Life on the Edge: The Eastern screech-owl in winnipeg

by

Chdstian Artuso

A Thesis submitted to the Faculty of Graduate Studies of

The University of Manitoba

in partial fulfilmenf of the requirerrents of the degee of

Doctor of Philosophy

Departrrrnt of Environment and Geography

University of Manitoba

Winnþeg

copyright @ 2w by christian Arhso THE UNIVERSITY OF' MANITOBA FACULTY OF GRADUATE STUDIES ,! rl *,r * COPYRIGHT PERMISSION

Life on the Edge: The Eastern Screech_Owl in Winnipeg

BY

Christian Artuso

A ThesisÆracticum submitted to the F'acurty of Graduate studies of The university of Manitoba in partial fulfirment of the requirement of the degree of

Doctor of philosophy

Christian Artuso @ 2009

Permission has been granted to the university rhesis/pracricum,ro of Manitob¿ I.ibraries urrary anå to rend a copy of this thesis/pracricum, all¡y., c"""äîoAc) and ro ;-"î, ruqlæ*a;;) ro lend a copy of rhis publishHt; microfirm, se, copies an absrract of this ,iÀirlp"u.ticum." and ro This reproduction or copy of this copyright thesis has been made available owner sorery f9; th"pu"p*e by authorify of the reproduced of privat" rrr;;;od research, and copied ur p""-üt.d;;^;;;;,;;;í"*, uoJ_uy only be authorization or wirh express wrirúen from rh, .ånñËü owner. ABSTRACT

cavity-nesting birds may be sensitive to urbanization due to changes in the availability of breedingresources. Nonetheless, the EasternScreech-owlhas higherproductivityand

survival in suburbs than rural areas in the southern portion of its range (Texas). Densities

and life-history haits of birds nray differ in range peripheral areas, rendering

generalization aborf the effect of urbanization difficult. I examined the population

density, reproductioq habitat selectioq and diet of Eastern Screech-owls in Manitoba to

test whether the patterns observed in Texas held at the northernperþhery of the range,

and whether such patterns would be npre or less pronounced. I conducted a random-

sftatified survey across a rural to urban gradient and monitored nests in natural cavities

and nest boxes' I collected data on diet and measwed variables related to the cavity tree

and at the habitat level at nests and unused cavities.

Eastern Screech-owl dersity was positively correlated with human density, peaking in

medium to high-density suburbs. Screech-owls preferred rþarian habitat brt densities

and breeding in residential areas versus greenspace did not differ. Brood sizes peaked in suburban areas' where fledging averaged five days in advance of rt¡ral areas. owls

selected habitat where canopy cover was relatively high, and with sufficient potential nest-sites close to some coniferous ffees. They selected nest sites in taller ftees, closer to water and with lower slÏub density below the nest than unused cavities. They avoided potential nest sites with npre buildings and higher domestic cat activity. Screech-owls had a more diverse diet with nrore vertebrate prey in low-derxity suburban areas as opposed to rtnal and high-dersity suburban areas. The rrpchanisms behind this pattern

lt appear to be the urban heat island, redrrced predation, greater diversity of prey, and favorable habitat alteration including planted conifers and buildings for roosting and

Írore open vegetation The percentage of rufor,rs morph screech-owls in Manitoba has fallen from 6- lIVo nthe 1920s to

Miruresota or North Dakota. I argue that, based on differences in swvival rates between color nnrplu in exteme cold, this is evidence of a northward range expansion facilitated by the anthropogenic influences noted above.

The differences in density, reproductivity, and breeding phenology between suburban and ruralareas, as well as the nechanisms involved, are similar to those found in Texas, suggesting that" at least in this case, the influence of urbannation is not strongly mitigated by latitude.

lll ACKNO\ryLEDGEMENTS

I would like to thank my funding sources: The University of Manitoba Graduate Fellowshþ, The Special Conservation Fund of Manitoba Conservatior¡ Manitoba Hydro, The Winnipeg Foundation, and the Great Gray Owl Fund administered by R. W. Nero, as well as a kind donationby Pat Cooper.

I wish to thark the members of my committee, Merlin w. Shoesmith Richard K. Baydack, James R. Duncar¡ Nicola Koper for their sìryport and guidance tluoughor.ú the degee process. I also thank other faculty rnembers of the University of Manitoba whose instruction and assistarrce have been invaluable to me including Stéphane Mcl-achlaq Spenser G. Sealy, David Walker, and James F. Hare. These chapteri were greatly improved by the thoughtful comments of Richard K. Baydack, Robert P. Berger, Brigitte De March, I¿ny De March Janes R. Duncan, John M. Eadie, Frederick R. GehhacL C. stuart Houston, Karla Kinstler, Nbola Koper, Erkki Korpim?iki Robert w. Nero, Spenser G. Sealy, Peter Taylor, Merlin W. Shoesmith, and various anonyrrlous reviewers. I thank Brigitte De March Nbola Koper, and DaviC Walker for their heip with the various statistical applicatiors u,sed. I thank Robert P. Berger, Jarres R. Duncar¡ Teny Galloway, Randy Mòoi, Justin Rasmusser¡ Rob Roughty, ãnd speruer G. sealy for assistance in identifying prey remains. I am grateful to Dale Marciski the Outreach Officer at the Meteorological Service of Environment Canada, for his assistance in gathering c limate data.

The following people sr¡bmitted observations of Eastern Screech-Owl and Great Horned Owl in Manitoba, particþated by listening for owls in their area, conftibrfed sight records for the historical database (even indirectly, e.g. by taking birds found injured to rehabilitation), heþed collecting pellets, helped to monitor roost sites, gave nre permission to enter their property, or otherwise assisted ne in various aspects of fieldwork or preparing rnanuscripts. I thank themall immensely: Linda Anderson" Nancy Andersor¡ læslie Andrews, Nadine Andrusiak, Terry Angus, Alfred Aug, Ron Arstiru Melissa Baer, Ken Baker, David Barnes, Pat Bar¡res, Mary Barrett, Derek Bartray, Nathalie Bays, Barny Beaulier¡ Krnt Beltor¡ [æane Beltor¡ Gloria Benhenl Keith Bersor¡ KeithBensoq Ballie Berger, Herry Berger, RonBerglund, Jim Bertwhistle, Patty Best, Brad Bird, Marcia Blair, Muray Blair, Luc Blanchette, Frank Boilear¡ Jim Boult, Chris Boumford, DaviC Bowles, Erl Braater¡ Jarres Bradley, Joy Bradley, Louise Brisebois, Jack Brown, Paul Buchanan, Bill Buhler, GaryBudyck, GregBuzza, ri- Byers, Alisoncanpbell, Kevincampbell, Brad carey, Læslie carignan, Brendan Camrthers, Silvio E. Cascino, Michelle Caughy, David Chevalier, DanCluanowski, phil Christiarso4 Ward Clristiansor¡ Dave Christie, John Christie, Jamie Chuback, Ed chudrick, Redmond Clarke, steve clubb, Michael cobus, colinCornad, Kyle conrad, Michael cornad, Pat cooper, Andy courcelles, [æa craig, Jack crolly, John cunningham, charles curtis, cal cuthbert Edgar Dandridge, Steve Davidge, Rob de Graaf, Brigitte De March, t-any de Marctr, chris De Ruyck, Ben DeBeer, Angel Delorme, Steve Demmings, KenDe smet, Paul Desrosiers, Nick Devine, wilbewit, Jeddich Doerksen, Cindy Dowse, Ronald Dueck, James R. Duncan, patsy Duncan, Alvin Dyck, Bill Eaton, Estelle Eaton, Kyle Elliott, Richard Ellis, Adolf Ens, Iiristen Eriksor¡

iv Dennis Fast, DebiForlanski, MargaretFrederikso4 PaulFrieser¡ MaureenFrolick, Barbara fuller, Vern Gaddows, Ian Gagnor¡ Wendy Garson" Corinna Gascho, Margaret Gibson, Randy Giesbrech Doreen Gbsbrecht, Mary Giesbrecht, Marlene Gifford, Mfte Gigiaa Ed Gilroy Micheline Girard, Miles Goolie, Paul Goossen, Gordon Grieef, Jaye Grieef, Paula Grieef, Susan Grieef, Jim Gyselinck, Keith Hamblin, Cary Harrel, Derrick Hanpton, I¿vin Hanke, Hao He, Bob Hart, David Hatch Jerry Hayes, iohnHays, CarolynHerrxandez, Ruth Hiebert, Jeff Higdon, Chris Higgs, tæane Hill, Heather Hinant Frank Hinings, Debbie Hobbs, Janice Hobbs, George Holland, Jean Hortoq Ray lversorL Jang Byoung-Soon, Liesel Janssor¡ Reinhard Janssón, Maricia Johrnon, paul Jones, Bernice Kennedy, Janet Kennedy, Fawad Khan, Mark King Sally King Todd King, Janis Klapecki, l-orne Klasser¡ Rudolf Koes, Hilton Kotyk, Brenäa Krõeker, Roy Laham, Linda r-ahzm, Heather I¿ild, Bill t apuck, Georgina Larsoq Raynnnd l¿rsorl Betty I¿vender, Anne tayma+ læw Layrnarq Kathleen læathers, yun-Kyoung Iæe, Gladys rngal Nick læone, sid læverick, Michael Iæwyc, pat Iæwyc, Li chunxiao, cory Lindgre4 Alain l-ouer, DeanMacDonald, Frank Machovec, Irene Magnesor¡ Nia Massey, Brenda Maxwell, christian McBride, Les Mccanr¡ Hope MJGaha, stephane Mcl-achlan, Ted Mcl¿chlan, Ardylle McMaster, DanMcMaster, GlenMcMaster, Laura McMaster, DeanMedeiros, Sarah Medill, Norm Melnychuk, GwenMewendorss, Al Mickey, Don Middletor¡ Kim Middleton, Corey Millar, Lucille Miller, Marthat Moffat, R. Moodie, Fred Mooi, Randy Mooi Janet Moore,LuMorash lan Morrow, John Murray, 'wayne Joan Murray, Shirley Murray, Gwen Mwenendorf, p. Neilly, John Neils, Robert w. Nero, Darrel Neufeld, Hany Neufeld, t ori Nichols, Robb Nichoí1, Kevin Nixon, Bruce Norvey, George Nykureik, Gerry oliver, cheryl orr, phil ould, Rachel parsons, Pantear¡ Brian Parkers, Robert Terry pastrick, Freðpawlyszyq Monica Pazernitlk, Linda Pearr¡ Carole Penner, Jake. E. Peters, Jennifer páloct, Barry pomeroy, Ryan Porteous, Angelé Pradaehi, procner, prouse, s usan Gordon Mike e uigley, Justin Rasmusser¡ Jacklyn Regeþ Don Reimer, Mary Reimer, Elvira Roberge-, aïúer Robinson, Robert RobirÌsoq Rong Rong Sheni Routly, Alex Sardersor¡ Chris Sands, Linda Sands' Linda Scarrow, Wayne Scarrow, Allan Scarth, Al Schirtt, Dorothy Schritt, Inara Schwartz, Gerald Scott, Sue Scott, Jean Scriver¡ Spenser G. Seal¡ Bob Sellars, Bud Shadholt, Bill Shestopalka, Courtenay Shuhz, Richard Silvermar¡bu*r, Simmons, Ken Sirncnite, Bob Sinclair, Lynn Sinclair, Rob Sinclair, Janet Skavinski, John Skene, JeffSleno, Mke Slobodiaq Alex Small, Borden Smid, Monba Smith, penny Smith, Rick Smith Sheila Smith, Shirley Snyder, Rbhard Sobkowich, Grace Sommerhoider, Marlene Sommerville, Doug Southanl Barbara Stacsbwicz, Diana Staniforth, Richard Staniforth" Bob stewart, Bill stokes, Bob stokes, Angela stoppa, Thaila stoppa, Brian sftaub, Glen Suggett, Jo swarta Dennis Swayze, Eugene Szach, George raybr, peter Taylor, Robert Jaylor, Yvette Taylor, Ruby Tekhaus, catherine Thextorl susan Thornas, Bðth Thorrpsor¡ Ken Tilling, Peter Tkachr:k, Teo Tkachuk, paul Tougas, Jeff Turner, Bob Tyler, Todd underwood, David unger, Jakonunger, Donald u*ur,, Marg unrutu Hersh Upadhyay, Katine Van den Meuler¡ Fran Vanderneuter¡ Liis Veelma, Aîdrey Veitclu Annette Verniest, Sherrie Versluis, Marlene Waldroru William J. Walley, Adam Walley* Glenwalleyq Gene_walz, Raþhwang 'watson, Dave warrenchuk, Sigrid warrenchuk, Barb Emmy weibe, Ron wiebe, John'weier, Jerry weshnoweski, ze,mwhjte, James whitelaw, Jarres whitman, David'wiebe, Ronwbbe, Jonathanwierx, Diane wilks, Judy winslow, I-arry winslow, Brent young, Reto Zaclu l¿urie zaporzag Tom Zaporzav Jeny Zaste. Sigrid Ztseff, and Geõrge Zulliclt

I also tharik the following people who submitted observations of Easternscreech-owls fromMinnesotå, North Dakota and SouthDakota: Phil Ahan, Lowell Andersor¡ Deric Andol, Rr¡h Audbery, DougBackland,_Dave Bartkey, Betsy Batstone-cu*inlhurr,, Patrick Beauzay, Gayle Beecher, Rr¡h Beecher, Robårt p. Èerger, Judd Brink, David cahlander, Elizabeth carrpbeil, Mark s-. Kathy connely, Keith corriss, Hugh Curtler, Linda Curtler, litsay, Jeff Darkert, Rob Daves, Alyssa DeRubéis, Roger olenich rterb Dingmar¡ Teny Dorsey, Tim Driscoll, Bob Dunlap, wayne Easle¡ t

The following museums provided information on Eastern Screech-owl specimern in their collections or sightings from record card schemes: The AmericanMmeum of Natural History, The Bell Museum of Nattnal History, carnegie Museum of Natrnal History, The Cornell University !9 Museum of vertebratãs, The FieldMuseum of Natural History, The Harvard University Museum of conpar ative znology, The Heritage (Dawhru North Museum Manitoba), The Manitoba Musèurn, The Mbñþan state unìíersityMur"urr,, The Provincial Museum of Aherta, The Royalonørio Mirseurq The Royal SaskatchewanMueum, The Science Museumof Minnesota, The SmithsonianNational Museumof Natural History, The Stewart Hay Memorial Museum in the University of Manit'oba's Departmerf of Biobgical Sciencãs, The University of MbhiganMuseum of Tnology, and The Yale peabody Mmeum.

VI To Youn-Young Park for her exftaordinary patience and continual vision and support, without which i could never have risked such change and withor¡ which this project would never have even got offthe ground, let alone bear fruit.

v11 CONTENTS

1. Introduction...... 1 BACKG.oUND:...... 1 AvianPreadaptations to the Process of Urbanization...... 1 The Effect of Urbanization on 8irds...... I2 I¡cRl- covruil: A cmrrunv oF CHANGE Iw wnrnvpnc's AvFATTNA ...... 30 SrRucrtnnoFTHETHESrs ...... 46 Irr¡RRrunsCrrm...... 52 2. Breeding and population density of the Eastern screech-owl at the northern periphery of its range ...... 86 IvrRonucrroN ...... g6 I\{¡-ürons...... gg Rnsul-rs ...... g2 Survey Data...... 92 Breeding ...... 97 DrscussroN ...... 103 Survey Data...... 103 Breeding ...... 105 IrrnRerunnCrrm...... 109 3. rrabitåt selection by the Eastern screech-owl along a rural - urban gradient ...... 119 IvrRooucrroN ...... ;...... 119 ME-IHoDS...... I25 R¡surrs ...... 132 SelectionofUsed Verst¡s Um¡,sed Cavities ...... 133 Differences betweenSuburbanand RuralNests...... I37 DlscussroN ...... 146 SelectionofUsed versus Unused Cavities...... 146 Differences between Suburban and Rural Nests ...... I54 IrrnnerunsCnm ...... 159 4. The diet of the Eastern Screech-owl at the northern periphery of Íts

INrRoDUsrroN ...... I73 M¡-rHons...... I75 R¡surrs ...... 1g0 DISCUSSIoN ...... Ig4 [.rrERATr.rRECrrm ...... 204 5. Eastern Screech-Owl hatches Wood Duck eggs...... ,...... 212 ACKNOWLÐGEMENTS...... 216 IJTERATURECITÐ ...... 2I7 6. Eastern screech-owl in Manitoba: Evidence of historical range exp:rnston ...... 220

vllt DlscussroN ...... 228 7. Conclusion: ...235 MncnamsMs oF SUBURBAN ADApTATIoN rN TrrE EesrmN Sm.nnc¡t-OlvI...... 235 Urban Heat Island ...... 235 Reduction inPredation...... 237 Habitat Change...... 238 Prey Availability ...... 239 PnmlcrloNs exo FnDnICrs...... 247 Il,rpucRrtoNs FoR CoNseRverroN AND MeNemlENT...... 243 l:rsn¡'rnnsCrrm ...... 247 8. AppendÍces.... t<)

TABLES

Table 2.1. Snatification categories for the Eastern Screech- Owl survey ...... 88 Table2.2. Eastern Screech-Owl detections on random stratified survey...... 94 Table2.3. Great Horned Owl detectiors on random statified survey...... 94 Table2.4. Averages of nurrerical values and ouþut of the GLM model for Eastern screech-owl survey with corresponding categorical variables ....96 Table2.5. EasternScreech-Owlbreeding2004-2007...... 98 Table2.6. EasternScreech-Owlbreeding by category ...... 100 Table2.7. Eastern Screech-Owl breeding from six different studies. ....101 Table 3.1. Disnibution of natural cavities by ftee species. ...133 Table 3.2.I. Averages + standard error of the mean of habitat variables with coefficient of variation and ouþut of the conditional regression model .134 Table 3.2.2. Roost site by period ...... 135 Table 3.2.3. Averages + standard error of the mean of nest-site variables with coefficienf of variation and ouþut of the conditional regression model 135 Table 3.2.4. Averages + standard error of the mean of disturbance variables with coefficient of variation and ouþut of the conditional regression model 136 Table 3.3.1. Averages + standard error of the mean of habitat featues around used and unused nest sites (plots of 50mradius) in three categories...... 138 Table 3.3.2. Pooled within-class standardized canonical coefficients and total canonical sftuctlne from MDA for habitat variables ...... 139 Table 3.4.I. Averages + standard error of the mean of nest-site features around used and unused nest sites (plots of L2m radits) in three categories ...... I41 Table 3.4.2. Pooled within-class standardized canonical coefficients and total canonical sftucture from MDA for nest-site variables...... I42 Table 3.5.1. Averages + standard error of the mean of distwbance and anthropogenic features around used and unused nest sites (ploß of 12m radius)...... 144 Table 3.5.2. Pooled within-class standardized canonical coefficients and total canonical structure from MDA for distrnbarrce variab les ...... I44 Table 4.1. Prey of Eastern Screech-Owl from2004 - 2008: sources of the data .....180

lX Table 4.2.1. Prey of Eastern Screech-Owl from 2004 - 2008: prey capttrres by period ...... 1g3 Table 4.2.2. Prey of Eastern screech-owl from 2004 - 2008: biomass consuned by period ...... 1g4 Table 4.2.3. Pooled within-class standardized canonical coefficients and total canonical sfructwe from MDA for prey type differences between biologically significant periods ...... 1g5 Table 4.3.I. Percentage prey disnibutior¡ diversity and niche breadth in breeding and non-breeding seasons ...... 1gg Table 4.3.2. Pooled within-class standardized canonical coefficients and tot¿l canonical süucture from MDA for prey type differences arrnng rural, low-density sr-rburban, and high-density suburban sites ...... 1g9 Table 4.4. Selectivity index for various guilds of birds at rural low-density suburbar¡ and high-density suburban sites ...... 191 Table 6.1. Break down of Eastern screech-owl records in Manitoba, northern Minnesota, and North Dakota from 1886 -2007...... 225

FIGT]RES

Fig. 1. Range of the Eastern Screech-Owt ...... 47 Fig.2.l. Average detection per transect of Eastern screech-owl and Great Horned Owl by hurnan density category...... 95 Fig.2.2. l,ocations of Eastern screech-owl nests, territories, and sites where birds were detected fewer than 3 times and did not form a stable territory 2004 -2007 ...... 95 Fig.3.1. Boxplots of canonical l scores for habitat variables...... 140 Fig.3.2. Box plot of canonical 1 scores for nest-site variables...... 143 Fig.3.3. Box plot of canonical 1 scores for disturbance variables...... J45 Fig.4.1. Scatter plot of canonbal I and2 scores from MDA for prey types differencesbetweenbiologicallysignificantperiods...... 1g6 Fig.4.2. Box plot of canonical 1 scores from MDA for prey types differences between rural, low-dersity suburban and high-density suburban sites....l90 Fi9.4.3. Niche breadth versus latitude from seven studies ...... 193 Fig.5.1. Three Eastern Screech-Owl eggs and two Wood Duck egp on24 April 2005, one Eastern Screech-Owl egg and three Wood Duck eggs on25 April2005 ...... 214 Fig.6.1. Percentage of rufous morph Eastern Screech-Owls recorded in Manitoba and in North Dakota and northern Minnesota by decade...... 226 ABBREVIATIONS Av average (also X) Br breeding season* CanI, CarQ canonical axis 1 and} (from multþle discriminant function analyses) DBH diarrpter at breast height (of nees) Dist. bldg distance to nearest building Dist. road distance to nearest road Dist. water distance to nearest river or creek DmCavH diarreter of trunk or brarrch at the height of the cavity EASO EasternScreech-Owl(Megascops asio) GHOW Great Horned Owl(Bubo virginianus) GHOW_Av average detections of Great Horned Owl per survey fransect* GLM general linear nrcdel H. height HDcat human dernity categoryx High bird species that tolerate high antltopogenic disturbance MDA multiple discriminant function analyses Mig passage migrants and winter vlsiting bird species Min bird species that tolerate only minimal anthropogenic disturbarrce Minfþath minimum flight path from cavityx Mod bird species that tolerate rmderate antlropogenb disturbance ms morph specified* nd no date Non-br non-breeding seasonx Non-p non-passerines NRCT native riparian cavity trees (American elm, green as[ Manitoba maple, btrr oak, basswood, and eastern cottonwood) other dec other deciduorìs tree (species other than those in NRCÐ open ugr open understory (ground layer) Vo catÅerse Vo canop! density above nest tree* Pass passerines plha number of people per hectare Relcanheight relative canopy heightx Res resident and locally breeding bird species Rþ rþarianx sp / spp species (singular / plural) tr üees w wildlands areas (<1 p/ha) R mral areas (1 - 10 p/ha) SI low-density suburban areas (>10 - 20 plIn) S2 nnderate dersity sububan areas (>20 - 30 p/ha) S3 high-density suburban and urban areas (>30 p/ha) s-h high-density suburban and urban areas (>30 p/ha) s-l low-density and nediun¡density suburban areas (>10 -30 plln) t See text for operational definition used in this thesis

KI L. Intrcduction

BacxcRouvn:

Avian Preadaptations to the Process of Urbanization

Birds react to the dramatic changes wrought on the landscape by urbanization in a great

varietyof ways. Birds are sometimes classifîed into broad categories depending ontheir

tolerance for human activity, e.g. predevelopment species which tolerate little or no

anthropogenic disturbance, suburban-adapted species which can benefit from sorrp level

of antlropogenic habitat change and human activity, and urbanexploiters whichbenefit

geatly from urbanization and which are either human cornmensals or near-obligate

human s ymb io nts (B la ir 199 6). S o me o f the b est kno wn exarrp les o f s ymb iotic

aviary'humanrelationshþs include the European Starling (Sturnus vulgaris), the House

Sparrow (Passer domesticus) and the Chimney Swtft (Chaetura pelagica), which forrrerly roosted and nested in large hollow ftees but which subsequent to broad scale habitat change in North America now relies almost exclusively on hunran süuctures

(Johrston 2001).

Such flexibility in the way birds respond to human activities makes summarizing the effects of urbanization on the world's aviåuna a challenge; however, sonre gener¿rl nends have been reported. Urban areas are sponsored environments for many species, e.g. certain corvids, at least in part became of the prevalence of anthropogenic food sources.

Therefore, despite their lower avian biodiversity, urban areas often have greater avian abundance (Emlen 1974,Walcott1974, Gavareski 1976,l-ancaster and Rees 1979, Beissinger and Osborne 1982, Cicero 1989, Blair 1996, Jokimäki et aL 1996, Bolger

1997, Germaine et aL 1998, Cam et a1.2000, Blair 2001, Marzluff 2}}l,McKinney 2006,

Palomino and Carrascal2006, Smith and Wachob 2006). Peak biodiversity, at least on a local scale in North America, may occur at intermediate levels of human disturbance.

This is perhaps because interrrpdiate human disturbance levels appear to create conditions where the gain of suburban-adapted avian species exceeds the loss of predevelopment species (Batten 1972, Jokimåiki and Suhonen 1993, Blair 1996, Blair

1999, Blair 200I, Adarrs 2005, Marzluff 2005, Donnelly and Marzlutr 2006, Palomino and CarrascaI2006). Sone have related this pattern to a theoretbal relatiorshþ between net primary productivity (highest in urbanized environnents) ard species richness with species richness peaking at intermediate levels of productivity (Morin 1999, Mittehach et aL 200I, Shochat etaL 2006). The relationshþ between net primary productivity and urbanization seems counterintuitive since impervious surfaces reduce primary prodrrctivity to near zero; however, this is compensated for by the above average primary productivity in several urban environments, in particular urban parks and greerspaces

(Imhoffet al. 2000, Kaye et aL.2005, Shochat et al. 2006). In Europe and Africa, at least on a regional level, evidence for the associationbetween internpdiate human distrnbance and avian biodiversity is lacking. Instead regions with high avian diversity correspond to relatively dense hurnan populations, perhaps due to the distribr¡tion of resor¡rces and continental climatic patterns (Balmford et aL Z}}l,Chown et al.2003,Gaston and Evans

2004,Nonis and Harper 2004,Pidgeon etaL2007). These observed differences illustate the need to consider multiple spatial scales when seeking generalizations about urban environrrents (Clergeau et aL 2006).

2 Even if avian biodiversity does peak at intermediate levels of distrnbance, this does not

represent the whole picture, as the suite of species occupying heavily impacted areas tend

to be very similar, whereas the specbs excluded from developing areas may be

threatened or limited in their distribution On the whole, urbanization appears to pronnte

a decline in native species and increased abundance of alien species, since the latter are

rnore often human symbionts (Jokimäki and SuhonenI993, Zalewski L994, Bla:l. 1996,

Rolando etaL 1997, Clergeauet aL 1998, Adams 2005, vanHeezk et al 2008). Many

authors have thus argued that the net effect of urbanization is homogenizatior¡ i.e. a

decrease in the avian biodiversity of a region despite possible increases at localized sites

augmented by inroduced species or species that have been able to use anthropogenic

habitat change as a vehicle for range expansion (Jokim¿iki et al. 1996, Olden et aL 2004,

Crooks et aL 2004, Adams 2005, Marzluff 2005, McKinney and Lockwood 2005,

McKinney 2006). One possible mechanismbehind this patternof homogenization maybe

corrpetitior¡ with rnban-adapted species with excellent colonizing abilities competitively

excluding less disturbance-adapted species (Bennett 1990, Shochat et aL 2004, Shochat et

aL 2006); however, it is important to note that competitive exclusion is likely a secondary

mechanism under these circumstances with primary mechanisms such as higher productivity and resource abundance (Imhoff et al. 2000, Bolger 2001, Shochat et aL

2004, Kaye et aL.2005, Shochat et al. 2006), reduced predation (Marzluff 2001, Shochat et aL 2004, Shochat et al. 2006), the urban heat island (Shochat et aL 2006) and habitat change (Shochat et aL 2004) first increasing the carrying capacity of urban areas (Emlen

L974, Shochat et al. 2006) and thereby altering the dynamics of interactions and time- energy budgets. This creates a competitive advantage for those disturbance-adapted specbs that are best suited to the characteristics of urban settings either through more efficient resource exploitation or interspecific aggression (Copley et al. 1999;

Tryjanowski and Kuczynsk i 1 999, Sho chat et aL 2004, S hochat et al. 2006).

Although attenpts to provide generulizations about urbanization often fail, many authors have sought to provide sorne clarity and predictability by examining the irrpact on higher taxonomic grorpings (genus or family) or avian guilds - groups of species with similar habitat preferences, foraging strategies, migratory sftategies or some other uniting feature. Dispersal capab ility is one of the features that may distinguish birds from many other taxa and yet there is considerable variability in the dispersal capability of different avian species and gror-ps (Donnelly and Marzluff 2006). In the fragnrcnted urban matrix, an ability not only to utilize small fragnrcnts but also for effective dispersal between fragnnnts is essential. Dispersal is biologbally stressful, and barriers to dispersal in urban areas such as roads can limit survivorshp and gene flow (Pearce et al 2007). For this reason good dispersers are favored over poor dispersers in urban environments

(Garden et al. 2006) and the dispersal of some urban-exploiting species is facilitated by human activities and landscape changes (McKinney 2008). Nonetheless, long-distance dispersers (migrants) are disadvantaged in urban areas that they might exploit for only a limited portion of the year (Friesen et al. 1995, Lindsay et aL 2002, Green and Baker

2003, Kark et aL 2006). Resident species rnay have advantages over migrants in urban environrrents by acquiring local knowledge that enables themto adapt to fluctuating

(ruman influenced) food supplies, altering distwbances regimes and localized predation threats (Stratford and Robinson 2005, Chace and Walsh 2006, Marzluff et aL 2007 ). One study of a Nearctic to Neoftopical migrant, the AcadianFlycatcher (Empidonax virescens) found that urbanization delayed nesting, lowered productivity (possbly via increased brood parasitism with fewer nesting attenpts), and increased turnover in site occupancy, implicating both individual responses and population-levelresponses in this miglant's urban disadvantage (Rodewald and Shustack 2008).

L-arge body size canbe a disadvantage in the patchy urban matrix where an ability to utilize small fragments is key (Chace and Walsh 2006). This may explain why top predators rrny be excluded from urban areas, providing rrrcsopredators with reduced conpetition and access to anple resources (Crooks and Soulé 1999). Perhaps even Íþre important than body size is home range size, sirrce species with small territories and home ranges are more likely to be able to utilize urban fragments (Jokim¿iki 1999, Chace and Walsh 2006). A possible exception to this generaluation may be species that do not necessarily reside in the urban rnatix bú which are able to move into urban areas for specific activities or to exploit temporary resources such as gulls that visit dump sites far from nesting colonbs (Chace and Walsh 2006).

Habitat selection is another key factor in urban environrrpnts. Urban environments with their manicured lawns and shrubbery and profision of roads and other impervious surfaces more closely resemble open habitats (e.g. early successional habitats) than mature forests. For this reason forest interior species tend to be excluded from urbanizing areas whereas open-country or early successional (disturbance-adapted) species may not be @iarnond \986,Kltzz et aL 2000). Habitat specialists are unlikely to find the resources to match their foraging and breeding requirements in the urban nnsaic, whereas habitat generalists, by virtue of their ability to exploit different habitats and resources, are rnore likely to be able to meet their needs (Aldrich and Coffin 1980).

Ground-nesting species are frequently excluded from suburban areas with npwed lawns and easy access by predators and pedesftians; however, shrub-nesting species such as the

Nortlrern Cardinal (Cardinalís cardinalis) may have an advantage in such habitats where shrubs grow in rows and may be very dense as a result of pruning (Jokim¿iki 1999). Often related to habitat preferences, certain aspects of behavior are also important for a species ability to exploit urban landscapes. Neophobic species that for evolutionary reasons are geared towards stability of environment and resources are less likely to adapt to the rapid changes irrposed on them in the urban matrix than species whose evolution has forced sorre degree of behavioral plasticity (e.9. in response to disturbances or environrrental flux) (Sol and Iæfebvre 2000, Kristan 200I, Sol et aL 2002).

Lke habitat specialists, diet specialists are more likely to be excluded from developing areas whereas diet generalists, in particular omnivores, specbs that can capitalize on anthropogenic food sources such as bird seed or refi.rse, and sone opportunistic nest predators are more likely to be sr.iburban-adapted (DeGraaf and Wentworth 1986,

DeGraaf 1991, Andren1992, Zalewski 1994, Nilon et al 1995, Jokimäki and Suhonen

1998, Haskell et aL 200I, Garden etaL2006, Chace and Walsh 2006, Kark et aL2006,

Smith and Wachob 2006). Increased water availability around human settlenpnts in dry habitats provides an advantage to many species, especially granivores, since this facilitates the digestion of dry seed permitting greater resource exploitation (Kotler et al. 1998). Sone species have been shown to switch the composition of their diets in urban areas to favor seeds whereas they rely on food with higher moisture content such as fruit and flowers in undisturbed xeric habitats (Wolf and del Rio 2000).

None of the above generalizations offer the ability to predict whbh species will benefit and which will suffer from urbanization A species' particular preadaptations, which can form the basis for microevolution, must also be considered (Fuller et aL 2006). The most obvious exarrples of microevolution include altering habitat selection from natural microhabitats to humanstructures, e.g. Metallic Starlings (Aplonis metallica) nesting in vents not tee cavities (Diannnd 1986) or Chimney Swift switching from hollow trees to chimneys as nest and roost sites (Graves 2004). However, despite the general trend of increased likelihood of colonization success with increased behavioral plasticity (Sol et aL 2002), other secondary cavity nesting species have either not undergone a similar microevolution or not altered habitat selection to the same extent, necessitating a case-by- case examination of preadaptations and the plasticity of individual species.

An excellent exanple of species with different preadaptations at different points along the confinuum of urban avoidance to suburban preadaptation is provided by three owl species, viz. Northern Spotted Owl (Srrur occidentalis caurina), Great Horned Owl (Bubo v irginianus) and Eas tern Screech- Owl (Me gascops a sio).

The Northern Spotted Owl is the least suburban-adapted of these species. Despite being a relatively small-bodied (45cm) resident predator, the average honp range of the Northern Spotted Owl is exüemely large, averaging over 10km2 (Carey et aL 1990) and sonetimes larger than 40knf (Harrer 1988). The Spotted Owl is a forest-dweller that avoids non- forested areas (Carey et aL 1990) and a specialist of old-growth forest (Gutiérrez et a[

1995) with old growth corstituting as much as 7 5Vo of indivi1ual home ranges (Carey et at 1990). Dispersing juveniles must be able to travel long-distances from their parent's territory to find suitable habitat and can therefore be compromised by deforested landscapes (Gutiénez et aL 1995). As such, this species is unlikely to be able to neet its hab iøt require nents in human- a ltered landscapes.

The breeding biology of the Northern Spotted Owl is geared toward environrrental stability in a relatively unchanging habiøt (old-growth forest). The Northern Spotted Owl is long- lived, has somewhat delayed sexual maturity for a medium-sized owl (at leastZ years old), a low reproductive rate with few young per brood, and deferred breeding when food supplies are inadequate (Gutiénez et aL 1995). This diet niche of the Northern

Spotted Owl is relatively ruurow as it preys rnainly on squirrels, rodents ard lagomorphs and to a lesserextentbirds ard invertebrates (Gutiérrezetal.1995). As aconsequence, the Northern Spotted Owl has not adapted well to human activity, and is corsidered an endangered species (U.S. Fish and Wildlife Service 1990). The Northern Spotted Owl is cwrently outcompeted in nnst habit¿ts except large tracts of old growth forest by its closely related sister species, the Barred Owl (Strix, varia), which is a generalist with smaller spatial requirements, a broader diet niche, and less specialized habitat requirements (Mazur and James 2000). The Barred Owl has benefited from antlropogenic habitat change and spread into the range of the Northern Spotted Owl

(Hamer 1988, Mazur and James 2000).

The Great Horned Owl (56cm) is larger than the Northern Spotted Owl and yet it is rrnre

distubance-adapted and has not suffered the declines that the Northern Spotted Owl has.

Of course these owls share sorrp similarities such as cryptic plumage and a sit-and-wait

hunting style that makes themrelatively inconspicuous to hurnan eyes - a corniderable advantage to a predator around human settlenents. They are both year-round residents in

their respective habitats and reach sexual maturity at a similar age. However other

biological differences ftanslate in very different relationships with humanbeings. The

Great Horned Owl is found throughout the Anericas in almost every habitat type

available from deserts to forests. This species even hunts over large unteed areas and can breed in small copses of tees in fragmented habitat maftices (Houstonet al. 1998). As

such the Great Horned Owl is able to utilize human-altered habitats and breeds in suburban areas with sufficient green space (Smith et aL 1999, Artuso 2007b).

The Great Horned Owl has other preadaptations to human-altered habitats. In addition to being a habitat generalis! this species is also a diet generalist and an opportunistic hunter able to exploit a wide variety of prey items. Unlike the Northern Spotted Owl, the Great

Horned Owl also exhibits broader nest-site selection, utilizing stick nests rnade by other birds, cliff ledges, caves, hollow logs, goose rrcsts on the ground, and even buildings or other human süuctues such as pipes or bridges (Hou,ston et aL 1998). Such behavioral plasticity conpensates for this owl's large sizn, giving access to a wide variety of

9 resources not available to a specialist. Therefore, despite its larger size, the home range

size of the generalist Great Horned Owl (1 - 9km2¡ averages much smaller thanthe

specialist Northern Spotted Owl (Houston et al. 1998). Living in rnore open habitats, the

Great Horned Owl is not only able to disperse better across human-altered landscapes but

is also more geared towards environrrpntal flux than the Northern Spotted Owl, although

northern populations do show periodic population tends following lagorrnrph cycles

(Hotston et al. 1998). The Great Horned Owl also commences breeding very early in the

spring, perhaps another distrnbance-adapted feature that permits renesting if required

(Marti 1969).

The F¿stern Screech-Owl (22cm) is by far the smallest of the tluee owl species discussed

here and the most suburban-adapted of these specbs. The small body size and small

spatial requirements (horrrc-range size) of this species permit the utilization of relatively

small uban habitat patches that could not provide sufficient resources for larger species.

In the breeding season, the home ranges of Eastern Screech-Owls in southern Manitoba,

at least as far as can be determined from repeat observatiors of territorial pairs, appear to be <1km2, although this greatly increases during the non-breeding season (C. Artuso, unpublished data). In addition to its small body size, the Eastern Screech-Owl's cryptic plumage and nocturnal sit-and-wait hunting style make it extrerrnly inconspicuow and thus less likely to experience human disturbance than larger predators. Lfte the Great

Horned Owl, the Eastern Screech-Owl is a remarkable diet generalists and is capable of exploiting a variety of seasonal resources (Chapter 4), rnaking it highly suited to suburban areas where it may benefit from concentrations of prey around bird feeders with

10 fallen seed or invertebrate abundance created by watering and fertilizing sububan garders (Chapter 4).

Although the nest-site selection of the Eastern Screech-Owl (cavity nesting) is much rnore restrictive than the Great Horned Owl, its small size affords it a wider selection of suitable cavity nest-sites and the ability to use some human-rnade nest-boxes (Chapter 3).

The Eastern Screech-Owl is also rnore resticted in habitat-selection than the Great

Horned Owl, generally preferring matwe but relatively open woodlands (Gehhach

1995); however, in the open habitats of sr-rburbia with tree-lined boulevards, city parks and open spaces such as lawns it finds a human-altered habitat that closely mimics its preferred open woodland habitat. l¿wrs and parks with open undergrowth are particularly well suited to the hunting style of the Eastern Screech-Owl, which does rnost of its hunting by pouncing on prey on the ground (Gehlbach 1994).

The Eastern Screech-Owl also has sone preadaptations to life in suburban areas, associated at least in part with srnall body size. Unlike the Northern Spotted Owl, Eastern

Screech-Owls breed annually, ahhough ch¡tch size appears to vary with food availability

(Chapter 2). They have relatively large broods (2 - 6) in Manitoba (Chapter 2) añ possibly as many as seven (Gehlbach 1995) with a relatively short incubation period that permits renesting if required (smaller specbs are likely to require renesting after predation events) (Gehhach 1994). This species is also polyterritorial. In this thesis I use the term polyterritorial to me defending multiple cavities within a single territory, not to indicate a polygynou specbs. The presence of several potential nest-sites rrpars that

11 renesting following a disturbance does not require abandoning a resource-rich territory.

Reaching sexual maturity at one year old, this species is capable of breeding rapidly and

can exhibit a faster gene ttnnover than larger-bodies species (Gehhach 1988, 1995),

allowing for recovery following disttnbance and exploitation of newly-available habitats,

e.g. following the erection of nest-boxes. In this thesis, I will explore how the Eastern

Screech-Owl exploits these preadaptatiors in suburban areas.

The Effect of Urbanization on Birds

Preadaptations and avian guild characteristics alone do not adequately describe the way

urbanization affects birds. Avian ecologists have been searching for patterns specific to

the types of fragmentation and disttnbance that urbanization produces and the resuhing

influences on biodiversity (Fahrig 2003) and habitat connectivity (Sasvári and Moskät

1988, Jokimäki 1999, Platt and Lill 2006) in order to understand the various mechanisms

behind such patterns and ultimately to inform the process of development such that

impacts onavian communities are minimized (Rodewald 2006). Birds in rnban

environrrpnts have also been changing their behaviors and adapting in order to contend

with human-ahered conditions that affect their ability to reproduce and survive. Birds in

urban or urbanizing areas face varioru forms of direct and indirect human disturbance,

altered food sowces and resource availability, altered conpetition and predation regimes

due to changes in species conpositioû rapid habitat change, altered terperature (urban heat island) and mitigation of various natural cycles such as floods and droughts and

other processes or of climatb variability (Shochat et aL 2004, Yeh and Price 2004, Parris

atñ Hazel2005, S hochat et aL 2006).

t2 One possible consequence for birds living in urban settings is exposure to pollution and disease. Various effects of toxicity on birds have been documented in urban areas including the accumulation of lead (Getz et al. 1977), altered blood indices as a consequence of pollution (Kostelecka-Myrcha et al. 1997), and bioaccumulation of goup elenents srch as platinum (Jensen et aL.2002). Noise pollution has also been shown to affect birds in urban areas causing them to expend energy by vocalizing louder and more frequently (Estes and Mannan2003, Brumm 2004), to adjust calling fre{uencies

(Slabbekoorn and Peet 2003) or to alter song phenology to favor times with less antlnopogenic noise (Fuller et aL2007). Since song is a fundamental mechanism of reproductive isolation in many birds, adapting to noise pollution may have evolutionary consequences for sorrn species, leading to developnrent of wban "subspecies"

(S labberkorn and Riprrrcester 2008). As disct¡ssed above, urban conditions can prornote higher avian densities but this has the potential to augnpnt disease trarsmission (Johnson and Glahn 1994). One well-documented exarrple of this pertains to Cooper's Hauil

(Accipter cooperii) in suburban Tucson, Arizona, whbh suffers 50Vo juvenile rrnrtality as conpared to SVo in rural populations due to a disease canied by their main prey, pigeons and doves (Estes and Mannan 2003), although this does not necessarily rrean that urban areas are acting as population sinks for this species (Mannan et aL 2008). Red-winged

Blackbirds (Agelaius phoeniceus) nesting in stormwater wetlands in residential and comrrercial areas with high human distrnbance and rnoderate water quality were conpromised in their nesting success resulting in a population sink one year but a source

ín another (Sparling et aL 2007).

t3 The urban heat island is another factor that birds in citþs and towrs must contend wittl

For sone species, the uban heat island provides benefits including permitting earlier nesting (Gehlbach 1994, Korpimäki 1978, Rollinson and Jones 20OZ), higher overwinter survival (Gehhach 1994) and a potential increase in invertebrate prey (Gehhach 1994,

Rollinson and Jones 2002but see Burke and Nol 1998). In association with these three factors, the urban heat island is also implicated ìnreduced migratory propensity and increased sedentariness (Adriaersen and Dhondt 1990, Luniak et al. 1990, Partecke and

Gwinner 2007), which in turn provides individuals that overwinter in urban areas a reproductive head start over migratory conspecifics (Partecke and Gwinner 2007). On the other hand, the urban heat island may also create süesses such as heat stress, whbh can be highly detrimental to nestlings and adults in the nesting period (Korpimäki 1985,

McCollin 1998, Kristan 2001). A further complication is altered song phenology from artificial lights (Adams 2005).

Direct antlropogenic disturbance is typically pronounced in wban areas (Forman and

Alexander 1998). Activities such as developnent and land clearing, rrnwing, pruning, tree removal or removal of dead branches and other forms of manburing can create ecologbal traps for birds that establish territories or begin nesting only to have their territories or nests disturbed during criticalperiods (Dwernychuk and Boag1972. Gates and Gysel1978, Best 1986, Rernes 20Õ0, Batfin 2004). Vehbles also carne considerable mcrtality as do collisiors with buildinp, especially glass structures (Klem 1989, Millsap

2002, Rollinsonand Jones 2002). Otherdisturbancetypes withsignificantconsequences include collisison with power lines and electrocrfion (Bevanger 1998). Eurasian Eagle

l4 Owß (Bubo bubo) abandon territorþs and show higher risk of nest failure within 200m of power lines, as well sufferrng a L77o loss of fledged young from electocution (Sergio et aL 2004). Seemingly innocuous activities such as pedestians or cyclists using nails may also have negative impacts onbreeding (Ferruández-Juicic and Tellería 2000).

Nonetheless, ffail use by pedestrians was not found to negatively influence reproductivity in the Northern Cardinal in Ohio, but rather influenced their nest-site selection (Smith-

Casúo 2008). Lkewise, EuasianMagpies (Pica pica) alter their nest-site selection in urban areas, selecting higher nest sites in proximity to pedestrian activity (Wang et al.

2008).

Aside fromdirect ecological ftaps and rrnrtality there are many rnore subtle effects of anthropogenic disturbance. One notable problem is the way animals are forced to alter their time/energy budgets in urban settings. Roads can constrain foraging or limit tenitory size or access to other resolnces, e.g. the Spanish Imperial Eagle (Aquila adalbertl avoids certainkey foraging areas on weekends and at times when vehicular traffic is high (Bautista et aL 2005). Some birds will adjust their dispersal patterrs to avoid areas of high human use (Chace and Walsh 2006). Sone mammals have been shownto become increasingly nocturnal inorder to minimize distubance from hurnan beings (McClennen et al. 200I, Tigas et aL.2002, Riley et aL 2003). This has not yet been dernonstrated in birds; however, EuropeanRobns (Erithacus rubecula) nnoisy urban environrrents sing at night rather than during the day @uller et aL 2OO7). Birds adjust their time/energy budgets in npre subtle ways, e.g. suburban Ferruginors Hawks (Buteo regalis) in Colorado perched less frequently on poles or on the ground (conspicuous

15 perches) in winter than in less distubed sites and spent rncre time at roost sites

(Plunpton and Andersen 1998). Red-tailed Hawks (Buteo jamaicensis) reduced their

alarm call and dive-bombing rates as they becane accustomed to increased hurnan

presence (Knight et al. 1989).

In most cases however, anthropogenic presence and habitat change have complicated

effects that are likely to benefit sorrre species while disadvantaging others. The number of

buildings in urban parks has been shown to be detimental to the breeding of Willow

Warbler (Phylloscopus trochilus), Hooded Crow (Corvus comix), and Spotted Flycatcher

(Muscicapa striata) (Jokimäki 1999). On the other hand, some birds derive advantages

under suchconditions, e.g. EurasianBlackbirds (Turdusmerula) nesting inproximityto

buildings benefit fromreduced nest predation from the EurasianMagpie (Møller 1988).

Antlropogenic featues can affect birds in surprising ways, e.g. the Spotted Towhee

(Pípilo maculatus) in Washington parks has better breeding success close to trails,

perhaps because of increased feeding opportunities or by using pedestrians as a form of

predator shþld (Bartos Smith et al 2006). Mourning Doves (Zenaida macroura) in

Texas selected nest sites in proximity to roads br-n away frombuildings and had better

nesting success close to roads and far frombuildings. To this species, roads represent an

excellent foraging opportunity to obtainboth grit and water, whereas buildings present a

distwbance (Muñoz et al. 2008).

As denpnstrated by the exanple of Red-tailed Hawks reducing their alarm call and dive- bombing rates, birds are sonetimes capable of relatively rapid adjustments that alleviate

T6 potential süesses or disturbances. Acquired tameness is beneficial in uban settings in reducing stress and energy expenditure (Chace and Walsh 2006), although there are risks associated with taneness such as the exploitation of tane birds by some human beings or the abandorunent of breeding efforts following a level of distwbance that has been tolerated drning less critical tìmes of the year (Adams 2005). Anthropogenic disturbance can have profound effects on population dynamics @brotti and Annett 2001, Shochat et aL 2OO6), e.g. juvenile Burrowing Owls (Athene cunicularí.a) in urban settings have greater survivalrates than adults because the highly mobile adults are more likely to be killed by collisions with vehicles or other similar accidents. The resuh is that juveniles are recruited into the breeding population faster than norrnal and gain nìore access to the best quality breeding sites. This in turn reduces breeding success because rnost of the breeders are relatively inexperienced (Millsap 2002). Suburban Eastern Screech-Owls in

Texas prodrce rnore offspring thanruralbirds but high mortality rates in the flrst year of life due to vehicle collisions and predation by domestic cats produces faster gene turn- over, higher genetic diversity and a rnore r-selected breeding biology (Gehhach 1988).

On the other hand, the Western Screech-Owl(Megascops kennicottü) l:n;s been shown to avoid subrnban habitats in Baja California unlike its closely related sister species the

Eastern Screech-Owl and possibly unlike some other corspecific populations (Rodríguez-

Estrella and P eIáez Carea ga 2003 ).

The diet of birds in urban and suburban areas can be very different from typical diets

(Robbins L993). Urban areas sometimes offer concentrated food strpplies although these pose risks of toxins and disease as discussed above. One of the best-known exanples of a

t7 concenftated food sryply is rubbish dunps, whbh encourage gulls, crows and other birds to forage preferentially inproximity to hurnan settlements. Bfud feeders have a similar effect for a different guild of bids. Sewage and organic waste provides a rich food source for sone gull species and their numbers have been shown to decline when sewage is properly treated (Raven and Coulson 2001). Anthropogenic food often results in increased concentrations and densities of certain species (Botelho and Arrowood 1996).

The indirect consequences of such foraging behavior are more complex, e.g. this can lead to increased innaspecific and interspecific interactions altering socialdominance hierarchies and feeding schedules (Ditchkoff et aL 2006). Sorrp concentated urban food sources offer poor nuúitional value with potentially significant consequences, e.g. sr¡burban American Crow (Corvus brachyrhynchus) nesthngs are smaller than rual nestlings and have reduced protein and calcium levels (Heiss et aL 2006), sr.rburban Blue

Tiæ (Parus caeruleus) and Great Tits (P. major) suffer from poor nutrition and fledge

50Vo fewer young than woodland conspecifics (Cowie and Hinsley 1987), and House

Sparrows that rely on human food sources exhibit altered physiology with greater nirogen & cholesterol levels (Gavett and Wakely 1986).

As with artificial food sources, artificial water sources can also have profound effects, e.g. Common Ravers (Corvus corax) in the Mojave Desert cluster nest around antlropogenic water and food and produce more fledged young, potentially impacting other species breeding in the area (Kristan 2001). Proximity to irrigation systens is implicated inthe colonizationbyDark-eyed Juncos (Junco hyemalis) of some urbanareas of southern California (Yeh and Price 2004). The watering of suburban lawrn (along with

18 reduced predation) has extended the breeding season of the Rufous-banded Honeyeater

(Conopophila albogularzs) (Noske 1998). Open water in winter as a result of dams, sewage works or effluent from factories has enabled rnany waterfowl to winter either fuither north than they previously did or in greater numbers than previously possible

(Sugden et aL I974, Wobeser 1997), although this behavior is associated with additional diseases and pollution effects (Wobeser 1997). Open water combined with the nutritionally rich food supply fromcrops has meant that sone goose species for whbh suiøble overwintering sites may formerly have been limiting (Alisauskas 1998, Gauthier etaL2005, Sherryet al. 2005) have now greatly expanded theirpopulations (Ankney

1996, Alisauskas 1998, Abrahamet aL 2005, Garfhier et aL 2005). In some cases this results in potentially detrimental effects to the floral subsftate of arctic breeding grounds

(Ankney 7996, Abraham et al. 2005).

Along with dietary changes conæS an array of potentially profound biological changes.

Ahered winter and pre-nesting diets, in particular, can alter both the time and ouþut of breeding activity, e.g. urban Australian Maglies (Gymnorhina tibicen) have advanced breedingdue to food and wateravailability (Rollinsonand Jones 20OZ). SubrnbanFlorida

Scrub-Jays (Aphelocoma coeruIescens) breed three weeks earlier than rural jays due primarily to increased foraging efficiency (Fleischer et al. 2003). This advanced breeding may in fact result in lower fecundity due to different food resowce availabilities at different times of year (Schoech and Bowman 2001). Food concentrations alter other behaviors suchas roostingpatterns (Robbins 1993, Kristanet aL 2003), e.g. urban

19 European Starlings behave differently from rural starlings, exhbiting high fidelity to roost sites near food (Johruon and Glahn 1994).

The starling example provided above highlights an apparent tend of increasingly sedentary behavior in urban settings due to supplenrentary resources and climate mitigatio n thro ugh the urban heat is land and a reductio n of the impact o f wet/dry c yc les due to water provisioning (Adams 2005). With increasing sedentariness comes other population changes such as higher densitbs (Tomialojc and Gehhach 1988, Rosenfield et aL 1996) and altered territoriality and habiøt selection (Newton 1986). Interesting examples of this phenomenon include Eurasian Sparrowhawks (Accipiter nisus) tn villages in The Netherlands, which enjoy a different diet than rural birds, utilizing a

gteater variety of nest types and producing larger clutches @iermen 1996); Ospreys

(Pandion haliaetus) sponsored by urban fish stocks altering their nest-sle selection

(Spitzer et al. 1985); and urban Cooper's l{awks with access to concentated prey sources reducing their territory size and altering nest-site selection with a preference for non- native ftees as nesting subsúate (Chiang et aL.2006).

Diet is also implicated in some Íþre subtle changes, e. g. sububan Arrprican Crows subjected to a low quality food regime occupy srnaller territories bú with rnore heþers at each nest, producing higher nesting densities in these areas. This leads to higher nesting success and post-fledging survival despite the fact that these crows produce srnaller young with fewer offspring per nest (McGown 2001). Anerican Crows also exhibit a high degree of suburban to urban dispersal producing what Marzluff et at (2001) terned

20 a"habitat sponge". The sponge is created by continual developrrent in the sgbrnbs

corstantly producing new early successional habitats favored by the crows and the

subsequent dispersal of the large numbers ofjuveniles born in these suburban habitats

into urban feeding areas (Marzluff et al. 2001). Urbanization affects much rnore than

species-specific dietary regimes and foraging strategies. It influences the whole ftophic

web such that the relative inportance of top-down conüols such as predation versus

bottorrrup limitations such as the limited availability of resources canbe altered

significantly in urban areas (Faeth et al. 2005).

Changes to the process of predation in urbanizing areas are more disputed than other

aspects discwsed above. Several studies have reported increased predation in urban areas

(see Flaskell et al. 2001) dr:e to "sponsored" populations or srpplenpntal food resources

(Chace and Walsh 2006), abundance of mesopredators due to exclusion of larger predators, irrcluding non-native predators sponsored by humans as in the case of domestic cats (Crooks and Soulé 1999, Rollinson and Jones 2002, Sorace zo0z, Woods et aL

2OO3). Other studies have reported decreased predation in ubanareas, e.g. Northern

Mockingþirds (Mimus polyglottos) (srracey er al. 2006), commonwoodpigeon

(Columbiø palumbus) (Tomialojc 1980) and Lesser Kestrel (Falco naumanni) (Tella et aL 1996) all experience lower predation in subwban areas. The F¿stern Screech-Owl apparently benefis from reduced Great Horned Owl densities (Gehlbach1994, see also

Chapter 2). The Rufous-banded Honeyeater e4joys higher reproductive srrccess in the city of Darwiru Arsfralia than in the surrounding counfryside due at least in part to a reduction in predators and a lengthened breeding season (Noske 1998). Yet other studþs

21 have reported no significant difference in (sub)urban versus rural predation rates

(Coleman and Temple 1993) or suggested that domestic cats and dogs have negligible impacts on the presence of srnall birds (Parsons et al.2O06). At least some examples of mesopredator release nray be only temporary as large predators that desert developing areas may returnto those same areas as habitats rnature (Munyenyembe et al. 1998).

Even rnore controversial thanpredation rates are the effects on the process of brood parasitism in developing areas. The debate around parasitism exterds beyond the context of urbanization and pertains mostly to the controversial hypothesis that fragnentation

(increased edge) increases both parasitism and predation rates. This hypothesis has received some slpport (e.9. Wilcove 1985, Robinson et aL 1995, Hobson and Bayne

2000, Rodewald and Shr¡stack2}}7) while other arÍhors have found no supporting evidence @aton 1994,Yahner 1996, Tewksbury et aL 1998, Keyser et aL 1998, Harrison and Bruna 1999,l-.a,hti 2001). One recent study found sftong support for increased parasitism from the Brown-headed Cowbírd (Molothrus ater) with increased human presence (density of farns and houses) but suggested that the relationshþ berween fragrrnntation and predation was complex and varied at different spatial scales

(Tewksbuy et aL 2006). Several recent review papers have suggested that studying fragrrentation rates per se cannot produce meaningful generalizatiors but that rather species and conteK specific interactions must be explored (Haskell et al. 2001, t¿hti

2001, ChaHoun et aL.2002). Furthermore, the effects of fragmentation on a given species may not be sftaightforward, e.g. the Northern Saw-whet Owl (Aegolius acadicøs) may

22 benefit from small amounts of fragnentation but suffer reduced reproductive success as

the level of fragrrrentation increases (Hinamand CassadySt. Clair2008).

Alongside the discussion on the impacts of fragnrcntatioq a similar debate regarding the

influence of rnbanization on brood parasitism and predation has begun The hypothesis

that urbanization increases brood parasitism and predation rates has received some

support (e.g. Nilon et aL 1995, Rerrps 2000, Haskell et al. 2001) but has also been

criticized (t¿hti z}OI,Marzluff et al. 2001, Smith & Wachob 2006). Some artificial nest

experiments found increased predation with irrcreasing human density and/or agrbultural

land use (Wilcove 1985, Andren1992, Major etal.1996, Bayne and HobsonI997, Wong

et al 1998, Hogrefe et al. 1998, Saracco and Collazo 1999, Jokimäki and Huhta 2000,

Haskell et aL 2001, Thorington and Bowman2003); however, other experimental studies

found no significant trend (Russo ard Young 1997, Danre,lson et aL 1997 ,DeGraaf et aL

1999, Gering arld Blair 1999, Melarrpy et aL 1999, Tegers et al. 2000). Other studies have suggested that avian nest predators may benefit from urbanization (Kluza et aL

2000, Smith and Wachob 2006) although this is countered by Marzluff et aL (2001).

Sone recent evidence suggests that any effect of human density is unlikely to be uniform br¡t rather strongly influenced by regional processes (Pllgeon et al. 2007). As with the fragnrentation debate therefore, it is extrenely difficult to make generalizatiorn on the effect of rnbanization on the processes of predation and parasitism and wiser to consider context specific examples.

23 Changes in predation and parasitism interactions derr¡cnstate the influence of other

aspects of avian ecology such as habitat selection and productivity in urban areas.

Eurasian Blackbirds that breed in small fragnents in urban settings have been shown to

alter their nest-site selection to nest preferentially in conifers and hedgerows (dense

cover) and in proximity to buildings (Møller 1988). A particularly interesring aspecr of predation in urban settings is the possibility of ecological ftaps and population sinks, for example, House Sparrow, Dunnock (Prunella modularis), and European Robin all experbnce ecological ftaps as a result of high predation by domestic cats on juveniles

(Baker et al. 2005).

Urbanization in fact creates a distinctive pattern of fragnentation (Bierwagen 2007).

Altered plant species composition (including the presence of many exotics) and süuctural corrplexity (RottenbornI999, Savard and Falls 2001) as well as different age and maturation profiles (Munyenyembe et al. 1998) make urban environmenfs sfructurally different from any surrounding landscape with different avian community composition

(Batten 1972,Walcottl974,Aldrich and Coffin 1980, Horn 1985). InFlorida rhis produces increased shrub density in sone urban areas to the benefit of Northern

Mockingbirds and Northern Cardinals but to the detriment of Acadian Flycatchers

(Stracey and Robinson 2006). Some species have been able to exploit these differences and used urban areas as stepping stones for range exparnior¡ e.g. by exploiting non- native fruiting nees with different fruiting phenology from native species (Reichard et aL

2001, Corlett 2005). I will argue in Chapter 6 that the Eastern Screech-Owl has benefited fromanthropogenic activities to expand its range into Manitoba, Canada. Differences in

24 urban environments canproduce more subtle effects such as conpetitive exclusion e.g.a

checkerboard distribution pattern aÍþng several species of Ptilinopøs fruit doves has

errerged in Papua New Guinea as a result of the availability of different fruiting frees in

different towns (Diannnd 1986). On the Canary Islands, the House Sparrow dominates in

urban areas, the Spanish Sparrow (Passer híspaniolensis) dominates in towrs and

villages, and the Rock Spanow (Petronia petronia) is thereby confined to the

surrounding countryside unless its competitors are absent (Cullen et aL 1952, Summers-

Smith1988, Newton 2003). Urbanization also influences breeding success, for example,

despite the fact that raptor diversity is correlated to wilderness area (Bosakowskiand

Smith 1997), several raptor species have higher breeding success in urban areas (Bloom

and McCrary 1996, Botelho and Arrowwood 1996, Parker 1996, Tella etal. 1996,

Rosenfield et aL 1996).

A particularly complex and fascinating aspect of this phenomenon is the way urbanization influences the process of habitat selection often in indirect ways such as by altering foraging opportunities or nesting subsftates (Haskell et aL.200I, Chace and

V/alsh 2006). Interestingly, such effects are not confined to urban areas but may also spill over into surrounding forests (Rottenborn 1999). Well known exanples include Peregrine

Falcon (Falco peregrinus) nesting on urban high-rises as opposed to cliffs and the accorrpanying change in diet that this entails (Horak 1986, Septon et aL 1996, Tenple

1988). The link between ahered habitat selection and dietåry changes is a common theme and has been noted in other species such as the Anerban Kestrel (Falco sparverius) which accepts artificial nest boxes as a nesting resource when these occur alongside

25 concenfrated food sources (Beny et al 1998). Great Horned Owls in suburban areas exhibit higher nest-site selectivity thanruralbirds dræ to the increased sfructural corrplexity of subtuban environments (Smith etal. 1999). Monk Parakeets (Myiopsitta monachus), which are habitat generalists in their native Sotúh Anerica, exhibit very specific nest-site selection for tall palms in the city of Barcelona where they have been introduced (Sol et aL 1997). Tawny Owls (,Srrux aluco) in northern Italy settled in adjacent territories forming an aggregation in large wooded patches in rural areas but settled at a distance fromearlier arrivals in fragmented urban areas (Galeotti 1994).

Another much discussed attribute of urbanization is landscape connectivity and how birds

Íþve between habitats in the urban mafrix. Urban obstacles such as roads can have strong negative impacts on many species by blocking or slowing their moverrent (Knight

1990, S oulé I99 1, Shafer 1 997, S o ul é et al. 2005) ; howe ver, other d is turbance - adapted species such as crows and some non-native species are capable of using roads and other feattnes of the urban mahix as conidors that aid dispersal (Crooks and Soulé 1999,

D'Antonio and Meyerson2}l2, McKinney 2006). The issue of conrpctivity is argu,ably errrrging as less of a concern to avian ecologists than to mammalogists and herpetologists as many authors have concluded that patch size is more significant than connectivity to birds in urban landscapes (Bunnell 1999, Hanison and Bruna 1999,

Clergeau et al. 2001, Thornas et al. 2001, Lichstein et aL.2002, Alberti and Marzluff2}} ,

Bierwagen 2005, Donnelly and Marzlutf 2006). Some authors have argued that the reason that connectivity in the urban matrix poses less of a problem to birds may be because the mobility afforded to themby flight permits detecting environmental featwes

26 readily, allowing birds to nnve tllough non-habitat into suitable habitat (Zollner and

Lima 1997, Donnelly and Marzlutr 2006).

When the degee of connectivity in a landscape as perceived by a given species is altered,

this has multiple conplex effects on nurnerous aspects of that species' life cycle. These

effects are not confined to dispersal Increased connectivity can alter immigration /

emigration rates, leading to greater genetic honngeneity (Bierwagen2007); increase

disease trans mis s ion rates (S imberloff and Cox 1987) ; hasten paras ite infection rates

(Simberloffand Cox 1987); facilitaæ the spread of invasive species (Simberloff and Cox

1987); and increase the impact of distrnbances such as flre (Bierwagen 2007). The

strongest effects on cormectivity are argued to be in areas where large cities have arisen

in formerly contiguous habitat and the smallest effects where urbanization has occured in formerly disaggregated habitats (Bierwagen 2007).

Urbanization is proving to have evolutionary consequences for many birds by rreans of a variety of nechanisms. Urban Herring Gulls (Larus argentatus) have faster heritable mutation rates (Yauk and Quinn 1996) and sububan Eastern Screech-Owls have ahered rates of genetic input and outflow (Gehhach 1988). In other cases, sexual selection is altered due to changes in contextual biological demands, e.g. following colonization of an urban area in California, a populationof Dark-eyed Junco enjoying mollified climate (in terms of both terrperature and water provisioning) extended their breeding season and e4perbnced reduced competition for mates but increased selection pressure for parental investment. The result was a 22Vo decrease in the amount of white in their tail feathers

27 (used for sexuaUterritorial signaling) in less tl:ø,nZí years (Yeh 2004, Yeh and Price

2004). Urban House Sparrows in Hungary are smaller in size and peqpetually in worse physiological condition than rural conspecifics. This is not due to short-term resorirce availability but rather to adaptive divergence in urban environments (Lker et al. 2008).

Urbanization is also producing genetically determined sedentariness in Eurasian

Blackbirds (Partecke et al. 2005, Partecke and Gwinner 2007).

The pictue emerging fi'om the above discussion is the great conplexity of interrelated factors and the potentially profound consequences of even very srrall habitat changes or avianresponses to small changes. This rneans that when assessing the patterns of urbanization and the conseqr¡ences for birdlife, cause and effect is often difficult to ascribe, e.g. the absence of certain passerine species in sububan areas of Sydney,

Australia might be related to characteristics of suburban gardens in those areas, conpetition and aggession fromdominant and innoduced species, or irrcreased predator dersities (Parsons et aL 2006). In other cases, different populatiors of the sanre species appear to have very different responses to urbanization in different geographic areas, e.g. the Red-shouldered Hawk (Buteo lineatus) avoids suburban habitat in New Jersey

(Bosakowski and Smith 1997), uses non-native üees as a nesting substrate in California

(Bloom and McCrary 1996) and shows no significant difference in habitat selection in urban versus rural Ohio other than using residential areas less than expected (Dykstra et aL 2000, Dyksra et aL 2001).

28 The Red-shouldered Hawk exarrple illustrates both the difficulty in finding meaningful generaltzations and the irrportance of understanding context-specific conditions. In other words, urbanization is a conplex process (or processes) further complicated by the fact that it functions in concert with other factors such as climatic and geographical clines, metapopulation structure, and other biological and ecological characteristics under regionally variable selectionpressures. In this thesis, therefore, I chose to research the

Eastern Screech-Owl, a reasonably well-studied specbs (VanCamp and Herury I975,

Gehhach 1994) which is known to be suburban-adapted at least in the southerly portion of its range (Gehhach 1994). Suburban Eastern Screech-Owls possess advantages in terms of fecundity and population recruitnent over nral populations (Gehlbach 1994).I conducted my research at the very northern perþhery of the Eastern Screech-Owl's range in order to test whether the patterns and nechanisms of subrnban adaptation observed in the south might hoH tue in the north The study area for this thesis (Winnipeg,

Manitoba) is above 49'N and experiences a radically different climate from southern

Texas where the pattern of suburban-exploitation by the Eastern Screech-Owl was first documented (Gehlbachl994). Since Gehlbach (1994) denpnstrated the benefit of the urban heat islard to Texan owls in terms of advanced breeding phenology, my initial prediction was that population density and reproductive success of Eastern Screech-Owls in V/innþeg, where winter temperatures approach the cold tolerance threshold of these birds (Mosher and Henny 1976), would also be influenced by the urban heat island. I also predicted that antlropogenic habitat features including the use nest-boxes and buildings as roost sites would benefit this specbs as has been suggested for the Barn Owl (Tyto alba) at the northern edge of its range in North Arrerica (Andrusiak and Cheng 1997).

29 Locar, CoNrpxr: A CBvrunyop CnaNcr IN Wrr.wrppc's Avmnwn

Corsidering the difficulty in providing generalization on the effect of urbanization on birds and noting the variation in the way birds have reacted to urbanization in different contexts, it is worth outlining the changes that urbanuationhas brought to the avifauna of

junction the study area of this thesis - Wirmipeg, Manitoba. Centered on the of the Red and Assiniboine Rivers, Winnþeg is situated in an ecological nansition zone, the aspen parkland belt that lies betr'veen the boreal shield to its north and east and the grassland ecosystems to its south and west. Despite the establishment of a small Hudson Bay

Conpany settlement in 1811, Winnipeg did not experience significant urbanization until well after its incorporation in 1872 (Boyens 2007). Nonetheless, considerable changes to the area's åuna and flora have occr¡rred within that time span

European settlement of North America brought several innoduced birds of Palearctic origin to the Winnipeg area, rnostly indirectly, ie. following range expansion subsequent to introduction ineasternparts of the continent. These include Rock Pigeon(Columba livia), House Sparrow, EuropeanStarling and GrayPartidge (Perdíx perdix) (Manitoba

Avian Research Committee 2003). The first three of these species have beconp the dominant residents of Winnipeg's urban area. This is mt surprising since urbannation creates suitable conditions for alien species, of which many are human symbionts

(Jokim¿ikiand Suhonen1993, BIah 1996, Adams 2005). These introduced species have likely increased conpetition for food and nesting and roosting sites, with a profound impact on many native species. Nonetheless, they are important prey items from several urban raptor species including Peregrine Falcon (R. W. Nero, unpublished data), Merlin

30 (Falco columbarius) (Oliphant and McTaggart 1977, Manitoba Avian Research

Committee 2003), and Eastern Screech-Owl (Chapter 4). House Finch (Carpodacus mexicanus) has also become a commonresident in the Winnipeg area and a regular visitor to feeders, after being brought from sor¡hwestern North Anerica to New York and then spreadìng westward across the continent (Aldrich and Weske 1978), reaching

Manitoba in the early 1980s (Koes 1985). Ground-dwelling birds are relatively poor dispersers and, along with ground-nesting species, are thought to be particularly vulnerable to disturbance from human activity and donestic pets, leading to their exclusion from most urban areas (Rottenborn 1999, Cicero 1989, Winters and Wallace

2006).It is therefore noteworthy that the innoduced Gray Partridge is one of only two ground bird species (Phasianidae, Odontophoridae) that breed within the city limits. The other is the Wild Turkey (Meleagris gallopavo), also intoduced into Manitoba (Manitoba

Avian Research Committee 2003). The Wild Turkey breeds in the Saint Norbert area jmt south of the city and apparently moved into V/innipeg during the 1997 Red River flood, whenbirds were pushed northward by the advancing floodwater and subsequently established themselves in certain suburbs such as Wildwood Park (G. Ball, pers. comm.).

Native ground birds have fared more poorly with urbanuation, both the Ruffed Grouse

(Bonasa umbellus) and the Sharp-tailed Grouse (Tympanuchus phasíanellus) having disappeared from Wirmipeg. The former was last recorded on a Winnipeg Christmas Bird

Count (CBC) in 1978 ard the latter in 1998 (National Audubon Society, nd.).

In addition to intoduced species, several suburban-adapted native species are common in

Winnþeg. The Chimney Swift is highly urbanized in Manitoba and elsewhere inNorth

37 Anerica (Johnston 2001). Common Grackles (Quiscalus quiscula) utilize feeders and

sorretimes form large flocks on suburban lawns where they probe for food. They also use

many hurnan sftucttnes as a nesting substate (C. Artuso, pers. obs.). This specbs reaches

its highest numbers in Manitoba in residential areas but despite increases in the 1980s

appear to have declined since 1995 (Manitoba Avian Research Committee 2003). Other

interesting adaptations to suburban life include rubbing their feathers with granular lawn

chemicals as a substitute for anting (Nero 1996). Black-capped Chickad ee (Poecíle

atricapillus) and V/hite-breasted Nuthatch (Sitta carolinensis) are other exanples of

native specbs that are common in the suburban areas.

Winnþeg's geographicalpositionplaces it onthe northernperiphery ofthe range of a

suite of species of southeastern woodlands, sorrrc of which have apparently expanded

their range over the course of the late 19th and 20ú Centuries, e.g. Wood Duck (Alx

sponsa),AmericanWoodcock (;Scolopax minor),Barred Owl(Srrlx varía),Eastern

S creech- O w I, Red- headed Woodpe cker (M e lane rp e s e ry t hro c e p halu s), Ye llow- throa ted

Vireo (Vireo flavifrons), Ptuple Martin (Progne subis), Golden-winged Warbler

(Vermívora chrysoptera), EasternBluebird (Sínlin sinlis), and Indigo Bunting (Passerina

cyanea)(ManitobaAvianResearchCommittee2003,Chapter6). Inadditiontobirds,

sonrc mammals such as the raccoon (Procyon lotor) have undergone a similar exparnion

(l-arivière 2004). Some of these species now corprise a regular component of

Winnþeg's avifaun4 e.g. Wood Duck, Yellow-throated Vireo and Indigo Bunting,

whereas others such as Anerican Woodcock and Eastern Bluebird utilize suburban habitats as migration stopover sites (C. Artuso, pers. obs.). Other open-country or edge

32 species whose population shifu, range expansions or movenænts have brought them to the V/innipeg area in the late 19th and 20th Century include Mor-nning Dove (Zenaida macroura), Western Kingbird (Tyrannus verticalis), and Barn Swallow (Hirundo rustica)

(Thompson [Seton] 1890, Speechly 1922, Horston 1979, Houston 1986, McNicholl 1988,

Houstonand Schmutz1999). Other species that have not undergone range expansionper se have shown increasing adaptation to urbanized environments and becone a regular part of Wirrnipeg's avifauna including American Crow (Corvus brachyrhynchos)

(Houston 1977a) and the so-calbd "prairie" sr,rbspecies of Merlin (F. c. ríchardsonü), which first colonized Winnipeg in the 1980s (McCowan 1978, Bancroft 1989, Houston and Schmutz 1999). After a pattern of range collapse and subsequent expansion Black- billed Magpie (Pica hudsonica)breeds in certain areas of the periphery of the city

(Houston 1977a, Manitoba Avian Research Committee 2003) such as the University of

Manitoba caÍpus (C. Artuso, pers. obs.).

Neotropical migrants are generally thought to be disadvantaged by ubanization (Friesen et aL 1995, Lindsay etaL.2002, Greenand Baker 2003. Kark et aL2006); however several species breed in Winnipeg. Among Neotropical migrant passerines that breed in

Winnþeg suburbs and parks are Great Crested Flycatcher (Myiarchus crínitus), Eastern

Kingbird (Tyrannus tyrannus), Eastern Phoebe (Sayornis phoebe), Red-eyed Vireo (Vireo olivaceus), Yellow-throated Vireo, House Wren (Troglodytes aedon), Gray Catbird

(Dumetella carolinensis), Cedar'Waxwing (Bombycilla cedrorum),Indigo Bunting,

Yellow Warbler (Dendroica petechin), Chþping Spanow (Spizella passerina), Clay- colored Sparrow (Spizella pallida), and to a lesser extent Rose-breasted Grosbeak

JJ (Pheuticus ludovicianus) (C.Artuso, unprúlished data). Some of these use human stuctltres to varying degrees, e.g. EasternPhoebes readily nest onbridges, docks, abandoned and even occupied buildings (Manitoba Avian Research Committee 2003).

The Yellow-throated Vireo reaches its northwestern range limit in Manitoba and appears to reach its highest dersities in suburban parks in Winnipeg, e.g. Assiniboine Park,

Kildonan Park, Saint Vital Park and King's Park or jrst orÍside the cify, e.g.I-a Barrière

Park (Manitoba Avian Research Committee 2003). These parks all possess very large deciduots riparian trees and are also attractive to a suite of open woodland resdents including Eastern Screech-Owl, Downy Woodpecker (Pícoides pubescens), Hairy

Woodpecker (Picoides villosus), Black-capped Chickadee, and White-breasted Nuthatch

The eastern subspecies of the Loggerhead Shrike (Lanius ludovicianus migrans) is unusual in that it is a threatened subspecies (Government of Canada, nd.) that, at least in

Manitoba, is associated with low-density human development, curently being restricted to areas on the northern boundary of V/innipeg (Blouin et aL 2001, Manitoba Avian

Research Committee 2003).

Nurrprous other species have undergone relatively recent and sometimes remarkable colonization of the urban and subtnban areas. Some of these cases may be linked to the provisioning of nest boxes, e.g. the Wood Duck was once rare in Manitoba and elsewhere

(Manitoba Avian Research Committee 2003); however, the suburbanbreeding populatio4 many of which use nest-boxes, reach seemingly large concenftations in parts of the city such as Saint Vital and Kildonan Park, where they are also fed by park visitors and nearby hone owners (C. Artuso, unpublished data, Artuso 2006). F¿stern Screech-Owls

34 also readily use nest-boxes and 347o of.the nest sites found between 2004 and2007 n

Winnþeg were in boxes designed for Wood Ducks (Chapter 2). Eastern and Mountain

Bluebirds (Sialia currucoides) regularly use nest-boxes in Manitoba and the former once nested in Winnipeg after range exparsion in the late 19th century (Hou:ston lg77b) although they no longer appear to do so (C. Artuso, pers. obs.). House Wrers have frequently rsed nest boxes and other human structures in V/iruripeg since l92I and are typically encountered near human habitation (Manitoba Avian Research Committee

2003). Like Chimney Swifu, Cliff Swallows (Petrochelidon pyrrhonota) now use anthropogenic structures almost exclusively as a nesting substate, especially bridges and darns, and are readily found in suburban areas (Manitoba Avian Research Committee

2003).

Urbanization entails habitat change that typically favors open-country or light woodland species (Aitken et aL 2002, Diamond 1986, Kluu et aL 2000). Not surprisingly, some open-woodland species such as Downy Woodpecker are rrnst likely to be encountered in

Manitoba in suburban or rural areas with large ftees (Manitoba Avian Research

Committee 2003). Many suburban areas differ from the surrounding aspenparkland in that they contain a more open understory and exotb specbs including fruit-bearing and coniferous frees. Conifers have become an important nesting substrate for several specbs including Amerban Crow, Black-billed Mag¡lb, Merlin (which reuse crow and occasionally magpie nests in conifers) and Chipping Sparrow (Manitoba Avian Research

Committee 2003). Conifers are implicated in the colonization of some prairie cities by the

Merlin, which only nest in conifers in these cities (McCowan 1978, Bancroft 1989).

35 Conifers have apparently facilitated the boreal forest-dwelling Red-breasted Nuthatch

(Sitta canadensis) to nnve into certain parts of the city in very recent years to attenpt

breeding (R. W. Nero, pers. comm.). This pattern also appears to be increasing in North

Dakota (C. D. Ellingson, pers. comm.) and Minnesota (Wann et al. 1981). Pine Siskins

(Carduelis pinus) occasionally breed in Winnipeg and other cities and towns. According

to the Prairie Nest Record File, their nests in these areas are almost invariably in planted

spruces, matching their general preference for conifers (Middleton 1998). Urban conifers

also appear to be a key elenent in the nest-site selection of Chipping Sparrows in

Winnþeg. Thorrpson lSeton] (1890) considered the Chþping Sparrow rare in Manitoba

b¡f "tolerably comrmn" in Winnipeg in the late 19th cerìtury. This suggests that this

species has enjoyed a relatively long association with ubanizatioq fitting with its preference for edges, clearings and open habitaa (Middleton 1998). Eastern Screech-

Owls select nest-sites close to conifers that provide a concealed and protected roost site for the male in the early nesting period before deciduous ftees have foliated (Chapter 3).

The planting of conifers therefore can have a corniderable effect on the avifauna of northern te nperate c ities.

Sorrp species that may have been excluded from the Winnipeg area as it was developed due to persecutior¡ disturbarrce or habitat alteratior¡ may move back to the city as habit¿ts rrìature. Due to their large spatial requirerrents, many large predators carurot easily establish home ranges in fragnented uban areas (Crooks and Soulé 1999

Jokimäki 1999, Chace and Walsh 2006). Some raptors are therefore sensitive to distubance and have also been the victims of direct persecution The Great Horned Owl

36 is not common in Winnipeg; however, the Prairie Nest Record File shows that some pairs

have nested in the city since 1958. I have found 4 - l0 nests annually in the city and

surroundingarea since 2005. Sonp pairs have used human structures as nest sites, e.g. a

steel tower in a Winnipeg oil refinery (Manitoba Avian Research Committee 2003) and a

sewer pipe no longer in we (C. Artuso, pers. obs.). Great Horned Owls exhibit a greater

degee of nest-site selection in suburbanareas than in nnal areas essentially because they have a greater variety of options available (Smith et aL 1999). CBC data show that the highest detection rates, at least in winter, occur in urban and agricultural areas (Manitoba

Avi¿n Research Committee 2003) although the rate of detection in Winnipeg has rennined fairly constant since the 1960s (0.04 individu,als per party hour in the 1960s,

0.02 in the 1970s, 1980s, and 1990s, and 0.03 from2000 -2007 (National Audr¡bon

Society, nd.). Cooper's Flawk have an increasing breeding presence in suburban areas,

and since the 1990s have nested annually in Assiniboine Park (Manitoba Avian Research

Committee 2003).I now find or am informed of at least a dozen nests every in year in

Winnþeg greenspaces. Occasionally two pairs have nested simultaneously within

Assiniboine Park ard Saint Vital Park as close as 7Z5mapart (C. Artuso, urpublished

data). After a significant recovery from near population collapse (Gerrard and Bortolotti

1988, Koonz 1988), the Bald Eagle (Haliaeetus leucocephalus) now breeds along the Red

River just north of Winnipeg (G. Machrrce, pers. comm.). Bald Eagles attenpted nesting in several consecutive years along the Red River in Kildonaru however, they abardoned this site aftet a new housing developrrrcnt was erected (R. Koes, pers. comm.).

37 ln addition to raptors other bids have been frequently disturbed by human activity.

Common Raven (Corvus corax) was also once persecuted and disappeared as a breeding

species from the Canadian prairies in the 19th century (Houston 1977a). Nonetheless,

ravens have recently begun breeding in the Canadian prairies including several locations

in WirLnþeg (Reaurrn 2006, C. Artuso, pers. obs.). American Crow has also dramatically

increased its breeding presence in Winnipeg and other cities and towns on the Canadian

prairies (Houston 1977a). The Pileated Woodpecker (Dryocopus pileatus) was once

frequently shot by hunters and trappers and is relatively intolerant of disturbance (Bull

and Jackson 1995). A bird of mature woodlands with large ftees, requiring relatively

large forest fracts (Bull and Jackson 1995), few would expect to find this specbs in urban

areas. Nonetheless, the number of sightings in the city appears to be increasing in recent

years as Winnipeg's wban forest rnatures. This species was first dete cted on the

Winnþeg CBC in 1947 bú. then not agatn until 1999. From 1999 -2002 and n2004,

one or two were recorded on each Winnipeg CBC. In 2003, 2005, 2006, and 2007 tluee

or four were recorded on each count (National Audubon Society, nd.). Pileated

Woodpeckers are also occasionally encountered in large suburban parks in the summer

and reprodrced at the St. Charles Golf and Counüy Club in 2000 (R. Koes, pers. comm.).

Urban areas provide concentrated food resources for numerous birds, e.g. nrbbish dumps

and sewage outlets are frequented by corvids and gulls (Robbins 1993, Botelho and

Arrowood 1996, Berry et aL 1998, Raven and Coulson 2001). There are several large

durrps on the outskirts of Winnipeg, including the Brady Road l¿ndfill just south of the

city's Perimeter Highway, where thousands of Ring-billed (Larus delawarenszs), Herring

38 Gulls and other species congregate to feed, roosting in either flooded fields or on the lakes of the nearby FortWhyte Alive (C. Artuso, pers. obs.). Urban areas are often associated with increased avian abundance but lower avian diversity (Blair 2001,

Marzluff 200T, McKinney 2006, Palomino and Carrascal2006, Smith and Wachob 2006).

As nentioned above, part of this abundance is made up from non-native specbs, freqræntly preyed tponby W'innipeg's raptor commtrniry e.g. Rock Pigeons are the rnost important single prey species for Peregrine Falcons, constituting 237o of their dbt

(Sliworsky and Nero 2003).

The various parks, greenways, retention ponds, and even lawns or vacant lots alongside major thoroughfares in Winnipeg now accommodate increasingly Ìarge breeding populations of waterfowl. Despite almost disappearing as a breeding species from southern Manitoba in late 19th Century (Manitoba Avian Research Committee 2003), the so-called "giantCanadaGeese" (Branta canadensis maxima) have adapted to uban greenspaces, especially golf courses and parks with water sources, where they have increased dramatbally in abundarrce to the extent that they are now considered problem wildlife (Manitoba ConservationWildlife and Ecosystem Protection Brancl¡ nd.). At

FottWhyte Alive they nest within a few meters of the entrance to the main interpretive building and are sometimes aggressive towards visitors (C. Artuso, pers. obs.). In addition to being a common subtnbanbreeder, their numbers are swollenby migrants srch that, at the peak of their fall migration, Winnipeg is host to 120,000 Canada Geese

(Manitoba Conservation Wildlife and Ecosystem Protection Brancl¡ nd.). Mallards

(Anas platyrhynchos) also breed in suburban areas, often using duck nesting tunnels

39 supplied at local nature centers or on ponds in greerspaces. They are also able to nest

directly on the ground in sorre suburban areas in secluded sites such as under bridges (C.

Artuso, pers. obs.). Wood Duck numbers have dramatically increased in the latter half of

the 20th Cenhly and they are now commonbreeders in suburban areas (Manitoba Avian

Research Committee 2003). Arrerican White Pelican (Pelecanus erythrorhynchos) are

now regularly observed along the Red and Assiniboine Rivers within the city's perimeter

and also on retention ponds (C. Artuso, pers. obs.), although I am unaware of any

breeding within the city, perhaps because these birds are highly sensitive to hunnn

disturbarrce at nesting colonies (Knopf and Evans 2004). Pied-billed Grebes (Podilymbus podiceps), Arrprican Coot (Fulica americana), and to a lesser extent Hooded Mergansers

(Lophodytes cucullatus) also breed within the city limits where suitable wetlands are

found and disturbance levels are relatively low, sirrce all three are more likely to abandon

nests due to hurrnn disturbance than suburban Mallards or Canada Geese @ugger et al

1994, Muller and Storer 1999,Ißhr and Mowbruy2002).

Argu,ably the nnst drastic changes to Winnipeg's avifauna ha'¿e occurred in winter dæ to

the combined effects of climatic warming and the urban heat island. Some species that

would not normally swvive this far north linger at feeders in the city tlroughoú the

winter. Thirty-eight of 54 marginal or rare winter passerine species recorded from 1973

to 1995 were typically found near human habitation including artificial food sources

(Taylor and Koes 1995). Pine Siskins rarely overwintered in Manitoba until the 1970s brf

then increased dramatically, relying on feeders throughout the winter, such that over 200

have been recorded on the Winnipeg CBC and, n 1994, over 3000 were recorded at

40 Glenboro (Manitoba Avian Research Committee 2003). Mourning Doves were flrst recorded on the Winnipeg CBC in I97I, atñ. then not again until 1976 but they have been recorded on at least one Manitoba CBC every year since except 1983 (National Audubon

Society, nd.). This species provided direct evidence of the effect of feeders on the winter distibution of birds, occurring annually throughout the winter at the sane feeders in

Saint Vital and North Kildonan since the 1990s until there was a change of honp owner

(Taylor and Koes 2007). The northward expansion of the wintering range of this species, with the aid of feeders, has also been docurrpnted elsewhere (Armstrong and Noakes

1983). Anerican Robins (Turdus migratorius) visit feeders less regularly than Mouning

Doves; however, their presence in winter is rnost commonly associated with suburban areas and rural towns with planted ornamental fruit tees such as crab apples (Malus sp)

(Taylor and Koes 1995). Other frugivores such as Cedar Waxwing, which do not comrrnnly winter in southern Manitoba, sonptimes linger in December and January in

Winnþeg when crab apples and mountain-ash (Sorbus sp) berrbs are available, occasionally alonpide the winter-visiting BohemianWaxwing (Bombycilla garrulus) (C.

Artuso, pers. obs.). In other parts of the continent, Cedar'Waxwings also regularly feed on urban fruit trees (Luther et al 2008). Blue Jays (Cyanocitta cristata) are frequent feeder visitors in winter and their numbers have apparently increased in recent winters

(Manitoba AvianResearch Committee 2003). Common Grackles now linger into

December and they are detected almost annually on the Winnipeg CBC, including a high count of23 n1997 (NationalAudrùonSociety, nd.). The flrst winter records of the migratory Anprican Crow from Winnipeg occurred in the early 1970s (Whaley 1972)bvt they have increased drarnatbally in the city such that flocks 50 -100 strong are present

4t rnost winters, with as many as 408 being recorded on one CBC (Taylor and Koes 1995,

National Audubon Society, nd.). Other species that rarely overwinter in Manitoba are sonptimes found in s upris ing concenff ations in Winnipe g, e. g. 24 Bro wn Creeper

(Certhia americana) were found on the Winnipeg CBC in 2001 (Nation¿l Audubon

Society, nd.). Raptors have apparently benefited from warm winter tenperatures and concenftated prey around bird feeders. Merlins and to a lesser etent Sharp-shinned

Hawks (Accipiter striatus), Cooper's Hawks, and Northern Saw-whet Owls are recorded with increasing frequency in the city in winter (Taylor and Koes 2007). Bald Eagles have

also increased their winter presence in Manitoba, apparently benefiting from milder winter and anthnopogenic food sowces such as poultry farms (Taylor and Koes 2007).

They are most lkely to be found in the city in winter along stetches of open water. This species was firstrecorded onWinnipeg's CBC in2001 ard has beenrecorded every year since except"2003 (National Audubon Socbty, nd.).

Open water in winter is another valuable resource and its presence in urban areas at sewage works or sources of effluent from factories has enabled waterfowl to winter in cities (Sugden et aL 1974). Fifty-two rare or marginal species of waterbirds or water- dependent species (Anatidae, Gaviidae, Podicipedidae, Pelecanidae, Phalacrocoracidae,

Ardeidae, Rallidae, Gruidae, l¿ridae, Alcidae, and Alcedinidae) have been recorded in winter in Manitoba and these are often linked to water sources that would norrnally be frozen without human intervention (Taylor and Koes 2007). Some of these were recorded in Winnþeg including Cackling Goose (Branta hutchínsii) and Snow Goose (Taylor and

Koes 2007). Canada Geese and Mallards do not commonly overwinter in Manitoba;

42 however, most examples of successful overwintering are fromWinnipeg, especially from a fast-flowing stetch of the Assiniboine River and the Charleswood Sewage Lagoors

(Taylor and Koes 2007). These species were recorded very sporadically before the early

1970s; however, since 1975 they have both been recorded nearly every year of

Winnþeg's CBC (Taylor and Koes 2007).

In addition to changing species composition other changes are more subtle. The urban heat island has apparently changed the breeding phenology of nurrrerous species, e.g. the

Wood Duck's peak hatching period inManitoba is second week of June (Manitoba Avian

Research Committee 2003); however, since 2006 I have observed females with fledged broods as early as May 24th (C. Arflso, unpr:blistæd data). Cooper's llawks inWinnipeg have laid eggs at end of April, well in advarrce of usual phenology in the second week of

May (Manitoba Avian Research Committee 2003). There is eviderrce of advanced breeding phenology of Great Horned Owls in suburban areas by as much as six weeks owing to the uban heat island (Artuso 2007b). Eastern Screech-Owls in subtnban areas fledge broods almost a week in advance of their rural count"tp*1 (Chapter 2). Spring arrival dates for migrants have also advanced in Manitoba in general under the influence of climate change (Murphy-Klassen et al. 2005).

Despite the many increases in Winnipeg's avifauna, there are also some recerÉ losses.

Baltimore Orioles (lcterus galbula) med to breed within the city of Winnipeg (R. Koes, pers. comm.) brr they no longer appear to do so, despite occurring in the city on passage and maintaining a breeding presence in some rural towrs (C. Artuso, pers. obs.). This

43 species has disappeared fromsome sites afterpesticides were sprayed, whereas populations remained intact in nearby uruprayed areas (Walcotf 1974, Rising and Flood

1998) and, therefore, spraying for nnsquitoes in Winnþeg warranfs investigation for its effect on this and other insectivores. Anprican Kestrel was fornerly a relatively common breeder in citybut is now almost absent (R. Koes, pers. comm.). This species has used nestingboxes and even nested onbuildings in downtown Winnþeg (Manitoba Avian

Research Committee 2003). Although nesting sites would not therefore seem to be limiting in the city, the rernoval of dead trees and suitabþ nesting snags in open areas, conpetition for cavities from Ewopean Starlings and other cavity nesters, and decreasing sr.pplies of certain grass-dwelling irsects such as grasshoppers are all potential factors in this decline that warrant investigation Similarly to the kestrel, the Red-headed

Woodpecker, another cavity-nester of open woodlands with a sparse understory (S mith et aL 2000), has disappeared from many V/innipeg golf courses and parks where it fornerly bred (Manitoba Avian Research Committee 2OO3). It is possible that a similar suite of mechanisms are influencing both this species and the Anprican Kestel Common

Nighthawk (Chorde ile s minor) formerly bred on gravel rooftops in cities s rrch as

Winnþeg but this has become increasingly rare as constuction materials have changed and this species is now only recorded in the city on passage (Manitoba Avian Research

Committee 2003).I-ark Spanow (Calamospiza mel.anocorys) was reportedly once conlnron around Winnþeg (Thorrpson fSeton] 1890) and were reported breeding in

North Kildonan (Cartwright 1931). They no longer breed in the city but are found in nearby Bird's Hill Park (G. Budyck, pers. comm.).

44 True grassland species, or at least those which are area sensitive, are generally absent from the city despite some remnant patches of native prairie, which are presumably too small to support populations of grassland specialists (Munyenyembe et al. 1989, Sodhi

1992). Related perhaps rnore to agriculture and non-native specbs than urbanization per se, a suite of true grassland birds that formerly bred in or around Winnipeg have all undergone a20th Century range collapse to the westernmost parts of the province, including Ferruginous Hawk (Buteo regalis), Bunowing Owf Sprague's Pipit (Anthus spragueä), Baird's Sparrow (Ammodramus bairdü), and Chestnut-collared l-ongspur

(Calcarius omatus) (Thorrpson [Seton] 1890, Manitoba Avian Research Committee

2003, Government of Canada, nd.). Short-eared Owl (Asioflammeas) was once a comrrnnbreeder in Winnþeg (Thonpson [Seton] 1890) and still breeds at Oak

Hammock Marsh north of the city; however, recent sightings in Winnipeg tend to be in winter when these owls may hunt over agricultural fields or where hay or manure piles provide access to roderÍ prey (J. Durcar¡ pers. coÍrm., C. Artuso, pers. obs.). Winnipeg's avian history therefore reflects that of southern Manitoba as a whole, with an increase in parkland and woodland species at the expense of grassland species. Nonetheless,

Winnþeg still possesses a few patches of remnant grassland habitat with varying degrees of native versus exotic species. Le Conte's Sparrows (Ammodramus leconteii ) persist in pockets around the Living Prairie Museurn, Henteleff Park and jurst south of the Perimeter along Clorfier Drive (C. Artuso, pers. obs.). Sedge Wrens (Cistothorus platensis) are also sometimes found in similar locations although these are perhaps rrnre dependent on water level (C. Artuso, pers. obs.). Western Meadowlarks (Sturnella neglecta) and

Bobolinks (Dolichonyx oryzivorus) are found in some agricultural areas bordering the

45 city (C. Artuso, pers. obs.). These species appear rnore resilient to hurnan activity than many other grassland specialists and can be coÍrmon in agricultural areas; however, they are now declining apparently as a result of changes in agicuhural practices (Brewer et al.

1991, [:nyon1994, Martin and Gavin 1995, Manitoba Avian Research Committee

2003). Edge specialists and species that use shrub vegetation are nìore likely to persist in the city, e.g. Song Sparrow (Melospiza melodia) is common along wooded riparian sections (C. Artuso, pers. obs.). Song Spanows can also be found in intensely farmed cropland as long as there are wet ditches with sufficient shrubbery (Manitoba Avian

Research Committee 2003).

If stopover migrants as well as breeding species are considered, some Winnipeg parks and suburban areas with mature trees curently have a considerable avifauna. One such example is Saint Vital Park, where from 2001 to 2008 I have observed 197 specbs in the park and on the Red River beside it. Of these 197 specbs, only 19 (I)Vo) are residents. I have recorded 35 specþs breeding in the park on at least orre occasion (nest located or adults seen feeding juveniles) and 18 others breeding in nearby suburbs, along the Bishop

Grardin Greenway or in King's Park. It remairs to be seen if such siûes can maintain theirbiodiversitywithanincreasinghumanpopulationardplans forconstructionin currently undevelop ed areas.

SrnucrunEoF TIrE THES rs

This thesis is structured in self-contained chapters. These chapters deal with the population density, breeding biology, habitat selection and diet of the Eastern Screech-

46 Owl across an urban to rual gradient in Winnipeg, Manitoba and surounding area, at the

Èr northofthis species' geographical range. Winnipeg is one of the northernnrcst location where this species breeds regularly, with the harshest climate that it experiences anywhere in its regular range (Fig. 1).

I J

¡ l¿

:'r-..ffif pIa'f,,- )'-'Y:-h f^7 ; * Fig. 1. Range of the Eastern Screech-Owl (adapted withpermission from Gehlbach 1995, Fig. 1). Light green shading represents areas where recent sightings have occurred but the species appears to be either rare or sporadic in occunence. The blue circle (also indicated by the blue arrow) represents the greater Winnipeg study area.

Chapter 2, Eastern Screech-Owl breeding and population density at the northern perþhery of its range, addresses aspects of the population density and the breeding biology of Eastern Screech-Owl in the Winnipeg study area. The population dersity aspect is based on a random sfratified survey conducted from 2004 -2007 with a gradient approachof humandensity, viz urban-sububan-rural-wildlands. Differences in

47 density at different points along this gradient are discrssed. This discussion pertains to

Johrson's (1980) second-order selection in that landscape level patterns are described.

The breeding biology is described from nests located during the sarre period using

taditional nest-searching techniques and subsequently monitored throughout the

breeding season The breeding section also examines difference between suburban and

rural nests. Very little is known of the way urbanization affects the reproductive ecology

of breeding species. Sorre studies (Osborne and Osborne 1980, Tomialojc 1980, Tella et

aL 1996) have found reduced predation in uban areas can increase breeding success of

sorre birds, whilst other studbs report reduced breeding success (Cowie and Hinsley

1987, Hõrak 1993) or sernitivity to fragrrent (park) size (Fernández-Juricic and Tellería

1999). This chapter therefore contributes to our understanding of how an urban gradient

can affect a nesopredator. Both the population dersity aspects and breeding biology are conpared with studies in more southerly locatiors.

Chapter 3, Habitat Seþctionby the Eastern Screech-Owl along a rural- urban gradient, addresses the question of how Eastern Screech-Owls select breeding habitat in the

Winnþeg study area. This discussion of the habitat context of breeding fi.nthers the understanding of the breeding biology in northern regiorn as discussed inChapter 1.

This chapter is based on habitat analysis that was conducted around all known nesting sites in the study area from 2004 -2007, compared with habitat analysis of potential nest sites that were not used by Eastern Screech-Owls in the sarre period. This chapter therefore pertains to Johrston s (1980) third-order selectior¡ i.e. which specific habitat

48 coÍponents are exploited; and fourth-order selection, i.e. whichspecific resources utilized, in this case which available cavities are used for breeding.

This chapter fi;rthers the theme of differences between suburban and rural birds by corrparing nest-site selection in relation to hurnan density. Finally, this chapter also attempts to address sonp of the mechanisms that may play a role in any of observed patterns of habitat selection Habitat selection in humanaltered landscapes is still poorly understood (Sodhi and Olþhanf.1993, Jokimäki and Suhonen 1998, Smithet aL 1999).

The ways in which habitat selection in range peripheral populations may differ from population ckiser to the center of the range is also poorly studied, except in theoretical terms of expected differences of population dynamics (Pulliam 2000, Swihart et aL

2003). Likewise, how the influences of urbanization on avian populations might be enhanced or mitigated in range-perþheral areas is poorly understood. Chapters 2 and3 therefore offer a significant contribution to our understanding of range-perþheral populations.

Chapter 4, The diet of the Eastern Screech-Owl at the northern perþhery of its range, addresses the question of the diet of Eastern Screech-Owl in the study area in terrrs of biologically significant period and in terms of differences in the diets of suburban and rural pairs. As with Chapters 3 and 4, this paper discusses the differences in diet in a range-perþheral context and makes conparisons with other studies in more southerly locations. The selectivity of avian prey is also discussed based on point counts in the study area. This chapter is based on the identification of prey by direct field

49 observations, video footage, dissection of pellets and the inspection of nest boxes in the

post-fledging period. This chapter pertains to Johßton s (1980) fourth-order selection in

that it examines the selection of prey resources. The conparison of diet and degree of

specialization in a northern context with diet proportions found in other studþs fuithers

the theme of the biological inportance of range-peripheral contexts arid contibutes to our

understanding of how this species is able to survive in Manitoba and the ìmportance of

anthropogenic influences on diet and hence suvival

Chapter 5, Eastern Screech-Owl hatches Wood Duck egp, is an account of an

observationof an interactionbetweenthe Wood Duck and the EasternScreech-Owl in

conpetition for a nesting site. This chapter disct¡sses one aspect of resornce conpetition

for suitable cavities for nesting sites. It therefore fuithers the theme of habitat selection

discussed in Chapter 3, in particular, providing eviderrce of the potentially limiting effect

of cavity availability discussed in that chapter. This paper was published in 2007 in The

W ilson Jo urnal o f Ornitho lo gy 1 1 9 : 1 l0 - I 12 (Artuso 2007 a).

Chapter 6, Eastern Screech-Owl in Manitoba: Eviderrce of historical range expansior¡

addresses the question of whether or not the Eastern S'creech-Owl was present in

Manitoba before European settlement. This paper is based on a database from multiple

sources of Eastern Screech-Owls recorded in Manitoba from I92O to 2007. A similar database of records from northern Minnesota and North Dakota is rsed for conparative purposes. This chapter draws heavily on the discussion of the previous chapters in evaluating the degree to which Eastern Screech-Owl is dependerit on human-ahered

50 habitat. In so doing, this chapter offers evidence for the importance of anthropogenic

habitat change in influencing population density and ultimately geographbal range of

range perþheral populations.

InChapter 7,the concluion,I discuss the overall inplications of findings of the

preceding chapters in both theoretical terms and their implicatiorx for conservation in

areas of human settlerrpnt and in human-altered landscapes. I also summarize and

review the mechanisms behind the patterns observed in Chapters 2,3, and 4 based on the

data collected in this study and in the literature.

With this research I have aimed to corrpare the population density, breeding srrccess, diet, and habitat selectionof EasternScreech-Owl in urban, suburbanand rural areas. I wanted to test whether human dersity affected these variables in the northernnnst part of

'Where the species' geographical range, as they do in the southernpart (Gehlbach 1994). possible, I endeavored to investigate the rrechanisms behind any potential patterrn by looking at predator dersities, prey availability, habitat differences, human influences such as pedestrian activity or proximity to buildings and roads, and other factors such as the urban heat island and how these related to breeding success and habiøt selection Before comrrerrcing this research I made two predbtions: 1) that Eastern Screech-Owls in the

Winnþeg area would demonstrate advantages at the upper end of the rural - urban gradient, i.e. in high-density suburban and urban areas, and 2) that these changes would be nnre pronounced than in Texas due to the harsher climate in Manitoba, at the limit of this species' cold tolerance (Mosher and Henny 1976), hence increasing the importance

51 of the urban heat island and antllopogenic sources of food and shelter. The following

chapters test those predictions.

LrrpnlrunE CmED

Able K. P., and J. R Belthoff. 1998. Rapid 'evolution' of migratory behaviour in the

intoduced house finchofeasternNorthAmerica. Proceedings of the Royalsociety

of l-ondor¡ Series B2652063-207L

Abraham K. F., R. L. Jefferies, and R. T. Alisauskas. 2005. The dynamics of landscape

change and snow geese in mid-continent North Amerba. Global Change Biology

11:841-855.

Adans, L.V/.2005. Urbanwildlifeecologyandconservation:Abrief historyof the

discþline. Urban Ecosystems 8 :139-l 56.

Adriaenser¡ F., and A. A. Dhondt. 1990. Population dynamics and partial migration of

the European robin (Erithacus rubecula) in different habitats. Journal of Animal

Ecology 59:1077-1090.

Aldrich, J.W., and J. S. Weske. 1978. Origin and evolutionof the easternHouse Finch

population Auk 95 :528-536.

Andren, H. 1992. Corvid Density and Nest Predation in Relation to Forest

Fragmentation: A I:nds cape Perspe ctive. Eco lo gy 7 3 :7 9 4-804.

Andnsiak, L. 4., and Cheng, K. M. 1997. Breeding biology of the BarnOwl (Tyto alba)

in the lower mainland of British Columbia. Pages 3846 ¿n J. R. Duncan, D. H.

Johnson, and T. H. Nicholls, editors. Biology and corservation of owls of the

52 Northern Hemisphere. North Central Research Statioq U.S. Departnent of

Agiculture, St Paul, Minnesota, U.S.A.

Ankney C. D. 1996. An embarrassment of riches: too rnany geese. Journal of Wildlife

Mana ge rrre nt 60 :2I7 123 .

Aherti, M., and J. M. Marzluff. 2004. Ecological resilience in urbanecosystems: linking

urban patterns to human and ecological functions. Urban Ecosysterrs 7 :24I--265.

Aldrich J. W., and R. W. Coffin. 1980. Breeding bird populations from forest to suburbia

after thirty-seven years. American Birds 34:3-7.

Andrerl H. 1992. Corvid Density and Nest Predation in Relation to Forest

Fragmentation: A I-ands cape Perspe ctive. Eco lo gy 7 3 :7 9 4-804.

Armstrong, E. R., and D. L. G. Noakes. 1983. Wintering bblogy of Mourning Doves,

Zenaida macroura, in Ontario. Canadian Field Naturalist9T :434438.

Artuso, C. 2007 a. Eastern Screech-owl hatches Wood Duck eggs. The Wilson Journal of

Omithology I 19 : 1 l0-I12.

Artuso, C.2007b. January initiationof suburbanGreat Horned Owl nests in Manitoba.

The Blue Jay 65',26-33.

Baker, P. J., A. J. Bentley, R. J. Ansell, and S. Hanis. 2005.Impact of predationby

domestic cats Felis catus'tnan urban area. Mammal Review 35:302112.

Balmford, 4., J. L. Moore, T. Brooks, N. Bugess, L. A. Hansen, P. Williams, and C.

Rahbek. 2OOl. Corservation confl icts across Afr ba. Science 291 :26161619.

Bancroft, J. 1989. Merlin nesting in southern Manitoba. Blue Jay 47:205-207 .

53 Bartos-Smith, S., J. E. McKay, and M. T. Murphy. 2006. Towhee or not to be: Are urban

Spotted Towhee populations self-srstaining? Fourth North American Ornithological

Confererrce, Veracruz, Mexbo, October 41, 2006.

Batten, L. A. 1972. Breeding bird species diversity in relation to increasing urbanization

Bird Study 19:157-166.

Battin, J.2004. When good animals love bad habitats: Ecological traps and the

conservatio n o f animal populatio rs. Co nservatio n B io lo gy 18 :1482-149 I.

Bautista, L.M., J. M. García, R. G. Calmaestra, C. Palacín, C. A. MartúI M. B. Morales,

R. Bonal, and J. Viñuela. 2005. Effect of weekend road traffic on the use of space by

raptors. Conservation B iology 19 :7 261 32.

Bayne, E. M., and K. A. Hobson 1997. Corrparing the effects of landscape

fragnnntation by forestry and agiculture on predation of artificial nests.

Cornervation B iolo gy II :1418-1429.

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

organization Condor 8 :75 -83.

Bennett, W. A. 1990. Scale of investigation and the detection of corrpetition: an example

from the home sparrow ard house finch introductions in North Arrprica. Arrprican

Nannalist 135:725147.

Berry, M. E., C. E. Bock, and S. L. Haire. 1998.Abundance of diurnal raptors on open

space grasslands in an urbanized landscape. Condor 100: 601-ó08.

Best, L. B. 1986. Conservation tillage: ecological taps for nesting birds? Wildlife

Society Bulletin L4 :308-317 .

54 Bevanger, K. 1998. Biologbal and corservation aspects of bird nnrtality caued by

e lectrb ity po wer lines : a review. B io lo gic a I Conservatio n 86 :67 --7 6.

B ierwager¡ B. G. 2005. Pred icting eco lo gical connectivity in urbanizing landscapes.

Environnent and Planning B: Planning and Design 32:763176.

Bierwager¡ B. G. 2007. Connectivity in urbanizing landscapes: the irrportance of habitat

configuratior¡ urbanareas size, and dispersal Urban Ecosystems 10:2942.

Blah, R. B. 1996. l¿nd use and avian species diversity along an urban gradient.

Eco lo gical App licatiors 6 :506-5 1 9.

Blair, R. B. 1999. Birds and butterflies along an urban gradient: surrogate taxa for

assess ing b iod ivers ity? Eco lo gba I App licatio rs 9 : I 64-1 70.

Blaft, R.8.2001. Creating a horrngeneous avifauna. Pages 440486 inJ.M. Marzluff,

R. Bowman, and R. Donnelly, editors. Avian ecology and conservation in an

urbanizing world. Kluwer Academic, Boston, Massachusetts, U.S.A.

Bloom, P. H., and M. D. McCrary.1996. The urban Buteo: Red-shouldered Hawks in

southernCalifornia. Pages 3143 ¿r? D. M. Bird, D. E. Varland, and J. J. Negro,

editors. Raptors in hurnan landscapes: adaptations to built and cultivated

environrrpnts. Academic Press, L,ondorl U.K.

Blouin, F. 2001. EasternManitoba loggerhead Shrke Survey, May 2001. Summary

report, Eastern Loggerhead Shrke Recovery Action Group.

Bolger, D. T. 2001. Urbanbirds: population, cornmunity, and ìandscape approaches.

Pages I55-I77 inJ.M. Marzluff, R. Bowman, andR. Donnelly, editors. Avian

ecology and cornervation in an ubanizing world. Kluwer, Boston, Massachusetts,

U.S.A.

55 Bolger, D. T., T. A. Scott, and J. T. Rotenbeny. 1997. Breedingbird abundance in an

urbanizing landscape in coastal soúhern California. Conservation Biology II:406-

42t.

Bosakowski, T., and D. G. Smith. 1997. Distribution and species richness of a forest

raptor community in relation to urbanization Journal of Raptor Research 3l:26-33.

Botelho, E. S., and P. C. Arrowood. 1996. Nesting success of western Burrowing Owls in

natural and hurnan-ahered environnents. Pages 61-68 inD.M. Bkd, D. E. Varland,

and J. J. Negro, editors. Raptors in human landscapes: adaptations to built and

cultivated environments. Academic Press, t ondon" U.K.

Boyen:s, I., editor. 2004. The encyclopedia of Manitoba. Great Plairs Publishing,

V/innþeg, Manitoba, Canada.

Brewer, R., G. A. McPeek, and R. J. Adams, Jr. 1991. The atlas of breeding birds of

Michigan MichiganState University Press, East Lansing, Michigaq U.S.A.

Brumm, H.2004. The impact of environmental noise on song anplitude in a territorial

bird. Journal of Animal Ecology 73:434440.

Bull, E. L. and J. A. Jackson 1995. Pileated'Woodpecker (Dryocopus pileatus). The

Birds of North Arrerica Online. A. Poole, editor. Ithaca: Cornell I¿b of Ornithology;

Retrieved January 2009 from the Birds of North Arrprica Online:

http ://b na.b irds.cornell.edu/bna/spec ies/148d oi:10.2I7 3 /brn.I48

Bunnell, F. L. 1999. What habitat is an island? Pages l-31 in J. A. Rochelle, L. A.

Læhman¡, and J. Wisniewski, editors. Forest Fragmentation: wildlife and

management imp lications. Brill, Bos ton, Mass achusetts, U.S.A.

56 Burke, D. M., and E. NoL 1998. Influence of food abundance, nest-site habitat, and forest

fra grrentatio n o n breed ing Ove nb irds. Auk 1 15 :96-L04.

Carn, 8., J. D. Nichols, J. R. Sauer, J. E. Hines, and C. H. Flather. 2000. Relative species

richness and community conpleteness: birds and urbanization in the Mid-Atlantic

States. Eco lo gbal Applicatiors 1 0 :1 1 96-12 1 0.

Carey, A. 8., J. A. Reid, and S. P. Horton 1990. Spotted owl hone range and habitat use

in soúhern Oregon coast ranges. Journal of Wildlife Managerrrnt 54:lI-17 .

Cartvnight, B. W. 1931. Notes and observations onsoûre Manitobanbirds. Canadian

Field- Naturalist 45 : 1 8 1-l 87.

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

and Urban Planning 74:4649.

Chalfouru A. D., F. R. Thompson, and M. J. Ratnaswamy. 2002. Nest predators and

fragnentation: a review and meta-analysis. Conservation Biology 16:306-318.

Chiang S. N., P. H. Bloom, and S. E. Thomas. 2006. Horrp range and habiøt use of

Cooper's Flawk's in r¡ban and natr¡al areas of California. FourthNorth Arnerban

Omitholo gical Conference, Veracruz, Mexico, October 4J, 2006.

Chowru S. L., B. J. VanRensburg, K. J. Gaston, A. S. L. Rodrigues, and A. S. Van

Jaarsveld. 2003. Energy, species richness, and humanpopulationsize: conservation

implications at a national scab. Ecologbal Applications 13:L233-124L.

Cicero, C. 1989. Avian community stuctrÌre in a large urban park: Controls of local

richness and diversity. l-andscape and UrbanPlanning 17:22L140.

57 Clergeaq P., J. Jokimäki, and J-P. L. Savard. 200I. Are urban bird communities

influenced by the bird diversity of adjacent landscapes? Journal of Applied

Ecology 38: lL22-1134.

Clergear¡ P., J. Jokimäki, and R. Snepp. 2006. Using hierarchical levels for urban

ecology. Trends in Ecology and EvolutionZl: 660-661.

Clergear¡ P., J.-P.L. Savard, G. Mennechea and G. Falardeau 1998. Bird abundance and

diversity along an urban-rural gradient: a corparative study between two citbs on

different continents. Condor 100 :413425.

Colemaru J. S., and S. A. Temph. 1993. Ruralresidents' free-ranging domestb cats: a

survey. Wildlife Society Bulletin 2l:381190.

Copley, D., D. Fraser, and J. C. Finlay. 1999. Purple Martins, Progne subís: a British

Columbian success story. Canadian Field-Nan¡ralist 113 :226-229.

Corlett, R. T. 2005. Interactions between birds, fruit bats and exotb plants in urban Hong

Kong, South China. Urban Ecosystems 827 5183.

Cowb, R. J., and S. A. Hinsley. 1987. Breeding success of Blue Tits and Great Tits in

suburban gardens. Ardea 75:81-90.

Crooks, K. R., and M. L. Soulé. 1999. Mesopredator release and avifaunal extirrctions in

a fragnrented system Nature 400:563-566.

Crooks, K. R., A. V. Suarez- arfl D. T. Bolger.2OO4. Avian Assemblages along a

gradient of urbanization in a highly fragmented landscape. Biological Conservation

lI5:451462.

Cullen, J. M., P. E. Guiton, G. A. Horridge, and J. Pierson 1952. Birds of Palma and

Gorrpra (Canary Islands). Ibis 94:68-84.

58 Danþlson, E. R., R. M. DeGraaf, and T. K. Fuller. 1997. Rural and suburban forest

edges: effect on egg predators and nest predation rates. l¿ndscape and Urban

Planning 38:2546.

D'Antonio, C., and L. A. Meyerson 2002. Exotb plant species as problems and solutiors

in ecological restoration: a synthesis. Restoration Ecology 10:708-713.

DeGraaf, R. M. 1991. Winter foraging guild structure and habitat associations in

sr¡b urban b ird co mmunities. l¿ndsc ape and Urban P lanning 2l :I7 3-180.

DeGraaf, R. M., and J. M. Wentworth" 1986. Avian guild structure and habitat

associations in suburban bird communities. Urban Ecology 9:399412.

DeGraaf, R. M., T. J. Maier, and T. K. Fuller. 1999. Predation of small eggs in artificial

nests: effects of nest positior¡ edge, and potential predator abundance in extensive

forest. Wilson Bulletin lIl :236-242.

Diamond, J. M. 1986. Rapid evolution of urbanbirds. Nature 324:IO7-108.

Diermen, I. 1996. Sparrowhawk (Accipiter nisus) breeding in village, rural landscape and

forest. De Iævende Matuur 97:43-5I.

Ditchkoff, S. S., S. T. Saalfeld, and C. J. Gibson 2006. Animal behavior in urban

ecosysterns: Modifications due to human-induced stess. Urban Ecosysterrs 9:5-12.

Donnelly, R., and J. M. Marzluff. 2006. Relative importance of habitat quantity,

sftucture, and spatial pattern to birds in urbanizing environnnnts. Urban Ecosystems

9:99-117.

Dugger, B. D., K. M. Dugger, and L. H. Fredrickson 1994. Hooded Merganser

(Lophodytes cucullatus), The Birds of North America Online. A. Poole, editor. Ithaca:

59 Cornell I¿b of Ornithology; Retrieved January 2009 from the Birds of North Anerica

Online: htç://bna.birds.cornell.edu/bna/species/098doi:I}.2173/bna.98

Dwernychr.rk, L., and D. Boag. 1972. Ducks nesting in association with gulls - an ecological trap? Canadian Journal of Zoology 50:559-563.

Dykstra, C. R., J. L. Hays, F. B. Daniel, and M. M. Simon 2000. Nest site selection and

productivity of subrirban Red-shouldered Hawks in southern Ohio. Condor 103 401-

408.

Dykstra, C. R., J. L. Hays, F. B. Daniel, and M. M. Simon 2OOl. Home range and habitat

use of suburban Red-shouldered Hawks in southwestern Ohio. Wilson Bulletin

113:308-316.

EmlerL J. T.lg74.An urbanbird community in Tucson, Arizona: derivation, stucture,

regulation Condor 7 6 :184-191 .

Estes, W. ,A'., and R W. Mannat 2003. Feeding behavior of Cooper's hawks at urban

and rural nests in southeastern Arizona. Condor 105:107-116.

Faetb S. H., P. S. Warren, E. Shochat, and W. Man¡ssich- 2005. Trophic dynamics in

urban communities. B ioScierrce 55 :399 407 .

Femândez-Juricic, E., ard J. L. Tellería,. 1999. Recruitment patterns of blackbirds

(Turdus merula) in tnban fragmented populations. Ardeola 46:6L10.

Fleischer, A. L, Jr., R. Bowman, ard G. E. Woolfenden 2003. Variation in foraging

behavior, diet, and time ofbreeding of Florida scnrb-jays in suburban and wildlife

hab itats. Condor I05 :515-527 .

Forrnarì, R. T. T, and L. E. Alexander. 1998. Roads and their rnajor ecological effects.

Annual Review of Ecology and Systematics 29'207-23L.

60 Frieser¡ P, F., J. Eagles, and R. J. Mackay. 1995. Effects of residential development on

forest dwelling Neotropbal migrant songbirds. Corservation Biology 9:L408-1414.

Fuller, R.4., P. H. Warren, P. R.A¡nsworth, A. Barbosa, R. G. Davies, J. Tratalos, and

K. J. Gaston 2006. Separating the effects of different components of urbanization on

avian assemblages. Fourth North American Ornithological Conference, Veracruz,

Mexico, October 4-7, 2006.

Fuller, R. 4., P. H. Waner¡ and K. J. Gaston 2007. Daytime noise predicts nocturnal

s inging in urban rob ins. B io lo gy Letters 3 :36817 0.

Galeotti, P. 1994. Patterns of territory size and defence level in rtnal and uban tawny owl

(Strix aluco) populations. Journal of Zoology 234:641458.

Garder¡ J., C. McAþine, A. Petersor¡ D. Jones, and H. Possingham.2006. Review of the

ecology of Australian urban fauna: A focus on spatially explicit processes. Amtralian

Ecology 3t:126-148.

Gaston, K. J., and K. L. Evans .2004. Birds and people in Europe. Proceedings of the

Royal Society of I-o ndoq Serb s B 27 I :1649-1655.

Gates, J., and L. Gysel. 1978. Aviannestdispersionand fledgingsrrccess in field-forest

ecotones. Ecolo gy 59 :87 1-883.

Gaúhier, G., J-F. Girowi; A. Reed, A. Béchet, and L. Bélanger. 2005. Interactiors

between land use, habitat rse, and population increase in greater snow geese: what

are the consequences for natural wetlands? Global Change Biology 11:856-868.

Gavareski, C. A. 1976. Relation of park size and vegetation to urban bird population:s in

Seattle, Washington Condor 7 8 :37 5482.

6T Gavett, A. P., and J. S. Wakely. 1986. Blood constituents and their relation to diet in

urban and nnal House Sparrows. Condor 88:279-284.

Gehhaclu F. R. 1988. Popuìation and Environmental Features that Prornote Adaptation

to Urban Ecosystems: the Case of Eastern Screech-Owß (Otus asio) n Texas. Pages

1809-1813 in H. Ouellet, editor. Acta XD( Congressu Internationalis Ornithologici.

National Museum of Natr-nal Sciences, University of Ottawa Press. Ottawa, Ontario,

Canada.

Gehhach F. R. 1994. The Eastern Screech Owl: Life history, ecology, and behavior in

the suburbs and counffyside. Texas A&M University Press, College Statiorl Texas,

U.S.A.

Gehhactu F. R. 1995. Eastern Screech-Owl(Otus asio). Account 165 in A. Poole and F.

Gill, editors. The Birds of North America. The Academy of Natr¡ral Sciences,

Philade þhi4 Penrsylvani4 and The Amerb an Omitho lo gists' Unb r! Washingto rt

D.C., U.S.A.

Germaine, S. S., S. S. Rosenstock, R. E. Schweirsburg, and W. S. Richardson 1998.

Relationships among breeding birds, habitat, and residential developnpnt in greater

Tucso n, Arizona. Eco lo gical App lications 8 :680{9 1.

Gering J. C., and R. B. Blair.1999. Predationonartificialbird nests alonganurban

gadient: predatory risk or relaxation in urban environments? Ecogaphy 22:532-54I.

Gerrard, J. M., and G. R. Bortolotti. 1988. The Bald Eagle: haunts and habis of a

wilderness monarch. Western Producer Prairie Books, Saskatoor¡ Saskatchewa4

Canada.

62 Getz L. L., L.B. Best, and M. Prather. 1977 . I-ead in urban and rural songbirds.

Enviro nnenta l Po lluti on 12 :23 5-23 8.

Government of Canada. Species at Risk Prùlic Registry. Reftbved October 2008 from:

http ://www. sarare gistr y. gc.ca/ sar/ inde>r/default_e.cfm

Graves, G. R. 2004. Avian cornmensals in Colonial America: when dtd Chaetura

pelagica become the chimney swift? Archives of Natural History 31:300-307.

Greer¡ D. 4., and M. G. Baker. 2003. Urbanization impacts on habitat and bird

communities ina SonoranDesert ecosystem l-andscape and UrbanPlanning 63:225-

239.

Gllúiérrez R. J., A. B. Franklin, and W. S. I^ahaye. 1995. Spotted Owl (Stríx

occidentalis). Account L79 in A. Poole and F. Gill, editors. The Birds of North

Arrprica. The Academy of Natural Sciences, Philadeþhia, and The American

Omithologists' Unior¡ Washingtoq D.C., U.S.A

Harrnr, T. E. 1988. Hone range size of the northern Bared Owl and the Northern

Spotted Owl in western Washington Thes is, Western Washington Univers ity,

Bellingharn, Washingtoû U.S.A.

Harrison, S., and E. Bruna. 1999. Habitat fragnrcntation and large-scale conservation:

what do we know for sure? Ecography 22:225132.

Haskell, D. G., A. M. Knupp, and M. C. Schneider. 2001. Nest predator abundance and

urbanization Pages 244158 in J. M. Marzluff, R. BowmarL and R. Donnelly,

editors. Avian ecology and conservation in an urbanizing world. Kluwer Academic,

Boston, Massachusetts, U.S.A.

63 Heiss, R. S., A. B. Clark, and K. J. McGowan 2006. American Crow nestling growth

and nuritional state vary with habitat. For¡th North American Ornithological

Conference, Veracru4 Mexbo, October 4-:7 ,2006.

Hinam, H. L., and C. Cassady St. Clair. 2008. High levels of habitat loss and

fragnrntation limit reprodtrctive success by redrcing home range size and

provisioning rates of Northern saw-whet owls. Biological Cornervation l4l:524-535.

Hobson, K. ,A.., and E. Bayne. 2000. Effects of forest fragrrrcntation by agricultue on

avian communities in the southern boreal mixed woods of western Canada. Wilson

Bulletin lt2:373-387.

Hogefe, T. C., R. H. Yahner, and N. H. Piergallini. 1998. Depredation of artificial

ground nests in a suburban versus a rural landscape. Journal of the Perusylvania

Academy of Sciences 72:3-6.

Horak, G. C. 1986. Cumulative effects of rapid urbanization on winter avian diversity in

northeastern Colorado. Thesis, Colorado State University. Fort Collins, Colorado,

U.S.A.

Hõrak, P. 1993. I-ow fledging success of urban Great Tits. Ornis Fennica 70 168-172.

Horn, D. J. 1985. Breeding birds of a central Ohio woodlot in response to succession and

r¡rbanization Ohio Journalof Scþnce 85:3440.

Howton, C. S. 1977a. Changing patterns of Corvidae on the prairies. Blue Jay 35:149-

156.

Horìstoq C. S. 1977b. The prairie bluebird trail. Nature Canada 6:3-9.

Houston, C. S.1979. The spread of Western Kingbird across the prairies. Blue Jay 37:53-

54.

64 Houston, C. S. 1986. Mouning Dove numbers explode on the Canadian prairies.

Arrerican B irds 40 :52-54.

Houston, C. S., and J. K. Schmut z. I999.Changes in bkd populatiors on Canadian

grasslands. Studies in Avian Biology 18:87-94.

Houston, C. S., D. G. Smith and C. Rohner. 1998. Great Horned Owl (Bubo virginianus).

Account 372 in A. Poole and F. Gill, editors. The Birds of North America. The

Academy of Natural Sciences, Philadeþhia, Pemsylvania and The Anerban

Omitho lo gists' Unio r¡ Washingto rL D. C., U. S. A.

Imhoff, M. L., C. J. Tucker, W. T. Lawrence, and D. C. Stutzer. 2000. The use of

multisource satellite and geospatial data to study the effect of urbanization onprimary

productivity in the United States. IEEE Transactions on Geoscierrce and Remote

Sensing 382549-56.

Jensen, K. H., S. Rauch, G. M. Monison" and P. Lindberg. 2002. Platinum goup

elenents in the feathers of raptors and their prey. Archives of Environmental

Corramination and To xicolo gy 42 :3381 47 .

Johrson, D. H. 1980. The conparison of wage ard availability measurements for

evaluating resource preference : Eco lo gy 6L :651 T.

Johrsorl R. J., and J. F. Glahn 1994. European starlings. Pages 109-120,n S. E.

Hygenstron¡ R. M. Timrn, and G. E. Larson, editors. Prevention and contol of

wildlife darnage. Cooperative ExtensionDivision, Institrre of Agricultue and Natural

Resources, University of Nebraska, Lincoln, Nebraska, U.S.A.

65 Johrston, R. F. 2001. Synantltopic birds of NorthAmerica. Pages 4947 inJ.M.

Marzluff, R. Bowmar¡ and R. Donnelly, editors. Avian ecology and conservation in

an rnbanizing world. Kluwer Academic, Bostoq Massachusetts, U.S.A.

Jokimäki, J.1999. Occurence of breedingbird species in wbanparks: effects ofpark

sffucture and broadscale variables. Urban Ecosystems 3:2L-34.

Jokirrxiki, J., and E. Huhtå. 2000. Artificial nest predation and abundance of birds along

an urban gradient. Condor 102:838-847.

Jokimäki, J., and J. Suhonen 1993. Effects of urbanizationonthe breeding bird specbs

richness in Finland: a biogeogaphical corrparison Ornis Fennica 70:7117.

Jokimåiki, J., and J. Suhonen 1998. Disribution and habitat selection of wintering birds

in urban environnents. Landscape and Urban Planning 39253163.

Jokimliki, J., J. Suhonen, K. Inki, and S. Jokinen 1996. Biogeographicalcomparisonof

winter bird assemblages in urban environenents inFinland. Journal of Biogeography

23:379-386.

Kark, S., A. Iwaniuk, A. Schalimtzek, and E. Bariker. 2007. Living in the city: can

anyo ne beco me an' urban ery lo iter' ? Jo tunal o f Bio geo grap hy 3 4 :63 8-65 1 .

Kaye, J. P., R. L. McCulley, and I. C. Burke . 2005. Carbon fluxes, ninogen cycling, and

soil microbial communities in adjacent uban, native and agricultural ecosystens.

Global Change Biology II:575-581.

Keyser, A. J., G. E. Hill, and E. C. Soehren 1998. Effects of forest fragment size, nest

density and proximity to edge on the risk of predation to ground-nesting passerine

birds. Conservation B iolo gy 12986-99I.

Klem, D. Jr. 1989. Bild-window collisions. WilsonBulletin 101 :606420.

66 Kluza, D. 4., C. R. Griffin, and R. M. DeGraaf. 2000. Housing developments in rural

New England: effects on forest birds. Animal Conservation3:15-26.

Knight, R. L. 1990. Ecologicalprinciples applicable to the managementof urban

ecosystems. Pages 2¿p-34 in E. A. Webb, and S. Q. Foster, editors. Perspectives in

urban ecology, Denver museumof n¿tural history and Thorne Ecological lrntitute,

Denver, Colorado, U.S. A.

Knight, R. L., D. E. Andersen, M. J. Bechard, and N. V. Marr. 1989. Geographic

variation in nest-defence behaviour of the red-tailed hawk Buteo jamaicensis.Ibis

13l:2216.

Knop{ F. L., and R. M. Evars. 2004. American White Pelican (Pelecanus

erythrorhynchos). The Birds of North America Online. A. Poole, editor. Ithaca:

Comell I¿b of Ornithology; Retrieved January 2009 from the Birds of North Arrerica

Online: http:i/bna.birds.cornell.edu/bna/species/057doi:10.2I731bm.57

Koes, R. F. 1985. Additions to the Manitoba bird list (1975 - 1984). Blue Jay 43:224- 23t.

Koonz, W. H. 1998. The Bald Eagle in Manitoba. Manitoba Departnrnt of Natural

Resources, Wildlife Branch Technical Report No. 88-1.

Korpimäki, E. 1978. Breeding biology of the Stffling Sturnus vulgarts in western

Finland. Ornis Fenníca 55:93-104.

Korpim¿iki, E. 1985. Clutch size and breeding success in relation to nest-box size in

Tengmalm' s owl Ae g o lius funereus. Ho larctic Eco lo gy 8 :1 75-1 80.

67 Kostelecka-Myrcha, 4., J. Zukowski and E. Oksiejczuk. 1997. Changes in the red blood

indices during nestling developnnnt of the Tree Sparrow Passer montanus inan

urban environment. Ibis 139 :92-96.

Kotler, B. P., C. R. Dickmarl and J. S. Brown 1998. The effect of water on patch use by

two Simpson desert granivores (Corvus coronoiães and Pseudomys

he rm ann s b ur g e n s is ). Austra lian Jo urna I o f Eco lo gy 23 :57 4-57 8.

Kristarì, W. B. 200I. Effects of Habitat Selectionon AvianPopulation Ecology in

Urbanizing l-andscapes. Dissertation, Universityof California, Riverside, California,

U.S.A.

I¿hti, D. C. 2001. The "edge effect on nest predation" hypothesis after twenty years.

B io lo gical Conservatio n 99 :3 6 5-37 4. f-ancaster, R. K., and W. E. Rees. 1979. Btd communities and the structwe of urban

hab itats. C anad ian Jo urn;al of Tnolo gy 57 :23 581368. l-anyor¡ W. E. 1994. WesternMeadowlark(Sturnella neglecta),The Birds of North

Anerica Online. A. Poole, editor. Ithaca: Cornell Lab of Ornithology; Refrieved

January 2009 from the Birds of NorthAmerica Online:

htç :/lbna.b irds.cornell.edu/bnalspec ies/1 04d o i:I}.2ll 3 /bm.l04

I¿rivière, S. 2004. Range expansion of raccoons in the Canadian prairies: review of

hypotheses. Wildlife Soc iety B ulletin 32 :955-963.

I-ehr, B. I., Jr., and T. B. Mowbray.2002. American Coot(Fulíca amerícanø), The Birds

of North America Online. A. Poole, editor. Ithaca: Correll t¿b of Ornithology;

Retrieved January 2009 from the Birds of North Arrprica Online:

http :/tbna.b irds.cornell.edu/bna/sp ecies 1697 adoi'L}.217 3 lbrn.697 a

68 Lichstein, J. W, T. R. Simons, and K. E. Franzreb.2002.landscape effects onbreeding

songbird abundance in rnanaged forests. Ecological Applications 12:83G857.

Lker, A., Z. Papp, V. Bokony, and A. Z.Intñ,vat 2008. Læanbirds in the city: body size

and condition of house sparows along the urbanization gradient. Journal of Animal

Ecology 77:789-:795.

Lindsay, A. R., S. S. Gillum, and M. W. Meyer. 2002.Influence of lakeshore

developrrent on breeding bird communities in a mixed northern forest. Biological

Conservation 107 : 1-1 1.

Luniak, M., R. Mulsow, and K. Walasz. 1990. Urbanization of the European Blackbird:

exparsion and adaptations of urban population Pages 187-198 ¿¡e M. Luniak, editor.

Urban ecological studies in Central and Eastern Europe; international syrrposium

'Warsaw, Poland. Polish Academy of Sciences, Warsaw, Poland.

Luther, D., J. Hilty, J. Weiss, C. Comwall, M. Wipf and G. Ballard. 2008. Assessing the

impact of local habitat variables and landscape context on rþarian bírds in

agricultural, urbanized, and native landscapes. Biodiversity Conservation 17:1923-

1935.

Major, R. E., G. Gowing, and C. E. Kendal. 1996. Nest predation in Austalian urban

environnents and the role of the pied currawong, Strepera graculina. Australian

Journal of Ecology 2I:399409.

Manitoba Avian Research Committee .2003. The Birds of Maníoba. Manitoba

Nahnalists Society, Manitoba Avian Research Committee, Winnipeg Manitoba,

Canada.

69 Manitoba ConservationWildlife and Ecosystem Protection Branch- Problem wildlife

page. Retrieved October 2008 from:

http://www.gov.mb.calconservatior/wildlife/problem_wildlife/waterfowl_cranes.html

Mannan, R. W., R. J. Steidl, and C. W. Boal. 2008. Identifying habitat sinks: a case study

of Cooper's hawks in an urban environnpnt. Urban Ecosysterns I I : I 4I -148.

Marti, C. D. 1969. Renesting by Barn and Great Horned Owls. V/ilson Bulletin 8I:467-

468.

Martin, S. G., and T. A. Gavin. 1995. Bobolnk (Dolichonyx oryzivorus), The Birds of

North Anerica Online. A. Poole, editor. Ithaca: Cornell Lab of Ornithology;

Retrieved January 2009 from the Birds of North Arrprica Online:

http ://bna.b irds.co rnell.edu/bna/sp ecies / 17 6doi:I}.2I7 3 /bm.I7 6

Marzluff, J. M. 2001. Worldwide urbanizatbn and irs effects on birds. Pages 19-38 in J.

M. Marzluff, R. BowmarL and R. Donnelly, editors. Avian ecology and conservation

in an ubanizing world. Kluwer Academic, Boston, Massachusetts, U.S.A.

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

colonization may determine biological diversity in human-dominated landscapes.

Urban Ecosystems 8 :157 -I77 .

Marzluff, J. M, K. J. McGowan, R. Donnelly, and R. L. Knight. 2001. Causes and

consequences of expanding Arrerican Crow populatiors. Pages 33I-363 in J. M.

Marzluff, R. Bowman, and R. Donnelly, editors. Avian ecology and conservation in

an urbanizing world. Kluwer Academic, Boston, Massachusetts, U.S.A.

Marzluff, J. M., J. C. Withey, K. A.Whittaker, M. D. Oleyar, T. M. Unfried, S. Rullmarl

and J. Delap. 2007 . Consequences of habitat utilization by nest predators and

70 breeding songbirds across multiple scales in an urbanizing landscape. Condor

109:516-534.

Mazur, K. M., and P. C. Janps. 2000. Baned Owl (Srrux varia). Account 508 in A. Poole

and F. Gill, editors. The Birds of North America. The Academy of Natual Sciences,

Philadeþhi4 Pennsylvania and The Amerban Omithologists' Unbq WashingtorL

D.C., U.S.A.

McClenner¡ N., R. R. Wigglesworth, and S. H. Anderson 2001. The effect of suburban

and agrbultrnal developnent on the activity patterns of coyotes (Canis latrans).

Anerican Midland Naturalist 146 :27 -36.

McCollin, D. 1998. Forest edges and habitat selection inbirds: A functional approach-

Eco graphy 2I :247 --260.

McCowan, M. 1978. Merlins wintering and breeding at Brandon, Manitoba. Blue Jay

36:39.

McGowar¡ K. J. 2001. Demographic and behavioral comparisons of subrnban and rural

Anerican Crows. Pages 365-381 in J.M. Marzluff, R. Bowmaû and R. Donnelly,

editors. Avian ecology and corservation in an urbanizing world. Kluwer Academic,

Boston, Massachusetts, U.S. A.

McKinney, M. L. 2006. Urbanization as a rnajor cause of biotic homogenization

B io lo gical C onservatio n L27 :247 -260.

McKinney, M. L. 2008. Effects of urbanization on species richness: A review of plants

and animals. Urban Ecosystens Il'161-176.

7t McKinney, M. L., and J. L. Lockwood. 2005. Pages 365-380 inD. F. Sax, J. J.

Stachowicz, and S. D. Gaines, editors. Species invasions: insights into ecology,

evolutior¡ and biogeogaphy. Sinauer Associates, Sunderland, Massachmetts, U.S.A.

McNicholl, M. K. 1988. Ecological and human influences on Canadian populations of

gass land b irds. Pa ges I - 25 ín P. D. Goriup, editor. Eco lo gy and conservation o f

Grassland Birds. International Council for Bird Preservation Technical Publication

No. 7, Cambridge, U.K.

Melarrpy, M. N., E. L. Kershner, and M. A. Jones. 1999. Nest predation in suburban and

rural wood lo ts o f no rthern O hio. American M id land Naturalis t I4l :282192.

Middleton, A. L. 1998. Chþping Sparrow (Spizella passerina), The Birds of North

Anerica Online. A. Poole, editor. Ithaca: Cornell Lab of Ornithology. Retrbved

January 2009 from the Birds of NorthAmerica Online:

http :/ lb m.birds. co rne ll.edu/b nals pecies /33 4do i:10.217 3 lbm.33 4

Millsap, B. A.2002. Survival of Florida burowing owls along an urban-development

gradient. Journal of Raptor Research 36:3-10.

Mittehach G.G., C. F. Steiner, S. M. Scheiner, K. L. Gross, H. L. Reynolds, R. B.

Waide, M. R. Willig, S. I. Dodsoq and L. Gough- 200L. What is the observed

relationship between species richness and productivity? Ecology 82:2381-2396.

Møller, A. P. 1988. Nest predation and nest site choice inpasserine birds in habitat

patches of different size: a study of mag¡lies and blackbirds. Okos 53:215-221.

Moriru P. J. 1999. Community ecology. Blackwell Science, Oxford, UK.

Mosher, J. A. and C. C. Henny. 1976. Thermal adaptiveness of plumage color in Screech

Owls. Auk 93:6I44L9.

72 Muller, M. J. and R. W. Storer. 1999. Pied-billed Grebe (Podilymbus podiceps),The

Birds of North Arrerica Online. A. Poole, editor. Ithaca: Cornell [¿b of Ornithology;

Retrieved January 2009 from the Birds of North Anerica Online:

http t lbm.birds.co rne ll.edu/b nalspec ies/4 1 0d o i:I}.2I7 3 /bm.4I0

Muñoa A. M., R. A. McCleery, R. R. L,opez, and N. J. Silvy. 2008. Nesting ecology of

rrnurning doves in an urban landscape. Urban Ecosysterrs 11.-251167.

Munyenyembe, F., J. Harris, and J. Hone. 1989. Determinants of bird populations in an

urban area. Austral Ecology 14:549-557.

Murphy-Klassen, H. M., T. J. Underwood, S. G. Sealy, and A. A. Czyrny. 2005.I-ong-

term rends in spring arrival dates of migrant birds at Delta Marsb Manitoba, in

relatio n to climate change. h:k 122 :I1 30-l 148.

National Audubon Society. Christmas Bird Count Historical results. Retrieved January

2009 fro m: http ://www. a ud ubo n o r gþ trd I cb c/hri inde x htrnl

Nero, R W. 1996. Comrnon Grackbs anting with'Weed and Feed" lawnchemicals.

Blue Jay 5431-34.

Newton, I.2003. The speciation and biogeogaphy of birds. Academic Press, I¡ndoru

U.K.

Nilon, C. H., C. N. Inng, and W. C. Zippercr. 1995. Effects of wildland development on

forest bird communities. l¿ndscape and Urban Planning 32:8L-92.

Norris, K., and N. Harper. 2004. Extinctionprocesses in hot spots of avianbiodiversity

and the targeting of pre-enptive conservation action Proceedinp of the Royal

Society of I-ondon, Series B 271:123-130.

73 Noske, R. 1998. Breeding biology, demography and success of the rufous-banded

honeyeater, Conopophila albogularls, in Darwin, a monsoonal topical city. Wildlife

Research 25:339-356.

Olden, J. D., N. L. Poff, M. R. Douglas, M. E. Douglas, and K. D. Fauch- 2004.

Ecological and evolutiorury consequences of biotic homogenization Trends in

Ecology and Evolution 19:18-24.

Oliphant, L. TV., and S. McTaggart¡. 1977.Prey utiìized by urbanMerlins. Canadian

Field- Naturalist 9 I :190-192.

Osborne, P., and L. Osborne. 1980. The conftibution of nest-site characteristics to

breeding success among blackbirds Turdus merula.Ibis 122:512-517 .

Palomino, D., and L. M. Carrascal. 2006. Urban influence onbirds at a regional scale: A

case study with the avifauna of northern Madrid province. l:ndscape and Urban

Planning 77:276190.

Parker, J. W. 1996. Urban ecology of the Mississþpi Kite. Pages 45-52 inD. M. Btd, D.

E. Varland, and J. J. Negro, editors. Raptors in human landscapes: adaptations to built

ard cultivated environments. Academic Press, I-ondorl U.K.

Parsors, H., R. E . Major, and K. French 2006. Species interactions and habitat

associations of birds inhabiting urbanareas of Sydney, Australia. Austral

Ecology 31.,217-227.

Parris, K. M., and D. L.Hazell.2OO5. Biotic effects of climate change in urban

environnents: the case of the grey-headed flying-fox(Pteropus poliocephalus) n

Me ho tnne, Australia. B io lo gbal Conservatio n I24 :267 17 6.

74 Partecke, J., arrd E. Gwinner.2007. lncreased sederrtariness in Europeanblackbirds

followin-e urbanization: A consequence of local adaptation? Ecology 88:882-890.

Partecke, J., T. Van't Hof ard E. Gwinner. 2005. Underlying physiological corÍrol of

reproduction in urban and forest-dwelling European blackbirds Turdus merula.

Journal of Avian Biology 36:295-305.

Pator¡ P. C. W. 1994. The effect of edge on avian nest success: how strong is the

evidence? Conservation B iolo gy 8 :I7 -26.

Pearce, C. M., M. B. Greerl and M. R. Baldwin.2007 . Developing habitat models for

waterbirds in urban wetlands: a log-linear approaclr- Urban Ecosystems 10:239154.

Pidgeon, A. M., V. C. Radeloff, C. H. Flather, C. A. tæpczyk, M. K. Clayton, T. J.

Hawbaker, and R. B. Hammer.2007. The effects of housing density and landscape

patterns onbird species richness across the United Kingdom. Ecological Applications

17:1989-?.010.

Pierotti R., and C. Annett. 200I. The ecology of Western Gulls in habitats varying in

degree of urban influerrce. Pages 307129 inJ.M. Marzluff, R. Bowman, and R.

Donnelly, editors. Avian ecology and corservation in an urbanizing world. Kluwer

Academic, Boston, Massachusetts, U.S. A.

Platt, 4., and A. Lill. 2006. Corrposition and conservation value ofbird assemblages of

urban 'habitat islards': Do pedestiantraffic and lardscape variables exert an

influence? Urban Ecosystens 9:83-97 .

Plumptor¡ D. L., and D. E. Andersen 1998. Anthropogenic effects on winter behavior of

fe rrugino us hawk s. Jo urnal o f W ild life Mana ge me t 62 :3 40 -3 46.

75 Pulliam, H. R. 2000. Onthe relationship between niche and disfribution Ecology lætters

3:34916I.

Raven S. J., and J. C. Coulson 2001. Effects of cleaning a tidal river of sewage on gull

numbers: a before-and-after study of the River Tyne, northeast England. Bird Study

48:48-58.

Reichard, S. H., L. Chalker-Scott, and S. Buchanan 200I.Interactions afiong non-native

plants and birds. Pages 179124 in J. M. Marzluff, R. Bowman and R. Donnelly,

editors. Avian Ecology and Corservation in an Urbanizing World, Kluwer Academic,

Bostorl Massachusetts, U.S. A.

Rerres, V. 2000. How can maladaptive habitat choice produce source-sirk population

dynamics? O kos 9l :57 9-582.

Reaurrrc, T. 2006. Common Ravens nesting in Winnipeg. Blue Jay 64:200--202.

Riley, S. P. D., R. M. Sauvajot, T. K. Fuller, E. C. York, D. A. Kamradt, C. Bromley, and

R. K. Wayne. 2003. Effects of urbanizationand habitat fragrnentationonbobcats and

coyotes in soúhern California. Cornervation Biology 17:566-576.

Rising, J. D., and N. J. Flood. 1998. Baltirnore Oriole (Icterus galbula), The Birds of

North Anerica Online. A. Poole, editor. Ithaca: Cornell l-ab of Ornithology;

Retrieved Janu,ary 2009 from the Birds of North Anerica Online:

htþ ://bna.b irds.cornell.edr-r/bna/spec ies/3 84d oi:10.2I7 3 lbm.384

Robbirs, C. T. 1993. Wildlife feeding and nutrition Academic Press, San Diego,

California, U.S.A.

Robinson, S. K., F. R. Thompson III, and T. M. Donovan 1995. Regional forest

fragnrntation and the nesting success of migratory birds. Science 267:1987-L990.

76 Rodewald, A. D. 2006. Towards a mechanistic understanding of urban-associated

changes in b ird communities. Fourth North American Ornithological Conference,

Veracruz, Mexico, October 4-7, 2006.

Rodewald, 4.D., and D. P. Shu,stack. 2008. Urban flight: understanding individual and

population-level responses of Nearctic-Neotropical migratory birds to urbanization

Journal of Animal Ecolo gy 7 7 :83-91.

Rodríguez-Estrella, R., and A.PelâezCareaga. 2003. The western screech-owl and

habitat alteration in Baja California: a gradient from urban and rural landscapes to

natural habitat. Canadian Jounal of Zoology 8I:916-922.

Rolando, 4., G. Maffei, C. Pulcher, and A. Giuso. 1997. Avian community structure

along an urbanization gradient. Italian Journal of Tnology 64341149.

Rollinson, D. J., and D. N. Jones. 2002. Variation in breeding parameters of the

Australian magpie Gymnorhina tibicen in suburban and rural environrnents. Urban

Ecosystems 6257--269

Rosenfield, R. N., J. Bielefeldt, J. L. Affeldt, and D. J. Beckmann 1996. Urban nesting

biology of Cooper's FIav¡ks in'Wisconsin Pages 4144 in D.M. Bird, D. E. Varland,

and J. J. Negro, editors. Raptors in human landscapes: adaptations to built and

cultivated environments. Academic Press, I¡rdor¡ U.K.

Rotteriborn, S. C. 1999. Predicting the impacts of ubanization onriparianbird

co mmunitie s. B io lo gical Co nservation 8 8 :289-299.

Russo, C., and T. P. Young. 1997. Egg and seed rernoval at urban and suburban forest

edges. Urban Ecosystems I :I7 I-17 8.

71 Saracco, J. E., and J. A. Collazo. 1999. Predation on artificial nests along three edge

types in a North Carolina bottomland hardwood forest. Wilson Bulletin 111:54I-

549.

Sasvári, L., and C. Moskät. 1988. Species richness, bird density and niche shifts in

cental- European marunade hab itats. Eko ló gia 7 :I2l - 146.

Savard, J-P. L., and J. B. Falls. 200I. Survey techniques and habitat restoration of

breeding birds in residential areas of Toronto, Canada. Pages 543-568 in J. M.

Marzluff, R. Bowman, and R. Donnelly, editors. Avian Ecology and Conservation in

an Urbanizing World..Kluwer Academic, Bostoq Massachusetts, U.S.A.

Schoech, S. J., and R. Bowman 2001. Variation in the timing of breeding between

suburban and wildland Florida Scrub Jays: do physiologic measures reflect different

environrrents? Pages 289-306 inJ.M. Marzluff, R. Bowman, and R. Donnelly,

editors. Avian ecology and conservation in an urbanizing world. Kluwer Academic,

Boston, Massachusetts, U.S. A.

Septon, G. 4., J. Bielefeldt, T. Ellestad, J. B. Marks, and R. N. Rosenfield. 1996.

Peregrine falcons: power plant nest structures and shoreline movements. Pages 145-

153 in D. M. Bird, D. E. Varland, J. J. Negro, editors. Raptors in Human l-andscapes:

Adaptatiorn to Built and Cultivated Environments. Academic Press, New York,

U.S.A.

Sergio F., Marchesi L., Pedrini P., Fener M. & Penteriani V. 2004. Electrocution alters

the distribution and derrsity of a top predator, the eagle owl Bubo buho. Journal of

Applied Ecology 4I :836-845.

t8 Shafer, C.L. 1997. Terresftial nature reserve design at the r-nban/rural interface. Pages

345-378 in M. W. Schwarta editor. Conservation in highly fragmented landscapes,

Chaprnan and Hall, New York, U.S.A.

Sherry, T. W., M. D. Johnson, and A. M. Stong. 2005. Does winter food limit

populations of migratory birds? Pages 414425 ¿n R. Greenberg and P. P. Marra,

editors. Birds of two worlds: the ecology and evolution of migration Johns Hopkins

University Press, Baltimore, Maryland, US A.

Shochat, E., S. B. lærmarL M. Katti, and D. B. tæwis. 2004. Linking optimal foraging

behavior to bird community sftucture in an urban desert landscape: Field experiments

with artific ial food patches. The American Natura list 164 :232142.

Shochat, E., Warrer! P., FaetlL S., Mclntyre, N., Hope, D. 2006. From patterns to

enrerging processes in mechanistic wban ecology. Trends in Ecology and Evolution

2I:186-19I.

Simberloff, D., and J. Cox. 1987. Corsequerrces and costs of conservation corridors.

Corservation B iolo gy I :631 t.

Slabbekoorr¡ H., and M. Peet. 2003. Birds sing at higher pitch in urban noise. Nature

424:267.

Slabbekoorn, H., and E. A. Rþmeester. 2008. Birdsong and anthropogenb noise:

implicatiors and applications for conservation Molecular Ecology 17:72-83.

Smitb C. M., and D. G. Wachob .2006. Trends associated with residential developrrpnt

in rþarianbreedingbird habitat along the Snake River in JacksonHole, WY, USA:

Irrplications for conservation planning. Biological Conservation 128:431446.

79 Smith, D. G., T. Bosakowski and A. Devine. 1999. Nest site selection by urban and rt¡ral

GreatHorned Owls in the northeast: Journal of Field OrnithologyT0:535-542.

Smith K. G., J. H. Withgott, P. G. Rodewald. 2000. Red-headed Woodpecker

(Melanerpes erythrocephalus). Account 518 ín A. Poole and F. Gill, editors. The

Birds of North Arrprica. The Academy of Natural Sciences, Philadelphia,

Pennsylvania and The Anrerican Ornithologists' Unior¡ Washingtoq D.C., U.S.A.

Smith-Casto, J. R. 2008. Inpacts of recreational trails onbreeding birds in forested

urbanparks. Thesis, Ohio State University, Columbus, Ohio, U.S.A.

Sodhi N. S. and L. W. Olþhant.1993. Prey selectionby urban-breedingMerlins. Auk

II0:727J35.

Sol, D., and L. læfebwe. 2000. Behavioural flexbility predicts invasion success inbirds

innoduced to New 7.e,aIand. Okos 90:599{05.

Sol, D., D. M. Santos, E. Feria, and J. C lavell. 1997 . Habitat Seþction by the Morik

Parakeet during colonization of a new area in Spain. Condor 99:3946.

Sol, D., S. Timmermans, and L. læfebvre. 2002. Behavioural flexfuility and invasion

succ ess in b irds. Animal Behavio ur 63 :49 5 -502.

Sorace, A.2002. High density of bird and pest species in urban habitats and the role of

predator abundance. Ornis Fennica 79:6011.

Soulé, M. E. 1991. l¿nd use planning and wildlife maintenance: guidelines for

conserving wildlife in urban landscapes. Journal of the Anerican Planning

Assoc iation 57 :3 13 123 .

80 Soulé, M. E., D. T Bolger, A. C. Alberts, J. ïVrights, M. Sorice, and S. Hill. 2005.

Reconsftucted Dynamics of Rapid Eninctiors of Chapanal-Requiring Birds in Urban

Hab itat I s lands. Co nservatio n B io lo gy 2 :7 5-92.

Sparling, D. W., J. Eisemanru and W. Kuenzel. 2007 . Nesting and foraging behavior of

red-winged blackbirds in stormwater wetlands. Urban Ecosystems 10:1-15.

Speechly, H. M. 1922. Birds that are little known in Manitoba. CanadianField Naturalist

36:177.

Spitzer, P. R., A. F. Poole, and M. Sceibel 1985. Initial population recovery of breeding

ospreys (Pandion haliaetus) in the region between New York City and Boston Pages

705-i.14 inY. D.Ilyichev, and V. M. Gawilov, editors. Proceedings of the 18th

International Ornitho lo gica I Co nges s, Mosco w, Rr-rss ia.

Stracey, C. M., and S. K. Robinsorr 2006. Do urban areas act as a refuge from nest

predation for an urban '\vinner", the Northern Mockingbtrd (Mimus polyglottos)?

Fourth North Anerican Ornithological Conference, Veracruz, Mexico, October 41,

2006.

Statford, J. 4., and W. D. Robinson 2005. Disftibution of Neotropical migratory bird

species across an urbanizing landscape. Urban Ecosystems 8:59-:77.

Sugde4 L. G., W. J. Thurlow, R. D. Hanis and K. Vernper. IgT{.Investigations of

Mallards overwintering at Calgar¡ Alberta. Canadian Field-Naturalist 88 303-31 1.

Summers-Smith J. D. 1988. The Sparrows. CaltorL Poyser, U.K.

Swihart, R. K., T. M. Gehring, M. B. Kolozsvary, and T. E.NW. 2003. Responses of

'resistant' vertebrates to habitat loss and fragmentation: the importance of niche

breadth and range boundaries. Diversity and Distributions 9:1-18.

8l Taylor. P., and R. F. Koes. 1995. Rare and marginal winter passerine birds in Manitoba.

Bltæ Jay 53:148-158.

Taylor. P., and R. F. Koes.2007. Rare and rnarginal winter non-passerine birds in

Manitoba. Blue Jay 65:179100.

Tella, J. L., F. Hiraldo, J. A. Donázar-Sancho, and J. J. Negro. 1996. Costs and benefits

of urbannesting inthe læsserKestrel Pages 53-60 inD.M. Bird, D. E. Varland, and

J. J. Negro, editors. Raptors in human landscapes: adaptations to built and cuhivated

environnpnts. Academic Press, London, U.K.

Terrple, S. A. 1988. Future goals and needs for the management and conservation of the

peregrine falcon Pages 843-848 in T. J. Cade, J. H. Enderson, C. G. Thelander, and

C. M. White, editors. Peregrine FalconPopulatiors: Their Managerrent and

Recovery. The Peregrine Fund Inc., Boise, Idaho, U.S.A.

Tewksbury, J. J., L. Garner, S. Garner, J. D. Lloyd, V. Saab, and T. E. Martin 2006.

Tests of landscape influence: nest predation and brood parasitism in fragmented

ecosystems. Ecolo gy 87 :7 591 68.

Tewksbuy, J. J., S. J. Hejl, and T. E. Martin 1998. Breedingproductivitydoes not

declirp with increasing fragmentation in a western landscape. Ecology 79:2890-

2903.

Thomas, J. A, N. A. D. Bourn, R. T.Clarke, K. E. Stewart, D.J. Sìmco¿ G. S. Pearmarl

R. Curtis, and B. Goodger. 2OOI. The quality and isolation of habitat patches both

determine where butterflies persist in fragmented landscapes. Proceedings of the

Royal Society of l-ondon, Serþs B 268:179L-1796.

82 Thonpson [Seton], E. E. 1890. The birds of Manitoba. Proceedings of the U.S. National

Museum 13:457443 [No. 841].

Thoringto& K. K., and R. Bowman 2003. Predation rate on artificial nests increases with

human housing density in suburban habitats. Ecography 26:188-196.

Tigas, L. 4., D. H. VanVuren, and R. M. Sauvajot.2002. Behavioralrespornes of

bobcats and coyotes to habitat fragnentation and corridors in an urban environrrent.

B io lo gical Conservation 108 :299-306.

Tomialojc, L. 1980. The impact of predation on urban and rural woodpigeon(Columbia

palumbus L.) populations. Polish Ecological Studies 5:I4I120.

Tomialojc, L., and F. R. Gehhach. 1988. Avianpopulation responses to man-made

environrrpnts. Acta )(IX Congressus Internationalis Ornithologici2:1177-1830.

Tryjanowski, P., and L. Krrczynski 1999. Shifting from outdoor to indoor breeding:

house martin's (Delíchon urbica) defense against houe sparrow (Passer domesticus).

Folia Zoolo gica 48 :10 1-106.

U.S. Fish and Wildlife Service. 1990. Federal register part tV 50 CFR Part 17 - RIN

1O18-4332. Endangered and Threatened tü/ildlife and Plants: Determinationof

Th¡eatened Status for the Northern Spotted Owl.

van Heezik, Y., A. Smytha, and R. Mathieub. 2008. Diversity of native and exotic birds

across an urban gradient in a New Z.ealand city. l-andscape and Urban Planning

87:223132.

VanCarrp, L. F., and C. J. Henny. 1975. The Screech Owl: Its life history and popuìation

ecology in northern Ohio. North American Fauna: 71. Unites States Department of

the Interior Fish and Wildlife Service.

83 Walcott, C.F. 1974. Changes in bfud life inCambridge, Massachmetts from 1860 to

1964. Auk 91:51-160.

W*g, Y., S. Chen, and P. Jiang. 2008. Black-billed Magries (Pica pica) adjust nest

characteristics to adapt to urbanization in Hangzhor¡ China. Canadian Journal of

Tnology 86:676484.

Wann, J., R. Sorenson, K. IVinkler, and T. Hoffman 1981. Successful Red-breasted

Nuthatch nesting in Ramsey County. I¡on 53:217-218.

Whaley, K. R. 1972. Crows overwinter. Bhæ Jay 30:133.

Wilcove, D. S. 1985. Nest predation in forest fracts and the decline of migratory

songb irds. Eco lo gy 66 :I2II-12I4.

Winter, L., and G. E. Wallace.2006.Impacts of feral and free-ranging cats on bird

species of corservation concern: A five-state review of New York, New Jersey,

Florida, California, and Hawaii American Bird Conservancy, The Plains, Virginia,

U.S.A.

Wobeser, G. A.1997. Diseases of Wild Waterfowl. Second edition PlenumPress, New

York, U.S.A.

Wolf, B. O., and C. M. del Rio. 2000. Use of saguaro fruit by white-winged doves:

isotopic evidence of a tight ecological association Oecologia 124:536-543.

Wong, T. C. M., N. S. Sodhi, and I. M. Twner. 1998. Artificial nest and seed predation

experimenfs in nopical lowland rainforest remnants of Singapore. Biological

Corservation 85 :97-104.

Woods, M., R. A. McDonald, and S. Harris. 2003. Predationof wildlife bydorrrstic cats

Felis catus in Great Britain Mammal Review 33:174-178.

84 Yauk, C. L., and J. S. Quinn 1996. Mutilocus DNA fingerprinting reveals high rate of

heritable genetic mutation in hening gulls nesting in an industialized urban site.

Proceedinp of the National Academy of Science, U.S.A 93:l2l3t-l2t4l.

Yeh, P. J.2004. Rapill evolutionof a sexually selected trait following population

establishment in a novel habitat. Evolution 58:166-174.

Yeh, P. J., and T. D. Price.2004. Adaptive phenotypic plasticity and the successful

coloniza tion of a novel environment. Anerican Naturalist 164:531-542.

Zalewski, A. 1994. A conparative study of breeding bird populations and associated

landscape character, Torun, Poland. landscape ard UrbanPlanning 293141.

Tngers, D. 4., S. May, and L. J. Goodrictl 2000. Identification of nest predators at

farm/forest edge and forest interior sites. Jornnal of Field Ornithology 7I:207116.

Tallner, P. 4., and S. L. Lima. 1997.1-andscape-levelperceptual abilities in white-footed

mice: perceptual range and the detection of forested habitat. Okos 80:51-60.

85 2. Breeding and population density of the Eastern Screech- Owl at the northern periphery of its range

Ansrnacr.-Suburban-adapted species may be advantaged or disadvantaged by a variety of npchanisms and have different population dernities at different points along the rural - urban gradient. I examined the density and nesting success of the Eastern Screech-Owl (Megascops asio) in Winnipeg Manitoba, Cannda, at the rnrthern periphery of this species' range, in response to 1) human density, 2) riparian versus non-rþarian habitat, and 3) the presence or absence of urban greenspace. Eastern Screech-Owl dersities peaked in upper suburban and urban areas (>20 plIn - people per hectare) and were lowest in wildlands (<1 pÆra). Brood size per nesting atterrpt was highest in suburban areas and fledging dates averaged 5 days in advance of rural areas. Great Horned Owl (Bubo virginianus) showed the opposite detection patterr¡ peaking in wildlands and rural areas (<10 p/ha) and being lowest in nedium - high-density suburban a.reas. Ninety-three percent of Eastern Screech-Owls detected and 887o of nests found were in riparian areas and 36Vo of detections atñ, 567o of nests were in urban greenspace. Rþarian nests fledged npre young and fledged well in advance of non-rþarian nests, whereas nests in greenspaces showed minimal differences to nests in residential areas. Eastern Screech-Owls are closely tied to riparian habitat in the study area. In suburban areas they have geater population dersity, a larger average brood size and an earlier average fledging date than nnal areas due at least partly to the urban heat is land and reduced predation pressure.

Keywords: Eastern Screech-Owl Megascops asío maxwelliae, playback, survey, gradienl human density, range periphery, population density.

IxrnonrrcuoN

Sorre disturbarrce-adapted bird species benefit from human activity and achieve high population densities in areas where anthropogenic habitat alteration is not extrerrrc (Blair

1996, Adans 2005, Donnelly and Marzluff2006, Palomino and Canascal2006), for example, suburbs that retain highly productive greenspaces or remain close to natural habitats (Imhoffet al. 2000, Marzluff 2005, Dorurelly and Marzluff 2006) and which may therefore sr-pport an increased carrying capacity for certain species (Emlen 1974, Shochat et aL 2006). Altered conditions under anthropogenic disturbance influerrce interspecific

86 dynamics such that large predators with large-area requirements may be excluded to the advantage of smaller predators (Crooks and Soulé 1999, JokimäkíL999, Chace and

Walsh 2006). Species that can exploit fluctuating urban resources more efficiently may be able to exclude others competitively (Copley etaI.1999, Shochat etaL.2004, Shochat et aL 2006). Disturbed habitats can also create ecological traps for exploiters. The risks include increased exposure to and transmission of diseases (Estes and Mannan 2003), mortality from elecftocution (Bevanger 1998, Marchesi et aL.2004) or collisions (Klem

1989, Millsap2002, Rollinsonand Jones 2002), and direct humandisturbance or persecution (Forman and Alexander 1998, Best 1986, Remes 2000, Battin 2004). In addition dispersal dynamics may be significantly ahered by urban conditiors (Marzluff et aL 2001b) and processes such as habitat selection rnay be affected (Wang et aL 2008).

Urbanized habitats are in fact the most likely place to find a disconnect between population dersity and reproductive success in birds (Bock and Jones 2004).

As a small-bodied resident species with small spatial requirerrents and a preference for relatively open woodland habitats, the Eastern Screech-Owl is well suited to exploit well- teed suburban habitats (Gehlbach 1995). This owl is also incornpicuous (cryptic plumage and noctt¡rnal behavior) and is a dbt generalist capable of exploiting a variety of seasonal resources (Gehhach I994b). The breeding strategy of Eastern Screech-Owl also makes it suited to suburban areas since it utilizes human-made nest-boxes and other human structures and lays relatively large clutches with early breeding short incubation, and an ability to renest following a disturbance (Gehhachl994b). Rapil breeding and a relatively fast gene turnover allow for population recovery following disturbance and

87 exploitation of newly available habitats under anthropogenic habitat change (Gehlbach

1988, 1995). Gehhach (1994b) demonsúated significant differences between a suburban

and a rural population of Eastern Screech-Owls in Texas in terms of fecundity and

survivorshþ. However, Ellison (1980) found that Eastern Screech-Owl in Massachuetts

avoided urban areas. This suggests that there may be a positive correlation between the

dersity of Eastern Screech-Owl and human-altered landscape, but that this relationshþ

may be non-linear or rnay be influenced by regional variables. I therefore examined the

patterns of population dersity and breeding biology at the northern perþhery of the

species' range to determine whether these patterns differ from more southerly

populations. I investigated correlations between Eastern Screech-Owl dernity and breeding and human density, riparian versus non-rþarian habitat, and the presence or

absence of tnban greenspace.

MurHons

Random stratified surveys for Eastern Screech-Owls were condrcted each spring from

2004 -2007. The study area was defined as a circle with a radius of 40km from the center of Winnþe 9(49'53 'N 97o09'E), which encompasses the cþ limits ard also areas orfside the city with lower human densities. The study area was randomly stratified by three variables (Table 2.1):

Table 2.1-. Snatification cate for the Eastern Screech-Owl survey Greers (<1 i. Rþarian t nspace / park area Rural (1-10 p/ha) ii Non-riparian ii Residential area Suburban low-densþ (>10-20 p/ha) (S1) Suburban moderate densþ (>20-30 p/tra) (S2) Suburban high-densiry & urban (>30 p/tra) (S3)

88 The human density categories were adapted from Marzluff et al. (2001a). Human dersity was calculated from 2001 census data for Winnipeg subdivisiors and surrounding rtnal districts (Statistics Canada 2004). For the purposes of classification, rþarian areas were defined as within 500m of a river or permanently flowing watercourse, based on observatiors in the study area which suggest this to be the average radius from the nest of a breeding territory. In the study area, riparian areas are dominated by American elm

(Ulmus americana), green ash(Fraxinus pennsylvanica), Manitoba maple (Acer negundo), and bur oak (Quercus macrocarpa), as well as eastern cottonwood (Populus deltoides), basswood (Tilia americana) and to a lesser extent peachleaf willow (Salix amygdaloides). Many planted native and exotic species also occur throughout the study area, in particular conifers such as white spruce (P icea glauca), eastern white cedar

(Thuja occidentalis) and the non-native Colorado blue spruce (Picea pungens).

Greenspace was defined as parks, ceneteries, golf cowses, and green corridors such as rþarian buffer sfrþs and forested areas without human residents but used by people to varying degrees.

Since wildlands are by definition greenspace, the wildlands category was not partitioned by the variable of greenspace (greenspace versus residential). Hence there were 18 categories. Six areas conesponding to each category were selected at random, except for the wildlands riparian and wildlands non-rþarian categories for which 12 areas were selected to ensure equal coverage of all levels of humandensity, resulting :-r:'I20 survey ftansects. Each transect was surveyed twice in a random order in two consecutive springs, half in 2004 arñ 2005 and the remaining areas lrl,2006 and2001. Surveys were

89 conducted in the peak local calling period of late February to April (personal observation).

Random statified surveys used tape-recorded song, arguably the most effective means of locating owls (Johnson et al. 1981, Lynch and Smith 1984). The timing of the survey, the rse of playback, and repetition of each ftansect was designed to reduce the likelihood of false negative bias, i.e. that a positive detection indicates presence, whereas a. zÊro detection does not necessarily indicate absence (Tyre et al. 2003). Each survey consisted of eight listening stations separated by 250rr¡ following a random direction from the randomly selected start point. Care was taken to enstre that each trarsect remained within the human population and habitat categories being surveyed (where necessary, a second and, rarely, a third randomdirection were selected at critical junctues). At each statiorL 2 minutes of listening were followed by playing a recording of forn Eastern

Screech-Owl nronotonic trill calls, each separated by 17 seconds (average pause between trill calls calculated from local males) and then 5 minutes of listening. The number of screech-owls detected was then recorded as being seen and/or heard either before and./or after playback. Srnveys were only conducted onrelatively calm nights (wind <20km/h) from30 minutes after sunset to one hou before sunrise.

The distance to the nearest permanent watercourse and the percent greenspace within a radius of 100m were calculated for each of the survey points using a combination of estimation at the site and calculations in GIS Arcview. These were then averaged for each

üarsect, providing numerbal variables enabling a comparison of averages for transects

90 where screech-owls were or were not detected. A general linear model (GI-IvI) was used in the Statistical Analysis System (SAS Version 9.1) software to analyzn conelations between the density of Eastern Screech-Owl detected and human density, rþarian habitat, urban green space and the presence of Great Horned Owls (Bubo virginianøs), which were also recorded on survey üansects despite the fact that their calls were not broadcast.

A single interaction term of hurnan dersityxriparian was also included to test for mitigating effects. The GLM was frst tested with a year variable (2004/2005 versus

2006/2007); however, because this showed no significant effects, nor did it affect the significance of other variables, it was rernoved from the final analysis. On most tansects, zeÍo, one or two screech-owls were detected (averaging either 0, 0.5, 1, or 1.5 over the 2 years) and only one transect had three and one trarisect fow owls (both avera gngto 2.5 over the two years). Therefore, after examination of the n scores plot and quantile- quantile plot, I used the square root ftansformation of the average number of screech- owls detected per transect as the response variable, providing a ne¿ìr normal distributioru enabling the GLM analysis, and increasing the .d value of the nndei

Following spring surveys, roost and cavity searches were conducted around all screech- owl detection points. Repeat visits to these areas were made to listen for spontaneous calling for a more accurate picture of territorial boundaries. Daytime roost and cavity searches further narrowed the searct¡ especially as nests were typically within 100m of a male regular roost site after late March Nests were then located by observing activity around poterìtial cavities in particular at dusk and dawn, when fennles typically exit the cavity briefly (Gehlbach 1995). Nests were then rrnnitoring every 4 -7 days (every

91 second day close to expected fledging) to determine breeding phenology and brood size;

however, nest cavities were only inspected where possible after all young had fledged to

minimize distrnbance. Failed nests were those where unhatched or damaged eggs were

found and no chicks fledged.

Data from two Environrrent Canada weather stations (Environment Canada, nd) were

used to assess the effects of tenperature dates (49' on fledging - Wiruripeg The Forks 53.400' N, 97o 7.800'W), and Winnipeg Richardson International Aþort (49' 55.200'N,

97' 13.800'W). The average March tenperature was med as this would be the rnost

significant temperatue period to influerrce egg laying, whbh according to back

calculatiors from fledging dates, begins in early April in the study area. Data from four

Environnent Canada weather statiors (Environment Canada, nd) were used to assess the

effects of snow cover (last day that snow on the ground was fiEasurable) on fledging dates (49' - Winnipeg WTC Stapon 54.000' N, 97o 3.600' W), Winnipeg Cangene (49o 48.000'N, 97o 9.000'W), WinnpegFederal l-ab (49" 54.600'N, 97o 9.600'W), and

Winnþeg South (49'46.800'N, 97o 7.800'\Ð. Inall cases I used data fromthe weather stationthat nrost chsely matched each nest's distarrce to the city cenûer.

Rnsur.rs

Survey Data

In the 4 years of this study, a totalof 55 EasternScreech-Owls were detectedon}g

(24Vo) transects. Detection rates in eachof the four years ranged from0.12 -0.3 screech- owls per transect. In some transects, detections were made on only one of the two years

92 surveyed. Screech-owls were detected in 12 of the 18 survey categories. The most productive categories were suburban low-density (S1) rþarianresidential, suburban low- dernity (S1) riparian greenspace, subtnban moderate-density (S2) riparian greenspace and suburban high-density and urban (S3) riparian greenspace, each with detection on four of the six fransects. Eastern Screech-Owl dersities peaked in upper suburban areas (>20

plha - people per hectare) where 36 (667o) detections occurred and were lowest in wildlands (<1 p/ha) where no detections occured (Fig. 2.1). Nonetheless, there were few screech-owl detections in the downtown core (Fig.2.2).

Fifty-one (93Vo) screech-owls were detected in riparian areas. Screech-owls were detected on all but one of the riparian categories (wildlands riparian) and on 42Vo of rþarian transects brr in only four of the nine non-rþarian categorþs on 7 Vo of ftansects.

The strong riparian preference is also evident when sightings are mapped (Fig2.2.). The difference between the greenspace and residential categorbs was much less pronounced, with detections on six of the 10 geenspace categories and five of the eight residential categories (l8%o and 257o of ftarnects suveyed in these categorbs respectively). Only 20

(36Vo) screech-owls were detected in urban greenspace (Table2.2).

93 Table 2.2. Eastern Screech-Owl (EASO) detectiors on random sftatified sr.lrve Category # EASO # transects with detections 7o detectiorn WildtanÈ(24) 0 0 0 Rural (24) 9 5 20.8 Suburban 1 (24) 10 8 33.3 Suburban2 (24) 18 8 JJ.J Suburban3 - Urban (24) 18 8 JJ.J

Riprian (60) 51 25 4t.7 Non-riparian (60) 4 4 6.7

Greenspace (72) 20 13 18.1 Residential (48) 35 t2 25 Ail (120) The number of ftansects are indicated beside each category in brackets.

ln addition to Eastern Screech-Owl Great Horned Owl (48), Northern Saw-whet Owl

(14), Long-eared Owl(3), and Bared Owl (1) were also detected. Great Horned Owl

showed the reverse detection pattern to Eastern Screech-Owl peaking in wildlands and

rural areas (<10 p/ha) where 37 (777o) detectiors occurred and being lowest in upper

suburbanand urbanareas where two (4Vo) of detections occured Fig. 1).Thirty-four

(ll%o) Great Horned Owls were detected in rþarian areas. Only eight of the 48 (l7%o)

Great Horned Owls were not detected in urban greenspace (Table 2.3).

Table 2.3. Great ftr""¿ O*t (CHO\Ð ¿"tr.

\{ildta¡rds (24) 18 8 33.3 Rurat (24) 19 1t 45.8 Suburban I (24) 9 6 25 Suburban2 (24) I 1 4.2 Suburban3 - Urban (24) I 1 4.2

Riparian (60) 34 18 30 Non-riparian (60) t4 9 15

Greenspace (72) q 6 29.2 Residentiat (48) 8 2I t2.5 aI (120)

94 0.55

0.5

845

(J E4 t'o E år{ 0.35 èû, 0.3 IJl o 8.2ã F(t o õü 8.2 > d 0.1 5

0.1

0.05

0 Wildlands Rural Suburban 1 Suburban 2 Suburban 3 <1 piha 1 - 10 p/ha >10 - 20 p/ha >20- 30 p/ha >30 p/ha

Fig. 2.1. Average detections per transect of Eastern Screech-Owl and Great Horned Owl by human density category. Error bars indicate the standard error of the mean plha,= number of humanresidents per hectare.

/ P .t-l sllrn '-.1 * '1.:ì Fig.2.2. l¡cations of Eastern Screech-Owl nests (N), tenitories (T), and sites where birds were detected fewer than 3 times and did not form a stable territory () from 2004

95 -2007. Great Horned Owl nests from2004 - 2008 are marked (Ð. The Perimeter Highway around WirLnipeg and major rivers and large permanent creeks are shown The city center is marked with an asterix. Scale 1:185000.

There were significantly more Eastern Screech-Owls in rþarian areas than in non-

rþarian areas and the number of screech-owls increased significantly with increasing

human density (Table 2.4). Therc was no significant difference in the number of screech-

owls detected in greerspace versus residential areas.

Table 2.4. Averages of nunerical values and ouþut of the GLM model for Eastern S creech- O wl s urvey with correspo nd ing cate gorical variab les

prcsent absent valæ Disrance water (m) Rip 311.26 + 110.75 22M.38 + 412.07 30.66 <0.0001* Human Density (p/ha)-- HDcat 25'8 + 3'58 15.42+ 1.82 3'84 0.01* Greenspace- GN 58x.6.07o 60+5.0Va 3'01 0'09 GHOW-Av (per transect) 0.12 + 0'05 0.23 + 0'05 2'25 O.l4 RipËHDcat 2.2 0'07 GLM rnodel: square root average detection Eastern Screech-Owl = Riparian versus non- rþarian (Rp) + Human density category (HDcat) + greenspace versus residential (GN) + average detections of GHOW (GHOW-Av) + RþxHDcat (GLM: F = 5.46, df = 11, ll9, p < 0.01, tr = 0.36). * indicates significance, p<0.05. Greenspace is a categorical variable; however the figures provided here are a percentage of greenspace calculated and each srnvey point in a radius of 50m around the point and averaged over each transect.

Four orthogonal conhasts were used to examine significant differences in the number of

screech-owls detected in the five levels of the human density variable. The wildlands

category had significantly fewer screech-owls than all other categories (F = 17.84, p <

0.01) and the rural and S1 categories had significantly fewer screech-owls than 52 and 53

(F = 4.25, p =0.04). The rural category did not differ significantly from the Sl category

(F = 0.93, p = 0.34) and the 52 category did not differ significantly from the 53 category (F=0.0,p=0.98).

96 Breeding

ln the 4 years of this study I located a total of 46 successful Eastern Screech-Owl nests, 6

failed nests, and37 stable territories (where a pair was active brl did not breed or where

one or more caviies was advertised by an unpaired male over a period greater than 6

weeks, Galeotti 1990) (Table 2.5). Only sites where unhatched eggs or remains were

found were treated as failed nests. It is possible that some seemingly territorial pairs

attempted nesting at an undetected site or that unhatched eggs were rernoved before I

inspected potenfial cavities. Therefore, the brood size per nesting attempt may be lower

than reported here. It is also possible that some of the later nests were renesting attempts,

although I found no evidence of renesting and later brood sizes differed little from earlier broods: nests where first fledging occurred before June22 averaged 4.3 fledglings per nesting attempt versus 4.I n broods where flrst fledging occurred after June 22. AnrlnI

variation (Table 2.5) appeared at least partly influenced by nnistue levels since brood

sizes in successful nests were positively correlated with total precipitation from March -

May (linearregressior¡ t =3.97,df =39,p < 0.01, É =0.29).

97 Table 2.5. Eastern Screech-Owl breeding data 2004 - 2007. Bmod size - Bmod size - per Ftedging successful only nesting attempt Successful: 12 3-6 3-6 28 May- 28 Jun Fatled 2 4.88 + 0.39 3.9 x.0.72 12 June + 3.49 days Territories:5 Successful: 10 3-6 2-6 3 June - 30 June Failed:2 4.1 + 0.18 3.4 x.0.48 15 June + 2.44 days Territories: 13 Successful: 14 2-5 2-5 2June-3 July Failed:0 3.57 x.0.25 3.57 x.0.25 17 June + 2.98 days Territories:9 Successful: 10 3-6 r-6 2June-3 July Failed:2 4.6 + 0.37 3.83 + 0.6 16 June + 3.37 days Territories:5 AII Successful:46 2-6 2-6 28 May- 3 July years Failed:6 4.19 + O.16 3.67 + 0.25 15 Jun + 1.38 days Territories:37 Range is given on the first line and average + standard error of the mean on the second line. Successful nests are those where at least one chick fledged. The brood size of failed nests istakenas zero forcalculationpr-uposes. The numberof territorbs includes unpaired males advertising cavities and pairs that were not known to have egp. These were not included when calculating averages.

I investigated the influence on breeding phenology of the urban heat island by regressing the fledging date of the oldest chbk in each nest against distarrce to city center. Fledging date was later with increasing distance from the city center (linear regession: f = 3.38, df

- 37 , p = 0.047 , r* = 0.1). I also wished to investigate the effect of terrperature and

snowfall and chose average March tenperature as an independent variable because

March temperatwes should be the rnost biologically significant, immediately prior to egg laying (limited localized weather data prevented the calculation of a shorter period such as two weeks prior to nesting). The last calendar day with snow on the ground was used as the nteasure of snow cover. For these analyses, I only used nests where the flrst chick fledged before June 21 becarse later nests would be much less influenced by March

98 teffperature. Since colinearity prevented combining tenperatue and snow dates with

distance from the city center in one rnodel, I substituted each variable in a new rnodel.

Fledging dates advanced with higher average March tenperatures, although this was

marginally insignificant (t = -1.79, df = 30, p = 0.08, P = 0.1). Fledging date advanced

significantly with earlier absence of snow cover (t =3.09, df = 30, p =0.04, d = 0.13).

Excluding repeat cavity uses, a total of 61 nest sites and tenitory centers were located, of

which 38 (62Vo) were in natural cavities, 2l (34Vo) in nesting boxes for Wood Duck (Arx

sponsa), and two (37o) at sites where I could not determine which cavity had been

selected. Human density around used territories ranged from0.24- 98.68 p/ha (X + SE:

26.01 + 2.01 plha), with 15 nests (25Vo) found in r,pper suburban and urban areas (S3), 26

sfres (43Vo) in rredium-density suburban areas (S2), 12 (20Vo) in lower suburban areas

(S1) and 8 nests (l3Vo) in wildlands and rural areas. Percentage greenspace in a radius of

100m around a used cavity ranged from0.6 -75Vo (16.9 t 2.4Vo), with29 nesrs (487o) centered in a greenspace and 32 nests (52Vo) centered in residential areas. Distance from med cavity to nearest permanent watercou$e ranged from 4.6 - 3477 m (255.3 + 69.1m), although only eight sites (13 7o) (of which only three nests were successful) were >500m fro m a permanent waterco urse.

The average brood size per nesting attenpt shows an increnpntal increase through the human dersity categories, with highest success in high-density suburban areas (S3) (4.25

+ 0.49) and the lowest strccess in the combined rwal and wildlands categories (3.44 +

0.58) (Table 2.6), although this was not statistically significant. Average fledging dates

99 did not significantly differ anþng categories. Average broods per nesting atteÍpt were

smaller and average fledging 12 days later in non-rþarian areas than in rþarian areas;

however, the fact that there were only six nesting attempts in non-rþarian areas precludes

meaningful corrparison

Table 2.6. Eastern Screech-Owl breeding by survey category. Av brood Av brood per n Successful onlv nesting attempt Av fledeins Rural and Wildlands 9 3.88 + 0.44 3.44+0.58 20-Jun+ 1.86 Suburban L T2 3.89 + 0.31 3.5 + 0.48 15-Jun+ 1.7 Suburban 2 23 4.47 + 0.21 3.62 + 0.43 14-Jun+ 3.08 Suburban 3 8 4.25 + 0.49 4.25 + 0.49 17-Jun+ 1.8 Riparian 46 4.L6 + 0.I7 3.76 + 0.25 15-Jun + 1.42 Non-ripa.rian 6 4.5 + 2.0 3.0 + 2.45 27-Jw+ 1.5 Greenspace 29 4.I2 + 0.22 3.8I + 0.29 l4-Jun + 1.69 Residential 23 4.29 + 0.25 3.48 + 0.43 l8-Jun + 2.41

Ail 52 4.19 + 0.16 3.67 + 0.25 l5-June + 1.38 Only one nest was found in the wildlands category so it was combined with the nnal category. The sample sizes given are for the data set overall; however, in a few cases either the exact fledging date or the exact number of fledglings was not known and these are excluded from the relevant calculations.

Predation at the nesting stage was minimal in the study area, with darrnged eggs being found ononly one occasion probably by Raccoon(Procyon lotor). One clutchof five eggs was renpved by a fernale Wood Duck (Artuso 2007a). No evidence of chick nnrtality was found when cavities were irspected post fledging. No nest site was used successfully in all 4 years of the study and only one cavity produced young in 3 out of 4 years. Six sites were rsed successfully in two years, 31 nest sites produced successful broods only orrce, and23 nest sites never produced young, of which 15 (65Vo) were held by unpaired males. In at least three cases, territorbs where breeding did not occur between 2004 and200l were sites ofprevious breeding according to local residents.

100 Average brood sizes in the Winnipeg study area support the trend of increasing clutch

size with increasing latitude observed by Munay (1976) (Table 2.7) and also noted in

other owl species (Overskaug and Bolst¿d 1998, Sunde et aL 2001). There remain two

caveats, however, with any conparisor¡ viz. brood sizes could differ from clutch sizes

and the fact that rnost long-term studþs have been based exclusively on nest boxes,

which are easily inspected and rnonitored (VanCamp and Henny 1975, GehlbachI994b).

In this study, larger broods of successful nests occurred in nest-boxes (4.47 + 0.26) than

in cavitþs (4.0 + 0.2I), although the difference is not significant (2-tailed t-test, t = I.29,

df = 39, p =0.21). The difference is 3.8 t 0.43 versus 3.57 + 0.3 per nesting attenpt.

Production in Texas did not differ between natural cavities and nest-boxes (Gehhach

I994a); however, the number and size of cavities rnay differ between the two regiorn.

Table 2.7. Eastern Screech-Owl breeding data from six different studies. n Clutch Bmod size Average brcod Fledging Dates size successfr¡l per nesting nests only attempt

(estimate) 3.44rwal* Ontario 58 1-6(3.83) 1-6 rila -May 19 - Jnly 5 Michigan 13 3-5(4) 3-5 2.6 13 May - 19 June (l June) Ohio 4Ð 2-6(4.43) 2-7 (3.8) 2.55 -1 May - 20 June (-31 May) Tennessee 25 3-6(4.1) 2-5(3.6) 2.5 N/A Texæ 97 2-6(3.8) 2-6(3.29) 1.9subu¡ban 1lFeb-15June(18-21 May) 1.0 rural Ranges given flrst then averages in brackets where available. Data from Ontario from Peck and Jarres (1983), fromMichigan from Craighead and Craighead (1956), fromOhio from VanCanp and Henny (1915), from Tennessee from Duly 1979, from Texas from Gehhach (I994b). * To aiC comparison here, the average brood per nesting attempt in Manitoba is given for suburban nests (here refers to all nests where human density >10 pitra) and rural (here refers to all nests where human density <10 pfta). Where fledging dates are prefixed with "-", the e¡øct range and average fledging dates were not provided so these are approximatiors based on cornments in the text. Fledging dates for Texas are for first nests only, differences in averages being between suburban and rwal sites. The latest fledging date for a renesting attempt in Texas was July 24.

101 Based on the random stratified survey and assuming a maximum detection dist¿rrce of

I25m around each listening statioq the density of tenitorial nnles detected ranged from

0.2 - 0.8 males/km2 per year. When data were pooled over the four years of the suvey

and repeat detectiorn of the same male (i.e. detected at the sanp location in consecutive

years) were rerÐved, this gave a density of 0.7 males/km2. I used the pooled data to

calculate a density for each of the 18 survey categories and multiplied this by the amount

of land in each category in the study area as a whole. This produced a population estimate

of 59.77 + 3.95 males (standard error of the mean of area x density per category) or 0.01

territories/km2 ouer the entire study arca (5026.4km2¡. This estimate needs to be

interpreted with caution as not all males were paired and this can therefore not be

extrapolated to the dersity of breeding pairs. Furthermore, it is possible that some owls

approached the observer from greater than 125rr¡ althougtr, with the high level of background noise at rnost listening points, this is unlikely. Another problem with this

exftapolation is that riparian and non-rþarian areas were given equal survey effort

whereas the total area of rþarian habitat is many times less than non-rþarian habitat.

Every year,25 - 35 territories were located, although sone males on territory did not have a partner and no npre than 14 pairs attempted breeding in any one year (Table 2.5).

Taking just the territories found within the perimeter of the city of Winnipeg (534km2),

produced a density estimate of 0.05 - 0.07 tenitories/km2. Suruey data (Table 2.1, Figure

2.2) showed that screech-owl densitþs are much higher in sr.rburban areas within the city than otfside the city. The survey dernity estimate of 0.11 tenitories/km2 should therefore be revised up if just suburbanareas are considered. Therefore, if the wildlands category

r02 (areas <1 p/ha) is removed from the survey calculation (no screech-owls detected) then

screech-owl density is 0.15 males/kmz across the rural - urban gradient.

DrscrssroN

Survey Data

ln the study area, screech-owls show a strong preference for riparian areas both in terms

of survey detections and in number of nest sites located. This was also suggested by

Kelso (1944) in New York. In the study area, the tallest trees and those rnost likely to produce suitable cavities (American elm, Manitoba maple, green ast¡ eastern cottonwood

and basswood Table 3.1) are typically found in the rþarian zone (Bird 1930, -see

Scoggan 1957 , Moffat et aL.2004). Taller [ees of urban habitats are selected b y some raptors, e.g. Great Horned Owl (Smith et aL 1999) and nny have positive effects on population density, e.g. Cooper's l{av¡k (Accipiter cooperü) (Boal and Mannan 1998). In the study area, riparian woodlands also contain a more open understory than aspen or bur

oak copses found fuither from the river (Staniforth z}}3),especially where flooding regularly deposits silt. This is particularly suited to the hunting style of the Eastern

Screech-Owl, which rrnst frequently hunts the ground layer from a perch (Gehhach

1994). Rþarian areas are also noted for their diversity and abundance of potential prey

(Doyle 1990, Naiman and Décamps 1997,1-nck and Naiman 1998, Sanders 1998). The

Eastern Screech-Owl appears less dependent on rþarian habit¿t fi-nther south towards the core of the range, where they utilize a wide variety of forest and woodland types

(Gehhach 1995).

103 Fig.2.I suggests that Eastern Screech-Owl density may peak at20 plln; however, the

120 ransects of this survey proved insufficient to adequately assess this pattern The 20 p/tra figure represents an intermediate human density and rnost suburbs with this density

in the study area were well-need. Other studies have also found cavity-nesting birds to have higher diversity and reproductive success in sr-rbrnbs with relatively contiguous free

cover (Blewett and Marzluff 2005). The number of nesting attenpts was highest in the 52

category (44Vo) as coÍpared to only l5%o n the 53 category, providing fuither evidence

of high reproductivity in rredium-density suburban areas. The paucity of sightings in the

downtown core with clusters of territories in suburban areas along the major rivers (Fig.

2.2) aßo indicate reduced usage of frue urban areas'with the highest anpunts of

impervious surfaces. Nonetheless, one successful nest was situated alongskle several high-rise condominiuÍN near the downtown core with a human density of 99 p/ha and

corrparatively few trees except ina narrow sftip along the river.

This study found minimal differences in EasternScreech-Owf s use of urban greenspace

as opposed to residential areas, suggesting a high tolerance for human activity. Eastern

Screech-Owl's small sun atÅ, highly nocturnal habits mean that even whenbreeding in

densely-populated areas, interactiorn with humans may be minimal Conversely, all 10

Great Horned Owl nests located in the study area in the same period were situated in rural

or suburban greenspace, in particular in or on the edge of golf courses that were seldom visited by hunnns at night. The small open lawns of residential neighborhoods with large

trees may be more suitable to screech-owls, while larger open spaces are nþre readily

exploited by Great Horned Owls. The Great Horned Owl is a rrnderately disturbance-

r04 adapted species (Bosakowski and Smith 1997) and the fact that peak densities were recorded in the rural category is indicative of this. When transects were separated by year, Great Horned Owls were found on only five transec ts (IzEo)where screech-owls were detected, suggesting that their presence may deter screech-owls; however the average detection rate of Great Horned Owl was not statistbally significant in predicting screech-owl density (Table 3 ). This may be because Great Horned Owl calls were not broadcast and detection rates thus not acctnately reflective of population dernities.

Brceding

Although I found differences between high-density suburbar¡ low-density suburbar¡ and rural breeding pairs in terrrs of brood size and fledging dates, these were not significant due to srrnll sanple sizes in the rural category. High-density suburban areas show the highest average brood size per nesting atterrpt and rural pairs the lowest; however, lifetime productior¡ as opposed to brood size, would be a better indbator of biologically significant differences. The combination of breeding data with the least squares means test and orthogonal corìtrasts from the survey data suggest that the high and nrcderate dersity suburban areas pattern similarly in terrrs of Eastern Screech-Owl population density, brood size and fudging, and differ from lower suburbar¡ rural and wildland areas in these aspects. Srnaller differences exist between lower subr¡rban and rural areas.

There is considerable evidence for a lfuk between warm tenperatues and early breeding in owls, especially in northern teÍperate climates (Watson 1933, Elder 1935, Martilgg4,

Andrrsiak and Cheng 1997,Holt and Drasen 2001, Artuso 2007b). The urban heat island in combination with unusually warm winters can advance breeding phenolo gy in owls by

105 up to six weeks (Arttso 2007b). Earlier fledging closer to the city center and with warmer

March temperature and earlier melting of snow cover point to the infiuence of the urban heat island in advancing breeding phenology. The five-day difference between the average fledging dates of suburban versus rural fledging dates in this study was closely matched by the six-day difference noted in Texas (Gehhach 1994b). Early breeding is presunubly advantageous to owls to reduce predation and competition from other raptors

(Sunde 2005) and increase fledgling survival (Gehlbach 1994). The urban heat island mayalso increase the carrying capacityof urbanareas (EmlenL974, Shochat etaL2006), proving greater food resources (Gehhach1994, Rollinson and Jorps 2002) or earlier access to sorrrc food resources (Chapter 4), as well as a head start to breeding (Korpimäki

1978, GehhachI994b).

The correlation between brood sizes and precþitation suggests that increased nnisture provides either increased prey abundance and/or improved access to prey, e.g. earthworm capture was highest in the wettest year (Chapter 4). The watering of suburban lawns and gardern is therefore an artificial form of climate mitigationthat canbenefit certain species (Shochat etaL2O04, Adams 2005, Parris and Hazel2OO5, Shochat etaL2006).

Increased access to water and increased rnoisture in soils have been factors in the colonization of urban areas (Yeh and Price 2004) and have had a positive influence on the breeding of other birds (Noske 1998, Kristan 2001, Rollinson and Jones 2002).

Screech-owl dersity in the study area appears to be much lower than other parts of the geogaphical range. The dersity calculated for the city of V/innipeg (0.15 males/km2) is

106 only a fraction of densitþs in suburban Connecticut (1.2 pairs/km2) and in a suburban

area of cenfral Texas (4.4 -7.4 panslkm2¡ lcehhach 1995). The nearest neighbor

distances of nesting pairs in Winnipeg was 612m versus 30m (suburbia) and 190m (rural

area) in Texas (Gehhach 1995). Of the 61 cavities, no nrore t\nn4T%o of those available

(not within the tenitory of another owl clogged by squinel activity or otherwise

damaged) were used in any given year. These data suggest that the population in the

study area is unsaturated. Cavity use rates suggest that the number of potential nest-sites

is not limiting (Chapter 3). The larger brood sizes than elsewhere in the range recorded

here, combined with large distances between territories suggest that this population is

limited by mechanisrns that irrpact survival rates (Greene and Starrps 2001, Sunde et al.

200I, Brown et al. 2002). Given the low cold tolerance of this species (Mosher and

Henny I976) and the lack of invertebrate food sources in winter (Chapter 4), winter

survival rates, especially of young birds, may be very low

Screech-owl territory occupancy was not consistent over the fotn years of this study, with

several pairs occupying fornerly unoccr.pied territories or abandoning them This was also the case for the Tawny Owl (Srrn aluco) in a range-perþheral context (Sunde et al.

2001). The reasons why screech-owls abandoned cavities and./or territories are not clear; however, in forn cases (of 39) abandonment coincided with the appearance of a larger owl species moving into the area. Great Horned Owl breeding appeared to cause screech- owls to vacate tlree territories and in one case the presence of a Baned Owl in the early spring coincided with the abandonment of a territory after 2 years of successful breeding.

Sonp unpaired males may have abandoned territories after failing to attact females or

107 replacement partners (9 cases) and increased human disturbance such as constuction activity or tree- felling close to a forner nest site may have resulted in four cases of abandonment. Although these cases represent anecdotal evidence, they suggest that predator avoidance is at least one cause of the density pattern observed in the study area

(positive correlationbetween screech-owl density and human density versus negative correlation between Great Horned Owl density and human density). This is not surprising since large owls have been shown to significantly impact the breeding and site occupancy of other smaller synpatic raptors (Sergio et aL 2003, Brambilla et aL 2006, Elliott 2006).

Boreal Owls (Aegoliusfunerezs), which are similar in size and weight to Eastern

Screech-Owls, appear to avoid nesting in proximity to the larger Ural Owl (Strix uralensis) but not the very large Eurasian Eagle-Owl(Bubo bubo) (Hakkarainen and

Korpimäki 1996). Despite being closely related to the Eurasian Eagle-Owl (cited as 2.8kg in Hakkarainen and Korpimäki 1996), the Great Horned Owl (average of Manitoba specirrens in Manitoba Mr¡seum = l.zkg, n = 16 of which 9 male, 2 fennle and 5 undetermined) is much rnore similar to the Ural Owl (cited as 0.9kg in Hakkarainen and

Korpimåiki 1996) in size and weight.

Both density and breeding success may peak above 2O plIn but it is possible that increasing human occupancy and developrrent above a threshold higher than 30 p/ha could negatively influence screech-owls. Owls that benefit from suburban conditions may still be disturbed if the human presence becones too high, e.g. Burrowing Owls in

Florlla benefited from high prey densities around human horres but increasing developnnnt lead to a reduction in nesting success (Millsap and Bear 2000). Powerful

108 Owß Q,linox strenua) that breed in sr-rbwban areas have abandoned nests, even eating their own young, when hurnan activity increased (Cooke et aL.2002).

The differences between suburban and rural broods in Manitoba was less pronounced than in Texas (Gehlbach 1994b); however this may be an artifact of the gradient approach of this study versus the corrparison of different study sites in Texas and different definitions of 'lural" versus "sr:burban" areas, e.g. Gehlbach's (1994b) suburban sh¡dy sites had only 4 p/ha and 5 plln, versus

(Gehhach L994b). Nonetheless, the urban heat island appears to be a significant factor in the higher densities of Eastern Screech-Owls in suburban Winnipeg, as evidenced by the correlation of fledging date to distance from city center and the larger brood sizes in suburban areas.

Lnrn¿,runn CrrED

Adans,. L. W. 2005. Urban wildlife ecology and conservation: A brief history of the

discþline. Urban Ecosystems 8 :139-156.

Andrmiak, L. 4., and K. M. Cheng. 1997 . Breeding biology of the Barn O wl (Tyto alba)

in the lower mainland of British Columbia. Pages 3846 ¿r? J. R. Duncar¡ D. H.

Johrnoru and T. H. Nbholls, editors. Biology and conservation of owls of the

Northern Hemisphere. North Cenftal Research Station, U.S. Departnent of

Agriculture, St Paul, Minnesota, U.S.A.

109 Artuso, C. 2007 a. Eastern Screech-owl hatches Wood Duck e ggs. Wilson Journal of

Omitho lo gy I L9 :I I0-I 12.

Artuso, C.2001b. January initiation of suburban Great Horned Owl nests in Manitoba.

Blue Jay 652613.

Battirl J.2A04. When good animals love bad habitats: Ecological traps and the

corservation o f anima I pop ulations. Co rservation B io lo gy 18 :1482-1 49 I.

Best, L. B. 1986. Corservation tillage: ecological taps for nesting birds? Wildlife

Society Bulletin 14:308-317 .

Bevanger, K. 1998. Biologbal and conservation aspects of bird rnortâlity caued by

electrbity power lines: a review. Biological Conservation36:6716.

Bird, R. D. 1930. Biotic communities of the aspen parkland of cenftal Can¿da. Ecology

LI:356441.

Blak, R. B. 1996. l¿nd use and avian species diversity along an urban gadient.

Eco lo gical App lications 6 :506-5 1 9.

Blewett, C. M., and J. M. Marzlutr 2005. Effects of urban sprawlon snags and the

abundance and productivity of cavity-nesting birds. Condor I07:6i8493.

Boal C. W., and Mannan R. W. 1998. Nest-site selection by Cooper's Hawks in an urban

environnpnt. Jotnnal of Wildlife Managemert 62:864-87I.

Bock, C. E., and Z. F. Jones. 2004. Avian Habitat Evaluation: Should Counting Birds

Count? Frontbrs in Ecology and the Environmerf2:403-4I0.

Brambilla, M., D. Rubolini, and F. Guidati. 2006.Eagle Owl Bubo bubo proximity can

lower productivity of cliff-nesting Peregrrnes Falco peregrinus. Omis Fennica 8320-

26.

110 Bosakowski T., and D. G. Smith. 1997. Distribution and species richness of a forest

raptor community in relation to urbanization Journal of Raptor Research 3I:26-33.

BrowrL D. R., C. M. Stong, and P. C. Stouffer.2002. Demogaphic effects of habitat

selectionby Hermit Thrushes wintering in a pine plantation landscape. The Journal of

Wildlife Mana gene nt 66 :407 4L6.

Chace, J. F., and J. J. V/alsh 2006. Urban effects on native avifauna: a review. Landscape

and Urban Planning 7 4:4649.

Cooke, R., R. Wallis, and A. Webster. 2002. Urbanisation and the ecology of Powerful

Owß (Ninox strenua) in outer Melbourne, Victoria. Pages 100-105 irz I. Newtoru R.

Kavanagl¡ J. OIsen, and I. R. Taylor, editors. Ecology and Conservation of Owls.

CSIRO Pub lishing, Co llingwood, Victoria, Ausûalia.

Copley, D., D.Fraser, and J. C. Finlay. 1999. Purple Martins, Progne subis; a British

Columb ian s uc cess s tory. Canad ian F ie ld- Nannalis t 1 1 3 :226--229 .

Craighead J. J., and F. C. Craighead. 1956. Hawks, owls, and wildlife. Stackpole

Conpany, Harrisb urg, Penns ylvania.

Crooks, K. R., and M. L. Soulé. 1999. Mesopredator release and avifaunal extirrctions in

a fragnented system Nature 4O0:563-566.

Donnelly, R., and J. M. Marzluff. 2006. Relative importance of habitat quantity,

stucture, and spatial pattern to birds in urbanizing environrrrcnts. Urban Ecosystems

999-rr7.

Doyle, A. T. 1990. Use of riparian and upland habitats by small mammals. Journalof

Mamrnalogy 7I:14-23.

lll Duly, L. J. 1979. Life history aspects of the screech owl (Otus asio) n Tennessee. Thesis,

University of Tennessee, Knoxville, Teruressee, U.S.A.

Elder, W. III. 1935. Early nesting of the Great Horned OwL Auk 52:309110.

Emlen, J. T.I974. An urbanbird community in Tucson, Arizona: derivation, süucture,

regulation Condor 7 6 :L84-I97 .

Elliott, K.2006. Declining numbers of WesternScreech-owl inthe lower mainland of

British Columb ia. British Columb ia B irds 14 :2-1 1.

Ellison P. T. 1980. Habitat use by resident Screech Owls (Otus asio). Thesis, University

of Massachusetts, Amhers t, Ma ssachusetts, U. S. A.

Environnent Canada. Weather Pages. Retrþved September 2004 - December 2008 from:

http : / I w eathero ffice. e c. gc.ca.

Estes, W. 4., and R W. Mannan 2003. Feedingbehavior of Cooper's Hawks at urban

and rural nests in southeastern Arizona. Condor 105:107-116.

Forrnar¡ R. T. T, and L. E. Alexander. 1998. Roads ard their rnajor ecological effects.

Annu¿l Review of Ecology and Systematbs29:20713I.

Galeotti, P. 1990. Territorial behaviour and habitat selection in an urban population of the

tawny owl Srr¿x aluco L. Bolletino di Zoologia 57 :5946.

Gehlbach F. R. 1994a. Nest-box and natural-cavity nests of the Eastern Screech-Owl: an

exploratory study. Journal of Raptor Research 28:154-T57.

Gehhacb F. R. 1994b. The Eastern Screech Owl: Life history, ecology, and behavior in

the suburbs and counftyside. Texas A&M University Press, College Statiorl Texas,

U.S.A.

112 Gehhactu F. R. 1995. Eastern Screech-Owl(Otus asio). Account 165 in A. Poole and F.

Gill, editors. The Birds of North America. The Academy of Natural Sciences,

Philadeþhia Pennsylvania and The American Omithologists' Unbr¡ WashingtorL

D.C., U.S.A.

Greene, C. M., and J. A. Stanps. 200I. Habitat Selection at [¡w Population Densities.

Ecology 82.,2091-2100.

Hakkarainerl H. and E. Korpimäki.1996. Conpetitive and predatory interactions annng

raptors: an observational and experimental study. Ecology 77 :1134-1142.

Holt, D. W., and S. Drasen. 2001. Early nestingby Great Horned Owls inMontana.

Journal of Raptor Research 35:6647 .

Imhoff, M. L., C. J. Tucker, W. T. Lawrence, and D. C. Stutzer. 2000. The use of

multisource satellite and geospatial dat¿ to study the effect of urbanization on primary

productivity in the United States. IEEE Transactions on Geoscience and Remote

Sersing 382549-56.

Johnson R. R., B. T Browr¡ L. T. Haight and J. M. Sirrpson 1981. Playback recordings

as a special avian censtsing technique. Pages 68-75 in C. J. Ralpt¡ and J. M. Scott,

editors. Estimating numbers of terrestrial birds. Studies in Avian Biology No. 6.

Cooper Omithological Society, New Jersey, U.S.A.

Jokimliki, J.1999. Occurrence ofbreedingbird species in urbanparks: effects of park

structure and broadscale variables. Urban Ecosystems 3:2114.

Kelso L. H. 1944. Behaviou of the eastern screech owl (Otus asio naevius). Biological

Iæaflet 23, Cornell University, New York, U.S.A.

Klem, D. Jr. 1989. Bird-window collisions. WilsonBulletin 10I:606420.

I13 Korpimäki, E. 1978. Breeding biology of the Starling Sturnus vulgaris in western

Finland. Ornis Fennica 55:93-104.

Kristan" W. B. 200I. Effects of Habitat Selection on Avian Population Ecology in

Urbanizing I-andscapes. Dissertation University of California, Riverside, California,

U.S.A.

[-ock, P. ,A,., and R. J. Naiman 1998. Effects of stream size onbird community structure

in coastal temperate forests ofthe Pacific Northwest, U.S.A. Journal of Biogeography

25:773J82.

LynchP. J., and D. G. Smith. 1984. Censrs of Eastern Screech-Owß (Otus asio) n urban

open-space areas wing tape-recorded song. American Birds 38:388-391.

Marti, C.D.1994. BarnOwl reproduction: Patterns and variation near the limit of the

species' distribution Condor 96 : 468-484.

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

colonization may determine biological diversity in human-dominated landscapes.

Urban Ecosystems, 8 :I57 -I7 7 .

MarzluffJ. M., R. Bowman, and R. Donnelly. 200Ia. A historical perspective on urban

bird research: trends, ternß, and approaches. Pages I-17 in J. M. Marzluff, R.

Bowman, and R. Donnelly, editors. Avian ecology and cornervation in an urbanizing

world: Kluwer Academic, Bostorl Massachusetts, U.S.A.

Marzluff, J. M, K. J. McGowar¡ R. Donnelly, and R. L. Knight. 200lb. Caues and

consequences of expanding ArrnricanCrow populations. Pages 331-363 inJ.M.

Marzluff, R. Bowman, and R. Donnelly, editors. Avian ecology and conservation in

an wbanizing world. Kluwer Academic, Bostoq Massachusetts, U.S.A.

TT4 Millsap, B. A.2002. Survival of Florida burrowing owls along an uban-development

gadient. Journal of Raptor Research 36:3-10.

Millsap, B. 4., and C. Bear. 2000. Densityand Reprodrrctionof Burrowing Owls along

an Urban Development Gradient. Journal of Wildlife Managerrerrt64334I.

Moffatt, S. F., S. M. Mcl¿chlar¡ and N. C. KenkeL 2004. Irrpacts of land use on rþarian

forest along an urban - rural gradient in sorfhern Manitoba. Plant Ecology I74:LT9-

135.

Mosher J. 4., and C. J. Henny. 1976. Thermal adaptiveness of plumage color in Screech

Owls. Auk 93:614419.

Munay G. A. 1976. Geographic variation in the clutch-sizes of seven owl species. Auk

93:602413.

Naiman, R. J., and H. Décamps. 1997. The ecology of interfaces: rþarian zones. Annual

Review of Ecology and Systematics23:621-ó58.

Noske, R. 1998. Breeding biology, demogaphy and success of the rufous-banded

honeyeater, Conopophila albogulans, in Darwin, a monsoonal tropical city. Wildlife

Research 25:339156.

Overskaug K., and J. P. Bolstad 1998. Geographical variation in female mass and

reproductive effort in the Tawny Owl Srr¿x aluco in Ernope. Fauna Norvegica Seria C

2I:l-6.

Palomino, D., and L. M. Carrascal. 2006. Urban influence onbirds at a regional scab: A

case study with the avifauna of northern Madrid province. landscape and Urban

Planning 77:276190.

115 Panis, K. M., and D. L. Hazell.2005. Biotic effects of climate change in urban

environrrents: the case of the grey-headed flying-fo x(Pteropus poliocephatus) n

Me ho urne, Australia. B io lo gbal Conservatio n L24 :267 17 6.

Peck G. K., and R. D. James. 1983. Breedingbirds of Ontario: nidology and distrbution

Volume 1: Non-passerines. Royal Ont¿rio Museum priblicatiors in life sciences,

Toronto, Ontario, Canada.

Renns, V. 2000. How can maladaptive habitat choice produce source-sink population

dynamics? O ko s 9t :57 9-582.

Rollinson, D. J., and D. N. Jones. 2002. Variation in breeding parameters of the

Australian magpie Gymnorhina tibicen in subrnban and rural environnents. Urban

Ecosystems 6257169.

Sanders, T. A. 1998. Breeding bird community conposition inrelationto rþarian

vegetation structure in the western United States. Journal of Wildlife Manage 62:461-

473.

Scoggar¡ H. J. 1957. Flora of Manitoba. National Museum of Canada, Biological Series

No. 47, Bulletin No. 40. Ottawa, Canada.

Sergio, F., L. Marchesi, and P. Pedrini. 2003: Spatial refugia and the coexistence of a

diurnal raptor with its intaguild owlpredator. Journal of Animal Ecology 72:232-

245.

Sergio F., Marchesi L., Pedrini P., Ferrer M. & Penteriani V. 2004. Electrocution alters

the disnbution and density of a top predator, the eagle owl Bubo bubo. Journal of

Applied Eco lo gy 4l :836-845.

TI6 Shochat, E., S. B. Iærmar¡ M. Katti and D. B. tæwis. 2004. Linking optimal foraging

behavior to bird community stucture in an urban desert landscape: Field experiments

with artificial food patches. The American Naturalist 164:232142.

Shochat, E., Warrer¡ P., Faet[ S., Mclntyre, N., Hope, D.2006. From patterns to

errrrging processes in mechanistic uban ecology. Trends in Ecology and Evolution 2I:186-19I.

Smith D. G., T. Bosakowski and A. Devine. 1999. Nest site selection by urban and rual

Great Horned Owls in the northeast: Journal of Fbld OrnithologyTO:535-542.

Ståniforth R. J. 2003. Ecosystems and bird habitat in Manitoba. Pages 34-55 in

Manitoba Avian Research Committee, P. Taylor, editor- in-chief. The Birds of

Manitoba. Manitoba Naturalists Society, Manitoba Avian Research Committee,

Winnþeg, Manitoba.

Statistics Canada. 2004. Profile of Cersus Tracts in Winnþeg,2001 Census. Catalogue

No. 95-241-)GB. Retieved February 2004 from:

http:/ldatacentre.chass.utoronto.ca.proxyl.lib. umanitoba.calcernus/index.html

Sunde, P., K. Overskaug, J. P. Bolstad, and l. J. Øien 2001. Living at the limit: ecology

and behaviour of Tawny Owls Strix aluco in a northern edge population in Cenftal

Norway. Ardea 89 495-508.

Sunde, P. 2005. Predators control post-fledging mortality in tawny owls, ^!rnx aluco. Okos IIO:461472.

Tyre J. 4., B. Tenhumberg, s. A. Field, D. Niejalke, K. parris, and H. p. possingham.

2003.Inproving precision and reducing bias in biological surveys: Estimating false-

ne gative error rate s. Eco lo gic a I App licatiorn 1 3 :1 79G-l 80 1 .

r77 VanCanp L. F., and C. J. Henny. L975. The Screech Owl: Its life history and population

ecology in northern Ohio. North American Fauna: 71. Unites States Department of

the Interior Fish and Wildlife Service.

Wang, Y., S. Chen, and P. Jiang. 2008. Black-billed Magpies (Pica pica) adjust nest

characteristics to adapt to wbanization in Hangzhor¡ China. Canadian Journal of

Zoology 86:676484.

Watson, G. H. 1933. Early nestingof the Great Horned OwL Auk 50:220-221.

Yeh, P. J.2004. Rapid evolutionof a sexually selected trait following population

establishment in a novel habitat. Evoh¡tion 58:166-174.

118 3. Habitat selection by the Eastern Screech-Owl along a rural - urban gradient

Ansrnacr.- Secondary-cavity nesting birds rely on a potentially limiting resource that governs their nest-site selection Cavity-nesting birds may be sensitive to urbanization due to changes in the availabilityof this resource. Altered selectionpressures in urban environnents may also produce changes in the process of habitat selection I examined the nest-site selection of the Eastern Screech- Owl(Megascops asio) n Winnþeg, Manitoba and surrounding area and compared habitat featues in suburban and rural nest sites. Conditional regression analysis indicated that canopy cover, the number of potential nest cavities, percent conifer and the DBH of ' trees taller than 10m were selected for at the level of habitat around the nest. Nest-site variables selected for were the heightof the nest tree, distance to river, and shrub densitybelow the nest. Screech-owls also selected sites with fewer buildings within a 50m radius and less domestic cat (Felis catus) activity. Rural nests differed from low- density suburban and high-density suburban nests in having denser canopy and vegetation but with rnore smaller and slender trees (rnore convoluted middle story), fewer coniferous trees, more cavities but fewer nest boxes, and rnore raccoons (Procyon lotor) and Great Horned Owls (Bubo virgininnøs) but less dorrestic cat activity. In the study area, Eastern Screech-Owls prefer rþarian habitat. Sone selected featues such as the presence of coniferous fees close to the nest and minimal shnrb density below the nest were Íþre lkely to be found in sububan areas than rural areas. Eastern Screech-Owls nest in higher densitbs in sr¡burban areas; which may offer more preferred hab itat characteristics than rural areas.

Keywords: Eastern Screech-Owl Megascops asio, Winnþeg, Manitoba, habitat preference.

IvrnoructroN

Secondary cavity-nesting birds are dependent on the availability of suitable cavities, a potentially limiting resource (Lundberg 1979, Korpimäki 1987, Brawn and Balda 1988,

Newton I994b, Holt and Martin 1997,bùt see SoúhernI97}, Mossop 1997,

Severinghaus 2007, Wesolowskí2007). Most cavity-nesting owls show stong territoriality and site fidelity, believed to be a response to secuing access to limited

119 nesting resources (Taylor 2002). As a polyterritorial species, the Eastern Screech-Owl

(Megascops asio) is further limited by this resource because each territorial pair defends several cavities (Gehlbach 1994b). Secondary cavity nesters may not select for certain features that are important to primary cavity nesters, e.g. many woodpeckers exhibit selection for cavity orientation with associated benefits in terms of microclimate (Conner

L975, Crocket and Hadow 1975, Inouye et al. 1981, Hooge et aL 1999, Wiebe et aL 200L) whereas cavity-nesting small owls oftendo not (McCallum and Gehlbach 1988, Belthoff and Ritchison 1990a, Gehlbach 1994b). However secondary cavity nesters still e4perbnce context-specific selection pressures for such features, e.g. Northern Spotted

Owls (Srrrx occidentalis caurina) onthe Olympic Peninsula of Washington state avoid cavities facing the prevailing direction of inclerrent weather (Forsman and Giese 1997) and Elf Owß (Micrathene whitneyi) in the Sonoran desert select for cavity orientation, either reflecting the preferences of the woodpeckers that excavated their cavities or the heightened importance of microclimate in an arid climate (Hardy and Morrison 2001).

The latter hypothesis is sftengthened by the fact that Elf Owls nesting in cooler upland areas of the Sonoran desert do not select for cavity orientation (Goad and Mannan 1987).

The princþal selection pressures that influence nest-site selection by cavity-dependant species include, but are not limited to, predation risk (Nilsson 1984, Sonerud 1985b,

Martin 1988, Rendell and Robertson 1989, Severinghaus 2007), interspecific and intaspecific competition (van Balen et aL 1982, Nilsson 1984, Artuso 2007 , Aitken and

Martin 2008); parasite load (Mappes et al. 1994, Aitken et al. 2002), and weather-related factors (Forsman and Giese 1997, Severinghaus 2007). Ftnthermore, socialbehavior

t20 such as the degree of intaspecific interactions, territoriality, dominance, and the likelihood of facultative polygamy are also important (Rendell and Robertson1994,

Aitken and Martin 2008). Such pressures are manifest in selection preferences related to cavity enüance size (Sonerud 1985b, Korpimäki 1987, Belthoff and Ritchison 1990a,

Yetter et aL 1999), cavity depth (Belthoff and Ritchison 1990a), cavity volume

(Korpimäki I98l , Carlson et al. 1998, Aitken and Martin 2004), cavity height (Søuffer and Best 1982, McCallum and Gehhach 1988), cavity floor dryness (Severinghaus

2007), cavity wall thickness (Carlsonet al. 1998), dianpter of branchor stem at cavity height (Carlson et aI 1998), degee of concealment (Serrel et al. 1988), the proximity of alternate cavities (van Balen etaI.1982, Martin 1988), and the species-specific likelihood of cavfy reuse (Aitken et aL 2002).

Many recent studies point to differences in the habitat selection process at different spatial scales (Johrnon 1980, Jones 2001) and, inadditionto cavity-specific features, vegetation characteristics surrounding the nest and at the landscape- level are also important. In the case of secondary cavity nesters, nest-site vegetation features have been found by sone studies to be less significant than cavity-specific features (Gutzwiller and

Anderson 1987, Belthoff and Ritchison I99Oa, Sedgwick and Knopf 1990, Severinghaus

2007). Nonetheless, cavity-nesting owls have been shown to select for features such as the arnrunt of shrubbery in front of the nest úee (McCallum and Gehhach 1988,

Flammulated Owl Otus flammeolus), relative canopy height (McCallum and Gehlbach

1988, Flammulated Owl), moist deciduous forest (Mftkola 1983, Tawny Owl.Srru.r aluco), open matue forest (McCallum and Gehlbach 1988, Flammulated Owl), Iarge

rzt trees with considerable understory vegetation (Devereux and Mosher 1984, Barred Owl

Strìx varia), and the density of mature saguaro cacti (Hardy and Morrison 2001, EH

Owl). These authors linked these selection features to predator avoidance, prey

availability, and hunting style.

The range of the Eastern Screech-Owl covers nrcst of eastern U.S.A, extending from

approximately 22'N in northeast Mexico to 50"N degees in Manitoba and southeastern

Saskatchewar¡ Canada (Gehhach 1995), and occasionally as far as 52oN (Walley and

Clyde Iggl).Different populatiorn of this species experience very different climates and habitats, which could influerrce habitat selection This species is resident in parts of

Manitoba that lie between the - 10oC and -20"C average January temperature isotherms

(Bartholomew 1985), which approaches the cold tolerance threshold of these birds

(Mosher and Henny 1976). There are numerous ways that such a range-peripheral context could influence habitat selection (Fuller et aL 2007). If cold temperatures ard limited resources resulted in lower survival rates and lower population dersities (Brown et al.

1995), then density-dependent competition and territoriality could conceivably be sonpwhat relaxed, resulting in either larger territories or greater distances to nearest neighbor (Greene and Stanps 200I, Sunde etal.200l, Brown età1.2002) or more pronounced seasonal shifu in areas occupied (Sunde et aL 2001). On the other hand, climate, food supply, or the limited availability of trees large enough to contain suitable nesting cavities could limit these bírds to very specific habitats, e.g. rþarian zones or areas influenced by the urban heat island, resulting in a rnore clustered population (Sasaki

1997). Such context-specific dersity consüaints influence habitat selection (Bult et al

r22 1999, Femandez-Juricic 2001), e.g. selection for vegetation around the nest-site was

indicated in a population of Flammulated Owls believed to occur in low dersity

(McCallum and Gehhach 1988), whereas such selection was not found in a population of

Eastern Screech-Owls believed to occrn at higher density (Belthoffand Ritchison 1990a).

Eastern Screech-Owls in Manitoba might also be influenced by a different predation regime than corspecifics in rn¡re soúherly locatiors, contending with raccoons (Procyon

lotor) but not the suite of smaller nest predators found further south including the black ratsnake (Elaphe obsoleta), Virginia opossum (Didelphß virgininna) and ringtail

(Bas s aris cu s ast ut us) (Gehhach I99 4b).

ln addition to broad geographical variables, urbanization and other forms of hunran

alteration of landscapes can influence avian habitat selection Sorrp species have expanded their chobe of nest site to include human sffuctures, the best-known exarrple being the Peregrine Falcon (Falco peregrínus) nesting on urban high-rises as opposed to cliffs and the accorrpanying change in dbt that this entails (Horak 1986, Septon et al.

1996, Tenple 1988). Great Horned Owls (Bubo virgini.anas) have also come to use buildings, pþes and bridges for nest-sites (Houston et al. 1998). Habitat changes

associated with urbanizztionareas have other consequences for Great Horned Owls, since

in suburban areas they exhbit higher nest-site selectivity than rural birds due to the

increased structual corrplexity of suburban environrrpnts (Smith et al. 1999). Ernasian

Sparrowhawks (Accípiter nísus) in villages in The Netherlands, which enjoy a different

diet than rural birds, ú.tllze a greater variety of nest types and produce larger clutches

(Diernren 1996). Urban Cooper's Flawks (Accipiter cooperü) with access to concenftated

123 prey sources have srnaller territories and alter their nest-site selection to favor non-rative nees (Chiang et al. 2006). For cavity-nesting birds, the availability of natural cavities may be much more limiting in habitats that are altered by human activity to varying degrees (Newton 1994a) than in pristine habitats (Wesolowski2007), and the provisioning of nest-boxes rtsy alter broad habitat-selection or reproductive success

(Mãnd et al. 2005).

Human activity can influence avian habitat selection in subtle ways; for exarrple,

Eurasian Magpies (Pica pica) select higher nest-sites in proximity to pedestrian activity to minimize distubance (Wang et aL 2008). Eurasian Blackbirds that breed in small fragrnents in urban settings have also altered their nest-site selection and prefer conifers and hedgerows (dense cover) in proximity to buildings for predator avoidance (Møller

1988). Northern Cardinals (Cardinalis cardinalís) selected nest-sites that were higher from the gound and more peripheral in the nest plant when near trails than when in forest interior possbly as a resuh of different predation risks (Smith-Castro 2008). The effects of urbanization and broad geogaphical variables can also interact, e.g. the Red- shouldered Hawk (Buteo lineatus) avoids suburban areas in New Jersey (Bosakowski and

Smith 1997), nests in groves of non-native ftees in California (Bloom and McCrary 1996) and shows no difference in habitat selection in urban versus rwal areas of Ohio (Dyksna et aL 2000, Dykstra et al. 2001).

Cavity-nesting birds are thought to be particularly sensitive to rnbanization (DeGraffand

'Wentworth 1986, Natuhara and Imai 1996, Bull et al. 1997, Rottenborn 1999, Fraterrigo

r24 and Wiens 2005, Blewett and Marzlutr 2005) because hurnan intervention and increased exposure to weather events may renþve potential nesting, roosting and foraging sites

(Blewett and Marzluff 2005) and in some cases they may be susceptible to increased corrpetition from a non-native cavity nester, the European Starling (Sturnus vulgaris)

(Ingold 1989,1994,1996). Nonetheless, the Eastern Screech-Owl has higher population density and fecundity in suburban areas in the southern portion of its range (Gehlbach

1994b). The influences of urbanization on this species' habitat selection is not well understood (see Belthoffand Ritchison 1990a). I examined the nest-site selection of

Eastern Screech-Owls across an wban - rural gradient (defined by human density) in

Wfumþeg; Manitoba" at the northern edge of this species' geographical range. I tested for selection by conparing cavities used by the owls with those that were available but unoccupied by owls at two spatial scales, flrstly at the cavity and cavity tree-specific level and secondly at the level of habitat surrounding the nest (50m radius). I also investigated potential influences to habitat selection including predation pressrne, the aÍþunt of urban greenspace, human dernity, building density and distance to nearest building, distarrce to roads and pedesftian activity.

METTronS

The study area was defined as a circle with a radius of 40km from the center of

Winnþeg, whbh encompasses the city limits and also areas outside the city with lower human densities. In the study area, riparian areas are dominated by American elm

(Ulmus americana), geen ash(Fraxinus pennsylvanica), and Manitoba maple (Acer negundo), as well as btn oak (Quercus macrocarpa), eastern cottonwood (Populus

r25 deltoi.des), basswood (Tília americana) and peachleaf willow (Salix amygdaloides) to a lesser eKent. Further away from the rivers, nembling aspen (Populus tremuloides) and bur oak becone dominant. Many planted native and exotic species also occur throughout the study area, in particular conifers such as white spruce (Picea glauca), eastern white cedar (Thuja occidentalis) and the non-native Colorado blue spruce (Picea pungens) and

Scots pine (Pinus sylvestris). Ahhough sonp relatively "natural" areas with mixed tee and shrub corrposition occur, in rnany parts of the city only one or two üee species were planted, e.g. Annricanelm may line streets in rows and conifers are oftenplanted as shelterbelts or tall "Ënces" þers. obs.). Riverbanks may be wooded, converted to open lawn or barked with rock. The shrub layer is particularly variable with rnany non-native ornamentals and/or cultivars, depending greatly on the preference of individual homeowners. For this reason I measured shrub and tree density but did not use plant species diversity counts as a variable. Tree composition by species was therefore chosen rather than a large-scale habitat classification schene.

Nests were located u,sing a three-step process. I frst narrowed the search area by triangulation from spontaneous calling sites and then undertook a ground search for the roost sites of males. Once a roosting male was located, I searched for all potential nest- sites within 100m of his roost and then waited at these cavities at dusk or dawn to observe the female. Females typically left the nest for l5-minute periods at dusk and dawn to defecate, drirk, bathe or accept a food item from their partner on a transfer perch close to the nest (Gehhach 1994b), enabling detection inthis manner. Nests were then rrnnitored and the number of chbks and fledging dates determined (Chapter 2).

126 Two randomdirections in degrees and dist¿nces 350 - 700m were then selected tning a random number generator and a cavity search found suitable cavities closest to these points. This distance range was chosen based on observation of local breeding pairs, the closest nests I had located being 7}4mapart l¿ter in the study I found two nests that were 6I2m apart across the Red River (150m wide in the relevant area). Observations of males ftilling were within 300m of the nest site or defended cavity. The cavities chosen for conparisons were therefore far enough away from a nesting cavity to be considered

'bff' the defrnded tenitory eventhough they would å11 well within the owl's honp range, hence irrcreasing the likelihood that they were selected against (McCallum and

Gehhach 1988, Hardy and Monison 2001), i.e. these cavities were available to the owls but not used. This nethod risks a bias of spatial autocorrelation since adjacent cavities may produce similar values for certain variables due to specific characteristics of the local area rather thandescribing differences between used and unmed cavities (Iænnon

2000); however, in site-atfribute design studies, the comparison of used versus unused habitas can only speak to selection if unused habitat is derrpnstrably available (Jones

2001). The key advantage of this technique is therefore that it focuses on sites that the owls canbe reasonably expected to have access to. This technique also permitted selection of cavities within areas of similar human density and similar distance to a river, which are both important to the owls (Chapter 2). Where possible, natural cavity nests and nest-boxes were matched with off-territory cavities and boxes respectively. Off- territory cavities were also chosen to match nests in terns of being situated in a geen space or in a residential neighborhood, although this distinction was less s ignificant to density or breeding (Chapter 2). These measures were designed to facilitate assessment of

t27 selection of microhabitat. Off-tenitory cavities were inspected with a Sandpiper teetop peeper video inspection system (TT4W) to ensue that they were not occupied and of suitable dimersions for nesting. Only cavities with an enüance 8 - 16cm in diarrpter and

25 - 60cmdeep (pers. obs., see also Belthoff and Ritchison 1990a, Gehlbach 1995) and not clogged with vegetative matter were considered available to the owls for nesting.

Dwk observations and tape playback around the off-tenitory cavity were also conducted on three nights 8 - 14 days apart to ensure that it was not within the territory of another screech-owl or larger owl Habitat measurements were then máde in the posrfledging period in order to minimize potential disturbarrce to nesting owls and to ensure cornistency in vegetation næasurements, since all tree and shrub species were fully foliated by this time.

Two habitat rnodels were planned - 1) broad habitat variables collected in plots of a radius of 50m around the nest tree or off-territory cavity tree, and 2) nest-site context variables, rnost of which were measurerrents made on the cavity ftee itself or else collected in a srnaller plot of l2mradits around the ftee (e.g. shrub density). The variables rreasrned for each model were 1) habitat variables: Vo carßp! cover, number of natual cavities and nest boxes, number of trees >10m tall, number of trees 5 - 10m tall,

average diameter at breast height (DBÐ of trees >10m tall, average DBH of üees 5 -

10m tall, Vo rntive rþarian canopy ftees (percentâges of American elm, Manitoba maple, bur oak, eastern cottonwood, and basswood independently counted and then combined),

7o other deciduous tees, Vo coniferous tees, 7o open understory, and 7o impervior,rs

surface andZ) nest-site variables: nest tree species, DBH nest tree, cavity height, height

r28 of nest tee, ftee dianeter at cavity height, diameter of cavity enffance, orientation of cavity, distance to nearest perrnanently flowing watercous e, Eo caîopy cover of nest tree, relative canopy height (defÏned as the average height of the three tallest canopy trees nearest to the nest tree), Vo shrub density, and minimum flight path (defìned as the minimum distance unobstructed by vegetation in a 45o angle forward from nest entrance, at or below nest height and >30cm off the gound). A maximum of eight variables were used per rnodel and some variables were omitted due to colinearity (see Tables 3.2.I,

3.2.3 arÅ3.2.4).

For statistical calculations, only independent cavities were considered (several cavities with repeat usage were only entered once). Measurements were made with a fape measure where possible, otherwise with a Suunto clinometer. L-arge horizontal distances were measured by tape measwements where possible or by reading from a nrrap in GIS

Arcview 3.2. Percent density measurements are all estimations made by subdividing the appropriate area into quadrants and then calculating total dersity.

In addition to the habitat models, a predator and disturbance nr¡del was pìanned based on information gathered around used and unused cavities. A series of predator-board trials were conducted at all sites in the posrfledging period. Fine-grained sand was spread over

45 x39cmtays, wet with a mist spray and then srnoothed to produce an even srnface.

The center of the tray was baited with a srnall amount of fish- flavored cat food. At each site, frays were laid at dusk at the most secluded site within 50mof the cavity and collected one hour after dawn Sampling was repeated on three separate occasions at least

r29 one week apart. This was to ensure minimal risk of disturbance by humans. If a nest was used over two consecutive years, the tays were laid in the same positions in the second

year. Raccoon and donestic cat indices were generated as the number of animals detected

divided by the number of nights where trays were laid. In additiorì. a Great Horned Owl

index was generated by calculating the proportion of Great Horned Owls detected on

survey üansects within the same census subdivision (Chapter 2) as the cavity. Other

variables, including distance to road, number of buildings wholly or partly within 50m of

nest, distance to nearest building, and human density (see Chapter 2) were also used in

the distubance model For the purposes of classifîcation, riparian areas were defined as

within 500mof a permanerìt river or watercourse. Greenspace was defined as parks,

ceneteries, golf courses, and green corridors such as rþarian buffer strips as well as

naturally occurring uninhabited copses (see Chapter 2).

I used a conditional logistic regression analysis for each of the three models because this

techniqrc permitted the comparison of two available sites per used site, increasing

sanple size and st¿tistical power. Futherrnore, the regression approach would allow for

the possibility of including both categorical and continuous variables (Hull Sieg and

Becker 1990). However, since homogeneity of variances is assurrpd, I flrst examined box

plots (Quinn and Keough2002) and used an F-test for variances between used and

available sites for each variable and then log-ftansformed any variable with

heterogeneous variances, except for percentage dala where the arcsine úarsformation

was used (7ar 1974), and then reexamined box plots. Correlation mafrices were examined

to ensure that collinear variables were not included in the same npdel. The nethod of

130 selecting two umsed cavities per used site and using tightly controlled selection paraneters meant than unmed sites often exhibited less variation (larger sample size with similar range) than used sites for many variables as indicated by the standard enor of the means and the coefficients of variation (Tables 2.I,2.2, and2.3). The rrethod of conparing variances between used and unused cavities across a range of variables as pioneered by McCallum and Gehlbach (1988) is therefore unsuited to this datå set.

ln order to ar:r,lyze the effects of human activity on habitat selection, I grouped independent cavity uses into three types: high-dersity suburbar¡ low-density suburban, and rural nests for three multiple discriminant function analyses (MDA) following the tfuee regression nndels. High-density suburban areas were defîned as having a human density of >30 people per hectare (p/ha) (the 53 category, Chapter 2) and averaged 4.37 +

0.57km from the city center. Two on-territory plots and three off-territory plots with slightlylessthan30p/ha (>27 plln)butwhichwere less than3kmfromcitycenterand in areas with high building density (12 private dwellings per hectare, plus non-residential buildings, Statistics Canada 2004) were reclassified as high-den:sity suburban areas. Low- dersity sr-lburban areas were def,rned as having 10 - 30 p/ha (the S1 and 52 categories,

Chapter 2); however, two on-territory plots and five off-territory plots with slightly greater than 30 plha (<34 plln) tlntwere >4km from the city center and few houses in proximity to the cavity in question were reclassified as low-density suburban areas. These sites all had approximately !2 private dwellings per hectare but without non-residential buitdings. Rural sites were defîned as having <10 p/ha (the wildlands and rual categories, Chapter 2); however, it is imporúant to note that the definition of the rural

131 category here is based on hurnan density and hence these areas are not necessarily associated with agriculture. The greatest number of nest-sites was in low-density suburban areas (n = 38, compared to 15 high-density sites and 8 rural sites).

MDA is not robust to heteroscedacity so I examined box plots and either log tansformed variables or, where transformations were insufficient to solve the problem of heteroscedacity, rerrnved the variable from the MDA analysis. One variable, the pedestian index was added to the disturbance MDA. The pedestrian index was only calculated on active territories because I only made three nighttime visits to off-tenitory sites, which was insufficient to produce a nreaningful index. The pedesftian index was cahulated by dividing the night into two-hour periods, P1: 8 - 10po,, P2: I0 - 12pr+ P3:

12 - Zan, P4: 2- 4ar\ andP5 : 4- 6am On each visit to a territory the number of pedestriars within 50m of the nest or fledged young was recorded. The number of pedestrians was divided by a correction factor for each period: Pl correction= 4,P2 conection= 3, P3 correction =2,P4 = 1, and P5 = 1.5. This purpose of the correction factor was to avoid biasing pedestrian counts by time of night (the number of people averaged approximately forn times higher near dusk than at 3am at any given site). The corrected number of pedesnians were then summed and divided by the total number of minutes from all visits and multiplied by 100.

Rnsur-rs

From2004 -2007,46 successful Eastern Screech-Owl nests, 6 failed nests, and37 stable territories (where a pair was active but did not breed or where one or rnore cavities was

t32 advertised by an unpaired male over a period greater than 6 weeks) were located (Chapter

2). Excluding repeat cavity uses, a total of 61 nest sites and territory centers were located, of which 38 (627o) were in natural cavitbs, 2I (34Vo) in nesting boxes for Wood Ducks

(Aix sponsa), and two (3Vo) at sites where cavity choice was undetermined. The majority of natrnal cavitbs were found in Americanelm and Manitoba maple (Table 3.1). None of the nest-sites used in this study were excavated by woodpeckers, presumably because the only two woodpeckers species in soúhernManitoba that excavate cavities of sufficient size to be used by screech-owls, Northern Flicker (Colaptes auratus) and Pileated 'Woodpecker (Dryocopus pileatus) (Gehlbach 1995) do not regularly breed in the suburbanand urbanpars ofthe study area.

Table 3.1. Distibution of natural cavities by tree species. Used cavities include only sites where at least one young fledged used unused Ulmus americana - Anerican elm 5 (25Vo) 23 (3OVo) Acer negundo - Manitoba maple 5 (25Vo) 29 (38Vo) Quercus mlcrocarpa - bur oak 3 (157o) 7 (9Vo) Fraxinus pennsylvanica - green ash 2 (IÙVo) 5 (6Vo) Populus deløides - easþrn cottonwood 2 (IÙVo) l0 (I3Vo) Tilia American¿ - basswood 2 (IOVo) 3 (4Vo) Populus balsamifera - balsam poplar I (5Vo) 0 (ÙVo)

Selection of Used Versus Unused Cavities

HRSITRTFEATURES

Four variables differed significantly between used and unu,sed cavity habiøt plots (Table

3.2.I). Screech-owls selected territories with nrore potential nest sites, averaging 3.4 potential nest sites on-territorybut only 1.8 off-tenitory (z =3.52, df = 8, p < 0.001). Nest sites selected by screech-owls also had significantly geater canopy cover within a 50m radius around the nest tree (40Vo versus 307o rcspectively, z = 2.7 , p = 0.007) although a

133 few nests were in snags with no canopy foliage directly above the nest. Screech-owls selected sites wiih a higher percentage of coniferous ffees (157o versus I2Vo respectively, z=2.13,p = 0.03) and where the DBH of tall trees (>10m) averaged slightly lower than unused sites (38.7cm versus 39.5crrì, z = -2.03, p = 0.04).

Table 3.2.1,. Averages + standard error of the mean of habitat variables with coefficbnt of variation (CV) and output of the conditional rePtression rnodel Used cavities Unused cavitþs X+SE CV XrSE CV z p Vo Canopy cover # cavities and boxes$ 3.35 + 0.31 68.76 1.83 t 0.15 71.84 3.52 <0.001* Vo contfer I5.ll + 1.72 89.57 12.4 + l.l7 88.8 2.13 0.03* DBHtrees >l0m(cm) 38.7 +O.99 20.45 39.5 +0.92 2L73 -2.03 0.04* # trees >l0m 50.22 + 3'47 53.22 53'95 + 3.31 58.24 -1.51 0' 13 %onativeriparian cavity spt 74.53 + 2.6 26.78 74.76 x'2.A 26.31 0.65 0'52 vo opeîunderstory 78'8 + 2.1 20.85 79.06 x.1.77 20.89 l.I2 0.26 o/o irnrervious surfacef 14.58 + 1.95 105.16 23.37 x.2.O3 82.03 -1.49 0.14 Conditional regression: (n=151): $ indicaûedthatthe variable was bg ta^nsfornæd. t indicates the arcsine transforrnation x indicates significance, p <0.05. lf = 0.33 of maximum possible 0.51. Likelihood ratio test = 59.9 on 8 dl p < 0.001.

I fi¡rther investigated the importance of conifers by conpiling a database of roost sites and then reducing the sarrple to reflect independent uses by deleting repeat entries of

s ites used rrnre than once per rrnnth (Tab le 3 .2.2). In winter and in the early part of breeding before trees are foliated (November - April), conifers constitúe 307o of roost sites (versus l4%o deciduous), but this drops to l4%o from May - October (versus 767o decidrnus). This difference is significarÍ" e =74.5 with Yates' continuity correction, df

= L, p < 0.001). Conifers and cavities combined constitute 70Vo of roost sites from

November - April suggesting their importance for efficient thermoregulation and concealment in northern regions. Roosting in or on buildings likely provides a similar therrrnregulatory advantage. Buildings were rarely used as roost sites in the suÍtmer

(only 5Vo of roost sites fromMay to October).

t34 Table 3.2.2. Roost site by period Conifer Cavity Building Deciduous Total

Mar - Apr 35 (45Vo) 25 (32Vo) 7 (97o) l0 (137o) 77 May - June 37 (l4Vo) IO (4Vo) t2 (4Vo) 210 (78Vo) 269 JuIy - Oct 5 (líVo) 4 (l2Vo) 4 (lZ%o) 20 (617o) 33 Total 86 (l9%o) 74 (167o) 4O (97o) 250 (56Vo) 450 Birds roosting in cavities were sometimes perched at the rnouth (sunning). Cavitþs and tree hollows (open to the floor of the hollow area) are not distinguished here.

The ground layer around nesting cavities wasT9Vo open (Table3.2.I), with few shrubs

and tall forbs or grasses. l¿wns conprised between zero and 787o of the area in a 50m

radirs around nest-sites, averaging 35Vo of thatarea(34.7 x.2.4Vo).

NEST-SITEFnerrrnrs

Nest-site specific variables showed less significance than habitat variables. Screech-owls

selected taller nest ftees (average 13.3 m on-territory versus I 1.4m off- territor!, z = 3.35,

df = 8, p < 0.01), lower shrub density at the base of the nest üee (107o versrs I6Vo, z -

- 1.98, p = 0.05), and chose trees closer to rivers or creeks even though the category

averages were exffemely similar (z= -1.2,p = 0.05) (Table 3.2.3).

Table 3.2.3. Averages + st¿ndard error of the mean of nest-site variables with coefficient of variation (CV) and output of the conditional regession model Used cavities Unused cavitþs X+SE CV i+SE CV z p Ileþhtof nest tree (m) ç Cavity tteight (m) 5.66-+- 0.4 53.26 5.5 + 0.27 46.08 -0.405 0.69 DBH nest tree (cm) $ 50.96 + 2.57 37 .19 54.87 + 2.69 45.25 -1.77 0.08 Diameúerentrance (cm) $ lO.7 + 0.42 29.n 10.54 + 0.21 18.0 -0.5 0.61 Dbtance to river (m) 255.3 + 69.1 211.42 254.8 x.35.9 166.57 -1.99 0.M7* Vo canopy cover 41.77 +2.96 52.5 42.25 x.2.31 50.63 0.I7 0.87 7o shrub dereitSrf 10.2+2.39 175.43 15.69 x.2.7 160.4 -1.98 0.048* Minimum flight path (m) 6.69 +O.74 81.1 5.71x.0.49 80.32 1.11 0.n Conditional regression: (n=143): $ indicated that the variable was bg transfornæd. t indicates the arcsine trarsformation * indicates significance, p <0.05. P =0.17 of maximum possible 0.5. Likelihood ratio test = 27 on 8 df, p < 0.001.

135 DNTURSaNCE exp PNMRTION RISr

This model evaluates variables related to the threat of predation and anthropogenic and natual disturbance (Table 3.2.4). Distance to building was not included in the disturbance rnodel because of colinearity with the number of buildings. The number of buildings is also correlated withpercentage impervious surface; hence they were not used in the same model. Two significant variables energed from this analysis. Firstly the domestic cat index averaged significantly lower at on-territory nest sites versus off- territory sites (0.2 versus 0.3 respectively, z= -2.5, df = 6, p = 0.01) suggesting that high domestic cat activity might be a deterrent in the nest-site selectionprocess. Secondly, the number of buildings wholly or partly within a radius of 50m around the nest-site averaged significantly lower at on-territory sites (4.5 versus 6.9, z= -2.2, p = 0.03).

Table 3.2.4. Averages + standard error of the mean of disturbance variables with coefficient of variation (CV) and output of the conditional regression model Used cavities Unused cavities X+SE CV X+SE CV z p

Cat index 0.2+0.M 131.95 0.34 + 0.03 85.21 -2.89 0.0ù1+ Raccoon index 0.13 + 0.03 193.27 0.14 + 0.03 t79.6 -1.21 0.22 # buildings 4.53 + 0.79 135.27 6.9 + 0.73 99.82 -2.55 0.01x Perrcentgrcenspæef 16.87 +2.38 110.1 17.55 + 1.98 lO7.l2 -1.11 0.27 Dbtance to road (m) 58 + 8.16 107.2 47.31 + 7.55 146.25 -0.44 0.66 Conditional regression: (n=151): t irdicates the arcsine transñrrnation t indicates significanca, p 10.05. F = 0.15 of rnaximum possible 0.51. Likelihood ratio test = 25 on 6df,p<0.001.

I fi¡ther investigated the potential effect of variables that might be related to predator avoidarrce using a simple classifying scheme of the success rate of all pairs (n = 44). Pairs tlrat had produced young for every year of occupancy were scored 1007o, pairs that produced young intwo of tlnee years of occupancy 677o, one of two years occupancy

50Vo,oneof three yearsoccræancy33Vo, and neverproduced youngOVo. Aregression

136 analysis (residual standard error= 29.33,df =22, Ê =0.42)of success as a functionof distance to building, shrub density, pedestian index, raccoon indet cat index, Great

Horned Owl index, cavity deptl¡ height of cavity, dianeter of cavity, number of buildings within 50rn, distance to water, minimum flight patt¡ and the number of cavities

+ boxes found that shrub density was significantly lower below more successful nesß (r =

-2.39,p = 0.03) and distance to building was significantly greater from strccessful nests (r

- 2.I7, p = 0.04). Nests w th I00 7o s uccess rate averag ed 55.62m from the neare st building whereas unsuccessful nests averaged 38.24m fromthe nearest building.

Differences between Suburban and Rural Nesß

Habitat features around nest sites differed considerably between high-density suburbarì, low-density suburban and rual nests (Table 3.3.1).

137 Table 3.3.1. Averages + standard error of the meanof habitat features around used and unused nest sites (plots of 50m radius) in three categories. High-dersity suburban Low-dersity suburban Rural Used Unused Used Unused Used Unused (n=15) (n=ã)) (n=38) (n=58) (n=8) (n=12)

# cavities 2.6 x.0.49 1.80 + 0.3 2.0 + 0.31 l.& + 0.2 3.5 + 0.76 1.75 + 0.37 # boxes 0.67 x.0.3 0.2 + 0.16 1.32 x.0.39 0.12 + 0.04 0.25 + 0.25 0.08 + 0.08 Cavities + boxes 3.27 + 0.51 2.0 + 0.33 3.32 x.0.43 1.76 + 0.19 335 x.0.62 1.83 + 0.34 # Trees>l0m 45.67 + 5.91 43.3 + 3.17 49.55 + 4.29 57.16 + 4.35 60.88 + 12.07 51.58 + ll.2l Tp10m DBH 39.82+ 1.7 q.72+ I.4I 37.42+ 1.35 39.2+ 1.15 39.63 +2.74 38.39 t 3.16 #Trees >5m 23.6 + 6.52 33.I + 4.28 33.37 +3.58 44.33 + 5.8 41.5 x.8.29 39.83 x.6.07 Tp5mDBH 14.76+ 0.90 13.91 +0.78 14.58 t 0.65 15.12+0.46 15.4+ l.4O 16.2+ 1.08 Vo C-onifer 18J7 + 3.83 17.16 + 2.52 14.96 ¡2.15 12.16 + l.4L 7.27 + 2.62 6.21 + 2.72 7o NRCT 73.78 x.3.93 72.83 + 3.45 75.99 + 3.03 76.83 + 2.36 68.51 + 11.48 64.95 + 8.53 %o elm 32.59x.4.93 31.16+3.1 20.78 +2.36 23.5x.2.38 23.49 +6.17 18.Mx.4.76 Vo maple 16.86 + 3.47 19.81 ¡2.56 15.8 *2.4 16.01 + 1.59 9.08 + 4.72 17.7I x. .TI Vo æh 7.96 + 1.63 9.M + 1.34 15.38 t 2.33 11.46 + 1.16 2O.06 + 4.3 13.44 x.4.24 7o od< 12.03 + 4.89 5.74 +2.47 18.74 x.3.48 19.0 +2.99 8.68 t 3.4 6.6 x.2.74 7o cotúonwood 3.34 + 1.26 4.53 + L5I 1.92 t 0.8 1.62 + 0.81 4.66 + 2.32 8.41 + 2.94 7o bassr¡ood 1.01 +0.58 2.55+0.87 3.36+2.01 5.24+1.81 2.53+1.36 0.75+0.54 7o Otber&c 7.45 + 2.38 10.0 + 1.83 9.05 t 1.9 11.02 + 1.56 24.21, + 12.05 28.84 + 8.85 7o opnagr 79.13 + 4.79 80.1I x.3.7 78.26 x.2.49 78.78 +2.2 78.0 + 6.66 77.75 + 4.75 7o imperviors 15.72+4.3233.95+4.25 15.32x.2.47 23.45+2.42 7.09+3.37 5.5+1.97 Abbreviations: ff = trees, DBH = dianeter at breast height, NRCT = native rþanan cavity tee (American elm, green ash, Manitoba maple, bur oak, basswood, and eastern cottonwood), other dec = other deciduous trees other than those in NRCT, open ugr - open understory (ground layer).

The MDA comparing the habitat variables from tsed cavities in three groups (high- dersity suburbar¡ low-density suburban" and rwal) identified one significant axis and one non-significant axis, viz. canonical I : canonical conelation = 0.7 ,likelihood ratio = 0.38,

F = I.65, df = 32,86, p - 0.04 and canonical 2: canonical correlation = 0.5, likelihood ratio = 0.75, F =0.97, df = 15, 44, p - 0.5. The Vo other deciduous trees variable (Table

3.3.1) and the 7o natíve rþarian cavity ffee, as well as the total number of cavities + boxes were rernf,ved to avoid producing an ill-conditioned matrix (number of boxes and

cavities were included separately). The MDA correctly classified 88Vo of rwalsites,T4Vo

138 of low-density suburbanandS0To of high-density suburban sites, suggesting süong habitat differences between these categories.

Table 3.3.2. Pooled within-class standardized canonical coefficients and total canonical snucture fromMDA for habitat variables Pooled Within-Class StandardizedCanonical Coefficienb TotalCanonicalStructure Canl Can2 Canl C^n2 Eo cartopy cover 0.23 -0.01 0.35 -0.22 # cavities 0.42 -0.0r 0.10 -0.51 # boxes 0.37 0.70 -0.11 0.43

# Trces>l0m $ -0.47 -0.17 o.25 -0.06 T>10m DBH -0.n -0.20 -0.04 -0.34 #Trees >5m 0.95 0.n 0.45 0.21 Tp5mDBH -0.2r -0.08 0.07 -0.12 Vo C-antfer I -0.76 0.01 -0.35 0.u2 Vo elml -1.00 -0.31 -0.33 -0.40 7o maplef -t.02 0.45 -0.29 0.20 7o æ}nf 0.29 0.31 0.38 0.10 Vo oak I -0.69 o.4l -0.02 0.45 7o cottonwood t 0.25 -o.22 0.08 -0.48 7o bassuoodt 0.03 0.n 0.15 0.14 7o openugr t 0.86 -0.03 -0.03 -0.09 7o imperviors T 0.29 o.45 -0.24 0.20 $ indicated that the variable was log tansforrrrd. t indicates the arcsine tansformation

The MDA for habitat variables found canonical axis 1 to be most sftongly correlated

(total canonical structure) with the number of ftees 5 - 10m (0.45), the 7o green ash (0.38)

¡he Vo canopy cover (0.35), and the number of frees >10m(0.25). It was negatively correlated wthVo conifer (-0.35), Vo Americ,anelm (-0.33), VoManítoba maple (-0.29),

7o inpervious surface (-0.24). Canonical axis 1 was also less süongly correlated with the numberof cavities (0.11) and negatively withnumberof nestboxes (-0.I2). Canonical I is therefore a gradient of denser canopy with rnore npdium- sized nees 5- 10m (and to a lesser extent tall trees >10m) with rnore green ash but less conifer, Manitoba maple and

139 Anerican elm (the latter are often planted in suburban areas as boulevard trees).

Anerican elm and Manitoba rnaple interestingly were the nee specbs with the rnost cavities (Table 3.1). In addition, canonical axis 1 is conelated with lower irrpervious surface area, rnore cavities and less nest boxes. Rural sites score highest on canonical variable 1 and high-density suburban sites lowest @ig 3.1).

Rural Low Suburban High Suburban Fig.3.1. Boxplots of canonical l scores for habitat variables.

Nest-site features around nest sites also differed between high-density suburbar¡ low- dersity suburban and rural nests (Table 3.4.1).

t40 Table 3.4.1. Averages + standard error of the mean of nest-site features around used and unused nest sites (plots of l2mradius) in three categories High-dersity Suburban Low-dernity Suburban Rural Used Unused Used Unused Used Unused (n=15) (n=Z)) (n=38) (n=58) (n=8) (n=12) DBH nesttree (cm) 55.67 + 4.55 59.26+ 3.06 50.97 +3.39 52.39 +3.07 4.75 +7.25 58.25 + 12.2 Cavity heisht (m) 5.87 r 0.65 5.81 t 0.79 5.37 + 0.52 5.13 + 0.28 5.87 + 1.36 6.38 + 0.76 H. rrest trce (m) 12.67 + 0.75 13.21+ 0.66 13.6 + 0.62 10.79 + 0.32 13.06 + 1.88 11.05 t 1.2 DmCavH (cm) 32.O7 + 1.73 37.21 + 2.0 37.77 + 2.4 36.05 t 1.1 33.5 + 2.26 37 .08 + 4.15 Diameúer(cm) 11.55 + 1.06 10.83 ¡.0.43 10.59 + 0.5 10.37 + 0.25 9.44 + 0.46 10.67 + 0.63 180.8 + 184.94 + 182.29 t 2TL52 + 741.44 + 755.83 x. Dbt. water (m) 53.9s 39.69 46.99 31.8 454.75 320.51 Relcanheþht (m) 13.77 + 0.71 15.6 + 1.68 13.77 + 0.55 13.06 + 1.39 14.03 x.l.Ø. 11.26 + 0.61 7o Canderse 47.05 + 5.64 45.86 x.4.75 37.77 + 3.57 38.03 t 2.0 52.53 + 9.62 55.52 + 4.M 7o slrrub dersity 8.12 + 2.84 9.94 + 2.9 8.62 x.2.81 14.39 + 3.24 19.54 + 10.54 28.Ø + 10.44 Minflpath(m) 7.9+0.96 5.76x.0.54 5.95+ 1.01 6.19+0.75 7.9+2.4 4.62x.0.74 Abbreviations: H. = height, DmCavH = dianeter of rurk or branch at the height of the cavity, Dist. water = distance to nearest river or creek, Relcanheight = relative canopy height, Eo candense - Eo canopy density above nest tree, and Minflpath = minimum flight pattr-

The MDA of the nest-site variables identified one significant axis and one non-significant

axis, viz. canonbal 1: canonical correlation = 0.58, likelihood ratio = 0.51, F = 1.J4, df =

20, 88, p = 0.04 and canonical?; canonical correlation = 0.47 ,likelihood ratio = 0.78, F

= I.43, df = 9, 45, p = 0.21. Nests-sites differed conparatively little between areas and

the MDA thus only conectly classified 75Vo of rwal sites, 707o low-density sububan

sites and 67Vo of high-density suburban sites.

141 Table 3.4.2. Pooled within-class st¿ndardized canonical coefficients and total canonical sffucture fromMDA for nest-site va¡iables Pooled Within-Class StandardizedCanonical Coefficienb Total CanonicalStructure Canl Canz Canl CarO DBH nesttrce (cm) -0.65 -1.11 -0.17 -0.31 Cavity height (m) 0.37 0.05 0.11 -0.t2 IL nest trce (m) -0.12 0.94 -0.10 0.20 DmCavH (cm) 0.10 0.63 -0.22 0.35 DÍameúer(cm) -0.31 -0.46 -0.19 -0.38 Dbt. waúer (m) $ 0.86 0.18 0.61 0.01 Relcanheþht (m) 0.43 -0.08 0.05 0.42 7o Canderse 0.77 -0.36 0.41 -0.23 7o shrub dersity $ 0.00 0.19 0.31 0.08 Minflpath (m) 0.52 -0.35 0.23 -o.25 $ indicated that the r¿ariable was log tansfornæd. t indicates the arcsine tansformation

In the analysis of nest-site features, canonical one was very strongly correlated to dist¿nce to water (0.61), canopy density directly above the nest tree (0.41), and shrub density directly below the nest tree (0.31). It is also less strongly correlated with minimum flight path (0.22) and negatively correlated with diameter at cavity height

(-0.22), DBH of nest tree (-0.17), and dianpter of the cavity's erúrance (-0.19). In addition to distance to water, canonical one represents a gradient of derser vegetation at the nest (canopy, middle story, and sltub layer) and decreasing width of the nest tree

(DBH and dianeter at cavity height) with a smaller enüance. Rural sites score highest on this axis (Fig.2), as they tend to show derser vegetation and slightly smaller ftees.

r42 I IT.--::-::t;i:.-__ r. "'.".:'d ¡ r ,::',';i{i:,:iiìÅ l''t: ,,,,,'iï¡Å __l_ +-

-4 Rural Low Suburban High Suburban Ftg.3.2. Box plot of canonical 1 scores for nest-site variables.

Predation risk, disturbance and anthropogenic featues around nest sites showed strong differences between high-density suburba4 low-density subrnban and nnal nests (Table

3.5.l).

143 Table 3.5.1. Averages + standard error of the mean of disturbance and anthropogenic featues around used and unused nest sites (plots of 12m radius) in tkee categories. High-dercity Suburban Low-dernity Suburban Rural On Otr On Otr On Otr (n=15) (n-Ð) (n=38) (n=58) (n=8) (n=12) Raccoon index 0.11 + 0.05 0.14 + 0.04 0.09 + 0.03 0.13 + 0.04 0.38 + 0.13 0.21 + 0.09 Cat index 0.32 + 0.09 0.43 t 0.07 0.18 + 0.04 0.31 + 0.04 0.08 + 0.05 0.37 + 0.09 GHOW index 0.03 + 0.02 0.04 + 0.02 0.09 + 0.03 0.11 + 0.02 0.21 + 0.05 0.19 + 0.04 Dist road (m) 51.45 + 12.7237.35 + 15.32 51.85 + 7.8634.29 + 4.7198.48 + Æ.84126.61 + 36.77 # buildings 6.67 + 2.45 lO.4O + l.7l 4.39 + 0.75 6.88 + 0.83 1.0 + 0.46 0.50 + 0.23 Dbt. bldg (m) 69.89 + 31.41 29.71 + 9.21 43.34 x.7 .73 31.99 + 3.&87.92 x.47.29 130.76 + 42.41 Vo greenspaoe 11.63+3.87 16.92x.4.83 14.14x.1.9113.79x.1.3839.65+ 11.63 36.8+9.14 Pedestrianindex 2.8+0.41 N/A 1.82x.0.19 N/A 1.03 +0.48 N/A Human dersity 44.63 + 4.69 49.02 + 4.6 22.94 + 0.8923.52 + 0.83 5.69 + 1.58 5.75 + 1.23 Human density was not included in the discriminant function analysis as it predefines the there categories. Abbreviations: Dist. road = distance to nearest road, dist. bldg = distance to nearest building.

The MDA of the disturbance variables identified one significant axis and one non- significant axis, viz. canonical 1 : canonical correlation = 0.57 ,likelihood ratio = 0.57 , F

= 2.4, df = 14, 104, p = 0.006 and canonic al2: canonbal correlation = 0.3 8, likelihood ratio = 0.86, F = 1.47, df = 6, 53, p - O.2I.The MDA correctly classified 887o of rural sites, 61 Vo of low-density suburban sites and 8O7o of high-density suburban sites (Table

3.s.2).

Table 3.5.2. Pooled within-class standardized canonical coefficients and total canonbal structure fromMDA for disturbance variables Pooled Within-Class StandardizedCanonical Coefficþnb TotalCanonicalStructure Canl Can2 Canl Can2 # buitdings 0.52 0.32 0.47 -0.08 Dist Bldg (m) $ 0.29 0.& -0.11 0.54 Dist Road (m) $ 0.23 0.28 -0.41 0.00 Raccoon index -0.31 0.83 -0.44 o.74 Cat index o.n 0.38 0.48 0.12 GHOW index -0.39 -0.06 -0.59 0.20 Pedestrian index 0.68 o.23 0.76 0.23 $ indicated that the variable was log ftarsforrrpd.

IM Canonical one was very süongly correlated to the pedestian index (0.76), the domestic cat index (0.48) and the number of buildings (0.47), and negatively correlated to the

Great Horned Owl index (-0.59), the raccoon index (-0.44), and disønce to road (-0.41).

This gradient typifies the urban - rural distinction with nrore people, buildings, and domestic cats but fewer natural predators (Great Horned Owl and raccoon).

Unsurprisingly, high-density suburban sites score highest on this axis and rual sites lowest (Fig. 3.3).

______o_

?a) rt) X

oCB É Êo (JCB

-4 Rural Low Suburban High Suburban Fig. 3.3. Box plot of canonical 1 scores for disturbance variables.

t45 Dsctssror.¡

Selection of Used versus Unrned Cavities

The preferred nesting habitat of screech-owls in the study area contains a rnoderately closed canopy (average 40Vo canopy cover), consistent with the range-wide preference for open areas in the subcanopy (Gehlbach 1995). Taller ftees were selected for nests; however, screech-owls did not select for cavities that were higher above the ground, which has been suggested as a predator avoidance strategy (Nilsson 1984, Rendell and

Robertson 1989, Belthoffand Ritchison 1990a), nor were their nests lower than unused cavities, which has been suggested as an energy saving Íreasure (Collias and Collias

1984). An open subcanopy thus appears important and may permit maneuverability around the nest and the opportunity to detect predators easily and attack them fromabove when necessary (Gehhach 1995). These owls are rnostly breeding in areas that can be described as "disturbed" and often close to buildings, roads and other antlnopogenic features but still require a minimal level of canopy cover. The lowest canopy closue of any successful nest was l6Vo and there were several cases where nesting territories were in areas with only a few tall trees and large open lawns.

Eastern Screech-Owls chose tenitories with an average of three potential nest-sites and avoid sites with fewer potential nest sites (Table 3.1). Alternate cavities are thus a significant habitat selection feature for this polytenitorial species (Gehlbach 1995).

Alternate cavities may serve as the site for renesting attempts (Gehlbach 1995) or occasionally for food-caching (pers. obs). Males also occasionally roosted at the entrance of alternate cavities when these were in line of sight with the nest cavity (pers. obs). In

r46 soÍre cases, screech-owls were observed alternatingbetweentwo similar nest-sites over a three-year period. Although some predators may increase their search intensity when

many cavities are present (Martin 1988), the switching of cavities rnay reduce the

likelihood of nest predation from predators that rely on mernory (Sonerud 1985a, 1989,

Gehhach L994b) or reduce the effect of ectoparasites (Aitken et al. 2002)'

The presence of coniferous trees around a nest-site rnay not be a significant factor over

rnost of the range of the Eastern Screech-Ow[ however, in Manitoba where there are no

native non-coniferous evergreen shrubs, conifers provide excellent roosting sites close to

the nests for both shelter and concealment especially early in the nesting season Male

screech-owls have been shown to roost close to their nests early in the breeding seasorl,

perhaps as an anti-predator behavior (Sproat 1997). Many paired males in this study used

conifers close to the nest as roost sites, especially in the early stages of nesting before

deciduors trees had fully foliated. It is not surprising therefore that the owls select nest

sites with an average of t5%o coniferous trees within a 50m radius of the nest (Table

3.2.I). Nonetheless there were five successful nests where no conifer was present within

50m of the nest, although in these cases conifer was present within 100m of the nest. The

presence of roost sites with suitable thernnregulatory properties (coniferous tees,

cavities, and buildings) in the early period of the breeding season therefore appears to be

an important habitat selection feature at very northern latitudes. Further south cavitbs

were found to be important winter roost sites in Virginia (Merson et al. 1983) and in

Texas for therrnoregulatory reasons (Gehhach 1994b). Conifers may also offer

concealment at other times of year, e.g. in Kentucky, 17 .4Vo of aduh and juvenile Eastern

r47 Screech-Owl roost sites in the post-fledging period were in conifers, when these family units were roosting on average 252m from their nest s ite (Belthoff and Ritchison 1990b).

The rnost unusu,al fìnding of this habitat model was that the DBH of trees >10m tall averaged significantly lower on-territory than off. A possible explanation is an effect created by the placennnt of nest boxes in areas with smaller trees that would increase the attactiveness of those areas to secondary-cavity nesters (Penins 1979, Tta;rlren et aL

1984), although nestboxes only constittúed34Vo of nest-sites (217o of total sample size).

Despite an effort to match used boxes with unused boxes and used cavities with unused cavities, boxes constitúed 387o of used nest sites versus only l l%o of unused sites. The

DBH of ftees on whbh boxes were placed was 409cnl16cm less than the DBH of trees with natwal cavities (56.8cm). The overall average of the DBH of all trees within a 50m radius of natwal cavities was 39.3cm versus 37 .6cm around nest-boxes. The latter small difference suggests that boxes had minimal influerrce on broad-scale habitat selection

Furthermore, inbothManitoba (Chapter 2) and Texas (Gehhach I994b), no significant differences in productivity were found betweenboxes and natural cavities, despite the fact that larger broods may occur inboxes than in natural cavities in other owl species

(Korpimåiki 1981).

The nest-site model suggests that screech-owls select cavities in øller nest ftees which are closer to water and with low shrub density around the base of the nest tee. This rnay imply less selection regarding the characteristbs of the cavity than the surrounding vegetation; however, this stems at least in part from the fact that variability of the off-

t48 territory cavities was reduced by only selecting cavities that conformed to the

characteristics of a potential nest-site in order to avoid biologically ftivial fìndings

(McCallum and Gehhach 1988). Cavity availability could influence cavity selectior¡ e.g.

low use of suboptimal cavities would imply high nest-site availability in proportion to

owl dernity (Severinghaus 2007) and the relatively low variance in certain features such

as cavity entrance size in this data set suggests that this may be the case in Winnipeg. In

2007, only 11 of the 55 (ZOVo) nest sites where breeding was confirned between 2004

and2}07 that were available (not clogged by squinels or within the territory of another

pair) were used. This percentage would be much lower if other cavities that appeared

suiøble but which were never found to be used by screech-owls in these four years were

included. The cavities in question were found vacant on at least three repeat visits and not

used by squirrels or Wood Ducks or other secondary-cavity nesters. These low

occupancy rates suggest an unsaturated populatiorL with low density in proportion to the number of cavities available, even if polyterritoriality is considered. This suggests that

cavities are not limiting in the study area, at least when supplemented with nest-boxes at the present rate. In addition, high nesting success (Chapter 2) añ apparent lack of corrpetition with non-native specbs such as Euopean Stårling are further suggestive of high cavity availability inproportionto owldensity (Blewett and Marzluff2005), although some colrpetition for nest-boxes with Wood Duck (Arr sponsa) was recorded

(Artuo 2007). Given the local climate and the cold tolerance of this species (Mosher and

Henny I976), winter food stpply or energetic demands are likely to be limiting factors in this region Several authors have argued that an abundance of a resource with prefened characteristics and unsaturated conditiors will result in geater selection for those

149 characteristics (Stephens and Krebs 1986, McCallum and Gehlbach 1988, Greene and

Stanps 200L, Hardy and Morrison 2001, Brown et aL.2002) which might explain why in this study I found npre selectior¡ at least at the habitat level around nests in Manitoba thanwas found inKentucky (Belthoffand Ritchison 1990a).

Screech-owls in the study area selected nest-sites with less shrubbery at the base of the nest tree and had higher nesting success where shrub density was low, as was the case in

Texas (Gehlbach 1994b). Although some shrubbery or thick vegetation on a territory might be important for certain behaviors such as hunting, roosting and territorial

'3inging" (Gehlbach 1995), a convolúed lower or middle story produces certain risks.

Shrubs below a nest can provide opportunities for ambush predators such as dorrpstic cats, ard high shrub density decreases the owls' ability to detect predators (Gehlbach

I994b). The selection of nest-sites with lower shrub dersity at the base of the nest tree may also be related to the owPs habit of approaching the nest by flying low to the ground and swooping upward (Gehlbach I994b). An open area. at the base of the nest tree permits such behavior and reduces the likelihood of detection (Gehhach 1994b). More generally, sparse shrubbery permits greater maneuverability and more open areas to hunt prey at ground level (Robbins et aL 1989, Gehhach 1995). Lkewise, a Íþre open middle story might not only improve the ability to detect potential predators (NiJsson 1984,

Belles-Isles and Picman 1986, Martin 1988, Martin and Roper 1988, Finch 1989, but see

Martin 1991 for the opposing view that dense vegetation improves nest concealment) but also provide flying room to swoop at predators should they approach too closely.

150 ln Texas, proximity to buildings was correlated with higher nesting success (Gehlbach

1994b).In Winnipeg, screech-owls selected nest-sites with fewer buildings within a radius of 50m and had higher nesting success further frombuildings. Human density was much higher in the Winnþeg study site, where 83Vo of nests occr¡rred in suburban areas with>lO plha" than in Gehlbach's (1994b) study site in Texas with <5 p/ha. A possible explanation therefore is a non- linear response to human activity: a certain proximity to buildings may offer advantages where human density is low, but there is a greater risk of disturbance in areas of higher human density. An alternative explanation is that screech- owls may avoid nest-sites that contain too high a proportion of non-habit¿t. Screech-owls will use houses as perches or roost sites but large areas of irrpervious swfaces offer poor- quality habitat. If nroderate proximity to buildings does provide an advantage, two possible mechanisms for this are increased prey accessibility and reduced predation

(Gehhach I994b). Sone important prey species such as Horse Sparrow (Passer domesticus) and house rnouse (Mus musculus) were routinely captured close to buildinp and screech-owls were also observed using window ledges as hunting perches and hunting near bird feeders (pers. obs). Screech-owls are known to use artificial lights and the light from wirdows when hunting (Getrlbach 1994). Cooper's Flar¡dcs also benefit from increased density of avian prey items correlated to housing density (Germaine et al.

1998, Mannan and Boal2000). Higher prey densities around horrps appear to benefit

Burrowing Owls in Florida; however, increasing developnrcnt lead to a reduction in nesting success (Millsap and Bear 2000). Eastern Screech-Owls, on the other hand, enjoyed both higher densities and higher reproductive success in high-density suburban areas (Chapter 2). Nesting in proximity to buildings may alter the predation risk or type

151 since as building dersity increased (also conelated with pedesfrian index and human

density), raccoon and Great Horned Owl indices decreased, whereas cat index increased

(Table 3.5.1). However, the relative importance of natural versus domestic predators is

not currently known There is general support for a decrease in the density of large predators with increasing human density (Crooks and Soulé 1999, Chace and V/alsh

2006, Chþman et al. 2008), and nesting near buildings to reduce predation has been

documented at least one species, the Eurasian Blackbird (Møller 1988).

In the study area, screech-owls exhbit a very strong preference for riparian habitat. Only

eight nest sites or territory centers (I3Vo) were located >500m from a river or creek.

Str-rdies using radio trarsmitters have also found strong habitat associations with running

water and wet woodlands and avoidance of dry rpland woods (Ellison 1980, Smith and

Gihert 1984). The majorityof screech-owl territories in the study area are inriparian

strþs strongly characteruedby American elm, Manitoba maple, green ash and basswood

(Moffat et aL.2004). These specbs are all relatively close to the northern limit of their

range in the study area (Hosie 1979), where they are associated with water courses,

especially the floodplains of larger rivers (Bid 1930, Scoggan 1957). Another cavity

free, eastern cottonwood, is rnost strongly associated with the shelf immediately adjacent

to rivers (Bfud 1930). Especially within the city of Winnipeg, where the extent of

seasonal flooding is partially contained, riparian strþs in the flood plain of the Red and

Assiniboine Rivers are rnost likely to rneet the habitat requirements of tall mature trees

with an open understory (sparse shrub cover) due at least in part to silt deposition from

seasonai flooding (pers. obs.). Orfside of Winnipeg in undisturbed areas, screech-owl are

152 also likely to hunt close to water because shrubs becorre denser and more diverse

towards the dry outer margins of rþarian forests (S taniforth 2003). Riparian areas also

have very high diversity and abundance of various taxa that are potential prey items for

screech-owls (Doyle 1990, Naiman and Décarrps 1997, l¡ck and Naiman 1998, Sanders

1998) and screech-owls in the study area frequently hunt along river edges for aquatic

prey as well as sone voles and sluews (Lynch and Smith 1984, Belthoff and Ritchison

1990a). Eastern Screech-Owls are not absolutely dependent on riparian areas in the study

site perhaps because planted shade ftees and lawns also create open-forest conditions and

bird feeders with fallen seed may sometimes produce prey concenftations.

The predator and disturbance model shows that screech-owls also select areas with low

domestic cat activity. In the suburban/urban parts of the study area domestb cats have

been observed attacking screech-owls when they cane to the ground to hunt (pers. obs.)

and even killing an entire brood of fledglings (G. Wala pers. comm.). Nonetheless, the relative impact of cats on screech-owls as conpared to natural predators such as Great

Horned Owls is unknown Raccoons may be the only nest predator in the study area although Anprican mink (Mustela víson) may be able to access some nests drning spring floods. Other known nest predators such as Virginia opossum (Dídelphis virginiana), ringøil (Bassariscus astutus),and black ratsnake (Elaphe obsoleta)(Gehlbach 1995) do not occur this far north The relative abundance of raccoons does not appear to influence screech-owl habitat selection directly. The Great Horned owl index was also non- significant; however, this finding may stem from the difficulty of assessing proportional use of any given area by Great Horned Owls. Great Horned Owls exhibit the opposite

153 distribution pattern to Eastern Screech-Owls in the study area, being detected rnost often on survey tansects in rural areas are much less often in suburban and urban areas

(Chapter 2). The habitat selection of Eastern Screech-Owls could therefore be influenced by Great Horned Owls, as has been shown for Tawny Owls, which avoid the larger

Eurasian Eagle Owl(Bubo bubo) (Sergio et aL.2007). More detailed data, e.g. radio telemety data on proportional habitat use, would be needed to address the impact of predatory owls on nest-site selectionby small owls.

Differences between Suburban and Rural Nesß

Suburban sites in the study area have higher densities of Eastern Screech-Owls than rural sÍes and this higher dersity is linked to rnore nesting attempts with higher breeding success (Chapter2). Discriminantfunctionanalysis (Fig.3.1, Fig.3.2, andFig.3.3) suggests a distinction between high-density suburban and rural sites br.t with low-density suburban sites being intermediate in many variables. There are several key differences between rwal and suburban sites including the number of ftees around the nest and canopy cover, the number of coniferous trees, the arnount of imperviotts surface, shrub density below the nest ftee, distance to water, the number of pedesfrians, the number of buildings, and the number of natural and domestic predators. Suburban sites also had npre American elm and Manitoba maple but less green ash than rural sites. This suggests that anthropogenic features and different levels of exposure to human activities are influencing the owls' habiøt selection and the predationregime they eryerience.

r54 Although screech-owls select for canopy cover (Table 3.2.I), the low average canopy closure on tenitories (407o) means that the reduction in canopy cover along the n¡ral - urban gradient is unlkely to deter screech-owls from occtpying high-density suburban areas. Screech-owl tenitories on rural sites had rnore tall úees >10mthan suburban sites, and, perhaps correspondingly, rnore natural cavities. Other studies have also noted a reduction in the number of cavitþs in suburban areas (GehhachI994b, Isaac et aL 2008) due perhaps to the removal of dead or dying üees or branches of mature trees or the plugging of cavities with cenent (Gehhach 1994b). Anthropogenic habitat aheration can negatively influence the breeding of owls in rnanaged forests (-õhmus 2003) presumably because human-altered habitats contain many fewer cavities than pristine ones

(Wesolows ki2007).In the study area however, suburban areas contained rrpre nest boxes and there were thus minimal differences in the overall number of potential nest sites.

Rrnal sites differed in having a slightly different tree composition and a denser middle story due to rncre tees in the 5 - 10m height range than suburban sites. These trees tend to produce a denser middle story than trees above 10m. The other deciduous frees variable had to be rernoved from the analysis became of heteroscedacity; however, rrnal sites also had the highestpercentages of tree species that typbally formdenser stands, such as trembling aspeq which creates a derser middle story becar-rse it gows in clones of sane-aged trees. Trembling aspen contained very few cavities in the study area, perhaps because several of the woodpecker species that regularly rnake holes in this

155 species, for example the Pileated Woodpecker, are rare (pers. obs). Rural sites also have less planted conifer, which the owls select (Table 3.2.I) and use as roost sites.

In the study area, tees in older suburbs are often larger than in the surrounding rtnal area

(pers. obs.) and rnore suitable to screech-owls both as. nest sites and in terms of general habitat preference. Two significant trends emerged: the DBH of cavity úees is negatively conelated with the distarrce to the city center (linear regression: t = -2.4'7, df = I4I, p =

0.01, P 0.04) or ahernatively the DBH of ftees above 10m is also negatively correlated to distance to city center (linear regression: t = -2.09, df = 149, p = O.03, rf = 0.03). This finding is unexpected since urban tees often have slower growth rates in proximity to impervious surfaces and with increased soil corrpaction (Quigley2004).It is possible that the suburban nest trees in this data set are older and hence have wider girth or that other factors such as the reduced competition for light in the canopy (Quigley 2004) or even the urban heat island in a northern latitude may balance other factors that hinder gowttr- The urban heat island has also been documented to advance flowering and leafing phenologies (White et aL 2002, Neil and Wu 2006) and earlier foliated fees would be beneficial to nesting owls.

Rural sites are charucterized by rnore Great Horned Owls and raccoons suggesting greater predationpressure. Other studies have also found that small owls face nnre predators in rural areas as opposed to suburban areas (Gehhach1994, Chipman et aL 2008). Although the donestic cat index is lower in rural areas, the conditional regression analysis suggests that screech-owls avoid sites with a large number of cats (Table 3.2.3), which could

156 minimize the negative influence of cats in suburban and urban areas. In rural areas screech-owls nests were closer to buildings (average 88m) than unused cavities (average

13lm from a building). Both high and low-dersity suburban nests averaged closer to buitdings thanrural nests; however, the opposite pattern occurred with nest-sites much fuither from buildings than unused cavities (Table 3.5.1). Furthermore, nest-sites were closer to buildings in low-density suburban areas than in high-dersity suburban areas.

One possible explanation is that cavities were nnre likely to be close to buildings in low

density sr-rburbs that usually had larger yards, whereas the houses in higher denser

suburbs were often closer together with space for ftees being rnore limited. Proximity to

buildings might therefore offer a greater advantage to screech-owls in rural areas for prey

concenftations or predator avoidance, since Great Horned Owls and raccoons are nþre

corrunon in rural areas (Chapter},Table 3.5.1). A geater need for predation avoidance in

rural areas is also suggested by the dianpter of nest entrances since rural pairs chose

cavities with the smallest enfrance (Table 3.4.I), a defense against nest predators

(Sonerud 1985b, Belthoff and Ritchison 1990a).

Even at the northern limit of its geogaphical range, the Eastern Screech-Owl exhibits

patterns of habitat selection that are similar to rrcre southerly populatiors. When nest

sites are compared with unused sites, Eastern Screech-Owls selected geater canopy

closure; rnore potential nest cavities, also found in Texas (Gehlbach 1994a); rnore

coniferous nees within 50mof the nest; lower DBH of tall trees around the nest-site;

taller nest trees; nest sites closer to running water, similar to Texas (Gehhach 1994) arñ

habitat use in Massachusetts (Ellison 1980); lower shrub density below the nest, also

r57 important in Texas (Gehhach L994) and selected for by Flammulated Owls in New

Mexico (McCallum and Gehlbach 1988); less domestic cat activity; and fewer buildings within 50m of the nest. The degree of selection for vegetation features was greater than in

Kentucky (Belthoff and Ritchison 1990a), perhaps related to sanple size. The greatest difference found for this northerly population was the selection for coniferous trees close to nest-sites for concealment and therrnoregulation, which does not appear necessary in rnost parts of the range, confers often being avoided in habitat selection (Ellison 1980,

Smith and Gilbert 1984) with the possible exception of southern pine forests (Gehlbach

1995). The preference for rþarian habitats and nest-sites within 500mof a river or perrnanently flowing creek is highly pronounced in Manitoba. Though described as a habitat generalist, the preference for rþarian woods appears more general throughout the range of the Eastern Screech-Owl (Gehhach 1995) and was also noted in New York

(Kelso 1944), and Massachusetts (Ellison 1980) but may be less pronounced in rnore rnoderate climates, e.g. upland woods were used in greater proportion than their availability in Connectbut (Smith and Gilbert 1984).

Linked with geater breeding success in subr¡rban areas, suburban rpst sites differed from rural nests in having fewer frees, especially fewer ftees less than 10m tall and correspondingly lower canopy closure (more open middle story and subcanopy), less green ash fewer raccoons and less Great Horned Owl activity but rnore coniferous trees, rrpre American elm and Manitoba maple (the species with the rrnst cavities), more buildings and a larger area of impervious surfaces, closer to running water and roads, lower shrub density below the nest tree, and rnore pedestrian and donpstb cat activity. In

158 suburban areas, the mature but nnre open vegetation facilitates hunting and predator detection planted conifers create suitable roost sites close to nests before deciduous trees have begun foliating, the lower density of natual predators despite the increased dersity of domestic cats appears to allow increased nesting success (Chapter 2), the urban heat island enables earlier breeding, which may be linked to higher survivorshþ (Chapter 2)' and prey accessibility and density is increased (Chapter 4). Several anthropogenic features influence the habitat selection process such as the preference for planted conifer in proximity to a nest site and the avoidance of domestic cats and nest-sites surrounded by manybuildings.

Lrrnn¡runnCITED

Aitkeq K. E. H., and K. Martin. 2004. Nest cavity availability and selection in aspen-

conifer groves in a grassland landscape. CanadianJournal of Forest Research

34:2099-2109.

AitkenK. E, and K. Martin 2008. Resource selectionplasticity and community

resporìses to experimental reduction of a critical resource. Ecology 89:971-980.

Aitken, K. E. H., K. L. Wiebe, and K. Martin. 2002. Nest-site reuse patterns for a cavity-

nesring bird community in interior British Columbia. Aìrk 119:391402.

Artuso C.2007 . Eastern Screech-owl hatches Wood Duck egp. Wilson Journal of

Omithology I 19 : 1 10-112.

Bartholomew, J. Editor. 1985. The Times atlas of the world. Corrprehersive edition

Times and John Bartholomew and Sons, I-ondorL U.K.

r59 Belles-Isles, J., and J. Picrnan 1986. Nesting losses and nest site preferences in House

Wrens. Condor 8 8 :483-486.

Belthoff, J. R., and G. Ritchison 1990a. Nest-site Selectionby EasternScreech-Owls in

Central Kentucky. Condor 92 :982-990.

Belthoff, J. R., and G. Ritchison 1990b. Roosting behavior of postfledging Eastern

Screech-Owls. Auk 107 :567 -57 9.

Bird, R. D. 1930. Biotic communities of the aspen parkland of cenftal Canada. Ecology Il:35644I.

Blewett, C. M., and J. M. MarzlufT. 2005. Effects of urban sprawl on snags and the

abundance and productivity of cavity-nesting birds. Condor 101:678493.

Bloom, P. H., and M. D. McCrary.1996. The urban Buteo: Red-shouldered Hawks in

southernCalifornia. Pages 3143 ínD. M. Bird, D. E. Varland, and J. J. Negro,

editors. Raptors in human landscapes: adaptations to built and cultivated

environnents. Academic Press, I-ondon, U.K.

Bosakowski, T., and D. G. Smith.1997. Distribution and species rbhness of a forest

raptor community in relation to urbanization Journal of Raptor Research 3l:26-33.

Brawn, J. D., and R. P. Balda. 1988. Populationbiology of cavity nesters in northern

Arizona: do nest sites limit breeding densities? Condor 90:6lJL.

Brown, D. R., C. M. Strong, and P. C. Stouffer.2002. Demographic Effects of Habitat

Selectionby Hermit Thnushes V/intering in a Pine Plantation Landscape. The

Journal o f W ildlife Mana geme nt 66 :401 416.

Brown, J. H., D. W. Mehtmar¡ and G. C. Stevens. 1995. Spatial variation in abundance.

Ecology 76:20281043.

160 Bull, E. L, C. G. Parks, and T. R. Torgerset 1997 . Trees and logs important to wildlife

in the interior Columbia River basin. USDA Forest Servbe General Technical

Report PNW-GTR-391.

Bult, T. P., S. C. Riley, R. L. Haedrich, R. J. Gibson and J. Heggenes. 1999. Dersity-

dependent habitat selectionby juvenile Atlantic salmon (Salmo salar) tn

experimental riverine habitats. CanadianJournal of Fisheries and Aquatic Sciences

56:1298-1306.

Carlsort 4., U. Sandstron¡ and K. Olsson 1998. Availability and use of natural ftee

holes by cavity nesting birds in a Swedish deciduous forest. Ardea 86:109-119.

Chiang, S. N., P. H. Bloom, and S. E. Thomas. 2006. Horre range and habitat use of

Cooper's Flau¡k's in urban and natual areas ofCalifornia. FourthNorth Anærican

Omithological Conference, Veracruz, Mexico, October 4J ,2006.

Chþman, E. D., N. E. Mclntyre, R.E. Strauss, M. C. Wallace, J. D. Ray, and C. W. Boal.

2008. Effects of Human l¿nd Use on Western Burrowing Owl Foraging and

Activity Budgets. Journalof Raptor Research 42:87-98.

Collias, N. E., and E. C. Collias. 1984. Nest building and bird behavior. Princeton

University Press, Princetor¡ New Jersey, U.S.A.

Conner, R. N. 1975. Orientationof enftances to woodpecker nest cavities. Auk 92:371.

374.

Crockeff, 4.B., and H. H. Hadow. 1975. Nest site selectionbyWilliannonand Red-

naped sapsuckers. Condor 7 9 :365-368.

DeGraff, R. M., and J. M. Wentworth 1986. Avian guild sftucture and habitat

associatiors in suburban bird communities. Urban Ecology 9:399412.

161 Devereux, J. G., and J. A. Mosher. 1984. Breeding ecology of Baned Owls in the cental

Appalachiars. Raptor Research 1 8 :49-58.

Diermen, J. 1996. Sparrowhawk (Accipiter nisus) breeding in village, rtnal landscape and

forest. De lævende Matuur 97:43-51.

Doyle, A. T. 1990. Use of riparian and upland habitats by small mammals. Journalof

MammalogyTl:IÇ23.

Dykstra, C. R., J. L. Hays, F. B. Daniel, and M. M. Simon 2000. Nest site selection and

productivity of suburban Red-shouhered Hawks in southern Ohio. Condor 103 401-

408.

Dykstra, C. R., J. L. Hays, F. B. Daniel, and M. M. Simon 2001. Home range and habitat

use of suburban Red-shouldered Hawks in southwestern Ohio. Wilson Bulletin

1 13:308-316.

Ellison P. T. 1980. Habitat use by resident Screech Owls (Otus asio). Thesis, University

of Massachusetts, Amhers t, Ma ssachusetts, U. S . A.

Femandez-Juricic, E. 2001. Density-dependent habitat selection of corridors in a

fragrrcnted landscape. Ibis 143 :27 8{,87 .

Fraterrigo, J. M., and J. A. Wiens. 2005. Bird communities of the Colorado Rocky

Mountains along a gadbnt of exurban development. I-andscape and Urban

Planning 7l:263-275.

Finch, D. M. 1989. Relationshþs of surrounding riparian habitat to nest-box we and

reproductive oúcome in House Wrens. Condor 91:848-859.

Forsman, E. D., and A. R. Giese. 1997. Nests of Northern Spotted Owls onthe Olympic

Penins ula, Washington Wilson Bulletin L09 :284L

t62 Fuller, R. J., K. J. GastorL and C. P. Quine. 2007. Living on the edge: British and Irish

woodland birds in a European context. Ibis 149:53-63.

Galeotti, P. 1990. Territorial behavior and habitat selection in an urban population of the

Tawny Owl,Srr¿¡ aluco L. Bolletino diZnologia 57:5946.

Goad, M. S., and R. W. Mannan 1987. Nest site selectionby Elf Owls in Saguaro

Natio nal Mo nument, Arizot:r,. C o ndor 89 :659462.

Gehhach F. R. 1994a. Nest-box and natural-cavity nests of the Eastern Screech-Owl: an

exploratory study. Journal of Raptor Research 28:154-157 .

Gehhach F. R. 1994b. The Eastern Screech-Owl: Life history, ecology, and behavior in

the suburbs and counüyside. Texas A&M University Press, College StatiorL U.S.

Gehhach F. R. 1995. Eastern Screech-Owl Account 165 in A. Poole and F. Gill, editors.

The Academy of Nattnal Sciences, Philadeþhia, Pennsylvania and The Arrerican

Omitho b gists' Unio n, Washingto n, D.C., U. S. A.

Germaine, S. S., S. S. Roserstock, R. E. Schweinsburg, and W. S. Richardson 1998.

Relatio ns hips among breed ing b irds, hab itat, and res idential de ve lop nent in greater

TucsorL Arizona. Eco lo gical App lications I :680{9 I .

Greene, C. M., and J. A. Stanps. 200I. Habitat Selection at Low Population Densities.

Ecology 82:2091-2100.

Gutzwiller K. J., and S. H. Anderson 1987. Multiscale associations between cavity-

nestingbirds and features ofWyoming sftearrside woodlands. Condor 89:534-548.

Hardy, P. C., and M. L. Morrison 200I. Nest site selection by Elf Owls in the Sonoran

Desert. Wilson Bulletin II3 :23-32.

r63 Holt, R. F., and K. Martin 1997 . I-andscape nndification and patch selection: the

denngaphy of nvo secondary cavity nesters colonizing clearcuts. Auk I 14:443-

455.

Hooge, P. N., M. T. Stanback, and W. D. Koenig. 1999. Nest-site selection inthe acorn

woodpecker. Ar:k 116:45-54.

Horak, G. C. 1986. Cumulative effects of rapid urbanizationonwinter aviandiversity in

northeastern Colorado. Thesis, Colorado State University. Fort Collins, Colorado,

U.S.A.

Hosie, R. C. 1979. Native trees of Canada. Fitzhenry and Whiteside Ltd, DonMills,

Ontario, Canada.

Hull Sieg, C., and D. M. Becker. 1990. Nest-site habitat selected by Merlins in

southeastern Montana. The Condor 92:688{94.

Ingold, D. J. 1989. Nesting phenology and competition for nest sites arrnng Red-headed

and Red-bellied V/oodpeckers and Ewopean Starlings. The Auk 106:209117.

Ingold, D. J.1994.Influence of nest-site conpetitionbetweenEuropeanStarlings and

woodpeckers. Wilson B ulletin 106 :227 14L. lngold, D. J. 1996. Delayed nesting decreased reproductive success in NorthernFlickers:

implicatiors for corrpetition with European Starlings. Journal of Field Ornithology

67:321-326. lnouye, R. S., N. J. Huntley, and D. W. Inouye. 1981. Non-randomorbntationof Gila

Woodpecker nest enftances in saguaro cacti Condor 83:88-89.

164 Isaac, 8., R. Cooke, D. Simmons, and F. Hogan 2008. Predictive rmpping of powerful

owl(Ninox strenua) breeding sites using Geographical Information Systems (GIS)

in urban Melbourne, Autralia. I-andscape and Urban Planning 84:212-218.

Johrsorl D. H. 1980. The conparison of usage and availability measrnements for

evaluating reso urce preference : Eco lo gy 6I :65 -7 I .

Jones, J.20Ol. Habitat selection studies in avian ecology: A critical review. Auk

ll8:557-562.

Kelso, L. H. 1944. Behaviour of the eastern screech owl (Otus asío naevius). Biological

I-e,aflet23, Cornell University, New York, U.S.A.

Korpimfüi E. 1984. Cluûch size ard breeding success of Tengmalm's Owl Aegolius

funereus in natual cavities and nest-boxes. Omis Fennica 61:80-83.

Korpim¿iki 8.1987. Selection for nest-hole shift and tactics ofbreeding dispersal in

Tengma lm' s owl Ae g o lius fune reus. Jo unal o f Animal Eco lo gy 56 : 1 85-l 96.

I-eruroq J. J.2000. Red-shifu and red herrings in geographicalecology. Ecogaphy

23:I0l-1I3.

I¡ck, P. 4., and R. J. Naiman 1998. Effects of stream size on bird community structure

in coastal temperate forests of the Pacific Northwest, U.S.A. Journal of

B io geo grap hy 25 :7 7 31 82.

Iõhmus, A.2003. Do Ural owls (,Srrlx uralensis) suffer from the lack of nest sites in

mana ged forests ? B io lo gica I Co nservatio n 1 1 0 :1 -9.

Lundberg, A.1979. Residency, migrationand a compromise: Adaptations to nest-site

scarcity and food specìalization in three Fennoscandian owl species. Oecologia

4l:273-281.

165 LynchP. J., and D. G. Smith. 1984. Census of EasternScreech-Owß (Otus asio) n urban

open-space areas using tape-recorded song. American Birds 38:388-391.

Mãnd, R., V. Tilgar, A. lõhmus, and A. læivits. 2005. Providing nest boxes for hole-

nesting birds - Does habitat matter? Biodiversity and Corservation I4:L823-1840.

Manna4 R W., andC. W. Boal 2000. Home range characteristics of male Cooper's

hawks in an urban environrrent. Wilson Bulletin Ll2:2I17.

Mappes, T., J. Mappes, and J. Kotiaho. 1994. Frtoparasites, nest site choice and breeding

success in the Pbd Flycatcher. Oecologia 98:147-L49.

Marks, J. S. 1986. Nests-site characteristics and reproductive success of l-ong-eared Owls

in sorfhwestern ldaho. Wilson Bulletin 98:547-560.

Martin, T. E. 1988. On the advantages of being different: nest predation and coexistence

of bird specbs. Proceedings of the Natural Academy of Science 85:2196--2199.

Martin, T. 8., and J. J. Roper. 1988. Nest predation and nest site selection of a western

populationof the Hermit Thrush. Condor 90:5I-57.

Martin, T.8.1992. Breeding productivity considerations: what are the appropriate habitat

featwes for management? Pages 455473 inJ.M. Haganand D. W. Johnstoru

editors. Ecology and conservation of Neotropical migrant land birds. Smithsonian

lnstitution Press, Washington, D.C., U.S.A.

Marzluff J. M., R. Bowmar¡ and R. Donnelly. 2001. A historical perspective on urban

bird research: trends, terns, and approaches. Pages I-17 in J. M Marzluff, R.

BowmarL and R, Donnelly, editors. Avian ecology and conservation in an

urbanizing world: Kluwer Academic, Bostorl Massachusetts, U.S.A.

t66 McCallurn, D. 4., and F. R. Gehhach 1988. Nest-site preferences of Flammulated Owls

in westernNew Mexico. Condor 90:653461.

Merson, M. H., L. D. [€ta, and R. E. Byers. 1983. Observatiors on Roosting Sites of

Screech-Owls. Journal of Field Ornithology 54:4t942I.

Mikkola, H. 1983. Owls of Europe. T & A D Poyser, Caltor¡ Staffordshire, U.K.

Millsap, B. 4., and C. Bear. 2000. Dersity and reproduction of Burrowing Owls along an

urban d eve lop ment grad ient. Jo urnal o f V/ild life Mana geme nt 64 :33 41.

Moffatt, S. F., S. M. Mcl-achlan, and N. C. KenkeL 2004. Irrpacts of land use on riparian

forest along an urban - rural gradient in soúhern Manitoba. Plant Ecology 174:lI9-I35.

Møller, A. P. 1988. Nest predation and nest site choice in passerine birds in habitat

patches of different size: a study of maglies and blackbirds. Okos 53:215121.

Mosher, J. 4., and C. J. Henny. 1976. Therrnal adaptiveness of plumage color in Screech

Owls. Auk 93:6t44I9.

Mossop, D. H. 1997. The importance of old-gowth refugia in the Yukon boreal forest to

cavity-nesting owls. Pages 584-586 in J. R. Duncan, D. H. Johrson, and T. H.

Nicholls, editors. Biology and conservationof owls of the NorthernHemisphere.

North Cental Research Station, U.S. Departnpnt of Agriculture, St Paul,

Minnesota, U.S.A.

Naiman, R. J., and H. Décamps.1997. The ecology of interfaces: rþarian zones. Annual

Review of Ecology and Systematics 28:621458.

Natuhara, Y., and C. Imai 1996. Spatial structure of avifauna along urban-rural

gradients. Ecological Research 11 :1-9.

167 Neil, K., and J. Wu 2006 Effects of urbaniza tion on plant flowering phenology: A

review. Urban Ecosystems 9 : 243157 .

Newton, I. I994a. Experiments on the limitation of bird breeding densities: a review. Ibis

136:397411.

Newtor¡ I.1994b. The role of nest sites in limiting the numbers of hole-nesting birds: A

review. B io lo gical Conservatio n 7 0 :26517 6.

Nilsson, S. G. 1984. The evolution of nest-site selection among hole-nesting birds: the

importance of nest predation and conpetition Ornis Scandinavica 15:165-115.

Perrins, C. M. 1979. British tits. Collins, I-ondon, U.K.

Quigley, M. F. 2004. Street ftees and rural corspecifics: Will long- lived tees reach fulI

size in urban conditions? Urban Ecosysterrs 7:29-39.

Qutnr¡ G. P., and M. J. Keougtr- 2002. Experimental design and data analysis for

biologists. Cambridge University Press, Cambridge, U.K.

Rendell, W. 8., and R. J. Robertson 1989. Nest-site characteristics, reproductive success

and cavity availability for Tree Swallows breeding in natural cavities. Condor

91:875-885.

Rendell, W. 8., and R. J. Robertsott 1994. Defense of extra nest-sites by a cavity nesting

bird, the Tree Swallow (Tachycineta bicolor). Ardea 82:273185.

Robbirs, C. S., D. K. Dawson, and B. A. Dowell. 1989. Habitat area requirements of

breeding forestbirds of the Middle-Atlantic states. V/ildlife Monographs L03:l-34.

Rottenborr¡ S. C. 1999. Predicting the impacts of urbanizationonriparianbird

co mmunitie s. B io lo gica I C o nservat io n 88 289 -?;99 .

168 Sanders, T. A. 1998. Breeding bird community conposition in relation to rþarian

vegetation structure in the western United States. Journal of Wildlife Manage

62:461473.

Sasaki A. 1997 . Clumped distribution by neighborhood conpetition Journal of

Theoretical B io lo gy 186 :4L5430.

Scoggan, H. J.1957. Flora of Manitoba. National Museum of Canada, Biological Series

No. 47, Bulletin No. 40. Ottawa, Canada.

Sedgwick J. 4., and F. L. Knopf. 1990. Habitat relationshþs and nest site characteristics

of cavity-nesting birds in cottonwood floodplains. Journal of Wildlife Management

54:lL2-L24.

Semel, 8., P. W. Sherman, and S. M. Byers. 1988. Effects of brood parasitism and nest-

box placenent onWood Duck breeding ecology. Condor 90:920-930.

Septor¡ G. 4., J. Bielefeldt, T. Ellestad, J. B. Marks, and R. N. Rosenfield. 1996.

Peregrine falcons: power plant nest structures and shoreline nþvemerìts. Pages

145-153 inD.M. Bird, D. E. Varland, and J. J. Negro, editors. Raptors in Hurnan

Landscapes: Adaptations to Built and Cultivated Environments. Academic Press,

New York, U.S.A.

Sergio, F., L. Marchesi P. Pedrini, and V. Penteriani. 2007. Coexistence of a generalist

owl with its innaguild predator: distance-sensitive or habitat-npdiated avoidarrce?

Animal Behavio ur 7 4 :1607 -16 16.

Severinghaus, L. L.2007. Cavity dynamics and breeding success of the l-anyu Scops

Owl (Ot us e Ie g ans). Jo urnal of Ornitho lo gy 1 48 :5407 -5416.

r69 Smitlu D. G., and Gilbert, R. 1984. Eastern Screech-Owl home range and use of suburban

habitas in Connecticut. Jornnal of Field Ornithology 55:322-329.

Smittu D. G., T. Bosakowski and A. Devine. 1999. Nest site selection by urban and rural

Great Horned Owls in the northeast: Journal of Fþld Ornithology 70:535-542.

Smith-Castro, J. R. 2008. Inpacts of recreational trails onbreeding birds in forested

urbanparks. Thesis, Ohio State University, Columbus, Ohio, U.S.A.

So[ D., D. M. Santos, E. Feria, and J. Clavell. 1997 . Habitat Selection by the Mork

Parakeet during colonization of a new area in Spain. Condor 99:3946.

Sonerud, G. A. 1985a. Nest hole shift in Tengmalm's Owl" Aegolíus funereLts, as defence

against nest predation involving long-term mefiþry in the predator. Journal of

Animal Ecology 54: 179-192.

Sonerud, G. A. 1985b. Risk of nest predation in three species of hole nesting owls:

influence on choice of nesting habitåt and incubationbehaviour. Ornis

S cand inavic a I 6 :26 L--269 .

Sonerud, G. A. 1989. Reduced predationbypine ffartens on nests of Tengmalm's Owl in

relocated boxes. Animal Behaviour 37:332134.

Southern, H. N. 1970. The natural control of a population of Tawny Owls (.Srrzx aluco).

Journal of Zoolo gy, l-ondon 162:197 -285.

Sparks, E. J. 1990. Home range and habitat utilizationby Eastern Screech-Owls in central

Kentucky. Thesis, Eastern Kentucky University, Richmond, Kentucky, U.S.A.

Sproat, T. M. 1997 .Male Eastern Screech-owl (Otus asio) roosting behavior: Possible

effects from nesting stâge and nest type. Pages 408410 ,i? J. R. Duncar¡ D. H.

Johrson, and T. H. Nicholls, editors. Biology and corservation of owls of the

t70 Northern Hemisphere. North Cenftal Research Station, U.S. Departrrent of

Agiculture, St Paul, Minnesota, U.S.A.

Sraniforrb R. J. 2003. Ecosystems and bird habitat in Manitoba. Pages 34-55 ín

. Manitoba Avian Research Committee, P. Taylor, editor-in-chief. The Birds of

Manitoba. Manitoba Naturalists Society, Manitoba Avian Research Committee,

Winnþeg, Manitoba.

Statistics Canada. 2004. Profile of Census Tracts in Winnþe g'2001Census. Catalogue

No. 95-24l-)(PB. Reftieved March2004 from:

http:l/datacenfte.chass.utoronto.ca.proxyl.lib.umanitoba.calcensus/index.htrnl

Stauffer, D. F., and L. B. Best. 1982. Nest-site selectionby cavity-nesting birds of

rþarian habitats in Iowa. Wilson Bulletin 94:329137.

'W., Stephens, D. and J. R. Krebs. 1986. Foraging theory. Princeton University Ptess,

Princeton, New Jersey, U.S.A.

Sunde P., K. Overskaug, J. P. Bolstad, and I. J. Oien 2001. Living at the limit: Ecology

and behaviour of Tawny Owls Strix aluco in a northern edge population in cental

Norway. Ardea 89 495-508.

Taylor, I. R. 2002. Occupancy inrelation to site quality in BarnOwls (Tyto alb¡a) n south

Scotland. Pages 30-41 ¿rl L Newtor¡ R. Kavanagh" J. Olsen, and I. R. Taylor,

editors. Ecology and Cornervation of Owls. CSIRO Publishing, Collingwood,

Victoria, Austalia.

Tenple, S. A. 1988. Futt¡re goals and needs for the manageÍEnt and conservation of the

peregrine falcon Pages 843-848 in T. J. Cade, J. H. Enderson, C. G. Thelander,

17l and C. M. White, editors. Peregrine FalconPopulations: Their Managenrnt and

Recovery. The Peregine Fund Inc., Boise, Idaho, U.S.A.

Tiainen J., P. Saurola, and T. Solonen 1984. Nest disribr.tion of the Pied Flycatcher

Ficedula hypoleuca in an area saturated with nest-boxes. Annales Zoologici

Fennici 2l]99104. van Balen, J. H., C. J. Booy, J. A. Van Franeker, and E. R. Osieck. 1982. Studies of hole-

nesting birds in natural nest sites. Ardea 7O:l-24.

V/alley, W. J., and C. F. Clyde. 1996. Occurrence and breeding of the Eastern Screech-

Owl north of the Riding Mountairs, Manitoba. Blue Jay 54:89-100.

Wang, Y., S. Chen, and P. Jiang. 2008. Black-billed Magpies (Píca pica) adjmt nest

characteristics to adapt to urbanization in Hangzhor¡ China. Canadian Journal of

Tnology 86:676484.

Wesolowsk \ T. 2007 . læssons from long-term hole nester studies in a primeval

tenperate forest. Journal of Ornithology 148:5395-405.

Whire, M. A, R. R. Nemani P. E. Thornton, and S. W. Running.2002. Satellite evidence

of phenological differences between urbanized and rtnal areas of the eastern United

States deciduous broadleaf forest. Ecosystems 5260-273.

Wiebe, K. L. 2001. Microclimate of free cavity nests: is it important for reproductive

success in Northern Flickers? Auk 1 I8:4I242I.

Yetter A. P., S. P. Havera, and C. S. Hine. 1999. Natural-cavity use by nesting Wood

Ducks in Illinois. Jor¡nal of Wildlife Management 63:630-638.

7Ar, J. H. I974. Biostatistical analysis. Prentice-Hall, Englewood Cliffs, New Jersey,

U.S.A.

r72 4. The diet of the Eastern Screech-Owl at the northern periphery of its range

Ansrn¡,cr.-A total of 2323 prey items of Eastern Screech-Owl (Megascops asio) were analyzed over a four-year period in Winnipeg, Manitoba at the northern limit of species' range, using four techniques: pellet analysis, post-fledging nest inspectio4 video nnnitoring, and direct observation Invertebrates corrprised6TVo of prey captures but only l IVo ofbionnss corìsumed. Seasonal shiffs in diet between biologically significant periods were significant such that multiple discriminant analyses correctly classified 6OVo of pre-breeding season, 9I7o of incubatiorU 67 Vo ofbrooding, 86Vo post-fTedging and l00Vo of winter corsumption Rural pairs consurrpd a higher percentage of invertebrates and fewer vertebrates than sr-rburban pairs. Among avian prey, srnall passerines (<209), migrants and winter visitors, foliage feeders and omnivores were selected. Only 37 vertebrate prey species were recorded and the total niche breadth was 6.5, reflective of a much higher degee of dbt specialization than more southerly populations.

Keywords: Eastern Screech-Owl Megascops asio maxwelliae, Winnþeg, Manitoba, diet.

INirRouucrIoN

The diet of the Eastern Screech-Owl (Megascops asio) is corrparatively well studied (e.g.

Allen 1924,YanCanp and Henny 1975, Turner ard Dimmick 1981, Ritchisonand

Cavanagh 1992, Gehhachl994); however most studies have utilized only one or two methods such as pellet analysis, identification of prey rernains in nest boxes, and stomach contents. This is problematic because these different rrpthods lead to different conclusions about the relative importance of different prey types, in particular invertebrate versus vertebrate prey (Ritchison and Cavanagh 1992). Of the 11 studies included in Ritchison and Cavanagh's (1992) review, only thee examined stomach contents and only one (Allen 1924) contained an observational corrponent. Pellet sarrples of Eastern Screech-Owl have typically underreported invertebrates (Errington

t73 1932, Cruighead and Craighead 1956) and have shown the percentage of mammals

cornumed to be very high (e.g. Cahn and Kerrp 1930, Wilson 1938, Craighead and

Craighead 1956, Korschgenand Stuart 1972) whereas sone studies of nestboxcontents

also underreported invertebrates (e.g. VanCamp and Henny I975) and suggested a higher

proportion of birds (e.g. Stewart 1969, Duly T979) in the diet. Studies based on the

examination of stomach contents found high predation on invertebrates such as

arthropods, especially in the sumrner months (Fisher 1893, Hanebrink et al1979, Brown

1989). Duly (1979) found stomach contents contained 93Vo trwertebrates but cached prey

only 8Vo invertebrates. Furtherrncre, most, if not all of these studbs underreported soft-

bodied invertebrates such as earthworms which leave no remains in pellets and which are

rarely cached (Ritchison and Cavanagh1992, C. Artuso, pers. obs.). This study therefore

aims to give a more comprehensive porftait of the diet of a northern population of Eastern

Screech-Owl by using a combination of two observational techniques (direct observation

and video recordinp) and two non-observational techniques (pellet analysis and nest-box

inspection).

In the Wiruripeg area, the Eastern Screech-Owl resides at various points along the nlal - urban gradient and reaches is highest densities in suburban areas (Chapter 2).

Urbanization can greatly influence the diet of birds, sonetimes offering either increased or Íþre varied resources (Botelho and Arrowood 1996, Diermen 1996) and sometimes causing profound changeS in dietary regimes (Robbirs 1993, Wolf and del Rio 2000,

Faeth et al. 2005). Such resources may be accompanied by increased risks such as pollutiorL toxins or diseases (Getz et aL 1977 , Kostelecka-Myrcha et al. 1997 , Jensen et

174 aL 2002) and may even result in ecological traps or population sinks (Rerrrcs 2000, Battin

2004,Baker et aL 2005). I therefore examined differences in the diet of nesting pairs of

screech-owl in rural, low-density suburban, and high-density subr-uban areas.

Mn'ruons

The study area was defined as a circle with a radius of 40km from the center of

Winnþeg, at the northern perþhery of the range of Eastern Screech-Owl, which errcompasses the city limits and also areas outside the city with lower human densities.

Rþarian areas are dominated by Annrban elm (Ulmus americana), green ash(Fraxinus pennsylvanica), and Manitoba maple (Acer negundo), with some bur oak (Quercus macrocarpa), eastern cottonwood (Populus deltoides), basswood (Tilin americana) and peachleaf willow (Salrx amygdaloides) also present. Trembling aspen (Populus tremuloi.des) and bur oak are often the dominant species in non-rþarian areas. Many species are also planted in garders and parks including white spruce (Picea glauca), eastern white cedar (Thuja occidentalis) and the non-native Colorado blue spruce (Picea pungens) and Scots pne (Pinus sylvestris).

Each screech-owl territory was classified as belonging to one of three categories pertaining to human dersity: high-density suburban (>30 people per hectare), low-density suburban (>10 - 30 people per hectare), and rural areas (<10 people per hectare). This follows a random-stratified survey of the study area (Chapter 2) but with the wildlands and rural categories r¡sed in that survey rrrrged under the n¡ral category here and with the mediunrdersity suburban category and low-density suburban category merged here and labe led low- de rs ity s ub urb an for co nve nie nce.

115 A total of 2323 prey items were identified between March 2004 and February 2008 from

föur sowces: field observatiors (n = 789) of nesting pairs and tenitorialbirds,

occasionally with the aid of flash photography, video footage from irnide a single nest box over two nesting seasons (n= 225), analysis of pellets collected under roost sites (n =

837), and the inspection of nest boxes after all young had fledged (n= 472) (Table 1). A total of 637 pellets ranging in size from 8x4mm to 51x12mm (but sometimes as much as

2}mmwide in part due to flattening) and averaging 25.4x10.5mm. were analyzed. The

sarrple from the nest with a video camera was supplerrented with direct observational

data and pellet data and care was taken to cross reference all prey iterns such that itens

observed in the field or video were not double counted in pellets or nest box remains. I

also conpared percentages of avian, mammalian and invertebrate prey from the video

recordings with those of nearby nests with direct observational data only to ensure that no

biases arose fromthis single sarrple. If a prey item was first observed being captured or

eaten and item rernains were later recovered in a pellet or nest box it was fteated as

observed once for the purposes of this study. The number of prey items from

observations and identified from remains were simila r,l0I4 (44Vo)and 1309 (56Vo)

respectively, yielding a reasonably unbiased database. However, the ratio of observations

to pellets was not constånt at all sites and therefore care is needed when drawing

corrclusions about comparative capture rates of individuals or pairs.

Mammal and bird remairs were identified by consulting the collection at the Manitoba

Mweum and suitable reference material (Bansfield I974,Elbroch and Marks 2001, Kays

and Wilson 2002, U.S. National Fish and Wildlife Forensics l:boratory, Wageningen

fl6 University Experimental Tnology Group). Invertebrates were identified to family or gemrs levelby examining head capsules, elyüa and legs with the assistance of Dr. Robert

Roughleyand Dr. TerryGalloway(Universityof Manitoba, Entornology). Dr. Richard

Westwood (Universityof Winnipeg, Biology) assisted inthe identificationof moths to family level fromphotographs. The number of beetles consumed was cabulated from head capsules. Feather rernains in nest boxes were always assurrpd to represent only one individual of each species identified unless the total number of remiges or rectrices exceeded the number found on one bird. Single feathers of Wood Duck (Au sponsa),

Northern Flicker (Colaptes auratus) and European Starling (Sturnus vulgaris) without other remains found in nest boxes were not assuned to be prey items as these may have been shed by these birds visiting the boxes prior to occupancy by the owls. The mass of prey allbirds was taken fromSibley (2000), and mammals as the average of the two values (range) given by Kays and \Wilson (2002). The average weight of invertebrate species was calculated from weighing similar individuals found in the study area. When a prey item was identified to genus, family, or group only, the average weight of all known rrembers of that group in the sample was calculated and assigned to unidentified items.

The two broad periods of breeding and non-breeding roughly followed Ritchison and

Cavanagh (1992) with an adjustment for the northern location of this study such that the breeding season was treated roughly as April through September or, where knowr¡ more precisely calculated for individual pairs as the period extending from onset of incubation until 10 weeks after the fledging date of the oldest chick, which corresponds to typical

t77 natal dispersal (Gehhach 1995). The non-breeding season was thus October through

March Prey consumption was also calculated by significant biologbal period for

individu,al pairs based on back calculation from fledging dates (Chapter 2). The nnst typical dates for these periods were: pre-egg laying (Mar 1 - Mar 31), incubating (April I

- May 8), brooding (May 9 - July 3), post-fledging (July 4 - Sept 30). TtE "summer" period refers to iterns collected from nest boxes post-fledging for which the exact time of

captìire is not known

I calculated the percentage consumptionof mammals, birds, amphibiars and fish and

invertebrates per neslyear as well as subdividing bird species consumed into resident and

locally breeding species versus winter visitors and passage migrants, mammals into rodents and non-rodents, and invertebrates into hard-bodied (detectable inpellets) versus

soft-bodied (only recorded from observations). Niche breadth and overlap were

calculated following Gehhach (1994) for the four prey classes @irds, mammals,

anphibians and fislu and invertebrates) to ail comparison with three other significant

studbs (VanCamp and Henny 1975, Turner and Dimmick 1981, Gehlbach L994;

summarized in Gehlbachl99í, Table 1) as well as for individual prey species. It is nonetheless important to note that the rrethodologies of those studbs, whbh relied on the

inspection of prey remains in nest boxes, are not directly corrparable to this study with

multiple data sources. I also calculated the total number of prey species identified at each

site. For this calculation invertebrates were tallied at the family level only and items

identified only to genus or group were only included if no other npmber of that group had been tallied as a species. Because differences in sample size prevented direct

178 comparison, I divided all species counts by the square root of the sample size to produce a sirrple prey diversity index.

ln addition to overall calculations by human category and biological period, I calculated the sane percentages, diversity index and niche breadth for each site; however, I teated each nesting attenpt as independent (i.e. I did not pool data across years). A second data set was obtained by calculating these same percentages at each site/year by significant biological period. These data were arnlyzed with a Chi-sqtrare test (Ritchison and

Cavanagh 1992) and multiple discriminant analysis (MDA) for significant differences between biological periods and between high-density suburbar¡ low-density suburbaq and rural sites. I examined box plots and log üarsforned diversity index and arcsine

üarsfornpd percentage variables as necessaryto reduce heteroscedacitybefore performing the MDA.

In order to examine prey selection, I also conducted a total of 58 five-minute fixed distance (100m) avian point counts on all known Eastern Screech-Owl territorbs in the study area in the frst four days of May 2005, and again on the sane sites in the flrst four days of June 2005. The resuhs of these points were then reduced to only those species or genera that were recorded in the diet of the owls and used to calculate the proportion of those species in the environnent. To determine their prey value, avian prey species were grouped in six ways: passerine versus non-passerine;resident and locally breeding species versus passage migrants and winter visitors; small (<20 grams), medium (20-a0g) and large (>40g) species; predominantly arboreal or sluub level feeders versus

t79 predominantly ground feeders; predominantly ganivorous, predominantly in:sectivorous

and omnivorous (mixture of insects, fruit and/or nectar) species; and species tolerating only minimal anthropogenic disturbance, moderate anthropogenic disturbance and high

anthropogenic disturbance. I then used Jacob's (1974) sebctivþ inde¿ which gives a

score between-1 (abundant in environment but not consured) to 1 (uncomrrnn in the environrrpnt but often consunnd), to examine preferences. Selectivity indices of non-

avian taxa were not estimatable.

Table 4.L. Prey of Eastern Screech-Owl from 2004 - 2008: sorrces of the data Bid Mammal Amphibian & Fbh Inverúebraúe Total Obserryatiom Field observations 15 24 4 746 789 Video 62 30 4 t29 225 Total observations 77 54 8 875 LOIA (447o) Remains Pellet Analysis 22r 367 0 249 837 Nest inspection 46 8 0 418 472 Totalremains 267 375 0 667 13ú (56Vo\ TotaI 3M 429 1542 2323

Rnsur,rs

A total of eight mammal species were recorded as prey in the study area. Meadow vole

(Microtus pennsylvanicas) was by far the most common mammalian prey (,637o of

mammalianprey and 757o of mammalianbiomass). Voles constituted 68Vo of all

mammalprey and SlVo of the mammalianbiomass consumed. Meadow vole capture rates

ranged from9%o (rural) - lZ%o (low and high-density suburban) of all prey captures in the

breeding season and from ZVo (rwal) - 20Vo (low-density suburban) - 457o (high-density

suburban) in the non-breeding season House ÍÐuse (Mus musculus) (I3Vo) and North

Arrerican deerrnouse (Peromyscus manículatus) (6Vo) were also regular prey items,

whereas little brown bat (Myotis lucifugus) (IVo) and northern short-tailed shrew (Blarina

180 brevicauda) (IEo) were consurred less frequently. House mouse was absent from rural

sites and ranged from 3 Vo (tow-density suburban) and LVo (high-density suburban) in the

breeding season to îVo(low-dersity suburban) arñ.I}Vo(high-density suburban) in the

non-breeding season

A totalof 26 bird species were recorded as prey; however, this figure is undoubtedly

higher since many bird remains could not be identified to species. Birds corsumed were

largelypasserine (82Vo) and the majoritywere small species under 40g(55Vo <20g,34Vo

20 - 40g, IIVo >409), contrary to Gehlbach(I994). The most commonly recorded avian

prey species were House Sparrow (Passer domesticus) (l3%o of all birds) and Black-

capped Chickadee (Poecile atricapillus) (9Eo). House Sparrow was much less common in high-dersity suburban and rural areas (<1 Vo) than in low-density suburban sites. Yellow- rumped Warbler (Dendroica coronata) (37o) was the npst commonly captured species that does not breed in the study area, except perhaps for a few isolated boreal pockets

(Holland et aL 2003) such as in Bird's Hill Provirrcial Park. The largest avian prey

captured were Rock Pigeons (Columba livia) andBlue Jays (Cyanocitta crßtata),as well

as a single Virginia Ratl (RaIIus limicola), a species seldom recorded within the city of

V/innþeg (pers. obs.). The only trvo amphibianprey items confrmed to species were northern leopard ftog(Lithobates pipiens) and wood frog(Rana sylvatica) and only a

single unidentified fish was found in this sanple.

Fifty-seven percent of invertebrates consumed were detected using observational techniques, corrpared to 22Vo for birds and 13 Vo for mammals. Invertebrates cornumed

181 included insects, earthworms, crustaceans, arachnids, gasftopods, and many unidentified

small invertebrates. The bulk of invertebrates consumed were insects and of these beetles

were most commonly recorded (43Vo of invertebrates), presumably because their remains

are rrìore readily found in pellets and nest boxes. The family Scarabidae constituted 8370

of the beetles identified, of which many appeared to be in the genus Phyllophaga. After beetles, the next nnst important invertebrate prey were both soft-bodied, the earthworm

Lumbricus terrestris (5Vo) and variou caterpillars (4Vo). Earthworms were most

comrrnnly captured in low-density suburban areas (IVo of aIl rural prey in the breeding

seasorì, 5Vo low sr.rburbar¡ 3Vo high suburban). The proportion of invertebrate prey captrned by three unpaired males (X + SE: 27.3 + l4.2Vo) was much lower than for breeding pairs (X + SE: 75.7 x.5.6Vo). Unpaired males caught mammals (56 t- 20.97o) nmre frequently thanbreeding patls (12.2*3.8%) MANOVA' Wilk's Lan:ùa:0.6, F =

6.38, df = 2,I9, p = 0.008); however, the data on the diets of unpaired males were collected primarily frompellets found at roost sites and are therefore biased due to a lack of ob servatio nal info rmation

Despite being unavailable to the owls in winter in the study area, invertebrates corstitute by farthe largestproportionof totalpreycaptures(667o overall, TIVo nthe breeding season, 20Vo nthe non-breeding season) (Table 4.2.1). Mammals are the next highest item corsuræd (I\Vo overall, 167o dtxng breeding arñ 40Vo in the non-breeding season) but only slightly higher than birds (I5Vo overall, IZVo during breeding and 40Vo in the non-breeding season). Anphibians and fish, which are also unavailable to the owls in winter in study area, were a very small corrponent of prey consurred (<77o overall). lhe

r82 The number of birds, mammals, amphibians and fish and invertebrates consumed (Table

.z.|)varied significantlybybiologicalperiod (t =854.2, df = 15, p < 0.0001).

Mamrnals were captured more thanbirds overall; however, birds and nrammals were consumed in equal proportions in the non-breeding season (Mammals captwed: I6Vo breeding, 407o non-breeding, 50Vo wtnter;birds captued: IZVIbreedng,40Vo non- breeding 50Vo wtnter only).

Table 4.2.I. Prey of Eastern Screech-Owl from 2004 - 2008: prey captures by period Period Bird Mammal Amphibian and fish Invertebrate Total Winter 22 (50Vo) 22 (50Vo) 0 OM Pre-egg laying 68 (38Vo) 66 (37Vo) 0 44 (25Vo) 178 Incubating 82 (t9%o) 249 Ø9Vo) 3 (r%o) l3I (3IVo) 425 Brooding 124 (l1%o) ll0 (líVo) 5 (l%o) 476(67Vo) 715 Post-fledging 25 (5Vo) 18 (3Vo) 0 491(92Vo) 534 Summer 23 (5Vo) 4 (IVo) 0 4N(94%o) 427 Total 344 (157o) 429 (t9Vo) 8 (

Despite the high percentages of invertebrates captued, invertebrates constitúed only llVo of the biomass consumed overall (l3%obreedtng,ZVo non-breeding) (Table 4.2.2).

Although the capture rate of mammals was only slightly higher than birds, they contibuted rrpre to biomass consumed (587o versus 3I7o overall, 577o versvs 29Vo breeding 597o versus 40Vo non-breeding). Although the ratio of vertebrates to invertebrates did not vary significantly by year (t = 0.36, df = 3, p = 0.95), there is annual variation in the type of invertebrates consumed, e.g. earthworrns were highest in the wettest year (I77o of invertebrate captures in 2004 when precþit¿tion from March - September totaled 531mm down to zero :r:.2006 whenMarch - September precipiøtion was only 23Imm) (Environnnnt Canada, nd).

183 The average rnass (+ standard error of the mean) of prey overall was 13.2 + 0.4E,lighter than2S.3greported by Ritchisonand Cavanagh (1992). Categoryaverages were birds

(27.9 x.1.69), mammals (41.3 + 0.69), amphibians and fish (I4.9 x.1.3g), and invertebrates (2.1 + 0.01g).

Table 4.2.2. Prey of Eastern Screech-Owl from2004 - 2008: biomass consurrrcd by period Period Bird Mammal Amphibian and fish Invertebraúe Total Winter 588e(40%o) 8789(607o) 0g 0g r466e Pre-egglaying l7l0g(397o) 25cÐg(59%o) 0g 9lg(2%o) 4399e Incubating 26899(237o) 86909 (74Vo) 349(

The MDA comparison of prey type by period (not including the unspecified surrìmer period) identified two significant axes, viz canonical 1 : canonical correlation = 0.91, likelihood ratio = 0.03, F =3.57, df = 48, I13.75,p <0.001 and canonical2: canonical correlation=0.S4,likelihoodratio=0.17,F=2.21,df=33,89.09,p=0.001 (Table

4.2.3). The MDA conectly classified 60Vo of the pre-breedingperiod, 9lVo of the incnbation period, 67 Vo of the brooding period, 86Vo of the post- fledging period, and

IOOTo of the winter period.

t84 Table 4.2.3. Pooled within-class standardized canonical coefficients and total canonical snuctwe from MDA for prey type differences between biologically significant periods Pooled Within-Class Standardized Canonical Coefficþnb Total Canonical Structure CanL Canz Canl Carâ

Birù T 0.34 2.24 0.78 -0.13 Resident birds T -0.36 -0.32 0.33 -0.19 Mþant birds T -0.08 -0.04 0.28 0.13 Mammals t -0.47 4.07 0.85 0.31 Rodents t 0.07 -0.31 0.54 0.23 Non rodents t 0.54 0.18 0.28 0.39 Amphibian/ fish 0.26 0.90 0.03 0.31 Invertebraúes t -t.62 3.85 -0.96 -0.02 Hard bodbd t 0.14 0.30 0.19 0.60 Soft bodied t 0.24 0.53 -0.51 -0.05 Divenity $ -0.68 -0.41 0.47 0.26 Niche breadth -0.06 -0.38 0.54 0.47 $ indicated that the variable was log tansfornred. t indicates ttre arcsirp tansformation

The frst two canonical axes from the MDA (Fig. a.1.) show stong separationof the diet in biologically significant periods primarily by the distribution of the prey classes and diversity. Canonical axis I is highly correlated to percent mammals (0.85, total canonical süuctue) and birds (0.78) and negatively correlated to percent invertebrates (-0.96).

Within the prey classes, this axis was also correlated strongly to rodents (0.54) and negatively to soft-bodþd invertebrates (-0.51). Niche breadth (0.54) and prey diversity

(0.47) are also important. Canonical axis one therefore represents a gradient of increasing vertebrate consumption and decreasing invertebrate corsunption (especially soft-bodied invertebrates) and the winter and pre-breeding periods showed the highest scores (Fig.

4.1). Canonical axis 2 was rnost strongly correlated with hard-bodied invertebrates (0.6), niche breadth (0.47), non-rodents (0.39), arrphibians and fish (0.31), mammals (0.31), and diversity (0.26).It was negativelycorrelated to birds (-0.13), inparticular resident birds C0.19). Canonical axis 2 is thus a gradbnt of increasing mammal consumption

185 including increasing diversity of mammals (addition of non rodents such as bats and

shrews to the diet) and higher niche breadth but slightly lower bird consurrption in

particular fewer migratory species. Accordingly the incr-rbation and brooding periods

scored highest onthis axis whereas the winter and pre-breeding periods scored lowest

(Fig.4.1)

o €lo o o

ôt rt) x0

ßl ()-r oÉ 11 6l r\ -2

-t0l2S4 Cauonical Axis I p¡s-b¡sgdi¡g }ts*x e+olncubatinE +++ Brooding *-ç"+post-breeding ¡È--;-rå Winter

Fig. 4.1. Scatter plot of canonbal I andT scores from MDA for prey types differerrces between b io lo gically s ignificant periods.

The percentages of birds, mammals, amphibians and fish and invertebrates consumed overall varied significantly between rural, low-density suburban and high-density suburban sit€s (?(2 = 1':- .2, df = 6,p = 0.008 ) (Tab le 4.3.1). Owls on rural s ites cons unpd

186 fewer birds thanowls in low and high-density suburban sites inboththe breeding and non-breeding season (9Vo fewer than low-density suburban owls in the breeding season

arñ'3o%o fewer in the non-breeding season). The ratio of resident and locally breeding species to passage migrant and non-breeding visitors was approximately 2:I onall sites

(Table 4.3.I). Owls in low-density suburban sites consumed almost double the arnount of birds in the non-breeding season as in high-density sites and triple that of birds in rural sites. Owls in nnal areas consumed fewer mammals in the breeding season than either goup of suburban owls. However, owls in high-density suburban areas consuned more mammals in the non-breeding season Mammalian diversity in the diet was greater on low-density suburban sites, there being no slrews or bats found in the diet anywhere else.

Owls on rural sites corsurrpd 167o nnre invertebrates than low-density suburban owls,

20Vo n:o e than high-density suburban owls in the breeding seasorL andTlTo and 437o

ûþre respectively in the non-breeding season (ï3 = 27 .6, df = 2,p < 0.001). Rual owls also consumed a higher proportion of hard-bodied invertebrates and fewer earthworrns and caterpillars. Prey species diversity was highest at low-density suburban sites in both the breeding and non-breeding season Differences in sample size do not rrnke sites directly corrparable; however, the adjusted diversity count (divided by the square root of the sanple size) produced the sarrp ratio (ruaf low-density suburbaq high-density suburban: 1.4, 1.8, 1.0 in the breeding seasonand 1.6, 2.0,11in the non-breeding season). Niche breadth was highest at high-dersity suburban sites in the breeding season; however, it was highest at low-density suburban sites overall. Niche breadth was higher in non-breeding season than in the breeding season for both rural and low-density suburban sites but not at high-density suburban sites (Table 4.3.I).

t87 Table 4.3.L. Percentage prey disnibution, diversity and niche breadth in breeding (Br) and non-breeding (Non-br) seasons hy human deruity category Rural Suburban - low Suburban - high Br Non-br Br Non-br Br Non-br Total sample size 490 7 998 186 613 29 Birù 43 14.3 12.9 44.1 17.0 24.1 Resident birds 66.7 100.0 63.2 70.3 65.9 83.3 Mþant birds 33.3 0.0 36.8 29.7 34.r t6.7 Mammals 10.6 42.9 18.0 33.9 17.8 75.9 Rodents 100.0 100.0 92.6 96.7 100.0 100.0 Non rodents 0.0 0.0 5.1 J.J 0.0 0.0 Amphibians and fish 0.4 0.0 03 0.0 0J 0.0 Invertebrates 84.7 42.9 68.7 22.0 64.6 0.0 Hard bodþd 94.9 0.0 66.1 85.0 65.5 0.0 Soft bodied 4.8 100.0 33.3 15.0 33.r 0.0 Average prey sp/site 14.o 4.0 t5.2 10.5 11.3 7.0 Niche brcadth - clæs t.4 2.6 t.9 2.8 2.r 1.6 Niche breadth -sæcþs 2.2 4.5 6.r t2.4 4.9 3.8 Percentages of birds, mammals, amphibiarn and fislu and invertebrates are calculated against the total number of prey items for each of the three hurnan density categorbs. Other percentåges are of against the total number of birds, mammals, or invertebrates that could be identified to species or genus level and were thus classifiable into subgroups.

The MDA comparison of prey tlpe among ruraf low-density suburban, and high-density

suburban sites identified one significant axis and one non-significant axis, viz. canonical

1: canonbal correlation= 0.94,likelihood ratio = 0.05, F = 2.58, df = 24, 18, p - 0.02

and canonicalZ: canonical conelation = 0.76, likelihood ratio = 0.42, F = L.26, df = 11,

I0, p - 0.36. The MDA conectly classified IOOVo of rtnal sites, 1007o of low-dersity

srù urban s ites and I00 Vo o f high- dens ity s ub urban s ites.

r88 Table 4.3.2. Pooled within-class standardized canonical coefficients and tot¿l canonical sftucture from MDA for prey type differences among nnal, low-density suburba4 and high- dens ity sub urban s ites Pooled Within-Class Standardize d Canonical Coefficienb Total Canonical Structure Canl CarO Canl Carû Birù I -0.37 1. l8 0.25 0.6 Resident birds t 2.O8 0.37 0.11 -0.15 Mþant birds T 1.48 -0.08 0.41 0.09 Mammals t 2.36 0.81 0.5 -0.05 Rodents t r.82 0.11 0.5 0.06 Non rodents t t.o2 0.43 0.14 0.38 Amphibian/ fish 0.20 -0.57 0.23 -0.2 Invertebraúes t 0.82 t.57 -0.46 -0.03 Hard bodied t -1.r7 -0.09 0.31 0.34 Soft bodied t 0.98 t.07 -0.12 0.23 Diversity $ 3.51 0.11 0.3 0.58 Niche brcadth -3.83 0.50 0.18 0.5 $ irdicated that the variable was log trarsfornpd. t indicaûes the arcsine tansformation

The flrst canonbal axis from the MDA (Table 4.3.z,Fig. 4.2.) shows strong separation of the diet between rural, low-density suburban, and high-density suburban sites. Canonical axis 1 is highly correlated to the percentage of mammals consuned (0.5), inparticular rodents (0.5); the percentage of birds consuned (0.25), in particular passage migrants and winter visitors (0.a1);and amphibians and fish (0.23). It is also conelated withprey diversity (0.3). This axis is strongly negatively correlated with invertebrate consurrption

(-0.46). Canonical axis 1 is therefore a gradient of decreasing invertebrate consumption and conespondingly increasing vertebrate corsumption (as percentages). Since the diversity of invertebrates was only measured to the family level, the higher diversity score on this axis is in nnstly explained by diversity of vertebrates consurrcd. Rural sites scored highest onthis axis, ie. high invertebrate consunptionand low vertebrate consumption, and high-dersity sububan sites the lowest (Fig. a.Ð.

189 .t) <0x oat, É 3-z (JÉË

Rural Low Suburban High Suburban Filg.4.2. Box plot of canonical 1 scores from MDA for prey types differences between rurat low-density suburban and high-density suburban sites.

Jacob's (1974) selectivity indbes were calculated for avøngrorrys only (Iable 4.4). In all areas, passerines were captured in greater proportion than their availability, whereas non-passerines had negative selectivity indices. Passage migrants and winter visitors had positive selectivity indices, whereas locally breeding species had negative indices. Srnall bird less ttp'n}} graÍB were selected for in all areas whereas nedium- sized birds 20 -

40g had negative selectivity indices everywhere except high-density suburban areas and large birds above 40g also had negative selectivity indices everywhere except rural areas.

Arboreal feeders were generally selected over ground feeders and insectivores over granivores in rnost cases. Species with a high tolerance to anthropogenic distwbance

(urban-adapted species such as European Starling and House Sparrow) had positive

190 selectivity indices except in rual areas, whereas sububan-adapted species with moderate

tolerance levels, many of which breed in the study area such as American Robin (Turdus

mígratorius) and Cedar Waxwing (Bombycitla cedrorum),had negative selectivity

indices. Species with lower tolerance to anthropogenic disturbance, a few of which breed

in large suburban parks but which are generally uncommon in the urban environnpnt

except on passage, e.g. Rose-breasted Grosbeak (Pheucticus ludovicianus) and Baltimore

Oriole (Icterus galbula), had positive selectivity indices in low-density suburban areas.

rable 4.4. Selectivity indices for various guilds of birds at rural (R), low-density suburban (S-l), and high-density suburban (S-h) sites n ervironnnnt (p) G) R S-I S-h R S-I S-h AII s-h All 'zr r9Z t l0 '323 (5) (23) (8) (36) (4) (37) (r7) (58) Taxonomy Pass 0.9 1.0 1.0 1.0 0.9 0.9 0.9 0.9 0.5 0.1 0.7 0.4 Non-p 0.1 0.0 0.0 0.0 0.1 0.1 0.1 0.1 -0.5 -0.1 -0.7 -0.4 Breeding statw Res O.7 0.8 0.7 0.8 0.9 0.9 0.9 0.9 -0.6 -0.1 -0.6 -0.3 Mig 0.3 0.2 0.3 0.2 0.1 0.1 0.1 0.1 0.6 0.1 0.6 0.3

Size Small 0.4 0.5 0.6 0.6 0.2 0.4 0.4 0.4 0.5 0.3 0.4 0.3 Med 0.3 0.3 0.3 0.3 0.5 0.4 0.3 0.4 -0.5 -0.1 0.2 -0.1 Large 0.3 0.1 0.0 0.1 0.0 0.2 0.3 0.3 0.9 -0.4 -0.9 -0.4 Foraging stratum Tree 0.7 0.8 0.8 0.8 0.7 0.ó 0.6 0.6 0.0 0.4 0.5 0.4 Ground 0.3 0.2 0.2 0.2 0.3 0.4 0.4 0.4 0.0 -0.4 -0.5 -o.4 Dþt Grain 0.3 0.5 0.3 0.4 0.4 0.4 0.4 0.4 -0.3 0.0 -0.2 -0.1 Insect+ 0.6 0.2 0.3 0.2 0.4 0.3 0.3 0.3 0.3 -0.3 -0.1 -0.2 krsect 0.1 0.4 0.4 0.4 0.2 0.3 0.3 0.3 -0.1 0.2 0.4 0.2 Dbturbance toleranoe Min 0.2 0.1 0.3 0.2 0.1 0.1 0.1 0.1 0.5 0.3 0.5 0.4 Mod 0.5 0.5 0.5 0.5 0.6 0.7 0.7 0.7 -0.2 -0.3 -0.4 -0.2 Hish 0.2 0.3 0.2 0.3 0.3 0.3 0.2 0.2 -0.1 0.2 0.1 0.1 The sample sizes given are the number of avianprey found from March to September (proportion in diet) and the number of individual birds of species or genera that screech- owls have been docunented eating in the study area (proportion in environnent). The number of sites where prey was detected and the number of point counts conducted are

19t given in brackets. Category definitions: Pass = passerine, Non-p = non-passerine, Res = resident and locally breeding specbs, Mig - passage migrants and winter visitors, Small = species averaging <2Ograms, Med = species averaging 20 - 40g,I-arge = species averaging >409, Tree = predominantly arboreal or sluub level feeders, Ground = predominantly ground feeders, Grain = predominantly granivorous, Insect + = consumes a mixture of insects, fruit and/or rrcctar, Insect - predominantly insectivorots, Min = tolerates only minimal anthropogenic distrnbance, mod = tolerates moderate anthropogenic disturbance, High = tolerates high anthropogenic distrnbance.

The total niche breadth calculated by species (B) for this study was 6.5, however the ftue niche breadth is higher becarise of some of the diversity in the uniCentified birds and invefebrates from this data set is masked. To facilitate corrparisons between studies, I renroved invertebrates except crayfish (following Gehlbach 1gg4) leaving 38 species

(including some identified only to genus) and a niche breadth of 5.9, much lower than

Ohio (69 spp, B = 16.6, VanCarrp and Henny I975) and Texas (72 spp, B = 18.0,

Gehhach L994). The niche overìap betweenManitoba and Ohio (VanCarrp and Henny,

1975, Table 2) ß 0.7 (70Vo), much higher than the overlap between Manitoba and Texas of 0.2I (2I7o) (Fig. 3). Gehlbach (1994) calculated the niche overlap between Ohio and

Texas as 0.31. A recent study in the northeastern most portion of the species range in

Québec identified 26 prey iterrs (Richards et al. 2006). The total niche breadth In

Richard's strdy was only 3.4 (calculated fiom Richards et aL.2006, Table 5); however diet was not the atúhors' focrs and the only method t¡sed was periodic inspection of rpst boxes. Almost no invertebrates were recorded in Quebec due to methodology, so I calculated niche overlap with Manitoba excluding invertebrates as 0.87, indicating substantial similarities in mammalian and avianprey.

t92 I fi-nther investigated latitudinal trends by selecting seven studies where sufficbnt

information on prey composition was provided and calculating niche breadth, including

crayfish br¡t excluding all other invertebrates to facilitate comparison (Fig. 4.3).I found a

non-significant decrease in niche breadth with increasing latitude (linear regression: / =

3.47, df = 5, p = 0.09, .É = 0.46). This nend would likely be significant with greater

cornistency in methodologies and sarrple sizes.

Niche Overlap l7o%l t-ËlY:-i 20

15

ïl (õ o û¡ 10 o z.g

3035ñ4550 Latitude oN Fig. 4.3. Niche breadth (vertebrates and crayfish only) versus latitude from seven studies with niche overlap between Manitoba and Québec, Ohio, and Texas. Data from Manitoba (this study), Quebec (Richards et al. 2006), New York (Allen1924), Ohio (VanCamp and Henny 1975), Kentucky S.itchison and Cavanagh1992), Tennessee (Duly 1979), and Texas (Gehlbach 1994)

Sone species were identified as prey items in several studies, although their relative importance differed. The rrpadow vole is particularly important in northern diets (687o of mammals, I3TooveralTand32Vo of allwintercaptures inManitoba;'75Vo of mammals

t93 and 52Vo overall in Quebec; and3TVo of mammals, I37o overall and237o of all captures

in the non-breeding season in Ohio). The degree of specialization in Manitoba is

conparatively high for a generalist predator, with the tlree most frequently consurrnd

mammals (nnadow vole, house Írouse, deermouse) comprising at least 867o of mammalianprey (the true percentage may be higher because some prey were not

identified to species and if these are excluded is 977o), and the top three birds (Hou:se

Spanow, Black-capped Chickadee, and either Cedar Waxwing or Yellow-rurrped

Warbler) 237o of avian prey or 48Vo of avian prey identified to species. These percentages are higherthanOhio (top 3 mammals 857o,top 3 birds 32Vo) and Texas

(76Vo,39Vo).

DscrssroN

The high percentage of invertebrates in the diet of the Eastern Screech-Owl in the study

area and the fact tlnt5TVo of invertebrates were detected using observationaltechniques dennrstrate the need for observation in assessing prey ratios of diet generalists such as the Eastern Screech-Owl The high consurrption rate of invertebrates in the breeding

seasonrecorded in this study has only been matched by studbs examining stomach conteffs (Duly I9T9,Hanebrink et aL 1979, Brown 1989). Biases in diet information collected solely frompellets have also been demonstrated in other species that readily consume invertebrates, such as the Burrowing Owl (Athene cunicularia) (Haug et al.

1993, Plumptonand Lutz1993, York et aL.2002). Nonetheless, soft-bodied invertebrate groups, such as moths, worÍN and caterpillars, were recorded almost only by observation

It is possible therefore that soft-bodied prey are undeneported even in this study using

194 multiple methods and that only pure observation could determine their relative importance. The general pattern observed in the study area is that invertebrates are by far the nrost common prey during the breeding season, despite their relatively srnall contibution to biomass consunæd, whereas birds and mammals become rnore common in the non-breeding season This pattern is also found in southernparts of the range

(Gehhach 1994,1995). Variation in the type of invertebrates consumed by year, e.g. the ratio of ea¡thworms to beetles, is likely due to accessibility at the surface related to factors including humidity, soil npisture and annual variation in precþitation (Gehhach

1994).

Fish appear to constitute a smaller proportion of the diet in the study area than elsewhere, with only a single unidentified fish found in this sanple, constituting0.05Vo of total prey capture in Manitoba versus I.37o of prey capture in Kentucky (Ritchison and Cavanagh

1992),2.IVo nTexas (Gehlbach 1994), and3.57o inOhio (VanCamp and Henny 1975).

Since screech-owls in the study area select riparian habitat (Chapter 3 ) and 88Vo of all nests were within 500m of a river of creek (Chapter 2), it remains unclear why this is the case. Screech-owls in the study area caught crayfish (0.5Vo of prey captures) more regularly than fish and were also observed catching frogs in shallow water.

Gehhach (1994) noted larger birds above 41g were more likely to be cached. However, because larger items are more likely to be delivered to nests and/or cached, estimating hunting selectivity from caches may be biased towards larger items (Bull et al. 1989).

The majority of birds captured in this study were small species under 20g, which were

195 selected for, unlike species weighing 20 - 4}gand species >40g (Table 4.4), although

larger species were selected in rwal areas. Gehlbach (1994,1995) suggested that rnale

birds were captured more often than females. There was irnufficient data to assess this in

this study; however, an examination of the selectivity index of individual species

(Appendix 1) suggests some selectivity towards colorful plumage. In this study there are

only seven species with a selectivity index > 0, and these are nnstly small colorful

species such as American Redstart, Yellow Warbler, House Finch and Baltimore Oriole,

the exceptions being Hairy Woodpecker, Tennessee Warbler and House Wren Gehlbach

(1994, 1995) noted also noted a prevalence of ground-feeding birds over foliage-feeders.

The opposite pattern errrrged from this study with foliage feeders being selected for and

ground feeders selected against (Table 4.4). Peculiar was the absence of Mourning Dove

(Zenaida macroura) as a prey item in the study area because this species was a frequent

avian prey item in other studies at northern locations (VanCarrp and Henny 1975,

Richards et aL 2006).

Niche breadth calculations suggest very different diets in the northern ard southern portions of the range of Eastern Screech-Owl with narrower niches and greater specialization in the north- The northward gradient of specialization in the Eastern

Screech-Owl dernonsftated by a conparison of different studies is reflected in other owls, for exarrple, Boreal Owls in Finland (Korpimåiki 1986), whose total niche breadth in

oN western Finland at 63 is only 4.4 (from 40 species of avian and mammalian prey) and whose 3 most frequently consumed rrnmmal species comprise 807o and the top three bird species 59Vo (Korpim¿iki 1988, calculated by Gehhach 1994).In terms of niche breadth

1,96 and diet specializatior¡ Manitoban Eastern Screech-Owls, with their high corsurrption of

meadow voles and a few commonbird species such as the House Sparrow, are more

similar to Finnish Boreal Owls than their Texan conspecifics. The meadow vole appears

to be particularly important in permitting Eastern Screech-Owls to persist in this northern

study area. Although a native species, meadow voles were interestingly more commonly

captued in suburban areas (9.57o of all captured prey in rural areas, 13.27o in low density

suburban and, 13.9Vo in high-density suburban areas). The üend of decreasing niche breadth with increasing latitude (Fig. 4.3), suggests that this pattern of specialization is driven by the lower diversity of prey types available at northern latitudes. The Long- eared Owl (Asio otus), often descrbed as a diet specialist, has a rrpre diverse dbt in

southern locations (Bertolino et al. 2001)

Eastern Screech-Owls in the study area are clearly able to shift their diet with seasonal availability. Invertebrates are unavailable and thus absent from the diet during winter, but invertebrate consurrption increases steadily from pre-laying period to the post-fledging period. Conversely, mammals and birds are equally important in the winter but steadily decline in proportion through the breeding season Although mammals, inparticular rodents, were captured more than birds and contributed rnore to the biomass overall, birds appear to become increasingly important in late fall and the winter rnonths in this northern study area. Mazrn (Igg2) also reported birds and mammals in equal proportions in the late fall near Winnipeg, Manitoba. Birds were also corsumed in much higher percentåges dwing the non-breeding season than the breeding season in Texas (Gehlbach

1994) but were only slightly higher in Tennessee (Trnner and Dimmick 1981) and

t97 Kentucky (Ritchisonand Cavanagh1992) and decreased sharply inOhio (VanCanp and

Henny L975) and Michigan (Craighead and Craighead 1956). Based onthe latter two

studþs and Allen (1924), Ritchison and Cavana gh (1992) concluded that birds were

corsumed rrnre frequently in the breeding season at northern locations perhaps due to

increased availability from an influx of migrants (VanCamp and Henny 1975). This is not

srpported by this study; however, there is some srpport fur VanCamp and Henny's

(1975) suggestion that avianprey increases with the arrival of spring migrants, viz. birds were taken far less than mammals in the incubationpeúod (19Vo versus 49Vo) but becarre

dominant over fft¿rmmals in the brooding period (177o versus L5Vo) and remained slightly higher in the post-fledging period (57o versus 3Vo). lnManitoba the peak arrival of

Neotropical migrants is in mid- to late May during the middle of the brooding period for most Eastern Screech-Owl pairs. Furtherrnore, aÍþng the birds captured, the percentage of migrants increased in the breeding season and migrants had higher selectivity indices than resident and locally breeding species (Table 4.4).

The higher importance of avianprey in the winter in Manitoba is likely due to accessibility of prey with increasing snow cover. Eastern Screech-Owls have symmetrbal ears (Marshall L967) and are believed to use vision in hunting (Gehhach 1995). They are thus less adapted to the snow-plunging technique of boreal forest species such as the similar sized Boreal Owl (Aegolíus funereus), although even this species can be adversely affected by deep snow (Sonerud 1984). Instead Eastern Screech-Owls in Manitoba prefer to hunt mammals in winter at the base of large coniferous trees where snow cover is reduced or when they surface, for example, near bird feeders with fallen seed or when

198 ftaveling between subnivean tunnels, crossing areas such as driveways where snow has

been cleared (pers. obs.). House Sparrows and other shrub roosting birds are often hunted

by flushing them from roosts (pers. obs.).

Screech-owls in rural areas consume more invertebrates thando suburban screech-owls.

The higher proportion of invertebrates in the diet in rural areas was also noted in Texas

(Gehhach 1994) duing the breeding season However, Burrowing Owls cornumed rnore

aerial irsects at urban sites than at rwal sites in Florida (Chipnran et aL 2008).

Earthworrrs were most frequently captued in low dersity-subwban areas, presumably because the only species recorded, Lumbricus terrestris, is typically associated with human activity (Reynolds 2000). In additio4 the watering of lawns at night with

sprinkler systems rnay provide greater access to this prey item in suburban areas.

Invertebrate consunption was highest in the post fledging period when fledglings are

learning to hunt and their captures are almost exclusively invertebrate (pers. obs).

Unpaired males apparently capture larger prey items nnre regularly and rely less on

invertebrate prey than breeding pairs, although a lack of observational data for urpaired males means that this result must be interpreted with care. Invertebrates contribrfe less biomass to the diet but are important when feeding chbks, as they are small, canbe captued close to the nest, and delivered very frequerfly (as often as every 45 seconds in the case of earthwornn, pers. obs.). The Eastern Screech-Owl has a higher hunting success rate for invertebrate prey (Abbr:uzzÊse and Ritchisonl99T) permitting feeding at rrnre regular intervals. Invertebrate capture is therefore important for the provisioning of broods. Consumption of invertebrates began as much as two weeks earlier in sr¡buban

199 areas with the flrst invertebrate prey being recorded on 30 March in suburban areas as

opposed to 16 April in rural area (earliest dates all occured :nr2007). The percentage of

invertebrates captured in March and April against the total invertebrate capture was I7o

in rwal areas, lOVo nlow-density suburban and.4Vo in high-density suburban areas.

Although the earlier availability of invertebrate prey is unlikely to be the sole factor

permitting earlier nesting in the subrnbs, there are additional dietary benefits to early

nesting in that recently fledged young who leave the nest on average 5 days earlier in

suburban areas than rural areas (Chapter 2), would have increased availability of avian

passage migrants. Passage migrants are not only selected for, the breeding cycle of this

species may be timed to coincide with maximum prey availability for hatchlings and

fledglings, in particular the spring arrival of migrants (VanCanp and Henny 1975,

Gehhach 1994). The earlier availability of invertebrate prey in sr-rburban areas may be

related to factors including snow clearance, fertilized garders, or the urban heat island,

which increases invertebrate diversity in cold climates @eiclrsel2006).

Unlke in Texas where mammal consurrption was higher in rural areas, in Manitoba

mammals constituted the highest percentage of the dbt in high-density suburban areas

and were very similar in rural and suburban areas despite seasonal differences. Prey

diversity and niche breadtls were highest at low-density sr¡burban sites in both breeding

ard non-breeding season Rural sites had higher diversity in the breeding season than high-density suburban sites and vice-versa in the non-breeding seasons. Nbhe breadth

was higher overall in high-density suburban areas compared to rural areas but only

marginally. Seasonal differences in niche breadth in these categories reflect srnaller

200 sarple sizes. The higher diversity of prey in the non-breeding season in suburban areas may relate to the presence of bird species that, despite being uncoÍrmon in urban areas in surilner, linger or overwinter in the city due in part to the urban heat island, regularly replenished anthropogenb food sources, and possibly also due to the protective benefits of planted conifers (Taylor and Koes 1995). Such birds may also be concenfrated in small areas with reliable food sources, increasing their accessibility to owls. Species such as

Dark-eyed Junco (Junco hyemalis) and White-tluoated Spanow (Zonatrichía albícollis), as well as several other Emberizidae and Fringillidae species that are scarce in winter in the study area are most lkely to occur around feeders and often in suburban/urban areas

(Taylor and Koes 1995). Nonetheless, contrary to the generalized remarks of Gehlbach

(1995), high-density suburban owls in Winnipeg consuned by far the greatest percentage of mammals inthe non-breeding season (76Vo), and despite having the highest niche breadth by class in the breeding seasorì, had the lowest niche breadth by class in the non- breeding season because of the dominance of rodents in the diet and the absence of invertebrates (niche breadth is highest when all prey classes are in equal proportions).

Rodents in particular are accessible around suburban food sources. The fact that niche breadths by class are higher in the non-breeding than breeding season in both rural and low-density suburban areas may relate to the seasonal shifu in diet which render the prey classes less evenly distributed, inparticular that invertebrates are nearly 4 times higher than the next nearest class (mammals) in low-density suburban areas and nearly eight tirrns higher in rural areas in the breeding season

20r Severalprey species were less frequently captured inruralareas. House Sparrow, the only avian specbs to conpose >I}Vo of the total number of birds captured was rncst frequently found in the diet of owls in low-density sububan areas. In the study area, this species frequently roosts in derne shnrbbery close to buildings often near feeders or grain sources (pers. obs, l-owther and Calvin 2006), and owls therefore had excellent access to this prey item in suburban areas. Another species of owl, the Tawny Owl, is known to take advant¿ge of easy accessibility of avian prey at urban roost sites and correspondingly increase the proportion of birds in their diet in urban areas (Galeotti l99O,l99I). The house rnouse was completely absent from rural diets, whereas the meadow vole was Íþre commonly recorded as a prey item in suburban areas. These species were the two nrost significant individual prey items in terms of biomass conftibution and were particularly important to the owls in the non-breeding season If these captue rates are reflective of either availability or access then their apparent increased abundance in subr¡rban areas would convey an important advantage. Since the house mouse in North Anerica is the commensal form that lives fiþstly in buildings

(Barsfield 1974) increased availability in suburban areas is not suprising. Meadow voles are found rrnstly in grassy habitats within the study area (pers. obs, Bansfield 1974). In

Penrsylvania, meadow voles were more common in some suburban rþarianparks than in matrne rþarian forest (Mahan ard O'Connell2005) and in Ontario their density increased with greater cottage development (Racey and Euler 1982). V/ith their small home ranges, rodents like meadow vole can tluive in disturbed suburban habitats

(Dickman and Doncaster 1987, Nilon and VanDruff 1987); however, rodent diversity declines with irrcreasing arrnunt of imperviors suråce and bare gound (VanDruff and

202 Rowse 1986) and disturbed areas such as parks surrounded by industial sites are

characterized by non-native species such as the home rnouse (Nilon and VanDruff 1987).

Lfte meadow voles, sorre other small mammal species such as the northern short-tailed

shrew (Blarina brevicauda) are also most likely to occur at points of intermediate

distr¡rbance levels (Racey and Euler t982) and in this study were only found at suburban

sites. Small rnarnmal aburdance is often higher in small urban patches (Ekernas and

Mertes 2006), a phenomenon that may be related to limited dispersal (Barko et aL 2003).

This suggests that low-density suburban areas may offer the most diversity and

abundance of prey species (Blair 1996) and the wider niche breadth and prey diversity of

suburban screech-owls does therefore reflect availability. Likewise, Burrowing Owls in

Florlla enjoyed higher prey densities close to buildings (Millsap and Bear 2000). In

addition to diversity and density, access to rodent prey in subwban areas might be higher

either due to nocturnal feeding on fallen seed under bird feeders, snow clearance, and the

greater number of coniferous ftees under which little snow accumulates and which therefore facilitate hunting (Chapter 2).

The preference for larger birds in rural areas may be related to the increased anrcunt of invertebrate and decreased availability of mammalian prey, ie. the prevalence of small items and lack of mammals in the diet increasing the importance of selecting avian prey of larger biomass to ÍEet energy requirenents. Rural sites also proved different to suburban sites in terms of selectivity for both arboreal feeders and ground feeders.

Ground- feeding birds are often diet generalists that may benefit from urbanization

(DeGraaf 1991, Smith and Wachob 2006) and were ÍÐre coÍrmon in suburban areas than

203 rural areas. This fact and a slightly higher capture rate in rural areas (0.3 versus 0.2 elsewhere) produced the difference in selectivity. Insectivores were generally selected over granivores (Table 4.4). Foliage feeding insectivores and omnivores tended to be npre colorful than ground feeders in the study area; however, roosting preferences would likely be a better indication of the ease of capture since rnost avian prey are not active whencaptured by screech-owls. The disturbance tolerance of avianprey mostly mirrored disftibution with urban-adapted species selected for in suburban areas but not in rwal areas, whereas urban-avoiding species were selected for. When they do occur in suburban and urban areas while migrating such species rnay be particularly conspbuous.

The dietary regime of the Eastern Screech-Owl at the northern perþhery of its range is similar in overall compositionto the diets of southerly populations. Nonetheless, niche breadth and prey diversity decrease northward corresponding to availability and seasonal shifts in invertebrate versus vertebrate corsunption appear rnore nnrked. Despite their narrow niche breadth in a northern location, screech-owls in sr-rburban areas had more diverse diets thanruralowls. Rural screech-owls in the study area rely Íþre on invertebrates and corsume less vertebrate prey than suburban owls, despite the fact that invertebrate prey were available to suburbanpairs two weeks earlier in the nesting season Subrnban areas, therefore, can offer small predators with a rnore diverse diet, especially in biologically sftessful periods, than might otherwise be available.

Lrrpnarr-nn CnED

Abbruzzese, C. M., and G. Ritchison 1997 . The hunting behavior of Eastern Screech-

owls (Otus asio). Pages 2l-32 inJ.R. Duncar¡ D. H. Johnson, and T. H. Nicholls,

204 editors. Biology and conservation of owls of the Northern Hemisphere. North Central

Research Statiorì, U.S. Department of Agriculture, St Paul, Minnesota, U.S.A.

Allen" A. A. 1924. A conüibution to the life history and economic status of the screech

owl(Otus asio). Auk 41 :l-17.

Baker, P. J., A. J. Bentley, R. J. Ansell, and S. Harris. 2005.Impact of predationby

domestic cats Felis catus inan uban area. Mammal Review 35:302-312.

Bansfield, A. W. F. 1974. The mammals of Canada. University of Toronto Press,

Toronto, Ontario, Canada.

Barko, V. 4., G. A. Feldhamer, M. C. Nicholson, and D: K. Davie. 2003. Urban habiøt:

A determinant of white-footed mourse (Peromyscus leucopus) abundance in southern

I llino is. S o utheas tern N atura li st 2 :3 69 -37 6.

Battiq J.2004. When good animals love bad habitats: Ecological traps and the

corservatio n o f anima I pop ulations. Co nservatio n B io lo gy 18 :1482-1 49 I .

Bertolino, S., E. Ghiberti, and A. Perrone .2001. Feeding ecology of the long-eared owl

(Asio otus) in northern ltaly: is it a dietary specialist? Canadian Jounal of Tnology

79:21921198.

Blair, R. B. 1996. I¿nd use and avian species diversity along an urban gradient.

Eco lo gical App lications 6 :506-5 1 9.

Botelho, E. S., and P. C. Arrowood. 1996. Nesting success of western Br.nrowing Owls in

natural and human-altered environnents. Pages 61-68 inD.M. Btd, D. E. Varland,

and J. J. Negro, editors. Raptors in human landscapes: adaptations to built and

cultivated environments. Academic Press, London, U.K.

Browq R. K. 1989. Food habits of Kentucky owls. Kentucky Warbler 65:38-48.

205 Bull, E. L., M. G. Henjum, and R. S. Rohweder. 1989. Diet and optimal foraging of great

gray owls. Journal of Wildlife Managemert5347-50.

Caltn, A. R., and J. T. Kenp. 1930. Onthe food of cert¿in owls in east-central Illinois.

A.tk 47:323-328.

Chþman, E. D., N. E. Mclntyre, R. E. Sftauss, M. C. Wallace, J. D. Ray, and C. W. Boal.

2008. Effects of hurnan land use on western Burrowing Owl foraging and activity

budgets. Journal of Raptor Research 42:81-98.

Craighead J. J., and F. C. Craighead. 1956. Hawks, owls, and wildlife. Stackpole

Corrpany, Harrisburg, Pennsylvania, U.S.A.

DeGraaf, R. M. L991. Winter foraging guild sructure and habitat associations in

sub urban b ird co mmunities. I¿ndsc ape and Urban P lanning 2l :17 3-180.

Deichsel, R. 2006. Species change in an urban setting-ground and rove beetles

(Coleoptera: Carabidae and Staphylinidae) in Berlin. Urban Ecosysterrs 9:16I-I78.

Dickman, C.R., and C. P. Doncaster. 1987. The ecology of small mammals in urban

habiøts. Volume 1. Populations in a patchy environment. Jornnal of Animal

Ecology 56:629440.

Diermen, J. 1996. Sparrowhawk (Accipiter nisus) breeding in village, rural landscape and

forest. De lævende Matuur 97:43-5I.

Duly, L.J.1979. Lifehistoryaspectsof thescreechowl(Otusasio)nTennessee. Thesis,

University of Teruressee, Knoxville, Tennessee, U.S.A.

Ekernas, L. S., and Mertes, K. J. 2006. The influence of urbanization, patch size, and

habitat type on small rnarnmal communities in the New York Metropolitan Region

WildMefto, New York, New York, U.S.A.

206 Ehroclu M., and E. Marks. 2001. Bird tracks and sign. Stackpole, Mechanicsburg

Penrsylvania, U.S.A.

Environnpnt Canada. Weather Pages. Retieved September 2004 - December 2008 from:

http ://weathero ffice. ec. gc.ca

Erringtor¡ P. L. 1932. Food habits in southern Wisconsin raptors. Part I. Owls. Condor

34:176-186.

Faeth, S. H., P. S. WarrerL E. Shochat, and W. Marussich. 2005. Trophic dynamics in

urban communities. B ioScience 55 :399407 .

Fisher, A. K. 1893. Hawks and owls of the U.S. U.S. Department of Agriculture,

Division of Ornithology and Mammalogy Bulletin 3:163-174.

Galeotti, P. 1990. Tenitorial behaviour and habitat selection in an urban population of the

tawny owl,Srr¿x aluco L. Bolletino di Zoologia 57:5946.

Galeotti, P. 1991. Feeding ecology of the tawny owls (Srnx aluco) in urban habitats

(northern Ita ly). Bo lletino di Zno\o gn 5 8 : 1 43-1 50.

Gehhach, F. R. 1994. The Eastern Screech Owl: Life history, ecology, and behavior in

the sububs and countyside. Texas A&M University Press, College Station, Texas,

U.S.A.

Gehhach F. R. 1995. Eastern Screech-Owl Account 165 in A. Poole and F. Gill, editors.

The Birds of North America. Arrnrican Ornithological Union and the Academy of

N atrna I S c iences o f P hilade lphia, Philade lp hia, Penns ylvania, U. S. A.

GetZ L. L., L. B. Best, and M. Prather. l9l7 . I-e,ad in urban and rural songþirds.

Environrrpntal Pollution 12 :23 5-238.

207 Hanebrink, E. L., A. F. Posey, and K. Sutton 1979. A note on the food habits of selected

raptors from northeastern Arkansas. Arkar¡sas Academy of Science Proceedings

33:79-80.

Haug E.4., B. A. Millsap, and M. S. Martell. 1993. Bunowing OwI(Speotyto

cunicularia). Account 6l ín A. Poole and F. Gill, editors, The Birds of North

Arrprica. American Ornithological Union and the Academy of Natural Sciences of

Philade þhia, Philade þ hia, Pe nrs ylvania, U. S.A.

Holland, G. E., C. E. Curtis, and P. Taylor. 2003. Ecosystems and bird habitat in

Manitoba. Pages 324-325 ¿n Manitoba AvianResea¡ch Committee, P. Taylor, editor-

in-chief. The Birds of Manitoba. Manitoba Naturalists Society, Manitoba Avian

Research Co mmittee, V/innipe g, Manitoba.

Jacobs, J.1974. Quantitative measurement of food selection: a rnodificationof the forage

ratio and IvleVs electivity index. Oecologia 14:413417.

Jensen, K. H., S. Rauch, G. M. Morrison, and P. Lirdberg. 2002. Platinum goup

elenpnts in the feathers of raptors and their prey. Archives of Environmental

Contamination and Toxico lo gy 42 :338{47 . 'W., Kays, R. and D. E. Wilson 2002. Mammals of North Anerba. Princeton Field

Guides. Prirrceton University Press, New Jersey, U.S.A.

Korpimåiki, E. 1986. Gradients in populatbn fluctuation of Tengmalm's owl Aegolius

funereus in Europe. Oecolo gia 69 :195101.

Korpiruiki, E. 1988. Diet of breeding Tengmalm's owl Aegolius funereus: long-term

changes and year to year variation under cyclic food conditions. Omis Fennica

65:21-30.

208 Korschgeq L. J., and H. B. Stuart. 1972. Twenty years of avian predator-small mammal

re lat io rx hip s in M isso uri Jo urna I o f W ild life Mana ge rrre nt 3 6 :269 --282.

Kostelecka-Myrcha, A., J. Zukowski, and E. Oksbjczuk. 1997. Changes in the red blood

indices during nestling developrrrnt of the Tree Sparrow Passer montanus in an

urban environment. Ibis 1 39 :92-96.

I¡wther, P. E., and C. L. Chk.2006. House Sparrow (Passer domesticus), The Birds of

North Anerica Online. A. Poole, editor. Ithaca: Cornell I¿b of Ornithology;

Retrieved September 2008 from the Birds of NorthAmerica Online:

http :/ lb m.birds. co rnell. ed u/b nalspec ies/0 1 2 do i:10.217 3 lbm.L2

MaharÌ, C. G., and T. J. O'Connell. 2005. Small mammal use of suburban and urban parks

in cental Pennsylvania. Northeastern Natur alist 12:3071I4.

Marshall, J. T., Jr. 1967. Parallel variation in North and Middle American screech-owls.

Monograph of the Western Foundation of Vertebrate Znology, Number 1.

lMàztn, K. M. 1992.FallFood of the Eastern Screech-Owl in Manitoba. Blue Jay 50:33-

35.

Millsap, B. 4., and C. Bear. 2000. Dersity and Reproduction of Burrowing Owls along

an Urban Development Gradient. Journal of Wildlife Managenrcnt 64:3341.

Niloq C. H., and L. W. VanDruff. 1987. Analysis of small mammal community data and

applications to managenrcnt of urban greenspaces. Pages. 53-59 tnL. W. Adams and

D. L. Iæedy, editors. Integrating Man and Nature in the Menopolitan Environment.

" National Institute Urban Wildlife, Columbia, Maryland, U.S.A.

Plumptor¡ D.L., and R. S. Lutz. 1993. Prey selection and food habits of Burrowing Owls

in Co lorado. Great Ba s in Naturalist 53 :299 104.

209 Racey, G. D., and D. L. Euler. 1982. Small nnmmal and habitat response to shoreline

cottage developnrnt in cenfal Ontario. Canadian Journal of Zoology 60:865-880.

Rerrrs, V. 2000. How can maladaptive habitat choice generate source-sink population

dynamics? O fto s 9I :57 9-582.

Reynolds, J. W. 2000. A conftibution to our knowledge of the earthworm fauna of

Manitob a, C anada (O ligochae ta, Lumb r ic idae), Me gadr ilo g ica 8 :9 -I2.

Richards, N. L., P. Mineau, D. M. Bird, P. Wery, J. l¿rivée, and J. Duffe. 2006. First

observations of an Eastern Screech- Ow\ Megascops asio, population in an apple-

produc ing re gio n o f so uthern Q uebec. C anad ian F ie ld Naturalis t 1 20 28 9 --293 .

Ritchison G., and P. M. Cavanagtr- 1992.Prey use by Eastern Screech Owls: seasonal

variation in cenftal Kentucky and a revþw of previous studies. Journal of Raptor

Research. 26.6613.

Robbirs, C. T. 1993. Wildlife feeding and nutrition Academic Press, San Diego, CA.

Sibley, D. A. 2000. National Audubon Socbty - The Sibley Guide to Birds. Knopf, New

York, U.S.A.

Smith C. M., and D. G. Wachob .2006. Trends associated with residential developrrnnt

in rþarian breeding bird habitat along the Snake River in Jackson Hole, WY, USA:

Irrplications for conservation planning. Biological Conservation 128:431446.

Sonerud, G. 1984. Effect of snow cover on seasonal changes in diet, habitat, and regional

distribution of raptors that prey on small mammals inboreal zones of Fennoscandia.

Holarctic Ecolo gy 9 3347 .

Stewart, P. A. 1969. Prey in nvo screech owl nests. Auk 86:141.

210 Taylor, P., and R. Koes. 1995. Rare and marginal winter passerine birds in Manitoba.

Blue Jay 53:148-158.

Turner, L. J., and R. W. Dimmick. 1981. Seasonal prey capture by the screech owl in

Tennessee. Journal of Tennessee Academy of Science 56:56-59.

U.S. National Fish and V/ildlife Forensics l¿boratory. 2008. Feather Atlas Website.

Retrieved September 2004 - December 2008 from

http ://www. lab. fivs. go v/featheratlas.

VanCanp, L. F., and C. J. Herury. 1975. The Screech Owl: Its Life History and

Population Ecology in Northern Ohio. North Arrerican Fauna: 71. Unites States

Departrrent of the Interior Fish and Wildlife Service.

VanDruff, L. W., and R. N. Rowse. 1986. Habitat association of rnammals in Syracuse,

New York. Urban Ecology 9 :413434.

ïVageningen University E4perimenfalTnology Grorp. Skull site. Retieved September

2004 - Dece mber 2008 fro m: htrp ://www. skulls ite. co m/index. htm

Wilsorl K. A. 1938. Owl studbs at Ann Arbor, Michigan Auk 55 :L87-197 .

WoH, B. O., and C. M. del Rio. 2000. Use of saguaro fruit by white-winged doves:

isotopic evidence of a tight ecological association Oecologia 124:536-543.

YonrTov, Y., and D. Wool. 1997.Do the contents of Barn Owlpellets accurately

represent the proportion of prey species in the field? Condor 99:972-976.

York, M. M., D. K. Rosenberg and K. K. Stum 2002. Dþt and food-niche breadth of

Burowing Owls (Athene cunicularia) n the Imperial Valley, California. Western

Nodr Anerican Naturalist 62: 280187 .

2rl 5. Eastern Scrrech-Owl hatches Wood Duck eggs

Anstnacr.- I describe an Eastern Screech-Owl (Megascops asio) hatching three Wood Duck (Aix sponsd) eggs in a suburban nest box. Wood Duck(s) rerrnved all five eggs of a completed screech-owl clutch the earliest of which had already been incubated for at least 19 days, and laid tluee eggs in their place. The female screech-owl hatched the Wood Duck eggs, preened the ducklings, and attempted to feed them until they exited the nest box within 48 hours of hatching. Received 22December 2005. Accepted 27 JuIy2006.

Wood Ducks (Aix sponsa) arc well known to lay parasitbally with corspecifics (Hartrnan

L972, Semel and Sherman 1986) and other cavity-nesting ducks (Bouvier 1974, Eadie et

aL 1998). Wood Ducks occasionally remove conspecific eggs, however, such records usually involve damaged eggs (Serrnl and Sherman 1986). Wood Ducks are reported to

"evict" other bird species including screech-owls from nest boxes (Bellrose and Holm

1994). However, Semel and Sherman (2001) report that when returning female Wood

Ducks found boxes in which they had previously nested occupied by heterospecifics, including Eastern Screech-Owl(Megascops asio), they switched to another box (n = 10).

Here I record Wood Drck(s) removing an entire clutch and laying in a nest of Eastern

Screech-Owls.

While studying the reproductive ecology of Eastern Screech-Owls in suburban Winnipeg

Manitoba, I irstalled a miniatrne video camera in May 2004 inside a nest box in which a pair had successfully reared broods in nvo previous years. Five chicks fledged from this box in 2004. A female began laying on 3 April 2005 and by 10 April was incubating a clutch of five eggs. There was no sign of any unusual activity at the nest until 22 April'

212 when observers noted there were only three screech-owl eggs and one much larger egg.

On23 April, there was a second large egg and the three remaining owl eggs (Fig. 5.1a).

On24 April no change was noted; however, onthe morning of 25 April, two nnre owl

egp had beenremoved and a third larger egg was present (Fig. 5.1b). Video recordings

were made on a nightly basis but unfortunately did not extend sufficiently into the rrnrning to record the rernoval of eggs. I rncnitored the nest box from 0600 to 0730 hrs

CST forthe next3 days, and, oneachrnorningapatr of WoodDucks landedclose tothe box. The female Wood Duck then flew to the roof of the box and stepped repeatedly and heavily on it before rnaking a short circular flight and landing on the enftance hole. On each occasion I observed this behavior, the incubating owl jumped up to prevent the duck's entrance. On27 April, a fernale V/ood Drrck was recorded gining entrance to the nest box at 0638 hrs but was expelled by the owl which tied to bite the intruder on the back of the neck. On the morning of 2 May, the remaining owl egg was removed, though cuiotxly no new Wood Duck eggs were added. Unfortunately egg removal was not recorded due to a technical difficulty.

I concluded the larger egp in the box were Wood Duck eggs (confrmed upon hatching;

Helgeson Nelson 1993). There was no evidence of any dannged eggs in the box and no eggs had been buried in the nesting material The edge of the Red River was only a few meters from the base of the nest üee and, because there were no eggs or shells below the box, the rennved eggs Inay have been dropped over water or consumed (Semel and

Sherman 1986).

213 Despite the absence of her own eggs, the female owl incubated the three Wood Duck

egp. The flrst egg hatched at224o hrs on 25 May, the other two hatched later that

evening. Shortly after hatching, the owl preened the ducklings and ate pieces of eggshell.

She also brooded and attempted to feed the chicks. When the ducklings attempted to exit

the box the female owl gave whirury calls, which are "elicited particularly by dispersing juveniles" (Gehlbach 1995:7). The first chick exited the box jrst befure 2200 hß on26

May and the second shortly afterwards. The third chick departed the box at"0130lus on

27 May. The property owner took one drckling to a local nature reserve, but the other

fwo were not located. Wood Duck chicks are highly precocial and brood merging has

been recorded (Kirby 1990), but it is not known whether the chicks in question survived.

Fig. 5.1. A: Three Eastern Screech-Owl egp and two Wood Duck eggs on 24 April 2005, B: One F¿stern Screech-Owl egg and three'Wood Drck eggs on25 April 2005. In both images the wing and tail of the female screech-owl sitting at the box entrance is visible in the lower left corner;Winnipeg, Manitoba.

Raptorial birds occasionally incubate waterfowl egp. Dawson and Bortolotti (1997)

reported an American Kesftel (Falco sparveríus) incubating a Bufflehead (Bucephala

albeola) egg and four kesnel eggs (the Bufflehead and two kestrel chicks fledged).

Fannin (1394) reported a mixed clutch of Canada Goose (Branta canadensís) and Osprey

214 (Pandion halinetus). The Black-headed Duck (Heteronetta atrícapilla), anobligate brood

parasite, sometimes parasitizes diurnalraptors (Weller 1968, Höhn 1,975). Wood Duck

eggs have occasionally been found in nests of Western Screech-owls (Megascops

kennicottii) (J. M. Eadie, pers. comm.). In Winnipeg, there are several records from

volunteers of the FortWhyte Alive nature center ofjoint use of nest boxes by Eastern

Screech-Owls andWood Ducks (e.g., on 11 April 1997,one EasternScreech-Owlegg,

five Wood Duck eggs, and seven rrembranes were found in one box). This does not

imply syncluony of use because soÍìe may have been from the previous year or

sequential nesting and it is possible that dumping or usupation may have occurred.

Eastern Screech-Owls have been recorded incr¡bating the eggs of other species (Breen

and Parrish 1996).

The average incubation periods for both Eastern Screech-Owl and Wood Duck are approximately 30 days (Gehhach 1995, Hepp and Bellrose 1995). In this case, the female owl sat on the Wood Duck eggs for 3114 days, within the normal range of incubation for Wood Ducks (25-37 days) (Hepp and Bellrose 1995). However, because the owl initiated egg laying much earlier thanthe Wood Duck(s), her eggs would have hatched approximately 3 weeks before any of the duck's had they not been renroved. By accepting the Wood Dtpk egp, the Êmale owl's total incr¡bationperiod was extended to

55 days. Eastern Screech-Owls have been recorded irrcubating infertile eggs for as long as 78 days (Gehhach 1995).

215 Incidental egg dumping has been recorded in many avian specbs (e.g. Sealy 1989).

However, the rernoval of host eggs over a 10-day period suggests this was not a case of egg dumping. A failed attempt at nest usurpatior¡ however, cannot be dismissed. Unusual interspecific interactions of this nature have been attrbuted to conpetition for nest sites

(Eadie et al. 1988), but several studies suggest that nest parasitism in Wood Ducks is not related to a lack of cavities (Semel and Sherrnan 1986). Dawson and Bortolotti (1997) argued that certain desirable qualities of the nest site might be a factor. In this case, there was a vacant nest box in the sane yard with no discernable structural differences in which'Wood Ducks had previously reared several broods. Serrel et at. (1988) denprntrated that conspicuous placement of nest boxes increased innaspecific nest parasitism rates in Wood Ducks and the box in question was highly visible. This serves as a reminder, perhaps, of the caution required when artificial nesting sfructures are used as managenrent tools (Eadie et al. 1998)

AcrclowrnDGEMENTs

I am gateful to David and Sigrid Warrenchuk for assisting in the video recording and permitting observation on their property. Silvio E. Cascino and Barry Pomeroy produced the photographs. This manuscript was greatly improved by the thoughtful comments of

Robert P. Berger, James R. Duncan, John M. Eadie, Frederick R. Gehlbach Robert. W.

Nero, Spenser G. Sealy, Merlin W. Shoesmith and an anonymous revþwer. Funding was provided by small grants from the Special Conservation Fund of Manitoba Cornervatior¡

Manitoba Hydro, and the Great Grey OwlFund administered by Robert W. Nero.

216 LrrBn¡.ruRB Crrno

Bouvier, J. M. 1974. Breeding biology of the Hooded Merganser in southwestern

Québec, including interactions with Common Goldeneye and Wood Ducks. Canadian

Fie ld-Naturalist 88 :323-3 30.

Bellrose, F. C., and D. J. Holm 1994. Ecology and Managenent of the Wood Duck.

S tackpo le, Me chanicsb urg, Penrs ylvania, USA.

Breen, T. F., and J. W. Parrish" 1996. Eastern Screech-Owl hatches an American Kestrel

Journal of Field Ornitholo gy 67 :612413.

Dawson, R. S., and G. R. Bortolotti. 1997. Misdirected incubation in American Kestels:

a case of competition for nest sites? Wilson Bulletin 109 732J34.

Eadie, J. M., F. P. Kehoe, and T. D. Nudds. 1988. Pre-hatch and post-hatch brood

arnalgamation in North American Anatidae: a review of hypotheses. Canadian Journal

of Znolo gy 66 :17 09-17 21.

Eadie, J. M., P. W. Sherrnan, and B. Semel 1988. Conspecific brood parasitisrr¡

population dynamics, and the conservation of cavity-nesting birds. Pages 306140 ín

T. M. Caro, editor. Behavioral Ecology and Conservation Biology. Oxford University

Press, New York, USA.

Fannir¡ J. 1894. The Canada Goose and Osprey laying in the same nest. Auk 1L:332.

Gehhaclu F. R. 1995. Eastern Screech-Owl(Otus asio). Account 165 in A. Poole and F.

Gill, editors. The Birds of North Anerica. The Academy of Natural Scierrces,

Philadeþhia, Pennsylvania ard The Anrerican Ornithologists' Uniort WashingÍon,

D.C., U.S.A.

2t7 Hartman, G. W. 1972. The biology of dump nesting in Wood Ducks. Thesis, University

of Missouri, Columbia, USA.

Helgeson Nelson, C. 1993. The downy waterfowl of North America. Delta Station Press,

Deerfield, Illinois, USA and Delta Waterfowl and Wetlands Research Station, Portage

la Prairie, Manioba, Canada.

Hepp, G. R., and F. C. Bellrose. 1995. Wood Duck (Arx sponsa). Account 169 in A.

Poole and F. Gill, editors. The Birds of North America. The Academy of Natural

Sciences, Philadelphia Pennsylvani a arñ. The American Ornithologists' Unior¡

Washington, D.C., U.S.A.

Höhn, E. O. 1975. Notes on Black-headed Ducks, Painted Snþe, and Spotted Tinanrous.

Atk 92:566-575.

Kirby, R. E. 1990. Wood Duck nonbreeding ecology: fledging to spring migration Pages

61J6 inL. H. Fredricksoû G. V. Burger, S. P. Havera, D. A. Graber, R. E. Kirby, and 'Wood T. S. Taylor, editors. Proceedings of the 1988 North Arrerican Duck

Syrrposium. St l-ouis, Missouri U.S.A.

Seal¡ S. G. 1989. Incidental'egg dunping' bythe Horse Wren ina Yellow Warbler

nest. Wilson Bulletin l0l:49I493.

Semel, 8., and P. W. She¡man. 1986. Dynamics of nest parasitism in Wood Ducks. Auk

103:813-816.

Semel, 8., and P. W. Sherman 2001. Inftaspecific nest parasitism and nest-site

corrpetition in Wood Ducks. Animal Behavior: 6I:787-803.

Semel, 8., P. V/. ShermarL and S. M. Byers. 1988. Effects of brood parasitism and nest-

box placerrrent on Wood Duck breeding ecology. Condor: 90920-930.

218 Weller, M. W. 1968. The breeding biology of the parasitic Black-headed Duck. Living

BrdT169--207.

219 6. Eastern Screech-Owl in Manitoba: Evidence of historical range expansion

Arsrnacr.- A database of a total of 1954 sight records and specimens of Eastern Screech-Owl(Megascops asio) from 1883 to present was assembled for Manitoba, northernMinnesota, and South Dakota from numerous sources. The percentage of rufous nnrph Eastern Screech-Owls in these areas was calculated per decade to determine changes over the twentieth century. The percentage of rufous rrnrph Eastern Screech-Owls in Manitoba has fallen from approximately 6-117o in the I920s to less than I7o today. The percentage of rufous morph b ird in North Dakota and northern Miruresota lies between 8 and I97o (overall averages I4Vo) and shows increase or decrease over time despite some short-term fluctuatiors. Since rufous morph screech-owls have lower suvivorship in cold climates, this suggests that the Eastern Screech-Owl expanded its range into Manitoba in the late nineteenth or early twentieth centwy, at which tirre the proportion of rufous rncrph birds was similar to those of northern Minnesota and North Dakota but fell to present levels post-expansion due to the colder climate in Manitoba. Possible mechanisrrs of such expansion are discussed.

Ir.¡tnonucuoN

Southern Manitoba to 52"N represents the northern range limit of Eastern Screech-Owl

(Megascops asío); however, Manitoba may not have been a part of this species' historical range (Manitoba Avian Research Committee 2003). The flrst confrmed records of

Eastern Screech-Owl in Manitoba are from the early 1920s. Hamilton t¿ing found this

species at Oak L¿ke in 1921 (specimens CMNAV 17082 and 17193 at the Canadian

Museumof Nature) and NormanCriddle recorded screech-owls at Aweme nL922

(Taverner 1927). There is no species account for the (Eastern) Screech-Owl in Ernest

Thonpson fSeton]'s Bírds of Manitoba, although he recorded other secretive srnall owls,

e.g. Northern Saw-whet Owl, which he described as a rare resident; Boreal Owl, which

he described as a probable resident and winter visitor; and l,ong-eared Owl, which he

described as 'tolerably common" (Thonpson [Seton] 1890). Nonetheless, Thompson

220 [Seton] mentions, under the account of the NorthernSaw-whet Owl, that R. H. Hunter

'blaims" to have seen ard heard Screech-owl in Saboskong Bay and Point du Chêne in

1871, the irrplicationbeing that Thonpson [Seton] ûeated these records as hypothetical

(Thompson [Seton] 1890). The lack of early records of a species thatresides in close proximity to hurrnns suggests that screech-owls were either absent or rare in Manitoba in the early twentieth century. Taverner (L927) concluded that this specbs was a new arrival and others noted an increased density of screech-owls in Manitoba in the 1930s, e.g.

Rrfherford (1935) wrote "This 'little horned owf is becoming increasingly comrnon in the greater Wiruripeg district" ard Cartwright (1931) stated they were "now quifé coÍunon in the timber along the Red and Assiniboine Rivers". Despite these suggestions, the possibility remains that screech-owls were initially overlooked, perhaps becar-rse they were scarce.

V/ithin the last 50 years, Eastern Screech-Owls have expanded their range into the Swan

River area of Manitoba as far as 52"N, froma previous northern limit of c. 50"N, clearly dernonstrating the species' capacity ûo cobnize new areas in a conparatively short time

(Walley and Clyde 1996). There are a few recent records in coniferous or mixed forest in areas such as Bird's Hill Provirpial Park and even as far rnrth as the Duck Mourúains

(single record onI2 Aprrl2002, B. Walley and P. I-etatn, pers. comm.). Anextraordinary exnalimital record of a gray adult seenby nine observers occurred on June 14 1998 at

Pisew Falls (approximately 600km north of V/innipeg above 55'N) (S. Clubb, pers. comm). It is possible that this species is becoming increasingly adapted to the transition zone between the boreal bione and woodlands of a more deciduous nature. A similar

22t phenomenon has been suggested for Barred Owl (Boxall and Stepney 1982), a species

first recorded in Manitoba in 1886 (lThorrpson] Seton 1886), which may have arrived in

the province at a similar time to Eastern Screech-Owl (Houston and McGowan 1999) and

whose recerìt range expansion is well docunented and dramatic (Maztn and Janes 2000).

One source of evidence supporting the hypothesis of range expansion may lie inthe

relative abundance of the color morphs: gray, rufous, and internpdiate or'brown", the

forrrer being coÍrmon in Manitoba while the latter two are rare. This evidence relies on

the fact that rufous morph screech-owls are more susceptible to cold tenperatures and

have a higher mortality, by as much as 40Vo, when terrperatures fall to -5o and - 10'C

(Mosher and Henny 1976). In their study in northern Ohio, VanCanp and Henny (197 5)

noted a reductionof the perceffage of rufous birds from 23.3Vo to I4.77o in the severe

winter of 1951 (lowest terrperatures and highest snowfall during their 3O-year study).

Gehhach (1994) recorded a similar decline following a record freezn where in one study

area rufous birds fell from84%o of the nesting population in 1983 to 4.2Vo in 1987 and

fromZ3.3%o to 16.l%o in another. Grayplumage contains more melanin than rufous

plumage and has better thernnregulatory properties and greater resistance to abrasion by

dustpartbles (Gehhach1994, Gill 1995, VanCanp and Henny 1975).Inother

polychromatic specbs, e.g. Ruffed Grorue, gray rnorph birds also fare better than rufous

ones in cold climates (Gullion and Marshall 1968.). At the northern limit of its range in

Finland, gray morph Tawny Owls corstitute 707o of the population and gray males have

a longer breeding lifespan and higher lifetime recruitment (ofßpring that survive to breed) than brown males (Brommer et aL 2005). Gehlbach (1994) noted that although

222 gray owls are more difficult to see in full sunlight, rufous birds are more cryptic in cloudy

or humid conditions because red light is filtered out in subdued lighting and scattered by

water vapor. Becawe of the nocturnal and crepuscular habits of Eastern Screech-Owl,

rufom plumage should be selected for in warnL humid environnents. Rufous is dominant

to gray genetically because gray x gray pairings always produce all gray offspring

whereas rufous x rufous pairing produce c.257o gray young (Gehlbach 1994, VanCamp

and Henny 1975).

I hypothesized therefore that, if the Eastern Screech-Owl expanded its range into

Manitoba, then there should be a decrease in the percentage of rufous morph birds over tirrn, as rufous rnorph owls, better adapted to warmer and more humid environments, would diminish in the Manitoba population due to lower survivorship. I therefore calculated the percentage of rufous rnorph birds in Manitoba by decade and compared these percentage with the populatiors in northern Minnesota and North Dakota.

Msrrlons

I assembled a database Eastern of Screech-Owl records in Manitoba from l92l - 2007 , and a comparative database from northern Minnesota and NårthDakota from 1883 -

2007 . Northern Minnesota is defined here as the area above a line across the southern border of V/ilkin County to the soühern border of Grant - Douglas - Todd - Morrison -

Mille Lacs - Kanabec - Pine counties. This corresponds roughly in latitude to the border between North and South Dakota. I chose this area because of its location directly south of Manitoba. Records from southern Minnesota were excluded as the considerably milder climate there would be predicted to hold a higher percentage of rufom morph screech-

223 owls. The Manitoba database consists of over 1700 records representing 1399 individuals from numeror¡s sources inclding published references, oologists' sets, museum specinrns, nest cards, rehabilitated birds, applications for taxidermy permits, Cluistmas

Bird Counts, the Manitoba Nocturnal Owl Survey, personal communication from observers, and other reliable sources. The database from northern Minnesota and North

Dakot¿ contains records of 611 individuals from a similar set of sources. The lower number of records from Miruresota and North Dakota conpared to Manitoba is not reflective of lower owl densities but rather of difficulties in collecting data from afar.

Care was taken to cross reference all data gathered to ensure that sightings of individual owls were not duplicated. These records were then grouped by decade and the percentage of rufous morph birds calculated per period.

Rnsur,rs

Nearly tlnee quarters (74Vo) of the records collected were sight records from either published sources or from personal communication with the author, l)Vo are specimens or birds found dead, and the remainder were either heard only or are banding records

(Table 6.1). When the entire database is considered and records where color was not specified (the 'Unkno\ryn" column) are e:

224 Table 6.1. Break down of Eastern Screech-Owl records in Manitoba, northern Minnesota, and North Dakota from I 886 - 2007 after removal of Adults Juveniles Brown Rufous Unknown Totals Manitoba Sightrecords 709 438 78s (68.4) 17 (1.5) 27 (2.4) 3t6 (27.6) n47 Specimen / dead n8 6 9s (16.6) 14 (11.3) 6 (4.8) 9 (7.3) 124 Heard only 93 0 0 0 0 93 (lo0) 93 Banded or rehab I 0 l (100) 0 0 0 I Total 94s 4s4 887 (ú3.4 32 (23 33 (2.4 445 ß1.8 1399 MN and ND Sight records 263 8l tsz (44.2) I (0.3) 32 (9.3) tsg (46.2) 344 Specimen / dead 80 l s9 (67.8) 4 (4.6) 21 (24.t) 3 (3.4) 87 Heard only 175 0 00 0 r7s (100) 175 Banded or rehab 5 0 4 (80) 0 0 l (20) 5 Total 523 88 2ts ß5.2 s3 (8.7) 33E (ss3 611 1457 780 Numbers in brackets are percentåges of the color morphs agairst the total for each category. Eggs are not counted per se, only individual owls reported by the oologists upon collecting. See text for definition of northern Minnesota.

A total of 783 records (39Vo of the database) could not be assigned a color rnorph either

because they were heard only or because this inforrnation was not provided. This

necessitated the calculation of two percentâges: flrstly, the number of rufous birds as a

percentage of the total number of records of individual birds and secondly the number of

rufous birds as a percentage of the total number of records where the morph is clearly

specified (hereafter "m.s') (Fig. 6.1). The data set ñr Minnesota and North Dakota

suffers from a very small sample size, in particular from 1940 - 1960 and I have

therefore averaged the two percentages (rufous rmrphbirds against all individuals and

rufous morph birds against individuals where color was specified) as a way of obtaining a

fairer Íleans of corrparison with the Manitoba data.

225 -El-% a.ll MB records o/o F---r -.ts ms MB records / *.x-..%NrhMN&ND I I I I \ >ç-- \ 7"-' ^. I ',\l/- \

'19,+l-50 <1930 r93 l -40 t95t-60 1961-70 l97r-80 l98l-90 199t-2000 2001 - 07

MB:n-48 MBn-40 MB:n-7 MBn-29 MBn*17 ì\lB: n * 98 MB:.n- 227 lvß: n* 412 MEn*465

m.s: n = 20 m.s:n= 12 rLs:n=3 m.s:n=8 m,s: n = l0 m.s: n = 50 m.s: n = 149 m.s: n = 239 m-s: n = 399

Nth: n =56 Nth: n = 47 Nth:n=6 Nth: n=9 Nth: n = 35 Nth: n = 99 Nth: n = 196 Mh:n=65 Nth: n = 100 Fig. 6.1. Percentage of rufous morph Eastern Screech-Owls recorded in Manitoba by decade {Vo of total number of birds recorded and of total number of records where rnorph is specified (ms)) and in North Dakota and northern Minnesota (Nth) {average of the Vo of totâl number of birds recorded and of totål number of records where morph is specifiedÌ.

This graph suggests that rufor¡s morph birds have declined in Manitoba from approximately 6.25 - IO.63Vo (average of Vo all records and To ms) in the 1920s to lVo today. This decline appears steady and gradual from the 1960s to 2005, although there is an apparent increase in the percentage of rufor¡s birds for the period l93I - 1960. This anomaly may stem fromthe very low sample sizes and different data collectionpractices during and in proximity to WWII. During this period, specimen and egg collecting declined in popularity and the number of field observations is low. Many of the birds reported from this period were found dead or injured, often in barns or on farmsteads.

Since rufous birds are rrìore susceptible to cold, there numbers would be inflated by a

226 sanple collected in this manner. The percentage of rufors birds from specimens only for

this period is higher than the same percentage calculated from sight records only

(Manitoba ms data set;3.Z%o of sightrecords rufous versus 5.27o of specimern, U.S. ms

data set: 17 .3Vo of sight records rufous versus 257o of specimers). Unfortunately, the various types of records in this data set are not evenly distribmed, with rrnre specimens

in the earlier part of the 20th Century and rnore sight records in the latter half.

Becar¡se rufous Eastern Screech-Owls are unusual in Manitoba and to a lesser extent in northern Minnesota and North Dakota, there is a tendency for observers to report their color, whereas many fail to comment on the color of gray birds. Therefore, the majority ofbirds in the "morph unkrþwn" category are probably gay.I believe for this reason that a Íþre precise figure lies somewhere inbetween the two types of percentages calculated, probably closer to the forner (ie. the lower end). Brownbirds are also underreported, as it is difficult to distinguish this intermediate form from gray, even in specimens, where intermediate characteristics, feather wear, fading, and foxing (browrs becone brighter or redder over time) must be considered. In fact, rrcst gray specimens and living screech- owls I have seen in Manitoba have had a srnall amount of rufous in the plumage, especially onthe tarsi; as a cenftal vertical band thnough the ear tufu; in the baning and cross-barring on the breast, belly, and flariks; or as a subtle wash across the underparts, nuchal collar, mantle, or wing coverts.

Despite the caveat that the percentåge of rufous birds before 1960 might be slightly inflated, Figure 1 shows an unequivocal decline in the percentage of rufous morph screech-owls in Manitoba but no clear pattern of decline in northern Minnesota and North

227 Dakota. Inthe latter areas, the percentage of rufous birds appears to have fluctuated

considerably but not otherwise changed greatly between the period prior to 1930 (2IVo)

and the 2001 - 2007 period (IgEo).I believe the fluctuatiors shown here are not real but

rather a product of the small sample size and the incredibly high number of morph

urspecified records (53Vo) in the U.S data set as compared to the Manitoba data set

(33Vo). Despite the poor quality of the data, this suggests a relatively stable percentage of

rufow screech-owls in the areas south of the Manitoba border, with sorrp fluctuations

expected due to climatic variatior¡ in süong contast to the situation in Manitoba. These data are therefore consistent with the hypothesis that the Eastern Screech-Owl expanded its range into Manitoba in the late nineteenth or early twentieth century and that rufous morph birds then declined in number due to their poor suitability to the local climate.

DncwsroN

If the Eastern Screech-Owl has extended its range into Manitoba, the most significant factor in this e4pansion is likely to be antlropogenic habitat change. Larger human settlerrpnts have brought with them the planting of trees around honesteads and shelterbelts inareas of prairie grassland formerly inhospitable to a woodland species such as Eastern Screech-Owl (Manitoba Avian Research Committee 2003). Many tees in

Manitoba may now be larger than in previous centuries, especially in proximity to human habitation dræ to watering and heat- island effects. During the 20th century, there have been both increased plant growth at mid to high latitudes (45.N and 70"N) and a lengthened growing season (Warren et al. 2004). Tree phenology has changed with some species blooming or budding weeks earlier (Wanen et aI.2004), and with sonn species in southeastern Manitoba becoming hardier (McKenny et aL 2001). Larger frees are more

228 likely to produce cavities of sufficient size for screech-owls and the open understory they

prefer for hunting. Buildings on the landscape may also have assisted in range expansion,

providing both shelter and access to prey, ensuing greater likelihood of winter survival

in a range perþheral contefts (Adam 1989, Houston 1989, Walley and Clyde 1996). An

additional benefit to Eastern Screech-Owl is the placement of nest boxes (usually for

Wood Ducks), which they readily use as nest and roost sites. The earliest record of

Eastern Screech-owls breeding in a nest box in Manitoba is 192g (nest card).

The cwrent distribution of Eastern Screech-Owls in Manitoba is sonewhat, but not

entirely, linked with human settlenpnt. Eastern Screech-Owls have higher population dersities in suburban Winnipeg than in nnal areas, producing larger broods and beginning breeding earlier in the year (Chapter 2). Available habitat in suburbs close to rivers or creeks with matue ftees matches many of their habitat selectionpreferences, including a relatively open subcanopy and middle story, many potential nest sites, some planted conifers (frequently used as roost sites bybreeding males), and tall nees with little shrub dernity below them (Chapter 3). Sr-rburbs also have lower densitbs of Great

Horned Owls and raccoons despite higher densities of domestic cats (Chapter 3), and a greater diversity of prey with access to invertebrate prey earlier in the breeding season

(chapter 4). only one of 52 nests found by rhe aurhor from2004 -2007 was in

"wildlands", i.e. <1 humanresiderú per hecta¡e (Chapter 2).

Another factor that may have assisted the Eastern Screech-Owl to spread northward is the innoduction of certain prey items of Palearctic origin, in particular the House Sparrow

(Manitoba AvianResearch Committee 2003), in the late nineteenth centwy, frst recorded

229 in the province :r;,1892 ([Thorrpson] Seton 1908). The strongest evidence for this is the

apparent importarrce of the House Sparrow, and the Rock Pigeon(captured inbarns presunrably along withrodents) for winter swvival in the Dauphin area (Walley and

Clyde 1996). Nonetheless, in the Winnipeg area, no innoduced prey item is consumed nearly as frequently as the native meadow vole, which comprises 357o of all vertebrate captures year round and 32Vo in winter (house mouse: 77o year round, 7 Vo wtnter; House

Sparrow: 67o,I17o;Rock Pigeon: 0.5Vo,)Vo; and Euopean Starling: 0.3Vo,07o; n= 781,

44;Clnpter 4). The introduced earthworm Lumbricus tetestris (commonly called "night crawler'), oftencaptued in watered lawrs and gardens, is summer prey only (5% ofall invertebrates captured, n= 1542, Chapter 4). Eastern Screech-Owls are thus not dependent on inftoduced prey. Probably rrnre important to screech owls are corrcentrations of prey, as may occur around fallen birdseed or at avian communal roosts.

Human activity may increase access to birds and mammals since together they constitúed l5Vo of tlrc diet in the breeding season in rual areas (n = 490) versus 33Vo n suburban areas (n = 161 1) (Chapter 4).

Eastern screech-owls appear to benefit from the urban heat islands and warmer terrperatues in Winnipeg, e.g. fledging dates are significantly conelated with average

Marchtemperature and withthe earlier absence of snow onthe gound (Chapter 2).

Climatic warming could benefit these owls by permitting earlier nesting and increased winter survival rates. In southern Manitoba, nxrst areas have experienced a ftend of increasing ar¡rual average tenperature tluoughout the centrny, especially in spring

(average increase of 1.7"C per century across five weather stations in Birtle, Brandon,

Morden, Sprague, and Winnipeg) and winter (average increase of 1.3'C per century)

230 (Tuner and Blair 2005). Despite gradual change over the past centuy, the largest

increases in average temperatures across the Holarctic region have occurred since the

1970s (Mann et aL 1998)' Climate change may therefore have been less significant than

habitat change in pronnting range exparsion.

The decline in rufous morphbirds in Manitoba over the course of the twentieth cenrury

from similar percentages to those found in North Dakota and northern Minnesota to near rnonochromatismtoday supports the hypothesis of historical range expansion into southern Manitoba. In addition to Eastern Screech-Owl a number of eastern woodland species have apparently undergone similar northor northwestward exparniorl including: wood Duck, Americanwoodcock, Barred owl, Red-headed woodpecker, yellow- throated Vireo, Purple Martin, Golden-winged Warbler, Eastern Bluebird, and Indigo

Bunting (Manitoba Avian Research Committee 2003), as well as Írammals such as the raccoon (Procyon lotor) (Larivière 2004). Manyof these species have some similar habitat requirenents to Eastern Screech-Owl, including cavity nest sites and/or a partially open understory for foraging. Although the car¡ses may differ with individual species, a tend of north and northwestern exparnio+ possibly coupled with the disturbance of prairie habitat, is implied in the early twentieth centrny in southern Manitoba, and nnre generally inNorth Anprica (Johrson 1994). The context and nechanisms of range expansion in the northernpraiie regiontherefore warrant further investigation

231 LnERATURE CmED

Adanl C. I. G. 1989. Eastern Screech-Owl in Saskatchewan and Adjacent areas. Blue Jay

47:16Ç188.

Boxall, P. C., and H. R. Stepney. 1982. The distribution and status of Baned O wl in

Aherta. Canad ian F ield- Naturalist 96 46-50.

Brommer, J. E., K. Ahola, and r. Karstinen 2005. The colour of fitness: plumage

coloration and lifetime reproductive success in the tawny owl. Proceedings of the

Royal Society of I-ondorr, Serþs B 272:935-940.

Catwright, B. Vy'. 1931. Notes and observations on sonìe Manitobanbirds. Canadian

Fie ld- Naturalist 45 : 1 8 1-1 87.

Gehhach F. R. 1994. The Eastern Screech-Owl: Life history, ecology, and behavior in

the suburbs and countyside. Texas A&M University Press, College Statiorl U.S.A.

Gill, F. B. 1995. omithology. second Edition w. H. Freeman, New york, u.s.A.

Gullioq G.W., and W. H. Marshall. 1968. Survival of Ruffed Grouse in a Boreal Forest.

Living Bird 7:117-167.

HoustorL C. S. 1989. First Fully Docurrpnted Nesting of Eastern Screech-Owl in

Saskatchewan Blue Jay 47:210112.

Homton, C. S., and K. J. McGowan 1999. The westward spread of the Barred OwL Blue

Jay 57:190-195.

Johrsoru N. K. 1994. Pioneering and natural exparsion of breeding distibutions in

western North Arrprican birds. Studþs in Avian Biology 15:2744.

I¿rivière, S. 2004. Range exparsion of raccoons in the Canadian prairies: review of

hypotheses. Wildlife Society B ulletin 32 :955-963.

232 Manitoba Avian Research Committee.2003. The Birds of Manitoba. Manitoba

Natt¡ralists Society, Manitoba Avian Research Committee, V/innipeg, Manitoba.

Mazur, K. M., and P. C. Janes. 2000. Barred Owl. Account 508 ¿n A. Poole, and F. Gill,

editors. The birds of North Anerica. Academy of Natural Sciences, Philadeþhia, and

Anærican Ornitho lo gists' Unio r¡ V/ashingto n, D. C, U. S. A.

McKenn¡ D.w., M. F. Hutchinso4 J. L. Kestevenard L. A. venier.200l. canada's

plant hardiness zones revisited using nndern climate interpolation techniques.

Canadian Journal of Plant Science 8l:ll7-129.

Mosher, J. 4., and C. J. Henny. 1976. Thermal adaptiveness of plumage color in Screech

Owls. Auk 93:614419.

Rr.therford, G. 1935. Chbkadee Notes #7 40 (a forner column of the Winnipeg Free

Press)

Taverner, P . A. 1921 . Some recent Canadian records. Avk 44:217 1ZB.

[Thompson] Seton, E. E. 1886. The birds of western Manitoba. Auk 3 :145-156,320-

329.

[Thompson] Setor¡ E. E. 1908. Recent bird records for Manitoba. Auk 25:450454.

Thorrpson [Seton], E. E. 1890. The bi¡ds of Manitoba. Proceedings of the U.S. National

Museum 13:457443 [No. 841].

Turner, J. and D. Blair. 2005. Climate change inManitoba. Unpublished draft for the

Provincial S ustainab ility Report. vancarrp, L. F., and c. J. Henny. 1975. The screech owl: Its Life History and

Population Ecology in Northern Ohio. North Anerican Fauna: 71. Unites Søtes

Departnent of the Interior Fish and Wildlife Service.

233 Walley, W. J., and C. F. Clyde. 1996. Occurrence and breeding of the Eastern Screech-

Owl north of the Riding Mounrairn, Manitoba. Blue Jay 54:89-100. warre4 F. J., E. Barrow, R. schwartz, J. Andrey, B. Mills and D. Riedel 2004. climate

change impacts and adaptation: a Canadian perspective. Governnent of Canada,

climate change Irrpacts and Adaptation Directorate, otlawa, ontario.

234 7. Conclusion:

Mnc H¿,Nrs r\{,s oF s uB URB A}r ADApTA TIoN IN um Eas rm,N S cn npc n-Ow t

Suburban Eastern Screech-Owls (Megascops asio) in the Winnipeg area exhibit higher population density, Iarger average brood sizes, earlier average fledging, greater prey

diversity, and wider niche breadth than rural screech-owls. Likewise, Gehhach (1994) derionstrated significant differences between a subtuban and a rwal population of this species in Texas with the sububan population producing one and a half times npre young and reaching five times greater density per unit area than the nnal population because of greater fecundity and survivorshþ. Conparison of these two studies suggest that same processes apply across the rural - sr-¡buban gradient in Texas as in Manitoba, and that they do not appear to be more or less pronounced in either context. Gehhach

(1994) discussed variors mechanisrns to explain this pattern and I have presented data in this regard in the preceding chapters. In conclusion, it is pertinent to address several possible mechanisms in tun

Urban Heat Island

The urban heat island in concert with other forms of climate buffering, e.g. the reduction of the impact of droughts through human interventioq permit higher overwinter survival and earlier nesting @mlen l9l4,Korpimäki 1978, Gehlbach1994, Rollinsonand Jones

2002, Shochat et al. 2006). Earlier nesting gives the owls a significant head start over migratory secondary cavity nesters such as Wood Ducks (Aix sponsa) and provides rnore tirre for the fledglings to master hunting techniques before their frst winter (Gehlbach

235 1994). Early nesting may also facilitate the provisioning of young because the period of highest food demand, when the chicks are getting larger but the female is not yet leaving

the cavity to assist the mab in provisioning, would coincide with the spring anival of

Neotropical passage migrants. Arrnng birds, screech-owls select passage migrants

(Chapter 4) perhaps because their lack of familiarity with the local area makes them easier to catch The urban heat island can increase a suburb's carrying capacity of prey populations (Enrlen 1974, Rollinson and Jones 2002, Shochat et aL 2006) and may even inftoduce new potentialprey iterrs to an area where they were not previorsly found

(Panis and Hazel2005).

During this study (2004 - 2007), the mean tenperature in do wnto wn Winnipe g (The

Forks) in winter (December - February) was -L2.52"C, nearly 2oC warner than the average temperature for the sane period at the aþort, an open area towards the outskirts of the city (-14.44"C). The ÍEandaily minimum temperature for the sanp period was

oC rnore than 3 warrrpr ar The Forks than at the aþort (-16.52"C versus - 19.68'C)

(Environrrnnt Canada, nd). During the breeding season of Eastern Screech-Owls (March

- July), the nean tenperature was 1.3oC warfirer at The Forks (10.75'C versus 9.47"C) and the nran daily minimum temperature was 2.4"C warmer (5.71'C versus 3.28'C)

@nvironnent Canada, nd). The sftongest evidence that these differences are biologically significant to Eastern Screech-Owls in the Winnipeg area is that the fledging date of the frst chick in each nest is correlated with distarrce to the city center and with the last date with snow on the gound (Chapter 2). Suburbanpairs fledge young on average five days in advance of rural pairs (Chapter 2).In addition suburban pairs had access to

236 invertebrate prey two weeks earlier than rural pairs (Chapter 4). Invertebrates are conpletely unavailable in winter in Manitoba so their early appearance in the diet of suburban screech-owls implies warmer terrperatures and reduced snow cover.

Reduction in Predation

Eastern screech-owls in suburban and wban areas of Winnþeg enjoy reduced densities of predators such as Great Horned Owls (Bubo virgininnus) and raccoons (Procyon lotor) than rural screech-owls (Chapters 2 and 3). On the other hand, they may live with higher numbers of domestic cats (Felis catus), which they avoid (Chapter 3). Screech-owls may avoid Great Horned Owls and in survey were only detected on the same üansect five tines (Chapter2). Furthermore, whenGreat Horned Owls and Barred Owls (Srrrx varia) rrpved into a new area, screech-owls abandoned breeding territories soon after. This provides some anecdotal evidence for an inverse relationship between the density of larger predatory owls and screech-owls. Nonetheless, direct evidence of a causal relationship between reduced predator densities and higher screech-owl densitþs in mediun¡high density suburban areas is lacking.

Nesting success was correlated to distance to building in Texas (Gehlbach 1994), as has also been documented for other bird species (Møller 1988). In this study, screech-owls avoided potential nest sites close to buildings; however, the average disørrce of a successful screech-owl nest to a building was <50m This behavior rnay facilitate the avoidarrce of certain predators. Rural screech-owls used cavities with a smaller entrance on average, suggestive of higher nest predation pressure (Sonerud 1985, Belthoffand

237 Ritchison 1990, Yetter et aL 1999).In Texas, sububan screech-owl chicks stayed longer

in the nest presumably due to the reduced threat of nest predation (Gehlbach 1994).

Although cause and effect are difficult to determine, predation risk appears to influence

the habitat selection of Eastern Screech-owls.

Habitat Change

Many urban and suburban areas offer suitable habitat to Eastern Screech-Owls in

Winnþeg and they show their highest breeding success and population density in these

areas. Older sr¡burbs in particular have large deciduous trees along steets and in yards

and parks, with open lawns beneath them. This habitat greatly suits the screech-owl's hunting style and habitat selection (Chapter 3). In Wirurþeg, suburban areas afford screech-owls another luxwy with a higher percentage of coniferous trees than rural areas, which the owls select for on breeding tenitories (Chapter 2) and which males in particular use as roost sites in the early stages of breeding before deciduous trees have foliated when they need to be close to the nesting cavity (Chapter 2). Conifers and buildings offer screech-owls excellent roost sites to avoid detection as well as for thernnregulatory purposes. Buildings are also a factor in the habitat selection of Barn

Owls at the northern perþhery of their range (Andrusiak and Cheng 1997).

Shrub density around suburban nests is also lower than in rural areas, in part because many gardens and parks are manicured (Chapter 3 ). Shrub density was negatively correlated with breeding success. The middle story of suburban areas is also less dense than inrural areas (Chapter 3). The less dense, but mature, vegetation with an open

238 understory found in suburban areas improves the ease of detecting predators and

defending nests. This also allows the owl to approach or leave the nest by flying very low

to the ground to avoid detection (Gehlbach T994). A fînal advantage of suburbia is the

placerrrnt of nesting boxes for Wood Ducks, often along rivers and close to houses,

which screech-owls readily use for both nesting and roosting. Because screech-owls are

polyterritorial, nest boxes may improve the quality of territorþs or even make some areas

suitable for breeding that would otherwise lack sufficient cavities. They rnay therefore

contribute to the higher densities of screech-owls insuburbanareas. There was no

eviderrce of nest-boxes creating an ecological nap since brood sizes from nest-boxes did

not differ significantly from those of natural cavitþs (Chapter 2).

Prey Availability

Sr.lburban screech-owls showed broader niche breadth and higher prey diversity than their rural countelparts (Chapter 4). They were also able to consunrc more birds and mammals

and were less dependent on invertebrate prey in the breeding season (Chapter 4). This

suggests greater prey availability in suburban areas. Rodents in particular are accessible to screech-owls around bird-feeders with fallen seed, and feeders also attract birds providing greater concentrations of prey in a small area (pers. obs). Well-watered and fertilized lawrn provide screech-owls with another concentated food source in earthworms and beetles that are rrcst active at the surface after rains or watering (pers. obs). Suburban areas'thus provide these owls with an increased abundance, variety, and accessibility of prey.

239 Eastern screech-owls have several important pre-adaptations to highly fragmented and

disturbed habitaa such as suburban areas. As discussed above, their preference for open

habitat rrakes them well suited to the suburbs and their small spatial requirements enable

themto exploit small habitåt patches (Gehlbach 1995). Their small body size,

inconspicuous perch hunting style, and cryptic plumage enable them to live in close

proximity to hurnans with less persecúion than larger predators (Gehlbach 1995). They

are diet generalists and can therefore capital:r;e on fluctuating food supplies (Gehhach

1994). They have large broods (range 2 - 6, average 4.2 fledglngs) and breed annually

(as opposed to being dependent onprey cycles as are more stenophagous owls) and at an

early age (one year old) (Gehhach 1994). Their conparatively short incubation and

nesting period combined with their polyterritoriality and the fact that they are residents

and can commence breeding in the early spring facilitate renesting when disturbance is

enco untered (Gehlbach I99 4). Co nsequently, Eas tern S creech- Owls in s ub wban

Winnþeg have reached higher densities, achieved larger brood and earlier fledging, and

enjoy nnre diverse diets than owls in rural areas on the outskirts of the city with lower

human density. There is also evidence to suggest that it was this pattern of exploiting the

advantages of proximity to humans and their sfructures, as well as the changes they brought to habitat and climate, that enabled the Eastern screech-owl to expand its

geographical range northward into Manitoba over the course of the Iate nineteenth and

fwentieth centuries (Chapter 6). In terrns of climate and diet, this expansion has propelled themto the biogeogaphical edge, fuither fromthe core range and characteristics of their

predominantly tropical insectivorous genus than any of their congenitors.

240 h.norcrroNs AI\D Fn,lnn¡Gs

At the oúset of this researct¡ I predicted that suburban and urban Eastern Screech-Owls

in the Winnipeg area would have advarrced breeding phenology and higher reproductive

success than screech-owls in rural areas. I eryected that in Manitoba's northern temperate

climate, at the perþhery of the range of Eastern Screech-Owl, any differences in breeding

phenology and brood size between rural and suburban pairs would be even greater than in

Texas (Gehlbach 1994). The first of these predictiorn has been clearly demornftated.

Eastern Screech-Owls have denser populatiorn, advarrced breeding, rrnre offspring per

nesting attempt and a more varied diet in subrnban areas than less densely populated

areas. Breeding phenology is correlated with distance from the city center implicating the urban heat island. There are also differences in the habitat available to the owls in more

densely populated versus less densely populated areas, e.g. high dersity suburban areas have a more open canopy with trees and shrubs spaced farther apart, more conifers, Ílore

Anprican elm and Manitoba maple (the two nþst commonly used üee species for nesting). Suburban owls have to contend with fewer raccoons and Great Horned Owls but rnore dorrpstic cats and live in geater proximity to buildings with more pedesftian activity (po ss ib le predator deterrents ).

The second predictioq that such differences would be nnre pronounced in Manitoba than

Texas, was not sr,pported. The difference in breeding phenology was in fact slightly greater in Texas than Manitoba (6 versus 5 days). Lftewise, sububan owls in Texas produced 0.9 chicks more per nesting attempt than rural birds (Gehhach 1994), corrpared to a difference of 0.3 chicks per nesting attenpt in Manitoba. Unfortunately, it has proven difficult to assess the significance of these results due to methodological

24r differences (gradient approach in Manitoba versus comparison of two study sites in

Texas). Inboth Manitoba and Texas, the evidence suggests that the sane suite of

interacting mechanisms combine to provide an advantage to suburban screech-owls.

These are the urban heat island, lower densities of natural predators and nest predators,

sorre buffering of predator activity due to human presence, greater and nrcre varied prey

availability, and suitable open habitat that facilitates hunting. In Manitoba, the addition of

certain habitat features such as coniferous trees and human structures for roosting is also

important.

The data suggest that screech-owls in Manitoba occur in much lower densities than elsewhere in the geogaphical range, as predicted by the abundant center hypothesis

(Brown 1984, Brown et al. 1995) and also found in other owl species (Sunde et aL 2007), though not always realuledin other birds (Blackburn et al. 1999). This is despite the fact that they produce more young per nesting atterrpt than elsewhere in the range, following a úend of increasing clutch size with increasing latitude (Mrnray 1976). This suggests that wirúer nnrtality is very high in Manitob4 dæ at bast inpart to this species' limited tolerance to cold (Mosher and Henny I976).In this study, 84Vo of swvey detectiors and

83Vo of nests found were in sr¡burban areas (>10 p/ha) and 65Vo of survey detections and

60Vo ofnests found were in nedium - high-dersity suburban areas (>20 p/ha). The range-wide trends disct¡ssed above and the fact that such a large portion of the nests and territories were in suburban areas strongly suggest that the various subrnban advantages discussed above play a significant role in the maintenance of this species in this northern

242 study area, and, given the similarities with other studies (Gehhach 1994), possbly across

the range.

Irwr,rc ¿,rroNs FoR CoNs rnv¿,TIoN AND M ANIcBMEN r

With close to half of the human population now living in cities (United Natiorn 2006),

the eftct of urbanizationon the world's åuna is highly significant and the management

of wildlife in ubanareas is of increasing conservation inportance. Urbanareas are

important both in their ownright for the plants and animals that live with themand also

in terms of the way they are connected to the sr¡rrounding landscape, enhancing or

limiting the dispersal of wildlife (Marzluff 2001,2005). Human activities can make

conunon species rare (or even extirrct) and vice-versa. In addition to threatened species,

species that are common today therefore warant rrnnitoring and research to understand

how human activities canbenefit or disadvantage them. The distribution of the Eastern

Screech-Owl in Winnipeg affords several important lessorn in managerrent to ersure that

cities and towns maintain their biodiversity and reasonably intact food chains.

The Eastern Screech-Owl's stong association with rþarian habitats in the str.dy area

suggests that protecting riparian areas and preventing rernoval of tees from these areas in

cases where new development is planned is key to ensuring a healthy population of

Eastern Screech-Owl in the city. Screech-owls breed in some areas where the riparian buffer súip, e.g. behind a row of houses, is little more than 50m wide, as long as there are

also mature nees in the surounding yards and along streets. The size requirements in

summer of this species appear to be small, most pairs occrpying tenitories of a radius of

243 <500m; however, these birds appear to utilize a much greater area in winter. Following

literature onthe importance of patch size and connectivity (Bunnell 1999, Clergeauet aL

2001, Ahertiand Marzluff 2004, Bierwagen2005, Donnelly and Marzluff2006),

ensuring that riparian corridors are intact on both sides of rivers tlrough the city, with

minimalbreaks in connectivity, would go a long wayto maintaining highbiodiversity

and ersure that owls born in suburban areas have territories they can occupy.

The city of Winnipeg has been dealing with the problemof the loss of Americanelms

(Ulmus americana) due to Dutch elm disease. The nesting data of Eastern Screech-Owls

in Winnþeg suggest that this native riparian cavity species is a key corrponent of the

rþarian community. American elm was in fact the ftee species that was used most often

by nesting screech-owls. The work to preserve the elm in the city is therefore extemely

valuable for wildlife. As a significant cavity-producing tree, the Anerban elm is

important for many species in addition to the Eastern Screech-Owl

Survey data (Chapter 2) indicate that, although screech-owls in moderate density

sr¡burban areas possess many advantages, their population densities rnay peak in the upper

suburban znne of 20 - 30 plln. Unfortunately the survey design in this study proved

insufficient to assess whether screech owls populations may decrease above a certain high human dernity, e.g. 50 pÆla. This raises a fundamental question about how different

levels of human density affect wildlife and how best to optimize space for wildlife in human populatbn centers. At least one pair of Eastern Screech-Owls nested successfully

in anexfterrply densely populated arca(99 p/ha) close to downtownWinnipeg with

244 several high-rise condominiums. Conversely other areas along the very same river, both inside and outside of the city, where trees and shrubs were removed to create a vþw of the water, had no resident pair of screech-owls over the four years of this study. Those urban owls were able to raise young in that area because, in addition to all the impervious surfaces surrounding those high-rises, the area contained a wooded riparian green space and süeets lined with tall, mature native cavity-bearing trees. Other areas, with fewer residents but farther from a river and without a designated greenspace per se, also held a resiCent pair of screech-owls. In these cases, the sfeets and rrnst yards contained mature cavity-bearing trees such that if viewed from the air, they would give the impression of reasonably contiguous canopy cover. In these cases the small interconnected

"greenspaces" that are private yards provided enough food and shelter, as well as a nesting resource (rrnny natual cavitbs) for the owls. This suggests that even dersely populated areas can support native predators if certain spatial requirements and basic habitat needs are rrnt. The simplest recommendation arising from this thesis therefore must be that, even in densely populated cities, rþarian corridors should be protected as greenspaces. This would not prevent humans from using such rþarian corridors for recreation In fact, many have argued that such corridors in cities improve the qu,ality of life or wbanresidents (\Morster 1973, Nieme[á,1999, Shutkin2000). Secondly, greenspaces should be arranged spatially so as to enable wildlife to persist in all areas and designed so as to rnaximize native vegetation rnatching the natr.nal ecosystemof the region Furtherrnore, wherever possible, residential areas should match those areas specifically designed for wildlife in basic habitat characteristics, thereby permitting both occupancy and animal moverrpnts. In a city such as Winnipeg, this would require careful

245 consideratiorL since several ecosystens are, or, at least historically, were present. This thesis has not dealt with native gasslands, which were once present in certainpars of the city, and which also wanant consideration in urban planning.

Mrch recent discussion in ecology has foctsed on the notion of ecologicaltraps and sink populations (Best 1986, Remes 2000, Battin2004, Baker etaL.2005, MannanetaL

2008). It is therefore possible that providing nest-boxes ûtsy entice wildlife into areas where their productivity will be low or where high rrnrtality or parasitism may create siriks. At least in the case of the Eastern Screech-Owl in Wiruripeg, there is no evidence of any decoupling of populationdensity and reproductive success, ie. the areas of highest population dernity also have the highest reproductive success, regardless of whether nest- boxes are supplied or not. Nest-boxes do not appear to be creating a problem for screech- owls and the broods they raise in them are of the sane size as those in natual cavities

(larger broods are produced in boxes on average br-t the difference is not statistically significant). Nonetheless, this research has demonstrated a different problem with nest- boxes, vu. tl:uÍ the corspicuous placement of many boxes in close proximity can lead to increased dump-nesting from Wood Ducks, which even has an impact on screech-owls

(Chapter 5). I would therefore recommend that nest-boxes be npnitored and their positioning on the landscape be corsidered, ideally overseenby a local authority, so as to corrplement the distribution of natural cavities but to minimize the risk of dump nesting

(Serrrel et aL 1988, Semel and Sherman 2001) and interspecific competition (Artuso

2007).

246 Lrrnn¡,runn Crrnn

Alberti, M., and J. M. Marzluff. 2004. Ecological resilience in urban ecosystems: linking

urbanpatterns to humanand ecological functions. UrbanEcosystenrs 7:24I--265.

Andrtsiak, L. 4., and K. M. Cheng. 1997 . Breeding biology of the Barn Owl (Tyto alba)

in the lower mainland of British Columbia. Pages 3846 irz J. R. Duncan, D. H.

Johrsoru and T. H. Nbholls, editors. Biology and conservation of owls of the

Nortlrcrn Hemisphere. North Cental Research StatiorU U.S. Departnent of

Agriculture, St Paul, Minnesota, U.S.A.

Artuso C.2007 . Eastern Screech-owl hatches Wood Duck eggs. Wilson Journal of

Ornitholo gy 1 19 : 1 l0-1I2.

Baker, P. J., A. J. Bentley, R. J. Ansell, and S. Hanis. 2005.Impact of predationby

domestic cats Felís catus tnan urban area. Mammal Review 35:302-312.

Battin, J.2004. When good animals love bad habitats: Ecological traps and the

cornervatio n o f anima I pop ulatio ns. Corservatio n B io lo g y 18 :I 482-149 I.

Belthoff, J. R., and G. Ritchison 1990. Nest-site Selectionby EasternScreech-Owls in

Central Kentucky. Condor 92 :982-990.

Best, L. B. 1986. Corservation tillage: ecological ftaps for nesting birds? Wildlife

Society B ulletrr 14 :3081 17

Bierwager¡ B. G. 2005. Predicting ecological connectivity in urbanizing landscapes.

Environnpnt and Planning B: Planning and Design 32:763176.

BlackburnT. M., K. J. Gaston, R.M.Quinn, and R. D. Gregory. 1999. Do local

abundances of British birds change with proximity to range edge? Journal of

B io geo grap hy 26 :493-505 .

247 Brown, J. H. 1984. On the relationshþ between abundance and distibrÍion of species.

Arrerican Naturalist L24 :255-27 9.

BrowrL J. H., D. W. Mehlrnaq and G. C. Stevens. 1995. Spatial variation in abundance.

Ecology 76:2028-2043.

Bunnell, F. L. 1999. What habitat is an island? Pages 1-31 in J. A. Rochelle, L. A.

Iæhrnanq and J. Wisnbwski, editors. Forest Fragmentation: wildlife and

mana gement imp lications. Brill, Bos to rL Mass achusetts, U. S.A.

Clergear¡ P., J. Jokimäki, and J-P. L. Savard. 200L. Are urban bird communities

influenced by the bird diversity of adjacent landscapes? Journalof Applied

Ecology 38:1122-1134.

Donnelly, R., and J. M. Marzluff. 2006. Relative importance of habitat qu,antity,

sftucttne, and spatial pattern to birds in urbanizing environnents. Urban

Ecosystems 999-117.

ErnlerL J. T. 1974. An urbanbird community in Tucson, Arlzot:nl. derivation, shuctrne,

regulation Condor 7 6 :184-197 .

Environnent Canada. W'eather Pages. Retrþved September 2OO4 - December 2008 from:

http ://weathero ffice. ec. gc.ca.

Gehhach F. R. 1994. The Eastern Screech Owl: Life history, ecology, and behavior in

the sububs and countryside. Texas A&M University Press, College Station, Texas,

U.S.A.

Gehhach, F. R. 1995. Eastern Screech-O wl(Otus asio). Account 165 inA. Poole and F.

Gill, editors. The Birds ofNorthAmericu. ffr. Academyof NatwalScierrces,

248 Philadeþhia, Pennsylvania and The American Omithologists' Unio4 Washinglon,

D.C., U.S.A.

Korpim¿iki, E. 1978. Breeding biology of the Starling Stunw vulgaris in western

Finland. Ornis Fennica 55:93-104.

Mannan, R. W., R. J. Steidl, and C. W. Boal. 2008. Identifying habitat sirks: a case study

of Cooper's hawks in an urban environnent. Urban Ecosystens 11 :141-148.

Marzluff, J. M. 2001. Worldwide urbanization and its effects onbirds. Pages 19-38 inJ.

M. Marzluff, R. Bowmar¡ and R. Donnelly, editors. Avian ecology and

cornervation in an urbanizing world. Kluwer Academic, Boston, Massachusetts,

U.S.A.

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

colonization may determine biological diversity in human-dominated landscapes.

Urban Ecosystems, 8 :I57 -L7 7 .

MarzluffJ. M., R. Bowmar¡ and R. Donnelly. }OOL. A historicalperspective on urban

bird research: trends, terms, and approaches. Pages L-17 in J. M. Marzluff, R.

Bowmarl and R. Dorurelly, editors. Avian ecology and conservation in an

ubanizing world: Kluwer Academic, Boston, Massachusetts, U.S.A.

Møller, A. P. 1988. Nest predation and nest site choice in passerine birds in habitat

patches of different size; a study of magpies and blackbirds. Okos 53:21512I.

Mosher J. 4., and C. J. Herury. 1976. Thermal adaptiveness of plumage color in Screech

Owls. Auk 93:614419.

Murray G. A. 1976. Geogaphic variation in the clutch-sizes of seven owl species. Auk

93:602413.

249 Niemelä, J.1999. Ecologyand urbanplanning. BiodiversityandConservationS:11-131.

Parris, K. M., and D. L.Haznll.2005 Biotic effects of climate change in urban

environnents: the case of the gey-headed flying-fox(Pteropus políocephalus) tn

Me ho urne, Austra lia. B io lo gica I Conservatio n 124 :267 17 6.

Rerrrcs, V. 2000. How can maladaptive habitat choice produce souce-sirk population

dynamics? O fto s 9l :57 9-582.

Rollinsoru D. J., and D. N. Jones. 2002. Variation in breeding parameters of the

Australian mag¡lie Gymnorhina tibicen in suburban and rural environrrents. Urban

Ecosystems 6257169.

Semel, 8., and P. W. Sherman 200I.Intraspecific nest parasitismand nest-site

conpetition in Wood Ducks. Animal Behavior: 6l:787-803.

Semel, 8., P. W. Sherman, and S. M. Byers. 1988. Effects of brood parasitism and nest-

box placennnt on Wood Duck breeding ecology. Condor: 90920-930.

Shochat, E., S. B. Iærmarl M. Katti and D. B. Iæwis. 2004. Linking optimal foraging

behavior to bird community sftucture in an urban desert landscape: Field experiments

wíh artific ial food patches. The American Natwa list 164 :232142.

Shochat, E., Warren, P., FaetlL S., Mclntyre, N., Hope, D. 2006. From patterns to

energing processes in mechanistic uban ecology. Trends in Ecology and Evolution

2l:186-191.

Shutkin, W. A. 2000. The land that could be. MIT Press, Cambridge, Massachusetts,

U.S.A.

250 Sonerud, G. A. 1985b. Risk of nest predation in three species of hole nesting owls:

influence onchoice of nesting habitat and incubationbehaviour. Ornis

S cand inavic a 16 :26 1169 .

Sunde P., K.Overskaug, J. P. Bolstad, and I. J. Oþn 200I. Living atthe limit: Ecology

and behaviour of Tawny Owls Stix aluco in a northern edge population in cental

Norway. Ardea 89 495-508.

United Nations Habitat (United Nations Hurnan Settlenent Programme).2006.

Urbanization: Facts and Figures. Retrieved September 2007 from:

http J I wwZ. unhab iø t. or g/ is ta nb ul+5 /b g 1 0. htm

Worster, D. editor. 1973. Ãmerican environmentalism: the formative period, 1860- 1950.

Wiley, New York, New York, U.S.A

Yetter A. P, S. P. Havera, and C. S. Hine. 1999. Natural-cavity use by nesting Wood

Ducks in lllinois. Jownal of Wildlife Management 63:630-638.

251 8. Appendices

AppBNox 1: knv lrBrts

Prey by species of Eastern Screech-Owl from March 2004 - February 2008. The selectivity index (s) for certain species of birds are also given Prey Iúem Total Captured Biomass Bnns n=344 96019 Passerines 282 (82Vo) 5984s (627o) Non-passerines I I (3Vo) l44ts(lSVo) Unidentified bird 5l (15Vo) 2U3g (237o) Unidentif þd sma ll passerine % (267o) 17829 (l9%o) House Sparrow 43(137o)s=-0.2 12Mg(l3%o) Unidentifbd sparrow or finch 33 (lÙVo) 662eQVo) Black-capped Chbkadee 3O(9Vo)s=-0.2 33Og (37o) Unidentifþd warbler t9 (6Vo) l90g (2Vo) Cedar Waxwing 9(3Vo)s=-0.5 2889(3Vo) Yellow-rumped Warbler 9(3Vo)s=-0.1 llIeIVo) American Redsart 7 (2Vo) s=0.8 589 (I7o) Yellow Warbler 6(27o)s=0.3 579 (IVo) House Finch and Carpodncus sp 6(2Vo)s=0.1 1289(IVo) White-bre aste d Nuthatch 4(l7o)s=-0.8 849 (t%o) Rock Pþon 4(IVo)s=-0.4 10809 (IlVo) American Robin 3(IVo)s=-0.9 2319(2Vo) Cathnrus sp 3(I7o)s=-0.1 93eIVo) Hairy Woodpecker 3(IVo)s=0.2 l98g(2Vo) Downy V/oodpecker 3(IVo)s=-0.6 SleQVo) Tennessee Warbler 3(IVo)s=0.2 30g(0.3Vo) Northern Shrike 2 (IVo) r30g (IVo) Blue Jay 2(l%o)s=-0.5 t70g(27o) American Goldfinch 2(IVo)s=-0.7 269(0.3Vo) European Starling 2(IVo)s=-0.7 I&g(2Vo) House Wren 2(I7o)s=0.5 22s(0.2Vo) Dark-eyedJunco 2(l%o)s=-0.6 38e(0.4Vo) Virginia Rail | (0.3Vo) 859(|Vo) White-throated Sparrow l(0.3Vo)s=-0.3 26e(0.3Vo) Baltimore Oriole I(0.3Vo)s=0.2 33e(0.3Vo) Pine Grosbeak I (0.37o) 569 (l%o) Rose-breasted Grosbeak I(0.3Vo)s=-0.3 45g(0.57o) Regulus sp l(0.3Vo)s=-0.4 6e(0.17o)

Mamrars n= 429 176989 Rodmt 418 (97Vo) 172769 (987o) Non-rodent 11 (3Vo) 422s (2Vo) Meadow vole and unidentified voþ 292 (68Vo) 143089 (817o) House mouse 55 (137o) lI28g(6Vo) Unidentif þd small mamma I 26 (6Vo) 78Og (4Vo) Deermouse 24 (6Vo) 48Os(37o)

252 Unidentifþd rodent 2l (5Vo) 5809(3Vo) Little brown bat 6 (l%o) 24Og(IVo) Short-tailed shrew 4 (l%o) 829(0.5Vo) Red squinel (young) I (0.2Vo) lffig(lVo)

AwTTmIINS AND FISH n=8 119g Unidentified frog 3 (38Va) 459(38Vo) Leopard frog 2 (25Vo) 4Og(34Vo) Wood frog 2 (25Vo) 24eQ07o) Unidentifþd fish I (13Vo) l0g (8Vo)

INvontmnlrns n=l*2 3277g Hard (detectable in pellets) 690 (45Vo) 14259(43Vo) Soft (not detecnble in pellets) 179 (12Vo) 3389(lÙVo) (467o) U n id. en tif ied inv e rt eb rat e 673 (447o) 1514g Beetle: Scarabidae 554 (367o) 11089 (34Vo) Beetle:family unknown 76 (5Va) r52g(57o) Beetle: Carabidae 29 (ZVo) 589(2Vo) Beetle: Staphylinidae 3 (0.2Vo) 69(0.2Vo) Beetle: Coccinellidae 3 (0.27o) 3g(O.IVo) Earthworm (mostly Lumb r icus te rre stris) 79 (5Vo) t58g (57o) Caterpillar sp 56 (4Vo) ll2g(3Vo) Moth (mostly Noctuidae) 36 (27o) 54g(2Vo) Winged irsect 12 (l%o) 42eQ%o) Crayfish sp 7 (0.57o) 379(lVo) Dragonfly sp 5 (0.37o) r5g(0.57o) Larvae or pupae 4 (0.37o) 8e(O.ZVo) Slug sp 2 (O.IVo) 29(0.17o) Spider sp 2 (0.17o) 49 (O.l%o) Cockroach sp I (O.IVo) 4g (0.1E") Twenfy "Mlrr"t"i .p" are all rnost lkely to be Meadow voles hence combined. Beetle identifications are imperfect and the number of Scarabidae may be inflated due to other families with similar head or leg shapes. Most of the unidentified invertebrates were observed and were rmst lkely small beetles or other very small prey. The rnost comnrcn Scarabidae prey are in the genus Phyllophaga, the fixost common Carabidae prey are in the genus Calosoma.

253