Journal of Ornithology

, Volume 152, Issue 2, pp 461–471

Influence of exurban development on bird species richness and diversity

Authors

    • Appalachian LaboratoryUniversity of Maryland Center for Environmental Science
  • Peter Leimgruber
    • Smithsonian Conservation Biology Institute, National Zoological ParkSmithsonian Institution
    • Smithsonian Conservation Biology Institute, National Zoological ParkSmithsonian Institution
    • Institute of Experimental EcologyUniversity of Ulm
Original Article

DOI: 10.1007/s10336-010-0605-x

Cite this article as:
Suarez-Rubio, M., Leimgruber, P. & Renner, S.C. J Ornithol (2011) 152: 461. doi:10.1007/s10336-010-0605-x

Abstract

Exurban development is an accelerating land use trend in the United States with new housing units emerging in formerly closed forests. Conservation practitioners and planners suspect exurban development alters ecological processes and biodiversity to a considerable larger extent than suspected by inhabitants of exurban development areas, but empirical support for this assertion is lacking. To examine the consequences of exurban development, we studied forest bird communities in exurban development and forests located in and around Shenandoah National Park and Massanutten Mountain in northern Virginia, USA. We conducted point-count surveys for birds three times at 106 sample locations from April to July 2006. We recorded 44 species in total; 30 species were present in both exurban development and forests, 9 species were only found in exurban development, and 5 species only in forest. Bird species composition differed significantly between land-use types based on analysis in a multi-response permutation procedure (MRPP; P < 0.01). Relative bird abundance for forest specialist species changed significantly in exurban development versus forest (t test, P < 0.05). Three species, American Robin Turdus migratorius, Northern Cardinal Cardinalis cardinalis, and Common Grackle Quiscalus quiscula were indicators of exurban development. Indicators of forest were Carolina Chickadee Poecile carolinensis, Eastern Towhee Pipilo erythrophthalmus, and Wood Thrush Hylocichla mustelina. Our study demonstrates that exurban development alters bird community composition and relative abundance of forest specialist species.

Keywords

Community response land coverLand-use changeRural sprawlUrban fringeEastern deciduous forest

Zusammenfassung

Exurbane Landentwicklung (engl.: exurban development) ist ein rasant zunehmender Landnutzungstyp in den Vereinigten Staaten von Amerika, bei dem Wohngebäude in zuvor geschlossenen und naturnahen Wäldern konstruiert werden. Naturschutzorganisationen und Landschaftsplaner vermuten, dass Exurbane Landentwicklung die ökologischen Prozesse und Biodiversität bedeutend verändern, allerdings fehlen bisher empirische Daten dies zu belegen. Um den Einfluss von Exurbaner Landentwicklung zu ermitteln, haben wir Vogelgemeinschaften in Exurbaner Landentwicklung mit Vogelgemeinschaften naturnaher Wälder im und um den Shenandoah Nationalpark und Massanutten Mountain im Norden des US Bundesstaates Virginia untersucht. An 106 Stellen wurden alle Vögel mit Punkt-Stopp-Zählungen insgesamt dreimal in der Zeit von April bis Juli 2006 aufgenommen. Insgesamt wurden 44 Arten erfasst; 30 Arten waren in beiden Habitattypen anzutreffen, während neun Arten nur in exurbaner Landentwicklung und fünf ausschließlich im naturnahen Wald anzutreffen waren. Die Zusammensetzung der Vogelgemeinschaft unterschied sich signifikant zwischen beiden Vogelgesellschaften (MRPP, P < 0.01). Ebenso änderte sich die Relative Abundanz für Waldvögel signifikant zwischen Exurbaner Landentwicklung und naturnahem Wald (t-test, P < 0.05; Exurbane Landentwicklung geht in der Regel aus naturnahem Wald hervor). Drei Arten, Turdus migratorius, Cardinalis cardinalis, und Quiscalus quiscula wurden als Indikatorarten für Exurbane Landentwicklung identifiziert, während Poecile carolinensis, Pipilo erythrophthalmicus, und Hylocichla mustelina als Indikatorarten für naturnahe Wälder gelten können. Die Studie belegt, dass Exurbane Landentwicklung die Vogelgemeinschaften in der Artenzusammensetzung und die relative Abundanz der Waldvogelarten deutlich beeinflusst.

Introduction

Global changes resulting from human activities are among the most pressing environmental challenges of the twenty-first century (Vitousek et al. 1997; Czech et al. 2000; Renner 2005; Chace and Walsh 2006; McKinney 2008). One challenge of global concern is the conversion or modification of natural habitats into urban development. Seventy percent of the world’s population is projected to live in urban areas by 2050 (United Nations 2007). In addition, the number of households have increased more rapidly than the population growth in the last couple of decades (Liu et al. 2003). Rising demands of ever-growing human population is resulting in degrading ecosystem services, seriously threatening biodiversity, and increasing species extinctions (Dale et al. 2000; McKinney 2002; Chace and Walsh 2006).

Urban development is the major cause for federal listing of threatened and endangered species in the United States (Czech et al. 2000). As areas become increasingly urbanized, species richness and abundance are altered (reviewed by, e.g., Chace and Walsh 2006; McKinney 2008; Evans et al. 2009) and community composition changes. Forest-specialized species are lost rapidly (urban avoiders) and other species are gradually replaced by a few ubiquitous urban species (urban adaptors; Luniak 1994; Blair 2001; Marzluff 2001; McKinney 2002), a process called biotic homogenization (McKinney and Lockwood 1999; McKinney 2006).

In the United States, rural landscapes have changed dramatically in recent decades due to the rapid development of private rural lands into exurban development (i.e., land-use change beyond the urban fringe; Hansen et al. 2002; Theobald 2005). Exurban development is the fastest-growing type of development in the United States (Theobald 2001; Hansen et al. 2005; Radeloff et al. 2005a). Exurban development is defined as low-density residential development scattered outside suburbs, cities, and towns (Daniels 1999) but within the limits of commuting range to urban and suburban employment opportunities (Nelson 1992). Exurban development (in contrast to suburban sprawl) has low-density housing units and is often located adjacent to or nearby protected areas because of their attractive recreational and aesthetical amenities (Hansen et al. 2005; Radeloff et al. 2005a, b). However, existing definitions of exurban development are inconsistent in measurements and scale, and include: areas with 1–10 people/ha (Marzluff et al. 2001), 6–25 housing units/km2 (0.06–0.25 units/ha or 0.08–0.17 ha/unit; Hansen et al. 2005), or 0.68–16.18 ha/unit (0.06–1.47 units/ha; Theobald 2005). Forests with low density of exurban development typically retain a closed canopy or at least canopy that appears many times to be closed if observed with medium-resolution satellite sensors. However, these forests experience significant changes in the understory resulting from maintained trails and lawns, and the landscape becomes increasingly fragmented by paved roads (Hansen et al. 2005). Recent studies have demonstrated that one-tenth to a quarter of the area (differing numbers result from different definitions and classifications, see above) and one-third of the housing units in the United States are already in exurban development areas where wild-land vegetation is interspersed by housing (USDA and USDI 2001). Exurban development also seems to be increasing more rapidly in areas of high conservation value (Hansen et al. 2002; Radeloff et al. 2005a, b), i.e. inside forests (Radeloff et al. 2005a, b), next to protected lands (Rasker and Hansen 2000), and along lakefronts (Radeloff et al. 2001; Schnaiberg et al. 2002).

In general, many United States citizens consider exurban development a valuable alternative to urban sprawl, because exurban development is considered to have less impact on biodiversity than urban sprawl (J. Buff, personal communication). However, the general view among conservationists is that exurban development alters ecological processes and biodiversity to a greater extent than forestry and agriculture (Marzluff 2001), and bird communities are suspected to be particularly sensitive (Hansen et al. 2005). Exurban development may also have a profound influence on adjacent protected lands that may include the loss of native species, changes in disturbance regimes, and the spread of invasive organisms (Hansen et al. 2002; Theobald 2004). Other concerns regarding urban sprawl and exurban development in particular are patterns of resource consumption, effects on air and water quality, and pollution (Dale et al. 2005; Hansen et al. 2005).

Exurban development affects large geographic areas (Brown et al. 2005), but the effects of this development on wildlife and biodiversity are little known especially in the eastern deciduous forest region where exurbanized areas have increased since the 1950s (Brown et al. 2005; Theobald 2005). Birds’ response to urbanization has in general been well documented. Several studies have focused on bird diversity in and around cities (e.g., Emlen 1974; Beissinger and Osborne 1982; Fernández-Juricic and Jokimäki 2001; Palomino and Carrascal 2005) and along a rural to urban gradient (e.g., Blair 1996; Clergeau et al. 1998; Crooks et al. 2004; Blair and Johnson 2008; Ortega-Alvarez and MacGregor-Fors 2009). However, those studies do not clearly distinguish exurban areas from suburbia. Some studies have addressed the effects of exurban development on birds in the western US (e.g., Odell and Knight 2001; Fraterrigo and Wiens 2005; Bock et al. 2008; Merenlender et al. 2009) and a few have addressed these effects in eastern deciduous forest (Friesen et al. 1995; Nilon et al. 1995; Kluza et al. 2000). These studies provide general insights into the negative consequences of exurban development on bird communities, but little is known about exurban bird communities in eastern deciduous forest, whether exurban bird communities have a composition similar to communities of forested areas in this ecoregion, and whether modification of forest understory, typical in these areas, affects forest specialist species and ground-breeding birds.

The aim of our study is to assess the impact of exurban development on bird communities in northern Virginia, eastern United States, and contribute empirical data and analysis of exurban development on wildlife. In particular, we addressed whether species composition differs between exurban areas and forests, and if forest specialists and near-ground breeding birds decrease in exurban areas. Our presented results are relevant for conservation planning and contribute to the understanding of the impact of less intensive land-use activities (i.e., exurban development) on species distribution and species richness.

Methods

Study area

Our study area covered six sites around Front Royal, Warren County, northern Virginia (Fig. 1). Front Royal is the northern gateway to the Shenandoah National Park. It is linked to the Washington, DC, metropolitan area via Interstate 66 (about 110 km). The area experienced rapid exurban expansion between 2000 and 2006, and the population increased by almost 13% (U.S. Census Bureau 2007).
https://static-content.springer.com/image/art%3A10.1007%2Fs10336-010-0605-x/MediaObjects/10336_2010_605_Fig1_HTML.gif
Fig. 1

Point count locations and land cover of Warren County, northern Virginia (source land cover: Huang et al. 2001; Homer et al. 2004). Exurban development localities (black dots) included Apple Mountain, High Knob, and Harmony Hallow; forest localities (yellow dots) are Massanutten Mountain, Shenandoah National Park, and the Smithsonian Conservation and Research Center (CRC)

Three of our study sites were used as a reference for forests and were located in the Shenandoah National Park, on Massanutten Mountain (part of George Washington and Jefferson National Forests), and in Poesy Hollow at the Smithsonian Institution’s Conservation and Research Center (Fig. 1). The other three sites were under exurban development and were placed in local communities including High Knob, Harmony Hallow, and Apple Mountain. All sites had similar vegetation characteristics and elevation (400–500 m a.s.l.). Exurban development in the study area was located in forests. The mean housing density in our exurban development sites was 0.23, 0.25, and 0.77 ha/units in High Knob, Apple Mountain, and Harmony Hollow, respectively.

Bird surveys

Six major localities were selected based on whether exurban development or forest was present. Then, we established 53 point-count localities in each of the two land-use types, exurban development and forest (106 point counts in total; Fig. 1). Point-count localities were randomly placed at least 150 m apart (maximum distance: 19.25 km). To increase comparability of forests and exurban development areas, we purposely choose our point-count localities on unpaved, narrow roads in exurban development and trails in forested areas. In addition, we choose exurban development point-count localities only with a well-retained canopy cover. Thus, differences in point-count localities between the two land-use types were minimal and should not have biased our results.

Two observers collected abundance data for birds at each sample location using fixed-radius point counts (50 m radius) from late April to July 2006, the peak of the breeding season for passerines at our study sites (Kingery 1998). We assumed each singing bird represented an adult and potentially as breeding during our surveying period (same year birds are rarely displaying if at all before the post-breeding period). Surveys were performed from sunrise until 1000 hours on days without rain and when wind was less than 15 km/h. Each sample location was surveyed three times during the surveying period for 5 min to maximize count efficiency (Petit et al. 1995), and efforts were made to avoid double-counting individuals moving among points (Bibby et al. 1992). We visited a maximum of 15 points per day, and changed sampling sequence of point counts every visit. We used the maximum detection per all three repetitions of each point-count locality as relative abundance.

Data analysis

We determined detectability of birds using the ‘single-season occupancy model’ in the program PRESENCE 2.4 (Hines 2006; with assumptions by Mackenzie et al. 2002, 2003). We determined detectability for all species in all 106 sample locations for the first survey plus two consequent repetitions (i.e. each sample location was visited three times). Patch occupancy was high for 37 species (out of 44 in total), i.e. these species were detected in at least 60% of all the point-count localities (Table 1).
Table 1

Species detected in forest and exurban development in Warren County, northern Virginia, USA, including subsets used for analysis of impact of exurban development on forest species and understory breeders; sorted by maximum total detections

English name

Scientific name

Forest specialista

Understory breederb

Forestc (t = 0)

Exurban developmentc (t = 1)

Total detectionsc

Probability of detection (P)d

Proportions of sites occupied (ψ)e

Northern Cardinal

Cardinalis cardinalis

 

X

50

117

167

0.482

0.855

Wood Thrush

Hylocichla mustelina

 

X

81

64

145

0.473

0.818

Carolina Chickadee

Poecile carolinensis

  

84

59

143

0.409

0.891

Tufted Titmouse

Baeolophus bicolor

  

52

60

112

0.428

0.801

Eastern Wood-Pewee

Contopus virens

  

54

45

99

0.296

1.000

American Crow

Corvus brachyrhynchos

  

39

59

98

0.387

0.773

Blue Jay

Cyanocitta cristata

  

36

54

90

0.362

0.765

Swainson’s Thrush

Catharus ustulatus

X

X

49

39

88

0.239

0.775

American Robin

Turdus migratorius

  

18

54

72

0.271

0.801

Scarlet Tanager

Piranga olivacea

  

35

29

64

0.201

1.000

Red-eyed Vireo

Vireo olivaceus

 

X

28

24

52

0.219

0.918

Eastern Towhee

Pipilo erythrophthalmus

 

X

25

11

36

0.210

0.688

Mourning Dove

Zenaida macroura

  

17

19

36

0.146

0.775

Hairy Woodpecker

Picoides villosus

  

17

17

34

0.177

0.746

Worm-eating Warbler

Helmitheros vermivorus

X

X

19

10

29

0.104

0.911

Common Grackle

Quiscalus quiscula

  

2

25

27

0.177

0.213

Pileated Woodpecker

Dryocopus pileatus

X

 

15

9

24

0.085

1.000

Cerulean Warbler

Dendroica cerulea

X

 

13

10

23

0.072

1.000

American Redstart

Setophaga ruticilla

  

9

8

17

0.310

0.183

Sharp-shinned Hawk

Accipiter striatus

X

 

11

5

16

0.354

0.891

Acadian Flycatcher

Empidonax virescens

X

X

5

10

15

0.109

0.547

Downy Woodpecker

Picoides pubescens

  

10

4

14

0.123

0.436

Hooded Warbler

Wilsonia citrina

X

X

10

4

14

0.035

1.000

Eastern Phoebe

Sayornis phoebe

  

7

4

11

0.079

0.518

Red-bellied Woodpecker

Melanerpes carolinus

X

 

3

4

7

0.047

1.000

Indigo Bunting

Passerina cyanea

 

X

3

3

6

0.242

0.117

Northern Flicker

Colaptes auratus

  

1

5

6

0.016

1.000

Alder Flycatcher

Empidonax alnorum

 

X

0

5

5

0.009

1.000

Canada Goosef

Branta canadensis

  

2

3

5

0.009

1.000

Common Raven

Corvus corax

  

1

2

3

0.009

1.000

Ovenbird

Seiurus aurocapilla

X

X

2

1

3

0.006

1.000

Blue-gray Gnatcatcher

Polioptila caerulea

  

0

2

2

0.006

1.000

Carolina Wren

Thryothorus ludovicianus

  

2

0

2

0.177

0.639

Gray Catbird

Dumetella carolinensis

 

X

0

2

2

0.006

1.000

Northern Mockingbird

Mimus polyglottos

 

X

0

2

2

0.006

1.000

Tree Swallowf

Tachycineta bicolor

  

0

2

2

0.003

1.000

American Goldfinch

Carduelis tristis

 

X

0

1

1

0.006

1.000

Brown Thrasher

Toxostoma rufum

 

X

0

1

1

0.003

1.000

Brown-headed Cowbird

Molothrus ater

  

1

0

1

0.382

0.025

European Starling

Sturnus vulgaris

  

0

1

1

0.003

1.000

House Sparrow

Passer domesticus

  

0

1

1

0.003

1.000

Kentucky Warbler

Oporornis formosus

X

X

1

0

1

0.006

1.000

Red-winged Blackbird

Agelaius phoeniceus

 

X

1

0

1

0.003

1.000

Summer Tanager

Piranga rubra

 

X

1

0

1

0.003

1.000

Grand Total

44

10

18

704

775

1,479

  

aHabitat mainly forest and rarely if at all found outside closed forest; data adapted from Poole (2005)

bBird breeding on or near ground below 5 m; data adapted from Poole (2005)

cRelative abundance derived from maximum detection per point count

dProbabilities of detection (P of PRESENCE results protocol) for each species, calculated with PRESENCE 2.4

eProportion of sites occupied (106 sites in total, default model parameters), calculated with PRESENCE 2.4; ψ = 1.000 indicates that species S probably occupied all sites

fOver-flying

Point-count data were used to calculate species diversity (EstimateS 8.0; Colwell 2006). We used these diversity statistics as indices and to establish whether or not our data sampling effort was sufficient. For species richness, species accumulation curves of the Abundance-based Estimator of Species Richness (ACE) was most suitable; the requirement to apply ACE and not other estimators of species richness are medium species numbers and medium to high sample sizes (Colwell 2006).

We analyzed differences in bird composition between the two land-use types using a multi-response permutation procedure (MRPP; McCune and Grace 2002). MRPP is a nonparametric procedure and we used it to test the hypothesis of no difference between the two land-use types. Prior to MRPP analyses, we log-transformed relative abundance values to compress high values, spread low values, and express the values as orders of magnitude. We searched for outliers among sample locations, using Sørensen (Bray–Curtis) distance with two standard deviations from the mean as cutoff value (McCune and Grace 2002). To identify community composition differences, we also used Sørensen (Bray–Curtis) distance measure and a natural group weighting factor nini (where ni is the number of sample locations in each group). The MRPP test statistic (T) describes the separation between the groups: stronger separations have more negative values of T, and the chance-corrected within-group agreement (A) describes the within-group homogeneity, compared to the random expectation. If the items are identical within groups, then A = 1. If the groups are heterogeneous and have less agreement than expected by chance, then A < 0 (McCune and Grace 2002).

To assess the impact of exurban development on relative abundance of forest birds, we classified birds into two groups and analyzed subsets of the entire bird community. The classification for both subsets was based on Poole’s (2005) breeding habitat and nest placement information and classification. The first group (forest specialists) included birds occurring mainly in closed forest and breeding rarely if ever outside forests (class values are listed in Table 1). The second group (understory breeders) included birds breeding mainly in understory below 5 m and on the ground (values listed in Table 1). Both groups should be affected especially by exurban development because the structure of the forests changed and the understory is diminished in exurban development. We compared relative abundance for both groups (subsets) of species with a t test (relative abundance of birds in exurban development vs forest), because exurban development can be regarded as change from forest (t = 0) to development (t = 1).

We conducted indicator species analysis following the method of Dufrêne and Legendre (1997) to find representative species for the different land-use types. The procedure combines a species’ relative abundance with its relative frequency of occurrence in each sample location to provide an indicator value for the particular species for each land-use type. The indicator value of a given species is independent of other species values. The index ranges from 0 to 100, the latter indicating all individuals of a species in a single land-use type and in all sample locations of that land-use type. We used the Monte Carlo technique with 1,000 permutations to test for significance of indicator values (McCune and Mefford 1999). The outlier analysis, MRPP, and the indicator species analysis were performed using PC-ORD 5 for Windows (McCune and Grace 2002).

Results

Species richness

We recorded 44 species in 1,479 detections; 30 of these species were present in both land-use types; we found 9 species only in exurban development and 5 only in forest (Table 1). Species with the highest relative abundance were Northern Cardinal, Wood Thrush, Carolina Chickadee, and Tufted Titmouse (Table 1). We detected Brown-headed Cowbird, Carolina Wren, Kentucky Warbler, Red-winged Blackbird, and Summer Tanager only in forest, and American Goldfinch, Brown Thrasher, European Starling, House Sparrow, Blue-gray Gnatcatcher, Gray Catbird, Northern Mockingbird, Tree Swallow, and Alder Flycatcher only in exurban development.

Accumulative species estimates (i.e., ACE; Fig. 2) indicated that our sampling effort was appropriate, capturing most species in the system. Similarly, standard deviations were low (>0.01 for the maximum sample) for all the accumulated diversity indices we calculated (Table 2; Fig. 2). This demonstrated that our surveys effectively sampled the bird community and that there was only a small chance that we missed additional species.
https://static-content.springer.com/image/art%3A10.1007%2Fs10336-010-0605-x/MediaObjects/10336_2010_605_Fig2_HTML.gif
Fig. 2

Accumulative species-sample curve of estimated species numbers of the Abundance-based Estimator of Species Richness (ACE) for forest, exurban development, and 106 samples combined (the latter with ±SD error bars) of Warren County, northern Virginia, USA

Table 2

Observed species richness and diversity of bird communities in Warren County, northern Virginia, USA; statistics calculated applying Colwell (2006)

Indicator

Variable name as in Colwell (2006)

Overall

Exurban development

Forest

Sample points

106

53

53

Observed species (S)

44

39

35

Detections (n)

1,479

775

704

Sobs

Sobs (Mao Tau)

44

39

35

Individuals

Individuals (computed)

1,470

772

704

Jackknife (±SD)

Jack 1 mean

54.83 (2.39)

47.67 (1.76)

42.33 (1.76)

Bootstrap (±SD)

Bootstrap mean

48.76 (0.00)

43.22 (0.00)

38.41 (0.00)

Michaelis–Menten (±SD)

MMRuns mean

48.52 (48.43)

49.47 (49.75)

41.92 (41.08)

α of log-series (±SD)

Alpha mean

8.54 (0.58)

8.67 (0.69)

7.74 (0.65)

Shannon–Wiener (±SD)

Shannon mean

3.03 (0.01)

2.98 (0.00)

2.99 (0.00)

Simpson (±SD)

Simpson mean

15.98 (0.00)

14.75 (0.00)

15.94 (0.00)

Differences on bird composition between forest and exurban development

Three point-count localities were outliers and excluded in further analyses because the weather conditions in these sample locations turned too windy to appropriately detect all the species. In addition, their average Sørensen (Bray–Curtis) distances were high (0.91, 0.91, and 0.82) compared to all other locations (mean ± SD: 0.62 ± 0.07). Bird species composition differed significantly between the two land-use types, i.e. forest and exurban development (MRPP: T = −8.51, A = 0.02, P < 0.01).

Impact of exurban development on relative abundance of forest birds

Exurban development significantly affected relative abundance of forest-specialist species (as classified in Table 1). Ten forest species had significantly different relative abundance in exurban development versus forest (t = −2.437, df = 9, P = 0.038). Eight of these ten forest species showed lower relative abundance in exurban development (vs forest; e.g., Worm-eating Warbler, Hooded Warbler), while only two have slightly increasing relative abundance (Acadian Flycatcher and Red-bellied Woodpecker; Table 1). Relative abundance of 18 understory-breeding birds was not significantly different between the two land-use types (t = 0.255, df = 17, P = 0.8).

Indicator species

Three species were identified as indicators of forest (Carolina Chickadee, Eastern Towhee, and Wood Thrush; Table 3), and three species as indicators of exurban development (American Robin, Northern Cardinal, and Common Grackle; Table 3).
Table 3

Species identified as indicators of exurban development and forest in Warren County, northern Virginia. All indicator values were significantly larger than random values based on Monte Carlo tests (1,000 permutations, P < 0.05)

Species by land-use type

Indicator value

Observed

Random

±SD

Exurban development

 American Robin

44.7

27.0

3.45

 Northern Cardinal

61.7

40.8

3.22

 Common Grackle

13.5

7.2

2.29

Forest

 Carolina Chickadee

45.7

38.2

3.18

 Eastern Towhee

26.1

18.0

3.27

 Wood Thrush

42.6

36.5

3.28

Discussion

Our results show that exurban development significantly alters bird community composition and changes relative abundance of forest specialists (as classified in Table 1). The number of species in our forests (35; Table 2) is comparable to other studies in forests of the eastern United States (Holmes and Sturges 1975; Kluza et al. 2000). Although the species richness between forest and exurban development in our study sites were similar (35 vs 39 species), the species composition differed between land-use types. Species commonly associated with humans (i.e., urban adaptors) were only present in exurban areas (e.g., European Starling, House Sparrow, Northern Mockingbird), whereas forest-specialist species were only found in forests (Kentucky Warbler, Summer Tanager). These findings show that exurban development is changing bird community composition, even though canopy cover is largely retained. Previous studies have also demonstrated changes in species composition in exurban areas of Missouri (Nilon et al. 1995), Oklahoma (Engle et al. 1999), Colorado (e.g., Odell and Knight 2001; Fraterrigo and Wiens 2005), and California (Merenlender et al. 2009), where (as a general trend) the numbers of urban adapter species increased while urban avoiders decreased. For instance, low human densities (<0.07 people/ha) favored open-habitat species, while higher human density areas (>0.12 people/ha) favored species associated with human activities (Engle et al. 1999). Most notably, forest and edge species were replaced by species associated with human development, especially in higher housing density rural areas. Changes in bird species assemblages have frequently been linked to diminished habitat patch size (e.g., Tilghman 1987; McIntyre 1995; Park and Lee 2000; Fernández-Juricic and Jokimäki 2001; Suarez-Rubio and Thomlinson 2009) or modification of vegetation structure (e.g., Germaine et al. 1998; Miller et al. 2003; Renner et al. 2006; Croci et al. 2008). Alterations in habitat structure can have a marked impact on breeding species either by providing well-hidden nesting sites and decreasing predation by visually oriented predators, or increasing predation through reduced ability of parents to detect predators and defend the nest (Jokimäki and Huhta 2000; Jokimäki et al. 2005; Weidinger 2002; Evans et al. 2009). Therefore, housing development is detrimental for natural bird communities (Lepczyk et al. 2008), even at low housing densities.

Our results reveal how groups of species respond differently to exurban development. Relative abundance of forest-specialist species was significantly different in forests versus exurban development, but that was not the case for relative abundance of understory breeders. Changes in relative abundance of forest specialists support the general finding of alterations in species composition as discussed above. However, we also expected a significant difference in relative abundance of near-ground breeding birds because ground vegetation within exurban development areas is often removed for housing construction, while canopy trees need to be retained (High Knob Owners Association, personal communication). Also, the adverse impact of feral cats and domestic dogs in these areas is well documented (Coleman et al. 1997; Odell and Knight 2001; Gillies and Clout 2003; Lenth et al. 2006). The lack of detectable effect for understory breeders may represent an extinction debt which may lead to the loss of understory breeders over time (Tilman et al. 2002). However, the bird community needs to be monitored over longer time periods to assess this assertion.

Bird communities are frequently used as indicators of environmental quality (Mayer and Cameron 2003; Lepczyk et al. 2008) and are considered a useful proxy for assessing the human influence on biodiversity (Garson et al. 2002; Lepczyk et al. 2008; Evans et al. 2009). Our results identified species as indicators of both, exurban development and forest. For example, the American Robin typically prefers open areas and frequents forests and woodlands with lighter understory (Sallabanks and James 1999). Other indicators of exurban development were Northern Cardinal and Common Grackle. These results suggest that exurban development is being perceived by birds more as an open habitat although the canopy is maintained. Species indicative of forest were Carolina Chickadee, Eastern Towhee, and Wood Thrush. The Wood Thrush inhabits the interior of deciduous and mixed forests, especially well developed with moderate sub-canopy and shrub density (Roth et al. 1996). Exurban development is jeopardizing Wood Thrush populations by altering the forest structure, as their relative abundance decreases. Wood Thrushes responded dramatically to development in Waterloo, Ontario, where they practically disappeared from residential areas (Friesen et al. 1995).

Conclusion

Our fine-scale study of bird communities demonstrates that exurban development in the eastern deciduous forest region is changing bird community composition and affecting forest-specialist species. At broader spatial scales, such alterations will have significant impacts on bird species persistence, entire bird populations, and biodiversity. Our results hence show that, despite the general perception of exurban development as environmentally preferable to urban sprawl, this is not necessarily correct. Given that exurban development is differentially impacting bird species, it would be valuable to keep monitoring these bird communities. In addition, educating residents of these exurban communities about the impacts of exurban development might encourage the management of these areas to provide suitable habitat for native bird species.

Acknowledgments

We would like to thank Todd Lookingbill and two unknown reviewers for providing critical comments on previous versions of the manuscript. Our special thanks goes to the High Knob Owners Association (especially Kay Clement and Marcy McCann) for organizing access to the facility and contact to the citizens, who always welcomed us and allowed data collection in their community, talked about the regional and High Knob area history, and provided coffee in the ‘Bagels-Bunch group’ discussions. Jennifer Buff facilitated contacts to Warren County administration for data on exurban development in the county. Gary Somers of The National Park Service, Shenandoah National Park (permit #: SHEN-2006-SCI-0009), and James T Smalls (Forest Ranger of Massanutten Mountain/George Washington and Jefferson National Forests) provided further valuable information and permits to establish the forest point counts for our study. The study was supported with funds from the Smithsonian’s Conservation and Research Center GIS Lab.

Copyright information

© Dt. Ornithologen-Gesellschaft e.V. 2010