Type Your Frontispiece Or Quote Page Here (If Any)

Reproductive Activity of Olive-sided Flycatchers (Contopus cooperi) in Commercial

Forests of Central New Brunswick

Delaney R. Brooks

B.Env.St. (Hons.), University of Manitoba, 2011

A Thesis Submitted in Partial Fulfillment of the Requirements for the Degree of

Master of Science in Forestry

in the Graduate Academic Unit of Forestry and Environmental Management

Supervisor: Joseph J. Nocera, Ph.D., FOREM

Examining Board: Michelle Gray, Ph.D., FOREM, Chair Wendy Monk, Ph.D., Biology Graham Forbes, Ph.D., FOREM

This thesis is accepted by the Dean of Graduate Studies

THE UNIVERSITY OF NEW BRUNSWICK

October, 2020

©Delaney Brooks, 2021 ABSTRACT

Olive-sided flycatcher (OSFL; Contopus cooperi) is an avian aerial insectivore facing range-wide population declines. OSFL select breeding territories in open forest near waterbodies and are known to preferentially inhabit harvested forests. To determine if forest harvesting act as an ecological trap for OSFL in New Brunswick, I quantified and analyzed the reproductive activity of individuals in harvested and non-harvested sites. I used OSFL call patterns to determine if individuals were successfully breeding in each habitat. OSFL do select harvest sites over non-harvest sites for their breeding territory. Individuals are selecting breeding territories with older cutblocks within 4.9 and

45 ha and sites with a smaller proportion of the territory being clearcut within 19.6 ha.

Harvest sites in New Brunswick do not seem to act as an ecological trap for olive-sided flycatchers. My study indicates that forest harvesting in central New Brunswick does not adversely affect the reproductive activity of OSFL.

ACKNOWLEDGEMENTS

I would like to thank my supervisor Dr. Joe Nocera for his guidance and patience throughout the three years of this project. This project would not be possible without the assistance of Nicole Murtagh, Sarah Cusack, and Brooklyn Hillman both in the field and in the lab. Their patience, hard work, and eagerness to assist throughout the project was invaluable. I would also like to thank Dr. Tony Diamond and Dr. Stephen Heard for their roles on my committee. Thanks to John C. Brazner (Nova Scotia Department of Lands and Forestry) and Sean Haughian (Nova Scotia Museum) for their interest, encouragement, and educating me on other subjects. I am thankful for my support system for being with me through the three years. My partner, Matthew, for his unending patience and encouragement. My friends and lab mates in the Nocera lab, Forbes lab, and the CRI for the help with my project and for the adventures. A special thanks to Shaylyn

Wallace for her support in the lab and in the real world. And lastly to my family for their support, the time they spent to understand my work, for always listening when I needed to talk, the assistance with stats/R (Brendan), and the flow of papers/articles they send me

(Mackenzie).

iii

Table of Contents

ABSTRACT ...... ii ACKNOWLEDGEMENTS ...... iii Table of Contents ...... iv List of Tables ...... vii List of Figures ...... ix List of Symbols, Nomenclature or Abbreviations ...... xiv Chapter 1 General Introduction ...... 1 Bibliography ...... 8 Chapter 2 Using Autonomous Recording Units and Change-point Analysis to Determine Reproductive Activity in an Aerial Insectivore ...... 13 Abstract ...... 13

Keywords: Autonomous Recording Unit, Change-point analysis, Reproductive

activity, Vocalization rate ...... 14

Introduction ...... 14

Methods...... 16

Study Area ...... 16

Field Recordings ...... 17

Analysis of Field Recordings ...... 18

Statistical Analysis ...... 20

Results ...... 21

Discussion ...... 22

Tables and Figures ...... 26

Bibliography ...... 32 Chapter 3 Olive-sided Flycatcher (Contopus cooperi) Habitat Selection and the Role of Habitat Availability in Post-Harvest Habitat in New Brunswick ...... 38 Abstract ...... 38

Keywords: Clearcut, Habitat selection, Reproductive activity ...... 39

Introduction ...... 39

Materials and Methods ...... 42

Study Area ...... 42

Field Recordings ...... 43

Statistical Analysis ...... 44

Results ...... 45

Field Recordings ...... 46

Statistical Analysis ...... 46

Discussion ...... 47

Tables and Figures ...... 52

Bibliography ...... 56 Chapter 4 General Discussion ...... 63 Bibliography ...... 70 Appendix A Bumble bee (Bombus spp.) diversity differs between forested wetlands and clearcuts in the Acadian forest ...... 73 A.1. Abstract ...... 73

A.2. Introduction ...... 74

A.3. Materials and Methods ...... 75

A.4. Results ...... 79

A.5. Discussion ...... 81

A.6. Literature Cited ...... 84

Appendix B Autonomous Recording Unit Call Data for Olive-sided Flycatchers in Central New Brunswick ...... 99 Curriculum Vitae

List of Tables

Table 2.1 Change-point analysis results of successfully breeding olive-sided flycatchers

(Contopus cooperi). Numbers indicate the total number of days an individual spent in each breeding stage. Numbers in bold indicate breeding stages that are estimates due to lack of data. Q is the maximum number of segments (number of change-points + 1) that package ‘changepoint’ will search for when using the segment neighborhoods algorithm.

...... 26

Table 2.2 Change-point analysis results of unsuccessfully breeding olive-sided flycatchers (Contopus cooperi). Numbers indicate the total days an individual spent in each breeding stage. Numbers in bold indicate breeding stages that are estimates due to lack of data. Q is the maximum number of segments (number of change-points + 1) that package ‘changepoint’ will search for when using the segment neighborhoods algorithm.

...... 27

Table A.1 Characteristic plants of Wet Coniferous Forest 1 and Wet Coniferous Forest 2 as per the Forest Ecosystem Classification for Nova Scotia (Neily et al. 2011)...... 90

Table A.2 Bombus spp. found in forested wetlands and disturbed sites in 2018 and 2019.

...... 93

Table A.3 Parameters of a generalized linear mixed-model for abundance of Bombus species with habitat as a fixed factor. B. ternarius and B. terricola were found to have a positive relationship with forested wetlands...... 95

Table A.4 Plant species phenology (from USDA 2019) present in wet coniferous forest with a moderate to high herbaceous cover...... 96

vii

Table A.5 The Shannon index results of diversity, species richness, species evenness, and rarefaction scores for the four types of forested wetlands and harvest sites (clearcuts).

Harvest sites have the highest diversity and evenness (*)...... 97

viii

List of Figures

Figure 2.1 Location of the 22 olive-sided flycatcher territories used in change-point analysis. Of the 22 territories, four had olive-sided flycatchers breeding in 2018 and

2019...... 29

Figure 2.2 Change-point analysis and call rate of OSFL24 in 2019. The changes in call rate of OSFL24 throughout the breeding season indicate breeding success under the assumption that call rate is linking to breeding phenology. ARU recording displays arrival date to departure date (Q = 5). Red triangles indicates change-point locations. ... 30

Figure 2.3 Change-point analysis and call rate of OSFL06 in 2018. OSFL06 successfully produced a fledgling with only three detected changes in call rate. ARU recording displays nest-building stage to post-hatch (Q = 4). Red triangles indicate change-point location, black diamonds indicate dates removed due to playback during recording, and the blue square indicates the date a fledgling was documented...... 30

Figure 2.4 Change-point analysis and call rate of OSFL04 in 2018. The changes in call rate of OSFL04 throughout the breeding season does not indicate breeding success under the assumption that call rate is linked to breeding phenology. ARU recordings display arrival of OSFL to post-hatch stage (Q = 5). Red triangles indicate change-point locations...... 31

Figure 2.5 Change-point analysis and call rate of OSFL43 in 2019. The changes in call rate of OSFL43 throughout the breeding season does not indicate breeding success under the assumption that call rate is linked to breeding phenology. ARU recordings display incubation to post-hatch stage (Q = 3). Red triangles indicate change-point locations. ... 31

Figure 3.1 Map of the sites surveyed with ARUs and point counts for olive-sided flycatchers in 2018 and 2019 in central New Brunswick...... 53

Figure 3.2 Arrival dates of olive-sided flycatchers in non-harvested sites and sites harvested between 1998 and 2019 in New Brunswick, Canada. Olive-sided flycatchers

(Contopus cooperi) selected territories with cut blocks cut after 1998 over non-harvest sites and sites with cut blocks cut prior to 1998. The central line indicates median arrival date, the boxes indicate the 25th and 75th percentiles, the whiskers indicate 1.5× IQR, and the dots indicate outliers...... 54

Figure 3.3 Arrival dates of olive-sided flycatchers in non-harvested sites and sites harvested between 2008 and 2019 in New Brunswick, Canada. Olive-sided flycatchers

(Contopus cooperi) selected territories with cut blocks cut after 2007 over non-harvest sites and sites with cut blocks cut prior to 2008. The central line indicates median arrival date, the boxes indicate the 25th and 75th percentiles, the whiskers indicate 1.5× IQR, and the dots indicate outliers...... 55

Figure B.1 Change-point analysis and call rate of OSFL06 in 2018. ARU recording displays nest-building stage to post-hatch (Q = 4). Red triangles indicate change-point locations, black diamonds indicate dates removed due to playback during recording, and the blue square indicates the date a fledgling was documented...... 99

Figure B.2 Change-point analysis and call rate of OSFL09 in 2018. ARU recording displays incubation stage to post-hatch stage (Q = 4). Red triangles indicate change-point locations and black diamonds indicate dates removed due to playback during recording.

...... 99

Figure B.3 Change-point analysis and call rate of OSFL16 in 2019. ARU recording displays arrival of OSFL to post-hatch stage (Q = 5). Red triangles indicate change-point locations...... 100

Figure B.4 Change-point analysis and call rate of OSFL21 in 2019. ARU recording displays arrival date to departure date (Q = 5). Red triangles indicate change-point locations and black diamonds indicate dates removed due to playback during recording.

...... 100

Figure B.5 Change-point analysis and call rate of OSFL24 in 2019. ARU recording displays arrival date to departure date (Q = 5). Red triangles indicate change-point locations...... 101

Figure B.6 Change-point analysis and call rate of OSFL30 in 2019. ARU recording displays incubation stage to post-hatch stage (Q = 3). Red triangles indicate change-point locations...... 101

Figure B.7 Change-point analysis and call rate of OSFL02 in 2018. ARU recordings display arrival of OSFL to post-hatch stage (Q = 5). Red triangles indicate change-point locations...... 102

Figure B.8 Change-point analysis and call rate of OSFL04 in 2018. ARU recordings display arrival of OSFL to post-hatch stage (Q = 5). Red triangles indicate change-point locations...... 102

Figure B.9 Change-point analysis and call rate of OSFL11 in 2018. ARU recordings display arrival to departure of OSFL (Q = 5). Red triangles indicate change-point locations...... 103

Figure B.10 Change-point analysis and call rate of OSFL13 in 2018. ARU recordings display incubation stage to post-hatch stage (Q = 3). Red triangles indicate change-point locations...... 103

Figure B.11 Change-point analysis and call rate of OSFL17 in 2018. ARU recordings display nest building stage to post-hatch stage (Q = 4). Red triangles indicate change- point locations and black diamonds indicate dates removed due to playback during recording...... 104

Figure B.12 Change-point analysis and call rate of OSFL21 in 2018. ARU recordings display incubation stage to post-hatch stage (Q = 3). Red triangles indicate change-point locations and black diamonds indicate dates removed due to playback during recording.

...... 104

Figure B.13 Change-point analysis and call rate of OSFL28 in 2018. ARU recordings display arrival of OSFL to incubation stage (Q = 4). Red triangles indicate change-point locations...... 105

Figure B.14 Change-point analysis and call rate of OSFL01 in 2019. ARU recordings display arrival to departure of OSFL (Q = 5). Red triangles indicate change-point locations and black diamonds indicate dates removed due to playback during recording.

...... 105

Figure B.15 Change-point analysis and call rate of OSFL02 in 2019. ARU recordings display arrival to departure of OSFL (Q = 5). Red triangles indicate change-point locations...... 106

Figure B.16 Change-point analysis and call rate of OSFL07 in 2019. ARU recordings display arrival to departure of OSFL (Q = 5). Red triangles indicate change-point xii

locations and black diamonds indicate dates removed due to playback during recording.

...... 106

Figure B.17 Change-point analysis and call rate of OSFL16 in 2019. ARU recordings display arrival of OSFL to post-hatch stage (Q = 5). Red triangles indicate change-point locations...... 107

Figure B.18 Change-point analysis and call rate of OSFL17 in 2019. ARU recordings display arrival to departure of OSFL (Q = 5). Red triangles indicate change-point locations...... 107

Figure B.19 Change-point analysis and call rate of OSFL23 in 2019. ARU recordings display arrival to departure of OSFL (Q = 5). Red triangles indicate change-point locations and black diamonds indicate dates removed due to playback during recording.

...... 108

Figure B.20 Change-point analysis and call rate of OSFL31 in 2019. ARU recordings display nest building stage to departure of OSFL (Q = 4). Red triangles indicate change- point locations and black diamonds indicate dates removed due to playback during recording...... 108

Figure B.21 Change-point analysis and call rate of OSFL32 in 2019. ARU recordings display incubation stage to departure of OSFL (Q = 3). Red triangles indicate change- point locations and black diamonds indicate dates removed due to playback during recording...... 109

Figure B.22 Change-point analysis and call rate of OSFL43 in 2019. ARU recordings display incubation to post-hatch stage (Q = 3). Red triangles indicate change-point locations...... 109 xiii

List of Symbols, Nomenclature or Abbreviations

ARU – Autonomous Recording Unit

Clearcut – a logged area with all trees removed cm – centimeter

COSEWIC – Committee on the Status of Endangered Wildlife in Canada

Cutblock – an area of harvested land with defined boundaries dbh – Diameter at breast height

GLMM – Generalized Linear Mixed Model ha – hectares

Khz – Kilohertz m – meters ml – milliliter

OSFL – olive-sided flycatcher

PQL – Penalized Quasi-Likelihood

Q – maximum number of segments that the package ‘changepoint’ searches for when using the segment neighborhoods algorithm

Spp. – species

xiv

Chapter 1 General Introduction

Anthropogenic land use such as agriculture, forest harvesting, and urban development can result in drastic changes to the environment that cause severe declines in wildlife populations including invertebrates (Colla and Packer 2008), mammals

(Baisero et al. 2020), and birds (Bowler et al. 2019). In Canada, aerial insectivores are experiencing the most drastic declines in population among birds, primarily due to changes in land use and a generalized decrease in insect abundance (North American

Bird Conservation Initiative Canada 2019). As land changes increase and land use intensifies in the breeding and/or non-breeding grounds of bird species, continued declines in populations are to be expected (Carlson 2000).

Population declines resulting from land use change may affect a species’ use of habitat and the selection of territories. Habitat selection is evidenced by the disproportionate use of a habitat in relation to that habitat’s availability; this requires an individual to choose among known alternatives (Mayor et al. 2009). Ecological traps occur when individuals select low-quality habitat over available high-quality habitat resulting in reduced fitness (Donovan and Thompson 2001). Ecological traps most often occur when the cues used by an individual to assess and select habitat have been altered suddenly (Gates and Gysel 1978). This is frequently a result of anthropogenic land use

(Donovan and Thompson 2001). Land use, such as forest harvesting, may create an ecological trap for some species (Hollander et al. 2013), such as olive-sided flycatcher

(Contopus cooperi), which is well known for being susceptible to such traps (Robertson and Hutto 2007).

Study species: Olive-sided Flycatcher

Olive-sided flycatcher is an aerial insectivore with an expansive breeding range, reaching from the boreal forest of Alaska and Canada to the western United States

(Altman and Sallabanks 2020). Olive-sided flycatchers overwinter in South America, primarily in the Andean Range of Colombia (Altman and Sallabanks 2020). Individuals most often inhabit open forests near waterbodies during the breeding season

(Environment Canada 2016, Altman and Sallabanks 2020). Notably, olive-sided flycatchers require the presence of tall snags or trees from which they defend territory and hunt for prey (Environment Canada 2016, Altman and Sallabanks 2020). Olive- sided flycatchers are a post-disturbance species, often found in early post-burn sites

(Hutto 1995) and forest harvest sites (Robertson and Hutto 2007). Territory size generally ranges from 10-20 ha, but may be as large as 45 ha (Altman and Sallabanks

2020).

Olive-sided flycatchers capture prey by sallying, a technique where an individual leaves a perch to capture a prey item and then returns to its perch (Environment Canada

2016, Altman and Sallabanks 2020). Females will forage near the ground in flowering vegetation near the nest site but males have not been observed displaying this feeding behaviour (Hagelin et al. 2015). Olive-sided flycatchers primarily feed on flying species of Hymenoptera (e.g., wasps, bees) but also Diptera, Lepidoptera, Orthoptera, and

Odonata (Altman and Sallabanks 2020).

In the breeding season, males will arrive on the breeding territory first (14 May to 7 June in my study) and the females will arrive seven to nine days after the males 2

(Wright 1997). Incubation is performed solely by females with provisioning from males

(Altman and Sallabanks 2020). Olive-sided flycatchers typically nest once per year with a clutch of 3-4 eggs, but will re-nest if the first attempt is unsuccessful (Environment

Canada 2016, Altman and Sallabanks 2020).

In Canada, olive-sided flycatchers have been listed as Threatened under the federal Species at Risk Act (SARA) since 2010, but the species is under consideration for status change. The species is listed as Threatened in Newfoundland and Labrador,

New Brunswick, Nova Scotia, and Manitoba, Threatened or Vulnerable in Quebec, and as Special Concern in Ontario (Environment Canada 2016). They are not listed under any endangered-species legislation in British Columbia, Alberta, Saskatchewan, Prince

Edward Island, Yukon, Northwest Territories, and Nunavut (Environment Canada

2016). The Breeding Bird Survey has shown that olive-sided flycatchers are facing population declines in the Maritime provinces with a steep decline (3.4% annually, 79% from 1968 to 2006; Environment Canada 2016) in New Brunswick (Peckford 2015 in

Stewart et al. 2015). The greatest threats to olive-sided flycatchers include a reduced availability of their prey, the suppression of natural disturbances (fires), and habitat loss via forest harvesting, land conversion, energy and mining operations, and urban development (Environment Canada 2016).

In New Brunswick, forest harvesting is common throughout the province with several active commercial operations. New Brunswick’s land base is 86% forest (6.1 million ha), of which 3.2 million ha are on public land (Crown) and the remaining 2.9 million ha are privately owned (Province of New Brunswick 2020). Approximately

80,000 ha of public and private land are harvested annually (Natural Resource Canada 3

2020). Forest harvesting on public land is regulated by forest management agreements but that on privately owned land is not. However, all harvesting must be in accordance with the Clean Water Act, SNB 1989, c C-6.1 (Province of New Brunswick 2020). The

Clean Water Act, SNB 1989, c C-6.1 restricts forest harvesting within a 30 m buffer of waterbodies in New Brunswick (Clean Water Act 2019).

Detecting Olive-sided Flycatchers

Olive-sided flycatchers nests are often inaccessible to human observers, as they typically nest high in treetops or at the terminus of tall snags (Altman and Sallabanks

2020). As such, detection of olive-sided flycatchers is often limited by observing individuals off the nest and/or by using auditory means.

During the breeding season males have predictable patterns of call frequency that correspond with each breeding stage (territory selection, pairing, incubation, and post-hatch). The male call frequency declines when a female arrives on the territory, it then increases during incubation, and finally declines again during the post-hatch stage as the male assists in care of the hatchlings (Wright 1997).

As olive-sided flycatchers are suboscine passerines, they have simpler vocal organs than oscine species (Raikow and Bledsoe 2000) resulting in the inability to produce complex vocalizations commonly referred to as song. Olive-sided flycatchers have one three-note call but will often abbreviate it to a two-note call. The species will also vocalize with a “pip-pip-pip”. Males are the primary callers but both sexes pip and perform bill snaps when agitated (Wright 1997). As olive-sided flycatchers are

monomorphic, call behaviour can instead be used to assist in the identification of male and female in a territory.

Autonomous Recording Units (ARUs) are used in wildlife studies to monitor vocal species. The most common use of ARUs for land animals is to monitor presence/absence within an area (Aide et al. 2013; Sidie-Slettedahl et al. 2015; Pérez-

Granados et al. 2018). ARUs are often used alongside more traditional survey methods such as point counts (Pankratz et al. 2017). As the use of ARUs are becoming more commonplace in wildlife monitoring we begin to see other studies such as ones comparing the results of point counts and recordings (Van Wilgenburg et al. 2017).

My Study of Olive-sided Flycatchers in New Brunswick

Olive-sided flycatchers are in population decline throughout their breeding range

(Environment Canada 2016, Altman and Sallabanks 2020) and strongly so in New

Brunswick (Peckford 2015 in Stewart et al. 2015). Previous studies on this species have shown that post-harvest habitat can act as an ecological trap when compared to post-fire habitat (Robertson and Hutto 2007). In Montana, olive-sided flycatchers were shown to preferentially breed in selectively harvested forest resulting in a lowered nest success rate than those breeding in post-burn sites (Robertson and Hutto 2007). Given the commonness of post-harvest habitat in New Brunswick, it is a model area to study whether such activities are creating ecological traps. I hypothesized that changes to the breeding habitat of olive-sided flycatchers, such as tree removal from clearcutting, by commercial forest harvesting would adversely affect olive-sided flycatcher reproductive

activity. Forest harvesting may result in a reduction of suitable prey (Atlegrim and

Sjöberg 1995, Brooks and Nocera 2020) and increase nest predation (King et al. 2008).

My primary objective was to examine the degree to which habitat availability may represent an ecological trap for olive-sided flycatchers. In this study, I sought to determine if forest harvesting in the Acadian forest (a mix of rich tolerant hardwood and spruce-fir forest; Loo and Ives 2003) is affecting olive-sided flycatcher reproductive activity. To do so, I also sought to develop a method to monitor reproductive activity using ARUs and change-point analysis. Overall, the objectives of this study were to determine:

1. The habitat olive-sided flycatchers are preferentially settling in during the breeding season; 2. The habitat characteristics, in regard to forest harvesting, that olive-sided flycatchers are selecting; 3. If the presence of forest harvesting within an olive-sided flycatcher territory affected reproductive activity. To meet my objectives, I: (i) monitored the call rate of male olive-sided flycatchers using ARUs; (ii) used a change-point analysis to determine if individuals were successfully breeding in their territory; (iii) developed a method to passively monitor olive-sided flycatcher reproductive activity using ARUs; and (iv) tested if olive-sided flycatcher reproductive activity was dependent on habitat type. I predicted that:

1. Olive-sided flycatchers would preferentially settle in post-harvest habitat during the breeding season; 2. Olive-sided flycatchers would select post-harvest sites that had been recently clearcut as the species is a post-disturbance specialist; 3. The presence of forest harvesting would negatively affect olive-sided flycatcher reproductive activity. The remainder of this thesis is structed in the following manner:

In Chapter 2 I monitored olive-sided flycatcher daily call rate using ARUs to determine if I could develop a passive method to monitor reproductive activity. I then used change-point analyses to determine if an individual was breeding by comparing the changes to an expected pattern of call rate. This proved to be a successful method capable of allowing inference about reproductive activity.

In Chapter 3 I used the ARU recordings to first determine which habitats olive-sided flycatcher were selecting when arriving on the breeding grounds. I then determined what characteristics individuals were using to select habitat. Then I used the method from

Chapter 2 to determine if reproductive activity was significantly different in sites with post-harvest habitat to identify any ecological traps.

Chapter 4 is a general discussion that provides a synthesis of the two data chapters.

Chapter 4 is followed by Appendices A and B. Appendix A is a reprint of a published paper (Brooks and Nocera 2020) describing an attempt to determine prey availability in these habitats. However, the prey availability aspect of the study was later abandoned due to uncertainty of species composition of olive-sided flycatcher diet. Nonetheless, the effort yielded data on the availability of bumblebees, a known olive-sided flycatcher prey item (Altman and Sallabanks 2020). Appendix B contains the raw data of call rates of olive-sided flycatchers from Chapter 2.

Bibliography

Aide, T. M., C. Corrada-Bravo, M. Campos-Cerqueira, C. Milan, G. Vega, and R.

Alvarez. 2013. Real-time bioacoustics monitoring and automated species

identification. Peer J 1:e103. DOI: 10.7717/peerj.103.

Altman, B. and Sallabanks, R. 2020. Olive-sided flycatcher (Contopus cooperi), version

1.0. In Birds of the World (A. F. Poole, Editor). Cornell Lab of Ornithology,

Ithaca, NY, USA. https://doi.org/10.2173/bow.olsfly.01

Atlegri, O. and K. Sjöberg. 1995. Effects of clear-cutting and selective felling in

Swedish Boreal coniferous forest: Response of invertebrate taxa eaten by birds.

Entomologica Fennica 6(2-3): 79-90. DOI: 10.33338/ef.83843.

Baisero, D., P. Visconti, M. Pacifici, M. Cimatti, and C. Rondinini. 2020. Projected

global loss of mammal habitat due to land-use and climate change. One Earth 2:

578-585. DOI:10.1016/j.oneear.2020.05.015.

Bowler, D. E., H. Heldbjerg, A. D. Fox, M. de Jong, and K. Böhning-Gaese. 2019.

Long-term declines of European insectivorous bird populations and potential

causes. Conservation Biology 33: 1120-1130. DOI: 10.1111/cobi.13307.

Brooks, D. R. and J. J. Nocera. 2020. Bumble bee (Bombus spp.) diversity differs

between forested wetlands and clearcuts in the Acadian forest. Canadian Journal

of Forest Research, In Press. DOI: 10.1139/cjfr-2020-0094.

Brown, C. R., M. Bomberger Brown, and E. Danchin. 2000. Breeding habitat selection

in cliff swallows: The effect of conspecific reproductive success on colony

choice. Journal of Animal Ecology 69: 133-142.

Carlson, A. 2000. The effect of habitat loss on a deciduous forest specialist species: The 8

White-backed Woodpecker (Dendrocopos leucotos). Forest Ecology and

Management 131: 215-221. DOI: 10.1016/S0378-1127(99)00215-7.

Clean Water Act, SNB 1989, c C-6.1, http://canlii.ca/t/533rs> Retrieved on 2020-10-25.

Colla, S. R. and L. Packer. 2008. Evidence for decline in eastern North American

bumble bees (Hymenoptera: Apidae), with special focus on Bombus affinis

Cresson. Biodiversity and Conservation 17: 1379-1391. DOI: 10.1007/s10531-

008-9340-5.

Donovan. T. M., and F. R. Thompson III. 2001. Modeling the ecological trap

hypothesis: A habitat and demographic analysis for migrant songbirds.

Ecological Applications 11(3): 871-882.

Environment Canada. 2016. Recovery strategy for the Olive-sided Flycatcher (Contopus

cooperi) in Canada. Species at Risk Act Recovery Strategy Series. Environment

Canada, Ottawa. vii + 52 pp.

Gates, J. E., and L. W. Gysel. 1978. Avian nest dispersion and fledging success in field-

forest ecotones. Ecology 59(5): 871-883.

Hagelin, J. C., Busby, S., Harding-Scurr, A., and Brinkman, A. R. 2015. Observations

on fecal sac consumption and near-ground foraging behavior in the olive-sided

flycatcher (Contopus cooperi). Wilson Journal of Ornithology 127(2): 332-336.

DOI: 10.1676%2Fwils-127-02-332-336.1

Hollander, F. A., N. Titeux, and H. Van Dyck. 2013. Habitat-dependent prey availability

and offspring provisioning explain an ecological trap in a migratory bird.

Functional Ecology 27: 702-709. DOI: 10.1111/1365-2435.12077.

Hutto, R. L. 1995. Composition of bird communities following stand-replacement fires

in Northern Rocky Mountain (U.S.A.) conifer forests. Conservation Biology

9(5): 1041-1058.

King, D. I., R. M. DeGraaf, and C. R. Griffin. 2008 Edge-related nest predation in

clearcut and groupcut stands. Conservation Biology 12(6): 1412-1415. DOI:

10.1111/j.1523-1739.1998.97199.x.

Loo, J. and N. Ives. 2003. The Acadian forest: Historical condition and human impacts.

The Forestry Chrinicles 79(3): 462-474.

Mayor, S. J., D. C. Schneider, J. A. Schaefer, and S. P. Mahoney. 2009. Habitat

selection at multiple scales. Écoscience 16(2): 238-247. DOI: 10.2980/16-2-

3238.

Natural Resource Canada. 2020. Statistical Data: New Brunswick. Retrieved from:

https://cfs.nrcan.gc.ca/statsprofile/overview/NB. Accessed: October 25, 2020.

North American Bird Conservation Initiative Canada. 2019. The State of Canada’s

Birds, 2019. Environment and Climate Change Canada, Ottawa, Canada. 12

pages. www.stateofcanadasbirds.org.

Pankratz, R. F., S. Hache, P. Sólymos, and E. M. Bayne. 2017. Potential benefits of

augmenting road-based breeding bird surveys with autonomous recordings.

Avian Conservation and Ecology 12(2):18. DOI: 10.5751/ACE-01087-120218.

Peckford, M. 2015. Olive-sided Flycatcher, pp. 308-309 in Stewart, R. L. M., K. A.

Bredin, A. R. Couturier, A. G. Horn, D. Lepage, S. Makepeace, P. D. Taylor, M.

-A. Villard, and R. M. Whittam (eds). 2015. Second Atlas of Breeding Birds of

the Maritime Provinces. Bird Studies Canada, Environment Canada, Natural 10

History Society of Prince Edward Island, Nature New Brunswick, New

Brunswick Department of Natural Resources, Nova Scotia Bird Society, Nova

Scotia Department of Natural Resources, and Prince Edward Island Department

of Agriculture and Forestry, Sackville, 528 + 28 pp.

Pérez-Granados, C., Bota, G., Giralt, D., and Traba, J. 2018. A cost-effective protocol

for monitoring birds using autonomous recording units: a case study with a

night-time singing passerine. Bird Study 65(3): 338-345. DOI:

10.1080/00063657.2018.1511682.

Province of New Brunswick. Forest governance in the Province of New Brunswick.

Available from

https://www.sfmcanada.org/images/Publications/EN/New_Brunswick_info_Prov

inces_and_territories_EN.pdf [accessed 16 August 2020].

Raikow, R. J. and A. H. Bledsoe. 2000. Phylogeny and evolution of the passerine birds.

BioScience 50(6): 487-499.

Robertson, B. A., and R. L. Hutto. 2007. Is selectively harvested forest an ecological

trap for Olive-sided Flycatchers? Condor 109(1): 109-121.

Saher, D. J. and F. K. A. Schmiegelow. 2005. Movement pathways and habitat selection

by woodland caribou during spring migration. Rangifer, Special Issue 16: 143-

154.

Sidie-Slettedahl, A. M., Jensen, K. C., Johnson, R. R., Arnold, T. W., Austin, J. E., and

Stafford, J. D. 2015. Evaluation of autonomous recording units for detecting 3

species of secretive marsh birds. Wildlife Society Bulletin 39(3):626-634. DOI:

https://doi.org/10.1002/wsb.569. 11

Tellaría, J. L. and T. Santos. 1995. Effects of forest fragmentation on a guild of

wintering passerines: The role of habitat selection. Biological Conservation 71:

61-67.

Van Wilgenburg, S. L., P. Sólymos, K. J. Kardynal, and M. D. Frey. 2017. Paired

sampling standardizes point count data from humans and acoustic recorders.

Avian Conservation and Ecology 12(1):13. DOI:10.5751/ACE-00975-120113

Wright, J.M. 1997. Preliminary study of olive-sided flycatchers, June 1994-April 1997.

Alaska Department of Fish and Game. Final research report. Endangered species

conservation fund federal aid studies SE-3-3, 4 and 5. Juneau, Alaska. 34 pp.

Chapter 2 Using Autonomous Recording Units and Change-point

Analysis to Determine Reproductive Activity in an Aerial Insectivore

Abstract

Autonomous Recording Units (ARUs) are commonly used, alongside traditional methods, to study the presence/absence of vocal species. ARUs are useful in monitoring during adverse weather, and for species that are secretive or vocalize at night. I endeavored to develop a method to monitor avian reproductive activity using ARUs.

Olive-sided flycatchers (Contopus cooperi), an at-risk passerine, are an ideal species with which to develop this method as they have a loud distinct call and have large territories without much intrusion from conspecifics. Olive-sided flycatchers have a distinct call pattern during the breeding season. I used a change-point analysis to determine the dates of significant changes in their call pattern to determine if individuals were successfully breeding. I monitored 22 olive-sided flycatchers in central New

Brunswick in 2018 and 2019. I found that using a combination of ARUs and change- point analyses was a viable method for studying reproductive activity of olive-sided flycatchers. I found that 27% of olive-sided flycatchers were successfully breeding which, when considering erroneous classifications, is within the range of nest success

(30-65%) documented elsewhere. My method shows promise for studying other bird species as well as other vocal non-avian species.

Keywords: Autonomous Recording Unit, Change-point analysis, Reproductive activity,

Vocalization rate

Introduction

Wildlife studies are often limited by financial cost, evasiveness of study species, and human effort. The use of Autonomous Recording Units (ARUs) in wildlife research allows for greater sampling effort without a proportional increase in human effort.

However, ARUs can only be used to detect species that can vocalize or produce distinct sounds. As such, ARUs are becoming more commonplace in avian monitoring

(Shonfield and Bayne 2017) and have allowed for enhanced nocturnal studies

(Frommolt and Tauchert 2014), studies during adverse weather conditions, and better study of secretive or evasive species (Sidie-Slettedahl et al. 2015).

Generally, ARUs are included as a component of monitoring programs that also use traditional methods such as point counts or transect surveys (Sidie-Slettedahl et al.

2015). The combination of methods can improve monitoring efficacy, e.g., conducting transects to determine ideal deployment locations for ARUs (Pérez-Granados et al.

2018). Often studies are focused on species of conservation concern, such as the

Eurasian bittern (Botaurus stellaris), Savi’s warbler (Locustella luscinioides; Bardeli et al. 2010), and hihi (Notiomystis cincta) in New Zealand (Metcalf et al. 2019). Some studies focus on evaluating the performance and efficacy of ARUs and automated signal recognition programs (Colbert et al. 2015). However, new uses for ARUs in wildlife monitoring can be developed to complement traditional methods.

Although the use of ARUs has become commonplace in wildlife monitoring

(e.g., occupancy, population estimates), their use in hypothesis-driven wildlife research remains limited. I sought to determine whether the use of ARUs could be extended beyond simply determining the presence/absence of a species to also estimating its behavioural phenology within its territory. In this study, I focus on the possible use of

ARUs to study the breeding activity of a vocal avian species, the olive-sided flycatcher

(Contopus cooperi). To my knowledge only one study has previously attempted to use call rates to detect breeding activity of birds (Upham-Mills et al. 2020). In the latter study, five-minute in-person song counts were used with three modeling methods

(multinomial, logistic regression, hierarchical) to predict breeding status (Upham-Mills et al. 2020). I used ARUs, instead which allows for an increase in time analyzed for call rates without an increase in effort at the site.

I sought to develop a simple, reproducible, and non-invasive method of tracking reproductive activity of passerines using ARUs. Reproductive activity can be tracked this way if call or song rate is correlated with breeding stage, as has been seen in

American redstarts (Setophaga ruticilla; Staicer et al. 2006) and olive-sided flycatchers

(Wright 1997). I chose to use ARUs rather than point counts (based on short periods of human listening) as they allow for monitoring of each individual for long periods. I tracked the daily calling patterns of olive-sided flycatcher as the focal species of this study for three reasons: 1) the species’ status on Schedule 1 of the Species at Risk Act

(SARA) in Canada is currently threatened (under consideration of status change), 2) their call is loud and distinct with little variation, and 3) the species’ use of large territory without much overlap with conspecifics’ territories (Altman and Sallabanks 15

2020). The call rate of olive-sided flycatchers changes with the four breeding stages

(territory selection, pairing, incubation, and post-hatch) in a predictable manner (Wright

1997). Male call rate declines as females arrive on the territory, increases during incubation, and then declines again during the post-hatch stage (Wright 1997). In

Alaska, prior to female arrival, average male call rate is 11.46 calls per 3-minutes, then decreases to 4.05 call per 3-minutes during pairing/nest building, increases to 4.18 calls per 3-minutes, then decreases to 0.09 calls per 3-minutes (Wright 1997).

In regards to olive-sided flycatcher breeding phenology, we know that pairing of males and females may last up to two weeks, nest building 5-7 days, the incubation period 14-19 days, and the nestling period 15-23 days (Altman and Sallabanks 2020).

Fledglings may rely on their parents for an additional 7 days after leaving the nest and will remain near the nest for up to 19 days post-fledging (Altman and Sallabanks 2020).

I use a statistical procedure (change-point analysis) to identify changes in call rates and breeding phases. If ARUs can quantify call rates over time and change-point analysis can be successfully applied to identify changes in those rates, then this will be the first combination of these two techniques to quantify behavioural phenology.

Methods

Study Area

I collected data on male olive-sided flycatcher (Contopus cooperi) call rate in three central New Brunswick (Canada) counties: Kent (46° 26´N 65° 16´W),

Northumberland (46° 54´N 65° 59´W), and Sunbury (45° 57N 66° 21´W; Figure 2.1). I 16

only collected passive breeding data (call rates) and did not actively collect breeding data such as fecal sac carrying, feeding, and fledgling presence. Central New Brunswick is primarily Acadian forest, a diverse habitat of rich tolerant hardwoods and spruce-fir forests (Loo and Ives 2003). The Acadian forest is characterized by the presence of red spruce (Picea rubens), yellow birch (Betula alleghaniensis), sugar maple (Acer saccharum), and balsam fir (Abies balsamea; Loo and Ives 2003). This area is actively harvested by multiple forestry companies. The region also commonly features forested wetlands including bogs, fens, freshwater marshes, shrub wetlands, aquatic beds, and wet coniferous forests.

Field Recordings

I collected audio recordings using Autonomous Recording Units (ARUs;

Wildlife Acoustics, Inc. ® Song Meter SM2, Song Meter SM4). I deployed 20 ARUs in both 2018 and 2019. The ARUs were deployed in seemingly suitable habitat that olive- sided flycatchers might inhabit prior to their arrival for breeding (9 May to 16 May

2018, 10 May to 12 May 2019). I selected sites using data from eBird (eBird 2017), the

Maritime Breeding Bird Atlas (Stewart et al. 2015), and olive-sided flycatcher habitat maps (GeoNB 2020). ARUs began recording the day after deployment. If an olive-sided flycatcher selected a territory at the edge of the ARU detection range (~250-500m), then

I would reposition the ARU to be near a frequent perch site of the territorial individual. I determined the range of detection by noting the distance of the olive-sided flycatcher to the ARU and then logging the time of calls. I then reviewed the acoustic recordings to determine if the ARU detected the logged calls. If an ARU had to be moved within a

site, I combined the data from the two locations if they were within 100 m of each other.

If I did not detect an olive-sided flycatcher near an ARU, I relocated it to a territory with an olive-sided flycatcher present. In 2018, I relocated 11 ARUs in 2018 and six in 2019.

Individuals were recorded throughout the breeding season (mid-May to early-August), from arrival onto territories to the departure for fall migration. The ARUs recorded daily from 0530-0830h with a sample frequency of 24 Khz. Overall, I catalogued acoustic recordings in 54 olive-sided flycatcher territories: 28 in 2018 and 26 in 2019. In 2018, I recorded an average of 22.4 days per olive-sided flycatcher with the longest number of recordings being 59 days (2 birds). In 2019, I recorded an average of 28.6 days per olive-sided flycatcher with the longest number of recordings being 75 days (1 bird).

Analysis of Field Recordings

Analyses were completed in R (version 3.5.2 R Core Team 2018) using the package ‘monitoR’ (Hafner and Katz 2018) to develop templates (recognizers) and to analyze the field recordings. The use of an automated signal recognition program such as monitoR is more efficient than human listening to recordings when considering data sets as large as mine (Knight et al. 2017). I chose to use monitoR over other available automated signal recognition programs (e.g., SongScope, convolutional neural network) as it is relatively user friendly, reports relatively few false positives, and has a high recall at a low score threshold (similar to human listening; Knight et al. 2017). I also chose to use monitoR because of its flexibility in call recognition; the olive-sided flycatcher call consists of 3 notes which can be represented mnemonically as “quick, three-beer”, however individuals can often be heard calling the 2 note abbreviation

“three-beer”. I created three templates with the rectangle method and using a clip from

Macaulay Library (The Cornell Lab of Ornithology): one 3 note call, two 2 note calls.

To improve the performance of each template I set the score cutoff of each to 0.3 which increased the number of detections that were true positives without increasing the number of false positives. The score cutoff represents the threshold at which matches are found to be acceptable (detections) or unacceptable (Katz et al. 2016a). I used three templates as this allowed me to capture the greatest amount of true positive detections without increasing false positive detections. I initially created two templates for each call type but excluded a 3 note template as it detected <10% true positives. True positives are events that occur above or equal to the score cutoff with false negatives scoring below the cutoff (Katz et al. 2016b). I tested the efficacy of the templates using an audio recording with known olive-sided flycatcher calls.

I analyzed the field recordings using the spectrogram cross correlation method

(see Katz et al. 2016b). Each recording was analyzed using the three templates for the three hours or for a two hour sub-segment (0530-0730h). I initially analyzed the three hours but determined olive-sided flycatcher call rate was relatively low from 0730-

0830h. I produced correlation scores using a Pearson correlation for template-recording matching. The maximum score or peak was then identified, resulting in a detection. I determined if each detection was a true positive or a false positive by plotting the spectrogram and listening to the audio clip.

Statistical Analysis

I tested for changes in male olive-sided flycatcher daily call rate using a change- point analysis with the R package ‘changepoint’ (Killick and Eckley 2014). I used the segment neighborhoods algorithm as it has a higher accuracy than binary segmentation

(Killick and Eckley 2014). I analyzed the changes in mean and variance using a normal distribution with a Bayesian Information Criterion (BIC; Schwarz 1978) penalty. The change-point analysis was set to detect up to four changes (Q = 5), representing the four breeding stages from territory selection to pairing/nest building to incubation to the post hatch stage (Wright 1997). Individuals that do not successfully breed should have fewer than four changes in call rate as males that are not paired display a consistent call rate and those that have a failed nest will resume calling at a high, consistent rate (Wright

1997). If I did not have data from arrival to departure for an individual, then the change- point analysis was set to detect fewer changes to reflect the lack of data (e.g., Q = 3). I ran the change-point analysis for individuals with a minimum of 25 days of recordings to ensure detection of pairing/nest building, incubation, and post-hatch stage. I did not include days prior to arrival or after departure for the change-point analysis. I removed days from the change-point analysis if: 1) playback was used (for mist netting purposes) during the recording at that site, and; 2) when the recording was not the full recording time (two or three hours). Following Wright (1997), olive-sided flycatchers that have a successful breeding season should have significant changes in their call rate that aligns with the above predictable pattern. Changes in call rate that do not align with the predictable pattern likely did not successfully breed (e.g., incubation period longer than

3 weeks). Patterns of changes in call rate of olive-sided flycatchers were previously quantified for breeding behaviour of olive-sided flycatchers in Alaska (Wright 1997).

Results

In total, 2600 days were recorded resulting in 672,387 detections. Most often, false positives were detections of alder flycatchers (Empidonax alnorum) and American red squirrel (Tamiasciurus hudsonicus). This misidentification is due to alder flycatcher calls being visually similar to olive-sided flycatcher calls in a spectrogram and

American red squirrel vocalizations being exceptionally noisy. Of these detections,

196,796 (25%) were true positives.

I conducted a change-point analysis for 22 olive-sided flycatchers (ten in 2018,

12 in 2019) to determine reproductive activity (Figure 2.1). Of the 22 olive-sided flycatchers, I had data for arrival date to departure date for seven individuals (one in

2018, six in 2019). Of the olive-sided flycatchers analyzed, two were confirmed to have fledglings indicating nest success (one per year). Overall, I analyzed 477 days of the ten olive-sided flycatchers in 2018 and 890 days of the 12 olive-sided flycatchers in 2019.

I found six olive-sided flycatchers successfully breeding using the change-point analysis method. Of these birds, I had acoustic recordings from arrival to departure date for two (Figure 2.2), nest-building to post-hatch for one, incubation to post-hatch for two, and arrival to post-hatch for one. For all individuals, their call rate reflected the expected breeding phenology (Table 2.1). For two individuals, the change-point analysis only detected 3 changes (OSFL06 [Figure 2.3] and OSFL21). However, both birds were known to have a fledgling (sighted 20 July 2018 and 31 July 2019). One individual

(OSFL16) erroneously appears to have a relatively short post-hatch stage due to ARU malfunction.

I found the remaining 16 olive-sided flycatchers were likely unsuccessful at breeding (vocalization data for these and all birds in the study are in Appendix B). Of these, I had recordings of arrival to departure date for six individuals (one in 2018

[Figure 2.4], five in 2019). Of the remaining ten olive-sided flycatchers, I had acoustic recordings from arrival to post-hatch for three, incubation to post-hatch for three (Figure

2.5), and one for the following: arrival to incubation, incubation to departure, nest building to post-hatch, and nest building to departure. I determined olive-sided flycatchers had failed breeding attempts by using the length of their incubation (2-10 days being too short, 21-38 days being too long) and the length of the post-hatch stage

(2-11 days; Table 2.2).

Discussion

This study demonstrates that ARU data collection combined with change-point analyses is a feasible method to detect changes in call rate which this study assumes reflects breeding phenology; I detected variation in olive-sided flycatcher call rate at all sites. Individuals that followed the predicted pattern outlined by Wright (1997) were treated as breeding successfully. I had anticipated the call rate of individuals that were successfully breeding would decline as females arrive and nest building begins, to increase during incubation, and then decrease during post-hatch as males attend to the young (Wright 1997). Two individuals did not directly follow the predicted call rate

(OSFL06, OSFL21) but were known to have fledglings. The call rates of non-breeding

individuals also did not follow the predicted pattern. This suggests that some individuals that I determined as non-breeders could be breeding but do not have a call pattern that reflects breeding behaviour. It is also possible the relationship between call rate and breeding phenology of olive-sided flycatchers is weaker than previously believed. To remedy this issue, it is important to document other behaviour that may indicate breeding such as nest defense (Altman and Sallabanks 2020), nest presence, feeding of nestlings, fecal sac removal or consumption (Hagelin et al. 2015), and the presence of fledglings. I passively monitored territories for the presence of fledglings resulting in the sighting of two fledglings. More active monitoring of breeding behaviour would provide greater support for the ARU and change-point analysis method.

The flycatchers from which I had a full set of call rate recordings (arrival date to departure date) were the easiest to determine breeding activity. These individuals were the easiest to analyze as I could determine each change starting from arrival on the breeding territory instead of estimating changes without a full recording of call behaviour. In my study, 27% of olive-sided flycatchers are believed to successfully breed, which is similar to their nest success rate of 30-65% in other jurisdictions

(COSEWIC 2018). The breeding success is within the known nest success rate when considering erroneously classified individuals. It is important to note that breeding status of most birds in this study was determined inferentially as I made the assumption that call rate indicates breeding behaviour; very few olive-sided flycatchers were confirmed to be breeding by physical evidence.

Combining ARUs and change-point analysis is a straightforward process to determine the breeding activity of some bird species. The method I have described here 23

uses large sets of ARU recordings to ensure call rate from individuals were recorded. By analyzing two hours of ARU recordings daily for each individual, I could accurately track changes in call rate which may be due to changes in breeding activity. The use of change-points then ensures detection of significant changes in these call rates. The packages ‘monitoR’ and ‘changepoint’ are freeware, which improves accessibility.

Although change-point analysis is frequently used in ecology (Beckage et al.

2007, Thomson et al. 2010) and animal behaviour studies (Mustonen et al. 2012,

Gurarie et al. 2016), this study is the first, to my knowledge, that is used to detect changes in call rates of avian species. This study demonstrates the ability to monitor reproduction of avian species at risk, such as olive-sided flycatchers, without the need to use invasive methods such as removing blood samples to monitor prolactin levels

(Sintich et al. 1995). The change-point analysis method may also be used to study changes in other behaviours reflective in avian vocalizations such as mobbing behaviour in the presence of different predators (Griesser 2009).

I found that the largest source of false positives was the presence of alder flycatchers (Empidonax alnorum) in olive-sided flycatcher territories as the species’ calls are similar on a spectrogram. The analysis of field recordings could be improved in further studies by creating multiple templates from a variety of audio clips. This may improve the detection analysis and decrease the number of false positives detected by encompassing variation in calls. I recommend that the species’ call behaviour is considered when deciding the number of hours analyzed. Olive-sided flycatchers call at the highest rate around sunrise but can be detected throughout the morning (Wright

1997). However, for species such as Bicknell’s thrush (Catharus bicknelli), that are 24

known to actively call at dawn and after dusk, one-hour recordings at both dawn and after dusk would be recommended (Townsend et al. 2020).

Although this method of acoustic analysis was developed to study the reproductive activity of a bird species, it is possible it could be used for a variety of species and behaviours. For instance, ARUs and other autonomous acoustic monitoring systems have been used to study frog species (Willacy et al. 2015; Taylor et al. 2017).

However, adding the change-point method could enhance the study of frog breeding phenology as well as determining baseline calling information for understudied frog species (Willacy et al. 2015). ARUs have also been used to study a variety of whale species including minke (Balaenoptera acutorostata; Nikolich and Towers 2020) and humpback whales (Megaptera novaeangliae; Smith et al. 2008). Male humpback whales are known to sing while escorting females with the length of the singing dependent on the presence of other males. Thus, the change-point method could be used to determine male behaviour when escorting female-calf groups using the length of songs (Smith et al. 2008). In general, the change-point method can be used for any species that vocalizes in ways that reflect their behaviour.

Tables and Figures

Table 2.1 Change-point analysis results of successfully breeding olive-sided flycatchers (Contopus cooperi). Numbers indicate the total number of days an individual spent in each breeding stage.

Numbers in bold indicate breeding stages that are estimates due to lack of data. Q is the maximum number of segments (number of change-points + 1) that package ‘changepoint’ will search for when using the segment neighborhoods algorithm.

Site Year Recording Q Territory Nest Incubation Post- Post-

Time Selection Building hatch fledge

06 2018 Nest 4 Unknown 6+ 2 51 2+ Building to P-H 09 2018 Incubation 4 17+ 5 19 4+ Unknown to P-H 16 2019 Arrival to 5 14 6 21 8 Unknown P-H 21 2019 Arrival to 5 2 48 13 14 Departure 24 2019 Arrival to 5 12 3 13 38 11 Departure 30 2019 Incubation 3 Unknown 9(+/-) 9 29 3+ to P-H

Site Year Recording Q Territory Nest Incubation Post- Post-

Time Selection Building hatch fledge

02 2018 Arrival to 5 11 24 2 10 16 P-H 04 2018 Arrival to 5 20 3 38 3 7 P-H 11 2018 Arrival to 5 4 21 26 2 Departure 13 2018 Incubation 3 29 4 2 to P-H 17 2018 Nest 4 3 29 4 12 Building to P-H 21 2018 Incubation 3 18 5 9 to P-H 28 2018 Arrival to 4 22 6 7 2 Incubation 01 2019 Arrival to 5 2 40 4 31 2 Departure 02 2019 Arrival to 5 16 2 24 6 31 Departure 07 2019 Arrival to 5 15 3 30 5 21 Departure 16 2018 Arrival to 5 26 18 10 5 7 P-H 17 2019 Arrival to 5 7 8 32 10 10 Departure 23 2019 Arrival to 5 16 3 5 6 35 Departure

31 2019 Nest 4 3 23 4 33 Building to Departure 32 2019 Incubation 3 7 3 43 to Departure 43 2019 Incubation 3 21 11 8 to P-H

Figure 2.1 Location of the 22 olive-sided flycatcher territories used in change-point analysis. Of the

22 territories, four had olive-sided flycatchers breeding in 2018 and 2019.

Figure 2.2 Change-point analysis and call rate of OSFL24 in 2019. The changes in call rate of

OSFL24 throughout the breeding season indicate breeding success under the assumption that call rate is linking to breeding phenology. ARU recording displays arrival date to departure date (Q =

5). Red triangles indicates change-point locations.

Figure 2.4 Change-point analysis and call rate of OSFL04 in 2018. The changes in call rate of

OSFL04 throughout the breeding season does not indicate breeding success under the assumption that call rate is linked to breeding phenology. ARU recordings display arrival of OSFL to post- hatch stage (Q = 5). Red triangles indicate change-point locations.

Figure 2.5 Change-point analysis and call rate of OSFL43 in 2019. The changes in call rate of

OSFL43 throughout the breeding season does not indicate breeding success under the assumption that call rate is linked to breeding phenology. ARU recordings display incubation to post-hatch stage (Q = 3). Red triangles indicate change-point locations.

Bibliography

Altman, B. and Sallabanks, R. 2020. Olive-sided flycatcher (Contopus cooperi), version

1.0. In Birds of the World (A. F. Poole, Editor). Cornell Lab of Ornithology,

Ithaca, NY, USA. https://doi.org/10.2173/bow.olsfly.01

Bardeli, R., Wolff, D., Kurth, F., Koch, M., Tauchert, K. -H., and Frommolt, K. -H.

2010. Detecting bird sounds in a complex acoustic environment and application

to bioacoustics monitoring. Pattern Recognition Letters 31: 1524-1534. DOI:

10.1016/j.patrec.2009.09.014.

Beckage, B., Joseph, L., Belisle, P., Wolfson, D. B., and Platt, W. J. 2007. Bayesian

change-point analyses in ecology. New Phytologist 174: 456-467.

Colbert, D. S., Ruttinger, J. A., Streich, M., Chamberlain, M., Conner, L. M., and

Warren, R. J. 2015. Application of autonomous recording units to monitor

gobbling activity by wild turkey. Wildlife Society Bulletin 39(4): 757-763. DOI:

10.1002/wsb.577.

COSEWIC. 2018. COSEWIC assessment and status report on the Olive-sided

Flycatcher Contopus cooperi in Canada. Committee on the Status of Endangered

Wildlife in Canada. Ottawa. ix + 52 pp. eBird. 2017. eBird: An online database of bird distribution and abundance. eBird,

Cornell Lab of Ornithology, Ithaca, New York. Available: http://www.ebird.org

(Accessed: October 25, 2020).

Frommolt, K. -H. and Tauchert, K. -H. 2014. Applying bioacoustics methods for long-

term monitoring of a nocturnal wetland bird. Ecological Informatics 21: 4-12.

DOI: http://dx.doi.org/10.1016/j.ecoinf.2013.12.009. 32

Griesser, M. 2009. Mobbing calls signal predator category in a kin group-living bird

species. Proceedings of the Royal Society B 276: 2887-2892. DOI:

10.1098/rspb.2009.0551.

Gurarie, E., Bracis, C., Delgado, M., Meckley, T. D., Kojola, I., and Wagner, C. M.

2016. What is the animal doing? Tools for exploring behavioural structure in

animal movements. Journal of Animal Ecology 85: 69-84. DOI: 10.1111/1365-

2656.12379.

Hafner, S. and Katz, J. 2018. monitoR: Acoustic template detection in R. R package

version 1.0.7. URL:

http://www.uvm.edu/rsenr/vtcfwru/R/?Page=monitoR/monitoR.htm

Hagelin, J. C., Busby, S., Harding-Scurr, A., and Brinkman, A. R. 2015. Observations

on fecal sac consumption and near-ground foraging behavior in the olive-sided

flycatcher (Contopus cooperi). Wilson Journal of Ornithology 127(2): 332-336.

DOI: 10.1676%2Fwils-127-02-332-336.1

Katz, J., Hafner, S. D., and Donovan, T. 2016a. Tools for automated acoustic

monitoring within the R package monitoR. Bioacoustics 25(2): 197-210. DOI:

10.1080/09524622.2016.1138415.

Katz, J., Hafner, S. D., and Donovan, T. 2016b. Assessment of error rates in acoustic

monitoring with the R package monitoR. Bioacoustics 25(2): 177-196. DOI:

10.1080/09524622.2015.1133320.

Killick, R. and Eckley, I. A. 2014. changepoint: An R package for changepoint analysis.

Journal of Statistical Software 58(3): 1-19. URL:

http://www.jstatsoft.org/v58/i03/. 33

Knight, E. C., Hannah, K. C., Foley, G., Scott, C., Brigham, R.M., and Bayne, E. 2017.

Recommendations for acoustic recognizer performance assessment with

application to five common automated signal recognition programs. Avian

Conservation and Ecology 12(2):14. DOI: https://doi.org/10.5751/ACE-01114-

120214.

Loo, J. and N. Ives. 2003. The Acadian forest: Historical condition and human impacts.

The Forestry Chrinicles 79(3): 462-474.

Metcalf, O. C., Ewen, J. G., McCready, M., Williams, W. M., and Rowcliffe, J. M.

2019. A novel method for using ecoacoustics to monitor post-translocation

behaviour in an endangered passerine. Methods in Ecology and Evolution 10:

626-636. DOI: 10.1111/2041-210X.13147.

Mustonen, A.-M., Lempiäinen, T., Aspelund, M., Hellstedt, P., Ikonen, K., Itämies, J.,

Vähä, V., Erkinaro, J., Asikainen, J., Kunnasranta, M., Niemelä, P., Aho, J., and

Nieminen, P. 2012. Application of change-point analysis to determine winter

sleep patterns of the raccoon dog (Nyctereutes procyonoides) from body

temperature recordings and a multi-faceted dietary and behavioral study of

wintering. BMC Ecology 12: 27. DOI: 10.1186/1472-6785-12-27.

Nikolich, K. and Towers, J. R. 2020. Vocalizations of common minke whales

(Balaenoptera acutorostrata) in an eastern North Pacific feeding ground.

Bioacoustics 29(1): 97-108. DOI: 10.1080/09524622.2018.1555716.

Pérez-Granados, C., Bota, G., Giralt, D., and Traba, J. 2018. A cost-effective protocol

for monitoring birds using autonomous recording units: a case study with a

night-time singing passerine. Bird Study 65(3): 338-345. DOI:

10.1080/00063657.2018.1511682.

R Core Team. 2018. R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. Available from

https://www.R-project.org/.

Schwarz, G. 1978. Estimating the dimension of a model. Annals of Statistics 6(2): 461-

464.

Shonfield, J. and Bayne, E. M. 2017. Autonomous recording units in avian ecological

research: Current use and future applications. Avian Conservation & Ecology

12(1):14. DOI: https://doi.org/10.5751/ACE-00974-120114.

Sidie-Slettedahl, A. M., Jensen, K. C., Johnson, R. R., Arnold, T. W., Austin, J. E., and

Stafford, J. D. 2015. Evaluation of autonomous recording units for detecting 3

species of secretive marsh birds. Wildlife Society Bulletin 39(3):626-634. DOI:

https://doi.org/10.1002/wsb.569.

Sintich, S. M., Keagle, M. K., Harper, R. G., and Welsford, I. G. 1995. Use of ELISA

for determination of plasma prolactin levels in the House Wren. Condor 97(4):

1057-1061.

Smith, J. N., Goldizen, A. W., Dunlop, R. A., and Noad, M. J. 2008. Songs of male

humpback whales, Megaptera novaeangliae, are involved in intersexual

interactions. Animal Behaviour 76: 467-477. DOI:

10.1016/j.anbehav.2008.02.013

Staicer, C. A., Ingalls, V., and Sherry, T. A. 2006. Singing behavior varies with

breeding status of American redstarts (Setophaga ruticilla). Wilson Journal of

Ornithology 118(4): 439-451.

Stewart, R. L. M., K. A. Bredin, A. R. Couturier, A. G. Horn, D. Lepage, S. Makepeace,

P. D. Taylor, M. -A. Villard, and R. M. Whittam (eds). 2015. Second Atlas of

Breeding Birds of the Maritime Provinces. Bird Studies Canada, Environment

Canada, Natural History Society of Prince Edward Island, Nature New

Brunswick, New Brunswick Department of Natural Resources, Nova Scotia Bird

Society, Nova Scotia Department of Natural Resources, and Prince Edward

Island Department of Agriculture and Forestry, Sackville, 528 + 28 pp.

Taylor, A., McCallum, H. I., Watson, G., and Grigg, G. C. 2017. Impact of cane toads

on a community of Australian native frogs, determined by 10 years of automated

identification and logging of call behaviour. Journal of Applied Ecology 54(6):

2000-2010. DOI: 10.1111/1365-2664.12859.

Thomson, J. R., Kimmerer, W. J., Brown, L. R., Newman, K. B., Mac Nally R.,

Bennett, W. A., Feyrer, F., and Fleishman, E. 2010. Bayesian change point

analysis of abundance trends for pelagic fishes in the upper San Francisco

Estuary. Ecological Applications 20(5): 1431-1448.

Townsend, J. M., McFarland, K. P., Rimmer, C. C., Ellison, W. G., and Goetz, J. E.

2020. Bicknell’s Thrrush (Catharus bicknelli), version 1.0. In Birds of the World

(P. G. Rodewald, Editor). Cornell Lab of Ornithology, Ithaca, NY, USA. DOI:

10.2173/bow.bicthr.01.

Upham-Mills, E. J., Reimer, J. R., Haché, S., Lele, S. R., and Bayne, E. M. 2020. Can

singing rate be used to predict male breeding status of forest songbirds? A

comparison of three calibration models. Ecosphere 11(1): e03005. DOI:

10.1002/ecs2.3005.

Willacy, R. J., Mahoney, M., and Newell, D. A. 2015. If a frog calls in the forest:

Bioacoustic monitoring reveals the breeding phenology of the endangered

Richmond Range mountain frog (Philoria richmondensis). Austral Ecology

45(6): 625-633. DOI: 10.1111/aec.12228.

Wright, J.M. 1997. Preliminary study of olive-sided flycatchers, June 1994-April 1997.

Alaska Department of Fish and Game. Final research report. Endangered species

conservation fund federal aid studies SE-3-3, 4 and 5. Juneau, Alaska. 34 pp.

Chapter 3 Olive-sided Flycatcher (Contopus cooperi) Habitat Selection

and the Role of Habitat Availability in Post-Harvest Habitat in New

Brunswick

Abstract

Ecological traps occur when a low-quality habitat is selected over available high-quality habitat, resulting in reduced fitness. This phenomenon typically results in a population sink and can occur when sudden changes to the environment disrupt cues individuals use that would otherwise allow them to reliably assess suitable habitat. Olive-sided flycatchers (Contopus cooperi) are an aerial insectivore with a history of occupying ecological traps. They breed and forage preferentially in open forest. Within these sites, tall snags or trees are important for foraging and defending territory. Harvested sites can provide adequate substrate for perching and nesting, although such sites are often ecological traps as they lack other requirements to maintain positive fitness. New

Brunswick is an ideal location to determine if ecological traps occur for Olive-sided

Flycatcher as harvested sites and natural patches are abundant and interspersed. I examine the degree to which harvested habitats act as an ecological trap for Olive-sided

Flycatchers in New Brunswick using Autonomous Recording Units (ARUs). I compare the selection of breeding territories at 64 sites in 2018 and 2019. I compared the reproductive activity of 21 individuals using ARUs and a change-point statistical analysis. Of the 64 territories, 61 individuals had territories in clearcuts within a 45 ha landscape scale. I found that Olive-sided Flycatcher selected for territories with older 38

cutblocks and smaller proportions of clearcuts within their territory. Ultimately, I found that olive-sided flycatchers selected clearcuts but did not experience reduced fitness in these territories.

Keywords: Clearcut, Habitat selection, Reproductive activity

Introduction

Habitat that is attractive to an animal while selecting territories, but is not sufficient to maintain fitness, may act as an ecological trap for a species. An ecological trap occurs when a low-quality habitat is selected over available high-quality habitat, resulting in reduced fitness (Donovan and Thompson 2001). This may occur when sudden changes to the environment disrupt cues individuals have evolved to use to assess suitable habitat resulting in a population sink (Gates and Gysel 1978, Robertson and Hutto 2006). There are three conditions in which an ecological trap arises

(Robertson and Hutto 2006): (1) habitat suitability remains the same but a change in cues (intensity, type, or number) cause the habitat to appear more attractive, (2) habitat suitability has decreased but original cues remain unchanged, and (3) habitat suitability has decreased and a change in cues has caused the habitat to appear more attractive. An ecological trap may also arise if a novel element of the habitat mimics a cue often used by an individual to assess habitat (e.g., light pollution resulting in inland migration in hatchling sea turtles; Schlaepfer at al. 2002). Further, there are three criteria a habitat must meet to be an ecological trap: (1) an individual must select one habitat over another, (2) there is a difference in fitness among habitats, and (3) the fitness outcome is lower in the selected habitat over other available habitats (Robertson and Hutto 2006). 39

Landscape alteration by human activity may result in numerous changes to wildlife behaviour (Lent and Capen 1995), such as inducing changes in foraging behaviour (Lent and Capen 1995), site abandonment (Bayne et al. 2008), mating success

(Habib et al. 2007), and territory selection (van den Berg et al. 2001). Land use may also affect the quality of habitat in which individuals hold territory (Suarez et al. 1997).

Forest harvesting is a common landscape alteration that has a history of inadvertently creating ecological traps (Powell et al. 2010). These traps can range in spatial scale and the number of species they affect. For example, at the landscape scale, forest harvest mosaics can present ecological traps to numerous migrating bird species (Sánchez-

Clavijo et al. 2016). At a smaller (stand) scale, silviculture practices like plantations can present ecological traps as seen for Lesser Spotted Woodpeckers (Dendrocopos minor) in poplar (Populus spp.) plantations in the Tordera River basin in Catalonia (Camprodon et al. 2015). Even forestry practices at the site scale can induce ecological traps, such as when harvested timber is piled too close to extant forest and attracts the endangered beetle Rosalia longicorn, whose larvae will never have time to develop in these piles before the wood is collected for market (Adamski et al. 2016, 2018).

Olive-sided flycatcher (Contopus cooperi) is a species with a history of occupying ecological traps (Robertson and Hutto 2007). The species is in the aerial- insectivore guild, feeding by sallying near the canopy (Altman and Sallabanks 2020).

Aerial insectivores are experiencing population declines greater than passerines of different guilds (Böhning-Gaese et al. 1993, Nebel et al. 2010, Fraser et al. 2012,

English et al. 2018, North American Bird Conservation Initiative Canada 2019). The behavioural tendency to occupy ecological traps may be partly responsible for the 40

population decline the species has been experiencing over the past 60 years

(Environment Canada 2016). Globally, olive-sided flycatchers have been assessed as

Near Threatened by the International Union for Conservation of Nature (IUCN) since

2004 (BirdLife International 2012). Throughout North America, the species has demonstrated a population decline of approximately 76% since the mid-1960s (Butcher and Niven 2007). In Canada, the species’ population has declined throughout the breeding range, and the species is classified as federally Threatened (currently under consideration of status change; Environment Canada 2016). In southern Canada, the

Breeding Bird Survey (BBS) estimates a 3.4% annual rate of decline for olive-sided flycatcher populations (Environment Canada 2016).

Olive-sided flycatchers breed and forage preferentially in open forests, often near waterbodies (Altman and Sallabanks 2020, Environment Canada 2016). Within these sites, tall snags or trees are important for foraging and defending territory (Altman and Sallabanks 2020, Environment Canada 2016). Availability of nest trees may also be an important cue for habitat selection (Robertson and Hutto 2007). Olive-sided flycatchers are a post-disturbance species and are often found in territories shortly after a burn or cut has occurred (Robertson and Hutto 2007, Yahner 2008). Harvested sites may provide adequate substrate for perching and nesting, although such sites have often been shown to be ecological traps as they lack other requirements to maintain fitness

(Robertson and Hutto 2007). Diet quality may be lower in harvested sites, resulting in lowered nest success (Duguay et al. 2000). However, much of what we know about ecological traps for olive-sided flycatchers is from studies in montane coniferous forests of Montana, USA. The conditions that make for ecological traps likely vary among 41

forest types, and there has been no assessment of whether harvest practices in the

Acadian forest (a mixedwood forest consisting of spruce-fir forest and rich tolerant hardwoods; Loo and Ives 2003) creates ecological traps for olive-sided flycatchers. Such an assessment is needed to better manage sensitive forested wetlands in the Acadian forest.

In this study, I sought to determine if harvested sites in the Acadian forest act as ecological traps for olive-sided flycatchers. I compared the reproductive activity of individuals in forested wetlands (e.g., bog, freshwater marsh, wet coniferous forest) to those in harvested sites at three spatial scales of hypothetical territory sizes. I used autonomous recording units (ARU) to determine which habitats were selected and to monitor reproductive activity (see Chapter 2). Male olive-sided flycatchers demonstrate variation in call rate with the four stages of breeding: territory selection, pairing, incubation, and post-hatch (Wright 1997). The aural detectability of males declines when females arrive on their territory but increases during incubation. Call rate declines again during the post-hatch stage of breeding (Wright 1997). Thus, call rate of males can be used as an indicator of reproductive activity.

Materials and Methods

Study Area

In 2018 and 2019, I collected audio and presence data of olive-sided flycatchers in three counties of New Brunswick: Kent (46° 26´N 65° 16´W), Northumberland (46°

54´N 65° 59´W), and Sunbury (45° 57N 66° 21´W; Figure 3.1). The Acadian forest ecosystem in the center of New Brunswick is dominated by forested wetlands. Such 42

habitats include bogs, fens, freshwater marshes, shrub wetlands, aquatic beds, and wet coniferous forest all of which I considered forested wetlands. Several forestry companies have active harvest operations in the area. I selected candidate deployment sites by assessing data from the Maritime Breeding Bird Atlas (Stewart et al. 2015), eBird (eBird 2017), and olive-sided flycatcher habitat maps (GeoNB 2020). I deployed

ARUs (Wildlife Acoustics, Inc. ® Song Meter SM2, Song Meter SM4) prior to olive- sided flycatcher arrival in the breeding territory (9 to 16 May 2018, 10 to 12 May 2019).

Field Recordings

I collected acoustic data of olive-sided flycatchers using ARUs from 10 May to

16 August 2018, and 11 May to 15 August 2019. I used ARUs to monitor occupancy and to document male call patterns. I conducted at least one 15-minute point count in habitats that I suspected had olive-sided flycatcher presence but did not have deployed

ARUs. ARUs were deployed in areas in which olive-sided flycatchers were expected to select territories (based on data from eBird [eBird 2017], the Maritime Breeding Bird

Atlas [Stewart et a. 2015], and habitat data from GeoNB [GeoNB 2020]) allowing a determination of the timing of settlement and departure of individuals; habitats that are settled first should possess more or better cues used to select territories. Likewise, territories that have individuals departing later could indicate higher quality habitats. To determine if individual olive-sided flycatchers were exhibiting reproductive activity, I used the method outlined in chapter 2 (this thesis).

Statistical Analysis

Analyses were completed in R (version 3.5.2 R Core Team 2018). For all analyses, I set α = 0.05. I conducted two Welch’s two sample t-tests comparing arrival date to determine if olive-sided flycatchers were selecting harvest sites or non-harvest sites (e.g., fen, bog, freshwater marsh). In the first t-test, any site that had been harvested between 1998 and 2019 was classified as a harvest site. In the second t-test, only sites that had been harvested between 2008 and 2019 were classified as harvest sites (sites harvested 1998-2007 were considered non-harvest, as opposed to the previous t-test). I did not consider scale of territory size for the Welch’s two sample t-test as for all individuals the habitat was the same. I did not conduct any analysis to determine if olive-sided flycatchers departed from harvest or non-harvest sites earlier as I had fewer individuals with recorded departure dates and all but one was in harvest sites.

To determine habitat selection by olive-sided flycatchers, I built three Poisson generalized linear mixed models using a penalized quasi-likelihood estimator

(glmmPQL) using the package ‘MASS’ (Venables and Ripley 2002). I used ‘fitdistrplus’ to determine that the errors in the data were Poisson distributed (Delignette-Muller and

Dutang 2015). The first glmmPQL was built considering a landscape scale of 4.9 ha, the second considered a landscape scale of 19.6 ha, and the third considered a landscape scale of 45 ha. These scales were selected as they represent the minimum, mean, and maximum documented size (respectively) of territories (Bock and Lynch 1970, Altman and Sallabanks 2020). I did not include sites that were harvested in 2018-2019 (three sites) as I did not have access to accurate cut block size. I used habitat data from the geographic data layers (forest and wetland) from GeoNB (2020) and harvest data (2009- 44

2016) from the Department of Natural Resource and Energy Development of New

Brunswick. I only included cutblocks that were clearcuts and did not include cutblocks whose year of cutting was not recorded. I set olive-sided flycatcher presence as the response variable, whereas year, habitat type (harvest, non-harvest, harvest/non-harvest mix), proportion of territory that is cut block, proportion of territory that is cut block harvested from 2008 onwards, and cut block age were set as fixed effects, and site as a non-nested random effect. I used a reverse stepwise model selection procedure where parameter estimates were evaluated with the loss of each term until an optimized and final model was identified. A candidate variable would be removed from the model if p

> 0.05 and the t < 2. I then followed the same procedure described above to model whether olive-sided flycatcher reproduction (successful/unsuccessful) is dependent on habitat type.

Results

Of 64 total sites (Figure 3.1) in the study, 62 were confirmed to have olive-sided flycatchers present; 54 recorded via ARUs and eight recorded during 15-minute point counts. 38 of the sites were recorded in 2018 and 26 in 2019. The two sites recorded without olive-sided flycatchers were in 2018. Point counts that did not detect olive-sided flycatchers were not included as individuals may be present but undetected. Of the 64 sites, 56 (32 in 2018, 24 in 2019) contained clearcuts within 4.9 ha, 58 (34 in 2018, 24 in 2019) contained clearcuts within 19.6 ha, and 61 (36 in 2018, 25 in 2019) contained clearcuts within 45 ha. Three sites (one in 2018, two in 2019) were harvested in 2018-

2019 and were thus excluded in the habitat selection and olive-sided flycatcher reproduction analyses.

Field Recordings

I collected acoustic data for 54 olive-sided flycatchers (28 in 2018, 26 in 2019).

The ARUs detected arrival date for 31 birds (14 in 2018, 17 in 2019). Flycatchers arrived from 14 May to 02 June 2018 and 23 May to 07 June 2019. Most arrived on 28

May 2018 (3) and 25 May 2019 (5). The average date of arrival in 2018 was 25 May and the average date of arrival in 2019 was 27 May. I had sufficient acoustic data to analyze the reproductive activity of 22 olive-sided flycatchers (10 in 2018, 12 in 2019).

However, I only analyzed reproductive activity of 21 olive-sided flycatchers (10 in

2018, 11 in 2019) as I did not have cut block size data for one individual.

Statistical Analysis

Arrival date of olive-sided flycatchers at sites harvested between 1998 and 2019 is earlier compared to the arrival date of individuals in non-harvest sites (28 harvest sites, 4 non-harvest sites; t = 170.2, df = 31.3, p = 2.2×10-16; Figure 3.2). Arrival date in harvested sites is also earlier when comparing sites harvested between 2008 and 2019 to non-harvested and pre-2008 harvest sites (23 harvest sites, 9 non-harvest sites; t = 169.7, df = 31.6, p = 2.2×10-16; Figure 3.3).

At a smaller scale (4.9 ha), the age of the most recent cut block within the territory is statistically significant for habitat selection (t = 2.9, df = 59, p = 0.005; Table

3.1). I found cut block age of the most recent cut had a positive relationship with olive- sided flycatcher habitat selection. At the 19.6 ha landscape scale the proportion of the 46

territory that was harvested is significant (t = -2.1, df = 59, p = 0.040; Table 3.2). I found that cut block size has a negative relationship with habitat selection. At a large scale (45 ha), the age of the most recent cut block within the territory was also significant for habitat selection (t = 2.7, df = 59, p = 0.009; Table 3.3). Similar to the 4.9 ha scale, the cut block age of the most recent cut has a positive relationship with olive- sided flycatcher habitat selection.

I found olive-sided flycatcher breeding activity (in terms of male call rate) is not dependent on habitat type in any of the scales of territory in this study (4.9 ha, t = -0.6, df = 19, p = 0.524; 19.6 ha, t = -0.7, df = 19, p = 0.502; and 45 ha, t = -0.1, df = 19, p =

0.932). Only one flycatcher exclusively used harvest sites in their breeding territory. The remaining 20 flycatchers inhabited territories with a mix of harvest and non-harvest habitat.

Discussion

Olive-sided flycatchers in the Acadian forest appear to select harvested sites to establish territories during the breeding season despite a small sample size. The arrival dates of olive-sided flycatchers in harvested sites was earlier than the arrival dates in both non-harvest sites and non-harvest sites with clearcuts harvested prior to 2008.

Within both the 4.9 ha and 45 ha scales, olive-sided flycatchers are selecting territories with older clearcuts. At the19.6 ha scale, olive-sided flycatchers are selecting territories with a smaller proportion of the site being clearcut over territories with large proportions of clearcuts.

The first condition of an ecological trap – an individual must select one habitat over another (Robertson and Hutto 2006) – is demonstrated by olive-sided flycatchers in the Acadian forest. The second condition (individual fitness must differ among habitats;

Robertson and Hutto 2006) and the third condition (the fitness must be lower in individuals selecting the preferred habitat over other available habitats; Robertson and

Hutto 2006) are not demonstrated here when using reproductive activity as an indicator of fitness. I found no evidence that harvested sites (clearcuts) in the Acadian forest act as ecological traps for olive-sided flycatchers. However, I only considered clearcut logging in this study. In Montana, olive-sided flycatchers preferentially settle in selectively harvested forests with these individuals having significantly lower nest success than in less preferred habitat (post-burns; Robertson and Hutto 2007). Other harvest treatments present in the Acadian forest, such as salvage cut, seed tree cut, partial cut, commercial thinning, patch cuts, or strip cut may demonstrate results similar to those in Montana and warrant further study.

Olive-sided flycatcher tends to occupy post-disturbance areas with a preference for early post-fire conditions (Hutto 1995) and harvest sites including partial cut, patch cut, seed tree cut, and clearcut (Hutto and Young 1999; Spies et al. 2007). I found when olive-sided flycatchers are selecting territories for the breeding season, the age of the harvest sites was important in small (4.9 ha) and large (45 ha) territory scales with individuals selecting for older clearcuts. In the Northern Rocky Mountains, olive-sided flycatchers were more often detected in old clearcuts than in more recent clearcuts with the assumption that old clearcuts are more structurally similar to a young forest (most trees <20 cm dbh) in this region than a recent clearcut (Hutto and Young 1999). Olive- 48

sided flycatchers may be avoiding more recent clearcuts as 2 to 5 year old cuts have fewer snags, an important element of their habitat, than post-burn sites (Harper et al.

2004).

Olive-sided flycatchers select territories with a smaller proportion of clearcut in the Acadian forest at a landscape scale of 19.6 ha. Within the 19.6 ha territory, the largest clearcut size I observed is 19.0 ha and the smallest is 0.2 ha. Smaller proportions of clearcuts within a territory may be selected as food availability may be lower in clearcuts. Although olive-sided flycatcher prey is not well documented, we know a large portion of their breeding season diet is composed of flying insects in the order

Hymenoptera (Altman and Sallabanks 2020). There are also reports of olive-sided flycatchers consuming insects in the orders Diptera, Lepidoptera, Orthoptera, and

Odonata (Altman and Sallabanks 2020). Lepidoptera abundance is significantly lower in clearcuts than in selectively harvested or non-harvested sites (Summerville and Crist

2002). I have shown elsewhere that clearcuts in the Acadian forest have a lower abundance of Bombus spp. (“bumblebees”; Hymenoptera) which could be an important food source for olive-sided flycatchers in this area (Brooks and Nocera 2020; reprinted in Appendix 1). Territories with a larger proportion of uncut habitat may also be selected as they may have lower presence of known olive-sided flycatcher predators. For example, Canada jays (Perisoreus canadensis) are a known nest predator of olive-sided flycatchers and were observed in my study sites (Wright 1997). Canada jays have large territory sizes (average 69 ha) which they actively defend (Strickland and Ouellet 2020).

A larger proportion of clearcuts within a territory may result in an increase in

interactions with Canada jays as the species often associates with forest edges (Ibarzabal and Desrochers 2004).

In the Acadian forest, olive-sided flycatchers are selecting for and potentially becoming more reliant on the present of clearcuts. To better understand if clearcuts are adequate habitat for olive-sided flycatchers in the breeding season we must first better understand their diet and predation pressures in the Acadian forest. Although I did not determine that clearcuts are an ecological trap for olive-sided flycatchers in this study I only assessed after-effects. Species monitoring both before and after a disturbance is critical to understand temporal changes of fitness as they relate to ecological traps (Hale and Swearer 2016). One site in this study was actively harvested between study years.

However, I was unable to include this site in the reproduction analysis due to lack of information on the cutblock size. This is also made difficult as New Brunswick has some of the most intensively managed forests in Canada with most major changes to the forest resulting from forest management activities such as clearcutting (Etheridge et al.

2005). Thus, we may be noting a trend similar to chimney swifts (Chaetura pelagica), a species that once nested and roosted in tree cavities but now almost exclusively nests in man-made chimneys (Graves 2004). It is possible olive-sided flycatchers are now nesting almost exclusively in post-clearcut sites due to an increase in clearcuts over time in New Brunswick. Further studies, at a larger spatial (e.g., all New Brunswick) and temporal scale (e.g., lifespan of olive-sided flycatcher, multigeneration study) should be considered before stating that forest harvesting does not cause an ecological trap.

Olive-sided flycatchers are facing range-wide population declines with an estimated yearly decline of 3.4% (Environment Canada 2016). In the Acadian forest, I 50

found olive-sided flycatchers have an estimated nest success of 27% (see Chapter 2). In this study I considered at least one fledgling as a successful nest as I was monitoring breeding behaviour using call rate. Typically, olive-sided flycatchers have a clutch of three eggs and will not renest after a successful clutch (Altman and Sallabanks 2020).

Pairing success was shown to be 76% in harvested sites in Montana (Robertson and

Hutto 2007). If olive-sided flycatcher pairing success is similar to Montana and breeding pairs in the Acadian forest are only producing a single fledgling each year, the species’ population will continue to decline in this area. There are currently few management strategies for olive-sided flycatchers due to a clear lack of data (Altman and Sallabanks

2020). A better understanding of olive-sided flycatcher threats is necessary to develop an effective management plan for New Brunswick.

Tables and Figures

Table 3.1 Parameters of generalized linear mixed-model for olive-sided flycatcher presence with cut block age within a 4.9 ha territory as a fixed effect.

Bird Presence ~ Cut Block Age, ~ 1|Site Value Std. Error DF t-value p-value (Intercept) -0.003 0.024 59 -1.3 0.204 Cut Block Age 0.003 0.001 59 2.9 0.005

Table 3.2 Parameters of generalized linear mixed-model for olive-sided flycatcher presence with proportion of 19.6 ha territory clearcut as a fixed effect.

Bird Presence ~ Cut Proportion, ~ 1|Site Value Std. Error DF t-value p-value (Intercept) 0.072 0.028 59 2.6 0.013 Cut Proportion -0.001 0.001 59 -2.1 0.040

Table 3.3 Parameters of generalized linear mixed-model for olive-sided flycatcher presence with cut block age within a 45 ha territory as a fixed effect.

Bird Presence ~ Cut Block Age, ~ 1|Site Value Std. Error DF t-value p-value (Intercept) -0.027 0.024 59 -1.1 0.258 Cut Block Age 0.005 0.002 59 2.7 0.009

Figure 3.1 Map of the sites surveyed with ARUs and point counts for olive-sided flycatchers in 2018 and 2019 in central New Brunswick.

Figure 3.2 Arrival dates of olive-sided flycatchers in non-harvested sites and sites harvested between 1998 and 2019 in New Brunswick, Canada. Olive-sided flycatchers (Contopus cooperi) selected territories with cut blocks cut after 1998 over non-harvest sites and sites with cut blocks cut prior to 1998. The central line indicates median arrival date, the boxes indicate the 25th and 75th percentiles, the whiskers indicate 1.5× IQR, and the dots indicate outliers.

Figure 3.3 Arrival dates of olive-sided flycatchers in non-harvested sites and sites harvested between 2008 and 2019 in New Brunswick, Canada. Olive-sided flycatchers (Contopus cooperi) selected territories with cut blocks cut after 2007 over non-harvest sites and sites with cut blocks cut prior to 2008. The central line indicates median arrival date, the boxes indicate the 25th and 75th percentiles, the whiskers indicate 1.5× IQR, and the dots indicate outliers.

Bibliography

Adamski, P., A. Bohdan, J. Michalcewicz, M. Ciach, and Z. Witkowski. 2016. Timber

stacks: potential ecological traps for an endangered saproxylic beetle, the Rosalia

longicorn Rosalia alpina. Journal of Insect Conservation 20: 1099-1105.

Adamski, P., J. Michalcewicz, A. Bohdan, and M. Ciach. 2018. Potential range of

impact of an ecological trap network: the case of timber stacks and the Rosalia

longicorn. Journal of Insect Conservation 22: 209-219.

https://doi.org/10.1007/s10841-018-0054-z.

Altman, B. and R. Sallabanks. 2020. Olive-sided flycatcher (Contopus cooperi), version

1.0. In Birds of the World (A. F. Poole, Editor). Cornell Lab of Ornithology,

Ithaca, NY, USA. DOI: 10.2173/bow.olsfly.01.

Bayne, E. M., L. Habib, and S. Boutin. 2008. Impacts of chronic anthropogenic noise

from energy-sector activity on abundance of songbirds in the boreal forest.

Conservation Biology 22(5): 1186-1193.

BirdLife International. 2012. Contopus cooperi. The IUCN Red List of Threatened

Species 2012:

e.T22699787A38074065. https://dx.doi.org/10.2305/IUCN.UK.2012-

1.RLTS.T22699787A38074065.en. Accessed 17 July 2020.

Bock, C. E. and Lynch, J. F. 1970. Breeding bird populations of burned and unburned

conifer forest in the Sierra Nevada. Condor 72(2): 182-189.

Böhning-Gaese, K., M. L. Taper, and J. H. Brown. 1993. Are declines in North

American insectivorous songbirds due to causes on the breeding range?

Conservation Biology 7(1): 76-86. 56

Brooks, D. R. and J. J. Nocera. 2020. Bumble bee (Bombus spp.) diversity differs

between forested wetlands and clearcuts in the Acadian forest. Canadian Journal

of Forest Research, In Press. DOI: 10.1139/cjfr-2020-0094.

Butcher, G. S. and D. K. Niven. 2007. Combining data from the Christmas bird count

and the breeding bird survey to determine the continental status and trends of

North American birds. New York, USA: National Audubon Society.

Camprodon, J., J. Faus, P. Salvanyà, J. Soler-Zurita, and J. L. Romero. 2015. Suitability

of poplar plantations for a cavity-nesting specialist, the Lesser Spotted

Woodpecker Dendrocopos minor, in the Mediterranean mosaic landscape. Acta

Ornithologica 50(2): 157-169.

Delignette-Muller, M. and Dutang, C. 2015. fitdistrplus: An R package for fitting

distributions. Journal of Statistical Software 64(4): 1-34. Available from

http://www.jstatsoft.org/v64/i04/.

Donovan. T. M., and F. R. Thompson III. 2001. Modeling the ecological trap

hypothesis: A habitat and demographic analysis for migrant songbirds.

Ecological Applications 11(3): 871-882.

Duguay, J. P., P. B. Wood, and G. W. Miller. 2000. Effects of timber harvests on

invertebrate biomass and avian nest success. Wildlife Society Bulletin 28(4):

1123-1131. eBird. 2017. eBird: An online database of bird distribution and abundance. eBird,

Cornell Lab of Ornithology, Ithaca, New York. Available: http://www.ebird.org

(Accessed: October 25, 2020).

English, P. A., D. J. Green, and J. J. Nocera. 2018. Stable isotopes from museum

specimens may provide evidence of long-term change in the trophic ecology of a

migratory aerial insectivore. Frontiers in Ecology and Evolution 6: 14.

Environment Canada. 2016. Recovery strategy for the Olive-sided Flycatcher (Contopus

cooperi) in Canada. Species at Risk Act Recovery Strategy Series. Environment

Canada, Ottawa. vii + 52 pp.

Etheridge, D. A., D. A. MacLean, R. G. Wagner, and J. S. Wilson. 2005. Changes in

landscape composition and stand structure from 1945-2002 on industrial forest in

New Brunswick, Canada. Canadian Journal of Forest Research 35: 1965-1977.

DOI: 10.1139/X05-110.

Fraser, K. C. B. J. M. Stutchbury, C. Silverio, P. M. Kramer, J. Barrow, D. Newstead,

N. Mickle, B. F. Cousens, J. C. Lee, D. M. Morrison, T. Shaheen, P. Mammenga,

K. Applegate, and J. Tautin. 2012. Continent-wide tracking to determine

migratory connectivity and tropical habitat associations of a declining aerial

insectivore. Proceedings of the Royal Society B 279(1749): 4901-4906.

Gates, J. E., and L. W. Gysel. 1978. Avian nest dispersion and fledging success in field-

forest ecotones. Ecology 59(5): 871-883.

GeoNB. 2020. Data Catalogue. Available from

http://www.snb.ca/geonb1/e/DC/catalogue-E.asp [accessed 14 January 2020].

Graves, G. R. 2004. Avian commensals in Colonial America: When did Chaetura

pelagica become the chimney swift? Archives of Natural History 31(2): 300-

307.

Habib, L., E. M. Bayne, and S. Boutin. 2007. Chronic industrial noise affects pairing

success and age structure of overnbirds Seiurus aurocapilla. Journal of Applied

Ecology 44: 176-184.

Hale, R. and S. E. Swearer. 2016. Ecological traps: Current evidence and future

directions. Proceedings of the Royal Society B 283: 20152647. DOI:

10.1098/rspb.2015.2647.

Harper, K. A., Lesieur, D., Bergeron, Y., and Drapeau, P. 2004. Forest structure and

composition at young fire and cut edges in black spruce boreal forest. Canadian

Journal of Forest Research 34: 289-302.

Hutto, R. L. 1995. Composition of bird communities following stand-replacement fires

in Northern Rocky Mountain (U.S.A.) conifer forests. Conservation Biology

9(5): 1041-1058.

Hutto, R. L. and Young, J. S. 1999. Habitat relationships of landbirds in the Northern

Region, USDA Forest Service. Gen. Tech. Rep. RMRS-GTR-32. Ogden, UT:

U.S. Department of Agriculture, Forest Service, Rocky Mountain Research

Station. 72 p.

Ibarzabal, J. and A. Desrochers. 2004. A nest predator’s view of a managed forest: Gray

Jay (Perisoreus canadensis) movement patterns in response to forest edges. Auk

121(1): 162-169. DOI: 10.1093/auk/121.1.162.

Lent, R. A. and D. E. Capen. 1995. Effects of small-scale habitat disturbance on the

ecology of breeding birds in a Vermont (USA) hardwood forest. Ecography 18:

97-108.

Loo, J. and N. Ives. 2003. The Acadian forest: Historical condition and human impacts.

The Forestry Chrinicles 79(3): 462-474.

Nebel, S., A. Mills, J. D. McCracken, and P. D. Taylor. 2010. Declines of aerial

insectivores in North America follow a geographic gradient. Avian Conservation

and Ecology 5(2): 1. [online] URL: http://www.ace-eco.org/vol5/iis2/art1/.

North American Bird Conservation Initiative Canada. 2019. The State of Canada’s

Birds, 2019. Environment and Climate Change Canada, Ottawa, Canada. 12

pages. www.stateofcanadasbirds.org.

Powell, L. L., Hodgman, T. P., Glanz, W. E., Osenton, J. D., and Fisher, C. M. 2010.

Nest-site selection and nest survival of the rusty blackbird: Does timber

management adjacent to wetlands create ecological traps? Condor 112(4): 800-

809. DOI: 10.1525/cond.2010.100152.

R Core Team. 2018. R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. Available from

https://www.R-project.org/.

Robertson, B. A., and R. L. Hutto. 2006. A framework for understanding ecological

traps and an evaluation of existing evidence. Ecology 87(5): 1075-1085.

Robertson, B. A., and R. L. Hutto. 2007. Is selectively harvested forest and ecological

trap for Olive-sided Flycatchers? Condor 109(1): 109-121.

Sánchez-Clavijo, L., J. Hearns, and P. F. Quintana-Ascencio. 2016. Modelling the effect

of habitat selection mechanisms on population responses to landscape structure.

Ecological Modelling 328: 99-107.

Schlaepfer, M. A., M. C. Runge, and P. W. Sherman. 2002. Ecological and evolutionary

traps. Trends in Ecology & Evolution 17(10): 474-480.

Spies, T. A., McComb, B C., Kennedy, R. S. H., McGrath, M. T., Olsen, K., and Pabst,

R. J. 2007. Potential effects of forest policies on terrestrial biodiversity in a

multi-ownership province. Ecological Applications 17(1): 48-65.

Stewart, R. L. M., K. A. Bredin, A. R. Couturier, A. G. Horn, D. Lepage, S. Makepeace,

P. D. Taylor, M. -A. Villard, and R. M. Whittam (eds). 2015. Second Atlas of

Breeding Birds of the Maritime Provinces. Bird Studies Canada, Environment

Canada, Natural History Society of Prince Edward Island, Nature New

Brunswick, New Brunswick Department of Natural Resources, Nova Scotia Bird

Society, Nova Scotia Department of Natural Resources, and Prince Edward

Island Department of Agriculture and Forestry, Sackville, 528 + 28 pp.

Strickland, D. and H. R. Ouellet. 2020. Canada Jay (Perisoreus canadensis), version

1.0. In Birds of the World (P. G. Rodewald, Editor). Cornell Lab of Ornithology,

Ithaca, NY, USA. DOI: 10.2173/bow.gryjay.01.

Suarez, A. V., K. S. Pfennig, and S. K. Robinson. 1997. Nesting success of a

disturbance-dependent songbird on different kinds of edges. Conservation

Biology 11(4): 928-935.

Summerville, K. A. and T. O. Crist. 2002. Effects of timber harvest on forest

Lepidoptera: Community, guild and species responses. Ecological Applications

12(3): 820-835. van den Berg, L. J. L., and J. M. Bullock, R. T. Clarke, R. H. W. Langston, R. J. Rose.

2001. Territory selection by the Dartford warbler (Sylvia undata) in Dorset, 61

England: the role of vegetation type, habitat fragmentation and population size.

Biological Conservation 101: 217-228.

Venables, W. N., and Ripley, B. D. 2002. Modern applied statistics with S. Fourth

Edition, Springer, New York. ISBN 0-387-95457-0.

Wright, J.M. 1997. Preliminary study of olive-sided flycatchers, June 1994-April 1997.

Alaska Department of Fish and Game. Final research report. Endangered species

conservation fund federal aid studies SE-3-3, 4 and 5. Juneau, Alaska. 34 pp.

Yahner, R. H. 2008. Bird responses to a managed forested landscape. Wilson Journal of

Ornithology. 120(4): 897-900.

Chapter 4 General Discussion

My research focused on the effects of forest harvesting on olive-sided flycatcher

(Contopus cooperi) reproductive activity in the Acadian forest of central New Brunswick.

My primary goal was to determine if post-harvest habitat, specifically clearcuts, created an ecological trap for olive-sided flycatchers in the Acadian forest. I hypothesized that the presence of post-harvest habitat (clearcuts) within an olive-sided flycatcher’s territory would negatively affect the individual’s fitness by reducing suitable prey (Atlegrim and

Sjöberg 1995, Brooks and Nocera 2020) and increasing nest predation (King et al. 2008)..

However, the data did not support my hypothesis; I did not detect ecological traps for olive-sided flycatchers in the Acadian forest of central New Brunswick.

Using Autonomous Recording Units (ARU; Wildlife Acoustics, Inc. ® Song

Meter SM2, Song Meter SM4), I monitored individuals in their breeding territories from arrival to departure date in 2018 and 2019. Olive-sided flycatchers are difficult to capture using traditional mist netting techniques (Bonter et al. 2008) and are currently designated as threatened under the Species at Risk Act (SARA) in Canada (Environment Canada

2016). As such, developing a method to study their reproductive activity without handling the individual is important for understanding the species fitness and potential causes of population decline.

In Chapter 2, I developed a method to study olive-sided flycatcher reproductive activity by combining the use of ARUs, a free automated signal recognition program

(monitoR; Hafner and Katz 2018), and change-point analyses. The method I developed accurately tracks the call rate of olive-sided flycatchers in their breeding territory and

determines when a significant change in their call rate occurs. These changes either correspond with anticipated changes in reproductive activity (e.g., from incubation to post-hatch stage) or they do not correspond indicating a failure of reproductive activity.

When considering erroneously classified individuals, this method accurately detected nest success within the expected rate of success (30-65%; COSEWIC 2018).

In developing the method to monitor reproductive activity of olive-sided flycatcher I chose to use the automated signal recognition program ‘monitoR’ (Hafner and Katz 2018). This is one of many automated signal recognition programs (e.g.,

Avisoft-SASLab Pro, Raven Pro, Song Scope Pro, Syrinx). Although less labour intensive than manually processing, automated signal recognition programs are not fault- free and often inaccurate (Priyadarshani et al. 2018). Using ARUs limited the spatial scale of my study. As ARUs are stationary, to cover the same area as a human observer would require a large number of ARUs (Shonfield and Bayne 2017). Future studies should focus on refining the method developed in Chapter 2 for other species that exhibit behavioural changes with vocalizations.

In Chapter 3, I assessed the effects of commercial logging in central New

Brunswick on the habitat selection of olive-sided flycatchers in the breeding season. I sought to determine if harvest sites were acting as an ecological trap for olive-sided flycatchers by monitoring their reproductive activity in each site. I found olive-sided flycatchers were settling in territories containing clearcuts over other available territories without clearcuts. I determined, at spatial scales of 4.9 ha and 45 ha, olive-sided flycatchers selected territories within older cutblocks. Within a scale of 19.6 ha, they selected territories with a small proportion of the territory being a clearcut. Lastly, I 64

found that reproductive activity did not differ between habitat types and thus did not meet the conditions of an ecological trap.

To determine which habitats olive-sided flycatchers settled first, I used the presence/absence of individuals, detected using point counts and ARU recordings, to determine habitat selection. I analyzed habitat selection at the spatial scales of three hypothetical territory sizes (4.9 ha, 19.6 ha, and 45 ha). The three landscape scales used were selected as olive-sided flycatcher will typically use territories up to 20 ha and have been observed using territories up to 45 ha (Altman and Sallabanks 2020). I included cut block age and size, habitat type (harvested, wetland, of harvest-wetland), year, and cut size after 2008. The reproductive activity method developed in Chapter 2 was used to determine if individuals were successfully breeding. I then analyzed whether cut block age and size, habitat type (harvested, wetland, of harvest-wetland), year, or cut size after

2008 had an effect on reproductive activity.

I did not find that clearcuts in the Acadian forest of New Brunswick create an ecological trap for olive-sided flycatchers. This is contradictory to findings in Montana that found that forest harvesting acted as a “severe” ecological trap for olive-sided flycatchers when post-burn sites are also available (Robertson and Hutto 2007). Olive- sided flycatchers are a post-disturbance specialist that require the presence of tall snags in their breeding territory for feeding and defending the area (Altman and Sallabanks 2020,

Environment Canada 2016). However, my study did not include nor have access to any post-burn sites. Although I did not find evidence of an ecological trap, the nest success of olive-sided flycatchers in the Acadian forest may be a factor in the species’ population

decline. The nest success rate (27%), when considering a nest with one fledgling is successful, may not be high enough to maintain the population.

An issue with presence/absence studies is the inaccurate detection of individuals

(MacKenzie 2005). I included sites as absent only if I had ARUs present that did not detect the species' calls. Thus, I was able to ensure that olive-sided flycatchers were genuinely absent from these sites. However, I lacked sites where olive-sided flycatchers were absent as I was limited by the number of ARUs in the study. Although I did not find that post-harvest habitat in the Acadian forest is an ecological trap, I did not include any post-burn sites in the study. In this study, I did not focus on why individuals select specific territories. Snags and living coniferous trees are often noted as important in olive-sided flycatcher habitat (Wright 1997). However, to my knowledge, there is no clear understanding of what cues individuals use to select territory in the breeding season.

Future studies should focus on the cues individuals use to select territories with a focus on predator presence/absence or activity and prey availability. Further studies need to be conducted to develop an understanding of olive-sided flycatcher predators and diet as both are relatively understudied (Altman and Sallabanks 2020).

A significant limitation to this study is the true accuracy of the reproductive analysis method developed in Chapter 2. In this study I did not actively nest search as olive-sided flycatchers’ territories are large and nests can be up to 34 m above the ground making them difficult to detect and monitor (Altman and Sallabanks 2020). As well, there is limited knowledge of the microhabitat characteristics to assist in detecting these nests (Altman and Sallabanks 2020). Another issue, resulting in unknown accuracy, is that many of the recordings for an individual were incomplete, i.e., they lacked a 66

complete set of recordings including all days from arrival to departure. The best way to understand the accuracy of my method would be to actively monitor olive-sided flycatchers to assist in nest finding. The nest activity, such as incubation, hatching, and fledglings could be used to determine if the method is truly successful. In future research, it would be important to deploy ARUs prior to arrival and leave each ARU at a site even if an olive-sided flycatcher was not detected with point counts. A number of individuals were not detected with point counts but were detected with the using the R package

‘monitoR’ after the units were removed from the site.

I suggest that in the Acadian forest, post-commercial harvesting areas do not have a negative effect on olive-side flycatcher reproduction and may not be a factor in the population decline of olive-sided flycatchers. However, forest harvesting may affect other aspects such as diet quality, predation, or clutch size. In this study, I included only clearcuts within a territory. I did not include other types of commercial forestry such as pre-commercial thinning, partial cut, or patch cut. Olive-sided flycatchers may have a difference in fitness in territories with other commercial forestry operations.

Olive-sided flycatcher populations are declining throughout North America, yet forest harvesting is increasing (Altman and Sallabanks 2020). Several studies have shown that olive-sided flycatchers are responding positively to forest harvesting throughout their breeding range (Franzreb 1977, Medin and Booth 1989, Tobalske et al. 1991, Altman and

Sallabanks 2020). In North America, approximately 6.1 million ha are harvested annually

(Masek et al. 2011). As forest harvesting practices vary throughout olive-sided flycatcher breeding range there may be other factors affecting the species’ drastic decline (Altman and Sallabanks 2020). 67

Fire suppression, reduction in prey availability, and habitat loss in non-breeding territories are suspected to be large threats to the olive-sided flycatcher population.

Historically, fire was a dominant disturbance in the Canadian boreal forest (Stocks et al.

2003). Currently, fires are suppressed in Canada with approximately 97% of fires suppressed before they exceed 200 ha (Stocks et al. 2003). Post-burn sites are important habitat for olive-sided flycatchers. However, fires are relatively rare in the Acadian forest with an annual burn size of 12,000 ha (Wein and Moore 1977). Olive-sided flycatcher population declines may be affected by the global decline of insect diversity and abundance. Globally, Lepidoptera and Hymenoptera species are experiencing drastic declines (Sánchez-Bayo and Wyckhuys 2019). Both Lepidoptera and Hymenoptera are important prey for olive-sided flycatchers (Altman and Sallabanks 2020). As important prey species decline, olive-sided flycatcher populations will also decline.

As olive-sided flycatchers are a neotropical migrant, typically breeding in North

America and wintering in South America, conservation and management efforts must be a joint effort. Collaborative work between governments would encompass threats olive- sided flycatchers are facing year-round. In general, a greater understanding of olive-sided flycatcher life history and threats to the species is needed (Altman and Sallabanks 2020).

Further studies on olive-sided flycatcher diet are needed to determine how prey availability and quality are affected population declines in the breeding and non-breeding territories. I suggest a study focusing on the migration and over-wintering areas of olive- sided flycatchers to better understand where important stopover sites are for this species.

As olive-sided flycatchers have a large breeding range, a better understanding of how and

where they migrate to may assist in our understanding of threats to individuals in different areas.

Bibliography

Altman, B. and Sallabanks, R. 2020. Olive-sided flycatcher (Contopus cooperi), version

1.0. In Birds of the World (A. F. Poole, Editor). Cornell Lab of Ornithology,

Ithaca, NY, USA. https://doi.org/10.2173/bow.olsfly.01

Bonter, D. N., E. W. Brooks, and T. M. Donovan. 2008. What are we missing with only

ground-level mist nets? Using elevated nets at a migration stopover site. Journal of

Field Ornithology 79(3): 314-320. DOI: 10.1111/j.1557-9263.2008.00179.x.

COSEWIC. 2018. COSEWIC assessment and status report on the Olive-sided Flycatcher

Contopus cooperi in Canada. Committee on the Status of Endangered Wildlife in

Canada. Ottawa. ix + 52 pp.

Environment Canada. 2016. Recovery strategy for the Olive-sided Flycatcher (Contopus

cooperi) in Canada. Species at Risk Act Recovery Strategy Series. Environment

Canada, Ottawa. vii + 52 pp.

Franzreb, K. E. 1977. Bird population changes after timber harvesting of a mixed conifer

forest in Arizona. Fort Collins, CO: Rocky Mtn. Forest and Range Exp. Stn.

Haché, S., Solymos, P., Bayne, E.M., Fontaine, T., Cumming, S.G., Schmiegelow,

F.K.A., and Stralberg, D. 2014. Analyses to support critical habitat identification

for Canada Warbler, Olive-sided Flycatcher, and Common Nighthawk, Final

Report 1. Project K4B20- 13-0367. Prepared for Environment Canada on behalf of

the Boreal Avian Modelling Project. Edmonton, AB.

Hafner, S. and Katz, J. 2018. monitoR: Acoustic template detection in R. R package

version 1.0.7. URL:

http://www.uvm.edu/rsenr/vtcfwru/R/?Page=monitoR/monitoR.htm 70

MacKenzie, D. I. 2005. What are the issues with presence-absence data for wildlife

management? Journal of Wildlife Management 69(3): 849-860.

Masek, J. G., W. B. Cohen, D. Leckie, M. A. Wulder, R. Vargas, B. de Jong, S. Healey,

B. Law, R. Birdsey, R. A. Houghton, D. Mildrexler, S. Goward, and W. B. Smith.

2011. Recent rates of forest harvest and conversion in North America. Journal of

Geophysical Research 116: G00K03. DOI: 10.1029/2010JG001471.

Medin, D. E. and G. D. Booth. 1989. Responses of birds and small mammals to single-

tree selection logging in Idaho. United States Department of Agriculture.

Priyadarshani, N., S. Marsland, and I. Castro. 2018. Automated birdsong recognition in

complex acoustic environments: A review. Journal of Avian Biology 49: ejav-

01447. DOI: 10.1111/jav.01447.

Robertson, B. A., and R. L. Hutto. 2007. Is selectively harvested forest and ecological

trap for Olive-sided Flycatchers? Condor 109(1): 109-121.

Sánchez-Bayo, F. and K. A. G. Wyckhuys. 2019. Worldwide decline of the entomofauna:

A review of its drivers. Biological Conservation 232: 8-27. DOI:

10.1016/j.biocon.2019.01.020.

Shonfield, J. and E. M. Bayne. 2017. Autonomous recording units in avian ecological

research: Current use and future applications. Avian Conservation and Ecology

12(1): 14. DOI: 10.5751/ACE-00974-120114.

Stocks, B. J., J. A. Mason, J. B. Todd, E. M. Bosch, B. M. Wotton, B. D. Amiro, M. D.

Flannigan, K. G. Hirsch, K. A. Logan, D. L. Martell, and W. R. Skinner. 2003.

Large forest fires in Canada, 1959-1997. Journal of Geophysical Research

108(D1): FFR 5-1-FFR 5-12. DOI: 10.1029/2001JD000484. 71

Tobalske, B. W., R. C. Shearer, and R. L. Hutto. 1991. Bird populations in logged and

unlogged western Larch/Douglas-fir forest in Northwestern Montana. United

States Department of Agriculture, Forest Service, Intermountain Research Station.

Wein, R. W. and J. M. Moore. 1977. Fire history and rotations in the New Brunswick

Acadian Forest. Canadian Journal of Forest Research 7: 285-294.

Wright, J.M. 1997. Preliminary study of olive-sided flycatchers, June 1994-April 1997.

Alaska Department of Fish and Game. Final research report. Endangered species

conservation fund federal aid studies SE-3-3, 4 and 5. Juneau, Alaska. 34 pp.

Appendix A Bumble bee (Bombus spp.) diversity differs between

forested wetlands and clearcuts in the Acadian forest1

A.1. Abstract

Bumble bees (Bombus spp.) are important pollinators that are generally in population decline, but species presence and relative abundance are unknown in forested wetlands of the Acadian forest. To address this knowledge gap, we sampled bumble bees in forested wetlands and harvested (clearcut) sites using vane- and pan-traps. We collected 617 specimens representing 11 species. We also included observations from iNaturalist (n = 70) in disturbed sites. We found that species-specific abundance in

Acadian forested wetlands differ significantly from harvested sites. Wet coniferous forests with moderate to high herbaceous cover have greater overall bumble bee abundance than do harvest sites. Species interactions also may influence community structure; sites with higher abundance of Bombus borealis and B. ternarius had fewer B. fervidus, B. flavidus, and B. terricola. Differences in presence and abundance of bumble bee species may be explained by forested wetlands having a greater variety of flowering plants than forest harvest sites.

1 Reprinted from: Brooks, D.R. and J.J. Nocera. 2020. Bumblebee (Bombus spp.) diversity differs between forested wetlands and clearcuts in the Acadian forest. Canadian Journal of Forest Research, In Press.

KEYWORDS – Clearcut; diversity; forested wetland; Bombus; population decline

A.2. Introduction

Forested wetlands are important and diverse ecosystems, ranging from bogs to shallow open waters. These types of habitats are abundant in the core of the Acadian forest (e.g., New Brunswick, Canada) where wetlands make up approximately 4%

(300,000 ha) of New Brunswick’s land base; 41% of which is classified as freshwater inland wetlands and 49% is classified as inland bogs (Natural Resource and Energy

2002). These wetlands provide crucial habitat for breeding birds, herpetofauna, vegetation, and pollinator species (Johnson et al. 2016; Riffell et al. 2006; Vickruck et al.

2019).

New Brunswick hosts 6.1 million ha of forest (86% of the province’s land area),

3.2 million ha of this forest are on public (Crown) land and 2.9 million ha are privately owned (Province of New Brunswick 2019). While forest management agreements are in place for forested public land (Province of New Brunswick 2019), the privately owned forested land is free to be managed by the landowner however they wish, so long as all activities are in accordance with the Clean Water Act, SNB 1989, c C-6.1 (Province of

New Brunswick 2019). As a result, wildlife populations may face different pressures on public vs. private forested land in New Brunswick. Regardless of these potential differences, forest harvesting practices such as clearcutting remain common on both public and private land, and may remove important habitat for wildlife, including pollinator species. Clearcutting can affect the diversity of insect communities (Korpela et

al. 2015), particularly if the clearcuts receive extensive silviculture treatments (Johansson et al. 2020).

Bumble bees (Bombus spp.) are an important group of pollinator species found in the Acadian forest. Several bumble bee species (e.g., Bombus affinis Cresson, B. terricola

Kirby, B. fervidus Fabricius) are facing multiple localized/regional population declines, primarily towards the cores of their ranges (Colla and Packer 2008; Grixti et al. 2008;

Klymko and Sabine 2015). Major factors affecting population declines in bumble bee species include loss of habitat, habitat fragmentation, parasites, and pesticide use

(Committee on the Status of Pollinators in North America 2007). Forested wetlands can be rich in floral resources and may act as high-quality habitat patches for pollinators in the Acadian forest. Certain forest harvesting practices (e.g., clearcutting) in the Acadian forest, especially in forested wetlands, may be contributing to changes in bumble bee populations through habitat loss and fragmentation (Hines and Hendrix 2005; Richardson et al. 2019).

In this study, we sought to determine if bumble bee diversity differs between forested wetlands and other habitat types in New Brunswick. Secondly, to address the potential effect of forest harvesting, we sought to determine if there is a difference in bumble bee community parameters (e.g., abundance, diversity) between forested wetlands and harvest sites adjacent to forested wetlands in New Brunswick.

A.3. Materials and Methods

Study Area

We sampled bumble bees from 16 sites in four counties of New Brunswick: Kent

(46° 26Ń 65° 16´W), Northumberland (46° 54Ń 65° 59´W), Sunbury (45° 57Ń 66°

21´W), and Queens (45° 52´N 65° 56´W) from 17 June-30 July 2018 and 15 June-31 July

2019. This area of New Brunswick is a part of the Acadian forest region with coniferous and deciduous species present. The area is also actively clearcut by numerous commercial harvesting companies. We stratified our bumble bee sampling between harvest sites (clearcuts) and forested wetlands (beaver pond, black spruce (Picea mariana) bog, wet coniferous forest). The harvest sites in our study ranged from 24 to

140 acres whilst the forested wetlands ranged from two to 60 acres. All harvest sites in this study were clearcut between 2011 and 2013 except for one site that was clearcut in

2019. The harvest site that was cut in 2019 was not sampled in 2018.

To collect bumble bee specimens, we used blue vane traps (henceforth BVT), blue pan-traps, and yellow pan-traps. We included BVTs as they are efficient at capturing

Bombus spp. with relatively low capture rates of other insects (Stephen and Rao 2005).

Pan-traps were 25cm in diameter and approximately 2.5cm deep. One of each of the trap types was placed at each site. We chose to sample with one of each trap at a site to expand the spatial scale of the study, recognizing that this was at the expense of deeper replication at each site. We hung the BVT from a tree at an approximate height of 1m.

We deployed the pan traps by sinking them slightly in the ground to make them flush with ground level, filled them with water and liquid detergent (to break surface tension and allow for submersion capture; 2.5ml liquid detergent:500ml water), and made sure they were in an open area within 5m of the BVT. In 2018, we deployed BVT and pan- traps at seven sites (four forested wetlands, three harvest sites). In 2019, we deployed 76

BVT and pan-traps at fifteen sites (eleven forested wetlands, four harvest sites). Of the sites we sampled, four were sampled both years (three forested wetlands, one harvest site). In 2019, pan-traps alone were deployed at four additional sites (three forested wetlands, one harvest site) as we were limited in the number of BVT available to us. We monitored traps and removed specimens every 48 hours. We then preserved samples in

95% ethyl alcohol. We dried and pinned the bumble bee specimens following the protocol outlined by Mengis (2014). We identified whole specimens to the species level using Williams et al. (2014). In our classification and analysis, we included cuckoo bumble bee species, as their presence may reflect the presence of their host species which, a priori, we could not be sure would be captured in our study. We classified habitat as clearcut or forested wetland. Forested wetland was then further classified using the Forest Ecosystem Classification for Nova Scotia which uses the plant community to identify the habitat type; we chose this classification protocol for its ease of use, accessibility, and habitat similarity to New Brunswick (see Table 4.1; Neily et al. 2011).

Citizen Science

We gathered observational data on bumble bee species for the province of New

Brunswick from the citizen science application iNaturalist, recognizing that these data can sometimes be imperfect (e.g., misidentification, variable and biased search effort).

We used all bumble bee sightings from 19 June to 31 July 2018 and 15 June to 29 July

2019. To reduce some of the error associated with misidentification we used only bumble bee sightings that were considered by iNaturalist as being “research quality”, i.e., a category of observations where >66% of community identifiers agree on identification

(iNaturalist 2020). We included iNaturalist bumble bee sightings that had their location 77

obscured, due to conservation status, only when we could obtain the habitat of observed individuals as forested wetland or as a disturbed site (which comprise harvested sites and urban sites; see below) using geographic data layers provided by GeoNB (2020) and

Google Earth. Some iNaturalist data were collected in urban areas, a habitat type we did not sample in our trapping. When appropriate for the analysis that included iNaturalist data, we included urban and harvested habitat in the category of “disturbed”.

Statistical Analysis

We conducted all statistical analyses in R v.3.5.2 (R Core Team 2018). To determine if the presence/absence of each Bombus species in our study differed between forested wetlands and disturbed sites, we ran a chi-squared test. For this test, we combined our sampling data (trap data) and the citizen science data (observed data), and we categorized the habitat as either forested wetland (i.e., beaver pond, black spruce bog, wet coniferous forest) or disturbed site (e.g., harvest, urban).

Although abundance of bumble bees is more accurately measured using nest abundance or molecular methods due to their eusocial behaviour (Geib et al. 2015), we calculate abundance as a simple proxy of functional diversity and ecosystem service delivery (extent of pollination). Therefore, we conducted a second chi-squared test (again with trap and citizen science data combined) to determine if the overall abundance of all bumble bees differed between forested wetlands and disturbed sites. We ran a final chi- squared test, on trap data only, to determine if there was a significant difference in overall abundance of bumble bees between forested wetlands and harvest sites (i.e., clearcuts; the only form of disturbed sites we sampled with traps); the citizen science data incorporated urban sites as disturbed and were thus excluded from this analysis. For all analyses we 78

incorporated bumble bees captured in pan traps and vane traps into one as there was only

8 instances where pan traps captured bumble bees. For all analyses, we set α = 0.05.

To determine if species-specific abundance of bumble bees differed between forested wetlands and harvest sites, we used a Poisson generalized linear mixed model using a penalized quasi-likelihood estimator (glmmpql in the package ‘MASS’ [Venables and Ripley 2002]). We determined that the errors in our data were Poisson distributed using the package ‘fitdistrplus’ (Delignette-Muller and Dutang 2015) and by comparing fit among twelve distribution families. We set species-specific abundance as the response variable, species and habitat as predictor variables, and site and year as non-nested random effects.

We used the R package ‘vegan’ (Oksanen et al. 2019) to determine species diversity and richness in forested wetlands and harvest sites. We used the Shannon’s index to determine species diversity of the five habitat types (i.e., beaver pond, black spruce forest, wet coniferous forest 1, wet coniferous forest 2, clearcut). We chose to use

Shannon’s index over Simpson’s index as it is generally more sensitive to rare species, and there are several rare species in the Acadian forest. However, we do not necessarily know if that rarity is because of fewer nests/colonies or fewer individuals per nest/colony.

We also calculated species evenness, followed by rarefaction to standardize sample sizes.

A.4. Results

Bumble bee Sampling

In 2018, we collected 158 bumble bee specimens, representing eleven species (B. borealis Kirby, B. fervidus Fabricius, B. flavidus Franklin, B. impatiens Cresson, B.

insularis Smith, B. perplexus Cresson, B. rufocinctus Cresson, B. sandersoni Franklin, B. ternarius Say, B. terricola Kirby, B. vagans Smith). In 2019, we collected 459 bumble bee specimens, representing ten of the same species except for B. rufocinctus. In total, we collected 617 bumble bee specimens of which B. ternarius was the most common species.

In 2018, there were 19 bumble bee observations from iNaturalist. Given the small sample size we obtained from iNaturalist compared to our own trapping data, we are confident that any errors associated with misidentification and/or search bias are reduced.

These observations represented seven species (B. bimaculatus Cresson, B. borealis, B. flavidus, B. impatiens, B. ternarius, B. terricola, B. vagans). In 2019, there were 51 bumble bee observations from iNaturalist. These observations represented eight species

(the same seven species as 2018 and B. rufocinctus). The most common species observed was B. ternarius. In total, we obtained 70 bumble bee observations from iNaturalist.

Statistical Analysis

Species presence/absence in forested wetlands did not differ from disturbed sites

(χ2 = 2.9621, df = 12, p = 0.9958). However, we found the overall abundance of bumble bees in forested wetlands was significantly higher than in disturbed sites (χ2 = 55.819, df

= 12, p = 1.29×10-7). This relationship persisted even after removing citizen science data

(which included urban habitat) thus, using only our trap data in harvested sites (χ2 =

35.659, df = 10, p = 9.633×10-5).

We found evidence of species interactions influencing community structure. Sites with higher abundance of B. borealis and B. ternarius showed lower abundance of B. fervidus, B. flavidus, and B. terricola (Table 4.2). We found species specific abundance 80

to be greater overall in wet coniferous forests with moderate to high herbaceous cover compared with harvest sites (Table 4.3). Bumble bee abundance (species specific) was found to be lowest overall in clearcuts and open black spruce bogs (Table 4.3). As bumble bees are pollinators, their life-history is tied to flower phenology (Bächman and

Tiainen 2002; Mizunaga and Kudo 2017). To put our results into context, we obtained average floral phenology patterns for some of the most common flower species in our study sites and present them in Table 4.4. Our modelling revealed statistically significant relationships to habitat for B. ternarius (t = 2.50, p =0.01) and B. terricola (t = 2.02, p =

0.05), both of which showed a positive relationship with forested wetlands. We found species diversity and evenness was highest in harvest sites and lowest in wet coniferous forests with moderate to high herbaceous cover (wet coniferous forest 1; Table 4.5).

However, rarefaction showed evenness to be highest in beaver ponds and lowest in wet coniferous forests with high shrub and sphagnum moss cover (wet coniferous forest type

2; Table 4.5).

A.5. Discussion

We found that bumble bee communities were significantly different in forested wetlands than in clearcuts and urban areas (i.e. disturbed sites) in that overall abundance was greater in forested wetlands; however, we do not know if that reflects a greater abundance of nests/colonies or simply more members of nests/colonies. This was especially apparent in our trap data where bumble bee abundance was significantly greater in forested wetlands than in harvest sites. However, our results show that presence/absence of species does not differ between forested wetlands and disturbed

sites. We found that species specific abundance was greatest in wet coniferous forests with a moderate to high herbaceous cover (wet coniferous forest 1; refer to Neily et al.

2011) and lowest in harvest sites. Therefore, in habitats fragmented by anthropogenic land use, such as forest harvesting and urban areas, forested wetlands may provide refuge for pollinators such as bumble bees either by yielding more nests/colonies or by enhancing the size of nests/colonies.

The wet coniferous forests we studied that had a high abundance of bumble bees are dominated by black spruce, cinnamon fern (Osmunda cinnamomea), and sphagnum moss (Sphagnum spp.) (Neily et al. 2011). This habitat also hosts a variety of flowering plants a (Neily et al. 2011). The overall higher abundance of bumble bees in this habitat type may be due to the variety of flowering plants present. The plants at our study sites flower throughout the active bumble bee season (May to September) providing an abundance of food in the wet coniferous forest habitat. Lowbush blueberry (Vaccinium angustifolium) may be especially important in our study, as it is known to be a floral association for species such as B. borealis, B. flavidus, B. insularis, B. perplexus, B. ternarius, B. terricola, and B. vagans (Boulanger et al 1967; Bumble Bee Watch 2020).

We noted evidence of community interactions in our study. Sites with a high abundance of B. borealis and B. ternarius – two species with relatively stable populations

– generally had a lower abundance of B. fervidus, B. flavidus, and B. terricola. Both B. fervidus and B. terricola are known to be facing drastic population declines (Colla and

Packer 2008; Jacobson et al. 2018). The higher abundance of B. borealis and B. ternarius may be primarily due to their commonness as well as their habitat preferences. The two species prefer to be close to, or within, wooded areas with B. ternarius also showing 82

preference for wetlands (Williams et al. 2014). Although B. terricola also shows habitat preference for woodlands and wetlands, the species has been experiencing recent population declines throughout its range (COSEWIC 2015). The relatively low abundance of B. fervidus that we detected is likely due to its population decline as well as its preference for open grassland, farmland, and urban areas such as gardens and parks

(Williams et al. 2014). B. flavidus may have a relatively low abundance in our study sites due to its uncommonness and because it is a colony parasite of B. rufocinctus and bumble bees in the subgenus Pyrobombus, such as B. perplexus (Colla et al. 2011; Williams et al.

2014), both species that are present in the forested wetlands in our study but in low abundance.

The overall low abundance of bumble bees in harvest sites shows that there is a negative effect on pollinators created by forest harvesting in the Acadian forest of New

Brunswick, either by reducing abundance of nests/colonies or reducing individual nest/colony size. This is supported by Pengelly and Cartar (2010) who found that the presence of a logged area near a forested area – such as a forested wetland – will disrupt the plant-pollinator relationship in both the disturbed and the undisturbed habitat in the boreal forest. The fragmentation of land created by forest harvesting may be decreasing the presence and abundance of bumble bees at local and large spatial scales.

It is possible that we did not detect a significant difference in species presence between forested wetlands and disturbed sites because harvest sites we studied are not far away enough to be independent of forested wetlands. The area we studied is, and has been, under active commercial harvesting, meaning that few forested wetlands would not have a clearcut within the potential movement range of a bumble bee. Individuals have 83

been shown to use large areas (up to 43.53 ha) and travel long distances (Hagen et al.

2011). Further, commercial harvesting in New Brunswick can occur as close as 30m to a wetland (Clean Water Act 2019), and some bumble bee species may easily travel between the habitat types by crossing this 30m buffer. It is also possible that, similar to avian species, harvest sites (like clearcuts) can provide some habitat for many bumble bee species resulting in the equal presence of species in harvest sites and forested wetlands

(Costello et al. 2000).

Although we found species diversity and evenness was highest in harvest sites, we suggest this is likely due to harvested sites having a relatively low abundance of bumble bees compared to forested wetlands. We argue that, although species diversity and evenness are highest in harvest sites, the overall low abundance of bumble bees in harvest sites is more meaningful because pollination is an essential ecosystem service.

It is evident from our study that the management and conservation of forested wetlands is necessary to conserve bumble bees as native pollinators in the Acadian forest of New

Brunswick. This is especially necessary when clearcuts are nearby, that will host a smaller abundance of bumble bees than the forested wetlands that provide vital habitat to native pollinators and their floral associations.

A.6. Literature Cited

Bächman, J.-P. C., and Tiainen, J. 2002. Habitat quality of field margins in a Finnish

farmland area for bumblebees (Hymenoptera: Bombus and Psithyrus).

Agriculture, Ecosystems and Environment 89: 53-68.

Boulanger, L. W., Wood, G. W., Osgood, E. A., and Dirks, C. O. 1967. Native bees

associated with the low-bush blueberry in Maine and Eastern Canada. Maine

Agricultural Experiment Station Technical Bulletin 26.

Bumble Bee Watch. 2020. Bumble Bee Species [online]. Available from

https://www.bumble beewatch.org [accessed 3 February 2020].

Colla, S. R. and Packer, L. 2008. Evidence for decline in eastern North American bumble

bees (Hymenoptera: Apidae), with special focus on Bombus affinis Cresson.

Biodiversity and Conservation 17: 1379-1391. doi: 10.1007/s10531-008-9340-5.

Colla, S., Richardson, L., and Williams, P. 2011. Bumble Bees of the Eastern United

States. United States Department of Agriculture. Available from

https://www.fs.fed.us/wildflowers/pollinators/documents/Bumble

beeGuideEast2011.pdf [accessed 14 January 2020].

Committee on the Status of Pollinators in North America. 2007. Status of Pollinators in

North America. National Academy of Sciences, Washington, DC, USA. pp. 84-

88.

COSEWIC 2015. COSEWIC assessment and status report on the Yellow-banded Bumble

Bee Bombus terricola in Canada. Committee on the Status of Endangered

Wildlife in Canada. Ottawa. ix + 60 pp.

Costello, C. A., Yamasaki, M., Pekins, P. J., Leak, W. B., and Neefus, C. D. 2000.

Songbird response to group selection harvests and clearcuts in a New Hampshire

northern hardwood forest. Forest Ecology and Management 127: 41-54.

Delignette-Muller, M. and Dutang, C. 2015. fitdistrplus: An R package for fitting

distributions. Journal of Statistical Software 64(4): 1-34. Available from

http://www.jstatsoft.org/v64/i04/.

Geib, J. C., Strange, J. P., and Galen, C. 2015. Bumble bee nest abundance, foraging

distance, and host-plant reproduction: Implications for management and

conservation. Ecological Applications 25(3): 768-778.

GeoNB. 2020. Data Catalogue. Available from

http://www.snb.ca/geonb1/e/DC/catalogue-E.asp [accessed 14 January 2020].

Grixti, J. C., Wong, L. T., Cameron, S. A., and Favret, C. 2009. Decline of bumble bees

(Bombus) in the North American Midwest. Biological Conservation 142: 75-84.

doi: 10.1016.j.biocon.2008.09.027.

Hagen, M., Wikelski, M., and Kissling, W. D. 2011. Space use of bumble bees (Bombus

spp.) revealed by radio-tracking. PLoS ONE 6(5): e19997. doi:

10.1371/journal.pone.0019997.

Hines, H. M., and Hendrix, S. D. 2005. Bumble bee (Hymenoptera: Apidae) diversity and

abundance in tallgrass prairie patches: Effects of local and landscape floral

resources. Environmental Entomology 34(6): 1477-1484. iNaturalist. 2020. Frequently Asked Questions. Available from

https://www.inaturalist.org/pages/help#quality [accessed 14 January 2020].

Jacobson, M. M., Tucker, E. M., Mathiasson, M. E., and Rehan, S. M. 2018. Decline of

bumble bees in northeastern North America, with special focus on Bombus

terricola. Biological Conservation 217: 437-445. doi:

10.1016/j.biocon.2017.11.026. 86

Johansson, V., Gustafsson, L., Andersson, P., and Hylander, K. 2020. Fewer butterflies

and a different composition of bees, wasps and hoverflies on recently burned

compared to unburned clear-cuts, regardless of burn severity. Forest Ecology and

Management 463: 118033.

Johnson, B. A., Homyack, J. A., Barrett, K., and Baldwin, R. F. 2016. Factors influencing

herpetofaunal assemblages of aquatic systems in a managed pine forest. Forest

Ecology and Management 379: 124-132. doi: 10.1016/j.foreco.2016.08.012.

Klymko, J, and Sabine, D. 2015. Verification of the occurrence of Bombus affinis

(Hymenoptera: Apidae) in New Brunswick, Canada. Journal of the Acadian

Entomological Society 11: 22-25.

Korpela, E.-L., Hyvönen, T., and Kuussaari, M. 2015. Logging in boreal field-forest

ecotones promotes flower-visiting insect diversity and modifies insect community

composition. Insect Conservation and Diversity 8: 152-162.

Mengis, D. 2014. How to wash, dry, and pin bees. Available from

https://www.vanengelsdorpbeelab.com/student-updates/by-dejen-mengis

[accessed 26 November 2019].

Mizunaga, Y., and Kudo, G. 2017. A linkage between flowering phenology and fruit-set

success of alpine plant communities with reference to the seasonality and

pollination effectiveness of bees and flies. Oecologia 185: 453-464. doi:

10.1007/s00442-017-3946-9

Natural Resource and Energy. 2002. New Brunswick Wetlands Conservation Policy.

Government of New Brunswick Natural Resource and Energy.

Neily, P., Basquill, S., Quigley, E., Stewart, B., and Keys, K. 2011. Forest Ecosystem

Classification for Nova Scotia, Part I: Vegetation Types (2010). Nova Scotia

Department of Natural Resources. ISBN 978-1-55457-411-7.

Oksanen, J., Blanchet, F. G., Friendly, M., Kindt, R., Legendre, P., McGlinn, D.,

Minchin, P. R., O’Hara, R. B., Simpson, G. L., Solymos, P., Stevens, M. H. H.,

Szoecs, E., and Wagner, H. 2019. vegan: Community Ecology Package. R

package version 2.5-6. Available from https://CRAN.R-

project.org/package=vegan.

Pengelly, C. J., and Cartar, R. V. 2010. Effects of variable retention logging in the boreal

forest on the bumble bee-influenced pollination community, evaluated 8-9 years

post-logging. Forest Ecology and Management 260: 994-1002. doi:

10.1016/j.foreco.2010.06.020.

Province of New Brunswick. Forest governance in the Province of New Brunswick.

Available from

https://www.sfmcanada.org/images/Publications/EN/New_Brunswick_info_Provi

nces_and_territories_EN.pdf [accessed 26 November 2019].

R Core Team. 2018. R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria. Available from

https://www.R-project.org/.

Richardson, L. L., McFarland, K. P., Zahendra, S., and Hardy, S. 2019. Bumble bee

(Bombus) distribution and diversity in Vermont, USA: a century of change.

Journal of Insect Conservation 23: 45-62. doi: 10.1007/s10841-018-0113-5.

Riffell, S., Burton, T., and Murphy, M. 2006. Birds in depressional forested wetlands:

Area and habitat requirements and model uncertainty. Wetlands 26(1): 107-118.

Stephen, W. P., and Rao, S. 2005. Unscented color traps for non-Apis bees

(Hymenoptera: Apiformes). Journal of the Kansas Entomological Society 78(4):

373-380.

USDA. 2019. Fire Effects Information System (FEIS): Index of Species Information.

Available from https://www.feis-crs.org/feis/faces/index.xhtml [accessed 14

January 2020].

Venables, W. N., and Ripley, B. D. 2002. Modern applied statistics with S. Fourth

Edition, Springer, New York. ISBN 0-387-95457-0.

Vickruck, J. L., Best, L. R., Gavin, M. P., Devries, J. H., and Galpern, P. 2019. Pothole

wetlands provide reservoir habitat for native bees in prairie croplands. Biological

Conservation 232: 43-50. doi: 10.1016/j.biocon.2019.01.015.

Williams, P., Thorp, R., Richardson, L., and Colla, S. 2014. An Identification Guide:

Bumble Bees of North America. Princeton University Press, New Jersey. ISBN

9780691152226.

Table A.1 Characteristic plants of Wet Coniferous Forest 1 and Wet Coniferous Forest 2 as per the

Forest Ecosystem Classification for Nova Scotia (Neily et al. 2011).

WC1 WC2

Species Frequency Frequency

(%) (%)

Balsam fir Abies balsamea 60 25

Black spruce Picea mariana 91 94

Hybrid spruce Picea spp. 0 3

Red maple Acer rubrum 51 19

Tamarack Larix laricina 34 38

White pine Pinus strobus 17 6

False holly Nemopanthus mucronatus 89 94

Huckleberry Gaylussacia spp. 0 19

Labrador tea Rhododendron groenlandicum 38 81

Lambkill Kalmia angustifolia 85 100

Lowbush blueberry Vaccinium angustifolium 35 44

Mountain-ash Sorbus Americana 20 22

Rhodora Rhododendron canadense 0 53

Serviceberry Amelanchier spp. 23 38

Speckled alder Alnus incana 0 19

Velvet-leaf blueberry Vaccinium myrtilloides 49 72

Wild raisin Viburnum nudum 65 88

Winterberry Ilex verticillate 0 13

Bluebead lily Clintonia borealis 29 0

Bracken Pteridium aquilinum 40 47

Bunchberry Cornus canadensis 92 94

Cinnamon fern Osmunda cinnamomea 88 56

Creeping snowberry Gaultheria hispidula 80 75

Dwarf raspberry Rubus pubescens 22 3

Goldthread Coptis trifolia 82 75

Indian pipe Monotropa uniflora 22 0

Mayflower Epigaea repens 0 19

New York fern Thelypteris noveboracensis 20 0

Painted trillium Trillium undulatum 25 13

Partridge-berry Mitchella repens 0 6

Pink lady’s slipper Cypripedium acaule 34 34

Pitcher-plant Sarracenia purpurea 0 3

Round-leaved sundew Drosera rotundifolia 0 6

Sarsaparilla Aralia nudicaulis 46 13

Starflower Trientalis borealis 38 31

Teaberry Gaultheria procumbens 0 22

Three seeded sedge Carex trisperma 65 50

Three-leaved false Solomon’s seal Maianthemum 26 0 racemosum

Trailing blackberry Rubus ursinua 0 13

Twinflower Linnaea borealis 38 0

Wild lily-of-the-valley Maianthemum canadense 43 25

Table A.2 Bombus spp. found in forested wetlands and disturbed sites in 2018 and 2019.

Species Number in Percent of Number in Percent of

forested forested disturbed sites disturbed site

wetlands wetland bumblebees

bumblebee

B. 0 0% 3 1.4% bimaculatus

B. borealis 49 10.2% 25 12.0%

B. fernaldae 0 0% 2 <1%

B. fervidus 5 1% 0 0%

B. flavidus 22 4.5% 8 3.8%

B. impatiens 5 1% 7 3.3%

B. insularis 14 2.9% 2 <1%

B. perplexus 50 10.4% 13 6.3%

B. rufocinctus 2 <1% 2 <1%

B. sandersoni 57 11.9% 27 13%

B. ternarius 221 46.1% 63 30.3%

B. terricola 21 4.3% 22 10.6%

B. vagans 33 6.9% 34 16.3%

Abundance ~ Species + Habitat, ~ 1|Site, ~1|Year

Value Std. Error DF t-value p-value

B. fervidus -1.4589 0.9566 94 -1.5250 0.1306

B. flavidus -0.6061 0.4585 94 -1.3220 0.1894

B. impatiens -1.5078 1.0489 94 -1.4375 0.1539

B. insularis -0.5756 0.5829 94 -0.9876 0.3259

B. perplexus -0.0186 0.3721 94 0.0500 0.9603

B. rufocinctus -1.0430 1.2055 94 -0.8651 0.3892

B. sandersoni 0.0902 0.3488 94 0.2586 0.7965

B. ternarius 0.7292 0.2917 94 2.4999 0.0142

B. terricola -0.9521 0.4720 94 -2.0171 0.0465

B. vagans -0.2628 0.3743 94 -0.7020 0.4844

Black Spruce Bog -0.5399 0.5355 10 -1.0083 0.3371

Clearcut -0.7198 0.5093 10 -1.4132 0.1880

Wet coniferous 1.1036 0.5491 10 2.0100 0.0722 forest 1

Wet coniferous 0.3879 0.6458 10 0.6006 0.5615 forest 2

Table A.4 Plant species phenology (from USDA 2019) present in wet coniferous forest with a moderate to high herbaceous cover.

Flower Species Scientific Name Flowering Time

Lowbush Blueberry Vaccinium angustifolium June – late July

Lambkill Kalmia angustifolia June – early July

Labrador tea Rhododendron groenlandicum Late May – June

Wild Raisin Viburnum nudum June – July

Bunchberry Cornus canadensis May – June

Goldthread Coptis trifolia May – July

Creeping Gaultheria hispidula May – June

Snowberry

Sarsaparilla Aralia nudicaulis May – July

Twin Flower Linnaea borealis June - September

Three-leaved False Maianthemum trifolium May – June

Solomon’s Seal

Table A.5 The Shannon index results of diversity, species richness, species evenness, and rarefaction scores for the four types of forested wetlands and harvest sites (clearcuts). Harvest sites have the highest diversity and evenness (*).

Habitat Shannon Richness Evenness Rarefaction

Index

Beaver pond 1.811523 11 0.755464 8.918933

Black Spruce Bog 1.631484 9 0.74252 8.586478

Wet coniferous 1.375889 8 0.661663 7.898659 forest 1

Wet coniferous 1.43766 7 0.738811 7.000000 forest 2

Clearcut* 1.884498 10 0.818427 8.604527

Figure A.1 Total abundance of Bombus spp. for all trap types caught in forested wetlands and harvest sites (2018 and 2019).

Appendix B Autonomous Recording Unit Call Data for Olive-sided

Flycatchers in Central New Brunswick

Figure B.1 Change-point analysis and call rate of OSFL06 in 2018. ARU recording displays nest- building stage to post-hatch (Q = 4). Red triangles indicate change-point locations, black diamonds indicate dates removed due to playback during recording, and the blue square indicates the date a fledgling was documented.

Figure B.3 Change-point analysis and call rate of OSFL16 in 2019. ARU recording displays arrival of

OSFL to post-hatch stage (Q = 5). Red triangles indicate change-point locations.

100

Figure B.5 Change-point analysis and call rate of OSFL24 in 2019. ARU recording displays arrival date to departure date (Q = 5). Red triangles indicate change-point locations.

Figure B.6 Change-point analysis and call rate of OSFL30 in 2019. ARU recording displays incubation stage to post-hatch stage (Q = 3). Red triangles indicate change-point locations.

101

Figure B.7 Change-point analysis and call rate of OSFL02 in 2018. ARU recordings display arrival of

OSFL to post-hatch stage (Q = 5). Red triangles indicate change-point locations.

Figure B.8 Change-point analysis and call rate of OSFL04 in 2018. ARU recordings display arrival of

OSFL to post-hatch stage (Q = 5). Red triangles indicate change-point locations.

102

Figure B.9 Change-point analysis and call rate of OSFL11 in 2018. ARU recordings display arrival to departure of OSFL (Q = 5). Red triangles indicate change-point locations.

Figure B.10 Change-point analysis and call rate of OSFL13 in 2018. ARU recordings display incubation stage to post-hatch stage (Q = 3). Red triangles indicate change-point locations.

103

Figure B.11 Change-point analysis and call rate of OSFL17 in 2018. ARU recordings display nest building stage to post-hatch stage (Q = 4). Red triangles indicate change-point locations and black diamonds indicate dates removed due to playback during recording.

104

Figure B.13 Change-point analysis and call rate of OSFL28 in 2018. ARU recordings display arrival of OSFL to incubation stage (Q = 4). Red triangles indicate change-point locations.

105

Figure B.15 Change-point analysis and call rate of OSFL02 in 2019. ARU recordings display arrival to departure of OSFL (Q = 5). Red triangles indicate change-point locations.

Figure B.16 Change-point analysis and call rate of OSFL07 in 2019. ARU recordings display arrival to departure of OSFL (Q = 5). Red triangles indicate change-point locations and black diamonds indicate dates removed due to playback during recording.

106

Figure B.17 Change-point analysis and call rate of OSFL16 in 2019. ARU recordings display arrival of OSFL to post-hatch stage (Q = 5). Red triangles indicate change-point locations.

Figure B.18 Change-point analysis and call rate of OSFL17 in 2019. ARU recordings display arrival to departure of OSFL (Q = 5). Red triangles indicate change-point locations.

107

Figure B.20 Change-point analysis and call rate of OSFL31 in 2019. ARU recordings display nest building stage to departure of OSFL (Q = 4). Red triangles indicate change-point locations and black diamonds indicate dates removed due to playback during recording.

108

Figure B.21 Change-point analysis and call rate of OSFL32 in 2019. ARU recordings display incubation stage to departure of OSFL (Q = 3). Red triangles indicate change-point locations and black diamonds indicate dates removed due to playback during recording.

Figure B.22 Change-point analysis and call rate of OSFL43 in 2019. ARU recordings display incubation to post-hatch stage (Q = 3). Red triangles indicate change-point locations.

109

Curriculum Vitae

Delaney Brooks

Education Master of Science in Forestry University of New Brunswick, Fredericton, NB (October 2020) Bachelor of Environmental Studies, Honours Coop University of Manitoba, Winnipeg, MB (June 2017) University of New Brunswick, Fredericton, NB (Fall 2015) o Honours Thesis: A Comparison of Female and Male Nest Attentiveness and Visiting Behaviour in Chestnut-collared Longspur (Calcarius ornatus) Research Experience Waterbird Technician Nature NB, Fredericton, NB (October 2020 – Present) Wildlife Technician Canadian Wildlife Service, Winnipeg, MB (April 2020 – Present) Wildlife Technician Canadian Wildlife Service, Winnipeg, MB (April 2016 – Aug. 2017) Avian Technician Natural Resource Institute, University of Manitoba (May – Sept. 2015) Undergraduate Honours Thesis Natural Resource Institute, University of Manitoba (Sept. 2015 – Aug. 2016) Volunteer Experience Newsletter Editor Atlantic Society of Fish and Wildlife Biologists (Oct. 2019 – Present) Front Desk Staff Ducks Unlimited Conservation Centre (Nov. 2019 – Feb. 2020) Nocturnal Owl Survey Birds Canada (Feb. 2019 – Present) Vice President Forestry Graduate Student Association (Jan. 2018 – Jan. 2020)

Committee Member Graduate Funding Task Force (April – Sept. 2018) Lab Assistant Natural Recourse Institute (Sept. – Oct. 2016) President Environment and Geography Students’ Association (April 2014 – Sept. 2016) Co-Chair and Volunteer Coordinator Manitoba Environmental Industries Association (Feb. 2014 – April 2017) Teaching Experience Teaching Assistant University of New Brunswick o Population Ecology, 1 semester (2019) o Biological Diversity Lab, 3 semesters (2018-2020) o Autecology of Forest Vegetation, 2 semesters (2017-2018) o Social and Cultural Systems, 1 semester (2018) Manuscripts Brooks, D. R., and J. J. Nocera. 2020. Bumble bee (Bombus spp.) diversity differs between forested wetlands and clearcuts in the Acadian forest. Canadian Journal of Forest Research (In Press) Presentations and Proceedings Brooks, D. R., and J. J. Nocera. (2019, August). Reproductive activity of olive-sided flycatchers (Contopus cooperi): Effects of ecological traps in forested wetlands. Oral Presentation at Acadian Entomological Society, Entomological Society of Canada, Canadian Society of Ecology and Evolution Joint Meeting, Fredericton, Canada. Brooks, D. R., and J. J. Nocera. (2019, April). Reproductive activity of olive-sided flycatchers (Contopus cooperi). Oral Presentation at Forestry Graduate Seminar, University of New Brunswick, Fredericton, Canada. Brooks, D. R., and J. J. Nocera. (2019, March). olive-sided flycatchers (Contopus cooperi) in forested wetlands of New Brunswick: Ecological traps and the role of habitat availability. Oral Presentation at Forested Wetlands in Atlantic Canada Concluding Project Workshop, Halifax, Canada. Brooks, D. R. and J. J. Nocera. (2018, August). olive-sided flycatchers (Contopus cooperi) in forested wetlands of New Brunswick: Ecological traps and the role of habitat availability. Poster session presented at the International Ornithological Congress, Vancouver, Canada. Brooks, D. R. and J. J. Nocera. (2018, March). Ecological traps and olive-sided flycatchers (Contopus cooperi). Poster session presented at the Graduate Research Conference, Fredericton, Canada.

Brooks, D. R. and J. J. Nocera. (2018, March). Ecological traps and olive-sided flycatchers (Contopus cooperi). Poster session presented at the Canadian Section of The Wildlife Society AGM, Winnipeg, Canada. Brooks, D. R., and J. J. Nocera. (2017, October). Ecological traps and olive-sided flycatchers (Contopus cooperi). Oral presentation at the Atlantic Society of Fish and Wildlife Biologists AGM, Moncton, Canada. Professional Assessment and Review Reviews Northeastern Naturalist (2018) Graduate Research Conference (2020) Awards and Honours Awards Fraser Inc. Paper Presentation Prize in Forestry (November 2019) ASFWB Fish and Wildlife Grant (April 2019) Forestry Futures Graduate Scholarship (October 2017, 2018) BSC/SCO Student Travel Award (August 2018) CSTWS Student Travel Award (March 2018) UNB Travel Award (October 2017) ASFWB Student Presentation Award (October 2017) New Brunswick Innovation Foundation STEM Award (September 2017) NSERC – USRA (May 2015)