UNIVERSITY of CALIFORNIA Los Angeles Living For

Home , Gray hawk, Pseudastur, Roadside hawk, Stephanoaetus

UNIVERSITY OF CALIFORNIA

Los Angeles

Living for the city: Using community science and historical data

to understand avian response to urbanization

A dissertation submitted in partial satisfaction of the

Requirements for the degree Doctor of Philosophy

in Biology

Daniel Steven Cooper

2020

Daniel Steven Cooper

2020

ABSTRACT OF THE DISSERTATION

Living for the city: Using community science and historical data

to understand avian response to urbanization

Daniel Steven Cooper

Doctor of Philosophy in Biology

University of California, Los Angeles, 2020

Professor Daniel T. Blumstein, Co-Chair

Professor Pamela J. Yeh, Co-Chair

As the world urbanizes, wildlife species will be forced to adapt to changed environments to survive. Despite the variety of development patterns and urban design of cities, species must overcome the loss of open space and transformation of natural vegetation to survive. I investigated spatiotemporal patterns of bird distribution to reveal ways urban wildlife communities assemble and persist. I first explored five decades of changes in nest sites relative to ornamental tree usage and urban land cover in a raptor community near Los Angeles. I showed that nest site re-use varied by species over time, and that nest substrate choice has shifted from largely native to strongly non-native. The amount of urban cover around most species’ nest sites has increased, but Cooper’s Hawk nest sites have become more urban than expected. I then expanded the study area to include the entire Los Angeles Basin, assembling occurrence data for

ii more than 50 species from the 1990s to present. Using phylogenetically-informed models, I identified two ecological traits that were significantly associated with occurrence in urban areas then and now: the tendency to nest on artificial structures (positively), and the tendency to use natural cavities (negatively). Phylogenetic relatedness was uncorrelated with urban occurrence, suggesting that a variety of birds – not just those in a few families or genera –persist or reinvade as the landscape urbanizes. Finally, I expanded my analysis to the world’s hawks (Accipitridae), using community-science records from 62 cities. Modeling three urban occurrence indices with life history traits, I found each index negatively associated with body mass, and found positive associations with both nest substrate and habitat breath. Again, I found no evidence of phylogenetic signal in the models, which suggests that although urban hawks tend to be smaller- bodied generalists, multiple lineages may succeed in cities. All three analyses provide examples of the potential of community science data, including historical datasets, to reveal patterns that might not be apparent from single studies in particular geographical areas. They also illustrate the persistence of ecological traits associated with urbanization, even as natural communities vary around the world and change over time.

iii The dissertation of Daniel Steven Cooper is approved.

Thomas W. Gillespie

Thomas B. Smith

Daniel T. Blumstein, Committee Co-Chair

Pamela J. Yeh, Committee Co-Chair

University of California, Los Angeles

2020

Table of Contents

Acknowledgments ...... vi Vita ...... vii Chapter 1. Introduction ...... 1 Chapter 2. Tolerance and avoidance of urban cover in a southern California suburban raptor community over five decades...... 18 Chapter 3. Temporally separated data sets reveal similar traits of birds persisting in a United States Megacity ...... 28 Chapter 4. Is there a “global urban raptor”? Using community science data to identify traits associated with urban occurrence in Accipiteridae ...... 40 Table 1. Three urban index variables used to calculate raptor occurrence in urban areas...... 47 Figure 1. The relationship between three urban occurrence variables (urban abundance, species proportion, and urban preference) using 90 focal raptor species across 62 cities...... 49 Table 2. Traits used in analysis...... 51 Table 3. Comparison of phylogenetically signal of each of the urban indices using three modes of evolution (Brownian, OU, Pagel’s lambda) and one non-phylogenetically informed model...... 54 Table 4. Comparison of AIC scores of four models used to test three urban indices against six traits...... 55 Table 5. Results from the best model for each urban index using Generalized Least Squares tests, fitted to explain variation based on six traits analyzed...... 56 Supplemental Information ...... 63

v Acknowledgments

I wish to thank my family for supporting my “side job” as a UCLA grad student, and my graduate committee (Dr. Daniel Blumstein, Dr. Pamela Yeh, Dr. Thomas Smith and Dr. Thomas

Gillespie) for their encouragement and support through this process. Dr. Daniel T. Blumstein and the members of the Blumstein lab were particularly helpful, at every stage from orientation until now. I also wish to thank two UCLA faculty members, Drs. H. Bradley Shaffer and Ryan J.

Harrigan, and Tessa Villaseñor, Graduate Student Affairs Officer, for their support and assistance throughout.

Please refer to the Acknowledgments in Chapters 2 and 3 (reprints of published articles). This research was conducted and published at the direction of my Committee Co-Chair, Daniel T.

Blumstein.

For Chapter 4, I wish to acknowledge Daniel T. Blumstein, Allison J. Shultz, and Cagan

Sekerciglou for their assistance. Rachel Blakey and Adam Pingatore assisted with data analysis, and Thomas B. Smith provided helpful comments on an earlier draft.

Funding for my Ph.D. work at UCLA was provided in part by the Sustainable Los Angeles

Grand Challenge grant from UCLA, and I am particularly grateful to Mark Gold for seeing the value of this research.

vi Vita

EDUCATION

University of California, Los Angeles, PhD candidate (Biology) University of California, Riverside, MSc 1999 (Biogeography) Harvard University, AB 1995 (Biology)

SELECTED RESEARCH GRANTS

University of California, Los Angeles, “Grand Challenge”. Management tools to promote the coexistence of L.A.’s nesting raptors and herons. (2017-2019. $56,000). Co-PI with Dr. Daniel Blumstein and Dr. Pamela Yeh, UCLA.

California Department of Parks and Recreation, Off-Highway Motor Vehicle Recreation Division. An assessment of the current and historic breeding status of the Golden Eagle and other raptors within the Jawbone-Butterbredt Area of Critical Environmental Concern. $65,000; 2018/2019 An assessment of the breeding status of the Golden Eagle and other raptors on BLM lands within a portion of the southern Sierra Mountains with additional emphasis on determination of the distribution of the LeConte’s Thrasher at lower elevations within the study area. $196,000. Co-PI with Bill Haas, Pacific Coast Conservation Alliance.

TEACHING/RESEARCH POSITIONS

Research Associate, Natural History Museum of Los Angeles County, Department of Ornithology, 2020-present. Visiting Researcher; Harvard Forest, Petersham, MA. Breeding bird survey of forest property in 2011 and 2013; butterfly surveys in 2015 and 2016. Thesis Advisor; Master’s Students at California State University, Los Angeles, and Oregon State University (2015-present). Teaching Assistant; University of California, Los Angeles and Univ. of California, Riverside. EEB 114 (Ornithology) and Geography 113 (Humid Tropics), 2019; courses in Geomorphology, Natural Disasters, and Astronomy, 1998-1999. Fellow; Center for Urban Resilience, Loyola Marymount Univ., Westchester, CA. Co-taught BIO 398, a field biology course (with Dr. Eric Strauss); informally advised graduate students, 2012-2015. Instructor; UCLA Extension School, Los Angeles, CA. Developed and taught courses on conservation biology and bird monitoring, organized overnight field trips, 2001 - 2003.

WORK HISTORY

Cooper Ecological Monitoring, Inc. Los Angeles, CA. 2005 - present President. An independent ecological consulting firm specializing in land use, wildlife and biodiversity issues, we provide expertise in study design & analysis, ecological assessment, and management recommendations.

vii Puente Hills Landfill Native Habitat Preservation Authority, Whittier, CA. 2007 - 2008 Biologist. Managed $2M of restoration contracts in coastal sage scrub, oak/walnut woodland, and riparian habitats in western Puente Hills; also developed and reviewed plant palettes and restoration design, and oversaw bio-monitoring of restoration sites

National Audubon Society Los Angeles, CA. Director of Bird Conservation, Audubon California (2001 – 2005; Biologist, Audubon Center at Debs Park (1999 – 2001).

RECENT PEER-REVIEWED ARTICLES

Shuford, W.D., K.C. Molina, J.P. Kelly, T.E. Condeso, D.S. Cooper and D. Johgsomjit. Breeding status of the double-crested cormorant in the interior of California, 2009-2012. Western Birds (in press). Shuford, W.D., J.P, Kelly, T.E. Condeso, D.S. Cooper, K.C. Molina, and D. Jongsomjit. Distribution and abundance of colonial-nesting herons, egrets, and night-herons in California, 2009–2012. Western Birds (in press). Uchida, K., R.V. Blackey, J.R. Burger, D.S. Cooper, C.A. Niesner, and D.T. Blumstein. 2020. Opinion: Urban biodiversity and the importance of scale. Trends in Ecol. and Evol. Published online 6 November 2020. DeMarco, C., D.S. Cooper, E. Torres, A. Muchlinski and A. Aguilar. 2020. Effects of urbanization on population genetic structure of western gray squirrels. Conservation Genetics. Published online November 2020. Cooper, D.S., P.J. Yeh, and D.T. Blumstein. 2020. Tolerance and avoidance of urban cover in a southern California suburban raptor community over five decades. Urban Ecosystems. July 2020. Cooper, D.S., A.J. Shultz, and D.T. Blumstein. 2020. Temporally separated data sets reveal similar traits of birds persisting in a United States Megacity. Frontiers in Ecol. and Evol. 8:251. Cooper, D.S. 2019. Review of Belonging on an Island: Birds, Extinction, and Evolution in Hawai’i, by Daniel Lewis. Yale University Press. Quarterly Review of Biology 94:234. Ryan, T.P., S. Vigallon, D.S. Cooper, C. Dellith, K. Johnston, and L. Nguyen. 2019. Return of beach-nesting Snowy Plovers to Los Angeles County following 68-year absence. Western Birds 50:16-25. Cooper, D.S. 2017. A flora of Griffith Park, Los Angeles, California. Crossosoma 41(1&2):1-86 Cooper, D.S., J. Mongolo, and C. Dellith. 2017. Status of the California Gnatcatcher at the northern edge of its range. Western Birds 48:124-140. Cooper, D.S. and A.E. Muchlinski. 2015. Recent decline of lowland populations of the western gray squirrel in the Los Angeles area of southern California. Bull. Southern California Acad. Sci. 114(1):42-53. Remington, S. and D.S. Cooper. 2015. Bat survey of Griffith Park, Los Angeles, California. Southwestern Naturalist 59(4):471-477.

viii Chapter 1. Introduction

One of my most vivid memories growing up birding in the Los Angeles area was finding a pair of Cooper’s Hawk building a nest in a huge sycamore along the San Gabriel River floodplain in the Whittier Narrows Recreation Area. Back in the late 1980s, this hawk was not a usual sight in the suburbs as it is today, and it was one of the few modern records from the area, which had been a National Audubon Society nature center in 1930s before the surrounding area was developed with freeways and houses, and the river channelized and “tamed”. Cooper’s Hawks were even rare a century ago, but by the 1990s, more were finding their way into residential Los

Angeles, where they were often seen feasting on the introduced Spotted Doves that were living alongside the native Mourning Doves. At some point in the 2000s, Spotted Doves essentially vanished in the region (a handful remain on Catalina Island today), and numbers of Cooper’s

Hawks exploded, to the point where they are now (2020) the most numerous urban raptor in Los

Angeles (www.ebird.org).

How did this happen? What factors allowed this raptor to thrive in the city where none had before, and what other bird species were following this trajectory? Most birders recognized that several other bird species suddenly became common in residential areas around this same time, such as Allen’s Hummingbirds, Black Phoebes and Cassin’s Kingbirds. Did they share any characteristics with Cooper’s Hawks?

Here, I explore patterns of urban bird occurrence at different spatial and temporal scales, starting with the most local (the Malibu Creek Watershed), expanding out to the entire Los

Angeles Basin, then finally focusing on the global distribution of one group (hawks) to explore even larger patterns of urban occurrence. As the world urbanizes, we must understand how

1 wildlife species respond to increased disturbance and habitat transformation. There will be clear winners (species that thrive and expand their ranges) and inevitably, losers (those that contract their ranges and become extirpated). This latter group must be the subject of conservation efforts if we are to retain a high level of global biodiversity. But my work is primarily focused on this first group – the “winners”, including synanthropes that depend on humans, and those that eke out a living in spite of us.

Background

As more people move to cities around the world, a clear understanding of species’ tolerance to urbanization will be key to conserving biotic diversity (Vitousek et al., 1997, Marzluff, 2005).

The process by which species invade and exploit novel environments has been referred to as

“filtering” (Clergeau et al., 2001), whereby certain species pass through the urban filter successfully or invading following urbanization, and others fail to do so (Lowry et al., 2013,

Wingfield et al., 2015). Johnston (2001) recognized a gradient of tolerance from urban avoidance to synanthropy, or a dependence on the built environment, and in a landmark review of urban ornithology, Marzluff et al. (2001) recommended that this tolerance be re-assessed over time, because patterns of human activity are constantly changing, with cities adopting new architectural styles and landscaping palettes. A species’ basic behavior also may change as populations become more tolerant to human disturbance (e.g., Slabbekoorn and den Boer-Visser,

2006; Francis et al., 2009). For the most sensitive species, even slight increases in human disturbance may have lasting negative consequences (e.g., from recreational activity within natural open space areas, Pauli et al., 2016), leading to loss of biodiversity over time.

2 While urbanization tends to homogenize formerly complex ecological systems

(McKinney, 2006; Devictor et al., 2007), specialists may exploit urban sites preferentially, or may assemble into novel communities where urban habitats are more structurally complex than those replaced, such as grassland or low scrub (e.g., Emlen 1974, Gonzalez-Garcia et al., 2014,

Møller et al., 2015). Food/prey and nesting sites may be superabundant in urban areas (Chace and Walsh, 2006), though this availability may be offset by novel hazards such as feral cats

(Loss et al., 2013). Thus, not all species that thrive in urban areas are drawn to hardscape or modified vegetation; some may simply maintain populations in habitat fragments within an otherwise urbanized landscape, for example marsh-dwelling birds occurring at small urban wetlands, along flood-control channels.

Urban birds tend to display behavioral boldness and “innovation propensity”, which compels individuals to explore new habitats and become established in these areas (Atwell et al.,

2012; Blumstein, 2014, and Sol et al., 2017). They have shorter flight initiation distances (FID) and exhibit heightened predator avoidance (Blumstein, 2006; Møller, 2010), heightened territoriality and aggression (Evans et al., 2010), and reduced vocalizations (Estes and Mannan,

2003). They also tend to have a broader elevational tolerance (Bonier et al., 2007) and a larger geographical range (Møller, 2009). Morphological variables have also been found to be associated with urbanization in birds, including body size (small size for raptors; Chace and

Walsh, 2006), and wingspan (large wingspan for passerines; Croci et al., 2008).

Habitat preference studies have found that urban passerines are disproportionately represented by forest species (Croci et al., 2008), and by species exhibiting a wide habitat breadth (Sol et al., 2014). Urban birds also tend to be non-migratory both globally (Sol et al.,

2014) and regionally in Europe (Croci et al., 2008) and Israel (Kark et al., 2007). Diet studies

3 have consistently found positive associations between urbanization and granivory, and negative associations between urbanization and insectivory, including for ground-foraging insectivores

(Croci et al., 2008; Kark et al., 2007; Evans, 2011; reviewed by Chace and Walsh, 2006). Urban birds tend to initiate nesting earlier than non-urban popualtions of raptors (Boal and Mannan,

1999; Kettel et al., 2018), as do species that visit (urban) feeders (O’Leary and Jones 2006). And nests of urban birds are more often located in artificial structures rather than natural cavities

(Blewett and Marzluff, 2005; Chace and Walsh, 2006),) or on the ground (Evans, 2011; Sol et al., 2014).

Still, much remains to be learned about how natural communities respond to urbanization at different scales, in order to formulate effective conservation plans and simply to make sense of the natural world as we interact with it (Uchida et al. 2020). Many studies of urban birds have found contradictory results, including those comparing urban vs. non-urban patterns of nest productivity, clutch size, nest site preference and food-provisioning to young (Chace and Walsh

(2006; see also Lowry, 2013 and Marzluff et al., 2015). Likewise, there appears to be little difference in the cognitive abilities of urban vs. rural populations of the same species, as measured by problem-solving ability and relative brain size (e.g., Sol et al., 2014; Carrete and

Tella, 2011). And, while studies of bird assemblages across gradients of urbanization (“space for time”) date to the 1970s (Emlen, 1974; Beissinger and Osborne, 1982; Blair, 1996), those that investigate the same community over time are much less common (but see Aldrich and Coffin,

1980; Shultz et al., 2012).

4 Raptors of the Malibu Creek Watershed

Around the time I was starting my Ph.D., I learned about a long-term database of nest locations of raptors (Accipitridae, Falconidae and related families) within the upper Malibu Creek watershed maintained by the National Park Service at the Santa Monica Mountains National

Recreation Area. Initially mapped through the 1970s, these locations were revisted by various biologists, including Lena Lee, a student at UCLA (Geography), who revisited the nest sites to examine the influence of distance from urban edge in nest site selection for her Master’s thesis

(Lee 2004). I decided to re-visit these same locations in 2017-18 and analyze the dataset in terms of surrounding urbanization at different scales to determine changes in nest location, nest re-use, and nest substrate (tree type). I asked two main questions: 1. How does nest site re-use and ornamental tree use relate to raptors’ persistence in urban areas over time? 2. Assuming ample nest site availability across the study area, which raptor species are using sites that are more or less urbanized than would be predicted, and has this changed over time?

I hypothesized that raptors’ acceptance of artificial nest structures, or their use of ornamental (vs. native/naturally-occurring) vegetation for nesting (see Bloom and McCrary

1996) may enable their colonization into urban areas that did not historically support these features. The earliest nests in the dataset (1970s-1980s) were all in native and naturally-occurring trees, yet by 2018, most Cooper’s Hawk nests, and many Red-tailed Hawk nests, were located in ornamental vegetation such as pines (Pinus spp.) and eucalyptus (Eucalyptus spp.). Recreating the urban footprint in the study area using historical aerial photos and overlaying the nest sites, I found that the amount of urban cover around nests increased for Red-tailed, Red-shouldered, and

Cooper’s hawks during the study period, but not for American Kestrels, which were confined to

5 the least-urban areas. Cooper’s Hawks appear to now be selecting urban nest sites over wildland sites, even where urbanization in the landscape has not substantially changed. This analysis illustrates ways in which members of a natural community can adapt to the urbanizing environment over time, which allows them to expand in the face of massive habitat transformation.

Breeding Bird Atlas vs. eBird

For my next chapter, I wanted to explore ways in which birds adapted to urbanization over time, so I expanded my taxonomic focus to include many more species (53 bird species known to breed widely in the Los Angeles area) and broadened my geographic focus to include the entire

Los Angeles basin below c. 2,000’ elevation from the foothills to the coast. I analyzed nesting- season bird records in two separate datasets covering 173 survey blocks in the Los Angeles area during two time periods, 1995–99 and 2012–16 (Los Angeles County Breeding Bird Atlas and eBird). I calculated urban land cover in each survey block to develop an index of urban association, and modeled the relationship between species occurrence and life history traits likely associated with urban tolerance, incorporating phylogenetic information into models evaluated.

I found two traits to be significantly associated with urbanization in both eras: Structure- nesting (i.e., the tendency to build nests on human-built structures) was positively associated, and cavity-nesting (i.e., the tendency to build nests in natural tree cavities) was negatively associated. This analysis illustrated the way life history traits associated with urban areas may persist, even as the makeup of these species communities may change through species turnover.

Not surprisingly, the same species I had noticed exploding in abundance in the Los Angeles area

6 since the 1990s had a very high index of urban association by the later era evaluated, and many, such as Cassin’s Kingbirds and Allen’s Hummingbirds, are frequently seen nesting in artificial structures and non-native vegetation. I did not, however, find phylogentic signal in either the index of urban association from either era evaluated, nor in the models using these indices and life history traits. This suggested that the ability of a species to exploit and become common in the urban environment is unrelated to taxonomy – basically, a wide variety of species can become common in cities, indicating the flexibility of a variety of bird species to adapt quickly to novel environments.

Global Raptors

Of course, many species that do not nest in artificial structures do just fine in cities, and some are thriving there (including Cooper’s Hawk, a tree-obligate nester that had originally sparked my interest decades ago). I wondered whether insight into the success of birds in cities could be discovered by focusing on a single clade, such as hawks (Accipitridae), and so I expanded this trait analysis to include more cities around the world. Large, diverse, and widely-distributed, hawks provide an ideal taxonomic group to study wildlife use of urban environments. One trait of hawks that immediately strikes birders as likely urban-associated is small body size. Globally, the most common urban raptors tend to be smaller species (e.g., small hawks and kites) than those in wildland areas (e.g., eagles).

Yet, small body mass has not been empirically shown to be a universal pattern in urban raptors (e.g., Croci et al. 2008, Sol et al. 2014, Santini et al. 2019), and it was not among the traits associated with a high urban index in the Los Angeles breeding bird analysis in Chapter 2.

7 Still, from Chapter 1, I observed that Cooper’s Hawks nest far more frequently than Red-tailed

Hawks in the most highly-developed sites in suburban Los Angeles, and smallish hawks like

Accipiters seem to be thriving in cities in various parts of the world. I speculated that in addition to a smaller size, life history traits of small raptors such as broad diet, non-migratory habit, widespread distribution, and/or broad habitat tolerance might be allowing them to dominate urban raptor communities at a global scale.

For chapter 3, I analyzed the status of more than 100 species of hawks (Family:

Accipitridae) using community-science records from 62 cities around the world between 2014-

2018. I modeled indices of occurrence with six ecological traits. I calculated occurrence in three ways: urban abundance, the frequency of breeding season reports within 10 km of city centers, species proportion, or the relative abundance of each species within its city of occurrence, and urban preference, a ratio of urban abundance to sighting frequency within a “peripheral band” located 10 to 100 km from each city center. Based on results from the best of four models, each index was significantly negatively associated with body mass, indicating that small size was correlated with frequency of urban occurrence. I further found that urban abundance and urban proportion were positively significantly associated with nest substrate breadth, and urban abundance was also positively associated with habitat breadth. Overall, I found that large-bodied raptors are rare or absent from urban areas, and although some smaller-bodied raptors are also rare in cities, those that are present do tend to be generalists. This study provides another example of the power of community science datasets in contributing to our understanding of urban ecology, and helped connect the local insights from Chapter 1 with patterns on a global scale (using analytic methods tested in Chapter 2).

8 Summary

All three chapters represent ways to combine historical and modern “citizen science” data to understand how bird species persist in urban areas. Increased community participation in online data-collection platforms such as eBird and iNaturalist in countries outside the US and Canada, and particularly in urban areas, will inform future analyses and allow us to better gauge conservation success in the cities of the world (i.e., where most people live; Ballard et al. 2017).

Ultimately, phylogeny appears to play a minor role in the response of birds to urbanization as compared to behavior and ecology. The diverse breeding urban avifauna we documented in the

Los Angeles area is reflected in the global hawk community, which includes representatives from multiple genera managing to infiltrate many urban areas. Replicating these studies for other taxonomic groups would also be worthwhile, to understand if certain traits are universal across even broader taxonomic groups. For example, are there intrinsic traits and extrinsic landscape features that enable certain butterflies to persist in urban areas, leaving others are confined to remote habitats? More granular diet studies and demographic research (such as reproductive success) would help fill gaps in our knowledge of urban wildlife and allow better planning for future ecological changes.

The influence of geographical location must be acknowledged in urban ecology, since a featureless desert may support relatively few bird species compared to cities lush with introduced vegetation (e.g., Gonzales-Garcia et al., 2014). Yet, this scenario would hardly be considered a desirable conservation outcome to be replicated (otherwise, why not cover the earth in cities?).

As more cities surpass the 10-million population mark and their surrounding suburbs expand into each other, they will leave precious little habitat largely undisturbed by humans. It may be that

9 high local diversity may be easier to achieve within cities than high global diversity, which requires the conservation of rare and endemic species in situ (e.g., Enedino et al. 2017,

McDonald et al. 2018).

I would also encourage further reflection on ways to define “success” in the fight for biodiversity in urban areas. On one hand, cities may be considered successful if they include built features and landscaping from around the world that support a high local diversity of species (Filazzola et al. 2018). Yet cities must also allow the least-adaptable species – those most strongly associated with wildland rather than urban habitats – to find refuge within the urban matrix as they urbanize (Sol et al., 2014). Future work could further investigate the shared characteristics of urban-avoiding species, which could aid in their conservation in areas that have not yet been developed, and might help soften the urban landscape to allow for greater levels of biodiversity. These insights could, for example, be applied to redesigning greenbelts and road edges through cities that are frequently simply lawn and ornamental landscaping, and enabling them to support features that would be utilized by wildlife species that would otherwise be unable to survive far from the wildland edge. Preserving trees with natural cavities, rather than simply installing bird houses, would likely benefit a wider range of species, as would preserving as many (natural) microhabitats as possible that would support a greater range of prey types. The close association found between increasingly urban species, such as Cooper’s and Red-tailed hawks, and non-native vegetation should (Chapter 1) could investigated for a wider range of species, a topic that is now only beginning to be explored (Esaian and Wood 2020). For example, if non-native vegetation is allowing some birds to persist in cities, could other, more sensitive taxa be enticed in by planting with natives? And does this vary with geographical location and

10 scale? In this way, an understanding of the mechanics of urban biodiversity is merely a first step on the way to developing meaningful conservation goals, anywhere in the world.

Literature Cited

Aldrich, J.W., and Coffin, R.W. (1980). Breeding bird populations from forest to suburbia after thirty-

seven years. American Birds 34, 3–7.

Atwell, J.W., Cardoso, G.C., Whittaker, D.J., Campbell-Nelson, S., Robertson, K.W., and

Ketterson, E.D. (2012). Boldness behavior and stress physiology in a novel urban

environment suggest rapid correlated evolutionary adaptation. Behav. Ecol. 23, 960–969.

doi: 10.1093/beheco/ars059)

Ballard, H.B., Robinson,, L.D., Young, A.N., Pauly, G.B., Higgins, L.M., Johnson, R.F., and

Tweddle, J.C. (2017). Contributions to conservation outcomes of natural history

museum-led citizen science: Examining evidence and next steps. Biol. Cons. 208, 87–97.

10.1016/j.biocon.2016.08.040

Beissinger, S.R., and Osborne, D.R. (1982). Effects of urbanization on avian community

organization. Condor 84, 75–83.

Blair, R.B. (1996). Land use and avian species diversity along an urban gradient. Ecol. Appl. 6,

506–519. doi: 10.2307/2269387

Blewett, C.M. and Marzluff, J.M. (2005). Effects of urban sprawl on snags and the abundance

and productivity of cavity-nesting birds. Condor 107, 678-693. doi:

10.1093/condor/107.3.678

11 Bloom, P.H. and McCrary, M.D. (1996). The Urban Buteo: Red-shouldered Hawks in Southern

California. Pages 31-39 in: Raptors in Human Landscapes. D.M. Bird, D.E. Varland, and J.J.

Negro, eds. Academic Press.

Blumstein, D.T. (2006). Developing an evolutionary ecology of fear: how life history and natural

history traits affect disturbance tolerance in birds. Anim. Behav. 71, 389-399. doi:

10.1016/j.anbehav.2005.05.010

Blumstein, D.T. (2014). “Attention, habituation, and antipredator behavior: implications for

birds,” in Avian Urban Ecology, eds D. Gil and H. Brumm (Oxford Univ. Press), 41–53.

doi: 10.1093/acprof:osobl/ 9780199661572.003.0004

Boal, C.W., and Mannan, R.W. (1999). Comparative breeding ecology of Cooper’s hawks in

urban and exurban areas of southeastern Arizona. J. Wildl. Manage. 63, 77–84.

Bonier, F., Martin, P.R., and Wingfield, J.C. (2007). Urban birds have broader environmental

tolerance. Biol. Lett. 3, 670-673. doi: 10.1098/rsbl.2007.0349

Carrete, M., and Tella, J.L. (2011). Inter-Individual variability in fear of humans and relative

brain size of the species are related to contemporary urban invasion in birds. PLoS One,

6(4):e18859. doi: 10.1371/journal.pone.0018859

Chace, J.F., and Walsh, J.J. (2006). Urban effects on native avifauna: A review. Landsc. Urb.

Plan. 74, 46-69. doi: 10.1016/j.landurbplan.2004.08.007

Clergeau, P., Jokimäki, J. and Savard, J.-P. (2001). Are urban bird communities influenced by

the bird diversity of adjacent landscapes. J. Appl. Ecol. 28, 1122-1134. doi:

10.1046/j.1365-2664.2001.00666.x

Croci, S., Butet, A., and Clergeau, P. (2008). Does urbanization filter birds on the basis of their

biological traits? Condor 110, 223-240. doi: 10.1525/cond.2008.8409

12 Devictor, V., Julliard, R., Couvet, D., Lee, A. and Jiguet, F. (2007). Functional homogenization

effect of urbanization on bird communities. Cons. Biol. 21, 741–751. doi:

10.1111/j.1523-1739.2007.00671.x

Enedino, T.R., Loures-Ribeiro, A., and Santos, B.A. (2018). Protecting biodiversity in

urbanizing regions: The role of urban reserves for the conservation of Brazilian Atlantic

Forest birds. Persp. Ecol. Cons. 16, 17-23. doi: 10.1016/ j.pecon.2017.11.001

Emlen, J.T., (1974). An urban bird community in Tucson, Arizona: derivation, structure,

regulation. Condor 76, 184–197.

Estes, W.A. and Mannan, R.W. (2003). Feeding behavior of Cooper’s Hawks at urban and rural

nests in southeastern Arizona. Condor 105, 107-116. doi: 10.1093/condor/105.1.107

Evans, J., Boudreau, K., and Hyman, J. (2010). Behavioral syndromes in urban and rural

populations of song sparrows. Ethology 116, 588-595. doi: 10.1111/j.1439-

0310.2010.01771.x

Evans, K. L., Chamberlain, D.E., Hatchwell, B.J., Gregory, R.D., and Gaston, K.J. (2011). What

makes an urban bird? Glob. Change Biol. 17, 32-44. doi: 10.1111/j.1365-

2486.2010.02247.x

Filazzola, A., Shrestha, N., and MacIvor, J.S. (2019). The contribution of constructed green

infrastructure to urban biodiversity: A synthesis and meta-analysis. J. Appl. Ecol. 56,

2131-2143. doi: 10.1111/1365-2664.13475

Francis, C.D., Ortega, C.P., and Cruz, A. (2009). Noise Pollution Changes Avian Communities

and Species Interactions. Curr. Biology 19, 1415–1419. doi: 10.1016/ j.cub.2009.06.052

Gonzalez-Garcia, F., Straub R., Lobato Garcia, J.A., and MacGregor-Fors, I. (2014). Birds of a

neotropical green city: An up-to-date review of the avifauna of the city of Xalapa with

13 additional unpublished records. Urb. Ecosyst. 17, 991-1012. doi: 10.1007/s11252-014-

0370-3

Johnston R.F. (2001). “Synanthropic birds of North America,” in Avian Ecology and

Conservation in an Urbanizing World, eds J.M. Marzluff, R. Bowman, and R. Donnelly.

(Boston, MA: Springer), 49-67. doi: 10.1007/978-1-4615-153 1-9_3

Kark, S., Iwaniuk, A., Schalimtzek, A., and Banker. E. (2007). Living in the city: can anyone

become an ‘urban exploiter’? J. Biogeogr. 34, 638–651. doi: 10.1111/j. 1365-

2699.2006.01638.x

Kettel, E.F., Gentle, L.K., Quinn, J.L., and Yarnell, R.W. (2018). The breeding performance of

raptors in urban landscapes: a review and meta-analysis. J. Ornithol. 159, 1-18. doi:

10.1007/s10336-017-1497-9

Lee, L. 2004. The distribution of breeding raptors in urban and natural areas of southern California.

MSc. Thesis. University of California, Los Angeles.

Loss, S.R., Will, T. and Marra, P.P. (2013). The impact of free ranging domestic cats on wildlife

of the United States. Nature Comm. 4, 1396. doi: 10.1038/ncomms2380

Lowry, H., Lill, A. and Wong, B.B.M. (2013). Behavioural responses of wildlife to urban

environments. Biol. Rev. 88, 537-549. doi: 10.1111/brv.12012

Marzluff, J.M. (2005). Island biogeography for an urbanizing world: how extinction and

colonization may determine biological diversity in human-dominated landscapes. Urb.

Ecosyst. 8, 157-177. doi: 10.1007/s11252-005-4378-6

Marzluff, J. M., Bowman R., and Donnely, R. (2001). “A historical perspective on urban bird

research: Trends, terms, and approaches,” in Avian Conservation and Ecology in an

14 Urbanizing World, eds J. M. Marzluff, R. Bowman, and R. Donnely (Boston, MA:

Kluwer Academic), 1–17. doi: 10.1007/978-1-4615-1531-9_1

Marzluff, J. M., Clucas, B., Oleyar, M.D., and DeLap, J. (2015). The causal response of avian

communities to suburban development: a quasi-experimental, longitudinal study. Urb.

Ecosyst. 19, 1597-1621. doi: 10.1007/s11252-015-0483-3

McDonald, R.I., Güneralp, B., Huang, C.-W., Seto, K.C., and You, M. (2018). Conservation

priorities to protect certebrate endemics from global urban expansion. Biol. Cons. 224,

290-299. doi: 10.1016/j.biocon.2018.06.010

McKinney, M.L. (2006). Urbanization as a major cause of biotic homogenization. Biol. Cons.

127, 247-260. doi: 10.1016/j.biocon.2005.09.005

Møller, A.P. (2009). Successful city dwellers: A comparative study of the ecological

characteristics of urban birds in the Western Palearctic. Oecologia 159, 849-858. doi:

10.1007/s00442-008-1259-8

Møller, A.P. (2010). Interspecific variation in fear responses predicts urbanization in birds.

Behav. Ecol. 21, 365-371. doi: 10.1093/beheco/arp199

Møller, A.P., Diaz, M., Flensted-Jensen, E., Grim, T., Ibanez-Alamo, J.D. Jokimäki, J., et al.

(2015). Urbanized birds have superior establishment success in novel environments.

Oecologia 178, 943-950. doi: 10.1007/s00442-015-3268-8

O’Leary, R. and Jones, D.N. (2006). The use of supplementary foods by Australian magpies

Gymnorhina tibicen: Implications for wildlife feeding in suburban environments. Austral

Ecol. 31, 208-216. doi: 10.1111/j.1442-9993.2006. 01583.x

15 Pauli, B.P., Spaul, R.J. and Heath, J.A. (2016). Forecasting disturbance effects on wildlife:

tolerance does not mitigate effects of increased recreation on wildlands. Anim. Cons. 20,

251-260. doi: 10.1111/acv.12308

Santini, L, González-Suárez, M, Russo, D, Gonzalez-Voyer, A, Von Hardenberg, A and

Ancillotto, L. (2019) One strategy does not fit all: determinants of urban adaptation in

mammals. Ecol. Lett. 22, 365–376. doi: 10.1111/ele. 13199

Shultz, A.J., Tingley, M.W., and Bowie, R.C.K. (2012). A century of avian community turnover

in an urban green space in northern California. Condor 114, 258-267. doi:

10.1525/cond.2012.110029

Slabbekoorn, H. and den Boer-Visser, A. (2006). Cities change the songs of birds. Curr. Biol. 16,

2326–2331. doi: 10.1016/j.cub.2006.10.008

Sol, D., Gonzalez-Lagos, C., Lapiedra, O., and Diaz, M. (2017). “Why are exotic birds so

successful in urbanized environments?” in Ecology and Conservation of Birds in Urban

Environments, eds E. Murgui and M. Hedblom (Boston, MA: Springer), 75-89. doi:

10.1007/978-3-319-43 314-1_5

Sol, D., González-Lagos, C., Moreira, D., Maspons, J., and Lapiedra, O. (2014). Urbanisation

tolerance and the loss of avian diversity. Ecol. Lett. 17, 942-950. doi: 10.1111/ele.12297

Uchida, K., Blackey, R.V., Burger, J.R., Cooper, D.S., Niesner, C.A. and Blumstein, D.T.

(2020). Urban biodiversity and the importance of scale. Trends. Ecol. Evol. Published

online 6 Nov. 2020. doi: 10.1016/j.tree.2020.10.011.

Vitousek, P.M., Mooney, H.A. Lubchenco, J., and Melillo, J.M. (1997). Human domination of

Earth’s ecosystems. Science 277, 494–499. doi: 10.1126/science. 277.5325.494

16 Wingfield, J.C., Krause, J.S., Perez, J.H., Chmura, H.E., Németh, Z., Word, K.R., et al. (2015). A

mechanistic approach to understanding range shifts in a changing world: What makes a pioneer?

Gen. Comp. Endocrinol. 222, 44-53. doi: 10.1016/j.ygcen.2015.08.022

Wood, E.M. and Esaian, S. (2020). The importance of street trees to urban avifauna. Ecological

Applications. 30, e02149. doi: 10.1002/eap.2149

17 Chapter 2. Tolerance and avoidance of urban cover in a southern California suburban raptor community over five decades.

Urban Ecosystems https://doi.org/10.1007/s11252-020-01035-w

Tolerance and avoidance of urban cover in a southern California suburban raptor community over five decades

Daniel S. Cooper1 & Pamela J. Yeh1,2 & Daniel T. Blumstein1

# Springer Science+Business Media, LLC, part of Springer Nature 2020

Abstract We explored nest site placement and re-use relative to ornamental tree usage and urbanization level in a diurnal raptor community in southern California (USA) during three discrete time periods spanning five decades (1971–2018). Re-use of prior years’ nests varied among species, with Red-tailed Hawks (Buteo jamaicensis) and American Kestrels (Falco sparverius) showing moderate re-use rates (ca. 30%), and Red-shouldered Hawks (Buteo lineatus), and Cooper’s Hawks (Accipiter cooperii) showing almost none. Nearly all nests were in native and naturally-occurring trees during the 1970s, yet by 2018, most Cooper’s Hawk nests, and many Red-tailed Hawk nests, were located in ornamental vegetation such as pines (Pinus spp.) and eucalyptus (Eucalyptus spp.). The amount of urban cover around nest sites increased for Red-tailed, Red-shouldered, and Cooper’s hawks during the study period, but not for American Kestrels, which were confined to the least-urban areas. Cooper’s Hawks appear to now be selecting urban nest sites over wildland sites, based on the increase in surrounding urban cover, even as landscape urbanization has not substantially changed in the study area during the last two time periods. Our study illustrates the utility of long-term datasets in understanding how a species’ urban tolerance can change over time, and highlights species (including three extirpated taxa) that may be failing to adapt to local urbanization.

Keywords Red-tailedhawk .Red-shoulderedhawk .Cooper’shawk .Americankestrel .Urbanization .Urbantolerance .Change over time . Re-use

Introduction recognized “urban avoiders”, “suburban adaptable” taxa, and “urban exploiters”, which represent a gradient of tolerance Urban areas have been expanding in extent twice as fast as from outright urban avoidance to synanthropy, a strong de- population increases (Seto et al. 2011), and understanding pendence of the built environment (Johnston 2001). urban tolerance in species is crucial to conserving the Research on urban birds must be re-assessed over time, Earth’s biotic diversity (Vitousek et al. 1997; Marzluff 2005; since ecological forces (and human activity) are in constant Sol et al. 2014). Even as urbanization homogenizes complex flux (Marzluff et al. 2001; Marzluff 2016), and because cities ecosystems at the global scale (McKinney 2006; Devictor are constantly evolving new architectural styles and landscap- et al. 2007), certain species exploit urban sites, resulting in ing palettes. Data on bird species distribution and habitat us- novel communities (Møller et al. 2015). Blair (1996) age prior to a period of environmental change can provide an important baseline to compare with contemporary observations (e.g., Tingley and Beissinger 2009). Electronic supplementary material The online version of this article We examined the response ofraptors(Accipitridae, (https://doi.org/10.1007/s11252-020-01035-w) contains supplementary material, which is available to authorized users. Falconidae and related families) to urbanization over a nearly fifty-year period by leveraging historical data to track changes * Daniel S. Cooper in nest location, nest re-use, and nest substrate (tree type) [email protected] within the upper Malibu Creek watershed in southern California, USA. Raptors provide ideal subjects to assess tol- 1 Department of Ecology and Evolutionary Biology, University of erance of urban environments, being apex predators with read- California, Los Angeles, 621 Charles E. Young Drive South, Los ily detectable nests often re-used for years. They display a Angeles, CA 90095-1606, USA broad range of urban tolerance, with certain species such as 2 Santa Fe Institute, Santa Fe, NM, USA Peregrine Falcons (Falco peregrinus) drawn to urban habitats

18 Urban Ecosyst

(e.g., Cade et al. 1996), and closely related species (e.g., and a record of raptor nests mapped and documented since the Prairie Falcon Falco mexicanus) avoiding them (Steenhof early 1970s. We restricted our fieldwork to the Upper Malibu 2013). Raptor nest site choice may be related to the disposition Creek Watershed, which covers ca. 13,000 ha in western Los and outcome of the prior year’s nest (Jiménez-Franco et al. Angeles County and adjacent Ventura County (Fig. 1). During 2014), the availability of resources (Kreiderits et al. 2016), the this period, the study area was transformed from a mostly rural, presence of conspecifics and competitors (Sumasgunter et al. ranching landscape of grassland with scattered oaks and small 2016), human disturbance (Richardson and Miller 1997), subdivisions of tract homes, to a modern one of protected open weather during nesting season (Rockweit et al. 2012), and space interspersed by large expanses of homes. Since the 1970s, the use of rodenticide by humans in the area (Rattner et al. avast“urban forest” has developed and matured across the Los 2011). In wildland areas, territory persistence has been shown Angeles region, featuring large ornamental trees from around the to affect reproductive output (higher productivity in newly- world (Gillespie et al. 2011). These trees now support a diverse established nests on existing territories; Jiménez-Franco avifauna year-round, including woodland species that might not et al. 2014), and nest re-use may be correlated with nest pre- have been present until a few decades ago (see Wood and Esaian dation (higher in re-used nests; Otterbeck et al. 2019). 2020). However, data on nest re-use in urban areas are sparse, and No ranching exists today in the study area, though cattle relatively few studies have investigated the ability of birds and sheep ranching was prevalent prior to the 1990s, and land (including raptors) to persist within urban landscapes over ownership now includes federal, state and local park/open multiple decades (see Marzluff et al. 2001). space agencies, and private property. A development boom We asked two main questions: 1. Assuming ample nest site occurred much later here than in the San Fernando Valley and availability across the study area, which raptor species are central Los Angeles Basin to the east, with the population of using sites that are more or less urbanized than would be Agoura Hills roughly doubling from 1980 (11,399) to 1990 predicted, and has this changed over time? 2. How does nest (20,390), and remaining roughly constant since then (Los site re-use and ornamental tree use relate to raptors’ persis- Angeles Almanac 2018). Elevation within the study area tence in urban areas over time? Because we did not mark and ranges from 185 to 730 m.a.s.l., and the climate is track individual raptors over time, we infer the use of these Mediterranean, with April temperatures with average low of strategies by analyzing historical and current land cover and 9 °C (record 3 °C) and average high of 24 °C (record 38 °C) nest location and re-use data, rather than by directly measuring (“Agoura Hills, CA”; www.myweather2.com). Rainfall is the movements of the pairs themselves. From a conservation highly variable year-to-year, with the average since the late perspective, we suggest that urban-avoiding species – those 1970s being ca. 43 cm/year with nearly all precipitation falling using urbanized sites at a lower rate than would be expected – in winter. are of higher conservation concern than urban-tolerant ones, Please refer to Supplemental Materials for historical and with urban-preferring species being of least conservation con- modern photographs of the study area, and examples of raptor cern (while recognizing that certain urban-preferring species nest sites. may still have specific and often unique ecological requirements, particularly when nesting). We further suggest that Focal species raptors’ acceptance of artificial nest structures, and use of ornamental (vs. native/naturally-occurring) vegetation for Nine raptor species bred regularly in the Santa Monica nesting (see Bloom and McCrary 1996) may enable coloniza- Mountains into the 1980s (Garrett and Dunn 1981), includ- tion into urban areas that did not historically support these ing several owls, which we dropped from the analysis due features. Understanding changes in the pattern of nest site to the difficulty of locating nests. We also dropped three placement and re-use through time should provide insight into species from our urban cover analysis that have long oc- how species may respond in the future to a landscape that is curred in very low densities (<5 pairs/year) in the entire less wild, and more disturbed by humans. Santa Monica Mountains, and that are considered extirpated within the study area such that an analysis of their urban tolerance and preferences is not possible: White-tailed Methods Kites (Elanus leucurus), Golden Eagles (Aquila chrysaetos), and Prairie Falcons (Falco mexicanus) Study area (Willet 1912,Allenetal.2016, www.ebird.org). Thus, we focused on the four most common, widespread and We selected a coastal southern California study area that features extant diurnal raptor species, Red-tailed Hawks (Buteo human-modified (urban) land interspersed with large protected jamaicensis), Red-shouldered Hawks (Buteo lineatus), areas of open space, a history of ornithological investigation in Cooper’sHawks(Accipiter cooperii)andAmerican the region dating to the nineteenth Century (e.g., Grinnell 1898), Kestrels (Falco sparverius).

19 Urban Ecosyst

Fig. 1 Study area. The study area was established based on the historical −118.8228); Mulholland Hwy. and No. Kanan Dume Rd. (34.0965, distribution of nests monitored in prior studies (Lee 2004; NPS, unpubl. −118.8123); Mulholland Hwy. and Stunt Rd. (34.1020, −118.6600), data). It is roughly bounded by: Kanan Rd. and Westlake Blvd (34.1970, and the western terminus of Victory Blvd. (34.1850, −118.6685)

Study period effort, but reported revisiting all mapped historical nest sites, and conducting extensive field visits to both urbanized and We obtained observational data from the study area from three natural open space across the study area. Her methods gener- discrete time periods, which we refer to as “eras”: ally followed recommendations by Craighead and Craighead Jr. (1969), in that likely raptor nesting areas were visited on 1. Early. Opportunistic nest-mapping of raptors in the Santa foot, and all trees/substrate scanned with binoculars. Monica Mountains and Simi Hills (which includes our During the spring/summer of 2017 and 2018, we attempted entire study area), beginning during the preparation of to replicate Lee (2004)bysearchingforandmapping planning documents to support the Santa Monica previously-reported raptor nests in the study area, visiting each Mountains National Recreation Area in the 1970s, and reported nest location to confirm re-use, and carefully searching continuing as the parkland was purchased and protected the vicinity of each nest on at least two days during the breeding in the 1980s (NPS, unpubl. data). Nest records from this season, scanning in all directions from the original nest site, and era within the study era extend from 1971 to 1986 (medi- as necessary, from vantage pointsnearbywithbettersight-lines an year = 1979). to determine the currentnestingstatusofallraptorsinthearea. 2. Middle. Systematic and comprehensive mapping of rap- We conducted surveys in 2017–18 only from public roads/trails tor nesting sites (except American Kestrel) by Lee (2004) (following Lee 2004). One author (DSC) lives near the center of in the Simi Hills and north-central Santa Monica the study area and submitted 54 eBird checklists from days Mountains during 2002 and 2003, centered on and afield the study area (exclusive of home) during February – encompassing the study area; July 2018 (and 15 during the same timeframe in 2017), and 3. Late. Re-surveys of prior raptor nest locations in the study two interns devoted portions of 29 field days to nest-searching area during 2017–2018, with comprehensive nest- and monitoring here between 2 April and 1 June 2018. To searching throughout the study area by DSC and field augment our observations, we searched submissions of focal assistants (this study). species within the study area in online databases (www. iNaturalist.org, www.eBird.org) throughout 2017 and 2018, No survey effort data nor methods used for nest-searching and attempted to track down reports of paired birds (and nests) exist for the early era. Lee (2004) did not report hourly/daily in the field during this time. While Lee (2004)providedfledging

20 Urban Ecosyst information for all nests, we obtained these data for a portion of However, two western sycamores (or hybrids between syca- nests, and do not analyze it here. We re-plotted all nest locations more and London plane, Platanus x. acerifolia) that held nests in 2017 and 2018 using reported coordinates or those derived in the late era were almost certainly planted as ornamental from the iPhone 7 map application (OS v. 12.1.2), confirmed landscaping, located in residential tracts well away from ripar- using satellite imagery in Google Earth Pro, and photographed ian zones, and so we treated these as non-native. each nest site in situ. Defining “urban habitat” Nest assessment, including re-use Informed by prior analyses (e.g., Dykstra 2018;Whiteetal. We considered a nest “active” if it appeared occupied during 2018), we used percentage of urbanized land surrounding each the nesting season, being structurally sound (fresh material nest as our unit of comparison, which served as a surrogate for used) with at least one adult bird performing nesting activity various urban-associated features. For nests of the middle and at the nest (typically nest-building, incubating, or tending late eras, we used publicly availableshapefilesofstatewideland- young; Fuller and Mosher 1987). Lee (2004) relied on nest cover data developed in the 2000s (CALVEG 2009), because appearance rather than presence of birds to determine activity, relatively little new urbanization had occurred in the study area and we have inconsistent data on how early-era nests were between the middle and late eras (confirmed by Google Earth determined to be active, though many noted the number of Pro). This coverage was comprised of the “Urban or Developed” young produced. We occasionally located nests by the pres- category, defined as “landscapes that are dominated by urban ence of nearby nestlings, and counted these as active only if structures, residential units, or other developed land use elements the young appeared not to be capable of sustained flight. We such as highways, city parks, cemeteries and the like” (https:// did not revisit nests to document fledgling success in 2017– www.fs.usda.gov/Internet/FSE_DOCUMENTS/fsbdev3_ 18, so our late era data should be considered an analysis of 045405.pdf). We combined all other land cover categories that “breeding events” (per Jiménez-Franco et al. 2014), rather were not “Urban and Developed” to create a map of just two than necessarily successful breeding. categories, urban (human-modified) and wildland (largely natu- To find new nests, we checked large stick structures on trees, ral). For early era nests, we drew an urban boundary for the study transmission towers, and rock outcrops that appeared to be inac- area from historical aerial photographs from the mid-1970s tive raptor nests at least two times during the 2017–18 breeding (UCSB Library 2018), using Google Earth Pro to overlay these season, but we only analyzed these further if we detected breed- image files atop modern imagery. ing activity; otherwise we dropped them from the analysis (or in We used QGIS (QGIS Development Team 2018) to calcu- the case of formerly active nests, considered them inactive). Due late percentage of urban (vs. wildland) landcover at radial to the difficulty in documenting occupancy of cavity nests of distances from all nests, using two distance scales, 250-m, American Kestrels, we assumed that a potential nest cavity was which was suggested by White et al. (2018) as approximating active if we observed a kestrel pair near the cavity, and at least the “macrohabitat” of raptor nests in their study of a raptor one adult entering the cavity, during the breeding season (April– community in Reno, and 670-m (“nearest nest” per White June). We pooled nest locations within each era in an effort to et al. 2018), which approximates the midpoint between two minimize inter-annual variation that might occur due to excep- adjoining nests, recognizing that territory size among species tional weather conditions in a particular year. and pairs is highly variable and difficult to estimate. We did not analyze year-to-year occupancy (due to incom- Specifically, we used percent urban cover for nests in the plete data), so cannot say with certainty what constituted a early era using the shapefile of urban cover from the 1970s, “new” nest occupancy event (per Jiménez-Franco et al. and used the CALVEG (2009) shapefile for nests from both 2014). And, we did not devote enough observation time to the middle and late eras. define territorial boundaries of nesting pairs of our focal species (nor was this reported for the historical nests). Thus, our Urban tolerance vs. urban preference nest re-use categories include total number of active nesting sites during each era, and number of nesting sites (nest We recognized two main strategies used by nesting raptors, “ur- matching the reported coordinates and tree type) reoccupied ban tolerating” and “urban avoiding”,whichmaybeemployed from either prior era, for each species. This differentiated pairs by individual pairs, as well as by species. An urban-tolerant pair that have re-nested in the same site from those that selected could either remain at the same nest site in or near urban cover new nest sites. It did not, however, differentiate pairs that year after year even as urbanization expands, or it might shift its selected new nesting sites within existing territories. nest site to maintain a similar level of urbanization within its We identified to genus, and if possible, species, the trees in breeding territory. Pairs may also shift nesting sites toward more which nests were built. In nearly all cases, planted, ornamental urbanized habitat, if that habitat provides them with resources trees were non-native, and naturally-occurring trees native. they cannot find in wildland habitat. We apply the term “urban

21 Urban Ecosyst preferring” to this latter scenario, referring to species whose nest Because Lee (2004)didnotrecordAmericanKestrelnests,we sites appear to have shifted towards urban cover over time, uti- used a Wilcoxon rank sum test with continuity correction for the lizing urban habitat (including ornamental vegetation) at a higher two samples of American Kestrel nests (early vs. late era). rate than would be expected given the background level of ur- To test for “preference” in nest site choice, we used 100 banization across the study area.Anurbanavoidingpairwould randomly plotted points within the study area (using the ran- either maintain nest sites far from urbanization over time, or dom point generator in QGIS), and compared these to our would move nest sites even farther away from urbanization, observed nests, for each era. For this comparison, we again and would be occurring in urban habitats at a lower rate than calculated percent urban cover, but instead used these random would be expected given unlimited nesting options. points with 250-m and 670-m buffer distances, resulting in While urban tolerance implies some level of acceptable ur- two sets of means (observed vs. random) for each species, banization around a nest site, “urban preference” is nuanced and for each of the three eras. We then used Wilcoxon rank sum more difficult to assess, and requires that we show that species tests to test for significance in the differences of mean urban selected territories at a higher rate than would be expected by cover between observed vs. random points. As with the ob- chance. This is essentially impossible to determine with certain- served nests, we calculated percent urban cover for nests in the ty without knowing the distribution of suitable nest sites (e.g., early era using a separate shapefile of urban cover from the the number and distribution of potentially suitable nest trees and 1970s, and used the CALVEG (2009) shapefile for tests on territories in the study area) and by tracking marked birds. We data from the middle and late eras. addressed this indirectly, in two ways. First, we used the same We evaluated whether the proportion of native vs. non- urban cover shapefile for the middle and late eras since urban native nest trees changed over time for each species using a development has been limited since 2003. Thus, any increase in chi-squared test for all three eras. urban cover around nests between these two eras would have to be from a pair moving its territory to a more urban neighbor- hood. We then used a random point design to compare mean urban cover around observed vs. randomly-plotted points across Results the study area. While some of these random points themselves might be unsuitable for nesting, we assumed the surrounding Nest re-use territory (at 250 and 670 m) would support at least one nest tree. Notably, because we observed local raptor nest sites in such a Nest re-use within the study area over time varied greatly wide variety of locations, including backyards, freeway among species (Table 1). While Red-tailed Hawk nest re-use offramps, school parking lots, marinas, etc., we felt that 100 appears to have been low between the early and middle era randomly-placed territories (at twodistancecalculations)would (3.3%), it jumped to 25.7% by the late era (9 of 35 late-era capture a range of potential nest sites. nests were re-used from at least one prior era). Of 21 Red- shouldered Hawk nest sites, we found none active in more Data analysis than one prior era. Of 25 Cooper’s Hawk nest sites monitored, we found just two active in prior eras, and no late-era Cooper’s We used R (ver. 1.0.153, R Core Team 2017)toperforma Hawk nests had been active in a prior era. We found three of Kruskal-Wallis test on mean urban cover for 250-m and 670-m 11 late-era American Kestrel territories active in both the early buffers around nests, across all three eras, for each hawk species. and late eras.

Table 1 Patterns of nest site re-use, by era. We considered late-era nest sites re-used if active in either the early or the middle eras. We pooled data from multiple years within each era to determine the total nest sites

Species Total Active Early Total Active Middle Re-used Total Active Late Re-used (Middle) (Late)

Red-tailed Hawk 30 39 1 35 9 Red-shouldered Hawk 4 8 0 9 0 Cooper’s Hawk 10 11 2 6 0 American Kestrel 7 N/A N/A 11 3 White-tailed Kite 2 4 0 0 0 Golden Eagle 3 0 0 0 0 Prairie Falcon 1 0 0 0 0

22 Urban Ecosyst

Native vs. non-native tree use Table 2 Nest sites of focal raptor species, by era (where known). Asterisks indicate non-native tree species. Note that we considered certain sycamores (Platanus sp.), alder (Alnus sp.) and cottonwoods (Populus Nearly all raptor nests within the Upper Malibu Creek sp.) non-native if they were obviously planted as part of urban landscap- Watershed since the 1970s have been in trees, though the tree ing, or were not clearly native forms type (where known) has changed markedly in recent decades, Red-tailed Hawk Early Middle Late and among the four focal species (Table 2). Red-tailed Hawk nests were overwhelmingly in (native) oaks during the early Cliff 1 0 0 era (coast live oak Quercus agrifolia and valley oak Tower 2 3 0 Q. lobata). Significantly more nests were in planted/non- Oak (Quercus) sp. 23 26 15 native trees (especially eucalyptus Eucalyptus and related spe- Sycamore (Platanus racemosa)11 1 2 cies, and pines Pinus spp.) by the middle era (X = 4.75, df = Cottonwood (Populus fremontii)* 0 0 1 1, p = 0.029), and by the late era, more than half of all Red- Eucalyptus (Eucalyptus) sp.* 1 6 11 tailed Hawk nests were placed in non-native trees, providing Pine (Pinus) sp.* 0 3 7 2 an even greater contrast with the early era (X = 17.37, df = 1, Alder (Alnus) sp.* 0 0 1 p <0.001). A handful of early and middle-era Red-tailed Native/Non-native 24/1 27/9 16/20 Hawk nests were located in cliffs and transmission towers, Red-shouldered Hawk though we documented no active nests in either of these sub- Oak (Quercus) sp. 1 4 0 strates during the late era. Sycamore (Platanus racemosa)1 0 8 Every Cooper’s Hawk nest during the early era was found Cottonwood (Populus fremontii)0 1 0 in a native coast live oak, and native willows (Salix spp.) were Eucalyptus (Eucalyptus) sp.* 1 2 1 used along with native oaks by this species during the middle Native/Non-native 2/1 5/2 6/3 era. However, by the most recent era, five of six Cooper’s Cooper’s Hawk Hawk nest trees were non-native species. Nest substrate Coast live oak (Quercus agrifolia) 10 7 1 choice by Red-shouldered Hawks appears to be skewed to- Sycamore (Platanus racemosa)0 0 2 ward natives, particularly sycamores. All American Kestrel Cottonwood (Populus fremontii)0 0 1 nesting sites during both the early and late era were in native Willow (Salix) sp. 0 4 0 trees, including oaks and sycamore. Shamel ash (Fraxinus udhei)* 0 0 1 Among the extirpated species, two White-tailed Kite nests active in the early era were both in native oaks (presumably Pine (Pinus) sp.* 0 0 1 coast live oak), and all four White-tailed Kite nests active in Native/Non-native 10/0 11/0 1/5 the middle era were also in the native coast live oak. Up to American Kestrel three Golden Eagle territories were noted (to 1993), two in Oak (Quercus sp.) 5 n/a 9 remote cliffs and one on a transmission tower within extensive Sycamore (Platanus racemosa) 2 n/a 2 oak savannah (NPS data, unpubl.). The single Prairie Falcon Native/Non-native 7/0 n/a 11/0 territory was high on a rocky outcrop along the northern edge White-tailed Kite of the study area, active only in the early era. Coast live oak (Quercus agrifolia)2 4 0 Native/Non-native 2/0 4/0 0/0 Change in urban cover over time

The mean percent urban cover surrounding each raptor nest distances), though in different directions. Late era Red- was significantly higher by the late era for the three hawk shouldered and Cooper’s hawks were more urban than would species (Red-tailed Hawks, Red-shouldered Hawks, and be predicted by random points (Fig. 2b and Fig. 2c). American Cooper’s Hawks) at both the 250 and the 670 buffer distance. Kestrels showed a change in the opposite direction, nesting in American Kestrels, had lower mean urban cover values during less urbanized sites that would be expected by random points the late era than in the early era, though we did not find these (Fig. 2d). values to be significantly different (Table 3). Comparing mean urban cover around observed nests vs. those of randomly-plotted points, we found no significant dif- Discussion ference between observed and random nest sites for Red-tailed Hawks at either the 250 m or 670 m buffer distance, across Our focal species had contrasting responses in nest site place- each era examined (Fig. 2a). Urban cover around Red- ment as the study area urbanized over time, with territories of shouldered Hawks, Cooper’s Hawks and American Kestrels Red-tailed Hawks, Red-shouldered Hawks, and Cooper’s nest sites differed significantly across eras (at both buffer Hawks increasing their urban cover, and American Kestrels

23 Urban Ecosyst

Table 3 Changes in mean percent urban cover surrounding each rank sum test used for American Kestrel (two samples; data from middle species’ nest, by era. “N” refers to the total number of active nests (see era not collected), and Kruskal-Wallis test used for the other raptor spe- text) observed in the study area during that era. Note that data for cies (df = 2 for each) American Kestrels during the middle era were not collected. Wilcoxon

250 m 670 m

Species Era N mean sd mean sd

Red-tailed Hawk Early 30 2.90 8.37 5.85 8.63 Middle 39 19.82 30.46 28.45 26.47 Late 35 27.93 27.72 23.36 26.90 3-era comparison X2 = 23.31 P < 0.001 X2 = 18.80 P < 0.001 Red- shouldered Hawk Early 4 28.08 28.92 11.44 10.02 Middle 8 6.56 13.02 9.89 12.95 Late 9 66.77 32.11 69.68 20.14 3-era comparison X2 = 11.56 P = 0.003 X2 = 14.28 P < 0.001 Cooper’s Hawk Early 10 0.33 1.05 1.32 4.03 Middle 11 29.59 39.93 22.51 26.01 Late 6 79.02 13.71 58.33 15.43 3-era comparison X2 = 15.13 P < 0.001 X2 = 19.01 P < 0.001 American Kestrel Early 7 11.30 27.57 10.23 19.48 Late 9 3.16 4.49 8.52 10.66 2-era comparison W = 28.5 P = 0.761 W = 25 P = 0.519

showing no significant change over time. By analyzing nest Local Red-shouldered Hawks showed even less nest site placement as urbanization remained similar (i.e., middle vs. fidelity sites than Cooper’s Hawks, and may be responding to late era), we found that Cooper’s and Red-shouldered hawk factors not associated with our simple urban/wildland dichot- nests were much more urban, a bias that was confirmed by omy, such as the presence of riparian corridors; of 29 nests in comparing randomly-placed points to observed nests (such a Orange County, Wiley (1975:136) found that this species bias was not found for Red-tailed Hawks). We found the op- “nested close to permanent or seasonal water, with no nest posite pattern with American Kestrels, which are now trees found farther than 23 m from a creek bed.” Thus, Red- selecting significantly less-urban territories than would be ex- shouldered Hawks may use a wide range of urbanization pro- pected. This suggests that a species’ urban tolerance can vided these features (and suitable nest trees) are present. change over time, either as the landscape becomes dramatical- Red-tailed Hawks showed a clear tendency to re-use prior ly more urbanized (i.e., early vs. middle eras), or if it remains nest sites (including those inactive in the middle era yet active roughly the same (middle vs. late eras). in the early era). This behavior has long been noted in the Though data are sparse on their level of nest re-use, Cooper’s species (e.g., Fitch et al. 1946), and is still frequent in the local Hawk pairs studied in Albuquerque, New Mexico, frequently population elsewhere in the region (McCammon and Cooper moved in and out of nesting territories from year to year, with 2018). Thus, many Red-tailed Hawks in the study area may be fewer than half the territories active for all five years (Millsap tolerating some increased level urbanization while remaining 2017). This tendency of Cooper’sHawkstorotatenestsites at the same nests year after year. American Kestrels, with annually (also noted locally by McCammon and Cooper 2018), some territorial fidelity noted (though with a relatively low combined with the strong increase in urban cover around sample size), may be using a similar strategy, though at the Cooper’sHawknestsbetweenthemiddleandlateeras(when opposite end of the urbanization spectrum, remaining in the urbanization across the study area did not substantially increase), least-urbanized habitats year after year. suggests this species may be shifting its nesting sites toward The acceptance of planted/non-native trees (which were urban areas. Though we did not confirm this using marked birds, scarce historically) by all three hawk species may enable their it appears to be a recent change in strategy, as an urban bias was persistence in urban areas (see discussion in Chiang et al. not apparent in the prior eras (Fig. 2c). Indeed, Bloom and 2012). Red-tailed Hawks may simply select territories with McCrary (1996)reportedfewerthan5%ofCooper’sHawk trees of any type (either planted or native) at the edges of territories in “urban environments” in Orange and San Diego natural habitat used for daily foraging (see Chace and Walsh counties (California) from 1970 to 1995. 2006), as long as these areas support very tall nesting trees

24 Urban Ecosyst

Urban cover (%) around raptor nests over time (250m buffer)

a Red−tailed Hawk b Cooper's Hawk 100 100

75 75

Species Species

50 RANDOM 50 RANDOM RTHA COHA % Urban Cover % Urban Cover

25 25

0 0

Early Middle Late Early Middle Late Era Era c Red−shouldered Hawk d American Kestrel 100 100

75 75

Species Species

50 RANDOM 50 RANDOM RSHA AKES % Urban Cover % Urban Cover

25 25

0 0

Early Middle Late Early Late Era Era Fig. 2 a Mean urban cover of Red-tailed Hawk territories, random vs. Hawk territories, random vs. observed, by era Fig. 2d Mean urban cover observed, by era Fig. 2b Mean urban cover Cooper’s Hawk territories, of American Kestrel territories, random vs. observed, by era random vs. observed, by era Fig. 2c Mean urban cover of Red-shouldered

(Fitch et al. 1946; Wiley 1975). Cooper’s Hawks frequently has been noted in each year of the Griffith Park raptor survey, nest in non-native street trees in wholly-urban settings in which examined a study area roughly the same size as ours southern California, including areas with no undeveloped land closer to the urban core of Los Angeles (McCammon and for several kilometers around (D.S. Cooper, unpubl. data). Cooper 2018). It could be that natural nest cavities are scarce Elsewhere, they have become an urban bird since the 1990s within urban areas (both Los Angeles County and Ventura in places like Tucson, AZ (Boal and Mannan 1999), County require that homeowners remove dead trees from res- Albuquerque, NM (Millsap 2017), Reno, NV (White et al. idential properties, and dead native trees are scarce in and near 2018) and Milwaukee, WI (Stout and Rosenfield 2010). urban areas; pers. obs.). Because species’ ecology varies geo- Nest site choices have clear conservation implications. graphically, we caution against extrapolating our findings too Local American Kestrels are now nesting in significantly less broadly across the entire range of our focal species. For ex- urbanized sites than would be predicted by random points, ample, American Kestrels are known to nest in industrial areas which suggests that urbanizing regions – at least with the type elsewhere in the Los Angeles area (DSC, pers. obs.), and Red- of urbanization found in the study area – may not support shouldered Hawks may be both a riparian specialist and a them in the long-term. The species is known to be in decline suburban adaptor, depending on the location of the study area throughout North America, which some authors have corre- (Bloom et al. 1993). lated with loss of large habitat patches (e.g., Smallwood et al. Finally, the extirpated species in our study area are arguably 2009). Since kestrels have previously responded positively to more imperiled than any of our four focal species in the region. nest box programs (e.g., Steenhof and Peterson 2009), these Golden Eagles were extirpated over most of the Los Angeles declines may be reversible. Yet, if kestrels rely on some prey Basin, including most of the Santa Monica Mountains, by the type or nest feature absent from urban and urban-edge sites, mid-1990s (Allen et al. 2016). Prairie Falcons were probably they may not recover as areas become more urban. This may declining even earlier (e.g., Willet 1912), and while a single be happening in the region, as just a single potential territory breeding territory may have been present at the edge of the

25 Urban Ecosyst study area (Simi Peak) into the 1970s (NPS, unpubl. data), Our study provides an example of how to incorporate his- specific information on this pair is sparse, and the species is torical data and modern nesting observations, and how to use not considered an extant breeder today. Kites persisted very these data to understand how species persist in urbanizing locally as breeders into the 2000s, yet have not nested in the areas. Multi-decade studies may clarify species that may be study area since ca. 2010 (D.S. Cooper, unpubl. data; www. in need of continued conservation attention, such as Golden ebird.org), and rarely nest near urban areas in southwestern Eagles, and ones that may be at risk of future decline, such as California (Unitt 2004). While we have little data on Golden American Kestrels. Research into the shared characteristics of Eagle, White-tailed Kite or Prairie Falcon, their absence may urban-avoiding species would aid in their conservation in indicate a current level of urbanization above a particular areas that have not yet been subject to the type of urban ex- threshold within the study area,thecumulativeeffectoflack pansion of places like Los Angeles, but whose avifauna may of a food source, or loss of a key foraging habitat. White et al. be similarly threatened. (2018:57) found Golden Eagles to be among the most sensitive to urbanization in a survey of nesting raptors in and around Acknowledgments Funding for this research was provided by the UCLA Reno, NV, noting that during their study “residential develop- Grand Challenge Grant and UCNRS Stunt Ranch Reserve Research Award. We thank Lena Lee (National Park Service) for providing histor- ment encroached within 0.5 km of nesting Golden Eagles co- ical raptor nesting data from the study area. Peter Bloom provided helpful inciding with the nesting area being unused the following year comments early in the study. Ryan Harrigan, Noa Pinter-Wolman, Dana for the first time in recent years”.Alarger-scalesurveywould Williams, Gabriela Medeiros Pinho and Watcharapong “Win” be needed to adequately assess the needs of these species. Hongjamrassilp assisted with statistical analysis. Monica Dimson, Taylor Zagelbaum and Salvador Contreras assisted with GIS analysis. We acknowledge that many other factors, such as food Clare Riley assisted with field data collection. We thank the Blumstein availability and interspecific interactions, must also play a role lab and two anonymous reviewers for constructive comments on previous in nest site selection. Common Ravens (Corvus corax), which versions of the MS. harass raptors, are abundant in the study area year-round, and Data availability We are providing our data table of historical and recent we noted multiple raven nests in transmission towers and tall raptor nests, and the R code used to generate the plots and conduct the trees, including some that had been mapped as raptor nests in statistical analysis, as supplementary data earlier eras (when ravens were apparently less common, per Lee 2004). Regional climate may also play a role in determining nest site location. Average rainfall dropped during each of our three temporal eras examined; the total winter precipita- References tion for the three years preceding the average year of discov- ery of the early nests (1976–1979) was 64 cm of rain, then Allen L, Garrett KL, Wimer M (2016) Los Angeles County breeding bird 40 cm (2000–2003), and just 28 cm (2015–2018; Woodland atlas. Los Angeles Audubon Society Hills; http://www.laalmanac.com/weather/we137a.php). This Blair RB (1996) Land use and avian species diversity along an urban gradient. Ecol Appl 6:506–519 drop in rainfall may have had a strong effect on local nesting Bloom, PH and MD McCrary. 1996. The urban Buteo: red-shouldered raptors and prey levels by pushing raptors toward urban-edge hawks in Southern California. Pages 31-39 in: raptors in human habitats (with irrigated trees and abundant squirrels and rab- landscapes. D.M. bird, D.E. Varland, and J.J. Negro, eds. bits for prey), and away from wildland habitat and oaks strug- Academic press Bloom PH, McCrary MD, Gibson MJ (1993) Red-shouldered hawk gling with drought. We have no prey data for nests of any of home range and habitat use in southern California. J Wildl Manag the eras, though we recognize this would be a fruitful area of 57(2):258–265 study. We also recommend examining the effect of wildfire in Boal CW, Mannan RW (1999) Comparative breeding ecology of the study area, as several major fires have impacted the wild- Cooper’shawksinurbanandexurbanareasofsoutheastern land habitats in the study area since the 1970s, including the Arizona. J Wildl Manag 63:77–84 Cade TJ, Martell M, Redig P, Septon GA, Tordoff HB (1996) Peregrine Topanga Fire in 2005 which burned a large portion of the falcons in urban North America. Pages 3-14 in: raptors in human study area between the middle and late eras, and impacted landscapes. D.M. bird, D.E. Varland, and J.J. Negro, eds. Academic many mature oaks (the devastating Woolsey Fire burned press much of the study area in November 2018, just after our field- CALVEG (2009) South coast and montane ecological province. work ended). Finally, we note that our analysis of site fidelity CALVEG Zone 7. Accessed August 2018. Available online at: https://www.fs.usda.gov/detail/r5/landmanagement/ is based on discrete nest sites (generally trees), rather than on resourcemanagement/?cid=stelprdb5347192 the much larger territories used by our focal species. Estimates Chace JF, Walsh JJ (2006) Urban effects on native avifauna: a review. of site fidelity would be higher if we considered whether Landsc Urban Plan 74(1):46–69 whole territories, rather than specific nest sites, were re-used, Chiang SN, Bloom PH, Bartuszevige AM, Thomas SE (2012) Home but since individual birds were not marked, it is nearly impos- range and habitat use of Cooper’s hawks in urban and natural areas. In: Lepczyk CA, Warren PS (eds) Urban bird ecology and conser- sible to determine the boundaries of territories from our data, vation. Studies in avian biology (no. 45). University of California particularly when temporally separated by decades. Press, Berkeley

26 Urban Ecosyst

Craighead JJ, Craighead FC Jr (1969) Hawks, owls and wildlife. Otterbeck A, Selås V, Nielsen JT, Roualet E, Lindén A (2019) The par- Stackpole co., Harrisburg, Pa., and wild. Mgmt. Inst., Wash., D.C. adox of nest re-use: early breeding benefits reproduction, but nest re- 443 pp use increases predation risk. Oecologia 190:559–568 Devictor V, Julliard R, Couvet D, Lee A, Jiguet F (2007) Functional QGIS Development Team (2018) QGIS geographic information system. homogenization effect of urbanization on bird communities. Open Source Geospatial Foundation Project. Available online: http: Conserv Biol 21:741–751 qgis.osgeo.org Dykstra CR (2018) City lifestyles: behavioral ecology of urban raptors. R Core Team (2017) R: a language and environment for statistical com- Pages 18-35. In: Boal CW, Dykstra CR (eds) Urban raptors: ecology puting. R Foundation for statistical computing, Vienna, Austria. and conservation of birds of prey in cities. Island press URL http://www.R-project.org/ Fitch HS, Swenson F, Tillotson DF (1946) Behavior and food habits of Rattner BA, Horak KE, Warner SE, Day DD, Meteyer CU, Volker SF, the red-tailed hawk. Condor 48(5):205–237 Eismann JD, Johnston JJ (2011) Acute toxicity, hisopathology, and Fuller MR, Mosher JA (1987) Raptor survey techniques. Chapter 10. In: coagulopathy in American kestrels (Falco sparverius) following ad- Pendleton BG, Millsap BA, Cline KW, Bird DM (eds) Raptor man- ministration of the rodenticide diphacinone. Env toxicology and agement techniques manual. National Wildlife Federation, Chemistry 30(5):1213–1222 Washington, D.C. Richardson CT, Miller CK (1997) Recommendations for protecting rap- Garrett KL, Dunn J (1981) Birds of Southern California: status and dis- tors from human disturbance: a review. Wildl Soc Bull 25(3):634– tribution. Los Angeles Audubon Society 638 Gillespie TW, Pincetl S, Brossard S, Smith J, Saatchi S, Pataki D, Rockweit JT, Franklin AB, Bakken GS, Gutiérrez RJ (2012) Potential Saphores J-D (2011) A time series of urban forestry in Los influences of climate and nest structure on spotted owl reproductive Angeles. Urban Ecosyst 15:233–246 success: a biophysical approach. PLoS ONE:e41498 Grinnell J (1898) Birds of the Pacific slope of Los Angeles County. Seto KC, Fragkias M, Guneralp B, Reilly MK (2011) A meta-analysis of Pasadena Acad Sci 2 global urban expansion. PLoS One 6:e23777 Jiménez-Franco MV, Martinez JE, Calvo JF (2014) Patterns of nest re- Smallwood JA, Causey MF, Mossop DH, Klucsarits JR, Robertson B use in forest raptors and their effects on reproductive output. J Zool (2009) Why are American kestrel (Falco sparverius) populations declining in North America? Evidence from nest-box programs. J 292:64–70 Raptor Res 43(4):274–282 Johnston RF (2001) Synanthropic birds of North America. In: Marzluff Sol D, González-Lagos C, Moreira D, Maspons J, Lapiedra O (2014) JM, Bowman R, Donnelly R (eds) Avian ecology and conservation Urbanisation tolerance and the loss of avian diversity. Ecol Lett in an urbanizing world. Springer, Boston, MA, pp 49–67 17:942–950 Kreiderits A, Gamauf A, Krenn HW, Sumasgutner P (2016) Investigating Steenhof K (2013) Prairie falcon (Falco mexicanus), version 2.0. In the the influence of local weather conditions and alternative prey com- birds of North America (a. Poole F, editor). Cornell Lab of position on the breeding performance of urban Eurasian kestrels Ornithology, Ithaca, NY Falco tinnunculus. Bird Study 63:369–379 Steenhof K, Peterson B (2009) American kestrel reproduction in south- Lee L (2004) The distribution of breeding raptors in urban and natural western Idaho: annual variation and long-term trends. J Raptor areas of southern California. University of California, Los Angeles, Research 43(4):283–291 MSc. Thesis Stout WE, Rosenfield RN (2010) Colonization, growth and density of a Los Angeles Almanac. (2018). “General population by city, Los Angeles pioneer Cooper’s hawk population in a large metropolitan environ- County, 1960–2000.” Given Place Media, publishing as Los ment. J Raptor Research 44(4):255–267 Angeles Almanac. 20 Dec. 2018. Available online at: http://www. Sumasgunter P, Millan J, Curtis O, Koelsag A, Amar A (2016) Is multiple laalmanac.com/population/po27.php nest building an adequate strategy to cope with inter-species nest Marzluff JM (2005) Island biogeography for an urbanizing world: how usurpation? BMC Evol Biol 16:97 extinction and colonization may determine biological diversity in Tingley MW, Beissinger SR (2009) Detecting range shifts from historical human-dominated landscapes. Urban Ecosyst 8:157–177 species occurrences: new perspectives on old data. Trends Ecol Evol Marzluff JM (2016) A decadal review of urban ornithology and prospec- 24:625–633 tus for the future. Ibis 159:1–13 UCSB Library (2018) FrameFinder map web application. Accessed Marzluff, JM, R Bowman, and R Donnely. 2001. A historical perspective August 2018. Available online: http://mil.library.ucsb.edu/ap_ on urban bird research: trends, terms, and approaches. Pages 1–17 indexes/FrameFinder/ in: avian conservation and ecology in an urbanizing world (J. M. Unitt P (2004) San Diego County bird atlas. Proc San Diego Soc Nat Hist Marzluff, R. Bowman, and R. Donnely, Eds.). Kluwer academic, 39 Boston Vitousek PM, Mooney HA, Lubchenco J, Melillo JM (1997) Human McCammon C, Cooper DS (2018) Observing feathered families. The domination of Earth’s ecosystems. Science 277:494–499 Griffith reporter (newsletter of friends of Griffith Park). Winter White JH, Smith JM, Bassett SD, Brown JL, Ormsby ZE (2018) Raptor 2018-2019:6–7 nesting locations along an urban density gradient in the Great Basin, McKinney ML (2006) Urbanization as a major cause of biotic homoge- USA. Urban Ecosyst 21:51–60 nization. Biol Conserv 127:247–260 Wiley JW (1975) The nesting and reproductive success of red-tailed Millsap BA (2017) Demography and metapopulation dynamics of an hawks and red-shouldered hawks in Orange County, California, urban Cooper’s Hawk subpopulation. Condor 120(1):63–80 1973. Condor 77:133–139 Møller AP, Diaz M, Flensted-Jensen E, Grim T, Ibanez-Alamo JD, Willet G (1912) Birds of the Pacific slope of southern California. Pac Jokimaki J, Maend R, Marko G, Tryjanowski P (2015) Urbanized Coast Avifauna 7 birds have superior establishment success in novel environments. Wood EM, Esaian S (2020) The importance of street trees to urban avi- Oecologia 178:943–950 fauna. Ecol. Applic. 27 April 2020 (preprint)

27 Chapter 3. Temporally separated data sets reveal similar traits of birds persisting in a United States Megacity

fevo-08-00251 July 28, 2020 Time: 18:35 # 1

ORIGINAL RESEARCH published: 30 July 2020 doi: 10.3389/fevo.2020.00251

Temporally Separated Data Sets Reveal Similar Traits of Birds Persisting in a United States Megacity

Daniel S. Cooper1,2*, Allison J. Shultz2,3 and Daniel T. Blumstein1

1 Department of Ecology and Evolutionary Biology, University of California, Los Angeles, Los Angeles, CA, United States, 2 Ornithology Department, Natural History Museum of Los Angeles County, Los Angeles, CA, United States, 3 Urban Nature Research Center, Natural History Museum of Los Angeles County, Los Angeles, CA, United States

We present an analysis of life history and behavioral traits associated with urbanization for 52 breeding bird species on 173 survey blocks in the Los Angeles area of southern California, United States, across two time periods, 1995–1999 and 2012–2016. We used observational data from two community science efforts and an estimate of urban

Edited by: land cover in each block to develop an index of urban association, and then modeled David Andrew Luther, the relationship between species occurrence and eight traits likely associated with urban George Mason University, tolerance. We found two traits to be significantly associated with urbanization in both United States eras: Structure-nesting (i.e., the tendency to build nests on human-built structures) was Reviewed by: Phillip Cassey, positively associated, and cavity-nesting (i.e., the tendency to build nests in natural The University of Adelaide, Australia tree cavities) was negatively associated. Our analysis provides a template for mining Mark C. Mainwaring, University of Montana, United States historical community science data, and for “retrofitting” contemporary data to gain *Correspondence: insights into ecological trends over time, and illustrates the persistence of ecological Daniel S. Cooper traits of species associated with urban areas even as the makeup of these species [email protected]; communities may change. [email protected] Keywords: community science, citizen science, California, eBird, breeding bird atlas, life history traits, urban Specialty section: tolerance This article was submitted to Behavioral and Evolutionary Ecology, a section of the journal INTRODUCTION Frontiers in Ecology and Evolution Received: 15 April 2020 Understanding species’ tolerance to urbanization will be key to conserving biotic diversity as global Accepted: 10 July 2020 population increases and as more people move to cities (Vitousek et al., 1997; Marzlu, 2005). Published: 30 July 2020 Various external factors, including mechanical noise, anthropogenic light, windows, and outdoor Citation: cats represent direct, urban-associated influences on bird distributions (reviewed by Marzlu, Cooper DS, Shultz AJ and 2016). The process by which species invade and exploit novel environments has been referred to Blumstein DT (2020) Temporally as “filtering” (Clergeau et al., 2001), and may be applied to those bird communities in or near Separated Data Sets Reveal Similar Traits of Birds Persisting in a urban areas, with certain species passing through the urban filter successfully or invading following United States Megacity. urbanization, and others failing to do so (Lowry et al., 2013; Wingfield et al., 2015). Johnston (2001) Front. Ecol. Evol. 8:251. recognized a gradient of tolerance from urban avoidance to synanthropy, or a dependence on the doi: 10.3389/fevo.2020.00251 built environment, and this vocabulary has been expanded by numerous authors (e.g., “specialist”

Frontiers in Ecology and Evolution | www.frontiersin.org 1 July 2020 | Volume 8 | Article 251

28 fevo-08-00251 July 28, 2020 Time: 18:35 # 2

Cooper et al. Nesting Traits Explain Variation

vs. “mutualist” species, from MacGregor-Fors and Ortega- Mannan, 1999; Kettel et al., 2018), and for species that visit Alvarez, 2011) to describe the aliation between certain species (urban) feeders (O’Leary and Jones, 2006). Several authors have and urban areas. noted that urban areas would favor species that nest on human- While urbanization tends to homogenize formerly complex made structures tend (reviewed by Chace and Walsh, 2006), and ecological systems (McKinney, 2006; Devictor et al., 2007), would disadvantage those that use natural cavities (Blewett and certain specialist taxa may exploit urban sites preferentially, or Marzlu, 2005) as well as ground-nesting species (Evans et al., may assemble into novel communities there (Møller et al., 2015), 2011; Sol et al., 2014). Comparisons of nest productivity, clutch particularly where urban habitats are more structurally complex size, nest site preference and food-provisioning (to young) among than those replaced, such as grassland or low scrub (e.g., Emlen, urban bird populations have yielded contradictory results, as 1974; Gonzalez-Garcia et al., 2014). Certain types of food/prey noted by Chace and Walsh, (2006; see also Lowry et al., 2013; and nesting sites may be superabundant in urban areas, owing to Marzlu et al., 2015). Likewise, there appears to be little dierence the presence of lush, landscaped vegetation, anthropogenic water in the cognitive abilities of urban vs. rural populations of the same and supplemental feeding (Chace and Walsh, 2006), though species, as measured by problem-solving ability and relative brain this availability may be oset by novel hazards such as feral size (e.g., Carrete and Tella, 2011; Sol et al., 2014). cats (Loss et al., 2013). Thus, not all species that thrive in But do these patterns persist through time, in that the urban areas are drawn to hardscape or modified vegetation; same traits that connote success in urban areas do so year some may simply maintain populations in habitat fragments after year? Marzlu et al. (2001) recommended that tolerance within an otherwise urbanized landscape, for example marsh- to urbanization be re-assessed for species over time, because dwelling birds occurring at small urban wetlands, along flood- patterns of human activity are constantly changing, with cities control channels. adopting new architectural styles and landscaping palettes. Eorts to identify traits that allow species (or individuals of A species’ basic behavior also may change as populations the same species) to tolerate and even thrive with urbanization become more tolerant to human disturbance; for example, date to the early 1960s; more recently, the term “urban bird they may become habituated to elevated noise and city lights syndrome” has been coined to capture behavioral, physical, (e.g., Slabbekoorn and den Boer-Visser, 2006; Francis et al., reproductive, and ecological traits (see Møller, 2014; Samia 2009). Conversely, for the most sensitive species, even slight et al., 2015 for meta-analyses and summaries of prior findings). increases in human disturbance may have lasting negative Urban birds tend to display behavioral boldness and “innovation consequences (e.g., from recreational activity within natural propensity,” which compels individuals to explore new habitats open space areas, Pauli et al., 2016), leading to loss of and become established in these areas (Atwell et al., 2012; biodiversity over time. Thus, behavioral plasticity, as well as Blumstein, 2014; see review by Sol et al., 2017). They have tolerance, may also connote success in urban areas, where shorter flight initiation distances (FID) and exhibit heightened birds that readily alter their behaviors would thrive in predator avoidance (Blumstein, 2006; Møller, 2010), heightened cities, while those that cannot either decline and vanish, or territoriality and aggression (Evans et al., 2010), and reduced they never colonize (West-Eberhard, 1989; Sol et al., 2013; vocalizations (Estes and Mannan, 2003). They also tend to have Jokimäki et al., 2017). While studies of bird assemblages a broader elevational tolerance (Bonier et al., 2007) and a larger across gradients of urbanization (“space for time”) date to geographical range (Møller, 2009). Morphological variables the 1970s (Emlen, 1974; Beissinger and Osborne, 1982; Blair, have also been found to be associated with urbanization in 1996), those that investigate the same community over time birds, including body size (small size for raptors; Chace and are much less common (but see Aldrich and Con, 1980; Walsh, 2006), and wingspan (large wingspan for passerines; Shultz et al., 2012), and we are not aware of any that explicitly Croci et al., 2008). It is important to note that these studies investigate ecological traits associated with urbanization across include those that compared traits across multiple species, two temporal eras. as well as those that investigated traits of individuals within The Los Angeles metropolitan area of southern California, the same species. United States (which includes the city of Los Angeles), is an ideal Diet studies have consistently found positive associations place to study urban tolerance and persistence in bird species, due between urbanization and granivory, and negative associations to its long history of ornithological investigation (e.g., Grinnell, between urbanization and insectivory, including for ground- 1898; Swarth, 1900), its high human population, the large areas foraging insectivores (Kark et al., 2007; Croci et al., 2008; of open space present around its borders and even within the Evans et al., 2011; reviewed by Chace and Walsh, 2006). urban core, and its large and active birding and citizen-scientist Habitat preference studies have found that urban passerines are community (Higgins et al., 2019; Li et al., 2019). Its diverse disproportionately represented by forest species (Croci et al., avifauna is also in constant flux in terms of species abundance and 2008), and by species exhibiting a wide habitat breadth (Sol et al., distribution (Allen et al., 2016; Garrett, 2018); some local species 2014). Urban birds also tend to be non-migratory both globally have long been present and common in Los Angeles’ urban (Sol et al., 2014) and regionally in Europe (Croci et al., 2008) and environment, such as House Finches (Haemorhous mexicanus), Israel (Kark et al., 2007). while others, such as Dark-eyed Juncos (Junco hyemalis), appear Many breeding behaviors have also been associated with to be in a more recent process of shifting from wildland-favoring urbanization, and studies examining nesting phenology have and somewhat migratory, to ubiquitous year-round residents found earlier nest initiation both for urban raptors (Boal and (Yeh, 2004).

Frontiers in Ecology and Evolution | www.frontiersin.org 2 July 2020 | Volume 8 | Article 251

29 fevo-08-00251 July 28, 2020 Time: 18:35 # 3

Cooper et al. Nesting Traits Explain Variation

We developed two separate databases separated by nearly was dominated by low scrub and prairie-like grassland, now 20 years, “retrofitting” modern eBird data1 to an older essentially replaced by residential and commercial development. dataset from the breeding bird atlas eort in the 1990s We excluded the Santa Clara River valley/Santa Clarita area (Allen et al., 2016) to understand: (a) which ecological, behavioral north of the study area, because it is separated from the main and morphological traits of nesting birds are associated with Los Angeles Basin by a high pass (Newhall Pass) and features urban landscapes, and (b) whether this has changed in the a slightly dierent avifauna typical of more interior locations in past 15–20 years. We calculated an “urban index” for each the state. While the study area includes many microclimates (e.g., species based on its detections within each of 173 survey the coastal areas are cooler during the summer than the interior blocks, correlated with urban cover data. This index served as areas), no major natural impediments to bird dispersal exist. a measure of association with urbanization, and we used this value as a response variable in multiple models incorporating Urban Cover eight life history and behavioral traits, body mass, nest height Because urban areas may be defined at multiple scales, there (lowest), ground foraging, migratory status, natural cavity is neither global consensus on what constitutes “urban habitat,” nesting, artificial structure nesting, habitat breadth, and diet nor on how best to describe habitats modified by humans yet breadth. We fitted this model for both the early era and late era still retaining important natural elements (Croci et al., 2008; datasets, and examined whether the same traits were associated MacGregor-Fors, 2010; Evans et al., 2011; Beninde et al., 2015; with our urban index during each era. but see White et al., 2005; Li et al., 2019). As a measurement By examining a range of traits that may account for shifts of the degree of urbanization in our study area, we calculated in range across the region, we aimed to gain insights into urban cover using the “Urban/Built-Up” category in the statewide possible mechanisms behind species’ increases and decreases vegetation mapping dataset “CALVEG,” which was created in urban areas, and potentially resolve some of the previously between 2002 and 2003 (CALVEG, 2009; 1 ha mapping units). contradictory findings about species traits associated with urban CALVEG was found to be the most popular California vegetation areas. Our findings may have conservation implications, because layer in a recent online survey (Center for Geographical Studies, the presence of typically urban-avoiding species can be seen as an 2015), and is frequently used in species distribution studies at indication of ecosystem health, while conversely, the spread and the scale of ours (e.g., Santos et al., 2017; City of Los Angeles, prevalence of urban-tolerant species may indicate an ecosystem 2018). This catch-all Urban/Built-Up coverage includes human- that has been disrupted, or one that has changed from its former, made structures such as buildings and roads, but also manicured more natural state. By using two dierent datasets, separated by parks, golf courses and cemeteries, which, in the Los Angeles area, up to 20 years, we test the durability of these findings to explain tend to lack natural, native vegetation (as of 2003). Our urban patterns of urban association in birds. cover designation includes the habitat now commonly referred to as “urban forest” (Wood and Esaian, 2020), as distinct from natural open space, which may include native forest types, as well MATERIALS AND METHODS as many other natural habitat types. We overlaid the survey block boundaries onto the urban/built-up coverage using QGIS (QGIS Study Area Development Team,, 2016), and calculated the amount of urban We consider the “Los Angeles area” to be the entire southern cover in each of the 173 survey blocks (for a description of survey half of the ca. 10,000 km2 expanse of Los Angeles County, blocks see Breeding Bird Data). which includes all or portions of more than 80 incorporated We used the same urban cover values when modeling both the cities. The study area includes all coastal-draining land in the early and late era datasets (our CALVEG coverage was developed county below ca. 1,000 m above sea level, from the Santa in the years between the two eras), because separate land use data Monica Mountains and San Fernando Valley east through the at a suitable scale do not exist for each era. We recognize that both San Gabriel Valley to the San Bernardino County line, south housing density increases and localized development continues to the Pacific Ocean, including the Puente Hills and Palos to occur across the study area (ca. 3% increase in the county’s Verdes Peninsula, while excluding oshore islands (Figure 1). population between 2000 and 2010; Los Angeles Almanac,, 2019) The native habitats of the Los Angeles area, now largely and that absolute tree cover increased dramatically over the limited to its perimeter (but penetrating the central urban past century as the urban forest replaced a landscape that had core via the Santa Monica Mountains), include a diverse mix been dominated by arid scrub and grassland (Gillespie et al., of evergreen chaparral (dominated by large shrub species in 2011). However, relative urbanization within the study area have the Anacardiaceae, Rosaceae, and Rhamnaceae families), low, remained constant across our survey blocks, in that the most summer-deciduous scrub (including coastal sage scrub, featuring highly urbanized blocks were highly urban in both the early and sages Salvia spp.), patches of evergreen woodland (dominated by late era used here, and the least urbanized blocks in the 1990s are coast live oak Quercus agrifolia), plus numerous microhabitats still the least urbanized today, such as those in the Santa Monica such as riparian woodland and scrub, alluvial fan scrub, and Mountains (see maps in Lee et al., 2017). both seasonal and permanent wetlands (e.g., Schoenherr, 1992; Stein et al., 2007). Historically, the floor of the Los Angeles Basin Species Selection Of the 228 bird species in the Los Angeles County Breeding 1www.ebird.org Bird Atlas, we eliminated 176 species of these due to various

Frontiers in Ecology and Evolution | www.frontiersin.org 3 July 2020 | Volume 8 | Article 251

30 fevo-08-00251 July 28, 2020 Time: 18:35 # 4

Cooper et al. Nesting Traits Explain Variation

FIGURE 1 | Map of the 173 survey blocks across the study area. Thick solid lines indicate county boundaries (County boundaries downloaded from https://data.ca. gov). Gray-shaded areas indicate modern distribution of “urban” vegetation cover (as deﬁned by CALVEG, 2009) within the study area, and white areas are undeveloped open space comprised of various habitat types. The black shading to the north denotes urban vegetation cover outside the study area. Note that non-urban/wildland habitats are clustered toward the north and west, but also occur near the center of the study area.

factors that would interfere with an analysis of urban association, patchily in the study area, and likely contribute more to the including very low regional population size, specific microhabitat distribution of species than degree of urbanization. We further requirements (which may not be present throughout the study eliminated nocturnal species, as well as aerial foragers such as area), and tendency to wander during the breeding season. We swallows (Hirundinidae) and those species that travel widely, first excluded species that occur only in montane/desert areas often across multiple survey blocks, during daily foraging activity outside the Los Angeles area, and marine species found along (e.g., Psittacidae), to avoid counting the same individual birds in the immediate coast or on oshore islands. We then excluded multiple blocks and assuming they were breeding in these blocks. species due to regional rarity (i.e., those detected on <30 survey Finally, we eliminated species that tend to have such protracted blocks of the study area during the breeding season in both the migratory periods that it is dicult to tell when they are actually early and late eras), since we were interested in birds that could on breeding territories or simply moving through, such as Black- potentially occur anywhere in the study area, and that were not in headed Grosbeaks (Pheucticus melanocephalus), which frequently low numbers due to some other factor. We then eliminated those wander through the region for much of the late spring/summer associated with specific and localized habitats, such as wetlands, (see Unitt, 2004). riparian, specific types of scrub, and those known to be grassland- Our final list of 52 species thus includes those that were: (a) obligate species, since these habitats were found narrowly and widespread enough to be found (or expected) across the study

Frontiers in Ecology and Evolution | www.frontiersin.org 4 July 2020 | Volume 8 | Article 251

31 fevo-08-00251 July 28, 2020 Time: 18:35 # 5

Cooper et al. Nesting Traits Explain Variation

area, (b) likely nesting where they are detected in spring/early County eBird records for 2012–2016 (1.36 million records), We summer, and c) habitat generalists, occurring in woodland and used the software R (version 3.4.1., R Foundation for Statistical shrubby vegetation that represents the dominant habitat across Computing, Wien, Austria) to create a database of sightings that the Los Angeles area, and which is simulated by ornamental fit our criteria for analysis. Some “coarsening” was necessary to plantings such as hedges and street trees. We included several directly compare the breeding status of species from the BBA to introduced taxa that we knew to be tied to urbanized/modified that derived from eBird data, as BBA data were reported at the habitats, such as Scaly-breasted Munia (Lonchura punctata). Our level of atlas block, while eBird data is reported by point data. final focal species list thus represents a mix of resident and We assigned each eBird record to a particular atlas block using a migratory status, size classes (e.g., raptors to hummingbirds), and spatial join function in QGIS (QGIS Development Team,, 2016). a diversity of morphological and ecological attributes, with each To further refine the eBird data, records of each focal species species having the potential to occur as breeding species in all were filtered by “safe date,”a range of dates for which the presence regions of the study area, and whose presence on a survey block of that species within a set of dates would be indicative of at during the breeding season would strongly suggest local nesting least “possible” breeding in the County, as determined during the on that block. atlas eort. In certain cases, we used the reported dates of local breeding in lieu of safe dates (e.g., for “breeds late March to early Breeding Bird Data July,” we used March 15 to July 15) if they were not provided BBA (Early Era Dataset) by Allen and Garrett (1995) for the atlas. Records outside From 1995 to 1999, the Los Angeles County Breeding Bird Atlas these dates were discarded. Because observations of breeding was organized around 414 blocks based on USGS topo quads behavior are not frequently reported in eBird checklists, we could (each roughly 5.8 km E–W 4.6 km N–S, or 2,668 hectares; rarely distinguish between “probable” and “confirmed” breeding. ⇥ some blocks were larger or smaller along county lines). Each atlas Therefore, we considered each species “probable/confirmed” for volunteer was assigned one or more blocks and given detailed a given survey block if more than two individuals were observed instructions on how to confirm nesting for as many species as at a single location (i.e., eBird Hotspot or personal location) possible within that block, over the span of 5 years. Species during the safe (or designated breeding) dates for any year during were assigned three levels of breeding status for each block the 2012–2016 period. We assigned species as being a “possible” (i.e., “confirmed,” “probable,” or “possible” breeding) based on breeder in the block if just one individual was detected with the standardized breeding indicators used during the atlas eort (e.g., safe dates, and noted a species as “not breeding” if it was not singing male represented “possible” breeding, carrying nesting recorded at all within safe dates. material and feeding young represented “confirmed” breeding, Because we had no observer eort associated with the BBA etc.). All data were pooled into an overall “highest breeding data, we did not calculate observer eort for the late era (eBird) status” value, by block, and no specific eort data were collected data, but worked under the assumption that the most-visited during the atlas project (i.e., how much observational time was sites in the late 1990s were the same (or were in the same spent within each block). In all, 22,840 records were amassed survey blocks) as those from 2012–2016. Likewise, we maintained for 228 species (not all of them confirmed as breeding) by 98 a conservative approach in data analyses and did not attempt observers searching their blocks. An additional 5,320 “casual to calculate species abundance within blocks, nor number of observations” by 218 observers were submitted to the atlas project years when observed, but simply counted a bird as achieving during the atlas period, for a total of 28,935 breeding records the highest breeding category during a particular span of years analyzed and vetted by sta of the Natural History Museum of (i.e., replicating what was done for the BBA project). As reviewed Los Angeles County (Allen et al., 2016). We analyzed only the by Horns et al. (2018) eBird data, even while opportunistically 173 atlas blocks that fell within our coastal lowland study area collected, produces similar results to other forms of observational (see section “Study Area”). data collection across large geographical scales, so we felt comfortable comparing the two datasets (BBA and eBird). EBird (Late Era Dataset) Data from eBird2 are collected in a completely dierent way Breeding Level and Urban Index than the BBA data, with sightings submitted opportunistically We entered three “breeding levels” for each species, for each by birders from either a specific, georeferenced location, or from survey block, during each era (0 = no record, 1 = possible somewhere within a larger “hotspot” (typically a park or a trail). breeding or 2 = probable/confirmed breeding). We then EBird data prior to 2,000 are relatively sparse compared to more calculated an “urban index” for each species during each era, recent years (hence our incorporation of breeding bird atlas which was the correlation coecient between that species’ data), and the platform continues to gain in global popularity breeding level within each block (0–2) and the percent urban (as of December 2019, eBird “checklists” – observations of one or cover value within that block, using a Spearman’s rank test with more species by a registered eBird user for a particular location, the rcorr function in R using the Hmisc package (Harrell, 2004). date and time period – were being submitted at the rate of ca. A positive urban index would indicate a positive association 50,000 per year for Los Angeles County, one of the most actively between a species and urban cover, while a negative urban birded regions of the world). After obtaining all Los Angeles index would indicate a negative association with urban cover; an urban index near zero would indicate no association with 2www.eBird.org urban cover. The urban index served as our response variable,

Frontiers in Ecology and Evolution | www.frontiersin.org 5 July 2020 | Volume 8 | Article 251

32 fevo-08-00251 July 28, 2020 Time: 18:35 # 6

Cooper et al. Nesting Traits Explain Variation

and formed the basis for our trait analysis, below (see also TABLE 1 | Functional traits considered for this analysis.

Supplementary Table S1). Trait Description Source Because the survey blocks used are arranged in a grid, and urban development in the Los Angeles area includes large, Adult body mass Total weight (grams; of male if Dunning, 2007 continuous areas of urbanization (as well as large blocks of open different) space), urban cover was necessarily autocorrelated (Moran’s I Lowest nest height Meters; lowest average nest Ehrlich et al., 1988; height BNA observed: 0.019, expected: 0.006, SD = 0.009, P = 0.010). We Forage ground only Categorical (2; 0/1); forages Wilman et al., 2014 sought to reduce any error introduced via spatial autocorrelation (during breeding exclusively or mainly on the by using a single urban index value for each species, which was season) ground not spatially explicit, but reflected the association between bird Migratory status Categorical (2; 0/1); eBird distribution and urban cover. non-migratory, partially/fully We first tested for a phylogenetic signal in the urban index migratory values for both the early/BBA values and the late/eBird values, Cavity nest Categorical (2; 0/1); frequently Allen et al., 2016 using models that employed three dierent modes of evolution: uses tree cavities for nesting. Brownian motion, Pagel’s lambda, Ornstein–Uhlenbeck, as well Structure nest Categorical (2; 0/1); frequently Allen et al., 2016 uses human-made structures for as a non-phylogenetic model (see Münkemüller et al., 2012). We nesting (excluding bird boxes) used the ape (Paradis et al., 2004), geiger (Pennell et al., 2014), Habitat breadth Level (3; 1–3) Garrett and Dunn, and picante (Kembel et al., 2010) packages in R, and used the 1981 phylosignal function to analyze the focal species’ urban index Diet breadth Level (6; 1–6) Sekerciglou, unpubl. relative to their corresponding positions on the phylogenetic data tree described above. We first tested a Brownian motion, or BNA, Birds of North America (various authors, https://birdsna.org/). random-walk model, using a Blomberg’s K test (Blomberg et al., 2003), which compares the variance of phylogenetic independent contrasts to what we would expect under a Brownian motion We then tested the association between these traits and each (BM) model. Here, K = 1 means that relatives resemble one species’ aliation for urban cover using the urban index as the another as much as we should expect under BM; K < 1 means that dependent variable (using both the “early” and “late” values in there is less “phylogenetic signal” than expected under BM, while separate tests), and the eight traits as independent variables. K > 1 means that there is more. We then analyzed the urban In separate tests (early and late) we ran three phylogenetic index and tree data using Pagel’s lambda (Pagel, 1999). Here, if generalized least squares (PGLS) tests and one non-phylogenetic our estimated lambda = 0, then the traits are inferred to have GLS tests using each, and compared AIC values of each to select no phylogenetic signal. Lambda = 1 corresponds to a Brownian the model that best explained variation in the data. motion model; 0 < lambda < 1 is intermediate. Finally, we used We used the gls function in the nlme package in R (Pinheiro a model which employed the Ornstein–Uhlenbeck (OU) mode et al., 2019), and incorporated a Brownian motion mode of of evolution which incorporates stabilizing selection wherein the evolution using the corBrownian function in the phytools trait is drawn toward a fitness optimum, or long-term mean, package in R (Revell, 2012), along with our phylogenetic tree data. rather than being completely random and directionless (Martins, We conducted a second PGLS test using the Ornstein–Uhlenbeck 1994). To test for no phylogenetic signal, we also used a “no- (OU) mode of evolution using the corMartins function in the signal” generalized least squares model where lambda was set to 0. sde package in R (Iacus, 2016). We conducted a third PGLS test using Pagel’s lambda test with the corPagel function also in the sde package. We fitted a non-phylogenetic least squares Trait Analysis model to compare with the PGLS tests. For all analyses, best We identified eight life history and behavioral traits likely fit parameters of the phylogenetic model were estimated with associated with urban tolerance based on those identified in maximum likelihood. Lastly, we checked residuals for normality previous studies (e.g., Møller, 2014; Samia et al., 2015): body using QQ tests, and selected the analysis with the lowest AIC mass, nest height (lowest), ground foraging, migratory status, values as the best model. natural cavity nesting, artificial structure nesting, habitat breadth, and diet breadth (Table 1). We were limited in which variables we could use for subsequent modeling by data gaps (e.g., flight RESULTS initiation distance has been calculated for fewer than half the focal species; D.T. Blumstein, unpubl. data). To account for Urban Index phylogenetic relatedness among species in our analyses, we used Nearly all focal species (48 of 52 species) showed an increase an avian phylogeny from Bird Tree (Jetz et al., 2012, 2014). With (i.e., toward positive) in urban index over time (Supplementary our list of 53 species, we used the phylogeny subset tool (in Table S1), and while we cannot directly compare urban indices Bird Tree) to create 1,000 trees built with a Hackett et al. (2008) between the two eras due to the dierent methodologies used in backbone. For use in subsequent analyses, we created a majority- data collection, some of these species shifted from a negative or rule consensus tree, collapsing nodes that did not show up in at neutral urban index to a positive one, suggesting they may now least 50% of the 1,000 trees. be preferring urban habitats – or, at least, natural habitats near

Frontiers in Ecology and Evolution | www.frontiersin.org 6 July 2020 | Volume 8 | Article 251

33 fevo-08-00251 July 28, 2020 Time: 18:35 # 7

Cooper et al. Nesting Traits Explain Variation

urban areas – over blocks with a greater percentage of natural dataset). Cavity-nesting and migratory status were negatively vegetation. These “shifters” include representatives from diverse associated with the urban index (that is, cavity-nesting and families, including Cooper’s Hawks (Accipiter cooperii)( 0.34 to migratory birds were more associated with natural habitat), while 0.15), Allen’s Hummingbirds (Selasphorus sasin) (0.10 to 0.24), structure-nesting was positively associated with the urban index. and Hooded Orioles (Icterus cucullatus)( 0.15 to 0.15). At the While the AIC score of the non-phylogenetically informed GLS other end of the spectrum, those with the largest negative residual was not suciently dierent from the OU and Pagel’s lambda values include California Quail (Callipepla californica), Wrentits models (Table 3), these models had nearly identical associations (Chamaea fasciata) and Spotted Towhees (Pipilo maculatus). By with urban index in both the early and late era datasets. We contrast, very few species shifted from positive (i.e., more urban- summarize the results of the best model (non-phylogenetically associated) to negative (Supplementary Table S1). We plot informed GLS) in Table 4. species’ representation on survey blocks (Figure 2A) as well as Non-significant negative associations were detected in several the urban indices for each species (Figure 2B), showing that both models (including the best/non-phylogenetically informed GLS values are highly correlated across eras (rp = 0.80, P < 0.001 for model) for body mass, ground-foraging, migratory status and number of blocks where suspected/confirmed breeding; rp = 0.90, habitat breadth, and non-significant positive associations with P < 0.001 for urban index). nest height and diet breadth. While they did not rise to the level of We found no indication of a phylogenetic signal in the urban significance (i.e., P > 0.05), they were consistent in their direction index value using three phylogenetic models (Brownian motion, across temporal eras. O–U, and Pagel’s lambda), with the non-phylogenetic model returning the lowest AIC value (Table 2). DISCUSSION Trait Analysis We found that two traits (cavity-nesting and structure-nesting) Our study is one of very few to analyze the persistence of avian were significantly associated (p < 0.05) with urban index values traits using both historical and current community-science data, using both the early and late eras in most models examined. and adds to an ample literature on why some birds thrive in We found migratory status was significantly associated with urban areas and others avoid them. Our results suggest that urbanization in the early era dataset (but not in the late era nest site choice and migratory status may confer either an

FIGURE 2 | (A) A comparison of species occurrence in early and late eras by number of blocks where suspected/conﬁrmed breeding. (B) A comparison of urban index values between the early and late eras. We have labeled the species with the largest residuals from the blue best-ﬁt line.

Frontiers in Ecology and Evolution | www.frontiersin.org 7 July 2020 | Volume 8 | Article 251

34 fevo-08-00251 July 28, 2020 Time: 18:35 # 8

Cooper et al. Nesting Traits Explain Variation

TABLE 2 | Comparison of three phylogenetically informed Generalized Least functional diversity maintained in an urban bird community over Squares (PGLS) models (Brownian, OU, Pagel’s lambda) and one a century, despite changes in community composition as the area non-phylogenetically informed model (GLS). urbanized (see also Hagen et al., 2017). Indeed, the lack of a strong Model Test statistic AICc phylogenetic signal in patterns we documented (in either era) suggests that urban tolerance is not restricted to a few related Early era (BBA data) species, but rather occurs across unrelated taxa, as observed in Brownian K = 0.223; P = 0.239 Alpha: 2.718 66.727 both birds and other taxonomic groups (Martin and Bonier, 2018; OU Sigma squared: 0.544 34.391 Merckx et al., 2018; Santini et al., 2019). Pagel’s lambda L < 0.001; P = 1 34.406 The significant negative association between cavity nesting Non-phy. Sigma squared: 0.001 32.151 and urban index may be a result of urban tree species in urban Los Late era (eBird data) Angeles having been selected for their longevity, rapid growth, Brownian K = 0.189; P = 0.452 Alpha: 2.718 38.826 and resistance to boring pests (Gutzat and Dormann, 2018; Frank OU Sigma squared: 0.361 13.118 et al., 2019), and the tendency for large urban trees, especially Pagel’s lambda L < 0.001; P = 1 13.099 those with dead limbs (“snags”) to be removed in residential areas Non-phy. Sigma squared: 0.001 10.844 due to safety concerns (falling branches injuring people). It may Here we use urban index scores only as our response variable (i.e., no be that more successful urban nesters would be those species behavioral or life history traits). The lowest AICc score was found using the non-phylogenetic model. that are able to utilize a variety of built structures (as well as natural cavities), including eaves of buildings, parking garages, and overpasses, with this flexibility allowing them to switch TABLE 3 | Comparison of AIC scores of best model (non-phylogenetic GLS) using urban index as the response variable and eight behavioral and life history traits as between substrates when one is not available. It is also possible the predictor variables. that aggressive (urban-tolerant) cavity nesters such as European Starlings (Sturnus vulgaris) and/or parrots may be displacing Model Early era (BBA) Late era (eBird) natural cavity nesters within urban areas, though direct evidence Brownian 93.724 71.685 of this is lacking (Koch et al., 2012; Diamond and Ross, 2019). OU 66.988 51.695 Still, avian diversity in urban areas could be enhanced by Pagel’s lambda 68.619 51.522 provisioning artificial nesting structures, and maintaining natural Non-phylogenetic 66.979 49.695 nesting sites such as dead trees (see Tomasevic and Marzlu, 2017), as well as by retaining patches of natural habitat of Full model results are available in the Supplementary Materials. various sizes within the urban matrix (Silva et al., 2015) and by planting a diversity of trees and shrubs as part of landscaping TABLE 4 | Results from the best model in both time eras (non-phylogenetic GLS) (Wood and Esaian, 2020). fitted to explain variation in the urban index values based on eight life history and We found no significant correlations between urban index behavioral traits (includes standard error and P-value). and body mass, ground foraging, or either diet or habitat Early era (BBA) Late era (eBird) breadth, all of which have been found to be associated with urban in prior studies (e.g., Chace and Walsh, 2006; Evans Intercept 0.240 0.161; P = 0.142 0.038 0.131; P = 0.773 ± ± et al., 2011). It could be that the large scale of the atlas blocks Trait (ca. 2,668 ha) encompassed a variety of habitat types and Log(adult mass) 0.041 0.041; P = 0.330 0.031 0.034; P = 0.372 ± ± variety of urban conditions (which we did not analyze here), Nest height 0.009 0.007; P = 0.238 0.010 0.006; P = 0.093 ± ± so a finer-level analysis (e.g., eBird point data) might detect Ground-foraging 0.070 0.120; P = 0.563 0.071 0.098; P = 0.471 ± ± more significant associations (see Croci et al., 2008; Ferenc Migratory status 0.201 0.099; P = 0.047* 0.131 0.081; P = 0.105 ± ± et al., 2014 for discussions of scale). While not statistically Cavity-nesting 0.208 0.098; P = 0.040* 0.206 0.080; P = 0.014* ± ± significant, the consistent positive associations found between Structure-nesting 0.264 0.092; P = 0.006* 0.191 0.075; P = 0.014* ± ± urban index and nest height may indeed be “real,” as so many Diet breadth 0.060 0.038; P = 0.135 0.032 0.031; P = 0.301 ± ± structure-nesting birds nest atop towers, buildings, and other tall Habitat breadth 0.049 0.052; P = 0.355 0.051 0.043; P = 0.238 ± ± features of the urban environment, which are less prevalent in Residual standard 0.295 0.235 error wildland habitats and which would become more common over time in urban areas as infill hardscape development displaces Please refer to Supplementary Materials for full model results. Asterisk and bold vegetation (e.g., Lee et al., 2017). Likewise, the consistent negative font denote P < 0.05. association with migratory status across several models used, while (weakly) significant only during the early era, may become advantage (for artificial structure nesters and sedentary species) stronger with additional (migratory) species included in a future or a disadvantage (for cavity nesters and for migratory species) analysis (including data from multiple cities), or with finer-grain within urban areas, that these patterns may persist over time migration data (our binary “migratory status” trait does reflect (even if the makeup of the species community changes) using the range of long- and short-distance and partial migrants). two dierent data collection methodologies (i.e., a BBA dataset We also note that certain species are clearly modifying their vs. an eBird dataset). The durability of these traits through tolerance to local urbanization as they increase in distribution time was suggested by Shultz et al. (2012), who found levels of within the study area, which may lead to concurrent changes

Frontiers in Ecology and Evolution | www.frontiersin.org 8 July 2020 | Volume 8 | Article 251

35 fevo-08-00251 July 28, 2020 Time: 18:35 # 9

Cooper et al. Nesting Traits Explain Variation

in their ecological traits. For example, Dark-eyed Juncos were (White et al., 2005; Filazzola et al., 2019). Yet cities must also found on roughly three times as many survey blocks in the allow the least-adaptable species – those most strongly associated early vs. late era, and saw their urban index shift from strongly with wildland rather than urban habitats – to find refuge within negative in the early era ( 0.56) to weakly negative by the late era the urban matrix as they urbanize (Sol et al., 2014). Much of ( 0.15) (Supplementary Table S1). This species is now a year- this tension results from studies using dierent scales of analysis; round resident across the Los Angeles Basin, and is frequently high local diversity may be easier to achieve within cities than found nesting in structures, including within parking structures high global diversity, which requires the conservation of rare in urban Los Angeles (D.S. Cooper pers. obs.); decades ago it was and endemic species (e.g., Enedino et al., 2018; McDonald et al., largely a migratory ground-nester, restricted to montane areas for 2018). Place matters, too, and while a featureless desert may breeding (Garrett and Dunn, 1981; Allen et al., 2016). Likewise, support relatively few bird species compared to the oasis-like structure-nesting in Cassin’s Kingbirds was not mentioned in city that replaces it (e.g., Gonzalez-Garcia et al., 2014), this recent breeding bird atlases based on data from the 1990s (Unitt, scenario would hardly be considered a desirable conservation to 2004; Allen et al., 2016), but this tendency has since become be replicated everywhere (otherwise, why not cover the earth in a frequent sight around Los Angeles (pers. obs.), during which cities?). Thus, an understanding of the mechanics of urban bird time this species has increased its representation on survey blocks community development is merely a necessary first step on the roughly fourfold. We encourage more research on the dierential way to developing meaningful conservation goals. usage of cavities and artificial structures in urban areas and at the urban edge, as urban-colonizing species continue to utilize new substrates for breeding (see Reynolds et al., 2019). Our finding that overall, species were found more widely (i.e., DATA AVAILABILITY STATEMENT in more survey blocks) and with higher urban indices in the late era than the early one may be an artifact of the two dierent All datasets generated for this study are included in the methodologies used in data collection rather than a biological article/Supplementary Material. pattern. This is likely a result of the more inclusive approach assigning breeding status from the (later) eBird data (where only breeding season records of single birds or pairs was used to denote breeding) versus the more conservative approach used AUTHOR CONTRIBUTIONS in generating the BBA data, which required observers to justify their assessment of nesting with field observations. Thus, birds DC and AS: conceptualization, methodology, and data analysis. recorded only once in 5 years in a given block might not have DB: methodology and supervision. DC: Writing, including warranted a “possibly breeding” (i.e., code 1) assignment in the original draft. All authors: review and editing. early era, because these determinations were often made post hoc and somewhat subjectively by the atlas coordinators, based on suitable habitat, other nesting behavior, etc. (see Allen et al., 2016); yet for the late era dataset, a “one-o” sighting would have ACKNOWLEDGMENTS been counted as possibly breeding. Because reliable abundance data were not available for the Breeding Bird Atlas, we did not We thank the Sustainable L.A. Grand Challenge for providing calculate abundance using the eBird data, and simply used scores financial support for graduate work of DC (especially M. Gold between 0 and 2, summing them as a substitute for abundance and C. Rausser). We thank K. Garrett at the Natural History across all 173 blocks. Abundance should be more easily calculated Museum, Los Angeles County for providing atlas data, including in the future as community-science projects expand and the unpublished data tables and field notes, and for insight into amount of point data increases, allowing for more granular the interpretation of these data. We also thank the numerous studies into local and regional biodiversity, population and range contributors to eBird within Los Angeles County, including the shifts, and community organization (e.g., Ballard et al., 2017; observers, data reviewers, and the project team at the Cornell Callaghan et al., 2017; Jaric´ et al., 2020). Although we limited our Lab of Ornithology. Museum sta G. Pauly, J. Vendetti, and analysis to the breeding season, when our focal bird species would members of the Urban Nature Research Center provided helpful likely be tied to a specific territory and thus dependent on the comments on an early draft of the manuscript. R. Harrigan local resources available for themselves and their ospring (Mills assisted with statistical analysis, and M. Dimson, T. Zagelbaum, et al., 1989), a similar analysis could be performed for wintering and S. Contreras assisted with geo-spatial analysis. or even transient species using data collected at other times of year, a period when urban habitats are utilized by a diversity of native bird species (e.g., Wood and Esaian, 2020). Finally, as urbanization continues to expand globally, we SUPPLEMENTARY MATERIAL encourage further reflection on ways to define “success” in urban areas. On one hand, cities may be considered successful if they The Supplementary Material for this article can be found online include built features that can support a high diversity of species, at: https://www.frontiersin.org/articles/10.3389/fevo.2020.00251/ some of which would not have occurred prior to urbanization full#supplementary-material

Frontiers in Ecology and Evolution | www.frontiersin.org 9 July 2020 | Volume 8 | Article 251

36 fevo-08-00251 July 28, 2020 Time: 18:35 # 10

Cooper et al. Nesting Traits Explain Variation

REFERENCES Diamond, J. M., and Ross, M. S. (2019). Exotic parrots breeding in urban tree cavities: nesting requirements, geographic distribution, and potential impacts Aldrich, J. W., and Con, R. W. (1980). Breeding bird populations from forest to on cavity nesting birds in southeast Florida. Avian Res. 10:39. suburbia after thirty-seven years. Am. Birds 34, 3–7. Dunning, J. B. (2007). CRC Handbook of Avian Body Masses, 2nd Edn. Boca Raton, Allen, L., Garrett, K. L., and Wimer, M. (2016). Los Angeles County Breeding Bird FL: CRC Press. Atlas. Los Angeles, CA: Los Angeles Audubon Society. Ehrlich, P. R., Dobkin, D. S., and Wheye, D. (1988). The Birder’s Handbook: A Field Allen, L. W., and Garrett, K. L. (1995). Atlas Handbook: Los Angeles County Guide to the Natural History of North American Birds. New York: Simon and Breeding Bird Atlas. Los Angeles, CA: Los Angeles Audubon Society. Schuster/Fireside Books. Atwell, J. W., Cardoso, G. C., Whittaker, D. J., Campbell-Nelson, S., Robertson, Emlen, J. T. (1974). An urban bird community in Tucson, Arizona: derivation, K. W., and Ketterson, E. D. (2012). Boldness behavior and stress physiology in structure, regulation. Condor 76, 184–197. a novel urban environment suggest rapid correlated evolutionary adaptation. Enedino, T. R., Loures-Ribeiro, A., and Santos, B. A. (2018). Protecting biodiversity Behav. Ecol. 23, 960–969. doi: 10.1093/beheco/ars059 in urbanizing regions: the role of urban reserves for the conservation of Ballard, H. B., Robinson, L. D., Young, A. N., Pauly, G. B., Higgins, L. M., Johnson, Brazilian Atlantic Forest birds. Perspect. Ecol. Const. 16, 17–23. doi: 10.1016/ R. F., et al. (2017). Contributions to conservation outcomes of natural history j.pecon.2017.11.001 museum-led citizen science: examining evidence and next steps. Biol. Cons. 208, Estes, W. A., and Mannan, R. W. (2003). Feeding behavior of Cooper’s Hawks 87–97. doi: 10.1016/j.biocon.2016.08.040 at urban and rural nests in southeastern Arizona. Condor 105, 107–116. doi: Beissinger, S. R., and Osborne, D. R. (1982). Eects of urbanization on avian 10.1093/condor/105.1.107 community organization. Condor 84, 75–83. Evans, J., Boudreau, K., and Hyman, J. (2010). Behavioral syndromes in urban and Beninde, J., Veith, M., and Hochkirch, A. (2015). Biodiversity in cities needs space: rural populations of song sparrows. Ethology 116, 588–595. a meta-analysis of factors determining intra-urban biodiversity variation. Ecol. Evans, K. L., Chamberlain, D. E., Hatchwell, B. J., Gregory, R. D., and Gaston, Lett. 18, 581–592. doi: 10.1111/ele.12427 K. J. (2011). What makes an urban bird? Glob. Change Biol. 17, 32–44. doi: Blair, R. B. (1996). Land use and avian species diversity along an urban gradient. 10.1111/j.1365-2486.2010.02247.x Ecol. Appl. 6, 506–519. doi: 10.2307/2269387 Ferenc, M., Sedláèek, O., Fuchs, R., Dinetti, M., Fraissinet, M., and Storch, D. Blewett, C. M., and Marzlu, J. M. (2005). Eects of urban sprawl on snags and (2014). Are cities dierent? Patterns of species richness and beta diversity of the abundance and productivity of cavity-nesting birds. Condor 107, 678–693. urban bird communities and regional species assemblages in Europe. Glob. Ecol. doi: 10.1093/condor/107.3.678 Biogeogr. 23, 479–489. doi: 10.1111/geb.12130 Blomberg, S. P., Garland, T., and Ives, A. R. (2003). Testing for phylogenetic signal Filazzola, A., Shrestha, N., and MacIvor, J. S. (2019). The contribution of in comparative data: behavioral traits are more labile. Evolution 57, 717–745. constructed green infrastructure to urban biodiversity: a synthesis and meta- doi: 10.1111/j.0014-3820.2003.tb00285.x analysis. J. Appl. Ecol. 56, 2131–2143. Blumstein, D. T. (2006). Developing an evolutionary ecology of fear: how life Francis, C. D., Ortega, C. P., and Cruz, A. (2009). Noise pollution changes avian history and natural history traits aect disturbance tolerance in birds. Anim. communities and species interactions. Curr. Biol. 19, 1415–1419. doi: 10.1016/ Behav. 71, 389–399. doi: 10.1016/j.anbehav.2005.05.010 j.cub.2009.06.052 Blumstein, D. T. (2014). “Attention, habituation, and antipredator behavior: Frank, S. D., Backe, K. M., McDaniel, C., Green, M., Widney, S., and Dunn, R. R. implications for birds,” in Avian Urban Ecology, eds D. Gil, and H. (2019). Exotic urban trees conserve similar natural enemy communities to Brumm (Oxford: Oxford Univ. Press), 41–53. doi: 10.1093/acprof:osobl/ native congeners but have fewer pests. PeerJ 7:e6531. doi: 10.7717/peerj.6531 9780199661572.003.0004 Garrett, K., and Dunn, J. (1981). Birds of Southern California: Status and Boal, C. W., and Mannan, R. W. (1999). Comparative breeding ecology of Cooper’s Distribution. Los Angeles, CA: Los Angeles Audubon Society. hawks in urban and exurban areas of southeastern Arizona. J. Wildl. Manag. 63, Garrett, K. L. (2018). “Introducing change: a current look at naturalized bird 77–84. species in western North America in Trends and traditions: avifaunal change Bonier, F., Martin, P. R., and Wingfield, J. C. (2007). Urban birds have broader in western North America,” in Studies of Western Birds 3, eds W. D. Shuford, environmental tolerance. Biol. Lett. 3, 670–673. doi: 10.1098/rsbl.2007.0349 R. E. Gill, Jr., and C. M. Handel, (Camarillo, CA: Western Field Ornithologists), Callaghan, C. T., Lyons, M. B., Martin, J. M., Major, R. E., and Kingsford, 116–130. doi: 10.21199/swb3.5 R. T. (2017). Assessing the reliability of avian biodiversity measures of urban Gillespie, T. W., Pincetl, S., Brossard, S., Smith, J., Saatchi, S., Pataki, D., et al. greenspaces using eBird citizen science data. Avian Cons. Ecol. 12:12. (2011). A time series of urban forestry in Los Angeles. Urb. Ecosyst. 15, 233–246. CALVEG, (2009). South Coast and Montane Ecological Province, CALVEG doi: 10.1007/s11252-011-0183-6 Zone 7. Available at: https://www.fs.usda.gov/detail/r5/landmanagement/ Gonzalez-Garcia, F., Straub, R., Lobato Garcia, J. A., and MacGregor-Fors, I. resourcemanagement/?cid=stelprdb5347192 (accessed August 10, 2018). (2014). Birds of a neotropical green city: an up-to-date review of the avifauna Carrete, M., and Tella, J. L. (2011). Inter-Individual variability in fear of humans of the city of Xalapa with additional unpublished records. Urb. Ecosyst. 17, and relative brain size of the species are related to contemporary urban invasion 991–1012. doi: 10.1007/s11252-014-0370-3 in birds. PLoS One 6:e18859. doi: 10.1371/journal.pone.0018859 Grinnell, J. (1898). Birds of the Pacific slope of Los Angeles County. Pasadena, NL: Center for Geographical Studies (2015). A Shared Vision for The Survey of Pasadena Academy of Sciences. California Vegetation. Los Angeles, CA: California State University. Gutzat, F., and Dormann, C. (2018). Decaying trees improve nesting opportunities Chace, J. F., and Walsh, J. J. (2006). Urban eects on native avifauna: a review. for cavity-nesting birds in temperate and boreal forests: a meta-analysis and Landsc. Urb. Plan. 74, 46–69. doi: 10.1016/j.landurbplan.2004.08.007 implications for retention forestry. Ecol. Evol. 8, 8616–8626. doi: 10.1002/ece3. City of Los Angeles (2018). 2018 Biodiversity report: Measurement of the 4245 Singapore Index of Cities’ Biodiversity and Recommendations for a Customized Hackett, S. J., Kimball, R. T., Reddy, S., Bowie, R. C. K., Braun, E. L., Braun, M. J., Los Angeles Index. Available at: https://www.lacitysan.org/cs/groups/public/ et al. (2008). A phylogenomic study of birds reveals their evolutionary history. documents/document/y250/mdi0/~{}edisp/cnt024743.pdf (accessed July 2, Science 320, 1763–1768. 2020). Hagen, E. O., Hagen, O., Ibáñez-Álamo, J. D., Petchey, O. L., and Evans, K. L. Clergeau, P., Jokimäki, J., and Savard, J.-P. (2001). Are urban bird communities (2017). Impacts of urban areas and their characteristics on avian functional influenced by the bird diversity of adjacent landscapes. J. Appl. Ecol. 28, diversity. Front. Ecol. Evol. 5:84. doi: 10.3389/fevo.2017.00084 1122–1134. doi: 10.1046/j.1365-2664.2001.00666.x Harrell, F. E. (2004). Hmisc S Function Library. Available at: http://biostat.mc. Croci, S., Butet, A., and Clergeau, P. (2008). Does urbanization filter birds on the vanderbilt.edu/s/Hmisc (accessed October 22, 2019). basis of their biological traits? Condor 110, 223–240. doi: 10.1525/cond.2008. Higgins, L., Pauly, G. B., Goldman, J. G., and Hood, C. (2019). Wild LA. Portland: 8409 Timber Press. Devictor, V., Julliard, R., Couvet, D., Lee, A., and Jiguet, F. (2007). Functional Horns, J. J., Alder, F. R., and Sekercioglu, C. H. (2018). Using opportunistic citizen homogenization eect of urbanization on bird communities. Cons. Biol. 21, science data to estimate average population trends. Biol. Constr. 221, 151–159. 741–751. doi: 10.1111/j.1523-1739.2007.00671.x doi: 10.1016/j.biocon.2018.02.027

Frontiers in Ecology and Evolution | www.frontiersin.org 10 July 2020 | Volume 8 | Article 251

37 fevo-08-00251 July 28, 2020 Time: 18:35 # 11

Cooper et al. Nesting Traits Explain Variation

Iacus, S. M. (2016). sde: Simulation and Inference for Stochastic Dierential Donnely, (Boston, MA: Kluwer Academic), 1–17. doi: 10.1007/978-1-4615- Equations. R package version 2.0.15. 1531-9_1 Jaric,´ I., Correia, R. A., Brook, B. W., Buettel, J. C., and Roll, U. (2020). iEcology: Marzlu, J. M., Clucas, B., Oleyar, M. D., and DeLap, J. (2015). The causal response harnessing large online resources to generate ecological insights. TREE 35, of avian communities to suburban development: a quasi-experimental, 630–639. doi: 10.1016/j.tree.2020.03.003 longitudinal study. Urb. Ecosyst. 19, 1597–1621. doi: 10.1007/s11252-015- Jetz, W., Thomas, G. H., Joy, J. B., Hartmann, K., and Mooers, A. O. (2012). 0483-3 The global diversity of birds in space and time. Nature 491, 444–448. doi: McDonald, R. I., Güneralp, B., Huang, C.-W., Seto, K. C., and You, M. (2018). 10.1038/nature11631 Conservation priorities to protect certebrate endemics from global urban Jetz, W., Thomas, G. H., Joy, J. B., Hartmann, K., Redding, D., and Mooers, A. O. expansion. Biol. Constr. 224, 290–299. doi: 10.1016/j.biocon.2018.06.010 (2014). Distribution and conservation of global evolutionary distinctness in McKinney, M. L. (2006). Urbanization as a major cause of biotic homogenization. birds. Curr. Biol. 24, 919–930. doi: 10.1016/j.cub.2014.03.011 Biol. Constr. 127, 247–260. doi: 10.1016/j.biocon.2005.09.005 Johnston, R. F. (2001). “Synanthropic birds of North America,” in Avian Ecology Merckx, T., Soureau, C., Kaiser, A., Baardsen, L. F., Backeljau, T., Bonte, D., et al. and Conservation in an Urbanizing World, eds J. M. Marzlu, R. Bowman, (2018). Body-size shifts in aquatic and terrestrial urban communities. Nature and R. Donnelly, (Boston, MA: Springer), 49–67. doi: 10.1007/978-1-4615-153 558, 113–116. 1-9_3 Mills, G. S., Dunning, J. B. Jr., and Bates, J. M. (1989). Eects of urbanization on Jokimäki, J., Suhonen, J., Vuorisalo, T., Kover, L., and Kaisanlahti-Jokimäki, J. breeding bird community structure in southwestern desert habitats. Condor 91, (2017). Urbanization and nest-site selection of the black-billed magpie (Pica 416–428. pica) populations in two Finnish cities: from a persecuted species to an urban Møller, A. P. (2009). Successful city dwellers: a comparative study of the ecological exploiter. Landsc. Urb. Plan. 157, 577–585. doi: 10.1016/j.landurbplan.2016.0 characteristics of urban birds in the Western Palearctic. Oecologia 159, 849–858. 8.001 doi: 10.1007/s00442-008-1259-8 Kark, S., Iwaniuk, A., Schalimtzek, A., and Banker, E. (2007). Living in the city: can Møller, A. P. (2010). Interspecific variation in fear responses predicts urbanization anyone become an ‘urban exploiter’? J. Biogeogr. 34, 638–651. doi: 10.1111/j. in birds. Behav. Ecol. 21, 365–371. doi: 10.1093/beheco/arp199 1365-2699.2006.01638.x Møller, A. P. (2014). “Behavioral and ecological predictors of urbanization,” in Kembel, S. W., Cowan, P. D., Helmus, M. R., Cornwell, W. K., Morlon, H., Ackerly, Avian Urban Ecology, eds D. Gil, and H. Brumm, (Oxford: Oxford Univ. Press), S. P., et al. (2010). Picante: R tools for integrating phylogenies and ecology. 54–68. doi: 10.1093/acprof:osobl/9780199661572.003.0005 Bioinformatics 26, 1463–1464. doi: 10.1093/bioinformatics/btq166 Møller, A. P., Diaz, M., Flensted-Jensen, E., Grim, T., Ibanez-Alamo, J. D., Kettel, E. F., Gentle, L. K., Quinn, J. L., and Yarnell, R. W. (2018). The breeding Jokimäki, J., et al. (2015). Urbanized birds have superior establishment success performance of raptors in urban landscapes: a review and meta-analysis. in novel environments. Oecologia 178, 943–950. doi: 10.1007/s00442-015- J. Ornithol. 159, 1–18. doi: 10.1007/s10336-017-1497-9 3268-8 Koch, A. J., Martin, K., and Aitken, K. E. H. (2012). The relationship between Münkemüller, T., Lavergne, S., Bzeznik, B., Dray, S., Jombart, T., Schiers, K., introduced European Starlings and the reproductive activities of Mountain et al. (2012). How to measure and test phylogenetic signal. Method Ecol. Evol. 3, Bluebirds and Tree Swallows in British Columbia, Canada. IBIS 154, 590–600. 743–756. doi: 10.1111/j.2041-210x.2012.00196.x doi: 10.1111/j.1474-919x.2012.01242.x O’Leary, R., and Jones, D. N. (2006). The use of supplementary foods by Australian Lee, S. J., Longcore, T. L., Rich, C., and Wilson, J. P. (2017). Increased home size magpies Gymnorhina tibicen: implications for wildlife feeding in suburban and hardscape decreases urban forest cover in Los Angeles County’s single- environments. Austral. Ecol. 31, 208–216. doi: 10.1111/j.1442-9993.2006. family residential neighborhoods. Urb. Forestry Urb. Greening 24, 222–235. 01583.x doi: 10.1016/j.ufug.2017.03.004 Pagel, M. D. (1999). Inferring the historical patterns of biological evolution. Nature Li, E., Parker, S. S., Pauly, G. B., Randall, J. M., Brown, B. V., and Cohen, B. S. 401, 877–884. doi: 10.1038/44766 (2019). An urban biodiversity assessment framework that combines an urban Paradis, E., Claude, J., and Strimmer, K. (2004). APE: analyses of phylogenetics habitat classification scheme and citizen science data. Front. Ecol. Evol. 7:277. and evolution in R language. Bioinformatics 20, 289–290. doi: 10.1093/ doi: 10.3389/fevo.2019.00277 bioinformatics/btg412 Los Angeles Almanac, (2019). Historical General Population, City & County of Los Pauli, B. P., Spaul, R. J., and Heath, J. A. (2016). Forecasting disturbance eects on Angeles, 1850 to 2010.” Los Angeles Almanac. Copyright 1998-2019 Given Place wildlife: tolerance does not mitigate eects of increased recreation on wildlands. Media, publishing as Los Angeles Almanac. Available at: http://www.laalmanac. Anim. Constr. 20, 251–260. doi: 10.1111/acv.12308 com/population/po02.php (accessed 5 December 2019). Pennell, M. W., Eastman, J. M., Slater, G. J., Brown, J. W., Uyeda, J. C., FitzJohn, Loss, S. R., Will, T., and Marra, P. P. (2013). The impact of free ranging domestic R. G., et al. (2014). geiger v2.0: an expanded suite of methods for fitting cats on wildlife of the United States. Nat. Commun. 4:1396. macroevolutionary models to phylogenetic trees. Bioinformatics 30, 2216–2218. Lowry, H., Lill, A., and Wong, B. B. M. (2013). Behavioural responses of wildlife to doi: 10.1093/bioinformatics/btu181 urban environments. Biol. Rev. 88, 537–549. doi: 10.1111/brv.12012 Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D., and R Core Team, (2019). nlme: MacGregor-Fors, I. (2010). How to measure the urban-wildland ecotones: Linear and Nonlinear Mixed Eects Models. R Pack. Vers. 3.1-143. Redefining ‘peri-urban’ areas. Ecol. Res. 25, 883–997. QGIS Development Team, (2016). QGIS Geographic Information System. Open MacGregor-Fors, I., and Ortega-Alvarez, R. (2011). Fading from the forest: Source Geospatial Foundation Project. Available at: http://qgis.osgeo.org bird community shifts related to urban site-specific and landscape traits. (accessed September 4, 2019). Urb. Forestry Urb. Greening 10, 239–246. doi: 10.1016/j.ufug.2011. Revell, L. J. (2012). Phytools: an R package for phylogenetic comparative biology 03.004 (and other things). Method Ecol. Evol. 3, 217–223. doi: 10.1111/j.2041-210x. Martin, P. R., and Bonier, F. (2018). Species interactions limit the occurrence of 2011.00169.x urban-adapted birds in cities. PNAS 115, E11495–E11504. Reynolds, S. J., Ibáñez-Álamo, J. D., Sumasgutner, P., and Mainwaring, M. C. Martins, E. P. (1994). Estimating the rate of phenotypic evolution from (2019). Urbanisation and nest building in birds: a review of threats and comparative data. Am. Nat. 144, 193–209. doi: 10.1086/285670 opportunities. J. Ornith. 160, 841–860. doi: 10.1007/s10336-019-01657-8 Marzlu, J. M. (2005). Island biogeography for an urbanizing world: how Samia, D. S. M., Nakagawa, S., Nomura, F., Rangel, T. F., and Blumstein, D. T. extinction and colonization may determine biological diversity in human- (2015). Increased tolerance to humans among disturbed wildlife. Nat. Commun. dominated landscapes. Urb. Ecosyst. 8, 157–177. doi: 10.1007/s11252-005- 6:8877. 4378-6 Santini, L., González-Suárez, M., Russo, D., Gonzalez-Voyer, A., Von Hardenberg, Marzlu, J. M. (2016). A decadal review of urban ornithology and prospectus for A., and Ancillotto, L. (2019). One strategy does not fit all: determinants of the future. IBIS 159, 1–13. doi: 10.1111/ibi.12430 urban adaptation in mammals. Ecol. Lett. 22, 365–376. doi: 10.1111/ele. Marzlu, J. M., Bowman, R., and Donnely, R. (2001). “A historical perspective 13199 on urban bird research: trends, terms, and approaches,” in Avian Conservation Santos, M. J., Smith, A. B., Thorne, J. H., and Moritz, C. (2017). The relative and Ecology in an Urbanizing World, eds J. M. Marzlu, R. Bowman, and R. influence of change in habitat and climate on elevational range limits in small

Frontiers in Ecology and Evolution | www.frontiersin.org 11 July 2020 | Volume 8 | Article 251

38 fevo-08-00251 July 28, 2020 Time: 18:35 # 12

Cooper et al. Nesting Traits Explain Variation

mammals in Yosemite National Park, California, U.S.A. Clim. Change Responses Vitousek, P. M., Mooney, H. A., Lubchenco, J., and Melillo, J. M. (1997). Human 4:7. domination of Earth’s ecosystems. Science 277, 494–499. doi: 10.1126/science. Schoenherr, A. A. (1992). A Natural History of California. California Natural 277.5325.494 History Guides: 56. Berkeley, CA: Univ. of California Press. West-Eberhard, M. J. (1989). Phenotypic plasticity and the origins of diversity. Shultz, A. J., Tingley, M. W., and Bowie, R. C. K. (2012). A century of avian Ann. Rev. Ecol. Syst. 20, 249–278. community turnover in an urban green space in northern California. Condor White, J. G., Antos, M. J., Fitzsimons, J. A., and Palmer, G. C. (2005). Non- 114, 258–267. doi: 10.1525/cond.2012.110029 uniform bird assemblages in urban environments: the influence of streetscape Silva, C. P., García, C. E., Estay, S. A., and Barbosa, O. (2015). Bird richness and vegetation. Landsc. Urb. Plan. 71, 123–135. doi: 10.1016/j.landurbplan.2004. abundance in response to urban form in a Latin American city: valdivia, Chile 02.006 as a case study. PLoS One 10:e0138120. doi: 10.1371/journal.pone.0138120 Wilman, H., Belmaker, J., Simpson, J., de la Rosa, C., Rivadeneira, M. M., Slabbekoorn, H., and den Boer-Visser, A. (2006). Cities change the songs of birds. and Jetz, W. (2014). EltonTraits 1.0: species-level foraging attributes of Curr. Biol. 16, 2326–2331. doi: 10.1016/j.cub.2006.10.008 the world’s birds and mammals. Ecology 95:2027. doi: 10.1890/13-1 Sol, D., Gonzalez-Lagos, C., Lapiedra, O., and Diaz, M. (2017). “Why 917.1 are exotic birds so successful in urbanized environments?,” in Ecology Wingfield, J. C., Krause, J. S., Perez, J. H., Chmura, H. E., Németh, Z., Word, and Conservation of Birds in Urban Environments, eds E. Murgui, and K. R., et al. (2015). A mechanistic approach to understanding range shifts in M. Hedblom, (Boston, MA: Springer), 75–89. doi: 10.1007/978-3-319-43 a changing world: what makes a pioneer? Gen. Comp. Endocrinol. 222, 44–53. 314-1_5 doi: 10.1016/j.ygcen.2015.08.022 Sol, D., González-Lagos, C., Moreira, D., Maspons, J., and Lapiedra, O. (2014). Wood, E. M., and Esaian, S. (2020). The importance of street trees to urban Urbanisation tolerance and the loss of avian diversity. Ecol. Lett. 17, 942–950. avifauna. Ecol. Applic. (accessed July 19, 2020). doi: 10.1111/ele.12297 Yeh, P. J. (2004). Rapid evolution of a sexually selected trait following population Sol, D., Lapiedra, O., and González-Lagos, C. (2013). Behavioural adjustments for a establishment in a novel habitat. Evolution 58, 166–174. doi: 10.1111/j.0014- life in the city. Anim. Behav. 85, 1101–1112. doi: 10.1016/j.anbehav.2013.01.023 3820.2004.tb01583.x Stein, E. D., Dark, S., Longcore, T., Hall, N., Beland, M., Grossinger, R., et al. (2007). Historical Ecology and Landscape Change of the San Gabriel River Conflict of Interest: The authors declare that the research was conducted in the and Floodplain. Costa Mesa, CA: Southern California Coastal Water Research absence of any commercial or financial relationships that could be construed as a Project. potential conflict of interest. Swarth, H. S. (1900). Avifauna of a 100-acre ranch. Condor 2, 14–16. doi: 10.2307/ 1361152 Copyright © 2020 Cooper, Shultz and Blumstein. This is an open-access article Tomasevic, J. A., and Marzlu, J. M. (2017). Cavity nesting birds along distributed under the terms of the Creative Commons Attribution License (CC BY). an urban-wildland gradient: is human facilitation structuring in the The use, distribution or reproduction in other forums is permitted, provided the bird community? Urb. Ecosyst. 20, 435–448. doi: 10.1007/s11252-016- original author(s) and the copyright owner(s) are credited and that the original 0605-6 publication in this journal is cited, in accordance with accepted academic practice. No Unitt, P. (2004). San Diego County Bird Atlas. El Cajon, CA: Sunbelt Publications. use, distribution or reproduction is permitted which does not comply with these terms.

Frontiers in Ecology and Evolution | www.frontiersin.org 12 July 2020 | Volume 8 | Article 251

39 Chapter 4. Is there a “global urban raptor”? Using community science data to identify traits associated with urban occurrence in Accipiteridae

Abstract: Large, diverse, and widely-distributed, raptors provide an ideal taxonomic group to study wildlife use of urban environments. Based on prior studies and our own observations in the

U.S. and elsewhere, I hypothesized that small-bodied, generalist raptors would likely be more common than larger specialist species in urban areas, and that this pattern would be consistent in cities throughout the world. I analyzed the status of more than 100 species of hawks (Family:

Accipiteridae) using community-science records from 62 cities on five continents, modeling three indices of occurrence with six ecological traits. I calculated occurrence in three ways, based on records between 2014-2018: urban abundance, the frequency of breeding season reports within 10 km of city centers, species proportion, or the relative abundance of each species within its city of occurrence, and urban preference, a ratio of urban abundance to sighting frequency within a “peripheral band” located 10 to 100 km from each city center. Based on results from the best of four models including alternative phylogenetic approaches, each index was significantly negatively associated with body mass, while urban abundance and urban proportion were significantly positively associated with nest substrate breadth, and urban abundance was also positively associated with habitat breadth. Overall, large-bodied raptors appear to be rare or absent from urban areas, and while some smaller-bodied raptors are also rare in cities, those that are present tend to be small to mid-sized. This study provides another example of the power of community science datasets to reveal patterns that might be obscured by studies limited to a small number of cities, those using a simple binary value for urban/non-urban occurrence, or those testing only numerical abundance without considering relative abundance of species within a community. Supplementary materials included at the end of this manuscript, following the

40 Discussion.

Introduction

Intuitively, the dominance of a “mid-sized generalist” within urban wildlife communities seems to be a pattern repeated around the world. In California where I work, coyotes (Canis latrans) weigh approximately 20 kg and are among of the most common (wild) urban carnivores, rather than the much larger mountain lion (Puma concolor; 100 kg) or the smaller gray fox (Urocyon cinereoargenteus 5 kg). In heavily-visited parks in Costa Rica, the primate one frequently sees is the white-fronted capuchin (Cebus albifrons), an average-sized monkey (3.5 kg), rather the larger black howler-monkey (Alouatta caraya; 7 kg) and spider monkey (Ateles geoffroyi; 9 kg), or the tiny squirrel monkeys (Saimiri sciureus; 1.5 kg). In urban and suburban areas across North

America, the Cooper’s Hawk (Accipiter cooperii; 450 g), a mid-sized species within the genus

Accipiter, is much more likely to be found in common in urban areas than either the larger

Northern Goshawk (A. gentilis; 900 g) or the smaller Sharp-shinned Hawk (A. striatus; 150 g)

(sighting information from www.inaturalist.org; weights generalized from multiple sources, including Jameson and Peters 1988 and Emmons 1990).

Raptors (Family: Accipitridae) provide an ideal group to study urban adaptation, as they exhibit a vast variety of morphological traits, with species ranging in size from diminutive sparrowhawks (Accipiter spp.) to massive Old World Vultures (Gyps spp.), occupying a broad range of ecological niches on every continent except Antarctica. Many raptor species, such as

Cooper’s Hawks, are clearly thriving in urban landscapes, nesting in built structures and planted introduced trees, and feeding on human-subsidized urban prey (McCabe et al. 2018, Rosenfield

41 et al. 2018). Other species are restricted to wildland habitats, including many tropical species and several single-island endemics, their biology comparatively poorly-known (McClure et al. 2018,

Beuchley et al. 2019). While urban raptor communities have been little-studied, as the world urbanizes (Seto etc. 2011), many species will need to adapt to some level of human disturbance to survive. Conversely, species that avoid co-existence with humans must be assigned higher priority for conservation. Here I use insights from a recent analysis of functional traits of nesting bird species associated with urban cover in the Los Angeles, California (US) area (Cooper et al.

2020b) to both narrow our focus phylogenetically (to Accipitridae), and to broaden it geographically to include more than 60 cities on multiple continents.

Many studies have investigated ecological traits associated with urban life in birds (see reviews by Chace and Walsh 2006, Marzluff 2016), but only a handful have examined raptor communities explicitly (e.g., Marti et al. 1993). While numerous authors have investigated demographics and distribution of urban raptors (often through single-species studies, but see

White et al. 2018), studies of ecological traits of these urban raptor communities remain comparatively rare. Kettel et al. (2018) reviewed various reproductive traits in urban versus wildland raptor populations (e.g., clutch size) across various countries through a meta-analysis but found little consistent pattern, concluding that prey availability must play the greatest role in determining raptor success in urban environments. Boal (2018) analyzed eight traits associated with raptor occurrence in 14 US capitals (all with populations > 100,000), finding that both diet breadth and preferred “normal” (non-urban) habitat type were strong predictors of presence in urban areas during both winter and summer. However, this study was largely limited to the US, and did not take species’ abundance into account.

42 Globally, it would seem that the most common urban raptors tend to be smaller species than those in wildland areas (e.g., small hawks and kites, versus eagles), and while not part of the family Accipitridae, falcons (family Falconidae) occupy a similar niche, average smaller than hawks, and have many familiar urban representatives such as the Peregrine Falcon (Cade et al.

1996). In a recent analysis of a suburban raptor community around Los Angeles, California (US),

Cooper et al. (2020a) showed how as the region urbanized over five decades, active nests of one midsized raptor, the Cooper’s Hawk, had greatly increased, while both the largest diurnal raptor

(Golden Eagle Aquila chrysaetos) and smallest ones (White-tailed Kite Elanus leucurus,

American Kestrel Falco sparverius) in the study area had become extirpated or nearly so. In

Reno, Nevada (US), White et al. (2018) also found Cooper’s Hawks and Sharp-shinned Hawks

(Accipiter striatus; lumped as “Accipiters”) among the most urban-tolerant at various landscape scales of eight taxa examined, and found the largest species treated, the Golden Eagle, to be the least tolerant. Still, small body mass has not been empirically shown to be a universal pattern for raptors in cities, and enough exceptions exist that prior authors have struggled to find a clear link

(e.g., Croci et al. 2008, Sol et al. 2014, Santini et al. 2019). For example, at least until recently, massive Old World Vultures were a familiar sight at garbage dumps and other urban zones in

Africa and India prior to precipitous declines linked to poisoning (Cuthbert et al. 2011).

I speculate that in addition to a smaller size, life history traits of mid-sized raptors might be allowing them to dominate urban raptor communities, and suggest that all cities around the world might also have a “Cooper’s Hawk analogue” that is doing similarly well. I hypothesize that a broad diet (which includes birds and mammals, including non-native, urban-associated prey items), sedentary nature (most populations, particularly in urban areas, are non-migratory), widespread distribution, and an ability to utilize a variety of habitats may all be factors

43 contributing to raptor success in urban environments. Here I take advantage of the millions of records from birders that have been entered via the community-science platform eBird

(www.ebird.org), a resource not available historically, in gathering breeding-season records of raptors in and around cities with relatively high usage of eBird. I calculate urban occurrence in three ways for each focal species, and model these values using six morphological and ecological traits to explore the existence of a “global urban raptor” with shared life history characteristics that may be enabling adaptation to urban life.

Methods

Data preparation

I used GBIF (GBIF 2020) to download eBird sightings of hawks (family Accipitridae; here referred to as “raptors”) from nine countries encompassing a range of global biomes (USA,

Mexico, Colombia, Brazil, Spain, Great Britain, South Africa, India, and Australia) within a recent five-year time period (2014-2018). Countries were selected on the basis of having a high number of sightings submitted to eBird1, and their geographical separation (i.e., I selected countries where overlap in breeding species was likely to be low). I selected records for months when our target species would likely be breeding, which varied by country. I then selected the largest cities in each country based on total population, eliminating cities that had fewer than ca.

1M people, and those with fewer than ca. 50 eBird submissions of Accipitridae during the period. I visually estimated the rough center of the urban extent of these cities (including

1 https://ebird.org/region/world/regions?yr=all&m=&hsStats_sortBy=cl&hsStats_o=desc 44 adjacent “suburbs”) using the most recent aerial imagery on the Apple Maps application on the iPhone (ver. 14.2), and used the Geosphere package (Hijmans et al. 2015) in R (R Core Team

2020; version 4.0.0) to find records at two radial distances from the urban center (10 km, 100 km). I then subtracted records inside 10 km from those inside 100 km to find records only within the 10 to 100 km “peripheral band” each city. These distances are somewhat arbitrary, but were selected for comparison across continents to approximate the zone of the most highly-urbanized city core and the less densely-settled outskirts, which may include agricultural lands, arid scrub, humid forest, ocean, or any number of habitats depending on the city used (mixed with areas of urban development of various sizes).

I made every effort to exclude eBird reports of transients or lingering winter visitors prior to analysis. To avoid counting species that may have been vagrants to urban areas, I eliminated species with fewer than 15 records over five years within 100 km of the urban center of the cities used (n=25; see Boal 2018). I then eliminated “single-city species” (those found in only one of the 62 cities analyzed; n=20), which Ihypothesized might be due to unique local conditions rather than a generalizable response. This resulted in 62 cities, 90 raptor species, and 612 unique city-species status combinations (as an example of one city-species combination, Accipiter cirrocephalus had 283 records within 100 km of the center of Sydney, 21 records within 10 km).

Our final dataset of 90 focal species represents ca. 31% of the world’s 285 widely-recognized raptor (Accipitridae) species, and 51% of its genera (34 of 67; Del Hoyo et al. 2013, BirdLife

International 2019). I condensed these data into a dataset comprised of one entry per species, taking the average value for each urban index (abundance and preference) for each species across all cities in which they were recorded Please refer to Supplemental Material for lists of cities analyzed (Table S1) and raptor species evaluated (Table S2).

45 I recognize that community-science data present many unique challenges with respect to addressing survey/observer effort. Survey effort could impact distributional analyses in several ways. For example, insufficient effort may result in less common species being undercounted, or missed entirely. To address this, I intentionally selected countries with the highest levels of eBird participation, and eliminated the rarest species (see above). Second, survey effort across a region might be unevenly distributed, which could result in more observational effort outside the urban core of a city than inside, thus decreasing the actual urban preference of a given species.

Conversely, if more effort occurred within a city than in the peripheral zone outside it, a species’ preference for urban areas might be inflated. I addressed this second concern by amassing a large number of species-city combinations (n=612), recognizing that each city differs in the number and location of sites visited by birders during the breeding season, and each has a different ratio of urban versus rural birding sites. Each species used in our analysis was found in over 5 cities, on average. Finally, species’ behavioral differences could result in certain raptor species being more or less obvious than others, affecting sighting number and ratios. To address this, I intentionally selected only diurnal raptors as a focal group due to their conspicuousness and popularity with birders.

Measuring Urban Occurrence

I used three indices of occurrence to assess the presence of raptors in urban areas, as summarized in Table 1.

46 Table 1. Three urban index variables used to calculate raptor occurrence in urban areas.

Index Measures Calculation

Urban Abundance Numerical abundance of Density (N/area) of reports within 10

each species in urban km of urban center

center

Species Relative abundance of each Percentage of reports of given species

Representation species in urban center relative to number of reports of other

raptor species, within 10 km of urban

center

Urban preference Urban bias of each species Density of reports (N/area) within 10

km / density of reports in 10-100 km

band

Each of these metrics was averaged across cities for each species, resulting in a single entry per species. I note that some species may be numerically scarce overall, but may nonetheless have a high urban preference if they occur more frequently in urban areas than in the surrounding landscape. Thus, the urban preference is not a measure of abundance in urban areas per se, but of degree of preference for more urbanized habitat relative to less urbanized habitat (assuming roughly equal effort within a 100-km band around the urban center point). Of course, some species may be common in the city, and also common in the peripheral band surrounding it, and would thus could still have a fairly low urban preference if there were simply more sightings in

47 the peripheral band. Indeed, all three measurements are useful to express species’ status, and different species had the highest values of each (Figure 1).

48 Figure 1. The relationship between three urban occurrence variables (urban abundance, species proportion, and urban preference) using 90 focal raptor species across 62 cities. Note: I have labeled species with the highest values of each urban index.

49 Traits

I evaluated variables most likely to influence urban occurrence based on prior research on urban birds, and urban raptors in particular (e.g., Samia et al. 2015, Boal 2018, Cooper et al. 2020b;

Table 2). Trait values were taken from a database maintained by Cagan Sekerciglou, University of Utah (see also Table S3; Sekercioglu et al. 2004; updated with Del Hoyo et al. 2013 with nomenclature updates following Birdlife International 2019). In cases of missing trait values, I used Ferguson-Lees and Christie (2001) and accounts from the Global Raptor Information

Network (The Peregrine Fund 2020) to calculate estimated values (Table S4). Due to high collinearity between artificial nest substrate and nest substrate breadth (r > 0.6), I eliminated the former both because its correlation with urban occurrence is already well established (e.g.,

Cooper 2020b) and because substrate breadth seemed more informative for a wider range of species (<10 species Itreated are known to nest on artificial structures). Ultimately, I selected six trait variables for the models: mass, diet breadth, habitat breadth, migratory status, nest substrate breadth and number of cities where recorded (see Table S5 for correlation matrix of urban indices and traits).

50 Table 2. Traits used in analysis.

See Table S3 for detail on trait variables used.

Variable Type Description

Mass Numeric; grams LN transformed

Diet breadth Numeric; 1-6 Calculated from 9 major food categories.

Habitat breadth Numeric; 1-10 Calculated from 15 habitat types.

Nest substrate breadth Numeric; 1-6 Calculated from 12 categories.

Migratory status Factor; 0/1/2 Includes non-migratory (0), full migrants

(1) and partial migrants (2)

Number of cities where Numeric 1-10 cities

recorded

Statistical Analysis

To account for phylogenetic relatedness among species in our analyses, I used the latest phylogeny of Accipitridae from the Open Tree of Life (2019, ver. 3.1), which represents a synthetic tree derived from multiple sources of phylogenetic information. I matched species names and assigned labels using the rotl (ver. 3.0.10) interface to the Open Tree of Life in R. I first tested for a phylogenetic signal in each urban index by fitting a series of models that employed three different modes of evolution: (Brownian motion, Pagel’s lambda, Ornstein–

Uhlenbeck (OU), as well as a non-phylogenetic model (see Münkemüller et al., 2012). I used the ape (ver. 5.4-1; Paradis et al., 2004), geiger (ver. 2.0.7; Pennell et al., 2014), and picante (ver.

51 1.8.2; Kembel et al., 2010) packages in R, and used the phylosignal function to analyze the focal species’ urban preference relative to their corresponding positions on the phylogenetic tree described above. I tested a Brownian motion, or random-walk model, using a Blomberg’s K test

(Blomberg et al., 2003), which compares the variance of phylogenetic independent contrasts to what I would expect under a Brownian motion (BM) model. Here, K = 1 means that relatives resemble one another as much as I should expect under BM; K < 1 means that there is less

“phylogenetic signal” than expected under BM, while K > 1 means that there is more. I then analyzed the urban preference and tree data using Pagel’s lambda (Pagel, 1999). Here, if our estimated lambda = 0, then the traits are inferred to have no phylogenetic signal. Lambda = 1 corresponds to a Brownian motion model; 0 < lambda < 1 is intermediate. Next, I used a model which employed the Ornstein–Uhlenbeck (OU) mode of evolution which incorporates stabilizing selection wherein the trait is drawn toward a fitness optimum, or long-term mean, rather than being completely random and directionless (Martins, 1994). This model has two terms, alpha, which represents the strength of the pull toward the fitness optimum, where alpha = 0 is no pull, and the same as a Brownian motion model and the larger the alpha value the stronger the pull, and sigma2, which is the dispersion of the data (Martins, 1994). To test for no phylogenetic signal, I also used a “no-signal” generalized least squares model where lambda was set to 0.

I then tested associations separately between each of our three urban indices and the six traits by fitting three phylogenetic generalized least squares (PGLS) models in the phytools package in R (ver. 0.7-70; Revell, 2012), and one non-phylogenetic general linear model in the nlme package in R (ver. 3.1-147; Pinheiro et al., 2019), and compared AIC values of each to select the model that best explained variation in the data. Ifirst used the gls function and incorporated a Brownian motion mode of evolution using the corBrownian function. I then fitted

52 a second PGLS test using the Ornstein–Uhlenbeck (OU) mode of evolution using the corMartins function, and a third PGLS test using Pagel’s lambda test with the corPagel function. Finally, I fitted a non-phylogenetic least squares model to compare with the PGLS tests. For all analyses, best fit parameters of the phylogenetic model were estimated with maximum likelihood. Lastly, I checked residuals for normality using QQ tests, and selected the analysis with the lowest AIC values as the best model for each urban index tested.

Results

Testing the three indices of urban occurrence (abundance, proportion, and preference), I found no evidence of phylogenetic signal using three phylogenetically-informed models (Brownian motion, OU, and Pagel’s lambda). Pagel’s lambda proved the best model for species proportion, while non-phylogenetic models returned the lowest AICc value for the remaining two indices, urban abundance and urban preference (Table 3).

Modeling the six trait variables with each urban occurrence index, I found the OU,

Pagel’s Lambda, and non-phylogenetic models returned similarly low AIC values, with parameters suggesting little phylogenetic signal in the data for models incorporating phylogeny

(Table 4). Results were robust across phylogenetic and non-phylogenetic models, but Ireport the best model for each index below (Table 5). I found body mass to be significantly (negatively) associated with each index in the best model identified for each. Looking at the other trait variables, both urban abundance and species proportion were also significantly positively associated with nest substrate breadth, and urban abundance was significantly positively associated with habitat breadth (Table 5). None of the three urban indices were found to be associated with either migratory status nor number of cities where present.

53 Table 3. Comparison of phylogenetically signal of each of the urban indices using three modes of evolution (Brownian, OU, Pagel’s lambda) and one non-phylogenetically informed model.

Note: models with the lowest AICc values for each index in italics.

Model Test statistic AICc

Urban Abundance

Brownian K = 0.046, P = 0.254 216.424

OU Sigma squared: 8.241; Alpha: 2.718 192.907

Pagel’s Lambda Lambda = 0.008; P = 0.929 144.981

Non-phylogenetic Sigma squared: 0.273 142.848

Species Proportion

Brownian K = 0.059, P = 0.026* -83.438

OU Sigma squared: 0.304; Alpha: 2.718 -103.937

Pagel’s Lambda Lambda = 5.529; P = 0.019 -139.095

Non-phylogenetic Sigma squared: 0.012 -135.706

Urban Preference

Brownian K = 0.035, P = 0.658 808.198

OU Sigma squared: 5815.073; Alpha: 2.718 783.228

Pagel’s Lambda Lambda < 0.001; P = 1 707.892

Non-phylogenetic Sigma squared: 142.284 705.751

54 Table 4. Comparison of AIC scores of four models used to test three urban indices against six traits.

Note that the scores of OU, Pagel’s Lambda and the non-phylogenetic model are all very close, suggesting minimal influence of phylogeny in explaining variation.

Model Urban Abundance Species Proportion Urban Preference

Brownian 215.007 -1.624 749.933

OU 148.936 -91.87696 679.5575

Pagel’s Lambda 155.6126 -92.55431 677.6304

Non-phylogenetic 153.7329 -93.56996 680.2447

55 Table 5. Results from the best model for each urban index using Generalized Least Squares tests, fitted to explain variation based on six traits analyzed.

Index (Best model) Mean (±SE), P-value

Urban Abundance (OU)

Intercept 0.293 ± 0.404, P = 0.471

Trait

LN(body mass) -0.132 ± 0.063, P = 0.039*

Diet Breadth -0.042 ± 0.046, P = 0.369

Habitat Breadth 0.105 ± 0.035, P = 0.003*

Migratory status (1) 0.119 ± 0.098, P = 0.226

Migratory status (2) 0.042 ± 0.192, P = 0.827

Substrate breadth 0.209 ± 0.070, P = 0.004*

N Cities 0.016 ± 0.011, P = 0.161

Species Proportion (Non Phylogenetic)

Intercept 0.236 ± 0.088, P =0.009

Trait

LN(body mass) -0.044 ± 0.014, P = 0.002*

Diet Breadth -0.004 ± 0.011, P = 0.732

Habitat Breadth 0.013 ± 0.008, P = 0.1278

Migratory status (1) 0.023 ± 0.023, P = 0.312

Migratory status (2) 0.069 ± 0.045, P = 0.130

56 Substrate breadth 0.035 ± 0.017, P = 0.041*

N Cities 0.004 ± 0.003, P = 0.172

Urban Preference (Pagel’s Lambda)

Intercept 36.382 ± 9.912, P = 0.0004

Trait

LN(body mass) -5.305 ± 1.497, P = 0.0007*

Diet Breadth 2.217 ± 1.165, P = 0.061

Habitat Breadth 1.266 ± 0.879, P = 0.153

Migratory status (1) -2.872 ± 2.532, P = 0.260

Migratory status (2) 1.766 ± 4.901, P = 0.719

Substrate breadth 0.973 ± 1.832, P = 0.597

N Cities 0.075 ± 0.300, P = 0.804

Discussion

I explored urban occurrence in raptors using three different urban indices: urban abundance, species proportion, and urban preference. Using three phylogenetically-informed models and one non-phylogenetic model, I found little evidence of phylogenetic signal in either our three indices, nor in the full models using these indices with six life history traits. All three urban indices were significantly (negatively) associated with body mass in the best models for each. Given the lack of phylogenetic signal, this suggests that raptor taxa found in cities tend to be of small body mass regardless of relatedness, in terms of a species’ numerical abundance, its proportion of that raptor

57 community, and its preference of the urban core over peripheral areas surrounding the cities. I note that the largest raptor species (i.e., those above 1,000 g) are either missing from the urban core and/or show no indication of urban preference.

Theories about the drivers of body size variation date to Bergmann’s Rule (1847), which holds that within a broadly-distributed clade, animal species (and often populations) tend to have larger body sizes at higher latitudes and colder environments. This geographic pattern has been documented for birds in some studies (e.g., James 1970 and Evans et al. 2009), but has been questioned as applicable to all species (Geist 1987). More recently, declining body size has been linked to climatic warming across bird populations (e.g., Lurgi et al. 2012, Weeks et al. 2020).

Merckx et al. (2018) suggested that in part because urban areas are influenced by the heat island effect, smaller body size favors species with increased dispersal capability and reduced metabolic needs. However, mobility and dispersal ability would mostly apply to strongly terrestrial groups such as mammals and reptiles, so other factors are likely involved for raptors, which are all strong flyers. Future work could investigate other aspects of body size, such as wing loading, which might enable types of flight or foraging methods conducive to life in urban areas.

The additional significant (positive) associations with habitat breadth and nest substrate breadth found for both urban abundance and urban proportion (both of which measure numerical abundance) suggests that the most common breeding raptors in urban areas are habitat generalists, both in terms of absolute numbers (urban abundance) and relative abundance

(species proportion). These associations reflect prior findings that generalists thrive in cities while urban avoiders show a narrower habitat tolerance (Croci et al. 2008, Sol et al. 2014).

Habitat generalists would be able to exploit many types of vegetation communities wherever they occur, and their sheer abundance might allow them to readily find mates and thus occupy a broader geographic area than scarcer species would. It is also likely that the most common urban species, like Red-tailed Hawks in the U.S., are simply the most abundant raptor species overall in the regions where they occur (i.e., not just within cities), so their abundance within the urban core represents “spillover” into the city from peripheral areas (Red-tailed Hawk had the highest number of records within 100 km of the cities in which it occurred of any species evaluated).

Further investigation into actual habitat usage within urban areas would elucidate some of these patterns, given how variable habitat can be from city to city (see Dykstra 2018 for discussion).

Urban preference is more nuanced, since it is not a measure of abundance, but rather an index of how biased a species is in selecting urban areas over habitats on the periphery of the city which are assumed to be less urban. While only body mass was found to be significantly

(negatively) associated with urban preference, diet breadth was nearly significantly (positively) associated (P = 0.061). Prey type and availability are understood to be key to the maintenance of raptor communities, including those along an urban gradient (e.g., Rullman and Marzluff 2014).

Prior studies have shown that diet generalist birds not only tend to populate cities, but thrive there and may favor urban areas over the surrounding landscape (see Palacio 2019). Prey type (as opposed to diet breadth) may also influence an urban bias toward smaller raptor body mass, as the most common urban prey items tend to also be small or moderately-sized compared to all prey items consumed by raptors elsewhere (e.g., doves Zenaida spp., small birds, and rats Rattus spp., versus larger mammals). Intriguingly, these patterns may be reversed with some species by the provision of “predictable anthropogenic food subsidies” (Oro et al. 2013, Shochat 2004). For raptors, these novel anthropogenic food subsidies such as massive garbage dumps within urban 59 areas may have enabled massive Old World Vultures to persist in cities, at least prior to their decline due to poisoning (see Cuthbert et al. 2011), and larger sizes of urban carnivores have been documented in parts of Israel that receive a “garbage subsidy” unavailable to non-urban populations (Yom-Tov 2003).

A potential weakness of using trait breadth (of diet, habitat, and nest substrate) as a variable is that it does not distinguish between a species that is ubiquitous and flexible enough to live anywhere (including “simplified” habitats such as residential yards), and one that require a diversity of a resource within a territory (e.g., extensive grassland near tall trees along a stream).

As a source of data, I recognize that eBird has potential limitations that could not be controlled in our study, including observer bias (over-reporting the same individual, or under-reporting a familiar species due to its abundance). While eBird is excellent for determining seasonal status and distribution (and comparable to existing standardized survey methods; see Horns et al.

2018), it cannot be used for assessing demographics or nesting success. Yet, Iknow of no similar

“global portrait” of urban bird occurrence that approaches eBird in its comprehensiveness. I hope our conservative approach to community science records, which relied on culling records by season, and using a variable that reached beyond simple presence/absence, can serve as a model for future work.

I also caution against treating increased urban occurrence of any wildlife species

(including urban raptors) as an unqualified success. Sol et al. (2020) and Bregman et al. (2016) discuss the loss of functional diversity in urban species assemblages, which in the long term may lead to loss of global biodiversity as species fail to adapt faster than their habitats urbanize. And,

I note that while being a smaller-bodied generalist raptor may give a species a chance to thrive in urban areas, it does not guarantee their success there. Many small and mid-sized raptor species 60 were found to be scarce in urban areas (urban abundance values near zero), and many do not appear to prefer urban areas over non-urban ones (urban preference values near zero).

Furthermore, our study did not compare traits of individuals within the same species (where, for example, smaller individuals of the same species might have reduced fitness; see Liker et al.

2008). Thus, I cannot draw any conclusions about the long-term outlook for the health or productivity of urban raptors through this analysis.

Ultimately, there appears to be no single genus of raptor that succeeds in urban areas globally, and the lack of phylogenetic signal detected in our variables supports this. So, while

Accipiter (a familiar urban-occurring genus in many regions of the world) accounts for half of the 12 taxa with the highest urban preference values in our study, membership in the genus

Accipiter does not automatically confer an advantage in urban areas – in Latin America, two unrelated species, gray hawk (Buteo plagiatus) and roadside hawk (Rupornis magnirostris) have effectively “assumed the Accipiter role” in the largest cities of Columbia and Brazil, for example. Yet these patterns may not always be apparent at the city or even country scale due to the wide variation in the overall number of raptor species from region to region, and the habitat differences among cities. For example, I found only four nesting raptor species consistently near cities in the United Kingdom, versus several dozen in and around tropical cities in Colombia,

Brazil and India. Increased participation in eBird across all countries outside the US and Canada, particularly in urban areas (as opposed to, say, more remote national parks where many birders are likely drawn) would refine future analyses, as would a comparison with patterns found in winter raptor communities in cities, though this may be unlikely to change overall results since so many raptors are non-migratory (though several large groups have migratory representatives, including Accipiter, Aquila, Buteo and Circus). Repeating the study for other taxonomic groups

61 would also be worthwhile to test whether the patterns observed for raptors are universal. Finally, more research into the mechanisms affecting raptor occurrence in urban areas, such as those involving diet studies, nest-searching and monitoring, and demographic research (such as nesting success) would help fill gaps in our knowledge of urban wildlife and allow better planning for future ecological changes.

Supplemental Information

Table S1. Cities used, including coordinates used for urban center and months used for breeding season records (2014-2018).

City Country Coordinates Months used (range) Adelaide Australia 138.6002, -34.84805 Sept. – January Brisbane Australia 153.0400, -27.4521 Sept. – January Melbourne Australia 145.0182, -37.7545 Sept. – January Perth Australia 115.8612, -31.9493 Sept. – January Sydney Australia 151.1370, -33.8697 Sept. – January Belo Horizonte Brazil -43.940013, -19.911168 July – January Brasilia Brazil -47.953068, -15.840663 July – January Curitiba Brazil -49.233399, -25.441625 July – January Fortaleza Brazil -38.548099, -3.787465 July – January Manaus Brazil -59.986931, -3.062840 July – January Rio de Janeiro Brazil -43.3666653, -22.804755 July – January Salvador Brazil -38.476446, -12.974050 July – January Sao Paolo Brazil -46.638580, -23.548892 July – January Barranquilla Colombia -74.7973, 10.9724 April – July Bogota Colombia -74.0834, 4.6368 April – July Cali Colombia -76.5182, 3.4534 April – July Cartegena Colombia -75.483030, 10.39318 April – July Medellin Colombia -75.5721, 6.2653 April – July Ahmedabad India 72.587981, 23.0380 December – May Bangalore India 77.5891, 12.9683 December – May Chennai India 80.2459, 13.0865 December – May Delhi India 77.2312, 28.6122 December – May Hyderabad India 78.487531, 17.443445 December – May Jaipur India 75.790503, 26.902555 December – May Kolkata India 88.3664, 22.5433 December – May Lucknow India 80.931553, 26.849800 December – May Mumbai India 72.8749, 19.0964 December – May Surat India 72.835806, 21.206134 December – May Guadalajara Mexico -103.3545, 20.6751 April – July Mazatlan Mexico -106.4199, 23.2503 April – July Merida Mexico -89.6165, 20.9678 April – July 63

City Country Coordinates Months used (range) Mexico City Mexico -99.0974, 19.4092 April – July Monterrey Mexico -100.3159, 25.7043 April – July Puebla Mexico -98.2059, 19.0492 April – July Veracruz Mexico -96.1541, 19.1747 April – July Cape Town South Africa 18.521476, -33.949923 Sept. – January Durban South Africa 31.010796, -29.854664 Sept. – January Johannesburg South Africa 28.031096, -26.211758 Sept. - January Port Elizabeth South Africa 25.566828, -33.922731 Sept. - January Pretoria South Africa 28.186064, -25.750952 Sept. – January Barcelona Spain 2.0568, 41.3528 March – June Bilbao Spain -2.976682, 43.294854 March – June Madrid Spain -3.6919, 40.4084 March – June Malaga Spain -4.4575, 36.7138 March – June Seville Spain -5.9843, 37.3886 March – June Valencia Spain -0.4054, 39.5310 March – June Birmingham United Kingdom -1.892390, 52.486005 April – July Glasgow United Kingdom -4.254918, 55.866864 April – July Liverpool United Kingdom -2.982390, 53.409553 April – July London United Kingdom -0.098132, 51.513295 April – July Manchester United Kingdom -2.243045, 53.485304 April – July Boston US – East -71.0926, 42.3573 April – July Chicago US – East -87.7053, 41.8744 April – July Detroit US – East -83.0409, 42.3328 April – July Miami US – East -80.2414, 25.7981 April – July New York City US – East -73.9772, 40.752 April – July Los Angeles US – West -118.2374, 34.0580 April – July Sacramento US – West -121.4725, 38.5591 April – July Salt Lake City US – West -111.8983, 40.7653 April – July San Diego US – West -117.1110, 32.7332 April – July San Francisco US – West -122.4537, 37.7554 April – July San Jose US – West -121.8912, 37.3242 April – July

Table S2. Summary of focal species and traits values, averaged across the number of cities where recorded. Refer to Table S4 for excluded species. DB = Diet Breadth, HB = Habitat Breadth, Mig. = Migratory status, SBS_N = Number of nesting substrate categories, SBS_U = Urban nesting recorded, N_City = Number of cities where recorded (this study).

Latin Body DB HB Mig. SBS_N SBS_U N_City Urban Species Urban Mass Abundance Proportion Preference

Accipiter badius 160.360 3 5 1 1 0 10 1.148 0.209 49.646

Accipiter bicolor 331.140 2 3 1 1 0 3 0.002 0.009 3.929

Accipiter cirrocephalus 179.460 2 4 2 1 0 5 0.145 0.133 21.525

Accipiter cooperii 449.360 1 3 1 1 0 10 1.621 0.249 11.750

Accipiter fasciatus 443.020 3 4 2 1 0 5 0.534 0.292 24.566

Accipiter gentilis 879.150 1 3 1 1 0 8 0.035 0.017 10.419

Accipiter melanoleucus 666.700 1 2 2 1 0 4 0.039 0.219 23.371 Accipiter nisus 197.860 1 2 1 1 0 12 0.073 0.176 10.782

Accipiter novaehollandiae 533.650 3 3 0 1 0 3 0.016 0.011 2.701 Accipiter ovampensis 197.020 1 5 1 1 0 2 0.013 0.484 17.380 Accipiter striatus 140.640 2 4 1 1 0 18 0.168 0.090 17.324

Accipiter tachiro 281.360 2 5 0 1 0 2 0.068 0.073 50.000 Aegypius monachus 9625.000 3 2 1 2 0 4 0.018 0.008 1.892

Aquila adalberti 3108.430 1 1 0 1 0 2 0.006 0.003 0.308 Aquila audax 3538.210 2 5 0 3 0 5 0.009 0.008 0.504

Aquila chrysaetos 4352.640 2 4 1 2 0 13 0.024 0.007 0.806 Aquila hastata 1448.220 2 4 0 1 0 8 0.021 0.002 5.300

Aquila rapax 2312.440 4 3 0 2 1 4 0.013 0.003 12.542 Aquila verreauxii 3787.930 2 2 0 3 1 3 0.010 0.062 4.935

Aviceda subcristata 316.520 3 2 2 1 0 2 0.151 0.050 7.662 Busarellus nigricollis 744.930 3 3 0 1 0 3 0.021 0.052 10.030 Butastur teesa 325.000 2 2 0 1 0 7 0.020 0.002 2.249

Buteo albicaudatus 956.570 3 3 1 2 0 10 0.009 0.041 7.537

Buteo albonotatus 761.000 2 5 1 2 0 6 0.015 0.061 18.895

Buteo brachyurus 491.500 2 3 2 1 0 17 0.029 0.054 16.914

Buteo buteo 899.750 3 5 1 2 0 11 0.176 0.365 6.709

Buteo jamaicensis 1177.710 2 10 1 4 1 16 3.621 0.339 7.187 Buteo lineatus 620.960 2 3 1 1 0 10 1.080 0.164 5.217

Buteo magnirostris 276.170 3 4 0 2 0 15 0.226 0.554 25.903 Buteo nitidus 516.670 2 2 1 2 0 6 0.015 0.038 16.009

Buteo plagiatus 523.620 2 4 1 1 0 6 0.050 0.260 37.415

Buteo platypterus 452.710 2 2 1 2 0 4 0.162 0.021 2.136 Buteo rufofuscus 1221.480 3 1 0 2 0 5 0.033 0.066 2.571

Buteo swainsoni 992.120 2 4 1 3 0 2 1.447 0.199 5.681 65

Latin Body DB HB Mig. SBS_N SBS_U N_City Urban Species Urban Mass Abundance Proportion Preference

Buteogallus anthracinus 996.290 3 4 1 1 0 7 0.025 0.067 10.239

Buteogallus lacernulatus 960.000 2 1 0 1 0 3 0.001 0.001 0.569 Buteogallus meridionalis 888.300 4 6 1 1 0 12 0.003 0.028 2.333

Buteogallus urubitinga 1178.570 6 6 0 3 1 6 0.005 0.013 10.125

Chondrohierax uncinatus 278.540 2 3 0 1 0 6 0.004 0.006 6.692

Circaetus cinereus 2000.000 1 3 1 4 1 2 0.002 0.028 3.300

Circaetus gallicus 1700.000 2 5 1 2 0 15 0.020 0.026 2.325

Circaetus pectoralis 1675.200 2 5 1 2 1 2 0.000 0.000 0.000 Circus aeruginosus 658.750 4 4 1 1 0 10 0.132 0.082 13.714

Circus approximans 728.070 4 4 1 2 0 5 0.208 0.143 18.419

Circus assimilis 551.360 2 4 1 1 0 4 0.006 0.006 3.247 Circus buffoni 493.500 1 3 1 1 0 2 0.000 0.000 0.000

Circus cyaneus 440.430 3 4 1 1 0 5 0.001 0.001 2.558 Circus hudsonius 425.000 2 5 1 1 0 10 0.063 0.008 0.757

Circus pygargus 312.360 2 4 1 1 0 6 0.006 0.006 1.434 Circus ranivorus 493.000 3 3 0 3 0 4 0.000 0.000 0.000

Elanoides forficatus 470.770 3 2 1 1 0 8 0.024 0.060 1.703 Elanus axillaris 274.790 2 3 1 1 0 5 0.217 0.158 14.129

Elanus caeruleus 255.570 2 5 0 3 1 19 0.053 0.059 3.167 Elanus leucurus 310.180 1 4 1 1 0 15 0.094 0.130 8.068

Gampsonyx swainsonii 96.680 2 4 0 1 0 4 0.014 0.037 9.456 Geranospiza caerulescens 313.710 2 6 0 1 0 8 0.004 0.011 4.236 Gyps fulvus 8093.380 2 5 1 2 0 7 0.053 0.042 0.955

Gyps indicus 5720.000 1 5 0 2 0 2 0.002 0.000 0.369 Haliaeetus leucocephalus 4688.570 4 5 1 4 1 9 0.141 0.039 3.081

Haliaeetus leucogaster 2775.000 3 5 0 3 0 8 0.096 0.038 4.168 Haliaeetus vocifer 2821.250 4 3 0 3 0 5 0.008 0.027 2.860

Haliastur indus 515.750 4 4 0 1 0 8 1.115 0.055 14.958

Haliastur sphenurus 744.290 4 3 1 1 0 5 0.164 0.105 6.736

Harpagus bidentatus 202.710 2 1 0 1 0 4 0.059 0.090 36.630 Harpagus diodon 200.000 2 1 1 1 0 3 0.000 0.000 0.000

Hieraaetus fasciatus 1905.710 1 6 0 2 0 10 0.017 0.003 4.917 Hieraaetus morphnoides 802.570 4 4 2 1 0 5 0.019 0.012 6.012

Hieraaetus pennatus 841.710 2 2 1 2 0 9 0.208 0.148 14.243

Hieraaetus wahlbergi 857.640 2 2 1 1 0 2 0.000 0.000 0.000 Ictinaetus malayensis 1245.550 3 2 0 1 0 2 0.014 0.001 3.630

Ictinia plumbea 244.540 2 2 1 1 0 7 0.008 0.012 2.117

Leptodon cayanensis 438.000 3 2 0 1 0 6 0.010 0.027 18.040 Leucopternis polionotus 725.000 1 1 0 166 0 3 0.000 0.000 0.000

Latin Body DB HB Mig. SBS_N SBS_U N_City Urban Species Urban Mass Abundance Proportion Preference

Leucopternis princeps 1000.000 1 2 0 1 0 2 0.000 0.000 0.000

Lophoictinia isura 588.360 2 5 1 1 0 5 0.036 0.013 5.335 Melierax canorus 810.750 3 3 0 2 1 3 0.000 0.000 0.000

Milvus migrans 839.070 5 8 1 3 1 25 2.321 0.370 42.349

Milvus milvus 1111.790 4 5 1 1 0 10 0.034 0.044 0.943

Neophron percnopterus 2018.600 5 5 1 2 0 10 0.054 0.006 3.866

Parabuteo leucorrhous 339.500 2 1 0 1 0 2 0.003 0.005 5.824

Parabuteo unicinctus 876.140 3 4 0 2 1 5 0.071 0.282 33.969 Pernis apivorus 698.250 2 3 1 1 0 6 0.023 0.026 8.163

Pernis ptilorhynchus 1117.500 3 2 1 1 0 10 0.121 0.017 12.581

Polyboroides typus 728.070 4 3 0 2 0 3 0.089 0.149 56.090 Rostrhamus sociabilis 387.150 2 2 1 1 0 10 0.031 0.094 13.322

Spilornis cheela 1266.000 2 5 0 1 0 8 0.073 0.006 21.222 Spizaetus cirrhatus 1419.640 1 3 0 1 0 5 0.006 0.002 2.113

Spizaetus ornatus 1102.610 1 2 0 1 0 2 0.008 0.012 15.469 Spizaetus tyrannus 1023.320 1 2 0 1 0 6 0.001 0.001 0.162

Spizastur melanoleucus 772.080 1 3 0 1 0 6 0.010 0.045 12.892

Table S3.

Explanation of trait value calculation from Sekerciglou et al. (2004, including updates).

Body mass: Up to four values were averaged.

Diet breadth: “Number of major food types consumed”, of nine types; e.g., “Invertebrate”, “Fruit, drupes”, etc.

Habitat breadth: “Number of major habitats used”, of 15 types. Multiple similar habitats were combined, such that “Savanna, arid plains, wooded grassland, pampas, campos” represented a single habitat type.

Nest substrate breadth: Calculated from twelve categories, e.g., “Bamboo”, “Building”, “Stump”, etc.

Migratory status: “Yearly, regular, long-distance” migrants assigned “1”; partial migrants assigned “2”. Sedentary and largely sedentary species assigned “0”.

Table S4. Summary of edits to eBird records and trait database (Sekerciglou et al. 2004, including updates).

Action Species/Element Variable/Value New value calculated Accipiter collaris Nest = tree, non-urban New value calculated Clanga hastata Nest = tree, non-urban New value calculated Cryptoleucopteryx plumbea Nest = tree, non-urban New value calculated Geranoaetus albicaudatus Nest = tree/shrub, non-urban New value calculated Parabuteo leucorrhous Nest = tree, non-urban New value calculated Accipiter collaris Body mass = 125 New value calculated Spizaetus Isidori Body mass = 1750 New value calculated Accipiter poliogaster Body mass = 440; used A. fasciatus New value calculated Pseudastur polionotus Body mass = 725; used P. albicolis New value calculated Circaetus pectoralis DB: 2; used C. gallicus New value calculated Buteo plagiatus DB: 2; used B. nitidus New value calculated Circus hudsonius DB: 2; used C. cinereous Dropped species (unique cases) Nisaetus limnaeetus = ssp. of Nisaetus cirrhatus Aquila pomarina = ssp. of A. Dropped species (unique cases) hastata Dropped species (unique cases) Pandion is a unique species for several Pandion haliaetus reasons Dropped species (unique cases) Parabuteo Malaga (escapee) Dropped species (no breeding) Accipiter cooperii South of US, Miami Dropped species (no breeding) Accipiter nisus India ex. Delhi, Jaipur, Lucknow Southern California and Miami; Veracruz, Dropped species (no breeding) Accipiter striatus Mazatlan, Merida Dropped species (no breeding) Aquila chrysaetos eastern US Dropped species (no breeding) Aegypius monachus India Dropped species (no breeding) Aquila heliaca India Dropped species (no breeding) Aquila nipalensis all Dropped species (no breeding) Aviceda leuphotes India Dropped species (no breeding) Buteo albonotatus Los Angeles Dropped species (no breeding) Buteo buteo India, Africa (south of Europe) Dropped species (no breeding) Buteo jamaicensis Veracruz Dropped species (no breeding) Buteo lagopus all Dropped species (no breeding) Buteo lineatus Mexico Dropped species (no breeding) Buteo platypterus Outside eastern US Dropped species (no breeding) Buteo regalis California, Mexico Dropped species (no breeding) Buteo rufinus all Dropped species (no breeding) Buteo swainsoni All ex. Sacramento, Salt Lake City Dropped species (no breeding) Buteo trizonatus Interior S. Africa Dropped species (no breeding) Circus aeruginosus India, Africa Dropped species (no breeding) Circus cinereus all Dropped species (no breeding) Circus cyaneus India, Africa Dropped species (no breeding) Circus hudsonius Miami, Mexico, Colombia Dropped species (no breeding) Circus macrourus all Dropped species (no breeding) Circus melanoleucus all Dropped species (no breeding) Circus pygargus India, Africa Dropped species (no breeding) Circus spilonotus all Dropped species (no breeding) Clanga clanga (incl. Aquila c.) India, Spain Dropped species (no breeding) Elanoides forficatus US ex. Miami, Mexico Dropped species (no breeding) Gypaetus barbatus all 69

Action Species/Element Variable/Value Dropped species (no breeding) Gyps himalayensis all Dropped species (no breeding) Haliaeetus humilis all Dropped species (no breeding) Haliaeetus leucocephalus Mexico, Southern California Dropped species (no breeding) Hieraaetus ayresii South Africa Dropped species (no breeding) Hieraaetus pennatus South India Dropped species (no breeding) Ictinia mississippiensis all Dropped species (no breeding) Pernis apivorus Africa Dropped species (< 15 records) Accipiter collaris Dropped species (< 15 records) Accipiter poliogaster Dropped species (< 15 records) Accipiter trivirgatus Dropped species (< 15 records) Accipiter virgatus Dropped species (< 15 records) Aquila spilogaster Dropped species (< 15 records) Aviceda cuculoides Dropped species (< 15 records) Aviceda leuphotes Dropped species (< 15 records) Buteo albigula Dropped species (< 15 records) Buteo polyosoma Dropped species (< 15 records) Buteogallus aequinoctialis Dropped species (< 15 records) Buteogallus coronatus Dropped species (< 15 records) Circaetus fasciolatus Dropped species (< 15 records) Circus spilonotus Dropped species (< 15 records) Cryptoleucopteryx plumbea Dropped species (< 15 records) Elanus scriptus Dropped species (< 15 records) Gypohierax angolensis Dropped species (< 15 records) Gyps africanus Dropped species (< 15 records) Haliaeetus ichthyaetus Dropped species (< 15 records) Harpyhaliaetus solitarius Dropped species (< 15 records) Kaupifalco monogrammicus Dropped species (< 15 records) Lophotriorchis kienerii Dropped species (< 15 records) Melierax metabates Dropped species (< 15 records) Morphnus guianensis Dropped species (< 15 records) Necrosyrtes monachus Dropped species (< 15 records) Terathopius ecaudatus Dropped species (< 2 cities) Accipiter minullus Dropped species (< 2 cities) Accipiter rufiventris Dropped species (< 2 cities) Accipiter superciliosus Dropped species (< 2 cities) Buteo regalis Dropped species (< 2 cities) Buteo trizonatus Dropped species (< 2 cities) Circus maurus Dropped species (< 2 cities) Gyps bengalensis Dropped species (< 2 cities) Gyps coprotheres Dropped species (< 2 cities) Haliaeetus albicilla Dropped species (< 2 cities) Harpia harpyja Dropped species (< 2 cities) Helicolestes hamatus Dropped species (< 2 cities) Leucopternis albicollis Dropped species (< 2 cities) Leucopternis melanops Dropped species (< 2 cities) Leucopternis schistaceus Dropped species (< 2 cities) Leucopternis semiplumbeus Dropped species (< 2 cities) Lophaetus occipitalis Dropped species (< 2 cities) Micronisus gabar Dropped species (< 2 cities) Polemaetus bellicosus Dropped species (< 2 cities) Sarcogyps calvus Dropped species (< 2 cities) Stephanoaetus coronatus 70

Action Species/Element Variable/Value Spelling errors/mismatches Neophron percnopterus as “perenopterus” in eBird Spelling errors/mismatches Ictinaetus malayensis as “malaiensis”

Table S5. Correlation Matrix for urban occurrence variables and trait variables evaluated.

Urban Species Urban Body Diet Habitat Migratory Nest Urban Number Abundance Proportion Preference Mass Breadth Breadth Status Substrate Substrate of Cities Number Urban 1.000 0.503 0.229 -0.082 0.106 0.482 0.162 0.309 0.272 0.410 Abundance Species 0.503 1.000 0.530 -0.205 0.062 0.311 0.237 0.133 0.109 0.342 Proportion Urban 0.229 0.530 1.000 -0.268 0.180 0.152 0.046 -0.078 0.026 0.150 Preference Body Mass -0.082 -0.205 -0.268 1.000 0.043 0.087 -0.083 0.410 0.141 -0.082 Diet 0.106 0.062 0.180 0.043 1.000 0.273 0.036 0.264 0.247 0.178 Breadth Habitat 0.482 0.311 0.152 0.087 0.273 1.000 0.116 0.397 0.330 0.488 Breadth Migratory 0.162 0.237 0.046 -0.083 0.036 0.116 1.000 -0.126 -0.125 0.215 Status Nest 0.309 0.133 -0.078 0.410 0.264 0.397 -0.126 1.000 0.637 0.178 Substrate Number Urban 0.272 0.109 0.026 0.141 0.247 0.330 -0.125 0.637 1.000 0.146 Substrate Number of 0.410 0.342 0.150 -0.082 0.178 0.488 0.215 0.178 0.146 1.000 Cities

Literature Cited

Birdlife International. 2019. HBW and BirdLife Taxonomic Checklist v4. December 2019.

Available online: http://datazone.birdlife.org/species/taxonomy

Blomberg, S.P., Garland, T., and Ives, A.R. 2003. Testing for phylogenetic signal in comparative

data: behavioral traits are more labile. Evolution 57, 717-745. (doi: 10.1111/j.0014-

3820.2003.tb00285.x)

Boal, C.W. 2018. Urban raptor communities: Why some raptors and not others occupy urban

environments. In: Urban Raptors: Ecology and Conservation of Birds of Prey in Cities,

Boal, C.W. and Dykstra, C.R., eds., Island Press. pp. 36-50. (doi: 10.5822/978-1-61091-

841-1_3)

Bregman, T.P., A.C. Lees, H.E.A. MacGregor, B. Darski, N.G. de Moura, A. Aleixo, J. Barlow

and J.A. Tobias. 2016. Using avian functional traits to assess the impact of land-cover

change on ecosystem processes linked to resilience in tropical forests. Proc. R. Soc. B.

283: 20161289. (doi:10.1098/rspb.2016.1289).

Beuchley, E.R., A. Santangeli, M. Girardello, M.H.C. Neate-Clegg, D. Oleyar, C.J.W McClure,

and C. Sekercioglu. 2019. Global raptor research and conservation priorities: Tropical

raptors fall prey to knowledge gaps. Diversity and Distributions 25:856-869. (doi:

10.1111/ddi.12901).

Cade, T.J., M. Martell, P. Redig, G.A. Septon and H.B. Tordoff. 1996. Peregrine Falcons in

Urban North America. In: Raptors in Human Landscapes: Adaptations to Built and

Cultivated Environments, Bird, D., D. Varland, and J. Negro, eds, Raptor Research

Foundation and Academic Press. Pp 3-14.

Chace, J.F., and Walsh, J.J. 2006. Urban effects on native avifauna: A review. Landscape and

Urb. Plan. 74, 46-69. (doi: 10.1016/j.landurbplan.2004.08.007)

Cooper, D.S., P.J. Yeh, and D.T. Blumstein. 2020a. Tolerance and avoidance of urban cover in a

southern California suburban raptor community over five decades. Urban Ecosystems.

July 2020. (doi: 10.1007/s11252-020-01035-w).

Cooper, D.S., A.J. Shultz, and D.T. Blumstein. 2020b. Temporally separated data sets reveal

similar traits of birds persisting in a United States Megacity. Frontiers in Ecol. and Evol.

8:251. (doi: 10.3389/fevo.2020.00251)

Croci, S., Butet, A., and Clergeau, P. 2008. Does urbanization filter birds on the basis of their

biological traits? Condor 110, 223-240. (doi: 10.1525/cond.2008.8409)

Cuthbert, R and 9 co-authors. 2011. Effectiveness of action in India to reduce exposure of Gyps

vultures to the toxic veterinary drug Diclofenac. PLoS One 6(5): e19069. (doi:

10.1371/journal.pone.0019069).

Del Hoyo, J., Elliott, A. & Sargatal, J. 2013. Handbook of the Birds of the World Vol 1–16, Lynx

Edicions.

Dwyer, J.F., S. Hindmarch, and G.E. Kratz. 2008. Raptor mortality in urban landscapes. In:

Urban Raptors: Ecology and Conservation of Birds of Prey in Cities, Boal, C.W. and

Dykstra, C.R., eds., Island Press. pp. 199-213. (doi: 10.5822/978-1-61091-841-1_14)

Dykstra, C.R. 2018. City lifestyles: Behavioral ecology of urban raptors. In: Urban Raptors:

Ecology and Conservation of Birds of Prey in Cities, Boal, C.W. and Dykstra, C.R., eds.,

Island Press. pp. 18-36. (doi: 10.5822/978-1-61091-841-1_2)

Emmons, L.H. 1990. Neotropical Rainforest Mammals: A Field Guide. Univ. of Chicago Press.

281 pp.

Evans, K.L., K.J. Gaston, S.P. Sharp, A. McGowan and B.J. Hatchwell. 2009. Oikos 118:251-

259. (doi: 10.1111/j.1600-0706.2008.17092.x)

Ferguson-Lees, J. and D. Christie. 2001. Raptors of the World. Houghton Mifflin. 992 pp.

Kembel, S.W., Cowan, P.D., Helmus, M.R., Cornwell, W.K., Morlon, H. Ackerly, S.P., D.D. et

al. 2010. Picante: R tools for integrating phylogenies and ecology. Bioinformatics 26:

1463-1464. (doi: 10.1093/bioinformatics/btq166)

Kettel, E.F., Gentle, L.K., Quinn, J.L., and Yarnell, R.W. 2018. The breeding performance of

raptors in urban landscapes: a review and meta-analysis. J. Ornithol. 159: 1-18.

(doi: 10.1007/s10336-017-1497-9)

Geist, V. 1987. Bergmann’s rule is invalid. Can. J. of Zool. 65: 1035-10338. (doi:10.1139/z87-

164)

GBIF.org. 2020. GBIF Home Page. Available from: https://www.gbif.org [multiple custom

downloads, May 6 – October 13, 2020].

Hijmans, R.J., E. Williams, and C. Vennes. 2015. Geosphere: spherical trigonometry. R package

version 1.5-10. Available online: https://CRAN. R-project. org/package= geosphere.

Horns, J. J., Alder, F. R., and Sekercioglu, C. H. 2018. Using opportunistic citizen science data

to estimate average population trends. Biol. Cons. 221:151–159. (doi:

10.1016/j.biocon.2018.02.027)

Iacus, S.M. 2016. sde: Simulation and Inference for Stochastic Differential Equations. R package

version 2.0.15. Available online: https://CRAN.R-project.org/package=sde

James, F.C. 1970. Geographic size variation in birds and its relationship to climate. Ecology 51:

365-390. (doi: 10.2307/1935374)

Jameson, Jr., E.W. and Peters, H.J. 1988. California Mammals. California Natural History

Guides 52. Univ. of California Press. 404 pp.

Liker, A., Z. Papp, V. Bókony and Á.Z. Lendvai. 2008. Lean birds in the city: body size and

condition of house sparrows along the urbanization gradient. J. Anim. Ecol. 77: 789-795.

(doi: 10.1111/j.1365-2656.2008.01402.x)

Lurgi, M., B.C. López, and J.M. Montoya. 2012. Novel communities from climate change. Phil.

Trans. R. Soc. B 367:2913-2922. (doi:10.1098/rstb.2012.0238)

Marti C.D., Korpimäki E., Jaksić F.M. 1993. Trophic Structure of Raptor Communities: A

Three-Continent Comparison and Synthesis. In: Power D.M. (eds), Current Ornithology

vol. 10. Springer, Boston, MA. (doi.org/10.1007/978-1-4615-9582-3_2)

Martins, E.P. 1994. Estimating the Rate of Phenotypic Evolution from Comparative Data. Amer.

Nat. 144: 193–209. (doi: 10.1086/285670)

Marzluff, J.M. 2016. A decadal review of urban ornithology and prospectus for the future. Ibis

159: 1-13. (doi: 10.1111/ibi.12430)

McCabe J.D., Yin H., Cruz J., Radeloff V., Pidgeon A., Bonter D.N., Zuckerberg B. 2018 Prey

abundance and urbanization influence the establishment of avian predators in a

metropolitan landscape. Proc. R. Soc. B 285: 20182120. (doi: 10.1098/rspb.2018.2120)

McClure, C.J.W and 14 co-authors. 2018. State of the world’s raptors: Distributions, threats and

conservation recommendations. Biol. Cons. 227: 390-402. (doi:

10.1016/j.biocon.2018.09.012)

Merckx, T. and 26 co-authors. 2018. Body-size shifts in aquatic and terrestrial urban

communities. Nature 558(7708): 113-116. (doi: 10.1038/s41586-018-0140-0)

Münkemüller, T., Lavergne, S., Bzeznik, B., Dray, S., Jombart, T., Schiffers K. and Thuiller, W.

2012. How to measure and test phylogenetic signal. Meth. Ecol. Evol. 3: 743-756. (doi:

10.1111/j.2041-210X.2012.00196.x)

Open Tree of Life. 2019. Version 3.2, October 30, 2019. Available online at:

https://tree.opentreeoflife.org/opentree/argus/ottol@1036185/Accipitridae

Oro, D., M. Genovart, G. Tavechhia, M.S. Fowler and A. Martínez-Abraín. 2013. Ecological and

evolutionary implications of food subsidies from humans. Ecol. Letters 16:1501-1514.

(doi: 10.1111/ele.12187)

Pagel, M.D. 1999. Inferring the historical patterns of biological evolution. Nature 401: 877-884.

(doi: 10.1038/44766)

Palacio, F.X. 2020. Urban exploiters have broader dietary niches than urban avoiders. Ibis 162:

42-49. (doi: 10.1111/ibi.12732)

Paradis E., Claude, J. and Strimmer, K. 2004. APE: analyses of phylogenetics and evolution in R

language. Bioinformatics 20: 289-290. (doi: 10.1093/bioinformatics/btg412)

Pennell, M.W., Eastman, J.M., Slater, G.J. Brown, J.W. Uyeda, J.C. FitzJohn, R.G. et al. 2014.

geiger v2.0: an expanded suite of methods for fitting macroevolutionary models to

phylogenetic trees. Bioinformatics 30: 2216-2218. (doi: 10.1093/bioinformatics/btu181)

Peregrine Fund, The. 2020. Global Raptor Information Network. Available online at:

http://www.globalraptors.org/grin/speciesLookup.asp?lid=1

Pinheiro J., Bates, D., DebRoy, S., Sarkar, D., and R Core Team. 2019. nlme: Linear and

Nonlinear Mixed Effects Models. R Pack. Vers. 3.1-143. Available

online: https://CRAN.R-project.org/package=nlme.

R Core Team 2020. R: A language and environment for statistical computing. R Foundation for

Statistical Computing, Vienna, Austria. Available online: https://www.r-project.org/.

Revell, L. J. 2012. Phytools: An R package for phylogenetic comparative biology (and other

things). Meth. Ecol. Evol. 3: 217-223. (doi: https://doi.org/10.1111/j.2041-

210X.2011.00169.x)

Rosenfield, R.N., R. William Mannan, and B.A. Millsap. 2018. Cooper’s Hawks: The bold

backyard hunters. In: Urban Raptors: Ecology and Conservation of Birds of Prey in

Cities, Boal, C.W. and Dykstra, C.R., eds., Island Press. pp. 93-109. (doi: 10.5822/978-1-

61091-841-1_7)

Rullman, S. and J. M. Marzluff. 2014. Raptor presence along an urban-wildland gradient:

influences of prey abundance and land cover. J. Raptor Res. 48(3):257-272. (doi:

10.3356/JRR-13-32.1)

Samia, D.S.M., Nakagawa, S., Nomura, F., Rangel, T.F., and Blumstein, D.T. 2015. Increased

tolerance to humans among disturbed wildlife. Nat. Comm. 6:8877. (doi:

10.1038/ncomms9877)

Santini, L., González-Suárez, M., Russo, D., Gonzalez-Voyer, A., Von Hardenberg, A., and

Ancillotto, L. 2019. One strategy does not fit all: determinants of urban adaptation in

mammals. Ecol. Lett. 22: 365–376. (doi: 10.1111/ele. 13199)

Sekercioglu, C. H., Daily, G. C. & Ehrlich, P. R. 2004. Ecosystem consequences of bird

declines. Proc. Natl Acad. Sci. USA 101: 18042–18047. (doi: 10.1073/pnas.0408049101)

Seto, K.C., M. Fragkias, B. Guneralp, M.K. Reilly. 2011. A meta-analysis of global urban

expansion. PLoS ONE 6:e23777. (doi: 10.1371/journal.pone.0023777)

Shochat, E. 2004. Credit or debit? Resource input changes population dynamics of city-slicker

birds. 2004. Oikos 106: 622-626. (doi: 10.1111/j.0030-1299.2004.13159.x)

Sol, D., González-Lagos, C., Moreira, D., Maspons, J., and Lapiedra, O. 2014. Urbanisation

tolerance and the loss of avian diversity. Ecol. Lett. 17; 942-950. (doi:

https://doi.org/10.1111/ele.12297)

Sol, D. and 9 co-authors. 2020. The worldwide impact of urbanization on avian functional

diversity. Ecol. Letters 23: 962-972. (doi: 10.1111/ele.13495)

Weeks, B.C., D.E. Willard, M. Zimova, A.A. Ellis, M.L. Witynski, M. Hennen, and B.M.

Winger. 2019. Shared morphological consequences of global warming in North

American migratory birds. Ecol. Letters 17: 942-950. (doi:

https://doi.org/10.1111/ele.12297)

White, J.H., J.M. Smith, S.D. Bassett, J.L. Brown, and Z.E. Ormsby. 2018. Raptor nesting

locations along an urban density gradient in the Great Basin, USA. Urban Ecosystems

21:51-60. (doi: 10.1007/s11252-017-0705-y)

Yom-Tov, Y. 2003. Body sizes of carnivores commensal with humans have increased over the

past 50 years. Functional Ecology 17:323-327. (doi: 10.1046/j.1365-2435.2003.00735.x)