Title: Sprawl and Fragmentation: How Much Development Deters Native Bird Species? Mary Ann Cunningham, Jake Hoffman Vassar College
1Sprawl and Fragmentation How Much Development
Deters Native Bird Species? Mary Ann
Cunningham, Jake Hoffman Vassar College
Built Tax Parcels Dutchess County NY
2Questions 1. Are native, woodland bird species
inhibited by urbanization on the urban
fringe? 2. Which gradients of urbanization are
most important in predicting these
species? (road density, tree cover, built
density, proximate conditions) 3. Can
easily-calculated measures, e.g. focal
statistics, support multiple-scale
investigations? (as alternative to
digitizing/interpreting landcover data)
3Study Area Poughkeepsie, NY
Study area approx 14 x 6 km (85 km2)
4Representative of the peri-urban region
5Methods Point counts (5 min, 100 m
radius) Distributed at 500 m intervals on E-W
transects Explanatory variables Proximate
habitat cover (grass, shrub, trees, pavement)
feeder (p/a), tree hights Calculated road area
(moving window/focal statistics over study area)
6Euclidean distance to nearest road (distance
function)
7Average built parcel size (? built
density) Interpolated surface from tax parcel
centroids using parcel size as value (natural
neighbor interpolation)
8Digitized Tree cover ? trees at 4 scales
9Digitized Tree cover ? trees at 4 scales
300 m 600 m 900 m 1200 m
10Results 84 observations 34 species, using P/A
measures, most lt 10 occurrences
Road area, 300 m, 600 m, 900 m radius
0 50 100
0 50 100
0 50 100
Average Parcel Size (ha)
Distance to Roads (m)
0 800 1600
0 400 800
11ANOVA indicates difference in variable value
between presence and absence observations
12Results Bird counts 84 observations 34
species, using P/A measures, most lt 10
occurrences
13Results Bird counts 84 observations 34
species, using P/A measures, most lt 10
occurrences
ANOVA indicates difference in variable value
between presence and absence observations
14Significant Results (small sample sizes make
results provisional!) (using Chi2 or backwards
anova)
Species N Temp Feeder Grass Trees Shrubs Pavemt PctRoads (300 m) PctRoads (600 m) PctRoads (900 m) Distance to roads Parcel size
American Crow 56 .021 .008- .038- .015 .0006 .0016
Black-capped Chickadee 44
White-breasted Nuthatch 20 .016- .003
European Starling 14 .07 lt.0001
Tufted Titmouse 14 .086- .048-
Northern Cardinal 13 .005 .013 .102 .105 .033-
Blue Jay 12 .022
Dark-eyed Junco 12 .038- .010 .005 .011 .072- .041-
Downy Woodpecker 12 .012 .015-
Mourning Dove 10 .003 .020 .105 .086- .098-
White-throated Sparrow 9 .031 .086
Herring Gull 8 .002- .038- .003
Red-tailed Hawk 8 .080- .086- .086-
House sparrow 6
Rock Dove (Pigeon) 6 .023- .053 lt.0001 lt.0001 .0002 .055-
American Goldfinch 5
Northern Mockingbird 4
Pileated Woodpecker 4 .022 .078
Red-bellied Woodpecker 4 .092
15Significant Results (small sample sizes make
results provisional!) (using Chi2 or backwards
anova)
Species N PctTrees (300 m) PctTrees (600 m) PctTrees (900 m) PctTrees (1200m)
American Crow 56 lt.0001 - lt.0001 - lt.0001 - lt.0001 -
Black-capped Chickadee 44
White-breasted Nuthatch 20
European Starling 14 .103 - .-17 - .015 - .092 -
Tufted Titmouse 14 .044 .067 .072 .092
Northern Cardinal 13
Blue Jay 12
Dark-eyed Junco 12 .041 - .037 - .076 - .053 -
Downy Woodpecker 12
Mourning Dove 10 .049 - .053 - .168 - .189 -
White-throated Sparrow 9 .090 - .189 -
Herring Gull 8 .010 - .025 - .004 - .004 -
Red-tailed Hawk 8
House sparrow 6 .052 - .039 - .022 - .009 -
Rock Dove (Pigeon) 6 .001 - .001 - .001 - .001 -
American Goldfinch 5
Northern Mockingbird 4
Pileated Woodpecker 4 .048 .074
Red-bellied Woodpecker 4
16Urban vs. Suburban groups
Species Temp Feeder Grass Trees Shrubs Pavement PctRoads (300 m) PctRoads (600 m) PctRoads (900 m) Distance to roads Parcel size
Urban .0001- .002 - .0001 .0003 .0001 .0001 .055- .017-
Suburban .069-
Species PctTrees (300 m) PctTrees (600 m) PctTrees (900 m) PctTrees (1200m)
Urban lt.0001 - lt.0001 - lt.0001 - lt.0001 -
Suburban
17(No Transcript)
18Scale patterns
Species N PctTrees (300 m) PctTrees (600 m) PctTrees (900 m) PctTrees (1200m)
American Crow 56 lt.0001 - lt.0001 - lt.0001 - lt.0001 -
Black-capped Chickadee 44
White-breasted Nuthatch 20
European Starling 14 .103 - .017 - .015 - .092 -
Tufted Titmouse 14 .044 .067 .072 .092
Northern Cardinal 13
Blue Jay 12
Dark-eyed Junco 12 .041 - .037 - .076 - .053 -
Downy Woodpecker 12
Mourning Dove 10 .049 - .053 - .168 - .189 -
White-throated Sparrow 9
Herring Gull 8 .010 - .025 - .004 - .004 -
Red-tailed Hawk 8
House sparrow 6 .052 - .039 - .022 - .009 -
Rock Dove (Pigeon) 6 .001 - .001 - .001 - .001 -
American Goldfinch 5
Northern Mockingbird 4
Pileated Woodpecker 4 .048 .074
Red-bellied Woodpecker 4
19Ubiquitous species
20Feeder species
21Woodland species
22Urban species
23(No Transcript)
24Christmas Bird Count results
Number in 2004
25Conclusions 1. Native species, woodland species
show relatively little effect on the range of
urbanization gradients in the study area
Urban species, including invasives, show
relatively large effects ----------- 2.
Percentage road area easy to calculate, useful
factor Distance not useful in suburban
contexts (gradients too small) Parcel size
easy to calculate but less useful than road
area All of these are easier to assess
accurately than LULC or digitized tree
cover. BUT tree cover is still generally more
influential ----------- 3. Despite small
samples, trends were consistent over scales
Focal statistics is a useful, easy approach to
multiple-scale studies
26(No Transcript)