Microhabitat Factors Associated with Occupancy of Songbirds in Suburban Forest Fragments in the Eastern United States.
Forest cover in the eastern United States has been fragmented by suburban development and agriculture (Zhou et al., 2011), creating a landscape of remnant forest patches surrounded by a matrix of varying land uses (Saunders et al., 1991). The structural characteristics of remnant forest patches are different from those of nonfragmented patches (Saunders et al., 1991; Harper et al., 2005). The impacts of fragmentation on forest structure can extend up to 75 m into a patch and can include decreased canopy coverage, increased understory tree density, increased shrub abundance, and increased herbaceous cover (Harper et al., 2005). In addition to these impacts, suburban development at forest edges may decrease understory vegetation through increased mowing or habitat alteration (De Chant et al., 2010).
Fragmentation, which results in an increase in the amount of edge, typically causes remnant forest patches to have high levels of invasion by nonnative plants. Densities of nonnative invasive plants are often highest along forest edges (Yates and Levia, 2004; Harper et al., 2005), and this effect may be exacerbated along edges of suburban habitat that can act as sources of invasive plants via horticultural plantings (Lussier et al., 2006). As nonnative plants have been shown to be unpalatable to many invertebrate herbivores, a dominance of nonnative plants could cause a decrease in native arthropod biomass (Tallamy et al., 2010). Therefore, the replacement of native plants with nonnative invasives may trigger a trophic cascade (Paine, 1980) and lead to reduced availability of invertebrate prey for higher trophic levels in forest fragments (Cripps et al., 2006; Gerber et al., 2008; Wu et al., 2009; Yoshioka et al., 2010).
Many Eastern forest bird species rely on remnant forest patches as the only breeding habitat available in the landscape (Tilghman, 1987; Melles et al., 2003). Populations of many Eastern forest birds have declined steadily over the last 40 y (Sauer et al., 2014), and efforts to reverse these trends likely need to include management of remnant forest patches as breeding habitat for forest birds. However, effective management of remnant forest patches requires an understanding of how birds occurring in suburban forest fragments are affected by microhabitat factors such as forest structure, nonnative plants, and invertebrate biomass. While there are known differences in microhabitat factors in remnant forest patches such as forest structure (Saunders et al., 1991; Harper et al., 2005), nonnative plant densities (Yates and Levi a, 2004; Harper et al., 2005), and invertebrate biomass (Burke and Nol, 1998; Zanette et al., 2000; Doyle 2008) it is not known how these factors influence occupancy of birds in suburban forest fragments. Therefore, our objective was to examine the effects of these factors on nine species of common songbirds. Results may be used to guide habitat management and restoration of remnant forest fragments.
Our study was conducted in Delaware and Maryland, U.SA. in upland and riparian mature forest patches located in White Clay Creek State Park (39[degrees]43T9.40", 75[degrees]45'41.77"), Fair Hill Natural Resources Management Area (39[degrees]41'51.35", 75[degrees]50'08.82"), University of Delaware Ecology Woods (39[degrees]39'49.09", 75[degrees]44'47.73"), Iron Hill Park (39[degrees]37'58.7742", 75[degrees]45' 25.3362"), St. Andrew's School (39[degrees]25'56.67", 75[degrees]41T3.56"), Mount Cuba Center (39[degrees]47'18.19", 75[degrees]38'56.97"), Red Clay Creek within Brandywine State Park (39[degrees]48'21.57", 75[degrees]34'22.66"), and Ashland Nature Center (39[degrees]47'50.99", 75[degrees]39'32.54") (Fig. 1). Sampling plots were randomly placed within these seven natural area complexes using the Hawths Tools extension for ArcGIS 9.2 (Beyer, 2004). The tool was used to generate the maximum possible number of sampling locations within each natural area that were located in mature forest 25 m from a forest edge and separated from another point by at least 250 m. There were 30 sampling plots within White Clay Creek State Park, 39 within Fair Hill Natural Resources Management Area, four within University of Delaware Ecology* Woods, five at St. Andrew's School, five at Iron Hill Park, and 15 points within the complex consisting of Mount Cuba Center, Red Clay Creek State Park, and Ashland Nature Center. Patches were highly linear and often interconnected making it difficult to determine number of sampling locations per forest patch. Across all natural area complexes, forest patches containing sampling plots varied in width from 52 m to 1388 m with an average width of 415.75 m (sf. = 31.13 m).
The native land cover of the area is a mix of hardwood species including northern red oak (Quercus rubra), white oak (Q_. alba), American beech (Fagus grandifolia), red maple (Acer rubrum), sycamore (Platanus occidentalis), and yellow poplar (Liriodendron tulipifera) (Heckscher, 2004). The land use in the surrounding landscape includes both agricultural and residential properties (Conover, 2011). Nonnative plants present in the study areas include, but are not limited to: autumn olive (Elaeagnus umbellate), garlic mustard (Alliaria petiolata), Japanese barberry (Berberis thunbergii), Japanese honeysuckle (Lonicera japonica), Japanese stilt grass (Microstegium vimineum), Norway maple (Acer platerioides), oriental bittersweet (Celastrus oriculatus), multiflora rose (Rosa multiflora), and wineberry (Rubus phoenicolasius).
We sampled 98 plots in seven natural area complexes for the presence of forest songbirds (Fig. 1). Sample plots were located in mature forest 25 m from a forest edge and were randomly placed within selected forest patches using the Hawths Tools extension for ArcGIS 9.2 (Beyer, 2004). We separated plots by 250 m to minimize the likelihood that individual birds would be double counted between sample plots (Bibby and Burgess, 2000). At each of the 98 sampling plots, we conducted 25 m radius avian point counts to estimate occupancy (Bibby and Burgess, 2000). The use of small radius plots allowed for the measurement of vegetation characteristics in close physical proximity to locations of birds, increasing the accuracy of resulting bird-habitat relationships (Ralph et al., 1995). This sampling method does not account for landscape-scale factors such as percent forest cover. However, as these factors were not the focus of this study, the natural area complexes containing our sampling plots were selected to have high overlap at a landscape scale to limit difference between sampling plots at a landscape level.
We conducted point counts three times per summer between 15 May-7 August in 2009 and 2010. Each point was visited on a randomly selected day between 15 May-15 June, 16 June-15 July, and 16 July-7 August. A single observer (AMD) was trained to identify Eastern forest bird species by sight and sound and conducted all point counts across both seasons.
We selected an assemblage of Eastern songbird species that commonly forage on invertebrates, including within our surveyed vegetation zone of [less than or equal to] 2 m above the ground (De Graaf et al., 1985; Table 1). We conducted point counts between 15 min before sunrise and 5 h after sunrise, with >96% of the surveys taking place between sunrise and 4 h after sunrise. Surveys were only conducted on precipitation-free days when the wind speed was <6.5 km/ hr. At each plot, we recorded the date, time, and temperature, and estimated percent cloud cover. Time, date, percent cloud cover, and temperature were recorded for use as detection covariates, as these factors influence the probability that a bird was detected during a survey if it was present at a site (MacKenzie et al., 2006). A 1 min acclimation period of minimal observer movement preceded each survey to minimize effects from observer disturbance (Buckland et al., 2001; Rosenstock et al., 2002). Following the acclimation period, we recorded all birds seen or heard within the survey area during a 5 min period of passive observation.
Within each 25 m radius plot, we measured a number of environmental variables for use as site covariates. We assumed vegetation structure and composition were constant across sampling years and sampled vegetation only in 2009. We sub-sampled vegetation at distances of 0 m, 25 m, and 50 m from the forest edge (Fig. 2). At each distance we measured understory coverage via a Vegetation Profile Board (Nudds, 1977). The board is marked in alternate colors (red/white) and set in the ground and viewed from two cardinal directions (parallel to the forest edge) from a distance of 10 m. The proportion of each Vi m interval covered by vegetation is recorded as single digit between one and five, allowing for the calculation of an average density for each vertical layer of vegetation. We also measured basal area with a 10 factor prism and canopy coverage with a densiometer (Strickler, 1959). We recorded vegetation composition along a transect 5 m in length and 15 mm in width running parallel to the forest edge at 0 m, 25 m, and 50 m from the forest edge. We identified all understory vegetation [less than or equal to] 2 m in height intersecting each transect (where focal avian species commonly forage on invertebrates; De Graaf et al., 1985; Williamson, 1971). We identified plants to species (plants from the genera Rubus and Trifolum were only identified to genus) and as native or nonnative species. Due to difficulty in identification, the terms "ferns" and "grasses" were used to indicate all species within these groups, which were excluded from calculations of native species proportion. The proportion of native plants at each site was calculated by dividing the number of decimeter sections of the 5 m transects that contained at least one native plant by the number of sections that contained any vegetation, either native or nonnative. The proportions calculated for the transects at the forest edge, 25 m from the forest edge, and 50 m from the forest edge at each point were averaged to obtain a value of native plant proportion for each sample plot. Nonvegetated ground at each site was calculated by dividing the number of decimeter sections of the 5 m transects that did not contain any vegetation by the total number of decimeter sections of the transect (50). The proportions calculated for the transects at the forest edge, 25 m from the forest edge and 50 m from the forest edge at each point, were averaged to obtain a value of nonvegetated ground for each sample plot.
INVERTEBRATE SAMPLE COLLECTION
We collected 588 vacuum samples for invertebrates within the 98 plots at three points per site each year during June and July, for a total of six samples over the 2 y. We sampled for invertebrates within three I m radius circles at the center of the same transects used for the vegetation surveys, totaling 6.28 [m.sup.3] per plot. We vacuum sampled vegetation for invertebrates using a leaf blower (Craftsman 25cc Gas Blower/Vac Model #358794740) in the vacuum setting and fitted with a nylon mesh paint strainer bag. A single technician performed all vacuum sampling to minimize the effects of sampling technique. Following sampling, we searched the vegetation for any remaining Lepidoptera larvae. Specimens were frozen at -10C in plastic zip-top bags before being sorted to retain invertebrate taxa known to be preferred breeding songbird foods (Martin et al., 1951). These taxa include the orders Orthoptera, Hemiptera, Coleoptera, Lepidoptera, Araneae, Opiliones, Hymenoptera, Diptera, and Isopoda, and the classes Gastropoda and Diplopoda. To determine biomass, we dried samples at 55 C for [greater than or equal to] 48 h and weighed them to the nearest 0.0001 g.
We used the single season habitat occupancy modeling approach of MacKenzie et al., (2006) for the nine songbird species to examine bird habitat selection. This approach predicts the probability of a site being occupied and at least one individual being detected during the survey using site and survey specific variables (MacKenzie et al., 2006). Our models included the site covariates native plant proportion (native), understory coverage (uc), canopy coverage (canopy), basal area (basal), and invertebrate biomass (inverts), which were consistent over the duration of the season. A logl(l transformation was applied to insect biomass data and a square-root transformation was applied to basal area and nonvegetated ground data to meet the assumptions of normality and homoscedasticity. All means and 95% confidence intervals are reported as back-transformed values. Survey covariates, which changed between repeat visits, included time since sunrise (time), number of days since start of survey period (day), temperature (temp), and cloud cover (clouds) for each avian survey.
We used the 'unmarked' package in program R (Fiske and Chandler, 2011; R Development Core Team, 2010) to model single-season occupancy for candidate species. The models were evaluated using Akaike Information Criterion (AIC) to determine the most parsimonious model (MacKenzie et al., 2006). We first modeled detection of each individual species using the survey covariates (date, temperature, minutes since sunrise, and cloud cover) and constant occupancy to determine the variables influencing detection of the species. For each species, every single-variable detection model was considered, as well as a null model where detection was constant, and a global model containing all detection variables. Detection models with [DELTA]AIC values > 2 were rejected due to a lack of empirical support (Burnham and Anderson, 2002). If multiple detection models had [DELTA]AIC values [less than or equal to] 2, the higher ranked model of either the null or global model was selected for the species.
To reduce our model set, we incorporated the best supported model (s) of detection for each species into an a priori set of occupancy models, including single-variable models of all site covariates, a selected set of two-variable models representing biologically meaningful interactions, a null model where occupancy was considered constant, and a global model with all site covariates (Table 2). Again, models with a [DELTA]AIC > 2 were rejected (Burnham and Anderson, 2002). Variance-covariance matrix or model convergence errors, were resolved by providing different initial values or fixing the beta value for certain parameters. The beta value for each variable was used to assess whether a variable was significantly related to the occupancy of a species, and whether the variable was positively or negatively related to species occupancy. The relationship was considered nonsignificant if the 95% confidence interval for the beta value contained zero.
Naive occupancy estimates for the candidate species at our study plots ranged from 45.9% to 81.6%. Naive occupancy rates were calculated simply as the percentage of plots where each species was detected during at least one visit and do not account for imperfect detection. Modeled occupancy rates accounted for factors which affected the probability of detecting a species during a visit if it is present at the site. The modeled occupancy rates of the candidate species at study plots based on the best-supported model of detection for each species ranged from 65.8% to 96.7% (Table 3).
We observed 94 species of plants during vegetation surveys (Conover, 2011), with 78.7% considered native to the study region (U.S. Department of Agriculture, 2011). Vegetation composition and forest structure measurements varied across study points (Table 4).
Invertebrate sampling yielded 12,108 individuals and 18.69 g of preferred breeding songbird foods. Total invertebrate biomass collected from each invertebrate sampling point in 2009 ranged from 0.018 g/[m.sup.3] to 0.12 g/[m.sup.3] (mean = 0.049, 95% ci [0.046, 0.052]). In 2010 total invertebrate biomass collected from each invertebrate sampling point ranged from 0.022 g/[m.sup.3] to 0.18 g/[m.sup.3] (mean = 0.053, 95% ci [0.050, 0.056]).
Time of day was a significant factor influencing detection of six of the songbird species (Table 5). Of the 13 models included in our apriori set to model occupancy, nine were top-ranked models for at least one of the songbird species (Table 6). Null models were included in the top models of occupancy for six species, ovenbird (Seiurus aurocapilla), Northern cardinal (Cardinalis cardinalis), Eastern towhee (Pipilo erythrophthalmus), Carolina chickadee (Poecile carolinensis), red-eyed vireo (Vireo olivaceus), and Carolina wren (Thryothorus ludovicanus). Of these species Eastern towhee, ovenbird, and Carolina chickadee did not show significant relationships between occupancy and any variables we examined.
Nonvegetated ground appeared in the top models for four species, more than any other variable. It was positively related to occupancy in Northern cardinal and American robin (Turdus migratorius) and negatively related to gray catbird (Dutnetella carolinensis) occupancy. Proportion of native plants was the next most common variable, appearing in the top-ranked models for three species. However, the only significant relationship was in the occupancy of Carolina wren, which was positively related to the variable. Basal area also appeared in the top-ranked models for three species and had a significant positive relationship with only one species, wood thrush (Hybcichla mustelina).
Canopy coverage and invertebrate biomass each appeared in the top-ranked models for one species, American robin and Carolina wren respectively, and each were positively related to the occupancy of that species. Understory coverage did not appear in the top-ranked models for any of the songbird species.
Occupancy of six of our nine songbird species in suburban forest fragments was affected by forest structure, native plant density, or invertebrate biomass. Percentage of nonvegetated ground had a significant relationship with the occupancy of more of our songbird species than any other variable, including gray catbird, Northern cardinal, and American robin. Gray catbird was the only species that had a negative relationship with the variable, indicating the species prefers sites with increased ground coverage. Occupancy of American robin and Northern cardinal were positively related to percentage of bare ground, indicating these species prefer sites with a more open understory. However, none of the species showed a relationship with understory coverage, suggesting they are responding to the presence of vegetation cover closest to the ground (measured by percentage of nonvegetated ground), rather than vegetation density within the entire understory up to 2 m (measured by understory composition). A relationship between abundance and vegetation density has been identified in previous studies for gray catbirds (Lent, 1990), American robins (Blake and Karr, 1987), and northern cardinals (Leston and Rodewald, 2006), although our results distinguish the importance of cover specifically in lowest part of the understory for these species in suburban forest fragments.
We found basal area to be the most significant factor influencing occupancy of wood thrush in suburban forest fragments. Wood thrush occupancy was positively associated with basal area, a relationship that has also been well documented for this species in contiguously forested areas (Wang et al., 2006, Sargent et al., 2003, Annand and Thompson III, 1997). Our results provide support that occupancy relationships with forest structure for wood thrush within suburban forest fragments align with those in areas of more contiguous forest cover.
Two foliage gleaning species, red-eyed vireo and Carolina wren were the only species found to have positive associations with either native plants or invertebrate biomass. Given the implications of the replacement of native plants by nonnative invasive species on the availability of invertebrate prey for avian insectivores in forest fragments, a positive relationship between the occupancy of foliage gleaning species and native plants may indicate selection of habitats offering a greater abundance of invertebrate prey. If native plants are providing a greater abundance of invertebrate prey for foliage gleaners, a positive relationship would then be expected between the occupancy of these species and invertebrate biomass at a site. Results of a study of bird occupancy along a gradient of urban development in California recorded increasing numbers of foliage gleaning species as the level of urbanization decreased, a relationship the author attributed to low insect densities in the urban areas dominated by nonnative plants (Rottenborn, 1999).
While a positive correlation between invertebrate biomass and occupancy of all insectivorous species would be expected, this finding for only one species in our study is likely due to an alignment between the primary foraging substrate of the red-eyed vireo and the most effective range of our invertebrate sampling method. Red-eyed vireos forage most commonly among live leaves and twigs near branch ends (Barrow, 1990), from which invertebrates would have been effectively sampled using our vacuum sampling technique (Doxon et al., 2011). However, we did not sample the leaf litter, soil, or air column for invertebrates. If birds were responding to invertebrates that were not sampled, vacuum-sampled invertebrate biomass values may not be an accurate measure of the entire avian food supply within sampling plots. In future studies litter collection and a wider variety of insect sampling techniques may provide a more accurate assessment of the availability of invertebrate biomass at our sites for species with varying foraging strategies.
Eastern towhee, ovenbird, and Carolina chickadee occupancy did not show significant relationships with any variables we examined. The null model was well-supported for each of these species, suggesting they were responding to variables outside the scope of this study. The occupancy of all our songbird species was likely influenced by unmeasured factors in the suburban landscape, including surrounding land use and forest patch size. Previous work examining factors influencing avian species richness in remnant habitat patches within suburban development has found while forest structural characteristics contribute to determining bird species richness, landscape-scale factors, including patch size, were also significant determinants of avian species richness (Palmer et al., 2008; Melles et al., 2003).
Overall, the occupancy relationships with forest structure variables we identified for our songbird species in suburban forest fragments aligned well with those previously identified in studies examining habitat use in more forested areas. This supports management activities encouraging forest structural characteristics associated with these species in more heavily forested areas could improve habitat quality for these species in suburban forest fragments as well. However, our study was a micro-scale analysis of the habitat factors influencing bird occupancy at the site level, and some of our selected species appeared to be responding primarily to factors outside the scope of this analysis. With this in mind, it is important to also consider the role of landscape-scale factors before planning the management of suburban forest fragments.
Only one species had a significant relationship with native plants, suggesting for the species we examined, vegetation structure, rather than composition is a more important factor in occupancy. However, if the replacement of native plants with nonnative invasives does lead to reduced availability of invertebrate prey, the impacts of this trophic breakdown may not be easily detected using occupancy analysis. Species may be selecting for the vegetation structure offered by native plants, while suffering fitness impacts from a lack of invertebrates in these patches. Further study of fitness impacts such as nest success is needed to fully understand how best to manage suburban forest fragments as habitat for songbirds.
Acknowledgments.--Our project was funded through the USDA Forest Service's Northern Research Station, with the support of the University of Delaware. Our project was made possible by the cooperation of the following parks and natural areas: White Clay Creek State Park, Fair Hill Natural Resource Management Area, St. Andrews School, Ashland Nature Center, Red Clay Greek Preserve, and Mt. Cuba Center. We thank J. Zaccaria for field assistance, and D. Tallamy and W. G. Shriver for support and advice. The authors further declare they have no conflict of interest associated with this research.
Annand, E. and F. R. Thompson III. 1997. Forest bird response to regeneration practices in central hardwood forests. J Wildl. Manage., 61:159-171.
Barrow, W. C. 1990. Ecology of small insectivorous birds in a bottomland hardwood forest. Phd Thesis, Louisiana State Univ., Baton Rouge. 212 pp.
Beyer, H. I., [online]. 2004. Hawth's analysis tools for ArcGIS. http://www.spalialecology.com/htools (17 November 2008).
Bibby. C. J. and N. D. Burgess. 2000. Bird Census Techniques. Academic Press, Burlington, Massachusetts, U.SA. 302 pp.
Blake, J. G. and J. R. Karr. 1987. Breeding birds of isolated woodlots: area and habitat relationships. Ecology. 68:1724-1734.
Buckland, S. T., D. Anderson, K. Burnham, J. Laake, D. Borchers, and L. Thomas. 2001. Introduction to Distance Sampling: Estimating Abundance of Biological Populations. Oxford University Press, New York, U.S.A. 432 pp.
Burke, D. M. and E. Nol. 1998. Influence of food abundance, nest site habitat, and forest fragmentation on breeding Ovenbirds. Auk, 115:96-104
Burnham, K. P. and D. R. Anderson. 2002. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach, 2nd ed. Springer, New York, U.S.A. 488 pp.
Conoyer, A. 2011. The impact of non-native plants on bird communities in suburban forest fragments. M.S. Thesis, University of Delaware, New-ark, USA. 44 pp.
Cripps, M. G., M. Schwarzlvnder, J. L. McKenney, H. L. Hinz, and YV. J. Price. 2006. Biogeographical comparison of the arthropod herbivore communities associated with Lepidium draba in its native, expanded and introduced ranges. J Biogeog., 33:2107-2119.
De Chant, T., A. H. Gallego, J. V. Saornil, and M. Kelly. 2010. Urban influence on changes in linear forest edge structure. Land. Urban Plan, 96:12-18.
De Graaf, R. M., N. G. Tilghman, and S. H. Anderson. 1985. Foraging guilds of North American birds. Environ. Manage., 9:493-536.
Doxon, E. D., C. A. Dams, and S. D. Fuhlendorf. 2011. Comparison of two methods for sampling invertebrates: vacuum and sweep-net sampling. J. Field Omithol., 82:60-67.
Doyle, A. L. 2008. Effects of forest fragmentation and honeysuckle invasion on forest Lepidoptera in southwest Ohio. M.S. thesis, Wright State University, Dayton, OH. 89 pp.
Fiske, I. and R.B. Chandler. 2011. unmarked: An R package for fitting hierarchical models of wildlife occurrence and abundance. J. Stal. Soft., 43:1-23.
Gerber, E., C. Krebs, C. Murrell, M. Morftti, R. Rocklin, and U. Schaffner. 2008. Exotic invasive knotweeds (Fallopia spp.) negatively affect native plant and invertebrate assemblages in European riparian habitats. Biol. Conserv., 141:646-654.
Harper, K. A., S. E. Macdonald, P. J. Burton, J. Chen, K. D. Brosofske, S. C. Saunders, E. S. Euskirchen, D. Roberts, M. S. Jaiteh, and P. Esseen. 2005. Edge influence on forest structure and composition in fragmented landscapes. Conserv. Biol., 19:768-782.
Heckschfr, C. M. 2004. Yeerv nest sites in a mid-Atlantic Piedmont forest: vegetative physiognomy and use of alien shrubs. Amer. Mid. Nat., 151:326-337.
Lent, R. A. 1990. Relationships among environmental factors, phenotypic characteristics, and fitness components in the Gray Catbird (Dumetella rarolinensis). Ph.D. Thesis, Stale Univ. of New York at Stony Brook, Stony Brook. 89 pp.
Leston, L. F. V. and A. D. Rodewald. 2006. Are urban forests ecological traps for understory birds? An examination using Northern cardinals. Biol. Conserv., 131:566-574.
Lussier, S. M., R. W. Enser, S. N. Dasilva, and M. Charpfntier. 2006. Effects of habitat disturbance from residential development on breeding bird communities in riparian corridors. Environ. Manage., 38:504-521.
Mackenzie, D. I., J. D. Nichols, K. H. Pollock, J. A. Royle, L. L. Bailey, and J. E. Hines. 2006. Occupancy Estimation and Modeling: Inferring Patterns and Dynamics of Species Occurrence. Academic Press, Burlington, Massachusetts, U.S.A. 344 pp.
Martin, A. C., H. S. Zim, and A. L. Nelson. 1951. American wildlife and plants: A guide to wildlife food habits: The use of trees, shrubs, weeds, and herbs by birds and mammals of the United States. McGraw-Hill, New York, U.S.A. 512 pp.
Melles, S., S. Glenn, and K. Martin. 2003. Urban bird diversity and landscape complexity: Species-environment associations along a nmltiscale habitat gradient. Conserv. Erol. 7:5. [online] URL: http://www.consecol.org/vol7/issl/art5/. (31 March 2017).
Nudds, T. D. 1977. Quantifying the vegetative structure of wildlife cover. Wildl. Soc. Bull, 5:113-117.
Paine, R. T. 1980. Food webs: linkage, interaction strength and community infrastructure. J. Anim. Erol. 49:666-685.
Palmer, G. C., J. A. Fitzsimons, M. J. Antos, and J. G. White. 2008. Determinants of native avian richness in suburban remnant vegetation: Implications for conservation planning. Biol. Conserv., 141:2329-2341.
R Development Core Team. 2010. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Retrieved from http://www.R-project. org. (31 March 2017).
Rvlpii, C. J., J. R. Sauer, and S. Droege. 1995. Monitoring Bird Populations by Point Counts. Gen. Tech. Rep. PSW-GTR-149. Albany, CA: Pacific Southwest Research Station, Forest Service, US. Department of Agriculture; 187 p.
Rosenstock, S. S., D. R. Anderson, K. M. Giesen, T. Leukering, M. F. Carter, and F. Thompson III. 2002. Landbird counting techniques: current practices and an alternative. Auk 119:46-53.
Rottenborn, S. C. 1999. Predicting the impacts of urbanization on riparian bird communities. Biol. Conserv., 88:289-299.
Sargent, R. A., J. C. Kilco, B. R. Chapman, and K. V. Miller. 2003. Nesting ecology of wood thrush (Turdidae: Passeriformes) in hardwood forests of South Carolina. Southeast. Mai. 2:217-222.
Sauf.r, J. R.,J. E. Hines, J. E. Fallon, K. L. Pardieck, D.J. Ziolkowski, Jr., and W. A. Link. 2014. The North American Breeding Bird Survey, Results and Analysis 1966-2012. Version 02.19.2014 USGS Patuxent Wildlife Research Center, Laurel, Maryland, U.S.A. [online] URL: https://www. mbr-pwrc.usgs.gov/bbs/bbs2010.html. (31 March 2017).
Saunders, D. A., R. J. Hobbs, and C. R. Margules. 1991. Biological consequences of ecosystem fragmentation: a review. Consetv. Biol., 5:18-32.
Strickler, G. S. 1959. Use of the densiometer to estimate density of forest canopy on permanent sample plots. P.N.W. Old Series Research Notes, 180:1-5.
Tallamy, D. W., M. Ballard, and V. D'Amico. 2010. Can alien plants support generalist insect herbivores? Biol Invasions, 12:2285-2292.
Tilghman, N. G. 1987. Characteristics of urban woodlands affecting breeding bird diversity and abundance. Landscape Urban Plan., 14:481-495.
U.S. Department of Agriculture [online]. 2011. The PIANTS Database, http://plants.usda.gov. (22 June 2009).
Wang, Y., Lesak, A. A., Felix, Z., and C.J. Schweitzer. 2006. A preliminary analysis of the response of an avian community to silvicultural treatments in the southern Cumberland Plateau, Alabama, USA. Integr. Zool 1:126-129.
Williamson, P. 1971. Feeding ecology of the red-eyed vireo (Vireo olivaeeus) and associated foliage-gleaning birds. Erol. Monographs, 41:129-152.
Wu, Y. T., C. H. Wang, and X. D. Zhang. 2009. Effects of saltmarsh invasion by Spartina alterniflora on arthropod community structure and diets. Biol. Invas., 11:635-649.
Yvtfs, E. D. and D. F. Levia. 2004. Recruitment of three non-native invasive plants into a fragmented forest in southern Illinois. Forest Erol. Manage., 190:119-130.
Yoshiokv, A., T. Kadoya, S. I. Suda, and I. Washitani. 2010. Impacts of weeping lovegrass (Eragrostis curvula) invasion on native grasshoppers: responses of habitat generalist and specialist species. Biol. Invas., 12:531-539.
Zanette, L., P. Doyle, and S. M. Tremont. 2000. Food shortage in small fragments: evidence from an area sensitive passerine. Ecology, 81:1654-1666.
Zhou, VV. G., Huang, S. T., A. Pickett, and M. L. Cadenasso. 2011. 90 years of forest cover change in an urbanizing watershed: spatial and temporal dynamics. Landscape Ecol, 26:645-659.
Submitted 22 July 2016
Accepted 6 April 2017
AMANDA M. DUREN and CHRISTOPHER K. WILLIAMS (1), Department of Entomology and Wildlife Ecology, University of Delaware, 250 Townsend Hall, Newark 19716
United States Forest Service, Northern Research Station Unit 04, 246 Townsend Hall, Newark Delaware 19716
(1) Corresponding Author: e-mail: email@example.com
Caption: Fig. 1.--Map of study area in Maryland and Delaware, U.S.A. Shaded areas indicate forested patches within study areas. Black circles denote study plots where point counts for birds, vegetation sampling, and vacuum sampling for invertebrates were conducted from 15 May-7 August 2009-2010
Caption: Fig. 2.--Design of study site for avian and vegetation surveys and invertebrate sampling within Delaware and Maryland, U.S.A., forest fragments, 15 May-7 August 2009-2010
Table 1.--Nine forest songbird species selected for occupancy analysis using data collected during point count surveys in Delaware and Maryland, U.S.A., from 15 May-7 August 2009-2010 Species Scientific Name Wood Thrush Hylocirhla mustelina Oven bird Seiurus aurocapilla Eastern Towhee Pipilo erythrophthalmus Carolina Wren Thriothorus ludovicanus Northern Cardinal Cardinalis cardinalis Gray Catbird Du metella carolinensis American Robin Turdus migratorius Red-eved Vireo Vireo olivareiis Carolina Chickadee Poecile carolinensis Table 2.--Covariate models used to analyze occupancy of 9 forest songbird species in Delaware and Maryland, U.S.A., from 15 May-7 August 2009-2010 Model Covariate (s) null None global All native Native Plant Proportion canopy Canopy Cover basal Basal Area inverts Invertebrate Biomass non-veg Nonvegetated Ground tic Understory Coverage native+uc Native Plant Proportion, Understory Coverage native+in verts Native Plant Proportion, Invertebrate Biomass native+non-veg Native Plant Proportion, Nonvegetated Ground can opy+ n on-veg Canopy Coverage, Nonvegetated Ground canopy+uc Canopy Coverage, Understory Coverage basal+uc Basal Area, Understory Coverage non-veg+uc Nonvegetated Ground, Understory Coverage Table 3.--Occupancy rales of nine forest songbird species during point count surveys at 98 sites in Delaware and Maryland, U.S.A., from 15 May-7 August 2009-2010. Naive occupancy estimates calculated as the percentage of study sites where the species was detected. Predicted occupancy estimates and associated standard errors (se) incorporate imperfect detection using the best-supported detection model for each species Species Number of Naive Predicted plots with occupancy occupancy species estimate estimate (%) detected (%) [+ or -] se Red-eyed Vireo 80 81.6 96.7 0.06 Grav Catbird 80 81.6 82 0.04 Northern Cardinal 73 74.5 88.3 0.06 Wood Thrush 72 73.5 87 0.06 Eastern Towhee 70 71.4 79.6 0.05 Carolina Chickadee 57 58.2 84.2 0.1 American Robin 52 53.1 65.2 0.07 Carolina Wren 49 50 84.3 0.14 Ovenbird 42 42.9 65.8 0.1 Table 4.--Vegetation composition and forest structure covariates measured at study sites in June 2009 for use in developing occupancy models of nine forest songbird species in Delaware and Maryland, U.S.A. Means and 95% confidence intervals of the means are reported as back-transformed values Variable Minimum Maximum Mean (95% CI) value value Native Plant Proportion 0 1 0.616 (0.613, 0.619) Basal Density (m/ha) 0.82 11.08 5.25 (4.85, 5.66) Canopy coverage 66.02% 97.57% 87.54% (87.47%, 87.61%) Non-vegetated ground 0% 62.66% 16.70% (13.69%, 20.25%) Understory coverage 4.17% 75.69% 40.34% (40.16%, 40.52%) Table 5.--Top ranked models of variables affecting detection of nine forest songbird species during point count surveys in Delaware and Maryland, U.S.A, from 15 May-7 August 2009--2010. Table presents the difference in Akaike's Information Criterion value compared to the top-ranked model ([DELTA][AIC.sub.c]), the AIC model weight (W), and the number of parameters in the model (K). Global models contain all detection covariates. Models selected for use in modeling occupancy shown in bold Species Model [DELTA] W K [AIC.sub.c] Gray Catbird time 0 0.6524 3 Northern Cardinal time 0 0.3286 3 null 0.76 0.2247 2 Wood Thrush temp, time 0 0.3306 4 time 0.67 0.2365 3 Eastern Towhee clouds, temp 0 0.3796 4 temp 0.54 0.2898 3 global 1.78 0.1559 5 Carolina Chickadee Time 0 0.7229 3 global 1.92 0.2768 5 American Robin null 0 0.2604 2 clouds 0.37 0.2164 3 temp 1.1 0.1502 3 time 1.85 0.1032 3 Carolina Wren null 0 0.3978 2 Red-eyed Vireo time 0 0.3999 3 global 0.31 0.3416 5 temp 1.22 0.2172 3 Ovenbird time 0 0.2168 3 temp 1.05 0.1283 3 time: time since sunrise; temp: air temperature; clouds: cloud cover AIC model weight (W) indicates the strength of evidence of each model relative to other models in the set of models considered
|Printer friendly Cite/link Email Feedback|
|Author:||Duren, Amanda M.; Williams, Christopher K.; D'amico, Vincent|
|Publication:||The American Midland Naturalist|
|Date:||Oct 1, 2017|
|Previous Article:||Local-scale Habitat Associations of Grassland Birds in Southwestern Minnesota.|
|Next Article:||Winter Breeding and Summer Nonbreeding Home Ranges of Bald Eagles from Louisiana.|