snow were considered ‘missing visits’ (sensu MacKenzie et al. 2002).

We generated pooled detection histories for different combinations of camera numbers and

transect lengths, here after referred to as methods, as follows: method 1 = 1 camera, method 2 = 2

cameras short (100 m between cameras), method 3 = 3 cameras short, method 4 = 2 cameras long (150

m between cameras) and method 5 = 3 cameras long (Fig 1.2). We pooled the individual camera

detection histories so that for each method 0 indicates no detections and 1 indicates a detection at any

9
one or more camera in that method. This allowed us to investigate the impact that adding cameras to a

survey configuration would have on the probability of detecting a species. One caveat is that our

methods using all three cameras for each potential spacing encapsulate the detection histories for the

two-camera method, and for the short array, the single camera method. This results in mathematical

dependency between our methods, because for methods with three cameras to result in a detection,

that event is also captured by at least one other. However, this reflects the application of the

techniques, and because the magnitude of difference in detection success going from one or two

cameras to three can only be under-estimated we feel our conclusions are conservative and consistent

with the underlying detection process.

We fitted multi-scale occupancy models (Nichols et al. 2008) to detection history data for each

of the six species considered in this study. The parameters of the multi-scale model are: Ѱi, the large-

scale probability that a member of the species i occupies a home range and uses the landscape

surrounding the survey site; ϴit, the small-scale probability of a species being in the immediate area of

the detection methods on survey occasion t given that it occupies the overall area; and pit, the

probability of detection on occasion t given that both the overall area is occupied and the species is in

the immediate area of the survey devices. In our study, Ѱ is the overall occupancy surrounding all five

cameras comprising the survey site, ϴ is the daily probability that a species is within detection range of

the cameras, and p is modeled such that each of the different methods of pooling cameras can have a

different probability of detecting a species. In the example below, we show the formula for estimating

these parameters for two days of a detection history, allowing the method type to influence detection.

We have L = 5 methods, over a two-day detection history of H = 01100 00000. This indicates that on day

one, detection methods two and three recorded the species, and on day two, no methods recorded an

image:

10
 Pr(00101 00000) =

Ѱ [θ1 (1 − 𝑝11 )(𝑝12 )(𝑝13 )(1 − 𝑝14 )(1 − 𝑝15 ) [(1 − θ2 ) + θ2 ∏(1 − 𝑝2𝑠 )]]
𝑠=1

On day one, we know that the species was available for detection because methods two and

three were successful, and can therefore write the other three methods as one minus the probability of

detection (known “false-absences”). On day two, no methods detected the species, thus the probability

is modeled as the sum of two potential causes: the species was not in the immediate vicinity of the

survey sites or all five methods s (sensu Nichols et al. 2008) failed to detect the species. Because at each

survey site overall, the probability of large-scale occupancy and small-scale, daily availability is shared,

the variation in detection success depends only on each method and any method specific covariates.

Models were fitted through Program PRESENCE (Hines 2017) and were ranked based on the Akaike

Information Criteria corrected for small sample sizes (AICc).

We followed a three-stage approach to fit occupancy models. In the initial stage we compared

two models that differed only by the inclusion or exclusion of the method type on the detection process.

In the second stage we focused on additional covariates affecting detection probability. We included the

following covariates (Table 1.1) which are likely to impact the overall occupancy of a species at a survey

site or the detection process within an array: the study area (Area) in which the array was deployed

(Scraggly Lake, SL; Telos Road and Nahmakanta Public Reserve Lands, TN; Moosehead Lake, ML) to

control for differences due to latitude and longitude and any other unmeasured variables beyond the

scope of our camera surveys. The access type (Access) was included as a surrogate for the influence of

transportation infrastructures which has been demonstrated to influence detection (Kowalski et al.

2015, Rich et al. 2016, Sirén et al. 2017) and was defined as either road (RD) if the array was located off

of a plowed road or snowmobile (SM) if it was accessible only by snowmobile. We also included

11
variables for the distance to the access feature, either road or trail, of the 1-3 camera units within each

method. Distance values (DistAve, DistMax, DistMin) were standardized by z-scoring (Zuur et al. 2007).

Following Burnham and Anderson (2002) we considered variables included in the top ranked model and

any models within 2 ∆AICc (corrected for our small sample size of 32 independent stations) to be

important, and tested additive models for these. We retained one top model for the detection process

for each species to use in the third and final stages.

In the third stage we focused on covariates affecting occupancy probability at the large scale of

animal use (Ѱ) and the small scale immediately around the site (ϴ). We included the Area categorical

variable as above, and the degree of tree loss (Loss) to quantify forest harvest events. This was obtained

from the Global Forest Watch project cumulative raster data which indicates 30m pixels that over any

one year period from 2000 to 2015 experienced a stand-replacing loss of tree canopy cover (Hansen et

al. 2013). Because wildlife species span a wide range of body sizes, and therefore home range sizes, we

assessed tree loss at multiple spatial scales by using four circular buffer sizes: 100 m, 500 m, 1 km and 5

km. We used ArcGIS 10.4.1 (ESRI 2016) to generate buffers around each camera site and then merged

the individual camera buffers into polygons to provide information about landscape features at each

scale. Pixels categorized as open water (www.maine.gov/megis) were excluded from analyses, and the

proportion of remaining pixels which experienced tree loss within each buffer size was standardized

using z-scoring. Following Fahrig (2013) and Sozio and Mortelliti (2016) we compared the relative fit of

models with different buffer size and retained only the highest ranking buffer size in the final stage.

Lastly we modeled each variable for Ѱ only, for Ѱ and ϴ as a single effect, and for Ѱ and ϴ as separate

effects.

12
Table 1.1. Definitions of terms used in study design and modeling to assess the influence of
different numbers and spacing of camera traps on the detection of mammals in Maine, USA.

Study design

Survey site An arrangement of five baited cameras traps as shown in Fig 2. Each survey site is
considered independent in terms of species occupancy status.

Method Arrangement of 1, 2, or 3 camera stations spaced either 100 m (short) or 150 m

(long) apart. Five detection methods compared are method 1 = 1 camera, method 2
= 2 cameras short, method 3 = 3 cameras short, method 4 = 2 cameras long and
method 5 = 3 cameras long.

Camera One Bushnell HD motion-triggered camera trap, placed facing bait and lure for two
to three weeks.

Model parameters

Psi (Ѱ) The probability of occupancy, here defined as at least one individual of a species
being active in the habitat surrounding a station over the course of the survey (large
scale availability for detection).

Theta (ϴ) The probability that at least one individual of a species is using the area immediately
around the survey site on a given survey occasion (small scale availability for
detection).

Detection (p) The probability of detecting a species given that it occupies the habitat and is
present in the vicinity of the survey site (available at both the large and small scale).
This probability will vary with by survey method, ps.

Model covariates

Access The infrastructure type used to reach each survey site: plowed road (RD), or
snowmobile trail (SM).

Area Three different study areas: Scraggly Lake (SL), Telos Road and Nahmakanta Public
Reserve Lands (TN), and Moosehead Lake (ML).

Distance The distance from an access feature to each camera in a method. For methods with
more than one camera, DistAve is the mean distance, DistMax is the distance of the
farthest camera and DistMin is the distance of the closest camera.

Loss Timber harvest activity generated from raster data on decreases in tree height from
2000 to 2015. Loss100m, Loss500m, Loss1k and Loss5k are tree loss at increasing
spatial scales to match different species body and home range sizes.

13
 We checked for spatial autocorrelation between survey sites in the residuals of simple, single

season occupancy models in Program R, using package ‘unmarked’ (Fiske and Chandler 2011). We used

the most parameterized model for detection data at the full array to create spline correlograms in

package ncf (Bjornstad 2018) and found no evidence of spatial dependence for any species. To the best

of our knowledge there is no established method to assess model fit within the residual structure of

multi-scale analyses, thus we estimated c-hat values for the most parameterized single season model

(Nichols et al. 2008, Stokeld et al. 2015) and did not observe over-dispersion.

1.4 Results

We deployed 32 arrays of five cameras each for two to three weeks (mean 17.2 days, range 16-23) from

12 January to 19 March, 2017. Three cameras failed to record images, and of the 157 cameras included

in analyses three had trap nights treated as missing visits due to heavy snowfall or battery failure. We

recorded 2,748 total trap-nights of data and captured 8,826 PIR-triggered images. American marten

were the most frequently detected species (at n = 71 cameras, and n = 172 trap nights with detections),

followed by fisher (n = 47 cameras, n = 92 trap nights) snowshoe hare (n = 40 cameras, n = 134 trap

nights), short-tailed weasel (n = 36 cameras, n = 108 trap nights), coyote (n = 28 cameras, n = 34 trap

nights) and red squirrel (n = 22 cameras, n = 49 trap nights).

In the first stage of multi-scale occupancy modeling, the inclusion of the method type in the

detection process always performed better than the null model (∆AICc > 8 for all species, see Appendix

B Table B.2), confirming that the number and spacing of cameras had an important effect on the

likelihood of detecting all species. In the second and third stages, the important variables included in all

models within ∆AICc > 2 of the top model for detection and occupancy parameters differed between

species (Table 1.2).

14
Table 1.2. Top-ranked occupancy models for six species surveyed with camera trap transects in Maine,
USA, 2017.

Species Top model(s) ∆AICc AICw

Canis latrans Ѱ(Loss1k), ϴ(.), (Method+Access) 0 0.34

Ѱ(.), ϴ(.), (Method+Access) 0.02 0.33

Pekania pennanti Ѱ(Loss5k), ϴ(Loss5k), p(Method+Access+DistAve) 0 0.54

Ѱ=ϴ(Area), p(Method+Access+DistAve) 1.58 0.27

Martes americana Ѱ(.), ϴ(.), p(Method+Area+DistMin) 0 0.59

Mustela erminea Ѱ(Loss100m), ϴ(Loss100m), p(Method+Area) 0 0.99

Lepus americanus Ѱ(Loss500m), ϴ(.), p(Method+Area+Access) 0 0.38

Ѱ(.), ϴ(.), p(Method+Area+Access) 0.43 0.30

Tamiasciurus hudsonicus Ѱ(Area), ϴ(Area), p(Method+Area) 0 0.43

The models shown are the top ranked model using Akaike’s information criterion corrected for small

samples and any models within ΔAICc < 2. AICw is the model weight. Ѱ is the probability of occupancy

across all stations, ϴ is the daily probability of availability for detection, and p is the probability of

detection given availability.

For all six species, increasing the number of cameras in a method had a strong positive effect on

the detection probability (Fig 1.3, Table 1.3). The magnitude of the effect was greatest when increasing

from one single camera to two cameras (ranging from 40% increase in detection probability for fisher to

400% for coyote). An increase in detection success was still evident when increasing from two cameras

to three (7% to 85%, again for fisher and coyote). The impact of the transect length varied by species,

with American marten and snowshoe hare detected slightly more readily at shorter spacing (26% and

32%) and short-tailed weasel showing improved detection success with longer spacing (26%). For

coyote, fisher, snowshoe hare and red squirrel, the difference between short and long transects was

15
negligible. We then tested that the differences we observed resulted from the different methods, and

not from the arbitrary selection of which two cameras out of the three available were included in the

two camera methods. We repeated our analyses using the alternative possible pairs of cameras for two

short (units A and B in Fig 1.2) and for two long (units B and D). We observed very similar trends for all

six species in terms of the role that camera number has on detection success. Only the trend for marten

and the role of camera spacing varied, as shown in Appendix B Table B.3 and Table B.4.

16
Fig 1.3. Model-averaged estimated detection probabilities and standard errors for camera trap surveys
conducted in north and central Maine, USA, during winter 2017. The detection histories for six species
were analyzed for five different potential survey methods of 1-3 camera units spaced short (100 m) or
long (150 m). When derived from models including the study area parameter, results are shown for the
Telos Road/Nahmakanta Public Reserve Lands, and when models included access type, results are
displayed for plowed road access.

17
 ̂ is
̂ is the probability of occupancy, ϴ
Table 1.3. Parameter estimates for six mammal species in Maine: Ѱ
the probability of availability for detection on any survey day, and p̂ is the probability of detection with a
specific method conditional on availability.

Species ̂ (se)
Ѱ ϴ̂ (se) p̂s(se) p̂s · ϴ̂ (se)
Canis latrans 0.67 (0.17) 0.07 (0.02) p1: 0.13 (0.07) 0.009
2
p : 0.51 (0.09) 0.036
3
p : 0.65 (0.08) 0.046
4
p : 0.34 (0.09) 0.024
p5: 0.62 (0.09) 0.043
1
Pekania pennanti 0.73 (0.07) 0.12 (0.08) p : 0.60 (0.16) 0.072
2
p : 0.79 (0.11) 0.095
p3: 0.83 (0.09) 0.100
4
p : 0.71 (0.14) 0.085
5
p : 0.73 (0.13) 0.089
1
Martes americana 0.76 (0.08) 0.24 (0.02) p : 0.29 (0.07) 0.070
p2: 0.47 (0.06) 0.113
3
p : 0.79 (0.06) 0.190
4
p : 0.55 (0.07) 0.132
5
p : 0.58 (0.07) 0.140
Mustela erminea 0.51 (0.12) 0.33 (0.04) p1: 0.24 (0.05) 0.079
2
p : 0.47 (0.06) 0.155
3
p : 0.60 (0.06) 0.198
p4: 0.52 (0.06) 0.172
5
p : 0.77 (0.05) 0.254
1
Lepus americanus 0.50 (0.12) 0.36 (0.03) p : 0.37(0.07) 0.133
2
p : 0.49(0.08) 0.176
p3: 0.72 (0.06) 0.259
4
p : 0.49 (0.07) 0.176
5
p : 0.61 (0.07) 0.220
Tamiasciurus hudsonicus 0.53 (0.31) 0.19 (0.10) p1: 0.34 (0.10) 0.065
2
p : 0.66 (0.09) 0.125
3
p : 0.81 (0.06) 0.154
4
p : 0.57 (0.09) 0.108
p5: 0.69 (0.09) 0.131

18
 Each of five methods of pooling detection histories by different configurations of cameras are

compared: p1 is a single camera, p2 is two cameras spaced 100 m apart, p3 is two cameras spaced 150 m,

̂ is the estimated probability

p4 is three cameras spaced 100 m and p5 is three cameras spaced 150 m. Ѱ

̂ is the estimated daily probability of availability for detection, 𝑝

of occupancy, ϴ ̂𝑠 is the probability of

detection given availability for each of five camera configuration methods s, and se is the standard error.

̂ and 𝑝
The multiplied value of ϴ ̂𝑠 represents the daily detection probability of a single-scale occupancy

model with only the data from cameras pooled into that method. When detection was modeled by

study area, Telos Road/Nahmakanta Public Reserve Lands estimates are shown; when modeled by

access type, plowed road estimates are shown.

Of the parameters influencing the detection process, study area was most frequently included in

top ranked models (Table 1.4). The next most frequently included variable was the access type

dominating the camera array. For fisher and snowshoe hare, arrays set near plowed roads showed

higher probabilities of detection, while for coyote arrays set near snowmobile trails showed higher

detection. The distance to road or trail was important for marten and for fisher, with both species

showing an increase in detection probability as transects were placed farther from roads and trails.

19
Table 1.4. Influence of camera site features on detection probabilities for six mammal species in Maine,
USA.

Variable Species Beta estimate

Access type (road) Canis latrans -0.262
Pekania pennanti +0.169
Lepus americanus +0.177
Minimum distance to road or trail* Martes americana +0.015
Average distance to road or trail* Pekania pennanti +0.008
Study area Martes americana SL: Intercept
TN: -0.037
ML: +0.143
Mustela erminea SL: Intercept
TN: -0.005
ML: -0.529§
Lepus americanus SL: Intercept
TN: +0.057
ML: -0.192
Tamiasciurus hudsonicus SL: Intercept
TN: +0.297
ML: +0.292

Surveys occurred in winter, thus Access type was defined as either plowed road or snowmobile trail.
Distance to road varied for different camera arrays, with minimum distances as close as 10 m and
median distance up to 320 m. The three study areas were Scraggly Lake (SL), the Telos
Road/Nahmakanta Public Reserve Lands (TN), and Moosehead Lake (ML).
*Betas indicate the impact of an additional 10 m in distance from the access point.
§
Mustela erminea (short-tailed weasel) was only detected once in the Moosehead Lake study area.

1.5 Discussion

Our findings demonstrate that increasing the number of camera traps deployed within a survey site had

a positive effect on the probability of detection for six species of North American mammals, while the

distance between cameras was less important. Although the objectives of camera trapping studies can

vary widely, if detection probability at each site is low, estimates of occupancy or other parameters of

interest are likely to be biased (MacKenzie et al. 2002, Guillera-Arroita and Lahoz-Monfort 2012). We

20
found that the magnitude of the increase in probability of detection given availability varied, but across

all species it was greatest going from a single camera to two cameras at a site. The increase from two

cameras to three was most substantial for coyote, the species with the lowest probability of availability

for detection during any one survey occasion.

1.5.1 Camera trap number and spacing

As the price for trail cameras decreases, it is more economically feasible to place more than one unit at

distinct micro-sites within a single survey site, without sacrificing the total number of independent sites

(though see Newey et al. 2015). Because it is often time and labor intensive to access spatially

independent survey sites, we believe it is prudent to follow a study design incorporating multiple

cameras placed systematically to improve the effectiveness of each site and to safe-guard against

wasted effort in the event of camera failures. While increasing the number of camera units at a site has

been demonstrated to increase detection probability (Stokeld et al. 2015, Pease et al. 2016, O’Connor et

al. 2017), the magnitude of the increase is likely to vary between regions and target species. We found

that the increase in detectability was higher going from a single camera to two cameras than when

increasing cameras from two to three and, in particular, it was beneficial for species with lower

detection probabilities, such as coyote. Although many other components of study design will influence

detection success, such as duration of camera deployments and the use and type of baits and lure, the

addition of cameras can be considered alongside these. Increasing the number of camera units at each

site may help also reduce the need to place cameras for long periods of time (see Shannon et al. 2014)

which will both increase the overall number of sites that can be sampled within a season and avoid

violations of occupancy model closure assumptions (Efford and Dawson 2012, Otto et al. 2013,

MacKenzie et al. 2017). The above trade-offs in study design will also depend on the species or group of

species targeted and must be considered within the context of overall study objective (Wintle et al.

2010, Noon et al. 2012). Increasing the number of camera units is a valuable tool and can provide large

21
gains with low added cost when including the time and effort expended to reach each survey site, and

should be considered during the study design process. In particular, our study demonstrates that

generally the greatest gain in detection probability can be obtained by working with two cameras,

spaced either 100 m or 150 m, as compared to using a single camera.

We concede that the dense arrangement of cameras we deployed may have resulted in a

saturation of lure that could have inflated the detection probabilities we observed. The attractant could

have increased the daily availability (ϴ) of species with a home range overlapping the survey site, and

this would then influence the overall detection probability of our methods. However, we do not expect

this to bias the results between different configuration methods because all cameras faced bait and lure.

Furthermore, we emphasize that the multi-scale modeling approach accounts for non-independence of

different methods being compared (Nichols et al. 2008). Indeed the dependence between scales is

exploited to allow for comparisons between methods at each array thus attributing variation in the

detection process to the different number and spacing of camera units (Nichols et al. 2008).

We have expanded upon other recent studies showing the impact of increasing the number of

cameras by explicitly investigating how consistent arrangements affect detection success. We observed

the highest detection probabilities for American marten and short-tailed weasel in survey methods with

150 m spacing, and the highest detection probability for coyote with 100 m spacing. This may reflect

that although weasel and marten are smaller bodied than coyotes, they differ in their hunting strategy

and tend to investigate many features of the landscape they move through, rather than following a

cursorial route (Feldhamer et al. 2003, Elbroch and Rinehart 2011). Fisher, red squirrel and snowshoe

hare showed almost no difference based on transect spacing. Although our results suggest that transect

length has the potential to influence detection and warrants further consideration, the influence of the

transect spacing, either 100 m or 150 m, was not as strong as increasing the number of cameras.

Furthermore, the difference between 100 m and 150 m, which was the most feasible spacing given the

22
constraints of our study, is relatively small. Although this ensures dependence among the camera sites

in terms of animal availability for small- to medium-bodied species, future work investigating greater

differences in the lengths between camera units would be valuable. The ideal spacing of camera units is

likely to vary in different regions and between species of interest, and one trade-off to consider when

designing surveys will be maximizing detection success for one single species versus obtaining adequate,

unbiased data on multiple target species. In the context of our work that may mean selecting a transect

spacing that raises detection probability for the species with the lower inherent probability, though this

same arrangement may not maximize all target species.

1.5.2 Landscape and camera site features

We found that the site characteristics which we modeled (overall study area, access type, and distance

to access feature) influenced the detection process for species differently. For coyote, American marten,

and fisher, the access type nearest to the camera transect ranked highly as an explanatory variable for

the detection process. Sites off of plowed roads had higher likelihoods of detection for marten and

fisher while sites off snowmobile trails had higher likelihoods for detecting coyote. Although not the

focus of our study, recreational activity have been shown to influence carnivore behavior (Moriarty et al.

2011, Kowalski et al. 2015), and this may reflect how different carnivore species perceive and avoid

human threats. This is an important consideration when designing camera surveys especially when ease

of access to deploy sites could bias results. One further consideration is that season could impact ease of

movement for both researchers and wildlife species. Winter has been preferred for some survey

methods (e.g. snow track surveys [Hoving et al. 2004], camera traps and genetic sampling [Robinson et

al. 2017]) but results could vary by season. For both American marten and fisher, there was also a

positive relationship between detection success and increasing distance from the access feature, though

even sites close to roads and trails did detect these species. Our individual camera microsites ranged

from 10 m to 320 m away from a road or snowmobile trail (mean 133 m). Fisher especially had higher

23
probability of detection for methods located on average farther from the access route, suggesting that

researchers should avoid only placing sites close to roadways despite the logistic ease. Similar findings

were recently shown in European carnivores (Mata et al. 2017), suggesting that some, though not all,

species avoid close proximity to roadways and that placing survey sites at a range of distances will

increase the efficacy of results when working in a multi-species system.

The study areas in which the camera stations were placed spanned three regions in north-

central Maine, and this was the most frequently included variable. For most species the Telos

Road/Nahmakanta Public Reserve Lands region showed the highest detection probability. American

marten, which is a species of special interest in Maine due to possible impacts from land use change

(Simons 2009), was the only species which showed a lower detection probability in this study area.

While a detailed comparison between the three study areas is not the goal of this paper, we note that

for this species increasing the survey unit from one to two cameras increased detection probabilities to

45-60% across all three of our study areas. However, this does highlight that the detection process will

vary by species, and may be important particularly in regions with small populations. Therefore we

recommend pilot studies to determine a survey design that will be able to acquire robust data for any

high priority target species and avoid ambiguous findings stemming from inadequate study methods

(see Hayward et al. 2015).

1.5.3 Camera trapping survey design

As the importance of camera trap data continues to grow, optimizing the efficacy of survey design is in

the best interest of both wildlife research and applied management, and increases the potential for

collaboration (Steenweg et al. 2017). For all camera surveys, regardless of overall objective, increasing

the probability of detection for species of interest when it is indeed present (which can also be thought

of as decreasing false-absences) will improve data quality and enable better interpretation of underlying

biological processes. For monitoring efforts or research on the underlying causes of species distribution

24
at large scales, increasing the detection probability at each independent survey location by pooling the

data from multiple cameras in close proximity, but at unique microsites, will increase the reliability of

occupancy estimates (MacKenzie et al. 2017). To explicitly clarify, although we have analyzed our data

using a multi-scale occupancy model to understand how different configurations of camera can impact

detection, we would encourage researchers to use the pooled observation histories of all co-located

devices to maximize detection probability at each site as a whole. However it is important to note that

recent work (Efford and Dawson 2012, Neilson et al. 2018, Stewart et al. 2018) indicate that for mobile

organisms such as terrestrial carnivores, “occupancy” of an area may be better described as the

probability that an animal’s movement pattern will bring it in front of the relatively small field of few of

point sampling devices, such as trail cameras. Even for different research objectives, such as spatial

capture-recapture which will require dependence at survey sites themselves, increasing detection

success with multiple sites may be worth considering (Sun et al. 2014, Tucker et al. 2014).

Our findings may be valuable to researchers in North America where these species occur, as well

as in similar ecosystems in Europe. For example, the European pine marten (Martes martes) and beech

marten (Martes foina) are sympatric species that may experience competition and exclusion similar to

the American marten and fisher in parts of their range (Monterroso et al. 2016, Manlick et al. 2017).

Camera trapping research for these species will benefit from using study designs that ensure high

quality data on multiple species of interest (Rota et al. 2016). Further research on the role that camera

number and transect length can have in other ecosystems will be useful, and can be implemented in a

pilot season prior to broader scale survey efforts. Particularly if species have relatively high detection

probabilities, such pilot seasons could avoid unnecessary increases in survey effort if only two cameras

can provide a substantial increase over one camera. Our work illustrates the benefit of such pilot studies

to determine optimal survey design for camera trapping research, and highlights the substantial

increase in data quality that results from a study design using transects with multiple cameras.

25
1.6 Acknowledgments

We are grateful to W. J. Jakubas for his support initiating the research, to B. J. Cimpher for assistance

during winter field research, and to G. Archambault for data entry. This project was supported by the

USDA National Institute of Food and Agriculture, McIntire‐Stennis Project Number ME0‐41913 through

the Maine Agricultural & Forest Experiment Station and by the Maine Department of Inland Fisheries

and Wildlife. We thank Huber Resources Corp and Katahdin Forest Management for granting access to

their holdings, as well as the Maine Department of Inland Fisheries and Wildlife and The Nature

Conservancy for providing access and accommodations. We also thank M. L. Allen for early comments

on the manuscript, and J. T. Fisher and three anonymous reviewers for valuable feedback. Maine

Agricultural and Forest Experiment Station Publication Number 3669.

26
 CHAPTER 2: MARTEN, FISHER, AND FOREST DISTURBANCE IN MAINE

2.1 Chapter abstract2

Human land-use is a driving force of habitat loss and modification globally, with consequences for

wildlife species. The American marten (Martes americana) and fisher (Pekania pennanti) are forest-

dependent carnivores native to North America. Both species suffered population declines due to loss of

forested habitat and over-harvest for furs, and continued habitat modification is an ongoing threat.

Furthermore, the smaller marten may be susceptible to intraguild exclusion where the larger fisher are

abundant, and both habitat modification and climate change may reduce spatial refugia available to

marten. A detailed understanding of co-occurrence patterns of marten and fisher in landscapes

subjected to intense forest disturbance represents a key knowledge gap for wildlife ecology and

management.

Maine, in the northeastern United States, supports populations of both these species. It is an

extensively forested state, and the vast majority is managed as commercial timberland. We designed a

large-scale field study to understand the relative importance of three sets of predictions for marten and

fisher occupancy patterns where commercial silviculture is widespread: 1) The intensity of forest

disturbance primarily determined both marten and fisher occupancy rates, 2) fisher occupancy was

limited to areas of shallower snow and marten limited by fisher presence, or 3) both species responded

to the composition of tree species within forested habitat.

Note on citation 2
The work described below has been accepted for publication in a peer reviewed journal, and any
readers wishing to cite our findings should do so as follows: Evans, B. E. and A. Mortelliti. 2022. Effects of
forest disturbance, snow depth and intraguild dynamics on American marten and fisher occupancy in
Maine, USA. Ecosphere, accepted 12/7//2021, manuscript ID ECS21-0761. To receive a copy of the
journal publication at no charge, please visit: https://www.researchgate.net/profile/Bryn-Evans-2 or
email: [email protected].

27
We collected data to test these non-mutually exclusive hypotheses via camera trap surveys, using an

experimental design balanced across a gradient of forest disturbance intensity. We deployed 197

camera stations in both summer and winter over three years (2017-2020). We tagged over 800,000

images and found marten at 124 (63%) and fisher at 168 (85%) of the stations. By fitting multi-season

occupancy models to the data we found that the degree of habitat disturbance negatively influenced

detection, occupancy, and temporal turnover for both species. Contrary to our expectations, however,

we found no evidence of interspecific competition, and instead support for positive associations with

detection probabilities both spatially and temporally. Both species were positively associated with forest

stands containing deciduous trees. Our findings further illustrate the impact that land-use has on the

occupancy dynamics for these forest-dependent carnivores.

Open Research statement: Data are archived on Figshare:

https://figshare.com/projects/Marten_and_fisher_occupancy_in_Maine_USA/121533

2.2 Introduction

American marten (Martes americana) and fisher (Pekania pennanti) are mustelid predators native to

forests in Canada and the northern United States (Powell et al. 2003, Aubry et al. 2012). Following

European colonization of North America, unregulated trapping for pelts and the clearing or logging of

vast areas of forest led to population declines and local extirpations (Maser 1989, MacCleery 2011,

White et al. 2015). As silvicultural techniques advanced towards sustainability (Puettmann et al. 2008)

and states implemented trapping regulations (and in some cases, active reintroduction efforts), both

species began to recover across much of their historic range (Aubry and Lewis 2003, Williams et al. 2007,

Lancaster et al. 2008). However, habitat disturbance, often resulting from timber extraction, is still a

major threat to these species: activities that reduce forest patch size, remove large and senescent trees,

and simplify forest structure may severely compromise the long-term viability of their populations

(Patton 1992, Harrison et al. 2004, Moriarty et al. 2011, Sauder and Rachlow 2014, Bridger et al. 2016).

28
Marten and fisher are taxonomically close (Stone and Cook 2002, Koepfli et al. 2008, Sato et al. 2012)

with similar life history traits and hunting strategies, but they differ in body size (marten 0.5–2 kg, fisher

2–7 kg, Zielinski et al. 2003). Their geographic ranges overlap, though marten extend farther north than

fisher (Clark et al. 1987, Powell 1993) (Fig 3.1). Both species are dependent on forested habitat, and

although the effects of land-use change on each individual species have been studied and suggest

certain effects are consistent across their ranges, others may be highly variable. As an example for

fisher, logging activity that removes snags and decaying trees from the landscape is likely incompatible

with successful reproduction (Paragi et al. 1996, Weir et al. 2012). Although large trees and connectivity

between stands are still required to sustain populations in managed forests, fisher do tolerate more

disturbance than previously thought (Matthews et al. 2013, Zielinski et al. 2013). For marten, even

partial-harvest activities can diminish the canopy cover, structural complexity and overall basal area

they require and thus reduce habitat suitability (Chapin et al. 1998, Fuller and Harrison 2005). Although

home ranges can overlap areas of cut forest and edge habitat, harvest practices that maintain forest

cover by retaining larger patches of uncut trees, reducing edges, and keeping total cut areas below

twenty- to thirty-percent of the landscape are more viable (Hargis et al. 1999, Potvin et al. 2000).

A critical knowledge gap in the study of marten and fisher response to landscape change is the

role of intraguild competition between these two species (Ritchie and Johnson 2009, Lesmeister et al.

2015, Sévêque et al. 2020). Because of the similarities in their ecological niches, marten and fisher may

experience intraguild competition in areas of sympatry (Polis and Holt 1992, Palomares and Caro 1999,

Donadio and Buskirk 2006). The disparity in body size between fisher and marten may allow fisher to

suppress marten populations, either due to direct predation (McCann et al. 2010, Jensen and Humphries

2019) or competition (Fisher et al. 2013, Manlick et al. 2017). However, there are also many parts of

their range where the species do co-occur, suggesting that some form of niche partitioning facilitates co-

existence (Hardin 1960, Schoener 1974). One hypothesis is that the larger, heavier fisher is limited to

29
areas of shallower, wetter snow in winter, providing a spatial refuge for marten at higher elevations or

other areas with deeper snowpack (Raine 1983, Krohn et al. 1995). This may be an important

component of the species’ geographic distributions, where marten benefit from both increased

subnivean hunting access in winter and reduced competition from fisher in regions of deeper snow. It

does also explain spatial segregation in some areas of sympatry, such as along the elevation gradient in

the Sierra Nevada mountains (Sweitzer and Furnas 2016). However, there are many areas where the

species do overlap in space and time, even in winter seasons (Williams et al. 2009, Croose et al. 2019).

Understanding the mechanisms allowing marten and fisher to coexist is essential for the long-

term conservation of both species. This is undoubtedly a dynamic process, involving local climate,

habitat characteristics, prey and predator species (Zielinski and Duncan 2004), and direct and indirect

human perturbations (Gabriel et al. 2012). The overarching goal of our study was to investigate the

interacting roles of forest disturbance, habitat and local climate on marten and fisher co-occurrence

patterns in Maine, USA. Specifically, we aimed to test the relative support for three, non-mutually

exclusive hypotheses: H1: Habitat disturbance, resulting from timber harvest, is the main driver of the

distributions of marten and fisher. Based on this hypothesis we would expect to see occupancy patterns

for both species restricted to areas of low- to moderate-intensity disturbance, and that co-occurrence

would be most common in areas of lowest forest disturbance (Thompson 1988, Simons 2009, Sauder

and Rachlow 2014, Lewis et al. 2016). H2: Species-specific climate limitations and intraguild dynamics

drive distributions in Maine. In this case, we expect fisher to be limited to areas of shallower winter

snow, and marten limited to areas with deeper snow and low fisher probability of presence (following

Krohn et al. 1995). H3: Habitat characteristics, as captured by the composition of forest stands, drive

marten and fisher distribution. If traits such as forest structure, deciduousness in winter, and food

available for prey species are most important, we will expect to see relationships between the

predictors for composition (deciduous versus coniferous tree species, which we are using as an index for

30
multiple facets of forest habitat) and the occupancy patterns of marten and fisher, with occasional co-

occurrence in areas where habitat preferences overlap (Powell et al. 2003, Aubry et al. 2012).

To achieve our goal, we established a field study across the northern two-thirds of Maine,

wherein we balanced survey effort for terrestrial wildlife across a gradient of forest disturbance

intensity and latitude. Marten and fisher occupancy data were collected through transects of camera

traps optimized for these species (Evans et al. 2019) across 197 survey sites, sampled over a 4-year

period (Kays and Slauson 2008, Kays et al. 2011, Boitani et al. 2012, Rowcliffe 2017, Steenweg et al.

2017). By comparing the relative importance of forest management, snow depth, and species dynamics

in an information theoretic framework, we aim to contribute to filling important knowledge gaps in the

basic ecology of marten and fisher and inform optimal conservation strategies.

2.3 Methods

2.3.1 Study area

Our research was conducted in the state of Maine, in the northeastern United States (Fig 2.1). Maine is

primarily forested, with 89% of the state (17.6 million acres) covered by forest land (Butler 2018), and

located along the transition zone from Eastern Deciduous in the south of the state to Boreal Forest in

the north (Gawler et al. 1996, McWilliams et al. 2005). Common tree species include maple (Acer sps.),

birch (Betula sps.) and American beech (Fagus grandifolia), balsam fir (Abies balsamea), spruce (Picea

sps.) and white pine (Pinus strobus) (Butler 2018). Forests are structurally diverse with predominantly

mixed-age stands resulting from both natural disturbance (insect outbreaks and storm damage [Fraver,

White, and Seymour 2009]) and anthropogenic disturbance (typically partial harvesting [Seymour et al.

2002]). Our 197 sampling sites were distributed across the majority of the state, except for areas outside

of the marten geographic range (Downeast and southern regions which are influenced by coastal habitat

and human development, and the far eastern border which is predominately open, agricultural land)

(geographic coverage: 44.15968°N to 47.42166°N and -70.55842°W to -67.62582°W). Mean

31
temperatures in July, the hottest month in Maine, are 18.5 C (average low of 12 C to a high of 25 C).

Winter temperatures in January, the coldest month, average -10 C (-16 to -4.5). From May to October

roughly 55 cm of precipitation fall as rain, and another 51 cm fall from November to April predominately

as snow (NOAA 2020).

Fig 2.1. Camera trap survey stations (n=197) deployed in Maine, northeastern USA, at the border of
overlapping marten (Martes americana) and fisher (Pekania pennanti) geographic ranges (IUCN 2021).
Each station was deployed for at minimum one summer and the following winter season, and a subset
(highlighted in white) were resampled for up to 7 total seasons.

32
2.3.2 Study design

We established a large-scale field study (Garton et al. 2011) to collect data on wild carnivore occupancy

patterns across the range of forest disturbance regimes characteristic of Maine. We stratified our

sampling based on the following factors: 1) Low to high intensities of forest disturbance (Appendix C), 2)

latitude, and 3) reported furbearer take by trappers for marten and fisher (Appendix D.1 Fig D.1, D.2,

D.3).

Forest disturbance intensity was obtained from remote sensed Landsat image data (Wulder et al. 2019)

that were processed using a novel method to combine multiple individual change-detection algorithms

(Kilbride 2018). Image layers from 1948 to 2017 provided continuous raster indices describing the

magnitude, year and duration of the most recent disturbance event modeled at each 30 m pixel,

calibrated specifically to northeastern forests. We categorized each pixel by the year of most recent

disturbance (none or prior to 1989 = 1, 1989-2000 = 2, 2001-2012 = 3, 2013-2017 = 4) and the

magnitude of that disturbance (0-199 = 1, 200-639 = 2, 640-1,278 = 3, 1,279-10,185 = 4). We used raster

math to multiply these, giving a discrete score from 1 to 16 for each pixel across the state of Maine.

From this layer we calculated the mean disturbance index at multiple spatial scales: townships (for initial

course-scale selection of survey locations), and circular buffers around each surveyed site once

established (300 m and 1 k radius “local” disturbance, and 3 k and 6 k radius “landscape” disturbance

scores, the maximum area around each station while maintaining independence). For further details see

Appendix C.

We used a balanced study design to distribute our survey effort along a continuum of forest

disturbance. We prioritized townships where high, moderate, and low forest disturbance regimes were

present in both large, continuous blocks (3+ townships of a similar value) and smaller, isolated blocks (1-

2 townships) surrounded by differing disturbance history. We replicated each of these at least once in

both the northern half and southern half of our survey extent to ensure no systematic bias was

33
introduced. We further confirmed this by checking for correlation between our predictor variables and

found that only latitude and snow had correlation coefficient greater than 0.1 and were therefore never

included in additive models (Appendix D.1 Fig D.4).

To maximize spatial coverage while also collecting multiple-season data, we used a rotating

panel design (MacKenzie et al. 2017; pg. 218) in which a “permanent” subset of our sites were deployed

from five to seven seasons, while other sites were deployed for only two seasons. We selected

“permanent” sites which were representative of the broader study design (e.g., sites in areas of low

disturbance in large blocks; sites in areas of high disturbance surrounded by only moderate disturbance;

etc.) and we replicated these in both northern and southern portions of the overall study area (Fig 3.1).

Furbearer trapping intensity was accounted for with additional survey stations deployed in townships

with the highest reported fur trapping harvest (based on the 2014-2018 trapping season reports for

marten and fisher [MDIFW 2019]) to augment the survey effort already expended in townships of low

and moderate reported harvest (Appendix D.1 Fig D.3).

Survey stations were placed at randomly generated points within high priority townships as

described above, maintaining a minimum distance of 6 k between stations. This spacing meets the

assumption of independence between our detections for martens (home ranges 6 k2 to 18 k2 [Gosse et

al. 2005, Simons 2009, Shirk et al. 2014] and fishers (home ranges up to 10 k2 for females and up to 38

k2 for males [Clark 1986, Powell et al. 2003, Furnas et al. 2017]).

Snow depth data were obtained from the National Snow and Ice Data Center (NSIDC) Snow Data

Assimilation System (SNODAS) data products (Brown and Brasnett 2010). We took the daily mean snow

depth value of estimates within 1 k of each survey site, for a duration of two months in the peak of each

winter season that surveys were conducted (January 15 to March 15).

34
 Habitat data were collected once at each survey station for stand level tree species composition

(living trees within two, 20-factor variable radius plots), and we used LANDFIRE vegetation type raster

data for 300 m and 1 k buffers at larger spatial scales (LANDFIRE 2016).

Camera trapping was conducted by placing three Bushnell Trophy Cam E2/E3 passive infrared

cameras (Overland Park, KS, USA) at each survey station, spaced 100 m apart in a linear transect

perpendicular to the access road/trail: we found from prior research that this configuration would

maximize the probability of detection for marten and fisher (Evans et al. 2019, see also Mann et al.

2015, Pease et al. 2016). Each camera was set approximately 40 cm above the ground (or packed snow),

facing towards a suet cage with a piece of beaver (Castor canadensis) meat, bait which served as a local

attractant (Stewart et al. 2019) (Appendix D.1 Fig D.5). Scent lure designed for furbearers (skunk essence

and Vaseline based, private seller in Kenduskeag, ME, USA) was applied at the base of the bait tree and

at > 2 m height, to act as a long-distance attractant (Schlexer 2008, Stewart et al. 2019). A previous

study in this region demonstrated that the combined use of bait and lure increased the detection

probability of carnivore species without impacting non-carnivore mammals (Buyaskas et al. 2020). Each

year of our surveys, stations were first deployed in a summer season for 15+ days, then revisited in the

following winter season for 15+ days to collect two seasons of detection data. Each following year a

subset of previously surveyed sites were re-deployed to increase the total number of contiguous

seasons of data (up to seven) while the remaining effort was allocated to new sites to increase total

sample size. Seasons ran from June to September in summer and January to April in winter, for the years

2017-2018, 2018-2019, 2019-2020, and summer 2020. Our methods adhere to ethical research

standards for free ranging wild mammals (Sikes 2016) and were approved by the University of Maine

Institutional Animal Care and Use Committee, Protocol #A2018-05-06.

35
2.3.3 Occupancy models

We analyzed species detection data in two phases: using multi-season, single-species occupancy models

in R unmarked (MacKenzie et al. 2003, R core team 2020, Chandler et al. 2021) and then multi-season,

two-species models in program Presence 2.13.12 (MacKenzie et al. 2017, Hines 2021). Occupancy

modeling accounts for the potential bias introduced with imperfect detection (MacKenzie et al. 2002,

White 2005) and allows inclusion of habitat level and observation level variables that may influence the

probability of an animal being observed at a survey station separately from the occupancy process

itself. We constructed detection histories by pooling images from all three cameras within each station,

for each 24-hour period that the station was active in summer and in winter. We incorporated the

following covariates: latitude; year and season of surveys; forest disturbance within 300 m, 1 k, 3 k and

6 k radius buffers around each station, to account for the scale of effect of local, home-range and

landscape-level roles of disturbance on detection, occupancy, and turnover; and the proportion of

deciduous trees within the immediate stand, as well as 300 m and 1 k radius buffers. We ran two

phases of sequential models: 1) Single-species, multi-season models to determine key landscape

variables affecting marten alone and fisher alone. We first ranked models with all parameters held

constant except the one of interest, beginning with detection. We then retained the top model for the

detection process as we repeated this step for the initial occupancy probability, then colonization, and

finally extinction. We ranked models via Akaike Information Criterion corrected for small sample sizes

(Burnham and Anderson 2002). If one or more model ranked within 2 ΔAIC of the top model, and was

not modeling the same covariate at a different spatial scale (e.g. disturbance at 1 k followed by

disturbance at 300 m) additive models were also tested. 2) After models with habitat covariates were

fully parameterized, we ran two-species, multi-season models for marten and fisher to explicitly test for

the importance of interaction between species. We used the ψ, p, r parameterization for the two-

species models, which allow for initial occupancy probabilities to be modeled by whether or not the

other species is present; for colonization and extinction between seasons to be modeled by the

36
presence or absence of the other species; and for detection to vary both at sites where the other species

is present, as well as at the observation occasion, if the other species is simultaneously detected. This

modeling requires one species to be considered “dominant”, i.e. driving the dynamics of the other

species, and thus we ran models twice to include all the possible directional interactions between

marten and fisher. We tested for the impact of interspecific dynamics on initial occupancy, colonization,

extinction, and detection with an a priori set of models, described in more detail in Appendix D.2, Table

D.1, D.2.

2.4 Results

We obtained image data from 197 independent survey stations across seven seasons, for a total of 632

station/seasons. Each station was deployed for a minimum of 15 consecutive days (33-day maximum

survey period analyzed, mean = 26.2 ± 4.6SD) in one summer season and all but four stations again in

the following winter season, from 2017 to 2020. A subset of stations were redeployed for additional

summers and winters, with three stations out for 3-4 seasons total, 18 stations out for 5 seasons total

and 38 for 6-7 seasons total (Appendix D.3 Table D.3). Attrition in the number of camera stations was

due to access issues (some areas were impossible to reach in winter), changes in landowner or

unforeseen equipment failures. In total we recorded over 800,000 images, of which 12,000 contained

identifiable martens and 67,000 contained fishers. After collapsing images into 24-hour survey periods,

our data set contained 1,086 independent marten visits and 1,112 fisher visits, and of those 123 were

visits by both species to the same station within a single day. Marten were detected at 35% to 61% of

stations deployed in each of the seven seasons (naïve occupancy), and fisher were detected at 26% to

86%, with consistently greater detection rates in winter than in summer seasons (Appendix D.3 Table

D.4). After accounting for imperfect detection, the estimated seasonal occupancy rates for marten

across the survey extent ranged from 0.41 to 0.59 and fisher occupancy ranged from 0.54 to 0.73, with

both species detected across the entire spatial extent of our surveys (Appendix D.3, Fig D.6, D.7).

37
 The top ranking single-species, multi-season model for marten included the intensity of

disturbance, with negative effects on initial occupancy (β = -0.64 ± 0.19) and colonization (β = -0.51 ±

0.18), and a positive effect on extinction probability (β = -0.52 ± 0.20). The percentage of deciduous

trees had a positive effect on initial occupancy (β = 0.84 ± 0.18) and the average depth of winter snow

was positive with colonization and negative with extinction (β = 3.02 ± 1.11 and β = -3.00 ± 1.21).

Detection probability for marten decreased with the number of days in the survey, and also decreased

with more disturbance (β = -0.025 ± 0.004 and β = -0.23 ± 0.032) (Table 2.1) (Fig 2.2) (Appendix D.3

Table D.5).

For fisher, no covariate ranked above null for initial occupancy, while the probability of turnover

between seasons negatively affected by disturbance (colonization β = -0.47 ± 0.25, and extinction β = -

0.23 ± 0.17). The proportion of deciduous trees had a positive effect on colonization (β = 1.25 ± 0.57),

and snow depth had a negative effect on extinction (β = -2.64 ± 1.13). Detection probability for fisher

varied with survey session, with higher detection rates in the winter seasons than in the summer

seasons (see Appendix D.3, Table D.5).

Results of the two-species, multi-season models indicate that for marten, the occupancy status

of fisher was not included in any of top models, but it was included in the top model sets (within 2 ΔAIC)

as a predictor for initial occupancy (positive), colonization (negative), and extinction (positive) (Table

2.2). However, the top ranked detection model included positive associations for both species. For

marten, detection probability during a given survey period was higher if a fisher was also detected (rFf <

rFF). For fisher, both the overall occupancy status of marten at a site, and the detection of a marten in a

period, increased detection probability (pm < rMm < rMM). For fisher, the probability of marten

presence was also included in models ranked within 2 ΔAIC as a predictor of initial occupancy (positive),

colonization (positive), and extinction (positive).

38
Table 2.1. Ranking of single-species, multi-season occupancy models within 2 ΔAICc of the top model for American marten (Martes americana)
and fisher (Pekania pennanti) derived from 197 camera trap station in Maine, USA, over four summer and three winter seasons, 2017-2020. ψ is
the probability of occupancy in the first season; γ is the probability of a colonization (a station that was unoccupied becoming occupied from one
season to the next); ε is the probability of extinction (an occupied station becoming unoccupied); and p is the detection probability given that a
station is occupied. Wi is the individual model, k is the number of parameters and R2 is the r-squared value per model.

Species Model formula ΔAICc AICc Wi k R2

Martes americana
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(DI_300m+Snow) p(DI_3k+Days) 0 6875.46 0.579 12 0.58
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(DI_300m+Lat) p(DI_3k+Days) 2.071 6877.53 0.206 12 0.58
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(DI_300m) p(DI_3k+Days) 4.171 6879.63 0.072 11 0.57
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(Snow) p(DI_3k+Days) 4.564 6880.02 0.059 11 0.57
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(Lat) p(DI_3k+Days) 5.955 6881.42 0.029 11 0.57
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(DI_1k) p(DI_3k+Days) 7.259 6882.72 0.015 11 0.56
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(.) p(DI_3k+Days) 7.982 6883.44 0.010 10 0.56
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(DI_6k) p(DI_3k+Days) 8.752 6884.21 0.007 11 0.56
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(DI_6k) p(DI_3k+Days) 8.765 6884.23 0.007 11 0.56
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(DC_300m) p(DI_3k+Days) 9.445 6884.91 0.005 11 0.56
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(DC_Stand) p(DI_3k+Days) 9.605 6885.07 0.005 11 0.56
ψ(DI_3k+DC_300m) γ(DI_300m+Snow) ε(DC_1k) p(DI_3k+Days) 9.895 6885.36 0.004 11 0.56

Continued next page

39
 Table 2.1 continued
Pekania pennanti
ψ(.) γ(DC_300m+DI_6k) ε(Snow+DI_6k) p(Survey) 0.000 7257.44 0.207 15 0.67
ψ(.) γ(DC_300m+DI_6k) ε(Snow) p(Survey) 0.042 7257.48 0.202 14 0.67
ψ(.) γ(DC_300m+DI_6k) ε(Snow+DC_1k) p(Survey) 1.111 7258.55 0.119 15 0.67
ψ(.) γ(DC_300m+DI_6k) ε(Snow+DC_1k +DI_6k) p(Survey) 1.460 7258.90 0.099 16 0.67
ψ(.) γ(DC_300m+DI_6k) ε(DI_6k) p(Survey) 3.401 7260.84 0.037 14 0.66
ψ(.) γ(DC_300m+DI_6k) ε(DC_1k) p(Survey) 3.976 7261.42 0.028 14 0.66
ψ(.) γ(DC_300m+DI_6k) ε(DC_300m) p(Survey) 4.422 7261.86 0.023 14 0.66
ψ(.) γ(DC_300m+DI_6k) ε(.) p(Survey) 4.597 7262.04 0.021 13 0.66
ψ(.) γ(DC_300m+DI_6k) ε(DC_Stand.) p(Survey) 4.701 7262.14 0.020 14 0.66
ψ(.) γ(DC_300m+DI_6k) ε( DI_3k) p(Survey) 5.529 7262.97 0.013 14 0.66
ψ(.) γ(DC_300m+DI_6k) ε(Lat) p(Survey) 5.831 7263.27 0.011 14 0.66
ψ(.) γ(DC_300m+DI_6k) ε(DI_300m) p(Survey) 6.389 7263.83 0.008 14 0.66

DI: Disturbance intensity, at the most influential spatial scale

DC: Percentage of living trees that were deciduous species, at the most influential scale
Snow: Daily values recorded within a 1 k radius buffer by National Snow and Ice Data Center (NSIDC) Snow Data Assimilation System (SNODAS),
averaged over all winters the stations were active (date range held constant, January 15 to March 15)
Survey: The season (summer or winter) and year (2017-2020) of data collection

40
Table 2.2. Ranking of two-species, multi-season occupancy models for American marten (Martes
americana) and fisher (Pekania pennanti). Models for each species were parameterized to test how the
occupancy and detection status of the other species would impact initial occupancy (ψ), colonization (γ),
extinction (ε) and detection (p). All parameters included the top ranked habitat covariates from the
single-species models previously run (reported in Table 2.1). We tested ψA versus ψa (initial occupancy
at a site where the other species was present versus absent); the four possibilities of turnover between
seasons; and detection depending on pa, rAa and raa. For marten occupancy models, F and f indicate the
status of fisher, and for fisher models M and m indicate the status of marten. See Table B.2 for
descriptions of the a priori models tested.

Species Parameter Model formula ΔAIC AIC Wi k

Martes americana
Initial ψF = ψf 0.00 14157.64 0.791 26
occupancy
ψF ≠ ψf 2.66 14160.30 0.209 27
Colonization γFF = γFf = γfF = γff 0.00 14157.64 0.466 26
γFF = γFf = γff ≠ γfF 2.23 14159.87 0.153 27
γFF = γFf ≠ γfF = γff 2.45 14160.09 0.137 27
γFF = γfF ≠ γFf = γff 2.66 14160.30 0.123 27
γFF = γFf = γfF ≠ γff 2.67 14160.31 0.122 27
Extinction εFF = εFf = εfF = εff 0.00 14157.64 0.442 26
εFF = εFf ≠ εfF = εff 1.94 14159.58 0.168 27
εFF = εfF ≠ εFf = εff 2.35 14159.99 0.136 27
εFF = εFf = εfF ≠ εff 2.38 14160.02 0.134 27
εFF = εf F = εff ≠ εFf 2.61 14160.25 0.120 27
Detection pf = rFf ≠ rFF 0.00 14147.33 0.628 27
pf ≠ rFF ≠ rFf 1.55 14148.88 0.290 28
pf = rFF ≠ rFf 4.21 14151.54 0.077 27
pf = rFf = rFf 10.31 14157.64 0.001 26
pf ≠ rFF = rFf 12.96 14160.29 0.001 27

Continued next page

41
 Table 2.2 continued
Pekania pennanti
Initial
ψM = ψm 0.00 14149.34 0.729 26
occupancy
ψM ≠ ψm 1.98 14151.32 0.271 27
Colonization γMM = γMm = γmM = γmm 0.00 14149.34 0.304 26
γMM = γMm ≠ γmM = γmm 0.39 14149.73 0.250 27
γMM = γMm = γmm ≠ γmM 0.66 14150.00 0.218 27
γMM = γMm = γmM ≠ γmm 1.91 14151.25 0.117 27
γMM = γmM ≠ γMm = γmm 2.00 14151.34 0.112 27
Extinction εMM = εMm = εmM = εmm 0.00 14149.34 0.401 26
εMM = εmM = εmm ≠ εMm 1.94 14151.28 0.152 27
εMM = εMm ≠ εmM = εmm 1.97 14151.31 0.150 27
εMM = εmM ≠ εMm = εmm 1.98 14151.32 0.149 27
εMM = εMm = εmM ≠ εmm 2.00 14151.34 0.148 27
Detection pm ≠ rMm ≠ rMM 0.00 14114.97 0.838 28
pm ≠ rMM = rMm 3.29 14118.26 0.162 27
pm = rMm ≠ rMM 17.42 14132.39 0.000 27
pm = rMM = rMm 34.37 14149.34 0.000 26

42
Fig 2.2. [Next page] Predictions from the top ranked multi-season occupancy models from 197 camera
survey stations deployed in Maine, USA. (A) Marten (Martes americana) detection decreases over the
survey period, and with increasing forest disturbance. (B) Marten initial occupancy is negatively affected
by forest disturbance intensity, and positively affected by the percentage of deciduous trees. (C) Marten
colonization between seasons is also negative related to disturbance, and positively related to increasing
snow depth. (D) Marten extinction is more likely with increased disturbance, and less likely in areas of
deeper snowpack. (e) Fisher (Pekania pennanti) detection varied with each survey, with a consistent
trend of higher detection probabilities in the winter seasons. (F) Fisher initial occupancy was around
50% across sites, while (G) colonization was less likely in areas of large-scale disturbance and more likely
in stands with more deciduous trees, and (H) extinction was negatively related to disturbance and snow
depth.

43
44
2.5 Discussion

Through a large-scale field study conducted in Maine, USA, over a four-year period we found that

marten and fisher dynamic occupancy patterns were driven primarily by the intensity of forest

disturbance (caused by forest management), followed by other habitat features (proportion of

deciduous trees) and, to a limited extent, by interspecific dynamics. We did not find a negative influence

of fisher presence on marten, but rather a positive relationship between detection events for both

species.

2.5.1 Effects of forest disturbance on marten and fisher occupancy patterns

The effects of disturbance were always negative: models indicated that survey stations in areas that had

more recent, or more intense, timber removal activities had lower probability of initial occupancy

(marten), lower colonization probability (marten and fisher), and higher probability of extinction

(marten and fisher). This is consistent with literature indicating that forest modification is detrimental to

habitat quality for marten and fisher (Chapin et al. 1998, Fuller and Harrison 2005, Zielinski et al. 2013)

and supports our first hypothesis that it is a driver of occupancy patterns in Maine. Marten especially

showed lower initial occupancy probability in areas of increasingly disturbed forest, and had both higher

extinction rates and lower colonization rates in these areas. This suggests that marten have a lower

tolerance for intense forest harvest than fisher, and may be more impacted by continued silviculture

practices in the state. The need for both cover from predators and structural complexity to access small

mammal prey could be more limiting for marten, where fisher are able to take advantage of increased

snowshoe hare in many areas (Fuller and Harrison 2005, Siren et al. 2021). For both species, our data

support our first hypothesis, that the degree of disturbance is a key factor in explaining the occupancy

patterns of marten and fisher in Maine.

45
2.5.2 Effects of snow depth on marten and fisher occupancy patterns

Snow depth positively affected both species, indicated by higher colonization and lower extinction rates

for marten, and lower extinction rates for fisher, in areas of deeper winter snow. Although the historic

literature suggests fisher are limited by deep snow (Krohn et al. 1995), we found they were widely

distributed, occupying over 70% of our survey stations regardless of either snow or latitude predictors.

The positive association between snow depth and mustelid predators may be linked to access to

“subnivean” resources, both for prey and for thermal shelter, during winter (Aubry et al. 2012). We

included only one factor of the many that can describe snow quality, and research into new modeling

techniques including sinking depths and formation of icy crusts may provide valuable insight (W.

Jakubas, personal communication). Overall our results, which failed to detect a negative impact on snow

depth on fisher, mirror recent trends documented in the northeast, suggesting that warming winters

and abundant prey such as snowshoe hare may be expanding the areas available to fisher (Suffice et al.

2017, 2020, Sirén et al. 2021). To what extent this trend may continue is unclear, and it is also possible

that had our surveys extended farther north (into Canada) we may have captured an inflection point

where winter conditions still limit fisher.

2.5.3 Intraguild effects on occupancy patterns

We documented co-occurrence between the two species at 98 stations (50%)—these detected both

marten and fisher within the same season—while another 13 (6%) detected both species, though only in

different seasons. We modeled marten occupancy based on the presence or absence of fisher to assess

the potential risk of intraguild predation (McCann et al. 2010, Manlick et al. 2017), and we modeled

fisher by including marten occupancy and detection not because we suspect they act as a dominant

predator, but to examine whether habitat preferred by marten was used more-than-chance by fisher.

Detection probability was higher for both species at sites where the other was present, and even on

individual survey occasions (24-hour periods) when the other species was also detected. This indicates a

46
greater-than-random overlap in the spatial and temporal use of their habitat, as represented by the

probability of daily detections in the same location, and perhaps indicating both species respond to

weather conditions favorable for hunting.

Because the inclusion of fisher occupancy probability failed to outrank the habitat-only models

for marten, we do not find evidence that competition is a limiting factor for the smaller species. Sévêque

and colleagues (2020) recently proposed a conceptual framework for three potential ways human

disturbance may influence niche partitioning among carnivores: disturbance can reduce total available

resources and thus increase competition; it can modify resources and destabilize community dynamics;

or it can increase the abundance or diversity of resources, and so facilitate greater niche partitioning.

This third scenario may be facilitating the overlap, both in space and time, we observed between these

species: they are taking advantage of the heterogeneous landscape resulting from forest disturbance,

rather than avoiding potential competitors as expected in our second hypothesis, though these are

speculative explanations.

While we did not find that marten occurrence was suppressed by the presence of fishers, it is

possible that either or both species are impacted by the risks posed by other mammalian predators (Bull

and Heater 2001, Feldhamer et al. 2003, Wengert et al. 2014, Moriarty et al. 2015, Jensen and

Humphries 2019). Exploratory analyses of our dataset indicate the potential for intraguild dynamics

driven by coyote (Canis latrans), Canada lynx (Lynx canadensis) and bobcat (Lynx rufus) and warrant

further investigations.

2.5.4 Effects of forest composition on occupancy patterns of marten and fisher

The species composition of trees as a predictor was included in top ranked models, and the presence of

deciduous trees was always positive, with higher deciduous proportion linked to a higher probability of

initial occupancy (marten) and colonization (fisher). This runs counter to some literature that indicates

preference for conifers (Wright 1999, Powell et al. 2003, Fuller et al. 2016), and does not offer support

47
for our third hypothesis that alternative habitat preferences segregated marten and fisher in Maine. The

preference of both mustelid species for stands with a greater percentage of deciduous species is

potentially linked to the availability of suitable trees for denning and reproduction (Payer and Harrison

2003, Weir et al. 2012). These stands may also offer increased availability and access to prey (Yaeger

2005, Davis 2009, Purcell et al. 2009), or the apparent role of hardwood may be a legacy of forest age

itself. Given the commercial value for conifer trees, many conifer-dominated stands in Maine are also

regenerating from harvest (McCaskill et al. 2016). However, we did not see any correlation between our

predictor variables for deciduous and for forest disturbance, suggesting that the tree composition itself

may influence habitat use for marten and fisher.

2.5.5 Conclusions

We found evidence that marten and fisher do respond negatively to heavily disturbed forests in Maine.

Although both species were detected in a variety of forest types and ages, and across the entire spatial

extent of our study, our occupancy models were primarily driven by the degree of forest disturbance.

We note that interpretation of dynamic multi-season depends on many factors, and in our effort to

collect widely spatially balanced data, we rotated through which areas were surveyed in successive

seasons. We feel this presents the most holistic picture of both spatial and temporal dynamics for

marten and fisher in this region, but acknowledge that no subset of data can capture all the information.

We did not find that deeper snow limited the distribution of fishers, nor that fisher were an important

limiting factor for martens. Both species had a positive relationship with the presence of deciduous trees

in forest stands. In northeastern forests, preserving forest stands with only moderate timber harvest

activities across both small and large scales, and ensuring that hardwood trees in particular are retained,

is likely favorable for marten and fisher population persistence. The strength of our findings derive from

the balanced natural experiment design of our study, which included replicated survey stations across all

levels of silvicultural intensity. The history of forested habitat in Maine is representative of much of

48
North America, from extensive clearing for agriculture and unsustainable timber extraction, followed by

abandonment and young forest regeneration, and now experiencing the effects of climate change, pest

outbreaks and modern timber harvest practices (FPA 2013). Further research designed specifically to

unravel broader community interactions among carnivores would be helpful to understand the full array

of factors driving marten and fisher distribution, and to better predict how these dynamics may be

impacted by ongoing anthropogenic land-use.

2.6 Acknowledgements

We acknowledge the indigenous Wabanaki Tribal Nations on whose homeland we reside, and hope our

research may help honor their legacy of stewardship for apanakes (marten), wlanikw (fisher), and the

integrity of forests in Maine. Our research was funded by the Maine Department of Inland Fisheries and

Wildlife, the USDA National Institute of Food and Agriculture (McIntire‐Stennis Project #ME0‐41913) and

the Cooperative Forestry Research Unit. Special thanks to W. Jakubas, C. Mosby, S. Webb and B. Roth.

We appreciate additional funding awarded by the Honors College Charlie Slavin Research Fund, and the

Center for Undergraduate Research at the University of Maine. We thank B. J. Cimpher, D. Nadjkovic, P.

Detwiler, W. Thomas, and M. Hess for their work in the field, and especially G. Archambault for

steadfast assistance. And our thanks to innumerable volunteers who helped tag image data.

Thank you to our reviewers for helpful critiques of earlier versions of this manuscript. Finally, this

work would not have been possible without permission to access land across the state, generously

granted by numerous public and private entities: American Forest Management, Inc.; Baxter State Park;

BBC Land, LLC; Downeast Lakes Land Trust; Elliotsville Plantation, Inc.; Huber Resources Corporation; JD

Irving Ltd.; Katahdin Forest Management, LLC; King and Bartlett Fish and Game Club; Maine Bureau of

Parks and Land; Maine Department of Inland Fisheries and Wildlife; The Nature Conservancy; North

Maine Woods, Inc.; Northwoods Management, LLC; Passamaquoddy Tribe of Indian Township; Seven

Islands Land Company; Tall Timber Trust; Wagner Forest Management; Weyerhaueser Company.

49
 CHAPTER 3: FOREST DISTURBANCE AND OCCUPANCY PATTERNS OF SHORT-TAILED WEASEL AND

LONG-TAILED WEASEL IN MAINE

3.1 Chapter abstract3

Weasels are the smallest members of the order Carnivora and act as predators of small mammals in a

wide array of natural and human-modified ecosystems. However, their response to anthropogenic

disturbance, such as forest harvest for timber products, is seldom the subject of focused research and

management projects. Both the short-tailed weasel (Mustela erminea) and the long-tailed weasel

(Neogale frenata) are native to Maine, USA, though detailed information on their distribution is scarce.

Commercial timber harvesting continues to be widespread in Maine, resulting in a mosaic of forest ages

and disturbance histories. This forest disturbance has many potential impacts on wildlife, including

unknown effects on weasel species. To contribute in filling this knowledge gap, we conducted a four-

year large-scale field study on the current occupancy patterns of both weasel species in Maine.

Specifically, our objective was to assess the effects of forest disturbance caused by timber harvest on

short-tailed and long-tailed weasels’ large-scale occupancy patterns. We collected data from 197 survey

sites, each composed of three camera traps, over four years and analyzed over 7,000 images of weasels.

We categorized each daily visit at a survey site by a) species, and b) certainty of identification. Using

false-positive occupancy models, we found that short-tailed weasels were widely distributed across the

state (confirmed at 54% of sites), while long-tailed weasels were rarer (confirmed at 16% of sites) and

more likely to be detected farther south. Both species respond positively to forest disturbance, with

higher probabilities of initial occupancy and colonization, and lower rates of local extinction as

Note on citation 3
This work has been submitted to The Journal of Mammalogy, coauthored with Dr. Alessio Mortelliti, and
is currently undergoing peer review. We ask that any readers interested in citing our findings please
locate the most up-to-date version. To receive a copy of the journal publication once it is available,
please visit: https://www.researchgate.net/profile/Bryn-Evans-2 or email: [email protected].

50
disturbance increases, especially at the larger landscape scale. Short-tailed weasel were more likely to

occupy stands with a higher percentage of conifer trees, while no such relationship was found for long-

tailed weasels. We conclude that current forest harvest practices in Maine are not detrimental to weasel

populations, but that the potential difference in forest composition preferences and the latitudinal

gradient of long-tailed distribution warrant continued monitoring. Further, we propose that camera

trapping—if designed to obtain detections of carnivore species, such as the transect of three cameras

per site reported here—is a viable tool to monitor both these species of weasels over the long-term.

3.2 Introduction

Weasels are small carnivores distributed across the globe from temperate to tropical forests, savannas,

and most habitats in between (Wilson and Ruff 1999, Feldhamer et al. 2003). In North America, three

species of weasel occur: the short-tailed weasel, also called ermine or stoat (Mustela erminea), the least

weasel or common weasel (Mustela nivalis), and the long-tailed weasel (Neogale frenata) (King and

Powell 2007, ASM 2021, Patterson et al. 2021). Maine, located in the northeastern United States, is

home to short-tailed and long-tailed weasels. Although the range of the least weasel borders the far

northwestern portion of the state, they are currently not thought to be present (Henderson 1994,

Sheffield and King 1994).

Although both the short-tailed and long-tailed weasel are classified as ‘least concern’ by the

IUCN, there are currently many knowledge gaps regarding the abundance and status of weasel

populations in the US (Helgen and Reid 2016, Reid et al. 2016, ski et al. 2021). Populations could be at

risk as climate change causes phenology to shift resulting in mismatching camouflage (Atmeh et al.

2018), and the increasing degree and intensity of human land use are causing global wildlife declines

(Schipper et al. 2008, Hoffmann et al. 2011, Di Marco et al. 2014). Recent work from across North

America indicates that weasel populations are trending downward, even as trapping effort diminishes

(Jachowski et al. 2021). Although studies on the influence of forest disturbance and regeneration

51
following both fire and timber harvest on habitat preferences suggest that weasels are highly adaptable

and use stands of any age class, data are limited for these species (Fisher and Wilkinson 2005). See

below for further notes on the biology of these species.

In Europe, the occupancy probability of least weasels was very low in urban areas (Łopucki et al.

2019); however, Pyšková et. al. (2018) found weasel species in many natural habitat types using camera

traps in central Europe. In a recent study in western North America (Linnell et al. 2017a), short-tailed

weasels selected for areas of relatively young forest—however, timber management strategies that

intensively replant stands can have reduced understory complexity, resulting in less cover and lower

small mammal densities (Ashton and Kelty 2018). Predation may play a confounding role in habitat

selection (Linnell et al. 2017b). Although it appears highly likely that forest type and timber management

affect both long-tailed and short-tailed weasel populations, very little is known on how these species

respond to large-scale intensive timber harvest (but see Linnell et al. 2017a). Our objective was to assess

the effects of forest disturbance caused by timber harvest on short-tailed and long-tailed weasels’

occupancy patterns over a large (state-wide) scale.

To accomplish our objective, we conducted a large-scale camera trapping field study, covering

the northern two-thirds of Maine (approximately 50,000 km2) over a four-year period. Following a

natural-experiment design, we balanced survey effort across a gradient of forest disturbance intensity,

latitude and forest types. Following King (1983) and Sheffield and Thomas (1994) we hypothesize that

both species will show a preference for intermediate levels of disturbance, as younger stands provide

abundant prey resources and the denser cover provides protection from predators (Fisher and

Wilkinson 2005, Reid et al. 2016). Forest management for timber and other resource extraction is a key

factor in shaping the quality of habitat for these species in Maine, as well as globally, with approximately

71% of forested land actively managed (IPCC 2019). We provide analyses of a four-year dataset,

52
spanning areas of low, moderate, and high forest disturbance, to help fill knowledge gaps on how these

wide-spread forest management practices may impact weasel populations.

3.2.1 Species biology

Both short-tailed and the long-tailed weasel are voracious, obligate carnivores that consume a wide

range of prey species including animals much larger than their own body size (King and Powell 2007).

They are generalist predators, although preferred prey species vary with body size from mice to

lagomorphs (Simms 1979, Lisgo 1999) due to both the pronounced sexual dimorphism within species,

with males larger than females, and between species. This variation in size is an important component in

the co-existence of otherwise similar species, as dramatic differences in body size allow individuals of

both species to have their own competitive advantages (Aldous and Manweiler 1942, Proulx and

Drescher 1993, King and Powell 2007). Weasels are considered habitat generalists, but research in North

America suggests that younger forest stands are preferred by short-tailed weasel (Gehring 2000, Hajduk

2008, Linnell et al. 2017a). Short-tailed weasels are circumboreal and found in most of northern North

America, distributed across Canada excepting the arctic and into the northern United States excepting

the prairie bioregions (e.g. eastern Montana and western North Dakota) (King and Powell 2007, Reid et

al. 2016). Long-tailed weasels are a new world lineage (Nyakatura and Bininda-Emonds 2012), hence the

proposed revision of the genus from Mustela to Neogale (Patterson et al. 2021). Their distribution

extends into Canada, across the United States (though potentially limited to areas with access to fresh

water [Henderson 1994, Johnston et al. 2019]) and into western South America (Helgen and Reid 2016).

Habitat needs vary across their range, but long-tailed weasels are thought to prefer later seral stages

than short-tailed (Wilson and Ruff 1999, King and Powell 2007) and to avoid heavily disturbed, open

areas (e.g. agricultural fields [Gehring 2000]).

53
3.2.2 Study area

We collected data across the northern two-thirds of Maine (Fig. 3.1). The area is remote and heavily

forested, and the majority of forested land (16.7 of the 17.6 million acres as of 2017) is managed for

commercial timber extraction (Butler 2018). Timber harvesting and natural disturbances have resulted

in a mosaic of mixed-age stands, though large-scale clear-cutting was reduced in 1989 (Maine Forest

Service 2013). Tree communities transition from the Eastern deciduous biogeographic region in

southern Maine, where maples (Acer sp.), birches (Betula sp.), and American beech (Fagus grandifolia)

are abundant, to the Boreal forest region in the north, where conifers are more dominant: balsam fir

(Abies balsamea), eastern hemlock (Tsuga canadensis), white pine (Pinus strobus) and spruces (Picea

sp.) (Gawler et al. 1996, McWilliams et al. 2005). The climate is classified as humid continental, with

warm summers (12 C to 25 C in July) receiving 55 cm of precipitation as rain, and cold winters (-16 C to -

5 C in January) receiving 51 cm of precipitation as snow (NOAA 2020). The geographic range of our data

collection was from 44.15968°N south to 47.42166°N north, and -70.55842°W west to -67.62582°W

east.

54
Fig 3.1. (A) Map of Maine, a state in the northeastern United States and (B) 197 camera trap survey sites
deployed from summer 2017 to summer 2020. (C) Each survey site was 6 km apart from any other and
(D) was made up of a linear transect of three camera stations spaced 100 m. (E) Each camera station was
comprised of a Bushnell E2/E3 trail camera, set facing beaver meat housed in a suet cage, and scent
lure.

55
3.3 Methods

3.3.1 Data Collection

We conducted large-scale camera trapping surveys from 2017-2020, deploying 197 sites across 7

consecutive summer (4) and winter (3) seasons. Survey sites were placed 6 km apart to maintain

independence (i.e., low likelihood of an individual weasel visiting two adjacent sites), and distributed

according to a natural experimental design (Garton et al. 2011) balancing the influences of forest

disturbance intensity, tree species composition, and latitude (See Appendix D.1). We used remote-

sensed forest disturbance data, obtained from Landsat images (Wulder et al. 2019) and calibrated to

New England forests (Kilbride 2018) to calculate a per-township score that combined the intensity and

timing of the most recent disturbance event (typically due to commercial timber harvest [Butler 2018]).

For more details on the forest disturbance index, see Appendix C. We selected survey sites to balance

data collection in townships ranging from the lowest disturbance scores to the highest, with a minimum

of twenty replicates in north versus south and in areas of high versus low regulated trapping activity

(MDIFW 2019). We then generated buffers of varying sizes around each site (300m, 1 km, 3km and 6km

radii) to understand how the spatial scale of disturbance influenced weasel occupancy patterns. Home-

range sizes for weasels vary greatly by species, sex and location (e.g. short-tailed weasel females range

from 3 to 80 ha, male long-tailed weasels from 20 to 200 ha [Hajduk 2008, Linnell et al. 2017a]). Because

animals respond to habitat features at multiple spatial scales, allowing competition among models at

several potentially important scales, such as 300 m radius buffer representing a core home range versus

a 6 km buffer the entire landscape for a weasel, optimizes the scale for each parameter that disturbance

may influence (McGarigal et al. 2016). We also assessed the percentage of deciduous trees (which is

inverse to the percentage of coniferous trees) at three local scales (stand, 300 m, and 1 km) to

determine if weasel species responded to forest composition (LANDFIRE 2019).

56
 Each survey site was composed of three camera stations, spaced 100 m apart in a linear transect

designed to optimize detection success for mesocarnivores in Maine (Evans et al. 2019). At each camera

station a Bushnell Trophy Cam E2/E3 passive infrared camera (Overland Park, KS, USA) was set to record

one 8-megapixel image per trigger, at normal sensitivity, with a 3 second delay between triggers (Burton

et al. 2015). Cameras were set on trees 20-50 cm from the ground or packed snow, facing a suet cage

containing a piece of beaver meat (Castor canadensis) and two, pea-sized dabs of Vaseline and skunk-

based furbearer lure (J. Braley, Kenduskeag, ME, USA), one at 20-40 cm and another at 2-3 m (Schlexer

2008, Stewart et al. 2019, Buyaskas et al. 2020). Our pilot work conducted in the area showed that this

combination maximized the detection of mustelids (Buyaskas et al. 2020). Camera sites were first

deployed in a summer season for more than two weeks, and then redeployed the following winter

season for more than two weeks. A subset of camera sites were then repeatedly revisited in the

following summer, then winter, up to seven total deployments (Fig. 3.1B). We abided by the research

ethics standards of the American Society of Mammalogists and the University of Maine Institutional

Animal Care and Use Committee (Sikes 2016; IACUC Protocol #A2018-05-06).

3.3.2 Image Analysis

Weasels are elusive and cryptic animals, and where both short-tailed and long-tailed weasel co-occur,

there is potential to misidentify one species for the other. They can also be difficult to capture on

camera traps (Johnston et al. 2019), and so we analyzed our image data using false-positive occupancy

models which allowed us to tackle both the issue of potential misidentification and also correct for the

potential of false absences in our dataset (Miller et al. 2011, MacKenzie et al. 2017). These models allow

the detection history data to be classified as “certain” or “uncertain” based on its quality, even when

using one survey method (e.g. amphibian call surveys, Miller et al. 2011).

We categorized image data by first reviewing all photographs recorded and assigning a “tag” for

the trigger cause. We then reviewed all images tagged as weasel in detail (further information is

57
provided in Appendix D.4): 1) First we classified the image quality, based on both clarity and body

position of the animal, and 2) whenever possible we measured tail length and body plus head length (in

pixels) to calculate their ratio (Elsasser and Parker 2008). 3) We recorded any qualitative notes regarding

the apparent tail and body lengths, apparent size of animal, size relative to the suet cage, and any

distinctive features, and 4) assigned each image an initial assessment of both species and confidence

level (Fig 3.2). 5) We grouped all images recorded within 30-minutes together as a single detection

event, and used the highest confidence initial assessment for the entire group: short-tailed certain or

uncertain; long-tailed certain or uncertain; least weasel uncertain; or weasel unknown. 6) Detection

histories for each species were generated by pooling all events recorded in a 24-hour period at any of

the three cameras at a site and using the highest level of certainty, if multiple events had occurred

(Appendix D.4).

58
Fig. 3.2. Examples of trail camera images of weasels recorded in Maine, USA and categorized by image
quality; by suspected species, short-tailed weasel (Mustela erminea) or long-tailed weasel (Neogale
frenata); and by confidence level. (A) The image quality was only considered fair, as a slightly blurry
night image, but body posture was excellent for estimating tail length to body (plus head) length: ratio
37:94 pixels = 0.39, resulting in a ‘certain short-tailed weasel’ initial assessment for the image. (B) This
image quality is a good, crisp night image, and body posture is good for measurements: 91:137 = 0.66,
‘certain long-tailed weasel’ initial assessment.

3.3.3 Occupancy Models

Data were analyzed using occupancy models (MacKenzie et al. 2003, 2017) to account for imperfect

detection (false-negative results) and incorporate habitat covariates. Specifically, we used multi-season,

false-positive models in Program Presence (Hines 2015) to account for the possibility of species

misidentification, which can bias estimates of occupancy, colonization and extinction (Royle and Link

2006, Miller et al. 2013). This potential for false-positives could introduce bias if not accounted for

(Chambert et al. 2015, Clement 2016, Rojas et al. 2019), while eliminating a large proportion of data

59
points (i.e. using only certain detections) could reduce sample size and create another host of concerns

(MacKenzie et al. 2017). The detection data are presented as 0 (no detection), 1 (uncertain detection)

and 2 (certain detection). The detection process is then modeled as p1 = the probability of a correct

detection (at a site that is occupied), p10 = probability of an incorrect detection (at a site that is

unoccupied) and b = the probability of at least one certain detection at an occupied site (Miller et al.

2011).

We sequentially ran 1) a set of a priori models for detection p1, while holding other parameters

fixed, and selected the top-ranked model via Akaike Information Criterion corrected for small samples

sizes (Burnham and Anderson 2002). We then included this covariate (or covariates if multiple models

were within 2ΔAICc and the additive model then ranked highest) to repeat the process for 2) initial

occupancy, 3) colonization and 4) extinction probabilities. Covariates included the average forest

disturbance index at 300 m, 1 km, 3 km, and 6 km radii circular buffers; the proportion of deciduous

trees out of all live trees at the stand level, at 300 m and 1 km buffers; winter snow depth; latitude; and

days into the survey period. We estimated model fit using the equivalent top models in R using packages

unmarked (Chandler et al. 2021) and AICcmodavg (Mazerolle 2020), using only the certain detection

data to calculate Nagelkerke’s R2.

3.4 Results

3.4.1 Image Data

Of the 197 survey sites established, 136 were active over two seasons (a summer and the following

winter), and 57 were active for a range of three to seven seasons, while four sites were active for only a

summer season (winter access issues). 121 sites were first set in summer 2017, 48 in summer 2018, and

28 in summer 2019. We recorded over 800,000 individual images resulting from motion triggers, and of

these 7,473 were tagged as triggered by weasels. Our final dataset contained 593 short-tailed weasel

events (350 certain, 243 uncertain) and 112 long-tailed weasel events (54 certain, 58 uncertain). There

60
were no images that appeared to be of least weasels, and 45 detection events could not be classified

beyond unidentifiable weasel and were thus excluded from analyses. Overall, short-tail weasel were

detected (with a certain image in at least one season) at 107 out of 197 sites and with uncertain images

only at another 22 sites, and long-tailed weasel were detected with certain images at 32 sites, and with

only uncertain images at another 23 sites.

3.4.2 Occupancy Analyses

The top ranked false-positive, multi-season model for short-tailed weasel included the forest

disturbance index for all three components of occupancy: initial probability of presence (at 6 km radius

buffer, positive relationship), probability of colonization (at 3 km buffer, positive relationship) and

probability of extinction (at 300 m buffer, negative relationship) (Table 3.1, Fig 3.3, Appendix D.5 Table

D.6). Detection of short-tailed weasels included the days in the survey (negative relationship) and

disturbance (positive relationship). For long-tailed weasel, top ranking models included disturbance for

initial occupancy (negative relationship) and for extinction (positive relationship), and proportion of

deciduous trees for initial occupancy (at the stand scale, positive relationship) and for colonization (at

300 m radius scale, positive relationship) (Table 3.1, Fig 3.4, Table D.7). The detection probability for

long-tailed weasel was affected by latitude (lower probability of a certain detection at higher latitudes)

and by disturbance at a 300 m scale (positive relationship).

The top ranked model for short-tailed weasel had R2 = 0.51, while for long-tailed weasel R2 = 0.20.

61
Table 3.1. The top ranking AICc models (within delta = 2) for multi-season, false-positive occupancy
models of short-tailed weasel (Mustela erminea) and long-tailed weasel (Neogale frenata) in Maine,
USA. Camera trap images from 197 sites over seven seasons were included. ψ is the probability of
occupancy in the first season; γ is the probability of a colonization between seasons; ε is the probability
of extinction; p1 is the detection probability given that a site is occupied (correct detection); p10 is the
probability of an incorrect detection, and b is the probability of obtaining at least one correct detection
at an occupied site. ΔAICc is the difference in each models’ Akaike Information Criterion from the value
for the top model, Wi is the individual model weight, and k is the number of parameters. DI =
Disturbance intensity, DC = deciduous trees, Lat = latitude (all z-scored).

Species Model formula ΔAICc Wi k

Short-tailed weasel
ψ(DI.6k+DC.1k), γ(DI.3k+DC.300m), ε(.), p1(Days+Lat), p10(.), b(.) 0.00 0.1704 12
ψ(DI.6k+DC.1k), γ(DI.3k+DC.300m), ε(DI.300m), p1(Days+Lat), p10(.), b(.) 0.50 0.1327 13
ψ(DI.6k+DC.1k), γ(DI.3k+DC.300m), ε(Lat), p1(Days+Lat), p10(.), b(.) 1.06 0.1003 13
ψ(DI.6k+DC.1k), γ(DI.3k+DC.300m), ε(DC.300m), p1(Days+Lat), p10(.), b(.) 1.61 0.0762 13
ψ(DI.6k+DC.1k), γ(DI.3k+DC.300m), ε(DI.300m+Lat), p1(Days+Lat), p10(.),
1.75 0.0710 14
b(.)
ψ(DI.6k+DC.1k), γ(DI.3k+DC.300m), ε(DC.stand), p1(Days+Lat), p10(.), b(.) 1.78 0.0700 13
ψ(DI.6k+DC.1k), γ(DI.3k+DC.300m), ε(DI.1k), p1(Days+Lat), p10(.), b(.) 1.95 0.0643 13
Long-tailed weasel
ψ(DI.6k), γ(DI.300m+DC.300m), ε(DC.stand), p1(Lat+DI.300m), p10(.), b(.) 0.00 0.1011 12
ψ(DI.6k), γ(DI.300m+DC.300m), ε(.), p1(Lat+DI.300m), p10(.), b(.) 0.51 0.0783 11
ψ(DI.6k), γ(DI.6k), ε(.), p1(Lat+DI.300m), p10(.), b(.) 0.87 0.0654 10
ψ(DI.6k), γ(DI.300m), ε(.), p1(Lat+DI.300m), p10(.), b(.) 0.87 0.0654 10
ψ(DI.6k), γ(DI.1k), ε(.), p1(Lat+DI.300m), p10(.), b(.) 0.87 0.0654 10
ψ(DI.6k), γ(DI.3k), ε(.), p1(Lat+DI.300m), p10(.), b(.) 1.33 0.0520 10
ψ(DI.6k), γ(DI.300m+DC.300m), ε(DC.stand+Lat), p1(Lat+DI.300m), p10(.),
1.49 0.0480 13
b(.)
ψ(DI.6k), γ(.), ε(.), p1(Lat+DI.300m), p10(.), b(.) 1.58 0.0459 9
ψ(DI.6k), γ(DI.300m+DC.300m), ε(Lat), p1(Lat+DI.300m), p10(.), b(.) 1.88 0.0395 12
ψ(DI.6k), γ(DC.300m), ε(.), p1(Lat+DI.300m), p10(.), b(.) 1.92 0.0387 10

62
Fig 3.3. [Next page] Results of multi-season, false-positive occupancy modeling for short-tailed weasels
(Mustela erminea) detected via camera trapping in Maine, USA. (A) The probability of a certain
detection event decreased over the length of the camera deployments, and (B) was higher at higher
latitudes. (C) Initial occupancy was higher at sites in areas where the landscape-scale disturbance was
higher and (D) was lower in areas with greater proportions of deciduous trees. (E) The probability of
colonization was higher in disturbed stands and (F) lower in more deciduous dominated stands. (G) The
probability of extinction was lower in stands that had greater disturbance at the weasel home range
scale and (H) higher at higher latitudes.

63
64
Fig 3.4. Results of multi-season, false-positive occupancy modeling for long-tailed weasels (Neogale
frenata) detected via camera trapping in Maine, USA. (A) The probability of a certain detection event
was higher at sites with more disturbed forest at the scale of weasel home ranges and (B) lower at
increasing latitudes. (C) Initial occupancy increased with disturbance at the landscape scale, while the
probability of colonization was (D) negative with disturbance and (E) positive with deciduous proportion
at a local scale.

65
3.5 Discussion

Through a four-year large-scale study conducted in Maine, in the northeastern United States, we found

that occupancy patterns for both short-tailed and long-tailed weasel were positively associated (i.e.,

higher probability of initial occupancy, and more likely to colonize while less likely to become locally

extinct) with areas that have experienced forest disturbance, with the strongest relationships found for

larger spatial scales. Short-tailed weasels were distributed across the entirety of our study area (the

northern two-thirds of the state) and detected at 54% of sites, while long-tailed weasels were less

common, detected at only 16% of sites, and more likely farther south.

3.5.1 Habitat relationships

3.5.1.1 Disturbance.—Our top ranked multi-season models included variables describing the level of

forest disturbance, driven by the intensity and recency of timber harvesting, for multiple components of

weasel occupancy. The initial probability of presence, which describes the likelihood that a species

occupies a site during the first season surveyed, was positively associated with disturbance at a 6 km

radius buffer for both short-tailed and long-tailed weasels. Furthermore, the probability of colonization,

which is the likelihood that a site that was previously unoccupied becomes occupied between seasons,

was also positively associated with disturbance for both species (at the 3 km scale for short-tailed and

300 m for long-tailed). In addition, for short-tailed weasel, extinction—the likelihood an occupied site

becomes unoccupied—was negatively associated with disturbance at the 300 m scale. Taken together

these results suggest that weasel populations are more stable in forested areas experiencing

disturbance—that is, they are less likely to become locally extinct and more likely to colonize if an area is

not already occupied—than in undisturbed stands. This relationship probably reflects the increased

availability of prey in regenerating stands and the accessibility of ground cover and slash piles for

foraging and for cover (Lisgo 1999, Fisher and Wilkinson 2005). Young stands may also provide

protection from predation by larger mustelid carnivores, such as American marten (Martes americana)

66
and fisher (Pekania pennanti), that prefer mature forests (Palomares and Caro 1999, Zielinski et al. 2004,

Cheveau et al. 2013, Evans and Mortelliti 2021 submitted). There may be a less powerful positive

relationship between disturbance at small spatial scales, as indicated by the less dramatic slopes and

wider confidence intervals, and between-season turnover. Recent work (Sullivan et al. 2017) suggests

that maintaining windrows to connect unharvested forest patches, and keeping complex woody

materials (slash piles) on the landscape, could maximize the habitat quality for both prey and predator

species in commercial forests. Maine has forestry laws in place which minimize the total area of

clearcutting and require an equally wide “separation zone” between clearcuts, which results in fewer

large blocks of entirely mature forest but does maintain connectivity and heterogeneity (Maine Forest

Service 2013). This type of harvest configuration on the landscape may also drive the positive trend

between weasels and areas that, at large scales, have experienced recent forest disturbance.

3.5.1.2 Forest composition.—Short-tailed weasels had slightly lower probability of initial presence at,

and were also less likely to colonize, sites with higher proportions of deciduous trees versus conifers.

Long-tailed weasels also included the deciduous parameter in the top-ranking model for colonization,

but the relationship is only very weakly negative. Other studies have found a preference for conifer

stands for short-tailed weasels (Malcolm et al. 2004, Linnell et al. 2017a), which seems supported by our

findings as well. Regenerating conifer stands in Maine have been shown to provide excellent habitat for

many small mammal prey species (Homyack et al. 2005). Although long-tailed weasels also may prefer

more closed-canopy forests (Sheffield and Thomas 1997), their relationship with forest composition is

not as clear from the literature. However, if these two species, which are jointly managed in Maine, do

have opposing habitat needs, further research may be warranted.

3.5.2 Large-scale distribution and co-existence patterns

We found short-tailed weasels at 54% of survey sites, and long-tailed weasel at only 16% of sites. We

documented occasional instances of spatial overlap for the two species (13 events at 12 different sites),

67
but generally the long-tailed weasel were more likely to occupy the southern parts of our study area,

while short-tailed weasel were found across its entirety. Maine is near the northern range boundary for

long-tailed weasels, thus a latitude trend is consistent with existing literature (Sheffield and Thomas

1997). Indeed, the observed patterns likely reflect the northern range limit of long-tailed weasels

(Simms 1979, Helgen and Reid 2016) , which are widely distributed across southern North America but

do not extend far into Canada, rather than a consequence of human-induced land-use change. Though

long-tailed weasel were relatively rare compared to short-tailed weasel in our surveys, as the winters

continue to warm we may see their populations increase (Simms 1979, Suffice et al. 2017), warranting

continued monitoring of weasel species in Maine. Monitoring is also important to detect if least weasels,

which are not currently present in Maine, may expand into the state as the climate becomes more

suitable for these small-bodied species (Sheffield and King 1994).

3.5.3 Detectability of weasels

Detection probabilities for weasel were generally low, with a mean daily detection probability of 0.09 for

short-tailed weasels (though as high as 0.15 on the first day of surveys) and 0.03 for long-tailed weasel.

The estimated probability of obtaining at least one, high-certainty image at a site where short-tailed

weasels were present was 0.62, and long-tailed weasel was 0.56, suggesting that nearly half of sites

could fail to detect a species over the course of a survey (33 days). A useful future research direction

may be optimizing camera-trapping techniques to improve the likelihood of recording these species

where they are present in managed landscapes. Recent studies have tested the use of a specialized

camera traps housed within a box and accessible via tubes as a means to target weasels (Mos and

Hofmeester 2020, Croose et al. 2021). Although these “Mostela” devices may be helpful for collecting

detailed information on weasels (sex and individual identification for least weasels [Mos and

Hofmeester 2020]), it seems that external cameras were more effective for detecting short-tailed

weasels in Ireland (Croose et al. 2021). Single camera traps may have highly variable, and often un-

68
explained, detection probabilities based on micro-habitat at the site where they are deployed (Kolowski

et al. 2021). Increasing the number of camera traps in a single array, as our study protocol does, may be

a beneficial practice (Pease et al. 2016, Evans et al. 2019), but further research should be done to

optimize the configuration and design specifically for weasels.

3.5.4 Management Implications

Because of recent concern over long-term declining trends for both short-tailed and long-tailed weasels

(Jachowski et al. 2021), gaining more accurate insight into the current occupancy status of weasels

should be a priority for wildlife management and research. Although both species of weasel showed

high probability of occupancy in stands that have experienced recent disturbance, they do depend on

the ability to alternate between prey species and require a mosaic of habitat patches (Gustafson et al.

2007). We found that habitat preferences varied between the two species, with short-tailed weasel

more stable (i.e., less likely to become locally extinct) in stands containing higher proportions of

deciduous trees and the reverse for long-tailed weasels. Managers should thus maintain a diversity of

tree species, both deciduous and coniferous, in managed forest to provide habitat for both these weasel

species.

Besides revealing patterns of weasel occupancy in relation to land-uses, our data also shows

how camera traps can be an effective method to monitor weasels (Mortelliti et al. undergoing revision),

particularly if data are analyzed in a false-positive framework (Miller et al. 2011). Long-tailed weasel

were less-frequently detected with certainty (50% of detections classified as certain, versus 70% for

short-tailed weasel), thus introducing the potential for bias if all data points are treated as equal (see

Miller et al. 2013). We suggest managers invest in long-term monitoring projects to generate large,

continuous data sets as these may be essential for understanding population trends for this species.

69
3.6 Acknowledgements

We offer our appreciation of the indigenous Wabanaki Tribal Nations on whose ancestral homeland we

have conducted this study, and acknowledge their stewardship of the land and wildlife. This research

was funded by the Maine Department of Inland Fisheries and Wildlife, the USDA National Institute of

Food and Agriculture (McIntire‐Stennis Project #ME0‐41913) and the Cooperative Forestry Research

Unit, with additional support from the Honors College Charlie Slavin Research Fund and the Center for

Undergraduate Research at the University of Maine. Sincere thanks to B. J. Cimpher, D. Nadjkovic, P.

Detwiler, W. Thomas, G. Archambault and M. Hess for their assistance in the field, and to the students

who volunteered to tag image data. Our thanks to the many public and private landowners that granted

access for our research.

70
 CHAPTER 4: AMERICAN BLACK BEAR OCCUPANCY AND REPRODUCTION IN MANAGED TIMBER

FORESTS OF MAINE

4.1 Chapter abstract4

Species respond to human disturbance and land-use change in myriad ways, both negative and positive,

and these vary over space and time. The American black bear (Ursus americanus) is an adaptable

generalist in many areas where humans and wildlife coincide, but the longer-term impacts of intensive

forest harvest, particularly on sensitive demographic stages, are still worthy of investigation. Much of

Maine, in the northeastern United States, is covered in forested habitat that is hospitable to bears—and

to commercial timber harvest. To look at the potential impacts of differing intensities of disturbance on

black bears, and particular on sows with cubs, we collected camera trap data from 197 stations

dispersed over representative forest stands in north and central Maine. By using multi-state occupancy

models, we could distinguish the overall trends of adult bears from bears that are currently

reproductive. We found that forest disturbance in small patches benefits both demographic groups, and

that the availability of hardwood trees was an important driver for bear reproduction.

4.2 Introduction

American black bears (Ursus americanus) are widely distributed and abundant across much of North

America, having successfully recovered in many areas that experienced near-extirpation following

European colonization of the continent (Garshelis et al. 2016). Although in some regions of the United

States, particularly in the south, they are considered threatened due to small, isolated populations

(Servheen et al. 1998, 2005), in most states black bear populations are stable or increasing (Garshelis

Note on citation 4
This chapter is being prepared for submission to The Journal of Wildlife Management coauthored with
Amay Bolinjkar, Griffin Archambault and Dr. Alessio Mortelliti. We ask that any readers interested in
citing our findings please locate the most up-to-date version. To receive a copy of the journal publication
once it is available, please visit: https://www.researchgate.net/profile/Bryn-Evans-2 or email:
[email protected].

71
2016). While the ecology and behavior of bears is well studied (Pelton 2003), one area that warrants

further investigation is the factors affecting their reproduction and how these are influenced by

landscape change (Noyce and Garshelis 1994; Laliberte and Ripple 2004).

Because bears are slow to reach sexual maturity and females have limited reproductive

potential (Pelton 2003; Garshelis 2016), understanding the factors behind successful reproduction is

important for the long-term management of bears. An abundant supply of high-quality food during the

fall hyperphagia period prior to hibernation is important for female bears to produce cubs the following

spring (Elowe and Dodge 1989; Samson and Huot 1995; Pelton 2003). Previous work suggests that hard

mast from deciduous tree species is linked to several components of reproduction (e.g. age of first

reproduction, proportion of adult females producing litters). In areas of Maine where American beech

(Fagus grandifolia) is the dominant hardwood, the cycle of alternate-year masting leads to synchronous

reproductive success or failure in the population (McLaughlin et al. 1994; Schooley et al. 1994; Jakubas

et al. 2005). This sensitivity to hardwood tree species abundance and diversity warrants attention

particularly in the northeast US: as an example, Maine has already lost one important hardwood species

entirely (American chestnut [Castanea dentata] becoming ecological extinct in the early 1900’s [Paillet

2002]) and managed forests may increasingly shift towards conifer species as a result of human land-use

and economic interests (Nelson et al. 2011). Thus understanding the sensitivity of black bears, and

especially black bears with cubs, to the current—and future—availability of hardwoods is an important

issue for ecologist and managers.

Although bears are habitat generalists and well adapted to human-modified landscapes, the

long-term consequences of such extensive land-use change on the species, and particularly on

reproduction, represents a knowledge gap. Previous work has primarily focused on the effects of

silvicultural practices on several facets of bear behavior (Ulev 2007, Costello et al. 2003, Elowe and

Dodge 1989), however there is limited knowledge regarding the specific impact of silvicultural activities

72
on the reproductive behavior of black bears in managed forests (Koehler and Pierce 2003). While some

aspects may be positive (increased access to food being a key factor), some studies have found

avoidance of recently cleared areas by black bears (Samson and Huot 1998, Etcheverry et al. 2005) and

preference for older mature stands by denning females (Brodeur, Ouellet and Fortin 2008) and areas of

greater cover (Vitale et al. 2018). Our goal is to contribute in filling the knowledge gap regarding the

degree to which forest disturbance itself, versus the composition of the forest stand, may impact

reproductive black bears. It is clear from the literature that food availability, particularly in Maine, is

linked to fall-masting tree species, so our specific objective is to determine how large-scale, long-term

forest disturbance resulting from commercial timber harvesting, and the forest composition (e.g., the

proportion of deciduous to coniferous trees) together impact the occupancy and reproductive patterns

of black bears.

Our study was conducted in Maine, a heavily forested state in the northeastern U.S. Here, black

bears occur in all but the farthest south (and most urbanized) areas, and are managed as a big game

species (McLaughlin 1999). Maine has an extensive bear research and management program, including

educational programs, managing beech tree stands (an important food source), population monitoring

through hunter data including age of harvested bears (since 2010), and ongoing radio-collared research

(see Schooley et al. 1994; Seger et al. 2013; Vitale et al. 2018). Survival of adult bears is high, with the

leading cause of mortality over the age of four years being legal hunter harvest (McLaughlin 1999). Black

bear hunting is economically important in Maine, a state that relies heavily on recreation and tourism of

it natural resources (Honey et. al. 2016; MDIFW 2017). Black bear populations in the state have been

stable for several decades (McLaughlin 1999) and the heavily forested nature of the state provides an

excellent natural experimental design to address our study questions. The majority of forests in Maine

are managed for timber and other resource extraction—of the 17.6 million acres of forestland in the

state, 16.7 million acres were designated as commercial land in 2017 (Butler 2018). We designed our

73
surveys to cover a gradient of forest disturbance intensity from the very lowest (e.g. Baxter State Park)

to the very highest (recently harvested commercial timber properties), and replicated study sites across

multiple geographic regions.

To study large-scale occupancy patterns of reproductive black bears we used motion-triggered

trail cameras (here after camera traps), which have become an increasingly popular and cost-effective

tool for researching and managing terrestrial wildlife (Rowcliffe and Carbone 2008; Burton et al. 2015).

The use of camera traps to understand species occurrence across large spatial and temporal scales is

expanding, and they have been used to study bear species (Kelly and Holub 2008; Fisher et al. 2014;

Steenweg et al. 2016; Iannarilli et al. 2021), and even including occupancy of reproductive females

(Fisher et al. 2014). Camera trapping data has the advantage of being suitable for working within the

occupancy framework, which allows to account for false-absence in the detection histories (MacKenzie

et al. 2002, 2017); and to estimate the probability of presence of a species in multiple states, such as in

reproduction or not. Although careful study design and interpretation is important when using camera

traps in continuous habitat (Efford and Dawson 2012; Neilson et al. 2018) recent work conducted on

black bears in New Mexico has validated the use of occupancy models to assess habitat use for even

these highly-mobile species (Gould et al. 2019). Through this work we aim to contribute to filling

knowledge gaps on the effects of forest composition and disturbance on the occupancy, and

reproduction, patterns of American black bears in Maine through the combination of camera traps and

dynamic multi-state models occupancy models (MacKenzie et al. 2009, 2017).

4.3 Methods

4.3.1 Study area

Our study took place in Maine, in the northeastern United States, which is in a humid

continental climate regime with warm summers and cold winters, and falls at the junction of temperate

forest and boreal forest ecoregions (Fig. 1) (Hasbrouck and Connors 1987; Gawler et al. 1996; NOAA

74
2020). Our data was collected between 44.15968°N and 47.42166°N latitude, and -70.55842°W and -

67.62582°W longitude, covering all of the northern region of the state and much of the interior region

(Fernandez et al. 2020). In this area, temperatures average 18.6C in July (the warmest month, with

averages ranging from 15.9C to 21.1C from 1895 to 2021) and average -10.2C in January (the coldest

month, averages from -15.7C to -5.2C) (NOAA 2021). Precipitation falls as rain in summer (55 cm

average) and as snow in winter (51 cm average) (NOAA 2020).

The landscape is primarily forested—with many areas experiencing clearing and subsequent re-

forestation since the 1700’s—and the majority is now managed for commercial timber production

(Maine Forest Service 1995; Butler 2018). Timber harvesting and natural disturbances have resulted in a

mosaic of mixed-age stands, and tree communities often contain: maples (Acer sp.), birches (Betula sp.),

and American beech (Fagus grandifolia) as dominant hardwoods and white pine (Pinus strobus), firs

(Abies sp.), eastern hemlock (Tsuga canadensis), and spruces (Picea sp.) as dominant conifers (Gawler et

al. 1996; McWilliams et al. 2005).

75
Fig 4.1. Maine, located in the far northeastern United States (A) and the survey sites visited with
transects of trail cameras from 2017 to 2020 (B). Orange circles indicate sites that were first set in
summer 2017, pink squares were first set in 2018, and purple triangles set in 2019. The numbers
indicate the total number of visits to each site (1-4).

76
4.3.2 Study design and data collection

We followed a natural experimental design while selecting locations to conduct trail camera surveys. We

balanced survey effort along a gradient of forest disturbance intensity, measured from Landsat image

data processed using novel change-detection algorithms developed specifically for New England forests

(Kilbride 2018; Wulder et al. 2019) (Appendix C). Working at the township scales, we selected areas that

were representative of both extremes (heavily harvested forest with the highest disturbance scores,

versus unharvested areas with the lowest scores) and of intermediate values of disturbance. We

included areas of both large contiguous areas of a similar disturbance score (three or more townships)

and in smaller, disjunct areas (only one-two townships of that disturbance intensity, surrounded by

either all higher or all lower scoring townships) and then replicated these criteria in both the north and

south of our study area. Once townships were selected for surveys, we randomly generated target

points to establish camera sites (for further details on our balanced study design see Evans and

Mortelliti in press,).

Camera sites were set a minimum of 6 k apart, and each site was composed of a transect of

three camera ‘stations’. Stations were comprised of one Bushell Trophy Cam E2/E3 (Overland Park, KS,

USA), attached to a live tree and set facing a clear area, with bait (beaver [Castor canadensis] meat)

enclosed in a suet cage and Vaseline-based scent lure (J. Braley, Kenduskeag, ME, USA) applied to a tree

next to the bait directly across from the camera (height 20-40 cm) as well as 2-4 m up the bait tree

trunk. These stations were spaced in a linear transect 100 m apart, and followed the survey design that

earlier work found to maximize carnivore detection success in Maine (Evans et al. 2019; Buyaskas et al.

2020). All our methods complied with the American Society of Mammalogists standards for ethical

research (Sikes 2016), and were approved by the University of Maine Institutional Animal Care and Use

Committee (IACUC Protocol #A2018-05-06).

77
 Survey sites were first deployed in summer 2017 for a minimum active period of two weeks. In

the following summers 2018-2020, a subset of these locations were “revisited” to collect data across

multiple years at the same locations, and in summers 2018-2019 new locations were also set to expand

geographic coverage (Table 4.1). Winter data was not included in this study because black bears in

Maine typically hibernate from November to April (Schooley 1990). Our allocation of survey effort

followed the rotating panel design described by MacKenzie et al. (2017, pg. 218) to collect both the

spatially independent site information and the time-series data needed to fit dynamic occupancy

models.

Table 4.1. Survey effort across four summers in Maine, USA. Each site is comprised of three camera
traps, and of the 197 total sites, 138 were deployed for one summer, 3 for two summers, 29 for three
summers, and 27 for four summers.
Summer
2017 2018 2019 2020
Single visit 80 30 28 0
1st visit 41 18 0 0
2nd revisit - 41 18 0
rd
3 revisit - - 40 16
4th revisit - - - 27
Total sites per summer 121 89 86 43

4.3.3 Analyses

All images we recorded were tagged to the level of species (or other cause of the motion-trigger event,

e.g. false-triggers or crew while setting up and taking down cameras), and all images containing bears

were further reviewed and classified by the number of individuals present and the ages detected. For

each 24-hour occasion at each camera site, one of three possible events was recorded: 0 = no bears

detected, 1 = at least one bear detected, 2 = reproduction detected, by the presence of at least one

confirmed cub. From these detection histories the probabilities of detection (p1, p2, δ), the true initial

78
occupancy state (Ψ and R), and the likelihood of changing from one state to another can be modeled.

Table 4.2 describes these parameters in detail.

Table 4.2. Parameters of dynamic multi-state occupancy models applied to American black bear
detection histories from 197 sites over four summers in Maine, USA. In each summer season, a site can
be in one of three states: 0, no bears present; 1, only adult bears present, no reproduction; and 2, bear
cubs present, reproduction occurring.
Parameter Description

Ψ Probability a site is occupied by adult, non-reproductive bears in the first season

R Probability a site is occupied by reproductive bears in the first season
CΨ0 Probability of local colonization by adult bears only (state 0→1)
CΨ1 Probability of constant adult occupancy (state 1→1)
CΨ2 Probability of change in state, going from occupied with reproduction to occupied by adult
bears only (state 2→1)
CR0 Probability of local colonization by reproductive bears (state 0→2)
CR1 Probability of site experiencing a change in state, going from adults only to reproduction
(state 1→2)
CR2 Probability of constant reproductive occupancy (state 2→2)
p1 Probability of detecting adult bears on an occasion during a season when the true state is 1.
p2 Probability of detecting adult bears without cubs (no indication of reproductive activity) on
an occasion during a season when the true state is 2.
δ Probability of detecting cubs (indicating reproductive activity) on an occasion during a
season when the true state is 2.

The models assume that false-positives do not occur, such that an adult bears are never observed when
the true state is 0, and cubs are never observed when the true state is 0 or 1.

We conducted multi-phase modeling to first incorporate any important covariates into the

detection processes, then the initial state, and finally into the transition probabilities. In the first phase

we compared p1 and p2 as constant or varying, and then tested site level covariates (disturbance at all

spatial scales, hardwood at all spatial scales, latitude of the site, and site specific habitat features, see

SI1) as well as a yearly covariate, and ranked the resulting models via Akaike Information Criterion

79
corrected for small samples sizes (Burnham and Anderson 2002). We repeated the process for δ, the

probability of detecting reproduction. In the second phase, we included the top ranking covariates for

the detection processes and ranked models including covariates for first the probability of initial

occupancy by adults only, and then the probability of initial reproduction (covariates included were

disturbance at all scales, proportion hardwood at all scales, and latitude). In the final phase, we tested

the transition probability CΨ and CR against the same covariates as initial state and ranked these models

to determine what habitat covariates influences any changes in occupancy state between seasons. We

conducted step-wise model ranking via Aikake Information Criterion corrected for small sample sizes

(Burnham and Anderson 2002) to select important covariates affecting each component of dynamic,

multi-state occupancy models.

4.4 Results

We collected trail camera data from a total of 197 sites, active for 15-33 days (mean = 26.2 ± 4.6), during

one to four summer seasons from 2017 to 2020 (Table 2). We cataloged over 800,000 images, and

identified bears in over 65,000. Our final dataset comprised 1,015 detections of only adult bears (a

detection defined as at least one observation of an adult bear without cubs, at any of the three cameras

in a site, within a 24-hour period) and 106 detections of confirmed cubs (36 events with one cub

observed, 60 events with two cubs observed, and 10 events with three cubs observed). Adult bears were

detected at least once at 168 out of 197 sites (85%) and bears with cubs were detected at 52 sites (26%).

4.4.1 Occupancy models

The results shown in Table 4.3 are from the final stage and resulting top model produce by our stepwise

multi-state, dynamic occupancy model selection process. Testing for spatial autocorrelation on a single-

state (bears of any reproductive status) occupancy model via spline correlograms (R package ncf,

Bjornstad 2018) indicated the lack of spatial dependence in our dataset (Appendix D.6).

80
Table 4.3. AIC model ranking for the final step of parameterizing dynamic, multi-state occupancy models
of American black bear in Maine, USA. We compared the importance of different habitat covariates on
predicting the probability of a site with reproduction transitioning in the next summer to any of three
state: CR(0,1,2) where state 0 is no bears present, state 1 is only adult bears and state 2 is reproductive
bears. Model ranking was performed in earlier steps on Ψ, the probability of adult bear occupancy in the
first season; CΨ, the probability of transitions from adult bears to other states; R, the probability of
reproduction in the first season; δ, the probability of detecting reproduction; p1, the probability of only
detecting adult bears when the true state is only adult bears; and p2, the probability of only detecting
adult bears with the true state is reproductive bears. ΔAICc is the difference in the corrected AIC from
the top ranking model, AICwgt is the model weight, and k is the number of parameters.

Model ΔAICc AICwgt k

Ψ(Dist_300m), CΨ(.), R(HW_Stand), CR(.,Dist_1k,zlat), δ(Dist_300m), p1(year), p2(year) 0 0.266 17

Ψ(Dist_300m), CΨ(.), R(HW_Stand), CR(.,Dist_1k,.), δ(Dist_300m), p1(year), p2(year) 0.6 0.197 16

Ψ(Dist_300m), CΨ(.), R(HW_Stand), CR(.,Dist_300m,.), δ(Dist_300m), p1(year), p2(year) 1.66 0.116 16

Ψ(Dist_300m), CΨ(.), R(HW_Stand), CR(.,.,zlat), δ(Dist_300m), p1(year), p2(year) 3.05 0.058 16

Ψ(Dist_300m), CΨ(.), R(HW_Stand), CR(.,Dist_3k,.), δ(Dist_300m), p1(year), p2(year) 3.05 0.058 16

Ψ(Dist_300m), CΨ(.), R(HW_Stand), CR(.,.,.), δ(Dist_300m), p1(year), p2(year) 3.76 0.041 15

Top ranked multi-season multi-state occupancy models are shown in Table 4.4: occupancy by

adult bears was positively associated with increasing forest disturbance at the local scale (i.e. a radius of

300 m or 1 k around the survey sites) (Fig 4.2A), occupancy by reproductive bears was positively

associated with an increasing proportion of hardwood trees in the stand (immediate area) (Fig 4.2B).

The probability of a site occupied by only adults in one season transitioning into a reproduction state in

the next season was positively related with disturbance at a 1 k radius buffer (Fig 4.2C), while the

probability that a reproductive site would remain so in a following season was higher at higher latitudes

(Fig 4.2D). Detection of reproductive bears was higher in disturbed stands, while detection of adults only

81
varied by year, with summers 2017 and 2019 higher than summers 2018 and 2020. The estimated

probabilities of detecting bears on a given survey occasion are shown in Table 4.5 (and Fig. 4.3).

Table 4.4. Parameter estimates (betas) and standard errors for top ranked dynamic, multi-state model
occupancy model of American black bears in Maine, USA. Data collected at 197 camera sites, from 2017
to 2020.
Parameter Intercept (se) Covariate Beta (se)

Ψ +2.647 (0.626) Dist_300m +1.032 (0.505)

CΨ - - -

R -0.117 (0.268) HW_Stand +0.553 (0.267)

CR0 - - -

CR1 -0.123 (0.426) Dist_1k +0.831 (0.388)

CR2 +0.599 (0.331) Latitude +0.387 (0.254)

δ -2.067 (0.114) Dist_300m +0.404 (0.114)

p1 -2.494 (0.131)

p2 -0.932 (0.089)

Summer_2018 -0.713 (0.123)

p1 and p2 † Summer_2019 -0.146 (0.117)

Summer_2020 -0.849 (0.149)

† p1 and p2 modeled with Summer_2017 as the intercept, year effect fixed for both

82
Fig 4.2. Four occupancy probabilities from multi-state, multi-season models of American black bear
camera trap data. (A) The probability of initial occupancy by only adult bears modeled by the forest
disturbance in a 1 k radius buffer, (B) the initial probability of reproduction as predicted by hardwood
(100 m), (C) the probability of a site transitioning from adults only to reproduction in the following
season, modeled by disturbance in a 1 k buffer, and (D) the probability of remaining reproductive in
following summers, modeled by latitude.

83
Fig. 4.3. Detection probabilities for American black bears in Maine, across four summers of camera

surveys. p1 (green squares) is the probability of detecting adult bears when only adults occupy the site,

p2 (orange circles) is the probability of detecting adult bears when reproductive bears occupy the site,

and δ (purple triangles) is the probability of detecting reproductive bears.

Table 4.5. Detection probabilities for American black bears in Maine, USA: p1 is the probability, within a
24-hour-period, of detecting adult bears when the true state is only adult bears; p2 is the probability of
only detecting adult bears with the true state is reproductive bears; and δ is the probability of detecting
cubs, correctly identifying a site in the true reproductive state.

Summer p1 (se) p2 (se) δ (se)

2017 0.28 (0.018) 0.08 (0.009)

2018 0.16 (0.012) 0.04 (0.005)

0.11 (0.011)
2019 0.25 (0.015) 0.07 (0.007)

2020 0.14 (0.015) 0.03 (0.005)

84
4.5 Discussion

We collected data on the occupancy of American black bears across north and central Maine to help

understand how two components of forest structure and composition (forest disturbance caused by

timber harvest, and the proportion of hardwood to softwood trees) can affect bear distribution and

their probability of producing cubs. We found that bears in Maine are widely distributed, from our top

ranked multi-year, multi-state model we estimated that bears occupied 90.7% (CI 71.6-99.8%) of study

sites, and were in a reproductive state at 47.2% (CI 28.4-69.5%). Forest disturbance per se is positively

related to bear occupancy, while the proportion of hardwood trees in a stand is positively related to

reproduction. This reflects the literature, which indicates that the availability of fall food is a key

predictor of female reproductive success, and that in Maine masting deciduous trees are a critical

source for bears (McLaughlin et al. 1994; Jakubas et al. 2005; Seger et al. 2013)

The spatial and temporal patterns of human land-use have complex impacts on ecosystems,

from fragmentation of habitat to changes in species composition, and different species respond

differently over time since disturbance (Fischer and Lindenmayer 2006). In the short-term, the

immediate impact of human disturbance and habitat alteration can be acutely negative effects for black

bears, such as direct mortality (e.g. road kill) and abandonment of home ranges or of den sites leading

to cub death (Elowe and Dodge 1989). However the longer-term impacts can be nuanced, for example,

by resetting the natural successional stages of forest stands to provide increased food availability (e.g.

Costello 1992) and reduced competition from other carnivores (Comiskey et al. 2004, Elbroch et al.

2015). In our study we found that forest disturbance scores were included in models for both occupancy

of adult bears, and the probability of “colonization” by reproductive bears, with a positive relationship

with both. The spatial scales that ranked highest were small (300 m and 1 k radii), while larger scales (3 k

and 6 k) were not included as important in our models (i.e. falling within 2, or even 4, delta AICc). Thus,

it appears the overall landscape being harvested for timber or not is not a driving factor in black bear

85
distribution across northern and central Maine. We note that the locations of our survey sites were not

currently being harvested, so disturbance would have occurred at minimum two years prior, and we

cannot speculate on how quickly post-disturbance the positive impacts may begin to outweigh any

negative ones.

A potential long-term consequence of human modification of the forests of Maine that may be

especially important to black bears is the shift in tree species composition to favor economically

favorable species and forest structure (Nelson et al. 2011; Harris and Betts 2021). Monocrop and even-

age stands reduce biodiversity, and although true plantations are relatively rare in New England (about

4% of managed forests, Nelson et al. 2012), other silivicultural practices can have similar effects

(Puettman 2009). Another practice, which has been found to negatively impact the abundance of

herbaceous species immediately post-clearing in Maine, is the use of aerial pesticides to encourage

regrowth of preferred species (Raymond et al 1996). As we have seen, herbaceous and hardwood

species of important to black bears in Maine (Jakubas 2005, McLaughlin 1994, Elowe and Dodge 1989).

Our findings reflect the literature, with our top models showing the probability of reproductive

bears was positively related to the proportion of hardwoods in the forest. We make the caveat that we

did not collect data on the different species of hardwood present at our sites, and thus cannot say if

there are tree-species-specific trends in bear reproductive probability (as seen in synchronous breeding

success linked to beech in some areas of Maine [Spencer 1955, Schooley 1990, McLaughlin 1994]).

However, the strength of the effect of even non-specific hardwood presence makes it especially clear

that these tree species are important to black bear populations. Further research and monitoring is

warranted, as the species composition of hardwoods in Maine has already changed and will continue to

do so with land-use and climate change (Kelly and Goulden 2008).

86
4.5.1 Management implications

Although abundant and apparently stable now, black bears have historically been over-harvested prior

to regulation of hunting activity, and although not currently an issue in Maine, illegal poaching of bears

for the bile trade has been identified as a global concern (Servheen et al. 2005). The leading causes of

mortality in Maine for adult bears is hunting, while for cubs, yearlings and juveniles it is natural causes

(starvation, intraspecific killing, etc, McLaughlin 1999). Thus, keeping consistent track of reproductive

success can provide key data for population projections and species management. We found that

camera trapping provided an abundance of data, independent of hunter harvest (and thus comprised of

all age classes), which when analyzed in a multi-state, multi-year occupancy framework can provide

detailed regional data on the status of black bears. We also found variability in detection rates for adult

between years, speculate that this reflects masting events: in years following high availability of natural

food, detection rates were lower (2018 and 2020 were mast years for several tree species in central

Maine, Brehm personal communication, and in southern Maine, Holt Forest Project 2019). More

research would be needed to see if this trend holds, and if detection rates could be used to extrapolate

to food availability, and thus predict the future reproductive success.

87
 REFERENCES

Aldous, S. E., and J. Manweiler. 1942. The winter food habits of short-tailed weasel in northern
Minnesota. Journal of Mammalogy 23:250–255.

American Society of Mammalogists [ASM]. 2021. Long-tailed weasel Neogale frenata. Mammal
Diversity: Explore the Database. Volume 4. <https://www.mammaldiversity.org/explore.html
PERMALINK:>. Accessed 12 Oct 2021.

Atmeh, K., A. Andruszkiewicz, and K. Zub. 2018. Climate change is affecting mortality of weasels due to
camouflage mismatch. Scientific Reports 8:7648.

Aubry, K. B. and J. C. Lewis. 2003. Extirpation and reintroduction of fishers (Martes pennanti) in Oregon:
Implications for their conservation in the Pacific states. Biological Conservation 114:79–90.

Aubry, K. B., W. J. Zielinski, M. G. Raphael, G. Proulx, and S. W. Buskirk, editors. 2012. Biology and
Conservation of Martens, Sables, and Fishers: A New Synthesis. Cornell University Press, Ithaca,
New York, USA.

Bjornstad, O.N. 2018. Package ‘ncf’: Spatial Covariance Functions. Version 1-1.7 [software]. CRAN.R;
http://cran.r-project.org/web/packages/ncf/index.html

Boitani, L. and R. A. Powell RA, editors. 2012. Carnivore Ecology and Conservation: A Handbook of
Techniques. Oxford University Press. New York, New York. 490 p.

Boitani, L., P. Ciucci and A. Mortelliti. 2012. Designing carnivore surveys. Pages 8-29 in Carnivore Ecology
and Conservation: A Handbook of Techniques. Oxford University Press, New York, New York,
USA.

Bridger, M. C., C. J. Johnson, and M. P. Gillingham. 2016. Assessing cumulative impacts of forest
development on the distribution of furbearers using expert-based habitat modeling. Ecological
Applications 26:499–514.

Brown, R. D. and B. Brasnett. 2020. SNODAS: NASA National Snow and Ice Cover Database. Accessed
January 30 2021. nsidc.org/data/g02158

Bull, E. L. and T. W. Heater. 2001. Survival, causes of mortality, and reproduction in the American
marten in northeastern Oregon. Northwestern Naturalist 82:1–6.

Burnham, K. P., and D. R. Anderson. 2002. Model selection and multimodal inference: A practical
information-theoretic approach. Second edition. Springer-Verlag, New York, NY, USA.

88
Burton, A. C., E. Neilson, D. Moreira, A. Ladle, R. Steenweg, J. T. Fisher, E. Bayne, and S. Boutin. 2015.
Wildlife camera trapping: A review and recommendations for linking surveys to ecological
processes. Journal of Applied Ecology 52:675–685.

Butler, B. J. 2018. Forests of Maine, 2017. Resource Update FS-160. Newtown Square, PA. US
Department of Agriculture, Forest Service, Northern Research Station doi.org/10.2737/FS-RU-
160.

Buyaskas, M., B. E. Evans, and A. Mortelliti. 2020. Assessing the effectiveness of attractants to increase
camera trap detections of North American mammals. Mammalian Biology 100:91–100.

Chambert, T., D. A. W. Miller, and J. D. Nichols. 2015. Modeling false positive detections in species
occurrence data under different study designs. Ecology 96:332–339.

Chandler, R., K. Kellner, A. I. Fiske, D. Miller, A. Royle, J. Hostetler, R. Hutchinson, and A. Smith. 2021.
Package ‘unmarked’: Models for data from unmarked animals. Version 1.1.1. https://cran.r-
project.org/web/packages/unmarked/index.

Chapin, T. G., D. J. Harrison, and D. D. Katnik. 1998. Influence of landscape pattern on habitat use by
American marten in an industrial forest. Conservation Biology 12:1327–1337.

Cheveau, M., L. Imbeau, P. Drapeau, and L. Belanger. 2013. Marten space use and habitat selection in
managed coniferous boreal forests of eastern Canada. Journal of Wildlife Management 77:749–
760.

Clare, J., S. T. McKinney, J. E. Depue, and C. S. Loftin. 2017. Pairing field methods to improve inference in
wildlife surveys while accommodating detection covariance: Ecological Applications 27(7):2031–
47.

Clark, A. 1986. Fisher Assessment. Maine Department of Inland Fisheries and Wildlife, Bangor, Maine,
USA.

Clark, T. W., E. Anderson, C. Douglas, and M. Strickland. 1987. Martes americana. Mammalian Species
289:1-8.

Clement, M. J. 2016. Designing occupancy studies when false-positive detections occur. Methods in
Ecology and Evolution 7:1538–1547.

Croose, E., F. Bled, N. L. Fowler, D. E. Beyer, and J. L. Belant. 2019. American marten and fisher do not
segregate in space and time during winter in a mixed‐forest system. Ecology and Evolution
9:4906–4916.

89
Croose, E., R. Hanniffy, B. Hughes, K. McAney, J. MacPherson, and S. P. Carter. 2021. Assessing the
detectability of the Irish stoat Mustela erminea hibernica using two camera trap-based survey
methods. Mammal Research doi:10.1007/s13364-021-00598-z.

Davis, L. R. 2009. Denning ecology and habitat use by fisher (Martes pennanti) in pine dominated
ecosystems of the Chilcotin Plateau. Thesis, Simon Fraser University, Burnaby, British Columbia,
Canada.

Di Marco, M., L. Boitani, D. Mallon, M. Hoffmann, A. Iacucci, E. Meijaard, P. Visconti, J. Schipper, and C.
Rondinini. 2014. A retrospective evaluation of the global decline of carnivores and ungulates.
Conservation Biology 28:1109–1118.

Donadio, E. and S. W. Buskirk. 2006. Diet, morphology, and interspecific killing in carnivora. American
Naturalist 167:524–536.

Efford, M. G., and D. K. Dawson. 2012. Occupancy in continuous habitat. Ecosphere 3:1-15 Article 32.

Elbroch, M., and K. Rinehart. 2011. Peterson reference guide to behavior of North American mammals.
Houghton Mifflin Harcourt Publishing Company, New York, New York. 374 p.

Elsasser, S. C., and G. H. Parker. 2008. Morphometric criteria for distinguishing species and age-cohorts
of ermine (Mustela erminea) and long-tailed weasel (M. frenata). Acta Zoologica Academiae
Scientiarum Hungaricae 54:75–88.

Elzinga, C. L., D. W. Salzer, and J. W. Willoughby. 1998. Measuring and Monitoring Plant Populations.
Denver, Colorado: BLM Technical Reference 1730-1; 1998. 477 p.

Evans, B. E., C. E. Mosby, and A. Mortelliti. 2019. Assessing arrays of multiple trail cameras to detect
North American mammals. PLoS ONE 14:e0217543.

Evans, B. E. and A. Mortelliti. 2022. Effects of forest disturbance, snow depth and intraguild dynamics on
American marten and fisher occupancy in Maine, USA. [Manuscript accepted for publication
December 7, 2021, Ecosphere, John Wiley and Sons]

Fahrig, L. 2013. Rethinking patch size and isolation effects: The habitat amount hypothesis. Journal of
Biogeography 40(9):1649–1663.

Feldhamer, G. A., B. C. Thompson, and J. A. Chapman, editors. 2003. Wild Mammals of North America:
Biology, Management and Conservation. Second edition. John Hopkins University Press,
Baltimore, Maryland, USA.

90
Fernandez, I., S. Birkel, C. Schmitt, J. Simonson, B. Lyon, A. Pershing, E. Stancioff, G. Jacobson, and P.
Mayewski. 2020. Maine’s Climate Future 2020 Update. Orono, ME.
<climatechange.umaine.edu/climate-matters/ maines-climate-future/>.

Fischer, J., and D. L. Lindenmayer. 2006. Beyond fragmentation: The continuum model for fauna
research and conservation in human-modified landscapes. Oikos 112:473–480.

Fisher, J. T., and L. Wilkinson. 2005. The response of mammals to forest fire and timber harvest in the
North American boreal forest. Mammal Review 35:51–81.

Fisher, J. T., M. Wheatley, and D. Mackenzie. 2014. Spatial patterns of breeding success of grizzly bears
derived from hierarchical multistate models. Conservation Biology 28:1249–1259.

Fisher, J. T., N. Heim, S. Code, and J. Paczkowski. 2016. Grizzly bear noninvasive genetic tagging surveys:
Estimating the magnitude of missed detections. PLoS One. 2016. 11(9): e0161055.

Fiske, I. J., and R. B. Chandler. 2011. unmarked: An R Package for Fitting Hierarchical Models of Wildlife
Occurrence and Abundance. Journal of Statistical Software 2011:43:1–23

Fraver, S., A. S. White, and R. S. Seymour. 2009. Natural disturbance in an old-growth landscape of
northern Maine, USA. Journal of Ecology 97:289–298.

Fuller, A. K. and D. J. Harrison. 2005. Influence of partial timber harvesting on American martens in
north-central Maine. Journal of Wildlife Management 69(2):710-722.

Fuller, A. K., D. W. Linden, and J. A. Royle. 2016. Management decision making for fisher populations
informed by occupancy modeling. Journal of Wildlife Management 80:794–802.

Furnas, B. J., R. H. Landers, R. L. Callas, and S. M. Matthews. 2017. Estimating population size of fishers
(Pekania pennanti) using camera stations and auxiliary data on home range size. Ecosphere
8(3):e01747.

Gabriel, M. W., L. W. Woods, R. Poppenga, R. a. Sweitzer, C. Thompson, S. M. Matthews, J. M. Higley, S.

M. Keller, K. Purcell, R. H. Barrett, G. M. Wengert, B. N. Sacks, and D. A. L. Clifford. 2012.
Anticoagulant rodenticides on our public and community lands: Spatial distribution of exposure
and poisoning of a rare forest carnivore. PLoS ONE 7(7):e40163.

Garshelis, D. L., B. K. Scheick, D. L. Doan-Crider, J. J. Beecham, and M. E. Obbard. 2016. Ursus

americanus, American Black Bear. Mammalian Species. Volume 647. The IUCN Red List of
Threatened and Endangered Species, e.T41687A114251609.

91
Garton, E. O., J. S. Horne, J. L. Aycrigg, and J. T. Ratti. 2011. Research and Experimental Design. Chapter 1
in N. Silvy, editor. Techniques for Wildlife Investigations and Management. Seventh edition. The
Wildlife Society, Bethesda, Maryland, USA.

Gawler, S. C., J. J. Albright, P. D. Vickery, and F. C. Smith. 1996. Biological Diversity in Maine: An
assessment of status and trends in the terrestrial and freshwater landscape. Report prepared for
the Maine Forest Biodiversity Project. Natural Areas Program, Maine Department of
Conservation, Augusta, Maine.

Gehring, T. M. 2000. Ecology of Mammalian Predators in a Fragmented Landscape. Dissertation, Purdue

University, Indiana, USA.

Gehring, T. M., and R. K. Swihart. 2003. Body size, niche breadth, and ecologically scaled responses to
habitat fragmentation: Mammalian predators in an agricultural landscape. Biological
Conservation 109:283–295.

Gosse, J. W., R. Cox, and S. W. Avery. 2005. Home-Range Characteristics and Habitat Use By American
Martens in Eastern. Journal of Mammalogy 86(6):1156–1163.

Gould, M. J., W. R. Gould, J. W. Cain, and G. W. Roemer. 2019. Validating the performance of occupancy
models for estimating habitat use and predicting the distribution of highly-mobile species: A
case study using the American black bear. Biological Conservation 234:28–36. Elsevier.

Guillera-Arroita, G., M. S. Ridout, and B. J. T. Morgan. 2010. Design of occupancy studies with imperfect
detection. Methods in Ecology and Evolution 1(2): 131–139.

Guillera-Arroita, G. and J. J. Lahoz-Monfort. 2012. Designing studies to detect differences in species

occupancy: Power analysis under imperfect detection. Methods in Ecology and Evolution 3(5):
860–869.

Gustafson, E. J., D. E. Lytle, R. Swaty, and C. Loehle. 2007. Simulating the cumulative effects of multiple
forest management strategies on landscape measures of forest sustainability. Landscape
Ecology 22:141–156.

Hajduk, L. I. 2008. Space use and habitat selection of long-tailed weasels (Mustela frenata) in southern
Illinois. Thesis, Southern Illinois University Carbondale, Illinois, USA.

Hansen, M.C., P. V. Potapov, R. Moore, M. Hancher, S. A. Turubanova, and A. Tyukavina. 2013. High-
resolution global maps of forest cover change. Science 342(6160):850–853

Hargis, C. D., J. A. Bissonette, and D. L. Turner. 1999. The influence of forest fragmentation and
landscape patterns on American martens. Journal of Applied Ecology 36:157–172.

92
Harrison, D. J., A. K. Fuller, and G. Proulx, editors. 2004. Martens and Fishers (Martes) in Human-Altered
Environments: An International Perspective. Springer, New York, New York, USA.

Hayward, M. W., L. Boitani, N. D. Burrows, P. J. Funston, K. U. Karanth, D. I. MacKenzie, K. H. Pollock, and

R. W. Yarnel. 2015. Ecologists need robust survey designs, sampling and analytical methods.
Journal of Applied Ecology 52:286–290.

Helgen, K., and F. Reid. 2016. Mustela frenata, Long-tailed weasel. The IUCN Red List of Threatened
Species 2016:e.T41654A45213820.

Henderson, F. R. 1994. Weasels. Pages C119-C122 in S. E. Hygnstrom, R. M. Timm and G. E. Larson,

editors. The Handbook: Prevention and Control of Wildlife Damage. University of Nebraska-
Lincoln.

Hines, J. E. 2006. PRESENCE: Software to estimate patch occupancy and related parameters. USGS-
PWRC. Version 2.13.12. https://www.mbr-pwrc.usgs.gov/software/presence

Hines, J.E. 2017. [software]. PRESENCE-Software to estimate patch occupancy and related parameters.
Version 11.8. http://www.mbr-pwrc.usgs.gov/software/presence

Hoffmann, M., J. L. Belant, J. S. Chanson, N. A. Cox, J. Lamoreux, A. S. L. Rodrigues, J. Schipper, and S. N.

Stuart. 2011. The changing fates of the world’s mammals. Philosophical Transactions of the
Royal Society B: Biological Sciences 366:2598–2610.

Homyack, J. A., D. J. Harrison and W. B. Krohn. 2005. Long-term effects of precommercial thinning on
small mammals in northern Maine. Forest Ecology and Management 205(1-3):43-57.

Hoving, C. L., J. D. Harrison, K. B. Krohn, W. J. Jakubas, and M.A. McCollough. 2004. Canada lynx Lynx
canadensis habitat and forest succession in northern Maine, USA. Wildlife Biology 10:285–294.

Iannarilli, F., J. Erb, T. W. Arnold, and J. R. Fieberg. 2021. Evaluating species-specific responses to
camera-trap survey designs. Wildlife Biology 2021. doi.org/10.2981/wlb.00726.

Intergovernmental Panel on Climate Change [IPCC]. 2019. Climate change and land: An IPCC special
report on climate change, desertification, land degradation, sustainable land management, food
security, and greenhouse gas fluxes in terrestrial ecosystems. Geneva, Switzerland.
https://www.ipcc.ch/srccl-report-download-page/

Jachowski, D., R. Kays, A. Butler, A. M. Hoylman, and M. E. Gompper. 2021. Tracking the decline of
weasels in North America. PLoS ONE 16(7):e0254387.

93
Jakubas, W. L., C. R. McLaughlin, P. G. Jensen, and S. A. McNulty. 2005. Alternate year beechnut
production and its influence on bear and marten populations. Changing forests - challenging
times: Proceedings of the New England Society of American Foresters; 85th winter meeting Gen.
Tech.:19.

Jensen, P. G. and M. M. Humphries. 2019. Abiotic conditions mediate intraguild interactions between
mammalian carnivores. Journal of Animal Ecology 88(9):1305–1318.

Johnston, S. D., D. B. Sasse, and R. E. Kissell, Jr. 2019. Long-tailed Weasel (Mustela frenata) Distribution
Survey in Arkansas: Challenges in Detecting a Rare Species. Journal of the Arkansas Academy of
Science 73:83–92.

Karanth, K.U. 1995. Estimating tiger Panthera tigris populations from camera-trap data using capture-
recapture models. Biological Conservation 71(3): 333–338.

Karanth, K.U., and J.D. Nichols. 1998. Estimation of tiger densities in India using photographic captures
and recaptures. Ecology 79(8):2852–2862.

Kays, R. W. and K. M. Slauson. 2008. Remote cameras. Pages 110–140 in R. A. Long, P. MacKay, W. J.
Zielinski, and J. C. Ray, editors. Noninvasive Survey Methods for North American Carnivores.
Island Press, Washington, District of Columbia, USA.

Kays, R., S. Tilak, B. Kranstauber, P. A. Jansen, C. Carbone, M. Rowcliffe, T. Fountain, J. Eggert, and Z. He.
2011. Camera traps as sensor networks for monitoring animal communities. International
Journal of Research and Reviews in Wireless Sensor Networks 1(2):19-29.

Kilbride, J. B. 2018. Forest Disturbance Detection and Aboveground Biomass Modeling Using Moderate-
Resolution, Time-Series Satellite Imagery. University of Maine.

King, C. M. 1983. Mustela erminea. Mammalian Species 195.

King, C. M., and R. A. Powell. 2007. The Natural History of Weasels and Stoats: Ecology, Behavior, and
Management. Second edition. Oxford University Press, New York, New York. 446 p.

Koepfli, K.-P., K. A. Deere, G. J. Slater, C. Begg, K. Begg, L. Grassman, M. Lucherini, G. Veron, and R. K.
Wayne. 2008. Multigene phylogeny of the Mustelidae: Resolving relationships, tempo and
biogeographic history of a mammalian adaptive radiation. BMC Biology 6:10.

94
Kowalski, B., F. Watson, C. Garza, and B. Delgado. 2015. Effects of landscape covariates on the
distribution and detection probabilities of mammalian carnivores. Journal of Mammalogy
96(3):511–521.

Kolowski, J. M., J. Oley, and W. J. McShea. 2021. High density camera trap grid reveals lack of
consistency in detection and capture rates across space and time. Ecosphere 12(2):e03350.

Krohn, W. B., K. D. Elowe, and R. B. Boone. 1995. Relations among fishers, snow, and martens:
Development and evaluation of two hypotheses. The Forestry Chronicle 71(1):97–105.

Lancaster, P. A., J. Bowman, and B. A. Pond. 2008. Fishers, farms, and forests in eastern North America.
Landscape Ecology 42:93–101.

LANDFIRE. 2016. U.S. National Vegetation Classification. https://landfire.gov/nvc.php. Accessed January

10 2020.

Lesmeister, D. B., C. K. Nielsen, E. M. Schauber, and E. C. Hellgren. 2015. Spatial and temporal structure
of a mesocarnivore guild in midwestern North America. Wildlife Monographs 191:1–61.

Lewis, J. C., K. J. Jenkins, P. J. Happe, D. J. Manson, and M. McCalmon. 2016. Landscape-scale habitat
selection by fishers translocated to the Olympic Peninsula of Washington. Forest Ecology and
Management 369:170–183.

Linnell, M. A., C. W. Epps, E. D. Forsman, and W. J. Zielinski. 2017a. Space use, movements, and rest site
use by short-tailed weasels Mustela erminea in managed forests of western Oregon. Wildlife
Biology 2017:wlb.00270.

Linnell, M. A., C. W. Epps, E. D. Forsman, and W. J. Zielinski. 2017b. Survival and Predation of Weasels
(Mustela erminea, Mustela frenata) in North America. Northwest Science 91:15–26.

Lisgo, K. A. 1999. Ecology of the Short-tailed Weasel (Mustela erminea) in the Mixedwood Boreal Forest
of Alberta. Dissertation, The University of British Columbia, Canada.

Łopucki, R., D. Klich, and I. Kitowski. 2019. Are small carnivores urban avoiders or adapters: Can they be
used as indicators of well-planned green areas? Ecological Indicators 101:1026–1031.

MacCleery, D. W. 2011. American Forests: A History of Resiliency and Recovery. Fourth Edition. Forest
History Society, Durham, North Carolina, USA.

MacKenzie, D. I., J. D. Nichols, G. B. Lachman, S. Droege, J. A. Royle, and C. A. Langtimm. 2002.

Estimating site occupancy rates when detection probabilities are less than one. Ecology
83:2248–2255.

95
MacKenzie, D. I., J. D. Nichols, J. E. Hines, M. G. Knutson, and A. B. Franklin. 2003. Estimating site
occupancy, colonization, and local extinction when a species is detected imperfectly. Ecology
84:2200–2207.

MacKenzie, D. I., L. L. Bailey, and J. D. Nichols. 2004. Investigating species co-occurrence patterns when
species are detected imperfectly. Journal of Animal Ecology 73:546–555.

MacKenzie, D. I., J. D. Nichols, J. A. Royle, K. H. Pollock, L. L. Bailey, and J. E. Hines. 2017. Occupancy
Estimation and Modeling: Inferring Patterns and Dynamics of Species Occurrence. Second Edi.
Elsevier Academic Press, Burlington, Massachusetts, USA.

Maine Forest Service [MFS]. 1995. Forest Trees of Maine Twelfth Edition. Department of Conservation.
12th edition.

Maine Forest Service [MFS]. 2013. Maine Forest Practices Act. Augusta, Maine, USA.

Malcolm, J. R., B. D. Campbell, B. G. Kuttner, and A. Sugar. 2004. Potential indicators of the impacts of
forest management on wildlife habitat in northeastern Ontario: A multivariate application of
wildlife habitat suitability matrices. Forestry Chronicle 80:91–106.

Manlick, P. J., J. E. Woodford, B. Zuckerberg, and J. N. Pauli. 2017. Niche compression intensifies
competition between reintroduced American marten (Martes americana) and fishers (Pekania
pennanti). Journal of Mammalogy 98:690–702.

Mann, G. K. H., M. J. O’Riain, and D. M. Parker. 2015. The road less travelled: Assessing variation in
mammal detection probabilities with camera traps in a semi-arid biodiversity hotspot.
Biodiversity and Conservation 24:531–545.

Maser, C. 1989. Forest Primeval: The Natural History of an Ancient Forest. Sierra Club, San Francisco,
California, USA.

Mata, C., P. Ruiz-Capillas, and J. E. Malo. 2017. Small-scale alterations in carnivore activity patterns close
to motorways. European Journal of Wildlife Research 63(64).

Matthews, S. M., J. M. Higley, K. M. Rennie, R. E. Green, C. A. Goddard, G. M. Wengert, M. W. Gabriel,

and T. K. Fuller. 2013. Reproduction, recruitment, and dispersal of fishers (Martes pennanti) in a
managed Douglas-fir forest in California. Journal of Mammalogy 94:100–108.

Mazerolle, M. J. 2020. AICcmodavg: Model selection and multimodel inference based on (Q)AIC(c). R
package version 2.3-1, https://cran.r-project.org/package=AICcmodavg.

McCann, N. P., P. A. Zollner, and J. H. Gilbert. 2010. Survival of adult martens in northern Wisconsin.
Journal of Wildlife Management 74:1502–1507.

96
McCaskill, G. L., T. Albright, C. J. Barnett, B. J. Butler, S. J. Crocker, C. M. Kurtz, W. H. McWilliams, P. D.
Miles, R. S. Morin, M. D. Nelson, R. H. Widmann, and C. W. Woodall. 2016. Maine Forests 2013.
U.S. Forest Service, Resource Bulletin NRS-103, Newtown Square, Pennsylvania, USA.

McGarigal, K., H. Y. Wan, K. A. Zeller, B. C. Timm, and S. A. Cushman. 2016. Multi-scale habitat selection
modeling: A review and outlook. Landscape Ecology 31:1161–1175.

McLaughlin, C. R. 1999. Black Bear Assessment and Strategic Plan 1999. 650 State Street, Bangor ME
04401.

McLaughlin, C. R., G. J. J. Matula, and R. J. O’Connor. 1994. Synchronous reproduction by Maine black
bears. International Conference on Bear Research and Management 9:471–479.

McWilliams, W. H., B. J. Butler, L. E. Caldwell, D. M. Griffith, M. L. Hoppus, K. M. Laustsen, A. J. Lister, T.

W. Lister, J. W. Metzler, R. S. Morin, S. A. Sader, L. B. Stewart, J. R. Steinman, J. A. Westfall, D. A.
Williams, A. Whitman, and C. W. Woodall. 2005. The Forests of Maine: 2003. Resource Bulletin
NE-164. United States Forest Service, Northeast Research Station, Newtown Square,
Pennsylvania. 188 p.

Maine Department of Inland Fisheries and Wildlife [MDIFW]. 2019. Furbearer harvest reports [internal
data]. Accessed April 30 2019. https://www.maine.gov/ifw/hunting-trapping/trapping-laws

Miller, D. A., J. D. Nichols, B. T. McClintock, E. H. C. Grant, L. L. Bailey, and L. A. Weir. 2011. Improving
occupancy estimation when two types of observational error occur: Non-detection and species
misidentification. Ecology 92:1422–1428.

Miller, D. A. W., J. D. Nichols, J. A. Gude, L. N. Rich, K. M. Podruzny, J. E. Hines, and M. S. Mitchell. 2013.
Determining Occurrence Dynamics when False Positives Occur: Estimating the Range Dynamics
of Wolves from Public Survey Data. PLoS ONE 8(6): e65808.

Monterroso, P., P. Rebelo, P. C. Alves, and P. Ferreras. 2016. Niche partitioning at the edge of the range:
A multidimensional analysis with sympatric martens. Journal of Mammalogy 97(3): 928–39.

Moriarty, K. M., W. J. Zielinski, and E. D. Forsman. 2011. Decline in American marten occupancy rates at
Sagehen Experimental Forest, California. Journal of Wildlife Management 75:1774–1787.

Moriarty, K. M., C. W. Epps, M. G. Betts, D. J. Hance, J. D. Bailey, and W. J. Zielinski. 2015. Experimental
evidence that simplified forest structure interacts with snow cover to influence functional
connectivity for Pacific martens. Landscape Ecology 30:1865–1877.

Mortelliti, A., and L. Boitani. 2008. Interaction of food resources and landscape structure in determining
the probability of patch use by carnivores in fragmented landscapes. Landscape Ecology 23:285–
298.

97
Mos, J., and T. R. Hofmeester. 2020. The Mostela: An adjusted camera trapping device as a promising
non-invasive tool to study and monitor small mustelids. Mammal Research 65:843–853.

National Oceanic and Atmospheric Agency [NOAA]. 2019. National Climate Data Center. Climate
Normals. https://www.ncdc.noaa.gov/cdo-web/datatools/normals. Accessed January 2 2019.

National Oceanic and Atmospheric Agency [NOAA]. 2020. Climate Normals. www.ncdc.noaa.gov/cdo-
web/datatools/normals. Accessed March 15 2020.

National Oceanic and Atmospheric Agency [NOAA]. 2021. National Centers for Environmental
Information, Climate at a Glance: Statewide Time Series. <https://www.ncdc.noaa.gov/cag/>.
Accessed 13 Dec 2021.

Neilson, E. W., T. Avgar, A. Cole Burton, K. Broadley, and S. Boutin. 2018. Animal movement affects
interpretation of occupancy models from camera-trap surveys of unmarked animals. Ecosphere
9.

Newey, S., P. Davidson, S. Nazir, G. Fairhurst, F. Verdicchio, R. J. Irvine, and R. van der Wal. 2015.
Limitations of recreational camera traps for wildlife management and conservation research: A
practitioner’s perspective. Ambio 44:624–635.

Nichols, J. D., L. L. Bailey, A. F. O’Connell, N. W. Talancy, E. H. Campbell Grant, A. T. Gilbert, E. M.

Annand, T. P. Busband and L. E. Hines. 2008. Multi-scale occupancy estimation and modelling
using multiple detection methods. Journal of Applied Ecology 45:1321–1329.

Noon, B. R., L. L. Bailey, T. D. Sisk, and K. S. McKelvey. 2012. Efficient species-level monitoring at the
landscape scale. Conservation Biology 26(3):432–441.

Nyakatura, K., and O. R. P. Bininda-Emonds. 2012. Updating the evolutionary history of Carnivora
(Mammalia): A new species-level supertree complete with divergence time estimates. BMC
Biology 10:12.

O’Connell, A. F., J. D. Nichols, and K. U. Karanth, editors. 2011. Camera Traps in Animal Ecology:
Methods and Analyses. Springer, New York, New York.

O’Connor, K. M., L. R. Nathan, M. R. Liberati, M. W. Tingley, J. C. Vokoun, and T. A. G. Rittenhouse. 2017.

Camera trap arrays improve detection probability of wildlife: Investigating study design
considerations using an empirical dataset. PLoS One. 2017 12(4): e0175684

98
Otto, C. R. V., L. L. Bailey, and G. J. Roloff. 2013. Improving species occupancy estimation when sampling
violates the closure assumption. Ecography 36(12):1299–1309.

Palomares, F., and T. M. Caro. 1999. Interspecific Killing among Mammalian Carnivores. The American
Naturalist 153:492–508.

Paragi, T. F., S. M. Arthur, and W. B. Krohn. 1996. Importance of tree cavities as natal dens for fishers.
Northern Journal of Applied Forestry 13:79–83.

Patterson, B. D., H. R. Ramirez-Chaves, J. F. Vilela, A. E. R. Soares, and F. Grewe. 2021. On the

nomenclature of the American clade of weasels (Carnivore: Mustelidae). Journal of Animal
Diversity 3:1–8.

Patton, D. R. 1992. Wildlife Habitat Relationships in Forested Ecosystems. Timber Press, Portland,
Oregon, USA.

Payer, D. C. and D. J. Harrison. 2003. Influence of forest structure on habitat use by American marten in
an industrial forest. Forest Ecology and Management 179:145–156.

Pease, B. S., C. K. Nielsen, and E. J. Holzmueller. 2016. Single-camera trap survey designs miss
detections: Impacts on estimates of occupancy and community metrics. PLoS ONE
11(11):e0166689.

Polis, G. A. and R. D. Holt. 1992. Intraguild predations: The dynamics of complex trophic interactions.
Trends in Ecology and Evolution 7(5):151–154.

Potvin, F., L. Bélanger, and K. Lowell. 2000. Marten habitat selection in a clearcut Boreal landscape.
Conservation Biology 14:844–857.

Powell, R. A. 1993. The Fisher: Life History, Ecology, and Behavior. Second edition. University of
Minnesota Press, Minneapolis, Minnesota, USA.

Powell, R. A., S. W. Burskirk, and W. J. Zielinski. 2003. Fisher and Marten. Pages 635–649 in G. A.
Feldhamer, B. C. Thompson, and J. A. Chapman, editors. Wild Mammals of North America:
Biology, Management and Conservation. Second. John Hopkins University Press, Baltimore and
London.

Proulx, G., and R. K. Drescher. 1993. Distribution of the long-tailed weasel, Mustela frenata longicauda,
in Alberta as determined by questionnaires and interviews. Canadian Field-Naturalist 107:186–
191.

Puettmann, K. J., K .D. Coates and C. Messier. 2008. A critique of silviculture: Managing for complexity.
Island Press, Washington, District of Columbia, USA.

99
Purcell, K. L., A. K. Mazzoni, S. R. Mori, and B. B. Boroski. 2009. Resting structures and resting habitat of
fishers in the southern Sierra Nevada, California. Forest Ecology and Management 258:2696–
2706.

Pyšková, K., O. Kauzál, D. Storch, I. Horáček, J. Pergl, and P. Pyšek. 2018. Carnivore distribution across
habitats in a central-European landscape: A camera trap study. ZooKeys 2018:227–246.

Raine, R. M. 1983. Winter habitat use and responses to snow cover of fisher (Martes pennanti) and
marten (Martes americana) in southeastern Manitoba. Canadian Journal of Zoology 61:25–34.

Reid, F., K. Helgen, and A. Kranz. 2016. Mustela erminea, Short-tailed weasel. The IUCN Red List of
Threatened Species 2016:e.T29674A45203335.

Rich, L. N., D. A. W. Miller, H. S. Robinson, J. W. McNutt, M. J. Kelly, and M. Hayward. 2016. Using
camera trapping and hierarchical occupancy modelling to evaluate the spatial ecology of an
African mammal community. Journal of Applied Ecology 53(4):1225–1235.

Robinson L., S. A. Cushman, and M. K. Lucid. 2017. Winter bait stations as multispecies survey tools.
Ecology and Evolution 7:6826–6838.

Rojas, V. G., S. C. Loeb, and J. M. O’Keefe. 2019. False-positive occupancy models produce less-biased
occupancy estimates for a rare and elusive bat species. Journal of Mammalogy 100:212–222.

Rota, C. T., M. A. R. Ferreira, R. W. Kays, T. D. Forrester, E. L. Kalies, W. J. McShea, A. W. Parsons, and J. J.

Millspaugh. 2016. A multispecies occupancy model for two or more interacting species.
Methods in Ecology and Evolution 7(10):1164–73

Rowcliffe, J. M. 2017. Key frontiers in camera trapping research. Remote Sensing in Ecology and
Conservation 3:107–108. <http://doi.wiley.com/10.1002/rse2.65>.

Rowcliffe, J. M., and C. Carbone. 2008. Surveys using camera traps: Are we looking to a brighter future?
Animal Conservation 11:185–186.

Rovero, F., F. Zimmermann, D. Berzi, and P. Meek. 2013. Which camera trap type and how many do I
need? A review of camera features and study designs for a range of wildlife research
applications. Hystrix 24(2):148–156.

Royle, J. A., and W. A. Link. 2006. Generalized site occupancy models allowing for false positive and false
negative errors. Ecology 87:835–841.

100
Royle, J. A., and R. M. Dorazio. 2008. Hierarchical Modeling and Inference. Elsevier Ltd, Oxford, United
Kingdom.

Sauder, J. D., and J. L. Rachlow. 2014. Both forest composition and configuration influence landscape-
scale habitat selection by fishers (Pekania pennanti) in mixed coniferous forests of the Northern
Rocky Mountains. Forest Ecology and Management 314:75–84.

Schipper, J., J. S. Chanson, F. Chiozza, N. A. Cox, M. Hoffmann, V. Katariya, J. Lamoreux, A. S. L.

Rodrigues, S. N. Stuart, H. J. Temple, J. Baillie, L. Boitani, T. E. Lacher, R. A. Mittermeier, A. T.
Smith, D. Absolon, J. M. Aguiar, G. Amori, N. Bakkour, R. Baldi, R. J. Berridge, J. Bielby, P. A.
Black, J. J. Blanc, T. M. Brooks, J. A. Burton, T. M. Butynski, G. Catullo, R. Chapman, Z. Cokeliss, B.
Collen, J. Conroy, J. G. Cooke, G. A. B. Da Fonseca, A. E. Derocher, H. T. Dublin, J. W. Duckworth,
L. Emmons, R. H. Emslie, M. Festa-Bianchet, M. Foster, S. Foster, D. L. Garshelis, C. Gates, M.
Gimenez-Dixon, S. Gonzalez, J. F. Gonzalez-Maya, T. C. Good, G. Hammerson, P. S. Hammond, D.
Happold, M. Happold, J. Hare, R. B. Harris, C. E. Hawkins, M. Haywood, L. R. Heaney, S. Hedges,
K. M. Helgen, C. Hilton-Taylor, S. A. Hussain, N. Ishii, T. A. Jefferson, R. K. B. Jenkins, C. H.
Johnston, M. Keith, J. Kingdon, D. H. Knox, K. M. Kovacs, P. Langhammer, K. Leus, R. Lewison, G.
Lichtenstein, L. F. Lowry, Z. Macavoy, G. M. Mace, D. P. Mallon, M. Masi, M. W. McKnight, R. A.
Medellín, P. Medici, G. Mills, P. D. Moehlman, S. Molur, A. Mora, K. Nowell, J. F. Oates, W.
Olech, W. R. L. Oliver, M. Oprea, B. D. Patterson, W. F. Perrin, B. A. Polidoro, C. Pollock, A.
Powel, Y. Protas, P. Racey, J. Ragle, P. Ramani, G. Rathbun, R. R. Reeves, S. B. Reilly, J. E.
Reynolds, C. Rondinini, R. G. Rosell-Ambal, M. Rulli, A. B. Rylands, S. Savini, C. J. Schank, W.
Sechrest, C. Self-Sullivan, A. Shoemaker, C. Sillero-Zubiri, N. De Silva, D. E. Smith, C. Srinivasulu,
P. J. Stephenson, N. Van Strien, B. K. Talukdar, B. L. Taylor, R. Timmins, D. G. Tirira, M. F.
Tognelli, K. Tsytsulina, L. M. Veiga, J. C. Vié, E. A. Williamson, S. A. Wyatt, Y. Xie, and B. E. Young.
2008. The status of the world’s land and marine mammals: diversity, threat, and knowledge.
Science 322:225–230.

Schlexer, F. V. 2008. Attracting Animals to Detection Devices. Pages 263–292 in R. A. Long, P. MacKay,
W. J. Zielinski, and J. C. Ray, editors. Noninvasive Survey Methods for Carnivores. Island Press,
Washington DC, USA.

Schoener, T. W. 1974. Resource partitioning in ecological communities. Science, New Series 185:27–39.

Schooley, R. L., C. R. Mclaughlin, G. J. Matula, Jr., and W. B. Krohn. 1994. Denning Chronology of Female
Black Bears : Effects of Food, Weather, and Reproduction. American Society of Mammalogists

Seger, R. L., F. A. Servello, R. A. Cross, and D. H. Keisler. 2013. Body mass and mast abundance influence
foraging ecology of the American black bear (Ursus Americanus) in Maine. Canadian Journal of
Zoology 91:512–522.

Servheen, C., S. Herrero, and B. Peyton, editors. 1998. The Status Survey and Conservation Action Plan:
Bears. IUCN/SSC Bear and Polar Bear Specialist Groups. IUCN, Gland, Switzerland and
Cambridge, UK.

101
Servheen, C., S. Herrero, B. Peyton, and I. S. B. S. Group. 2005. Bears - Status Survey and Conservation
Action Plan.

Sévêque, A., L. K. Gentle, J. V López-bao, R. W. Yarnell, and A. Uzal. 2020. Human disturbance has
contrasting effects on niche partitioning within carnivore communities. Biological Reviews
95(6):1689–1705.

Seymour, R. S., A. S. White, and P. G. deMaynadier. 2002. Natural disturbance regimes in northeastern
North America—Evaluating silvicultural systems using natural scales and frequencies. Forest
Ecology and Management 155:357–367.

Shannon, G., J. S. Lewis, and B. D. Gerber. 2014. Recommended survey designs for occupancy modelling
using motion-activated cameras: Insights from empirical wildlife data. PeerJ. 2014; 2: e532.

Sheffield, S. R., and C. M. King. 1994. Mustela nivalis. Mammalian Species 454.

Sheffield, S. R., and H. H. Thomas. 1997. Mustela frenata. Mammalian Species 570.

Shirk, A. J., M. G. Raphael, and S. A. Cushman. 2014. Spatiotemporal variation in resource selection:
Insights from the American marten (Martes americana). Ecological Applications 24:1434–1444.

Sikes, R. S. 2016. 2016 Guidelines of the American Society of Mammalogists for the use of wild mammals
in research and education. Journal of Mammalogy 97:663–688.

Simons, E. M. 2009. Influences of past and future forest management on the spatiotemporal dynamics
of habitat supply for Canada lynx and American martens in northern Maine. Dissertation,
University of Maine, Orono, Maine, USA.

Simms, D. A. 1979. North American weasels: Resource utilization and distribution. Canadian Journal of
Zoology 57:504–520.

Sirén, A. P. K., P. J. Pekins, J. R. Kilborn, J. J. Kanter, and C. S. Sutherland. 2017. Potential influence of
high-elevation wind farms on carnivore mobility. Journal of Wildlife Management 81(8):1505–
1512.

Sirén, A. P., C. S. Sutherland, C. A. Bernier, K. J. Royar, J. R. Kilborn, C. B. Callahan, R. M. Cliché, L. S. Prout

and T. L. Morelli. 2021. Abiotic stress and biotic factors mediate range dynamics on opposing
edges. Journal of Biogeography DOI: 10.1111/jbi.14112.

Sozio, G., and A. Mortelliti. 2016. Empirical evaluation of the strength of interspecific competition in
shaping small mammal communities in fragmented landscapes. Landscape Ecology 31(4):775–
789.

102
Steenweg, R., J. Whittington, M. Hebblewhite, A. Forshner, B. Johnston, D. Petersen, B. Shepherd, and P.
M. Lukacs. 2016. Camera-based occupancy monitoring at large scales: Power to detect trends in
grizzly bears across the Canadian Rockies. Biological Conservation 201:192–200.

Steenweg, R., M. Hebblewhite, R. Kays, J. Ahumada, J. T. Fisher, C. Burton, S. E. Townsend, C. Carbone, J.

M. Rowcliffe, J. Whittington, J. Brodie, J. A. Royle, A. Switalski, A. P. Clevenger, N. Heim, and L. N.
Rich. 2017. Scaling-up camera traps: Monitoring the planet’s biodiversity with networks of
remote sensors. Frontiers in Ecology and the Environment 15:26–34.

Stewart, F. E. C., J. T. Fisher, A. C. Burton, and J. P. Volpe. 2018. Species occurrence data reflect the
magnitude of animal movements better than the proximity of animal space use. Ecosphere 9(2):
e02112

Stewart, F. E. C., J. P. Volpe, and J. T. Fisher. 2019. The Debate About Bait: A Red Herring in Wildlife
Research. Journal of Wildlife Management 83:985–992.

Stokeld, D., A. S. K. Frank, B. Hill, J. L. Choy, T. Mahney, A. Stevens, S. Young, D. Rangers, W. Rangers, and
G. R. Gillespie. 2015. Multiple cameras required to reliably detect feral cats in northern Australian
tropical savanna: An evaluation of sampling design when using camera traps. Wildlife Research
42(8):642–649.

Stone, K. D., and J. A. Cook. 2002. Molecular evolution of Holarctic martens (genus Martes, Mammalia:
Carnivora: Mustelidae). Molecular Phylogenetics and Evolution 24:169–179.

Suffice, P., H. Asselin, L. Imbeau, M. Cheveau, and P. Drapeau. 2017. More fishers and fewer martens
due to cumulative effects of forest management and climate change as evidenced from local
knowledge. Journal of Ethnobiology and Ethnomedicine 13:51.

Suffice, P., M. Cheveau, L. Imbeau, M. J. Mazerolle, H. Asselin, and P. Drapeau. 2020. Habitat, climate,
and fisher and marten distributions. Journal of Wildlife Management 84:277–292.

Sullivan, T. P., D. S. Sullivan, and J. H. Sullivan. 2017. Mammalian responses to windrows of woody debris
on clearcuts: Abundance and diversity of forest-floor small mammals and presence of small
mustelids. Forest Ecology and Management 399:143–154.

Sun, C. C., A. K. Fuller, and J. A. Royle. 2014. Trap configuration and spacing influences parameter
estimates in spatial capture-recapture models. PLoS One 9(2):e88025.

Sweitzer, R. A. and B. J. Furnas. 2016. Data from camera surveys identifying co-occurrence and
occupancy linkages between fishers (Pekania pennanti), rodent prey, mesocarnivores, and larger
predators in mixed-conifer forests. Data in Brief 6:783–792.

103
Thompson, I. D. 1988. Habitat needs of furbearers in relation to logging in Boreal Ontario. The Forestry
Chronicle 64:251–261

Tucker, J. M., M. K. Schwartz, R. L. Truex, S. M. Wisely, and F. W. Allendorf. 2014. Sampling affects the
detection of genetic subdivision and conservation implications for fishers in the Sierra Nevada.
Conservation Genetics 15:123–136.

Turcotte, Y., and A. Desrochers. 2003. Landscape-dependent response to predation risk by forest birds.
OIKOS 100(3):614–618.

Vitale, A. A., S. T. McKinney, and D. W. Linden. 2018. Maternal effect and interactions with philopatry in
subadult female American black bear, Ursus americanus, den selection. Animal Behaviour
138:131–139. Elsevier Ltd. <https://doi.org/10.1016/j.anbehav.2018.02.008>.

Weir, R. D, A. S. Harestad, and R. C. Wright. 2005. Winter diet of fishers in British Columbia.
Northwestern Naturalist 86(1):12–19.

Weir, R. D., M. Phinney, and E. C. Lofroth. 2012. Big, sick, and rotting: Why tree size, damage, and decay
are important to fisher reproductive habitat. Forest Ecology and Management 265:230–240.

Wengert, G. M., M. W. Gabriel, S. M. Matthews, J. M. Higley, R. A. Sweitzer, C. M. Thompson, K. L.

Purcell, R. H. Barrett, L. W. Woods, R. E. Green, S. M. Keller, P. M. Gaffney, M. Jones, and B. N.
Sacks. 2014. Using DNA to describe and quantify interspecific killing of fishers in California.
Journal of Wildlife Management 78:603–611.

White, G. C. 2005. Correcting wildlife counts using detection probabilities. Wildlife Research 32:211–
216.

White, H. B., T. Decker, M. J. O. Brien, J. F. Organ, and N. M. Roberts. 2015. Trapping and furbearer
management in North American wildlife conservation. International Journal of Environmental
Studies 72:1–14.

Williams, B. W., J. H. Gilbert, and P. A. Zollner. 2007. Historical perspective on the reintroduction of the
fisher and American marten in Wisconsin and Michigan. General Technical Report NRS-5 29.

Williams, B. W., D. R. Etter, D. W. Linden, K. F. Millenbah, S. R. Winterstein, and K. T. Scribner. 2009.

Noninvasive hair sampling and genetic tagging of co-distributed fishers and American martens.
Journal of Wildlife Management 73:26–34.

Wintle, B. A., M. C. Runge, and S. A. Bekessy. 2010. Allocating monitoring effort in the face of unknown
unknowns. Ecology Letters 13(11):1325–1337.

104
Wright, J. L. 1999. Winter home range and habitat use by sympatric fishers (Martes pennanti) and
American martens (Martes americana) in northern Wisconsin. Thesis, University of Wisconsin,
Stevens Point, Wisconsin USA.

Wulder, M. A., T. R. Loveland, D. P. Roy, C. J. Crawford, J. G. Masek, C. E. Woodcock, R. G. Allen, M. C.

Anderson, A. S. Belward, W. B. Cohen, J. Dwyer, A. Erb, F. Gao, P. Griffiths, D. Helder, T.
Hermosilla, J. D. Hipple, P. Hostert, M. J. Hughes, J. Huntington, D. M. Johnson, R. Kennedy, A.
Kilic, Z. Li, L. Lymburner, J. McCorkel, N. Pahlevan, T. A. Scambos, C. Schaaf, J. R. Schott, Y.
Sheng, J. Storey, E. Vermote, J. Vogelmann, J. C. White, R. H. Wynne, and Z. Zhu. 2019. Current
status of Landsat program, science, and applications. Remote Sensing of Environment 225:127–
147.

Yaeger, J. S. 2005. Habitat at fisher resting sites in the Klamath Province of Northern California. Thesis,
Humboldt State University, Humboldt, California, USA.

Zielinski, W. J. and N. P. Duncan. 2004. Diets of sympatric populations of American martens (Martes
americana) and fishers (Martes pennanti) in California. Journal of Mammalogy 85:470–477.

Zielinski, W. J., R. L. Truex, G. A. Schmidt, F. V. Schlexer, K. N. Schmidt, and R. H. Barrett. 2004. Home
Range Characteristics of Fishers in California. Journal of Mammalogy 85:649–657.

Zielinski, W. J., C. M. Thompson, K. L. Purcell, and J. D. Garner. 2013. An assessment of fisher (Pekania
pennanti) tolerance to forest management intensity on the landscape. Forest Ecology and
Management 310:821–826.

Zuur, A.F., E. N. Ieno, and G. M. Smith. 2007. Analysing Ecological Data. Springer, New York, New York.
672 p.

105
 APPENDIX A: ASSOCIATED PUBLICATION ABSTRACTS
Two other publications have been produced with data collected and curated during this project. Bryn
Evans was a co-author on both, and the findings may be relevant to readers of this dissertation. The first
has been published in Mammalian Biology, while the second is undergoing peer review for publication
by Scientific Reports. Full copies, once published, will be available at:
https://www.researchgate.net/profile/Bryn-Evans-2 or by emailing: [email protected], and the abstracts
are included below.

Assessing the effectiveness of attractants to increase camera trap detections of North American
mammals

Michael Buyaskas, Bryn E. Evans, Alessio Mortelliti

Abstract

Camera traps are a cost-effective tool for large scale and long-term population monitoring of
mammals. Either bait or lure are often used to attract animals in front of a camera, however the relative
efficiency of these two attractants in combination is not well understood. Our objective was to
determine the optimal attractant setup for maximizing detection probabilities of mammalian
communities in the northeast USA. We conducted a camera trapping project in northern Maine, USA,
from August to November 2018, and tested 3 distinct attractant setups. Sampling stations were a
minimum of 5km apart, and consisted of four camera units spaced 100 meters apart, and each sampling
unit constituted either a treatment or a control: 1) bait plus lure (treatment), 2) bait (treatment), 3) lure
(treatment), and 4) camera only (control). Detection data on 11 species of mammals was collected from
41 stations and analyzed through multi-method occupancy models. We recorded a total of 4,280 total
trap-nights and captured 37,781 images. Results showed that the combination attractant of bait plus
lure was the most effective for increasing detection probability of carnivores. Specifically bait plus lure
proved to be particularly effective for mustelid species, while lure was particularly effective for
American black bear (Ursus americanus). While attractant usage was shown to be ineffective for
increasing detection probability of non-carnivores, it also did not decrease effectiveness. Based on our
results we recommend the simultaneous use of both bait and lure as attractants when conducting
camera trapping work on mammals. The combination of bait and lure appears to maximize detection of
carnivore species while simultaneously having minimal effects on the detection of other taxa.

106
 Umbrella effect of monitoring protocols for mammals in the Northeast US

Alessio Mortelliti, Allison M. Brehm and Bryn E. Evans Bryn

Abstract

Developing cost-effective monitoring protocols is a priority for wildlife conservation agencies

worldwide. In particular, developing protocols that cover a wide range of species is highly desirable.
Here we applied the ‘umbrella species’ concept to the context of ecological monitoring; specifically
testing the hypothesis that protocols developed for the American marten would contextually allow
detecting occupancy trends for 13 other mammalian species. We conducted a large-scale four-year
camera trapping survey across a gradient of forest disturbance in Maine, USA. We sampled 197 sites
using a total of 591 cameras and collected over 800,000 photographs to generate detection histories for
the most common terrestrial species. By combining multi-season occupancy modelling and power
analyses, we estimated the required sampling effort to detect 10%, 25% and 50% declines in the target
species. By conducting a spatially explicit comparison of sampling effort, we found evidence that
developing monitoring protocols for American marten would provide an umbrella effect for up to 11
other mammal species. The efficiency of the umbrella effect varied among species, with fisher,
snowshoe hare, red squirrel, and black bear consistently covered under several scenarios. Our results
support the application of the umbrella species concept to monitoring (here defined as ‘umbrella
monitoring species’), providing empirical evidence for its use by management agencies.

107
 APPENDIX B: SUPPORTING INFORMATION FOR CAMERA ARRAY SELECTION (CHAPTER 1)

Appendix B.1: Comparison between 24-hour and 1-week detection windows

How camera trap data are delineated into “visits” will depend on study design and researcher
objectives. With a relatively short time window to conduct our field work, we elected to place our
survey stations for a minimum of two weeks and analyze each 24-hour window as a unique “visit”. From
a qualitative assessment of our detection history data, the occasions with positive detections did not
appear to be clumped in such a way that would suggest this 24-hour binning violated assumptions of
temporal independence. However, we further assessed this by comparing the model fit and the results
for Psi, theta and p_method from a weekly-binning approach instead, incorporating the same top model
covariates.

Generally, the trend in the effect that the method had on detection given availability was the
same (Fig B.1), although absolute values of estimates varied (Table B.1). For coyote, the model was
unable to converge on an estimate for Psi, and for weasel we see that the standard errors are noticeably
larger than when running with the 24-hour detections.

108
Fig B.1. Detection probabilities for five different camera transect arrays, comparing between 24-hour
and 1-week temporal bins.

Coyote Fisher
0.8 1
0.6 0.8
0.4 0.6
0.4
0.2
0.2
0
0
p1 p2 p3 p4 p5
p1 p2 p3 p4 p5
Coyote 24Hr Coyote 1Week
Fisher 24Hr Fisher 1Week

Marten Weasel
1 1
0.8 0.8
0.6 0.6
0.4 0.4
0.2 0.2
0 0
p1 p2 p3 p4 p5 p1 p2 p3 p4 p5

Marten 24Hr Marten 1Week Weasel 24Hr Weasel 1Week

Snowshoe Hare Red Squirrel

0.8 1
0.6 0.8
0.4 0.6
0.2 0.4
0 0.2
p1 p2 p3 p4 p5 0
p1 p2 p3 p4 p5
Snowshoe 24Hr Snowshoe 1Week
Squirrel 24Hr Squirrel 1Week

109
Table B.1. Comparisons by species for 24-hour temporal binning versus 1-week binning for detection histories.

Species Temporal bin ̂ (se)

Ѱ ̂𝟏 (se)
𝒑 ̂𝟐 (se)
𝒑 ̂𝟑 (se)
𝒑 ̂𝟒 (se)
𝒑 ̂𝟓 (se)
𝒑

24-hour 0.67(0.17) 0.13(0.07) 0.51(0.09) 0.65(0.08) 0.34(0.09) 0.62(0.09)

Canis latrans
1-week Did not converge 0.17(0.05) 0.45(0.11) 0.72(0.11) 0.54(0.11) 0.63(0.12

24-hour 0.73(0.07) 0.60(0.16) 0.79(0.11) 0.83(0.09) 0.71(0.14) 0.73(0.13)

Pekania pennanti
1-week 0.56(0.13) 0.41(0.09) 0.75(0.06) 0.86(0.03) 0.61(0.07) 0.74(0.06)

24-hour 0.76(0.08) 0.29(0.07) 0.47(0.06) 0.79(0.06) 0.55(0.07) 0.58(0.07)

Martes americana
1-week 0.79(0.09) 0.41(0.08) 0.79(0.05) 0.89(0.03) 0.72(0.05) 0.75(0.05)

24-hour 0.51(0.12) 0.24(0.05) 0.47(0.06) 0.60(0.06) 0.52(0.06) 0.77(0.05)

Mustela erminea
1-week 0.84(0.25) 0.21(0.15) 0.64(0.36) 0.55(0.33) 0.52(0.31) 0.55(0.33)

24-hour 0.50(0.12) 0.37(0.07) 0.49(0.08) 0.72(0.06) 0.49(0.07) 0.61(0.07)

Lepus americanus
1-week 0.51(0.14) 0.29(0.16) 0.59(0.13) 0.75(0.09) 0.44(0.16) 0.50(0.15)

24-hour 0.53(0.31) 0.34(0.10) 0.66(0.09) 0.81(0.06) 0.57(0.09) 0.69(0.09)

Tamiasciurus
hudsonicus 1-week 0.57(0.32) 0.32(0.13) 0.78(0.10) 0.75(0.12) 0.59(0.15) 0.63(0.14)

The different methods compared are: p1 is a single camera, p2 is two cameras spaced 100 m apart, p3 is two cameras spaced 150 m, p4 is
three cameras spaced 100 m and p5 is three cameras spaced 150 m.

110
Appendix B.2: Results of phase 1 modeling to compare detection by method versus constant

The first step in our multi-scale occupancy model approach was to determine if our approach to
pooling detection history data from different arrangements of cameras at the same survey site was
valid. We did this by comparing a multi-scale model with no covariates, but with the detection
parameter allowed to vary by each method p, to a model in which detection was held constant. In all
cases the model including method ranked much higher (>∆AICc 8, with model weight always rounded to
1.00) as shown in Table B.2.

Table B.2. Model ranking comparing the inclusion of the “method” covariate for detection probability
versus a null detection model.

Species Competing Models AICc ∆AICc AICw

Canis latrans Ѱ(.), ϴ(.), p(Method) 424.66 0 1.00

Ѱ(.), ϴ(.), p(.) 433.23 8.82 0.00

Pekania pennanti Ѱ(.), ϴ(.), p(Method) 683.86 0 1.00

Ѱ(.), ϴ(.), p(.) 698.10 14.97 0.00

Martes americana Ѱ(.), ϴ(.), p(Method) 1140.82 0 1.00

Ѱ(.), ϴ(.), p(.) 1179.25 38.43 0.00

Mustela erminea Ѱ(.), ϴ(.), p(Method) 875.75 0 1.00

Ѱ(.), ϴ(.), p(.) 911.23 35.48 0.00

Lepus americanus Ѱ(.), ϴ(.), p(Method) 1265.24 0 1.00

Ѱ(.), ϴ(.), p(.) 1324.72 59.08 0.00

Tamiasciurus hudsonicus Ѱ(.), ϴ(.), p(Method) 575.87 0 1.00

Ѱ(.), ϴ(.), p(.) 588.35 12.48 0.00

111
Appendix B.3: Model comparison between different possible pairs for methods with two cameras.

The “2-short” method, with 100 m spacing, composed of the “Original” pair used in our
analyses, cameras B and C in fig. 1.2, and the “Alternative” pair, cameras A and B. The “2-Long” method
with 150 m spacing is “Original” pair D and E, and “Alternative” pair B and C.

The Alternative pairs were used to create detection histories following the same methodology,
and replaced the Original methods for 2-short and for 2-long. We ran the same multi-scale analysis
(Nichols et al. 2008), and saw the same pattern for the support of “method” as a driving factor in the
detection process (Table B.3). We then compared the model output and observed the same trends in
terms of 1, 2 and 3 cameras at the different spacing using both the Alternative and the Original pair
(Table B.4). Because the Original pair fully separates the detection histories used between method 1
(camera A) and 2-short, and between 2-short and 3-short, we carried out the rest of our analysis only
using these pairs.

112
Table B.3. Comparison of model ranking output between the “original” pairing and the “alternative”
pairing configurations possible from out T-array of five camera traps.

Species Pairs Models ∆AICc

Ѱ(.), ϴ(.), p(mOriginal) 0
Original
Ѱ(.), ϴ(.), p(.) 10.23
Canis latrans
Ѱ(.), ϴ(.), p(mAlternative) 0
Alternative
Ѱ(.), ϴ(.), p(.) 8.82
Ѱ(.), ϴ(.), p(mOriginal) 0
Original
Ѱ(.), ϴ(.), p(.) 18.44
Pekania penanti
Ѱ(.), ϴ(.), p(mAlternative) 0
Alternative
Ѱ(.), ϴ(.), p(.) 10.43
Ѱ(.), ϴ(.), p(mOriginal) 0
Original
Ѱ(.), ϴ(.), p(.) 44.83
Martes americana
Ѱ(.), ϴ(.), p(mAlternative) 0
Alternative
Ѱ(.), ϴ(.), p(.) 37.78
Ѱ(.), ϴ(.), p(mOriginal) 0
Original
Ѱ(.), ϴ(.), p(.) 38.43
Martes erminea
Ѱ(.), ϴ(.), p(mAlternative) 0
Alternative
Ѱ(.), ϴ(.), p(.) 35.48
Ѱ(.), ϴ(.), p(mOriginal) 0
Original
Ѱ(.), ϴ(.), p(.) 26.97
Lepus americanus
Ѱ(.), ϴ(.), p(mAlternative) 0
Alternative
Ѱ(.), ϴ(.), p(.) 25.0
Ѱ(.), ϴ(.), p(mOriginal) 0
Original
Tamiasciurus Ѱ(.), ϴ(.), p(.) 17.02
hudsonicus Ѱ(.), ϴ(.), p(mAlternative) 0
Alternative
Ѱ(.), ϴ(.), p(.) 12.48

113
Table B.4. Comparison of beta estimates between the “original” pairing and the “alternative” pairing
configurations possible from out T-array of five camera traps.

Original Pairs Alternative Pairs

Species Method
Untransformed Beta Est. Untransformed Beta Est.
1 camera -1.632 -1.599
2 short 0.086 0.149
Canis latrans 3 short -0.307 -0.255
2 long -0.726 0.567
3 long 0.627 0.716
1 camera -0.468 -0.466
2 short 0.537 0.389
Pekania pennanti 3 short 1.462 1.467
2 long -0.125 0.314
3 long 0.779 0.783
1 camera -0.578 -0.575
2 short 0.122 0.442
Martes americana 3 short 0.776 0.782
2 long 0.438 0.651
3 long 1.582 1.593
1 camera -1.185 -1.179
2 short 0.001 -0.183
Martes erminea 3 short 0.402 0.413
2 long -0.191 0.362
3 long 1.077 1.094
1 camera -1.001 -1.018
2 short -0.520 -0.464
Lepus americanus 3 short 0.347 0.318
2 long -0.082 -0.648
3 long 0.384 0.342
1 camera -1.225 -1.221
2 short 0.0389 -0.132
Tamiasciurus
3 short 0.586 0.596
hudsonicus
2 long -0.318 0.135
3 long 0.785 0.797

114
 APPENDIX C: FOREST DISTURBANCE INDEX

Forest disturbance metrics were obtained from remote sensed Landsat image data that were processed

using a novel method of stacked generalizations from multiple individual change-detection algorithms

fused together with a secondary classifier (Kilbride 2018). We categorized the continuous raster product

describing the intensity of disturbance most recently experienced by each 30m pixel into four discrete groups

(Intensity), and the year of the most recent disturbance event (Year). We multiplied these two layers both

with values from 1 to 4 using raster math to create a single variable with six levels incorporating both the

magnitude and the recency of forest change (Disturbance Score). For example, a pixel experiencing low

magnitude disturbance (intensity value = 2) in a less recent time period (1989 – 2000, year = 2) would have a

disturbance score of 4, while a pixel with very recent, moderate intensity disturbance would be scored 12.

Disturbance Score
Intensity Year 1 None / historic
1 None / historic 1 None / historic (Prior to 1989) 4 Low
2 Low 2 Less recent (1989 – 2000) 6 Moderately low
x =
3 Moderate 3 Recent (2001 – 2012) 8 Moderate
4 High 4 Very recent (2013 – 2017) 12 High
16 Extreme

Although many methods to consolidate and discretize this detailed forest change information

are possible, this approach was discussed with the author of the novel dataset and we agreed that it

provided the simplest, least biased way to provide a single metric to encompass the incredibly complex

reality of forest disturbance.

Using multiple different spatial scales also allowed me to ensure that some elements of the

patch size of the disturbance category were represented in both the northern and southern parts of the

study. When selecting townships to focus survey effort, I looked for both “Contiguous” areas, where

multiple townships in a continuous block were all either “High” disturbance or “Low” disturbance (the

upper and lower 10% of all townships in the study area). I also selected townships that were in small

patches (one or two townships) of one extreme, surrounded by townships of the opposite disturbance
115
category, and called these “Disjunct” areas. These initial selection criteria was reflected in the finer-scale

buffers then used for modeling species occupancy. A subset are shown below, both in the northern half

(site names labeled as N) and in the southern half (site names with S).

STATION Origal cat as observed 300m MEAN 1k MEAN 3k MEAN 6k MEAN

S01_17 Low Contiguous 91.76419214 83.37009565 129.0187751 234.6926488
N06_14 Low Contiguous 18.82017544 19.88862323 101.5393378 164.4113859
N09_03 Low Disjunct 88.57700651 77.01655629 186.8627944 334.1638222
S10_03 High Disjunct 368.1786492 325.9986248 207.2510273 169.2629898
N01_02 High Disjunct 116.40625 290.9104267 299.3973102 385.4450435
N11_05 High Contiguous 181.2516411 378.3470472 358.1033104 321.6710525
N12_07 High Contiguous 314.6688453 493.9351562 296.3825477 88.61446059

I also collected qualitative habitat notes at the camera stations as they were deployed, and

matched these against the remote sensed layers. The only discordance occurred at sites that had been

clearly harvested in <1 year, and these were excluded from further analyses (e.g. cut from sampling

effort, as the disturbance layer would no longer correspond to the habitat that animals were

experiencing).

116
 APPENDIX D: SUPPORTING INFORMATION FOR LARGE-SCALE CAMERA SURVEYS (CHAPTERS 2-4)

Appendix D.1: Study design for large-scale camera trapping surveys

Fig D.1. Boxplots illustrating the independence of predictor variables across a balanced study design to
understand the roles of forest disturbance, latitude, and habitat characteristics on carnivores in Maine.

117
Fig D.2. Forest disturbance categories at the township level across Maine, USA. Townships were selected
as “high priority” to collect detection- non-detection data on carnivores in order to balance survey effort
across a) very low to very high disturbance, b) relative continuity in disturbance between a single
township and those surrounding, and c) latitude.

118
Fig D.3. Reported harvest success via both trapping and hunting over the 2014 to 2018 seasons. A)
Marten (Martes americana), B) fisher (Pekania pennanti), and C) coyote (Canis latrans).

119
Fig D.4. Correlation coefficients between predictor variables for occupancy models for carnivores in
Maine, USA.

120
Fig D.5. Example configuration of A) independent camera stations spaced at least 6 k apart, B) three
camera sites along a 200 m linear transect at each station and C) the setup of the camera and bait cage
at a single site.

121
Appendix D.2: Two-species models for marten and fisher interaction

The two-species models described by MacKenzie and co-authors1,2 and implemented in program
Presence3 allow occupancy parameters to vary in response to both the occupancy status and the
detection patterns of a second co-occuring species.

For initial occupancy, there are two possible scenarios for each species:

ψF versus ψf (marten initial occupancy at a site occupied by fisher, versus at a site without fisher) and ψM
versus ψm (fisher initial occupancy at a site occupied by marten, versus at a site without marten).

For colonization and extinction, there are four possible scenarios. The first letter in the superscript
indicates the status of the other species at time = t, and the second is the status at time = t+1.

Colonization for marten, with all four combinations of fisher status, is illustrated below:

γFF (fisher present at both times), γFf (fisher present at time = t and absent at time = t+1), γfF (fisher
absent at time = t and present at time = t+1), γff (fisher absent at both times).

For detection, there are three possible scenarios. p is used to describe the probability of detection at a
site where the other species is absent (and therefore not detected on any occasion) and r is used to
describe the probability of detection at a site where the other species is present, conditioned on if it was
detected in that same occasion:

pf (probability of detecting a marten at a site where fisher are absent)

rFF versus rFf (probability of detecting a marten at a site where fisher is present and during the same
occasion when fisher is detected, versus the probability at a site where fisher is present, but during an
occasion when only marten is detected).

1. MacKenzie, D. I., L. L. Bailey and J. D. Nichols. 2004. Investigating species co-occurrence patterns
when species are detected imperfectly. Journal of Animal Ecology 73:546-555.
2. MacKenzie, D. I., J. D. Nichols, J. A. Royle, K. H. Pollock, L. L. Bailey, and J. E. Hines. 2017. Occupancy
Estimation and Modeling: Inferring Patterns and Dynamics of Species Occurrence: Second Edition.
Elsevier Academic Press, Boston, Massachusetts, USA.
3. Hines, J. E. 2006. PRESENCE: Software to estimate patch occupancy and related parameters. USGS-
PWRC. Version 2.13.12. https://www.mbr-pwrc.usgs.gov/software/presence

122
Table D.1. Parameterization of marten (Martes americana) occupancy models used to compare the relative importance of interspecific dynamics
from fisher (Pekania pennanti) presence and detection, and descriptions of the interaction being tested.

Parameter Model Description of interaction

Initial ψF = ψf Probability of marten occupancy is the same at sites where fisher are present and where fisher are absent
occupancy F
ψ ≠ψ f
Probability of marten occupancy is different at sites where fisher are present versus where fisher are absent
FF Ff fF ff
Colonization γ =γ =γ =γ Probability of marten colonization (t = absent, t + 1 = present) is the same regardless of fisher presence in
time t and in time t + 1
γFF = γFf ≠ γfF = γff Probability of marten colonization is different when fisher are present in t versus fisher absent in t
FF fF Ff ff
γ =γ ≠γ =γ Probability of marten colonization is different when fisher are present in t + 1 versus fisher absent in t + 1
FF Ff fF ff
γ =γ =γ ≠γ Probability of marten colonization is different when fisher are present in any time versus when fisher were
absent in both times
γFF = γFf = γff ≠ γfF Probability of marten colonization is different when fisher also colonizes versus any other combination of
present / absent
Extinction εFF = εFf = εfF = εff Probability of marten extinction (t = present, t + 1 = absent) is the same regardless of fisher presence in time
t and in time t + 1
εFF = εFf ≠ εfF = εff Probability of marten extinction is different when fisher are present in t versus fisher absent in t
FF fF Ff ff
ε =ε ≠ε =ε Probability of marten extinction is different when fisher are present in t + 1 versus fisher absent in t + 1
εFF = εFf = εfF ≠ εff Probability of marten extinction is different when fisher are present in any time versus when fisher were
absent in both times
εFF = εfF = εff ≠ εFf Probability of marten extinction is different when fisher also becomes locally extinct versus any other
combination of present / absent
Detection pf = rFF = rFf Probability of marten detection is the same regardless of fisher presence or detection
f FF Ff
p ≠r =r Probability of marten detection is different at sites where fisher are present versus sites where fisher are
absent
pf = rFf ≠ rFF Probability of marten detection is different on an occasion when fisher is also detected versus an occasion
when fisher is not detected or at a site where fisher is absent
pf ≠ rFF ≠ rFf Probability of marten detection is different at sites where fisher are absent, and different on occasions
when fisher are detected versus not detected at sites where present

123
Table D.2. Parameterization of fisher (Pekania pennanti) occupancy models used to compare the relative importance of interspecific dynamics
from marten (Martes americana) presence and detection, and descriptions of the interaction being tested.

Parameter Model Description of interaction

M m
Initial ψ =ψ Probability of fisher occupancy is the same at sites where marten are present and where marten are absent
occupancy ψ ≠ψM m
Probability of fisher occupancy is different at sites where marten are present versus where marten are
absent
Colonization γMM = γMm = γmM = γmm Probability of fisher colonization (t = absent, t + 1 = present) is the same regardless of marten presence in
time t and in time t + 1
γMM = γMm ≠ γmM = γmm Probability of fisher colonization is different when marten are present in t versus marten absent in t
MM mM Mm mm
γ =γ ≠γ =γ Probability of fisher colonization is different when marten are present in t + 1 versus marten absent in t + 1
MM Mm mM mm
γ =γ =γ ≠γ Probability of fisher colonization is different when marten are present in any time versus when marten were
absent in both times
γMM = γMm = γmm ≠ γmM Probability of fisher colonization is different when marten also colonizes versus any other combination of
present / absent
Extinction εMM = εMm = εmM = εmm Probability of fisher extinction (t = present, t + 1 = absent) is the same regardless of marten presence in time
t and in time t + 1
εMM = εMm ≠ εmM = εmm Probability of fisher extinction is different when marten are present in t versus marten absent in t
MM mM Mm mm
ε =ε ≠ε =ε Probability of fisher extinction is different when marten are present in t + 1 versus marten absent in t + 1
εMM = εMm = εmM ≠ εmm Probability of fisher extinction is different when marten are present in any time versus when marten were
absent in both times
εMM = εmM = εmm ≠ εMm Probability of fisher extinction is different when marten also becomes locally extinct versus any other
combination of present / absent
Detection pm = rMM = rMm Probability of fisher detection is the same regardless of marten presence or detection
m MM Mm
p ≠r =r Probability of fisher detection is different at sites where marten are present versus sites absent
m Mm MM
p =r ≠r Probability of fisher detection is different on an occasion when marten is also detected versus an occasion
when marten is not detected or at a site where marten is absent
pm ≠ rMM ≠ rMm Probability of fisher detection is different at sites where marten are absent, and different on occasions when
marten are detected versus not detected at sites where present

124
Appendix D.3: Additional results of marten and fisher two-species models

Table D.3. Survey effort over seven seasons of camera trap transects deployed in Maine, USA.

Summer Winter Summer Winter Summer Winter Summer

2017 2018 2018 2019 2019 2020 2020
Total stations 121 119 88 85 84 78 43
New station 121 119 48 46 28 28 -
Revisits 3-4 - - 40 39 17 14 -
Revisits 5-6 - - - - 39 36 11
Revisit 7 - - - - - - 32

Table D.4. Naive occupancy rates and raw count of stations which detected marten (Martes americana),
fisher (Pekania pennanti), and both species together across seven survey periods.

Species Summer Winter Summer Winter Summer Winter Summer

2017 2018 2018 2019 2019 2020 2020
(n=121) (n=119) (n=88) (n=85) (n=84) (n=78) (n=43)

Martes americana 0.38 (46) 0.48 (57) 0.35 (31) 0.61 (52) 0.50 (42) 0.58 (45) 0.51 (22)
Pekania pennanti 0.26 (31) 0.76 (91) 0.41 (36) 0.62 (53) 0.39 (33) 0.86 (67) 0.49 (21)
Both species 0.08 (10) 0.39 (47) 0.17 (15) 0.39 (33) 0.23 (19) 0.51 (40) 0.26 (11)

125
Table D.5. Parameter estimates (betas) and standard errors for the top ranked single-species, multi-season occupancy models of marten (Martes
americana) and fisher (Pekania pennanti), untransformed.

Species Occupancy Colonization Extinction Detection

Martes americana Intercept: -0.457(0.174) Intercept: -2.62(0.673) Intercept: 0.469(0.711) Intercept: -1.614(0.065)
DI_3k: -0.645(0.186) DI_300m: -0.510(0.178) DI_300m: 0.524(0.203) DI_3k: -0.233(0.032)
DC_300m: 0.843(0.182) Snow: 3.02(1.22) Snow: -3.00(1.21) Day: -0.025(0.004)

Pekania pennanti Intercept: 0.16(0.262) Intercept: 1.394(0.554) Intercept: 0.438(0.672) Intercept (S17): -3.238(0.207)
DI_6k: -0.455(0.254) DI_6k: -0.253(0.168) W18: 1.573(0.216)
DC_300m: 1.291(0.607) Snow: -2.470(1.135) S18: 0.149(0.243)
W19: 1.161(0.226)
S19: 0.650(0.245)
W20: 1.714(0.215)
S20: 0.894(0.255)
DI: Disturbance intensity, at the most influential spatial scale

DC: Percentage of living trees that were deciduous species, at the most influential scale

Snow: Daily values recorded within a 1 k radius buffer by National Snow and Ice Data Center (NSIDC) Snow Data Assimilation System (SNODAS),
by season (winter seasons January 15 to March 15, summer seasons recorded no snow presence)

Survey: The season (summer or winter) and year (2017-2020) of data collection, here used as a categorical variable

126
Fig D.6. Examples of images positively identified as American marten (Martes americana) (A,C) and
fisher (Pekania pennanti) (B,D). The detections below also illustrate the spatial overlap observed
between these two species, both in summer and in winter surveys, and allow comparison between their
physical characteristics.

127
Fig D.7. Spatial distribution of detection for American marten (Martes americana) (red) and fisher
(Pekania pennanti) (blue) over 4 years of camera trap surveys in Maine, USA.

128
Appendix D.4: Weasel image tagging protocol

We collected data from 197 camera survey stations (three motion-triggered trail cameras with
bait and lure) over four years in Maine, USA. Of the 800,000 images collected, 7,473 were initially tagged
as triggered by a “weasel” without further identification to species. We followed the steps below to
categorize all weasel detections to a) one of the three species present in Maine and b) by the level of
confidence (certain or uncertain) in that species identification.

1) Group all weasel images by dependent events, either a single image tagged as weasel, or a series of
images recorded with <30 minute gaps between any two images tagged as weasel.
2) Review all weasel images* and assign a score to each image for
A) Image quality,
B) body measurements, and
C) qualitative assessment.
*For series longer than 40 images, images were inspected and assigned to each category until a
‘certain’ identity was confirmed, and the rest were only briefly viewed.
A) Image quality
A.1 Excellent = daylight, clear, weasel body all visible, posture straight, perpendicular to camera.
Measurements and assessment very reliable
A.2 Good = daylight or clear night image, weasel body all visible, posture straight or curved, angle
<30° from camera. Measurements good, consider reliable
A.3 Fair = blurry image, weasel body <25% obscured, posture curved, angled >30° away from
camera. CAN take measurements, but not reliable
A.4 Poor = Can see it is a weasel but measurements not possible. If it’s the only image, may attempt
suspected ID on qualitative metrics, but unreliable
A.5 None = Not enough to even confirm a weasel, or nothing in frame (e.g. images recorded after
weasel left the frame)
B) Body measurements
We used the tool SmallMeasure, freely available from DonationCoder member Skrommel’s One
Hour Software (https://www.dcmembers.com/skrommel/download/smallmeasure/).
B.1 Tail length was measured when tails were straight and extended perpendicular to the camara
B.2 Head plus body length was measured when the weasel was perpendicular to the camera, and if
not fully extended, would be measured in segments that were then added together
B.3 The bait cage, if the weasel was located at the same distance from the camera as it (typically
when climbing the tree to investigate the bait). We predominately used C&S EZ Fill Suet Baskets, or
when unavailable similar products, which measure 15 cm from bottom to top of the front face. See
Fig. SD2.1 for an example of the SmallMeasure tool.

129
All images were reviewed on the same Dell Latitude laptop with the image displayed at full screen.
The unit of measurement is pixels, and while it should be noted that these do NOT translate into
estimates of actual size (given that the distance from the camera is unknown), but as long as the
overall image quality and body position were appropriate, measurements provided a ratio of tail
length to head plus body length to help differentiate between weasel species (Elsasser and Parker
2008, Naughton 2012).

Fig D.8. The red and

blue triangles indicate
the points that
SmallMeasure
provides the distance
between, in pixels, at
any angle. A) Tail
length of 44 pixels, B)
body plus head length
of 117 pixels, and C)
the bait cage of 70
pixels, and was
included since the
weasel is positioned
at the same distance
from the camera.

130
4) The final step for generating detection histories was to collapse any events within a single 24-hour
period (midnight to midnight) and separate by species. For example, if within a single day a short-tailed
uncertain detection event occurred in the morning and a short-tailed certain event in the evening, that
24-hour period would be scored in the short-tail weasel dataset as “2”, short-tail certain. If on another
day, a short-tailed weasel uncertain event was recorded in the morning and then a long-tail certain
event that night, that date and site would be scored as “1”, uncertain, in the short-tail dataset and as
“2”, certain, in the long-tail dataset. The example below (Fig C.10) is actual data from winter 2019, from
a study area in the central north of Maine, where both short-tailed and long-tailed weasels were
detected.

Fig D.9. Examples based on real data (trimmed to 20 days out of 33 for ease of presentation), where A)
is the final clean table with all species and all levels of certainty color-coded for interpretability. B)
Detection histories for short-tailed weasel (top) and long-tailed weasel (bottom) where 0 indicates no
detection on that day of the survey, 1 indicates an uncertain detection, and 2 a certain detection. Note
that on day 17 at station N09_01, both C) and D) long-tailed and E) short-tailed weasels were detected.

131
Appendix D.5: Full model ranking for weasel multi-season false-positive occupancy models

Table D.6. Short-tailed weasel (Mustela erminea) model ranking: All multi-season, false-positive occupancy models tested, ranked via AICc
(corrected for small sample size, n = 197). ΔAICc is the difference from the top model AICc score, Wi is the model weight, k is the number of
parameters.
Model formula AICc ΔAICc Wi k
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(.), p1(days+lat), p10(.), b(.) 5234.28 0.00 0.1704 12
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(dist_300m), p1(days+lat), p10(.), b(.) 5234.78 0.50 0.1327 13
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(lat), p1(days+lat), p10(.), b(.) 5235.34 1.06 0.1003 13
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(dec_300m), p1(days+lat), p10(.), b(.) 5235.89 1.61 0.0762 13
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(dist_300m+lat), p1(days+lat), p10(.), b(.) 5236.03 1.75 0.0710 14
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(dec_stand), p1(days+lat), p10(.), b(.) 5236.06 1.78 0.0700 13
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(dist_1k), p1(days+lat), p10(.), b(.) 5236.23 1.95 0.0643 13
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(dec_1k), p1(days+lat), p10(.), b(.) 5236.33 2.05 0.0611 13
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(dist_6k), p1(days+lat), p10(.), b(.) 5236.34 2.06 0.0608 13
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(dist_3k), p1(days+lat), p10(.), b(.) 5236.42 2.14 0.0584 13
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(dist_300m+dec_300m), p1(days+lat), p10(.), b(.) 5236.52 2.24 0.0556 14
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(dec_300m+lat), p1(days+lat), p10(.), b(.) 5236.83 2.55 0.0476 14
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(dist_300m+lat+dec_300m), p1(days+lat), p10(.), b(.) 5237.66 3.38 0.0314 15
ψ(dist_6k+dec_1k), γ(dist_3k), ε(.), p1(days+lat), p10(.), b(.) 5248.90 14.62 0.0001 11
ψ(dist_6k+dec_1k), γ(dist_1k), ε(.), p1(days+lat), p10(.), b(.) 5251.56 17.28 0 11
ψ(dist_6k+dec_1k), γ(dec_300m), ε(.), p1(days+lat), p10(.), b(.) 5257.04 22.76 0 11
ψ(dist_6k+dec_1k), γ(dec_1k), ε(.), p1(days+lat), p10(.), b(.) 5262.18 27.90 0 11
ψ(dist_6k+dec_1k), γ(dist_6k), ε(.), p1(days+lat), p10(.), b(.) 5262.36 28.08 0 11
ψ(dist_6k+dec_1k), γ(lat), ε(.), p1(days+lat), p10(.), b(.) 5266.29 32.01 0 11
ψ(dist_6k+dec_1k), γ(dec_stand), ε(.), p1(days+lat), p10(.), b(.) 5269.06 34.78 0 11
ψ(dist_6k+dec_1k), γ(snow), ε(.), p1(days+lat), p10(.), b(.) 5270.22 35.94 0 11
ψ(dist_6k+dec_1k), γ(.), ε(.), p1(days+lat), p10(.), b(.) 5271.20 36.92 0 10
ψ(dist_6k), γ(.), ε(.), p1(days+lat), p10(.), b(.) 5273.55 39.27 0 9
ψ(dist_3k), γ(.), ε(.), p1(days+lat), p10(.), b(.) 5275.42 41.14 0 9

132
 Table D.6 continued
ψ(dec_1k), γ(.), ε(.), p1(days+lat), p10(.), b(.) 5276.26 41.98 0 9
ψ(dist_1k), γ(.), ε(.), p1(days+lat), p10(.), b(.) 5276.82 42.54 0 9
ψ(.), γ(.), ε(.), p1(days+lat), p10(.), b(.) 5278.45 44.17 0 8
ψ(dec_300m), γ(.), ε(.), p1(days+lat), p10(.), b(.) 5279.06 44.78 0 9
ψ(dist_300m), γ(.), ε(.), p1(days+lat), p10(.), b(.) 5279.21 44.93 0 9
ψ(snow), γ(.), ε(.), p1(days+lat), p10(.), b(.) 5280.26 45.98 0 9
ψ(lat), γ(.), ε(.), p1(days+lat), p10(.), b(.) 5280.58 46.30 0 9
ψ(dec_stand), γ(.), ε(.), p1(days+lat), p10(.), b(.) 5280.63 46.35 0 9
ψ(.), γ(.), ε(.), p1(days), p10(.), b(.) 5302.12 67.84 0 7
ψ(.), γ(.), ε(.), p1(lat), p10(.), b(.) 5320.21 85.93 0 7
ψ(.), γ(.), ε(.), p1(dbh), p10(.), b(.) 5338.00 103.72 0 7
ψ(.), γ(.), ε(.), p1(dec_1k), p10(.), b(.) 5342.22 107.94 0 7
ψ(.), γ(.), ε(.), p1(dec_300m), p10(.), b(.) 5343.21 108.93 0 7
ψ(.), γ(.), ε(.), p1(dist_3k), p10(.), b(.) 5343.29 109.01 0 7
ψ(.), γ(.), ε(.), p1(cwd_sticks), p10(.), b(.) 5345.39 111.11 0 7
ψ(.), γ(.), ε(.), p1(.), p10(.), b(.) 5345.78 111.50 0 6
ψ(.), γ(.), ε(.), p1(snags), p10(.), b(.) 5346.30 112.02 0 7
ψ(.), γ(.), ε(.), p1(dist_1k), p10(.), b(.) 5347.54 113.26 0 7
ψ(.), γ(.), ε(.), p1(dist_300m), p10(.), b(.) 5347.60 113.32 0 7
ψ(.), γ(.), ε(.), p1(dist_6k), p10(.), b(.) 5347.69 113.41 0 7
ψ(.), γ(.), ε(.), p1(dec_stand), p10(.), b(.) 5347.79 113.51 0 7
ψ(.), γ(.), ε(.), p1(cwd_logs), p10(.), b(.) 5347.90 113.62 0 7
ψ(.), γ(.), ε(.), p1(winter), p10(.), b(.) 5353.01 118.73 0 7
ψ(.), γ(.), ε(.), p1(snow), p10(.), b(.) 5368.73 134.45 0 7
ψ(dist_6k+dec_1k), γ(dist_3k+dec_300m), ε(snow), p1(days+lat), p10(.), b(.) 5469.87 235.59 0 13
ψ(.), γ(.), ε(.), p1(days2), p10(.), b(.) 5496.81 262.53 0 7
ψ(dist_6k+dec_1k), γ(dist_300m), ε(.), p1(days+lat), p10(.), b(.) 5609.39 375.11 0 11

133
Table D.7. Long-tailed weasel (Neogale frenata) model ranking: All multi-season, false-positive occupancy models tested, ranked via AICc
(corrected for small sample size, n = 197).
Model formula AICc ΔAICc Wi k
ψ(dist_6k), γ(dist_300m+dec_300m), ε(dec_stand), p1(lat+dist_300m), p10(.), b(.) 1394.01 0.00 0.1011 12
ψ(dist_6k), γ(dist_300m+dec_300m), ε(.), p1(lat+dist_300m), p10(.), b(.) 1394.52 0.51 0.0783 11
ψ(dist_6k), γ(dist_6k), ε(.), p1(lat+dist_300m), p10(.), b(.) 1394.88 0.87 0.0654 10
ψ(dist_6k), γ(dist_300m), ε(.), p1(lat+dist_300m), p10(.), b(.) 1394.88 0.87 0.0654 10
ψ(dist_6k), γ(dist_1k), ε(.), p1(lat+dist_300m), p10(.), b(.) 1394.88 0.87 0.0654 10
ψ(dist_6k), γ(dist_3k), ε(.), p1(lat+dist_300m), p10(.), b(.) 1395.34 1.33 0.0520 10
ψ(dist_6k), γ(dist_300m+dec_300m), ε(dec_stand+lat), p1(lat+dist_300m), p10(.), b(.) 1395.50 1.49 0.0480 13
ψ(dist_6k), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1395.59 1.58 0.0459 9
ψ(dist_6k), γ(dist_300m+dec_300m), ε(lat), p1(lat+dist_300m), p10(.), b(.) 1395.89 1.88 0.0395 12
ψ(dist_6k), γ(dec_300m), ε(.), p1(lat+dist_300m), p10(.), b(.) 1395.93 1.92 0.0387 10
ψ(dist_6k), γ(dist_300m+dec_300m), ε(dist_3k), p1(lat+dist_300m), p10(.), b(.) 1396.23 2.22 0.0333 12
ψ(dist_6k), γ(dist_300m+dec_300m), ε(dist_300m), p1(lat+dist_300m), p10(.), b(.) 1396.67 2.66 0.0267 12
ψ(dist_6k), γ(dist_300m+dec_300m), ε(dec_300m), p1(lat+dist_300m), p10(.), b(.) 1396.69 2.68 0.0265 12
ψ(.), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1396.75 2.74 0.0257 8
ψ(dist_6k), γ(dist_300m+dec_300m), ε(snow), p1(lat+dist_300m), p10(.), b(.) 1396.75 2.74 0.0257 12
ψ(dist_6k), γ(dist_300m+dec_300m), ε(dec_1k), p1(lat+dist_300m), p10(.), b(.) 1396.77 2.76 0.0254 12
ψ(dist_6k), γ(dist_300m+dec_300m), ε(dist_6k), p1(lat+dist_300m), p10(.), b(.) 1396.78 2.77 0.0253 12
ψ(dist_6k), γ(dist_300m+dec_300m), ε(dist_1k), p1(lat+dist_300m), p10(.), b(.) 1396.78 2.77 0.0253 12
ψ(dist_6k+dec_stand), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1397.21 3.20 0.0204 10
ψ(dist_6k), γ(snow), ε(.), p1(lat+dist_300m), p10(.), b(.) 1397.70 3.69 0.0160 10
ψ(dist_6k), γ(dec_1k), ε(.), p1(lat+dist_300m), p10(.), b(.) 1397.75 3.74 0.0156 10
ψ(dist_6k), γ(lat), ε(.), p1(lat+dist_300m), p10(.), b(.) 1397.76 3.75 0.0155 10
ψ(dist_6k), γ(dec_stand), ε(.), p1(lat+dist_300m), p10(.), b(.) 1397.79 3.78 0.0153 10
ψ(dist_6k+lat), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1397.81 3.80 0.0151 10
ψ(dec_stand), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1398.00 3.99 0.0137 9
ψ(dist_3k), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1398.21 4.20 0.0124 9

134
 Table D.7 continued
ψ(dec_300m), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1398.69 4.68 0.0097 9
ψ(lat), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1398.87 4.86 0.0089 9
ψ(dist_300m), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1398.89 4.88 0.0088 9
ψ(dist_1k), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1398.92 4.91 0.0087 9
ψ(dec_1k), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1398.93 4.92 0.0086 9
ψ(snow), γ(.), ε(.), p1(lat+dist_300m), p10(.), b(.) 1398.94 4.93 0.0086 9
ψ(.), γ(.), ε(.), p1(lat), p10(.), b(.) 1399.80 5.79 0.0056 7
ψ(.), γ(.), ε(.), p1(dist_300m), p10(.), b(.) 1403.07 9.06 0.0011 7
ψ(.), γ(.), ε(.), p1(dec_stand), p10(.), b(.) 1404.25 10.24 0.0006 7
ψ(.), γ(.), ε(.), p1(dec_1k), p10(.), b(.) 1404.59 10.58 0.0005 7
ψ(.), γ(.), ε(.), p1(snow), p10(.), b(.) 1405.47 11.46 0.0003 7
ψ(.), γ(.), ε(.), p1(cwd_sticks), p10(.), b(.) 1406.34 12.33 0.0002 7
ψ(.), γ(.), ε(.), p1(dist_3k), p10(.), b(.) 1406.99 12.98 0.0002 7
ψ(.), γ(.), ε(.), p1(dbh), p10(.), b(.) 1407.04 13.03 0.0001 7
ψ(.), γ(.), ε(.), p1(dec_300m), p10(.), b(.) 1407.55 13.54 0.0001 7
ψ(.), γ(.), ε(.), p1(winter), p10(.), b(.) 1407.97 13.96 0.0001 7
ψ(.), γ(.), ε(.), p1(dist_1k), p10(.), b(.) 1408.69 14.68 0.0001 7
ψ(.), γ(.), ε(.), p1(.), p10(.), b(.) 1409.38 15.37 0 6
ψ(.), γ(.), ε(.), p1(cwd_logs), p10(.), b(.) 1411.08 17.07 0 7
ψ(.), γ(.), ε(.), p1(dist_6k), p10(.), b(.) 1411.17 17.16 0 7
ψ(.), γ(.), ε(.), p1(snags), p10(.), b(.) 1411.25 17.24 0 7
ψ(.), γ(.), ε(.), p1(days), p10(.), b(.) 15323.71 13929 0 7
ψ(.), γ(.), ε(.), p1(days2), p10(.), b(.) 15323.71 13929 0 7

135
Appendix D.6: Correlograms indicating no spatial-autocorrelation for American black bear models

136
 BIOGRAPHY

Bryn Evans was born in 1988 in the Santa Cruz mountains of California and raised, as her parents lovingly

joke, partly feral running barefoot in the woods with her younger brother as a constant companion.

Though slightly unorthodox, her education was grounded in curiosity for the natural world: from

homeschooling in the garden and taking community college science classes for fun, to “dropping out” of

high school at 14 (by passing the California High School Proficiency Exam) to attend Cabrillo College full

time, and finally obtaining her bachelor’s in Biological Sciences (Ecology) from University of California at

Davis magna cum laude in 2009.

After graduating, Bryn worked for UC Davis, the US Forest Service, and the California

Department of Fish and Wildlife on several wildlife research projects. Catching neonatal Black-tailed

deer, tailing scat-detection dogs in their search for fisher poop, and tracking radio-collared mountain

lions cemented her passion for field work. She completed her master’s degree in Wildlife Ecology at the

University of Wisconsin Madison, comparing aerial survey and camera trap data to assess the recovery

of semi-aquatic mammals in the St Louis River Area of Concern.

Bryn moved to Maine in January 2017 to join the Mortelliti Lab and has been collecting camera

trap images of the wildlife here ever since. She is a candidate for the Doctor of Philosophy degree in

Wildlife Ecology from the University of Maine in December 2021.

137