UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effects of landscape and local habitat features on bird communities : a study of an urban gradient in… Melles, Stephanie J. 2000

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

Item Metadata


831-ubc_2001-0075.pdf [ 5.09MB ]
JSON: 831-1.0099590.json
JSON-LD: 831-1.0099590-ld.json
RDF/XML (Pretty): 831-1.0099590-rdf.xml
RDF/JSON: 831-1.0099590-rdf.json
Turtle: 831-1.0099590-turtle.txt
N-Triples: 831-1.0099590-rdf-ntriples.txt
Original Record: 831-1.0099590-source.json
Full Text

Full Text

E F F E C T S O FL A N D S C A P E A N D L O C A L H A B I T A T F E A T U R E S O N BIRD COMMUNITIES:  A S T U D Y O FA N U R B A N G R A D I E N T  G R E A T E R  IN  VANCOUVER.  by Stephanie J. Melles B . S c , The University of Toronto, 1994  A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF T H E REQUIREMENTS FOR T H E D E G R E E OF M A S T E R OF SCIENCE in T H E F A C U L T Y OF G R A D U A T E STUDIES T H E F A C U L T Y OF FORESTRY Department of Forest Sciences Centre for Applied Conservation Biology  We accept this thesis as conforming to the/equired standard  T H E UNIVERSITY OF BRITISH C O L U M B I A December, 2000 © Stephanie J. Melles, 2000  U B C Special Collections - Thesis Authorisation Form  Page 1 of 1  In p r e s e n t i n g t h i s t h e s i s i n p a r t i a l f u l f i l m e n t o f t h e requirements f o r an advanced degree a t the U n i v e r s i t y of B r i t i s h Columbia, I agree t h a t t h e L i b r a r y s h a l l make i t f r e e l y a v a i l a b l e f o r r e f e r e n c e and study. I f u r t h e r agree that p e r m i s s i o n f o r e x t e n s i v e c o p y i n g of t h i s t h e s i s f o r s c h o l a r l y purposes may be g r a n t e d by t h e head o f my department o r by h i s o r h e r r e p r e s e n t a t i v e s . I t i s u n d e r s t o o d t h a t c o p y i n g o r p u b l i c a t i o n o f t h i s t h e s i s f o r f i n a n c i a l g a i n s h a l l not be a l l o w e d without my w r i t t e n p e r m i s s i o n .  Department o f  f-mn^t,  ^  UJ^XJU^  The U n i v e r s i t y o f B r i t i s h Columbia Vancouver, Canada  httpV/vv^vvw.library.ubc.ca/spcoll/thesauth.html  12/11/00  11  ABSTRACT  Bird species diversity and abundance trends in urban areas can provide evidence to predict the relative importance of local bird habitat under different landscape contexts. I examined the hypothesis that selected species and nesting guilds should be more closely associated with landscape level features, such as proximity to large forested areas (< 100 ha), than with local scale habitat measures (< 1 ha). I collected avian community data during surveys completed over a two year period at 285 point count stations along four linear road transects located in Vancouver and Burnaby, British Columbia. Stations were located along transects bisecting three large parks (>324 ha) and proceeding away from these parks along residential streets into highly urban and suburban areas. A total of 49 breeding bird species were observed including 36 common species and 13 species that were sighted only once. Canonical correspondence analysis was used to view the main associations between measured habitat variables and species distributions. Species richness declined with increasing urbanization and the gradient (CCA axis one site scores) was dominated by landscape level habitat measures. Park area-by-distance metrics, developed using G.I.S., had the highest correlation with C C A axis one indicating the importance of park area in the vicinity for many species of birds breeding in marginal residential areas. Different land use zones did not neatly separate the urbanization gradient into simple bird habitat categories. Habitat models were created for five nesting guilds and three selected species (Song Sparrow, Melospiza melodia, Spotted Towhee, Pipilo maculatus, and American Robin, Turdus migratorius) with sequential block addition of landscape and local variables. Local variables significantly improved predictability of landscape variable only models, but the difference was slight. Landscape variables alone were often good predictors of presence or absence of most groups of species (guilds), but were less sensitive than local variables at predicting individual species presence. Incidence (percent stations occupied) of several bird species increased with park area in the vicinity as an inverse function of distance. The results of this study suggest that matrix areas surrounding parks and reserves should be integrated into urban planning and development designs.  iii  T A B L E OF CONTENTS  ABSTRACT  ii  T A B L E OF CONTENTS  iii  LIST OF T A B L E S LIST OF FIGURES  v '.  vi  A R C H I V E L O C A T I O N OF ORIGINAL D A T A  vii  ACKNOWLEDGEMENTS  viii  INTRODUCTION  1  METHODS Study Area Bird Surveys Habitat Characteristics Derived Park Area and Locality Measures Landscape Habitat Cover Local Vegetation Definition of Landscape Land use Categories Statistical Methods Correlation Analysis between Years Canonical Correspondence Analysis Logistic Regression Species Incidence Functions  10  ,  ;  10 13 14 14 15 17 19 20 22 23 25 26  RESULTS Bird Occurrence and Relative Abundance Between Year Comparisons of Nesting Guilds Bird Community - Habitat Relationships Land Use Habitat Positions along the Urbanization Gradient A) Characteristics of Habitat Categories - Landscape Level Features B) Characteristics of Habitat Categories - Local Level Habitat C) Characteristics of Habitat Categories - Nominal variables Bird Nesting Guild-Habitat Relationships Bird Species Incidence as a Function of Park-space by Distance  28 28 28 32 44 44 46 47 48 57  DISCUSSION Prediction 1, Bird Species along a Gradient of Urbanization Prediction 2, Landscape verses Local Habitat Predictors Prediction 3, Urban Land Use Categories, a Poor Indicator of Bird Habitat Types?  62 63 65 68  IV  Prediction 4, Species Incidence as a Function of Park Area and Distance Speculation - Dispersal from Natal Area? Study Design Critique  70 72 73  L I T E R A T U R E CITED  79  APPENDIX I  86  APPENDIX I cont  87  APPENDIX II  88  APPENDIX III  89  APPENDIX IV  90  APPENDIX V  95  V  LIST OF T A B L E S  T A B L E 1.  T A B L E 2.  T A B L E 3.  T A B L E 4.  T A B L E 5.  T A B L E 6.  T A B L E 7.  Detection frequency and nesting guild category for 36 species of breeding birds recorded at 285 point count stations in Greater Vancouver, B.C. - 19971998  29  Simple correlations among (LoglO(X+l)) detections of urban bird nesting guilds between the 1997 and 1998 seasons to examine for year effects  31  Canonical correlation coefficients, percents of variance, and redundancies between nesting guilds in 1997 and nesting guilds in 1998 with their corresponding canonical variates  33  Correlation analysis between (Log (X+l)) detections of urban bird nesting guilds for breeding seasons, 1997 and 1998, with their canonical variates to examine community associations between the two year sets of seven guilds  34  Correlation coefficients between habitat variables at local and landscape levels with the first and second canonical correspondence axes in an urban bird biodiversity study in Vancouver - Burnaby, British Columbia  40  Model selection: fit of species richness at road survey sites with C C A axis one (urbanization gradient) Vancouver - Burnaby, British Columbia, 1997-1998  42  Sequential logistic regression analyses for urban bird species and nesting guilds, presence/absence at 285 sites for three sets of predictor variables. Block 1) landscape variables (<100 ha) , block 2) local habitat variables (<1 ha) , and block 3) interactions  50  Curve model fit for three selected bird species of Park 500-760 m Index verses percent stations occupied, Vancouver - Burnaby, British Columbia, 1997-1998  61  10  1  2  T A B L E 8.  3  VI  LIST OF FIGURES  Figure 1.  Figure 2.  Figure 3.  Figure 4.  Figure 5.  Figure 6.  Figure 7.  Figure 8.  Figure 9.  Figure 10.  Study area, parks and urban greenspace (black areas), and transect locations (numbers 1 -4) for breeding bird and habitat sampling in the Greater Vancouver area, British Columbia  11  Representation of park-distance variable as the amount and proximity of park around each bird count station in the Greater Vancouver area  16  Example of a point count station (dot in centre of circle) and vegetation plot layout with 50 m radius, overlaid on a Vancouver, B.C. city zoning base map (1:2,000)  18  Schematic representation of urban landscape categories (1-6), point count stations along roadside transects through Vancouver and Burnaby  21  Canonical correspondence ordination diagram of 36 bird species to examine the strength of association between local and landscape level habitat variables in urban biodiversity study  38  Canonical correspondence ordination diagram showing the distribution of for point- count stations categorized by habitat along the axes. Points correspond to station scores as predicted by the axes (arrows are the same as in Figure 5)  39  Curvilinear and linear regression model fit with 95% confidence limits for C C A axis one versus maximum avian species richness (including rare species sighted during the breeding season) at point count stations along four roadside transects in Vancouver and Burnaby, 1997-1998  41  Star plots for landscape habitat and local level habitat characteristics recorded at 285 point count stations (grouped by landscape category) in the Greater Vancouver area, 1997-1998  45  A-B) Predicted incidence (% stations occupied) for common bird species associated with the Park 0 - 260m index (from canonical correspondence analysis, Figure 5). C) Incidence plots for common bird species associated with Park 500-760m (from canonical correspondence analysis, Figure 5)  59  Incidence (percent sites occupied) verses Park 500 - 760 m index for three selected bird species, Vancouver - Burnaby, British Columbia (refer to Table 8 for quadratic fit parameter estimates)  60  Vll  A R C H I V E LOCATION OF ORIGINAL D A T A  The original data used in this thesis are not reproduced here. However, archives of all the source data have been provided to the Data Services Department of Walter C. Koerner Library, University of British Columbia. This data may be requested directly from the Data Library or by phone: (604) 822-6742.  Vlll  ACKNOWLEDGEMENTS  This project was funded in part with generous support from the Asa -Yohal Graduate Fellowship in Forestry and the Bert Hoffmeister Scholarship in Forest Wildlife. The Centre for Applied Conservation Biology, the Geography department at U B C , and the Canadian Wildlife Service also provided many resources. I would like to thank my co-supervisors, Professor's Susan Glenn and Kathy Martin. Without Susan Glenn's enthusiasm, trust, generous advice, and theoretical savvy this project would not have been realized; and Kathy Martin, whose scientific ardor and expertise I admire and aspire to, thank you so much for your support and for caring. My other committee members, Professor's Brian Klinkenburg and Tony Sinclair - Brian you were always available to provide insight, humour, and surprisingly quick solutions to all of my GIS related problems. Your help and suggestions were always appreciated. Tony Sinclair, you were particularly instrumental in the beginning of this endeavour, sitting me down and getting right into the nature of how to ask a question. Thank you. Several people at the Centre for Applied Conservation Biology and the Statistical Department at U B C were able to assist me with statistical and data related inquiries, Andreas Hammond, Milosh Ivkovich, Glenn Sutherland, Pierre Vernier, Ralph Wells, and the graduate students of Statistical Consulting, STATS_500, as well as many others. When it comes to writing, Glenn Sutherland, I esteem your words and thank you for giving so generously. Finally, to _all_ my good friends, and particularly "The Girls" - Heather and Christine, Anne O', Lisa, Caitlin, Devon, Susan and Susan, and lastly to Milosh: It is with a tremendous amount of sadness that I write these words knowing that the time we have shared has come to pass, and that I may no longer have your friendship so near at hand. My respect for you has only increased with my knowing you and I trust and hope that our friendship will continue always. Special thanks to The favourite Molestics for supreme bits of Hokum.  INTRODUCTION  Worldwide, urban areas are expanding both in size and number with ever-increasing human populations. By the year 2030, the number of people living in cities is expected to increase by 2 billion inhabitants, reaching an estimated 4.9 billion or 60% of the expected global population (UN Population Division 1999). As a result of urban expansion, native vegetation is reduced and fragmented over a landscape matrix in which both the amount of impervious surface is increased, and the structure and composition of the remaining vegetation is progressively altered (Beissinger and Osborne 1982, Arnold and Gibbons 1996, Germain et al. 1998, Marzluff et al. 1998). Cities are typically located in coastal areas, close to rivers and estuaries, or near large bodies of inland water (40% of cities with populations greater than A million are located on X  coasts, WRI 1996). Therefore, urbanization is biased to bottomland and riparian systems that often support more species of breeding birds than extensive surrounding upland areas (Knopf al. 1988, Ohmart 1994). Large parks and reserves in urban areas often support high species diversity because these protected areas are the habitat 'fragments' of exceptional ecosystems (Schaefer 1994). Increasing urbanization adjacent to natural areas and parks often results in simplified habitats and this leads to a community of birds with fewer species dominated by superabundant flocks of exotic species (Marzluff et al. 1998). Conservation biologists have been predominantly interested in the protection of natural ecosystems and have placed little importance on urban areas or urban biodiversity overall (Vandermeer 1997, Jules, 1997). Some studies on urban birds have focused on species richness within a patch of native habitat (e.g., a large park) in relation to size of patch and degree of isolation from other contiguous areas of habitat (Tilghman 1987, Diamond 1988, Soule et al.  2  1988). Other studies of urban avifauna have focused on the habitat associations affecting birds at local spatial scales (2 to <25 ha, Emlen 1974, Weber 1972, Campbell and Dagg 1976, Lancaster and Rees 1979). Citywide surveys are rare (Hadidian et al. 1997), and little is known about how the juxtaposition of different habitats (i.e., high density housing and suburban) affects avian diversity both within each habitat type and at varying distances from the edge of a large forested area (Marzluff et al. 1998). However, land-use planners may benefit from greater understanding of how urban development affects birds in parks and reserves, in surrounding residential areas, and at the level of entire landscapes. Land use terms such as, rural, suburban, and urban, pose difficulties for urban bird research because they lack standard definitions and they attempt to define different habitat types along a complex environmental gradient that lacks distinct boundaries (Marzluff et al. 1998). There is a need for standard measures that accurately quantify the position of an urban or suburban site along this complex gradient, and to relate these findings to urban bird populations and community ecology (Marzluff et al. 1998, Clergeau 1995). Urban areas are generally defined by land use planners as areas having population densities over a minimum of 620 individual humans/km (McDonnell and Picket 1990) while suburban areas are typically defined as districts 2  located on the edge of a larger urban centre. When these land use categories are examined as different habitat types for birds, the findings from different urban areas are inconsistent. Some suburbs with low levels of development support quite varied bird communities with higher than expected bird species richness, including a mixture of native and non-native species (Emlen 1974, Rosenberg et al. 1987; DeGraaf 1991). Suburbs may have potential for land management practices that would enhance the value of these areas for birds (Blair 1996). However, vegetation is invariably altered with urbanization. Suburban areas rarely include the full complement of vertical strata found in natural forests (Beissinger and Osborne 1982), and native  3  plant species are often removed and/or replaced by exotic ornamentals (Rosenberg et al. 1987, Blair 1996). Birds may be responding directly to characteristics of the vegetation within suburban habitats or they may be responding to the proximity of both large forested areas and highly urbanized areas. Suburban areas may be replenished locally by continued immigration of individual birds from more-productive, forested areas nearby, so suburbs could be "sink" habitats where within-habitat reproduction is insufficient to balance local mortality (Pulliam 1988, Robinson et al. 1995). Urban bird community-level dynamics have been investigated in relation to at least three dominant ecological theories: 1) Stand level habitat modeling and habitat diversity theories, which examine species assemblage in relation to the structure and composition of resources at the local stand level scale (discussed in Gauch 1973, Shmida and Wilson 1985, McGuinness and Underwood 1986). 2) Island biogeography theory, which examines the effects of island isolation and size in relation to species richness, immigration, and extinction rates, and has been applied to habitat patch dynamics in fragmented urban areas (MacArthur and MacArthur 1961, MacArthur and Wilson 1967). 3) Spatial structuring theories of ecological communities where local scale habitat features are viewed in relation to larger scale surrounding landscape features (Bolger et al. 1997, Germain et al. 1998, Rottenborn 1999, Saab 1999). I followed this approach to examine bird communities in relation to features of the entire urban matrix. Urban bird communities have been examined in relation to the structure and composition of vegetation and food resources at the stand level with areas ranging in size from -16-50 ha city blocks (Weber 1972, Campbell and Dagg 1976, Lancaster and Rees 1979). Stand level habitat diversity findings have been fairly consistent, demonstrating that species richness (the number of  4  species) decreased while the relative abundance of common species increased in highly urbanized areas (i.e. commercial and industrial sites) compared to areas with lower residential housing densities (e.g., Emlen 1974, Blair 1996). Relative abundance and species occurrence in a habitat block or stand were related to individual species requirements, such as: weedy lawn cover, number of home feeders (Emlen 1974); foliage height diversity (Lancaster and Rees 1979); woody cover, water (Campbell and Dagg 1976, Edgar and Kershaw 1994); modification of vegetation, and park size (Gavareski 1976). However, the juxtaposition of habitats surrounding each habitat stand was not considered. Lower density residential areas may often be located in close proximity to regional parks, and may also have higher within-habitat diversity, vegetation structure and composition, than more highly urbanized areas with greater housing densities. Secondly, urban avian diversity in fragmented areas has been correlated with size of remaining habitat areas and, with measures of isolation distance from other 'islands' of habitat. In Springfield, Massachusetts, the diversity of birds correlated positively with size of woodland, water, and coniferous tree cover; and negatively with distance to the nearest park trail, and building density in areas immediately adjacent to 32 forest 'islands'. Avian richness in these urban forest patches was not related to (isolation) distance to the nearest extensive woods (>400 ha) in multiple regression models (Tilghman 1987). In canyons around the city of San Diego County, California, canyon occupancy by endemic chaparral scrub requiring birds was related to the size of remaining scrub habitat, fragmented by urban development (Soule et al. 1988). In addition, birds dispersing from nearby canyon habitat patches were expected to replace extirpated populations in small habitat patches; species richness was expected to decline inversely with distance from the nearest habitat patch; however, small patches in close proximity to the nearest occupied patch were found unoccupied. The poor dispersal abilities of these  5  native bird species apparently prevented them from migrating even short distances (<100 meters) through developed areas. However, some of these same species often breed in suburban areas (e.g. Bewick's Wren, Thryomanes bewicki, and Spotted Towhee, Pipilo erythrophthalmus). Therefore, the absence of these species could have been related to other, unmeasured or missing, habitat features of the larger landscape. Thus, the focus on 'island' isolation of strict habitat patches was perhaps too narrow. Finally, avian community dynamics can be related to habitat variation at different scales, features of the local habitat are sometimes inadequate to explain species occurrence and surrounding habitats should also be considered (Saab 1999). Previous research demonstrated that urbanization affected bird communities in habitats defined at the site-specific level of city blocks (e.g., Emlen 1974, Campbell and Dagg 1976, Lancaster and Rees 1979), but spatial scale studies indicated that the surrounding landscape features have more predictive power for understanding bird communities at the site level (Blair 1996, Germain et al. 1998, Saab 1999). Landscape level characteristics (<100 ha - <30,000 ha) were better predictors of bird species presence/absence, than local, site level (<3 ha) habitat characteristics (Smith and Schaefer 1992, Bolger et al. 1997, Germain et al. 1998, Rottenborn 1999). Similar studies in other urban areas are necessary to substantiate findings that both the spatial arrangements of the habitat, the physiognomy, and the habitat composition at the level of the landscape, influence the relative abundance and presence of bird species (Bolger et al. 1997). The ecological mechanisms generating community responses to landscape level spatial variability have not been well established (Bolger et al. 1997, Marzluff  al. 1998, Germain et  al. 1998). Research on bird dispersal and settlement patterns may help to explain spatial structuring in bird communities. Some evidence suggests that birds search for and select next year's nesting site during post-fledging exploration (Baker 1993). Therefore, habitat area and  6  proximity to the natal area will affect the occupancy of areas close to successful breeding sites (Baker 1993, Bolger et al. 1997, Helzer and Jelinski 1999, Sutherland et al. 2000). Dispersal of birds from larger forests of a region may determine bird assemblages of other smaller forest 'patches' (Freemark and Merriam 1986, Hinsley et al. 1996). Two ecological processes that may also help to explain the patterns of species breeding in complex landscapes are complementation and supplementation (Dunning et al. 1992). Complementation occurs when species require a complement of non-substitutable, fixed resources that are found in spatially discrete areas or patches. A patch may be defined as any non-linear surface area that differs in appearance from the surrounding areas (Forman and Godron 1986). Regions of the landscape where patches with different types of resources are in close proximity will support more individuals than regions where one patch type is relatively rare. Supplementation occurs when species survive in resource patches that are individually too small by supplementing similar resources from the surrounding areas, so long as the supplementary resources are accessible from the local area. Residential sites with foraging areas that are individually too small or of low nutrient quality may still attract birds if individuals can supplement their requirements with resources from nearby foraging patches, or by supplementing their standard food resources with food and complementary water and nesting sites provided by man. Evidently, there is considerable overlap between these ecological processes in urban systems with distance to resources being a potentially critical variable. Investigations of community assemblage necessarily encompass overlapping alternative hypotheses (Quinn and Dunham 1983). The variation observed in bird abundance and community composition due to urbanization is the result of many ecological factors operating at several observation scales (e.g., habitat area, habitat heterogeneity, isolation distance, food quantity and quality, landscape physiognomy, composition, predator abundance, observer bias,  7  Clergeau 1995, A M and Allen 1996). As a result, several interpretations and predictive models can be tested and applied. Past research relating local bird abundance to stand level habitat factors may be more appropriate under conditions where the local habitat is of excellent quality and where nesting success is high (Brawn and Robinson 1996, Arcese et al. 1992), but this perspective, if applied under conditions of poor quality local habitat, neglects to consider the influence of regional 'source pools' or landscape level factors (Sewell and Catterall 1998). Research relating local species occurrence and diversity patterns to island biogeography may apply when the landscape can be classified into suitable and unsuitable habitat. In a complex matrix, though, the presence and detection of individual birds is expected to vary probabilistically with different land use practices and with the overall amount of habitat or forest cover (Trzcinski 1999). In this study, I use the method of urban gradient analysis (McDonnell and Pickett 1990) to investigate bird species richness, species relative abundance, and groups of birds (by nesting habitat guild) in relation to changes in habitat along four urban transects within the municipalities of Vancouver and Burnaby. My main objective was to quantify the changes in bird abundance and diversity from large urban parks to downtown core areas, and to relate attributes of the bird community to habitat at two scales, the landscape scale (<100 ha) and the local or stand level (measured at <1 ha). I know of only one other study that has examined an urban bird community citywide at different scales (Haddidian et al. 1997). Although a few studies have examined the effects of surrounding urbanization on bird species richness in park and riparian fragments, none have looked at the effects of surrounding parks on the community of birds in residential areas. The municipalities of Vancouver and Burnaby are located within the Georgia Basin ecoprovince of British Columbia. The Georgia Basin lies between the Vancouver Island Mountains and the southern Coastal Mountains and represents 3% of the  8  British Columbia land base (2, 772,571 ha). The region supports the highest diversity of breeding birds in the province, contains three cities, and approximately three-quarters of the provincial population, 2.9 million people on the Canadian side (Demarchi 1996). Urban areas often have poor quality local habitat for birds and they may not maintain breeding populations of groups of sensitive species over time (Emlen 1974, Bolger et al. 1997, Rottenborn 1999). I examine the hypothesis that bird guilds should be more closely associated with landscape level habitat features, measured at scales < lOOha, than with local scale habitat measures, <1 ha, under the premise that surrounding habitats may act as resource areas for residential birds and 'source' areas for dispersing birds. I evaluate the following predictions for how patterns of urban bird communities should relate to local and landscape level habitat characteristics and urban land use patterns. Prediction 1) Bird species richness should decline and mean relative abundance of the remaining species should increase with increasing urbanization summarized by a habitat gradient. Prediction 2) Because residential areas outside of parks should have relatively poor quality, local habitat for native birds, they may be 'rescued' by surrounding regional parks with higher bird diversity. Broad scale habitat measures (<100 ha) should adequately describe the urbanization gradient and should be better predictors of bird species and nesting guild presence than local, site level habitat measures. Prediction 3) Because land use types are often defined by arbitrary boundaries, a continuous gradient should describe the matrix of bird habitat conditions better than categories of urban land use types.  9  Prediction 4) Species incidence (the proportion of sites occupied) should increase with proximity to parks of increasing area if birds are reaching high densities in parks and dispersing from these areas to more marginal nesting locations along residential streets. These predictions were tested using avian community data from surveys completed over a two year period along roadside transects through Vancouver and Burnaby. The local habitat within 50 meters (m) of point count stations was surveyed, landscape park-distance metrics were developed, and habitat cover variables were estimated at different landscape scales using a G.I.S. and aerial photographs. The urban gradient is summarized using the two scales of habitat variables, and the main associations between bird species distributions with this gradient are examined using canonical correspondence analysis ordination diagrams and regression of ordination axis one scores with species diversity. I produce nesting guild habitat models with logistic regression using sequential addition of local and landscape variables. The bird community was further examined in relation to different geographic areas or habitat types (e.g. suburban and urban parks) in order to determine their position along the urbanization gradient using local and landscape habitat measures. Next, I relate bird incidence (% stations occupied) to the amount of park in the vicinity with the park-distance metrics. I discuss the bird community response in relation to published literature on local and regional effects, avian dispersal, and community dynamics. Finally, I consider conservation implications of urbanization on a global scale.  10  METHODS  Study Area I conducted this research in the municipalities of Vancouver and Burnaby, hereafter "Greater Vancouver", British Columbia (49° 18' N , 123° 12' W, Figure 1) in the Fraser River Lowland area of the Georgia Depression. There are three large parks in this urban area. The campus of the University of British Columbia (UBC) is located on the westernmost peninsula of Vancouver surrounded by a 763 ha forested park, Pacific Spirit Park. Downtown Vancouver is predominantly commercial with the exception of another large park, Stanley Park (405 ha); this park also occupies a peninsula jutting into Burrard Inlet. Industrial areas are located east of the downtown core with residential areas and suburban areas in Burnaby (Figure 1). Simon Fraser University (SFU) campus, like U B C , is a developed piece of land (or 'encroachment') surrounded by another large park, Burnaby Mountain Regional Park (324 ha, G V R D Strategic Planning Department 1993). Directly across Burrard inlet, SFU campus faces mountainous watershed reserves along the north shore of Vancouver. The north shore has both residential and commercial development extending partially up the continuously forested mountains. Vancouver in the 1880's was a small settlement of sawmills, houses, and forest clearings surrounded by continuously forested land (Oke et al. 1992). The city has expanded over the past 119 years. The ocean and mountains have tended to constrain or direct Vancouver-Burnaby urban development. A consequence of development was the complete removal of forest and ground cover, later to be replaced with tree plantings along many city streets. Initially, the westend of Vancouver was planted with native tree species (Oke et al. 1992). However, these native species, such as Douglas fir (Pseudotsuga menziesii) and Western hemlock (Tsuga heterophylla)  12  were later deemed unsuitable to urban life since they grew too fast and their root systems buckled the sidewalks. Thus, as the city grew eastward and southward, the tree species planted were largely non-native species, over one-third being cherry or plum (Prunus sp.) trees. The result is an uneven distribution and composition of trees that are often larger and more likely native species on the west side than on the east side of the city, which has more deciduous, nonnative species in the 'newer' suburbs (Oke et al. 1992). These same trends continue to be followed today, although more diverse street tree plantings have replaced planted monocultures of trees. The Vancouver area has a cool, humid, mesothermal climate. It has cloudy, wet, and mild winters and sunny, dry, and warm summers (Weber 1972, Meidinger and Pojar 1991). Mean daily temperature values recorded during the study period ranged between 14.2-19°C (Appendix I) and were just slightly above normal mean daily temperatures for June-September. Precipitation means were also near normal levels (72 - 108mm at higher elevations or 40-76mm at lower elevations, Appendix I). My study area ranged in elevation from sea level to 370 m. The Coastal Western Hemlock zone is the dominant biogeoclimatic zone of this region (Green and Klinka 1994). The original vegetation of the area would have resembled a dense coniferous forest, with a shrub-dominated understorey. The climax vegetation of the C W H zone is generally dominated by a canopy of Western red cedar (Thuja plicata) and Western hemlock (Tsuga heterophylla), with Douglas fir (Pseudotsuga menziesii) in drier areas and minor amounts of Sitka spruce (Picea sitchensis), Yellow-cedar (Chamaecyparis nootkatensis) and Lodgepole pine (Pinus contorta). The most prominent species of the original understorey vegetation were Salal (Gaultheria shallon), Alaskan blueberry (Vaccinium alaskaense), False azalea (Menziesia ferruginea), and Red huckleberry (Vacciniumparvifolium), with Salmonberry (Rubus spectabilis) and Red elderberry (Sambucus racemosa) found on wetter sites. The supplanted  13  herb and dense moss layers were composed of Deer fern (Blechnum spicani), Bunchberry (Cornus canadensis), and False-lily-of-the-valley (Maianthemum dilatatum), Step moss (Hylocomium splendens), Lanky moss (Rhytidiadelphus loreus), and (Kindbergia oregana) Oregon-beaked moss (Weber 1972, Meidinger and Pojar 1991, Demarchi 1996). Pacific Spirit, Stanley, and Central Park, Vancouver, and Burnaby Regional Parks are relatively large areas within this urban context which still contain representative vegetation of the C W H zone; these parks also contain many species of non-native vegetation, such as English holly (Ilex aquifolium) and Himalayan blackberry (Rubus discolor). Botanical nomenclature follows Pojar and Mackinnon (1994) for vascular plants, Little (1980) for woody plants.  Bird Surveys I collected data on; species presence and absence, the total number of bird species detected, and the relative abundance of individual bird species at 285 point count locations along four road side transects in Vancouver and Burnaby. Point locations were sampled once within each of three periods (two breeding seasons and one fall season). Breeding season counts were conducted between 24 June-13 July 1997 and 1 May-1 June 1998 and the fall count was from 24 August-11 October 1997. A total of 46 bird surveys were completed, however, only bird census data from the breeding seasons were used for analyses (Appendix II). Point count stations were established along transects with an inter-station distance of 250 meters. Transect placement was subject to the constraint of road length and orientation. The fixed radius point count methodology, with count duration of five minutes, is a commonly used method for bird surveys and was employed to maximize sample size while still ensuring independence (Ralph et al. 1993). A 50-meter radius detection area was adopted as DeGraaf et al. (1991) found that in urban areas only 6.9% of birds were recorded at distances greater than 46 m from the centre of  14  street transect lines. I carried out all bird censusing myself, on clear days during the first four hours following sunrise to coincide with peak singing activity (Ralph et al. 1993, Robbins 1981). North American birds have well-known habitat affinities and can be grouped into guilds based on their breeding habitat associations (Haddidian et al. 1997, Ehrlich et al. 1988). Subsets of species, particularly ground and shrub nesting bird guilds, are of concern to land managers in urban areas because urban residential zones may be currently unsuitable for these groups of species (Rottenborn 1999). Birds were grouped into the following nesting habitat guilds: deciduous tree, coniferous tree, building, ground, shrub, cavity nesters, and ledge/cliff nesters (Ehrlich et al. 1988). The ledge/cliff, nesting guild was entirely comprised of gull species. Although, this guild nests on building ledges, it was maintained separately from building nesters since gulls are predominantly restricted to rocky terraces or sandy coasts. Birds were assigned to nesting guilds after Ehrlich et al. (1988) with the exception of the Violet-green Swallow (Tachycineta thalassina) and the European Starling (Sturnus vulgaris) which were assigned to the building nesting guild, since in urban areas they more commonly nest in buildings.  Habitat Characteristics Derived Park Area and Locality Measures I derived four distance by park variables at different radii, 0-260 m, 260-500 m, 500-760 m, 760-1000 m, to summarize the amount and proximity of park-space around each station. A digital map of land-use in the Greater Vancouver Regional District (GVRD Strategic Planning Department 1996) was used to develop a map of green-space in the study area, including parks greater than one hectare, cemeteries, and golf courses (IDRISI Version 4.0 1992, G.I.S. software). Universal Transverse Mercator (UTM) coordinates of the point count stations were registered to the map of urban green-space. Next, I created 20 m distance buffer intervals, in  15  concentric rings, around each station (Figure 2). Park area within each ring was divided by distance in order to weight parks closer to each station more highly than parks farther away creating a park exposure index. Thus, four composite, park-distance variables were calculated as follows:  1) 2) 3) 4)  Park0-260m Park260-500m Park500-760m Park 760-1000m  k  where P = park area within each 20 m 'donut' ring, and D = distance (i.e. 20 m, 40 m, 60 m....) and summed over k distance rings within each park-distance variable (e.g., for park index 0260 m, the total radius, 260 m, is divided by 20 m rings => 260/20 =13 rings, k=l3) Division by distance assumes that species presence should decrease non-linearly with distance from 'natural' areas. Research suggests that distances dispersed by birds and mammals from their natal areas follow the negative exponential distribution (Sutherland et al. 2000). Island biogeography predicts declining rates of immigration and increasing rates of extinction with distance from large natural areas, and spatial structuring predicts that close habitats are visited more frequently than distant ones (Dunning et al. 1992).  Landscape Habitat Cover Landscape cover variables were measured from digital aerial photographs to describe the landscape composition around each point. These photographs of Vancouver and the Fraser Valley (May/July, 1995 1:30,000), with an orthophoto accuracy registered to TRIM (Terrain Resource Information Management) (1:20,000), and a pixel size of one meter (m), were used to estimate cover of: salt water (SALT500), grass (GRASS500), impervious surface  Figure 2. Representation of park-distance variable as the amount and proximity of parkspace around each bird count station in the Greater Vancouver area. Contour lines are at 20 m intervals and the amount of park area within each 20m layer was divided by distance and summed across all radii, creating park exposure indices at different scales (see text).  17  (IMPERV500), coniferous (CONIF500) and deciduous (DECID500) trees within 500 m of each point count station. I centered an acetate dot-grid overlay representing a 500 m radius circle on each census point located on the aerial photo and recorded the percentage of each of these cover type (Blair 1996, Rottenborn 1999, Germaine et al. 1998, Bolger et al. 1997).  Local Vegetation City zoning base maps were used (1:2,000) to delineate the 50 meter boundaries for vegetation plots (Figure 3). I was able to identify the number of individual house lots (# HOUSE LOTS) within the plot boundary and to record the local habitat variables. The percentage of impervious surface cover changes from 60% for multi-family to 90% for commercial and industrial land uses (Dinicola 1990). When commercial and industrial buildings were present on a plot, fifteen house lot units were added to #HOUSE LOTS and ten units were added if apartment buildings were present. To assess the local habitat, I measured 30 additional variables around each point count station. The vegetation variables broadly describe land cover type (to a maximum of 100% total cover), and vegetation composition and structure. Composition was described simply by tree species richness (SPR), and by the relative proportion of deciduous to coniferous trees and shrubs. A separate category for berry producing trees and/or shrubs was used to assess food plant availability. I recorded the number of standing dead trees (DEAD) and fallen logs or downed wood (DOWN) greater than 15 cm diameter at breast height (dbh -1.3 m). The number of trees in four size classes were counted within 50 m of each station: <15 cm in diameter at breast height (dbh), >15 and <30 cm dbh, between 30 and 60 cm dbh, and > 60 cm dbh (DECID1-4, CONIF1-4). Shrub species > two m in height were considered 'trees' if their stems were distinct enough to be counted. Shrubs were similarly categorized as either deciduous or  18  Figure 3. Example of a point count station (dot in centre of circle) and vegetation plot layout with 50 m radius, overlaid on a Vancouver, B.C. city zoning base map (1:2,000).  19  coniferous and their numbers were counted within three size classes: < 1 m in height, 1 - 2 m, and > 2 m in height (D1-3shrub, CI-3shrub). I recorded additional local habitat characteristics of importance to birds with the following categorical variables around each 50 m radius point count: the number of street intersections (residential street, residential intersection, thoroughfare intersection), the presence or absence of nest boxes, bird feeders, and water. Water was classified as birdbath, fresh water streams, marine, or both. The percent cover of grass was estimated by sketching grass cover onto city zoning base maps for residential lots within the boundaries of each 50 m plot; these maps were later overlain with dot matrix acetate to estimate local grass cover. Developed sites that were completely surrounded by park-space (urban encroachments) were located at different elevations, elevation was estimated using TRIM (Terrain Resource Information Management) maps (1:20,000) with contours generated from a digital elevation model (Ministry of Crown Lands, British Columbia 1993). Finally, ASPECT (north, east, south, west) was recorded at each site.  Definition of Landscape Land use Categories Six different habitat types (e.g., urban, encroachments, suburban) are depicted (Figure 4), as they would occur along an urbanization gradient with decreasing landscape levels of native habitat or vegetation. All 285 point count sites were categorized into these groups in an attempt to simplify the entire gradient of urbanization into land use categories used by land planners, and to later investigate their ecological relevance in relation to bird species distributions. All counts were along roads, so they represent (road) edge habitat in different contexts. Roads through parks were surveyed and point count stations more than 250 meters inside a forested area larger than 300 ha in size were categorized as park sites; there were three large parks in the survey area,  20  Pacific Spirit (763 ha), Burnaby Mountain Regional (324 ha) and Stanley Park (405 ha). A minimum distance of 250m from the outside edge of these forested areas ensured that sites classified as park, although surveyed along a road edge, had a habitat context of interior forest in all directions (Murcia 1995, Robinson et al. 1995). Stations within 250m of the border of one of these large parks were classified as edge sites. Stations within a developed area more than 250m from a park edge but completely surrounded in all directions by park were classified as encroachment sites. Stations within urban parks, or parks completely surrounded by the developed matrix (>5ha in size, to avoid tree-lined baseball parks) were classified as urban parks. Within the City of Vancouver, stations were classified as urban sites, with an average population density of 4940 individuals/km and 13.5 properties/ha (GVRD Municipal and 2  Hospital Values 1999). Finally, stations within the politically defined suburban zone of Burnaby, with an average population density of 2150 individuals/km and 6.2 properties/ha (all types), 2  were classified as suburban sites (GVRD Municipal and Hospital Values 1999).  Statistical Methods A simple correlation matrix was examined and two of the landscape level park-distance variables (Park 260-500 m and Park 760-1000 m) were omitted from further analyses because they were highly correlated with Park 260 m and Park 500-760 m respectively, Pearson correlation coefficients >0.75. These two variables were chosen for omission, one, because the 260-500 m scale was already represented by landscape cover variables (e.g., IMPERV500, CONIF500). Two, the 760 to 1000 m scale level was omitted because it was highly correlated with the next inner ring and I did not want to leave a gap between measurement scales at 500 to 760 m. Next, I calculated descriptive statistics and created frequency distributions for all  250 m  1) Roadside park stations (n=35)  2) Edge (n=30)  5) Urban Park (n=14)  3) Urban encroachment, small developement area surrounded by park (n=23)  6)Urban(n=141)  4) Suburban (n=42)  Figure 4. Schematic representation of urban landscape categories (1-6). Stippled areas represent natural area parks. Point count stations along roadside transects through Vancouver and Burnaby were grouped (points within rectangles) into one of these six categories. Stations were 250 m apart along continuous transects.  22  environmental variables to check for data normality, skewness and kurtosis, and the necessity of data transformations (Sabin and Stafford 1990). Count data, with many values at or close to zero, were log transformed (Log (X+l)) to better meet the distribution requirements for the 10  explanatory habitat variables used in the analyses (ter Braak 1986, Hosmer and Lemeshow 1989, Jongman et al. 1995, Tabachnick and Fidell 1996). Variables with few discrete values were converted to presence/absence variables (i.e., fresh H 0 500 m, large berry producing trees 2  (BERRY1 & 2, 15-30 and > 30 cm dbh) and percentage cover variables with non-normal distributions were logit transformed (Log, (p/l-p)) following the recommendations of Sabin and 0  Stafford (1990). I removed species that were observed only once from the analyses to avoid potentially mistaking migrants or wandering individuals for breeding residents. The dimension of the response variables (number of bird species recorded within 50 m of each point count station) was reduced to 36 with the removal of singular sightings from the analyses (Appendix III).  Correlation Analysis between Years Correlation analysis (CO) was used to check for year effects in relative bird abundance, by nesting guild. Since birds were surveyed later in the 1997-breeding season (24 June to 13 July) than in the 1998 breeding season (1 May - 1 June), I used canonical correlation (CO) analysis to examine the relationships between the relative abundance of different nesting guilds in both years to test for year effects. Both of the year sets contained the following nesting guilds: deciduous tree, coniferous tree, building, ground, shrub, cavity, and ledge/cliff nesters. Canonical correlation finds linear combinations for each set (year in this case) of variables (nesting guilds) in terms of canonical variates such that the correlation between the two variates is maximized (CANCORR procedure, SAS Institute Inc. 1996, Tabachnick and Fidell 1996). The variables in one set, 1997-bird guilds, are linearly combined into single community attribute  23  or canonical variate and then these are compared with a twin community variate, summarizing the 1998-bird guilds. Thirty-six of the most common species were grouped by nesting guild and included in the analysis. Canonical variate scores produced by a preliminary canonical analysis were examined for normality, linearity, and homoscedasticity using scatter plots of pairs of variates to test for the assumption of multivariate normality. Evidence of failure of normality and homoscedasticity suggested that the data should be log (log, (X+l)) transformed 0  (Tabachnick and Fidell 1996, Sabin and Stafford 1990), and transformation improved the scatter plot test for data normality. The analysis was performed on species nesting guilds (Appendix III) since these guild groupings were used in later regression analyses.  Canonical Correspondence Analysis  To examine the prediction that bird species richness and relative abundance may be predicted by a gradient of urbanization, prediction 1,1 used canonical correspondence analysis (CCA) on the abundance distributions of the 36 species of birds. The method of direct gradient analysis can be used to relate species occurrences directly with landscape or local environmental variables of interest (ter Braak 1986). C C A is akin to direct gradient or regression analysis done in multivariate, species space. Resulting ordination diagrams provide a visual representation of complex relationships between community composition and habitat variables of interest. I examined prediction 2 with C C A to generate hypotheses for the relative importance of broad scale verses local scale habitat variables as predictors in linear models of the urban gradient system (ter Braak 1986, Jongman et al. 1995, Rottenborn 1999). With canonical correspondence analysis one can infer which variables, landscape and/or local, may best explain the variation in species distributions because the most important habitat variables load most highly on the first axis. Identified variables that were highly correlated with C C A axis one were further tested using logistic regression analyses to predict for nesting guilds and species. Axis one scores were  24  also used in simple linear regression as a predictor of species richness. Species richness for the simple regression analysis was calculated to include 'rare' species (singular sightings) that had been excluded from the C C A . I included these species in richness estimates to avoid underestimating the number of species with low detection probabilities. Land use categories such as urban and suburban are not expected to fully describe or simplify the urbanization gradient into habitat segments without overlapping features, prediction 3. I plotted these site classifications on the C C A urbanization gradient axes in a site ordination plot. The position of these average categories along the urbanization gradient was also quantified and depicted using the landscape and local level variables in radial star plots. Star plots use the average value of a variable in suburbs, for instance, and subtract that value from the mean for all stations sampled. Then, this value is standardized by dividing by the entire sample standard deviation for the variable and is compared as a relative radial length in relation to other variables on the star plot. For the C C A species ordination, the data consisted of the (Log (X+l)) number of birds 10  detected for each species recorded at point count stations and the 1997 and 1998 data were pooled by selecting the maximum abundance of the two years. Maximum relative abundance was used instead of an average to avoid the smoothing effect of an average that would result in less detectable bird-habitat trends. Maximum or peak counts are also not influenced by migrating birds because the counts were done during the breeding season. Since point count stations were surveyed only once per season, the number detected at a particular station does not represent an average for that year. Although the maximum value may be an optimistic estimate, this measure is likely to be a more accurate estimate of abundance at a particular site than the mean of one survey in each of two years (Vander Haegen et al. 2000). The program ADE-4, a multivariate analysis and graphical display software, was used to perform the C C A (Thioulouse  25  et al. 1997). A randomization test (Monte Carlo) was performed on the projected relationships to test the significance of the C C A ordination of species points; 1000 random permutations were performed on the distribution of species points along the axes using the Projectors: Subspace test (ADE-4 program, Fraile et al. 1993). C C A models assume that species have Gaussian (unimodal) response surfaces with respect to compound urbanization gradients that are constrained to be linear combinations of the two-scale habitat variables (ter Braak 1986), an assumption that has not been well tested empirically (Rotenberry and Wiens 1980).  Logistic Regression I chose a regression design for this study to identify urban bird community patterns in relation to the two scales of habitat investigation by direct gradient analysis. I used a stepwise selection procedure (p  forward  <0.20, p  backward  <0.25) to effectively screen a large number of  variables, and to simultaneously fit a number of logistic regression equations (Hosmer and Lemeshow 1989, Tabachnick and Fidell 1996). Following this stepwise procedure, variables deemed to be of importance either statistically (results from stepwise logistic procedure and CCA) or biologically were entered with sequential logistic regression in three different sets: 1) local level variables only OR landscape level variables only, 2) blocks of landscape variables were entered first followed by local level variables, and 3) interaction variables. Sequential addition of variables, landscape followed by local variable sets, answers the question, Do local habitat variables add to prediction of nesting guilds beyond that of landscape habitat variables alone, prediction 21 The third sequential block answers the question, does an interaction between habitat variables at the two scales add to prediction beyond that of landscape and local variables? The final logistic equation, the odds ratios, and log likelihood statistics at the end of sequential block entry are the same as would result from standard or direct logistic regression  26  with all variables entered at once. Thus, the final models are not sensitive to the order of variable entry and the contribution of each predictor can be evaluated as though it entered the equation last - over and above that of the other predictors (Tabachnick and Fidell 1996). Final model residuals were examined for deviations from the assumption of linearity in the logit. All regression analyses were done using SPSS (1996).  Species Incidence Functions To examine the expectation that species incidence increases in relation to the amount of park-area weighted by inverse distance prediction 4,1 plotted increasing park distance metrics verses the proportion of sites occupied by different species (incidence) with similar parkdistance values. The incidence of a species cannot be higher than one or 100% sites occupied so species incidence should asymptote to one. I used curve estimation regression methods to model these relationships, but for species with very low site incidence values that were not approaching the asymptote of 100% sites occupied, I used a linear regression fit. Bird species with adequate sample sizes (n>10 sites occupied) were selected on the basis of the C C A avian-habitat associations. The incidences of the three common species selected for logistic regression, American Robins, Spotted Towhees, and Song Sparrows, were also examined for the relationship between park-distance metrics and incidence. Grouping species into nesting guilds assumes that the nesting habitat of bird species is well known and that species with similar nesting habitat preferences will be found in similar locations. Nesting requirements are constrained for many species (e.g., cavity nesters, Mikusinski and Angelstam 1998), and ground nesters are absent from many urban bird study sites (Lancaster and Rees 1979, Soule et al. 1988, DeGraaf 1991, Rottenborn 1999). So, it is estimated that the best predictors of nesting guild presence in residential areas will be habitat characteristics related to nesting requirements. While grouping species into guilds loses some  27  information, an examination of individual species makes the assumption that species distributional patterns are completely independent from species to species. In a community of interacting species, this is clearly not the case. Logistic regression uses maximum likelihood methods for parameter estimation, a robust and flexible technique with no assumptions about the distributions of predictor variables (Tabachnick and Fidell 1996).  28  RESULTS  Bird Occurrence and Relative Abundance A total of 65 bird species were recorded at the 285 point count stations (1997-98, all seasons) on four roadside transects throughout Vancouver and Burnaby, including 62 native and three non-native species (Appendix III). The majority of individuals detected at all sites (-50%) were non-native species (average, 1997-1998), whereas several native species were detected only once (Table 1, Appendix III). The latter species were omitted from most analyses in order to avoid drawing conclusions based on transient or migrating birds and to increase the statistical power to detect bird-habitat trends. The most common species detected were European Starling (Sturnus vulgaris), Northwestern Crow (Corvus caurinus), House Sparrow (Passer domesticus), House Finch (Carpodacus mexicanus), American Robin (Turdus migratorius), and the Blackcapped Chickadee (Poecile atricapillus) (Table 1). Building nesters had the highest relative abundance per occupied site as they were detected up to three times more often than the other guilds. In sequence, the deciduous tree and coniferous tree nesters followed building nesters in the number of occupied sites and relative abundance per occupied site. A moderate number of ground and cavity nesters were detected per site occupied, and the ledge/cliff nesters were detected on the least number of sites, while shrub nesters had the lowest relative abundance per occupied site (Table 1).  Between Year Comparisons of Nesting Guilds The patterns of bird species diversity and relative abundance that occur in response to urbanization are not likely to change largely from year to year, at least over the two year period of this study, because the urban habitat gradient remains fixed and distinct (Germaine et al.  29  T A B L E 1. Detection frequency and nesting guild category for 36 species of breeding birds recorded at 285 point count stations in Greater Vancouver, BC - 1997-1998'  Guild  Code  Common name  Scientific name  # Sites detected  Relative abundance on occupied sites  3  DECIDUOUS TREE  Guild Total CONIFEROUS TREE  Guild Total BUILDING  American Robin Black-headed Grosbeak Bushtit Cedar Waxwing House Finch Red-eyed Vireo  Pheucticus melanocephalus Psaltriparus minimus Bombycilla cedrorum Carpodacus mexicanus Vireo olivaceus  NOCR PISI RCKI RUHU STJA  Northwestern Crow Pine Siskin Ruby-crowned Kinglet Rufous Hummingbird Stellar's Jay  Corvus caurinus Carduelis pinus Regulus calendula Selasphorus rufus Cyanocitta stelleri  BASW EUST HOSP RODO VGSW  Barn Swallow European Starling House Sparrow Rock Dove Violet-green Swallow  Hirundo rustica Sturnus vulgaris Passer domesticus Columba livia Tachycineta thalassina  BHCO CAGO DEJU FOSP OCWA SOSP SPTO WIWA  Brown-headed Cowbird Canada Goose Dark-eyed Junco Fox Sparrow Orange-crowned Warbler Song Sparrow Spotted Towhee Wilson's Warbler  Molothrus ater Branta canadensis Junco hyemalis Passerella species Vermivora celata Melospiza melodia Pipilo maculatus Wilsonia pusilla  AMGO COYE SWTH WCSP WIFL  American Goldfinch Common Yellowthroat Swainson's Thrush White-crowned Sparrow Willow Flycatcher  Carduelis tristis Geothlypis trichas Catharus ustalutus Zonotrichia leucophrys Empidonax trailii  BCCH NOFL RBNU WIWR  Black-capped Chickadee Northern Flicker Red-breasted Nuthatch Winter Wren  Poecile atricapillus Colaptes auratus Sitta canadensis Troglodytes troglodytes  GWGU MEGU RBGU  Glaucous-winged Gull Mew Gull Ring-billed Gull  Larus glaucescens Larus canus Larus delawarensis  4  4  Guild Total GROUND  Guild Total SHRUB  Guild Total CAVITY  Guild Total LEDGE/CLIFF  Guild Total  Turdus migratorius  AMRO BHGR BUSH CEWA HOFI REVI  4  186 2 57 22 190 2 255 219 19 8 12 12 240 31 223 205 76 105 257 38 4 32 3 15 38 87 5 124 25 3 23 50 6 87 160 3 4 13 166  1.06 0.50 1.66 0.57 1.34 0.75 2.75 1.35 0.68 0.75 0.63 0.63 1.89 1.35 2.76 3.09 1.83 1.42 6.95 0.82 0.88 0.64 0.50 0.60 1.08 1.02 0.50 2.12 0.62 0.50 0.80 0.77 0.50 1.45 1.22 0.50 0.63 0.65 1.83  54  2.31  Birds were assigned to nesting guilds according to Ehrlich et al., 1988, exceptions of VGSW and EUST, moved from the snag and coniferous tree nesting guilds respectively to building nesting guild since they more commonly nest in building cavities in urban areas. BHCO was moved to the ground nesting guild because it is a nest parasite and SOSP are the most common host species. Bolded species and guilds were used for logistic regression analysis. Average of both study years. Introduced species. 1  2  3  4  30  1998, Fernandez-Juricic 2000). To examine this assumption for birds over the two years of this study, I used correlation analysis. A simple correlation table comparing the relative abundance of all nesting guilds for 1997 and 1998 (log transformed to achieve multivariate linearity and normality) were examined both between and within years. Correlation between years was significant for six of the seven nesting guilds (Table 2). Within the same year, significant community associations were found between several guilds indicating that certain groups of species were associated with each other. The abundances of ground, shrub and cavity nesters were all positively correlated with each other, and deciduous tree and building nester abundance was also positively correlated. There were significant negative associations between building and three other nesting guilds (Table 2). Further analysis (CO) revealed that the first four pairs of linearly combined nesting guild variates accounted for significant relationships between years (Table 3, 4). There was consistent annual variation between bird nesting guilds, 1997 and 1998. The first test of significance (r 0.71, p<0.001) demonstrated that the associations between bird c  guilds were strong, year to year (Table 3, 4). Stations with a high abundance of building nesters (0.74) and low numbers of ground nesters (-0.81) in 1997 had high building nester abundance (0.80) and low ground nester abundance (-0.83) in 1998. The second test was for all pairs of variates with the first and most important linearly combined pair removed, and so on for the third and fourth tests. The correlation coefficients between variate pairs, two to four, were all significantly different from zero indicating further associations between years. The remaining three canonical correlation coefficients were not significantly different from zero (Table 3, p<0.05). With the exception of canonical variate pair three, there were no year effects either in the global data set or within various guilds. The relative abundance of similar nesting guilds in different years were significantly associated. The third variate pair combines high numbers of  31  05  C  ON  o  a  in  D  c/3  00  o  ON  >  ON ON  U  T—I  C  *  r-  o o  ON  CQ  ON ON  00  *#  <U  X!  N o C O cn d d d  -4—»  ON CD  rt  a  ** * * O O CN o o NO  c  S  ON  sg  OO T3 00 C a  c/3  7—  C 00  "9 u rt  D  •e « oo % c ri o <u  '§ "° £  o  T3  rt  ^ 13 ~ fa + o  >< 2 o 2  is .2 -e  •S ^ rt O O  o  CQ ON  H  u  <n «  ^  S  Js rt  to  ^  • ,o  <N W <+-<  ii  CQ a rt X  < ii  hi  o  di d d  o —i  —i  o d d  oo  > di d U  oo  o  o  o  o o  oo  * o  o  ON  o  >  U oo ON CQ  XT CN —< o  oo  o  o d  o  * —i  o  * — 1o — 1o Y—I < o di d  * cn * O od cn d ON  oo  O  * * * cn CN o o  #  —i  CN  *  ** CN cn CN O —< o  00  ** r--  CQ  o  o  o  o  co  r- o o o  NO  ON  00  t--  ON  *-i  H U  o  * #  ON  * * * *  ON  H Q  ON  ON  ^ oo -' 2? ^  00 00  ON  ON  H H O N O N pq > QUfflOwU  3 O u  00 xT  *CN CQ" ©  oo bo a  ON  3 X)  CQ"  o  cnm 00 hi >n o o .. o o ^ O d d di d d 8o TH  *  o  o  O CN . O ^ oo d oo -  •a c  T3 U  OO  o  *  H U  cnd d "S3 d i i oo  d d  d ^ 9 ^ d ^ 9 !2 do  o  * * o *CN cn  ON  ci ci ci ci d>  oo  o  ON  o d d  o d d  o  NO  o o cn d d d  —H  'S o o  O  *  * * co io oo O C N O o cn CN CN '—i  00  *  ci ci ci ci d d  * *  ON  a o  2 d d d x;  *  rt  ©  * *  O d d O • d d r-l  ON  >  rt o  NO  * CN 00 CO r-H C N  rON H Q  ON  o 3 S oo rt c  coo  i s VI  cn o o O d d I s  8  O T3 VI  r- r-~ O N °^ t-~ E-T O N pq > ON Q Q U CQ O K U J -IT  ON  ON  0O H H ON  J  ON  1  * *  ^  32  building, ground, and coniferous tree nesters with low numbers of shrub nesters in the 1997 variable set and correlates this variate with a combination of ledge/cliff, ground, and coniferous tree nesters in the second variable set, 1998 (Table 4). The two years of data were combined in further analyses because no strong year effects were detected.  Bird Community - Habitat Relationships Urban bird distributions were expected to follow an overall pattern organized predominantly along an urbanization gradient with decreasing species richness and increasing relative abundance {prediction 1). Canonical correspondence analysis was used to depict the main pattern in the relation between the urban bird species community and the observed habitat variables because this method is an ordination technique relating species occurrence and relative abundance to habitat variables (ter Braak 1986). Habitat variables measured or derived at the landscape level were expected to have more explanatory power because they summarize more information about broad scale habitat features. The local urban habitat, in comparison, was expected to have lower habitat value since local urban habitat should be more marginally valuable to native birds {prediction 2). None of the local habitat variables were highly correlated with the landscape level variables in simple correlation tables (not shown), and most significant Pearson correlation coefficients were between +/- 0.30 and 0.50, indicating no more than 25% of shared variation. Only four correlation coefficients were higher, local #House Lots correlated both with Park 0-260 m and Impervious surface cover within 500 m (  Pearson  . r = -0.54 and 0.72 s  respectively), also both local # deciduous trees (<30 cm dbh) and elevation correlated with deciduous tree cover within 500 m (  Pearson  . r =0.52 and 0.59 respectively). s  To examine the relative importance of habitat variables measured at different scales, the correlation coefficients of canonical correspondence analysis were calculated as estimates of the association between each habitat variable and the first and second correspondence axes (Table  33  o o o  O O O A O O O O O O  in  <M  T3  vo oo o cn i-H o  o  oo cn V") CN vo ON Q\ H 00 O cn oo r O O O O  H  CN  vo TJ-  00 VO I O  m (S ^ o\ Tf  ON CN cn ON H  \d m ri O  03 oo 03 r- Xi CN M 03 Xi 00  C 3 D  ca ca ON >  O  rM  O  CN  ^  CN  vo cn  ON CN  T3 C 3 T3 OJ Pi  ^H  rH O  O  O  O  c  U  O  O  O  ai -o  .2 fa -M  fe o > °  M  o  -M  o3 C  .2 x)  O  c  ai  00 00  03  aio o T3  cn CN  -H  cn CN cn CN i—H  T3  o  c  4—<  03  i—H  c d d d d d d d ai _g > 00-o eg « , *—H  *-H  *-H  »—H  .2 u  c fe  .2 § -M  O  o o o o o o d d O  cn  o  w  Oi  2 o 203 U_ in >n vo vo vo vo75 *_>• ffl o o o o o o o fe ai > o ~0 oo ci ci ci ci ci ci ci 8 « ^ M  OH  03  cn  *  03  o c  , u 03  oo  <  03  .S O  I ci ci ci ci ci ci ci  ca  'S  O  —<  c goo ai o •c ON ON M  ON  03 ai  ai  •c  00  T^- ,-H ON  d O  a ON CN  03  OH  H  2 '3 oo  3 o o o o o o o C 00 o o d o o o o E ON o o ON  T3  ai c ai o c  H  H  CM M  a  o  00 CN VO Tj"  C\ vo in vo H;  03 3  4—<  C U  W ^  M  «  CN £ <~ a o CN o o cn d d d d d d d  ^  ^ o o 03 U U OH  ni  03 cn  •  ca o  fe -G  3  3 al « ^  1-H  CN cn  vo  K  1)  s£  PH " O  *s  T3 C  rt bO  On  .5  r*. • * co cn CN 00 O d d  ON -Q ON x:  rt li c/5  o d  00  T3 > ai ai M  •° o ai « !-. rt U '3 < >, 60 O 00  (U 0  rt rt > _  -a cai  "a  oo rt  oo On o T—1 o d O d  vo O  d  rt Hi a  cn cN o CN vo vo vo CN cn cn d d d d d d i  CN  cn CN d"d d 00  •a % <a  § s X) ai 1,3  CN  cn o  o  o d  >  CN CN O O  X!  "O  C  I  VO o o d d d i VO 00 ON CN o d d d  on C  O  a O o o c C rt rt o  d  I  CN Ov cn O CN o d d d i i cn m cn d d  cn o d d  d  ON o d d  vo m VO d d d  CN  ai  w  i—l  o >/n CN d d o o CN r-d d  CN  o d d  aj o .fe «a o ai *s •*-» 60 _,: xl '3 ^ .fe ^  a  i  ^ rt  CQ OO  d  i  VO  rt li  c o rt  d  rt  rt c '5b 2 o 'C oo ai o o c ai  i—i -a r- rt  d  C  o M  e  cn  o\ o  c -3 «  .fe  cn t-o o d d d i i  On  3  2  d  o d d  t- CN d d  C  1  ^  in <o o VO d d i  CN O  c  X) -=> fcj M <U e xj O -J3  Td  I  TJ- cn o ON VD < CN -*t wn d d d d d  1  cn Si o o ai xi o  O  cn CN o d d  3  c ai .fe 0)  00 o o d d  Ci  S S1  &  1-H  CN  cn  VO  35  5). C C A can also be used to calculate the overall correlation between the species distributions of relative abundance and the first and second axes. These species - habitat correlations provide a measure of how well the variation in species scores can be explained by the combination of habitat variables in the axes. The first axis had a high overall species - habitat correlation (0.85), while the second axis correlation was not as strong (0.49) indicating that overall C C A axis one summarized the gradient of urbanization and the species distributions of relative abundance fairly well. In the following three paragraphs, first I examine the C C A model axes one and two, and then I describe general trends in the bird species distributions as they relate to these axes. High density sites, as indicated by the number of intersections (INTERSECTIONS) and number of house lots (# HOUSE LOTS) at the local level (< 1 ha) correlated positively with axis one, but with less relative importance than the two landscape park index variables (Park 260 m, Park 500-760 m, Figures 5, 6; Table 5). Moreover, the absence of water and number of small and large deciduous trees (DECID<15 cm and DECID>60cm dbh) correlated positively with this axis (Table 5, right side, Figures 5,6). The amount of downed wood (DOWN) and grass within 50 meters (LOCGRASS) correlated negatively with the first axis (left side, Figure 5, 6, Table 5). The relative strength of these associations indicate that axis one is an intensity of urbanization gradient characterized largely by landscape habitat variables, the park-distance indices, at the negative (left) extreme, and intersection size, landscape grass cover, large deciduous trees, absence of water, and housing density at the most urban, positive end of axis one. Axis two (vertical) was also partly dominated by landscape level variables, separating sites with high impervious surface cover (IMPERV 500m) from sites with high grass surface cover (GRASS 500m) in the surrounding area (Figure 5 and 6, Table 5). However, this axis explains little variation in the data set and due to the appearance of the 'arch effect' (see discussion) will  36  not be examined in detail (Eigenvalue = 0.09, Table 5). Eigenvalues associated with each axis give a relative indication of the ability for that axis to separate or order the species distributions, and can be used to indicate the importance of the variation explained by different sets of explanatory variables. The ordination diagrams provide a visual representation of complex relationships between community composition and habitat variables (ter Braak 1986, Jongman et al. 1995), and depict a joint plot of bird species scores (points) in relation to habitat variables (arrows, Figure 5). Both the length of the habitat arrow on ordination diagrams and the variable correlation coefficients (Table 5) provide an indication of variable relative importance in the model. Landscape level park-distance indices (Park 0-260 m and Park 500-760 m) load most highly on the most natural end of the gradient (axis one) and had the strongest, negative correlation with this axis; they dominated the left side of the ordination. Landscape variables were followed less strongly by local level variables, positively correlated with the most urban end of gradient, axis one (Water, # Intersections, # House Lots, and deciduous trees, <15 cm and > 60 cm dbh, Figure 5, Table 5). Both landscape (impervious surface and grass cover) and local habitat characteristics, such as dead trees, shrubs, and water dominate the vertical axis two. Axis one was four times better at explaining the variation in the data set than axis two (Figure 5). The ordination was significant (pO.OOl, Projectors - subspace (randomization) test, ADE-4, 1997) indicating that the ordination provides a good approximation of relationships between the observed data on bird species distributions and the two scales of habitat variables. A total of 25 of the 36 species were located on the natural end of the urbanization gradient indicating they are associated with landscape scale habitat (large parks) and supporting the prediction that landscape variables better explain urban bird species distribution patterns (Figure 5, Table 5, prediction 2). Two landscape park-distance variables were identified for further hypothesis testing.  37  C C A axis one scores for each station were used to predict species richness using simple linear regression and curvilinear model fit confirming a significant negative relationship (Figure 7, Table 6, F=71.86, p<0.001, pre diction 1). Species richness decreased with increasing urbanization along a gradient, however, the amount of variation around the predicted linear relationship was high as evinced by the low r-square (r = 0.23) and large, 95% confidence 2  limits. The relationship shows some evidence of non-linearity, but improvement in fit with quadratic and cubic terms was not substantially better (r  2 linear  = 0.20, Table 6), considering the  drawback of having additional parameters in the model (Table 6, i^and SSQ, Hilborn 1997). Overall, the bird community relationships depicted in these results showed generalist and specialist occurrences along an intensity of urbanization gradient {prediction 1). A l l ground (5), shrub (5), and cavity (2) nesting species were found on the left side of the ordination with the exception of Canada Goose (Branta canadensis) and Fox Sparrow {Passerella iliaca, Figure 5, Table 1). In general, habitat specialists were identified as species associated with very high values of Park 500-760 m, e.g., Winter Wren (Troglodytes troglodytes), Song Sparrow (Melospiza melodid), Common Yellowthroat (Geothlypis trichas), and Cedar Waxwing (Bombycilla cedrorum), left side of Figure 5. These species are associated with more (> 100 ha) proximal park area as indicated by high park 500-760 m values (Figure 5, Table 5). There were fewer urban associated species at the extreme right end of this urbanization gradient (Figure 5), and as expected these were also the species with the highest relative abundance per site on average (Table 1, prediction 1). They were identified as species associated with higher densities of development, small deciduous trees, less understory vegetation, and more impervious surface cover, e.g. House Finch (Carpodacus mexicanus), European Starling (Sturnus vulgaris), House Sparrow (Passer domesticus), and Rock Dove (Columba livid).  38  >-> o x i —<  -a <> i  o Q •s  c 3/ Q Z <  • £ SP  o LL  c  < u  bO ^ cu C  ^1  +  Si O  crto  O C OC a  1  e o  -a -SS  18  • •  •  •sit:  O  •  °  <D X!  • • •  rt  V g  e ^  •3  • • • •  O  a rt O  C O  xl ^»  e  oo ja  11 a Vegetat Axis 2  JO  o  • G  o « 0  a •  o  m  M  y «i <u "to o «S *G fi < 2*> o rt rt- <ou DO •rt _ 3 rt C fi3 O o oo T3  g  B  SS  2  ^ c c3 o  rt  9 .2  <U So s O CD to  •rt ° °  c o 2° c o D Hfi"fj O0 -rt 'rt <u o c  c  a  OH  «  -fi rt x fe .O O rt <+-< fi CO O 42 C O fi d  H->  60  C  S>  U c .SP •  .5 £ ND  rt o X! 04 r  H^  40 T A B L E 5. Correlation coefficients between habitat variables at local and landscape levels with the first and second canonical correspondence axes in an urban bird biodiversity study in Vancouver - Burnaby, British Columbia.  Eigenvalues  A) Landscape Level Environmental Variables 2  Park 260m Park 500-760m G R A S S 500m I M P E R V 500m SALT-H20 500m CONIF 500m  1 2 0.35 0.09 Variable Correlation Coefficients -0.76 0.15 Park-distance <260m -0.45 -0.33 Park-distance between 500-760m 0.13 -0.41 % cover grasses <500m 0.08 0.51 % cover impervious <500m 0.05 -0.01 % cover salt water <500m 0.04 -0.01 % cover coniferous trees < 500m  c  d  DECID500m  % cover deciduous trees <500m  -0.04  -0.02  F R E S H H 0 500m (+/-)  Presence of fresh water <500m  0.00  0.02  W A T E R (1-5) LOCGRASS DOWNwoocf D E A D (trees) # HOUSE LOTS SPR (trees)  Presence, water (bath, fresh, marine) % cover of grass at site level Amount of downed wood # of standing dead trees # of house lots Species richness of trees  0.21 -0.16 -0.15 0.05 0.14 0.08  0.16 0.09 -0.04  Feeders (+/-)  Presence of bird feeders  0.01  0.16  BOX(+/-)  Presence of nest boxes  0.00  0.11  d  2  B) Local Level Environmental Variables b  c  D E C I D 1 (<15 dbh)  -0.01  # of medium sized deciduous trees  0.01  -0.01  # of large deciduous trees  0.14  -0.15  # of sapling coniferous trees  0.00  0.10  # of small coniferous trees  0.03  0.11  -0.01  0.07  c  # of small deciduous trees  DECID3 (30-60 dbh)  c  C O M F 1 (<15 dbh)  c  c  CONIF2 (15-30 dbh)  c  CONIF3 (30-60 dbh)  c  C O N I F 4 (>60 dbh) DI S H R U B (<lm)  c  c  # of medium sized coniferous trees  0.01  -0.16  # of ground level deciduous shrubs  -0.09  -0.11  # of large coniferous trees # of small deciduous shrubs  -0.05  0.14  c  # of large deciduous shrubs  -0.09  -0.15  c  # of ground level coniferous shrubs  D2 S H R U B (l-2m)  c  D 3 S H R U B (>2m)  -0.05  0.16  # of small coniferous shrubs  0.00  -0.09  C3SHRUB (>2m)  c  # of large coniferous shrubs  0.01  -0.10  B l S H R U B (<lm)  c  # ground level berry shrubs  0.00  -0.10  # of small berry shrubs  0.04  -0.11  # of large berry shrubs  0.00  0.00  # of sapling berry trees  C 1 S H R U B (<lm)  C2SHRUB (l-2m)  B2SHRUB (l-2m)  c  c  B3SHRUB (>2m)  c  -0.06  0.01  # of small berry trees  0.00  0.08  B E R R Y 3 (>30 dbh) (1-3)  # medium sized berry trees  0.01  0.03  I N T E R S E C T I O N (1-3)  1,2, busy street intersection  0.20  0.01  -0.05  0.09  B E R R Y 1 (<15 dbh)  c  B E R R Y 2 (15-30 dbh) (1-5)  ELEVATION  Elevation (m)  0  T h e proportion of impervious surface, grass, salt water, deciduous and coniferous tree cover <a 500 m buffer radii, the presence/absence of fresh water, and 2 composite park variables as an index of park importance <260 m and 760 m radii of each point count station. Logl0(X+l) transformed, Logit transformed [p/(l-p)], L o g l 0 transformed. Variables were measured around each point count station <a 50m radius 0  b  -0.06  DECID2 (15-30 dbh) D E C I D 4 (>60 dbh)  a  0.15 -0.02  # of sapling deciduous trees  c  -0.23 -0.07 0.15  d  e  41  Quadratic fit Y=-1.21CCA1 - 0.30(CCA1) + 7.42 2  e o  •c OO  SH  •  'o  •  •  4»» •  <D 00 « Species richness - Quadratic curve fit - Upper 95% CI - Lower 95% CI  -4.00  -3.00  -2.00  -1.00  .00  1.00  2.00  C C A axis one urbanization gradient site scores  Figure 7. Cuvilinear regression model fit with 95% confidence limits, C C A axis one versus maximum avian species richness at point count stations along four road-side transects in Vancouver and Burnaby.  42 T A B L E 6. Model selection: fit of species richness at road survey sites with C C A axis one (urbanization gradient) Vancouver - Burnaby, British Columbia.  MODEL CCA1 vs. Richness Linear Fit  Variable (m) Parameter Standard estimate (B) error of the estimate (B)  SSQadjusted T  Sig  n  C C A Axisl  -0.95  0.11  -8.48  0.00  283  Constant  7.11  0.11  62.34  0.00  283  C C A Axisl  -1.21  0.14  -8.87  0.00  283  (CCA A x i s l )  2  -0.30  0.09  -3.20  0.00  283  Constant  7.42  0.15  49.79  0.00  283  C C A Axisl  -1.51  0.20  -7.73  0.00  283  Fit  (CCA A x i s l )  2  -0.03  0.16  -0.20  0.84  283  F=42.24***  (CCA A x i s l )  3  0.17  0.08  2.14  0.03  283  Constant  7.29  0.16  45.51  0.00  283  F=71.86*** Quadratic Fit F=30.05*** Cubic  R  2  SSQ  SSQ(m)/ n-2m  0.20  265.31  0.94  0.23  301.97  1.08  0.24  318.18  1.15  43  To assess the position of my habitat category sites, e.g., 'urban', 'urban encroachment', and 'suburban', in relation to these C C A model axes of the habitat gradient, the canonical correspondence score of each survey station was plotted along each axis (Figure 6). This figure represents a joint plot ordination of the 285 site score points, with arrows identical to those shown in Figure 5. Each station in this diagram is plotted at the centroid of the species points that were found there, and hence the location of each point indicates which species are likely to be found at a particular site (Figure 6, ter Braak 1986). Urban sites occur predominantly on the lower right hand side of the plot, whereas park, edge, and urban encroachment sites occur largely on the left side of the ordination indicating that some of these categories are useful in simplifying the urbanization gradient, but there is overlap (prediction 3). Two thirds of the large park sites occur on the lower left extreme side of the plot, whereas smaller urban parks and edge sites cluster in the mid to upper left half of the plot. This indicates that Park 500-760 m mainly represents the effects of the three large parks in the study area (Burnaby mountain, Pacific Spirit, and Stanley, > 324 ha in size) and Park 260m represents a combination of effects from edges, small (>5 ha) and large parks. So, birds that occurred predominantly along roads bisecting any of the large parks were located at the negative lower end of the C C A plot. Urban encroachment sites scored predominantly in the upper left half of the plot. Suburban stations have a fairly even distribution over the centre of the ordination, indicating that suburban sites constitute the most variability in habitat measures at both scales. This finding confirms that subjectively defined land use categories do not fully simplify the urbanization gradient (prediction 3), that is, they are dispersed across a large range of conditions with respect to urban gradient, C C A axis one. Interestingly, suburban stations do cluster more on the positive, upper end of axis two. If this axis can be viewed as a local vegetation or habitat resource gradient, then this indicates that the  44  best suburban stations had small deciduous and coniferous shrubs (1 to 2 m's), few house lots, and the presence of water, increasing the number of species at suburban sites.  Land Use Habitat Positions along the Urbanization Gradient To further quantify the position of an average 'urban' and 'suburban' station along the gradient of urbanization, landscape and local level habitat variables have been plotted in relation to each other in star plots for each of the six land use habitat categories (Figure 8). Star plots can be interpreted for each variable such that the magnitude of a star's spike represents a standardized value for a variable, plotted along radial axes in relation to other variables. Habitat types with a larger relative value for a variable will show a greater star spike than corresponding spikes on star plots for other habitat types (Cleveland 1993). Landscape level variables, derived to estimate the importance of parks and reserves close to each survey station, and estimates of landscape composition (i.e., percent cover of trees, impervious surface, water, and grass <100 ha), are shown in Figure 8A. Data on local vegetation, composition and structure, around each point count station (Figure 8B), and non-continuous variables (Figure 8C) are presented and the plots are described and compared in the following three sections.  A) Characteristics of Habitat Categories - Landscape Level Features Urban and suburban areas were defined by their political boundaries and by population density measures in this study (4940 individuals/km and 2150 individuals/km respectively). 2  2  The average amount of impervious surface cover (landscape, <100 ha) increased with increasing urbanization from parks, edge, urban encroachments, suburban, to urban. However, urban park areas had lower levels of impervious surface cover at the landscape level than suburban areas. The number of house units or lots at the local scale was NIL in parks, low at park edges, and  45 C  "o  OH  Az aa 2  —  at <  w  D  H  oo 03  o oQ t X" O T3  « u m u.  <D Vi O O (D UH  oo"  (N —  o A -V- c.  CQ CQ CO CQ D D P D OS OS Di OS X 35 35 35 « mw m (N ro —- (N CJ CJ CQ CQ  in  03  03 C  O C  —  ' (N ~  V C- A V  l i p  U T3  a  od  o3 00  ON  ON r — l  o oo no o 1/3  <u ^, 03 03 <u  +—< O  — <N ro —  Q Q Q CJ  o -a — •7 o A —• nA <3u oo Ol >- >- X Z as o. Q > OS os < ^ U M B O CO CO Q Q m in  03 o3  -s  fe > 03 »3  •R -2 §°  2 c  X! 03  13 > >  — rt  8 " ° ? v in m ^ r w  " Ml. J ft. S S 0-  sOOn g CJ CJ CO  *V O m in o — — ro A V CN ro ^ — < Q D Q ft.  uuuz uuuo Q Q Q V  I I I I in *o r~ oo  UH  o  pg -S  sa  3 3  03 o3 X! O <U 1 ) CH OH o3 03 O O oo oo -0 C —H  T3 C  JS  < X) ^  OH  oo  o  3  2  03 <£  oo u o oo *o S E-i O  00 —  OS - O W C~uDw O S J Q O U I OI — (N ro TJ-  00 o  Oi  u  1/1  t; 60 53 UH  46  highest in urban areas and in urban 'encroachment' areas surrounded by parks {italics, Figure 8A). Park sites were often coastal, marine sites; they had less grass cover than any other category (Figure 8A). Edge and park sites were quantified by high park index values (Park 260m and Park 500-760m). Suburban and urban encroachment sites also had high values for park index 500-760m but low values for the park 260m index (Figure 8A). Suburban areas were largely developed in their immediate surroundings, whereas at a larger spatial scale these areas typically had higher values of park index 500-760m than urban areas. The differences between urban encroachment and suburban areas lie mainly with the amounts of grass cover and fresh water, both of which were higher in suburban areas; encroachment sites also had slightly greater marine, coniferous and deciduous tree cover, and Park 500-760 m values (Figure 8A). Conversely, urban park sites had high values for park index 260 m, but had low values for the Park 500-760 m index, indicating that urban park areas were surrounded at larger spatial scales by other land uses and further development (Figure 8A).  B) Characteristics of Habitat Categories - Local Level Habitat Local level variable star plots summarize the microhabitat variability within habitat categories (Figure 8B). Suburban areas tended to have the most small deciduous shrubs <1 m to 2 m, similar numbers of coniferous shrubs as urban areas, and a majority of berry producing shrubs >2 m. They also tended to have higher tree species richness, higher numbers of coniferous trees 30 to 60 cm dbh, and lower housing densities (# HOUSE LOTS) than any other areas (Figure 8B). In comparison, urban encroachment sites were characterized by high elevations, large coniferous trees, smaller shrubs, moderate amounts of dead and downed wood, and higher housing densities at the local level (<lha, Figure 8B). Park areas had typically high levels of dead and downed wood, berry producing shrubs <1 to 2 m, and large numbers of  47  deciduous trees > 30 to > 60 cm dbh and comparable amounts of coniferous trees to edge areas. Edge sites were similar to park sites but differed in the amount of dead and downed wood, they also tend to have larger berry shrubs (>2m), and the presence of nest boxes (Figure 8B). Local urban area star plots had diametrically opposite habitat variable high points in comparison with park areas, where parks had habitat maxima, urban areas had habitat minima. Urban areas in this study had high housing densities, high tree species richness and often had bird feeders present. Moreover, they typically had berry producing trees and fewer deciduous and coniferous trees (Figure 8B). Shrub values for the urban stations were comparable to suburban ones, although urban areas tended to have more coniferous than deciduous shrubs (Figure 8B).  C) Characteristics of Habitat Categories - Nominal variables The nominal variables are potentially important local habitat variables indicating crucial water resources, the level of disturbance by road intersection size, and aspect; they were treated separately in order to examine them with ease and to reduce the number of variables presented on the local star plots. The most noteworthy differences among habitat categories in relation to aspect, intersection size, and the presence of water at the local scale (<lha) had to do with the presence of fresh and salt water on park, edge, and suburban sites. Water was typically absent in urban encroachments, urban parks and urban areas. Suburban, encroachment, and urban park areas had more large intersections (Large X , Figure 8C), whereas all of the habitat categories were roughly comparable in terms of the average number of small intersections (Sm X, Figure 8C).  48  Bird Nesting Guild-Habitat Relationships Known nesting guild groupings were confirmed by the species distribution patterns evident in the C C A ordination (Figure 5). Given that ground, shrub and cavity nester species were found on the left side of the ordination, nesting habitat requirements broadly separated the observed distributions of birds. I used groups of nesting guilds to further test the prediction that landscape level habitat features (measured at < 100 ha) should be stronger predictors of nesting guild occurrence than local habitat measures (<1 ha, prediction 2). I used logistic regression techniques to find the best fitting models that describe bird occurrence using habitat variables measured at landscape and local scales to examine the specific associations between nesting guilds and vegetation structure and composition at both the regional and stand levels. The predictive value and significance of sequential logistic models for seven nesting guilds and three individual species (see below) were compared following entrance of blocks of variables: Blockl) local variables only or landscape variables only, block 2) local variables were added to landscape variable only models from block 1), and block 3) landscape*local level variables were added. Of the final nesting guild and bird species models presented, five of seven guilds and all three selected species had good model fit at each block (Hosmer-Lemeshow goodness of fit test, p>0.05, Table 7) - indicating that model predictions were not significantly different from the observed data. The exceptions were the local variable only models for building nesters and Spotted Towhee, and the landscape variable only model for American Robins and shrub nesters (Table 7, X goodness of fit, p<0.05). The likelihood ratio test statistic is another test for model 2  goodness of fit comparing the difference in log likelihoods between a model with and without predictive variables to determine if the model fit was significantly improved. A l l final models in each block (with local and or landscape variables) were better at describing the observed data than null or constant only models (Table 7, X  2 t o r e m o v e  p<0.01).  49  For guilds that were present at most point count survey stations, such as coniferous and deciduous tree nesters, the models failed to converge and maximum likelihood logistic models could not be obtained. The model for building nesters had low model specificity, or ability to predict absence, probably because there were very few cases where building nesters were absent. Common individual bird species were selected, for which data were sufficient to build robust models. The American robin was selected to represent the deciduous tree nester guild. All of the species in the coniferous tree nester guild had either insufficient number of cases or they were present at almost all sites. I chose to model the occurrence of the two most common ground nesters, Song sparrow and Spotted Towhee, since ground nesters were estimated to have the most specialist habitat requirements. Variables chosen for entry in each model of the sequential regressions were selected based on canonical correspondence analysis results, statistical significance (p<0.20 in forward stepwise logistic regression), and known habitat associations and requirements. In seven cases out of eight (including 5 nesting guilds, 3 species), the first variable to enter stepwise screening models was a landscape level predictor. In six of those cases this variable was one of either impervious cover < 500 m, or a park-distance variable (Imperv 500 m, Park 260m, Park 500-760m, Table 7). The ledge/cliff nesters were the exception; the number of house lots entered as the first variable for this nesting guild. When only local level variables were entered into a model, the overall rate of correct classification for building, ground, cavity, and ledge nester guilds was lower than when only landscape variables were modeled (Table 7). This indicates that broad scale measured variables were more specific and/or sensitive predictors of the presence of these four guilds. Conversely, the local variable models for shrub nesters provided more accurate percent correct predictions than the landscape variable only models for these guilds (Table 7). When the five nesting guild models were compared to constant only models, improvement in log likelihood was better for  o  s  oo >/n oo oo oo  CN ON OO  vd  ca O N oo vd  ' vo w->  ON  IO  OO  ON  ON  1>  <>  rt  VO O0  00  >o r> >o r~  ON  OH  o is wi  <  vd vo  »rH  T3 O O  O  O O  1  OO  00  oo  vo O r~  Ti-  VO  ON  o o  ©  CN cn o  o o 00  ©  CN  oo  rt rt  OO  CN  CM  1-;  ON ON  cn cN  Tf  VO  CN  l>  Tf  CN  r>  VO  X ai  Ti# VO  rt  rt Tt  i•-—H1 0V 0O  s  1  Tt  H rt  Pi  * *V*D **  **  NO  ON  00 OO  H  ** o\  CN I>  05  Oi  d  s  * *#  #  VO OO  ** OS OO  rH  00  Oi  d  rH  rH o> oq rH  o  oo  o  o o o  o ©  CO O  u  ON ON -H  C£> 00  o> d  rtd  s  rt rt cn cs •* «n  *r- *rt  * *OO  I>  p CD  rH  N< I>  +  LO  < u 00 p  W O  Oi  r-i  d  LO  CO  CN  d  T< i in CO CO  p d  p  r-i  LO •*  d  d  in o io N< d  r-J  o p d  vd  LO  VO  U  o 05 O  Tt <N rt  00 CN  o  00  OH  ortrto  1  vo Ov  CN CD 00  -  # VO  cn r~  ON  rtd  p CO  **  CO  Oi  -J CO  p  u o  CO ( N CT> CO  Ti i rH  oi 00  CN  OJ  w OH <; u  CO  CD  CD I>  rH rH  r-i  rt)  CO 00  o d o d  CO  vj wl  #  o rH  VO  o  Tj-  1-1  #  i"—i  O O O O rt rt U-) ^  o  o\ CN oo vo O <N cn vd ON  CN  rt CN O rt rt  -o -a  un  1  ON  O  00  </-> >o cn r>  VO VO  o u  cn  VO  #  ON  <N cn  CN T3 TO O  O  oo  O  * * # * *# *# o O* <N  T3  CA  ON  c/3  .SP to  2  o  >/n  ci oo oo  o  Tl-  io rt CN rH  o d  rH LO rH  CO  Oi  d  LO LO  Oi  d p oi d  00 LO  d  VO  ON  CQ  ?  •1" ° o  ca  W OH  vN  &H  °H  5j  v  I I  A I  P"  H  N<l  Ol  ?  ?  03  U  CH  < o  33 B CQ  -  >  s~  ^ £  w  w ^  c-  "s e  53  c»  S J-l  2 £ tiI NSO c o o uOH O CH  vo  Q O  cn Till  o  6J3  e  o  o g  q  ^  WD  c4 cN  cn  C L ^ U O H O H O H O H C Z I I Z I (0 CO C  p  o  S S O ^ c n ^ & ^ c n c n  Sg£rti<t<:<;<tcQQ  B  i2  a o U  S o  cN d oo r~  CN d  ON  a  Tj-  NO  NO  NO  o  00  rt  p  >o cN >o CN  r-  ON  t>  NO  NO NO  CN  NO  00  i  o U  fl  o ON  oo  <0  NO  oo  cN cN cn t* *  00  oo d  ON  d  T 3  O O  O <rt  O O  CN  OO  <ri  ON  T3  o  ** * * — f1  * * ** **  CN* cn  ** ** ** * * oo  ** *#  #  NO  NO  d  ON  o  NO  NO  <o o  >o  ON  ON  V-)  P~  "O NO  r~  NO NO  ON  ON  ON  IT) ON  p oo NO  NO NO  NO  o  ON  NO  %  ON  rt  CN CN © IO  rt U  <  cO nOO NO rN tc iNo- c^ nl -c ono ci ON No Oo cn~rtrjcn cN >o  o  00  p  rtd-^cNcNddd  O &  o d  I  ON  d  i  d  cn oo o  r~  CN >r> ON  NO  d od  Ui  u o rtl pa  —  ON  w  #  #  cn 00rt rt cn •* CN CN CN00rt CNrt cN CN CN CNrt cn cn cn cn Ui <=>. u ~ o rtONOTtTtOOO opa 00OCNrtcNrt(~~ s OrtioNoooiocNcnr~ o ..;©--<-;—<©©©  ON  NO  «§  00 rt CN* * *  •*  p ON  * *# * ** ** * ** o rt  T3  CN d  ON  u  3  < PH  CN NO  U  00  cn cN r>  o NO ^ "* ON rt \n d rt © d  NO  NO 0O  ©  ©  ©  ON  ON  rt  ©  ^ to pa P* <o C T3 O  o  a P O  u na 35 u 00  Ml  2  Q O S  (H >, W  5  g  CO N-  u 8z>  oi  II  vo  315  s  00  vn  tvo.  C u 43  <  -o o o O  cn oo oo r~  oo  c- oo  vo >o  CN CON  CN  ON  -H  .SP  O  00  ©N rt  O  ON  vo oo  VO  00  00  O O O  CN d  CN d  _l  cn  VO  cn  —i o  CN  _l  CN cn — iH d  d  "H  .13**  # CN cN >o oo cn —* °) cn oi  X  rt vo  T3 T3 O  u  «  •a ca O Di <0  oOo  -*  vo O ©  ON  ON  oz uo  s s  vq  CO r~CN CN O  oo oo  O  ** *  * * vo CN vd cn  vo Ov CN O ON o — d —!  rt  CN rt OO ON O cn  u<  l> VO o >/->  ** ** ON * cn * ON  O -* o o  U  o\ vo •* cn ON VO ON o O © © CN p  vo VO p  CN Ov  * * ** #* ** * OSl cn •«* oo CN -*  oo oo t> ON cn p VO >o as oo VO oo CN O d d CN CN  ©  O  * # oo oo vd  OO  CN d d ON  ON >o c- o cn CN CN -4 >-i © i-i  cn vo cn O ON Ov d d  oo P  vo t- o O ON  ON  m cn c- <o NO o  cn vn r-  o o o  ©  O O  cN rt rt  -d o  T3 O O  o H3 "33  •c o  ~3  60  Ml  u  a s s  *  to cn  o 00  vo II c  c  S CN vS II  1  C3  a  a  00  PQ  so o  T3  C3  O =O3 xn vo C3 00 A  •  Q si O Q O  Ha  B  S  f i w VI W cn o Q CQ U  cII  a  o OH  Tow hee presen  ^1  * <N fvd i>  ec  •o o o  .57**  oo CN  t/u  X)  O ON  Q ui O Q S O  cn IT)  oo cn ON  OO  rH ON  00 oi  o  oi ON  •* vd f-  VD ON  Tf  VO VO  .1 TJ  a so  a a. <j co o U  B  OO  cn  o  o  S oo U  od in vo  q  o\  d  ON  oi oo  VO  oo  vo 00 vo p~ i>N O O N ON  vo vd C\  oo cn  OO  00 CN  oo oo cn O cn O d d  oo  O vo r-_ rtN cn O CN  vo oq  VO  VO  o  VO  00  t-  .SP  13 O O  d  OO  un d  vo C N d d  C O <rt 00  _,  c- vo o- oo cn vd vd d  Vi M M O a o  rt  vo  C O  •*  T>  1  O  o  > -a.  t~ o oo OooN cn •* d 00 r- Ol r> oo f-  2 E ^ a co ^3 ca X Oi J Di  <N  Ol  ON  cn vo CN o vd oi  VO  oo cn vo cn d  d  o  -H  vo  vo o  T3 TJ  ON ON  d  vo  0\ -j vq cn vd H  ©  ON  d  HH  ON  o  Ol olvoooolcnvOTt-—< — < oq ON. © vqvq -H vq "tf  U  o  d  d  —  •  od  H  d  d  H  ON vo  TJ  ^  •oO  d  -a  C O  vo V O d  oi cn cn Cvo >o cn vo ON o vo oq cn d d - i cn vo d - H oo  *  *  *  Tf  VO  Ol  cn  s U  rJ  ! #  #  #  vo C\ O cn cn rt O O N ol O N O N U —< d vo ON o J CQ  oa  ol O vo oo cn O 0\ \t m (S rt o d d —i  o  cn  Ol Ol  v ocvno O N cc nnoOo rO- cOn -oHo [r- -~ O -H  *  -HcnONcncnONVOONOI vOTt-oovovocnolOvo VO O N O — t © cn oo r»  < rt  o rt  cn o o  o\ ©  vo o d d  cn vo i n oo rt  vo ON  d  ol vo vq Tf d  «.sto _09 S  c  ^  o- co  •5 %  O  l 3  •s VOO E P Q  £ Q  •S  u  &o CQ  VD  C O S a Vi  >H  cd  (V  •C  C  ^  o  Ol  x l  vo ^  rt (]j II  £  < (rt 3 -S  ^rt  Q S  2 S Q P  54  landscape variable only models than for local variable only models (X  e> Table 7) with the  2  exception of shrub nesters but all were significant (Model vs Model fit  10  rem0V  constant  , p<0.001 Table 7).  This indicates that landscape variable only models improved fit more than constant only models with fewer variables (dfs, Table 7) than local variable only models for four of five nesting guilds tested. The comparison for species is somewhat different. Spotted Towhee, Song Sparrow and American Robin models had better overall rates of correct predictions with local variable only models than landscape variable only models - albeit the local model for Towhees failed the Hosmer-Lemeshow chi-square goodness of fit test (Table 7). When these three species models were compared to constant only models, improvement in log likelihood was again better for landscape variable only models than for local variable only models (X  2 10  r e m o v e  , Table 7), but the  difference was slight. In the case of the Song Sparrow, the difference between chi square loglikelihood to remove for local variables only verses landscape variables only was next to nil in the reverse direction. This indicates that when modeling for individual species as opposed to groups of species, local variables are more precise in terms of correct predictions. Birds should select local sites that provide adequate nesting habitat conditions during the breeding season and thus their occurrence at point count stations should be related to the local habitat conditions around each station. Following landscape level predictors, microhabitat or local level descriptors significantly improved the predictive ability of the landscape-variable only models, but the overall rate of correct classification increased only minimally (Table 7, Block 2 vs. Block 1). The addition of local variables to the landscape only models for building and cavity nesters was barely significant (Table 7, X  2 10  r e m o v e  p~0.10). Point count stations in  areas with both high quality local habitat and a large amount of surrounding regional habitat should have higher than expected guild and species occurrence (presence). A significant  55  interaction term between local and landscape scales was identified only for Song Sparrows (X  2  remove"  t0  6.23, p<0.05, df=8, N=265). The overall rate of correct classification for the sparrow  model with significant interaction term improved only marginally (Table 7). Six of the eight models included the landscape variable Park 500-760m, implying that distance and the amount of park area in the neighbouring vicinity is a strong predictor of guild and species presence. The odds ratio represents the increase (or decrease if the ratio is less than one) in the odds of finding a guild or species present for every one unit increase in the value of a variable, after adjusting for all other predictors in a model (Tabachnick and Fidell 1996). It is a multiplicative factor, 1.05 for instance represents a 5% increase in odds per unit of the associated predictor value. To make this value more meaningful, the odds ratio is often calculated for greater than one unit increases in predictor values. So, for ay?ve-unit (m) increase in the value of Park 500-760 m, the likelihood of ground, shrub, and cavity nester occurrence increased by 25%, 15%>, and 15% respectively (Table 7, Odds ratio  Park 5 0 0  .  7 6 0 m  = 1.05, 1.03, and 1.03, p.00Kp<0.05).  Only the ledge and building nester models did not include this variable as a significant predictor. This finding is consistent with results from canonical correspondence ordinations since all shrub, ground, and cavity nesters were associated with these two habitat variables on the left end of the ordination (Figure 5), while building and ledge nesters were not. Another consistently important landscape predictor was the percent cover of impervious surface (IMPERV 500m). It was included in five of the eight models (X  2 t o r e m o v e  , 0.001<p<0.05) supporting the use of this variable  as an average metric to characterize degree of urbanization (Arnold and Gibbons 1996 and Marzluff et al. 1998). American robin was the only species and shrub and cavity nesters were the only nesting guilds that did not include impervious surface as a significant predictor (Table 7).  56  In general, the best local habitat predictors of nesting guild presence were both habitat structure and composition variables and variables indicative of urban density. The likelihood of finding ground and shrub nesters significantly decreased with increasing housing densities and increasing cover of local grass (LOCGRASS and # HOUSE LOTS). There was an 10% decrease in the likelihood of shrub nester occurrence with each additional housing lot (<lha) after adjusting for other important local predictors such as numbers of deciduous and coniferous shrubs and water (Table 7, odds ratio house lots = 0.90, X  2 t 0  remove,  pO.OOl). The likelihood of  finding ground nesters decreased by 50% for every 10% increase in local grass cover (Table 7, odds ratio i  0 C  ai grass  =  0.95 a 5% decrease for every 1% unit change in local grass cover, X  2  remove,  10  p<0.001). The presence of gulls and robins were also found to decrease in likelihood with increases in these variables. Water was included in all of the models, with the exception of cavity nesters, and the likelihood of finding shrub nesters was four times higher when bird baths were present (Table 7, odds ratio ater W  bath  = 4.19, X  to remove P<0.05).  The likelihood of finding ledge nesters was seven  and one half times higher when salt water was present < 500 m (Table 7, odds ratiosait = 7.46, X to remove p<0.01).  2  Song Sparrows were almost six times more likely to occur at sites with fresh  water present (Table 7, odds ratio eam Block 2) = 5.85, X str  2 t 0  remove,  p<0.01); there were large  standard errors in the stream parameter estimate for Block 3), which included an interaction term coniferoussooX fresh water. This indicates that the interaction term and fresh water have overlapping conditions and since both are good predictors, the iterative process bounces back and forth estimating water and interaction parameters, resulting in large standard error terms (Hosmer and Lemeshow 1989, Tabachnick and Fidell 1996). Local structural variables such as large deciduous trees (DECID4 (>60dbh)) increased the likelihood of American Robin occurrence (Table 7, odds ratio = 4.95, X  2  to remove p<0.05),  but  57  decreased the likelihood of Spotted Towhee occurrence (Table 7, odds ratio = 0.07, X  2 t o r e m o v e  ,  pO.Ol), indicating that these two species may have divergent requirements. In addition, increasing numbers of local deciduous and large berry producing shrubs increased the likelihood of ground and shrub nesting species, but this was not true for coniferous and small < lm berry producing shrubs (Table 7, odds ratios  D1>D3  ,c3,B3  Ihrubs  ).  Bird Species Incidence as a Function of Park-space by Distance Species incidence, estimated by the percent of stations (with similar habitat characteristics) occupied, should increase both with area and proximity to parks if parks are 'source' areas for marginal nesting locations along residential streets {prediction 4). From the results of the canonical correspondence analysis and ordination, certain bird species were associated with the Park 0 - 260 m index and others were more strongly associated with Park 500-760 m (Figure 5). The percent stations occupied (incidence) of these bird species were plotted as a function of proximity to park if they occurred at more than 10 stations (n>10, Figure 9). These graphs indicated that for some species, incidence increases with park area as an inverse function of distance, but the trend, if any, is most apparent with the Park 500 - 760 m index (Figure 9C-D). According to the results of C C A , Cowbirds and Juncos were associated with Park 0 - 260 m, but their site incidence patterns showed no relationship with this variable (Figure 9A). There were slight increases in the species incidence of Rufous Hummingbirds, Orange-crowned Warblers, and Pine Siskins with increasing values of Park 0 -260 m, but low r-square values indicate that this variable had low predictive ability for these species in urban areas (Figure 9B, r = 0.14, 0.21, 0.18 respectively, p<0.05). The low amount of variation explained by these 2  regressions may also be a function of low sample size for these species. The Winter Wren, Swainson's Thrush, and Cedar Waxwing appeared to have higher incidence in association with  58  increased values of Park 500-760 m index (Figure 9C-D, rMJ.22, 0.35, 0.43 respectively, 0.05>p<0.005), indicating that these species occurred at sites with the highest amounts of park area per distance unit at a scale of 500-760 m.  The occurrence of American Robins, Spotted Towhees, and Song Sparrows modeled with logistic regression showed that the variable Park 500-760 m was significant at predicting their distribution. The percent stations occupied with similar values of Park 500-760 m was fit with simple regression (Figure lOa-c). Park 500-760 m explained 50% of the variation in Spotted Towhee incidence, followed by American Robin (45% variation explained) supporting the importance of regional park-space to the residential nesting occurrences of these species (Figure lOa-c, p<0.001). Song Sparrows were not adequately explained by Park 500-760 m index (20% variation explained, Figure 10), indicating the presence of Song Sparrows is not highly associated with this variable. This was not entirely surprising since logistic regression models for Song Sparrows suggested that their presence was more associated with local habitat conditions, particularly riparian habitat. Park 500-760m is a landscape level predictor with an unimpressive level of significance in relation to the other variables in the final logistic model for this species (Table 7). Overall, while Park 500-760 m is a good predictor for groups of nesting bird guilds and some common species such as robins and towhees, it cannot be used as an indicator for all species.  59  x i .2 ^ o J3 ^ 43  o  > 7 3  rt CD ' u 43 O * oo J  •S rt  > o  -a 'o CD CD ti OH S  O T3 O  00 -rt od 43 « 3  S o § CD  OH  43  s  O  e s  « o  ,0 ^  00 * O  * ^  OH in  CD  o 3 o T3  V OH  O V  OH  CD 3 ~  CD  'SH U 3 O ^  2  3  60  o o 2 oo 3 2 a 60  '•4—»  (paidnoDO suopeis %) aousppirj sapads  rt -»-» oo • ^ rn  g, CD CJ  C  CD  s  rt c  s CD rt T3 CD 3  °  s  C  OH  2 a c o o a, o —H JD rt R O o rt rt .2 o 2  •»—t"3 T3  CD  T3  CD  rt  OH  PQ r  < OS  C O  CD S »  (paidnoDQ suopBis %) aousppuj sapads  S rt  s  £  E o o  60 Y = 0.07Park500-760 - 1.9xl0" + 0.29  Y = 0.05Park500-760 - 5.5X10 - 0.06  3  -4  1.2  1.2  1.0  a) Spotted Towhee (n=87) incidence vs Park 500 - 760 m  1.0  -a  -a 'a.  .9 'cL 3  o  b) American Robin (n=186) incidence Park 500 - 760 m  4  3  O  oo t#  Rsq = 0.50***  n o  Rsq(adj) = 0.46  CO  oo i£  0.0  10  15  20  Park Area/Distance 760 Index (m)  Rsq = 0.45*** Rsq(adj) = 0.41  0.0 10  25  15  20  Park Area/Distance 760 Index (m)  Y = 0.02Park500-760 - 2.3x10"*- 0.02  c) Song Sparrow (n=38) incidence vs Park 500 - 760 m  T3 .3 SH  'a, 3  8  .2  o  c/> C  •2 .1  Rsq = 0.20* Rsq(adj) = 0.14  00  ^0.0 20  25  Park Area/Distance 760 Index (m)  Figure lOa-c). Incidence functions (percent sites occupied) verses Park 500 - 760 m index for three selected bird species, Vancouver - Burnaby, British Columbia (*p<0.05, ***p<0.001, refer to Table 8 for quadratic fit parameter estimates).  25  61  T A B L E 8. Curve model fit of Park 500-760 m index as a predictor for the incidence (percent stations with similar Park 500-760 m values occupied) of three selected bird species, Vancouver - Burnaby, British Columbia. Standard Variable (m) Parameter error of the estimate (B) estimate (B)  MODEL  T  Sig.T  n  1.93  0.06  29  R  2  SSQ  (R'adi) Spotted Towhee Quadratic Fit F=13.63*** American Robin Quadratic Fit F=11.09*** Song Sparrow Quadratic Fit F=3.46  ***p<0.001  Park 500-760 m (Park 500-760 m)  0.05  0.02  -5.5 x 10"'  1.2 x 10"  -0.44  0.66  Constant  -0.06  0.11  -0.54  0.59  Park 500 - 760 m  0.07  0.02  2.96  0.01  -1.77  0.09  (Park 500 - 760 m)  2  2  -1.9xl0"  3  3  l.lxlO"  3  Constant  0.29  0.10  3.03  0.005  Park 500 - 760 m  0.02  0.02  1.00  0.32  -0.26  0.80  -0.23  0.82  (Park 500 - 760 m)  2  Constant  -2.3xl0" -0.02  4  9.0xl0" 0.08  4  0.50  1.51  (0.47)  29  0.45  0.95  (0.41)  29  0.20 (0.14)  0.20  62  DISCUSSION  The avian community in Vancouver and Burnaby, B.C. showed specialist and generalist occurrences along an intensity of urbanization gradient with decreasing species richness and increasing relative abundance (Figure 5, Appendix V), consistent with previous work (Lancaster and Rees 1979, Blair 1996, Haddidian et al. 1997). This urbanization gradient was strongly dominated by landscape level measures, so that the amount of park area per distance segment away from point count stations provided the best separation of the observed species distribution patterns. Several local habitat variables, such as housing density, intersection size, water, small and large deciduous trees and tree species richness, dominated the most urban extreme of the gradient (Figure 5). Subjectively defined land use categories such as 'suburban' and 'urban encroachment' did not fully simplify this habitat gradient (Figure 6 and Figure 8). Suburban sites were dispersed across a large range of habitat conditions with respect to the gradient of urbanization.  I have presented several habitat models using logistic regression techniques for five nesting guilds and three selected species, the Song Sparrow, Spotted Towhee, and American Robin, that may be of interest to urban land planners, landscape architects and urban residents. Landscape variable-only models improved the fit more and used fewer variables than did local-variable only models (improvement in log likelihood vs constant only model) for all nesting guilds and species, with two exceptions - shrub nesters and Song Sparrows (Table 7). Contrary to my prediction, however, local models were often more sensitive and/or specific than landscape variables (chi-square goodness of fit test, Hosmer and Lemeshow 1989) in predicting individual species presence/absence. This indicates that when modeling for individual species as opposed  63  to groups of species, local variables were more precise. The local variable models for cavity and shrub nesters also provided more accurate percent correct predictions than did landscape variable only models (Table 7).  Prediction 1, Bird Species along a Gradient of Urbanization The spatially varying effects of urbanization seemed to pattern the corresponding bird communities along a gradient that gradually separated a wide distribution of birds in Greater Vancouver. The bird communities approximated the gradient paradigm described by McDonnell and Pickett (1990) and pioneered by Whittaker (1967) and is the view that spatial patterns in the environment organize the structure and function of populations, communities, or ecological systems. The rate of habitat change in space affects the steepness of the gradient in ecosystem structure and function (Karr and Freemark 1983, McDonnell and Pickett 1990, Keddy 1991). If the spatial patterns of birds were generated by dispersal and settlement biases (Bolger et al. 1997), then landscape habitat characteristics such as proximity to natal areas will affect the chances that a bird occurs at a site. The local matrix of habitat conditions will affect how quickly forest species turnover to edge and urban bird species. In my research in Greater Vancouver, parks and reserves augmented avian diversity in the surrounding residential areas along a gradient of urbanization that changed with conditions of the local habitat as well. Habitat specialists are likely to disperse from reserve areas in times of high regional nesting success (Hinsley et al. 1996, Helzer and Jelinski 1999,) to more marginal nesting areas where they may be exposed to increased predation and competition (Soule et al. 1988, Rottenborn 1999). Canonical correspondence analysis depicts the overall pattern in the relation between the community of birds and the observed habitat variables along multiple linear regression model  64  axes. Since the second and third axes are limited by the condition that they be uncorrelated with the first axis, it is possible that the second axis selected is a mathematical construct or polynomial gradient rather than a 'true' second habitat gradient. A modified and folded first axis is a type of polynomial distortion which creates the 'arch effect' (Jongman et al. 1995). The arch effect is apparent in my results and is a recognized fault of C C A (Jongman et al. 1995), so I have not interpreted the second axis in any detail. Species points at the edge of ordination diagrams, the Red-eyed Vireo and Black headed Grosbeak, are often species with very low abundance in the data set (Table 1, Figure 5) that ordinate at the extreme edge of the gradient either by chance or because they prefer extreme conditions (Jongman et al. 1995). One can only decide between these two possibilities with additional knowledge. Red-eyed Vireos an occasional breeder in the area, their presence at these sites may be a chance occurrence or may indicate breeding summer visitors (Ehrlich et al. 1988). The Black-headed Grosbeak is often found in open woodlands and at the forest edge (Ehrlich et al. 1988), so their presence may be related to park-space (Park 500-760m). However, species at the extreme ends of the C C A ordinations have little influence on the analysis (Jongman et al. 1995). Another shortcoming of the method of weighted averaging is that species at the very centre of the diagram (Steller's Jay, Violet-green and Barn Swallow) may either be unimodal with optima at the centre, or bimodal, or unrelated to the ordination axes. The preferred habitat of Steller's Jays is successional forest edges and neighborhoods with suburban vegetation characteristics, consistent with their position in the canonical correspondence ordination (Figure 5, Sieving and Willson 1998). Violet-green and Barn Swallows may be expected to have optima at the centre of the axes, as they are prevalent throughout many areas of the city and they nest in building cavities.  65  The results of C C A on these data support the hypothesis that landscape variables are better able to explain the variability in the Greater Vancouver urban bird community than the local habitat, but the resulting bird-habitat trends are exploratory. In general, the ordination diagrams of individual species (Figure 5) and point count stations (Figure 6), and the logistic regression analysis (Tables 7) on nesting guilds, are largely consistent. The same landscape and local variables showed significant relationships with bird species and nesting guild occurrence in logistic regression. Given that the results of the canonical correspondence analysis and the regression analysis are similar, we can be confident that the important environmental variables have been examined in the survey (Jongman et al. 1995).  Prediction 2, Landscape verses Local Habitat Predictors The distance from large potential 'source' and resource habitat areas was one of the most important variables in determining the distribution of breeding birds in Greater Vancouver. Similar results were found by Munyenyembe et al. (1989) and Germaine et al. (1998), but results from multiple scale studies have not consistently shown that landscape level effects are significant. Neither the occurrence of individual breeding bird species, nor species richness, was related to any measure of landscape context in foothill shrub avian communities of Colorado (Berry and Brock 1998). They suggested that bird species in the foothill shrub could have evolved tolerance for habitat fragmentation because it represents the natural state of the landscape in which these birds occur. Clergeau et al. (1998) found that local site level features were more important than landscape level features along an urban-rural gradient in two cities, Rennes, France and Quebec City, Canada. However, they did not quantify the surrounding landscape 'setting' and excluded all natural areas such as parks and woodlots from their analysis, examining only the percent of vegetated open areas under different land uses. So, it is possible  66  that they were unable to detect landscape level effects because they did not capture variation in habitat at the landscape level in their study. In urban areas, there is more support for landscape level effects, but the perspective is generally that urbanization around habitat patches and riparian areas affects birds within these areas, rather than residential bird communities being 'rescued' by their surroundings, as in this study. Bird species richness in urban riparian corridors in Alachua County, Florida was negatively correlated with housing densities in areas adjacent to these corridors (Smith and Schaefer 1992). Urbanization on lands adjacent (< 78 ha) to intact riparian woodlands (~1 ha width) in Santa Clara Valley, C A , had a substantial impact on the riparian bird communities. The number of bridges within 500 m of a plot was significantly related to decreasing species richness in regression models (Rottenborn 1999). Saab (1999) had similar findings for breeding birds in riparian forests, South Forks, Idaho. Examining the effects of land cover types around study plots, Germaine et al. (1998) found that the abundance of 17 of 21 bird species were associated with land cover variables (> 3 ha) such as housing density at the landscape level in Tuscon, Arizona. Landscape descriptors (< 20-3000 ha) were often better predictors than local habitat variables (< 3 ha) for the abundance of 10 of 20 bird species in San Diego County, California (Bolger et al. 1997). My local variable-only models for individual species and for cavity and shrub nesters were often more sensitive or specific predictors than landscape variable-only models. The cavity and shrub nesting guilds contain species that are known to nest under a variety of conditions along an urbanization gradient, Black-capped Chickadees, Northern Flickers and White-crowned Sparrows for instance. The scale of response for these groups of species acts more at a local level along the urbanization gradient probably because many of these species have adapted to landscape habitat fragmentation and are cueing into local habitat variation. American Robins are  67  a ubiquitous species, known to be very robust to changes in broad scale urbanization, so it is not surprising that they were more associated with local variables than landscape variables. Spotted Towhees were insensitive to landscape level habitat fragmentation in other studies as well (Bolger et al. 1997, Berry and Bock 1998). Although the occurrence of Towhees and Robins were better predicted by local habitat conditions, their incidence (% stations occupied) was associated with the amount of park area over distances within 500 to 760m at the landscape scale. It would be interesting to investigate nesting success and post-natal dispersal of these resident species in relation to distance to the edge of a large forested area. The incidence of Song Sparrow in my study was associated with local variables more than with the landscape variable, Park 500-760 m. Other research has shown that this ground nester appears to display a high degree of site fidelity - the observed maximum distances dispersed by Song Sparrows were shorter on average (72% less) than predicted values (Sutherland et al. 2000). In another urban bird study, Song Sparrow abundance was related to the (local) presence of, or increases in, shrub stratum on one ha sample plots in Montreal, Quebec parks over a 15-year period (Morneau et al. 1999), and this was confirmed in Vancouver as well. Demographic differences in reproduction and survival success at the local site level are strongly associated with nearby native habitat resources that may be acting as sources of species immigrants and visitors (Pulliam 1988). The overall avian community response may be related to a combination of several factors, such as, differences in rates of competition, predation, nest parasitism, immigration and emigration, nesting success, and resource abundance. Each of these mechanisms operates at the site level and each has features associated with the surrounding landscape, so landscape level variables may integrate a combination of factors. Landscape variables are likely to have more information than the variables measured at smaller scales. This would explain the inconsistency between the results for ground nesters (landscape models were  68  better predictors) and two individual species of ground nesters, Song Sparrows and Spotted Towhees, with occurrence patterns better predicted by local level variables. Landscape models seem to be better able to predict for presence species groups rather than for individual species. If the landscape variables explain more variation, it may not indicate that bird communities are mechanistically associated with landscape measures, it may simply mean that we can predict bird distributions better with measurements made at the level of the landscape. In a study on the impacts of urbanization on riparian bird communities, the abundance of ground nesters was highly correlated with measures of the local habitat such as native and total vegetation volume within 10 meters. Landscape measures such as the overall degree of urbanization or disturbance within 500 meters of a plot were also correlated with ground nester abundance (Rottenborn 1999). Cats and other predators were the suggested long-term mechanism keeping the numbers of ground nesters in his study low; although the density of cats was estimated, however, it was not a significant variable in his final multiple linear regression models (Rottenborn 1999). My results agree with those of Rottenborn, but I would suggest that habitat may be more important than the number of cats when it comes to nest site selection for some ground nesters (i.e., Towhees). Still, cat predation could be impacting nesting success after birds have already invested nesting time at a location, but further study would be required to examine cat predation of native birds in urban areas close to parks and in suburbs. A single pet cat was responsible for the deaths of at least 62 individual birds over a period of 18 months in Michigan (Bradt 1949 cited in Soule et al. 1988).  Prediction 3, Urban Land Use Categories, a Poor Indicator of Bird Habitat Types? Many studies have focused directly on 'suburban' birds, but fail to define the term, suburban, in more than a general sense (e.g., Vale and Vale 1976, Rosenberg et al. 1987, DeGraaf 1991,  69  Sodhi 1992, Zalewski 1994). Land use terms (suburban, rural, commercial) have different meanings in different cultural contexts, where they also change with societal gradients (i.e., changes in economic status, socio-cultural differentiation, land use and housing density changes) and societal gradients may or may not represent parallel habitat gradients of importance to birds. Nonetheless, the utility of the terms, urban and suburban, lies in their (presumed) ability to simplify a complex gradient into discrete groupings that represent an easily understood habitat type of significance to birds and land-use planners. So, standard measures that accurately quantify the position of an urban or suburban site along a complex urbanization gradient would be useful (Marzluff et al. 1998). However, according to my results in Greater Vancouver, land use categories were not accurate or precise ways to delineate bird habitat and perhaps such subjective habitat groupings should be avoided. Some studies have found that species diversity and bird biomass peaked in suburban areas with slight levels of development, rather than at the most natural site (Batton 1972, Lancaster and Rees 1979, Blair 1996). Star plots of Burnaby suburbs in this study showed that suburban areas had intermediate levels of impervious surface cover (54%) and intermediate levels of vegetation at different scales (Figure 8A-C). The finding that species richness peaks in suburbs or at low levels of diversity has been related to the hypotheses of intermediate disturbance theory (Connell 1978, Collins et al. 1997, as in Blair 1996). According to this theory, superior competitors are assumed to be susceptible to intermediate conditions of disturbance and, thus, they do not achieve dominance, while less competitive species co-exist because they are assumed to be relatively tolerant of intermediate disturbance conditions (spatial in this case, Collins et al. 1997). Although the results of this study agree with this hypothesis, because species richness seemed to peak at intermediate levels of the urbanization gradient, C C A axis  70  one, the relationship showed a large amount of variation and could be modeled both curvilinearly and linearly (Table 6). Combined with park-distance metrics (see below), impervious surface cover may be a better habitat quality indicators for bird species distributions. The amount of impervious surface cover in the surrounding area (<500 m) was an important and frequent predictor of bird occurrence in this study. American Robin was the only species and shrub and cavity nesters were the only nesting guilds that did not include impervious surface as a significant predictor (Table 7). However, Robins, Towhees and shrub nesters did include the number of local level house lots as a significant predictor, and the number of house lots had the highest correlation with landscapelevel impervious surface cover (  Pearson  . r = 0.72). Geologists and hydrologists have used percent s  impervious surface cover to assess water cycle and infiltration changes associated with urbanization (Stankowski 1972 cited in Arnold and Gibbons 1996). They suggest using impervious surface cover as an integrative environmental indicator of complex urban environmental issues, such as cumulative water resource impacts with many nonpoint sources of pollution. The relative amount of impervious cover seems to be a relevant indicator for bird species distributions as well, because many native species have declining abundance with the increasing percent area paved (e.g., Munyenyembe et al. 1989, Germaine et al. 1998).  Prediction  4, Species Incidence as a Function of Park Area and Distance  Area and distance from parks explained species incidence, and six of the eight final logistic regression models produced, included the landscape variable Park 500-760m. This park by distance index also explained 35 to 50 percent of the variation in the site occurrences (incidence) of Swainson's Thrush, Cedar Waxwing, American Robin and Spotted Towhee (Figure 9 and 10).  71  Thus, park area weighted inversely by distance away was a significant predictor for the presence of several bird species in Greater Vancouver, and species incidence increased with the amount and proximity to parks. Only the ledge and building nester regression models did not include Park 500 - 760 m as a significant predictor. This was not surprising given that building and ledge nester guilds included many urban generalist species. The building nester guild included non-native species such as European Starling, House Sparrow, and Rock Dove and these species generally nest on building ledges, eaves, and building cavities, within highly urbanized areas where they can find adequate food (Ehrlich et al. 1988). Although the presence of Violet-green and Barn Swallows should be related to park habitat nearby, given their known habitat requirements, the relationship may be a weak one in Greater Vancouver and could have been masked by the occurrence patterns of the other species included in the building nester guild. Barn Swallows are often found near open fresh water ponds, and Violet-green Swallows breed in areas with coniferous or deciduous open forests (Ehrlich et al. 1988). Three species of gulls were included in the ledge nester guild, however the most frequently sighted species was the Glaucous-winged Gull. The gulls were wide ranging and their presence was probably more an indication of food availability rather than nesting occurrence. In another urban bird study, decreasing weights were assigned to the amount of developed land within increasing concentric rings at different spatial scales around bird survey points, coastal southern California (Bolger et al. 1997). They assigned an inverse weighting scheme to the proportion of developed land within each concentric ring around point counts located in 'islands' of natural habitat, creating an urban exposure index. However, when this variable was entered in logistic regression models, urban exposure was significant in only two of twenty species occurrence habitat models. More significant variables were the non-weighted proportions of edge and natural areas around 'island' survey sites. M y results agree with those  72  of Bolger et al. (1997), but I did not use the proportion of natural area within concentric rings around each bird survey station, rather I used park area within each ring directly and divided that area by distance away. Dividing park area by distance weights habitat farther away with an exponentially declining importance. The use of the proportional natural area within each concentric ring around a site also gives habitat areas farther away a declining importance, because the outermost rings will be much larger than the innermost ring (as in Bolger et al. 1997). My inverse of distance weighting scheme perhaps relates more directly to the theoretical expectation that species and guild occurrence should decrease non-linearly with distance from 'natural' areas, given known natal dispersal trends for birds (Sutherland et al. 2000). Moreover, if species are using park habitat as a resource for complementary habitat, or if they are supplementing these areas with resources that are lacking in residential urban areas, then close habitats should be visited more frequently than distant ones.  Speculation - Dispersal from Natal Area? Several studies have found declining species richness and declining site incidence with increasing distance from native habitats (Munyenyembe et al. 1989, Bolger et al. 1997, Germaine et al. 1998). My research confirms these findings, but the decline in species richness may be an isolation effect or could it be related to the quality of the local habitat in urban areas. Many urban and forest dwelling bird species are migratory and are not likely to be severely limited by their dispersal abilities (Tilghman 1987, Soule et al. 1988). Birds can probably move with relative ease through sparsely treed and patchy areas to local sites with good nesting habitat characteristics. There is evidence to suggest that birds do not generally disperse far from their natal area and that the number of individuals dispersing decreases exponentially with distance from parental nesting sites (Sutherland et al. 2000). Published data on natal the dispersal  73  movements for 77 bird species indicates that most dispersers move relatively short distances (30% of species moved median distances of < 1 km and just over 60% of species examined moved < 10 km). In their survey of published dispersal literature, the maximum distances dispersed were 28.4 km for migratory bird species and 24 km for resident species. If the local (matrix) habitat is highly unsuitable and/or if the birds do not survive their dispersal movements, then the number of individuals at particular distances from natal areas would be low (Sutherland et al. 2000). If the local (urban matrix) habitat has an abundance of breeding sites, albeit of a lower quality, than potentially overcrowded, high quality natal sites, then birds may disperse to lower quality sites to breed (Pulliam 1988, Brawn and Robinson 1996). The finding that bird species incidence is related to inverse distance weighted park-space seems to agree with their findings. If dispersal decreases exponentially with distance from natal area, then you would expect the percent of sites occupied to be highest close to large forested areas (assuming that large parks are high quality natal areas for many species) and to decrease inversely with distance. High quality local habitat may remain unoccupied if it is too far away.  Study Design Critique I conducted point counts only once per breeding season over each of two years. This may limit inferences I can make about community trends. Because urban areas have large amounts of impervious surface cover, they have large distances over which resources experience their full range of variability. Many of the bird species living in these areas have small territories and relatively high densities in relation to the changes in habitat. This combination of factors is ideal for detecting species-habitat relations at the level of the landscape (Goodinson 2000). So, replication in space rather than performing multiple point counts at fewer locations can be an advantage rather than a disadvantage in urban areas because replication in space leads to more  74  certainty about the species - habitat associations at the expense of certainty about a particular species presence at any individual point (Bolger et al, 1997, Goodinson 2000). Data were collected earlier in the 1998-breeding season than in the 1997-breeding season and included some late spring migrants from May 1998. While data were screened by omitting species that were sighted only once, this screening process was not sufficient enough to screen migrant species such as Ruby-crowned Kinglets (n^S) and Fox Sparrows (n=3), species that are known to breed farther north in British Columbia (Campbell et al. 1997). I avoided underestimating the number of species with low detection probabilities but included some migrant individuals in the analysis. Vagrant individuals of two gull species, the Mew Gull and the Ring-billed Gull, were also included in the analysis although they do not breed in the area (Campbell et al. 1997). These species are known to occur in the Vancouver area year round, but do not breed here (Campbell et al. 1997). Twelve species were omitted from the breeding season analysis because they were recorded only once (Appendix III). Most of these birds were recorded within one of the three large parks and thus would have been placed at the left end of the C C A ordination. By removing these species, the ordination and logistic regression model results would have slightly underestimated the landscape effects of park area by distance. I classified species into nesting guilds according to their known habitat affinities across North America (Ehrlich et al. 1988), but species are known to change their nesting preferences depending on the availability of suitable nesting sites (Wiens 1989). The general nesting preference across North America may not apply directly to Vancouver and Burnaby. I will mention a few species in particular that could be grouped into different nesting guilds. House Finches, Bushtits and American Robins were grouped into the deciduous tree nesting guild, but also commonly nest in buildings. Song Sparrows were grouped into the ground-nesting guild,  75  but Song Sparrows are commonly a shrub nesting species as well. Black-capped Chickadees and Red-breasted Nuthatches are cavity nesters that commonly excavate cavities in coniferous trees. These two species could be grouped into the cavity nesting guild or the coniferous tree, nesting guild. Because of the inherent arbitrariness in the method of guild classification, resultant guild habitat 'patterns' are consequences of imposing an arbitrary arrangement on a community that may actually be structured ecologically in some other way altogether - or not structured at all (Wiens 1989). Given that the known nesting guild groupings were at least broadly confirmed by the species distribution patterns evident in my C C A ordination, I have some evidence that this is not a problem in my results. Moreover, when Chickadees and Nuthatches were moved from the coniferous tree, nesting guild to the cavity nester guild because this made more sense biologically, logistic regression model results were qualitatively similar. Landscape level variables, and in particular the park-distance variable, Park500-760m, was a significant predictor in both models. The parameter values in the model changed quantitatively when species with similar nesting requirements were added or removed from a guild, but the general trends remained consistent. This was also true when Brown-headed Cowbirds were moved from the deciduous tree, nesting guild to the shrub-nesting guild in order to place this nesting parasite with its most common host, the Song Sparrow. I was able to sample a large number of plots over an extensive area, allowing more analytical power to detect species - habitat trends at the level of the landscape. Urban Vancouver and Burnaby are exceptional cities due to three large (>324 ha) natural area park remnants in close proximity to different residential sections, providing resource variability at a landscape level. Transect placement was subject to the constraint of road length and orientation, yet it is unlikely  76  that road orientation and length would introduce any bias in the sampled populations of birds since these roads were typical of urban residential streets (Appendix IV). I found clear patterns of landscape and local level bird-habitat associations along an urbanization gradient, but it cannot be assumed that these patterns of species occurrence reflect ecological success along this gradient (Van Home 1983 cited in Berry and Bock 1998). It is possible that residential areas close to parks are attracting species to marginal nesting habitat in times of high regional species densities, but whether or not birds can nest successfully in these locations is unknown, especially considering potentially higher rates of predation and disturbance. Still, the patterns of habitat occurrence provide a framework for future investigations into breeding success, dispersal, and species-species relationships in urban areas.  CONSERVATION IMPLICATIONS  Ninety percent of all bird species known to occur in and sixty percent of the species known to breed in British Columbia (Demarchi, 1996) are found in the Lower Mainland and the Georgia Basin area, an area under extreme urban development pressure. In addition, the Basin may be of national significance to breeding landbirds with the highest number of species breeding and over-wintering in Canada (Demarchi, 1996). Currently, only five percent of the Lower Mainland (80,000 ha of 1.58 million ha) is set aside as parkland (GVRD Strategic Planning Department 1996). Although a large portion of the Lower Mainland maintains some natural character under forest management and agricultural reserves, more than 1.2 million additional people are expected to live in the Greater Vancouver Region alone by the year 2021 (GVRD Strategic Planning Department 1993).  77  The results of my thesis suggest that matrix areas surrounding parks and reserves should be integrated into urban planning and development designs. Recommendations for urban planners, particularly in Vancouver, developing on the verge of continuously forested areas would include minimizing impervious surface cover and house size, keeping grass cover to a minimum, maintaining native tree cover and berry shrubs, and integrating new ponds, and natural fresh water sources, into planning designs. Recommendations to home-owners living close to large parks and continuously forested areas may include keeping cats indoors because although cats were not investigated in this study, it is reasonable to assume that they could have a significant impact on ground and shrub nesters in these areas. Peripheral residential areas have a high likelihood of recruiting sensitive nesting species and will probably experience frequent visits by different species from nearby parks and could be managed as potential sink habitats. Investigations at larger spatial scales allow us to study the response of bird species to landscape level habitat heterogeneity, a scale often neglected in the past by local habitat land use classifications and diversity investigations. Ecologists have become increasingly aware of the importance of examining ecological processes at multiple spatial scales and it is perhaps nontrivial that this direction coincides with the use of geographic spatial analysis technologies. It may be pertinent to consider where these bearings are taking us because although landscape level studies detect and predict patterns, local level and demographic studies are necessary to determine many of the mechanisms involved in population and community change. I would recommend that urban bird studies examine a few key landscape measures, such as distance to, and size of, large natural park areas with potentially good habitat for birds. Then, ecologically meaningful landscape metrics can be developed using the negative exponential of distance as an  78  indicator of species occurrence patterns. I would also recommend that impervious surface be measured at different spatial scales. Habitat variability at different spatial scales is the most important, easily measured and controlled, factor affecting bird communities in urban systems. Nonetheless, changing ecosystem processes resulting from human development (altered fire, water, radiant energy flows, nutrient and pollutant cycling) may have the greatest, long-term impact (Marzluff et al. 1998). Higher ambient temperatures in urban environments, alone, have been related to shifts in the availability of invertebrate foods, affecting egg-laying dates and the breeding biology of magpies (Pica pica) in Britain (Eden 1985). In the face of inevitable urban expansion there are many local vegetation changes that we can make, but the impact on the bird community may be expected to decline with distance from large continuously forested areas and increasing impervious surface cover. Providing complementary habitat resources such as, small backyard habitat ponds, different types of feeding resources, berry producing shrubs, and cover options would increase the likelihood of several species of shrub and ground nesters, White-crowned Sparrows, Common Yellowthroats, American Goldfinch, Spotted Towhees and Dark-eyed Juncos. Perhaps we may even hear the swirling song of a Swainson's Thrush, but residential areas on the extreme end of the urbanization gradient may never have the pleasure.  79  LITERATURE CITED  Ahl, V. and T. F. H . Allen. 1996. Hierarchy Theory: A Vision, Vocabulary, and Epistemology. Columbia University Press, New York, N.Y., USA. Arcese P., J. N. M . Smith, W. M . Hochachka, C. M . Rogers, and D. Ludwig. 1992. Stability regulation and the determination of abundance in an insular song sparrow population. Ecology 73:805-822. Arnold, C. L . and C. J. Gibbons. 1996. Impervious surface cover: The emergence of a key environmental indicator. Journal of the American Planning Association 62:243-255. Baker, R. R. 1993. The function of post-fledging exploration: A pilot study of three species of passerines ringed in Britain. Ornis Scandinavica 24:71-79. Batten, L. A . 1972. Breeding bird species diversity in relation to increasing urbanization. Bird Study 19:157-166. Beissinger, S. R. and D. R. Osborne. 1982. Effects of urbanization on avian community organization. Condor 84:75-83. Berry, M . E. and C. E. Bock. 1998. Effects of habitat and landscape characteristics on avian breeding distributions in Colorado foothills shrub. The Southwestern Naturalist 43:453-461. Blair, R. B. 1996. Land use and avian species diversity along an urban gradient. Ecological Applications 6:506-519. Bolger, D. T., T. A . Scott, and J. T. Rotenberry. 1997. Breeding bird abundance in an urbanizing landscape in coastal southern California. Conservation Biology 11:406-421. Brawn, J. D. and S. K. Robinson. 1996. Source-sink population dynamics may complicate the interpretation of long-term census data. Ecology 77:3-12. Bradt, G. W. 1949. Farm cat as a predator. Michigan Conservation 18:25-26. Campbell, R. W., N. K. Dawe, I. McTaggart-Cowan, J. M . Cooper, G. W. Kaiser, M . C. E . McNall and G. E . J. Smith. 1997. The Birds of British Columbia. U B C Press, Vancouver. Campbell, C. A., and A. I. Dagg. 1976. Bird populations in downtown and suburban KitchenerWaterloo, Ontario. Ontario Field Biologist 30:1-22. Clergeau, P. 1995. Importance of multiple scale analysis for understanding distribution and for management of an agricultural bird pest. Landscape and Urban Planning 31:281-289.  80  Clergeau, P., J-P. L. Savard, G. Mennechez and G. Falardeau. 1998. Bird abundance and diversity along an urban-rural gradient: A comparative study between two cities on different continents. Condor 100:413-425. Cleveland, W. S. 1993. Visualizing Data. Hobar Press, Summit, New Jersey, USA. Connell, J. H. 1978. Diversity in tropical rain forests and coral reefs. Science 199:1302-1310. Collins, S. L. and S. M . Glenn. 1997. Intermediate disturbance and its relationship to withinand between-patch dynamics. New Zealand Journal of Ecology 21:103-110. Demarchi, D. A. 1996. An introduction to the ecoregions of British Columbia. Wildlife Branch, Ministry of Environment, Lands and Parks, Victoria, British Columbia, Canada. DeGraaf, R. M . 1991. Winter foraging guild structure and habitat associations in suburban bird communities. Landscape and Urban Planning 21:173-180. DeGraaf, R. M . , A. D. Geis and P. A. Healy. 1991. Bird population and habitat surveys in urban areas. Landscape and Urban Planning 21:181-188. Diamond, J. M . 1988. Urban extinction of birds. Nature 333:393-394. Dinicola, R. S. 1990. Characterization and simulation of rainfall runoff relations for headwater basins in Western King and Snohomish Counties, Washington. U.S. Geological Survey. Water Resources Investigations Report. 89-4052:51 pp. Dunning, J. B., B. J. Danielson, and H. R. Pulliam. 1992. Ecological processes that affect populations in complex landscapes. Oikos 65:169-175. Eden, S. F. 1985. The comparative breeding biology of magpies Pica pica in an urban and a rural habitat (Aves: Corvidae). Journal of the Zoological Society ofLondon 205:325-334. Edgar, D. R. and G. P. Kershaw. 1994. The density and diversity of the bird populations in three residential communities in Edmonton, Alberta. Canadian Field-Naturalist 108:156-161. Ehrlich, P. R., D. S. Dobskin, and D. Wheye. 1988. The Birders Handbook: A Field Guide to the Natural History of North American Birds. Simon and Schuster Inc., Toronto, Canada. Emlen, J. T. 1974. A n urban bird community in Tucson, Arizona: derivation, structure, regulation. Condor 76:184-197. Fernandez-Juricic, E. 2000. Local and regional effects of pedestrians on forest birds in a fragmented landscape. Condor 102:247-255. Forman, R. T. and M . Godron. 1986. Landscape Ecology. Wiley & Sons, New York.  81  Fraile, L., Y . Escoufier and A. Raibaut. 1993. Analyse des Correspondances de Donnees Planifiees: Etude de la Chemotaxie de al Larve Infestante d'un Parasite. Biometrics 49:11421153. Freemark, K. E. and H. G. Merriam. 1986. Importance of area and habitat heterogeneity to bird assemblages in temperate forest fragments. Biological Conservation 36:115-141. Gauch, H. G. 1973. The relationship between sample similarity and ecological distance. Ecology 54:618-622. Gavareski, C. A . 1976. Relation of park size and vegetation to urban bird populations in Seattle, Washington. Condor 78:375-382. Germaine, S. S., Rosenstock, S. S., Schweinsburg, R. E. and W. S. Richarsdson. 1998. Relationships among breeding birds, habitat, and residential development in Greater Tucson, Arizona. Ecological Applications 8:680-691. Goodinson, C. 2000. Limitations of the point count method for inferring stand-level species resource relationships: A sampling simulation approach. MSc. Thesis, Department of Forest Sciences, University of British Columbia. Greater Vancouver Regional District. 1993. Creating Greater Vancouver's Green Zone. The Livable Region Strategic Plan. Strategic Planning Department, Parks Department, B.C. Greater Vancouver Regional District. 1996. Creating Our Future. Steps to a More Livable Region. Strategic Planning Department, Parks Department, B.C. Greater Vancouver Regional District. 1999. Municipalities in the GVRD. Regional District General Municipal and Hospital Values, British Columbia Assessment Authority. Green, R. N. and K. Klinka. 1994. A Field Guide to Site Identification and Interpretation for The Vancouver Forest Region. Land Management Handbook. 28, B.C. Ministry of Forests, Victoria, B.C. Haddidian, J., J. Sauer, C. Swarth, P. Handly, S. Droege, C. Williams, J. Huff and G. Didden. 1997. A citywide breeding bird survey for Washington, D.C. Urban Ecosystems 1:87-102. Helzer, C. J. and D. E. Jelinski. 1999. The relative importance of patch area and perimeter-area ratio to grassland breeding birds. Ecological Applications 9:1448-1458. Hilborn, R. 1997. The Ecological Detective: Confronting Models With Data. Princeton University Press, Princeton, New Jersey. Hinsley, S. A., Bellamy, P. E., Newton, I., and T. H. Sparks. 1996. Influences of population size and woodland area on bird species distributions in small woods. Oecologia 105:100-106.  82  Hosmer, D. W. and S. Lemeshow. 1989. Applied Logistic Regression. John Wiley and Sons, New York, New York. IDRISI Project. 1992. User's Guide, version 4.0. J.R. Eastman, Clark University, Worcester, MA. Jongman, R. H. G., C. J. F. Ter Braak, and O. F. R. van Tongeren. 1995. Data Analysis in Community and Landscape Ecology. Cambridge University Press, UK. Jules, E. S. 1997. Dangers in dividing conservation biology and agroecology. Conservation Biology 11:1272-1273. Karr, J. R. and K. E. Freemark. 1983. Habitat selection and environmental gradients: Dynamics in the "stable" tropics. Ecology 64:1481-1494. Keddy, P. A. 1991. Working with heterogeneity: A n operator's guide to environmental gradients. Pages 179-201 in J. Kolasa and S. T. A. Pickett, editors. Ecological Heterogeneity. Springer-Verlag, New York, New York. Knopf, F. L., Johnson, R. R., Rich, T., Samson, F. B. and R. C. Szaro. 1988. Conservation of riparian ecosystems in the United States. Wilson Bulletin 100:272-284. Lancaster, R. K. and W. E. Rees. 1979. Bird communities and the structure of urban habitats. Canadian Journal of Zoology 57:2358-2368. Little, E. L. 1980. The Audubon Society Field Guide to North American Trees. Chanticleer Press, Inc., New York, New York. MacArthur, R. H. and J. W. MacArthur. 1961. On bird species diversity. Ecology 42:594-598. MacArthur, R. H. and E. O. Wilson. 1967. The Theory of Island Biogeography. Princeton University Press, Princeton, NJ. McDonnell, M . J. and S. T. A . Pickett. 1990. Ecosystem structure and function along urbanrural gradients: an unexploited opportunity for ecology. Ecology 71:1232-1237. McGuinnes, K. A. and A. J. Underwood. 1986. Habitat structure and the nature of communities on intertidal boulders. Journal of Experimental Marine Biology and Ecology. 104:97-123. Marzluff, J. M . Gehlbach, F. R. and D. A. Manuwal. 1998. Urban Environments: Influences on avifauna and challenges for the avian conservationist. Pages 283-296 in J. M . Marzluff and R. Sallabanks, editors. Avian Conservation: Research and Management. Island Press, Washington, D.C. Meidinger, D. and J. Pojar. 1991. Ecosystems of British Columbia. B.C. Ministry of Forests, Victoria, B.C.  83  Mikusinski, G. and P. Angelstam. 1998. Economic geography, forest distribution, and woodpecker diversity in Central Europe. Conservation Biology 12:200-208. Moraeau, F., R. Decarie, R Pelletier, D. Lambert, J. L. DesGranges, and J-P. L. Savard. 1999. Changes in breeding bird richness and abundance in Montreal parks over a period of 15 years. Landscape and Urban Planning 44:111-121. Munyenyembe, F., J. Harris, J. Hone and H. Nix. 1989. Determinants of bird populations in an urban area. Australian Journal of Ecology 14:549-557. Murcia, C. 1995. Edge effects in fragmented forests: Implications for conservation. Trends in Ecology and Evolution 10:58-62. Oke, T. R., M . North, and O. Slaymaker. 1992. Chaper 5 in Vancouver and Its Region. G. Wynn and T. Oke, editors. University of British Columbia Press. Ohmart, R. D. 1994. The effects of human-induced changes on the avifauna of western riparian habitats. Studies in Avian Biology 15:273-285. Pojar, J. and A. MacKinnon. 1994. Plants of Coastal British Columbia. Lone Pine Publishing, Vancouver, B.C.. Pulliam, H. R. 1988. Sources, sinks, and population regulation. The American Naturalist  132:652-661. Quinn, J. F. and A. E. Dunham. 1983. On hypothesis testing in ecology and evolution. American Naturalist 122:602-617. Ralph, C. J., G. R. Geupel, P. Pyle, T. E. Martin, D. F. DeSante. 1993. Handbook of field methods for monitoring landbirds. General Technical Report - Pacific Southwest Research Station, U S D A Forest Service, Berkley. Ricklefs, R. E. 1990. Ecology. W. H. Freeman and Company, New York. Robbins, C. S. 1981. Bird activity levels related to weather. Studies in Avian Biology 6:301310. Robinson, S. K., F. R. Thompson, T. M . Donovan, D. R. Whitehead and J. Faaborg. 1995. Regional forest fragmentation and the nesting suiccess of migratory birds. Science 267:19871990. Rosenberg, K. V., S. B. Terrill, and G. H. Rosenberg. 1987. Value of suburban habitats to desert riparian birds. Wilson Bulletin 99:642-654. Rotenberry, J. T. and J. A. Wiens. 1980. Habitat structure, patchiness, and avian communities in North American streppe vegetation: A multivariate analysis. Ecology 61:1228-1250.  84  Rottenborn, S. C. 1999. Predicting the impacts of urbanization on riparian bird communities. Biological Conservation 88:289-299. SAS Institute Inc. 1996. SAS/STAT User's Guide, release 6.12 edition. SAS Institute Inc., Cary, N C Saab, V . 1999. Importance of spatial scale to habitat use by breeding birds in riparian forests: A hierarchical analysis. Ecological Applications 9:135-151. Sabin, T. E. and S. G. Stafford. 1990. Assessing the need for transformation of response variables. Oregon State University, Corvallis, OR. Schaefer, V . 1994. Urban Biodiversity. Pages 307-318 in Biodiversity in British Columbia. L.E. Harding and E. McCullum, editors. Environment Canada, Canadian Wildlife Service. Sewell, S. R. and C. P. Catterall. 1998. Bushland modification and styles of urban development: their effects on birds in south-east Queensland. Wildlife Research 25:41-63. Shmida, A . and M . V . Wilson. 1985. Biological determinants of species diversity. Journal of Biogeography 12:1-20. Sieving, K. E. and M . F. Willson. 1998. Nest predation and avian species diversity in northwestern forest understory. Ecology 79:2391-2402. Smith, R. J. and J. M . Schaefer. 1992. Avian characteristics of an urban riparian strip corridor. Wilson Bulletin 104:732-738. Sodhi, N . S. 1992. Comparison between urban and rural bird communities in prairie Saskatchewan: Urbanization and short-term population trends. The Canadian Field Naturalist 106:210-215. Soule, M . E., D. T. Bolger, and A. C. Alberts. 1988. Reconstructed dynamics of urban habitat islands. Conservation Biology 2:75-91. SPSS Inc. 1996. User's Guide, Base 7.0 for Windows, Chicago, IL. Stankowski, S. J. 1972. Population density as an indirect indicator of urban and suburban landsurface modifications. U.S. Geological Survey Professional Paper 800-B:B219-B224. Sutherland, G. D., A. S. Harestad, K. Price, and K. P. Lertzman. 2000. Scaling of natal dispersal distances in terrestrial birds and mammals. Conservation Ecology 4:16. [online] URL: http://www.consecol.org/vol4/issl/artl6 Tabachnick, B. and L. S. Fidell. 1996. Using Multivariate Statistics, 3rd edition. HarperCollins College Publishers, New York.  85  ter Braak, C. J. F. 1986. Canonical correspondence analysis: a new eigenvector technique for multivariate direct gradient analysis. Ecology 67:1167-1179. Thioulouse, J., D. Chessel, S. Doledec, and J. M . Olivier. 1997. ADE-4: a multivariate analysis and graphical display software. Statistics and Computing 7:75-83. Tilghman, N . G. 1987. Characteristics of urban woodlands affecting breeding bird diversity and abundance. Landscape and Urban Planning 14:481-495. Trzcinski, M . K., F. Lenore and G. Merriam. 1999. Independent effects of forest cover and fragmentation on the distribution of forest breeding birds. Ecological Applications 9:586-593. United Nations. 1999. World urbanization prospects: The 1999 revision. Population Division, Department of Economic and Social Affairs, http://www.popin.org/popl998/ Nov. 3, 2000 Vale, T. R. and G. R. Vale. 1976. Suburban bird populations in west-central California. Journal of Biogeography 3:157-165. Vander Haegen, W. M . , F. C. Dobler and D. J. Pierce. 2000. Shrubsteppe bird response to habitat and landscape variables in Eastern Washington, U.S.A. Conservation Biology 14:11451160. Van Home, B. 1983. Density as a misleading indicator of habitat quality. Journal of Wildlife Management 47:893-901. Vandermeer, J. 1997. The agroecosystem: A need for the conservation biologist's lens. Conservation Biology 11:591-592.  1  Weber, Wayne C. 1972. Birds in cities: a study of populations, foraging ecology and nest-sites of urban birds. MSc. Thesis, Department of Zoology, University of British Columbia. Wiens, J. A . 1989. The Ecology of Bird Communities. Vol. 1, Foundations and Patterns. Cambridge University Press, New York, N.Y., USA. Whittaker, R. H. 1967. Gradient analysis of vegetation. Biological  Reviews 49:207-264.  World Resources Institute. 1996. The Urban Environment. Washington, DC. Zalewski, A . 1994. A comparative study of breeding bird populations and associated landscape character, Torun, Poland. Landscape and Urban Planning 29:31-41.  86 3  © © rNO od r-^  03 CD  cd  o  P  ON ON NO  00  >—< CO  O  CO  3  rt  CN KO CO iri h h  CD  O  o rn  NO  CO  ON  t-~  o o oo o oo d I N H o( f i CN CN CN —i —i  S Q 2  - 7  ON NO NO  ON I—i  o 6 CN  I — < S  CN CN oi n vi  CN  ON  O ON O CO CN r-; — ' < CN co — ' < r-- od  3  rt  o  — ' <  Q co ON  SQ rt  .5  ON  UO © NO ON (N - H  rrt' CN CN T-H  co  h ON rrt rrt  ' ' l t °i 00 rt  °i ^  t  t  SQ  rt  CO  ©  rtr^H V X rt ^ rt S Q~ ^  CO r-; © I/O NO CO CO ON '—< CN CO NO ON •H N CN! rn rt rt rt  <P  ON  - H  73  rt -5  CO  NO  00  ^5 CD O  X  rt-£?S « rt -5  vo h cn CN| d (N r i r H ^  £ Q  HH  Q  §  co  rt  SQ  OH PH  <  CN  vo  CN CN - H  rt rt vo h o  3 EJ HH  U ^  CD  ca a •2 2  s .a 00  00  oo ON ON  r-^ ON ON  Q  rt  3 8J  rt  HO PH  Q s CD  O  CM CD  Q 00  rt  Q s rt CD  CN  oo q  TJ-  NO ON  CN  00  <  HH  00  O)  oi vi  OJ  rt in co -sr' K  rt rt rt CN  in d  r~in od  r~- oo o oo co -rr' CN  © NO  CN 00 CN ON cn tN CN —i rt  rt  NO  ON  r—i cn  —i  in d  vo  cn © ON CN rj-' rt  2  w  inV  rt rt o  CN  © © p-^  ^  cn CN ON cn od rt cn ON rt H fN (N H H  rt 3  o  3 .3  i? .3  _f  O H -  < oo O  &  §  2 ^  vo o ^ oo vq vq d ' t '5 oi rn in d CN CN >  t-- CN  rt  o  ro i vo o co O CN CN O VO r- if vo in TI- vo r-»  VO  s  u Oi  H  o  ON i> CN CN <-< c--di (N (N O h 0\ in ro co tj- co in in  Is i O  ON  jg v  CN o q rt n H N 00 K CO "^t' ON rtoo ON o o m h o\ M  ro  CO  00 CN CN 00 ^ in r r t vo ov h in  TT_  •rl- co in co in rt  ^  U  L-H  rt rt  -^J-  CO CN rt rt  00 CN CN CO rt 00 OO ON fO 00 rt ON  oo ON ON ON  o ^r; CN 00 o VO 00 CN o cd ro o r-CN i-H  <4-H  r-r- 00 ON in CN 00 00 CN  CCJ  o  W PL,  Q  O H  o  00 vq 00 in 00 ON C so O ro O o ON 00 (SO CL CN >n ON o ro in © ON rt © m oo m CN oo CN 00  r~- r~- r~- ^ oo oo  as ON ON J > ON ON os 5^ as ON g; as as ON  1 o  2 >> <M H  rt  3  j-H  >_j  ro  88  APPENDIX II  TIMES A N D W E A T H E R CONDITIONS D U R I N G POINT C O U N T S U R V E Y S IN T H E G R E A T E R VANCOUVER AREA.  Date 24-Jun-97 25-Jun-97 28-Jun-97 29-Jun-97 30-Jun-97 2-M-97 3-Jul-97 4-Jul-97 5-M-97 7-Jul-97 12-Jul-97 13-M-97 16-Jul-97 20-Aug-97 22-Aug-97 23-Aug-97 24-Aug-97 2-Sep-97 3-Sep-97 4-Sep-97 6-Sep-97 7-Sep-97 8-Sep-97 9-Sep-97 1l-Sep-97 12-Sep-97 20-Sep-97 26-Sep-97 ll-Oct-97 12-Oct-97 l-May-98 2-May-98 3-May-98 6-May-98 10-May-98 12-May-98 16-May-98 18-May-98 19-May-98 20-May-98 24-May-98 29-May-98 30-May-98 3 l-May-98  Time 5:13-9:04 5:23-9:04 5:07:9:24 5:37-7:54 5:51-8:35 5:43-9:30 5:37-8:58 5:45-9:22 5:51-8:35 6:35-9:10 5:40-9:45 6:01-938 6:31-8:37. 6:27-9:32 7:52-9:36 6:47-10:22 7:00-10:30 9:17-9:22 7:40-9:25 7:25-9:01 7:27-8:42 6:15-9:17 7:05:9:45 7:52-9:17 6:43-9:13 7:18-10:28 8:18-9:12 7:37-8:33 9:26-10:10 9:25-10:50 6:28-8:57 6:23-9:13 6:07-9:34 5:19-8:52 6:11-9:50 5:50-9:20 6:36-9:05 6:30-9:00 7:05-8:35 5:58-9:15 7:40-8:52 6:06-9:40 6:00-9:28 5:58-9:33  Temperature -15-18C ~20C ~20C -15-18C ~18C ~16C ~18C ~20C ~22C -16-18C ~17C ~16C -22-24C ~22C ~20C ~18C ~18C ~15C ~16C ~17C -13-15C ~14C ~15C ~15C ~15C ~15C -13-14C ~15C ~10C ~8C -16-18C ~12C ~16C ~14C ~16C ~15C ~16C ~15C ~16C ~17C ~12C ~19C ~18C ~16C  Weather conditions Cloudy, very light rain Sunny and warm Sunny Overcast, light rain possible Sunny, Partial Cloud Sunny Sunny Sunny Sunny A bit cooler, overcast Partly cloudy Partly cloudy Sunny Fair Slightly overcast Cloudy, slight drizzle Cloudy Some clouds Partly cloudy Sunny with part cloud Sunny, Clear, Crisp Crisp & Clear Sunny Cloudy Cloudy & Windy Partly cloudy Cool, Crisp, Sunny Slightly overcast, chance of drizzle Sunny & Brisk Partly cloudy Sunny Mostly sunny, breezy Sunny Sunny and breezy Cloudy Cloudy and cool Sunny Overcast, light rain, breezy Partly cloudy Sunny Cloudy and overcast Sunny Sunny Some cloud, overcast  Survey points 1-30A 31-36A, 1-20B 37-65A 1-16C 17-36C 37-60C 61-77C 1-25D 26-44D 44a-63D, 66A 68-89A 90-111A 66-81D 1-6D, 44-57D 7-16D 17-44D 58-81D 37-49A 53-65A 50-52A, 33-36C 25-32C 1-24C 1-20B 1-12A 37-59C 67-85A 23-30A 13-22A 31-36A 60-77C 60-81D 1-4D 5-28D 30-36A, 1-18B 86-111A 62-85A 1-21A, 23A 19-20B.22A, 24A 29-39D 37-61A 1-8C 55-77C 9-16C, 43-54C 17-42C  89  APPENDIX III  LIST O F BIRD SPECIES ( C A M P B E L L ET AL. 1997), A N D N E S T I N G G U I L D ( E H R L I C H ET AL. 1988), R E C O R D E D A T POINT C O U N T STATIONS IN T H E G R E A T E R V A N C O U V E R A R E A , 1997-1998 (N=65).  4-Letter Code  Common Name  Scientific Name  AMGO AMRO BASW BCCH BHCO BHGR *BOWA *BRBL *BRCR *BTPI' *BTGWA BUSH CAGO *CBCH CEWA •COHA *CORA COYE DEJU EUST FOSP *GCSP GWGU *HETH HOFI HOSP *HOWR *HUVI •KILL •MALL •MERL MEGU •MODO NOCR NOFL OCWA *OSFL PISI RBGU RBNU RCKI REVI RODO *RTHA RUHU *RWBL *SORA SOSP SPTO STJA SWTH *VATH VGSW WCSP *WETA WIFL WIWA WIWR *WEFL' BAEA BLSW GBHE NOHA PIWO SAVS  American Goldfinch American Robin Barn Swallow Black-capped Chickadee Brown-headed Cowbird Black-headed Grosbeak Bohemian Waxwing Brewer's Blackbird Brown Creeper Band-tailed Pigeon Black-throated gray Warbler Bushtit Canada Goose Chestnut-back Chickadee Cedar Waxwing Cooper's Hawk Common Raven Common Yellowthroat Dark-eyed Junco European Starling Fox Sparrow Golden-crowned Sparrow Glaucous-winged Gull Hermit Thrush House Finch House Sparrow House Wren Hutton's Vireo Killdeer Mallard Merlin Mew Gull Mourning Dove Northwestern Crow Northern Flicker Orange-crowned Warbler Olive-sided Flycatcher Pine Siskin Ring-billed Gull Red-breasted Nuthatch Ruby-crowned Kinglet Red-eyed Vireo Rock Dove Red-tailed Hawk Rufous Hummingbird Red-winged Blackbird Sora Song Sparrow Spotted Towhee Steller's Jay Swainson's Thrush Varied Thrush Violet-green Swallow White-crowned Sparrow Western Tanager Willow Flycatcher Wilson's Warbler Winter Wren Western Flycatcher Bald Eagle Black Swift Great Blue Heron Northern Harrier Pileated Woodpecker Savannah Sparrow  Carduelis tristis Turdus migratorius Hirundo rustica Poecile atricapillus Molothrus ater Pheucticus melanocephalus Bombycilla garrulus Euphagus cyanocephalus Certhia americana Columba fasciala Dendroica nigrescens Psaltriparus minimus Branta canadensis Poecile rufescens Bombycilla cedrorum Accipiter cooperii Corvus corax Geothlypis trichas Junco hyemalis Sturnus vulgaris Passerella species Zonotrichia atricapilla Larus glaucescens Catharus guttatus Carpodacus mexicanus Passer domesticus Troglodytes aedon Vireo huttoni Charadrius vociferus Anas platyrhynchos Flaco columbarius Larus canus Zenaida macroura Corvus caurinus Colaptes auratus Vermivora celata Contopus borealis Carduelis pinus Larus delawarensis Sitta canadensis Regulus calendula Vireo olivaceus Columba livia Buteo jamaicensis Selasphorus rufus Agelaius phoeniceus Porzana Carolina Melospiza melodia Pipilo maculatus Cyanocitta stelleri Catharus ustulatus Ixoreus naevius Tachycineta thalassina Zonotrichia leucophrys Piranga ludoviciana Empidonax trailii rVilsonia pusilla Troglodytes troglodytes Empidonax difficilis Haliaeetus leucocephalus Cypseloides niger Ardea herodias Circus cyaneus Dryocopus pileatus Passerculus sandwichensis  2  2  2  3  2  2  3  3  2  2  3  3  3  2  3  2  3  3  3  2  2  4  4  4  4  4  4  GUILD Nesting SB DT B  cv  G DT CT CT CT CT CT DT G CV DT DT L SB G DT G G L G DT B DT DT G G DT L DT CT CV G CT CT L CV CT DT B DT CT R R G G CT SB CT CV SB CT SB G CV DT CT L DT G CV G  *Species excluded from the analysis (n=29). Species censused in: 1997 breeding only (n=2); 1997 fall migration only (n=l 1); 1998 breeding only (n= 10); Species recorded > 50 m from point count stations (n=6). N E S T is nesting guild where SB=shrub, CV=Cavity, G=ground, R=riparian, L=Ledge/Cliff, CT=coniferous tree, DT=deciduous tree, and B=building nesting species. Underlining indicates non-native species 1  2  3  4  APPENDIX IV  90  PHYSICAL CHARACTERISTICS OF TRANSECT POINT C O U N T STATIONS (N=285) IN T H E G R E A T E R V A N C O U V E R A R E A . Transect  Point #  UTM lat  UTM long  Elevation  Address  Urban Aspect Water Category  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 56 57 58 59 60 61  5456873 5456885 5456889 5456898 5456907 5456916 5456922 5456928 5456942 5456942 5456943 5456975 5456958 5456965 5456973 5456979 5456982 5456986 5456985 5457001 5457009 5457017 5456829 5456820 5456808 5456793 5456825 5457031 5457041 5456899 5457062 5457119 5456984 5456896 5456784 5456874 5456866 5456860 5456855 5456845 5456840 5456830 5456812 5456785 5456782 5456770 5456757 5456734 5456730 5456728 5456783 5456878 5456835 5456734 5456720 5456714 5456715 5456685 5456688  489903 489553 489270 489033 488676 488406 488203 487930 487665 487382 487180 486960 486720 486526 486315 486080 485738 485498 484854 484682 484450 484380 484272 484018 483773 483399 483073 482877 482679 482400 482262 482092 481772 481569 481332 490020 490270 490510 490760 491010 491280 491530 491780 492060 492260 492530 493050 493560 493793 494072 494233 494680 494780 494942 495290 495500 495750 496000 496300  35 35 30 40 35 30 30 30 30 30 40 40 40 40 60 65 70 75 75 80 95 95 95 95 100 95 90 95 100 95 90 90 90 90 85 35 40 40 40 35 30 25 20 20 25 30 40 40 30 20 20 25 25 25 45 45 45 45 55  8th and Granville 8th and Burrard Cypress and 8th 8th and Arbutus 8th and Yew 8th and Balsam 2545 W 8th 2745 W 8th 8th and Bayswater 3136 W 8th 3358 W 8th 3584 W 8th 8th at Dunbar 8th at Highbury 3894 W 8th 4026 W 8th 4186 W 8th 1/2 way up hill 8th at Sasamat 8th at Tolmie 8th at Blanca At Bus Loop University Blvd., outside Golf Course University Blvd., Club House University Blvd. University Blvd., outside Golf course University Blvd. University Chapel Acadia and Toronto Allison and Toronto On Allison Gate 1 UBC At East Mall Main Mall and University Blvd At Lower Mall 1443 W 8th at Granville 1245 W 8th 1081 W8th at Spruce 905 W 8th at Laurel 795 W 8th and Willow W 8th at Ash 8th at Yukon 8th at Manitoba 8th at Ontario 8th at Scotia 8th at Prince Edward 8th at St. George 8th and Fraser 8th and St. Catherine 8th and Glen 7th and Keith 1365 E. 7th at McLean 1523 E 8th and Woodland 8th and N. Greenview at Rail greenway 8th past Victoria 8th At Lakewood 8th and Garden 8th and Kamloops 2651 E8th  Urban Urban Urban Urban Urban • Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Edge Edge Edge Edge Edge Edge Edge Fragment Fragment Fragment Fragment Fragment Fragment Fragment Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban park Urban Urban Urban Urban Urban  n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 90 90 90 90 90 90 n.a. n.a. 90 180 180 180 180 180 180 n.a. n.a. n.a. n.a. n.a. n.a. 270 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 270 270 n.a. n.a. n.a. n.a. n.a. 270 270  None None None None None None None None None bath bath None None None None bath None bath None None bath None bath None None None None None None None None None None None None None None Fresh None None None None None None Fresh None None None None None None None None None None None None None None  APPENDIX IV cont.  Transect  91  Point #  UTM lat  UTM long  Elevation  Address  Urban Aspect Water Category  62 63 64 65 66 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111  5456683 5456682 5456684 5456682 5456649 5457303 5457265 5457256 5457246 5457483 5457438 5457277 5457240 5457073 5457020 5456921 5456825 5456740 5456545 5456055 5455592 5455458 5455456 5455344 5455341 5455341 5455340 5455339 5455333 5455337 5456097 5456010 5455761 5455636 5455627 5455630 5455625 5456000 5456000 5456000 5456000 5455972 5455963 5455955 5455944 5455944 5455938 5455935 5455932  496550 496800 497020 497240 497690 498853 499017 499353 499637 499811 500050 500295 500680 501040 501290 501530 501750 502000 502359 503292 503548 503813 504060 504260 504510 504730 505000 505240 505530 505770 506190 506508 506580 506847 507120 507380 507601 507855 508150 508400 508630 508880 509150 509400 509650 509900 510150 510400 510650  60 55 55 50 60 40 40 40 45 50 50 55 55 35 35 30 30 25 25 45 40 40 38 25 25 30 35 40 35 30 45 40 35 40 50 60 80 100 100 100 120 120 140 140 140 145 140 140 140  8th and Kaslo Urban 2936 E 8th at Renfrew Urban Urban 3098 E 8th at Lillouet 3198 E 8th at Windmere Urban Urban At 8th past Rupert 2nd and MacDonald Urban Urban Halifax and Chalet Cabinets Suburban 4305 Halifax Suburban 4430 Halifax Suburban 4560 Brentlawn 4645 Brentlawn Suburban 4806 Brentlawn Suburban Halifax and Woodway Suburban Suburban 54 Broadway 5695 Broadway Suburban Suburban Parkcrest Plaza 6205 Broadway Suburban 6385 Broadway Suburban Broadway and Kensington Suburban Coventry and Collister Suburban Buffalo and Lyndale Suburban Between Philips and Chrisdale - Colleen Suburban Suburban 7511 Colleen 7572 Government Suburban 7750 Government Suburban Suburban Government and Piper Suburban 7986 Government 8121 Government Suburban 8243 Government Suburban AT Dalebright Suburban Suburban East Lake atADT Suburban East Lake and Centaurus Suburban East Lake and turn off Cameron and Kenswick Suburban Cameron and Beaverbrook Suburban Cameron Towers Suburban Walmart and Cameron Suburban Suburban Foster and Ebert Foster almost at Austin Suburban 605 Foster Suburban Suburban Foster and Florence Suburban Foster and Sprice Suburban 827 Foster Foster and Hailey Suburban Foster and Macintosh Suburban Suburban Porter on Foster Suburban 1207 Foster Suburban 1410 Foster Suburban Foster and Berry  270 90 90 90 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 180 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 360 360 360 180 180 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a.  None None None None None None None None None None None None Fresh Fresh None None None None Fresh None None None None None Fresh None None Fresh Fresh None None Fresh None Fresh None None None None Fresh None None None Fresh None None None None None None  1 2 3 4 5 6 7 8 9 10 11  5457359 5457491 5457728 5457936 5458126 5458313 5458442 5458538 5458510 5458413 5458416  481173 481358 481515 481582 481790 481965 482156 482420 482697 482939 483189  70 70 65 60 40 30 10 10 10 10 10  Gate 4, Museum of Anthropology At P3 and Info Sign Newton on NW Marine NW Marine NW Marine Acadia NW Marine NW Marine / Parking Lot NW Marine NW Marine NW Marine  0 0 0 0 0 0 0 0 0 0 0  Marine Marine Marine Fresh Marine Marine Marine Fresh/Marine Fresh/Marine Fresh/Marine Fresh/Marine  Edge Edge Park Park Park Park Park Park Park Park Park  92  A P P E N D I X IV cont.  Transect  Point #  UTM lat  UTM long  Elevation  Address  Urban Aspect Category  Water  12 13 14 15 16 17 18 19 20  5458377 5458325 5458254 5458203 5458061 5457975 5458020 5458069 5457985  483460 483717 483951 484211 484536 484793 485012 485313 485526  10 7 7 7 7 7 7 7 10  NW Marine At Spanish Banks Food Stand NW Marine NW Marine NW Marine /Tolmie NW Marine at Sasamat At Corner NW Marine At Parking Lot/Jericho beach Jericho Park.  Park Park Park Park Park Urban park Urban park Urban park Urban park  0 0 0 0 0 n.a. n.a. n.a. n.a.  Fresh/Marine Fresh/Marine Marine Marine Fresh/Marine Fresh/Marine Fresh/Marine Fresh/Marine Fresh/Marine  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52  5460782 5460542 5460359 5460111 5460005 5459823 5459592 5459488 5459293 5459043 5458792 5458583 5458452 5458620 5459028 5458901 5458842 5458836 5458825 5458818 5458809 5458809 5458846 5458622 5458617 5458605 5458600 5458600 5458917 5458911 5458909 5458907 5458915 5458908 5459058 5458812 5458811 5458808 5458805 5458801 5458795 5458796 5458785 5458783 5458780 5458481 5458481 5458479 5458478 5458502 5458569  490453 490324 490119 489951 489886 490052 490231 490380 490586 490827 490988 491181 491432 491607 492004 492266 492579 492780 493087 493375 493642 493642 494150 494638 494899 495127 495398 494645 495925 496195 496471 496749 496998 497240 497459 497773 498289 498603 498855 499152 499398 499648 499867 500122 500369 500560 500815 501146 501369 501614 501785  30 20 10 10 15 15 15 25 40 35 30 30 30 25 15 15 15 15 20 20 15 15 15 15 15 20 25 30 40 45 50 45 35 35 35 60 80 95 95 90 90 85 75 90 100 100 100 100 100 100 100  Stanley Park Aquarium Parking lot fork On Bridge At Alberni and Chilco Chilco on Barclay Barclay and Denman Barclay at Cadero Barclay at Broughton Barclay and Bute Barclay at Thurlow Burrard and Hornby on Smithe Granville and Seymour On Homer between Smithe and Robson Homer at Georgia in front of library Homer and Cordova Cordova and Abbott Cordova and Columbia Cordova and Main Dunlevy and Cordova Cordova and Princess Cordova and Hawks Cordova and and Campbell Franklin and Glen E Pender and Mclean Commercial and Pender Pender and Salsbury Pender and Semlin Pender and Templeton Franklin and Nanaimo Penticton and Franklin Franklin at Slocan Franklin at Renfrew PNE Parking Lot PNE Park PNE Park 3400 Franklin at Cassiar Franklin and Boundary Albert and Ingleton Albert and MacDonald Albert at Carleton Albert at Madison Albert at Rosser Confederation Community Centre Beta and Albert Albert at Gama Frances and Delta Frances and Springer Frances and Howard Frances at Holdom At Gate to park on Frances Frances/Union Bike Route  Fragment Fragment Edge Edge Edge Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban park Urban park Urban park Urban Urban Urban Urban Urban Urban Urban Urban park Urban park Urban Urban Urban Urban Urban Urban park Urban park  90 180 n.a. 315 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 270 270 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 270 n.a. n.a. n.a. n.a. n.a. n.a. n.a. 270 270 n.a. n.a. n.a. n.a. n.a. 270  Fresh/Marine Fresh/Marine Fresh/Marine Fresh/Marine None None None None None None None None None None None None None None None None None None None None None None None None None None None None None bath None None None None None None None None None None None None None None None Fresh None  APPENDIX IV cont.  Transect  93  Point #  UTM lat  UTM long  Elevation  Address  Urban Aspect Category  Water  53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77  5458517 5458331 5458331 5458333 5458332 5458326 5458050 5458048 5458060 5457937 5457892 5457885 5457940 5458096 5458269 5458240 5458293 5458275 5458458 5458585 5458663 5458588 5458219 5458186 5458547  501991 502202 502458 502728 503008 503434 503704 504058 504368 504631 504879 505129 505376 505140 505134 505377 505618 505847 505959 505851 506115 506347 506745 506433 505560  100 100 100 90 90 100 110 130 190 200 220 240 250 260 280 290 300 300 300 300 300 300 300 300 300  At Kennsington/Burnaby Sec. School Union and Brooklyn Union and Duncan Union at Cliff Union and Calvin Union and Duthie Curtis and Duthie Curtis Cul de Sac Road to SFU Road to SFU Road to SFU Road to SFU at Discovery sign Road to SFU at SFU sign Road to SFU Road to SFU Road to SFU SFU campus Road SFU campus at Central stores Science Portables WAC Bennet Library Strand Hall sign Parking lot C23 Parking lot G24 Parking lot B26 Residences/Daycare  Urban Urban Urban Urban Urban Urban Urban Edge Park Park Park Park Park Park Park Park Park Park Fragment Fragment Fragment Fragment Fragment Fragment Fragment  n.a. n.a. n.a. n.a. n.a. 270 270 270 270 200 200 200 270 270 270 270 270 0 n.a. 180 n.a. n.a. n.a. n.a. n.a.  None None None None None None None Fresh Fresh Fresh Fresh Fresh Fresh Fresh Fresh None Fresh Fresh None None None None None None None  1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 24 25 26 27 28 29 30 31 32 33 34 35 36  5453970 488490 5453978 488182 5453982 487936 5453981 487689 5453981 487427 5453984 487166 5453986 486930 5453998 486630 5453998 486370 5454003 486140 5454007 485749 5454219 485680 5454426 485684 5454720 485688 5455090 485691 5454981 485373 5455222 485138 5455408 485302 5455652 485351 5455892 485271 5456052 485230 5456292 485220 5456300 484720 5456309 484470 5456239 484166 5456243 483906 5456243 493507 5456150 483184 5456073 483016 5455988 482850 5455838 482597 5455730 482373 5455611 482162 5455645 481956 5455848 481972  60 60 60 60 60 60 60 55 50 50 40 50 65 75 80 80 80 80 80 80 80 80 90 90 90 90 90 90 90 90 90 90 90 70 70  2292 W 37th at Yew 2527 W 37th at Larch 2736 W 37th 2915 W 37th 3091 W37th 3294 W 37th 37th and Collingwood 37th at Dunbar 3835 37th at Highbury 3959 W 37th 37th at Camosun 35th at Camosun 33 rd and Camosun 30th and Camosun 28th and Camosun 29th and Kevin Doncaster Way Imperial Along Imperial Along Discovery 5m and boy w/ ball sign 15th and Discovery 15th up from Sasamat 4348 W 15th Along 16th Along 16th Along 16th 16th and Salish Trail 16th and Wyndham Hall Along 16th Along 16th and Thunderbird field East Mall Outside Botanical Garden at GATE Over Botanical Garden tunnel Gate 8  Urban Urban Urban Urban Urban Urban Urban Urban Edge Edge Edge Edge Edge Edge Edge Edge Park Park Park Park Park Edge Edge Edge Park Park Park Park Park Park Edge Edge Edge Edge Edge  n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 90 90 90 90 n.a. 0 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 180 180 180 180 n.a. n.a. n.a. 270 180 270 270  None None None None None None None None None None None bath None None None None None None None None None None None None None None None None None None None None bath bath None  APPENDIX IV cont.  Transect  94  Point #  UTM Iat  UTM long  Elevation  Address  Urban Aspect Category  Water  37 38 39 40 41 42 43 44 44a 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81  5456075 5456726 5456336 5457281 5457386 5457554 5457835 5457444 5453940 5453961 5453951 5453932 5453929 5453914 5453897 5453887 5453870 5453853 5453836 5453821 5453857 5453846 5453826 5453830 5453821 5453760 5453758 5453749 5453790 5453782 5453775 5453776 5453801 5453799 5453857 5453851 5453676 5453485 5453155 5453139 5453148 5452500 5452500 5452500 5452501 5453166  481870 481653 481602 482070 482286 482222 482162 482512 488680 488969 489215 489775 490034 490281 490533 490799 491117 491368 491608 491868 492148 492451 492746 492952 493205 493496 493806 494095 494320 494590 494868 495121 495373 495630 495877 496139 496427 496544 496672 496949 497258 497707 498023 498275 498499 498678  80 85 90 90 95 95 90 85 70 80 80 100 108 100 90 90 90 100 100 100 100 95 90 85 80 85 90 100 95 90 85 85 85 85 85 85 90 90 95 95 105 115 120 120 125 125  Along West Mall Engineering Annex Brock Hall College Highroad and Wesbrook Cres Allison and College Way 1856 Allison 1616 Acadia College Highway and Acadia 37th and Maple 1868 W 37th at Pine 1705 W 37th at Marguerite 1530 W 37th and Granville 1408 W 37th and Cartier 1250 W 37th 1076 W 37th and Osier 928 W 37th Baillie and 37th Manson and 37th 37th and Cambie 280 W 37th and Elizabeth 88 W 37th and Manitoba 82 E 37th and Quebec 37th and Sophia 37th and Prince Edward Bone Yard 758 37th and Chester 966 37th and Sommerville 37th and Ross 1333 E 37th at Colloden 1488 E 37th 1726 E 37th 1928 E 37th 2097 E 37th 2262 E 37th St. Margarets and 37th Chambers and 37th Earles and 38th 40th and Killarney Lancaster and 43 rd McKinnon and 43rd 43 rd and Latta 3441 E45th 3550 E 45th Boundary and 45th Corner Parking lot/Swangard Stadium Central Park  Fragment Fragment Fragment Fragment Fragment Fragment Fragment Edge Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban Urban park Urban park  None None None None None None None None None None None bath bath bath None None None bath None None None None None None None bath bath None None None None None None None None None None None None None None None None None None Fresh  n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 270 n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. n.a. 270 0 n.a.  APPENDIX V  95  A) DISTRIBUTION OF M A X I M U M A V I A N SPECIES RICHNESS FOR POINT COUNT STATIONS A L O N G FOUR ROADSIDE TRANSECTS IN G R E A T E R V A N C O U V E R , 19971998. B) as in A), distribution of Simpson's diversity index (1/Zpi ) where pi is the proportional abundance of each species by station. Simpson's diversity index is relative measure of diversity that weights each species by its relative abundance (Ricklefs 1990). 2  A)  •  12-13  


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