UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Associations of songbirds with their habitat in the Garry Oak (Quercus garryana) ecosystems of Victoria,… Feldman, Richard E. 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2003-0006.pdf [ 3.37MB ]
JSON: 831-1.0075092.json
JSON-LD: 831-1.0075092-ld.json
RDF/XML (Pretty): 831-1.0075092-rdf.xml
RDF/JSON: 831-1.0075092-rdf.json
Turtle: 831-1.0075092-turtle.txt
N-Triples: 831-1.0075092-rdf-ntriples.txt
Original Record: 831-1.0075092-source.json
Full Text

Full Text

ASSOCIATIONS O F SONGBIRDS W I T H T H E I R H A B I T A T IN T H E G A R R Y O A K (QUERCUS GARRYANA) E C O S Y S T E M O F VICTORIA, BRITISH C O L U M B I A , C A N A D A by Richard E. Feldman B.Sc, Queen's University, Kingston, 1999  A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF M A S T E R OF SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES THE F A C U L T Y OF F O R E S T R Y Department of Forest Resources Management Centre for Applied Conservation Research  We accept this thesis as.conforming to the required standard  THE UNIVERSITY OF BRITISH C O L U M B I A October 2002 © Richard E. Feldman,  2002  In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, 1 agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department  or by his  or her representatives.  It is  understood  that  copying or  publication of this thesis for financial gain shall not be allowed without my written permission.  Department of  fafesV He&ift<rce& M<sHtA  The University of British Columbia Vancouver, Canada  Date  DE-6 (2/88)  QcY  Z9  .1^C*~L  ABSTRACT Urbanization fragments habitat, which affects bird abundance and distribution. This study tested the effect of urbanization on birds in fragments of Garry oak (Quercus garryana) and Douglas-fir (Pseudotsuga menziesii) ecosystems in Victoria, British Columbia. Birds were surveyed along transects in seven Garry oak patches in May and June of 2000. The patches were stratified by surrounding habitat type; four were surrounded by coniferous forest and three were surrounded by urban development. Birds were also surveyed in these adjacent habitats. Relative species abundance was determined using point counts along transects and territory mapping. Associations between bird abundances and urbanization in the landscape, patch size and Garry oak stem volume were tested. Linear regression models related environmental variables to the abundances of 17 common birds. The best model was chosen via Akaike's Information Criterion. For 42 birds, relationships between species composition, adjacent habitat type and vegetation composition were analyzed with detrended and canonical correspondence analysis and cluster analysis. For the 17 common species, 10 were strongly associated with the level of urbanization in the surrounding landscape. Seven species were associated with Garry oak stem volume, three positively and five negatively. Garry oak patches surrounded by urbanization had substantially different species assemblages than those surrounded by coniferous forest, in part due to the influence of coniferous trees in Garry oak patches. The results indicate that the level of urbanization in the broader landscape and tree composition in patches are significant factors associated with Garry oak bird assemblages. Conservation of Victoria's songbirds requires conserving and restoring a spectrum of habitats, from open savanna to closed-canopy coniferous forests. Management should be targeted at the landscape-scale and explicitly consider the extent of urbanization in the landscape.  ii  TABLE OF CONTENTS ABSTRACT T A B L E OF CONTENTS LIST OF TABLES LIST OF FIGURES PREFACE ACKNOWLEDGEMENTS  II III IV V VI Vll  CHAPTER 1  I  CHAPTER 2  5  ABSTRACT INTRODUCTION METHODOLOGY  5 6 8  STUDY SITES ENVIRONMENTAL MEASURES BIRD SURVEYS D A T A ANALYSIS  8 9 11 14  RESULTS DISCUSSION  15 18  CHAPTER 3  24  ABSTRACT INTRODUCTION METHODS  24 24 26  STUDY SITES BIRD SURVEYS VEGETATION SURVEYS D A T A ANALYSIS  26 26 27 28  RESULTS  30  BIRD SPECIES COMPOSITION VEGETATION COMPOSITION BIRD - HABITAT RELATIONSHIPS COMMUNITY COMPOSITION  30 30 34 42  DISCUSSION  42  CHAPTER 4  50  M A N A G E M E N T RECOMMENDATIONS  52  LITERATURE CITED  56  APPENDICES  63 iii  LIST OF TABLES Table 2.1: Characteristics of each site  12  Table 2.2: Best linear regression equation describing the relationship between bird abundance and environmental variables  16  Table 3.1: Tree species composition of transects in Garry oak sites  31  Tab;e 3.2: Understory composition of transects in Garry oak sites  32  Table 3.3: Results of Detrended Correspondence Analysis of bird species among Garry oak, coniferous and urban samples 41 Table 3.4: Correlation of each vegetation variable with the first and second axes derived from Canonical Correspondence Analysis 41 Table 3.5: Results of Canonical Correspondence Analysis of bird species and vegetation variables among Garry oak samples  41  iv  LIST OF FIGURES Figure 1.1: The range of Garry oak in North America and British Columbia  2  Figure 2.1: The study area (Victoria, BC)  10  Figure 3.1: Detrended Correspondence Analysis of bird species in Garry oak, coniferous and urban habitats  35  Figure 3.2: Canonical Correspondence Analysis of bird species and vegetation variables in Garry oak sites  38  Figure 3.3: Cluster analysis based on Jaccard's similarity index of bird species composition of Garry oak, coniferous and urban sites  43  Figure 3.4: Scatterplot of Jaccard's community similarity index versus distance (km) of all Garry oak, coniferous and urban site pairs  45  v  PREFACE Some of the data in this thesis has been previously published: Feldman, R. E. and P. G. Krarinitz, 2002. Does habitat matter in an urban landscape? The birds of the Garry oak (Quercus garryana) ecosystem of southeastern Vancouver Island, pp. 169 - 178 in Standiford, R. B., D. McCreary and K . L . Purcell, tech. coords. Proceedings of the fifth symposium on oak woodlands: oaks in California's changing landscape. 2001, Oct. 22-25, San Diego, C A . Gen. Tech. Rep. PSWGTR-184, Albany, C A , Pacific Southwest Research Station, Forest Service, US Department of Agriculture, 846 p. Feldman, R. E. and P. G. Krannitz, 2002. The birds of the Garry oak ecosystem of southeastern Vancouver Island, pp. 88 - 94 in Burton, P. J., ed. Garry oak ecosystem restoration: progress and prognosis. Proceedings of the Third Annual Meeting of the B.C. Chapter of the Society for Ecological Restoration, April 27-28, 2002. University of Victoria. BC. Chapter of the Society for Ecological Restoration, Victoria, B C , 109 p.  vi  ACKNOWLEDGEMENTS Funding for this project was supplied by the Georgia Basin Ecosystem Initiative of Environment Canada, Canadian Wildlife Service, the Nature Conservancy of Canada (thank you Tim Ennis), Science Horizons and the British Columbia Conservation Foundation. A tremendous thank you goes to my supervisor, Pam Krannitz. She was consistently available during the course of this thesis to answer questions and provide input. Numerous opportunities were and are available to me because of her affiliation with the Canadian Wildlife Service. Many key suggestions on this and earlier drafts were made by my committee. Past and present members include Jamie Smith, Brian Klinkenberg, John Innes, Sue Glenn and Cindy Prescott. I thank them for contributing their time to this project. I especially thank Jamie for being thorough, thought-provoking and challenging me to be my best. Peter Arcese also contributed helpful comments. Much of my love is devoted to my immediate family: Mom and Arnie, Dad, Becky and Jordan. Your past, present and future love is my guide. Much of this project could not have been done without Katie Christie, my field assistant but also a great friend. Because of her, I effortlessly learned my bird songs. Because of her, my time in Victoria was some of the best of the past few years. I also thank the many others who were a part of my Victoria life. Velvet, the Walbran and the Parse will forever be with me. In Vancouver, the "blender" gang and others at U B C have become cherished friends. Of these I would like to single out Jim Herbers who early-on became a friend and mentor and helped with some philosophical and statistical aspects of this thesis. Most of my education over the past few years has happened outside of school. For this I have to thank above all else, Gennady and Naz. I love you! Your support has been life-saving. Andrew and Chris have also been invaluable friends. Finally, this thesis is dedicated to Marilyn Fuchs, coordinator of the Garry Oak Ecosystems Recovery Team, the volunteers of the recovery team and its affiliated groups and anyone who has given their energy to the conservation of the biodiversity and beauty of Garry oak ecosystems.  CHAPTER 1 G E N E R A L INTRODUCTION The Garry oak ecosystems of southeastern Vancouver Island and the Gulf Islands are considered one of the most endangered habitats in British Columbia (Erickson 1993). Natural patchiness, a limited range and being at the northern edge of its range combined with on-going habitat loss and degradation threaten the ecosystem and its biota. Garry oak is a biologically rich habitat with many species found nowhere else in British Columbia (Fuchs 2001). Ninety species occurring in Garry oak ecosystems are considered at risk (Garry Oak Ecosystems Recovery Team 2002), 21 of which have been listed nationally by the Committee on the Status of Endangered Wildlife in Canada (COSEWIC 2002). The British Columbia government has 60 species associated with Garry oak ecosystems on its red (endangered) and blue (threatened) lists (URL: http://srmwww.gov.bc.ca/atrisk). Garry oak ranges from northern California to Vancouver Island and the Gulf Islands, British Columbia (Figure 1.1a, b). It is the dominant oak in Oregon and the only native oak species in Washington and British Columbia. Garry oak grows in the rainshadow of the Cascade, Olympic and Vancouver Island mountain ranges (Erickson 1993). Dry conditions distinguish Garry oak from the wetter temperate forests common to the Pacific Northwest coast. In Oregon, Garry oak primarily occurs in the Willamette valley. Habitat ranges from savanna to forest habitat depending on tree size and density and whether the understory is shrub or herb dominated (Thilenius 1968). In Washington, Garry oak is ecotonal between coniferous forest and prairie (Rogers 2000, Thysell and Carey 2001). British Columbia's Garry oak includes coastal bluffs and rocky outcrops on  1  Figure 1.1: The range of Garry oak in (a) North America and (b) British Columbia. Maps reprinted from Erickson (1993).  slopes with well-drained shallow soils (Roemer 1972, Erickson 1996). On flat ground with deep soil, Garry oak occurs in savanna or woodland (Roemer 1972, Erickson 1996). It occurs in isolated stands, surrounded by dry Douglas-fir {Pseudotsuga menziesii) forest or in transitional forests mixed with Douglas-fir and arbutus (Arbutus menziesii). Garry oak habitat has always been closely associated with human use and modification. Periodic burning by First Nation's peoples provided camas (Camassia spp.) bulbs as a steady food source and effectively maintained open canopies and grassland understories (Thilenius 1968, Beckwith 2002). European settlement introduced exotic species and changed disturbance regimes by suppressing fires and introducing grazing (Hebda 1993). Habitat loss and fragmentation has been the most dramatic ecological change. Over the past 100 years, greater Victoria has grown from 23 688 residents to 311 902 (Canada Census Office 1902, Statistics Canada 2002). During this period, Garry oak habitat has been reduced from 10 510 ha to 512 ha, a loss in cover of 95% (Lea 2002). Most of the threatened Garry oak species are vascular and non-vascular plants. Unique plant associations result from specific soil and disturbance conditions (Roemer 1972, Erickson 1996). Garry oak habitat is also important to wildlife, including birds. In the Garry oak of Oregon, shrub-associated and mid-canopy deciduous species were more abundant than in nearby coniferous forests (Hagar and Stern 2001). Overall more neotropical migrant species were found in oak woodlands than nearby coniferous forests (Anderson 1972, Hagar and Stern 2001). Garry oak bird composition is associated with successional changes in habitat composition and structure. Grassland/meadow type habitat is uniquely associated with streaked horned lark (Eremophila alpestris strigata), vesper sparrow (Pooecetes gramineus affinis) and western meadowlark (Sturnella  3  neglecta)  (Campbell  et al.  1997, 2001, Rogers 2000). Savannas with large, well-spaced  trees are habitat for Lewis' woodpecker (Melanerpes mexicana)  and other cavity-nesters (Campbell  et al.  lewis),  western bluebird  (Sialia  1990, Rogers 2000, Hagar and Stern  2001). Without fire or other disturbances, Garry oak density increases (Tveten and Fonda 1999, Thysell and Carey 2001). Woodland habitats are associated with deciduous tree and shrub birds (Hagar and Stern 2001). In some woodlands, Douglas-fir establishes and promotes succession to coniferous forest (Barnhardt et al. 1987). Forest birds are common in this habitat type (Anderson 1970, Hagar and Stern 2001). Garry oak bird assemblages are also expected to be related to landscape composition and structure, though empirical evidence is lacking. Hagar and Stern (2001) remark that an oak stand closest to an urban area had a bird community radically different than other stands. Birds are sensitive to habitat composition at multiple scales (Wiens et al. 1987, Saab 1999). Landscape composition is associated with bird abundance and distribution in other oak ecosystems (Sisk et al. 1997, Hannson 2000). Much of British Columbia's Garry oak are in highly fragmented urban and agricultural landscapes. Human induced changes to the landscape alter processes that affect bird abundances (e.g. predation rates, dispersal frequency [Fahrig and Merriam 1994]). The purpose of this thesis is to describe associations of the spring bird assemblage of Victoria's Garry oak ecosystem with habitat and landscape features. I hypothesize that urbanization in the landscape (approx. within 2 km of the oak stand) is associated with the abundance of individual species. I will test this hypothesis by describing the relative influence of urbanization in the landscape, patch size and Garry oak stem volume on the abundance of birds inhabiting Garry oak and coniferous habitat patches. I also hypothesize that the type of adjacent habitat is associated with bird composition of Garry  4  oak patches. I will test how associations with adjacent habitat are related to vegetation composition. I will also compare Garry oak bird assemblages to those in the adjacent habitat. The results of this thesis are intended to provide basic information on the Garry oak avifauna necessary for managing British Columbia's Garry oak ecosystems.  CHAPTER 2 T H E R E L A T I V E I N F L U E N C E O F URBANIZATION, P A T C H SIZE AND G A R R Y O A K S T E M V O L U M E T O BIRDS I N G A R R Y O A K A N D D O U G L A S FIR H A B I T A T P A T C H E S ABSTRACT In this study I tested whether bird abundance in remnant patches of Garry oak (Quercus garryana) and Douglas-fir (Pseudotsuga menziesii) was related to Garry oak stem volume, patch size and urbanization. Breeding bird populations were surveyed at seven Garry oak sites and four adjacent coniferous sites. Relationships between environmental variables and abundance of 17 species of birds were inferred by selecting the best linear regression model by the Akaike Information Criterion. Though individual species associations vary, human population density in the landscape was associated with the most species. Six species were uniquely associated with human population density in the landscape, four of which were negatively associated with urbanization. Four species were negatively associated with urbanization and with Garry oak stem volume. Three species were positively associated with Garry oak stem volume. One species was more abundant in larger patches, irrespective of habitat type. Four species were not associated with the measured variables. Given these associations between bird abundance and urbanization, future urban development is likely to affect bird composition of Garry oak remnants.  5  INTRODUCTION Urban habitat patches lie within a matrix that differs in structure, composition and function (Savard et al. 2000, Marzluff and Ewing 2001). Urban development around forest patches has been associated with negative edge effects and increased patch isolation (Wilcove 1985, Fernandez-Jurcic and Jokimaki 2001). Urbanization is correlated to bird abundance and presence (e.g. Bolger et al. 1997, Rottenborn 1999) and can cause local extinction (Soule et al. 1988). The extent to which the structure, composition and function of a habitat patch differs from its matrix influences bird abundances in fragmented landscapes (Fahrig and Merriam 1994, Wiens 1994). In a highly urbanized matrix, remnant habitat patches are analogous to islands (Fernandez-Jurcic and Jokimaki 2001). Positive associations between species richness and individual species abundances and patch size are evident (Tilghman 1987, Fernandez-Jurcic 2001). In most North American cities, the degree of urban development varies along an urbanization gradient (McDonnell and Pickett 1990). Along parts of the gradient, patches may not function as islands. Landscape composition may be better associated with changes in bird communities than patch size (Wiens 1994). For example, substantial differences in the abundance and distribution of birds were found between urban and suburban sections of an urbanization gradient in California (Blair 1996). Most studies of the effects of fragmentation on birds in the Pacific Northwest have been conducted in forested landscapes where fragmentation is a result of forest harvesting (e.g. McGarigal and McComb 1995). Generally, most species are not associated with patch size or landscape composition (McGarigal and McComb 1995, 6  Shieck et al. 1995). Instead, bird abundances are related to vegetation composition and structure (Hansen et al. 1995, Chambers et al. 1999). For example, open and closed canopy stands have different bird communities (Beese and Bryant 1999, Chambers et al. 1999). In eastern North America, there is stronger evidence for an influence of patch size and landscape composition on bird composition and abundances (e.g. Blake and Karr 1987). Fragmentation effects may not occur in western forests because fragmentation by forestry does not create a strong and permanent contrast between remnant and matrix habitat structure (Hansen et al. 1991). In Pacific Northwest cities, landscape structure influences bird distribution and abundance (e.g., Melles 2000). Increased urbanization is associated with a shift from species rich assemblages to those dominated by a few urban and non-native species (Gavreski 1976, Lancaster and Rees 1979, Melles 2000). Furthermore, many species present in large and non-urbanized forests drop out as urbanization increases. Similar results have been found in other regions (e.g. Beissinger and Osborne 1982, Blair 1996). The Garry oak ecosystem of greater Victoria, British Columbia is highly fragmented, consisting of patches isolated within an urban and suburban matrix (Hebda 1993, Fuchs 2001). Garry oak habitat co-occurs with Douglas-fir dominated forest remnants. Associations of birds with these habitats has been explored in Oregon (Anderson 1970, 1972, Hagar and Stern 2001). Associations with landscape composition and patch size have not been investigated previously. For this study, I surveyed the bird assemblages of Garry oak and coniferous patches along a gradient of human population density. I tested the relative strengths of association between the abundance of common species and human population density in the landscape, patch size and Garry oak stem volume.  7  METHODOLOGY Study sites The city of Victoria, British Columbia (48° 25' N 123° 19' W; pop. 311 902 [Statistics Canada 2002]) is located in the coastal Douglas-fir biogeoclimatic zone (Green and Klinka 1994). The zone is characterized by a Mediterranean-like climate. Mean daily temperatures range from a high of 15.6°C in July to a low of 5.0°C in January. Average annual precipitation is 607.6 mm with a drought in summer (Environment Canada U R L : http://www.msc.ec.gc.ca/climate/climate_normals/ results_e/cfm). This is the only region in British Columbia in which Garry oak ecosystems occur. Garry oak patches in this study occur on hill slopes with rocky soil and deeper-soiled flat sites. Garry oak is the dominant canopy tree in the patches. Arbutus and Douglas-fir are also present. Common native shrubs at my study sites are snowberry (Symphoricarpos albus) and oceanspray (Holodiscus discolor). Exotic shrubs include Scotch broom (Cytisus scoparius) and Himalayan blackberry (Rubus discolor). Patches of forbs and grasses are also found. Common camas (Camassia quamish) and Idaho fescue (Festuca idahoensis) are two of the abundant native species (Erickson 1996). Non-native grasses such as orchard grass (Dactylis glomeratd) and Kentucky blue grass (Poa pratensis) can also dominate. Adjacent to some Garry oak patches are Douglas-fir dominated forests. The forests have greater canopy closure and are cooler and wetter than Garry oak. Grass is nearly absent from the understory, which is instead dominated by salal (Gaultheria shallon) and sword fern (Polystichum munitum) (Pojar and MacKinnon 1994). The Sensitive Ecosystem Inventory is a GIS database of ecological attributes of southeastern Vancouver Island and the Gulf Islands of British Columbia (Canadian 8  Wildlife Service 1997). I queried the GIS to find sites with the following criteria: 1) patches classified as being dominated by Garry oak; 2) patches > 4 ha and 3) sites that were easily accessible. Seven sites met the criteria. They lay along an urban-rural gradient emanating from the core of the city of Victoria (Figure 2.1). Sites differed in size and level of development in the surrounding landscape as measured by human population density. Four of the seven Garry oak patches were situated adjacent to Douglas-fir dominated habitat. Since each Garry oak - coniferous pair shared the same landscape, landscape differences among the two habitat types were minimized. Bird communities were surveyed in the Garry oak and coniferous study sites. The sites are all protected. Five are city and regional parks and two (Mary Hill and Rocky Point) are owned by the Canadian Department of National Defence. They are all subject to human disturbances such as hiking trails and dog-walking. Environmental measures  The study site locations were entered into the Sensitive Ecosystem Inventory GIS database (Canadian Wildlife Service 1997). The area of the polygon containing the study site and the areas of all adjacent polygons of the same tree cover class (Garry oak or Douglas-fir) were summed to derive overall patch area. The GIS cover polygons were derived from air photos taken between 1989 and 1992. Human population density was measured to reflect the degree of urbanization. Population data was taken from the 1996 Canada Census . Enumeration areas (EA's) are 1  the smallest population unit. Study site locations were added to a GIS map of E A ' s . Two km radius buffers (approx. 1256 ha) were drawn around the study sites.  ' Population data obtained using GeoSuite 1.00 program downloaded from University of British Columbia Library Numeric Data Services web site (URL: http://fosthall.library.ubc.ca) Enumeration areas imported as Arc View polygons; downloaded from University of British Columbia Library Numeric Data Services web site (URL: http://fosthall.library.ubc.ca) 2  9  Figure 2.1: The study area (Victoria, BC). Seven Garry oak patches were chosen for this project. Rocky Point, Mary Hill, Thetis Lake and Mt. Douglas patches were adjacent to Douglas-fir forest while the remaining three were completely surrounded by city. Bird surveys were conducted in both the Garry oak and Douglas-fir habitats. Map courtesy Sensitive Ecosystem Inventory (Canadian Wildlife Service 1997).  A l l E A ' s where at least 50% of the polygon was contained within the buffer were summed to estimate the total human population for that study site. Water that fell within the buffer was assigned a population of zero and was retained in the calculation of total area. Buffers varied in size for each study site because of the irregular sizes and boundaries of the E A ' s . Total population was divided by total buffer area (average ± standard deviation = 1749.45 ha ± 462.16 ha, n = 11) to obtain human population density. In each site, two transects were established for bird surveys. Each of these transects contained 4-7 bird survey points (see bird surveys below).Tree measurements were made along 50 m transects that ran perpendicular to the bird survey transects at each bird survey point. A l l oaks greater than two metres in height and five centimetres in diameter at breast height within five metres of each side of the transect were counted. Diameter at breast height and height were measured on five randomly chosen individuals on each transect. Stem volume of Garry oak (Table 2.1) was estimated from the equation: log volume (m ) = -4.537 + 1.908 log diameter (m) + 1.12 log height (m) (British 3  Columbia Forest Service 1983). This measure was determined for each oak surveyed, averaged and multiplied by tree density to obtain a volume/ha measurement for each site. Bird surveys At each site, bird surveys were conducted along two parallel transects (Verner and Ritter 1985) separated by 100 m. The transects were placed to maximize the length of the patch while avoiding edges. A l l transects were 400 m in length except Rocky Point (350 m) and the coniferous portion of Mount Douglas (350 m) and Highrock (250 m). Bird survey points were placed every 50 m along the transect. Therefore, most transects  11  to  a s  o  to  R  o  a s o> ON UO oo O 00 © © © ©  o  s  s  O  c o c o OO <N 1/0 00 CO oo o CN © © ©  cR  CN CN CN  to  3 O  o O  r-in  UO  © CN >o CN CM CN l O — < CN -3-  •  i  i  o CN  <3  <3  _<° 00  o  CN  in  o  r-- CO o> o CO  CN  to  3  8 s o  o  o  3£  <3  1-  ^ m  MD  CO  ,—<  —H  in  Tf  —H  ON CN CN  CO  CO  o  00 CN  —<  DOS  O  cd  CO  —<  1/0  in  r—<  —<  CN CO  o o  g 3*^2 O J CO  00 .  5 1/3 °  contained seven points. The 350 m transects contained six points and the 250 m transect contained four points. Bird surveys were conducted between dawn and four hours after dawn from 5 May to 5 July 2000. Each site was surveyed by two observers. At first, the surveys were conducted together to ensure consistent reporting. After, each person separately surveyed a different transect. The order of site visits and the transect surveyed by a particular observer were randomized. During the survey, the observer walked the transect (starting point was randomized) and stopped at each bird observation point for five minutes. During the point-count all birds seen and heard within a 50 m radius of the point were recorded on a map of the site. Each notation on the map consisted of the American Ornithological Union four letter-code for the species, whether the bird was singing, calling or spotted visually and the distance and position of the individual relative to the observer. (The majority of detections were of singing males). Five visits were made to Mary Hill, Rocky Point, Mount Douglas and Thetis Lake while the remaining sites were visited four times. Visits to the same site were separated by at least five days. Surveys were not conducted when precipitation was more than a light drizzle. Upon completion of the surveys, every visit for every site had its own map of all species. These maps were compiled into maps of individual species for each site (Bibby et al. 1992). Every location for a bird was mapped according to a letter assigned to each visit. Two or more different letters in a cluster was assumed to represent repeated observations of the same individual. The cluster was considered an individual's territory (Bibby et al. 1992). The number of territories for each species was converted to a density value by dividing by the area surveyed. The index of abundance used in data analysis was territory density (# territories/ha).  13  Data analysis  A l l species that I detected with territories were used for calculations of species richness. Analysis of individual species was performed for species that were: 1) non flocking, so that the territory mapping method was appropriate and 2) relatively abundant, with occurrences in at least five patches. Simple and multiple linear regression models were used to relate the response variable (bird territory density) to one or more predictor variables (Garry oak stem volume [ O A K V O L ] , patch size [AREA] and human population density [HUMANS]). Each regression model was ranked by Akaike's Information Criterion with a correction for small sample size (AICc; Burnham and Anderson 1998). AICc was calculated from the maximum log-likelihood of the regression model (Burnham and Anderson 1998, StatSoft Inc. 2000). Log-likelihood, itself, was derived from the regression sums of squares. AIC/AICc selects the best approximating model for the data, considers parsimony in model-selection and does not depend on null hypothesis testing. Included in the series of models developed to explain bird abundance was a "null" model. This predicted abundance from mean population size and variance. A l l variables were logtransformed to normalize their distributions and improve homoscedasticity. The series of models for each species were ranked in order from lowest to highest AICc score. Model weights and parameter importance values were also calculated. In most cases the model with the lowest AICc value was chosen as the model with which to infer bird abundance relationships (Burnham and Anderson 1998). More than one model was used to describe relationships when separate models were differentiated by AICc scores less than two and weighted similarly (Burnham and Anderson 1998). For all chosen models, parameter estimates and 95% confidence intervals were calculated. The  estimates were derived from all the models in which a variable appeared and weighted accordingly, following the methods of Anderson et al. (2000).  RESULTS Garry oak patches had more bird species overall than coniferous habitat (37 vs. 32). On average, however, coniferous patches (18 spp.) were slightly more species rich than Garry oak (16.5 spp.). The American robin (Latin names are given in Appendix 1), chestnut-backed chickadee and pine siskin were the only species found at all sites. The most abundant species in the survey was the spotted towhee with 10.52 territories ha"  1  across the landscape and an average of (mean ± standard error of the mean) 0.957 ± 0.281 territories ha" in each patch (n=l 1). The towhee was also the most abundant bird in Garry 1  oak (average density = 1.13 ± 0.303 territories ha" in each patch, n=7). The Pacific-slope 1  flycatcher was the most abundant bird in the coniferous patches (average density =1.11 + 0.239 territories ha" in each patch, n=4). 1  Of the 42 species detected, 17 were used in the regression analysis. Five species were excluded because they were observed in flocks and/or did not have fixed territories (barn swallow, European starling, northwestern crow, pine siskin and red crossbill). Twenty species were found on less than five sites and were excluded for not meeting the minimum frequency criteria. A strength of evidence approach using AICc was used to choose the model in which to infer species - environment associations. The top models for each species and the statistics used to derive strength of evidence are presented in Appendix 2. The best model(s) for each species along with the unconditional parameter estimates are presented in Table 2.2. Some species were associated with more than one variable because 1) the 15  Table 2.2: The best linear regression equations describing the relationship between bird abundance and environmental variables. A l l models for which there is strong evidence for a relationship are shown. Parameter estimates (i.e. regression coefficients) for terms used in the final model were calculated along with the unconditional 9 5 % confidence interval (Anderson et al. 2 0 0 0 ) . Equations without a parameter correspond to a null relationship. Best approximating models  (Intercept ± 95% conf. int.) + (estimate. ±95% conf. int.) parameter  American robin  (0.200 ± 0.092) - (0.044 ± 0.056) humans  Anna's hummingbird  (-0.015 ± 0.078) + (0.051 ± 0.064) humans  Bewick's wren  0.105 ± 0.091  Brown-headed cowbird  (0.023 ± 0.193) + (0.100 ± 0.073) oakvol  brown creeper  (0.333 ± 0.194) - (0.134 ± 0.076) oakvol - (0.160 ± 0.109) humans  chestnut-backed chickadee  (0.330 ± 0.103) - (0.041 + 0.005) oakvol - (0.074 ± 0.036) humans  chipping sparrow  (0.012 ± 0.012) + (0.023 ± 0.035) oakvol (0.069 ± 0.094) - (0.033 + 0.059) area  dark-eyed junco  (0.131 ± 0.017) - (0.107 ± 0.106) humans  golden-crowned kinglet  (0.279 ± 0.183) - (0.099 ± 0.071) oakvol - (0.132 ± 0.102) humans  orange-crowned warbler  (0.136 ± 0.122) - (0.109 ± 0.090) humans (0.135 ± 0.023) - (0.109 ± 0.090) humans + (0.023 ± 0.036) oakvol  Pacific-slope flycatcher 0.156 ± 0.116 red-breasted nuthatch (-0.056 ± 0.080) + (0.080 ± 0.059) area spotted towhee (0.200 ± 0.179) + (0.200 ± 0.173) humans Townsend's warbler (0.110 + 0.089) - (0.104 ± 0.086) humans winter wren (0.251 ± 0.194) - (0.100 ± 0.079) oakvol - (0.093 ± 0.092) humans species richness 17.09 ± 3 . 2 8  16  selected model contained two variables (brown creeper, chestnut-backed chickadee, golden-crowned kinglet, orange-crowned warbler, winter wren) or 2) more than one model was selected as the final model for inference (chipping sparrow, orange-crowned warbler). For two species, the house wren and Wilson's warbler, each single variable model performed equally poorly (high uncertainty and low weights). Consequently a discernible pattern of abundance across fragments for these two species was not evident. Human population density was positively associated with the number of detections of spotted towhee and Anna's hummingbird and negatively associated with detections of American robin, orange-crowned warbler, Townsend's warbler, dark-eyed junco, brown creeper, chestnut-backed chickadee, golden-crowned kinglet and winter wren (Table 2.2). Patch size was positively associated with the number of detections of red-breasted nuthatch and negatively associated with detections of chipping sparrows (Table 2.2). Garry oak stem volume was positively associated with the number of detections of chipping sparrow, orange-crowned warbler and brown-headed cowbird and negatively associated with detections of brown creepers, chestnut-backed chickadees, golden-crowned kinglets and winter wrens (Table 2.2). The null model was the best fit for the number of detections of Bewick's wren and Pacific-slope flycatcher and overall species (Table 2.2). The four species negatively associated with Garry oak stem volume were also negatively associated with human population density. For brown creepers, goldencrowned kinglets and winter wrens, there was greater evidence for an association with habitat than with human population density. The reverse was true for the chestnut-backed chickadee (see parameter importance values in Appendix 2). The chestnut-backed  17  chickadee was more abundant in Garry oak habitat then the other species. Chickadees were also detected in urbanized Garry oak.  DISCUSSION The common birds I detected in Victoria's Garry oak and conifer fragments were associated with urbanization in the landscape and/or with Garry oak stem volume. Associations between bird abundance and habitat structure are well established (James and Warner 1982, Rice et al. 1984). More recent studies have shown statistical relationships between landscape-scale features and bird abundance (e.g. Saab 1999, Haire et et al. 2000). It is now widely recognized that bird habitat selection is influenced by features and processes acting at multiple scales (Wiens et al. 1987, Wiens 1994, Saab 1999). Both habitat and landscape scale relationships exist for birds in oak woodlands. In Oregon, Hagar and Stern (2001) found significant relationships between bird abundances and habitat features such as canopy cover, shrub cover and tree diameter. In California, the matrix habitat type was significantly correlated with bird richness, diversity and abundance (Sisk et al. 1997). Bird associations with both habitat and landscape features have also been observed in other urbanization studies. Rottenborn (1999) found significant relationships between urbanization and abundance for factors measured within 500 m of bird survey sites. He also found that small-scale habitat features (< 35 m from a bird survey station) were significantly related to species richness and abundance. Bolger et al. (1997) found that urbanization was correlated to bird presence/absence at scales ranging from 250 m to 3000 m from a survey site. For most birds, patterns of abundance were best explained by  18  logistic regression models containing small-scale (< 100 m) habitat cover and large-scale landscape composition variables. I used human population density to indicate the degree of urbanization in the landscape. Human population reflects a suite of factors associated with urbanization including cat density, movement barriers, human intrusion into habitat and traffic (Rottenborn 1999). Human population density is likely to be correlated with housing density. Several studies have shown significant associations between bird abundance and housing density or proximity to buildings (Friesen et al. 1995, Germaine et al. 1998, Rottenborn 1999). Urbanization is also associated with an increase in exotic vegetation cover, which is negatively associated with the abundance of native bird species (Bolger et al. 1997, Germaine et al. 1998, Rottenborn 1999). A bird's sensitivity to urbanization is likely related to its foraging and/or nesting requirements. In Vancouver, urbanization was associated with a decrease in the abundance of shrub and cavity nesting birds (Melles 2000). A shift in composition from foliage and bark gleaning insect eaters to ground foraging seed eaters is associated with increasing urbanization (Beissinger and Osborne 1982, Blair 1996, Rottenborn 1999). Urbanization is also associated with an increase in the abundance and dominance of nonnative species (Mills et al. 1989, Germaine et al. 1998). Of the eight species that I found to be negatively associated with urbanization, all were insect eaters (Ehrlich et al. 1988). However I did not find a concomitant increase in non-native species or ground foraging seed eaters. The two species I found to be positively associated with urbanization were the spotted towhee and Anna's hummingbird. Both species probably benefit from the increased food resources associated with urban areas such as ornamental shrubs and bird 19  feeders (Inouye et al. 1991, Jokimaki and Suhonen 1995). In California and Arizona, Anna's hummingbirds were positively related to exotic vegetation and urban development (Blair 1996, Bolger et al. 1997, Germaine et al. 1998). Spotted towhees were not adversely affected by urbanization in Seattle (Gavareski 1976). However, Rottenborn (1999) considered towhees urbanization sensitive because of a negative association with paved ground. The specific aspect of urbanization under study likely influences the observed response in bird abundances and distribution. One outcome of urbanization is the fragmentation of habitat into patches of various sizes (Marzluff and Ewing 2001). In this study, however, only the red-breasted nuthatch was strongly associated with large patches, irrespective of habitat type. Nuthatches were also associated with large patches in Vancouver Island's high-elevation forests (Schieck et al. 1995). At a 300 ha scale, nuthatch abundance was positively correlated with the amount of late serai forest (McGarigal and McComb 1995). Other studies have shown that red-breasted nuthatch habitat use reflects the large, mature trees this species use for foraging (Mariani and Manuwal 1990, Adams and Morrison 1993). This suggests that the area sensitivity exhibited by the red-breasted nuthatch in my study is a result of decreased resource availability rather than more indirect fragmentation effects (e.g. edge or patch island effects). This is further supported by the fact that redbreasted nuthatch habitat use is largely influenced by conifer seed availability. In the Boreal forest, nuthatches search for food beyond their normal range when good seed-crop years are followed by poor seed-crop years (Koenig and Knops 2001). This may extend to feeder use in urban areas. In Vancouver, nuthatch abundance is positively associated with urban edges (P. Arcese, pers. comm.).  20  The chipping sparrow, orange-crowned warbler and brown-headed cowbird were associated with Garry oak. Chipping sparrows were also found in Oregon's Garry oak (Anderson 1970) but declined as these stands increased in conifer density (Hagar and Stern 2001). The orange-crowned warbler is associated with shrubby and clearcut forest habitats (Hutto and Young 1999) and used deciduous habitat in Seattle (Gavareski 1976).The fact that I did not detect cowbirds in coniferous patches cannot be generalized to other areas. Cowbirds will penetrate large forest patches in fragmented landscapes (Robinson and Wilcove 1994, Winslow et al. 2000). However lower parasitism rates in closed-canopy forests relative to more open habitats have been observed in Illinois (Robinson et al. 2000), the Okanagan Valley of British Columbia (Ward and Smith 2000) and riparian habitat in Montana (Tewksbury et al. 1998). Cowbird habitat associations reflect host abundance and proximity to food resources (Donovan et al. 1997, Tewksbury et al. 1998) and therefore their association with Garry oak may not reflect habitat composition per se. I found that brown creepers, chestnut-backed chickadees, golden-crowned kinglets and winter wrens were associated with coniferous habitat. The kinglet nests and forages in forest canopies (Ehrlich et al. 1988) and is associated with young stands with high densities of medium-sized trees (Hansen et al. 1995). The winter wren and the brown creeper prefer mature forest canopies over open habitats such as shrublands (Schwab and Sinclair 1994), partially harvested stands and clearcuts (Hansen et al. 1995, Hutto and Young 1999). Brown creepers and chestnut-backed chickadees are cavity nesters associated with mature forests, high conifer tree densities and large diameter trees (Hansen et al. 1995).  21  In my study area, chestnut-backed chickadees were more ubiquitous than the other forest birds. Habitat use by chestnut-backed chickadees may be related to the presence or absence of black-capped chickadees (Parus atricapillus). In both Victoria and California, chestnut-backed chickadees commonly used oak woodlands in the absence of black-capped chickadees (this study, Wagner 1981). In Oregon's Garry oak stands, black-capped chickadees were abundant and chestnut-backed chickadees were rarely detected (Hagar and Stern 2001). Both chickadees used urban parks in Seattle, though chestnut-backed chickadees were less abundant than black-capped chickadees (Gavareski 1976). The red-breasted nuthatch and Pacific-slope flycatcher are considered forest species in the Pacific Northwest (McGarigal and McComb 1995, Schieck et al. 1995) but were not strongly associated with coniferous habitat in my study. This suggests that these species may be more flexible in their habitat use in urban landscapes than in forested landscapes. Compared to the brown creeper, the red-breasted nuthatch forages on a wider variety of tree species and substrates including oak (Airola and Barrett 1985, Adams and Morrison 1993). Similarly, the Pacific-slope flycatcher may not select for specific tree species. In Oregon, Pacific-slope flycatcher abundance was better predicted by tree density and canopy cover variables rather than by habitat type (i.e. forest serai stage) (Hansen et al. 1995, McGarigal and McComb 1995). I did not detect strong habitat associations for Bewick's wren, house wren and Wilson's warbler. These species may select habitat based on vegetation characteristics not measured in this study. For example, Bewick's wren specialized on open juniper woodlands with dense shrub cover in Wyoming (Pavlacky and Anderson 2001). In riparian habitats, Bewick's wren abundance was positively associated with vegetation 22  volume (Rottenborn 1999). Wilson's warblers were highly associated with shrub thickets in riparian habitat of the Rockies (Hutto and Young 1999). My study underestimated the effect of urbanization across the region's entire bird community. The majority of bird species I detected were present in only one or two habitat patches. Bird-habitat relationships were not assessed for these rare species. It is possible that the greater Victoria region is too urbanized to support large populations of most species; urban and agricultural development has been on-going for over 100 years (Canada Census Office 1902). Comparing Victoria to less urbanized Garry oak (Willamette Valley of Oregon [Hagar and Stern 2001]), finds that many species rare in Victoria (e.g. hairy woodpecker, Cassin's vireo, western tanager) are common in Oregon. Also the full effect of urbanization on forest birds could not be adequately assessed because coniferous forest patches were not found in the most highly urbanized landscape. M y study and those of Gavareski (1976), Lancaster and Rees (1979) and Melles (2000) show that increasing urbanization alters bird assemblages for major urban centres in the Pacific Northwest. The most sensitive species, as found in my study and these others, include migratory and insect eating birds. Even in protected areas, where habitat structure is more complex than in adjacent urban areas, bird populations are susceptible to urbanization. This serves as a warning that as the region's human population expands, some species may be lost from parts of the landscape. The fact that few species were strongly or positively associated with Garry oak habitat indicates that, for common birds, Garry oak woodlands and Douglas-fir forests should be managed together with emphasis on reducing urbanization effects at the landscape-scale.  23  CHAPTER 3 A S S O C I A T I O N S O F BIRDS W I T H A D J A C E N T H A B I T A T A N D H A B I T A T C O M P O S I T I O N IN V I C T O R I A ' S G A R R Y O A K E C O S Y S T E M ABSTRACT I tested associations between bird species composition and coarse habitat type for birds in Garry oak, coniferous and urban habitat in Victoria, British Columbia. I also tested microhabitat associations for birds in Garry oak remnants adjacent to either urban areas or coniferous forest. Using detrended and canonical correspondence analysis, I found a strong overlap in bird species composition between Garry oak habitat and its adjacent habitat. There was, however, less overlap between assemblages in the Garry oak of each surrounding habitat type. I found that conifer stem volume was the vegetation factor most strongly correlated with differences in bird assemblages among transects in Garry oak patches. The Garry oak bird assemblage consisted of three sub-types: species tolerant of urbanization and associated with the shrub snowberry (Symphoricarpus albus), open woodland species and species associated with conifer trees. In Victoria, the Garry oak ecosystem is becoming increasingly urbanized and encroached upon by conifer trees. I therefore predict that shifts in the Garry oak bird assemblage will occur in the future.  INTRODUCTION Edge creation has been implicated as a cause for the decline of forest interior songbirds in eastern and midwestern North America (Brittingham and Temple 1983, Ratti and Reese 1988). Martin (1992) remarked that predation and brood parasitism by brown-headed cowbirds account for 80% of nest failure in songbirds. It had been widely thought that higher rates of nest predation and brood parasitism occur at edges than in habitat interiors (Wilcove 1985, Robinson 1992). Recent reviews do not support this claim (Paton 1994, 24  Lahti 2001, Chalfoun et al. 2002). The type of edge and adjacent habitat contributes to the presence and magnitude of edge effects (Robinson et al. 1995, Tewksbury et al. 1998, Brand and George 2000). Moreover, the amount of suitable habitat in the landscape influences the likelihood of observing increased nest predation at edges (Lahti 2001). Predator abundance is more likely to vary among landscapes than among patches of different sizes or between interior and edge habitat (Chalfoun et al. 2002). The relationship between brood parasitism and edge is also landscape-dependent (Tewksbury et al. 1998, Brand and George 2000). Negative edge effects, however, may be more common in urbanized landscapes. Urban edges are associated with increased predation and brood parasitism rates (Wilcove 1985, Marzluff and Ewing 2001). In addition, predators are more abundant in urban/suburban and fragmented landscapes (Villard et al. 1992, Marzluff and Ewing 2001). Partly due to differences in biotic and abiotic edge effects, the type of surrounding landscape influences the distribution and abundance of birds in remnant woodlands and forests. For example, species richness and the abundance of neotropical migrants in southern Ontario woodlands are associated with the amount of surrounding residential development (Friesen et al. 1995). There are also differences in bird species richness and abundance between hardwood stands surrounded by agriculture and those surrounded by pine forest in South Carolina (Kilgo et al. 1997). The bird community of California oak woodland patches surrounded by chaparral differs from those of similar structure surrounded by grassland (Sisk et al. 1997). By using a model based on bird abundance at distance intervals from the edge, Sisk et al. (1997) further showed that a species' sensitivity to edge depends on the type of surrounding habitat.  25  Garry oak patches in Victoria are surrounded by urban development or coniferous forests. Conifer edges are associated with conifer encroachment, reflecting the succession of Garry oak savanna to coniferous forest (Thilenius 1968, Tveten and Fonda 1999). In the Garry oak of Oregon's Willamette Valley, conifer encroachment over 30 years was associated with changes in bird species composition (Hagar and Stern 2001). In the South Okanagan, British Columbia, Ponderosa pine (Pinus ponderosd) encroachment into grassland strongly influenced bird community composition, species richness and diversity (Krannitz and Rohner 2000). Tree expansion in oak woodlands of Minnesota influenced bird community composition (Davis et al. 2000). In all these studies, increasing tree density was associated with the reduced abundance of open-breeding species and the increased abundance of woodland and forest birds. The purpose of this study is to describe the differences in bird assemblages of Garry oak patches surrounded by urban development and Garry oak patches surrounded by coniferous forest. I explore the hypothesis that bird assemblages in Garry oak will consist primarily of species common in the adjacent habitat (urban or coniferous forest). I also hypothesize that bird assemblage differences within and among Garry oak remnants are associated with overstory and understory vegetation cover and composition.  METHODS Study sites The study area and methods of selecting sites are described in chapter 2. Bird surveys Birds surveys were conducted as described in chapter 2. In this analysis, I include results from surveys conducted in urban areas. Urban surveys were conducted along 26  sidewalks in the residential area surrounding Mount Tolmie, Highrock and Uplands parks. Urban transects were chosen using a random number generator to determine an angle between 0° and 360°. The closest 400 m stretch of a street that fell within the calculated direction was chosen for the survey. The survey was repeated on a parallel street - the methods remaining consistent with surveys in the non-urban areas. The two streets were separated by a distance of 70 m (Mount Tolmie), 95 m (Uplands) and 100 m (Highrock). Bird survey points were placed every 50 m along the transect. There were seven points on each urban transect. Since habitat heterogeneity within Garry oak patches was expected, analyses compared bird abundances at the transect level (i.e. two samples per site). Vegetation Surveys Tree volume was assessed as per the methods described in chapter 2. However, in addition to Garry oak, volume for all species was calculated. Stem volume for all deciduous trees was calculated via an equation developed for bigleaf maple trees (B.C. Forest Service 1983): log volume (m ) = -4.537 + 1.908 log diameter (m) + 1.120 log height (m) A similar equation was used to calculate conifer stem volume (B.C. Forest Service 1983): log volume (m ) = -4.358 + 1.692 log diameter (m) + 1.182 log height (m) 3  Shrub and grass cover was measured along the same transects used in the tree surveys. Shrub cover was determined using the line-intercept method (Brower et al. 1989). Each species was measured separately but all species were pooled for analysis, except for Scotch broom and snowberry, which were analyzed singly because each is expected to characterize different soil and climate conditions within the Garry oak ecosystem. Grass cover was measured in a 1.0 m x 0.50 m rectangular plot placed every  27  ten m along the transect. A l l cover values were measured as percentages and averaged across the vegetation transects to give a value for each transect. In addition, tree and shrub composition at each site was determined by dividing the measures for each species by volume and cover totals. Data analysis  To explore the relationships between coarse habitat type (Garry oak, coniferous and urban) and bird species composition, I used Detrended Correspondence Analysis (DCA [Hill and Gauch 1980]). D C A is an ordination technique for species x site multivariate data tables. The maximum gradient length for the ordination of bird abundances among samples was 3.348 standard deviations. This indicates that the birds are responding to the environmental gradient in an unimodal fashion. Therefore D C A is a more appropriate ordination method than Principal Components Analysis (Palmer 1993, Legendre and Legendre 1998). Detrending was done by segments (Palmer 1993) in the C A N O C O computer program (ter Braak and Smilauer 1998). Using bird density from each transect, which were categorized as Garry oak, conifer or urban habitat, I produced an ordination diagram that reflected bird species composition within and among habitat types. To assess how vegetation composition was related to bird species composition, I used Canonical Correspondence Analysis (CCA), a direct-ordination technique (ter Braak 1986, 1994, 1995). Since C C A examines directly the relationship between environment and species in a multivariate data set (ter Braak 1986) it has been used extensively for predicting avian responses to habitat variation (e.g. Blair 1996, Rottenborn 1999). I tested the association of bird species with vegetation composition along transects in Garry oak habitat only using the forward stepwise regression and Monte 28  Carlo randomization procedure available in C A N O C O (ter Braak and Smilauer 1998). The significance of the first axis and all canonical axes together were also tested with Monte Carlo permutation tests (ter Braak and Smilauer 1998). Since bird and vegetation composition was measured on two transects located in the same site (i.e. sample plots within whole plots), pseudoreplication and spatial autocorrelation could potentially confound the analysis (Borcard et al. 1998). To remove the influence of pseudoreplication and spatial autocorrelation, I used the split-plot design in C A N O C O (ter Braak and Smilauer 1998, Coppedge et al. 2001). This procedure restricts permutations to whole-plots and keeps the sample transect pair together (ter Braak and Smilauer 1998). The ordination analyses were carried out only for those species that could be mapped via the territory mapping method. Consequently, flocking species and those without fixed territories, such as pine siskins, northwestern crows and swallows, were excluded from the analyses. In addition, species occurring in only one sample were excluded from the analyses. A l l data were log-transformed and rare species were downweighted by C A N O C O to reduce their effect on the spread of species. Statistical testing via Monte Carlo was based on 199 different permutations and a p-value of 0.05. I used a cluster analysis to explore similarity in bird community composition among all Garry oak, conifer and urban study sites (Sisk et al. 1997, Legendre and Legendre 1998). The presence-absence of individual bird species was used to calculate Jaccard's community similarity index for each pair of sites. The resulting similarity matrix, S, was converted to a distance matrix (i.e. a dissimilarity matrix, D, where D = 1 - S). The matrix was input into the cluster analysis feature of the Statistica software package (StatSoft Inc. 2000). Hierarchical and agglomerative clustering algorithms were 29  used to set the rules for fusing sites into clusters. Since different clustering algorithms produce different results, single-linkage, complete-linkage, unweighted and weighted pair-group averaging, unweighted and weighted centroid and Ward's minimum variance methods were all applied to the data. (Legendre and Legendre 1998, StatSoft Inc. 2000). Patterns that are consistent across all cluster outputs are considered robust (Legendre and Legendre 1998).  RESULTS Bird species composition  There was little difference in species composition between Garry oak and its adjacent habitat. Of the 30 species detected more than once in either Garry oak or adjacent coniferous forest, only six were unique to one habitat type. Four were detected solely in Garry oak (California quail, warbling vireo, white-crowned sparrow and yellow warbler) and two were detected solely in the forest (Cassin's vireo, song sparrow). Yellow warbler and song sparrow presence was likely influenced by proximity to water (Ehrlich et al. 1988). Of the 10 species detected more than once in urban-surrounded Garry oak and/or urban habitat, three were present solely in the Garry oak. The house sparrow was the only urban species not detected in the Garry oak remnants. Comparing bird assemblages of Garry oak surrounded by forest to Garry oak surrounded by urban development shows less overlap. Of the 27 species detected in Garry oak, only 11 were present in both. Only one of these 11 (the house finch) was restricted to urban-surrounded Garry oak. Vegetation composition  Vegetation composition of Garry oak patches varied extensively within transects, between transects and between sites (Tables 3.1, 3.2). Thetis Lake had the highest conifer  Table 3.1: Tree species composition of transects in each Garry oak site Site  Transect  Tree species stem volume (m /ha): mean + 95% conf. intv (n = 7)* Garry oak Conifer Other deciduous  Rocky Point  1  495.341 829.63  206.91+ 181.71  0  2  75.27 + 62.21  627.00+ 644.46  72.77+ 142.63  1  96.35 + 130.03  314.79+514.74  15.67+21.29  2  118.26+ 175.02  84.68+ 121.08  30.50+ 42.16  1  28.75 + 28.97  536.31+505.66  3.17+ 6.21  2  6.20 + 2.95  387.79+ 293.85  9.23+ 13.62  1  7.18 + 4.55  0  0.04+ 0.08  2  62.33 ± 34.77  0.98+ 1.33  1.54+3.01  1  56.32 + 54.98  10.65+ 11.03  8.41+9.52  2  37.13 + 48.30  5.10+ 9.99  0.70+ 1.37  1  31.19 + 34.09  1.01+ 1.97  0  Mary Hill  Thetis Lake  Mount Douglas  Highrock  Mount Tolmie  Uplands  1  17.65 + 16.21  0  47.84 + 31.75  55.05+ 107.90  15.20+ 18.36  39.46+ 68.24  *Except Rocky Point (n = 6) and Highrock (n = 4)  31  Table 3.2: Understory composition of transects in each Garry oak site Site  Transect  Shrub cover (%): mean ±95% conf. intv. (n =7)*  Grass cover (%) mean ±95% conf intv. (n = 7)*  Rocky Point  Mary Hill  Thetis Lake  Mount Douglas  Highrock  Snowberry  Scotch broom  Other  1  0  0  22.13± 17.47  65.56± 18.95  2  0  0  31.07±9.20  37.38+ 17.15  1  0  9.4414.13  5.1817.55  36.181 10.55  2  0  18.8619.49  6.9816.33  57.17121.35  1  13.471 10.94  20.561 12.45  15.1718.14  41.151 12.95  2  5.421 3.95  16.5716.95  21.0713.96  26.9017.59  1  3.271 5.03  26.341 14.39  3.8213.05  46.401 13.54  2  33.35130.17  26.161 16.22  12.521 12.81  20.761 15.23  1  17.8115.77  0  24.4117.95  38.20123.76  28.281 15.46  1.83+2.62  22.181 12.73  35.85126.21  1.9712.75  3.7514.28  1.581 1.55  54.951 22.58  14.10115.31  1.7711.90  2.731 2.90  40.431 10.28  49.70125.77  1.3312.60  29.15122.39  32.88127.96  13.12120.09  9.9019.59  29.28130.20  42.43125.95  Mount Tolmie  Uplands  1  "Except Rocky Point (n - 6) and Highrock (n = 4)  32  volume (95% of total tree volume was coniferous; Table 3.1). This high volume reflects the large size of coniferous trees relative to Garry oaks. Also, the volume measure is averaged across the transect and does not reflect patchiness in where conifers occur. Consequently, Thetis Lake can still be considered as a Garry oak site, but one that is rapidly succeeding to a coniferous forest. Conifers dominated on portions of Rocky Point and Mary Hill, with average stem volumes greater than 50% (Table 3.1). At these sites, both Douglas-fir and grand fir (Abies grandis) were present, though the former was more abundant. By contrast, the remaining sites had little to no conifer presence (Table 3.1). Deciduous trees other than Garry oak formed a minor part of the canopy in all sites but Uplands. At all sites, most non-oak deciduous trees were arbutus and bigleaf maple. However, trembling aspen (Populus tremuloides) was a unique component of the deciduous canopy at Uplands. Snowberry was the dominant shrub species at three sites - Mount Tolmie, Highrock and Uplands - representing 42% to 68% of all shrub cover (Table 3.2). Other species common in the understory, especially at Highrock and Uplands were nootka and baldhip rose (Rosa nutkana, R. gymnocarpa), Himalayan blackberry (Rubus discolor), cascara (Rhamnus purshiana) and India plum (Oemleria cerasaformis). These urban sites did not have high Scotch broom cover. Rocky Point was devoid of Scotch broom and snowberry. Its understory was dominated by shrubs such as trailing blackberry (Rubus ursinus), bracken fern (Pteridium aquilinum) and Douglas-fir seedlings. Mary Hill, which also did not have snowberry, was heavily invaded by Scotch broom (67% of total shrub cover; Table 3.2). Thetis Lake and Mount Douglas were also heavily invaded by Scotch broom; cover of this shrub exceeded cover values for snowberry and the combined values for other shrubs (Table 3.2).  33  Grass cover was measured differently than shrub cover and therefore the two values can not be directly compared. For example, the absence of grass does not imply high shrub cover. Instead, ground cover could consist of moss, rock or litter. Since grass was not identified to species, the proportion of native to non-native species could not be determined. However, grass cover values provided a relative measure of the extent to which grass dominated the ground layer. The values ranged from a low of 20% along one of the Mount Douglas transects to a high of 65% at Rocky Point (Table 3.2). However there was high heterogeneity among vegetation transects. Bird - habitat relationships  The first axis of the D C A ordination, which accounted for 28.4% of the variation among bird species, shows a continuum of species associations (Table 3.3, Figure 3.1). The urban bird assemblage was at the far end of the gradient (highly positive on the xaxis). Urban samples were relatively species poor (total species richness = 7, average ± 95% conf. intv. = 4.0 ± 0.98, n = 6); only house finch and bushtit were highly associated with this habitat. (However, many of the flocking species that were not analyzed were abundant in urban areas. These included northwestern crow, house sparrow, European starling and pine siskin). The long gradient length indicates that species highly associated with urban samples are not likely to be found in Garry oak and coniferous samples. A second species grouping consisted of species common to Garry oak. Some of these were also abundant in the urban sites, such as American robin, Anna's hummingbird and Bewick's wren. Most Garry oak birds were uniquely associated with Garry oak (e.g chipping sparrow, white-crowned sparrow, yellow warbler and brownheaded cowbird) or were common in both Garry oak and coniferous habitats (e.g. yellowrumped warbler, northern flicker, dark-eyed junco). Another species grouping was 34  Figure 3.1: Detrended Correspondence Analysis of bird species in Garry oak, coniferous and urban sites. The position of the species indicate the v distances among them. The sites are 2  plotted at the species' centroids based on the relative abundance of each species for the site. Habitat types: A - Conifer; • - Garry oak; • - urban. Bird species: A M R O - American robin; A N H U - Anna's hummingbird; B E W R - Bewick's wren; B H C O - brown-headed cowbird; B R C R - brown creeper; B U S H - bushtit; CAQU - California quail; C A V I - Cassin's vireo; C B C H - chestnut-backed chickadee; CHSP - chipping sparrow; C O R A - common raven; DEJU - dark-eyed junco; DOWO - downy woodpecker; G C K I - golden-crowned kinglet; HOFI - house finch; HOWR - house wren; N O F L - northern flicker; O C W A - orangecrowned warbler; PSFL - Pacific-slope flycatcher; PUFI - purple finch; R B N U - red-breasted nuthatch; SOSP - song sparrow; SPTO - spotted towhee; T O W A - Townsend's warbler; W A V I - warbling vireo; W E T A - western tanager; WCSP - white-crowned sparrow; WIWA Wilson's warbler; WIWR - winter wren; Y E W A - yellow warbler; Y R W A - yellow-rumped warbler.  35  associated with coniferous habitat (e.g golden-crowned kinglet, Cassin's vireo, winter wren, downy woodpecker and common raven). Generally, there was little separation in bird assemblages of Garry oak surrounded by coniferous forest and the adjacent forest (Figure 3.1). Also, the species richness of both habitat types were similar. Coniferoussurrounded Garry oak contained a total of 28 species (including rare species) and the eight transects averaged (average ± 95% conf. intv.) 17.75 ± 4.26 species. The coniferous forest sites, on the other hand, had 26 species (average ± 95% conf. intv. = 15.25 ± 2.17 species). The Canonical Correspondence Analysis of Garry oak samples revealed that the distinction between urban-surrounded and coniferous-surrounded Garry oak assemblages evident from the D C A can be attributed to the presence of conifers (Figure 3.2). Conifer stem volume contributed significantly to the variation in the bird assemblage (F = 3.10, p = 0.01, d.f. = 12). The first axis, with which conifer volume was highly correlated (r = 0.85; Table 3.4), explained 23% of the variation in species composition but was not significant at the 0.05 alpha level (Table 3.4; F = 1.789, p = 0.09). The four canonical axes together were significant (F = 1.664, p = 0.04). Yellow-rumped warbler, Townsend's warbler, dark-eyed junco, winter wren, golden-crowned kinglet and purple finch were associated with Garry oak patches surrounded by forest and with high conifer stem volume. At the opposite end of this gradient were species related to urbanized patches with a snowberry understory (first axis correlation r = -0.55; Table 3.4). These included Anna's hummingbird, spotted towhee, bushtit, Bewick's wren and house finch (Figure 3.2). Not all species in Garry oak surrounded by coniferous forest were associated with conifer volume. The second C C A axis separated those species using high amounts of  Figure 3.2: Canonical Correspondence Analysis of bird species and vegetation variables in Garry oak sites. The arrows indicate the strength of the relationship of the vegetation variables with the two axes that describe variation in the bird communities. The proximity of a species to an arrow indicates the strength of association between the species and the vegetation factor. Along the first (horizontal) axis bird assemblages range from being associated with high conifer stem volume to being highly associated with snowberry. In between and at the bottom of the second (vertical) axis lie species associated with sites containing high Scotch broom cover, grass cover and Garry oak volume. Bird species: A M R O - American robin; A N H U - Anna's hummingbird; B E W R - Bewick's wren; B H C O - brown-headed cowbird; B R C R - brown creeper; B U S H - bushtit; C A Q U - California quail; C B C H - chestnut-backed chickadee; CHSP - chipping sparrow; DEJU - dark-eyed junco; G C K I - golden-crowned kinglet; HOWR - house wren; N O F L - northern flicker; O C W A - orange-crowned warbler; PSFL Pacific-slope flycatcher; PUFI - purple finch; R B N U - red-breasted nuthatch; SPTO spotted towhee; T O W A - Townsend's warbler; W A V I - warbling vireo; W E T A western tanager; WCSP - white-crowned sparrow; WIWA - Wilson's warbler; WIWR - winter wren; Y E W A - yellow warbler; Y R W A - yellow-rumped warbler.  38  Garry oak volume, Scotch broom and grass cover. Broom significantly influenced variation among bird densities (F = 2.13, p = 0.05) and was moderately correlated with the second axis (r = -0.37). Sites with high broom cover did not have much cover by other shrub species (second axis correlation r = 0.47). The community associated with the oak-broom-grass complex included western tanager, orange-crowned warbler, Pacificslope flycatcher, chipping sparrow, white-crowned sparrow, warbling vireo, Wilson's warbler and California quail (Figure 3.2). Some species, the northern flicker, red-breasted nuthatch, brown creeper and house wren, were equally associated with Garry oak and conifer trees. They were associated with both the first and second axes, indicating an equal relationship to both tree volume variables (Figure 3.2). Deciduous volume contributed little to overall bird community variation (for deciduous volume, first axis correlation: r = 0.09, second axis correlation: r = 0.18; Table 3.4). A group of species- yellow warbler, brown-headed cowbird, chestnut-backed chickadee and American robin - were also located at the origin of the C C A bi-plot (Figure 3.2). They were positioned at the origin because of their ubiquity across habitat types. The sum of the canonical eigenvalues was 0.954 out of a total inertia of 1.446, indicating that 34% of the variation was not explained by the vegetation variables (Borcard et al. 1998; Table 3.5). There was little intercorrelation among the vegetation variables; the variance inflation factors (VIF) of the vegetation variables never exceeded 3.9% VIF is a measure of goodness-of-fit and multicollinearity. It compares observed collinearity to that which is expected under no collinearity. Ter Braak and Smilauer (1998) suggest that models with a VIF greater than 20% can not be interpreted accurately. 40  Table 3.3: Results of Detrended Correspondence Analysis of bird species among Garry oak, conifer and urban samples (n - 28). Axes 1 2 3 4 Eigenvalues 0.546 0.157 0.07 0.043 Lengths of gradient 3.348 1.773 1.606 1.288 Cumulative percentage variance 28.4 36.6 40.2 42.5 of species data Sum of all unconstrained 1.922 eigenvalues Table 3.4: Correlation of each vegetation variable with each axis derived from Canonical Correspondence Analysis. Vegetation variable Correlation: Correlation: Correlation: Correlation: axis 1 axis 2 axis 3 axis 4 Conifer volume 0.8485 0.4013 -0.0124 -0.1809 Garry oak volume 0.3826 -0.3716 -0.4343 0.4848 Other deciduous volume 0.0871 0.1814 -0.5232 -0.2741 Scotch broom cover 0.0114 -0.3743 0.838 -0.0204 Snowberry cover -0.5523 0.2241 -0.022 0.3227 Other shrub cover 0.1602 0.4733 -0.4136 0.3981 Grass cover 0.1296 -0.2893 -0.2579 0.0361 Table 3.5: Results of Canonical Correspondence Analysis of bird species and vegetation variables among Garry oak samples (n = 14). Axes 1 2 3 4 Eigenvalues 0.332 0.204 0.151 0.114 Species-environment correlations 0.965 0.973 0.958 0.876 Cumulative percentage variance 23.0 37.1 55.4 47.5 of species data Sum of all unconstrained 1.446 eigenvalues Sum of all canonical eigenvalues 0.954  41  Community composition The results of the cluster analysis confirmed the D C A results. The most pronounced difference in species composition occurred between urbanized habitat (urban-surrounded Garry oak and urban samples) and forested habitat (forest-surrounded Garry oak and coniferous samples; Figure 3.3). Bird assemblages of urban-surrounded Garry oak and urban samples were also dissimilar. There was high similarity among Garry oak sites surrounded by coniferous forest and coniferous forest patches. This structure was revealed by six of the seven cluster algorithms. Only the single-linkage method did not show obvious partitions among study sites. The similarity between forestsurrounded Garry oak and conifer sites is not attributable to high similarity among individual Garry oak-conifer pairs. There was no evidence that a pair of sites from the same location was more similar than two sites at different locations. For example, Mount Douglas-Garry oak was more similar to three other sites than to Mount Douglasconiferous. There was little correlation between distance among sites and community similarity index (r = 0.15, n = 591; Figure 3.4).  DISCUSSION The bird species composition of Victoria's Garry oak patches differed depending on whether the surrounding habitat was urban or coniferous forest. Garry oak patches surrounded by forest contained species also common in the adjacent forest, such as winter wrens, brown creepers and golden-crowned kinglets. Within Garry oak, these species were more abundant where there was high conifer volume. There was little conifer tree cover in Garry oak sites surrounded by urbanization and forest birds were not detected in these sites. The strength and direction of the relationship birds have with conifer volume 42  Figure 3.3: Cluster diagram of the similarities in species assemblages of all Garry oak, Douglas-fir and urban (UR) study sites. The clusters depicted are a result of the Weighted Pair Group method of the Statistica program. Linkage distances express the dissimilarity among groups. Note three dominant clusters: 1) Mt. Douglas-Garry oak, Rocky Point-Garry oak, Thetis LakeDouglas-fir, Mt. Douglas-Douglas-fir, Rocky Point-Douglas-fir, Mary HillDouglas-fir, Thetis Lake-Garry oak, Mary Hill-Garry oak, 2 ) Uplands-Garry oak,, Mt. Tolmie-Garry oak, Highrock-Garry oak, and 2 ) Uplands-urban, Highrock-urban and Mt. Tolmie-urban.  o Ko  o w  O CD CD O  ST =3 O CD  o bi  o  o  o CO  0.7 0.6 H 0.5  x  CD  TD  S  0.4  CO  8 o  S • 0.3  CD  0.2  • ••  0.1  — i —  5  10  15  20  25  Distance (km)  Figure 3.4: Scatterplot of Jaccard's community similarity index versus distance (km) for all study site pairs. Site similarity is represented by Jaccard's similarity index (Legendre and Legendre 1998). There was little relationship between distance and community similarity (r = 0.15. n = 91V  45  best characterizes species composition among Garry oak sites. The importance of conifers to bird species composition in Garry oak was also found in Oregon (Hagar and Stern 2001). Oak can also be an important resource for forest birds in other temperate woodlands. In Sweden, Hansson (2000) found coniferous forest birds in remnant oak woodlands surrounded by conifer plantations. Black oak (Quercus kelloggi) was an important spring/summer foraging resource for the red-breasted nuthatch in Sierra Nevada coniferous forests (Adams and Morrison 1993). Szaro and Balda (1979) found that 15 species of forest birds used oaks for foraging in an amount greater than their availability during the breeding season. Some species that used mixed Gambel oak {Quercus gambelii) - ponderosa pine stands in Arizona were absent from pure pine stands (Rosenstock 1998). Species composition of Garry oak sites completely surrounded by urbanization was, for the most part, limited to species that were also found in the adjacent urban areas. The bird species associated with urbanized Garry oak were almost exclusively resident ground foraging species. On the other hand, less urbanized patches included migrants and birds from other foraging guilds. This is similar to the results from urban bird communities in California (Blair 1996, Bolger et al. 1997). Similarly, in Oregon, where Garry oak stands are not urbanized, migratory foliage gleaning insect eaters were common inhabitants (Anderson 1970, Hagar and Stern 2001). The birds of Garry oak patches surrounded by urbanization were also highly associated with snowberry. Extensive shrub cover may enhance habitat quality for birds in the face of urbanization pressures. Orange-crowned warblers and spotted towhees were common in urbanized remnants but were not detected in the adjacent urban areas. Dense shrub cover may provide a means of nest concealment against predators (Ratti and Reese 46  1988, Saurez et al. 1997). However, this does not improve habitat quality for all species. For example, white-crowned sparrow and chipping sparrow were not found in urbanized patches of Garry oak despite good cover of snowberry. For non-urbanized sites with less conifer presence, a third bird grouping emerged. I call it an "open-oak community" because these species were associated with large volume Garry oak combined with high grass and broom cover. Some species that commonly breed in open-canopied habitat, such as white-crowned and chipping sparrows (Ehrlich et al. 1988), reached their peak abundance in these locations. Hagar and Stern (2001) also found birds associated with shrub and deciduous trees to be common in Garry oak stands in Oregon. Chipping sparrows, however, disappeared from the Oregon oak stands between the late 1960's and the mid 1990's. Other open-woodland species also decreased in abundance over this time period. Hagar and Stern (2001) attributed the changes in species composition to changes in conifer density and shrub cover. On Vancouver Island, it is believed that succession of Garry oak savanna to woodland and coniferous forest partly contributed to declines of streaked horned lark, vesper sparrow, Lewis' woodpecker, western bluebird and western meadowlark (Campbell et al. 1990, 1997, 2001, Rogers 2000). The three Garry oak vegetation site types that I found, correspond closely to those found in the Puget Trough of Washington (Thysell and Carey 2001). Thysell and Carey (2001) classified their site types as 1) Garry oak with an understory of snowberry and Saskatoon (Amelanchier alnifolia), 2) Douglas-fir dominated with an understory of India plum and swordfern (Polystichum munitum) and 3) savanna-like where Garry oaks were few but large and the understory was mostly exotic grasses. These sub-types are directly homologous to my sites: 1) the urbanized sites dominated by snowberry, 2) the sites with  47  high conifer volume (although the understory is moderately different) and 3) the (formerly) savanna sites (now with Scotch broom in place of much of the grass cover). M y study goes a step forward by showing that bird species are also distributed between these three site types. M y results also compare well with a study of oak savannas along a restoration gradient in Minnesota where bird species composition was distinguished according to tree density and shrub cover (Davis et al. 2000). The latter grouping consisted primarily of low canopy insect gleaning species. This is the first study to compare the bird communities of remnant Pacific northwest oak woodlands in the context of surrounding landscape and urbanization. Except for a few of the "open-oak woodland" species, Garry oak remnants of Victoria did not have a distinct assemblage. Species composition and abundances were similar between Garry oak and the adjacent habitat. The recent history and the narrow and patchy distribution of Garry oak in British Columbia also likely contributes to the lack of a distinct assemblage. Bird assemblages of Garry oak of the two surrounding habitat types overlapped to a lesser degree. It is not known whether urban edge, urbanization at larger scales, vegetation composition or other mechanisms are responsible for the observed differences between Garry oak of the different surrounding habitats. Birds associated with conifer volume may increase in abundance if Garry oak succeeds to coniferous forest. Without historical information on the tree and bird composition of the Garry oak remnants I surveyed, I can only speculate on the effect of conifer encroachment. Similarly, the effects of Scotch broom invasion on bird community composition have not been studied. Considering that both types of habitat change are expected to continue to occur in the future (Fuchs 2001, Thysell and Carey 2001), the long-term effects should be monitored.  While the spring-time bird assemblages of Garry oak and coniferous forest are similar, Victoria's Garry oak bird community in other seasons has not fully been studied. Garry oak remnants may be important stopover habitat during migration because Victoria occurs on the Pacific flyway. Garry oak meadow habitat with early blooming wildflowers and their associated insects may be used during early spring migration. Wintering sparrows may also benefit from open oak habitat. Anderson (1970) found a seasonal pattern in bird abundance in Oregon's Garry oak woodlands. Some high elevation forestbreeding species become more abundant in Garry oak in winter (Anderson 1970, 1972). Two of these species, the band-tailed pigeon and the Steller's jay forage on acorns. Managing oak habitat for jays is potentially important because of their role in acorn dispersal and oak regeneration (Fuchs et al. 2000). Information on year-round bird use is needed to inform management of the Garry oak ecosystem.  49  CHAPTER 4 G E N E R A L DISCUSSION A N D C O N C L U S I O N S There were three general goals for this thesis: 1) to describe associations of Garry oak birds with urbanization in the landscape, 2) to describe associations of Garry oak birds with adjacent habitat type and vegetation composition and 3) to assess how these associations can be used in Garry oak ecosystem management. The results of the thesis also highlight similarities and differences in bird composition of Garry oak, conifer forest and urban habitats. Urbanization in the broader landscape (approx. 2 km surrounding each study site) is negatively associated with the majority of birds in Garry oak and conifer habitat fragments. More species were associated with urbanization than with patch size or Garry oak stem volume. Only two of the ten species associated with urbanization were more abundant in urban Garry oak. Forest birds that were detected in Garry oak were limited to less urbanized Garry oak patches (except the chestnut-backed chickadee). Forest bird associations need to be interpreted with caution because: a) coniferous forest patches were not sampled along the full urbanization gradient and b) urbanization may be confounded with habitat structure (see below). Since Garry oak birds are associated with urbanization in the broader landscape, it is possible that landscape-scale processes affect these species. The recent literature has considered how landscape structure may be linked to population dynamics (e.g. island, source-sink and metapopulation models [Dunning et al. 1992]) but there is little agreement on which elements of the landscape affect populations in different landscapes (Wiens 1994). Urbanization is generally associated with increased nest predation (Marzluff and Ewing 2001). 50  There were substantial differences in bird composition between Garry oak surrounded by forest and Garry oak surrounded by urban development. Nineteen species were associated with the former habitat compared to four species associated with the latter. Three species were common to both habitats. The difference in bird assemblages can be partly attributed to differences in habitat composition, particularly the presence of large conifer trees. Conifer tree volume was highly associated with nine species. Eight species were associated with habitat features indicative of oak woodland. Other biotic processes associated with edge type can influence bird composition. For example, increased nest predation and brood parasitism at edges is generally associated with reduced population size in some fragmented landscapes (Martin 1992, Robinson et al. 1995, Donovan et al. 1997). In southeastern Vancouver Island, cowbird parasitism and nest predation is high in deciduous-riparian habitat (J. N . M . Smith, personal communication). In my study, brown-headed cowbirds were more common in Garry oak than in coniferous forests. Many songbirds using Garry oak are common cowbird hosts. M y results are based on correlations between bird abundances and environmental variables. They do not explain how differences in urbanization affect the birds found in Victoria. I collected data during one season and one year only. In Oregon, species composition of Garry oak and coniferous forests changed during the year (Anderson 1970, 1972). Distinct migrant and wintering assemblages were found. Temporal variation in distribution and abundance of birds in fragmented landscapes is common (Boulinier et al. 2001). The bird abundances I analyzed were based on an index of relative abundance. Measures of abundance are not always correlated with habitat quality (Van Home 1983). It is not known how the environmental variables are related to the long-term persistence of these species.  51  The bird survey methods I used have many assumptions. Standard point counts do not take into account detection probabilities (Rosenstock et al. 2002). Without using a model of detection probability, it is questionable whether differences in observed abundances reflect actual differences in population sizes (Nichols et al. 2001, Farnsworth et al. 2002). Detection probabilities vary with species (Farnsworth et al. 2002), time of day and season (Robbins 1981, Skirvin 1981) and habitat (McShea and Rappole 1997). Detection probability also varies among observers (Sauer et al. 1994). I controlled for this last problem by using only two observers who trained together to ensure consistency in the identification of species and distances. The majority of bird detections were from song. The probability of a bird singing varies during the day and year and is related to breeding phenology (Best 1981, Wilson and Bart 1985). Since mated males sing less, counts based on song are biased toward detecting unpaired males (Gibbs and Wenny 1993). This compromises the estimation of territory density and pairing success. Inferences based on this study should keep in mind its biases. However, this does not preclude the need to make management recommendations or to carry out long-term monitoring of birds in this system. Management recommendations A recovery strategy for Garry oak ecosystems is already in place (Garry Oak Ecosystems Recovery Team [GOERT] 2002). The strategy includes both coarse-scale objectives that address ecosystem conservation and fine-scale components that target individual species whose needs are not addressed in ecosystem-level management. Given that I did not detect rare species requiring specific microhabitat features in Victoria's Garry oak, ecosystem-wide conservation and restoration suits avian conservation needs.  52  The recovery strategy defines three "essential ecosystem characteristics" that define its coarse-scale objectives (Fuchs 2001). The "spatial integrity" characteristic concerns habitat loss and fragmentation (Fuchs 2001). I found that many species are negatively associated with urbanization and urban edges. I did not find strong associations between species and patch size, indicating that fragmentation per se may not be an issue. The type of habitat that is replacing and fragmenting Garry oak may be as an important consideration as the amount that is being lost. Setting limits on the intensity of urbanization (e.g. human population density) in future development surrounding Garry oak may be an important policy to craft and implement for Vancouver Island municipalities. Defining an "ecosystem" as Garry oak dominated makes little sense for avian conservation. Most birds in Victoria use both Garry oak and coniferous forest habitat. Urbanization is a landscape-scale processes that affects both habitats. Coastal Douglas-fir forest is also at risk from habitat loss and fragmentation (Flynn 1999). "Spatial integrity" should consider both habitat types together. For example, conservation targets for the amount or size of habitat to be protected within a landscape habitat should consider Garry oak and coniferous forest combined. Management specific to Garry oak should be implemented only i f a causal link between specific changes within Garry oak compromise a species' long-term population persistence. The second essential ecosystem characteristic recognizes that frequent and lowintensity fires are an important disturbance factor in Garry oak ecosystems (Fuchs 2001). Without fire, Garry oak habitat can become dominated by Douglas-fir. In this thesis I have identified a suite of species associated with high levels of conifer tree volume. These species are common forest birds. Conifer encroachment may increase dominance 53  by forest birds as reported in Oregon's Garry oak (Hagar and Stern 2001) and elsewhere in British Columbia (Krannitz and Rohner 2000). Whether a shift toward more forest species comes at the expense of other species is not locally known. However, in Oregon, the decline of several species was associated with conifer encroachment (Hagar and Stern 2001). The "oak-woodland" suite of species I identified may be threatened by conversion of oak woodland to coniferous forest. (Two of these species, chipping sparrow and yellow warbler, disappeared altogether from Oregon's Garry oak [Hagar and Stern 2001]. Whether conifer encroachment alone caused their extirpation is unknown. Both species are also susceptible to cowbird parasitism). The third essential ecosystem characteristic concerns the effects of exotic species (Fuchs 2001). M y findings do not directly address the association between birds and exotic species. Scotch broom is an exotic shrub and is detrimental to the native flora. I found several birds positively associated with Scotch broom. Some of these species are shrub nesters. Whether there is a direct benefit accrued by broom for birds has not been tested. The use of broom may be coincident with other related habitat factors. The longterm effects of broom invasion on bird assemblages need to be studied. Alternatively, the effects of cutting and/or pulling broom, a current conservation measure (Ussery and Krannitz 1998), are not known. Exotic grasses are dominant in some Garry oak meadows (Erickson 1996). I did not test for bird associations with different plant communities. I do not expect most birds currently using Garry oak to be affected by grass species composition. Of all the species associated with grass cover, only the introduced California quail primarily feeds on seeds (Ehrlich et al. 1988). Species that are more closely associated with grasses (western meadowlark, vesper sparrow and streaked horned lark [Wiens 1974, Rogers 2000]) have been extirpated or nearly extirpated from  54  Garry oak (Campbell et al. 1997, 2001). Changes in grass species composition and grassland extent may have contributed to their decline. Conserving only woodland and forest species is not sufficient to fully represent the Garry oak bird community (sensu Margules and Pressey 2000). Restoring the Garry oak savanna bird community (western bluebird, western meadowlark, vesper sparrow, streaked horned lark, Lewis' woodpecker) is a desirable long-term conservation goal (GOERT 2002). These species are declining throughout the coast of the Pacific Northwest (Rogers 2000). Vancouver Island's Garry oak is an important component in conserving these birds throughout their range. Restoring and re-creating large tracts of suitable savanna and grassland for these birds is an expensive and complex endeavour. A n example of active management for coastal savanna birds is the Fort Lewis Prairie Oak Preserve in the South Puget Sound area of Washington state (Dunn and Ewing 1997, Rogers 2000). Four of the five savanna species (there are no Lewis' woodpeckers) have populations on the preserve. Management includes removing invasive shrubs, setting small fires to maintain open conditions and maintaining bluebird nest boxes (Dunn and Ewing 1997). The geology of South Puget Sound differs from Vancouver Island. However the structure and size of the prairie are an indication of the species' habitat requirements. Moreover, much of the oak savanna is reclaimed pasture land. Recently, the Nature Conservancy of Canada has established a Garry oak preserve on former agricultural land near Duncan, Vancouver Island. Management there includes broom removal. More land acquisition on Vancouver Island and the Gulf Islands followed by intensively managing for savanna conditions is needed. However, urbanization in the landscape will compromise the success of savanna bird restoration efforts.  L I T E R A T U R E CITED Adams, E. and M . L . Morrison, 1993. Effects of forest stand structure and composition on red-breasted nuthatches and brown creepers. Journal of Wildlife Management 57: 616-629. Anderson, D R., K. P. Burnham, W. L . Thompson, 2000. Null hypothesis testing: problems, prevalence and an alternative. Journal of Wildlife Management 64: 912 — 923. Anderson, S. H., 1970. The avifaunal composition of Oregon White oak stands. Condor 72: 417-423. Anderson, S. PL, 1972. Seasonal variations in forest birds of western Oregon. Northwest Science 46: 194-206. Airola, D. A . and R. H . Barrett, 1985. Foraging and habitat relationships of insectgleaning birds in a Sierra Nevada mixed-conifer forest. Condor 87: 205 - 216. Barnhardt, S. J., J. R. McBride, C. Cicero, P. da Silva and P. Warner, 1987. Vegetation dynamics of the northern Oak woodland, pp 53-58 in Plumb, T. R. and N . H . Pillsbury, eds. Proceedings of the symposium on multiple-use management of California's hardwood resources. San Luis Obispo, C A . U S D A Forest Service General Technical Report PSW-100. Pacific Southwest forest and range experimental station, Berkeley, CA. Beckwith, B . R., 2002. Colonial Eden or indigenous cultivated landscape: reconstructing nineteenth century camas meadows on southern Vancouver Island, pp. 64-72 in Burton, P. J., ed. Garry Oak Ecosystem Restoration: Progress and Prognosis. Proceedings of the Third Annual Meeting of the B.C. Chapter of the Society for Ecological Restoration, April 27-28, 2002, University of Victoria. B.C. Chapter of the Society of Ecological Restoration, Victoria, B C . Beese, W. J. and A. A . Bryant, 1999. Effects of alternative silvicultural systems on vegetation and bird communities in coastal montane forests of British Columbia, Canada. Forest Ecology and Management 115: 231-242. Beissinger, S. R. and D. R. Osborne, 1982. Effects of urbanization on avian community organization. Condor 84: 7 5 - 8 3 . Bender, D. J., T. A . Contreras and L. Fahrig, 1998. Habitat loss and population decline: A meta-analysis of the patch size effect. Ecology 79: 517-533. Best, L. B., 1981. Seasonal changes in the detection of individual bird species. Studies in Avian Biology 6: 252-261. Bibby, C. J., N . D. Burgess and D. A . Hill, 1992. Bird Census Techniques. Academic Press, San Diego, C A . Blair, R. B., 1996. Land use and avian species diversity along an urban gradient. Ecological Applications 6: 506-519. Blake, J. G. and J. R. Karr, 1987. Breeding birds of isolated woodlots: area and habitat relationships. Ecology 68: 1724-1734. Bolger, D. T., T. A . Scott and J. T. Rotenberry, 1997. Breeding bird abundance in an urbanizing landscape in coastal Southern California. Conservation Biology 11: 406421. Borcard, D., P. Legendre, P. Drapeau, 1992. Partialling out the spatial component of ecological variation. Ecology 73: 1045-1055. Boulinier, T., J. D. Nichols, J. E. Hines, J. R. Sauer, C. H . Flather and K . H . Pollock, 2001. Forest fragmentation and bird community dynamics: inference at regional 56  scales. Ecology 82: 1159 - 1169. Brand, L . A . and T. L . George, 2000. Predation risk of nesting birds in coast redwood forest fragments. Journal of Wildlife Management 64: 42 - 51. British Columbia Forest Service, 1983. Whole stem cubic metre volume equations, 1976 in Watts, S. B., ed., Forestry Handbook for British Columbia. 4 ed. Forestry Undergraduate Society, Vancouver, BC. Brittingham, M . C. and S. A . Temple, 2983. Have cowbirds caused forest songbirds to decline? Bioscience 33: 31 - 35. Brower, J. E., J. H . Zar and C. N . Van Ende, 1989. Field and Laboratory Methods for General Ecology. 3 Edition. W. C. Brown Publishers. Dubuque, IA. Burnham, K . P. and D. R. Anderson, 1998. Model selection and inference: a practical information-theoretic approach. Springer, New York, N Y . Campbell, R. W., N . K . Dawe, I. McTaggert-Cowan, J. M . Cooper, G. W. Kaiser and M . C. E. McNall, 1990. The Birds of British Columbia. Vol. 2. Royal British Columbia Museum, Victoria, British Columbia. Campbell, R. W., N . K . Dawe, I. McTaggert-Cowan, J. M . Cooper, G. W. Kaiser and M . C. E. McNall, 1997. The Birds of British Columbia. Vol. 3. University of British Columbia Press, Vancouver, British Columbia. Campbell, R. W., N . K . Dawe, I. McTaggert-Cowan, J. M . Cooper, G. W. Kaiser, A . C. Stewart and M . C .E. McNall, 2001. The Birds of British Columbia. V o l . 4. University of British Columbia Press, Vancouver, British Columbia. Canada Census Office, 1902. Fourth Census of Canada, 1901. V o l . 1. S. E. Dawson, Ottawa, Ontario. Canadian Wildlife Service, 1997. Sensitive Ecosystem Inventory: East Vancouver Island and Gulf Islands. Select Version 2. Environment Canada, Qualicum Beach, British Columbia. Chambers, C. L , W. C. McComb and J. C. Tappeiner III, 1999. Breeding bird responses to three silvicultral treatments in the Oregon coast range. Ecological Applications 9: 171 - 185. Chalfoun, A . D., F. R. Thompson III and M . J. Ratnaswamy, 2002. Nest predators and fragmentation: a review and meta-analysis. Conservation Biology 16: 306 - 318. Coppedge, B. R., D. M . Engle, R. E. Masters and M . S. Gregory, 2001. Avian response to landscape change in fragmented southern Great Plains grasslands. Ecological Applications 11: 47 - 59. COSEWIC, 2002. Canadian Species at Risk, May 2002. Committee on the Status of Endangered Wildlife in Canada, 34pp. Davis, M . A., D. W. Peterson, P. B . Reich, M . Crozier, T. Query, E. Mitchell, J. Huntington and P. Bazakas, 2000. Restoring savanna using fire: impact on the breeding bird community. Restoration Ecology 8: 30 - 40. Donovan, T. M . , P. W. Jones, E. M . Annand and F. R. Thompson III, 1997. Variation in local-scale edge effects: Mechanisms and landscape context. Ecology 78: 2064 2075. Dunn, P. and K. Ewing (eds.), 1997. Ecology and Conservation of the South Puget Sound Prairie Landscape Dunning, J. B., B . J. Danielson and H. R. Pulliam, 1992. Ecological processes that affect populations in complex landscapes. Oikos 65: 1 6 9 - 175. Ehrlich, P. R., D. S. Dobkin, D. Wheye, 1988. The Birder's Handbook: a field guide to the natural history of North American birds. Simon and Shulster, New York, N Y . th  rd  57  Erickson, W. R., 1993. Garry Oak Ecosystems. Ecosystems at Risk Brochure. B C Ministry of Environment, Lands and Parks, Victoria, B C . Erickson, W. R., 1996. Classification and interpretation of Garry Oak plant community communities and ecosystems in southwestern British Columbia. MSc thesis, Department of Geography, University of Victoria, Victoria, B C . Fahrig, L . and G. Merriam, 1994. Conservation of fragmented populations. Conservation Biology 8: 5 0 - 5 9 . Farnsworth, G. L., K . H . Pollock, J. D. Nichols, T. R. Simons, J. E. Hines and J. R. Sauer, 2002. A removal model for estimating detection probabilities from point-count surveys. Auk 119: 414-425. Fernandez-Juricic, E. and J. Jokimaki, 2001. A habitat island approach to conserving birds in urban landscapes: case studies from southern and northern Europe. Biodiversity and Conservation 10: 2023 - 2043. Flynn, S. 1999. Coastal Douglas-fir Ecosystems. Ecosystems at Risk Brochure. B C Ministry of Environment, Lands and Parks, Victoria B C . Friesen, L. E., P. F. J. Eagles and R. J. MacKay, 1995. Effects of residential development on forest-dwelling Neotropical migrant songbirds. Conservation Biology 9: 1408 1414. Fuchs, M .A., P. G. Krannitz and A. S. Harestad 2000. Dispersal of Garry oak acorns by Steller's jays. pp. 263-266 in Darling, L. M . , ed. Proceedings of a Conference on the Biology and Management of Species and Habitats at Risk, Feb. 1 5 - 1 9 , 1999 Kamloops, B C . Volume 1. Ministry of Environment, Lands and Parks, Victoria, B C and University College of the Cariboo, Kamloops, B C . Garry Oak Ecosystems Recovery Team, 2002. Recovery Strategy for Garry Oak and Associated Ecosystems and Their Associated Species at Risk in Canada, 20012006. Draft 20 February 2002. Garry Oak Ecosystems Recovery Team, Victoria, BC. Gavareski, C. A., 1976. Relation of park size and vegetation to urban bird populations in Seattle, Washington. Condor 78: 375-382. Germaine, S. S., S. S. Rosenstock, R. E. Schweinsburg and W. S. Richardson, 1998. Relationships among breeding birds, habitat and residential development in Greater Tuscon, Arizona. Ecological Applications 8: 680 - 691. Gibbs, J. P. and D. G. Wenny, 1993. Song output as population estimator: effect of male pairing status. Journal of Field Ornithology 64: 316 - 322. Green, R. N . and K. Klinka, 1994. A field guide to site identification and interpretation for the Vancouver Forest Region. B C Ministry of Forests, Research Branch, Victoria, BC. Hagar, J. C. and M . A . Stern, 2001. Avifauna in oak woodlands of the Willamette Valley, Oregon. Northwestern Naturalist 82: 12-25. Haire, S. L., C. E. Bock, B . S. Cade and B. C. Bennett, 2000. The role of landscape and habitat characteristics in limiting abundance of grassland nesting songbirds in an urban open space. Landscape and Urban Planning 48: 65 - 82. Hansen, A. J., T. A . Spies, F. J. Swanson and J. L . Ohmann, 1991. Conserving biodiversity in managed forests. Bioscience 41: 382-392. Hansen, A. J, W. C. McComb, R. Vega, M . G. Raphael and M . Hunter, 1995. Bird habitat relationships in natural and managed forests in the west Cascades of Oregon. Ecological Applications 5: 555 - 569. Hansson, L., 2000. Interactions by vascular plants and birds between isolated ancient 58  oak-hazel woods and a matrix of conifer plantations. Biological Conservation 95: 191 - 196. Hebda, R. J., 1993. Natural history of the Garry Oak. pp. 3-7 in Hebda, R. J. and F. Aitkens, eds. Garry Oak-meadow colloquium. Proceedings, Victoria, B C . Hill, M . O. and H. G. Gauch, Jr., 1980. Detrended correspondence analysis: an improved ordination technique. Vegetatio 42: 47-58. Hutto, R. L . and J. S. Young, 1999. Habitat relationships of landbirds in the Northern Region, U S D A Forest Service. General Technical Report RMRS-GTR-32. Ogden, UT. US Department of Agriculture, Forest Service, Rocky Mountain Research Station, 72p. Inouye, D. W., W. A. Calder and N . M . Waser, 1991. The effect of floral abundance on feeder census of hummingbird populations. Condor 93: 279 - 285. James, F. C. and N . O. Warner, 1982. Relationships between temperate forest bird communities and vegetation structure. Ecology 63: 1 6 9 - 171. Jokimaki, J. and J. Suhonen, 1998. Distribution and habitat selection of wintering birds in urban environments. Landscape and Urban Planning 39: 253-263. Kilgo, J. C, R. A . Sargent, K . V . Miller and B. R. Chapman, 1997. Landscape influences on breeding bird communities in hardwood fragments in South Carolina. Wildlife Society Bulletin 25: 878 - 885. Koenig, W. L . and J. M . H. Knops, 2001. Seed-crop size and eruptions of North American boreal seed-eating birds. Journal of Animal Ecology 70: 609-620. Krannitz, P. G. and C. Rohner, 2000. Habitat relationships of endangered grassland birds in the south Okanagan. pp. 823-829 in Darling, L. M . , ed. Proceedings on a Conference on the Biology and Management of Species and Habitats at Risk. University College of the Cariboo and British Columbia Ministry of Environment Lands and Parks, Victoria B C . Lahti, D. C , 2001. The "edge effect on nest predation" hypothesis after twenty years. Biological Conservation 99: 365-374. Lancaster, R K . and W. E. Rees, 1979. Bird communities and the structure of urban habitats. Canadian Journal of Zoology 57: 2358-2368. Lea, T. 2002. Historical Garry oak ecosystems of Greater Victoria and Saanich Peninsula - A 1:20000 map. pp. 24-27 in Burton, P. J., ed. Garry Oak Ecosystem Restoration: Progress and Prognosis. Proceedings of the Third Annual Meeting of the B.C. Chapter of the Society for Ecological Restoration, April 27-28, 2002, University of Victoria. B.C. Chapter of the Society of Ecological Restoration, Victoria, B C . Legendre, P. and S. Legendre, 1998. Numerical Ecology. 2 English Edition. Elsevier Scientific Publishing, Amsterdam. Margules, C. R. and R. L. Pressey, 2000. Systematic conservation planning. Nature 405: 243-253. Mariani, J. M . and D. A. Manuwal, 1990. Factors influencing brown creeper (Certhia americana) abundance in the southern Washington Cascade range. Studies in Avian Biology 13: 53 - 5 7 . Martin, T. E., 1992. Breeding productivity considerations: what are the appropriate habitat features for management? pp. 455 - 473 in Hagan III, J. M . and D. W. Johnston, eds. Ecology and conservation of Neotropical migrant landbirds. Manomet Bird Observatory, Smithsonian Institution Press, Washington, DC. Marzluff, J. M . and K. Ewing, 2001. Restoration of fragmented landscapes for the conservation of birds: A general framework and specific recommendations for nd  59  urbanizing landscapes. Restoration Ecology 9: 280 - 292. McDonnell, M . J. and S. T. A . Pickett, 1990. Ecosystem structure and function along urban-rural gradients: an unexploited opportunity for ecology. Ecology 71: 1232 1237. McGarigal, K . and W. C. McComb, 1995. Relationships between landscape structure and breeding birds in the Oregon coast range. Ecological Monographs 65: 235 - 260. McShea, W. J. and J. H . Rappole, 1997. Variable song rates in three species of passerines and implications for estimating bird populations. Journal of Field Ornithology 68: 367-375. Melles, S. J., 2000. Effects of landscape and local habitat features on bird communities: a study of an urban gradient in Greater Vancouver. MSc. Thesis, Centre for Applied Conservation Biology, University of British Columbia, Vancouver, B C Mills, G. S., J. B . Dunning, Jr. and J. M . Bates, 1989. Effects of urbanization on breeding bird community structure in southwestern desert habitats. Condor 9 1 : 4 1 6 - 428. Nichols, J. D., J. E. Hines, J. R. Sauer, F. W. Fallon, J. E. Fallon and P. J. Heglund, 2000. A double observer approach for estimating detection probability and abundance from point coutns. Auk 117: 393-408. Palmer, M . W., 1993. Putting things in even better order: the advantages of canonical correspondence analysis. Ecology 74: 2215-2230. Paton, P. W. C , 1994. The effect of edge on avian nest success: how strong is the evidence? Conservation Biology 8: 17 - 26. Pavlacky, D. C. and S. H . Anderson, 2000. Habitat preferences of pinyon-juniper specialists near the limit of their geographic range. Condor 103: 322 - 331. Pojar, J. and A. MacKinnon, 1994. Plants of Coastal British Columbia. Lone Pine Publishing, Edmonton, A B . Ratti, J. T. and K. P. Reese, 1988. Preliminary test of the ecological trap hypothesis. Journal of Wildlife Management 52: 484 - 491. Rice, J., B. W. Anderson and R. D. Ohmart, 1984. Comparison of the importance of different attributes to avian community organization. Journal of Wildlife Management 48: 895-911. Robbins, C. S., 1981. Effect of time of day on bird activity. Studies in Avian Biology 6: 275-286. Robinson, S. K., 1992. Population dynamics of breeding neotropical migrants in a fragmented Illinois landscape, pp. 408 - 418 in in Hagan III, J. M . and D. W. Johnston, eds. Ecology and conservation of Neotropical migrant landbirds. Manomet Bird Observatory, Smithsonian Institution Press, Washington, DC. Robinson, S. K . and D. S. Wilcove, 1994. Forest fragmentation in the temperate zone and its effect on migratory songbirds. Bird Conservation International 4: 233 - 249. Robinson, S. K., F. R. Thompson III, T. M . Donovan, D. R. Whitehead and J. Faaborg, 1995. Regional forest fragmentation and the nesting success of migratory birds. Science 267: 1987- 1990. Roemer, H . L., 1972. Forest Vegetation and Environments on the Saanich Peninsula, Vancouver Island. PhD dissertation, University of Victoria, Victoria, B C . Rogers Jr., R. E., 2000. The status and microhabitat selection of streaked horned lark, western bluebird, Oregon vesper sparrow and western meadowlark in western Washington. Thesis, Environmental Studies, Evergreen State College, Olympia, WA. Rosenstock, S. R., 1998. Influence of Gambel oak on breeding birds in Ponderosa pine 60  forests of Northern Arizona. Condor 100: 485-492. Rosenstock, S. S., D. R. Anderson, K . M . Giesen, T. Leukering and M . F. Carter, 2002. Landbird counting techniques: current practices and an alternative. Auk 119: 46-53. Rottenborn, R. C , 1999. Predicting the impacts of urbanization on riparian bird communities. Biological Conservation 88: 289-299. Saab, V., 1999. Importance of spatial scale to habitat use by breeding birds in riparian forests: a hierarchical analysis. Ecological Applications 9: 135-151. Sauer, J. R., B. G. Peterjohn and W. A . Link, 1994. Observer differences in the North American Breeding Bird Survey. Auk 111: 50-82. Saurez, A . V., K. S. Pfennig and S. K . Robinson, 1997. Nest success of a disturbancedependent songbird on different kinds of edges. Conservation Biology 11: 928 935. Savard, J-P. L., P. Clergeau and G. Mennechez, 2000. Biodiversity concepts and urban ecosystems. Landscape and Urban planning 48: 131-142. Schieck, J., K. Lertzman, B. Nyberg and R. Page, 1995. Effects of patch size on birds in old-growth montane forests. Conservation Biology 9: 1072-1084. Schwab, F. E. and A. R. E. Sinclair, 1994. Biodiversity of diurnal breeding bird communities related to succession in the dry Douglas-fir forests of southeastern British Columbia. Canadian Journal of Forest Research 24: 2034-2040. Sisk, T. D., N . M . Haddad and P. R. Ehrlich, 1997. Bird assemblages in patchy woodlands: modeling the effects of edge and matrix habitats. Ecological Applications 7: 1170- 1180. Skirvin, A . A., 1981. Effect of time of day and time of season on the number of observation and density estimates of breeding birds. Studies in Avian Biology 6: 271-274. Soule, M . E., D. T. Bolger, A . C. Alberts, J. Wright, M . Sorice and S. Hill, 1988. Reconstructed dynamics of rapid extinctions of chaparral-requiring birds of urban habitat islands. Conservation Biology 2: 75 - 92. Statistics Canada, 2002. Statistics Canada's Internet site U R L : httpW: www.statcan.ca StatSoft, Inc. (2000). STATISTICA for Windows [Computer program manual]. StatSoft, Inc, Tulsa, OK. Szaro, R. C. and R. P. Balda, 1979. Bird community dynamics in a ponderosa pine forest. Studies in Avian Biology 3: 1-66. Ter Braak, C. J. F., 1986. Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology 67: 1167-1179. Ter Braak, C. J. F., 1994. Canonical community ordination. Part I: basic theory and linear methods. Ecoscience 1: 127-140. Ter Braak, C. J. F., 1995. Ordination. Pages 91-173 in R. H . G. Jongman, C. J. F. ter Braak and O.F.R. van Tongeren, eds. Data Analysis in Community and Landscape Ecology. Cambridge University Press, Cambridge. Ter Braak, C. J. F. and P. Smilauer, 1998. C A N O C O reference manual and user's guide to Canoco for Windows. Software for community ordination (version 4). Microcomputer Power, Ithaca, N Y . Tewksbury, J. J., S. J. Hejl and T. E. Martin, 1998. Breeding productivity does not decline with increasing fragmentation in a western landscape. Ecology 79: 2890 2903. Thilenius, J. F., 1968. The Quercus garryana forests of the Willamette Valley, Oregon. Ecology 49: 1124-1133. 61  Thysell, D. R. and A. B . Carey, 2001. Quercus garryana communities in the Puget Trough, Washington. Northwest Science 75: 219 - 235. Tilghman, N . G., 1987. Characteristics of urban woodlands affecting breeding bird diversity and abundance. Landscape and Urban Planning 14: 481-495. Tveten, R. K . and R. W. Fonda, 1999. Fire effects on prairies and oak woodlands on Fort Lewis, Washington. Northwest Science 73: 1 4 5 - 158. Ussery, J. G. and P. G. Krannitz, 1998. Control of Scot's broom (Cytisus scoparius (L.) Link): the relative conservation merits of pulling versus cutting. Northwest Science 72: 268-273. Van Home, B., 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47: 893 - 901. Vemer, J. and L. V . Ritter, 1985. A comparison of transects and point counts in oak-pine woodlands of California. Condor 87: 47-68. Villard, M - A . , K. E. Freemark and G. Merriam, 1992. Metapopulation theory and Neotropical migrant birds in temperate forests: an empirical investigation, pp. 474 482 in Hagan III, J. M . and D. W. Johnston, eds. Ecology and conservation of Neotropical migrant landbirds. Manomet Bird Observatory, Smithsonian Institution Press, Washington, DC. Wagner, J. L., 1981. Seasonal changes in guild structure : oak woodland insectivorous birds. Ecology 62 : 973-981. Ward, D. and J. N . M . Smith, 2000. Interhabitat differences in parasitism frequencies by brown-headed cowbirds in the Okanagan valley, British Columbia, pp 210-219 in Smith J. N . M . , T. L . Cook, S. I. Rothstein, S. K . Robinson and S. K . Sealy, eds. Ecology and management of cowbirds and their hosts. University of Texas Press, Austin, T X . Wiens, J. A., 1974. Habitat heterogeneity and avian community structure in North American grasslands. American Midland Naturalist 91: 195-213. Wiens, J. A., 1994. Habitat fragmentation: island v. landscape perspectives on bird conservation. Ibis 137: S97 - SI04. Wiens, J. A., J. T. Rotenberry and B. Van Home, 1987. Habitat occupancy patterns of North American shrubsteppe birds the effects of spatial scale. Oikos 48: 132 - 147. Wilcove, D. S., 1985. Nest predation in forest tracts and the decline of migratory songbirds. Ecology 66: 1211 - 1214. Wilson, D. M . and J. Bart, 1985. Reliability of singing bird surveys: effects of song phenology during the breeding season. Condor 87: 69-73.  62  in ireqan - spuBjdfi  >[B0 AjiBQ -  spuBjdfi  J3JIU03  m to.  co o m  in  C M —; d o o in  o o o  d CD  in  d  d  m  o o  in  o o  d  00  CM co d  in  o o  o o  o o  o o  •*  o  o  o  |o-  d  o d  o d  o d  o o o d  CO  - JUIOJ X>JOO _ d  o o o d  d  d  o in  o o  o o  m  in  d  o d  o m  o in  o o o  in  d  CO d  co  o  d  o d  in  in  o  co d  co d  o o d  in  o o  o o  - a^Bq sijaiij,  J3JIU03  - niioj XJJOO^I  m  m  o o o  C M —' d  5(B0 AjJBQ  CM d  d  o o o  o d  - 3>[BT SIJ3UJ,  co o  o o in  to-  |o-  oo o —•  }[BO XjJB£)  ireqjfi  }[BO /QiBQ  |o-  |o-  d J3JIU03 - SBjSnoQ iunop\[  >[BO AjJBQ SBjSnoQ junop\[  J3JIU00  - II!H  m  oo d |o-  CM teen d  y\vo XXIBQ  - IUH ^ B J M UBqjn  - JJOOJUSIH >(BO AJUBQ  CD  in  o d o  CM d  |o-  o d m  10. oo d  CM d  in  o o o  CM _  "  o o  - 5JOOJU3IH d  o d o  o d  o  o d o o  q d  d  q d  o o o  •—  d  o d o  o d m  CM d O)  CM d  d  o o m  o o o  o o o  o o  m  o o  o o  o o in  o o  o in  m  d  q d  C M —• d  o m  o o  CM m  CM d T—  lo-  in  co d  o d  d  q d o o  o d  o  o  OT  o d  o d  o  o  o o  o o  o o m  o o o  o  o  o d  o d  Ix-  o  in  C M CD d o o  o d o m  d o  in  CM d T—  o m  CM d m i»  lo-  in 00  d  m lo-  CO d o o  q d co  o o  o o  o o  o o o  m  o o o  in CM T—  d  C M —• d  d  d  o o o  d  o  CO  o  o  co  o  q  T—  O  o d  o d  —•  o d  d o  in  o d  o d  o  o  o d  o  in  o d  o d  o  C M •— d  C M —• d  in  in  o  o o d  CM co d  in  o  o in  o o  o in  in  CM d  oo d  o o  d  q d  o o  o o  o o  o m  o o  m  |o-  d  o o  o o o  o o o  CM d  CM  o o  o d  CM d  d  o m  CM d  d  o o o  d  d  d  o o o  o o  |o-  o d  d  o o  o m  C M •<+ d  d  d  lo-  oo d  d  o d  q d  O O  CM d  q d  o d o o  o d  o o o d  CM d  o o  o o  o d  o o o d  CM co d  o d  o o  O) CM d  fo  d  o o o  o d  d  o d m  d  C M —• d q d  CM d  o d  |o-  |o-  oo d  o m d o d o o in  d  o o o  o o o  o o o  o o  o o o  o o o  o o  o o o  d  d  d  d  d  CD  d  q d  d  2 -2 § « _» « © -< J3  H  3  Q j  63  o o o d  o o o d  o o o d  o o o d  o o q  o o q d  o o o d  o o q d  o m CM CM  0) CD O) co  o o o d  o o o d  co CD o d  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  O O O d  o o o d  o o o d  o o o d  m CM  o to CM d  o o o d  o o q d  m iooq d  in CM  in CM  d  d  o o o d  o o o d  o o o d  o o o d  o o q  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  co  o o o d  o o o d  o o q  o o q d  o o q d  o o o d  o o o d  CD oo CM d  CO  CD oo o d  CD oo CM d  o o o d  o o p d  o o o d  o o o d  o o q d  oo CO CO  o o q d  o o o d  o o o d  o o q d  o o o d  o o o d  o o q d  CD 00 CM d  o o o d  CO  o o  CM  o o  o d o  o o d  o o  o d  —•  d  •«*  d  d  o d  o  o  T—  o o  o o o d  o o o d  o o o d  o m CM CM  o o o d  o o  o d  oo co C) d  o o o d  o o  o d  o o o d  CD O 03 d  o o o d  o o o d  o o o d  o o o d  o o in d  o o o d  o o o d  o o CD d  m loCO  o o o d  o m CM d  m loCO d  o o o d  o o in d  m CM  m  d  o o o d  o o o d  o o q d  o o o d  co  co  co  co  ^—  —  d  o  o  d  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  CD CM  CO  CM  d  d  o o o d  d  o o o d  o o o d  loin CM d  o o q d  o o o d  in io. 00 d  oo CO C)  o o o d  o o o d  00 |od  o o o d  o o o d  o o o d  CO  o o o d  o o o d  o o o d  in CM d  o o o d  o o o d  o o o d  m loOO d  o in CM d  o o o d  o o o d  O) m d  o o o d  0) CM d  o o q d  o o o d  o o o d  o o o d  CO oo CM d  o o o d  o o o d  o o o d  o o o d  co  d  o o q d  loin 00 d  CD 00 CM  o o  o o  o o  o o  o o  o o  o o  in CM  o o  m loCO d  o o  o o  d  o d  o o  o d  o o ro. d  O in CM d  o  o  o  o  d  d  d  d  m CM CD d  d  o d  o d  o d  —•  o d  o d  <*  o d  o d  •<*  d  IO-  CS) —  —  o  o  o  o  o  o  d  d  d  d  d  d  d  d  d  in CM cq d  o o  o o  o o  o o  o o  o o  o o  o d  o in CM d  o o  m CM  o o  o o  d  o d  o o  o o fo.  co CD in co  o o o d  o o o d  o o o d  o in r-. d  o o o in  o o o d  o o o d  in CM CD d  o o o d  o o o d  o o o d  co to.  o  o  o  o  o  o  o  o  o  o  o  o  o  d  o o  in CM  o o d  o o o d  o o o d  o o o d  o  o  o  o  o o d  o d  —•  o  o  o o o d  o o d  o d  —•  in CM  o m d  o d  co in d  o o d  o o d  o  o o  o d  o o d  —•  o d  o o d  o o  o d  o o d  o o  o d  qo d  o o  o d  o CM d  o o  o d  o o d  o o  o o d  o q  o o d  o q  o d  m CO d  —•  o o d  o q  o d  o o d  o o  o d  o o d  m  o CD  CO oo CM  |od o o o d  ~ '  o o  o o q d  o  o o d  . £ 8  o o o o  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  o o m d  in CN  d  o o o co  o o o d  o  m  m  CM d  o in CD d  co d  co d  o o o d  o o o d  m CN in d  o o o d  m CN  co  o o o d  o o o d  o o o d  co  o o o d  CO  T  d  m d  ^|-  d  d  co  co  ^_ oo CN d  o o o d  o o p d  o o p d  o o p d  o o o d  o o p d  o o p d  o o o d  o o o d  o o o d  o o o d  o o o d  in CM  in CN  d  d  o o o d  o o o d  o in CN d  o o o d  o o o d  o o o d  o o o d  o m CN d  in  o o o d  in CN  m  o o o d  o o o d  o o o d  m CN  o o o d  o o o d  m CN d  in CM CD d  o o o d  o o o d  o o o d  o o o d  CO  o o o  o o o d  o o o d  o o o d  CO  CO  d  CD oo CN d  T  d  o o o d  m d  o o o d  CD oo CN d  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  co d  d  T—  d  d  oo d  T—  T—  o T—  m  T—  c3 0 > u u J3 ID O  §3 T  C O C CM 3 X>  <  d  -6 No i> X> r*- -4o—» c  3o o  •<*  d  d  00 d  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  00 CO  d  o o o d  o o p d  o o p d  o o o d  o o o d  o o in  o o o d  o o o d  o o o d  o o o d  o o o d  o o p d  o o o d  o o o d  o o o d  o o o d  co CO CN d  05 CN  o o o d  hm 00  o o o d  o o o d  o o o d  o o o d  CO  d  o o o d  CD 00 CN  o o o d  o o o d  o o o d  CN  CN  o o o d  m  o m CM d  o o m d  in CN CD d  o o o d  m CN  d  o o in  m CN  d  o o o d  o o o d  o m CN  d  o  m  CN d  d  T  ~  d  o o o d  m  m  co d  CO d  oo d  o o o d  o o q d  o o p d  o o p d  o o o d  o in CN d  o o o d  o o o d  o o o d  in oo d  o o o d  o  o  o  d  o o o d  IX) i*CN "•~  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  05 CD  d  o p d  o o o d  o o o d  o o o d  o o o d  o o  o o o d  o o o d  o o o d  in CN T—  d  o p d o o o d  T  ~  in  d  oo d  o o o d  o o o d  o o o d  o o o d  o o p d  o o o d  o p d  o o o d  o o o d  o o o d  o o o d  o o o d  o o o d  o o p d  o o o d  o o o d  m CN  T—  T3 -a  CO CN  in —  3 o  13  C3  3 -o  •>  '> -3 .g w  JS a  X> oo c CD o o  c o  ctl—1 , o  c o  D  C3 R o5 .ts xi .22 u o c ^ s  O) 3 o  w  ID  CO o  $  ID  c 3  11  •8 3 o -5 ° o  65  APPENDIX 2: The top three linear regression models describing the relationship between bird abundance and environmental variables. The top models are ranked according to AICc. Model weight and parameter importance were used in conjunction with AICc to select the model used for inference. Best approximating models  # parameters  AICc  AAICc  Weight Importance  American robin humans oakvol + humans oakvol  3 4 3  -31.87 -30.81 -29.99  0 1.05 1.88  0.34 0.20 0.13  0.73  Anna's hummingbird humans area + humans oakvol + humans  3 4 4  -29.38 -27.38 -27.32  0 2.00 2.06  0.36 0.13 0.13  0.60 0.39 (area) 0.38 (oakvol)  Bewick's wren null oakvol + humans humans  2 4 3  -13.93 -11.71 -11.53  0 2.22 2.40  0.44 0.15 0.13  0.33 (oakvol) 0.38  Brown-headed cowbird oakvol oakvol + humans oakvol + area  3 4 4  -26.84 -25.18 -24.84  0 1.66 2.00  0.50 0.21 0.18  0.97 0.29 (humans) 0.26 (area)  brown creeper oakvol + humans oakvol + humans + area area + humans  4 5 4  -23.79 -22.57 -16.78  0 1.22 7.01  0.61 0.33 0.02  0.96 (oakvol) 0.95 (humans) 0.37 (area)  chestnut-backed chickadee oakvol + humans oakvol + humans + area humans  4 5 3  -24.23 -23.30 -23.26  0 0.92 0.97  0.31 0.19 0.19  0.67 (oakvol) 0.35 (area) 0.75 (humans)  chipping sparrow oakvol area area + humans  3 3 4  -25.56 -25.00 -24.70  0 0.56 1.86  0.29 0.22 0.11  0.54 0.47 0.37 (humans)  dark-eved junco humans area + humans oakvol + humans  3 4 4  -21.68 -19.47 -19.28  0 2.21 2.40  0.47 0.16 0.14  0.83 0.28 (area) 0.26 (oakvol)  0.46  66  oakvol + humans oakvol + humans + area oakvol  4 5 3  -22.20 -19.92 -17.71  0 2.28 4.49  0.63 0.20 0.07  0.87 (humans 0.26 (area) 0.91  house wren' humans area oakvol  3 3 3  -30.06 -29.68 -29.28  0 0.39 0.78  0.27 0.23 0.19  0.51 0.45 0.37  orange-crowned warbler humans oakvol + humans area + humans  3 4 4  -20.40 -19.62 -18.10  0 0.78 2.30  0.36 0.24 0.11  0.80 0.43 (oakvol) 0.26 (area)  Pacific-slope flvcatcher null oakvol + humans humans  2 4 3  -11.58 -7.55 -6.20  0 4.02 5.37  0.72 0.10 0.05  0.19 (oakvol) 0.19  red-breasted nuthatch area area + humans area + oakvol  3 4 4  -44.43 -43.55 -42.54  0 0.88 1.89  0.39 0.25 0.15  0.93 0.43 (humans 0.35 (oakvol)  spotted towhee humans null area + humans  3 2 4  -7.30 -6.31 -5.48  0 0.99 1.82  0.25 0.24 0.11  0.70  Townsend's warbler humans oakvol + humans area + humans  3 4 4  -30.60 -28.41 -28.16  0 2.17 2.44  0.55 0.18 0.16  0.95 0.27 (oakvol) 0.24 (area)  Wilson's warbler oakvol humans area  3 3 3  -25.26 -24.89 -24.71  0 0.38 0.55  0.27 0.23 0.21  0.48 0.40 0.40  winter wren oakvol + humans oakvol oakvol + humans + area  4 3 5  -16.75 -14.41 -14.06  0 2.34 2.69  0.46 0.14 0.12  0.64 (humans 0.78 0.28 (area)  species richness null humans oakvol + humans  2 3 4  20.32 65.28 67.68  0 44.96 47.36  1.00 -  -  0.23 (area)  1  -  These species have three equally likely single variable models. A final model could not be chosen for inference.  67  


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items