TY - JOUR
AU - Delap,, Jack
AB - Abstract Abstract. Nest predation may influence habitat selection by birds at multiple spatial scales. We blended population and community ecology to investigate this possibility for 15 species of forest songbirds and their diurnal nest predators (corvids and sciurids) in 28 1 km2 sites near Seattle, Washington, from 1998 to 2004. We determined whether songbirds were positively or negatively associated with nest predators at three spatial scales, and whether their co-occurrence affected reproductive success. At the largest ‘neighborhood’ scale (1 km2 areas that included suburban and exurban development and second-growth forest remnants), nest predators and their prey were positively or negatively correlated according to general species-specific habitat associations. At the intermediate ‘forest patch’ scale (among remnant forested areas 0.5 to 70 ha), associations between predators and prey were generally weak. At the smallest ‘within patch’ scale (multiple 50 m radius survey plots within each forest patch), some songbird species avoided areas with greater predator use, particularly by Steller's Jays (Cyanocitta stelleri). Failed nests and territories tended to be in locations of higher predator occurrence (especially of corvids) than successful ones, but at the largest 1 km2 neighborhood scale relative abundance of nest predators was not correlated with the fate of nesting attempts or annual reproductive success. Reproductive success was generally high, with 52% of all nests and 49% of all territories fledging at least one young (for all species and years combined). Nest predation influenced some species' use of resources, but was not a strong influence on overall reproductive success or community structure. Consecuencias De La Utilización De Hábitat Por Parte De Depredadores De Nidos Y De Aves Canoras Reproductivas a Través De Múltiples Escalas En Un Paisaje Urbanizado Resumen. La depredación de nidos puede influenciar la selección de hábitat de las aves a múltiples escalas espaciales. Combinamos conceptos de ecología de comunidades y de poblaciones para investigar esta posibilidad en 15 especies de aves canoras de bosque y sus depredadores diurnos (córvidos y esciúridos) en 28 sitios de 1 km2 en las cercanías de Seattle, Washington, entre 1998 y 2004. Determinamos si las aves canoras estaban asociadas de forma negativa o positiva con los depredadores de nidos a tres escalas espaciales diferentes, y si esta co-ocurrencia afectó sus éxitos reproductivos. A la escala espacial más grande de “vecindad” (áreas de 1 km2 que incluyeron tanto zonas de desarrollo suburbano y ex-urbano como remanentes de bosque secundario), los depredadores de nidos y sus presas presentaron una correlación positiva o negativa dependiendo de las asociaciones generales de hábitat específicas para cada especie. A la escala intermedia de “parche de bosque” (entre áreas de bosque remanente de 0.5 ha a 70 ha), las asociaciones entre depredadores y sus presas fueron en general bastante débiles. A la escala espacial menor de “intra-parche” (varias parcelas de 50 m de radio muestreadas dentro de cada parche de bosque), algunas especies de aves evitaron las áreas con un mayor uso por parte de los depredadores, particularmente las áreas utilizadas por el córvido Cyanocitta stelleri. Los territorios y nidos que fracasaron tendieron a estar en áreas con mayor ocurrencia de depredadores (especialmente de córvidos) que los que fueron exitosos. Sin embargo, a la escala espacial más grande de vecindad (1 km2), la abundancia relativa de depredadores de nido no se correlacionó con el destino de los intentos de nidificación ni con el éxito reproductivo anual. El éxito reproductivo fue generalmente alto; el 52% de los nidos y un 49% de todos los territorios produjeron al menos un volantón (combinando todas las especies y años). Si bien la depredación de los nidos influyó sobre la utilización de los recursos en algunas especies, ésta no tuvo una influencia fuerte sobre el éxito reproductivo general o sobre la estructura comunitaria. Introduction Habitat selection by birds is generally believed to be a hierarchical decision-making process dependent on an individual's ability to perceive environmental cues at different spatial scales (Hutto 1985, Block and Brennan 1993). Selection of breeding habitat in particular can be influenced by vegetation type or structure, food availability, nest site availability, presence of competitors or conspecifics, and the probability of nest predation (Cody 1985, Martin 1998). While nest predation is an important influence on avian productivity (Ricklefs 1969, Rotenberry and Wiens 1989, Martin 1993), parental behavior (Curio 1976, Marzluff 1985, Bradley and Marzluff 2003, Pietz and Granfors 2005), and nest-site selection (Martin et al. 2000, Forstmeier and Weiss 2004, Peluc 2006), its influence on general habitat selection is just beginning to be appreciated. Ovenbirds (Seiurus aurocapillus) in Pennsylvania, for example, exhibit atypical habitat selection for their species by breeding in forest edges and avoiding interior forests. This is attributed to the abundance of eastern chipmunks (Tamias striatus), a nest predator common in locations unused by Ovenbirds, but absent from Ovenbird territories near forest edges (Morton 2005). The existence of “ecological traps,” areas where cues like vegetation structure mislead birds to settle where predators are common and hence reproductive success is poor, indicate that some species do not directly assess predator occurrence when selecting breeding habitat (Dwernychuk and Boag 1972, Gates and Gysel 1978, Schlaepfer et al. 2002). Ignoring predator occurrence and settling in traps may be increasingly common in a human-dominated world where direct and indirect actions of people increase predator occurrences without grossly altering other cues used by birds to select habitats (Schmidt and Whelan 1999, Borgmann and Rodewald 2004, Millspaugh et al. 2006). The decoupling of proximate cues for habitat selection from ultimate factors determining reproduction and survival may also be affected by lag times in the response of birds to anthropogenic landscape change (Knick and Rotenberry 2000). Urban ecology may provide novel insights into the mechanisms and consequences of avian habitat selection, nest success, and the relationship of these factors to predation, competition, and vegetation. Since 1998 we have been studying bird populations and communities in the Seattle metropolitan area along a gradient of human density and settlement. This region has experienced dramatic urban growth, especially during the last 30 years (Hansen et al. 2005, Robinson et al. 2005), and remnant forests exist in a variety of sizes and settings, from small urban parks or undeveloped parcels to large blocks of contiguous forests (Donnelly and Marzluff 2004a, 2006). Songbird diversity peaks in landscapes with 50%–60% forest cover, because such areas gain more human commensal (synanthropic) and early successional species than the native forest species they lose (Marzluff 2005). This setting of remnant forest patches surrounded by differing amounts of urban development provides an opportunity not only to examine the effect of nest predators on habitat use and productivity of songbirds, but also to determine if predator assemblages change with removal of forest cover. Nest predators typical of Northwest forests include a variety of corvids, accipiters, sciurids, and small mammals (Sieving and Willson 1998, Luginbuhl et al. 2001, Raphael et al. 2002, Bradley and Marzluff 2003). If any particular predator species or suite of species exerts significant predation pressure on their prey, we would predict reduced nest success in areas with greater predator abundance, and possibly reduced use of those areas by prey species. To determine if such reduced use occurs, we predicted which predator and prey species were likely to have correlated abundances, based on habitat associations of these species in the general study area (Donnelly and Marzluff 2004a, 2006). We tested these predictions by developing a new analysis technique to relate observations of predators to songbird occurrence and nesting success. We expanded on previous studies of antipredator behavior at the scale of individual movements or nest site selection (Forstmeier and Weiss 2004, Peluc 2006) to larger scales of breeding territories within a forest patch and selection of forest patches in which to breed. Our research questions regarding the choices and consequences of avian habitat selection were: (1) are prey species less likely to use landscapes, forest patches, or areas within forested habitat with more predators? (2) if prey species use habitat with more predators, are there negative consequences for their productivity? Methods Study Area and Site Selection We chose 28 1 km2 suburban, exurban, and forested reserve sites (Marzluff et al. 2001) within a 3200 km2 area of temperate, moist forest around Seattle, Washington (study sites; Fig. 1). Elevation varied from sea level to near 300 m on the lower slopes of the Cascade Range. Four study sites were entirely forested reserves and 24 included developed (single-family residential) and forested portions. Forests were mostly coniferous, including western hemlock (Tsuga heterophylla), Douglas-fir (Pseudotsuga menziesii), and western red cedar (Thuja plicata), with lesser amounts of red alder (Alnus rubra), big-leaf maple (Acer macrophyllum), black cottonwood (Populus trichocarpa), and Oregon ash (Fraxinus latifolia; Franklin and Dyrness 1988). Figure 1 Open in new tabDownload slide Location of 28 study sites within the urbanizing landscape in and around Seattle, Washington. Each 1 km2 site (‘neighborhood’) is shown, and within each site the forested area in which songbird nests and territories were monitored, and songbird territories and predator locations were mapped, is shown to scale. Background image is from 1998 LandSat data classified by the University of Washington's Urban Ecology Research Lab. Figure 1 Open in new tabDownload slide Location of 28 study sites within the urbanizing landscape in and around Seattle, Washington. Each 1 km2 site (‘neighborhood’) is shown, and within each site the forested area in which songbird nests and territories were monitored, and songbird territories and predator locations were mapped, is shown to scale. Background image is from 1998 LandSat data classified by the University of Washington's Urban Ecology Research Lab. We randomly selected sites after stratifying the entire region on a per km2 basis by three landscape attributes: percent of urban and forest land cover types in the site, average patch size of urban land cover, and contagion—the probability that two randomly chosen adjacent areas belong to the same class (see Blewett and Marzluff 2005, Donnelly and Marzluff 2006 for details). Site contagion ranged from aggregated (similar land cover tended to occur in large patches) to interspersed (similar land cover tended to be in disjunct small patches). Landscape variables may be correlated with each other, thus we selected the fewest that adequately described important characteristics of study sites (Alberti et al. 2001, Rohila 2002). Songbird Relative Abundance and Productivity From April to August of 2002, 2003, and 2004, we used a modified behavioral mapping technique (International Bird Census Committee 1970), nest searching (Ralph et al. 1993), and the Vickery index (Vickery et al. 1992, Christoferson and Morrison 2001) to monitor the productivity of 15 species within the forested portions of each study site. In addition, nests (but not territories) were found and monitored from 1998 to 2001 (Donnelly and Marzluff 2004a, 2006, Blewett and Marzluff 2005). We focused our nest-searching efforts on ten ground-, shrub-, and cavity-nesting species whose nests we could find in sufficient quantities for subsequent analyses: the Pacific-slope Flycatcher (Empidonax difficilis), Bushtit (Psaltriparus minimus), Winter Wren (Troglodytes troglodytes), Bewick's Wren (Thryomanes bewickii), Swainson's Thrush (Catharus ustulatus), American Robin (Turdus migratorius), Spotted Towhee (Pipilo maculatus), Song Sparrow (Melospiza melodia), Dark-eyed Junco (Junco hyemalis), and Black-headed Grosbeak (Pheucticus melanocephalus). We monitored territories of the Black-throated Gray Warbler (Dendroica nigrescens), Wilson's Warbler (Wilsonia pusilla), and all those above except Bushtits, American Robins, and Black-headed Grosbeaks. In addition, some nests of the Black-capped Chickadee (Poecile atricapillus), Chestnut-backed Chickadee (Poecile rufescens), and Brown Creeper (Certhia americana) were found incidentally and monitored. The discrepancy between species whose territories or nests were monitored is due to differences in behavioral patterns and microhabitat selection. For example, American Robins were observed both in their breeding territories and at group foraging areas, which confounded spot-mapping observations when more than two breeding adults were present. In other species, such as Black-throated Gray Warblers, nests were extremely difficult to locate, so behavioral mapping in territories was the only feasible option for estimating productivity. From 2002 to 2004, we monitored 31 nests with small (about 6 cm long × 4 cm diameter) color and infrared video cameras (PicoCam II, Sandpiper Technologies, Manteca, California). The cameras were mounted 2–5 m from the nest and recorded video 24 hr per day on digital video recorders or VCRs. Nests were monitored until they either fledged young or failed due to predation, in which case we attempted to identify the predator species by reviewing video recordings. We spent 100–120 hr per site per year searching for nests, mist-netting and color-banding birds, conducting point count surveys, and mapping territorial activities between dawn and 15:00 (PST). The total area surveyed differed among sites (6.1 ± 0.9 ha, range  =  1.6–18.7 ha; Fig. 1) because of differences in forested area, bird density, and accessibility, but our effort was similar per unit area across sites. We netted at each site 4–6 times per year (4–8 hr per session) using both passive capture and playback of territorial songs and calls. Birds were color-banded with unique combinations to aid in the identification of different individuals and sexes. During weekly visits to each site, we searched for nests and observed each territory, typically for 3–6 hr per day (International Bird Census Committee 1970). We checked activity at known nests every 2–4 days after initial discovery. We discovered most nests early in the nesting cycle (building or egg-laying) and categorized them as ‘successful’ if they fledged at least one young (determined by observing fledglings, parental care, or feces on the nest rim on the expected fledging date), or ‘failed’ if they did not fledge young. Territory monitoring commenced each breeding season when males of each species first began to sing (or other prebreeding activity was observed), and territories were monitored throughout the spring and summer until the end of the final nesting attempt. Annual productivity was determined by repeated observations of parental behavior, mapping the locations of these behaviors, and assigning the respective Vickery rank to each territory (Vickery et al. 1992). Vickery ranks range from 1 (solitary male) to 5 (successful fledging) for single-brooded species and to 7 (successful fledging of two broods) for double-brooded species. Here, we refer to ‘failed’ territories as those with Vickery ranks <5, and ‘successful’ territories as those with Vickery ranks ≥5. Relative Use of Study Sites by Nest Predators We recorded the locations of all American Crows (Corvus brachyrhynchos), Common Ravens (Corvus corax), and Steller's Jays (Cyanocitta stelleri) seen or heard (‘corvid mapping’) in a subset of 16 sites used in a concurrent study of corvid habitat associations. In these sites we walked through an area that extended up to 400 m from the forest patch where songbird species were mapped and monitored. The 400 m buffer was chosen to include crow territories that might reasonably overlap the forest patches, based on an estimated 450 m territory diameter (Marzluff, McGowan et al. 2001). We visited each corvid mapping site six times per year between 15 April and 15 August in 2002, 2003, or 2004. Overall corvid abundance was significantly correlated between years (r  =  0.82 for five sites surveyed in both 2002 and 2003) so in 2004 we surveyed a different set of sites. Separately, in all 28 sites we noted locations of all potential diurnal nest predators while we mapped songbird behavior (‘incidental mapping’). These locations, and those from the corvid mapping efforts that overlapped the incidental mapping area, were used independently to estimate each predator species' and group's utilization distribution (Marzluff et al. 2004, Millspaugh et al. 2006; Fig. 2). The utilization distribution (UD) is a probability density function (Silverman 1986) that quantifies an individual's or group's relative use of space (van Winkle 1975, Kernohan et al. 2001). It depicts the probability of an animal occurring at each location within its traveled area as a function of relocation points (White and Garrott 1990:146). Utilization distributions can be estimated from point processes, such as observed locations of animals, using kernel techniques (Worton 1989, Kernohan et al. 2001). We used fixed kernel estimation with least squares cross validation (Kernohan et al. 2001) in the ANIMAL MOVEMENTS extension (Hooge and Eichenlaub 1997) of ArcView 3.1 (ESRI, Redlands, California) to estimate the UD from predator locations for a given site in a given year. We determined UDs for sciurid (Douglas' squirrel [Tamiasciurus douglasii], Townsend's chipmunk [Tamias townsendii], and eastern gray squirrel [Sciurus carolinensis]) and corvid (Steller's Jay, American Crow, and Common Raven) predators in total, for each predator group separately, and, when at least 10 locations were recorded, for individual predator species. Figure 2 Open in new tabDownload slide Determination of the predator load at specific points within a study area. Observed predator locations are mapped (A). A utilization distribution (UD) is derived by fixed kernel estimation from observed predator locations (B), reflecting areas of low to high predator load (low to high UD values). The UD value at any point of interest, such as the location of a nest (C), can be determined. The UD value is the predator load at a point in the site relative to other points. Figure 2 Open in new tabDownload slide Determination of the predator load at specific points within a study area. Observed predator locations are mapped (A). A utilization distribution (UD) is derived by fixed kernel estimation from observed predator locations (B), reflecting areas of low to high predator load (low to high UD values). The UD value at any point of interest, such as the location of a nest (C), can be determined. The UD value is the predator load at a point in the site relative to other points. We quantified the “predator load” at each nest or territory center (x, y) by determining the height of the UD (f ^UD[x, y]) at that location for each predator species or group (Fig. 2C). The height of the UD represents the amount of predator use at that location relative to other locations in the plane of the study site (Silverman 1986; Fig. 2C), which in our sample is actually our likelihood of detecting a predator at a given location. We measured the average height of the UD using an ArcView 3.1 extension (FOCAL PATCH, available at <http://gis.washington.edu/phurvitz/av_devel/focalpatch/; see Marzluff et al. 2004 for details). The effect size of interest in this part of our study was the difference in UD values (which range from 0 to 100) for individual predator species and groups of predator species at points (nest locations or territory centers) where birds successfully fledged versus failed to fledge young. Predator Mapping Analyses We explored different ways of calculating and using the predator mapping data collected by incidental and corvid mapping, to test whether incidental mapping methods were biased toward songbird nest and territory locations. We compared the UDs for corvids calculated from incidental mapping to the subset of observations in the same forest patch from corvid mapping by calculating the volume of intersection of each UD using FOCAL PATCH in ArcView 3.1. The incidental mapping UDs of corvids occupied 63% ± 4% of the volume of the corvid mapping (n  =  12 comparisons, sample size <16 because in four sites there was not a sufficient subset of observations in the forest patch from the corvid mapping method to calculate a UD), which reflected general concurrence of the two methods (Fig. 3). Due to the greater sample size, and the availability of data for both corvids and sciurids, we used UDs from incidental mapping locations for further analyses. Figure 3 Open in new tabDownload slide Comparison of utilization distributions (UDs) calculated from (A) ‘incidental mapping’ observations of corvids recorded during songbird spot-mapping, and (B) ‘corvid mapping’ observations recorded during separate visits to the field site. Both UDs were calculated for combined corvid observations made during 2002. The ‘volume of intersection,’ which reflects the amount of overlap of the two UDs, in this case was 61%. Figure 3 Open in new tabDownload slide Comparison of utilization distributions (UDs) calculated from (A) ‘incidental mapping’ observations of corvids recorded during songbird spot-mapping, and (B) ‘corvid mapping’ observations recorded during separate visits to the field site. Both UDs were calculated for combined corvid observations made during 2002. The ‘volume of intersection,’ which reflects the amount of overlap of the two UDs, in this case was 61%. We also tested whether the calculation of UDs was affected by recording observations of groups of predators (e.g., a flock of 10 crows) as a single detection or as separate individual detections. We compared UDs by calculating the volume of intersection (VI) between UDs at the same site derived from datasets with groups recorded as a single detection or multiple detections reflecting each member of the group. These two UDs were essentially identical (VI  =  92% ± 2%; n  =  10 comparisons, maximum group size varied from two to 50). For further analyses we calculated UDs based on datasets that recorded all predators in groups as multiple detections. An additional consideration when using UDs to represent predator utilization of an area is how to measure the height of the UD at a nest or territory location. We considered using an average (or weighted average) height of the UD in a certain radius around the nest or territory. However, the fixed kernel estimation using least squares cross validation assigns a ‘smoothing factor’ or H-value to the grid, which introduces spatial autocorrelation to UD heights up to the distance of the H-value. For the predator UDs we calculated, H-values were 34.1 ± 1.1 m (n  =  274, median  =  31), indicating that on average the UD height at a point was influenced by predator use (observations) within a 30–35 m radius. Since the territories of songbird species in our study are not likely to be much greater than 70 m in diameter, we used the UD height at a single point, centered on the nest or within the territory, for our analyses. Scale-specific Associations Between Predators and Prey We investigated the co-occurrence of 12 forest songbird species (prey) and their potential diurnal nest predators with data at three spatial scales: ‘neighborhood’ (relative abundance of predators and prey from all survey points in 1 km2 site), ‘forest patch’ (relative abundance of predators and prey from all survey points located in forested portions of each site, 0.5–70 ha), and ‘within patch’ (relative abundance of prey and concentration of predator use from individual survey points within each forest patch). At the neighborhood scale, we surveyed predators and prey using four to eight 50 m fixed-radius point counts (Ralph et al. 1993) placed in developed and forested portions of the 1 km2 site. Counts of all species heard or seen within 50 m lasted 10 min and were replicated four times per year at each point (Donnelly and Marzluff 2004a). At sites with interspersed developed and forested land, we allocated survey effort to six points in developed and two in forested portions (Fig. 4) because counts in developed portions are more variable than those in forest (Donnelly and Marzluff 2004b, 2006). At completely forested sites, we surveyed two to eight points per year. We standardized counts (detections per 10 min of survey effort), and used the average abundance per species per site per year as the experimental unit (n  =  71 “neighborhood years,” sample size <84 because not all 28 sites were surveyed in all three years). Figure 4 Open in new tabDownload slide Detailed view of two example study sites with point count locations and songbird and predator mapping areas shown within the 1 km2 ‘neighborhood.’ To the north is a control site with all survey points in a forested reserve. To the south is a typical suburban site with two survey points in the forested portion of the site (where nests and territories of songbirds, and point locations of predators, were also mapped) and six survey points in the developed portion of the site. Results from all eight points for each site were used for the neighborhood scale, while results from just points inside forested areas were used for the forest patch scale. Figure 4 Open in new tabDownload slide Detailed view of two example study sites with point count locations and songbird and predator mapping areas shown within the 1 km2 ‘neighborhood.’ To the north is a control site with all survey points in a forested reserve. To the south is a typical suburban site with two survey points in the forested portion of the site (where nests and territories of songbirds, and point locations of predators, were also mapped) and six survey points in the developed portion of the site. Results from all eight points for each site were used for the neighborhood scale, while results from just points inside forested areas were used for the forest patch scale. At the forest patch scale within each 1 km2 site, we selected only the point counts conducted within forest patches (0.5–70 ha in size; Fig. 4) and examined correlations among the average abundance per species per year at forested points as above. The annual survey of a site is again the experimental unit (n  =  62 “forest patch years,” sample size <71 because not all sites contained both developed and forested portions). At this scale we tested for co-occurrence of predators and prey in forest patches that differed in extent (3%–100% of each 1 km2 was forested) and configuration (forests in our study area varied from highly clumped to highly dispersed; Blewett and Marzluff 2005). We conducted a similar analysis using only data from point counts conducted outside forest patches (Fig. 4). However, results of this analysis were similar to the neighborhood scale results, and we could not examine correlations at the finer ‘within patch’ scale (see below) because we did not map predators outside the forested areas. At the ‘within patch’ scale (survey points within forest patches) we determined the predator load at each survey point within the forested portion of a study site, and correlated this with the relative abundance of songbirds detected per 10 min of survey effort at each point. Because predator loads are relative to points within a study site, we scaled them to also reflect site-specific predator abundance so that points could be compared within and among study sites. We scaled predator load by multiplying load by the abundance of predators (+0.016; half of the minimum nonzero value recorded) in the 1 km2 site. At this scale, the annual survey of a forested point is the experimental unit (n  =  159 for all predators combined, n  =  71 for crows, n  =  47 for jays, and n  =  107 for sciurids; sample sizes are lower for some individual predator species because a low number of detections made UDs impossible to calculate). To distinguish between potential avoidance of predators by prey and negative correlations due to selection of different habitat types, we categorized the 12 songbird species used for scale-specific correlations as likely to be more abundant in ‘forest’ or ‘open or edge’ habitats (Table 1). These expectations were based on relative abundance models for these species in this study area (Donnelly and Marzluff 2004a, 2006). ‘Forest’ species had positive coefficients for at least one of the variables ‘forest aggregation,’ ‘tree density,’ or ‘canopy closure.’ ‘Open or edge’ species had positive coefficients for at least one of the variables ‘exotic shrub cover,’ ‘exotic ground cover,’ or ‘urban land cover’ (see Donnelly and Marzluff [2004a, 2006] for variable definitions). Two species with positive coefficients for variables in both categories were labeled ‘mixed’ (Table 1). Table 1 General habitat associations of songbirds and nest predators near Seattle, Washington. Expected relative abundances were from previous study of the region, including our specific study sites, by Donnelly and Marzluff (2004b, 2006). Species assigned to ‘open or edge’ habitat were positively associated with the variables ‘exotic shrub cover,’ ‘exotic ground cover,’ or ‘urban land cover.’ Species assigned to ‘forest’ were positively associated with the variables ‘forest aggregation,’ ‘tree density,’ or ‘canopy closure.’ Species classified into the ‘mixed’ category were positively associated with at least one variable in each habitat. Open in new tab Table 1 General habitat associations of songbirds and nest predators near Seattle, Washington. Expected relative abundances were from previous study of the region, including our specific study sites, by Donnelly and Marzluff (2004b, 2006). Species assigned to ‘open or edge’ habitat were positively associated with the variables ‘exotic shrub cover,’ ‘exotic ground cover,’ or ‘urban land cover.’ Species assigned to ‘forest’ were positively associated with the variables ‘forest aggregation,’ ‘tree density,’ or ‘canopy closure.’ Species classified into the ‘mixed’ category were positively associated with at least one variable in each habitat. Open in new tab Statistical Analyses We used Pearson correlation analyses to test the two-tailed hypothesis that predator and prey counts were correlated. Significant correlations (positive or negative) were evaluated based on our expectations of predator and prey habitat use (Table 1). Of particular interest were negative correlations of species that were expected to select similar habitats, or negative correlations at only certain scales, because such results could be evidence of predator avoidance. We used a two-way repeated measures ANOVA to test whether our conclusions regarding the effect of predator load on songbird reproductive success varied by the method of predator mapping (incidental vs. corvid). We used this test for 12 sites where methods overlapped sufficiently. We measured whether the effect of predator load (UD values) on territory success (failed versus successful) changed depending on the source of the predator load data (incidental mapping versus corvid mapping), by looking for significant interaction terms in the ANOVA. To meet assumptions of normality and equal variances, we transformed all percentages or proportions using an arcsine square root transformation, and transformed the scaled predator load using a square root transformation (Zar 1996). All data are reported as means ± SE. We completed all statistical analyses using SPSS 10.1.3 software (SPSS 2001). Results Nest Predators Thirteen of 31 nests monitored by video were depredated. In seven cases the predator was not identified due to battery failure or the camera being knocked over before predation occurred. Predators observed on video were the American Crow (n  =  2 predation events), Cooper's Hawk (Accipiter cooperi; n  =  1), common garter snake (Thamnophis sirtalis; n  =  1), North American opossum (Didelphis virginiana&semi, n  =  1), and an unidentifiable nocturnal small mammal (likely Peromyscus spp.; n  =  1). We also directly observed predation of songbird eggs or nestlings as it occurred by the American Crow (n  =  3), Steller's Jay (n  =  1), and eastern gray squirrel (n  =  1). The amount of forest in the neighborhood influenced the occurrence of nest predators. American Crow abundance was negatively associated with percent forest cover at the 1 km2 scale (r  =  −0.55, P < 0.01, n  =  28) while Douglas' squirrel abundance was positively associated with percent forest cover (r  =  0.49, P < 0.01, n  =  28; Fig. 5). Common Ravens appeared to increase with more forest cover (r  =  0.37, P  =  0.052, n  =  28), while no other predator species were significantly correlated with percent forest cover (all P > 0.4). Overall, predator abundance declined slightly with increasing forest cover (r  =  −0.32, P  =  0.09, n  =  28). Figure 5 Open in new tabDownload slide Relationship between predator abundance and the amount of forest in the 1 km2 ‘neighborhood.’ Predator abundance at a site, calculated from all points at a site over all survey years, is shown separately for corvids and sciurids. Figure 5 Open in new tabDownload slide Relationship between predator abundance and the amount of forest in the 1 km2 ‘neighborhood.’ Predator abundance at a site, calculated from all points at a site over all survey years, is shown separately for corvids and sciurids. Scale-specific Associations Between Predators and Prey The pattern of association between nest predators and songbird abundance varied among predator species. Pacific-slope Flycatchers, Winter Wrens, and Black-throated Gray Warblers were negatively associated with American Crows regardless of scale (Fig. 6A). Bushtits and Dark-eyed Juncos were positively associated with American Crows at all scales. Crows were weakly or inconsistently associated with the other songbird species. Steller's Jays were positively associated with Winter Wrens, Swainson's Thrushes, and Spotted Towhees at all scales (Fig. 6B). More often, songbird associations with Steller's Jays were scale-dependent (see below). Sciurids were weakly associated with songbirds at most scales (Fig. 6C). Most songbird associations with the combined abundance of corvid and sciurid nest predators were consistently negative (Pacific-slope Flycatchers, Winter Wrens, Black-throated Gray Warblers, and Black-headed Grosbeaks), or consistently positive (Bushtits, Bewick's Wrens, Spotted Towhees, and Dark-eyed Juncos; Fig. 6D) across scales. Figure 6 Open in new tabDownload slide Co-occurrence of twelve forest songbirds and their diurnal nest predators: (A) American Crow; (B) Steller's Jay; (C) Douglas' squirrel, eastern gray squirrel, and Townsend's chipmunk; and (D) All of the above and Common Raven. Correlations between the relative abundance of songbirds and nest predators were calculated at three scales. Neighborhood scale  =  correlation among average detections of predators and prey at all survey points in the developed and forested portion of the 1 km2 study site; forest patch scale  =  correlation among average detections at all survey points in the forested portion of the study site; within patch scale  =  correlation among detections of songbirds at each point within the forest patch and point-specific value of each predator species' or group's utilization distribution scaled to also include landscape-specific relative abundance of each predator. Correlations large enough to reject the two-sided hypothesis that predator and prey abundances are not correlated are indicated by * (P < 0.05) or ** (P < 0.01). Correlations large enough for near-significant P-values (0.05 < P < 0.10) are indicated by †. Figure 6 Open in new tabDownload slide Co-occurrence of twelve forest songbirds and their diurnal nest predators: (A) American Crow; (B) Steller's Jay; (C) Douglas' squirrel, eastern gray squirrel, and Townsend's chipmunk; and (D) All of the above and Common Raven. Correlations between the relative abundance of songbirds and nest predators were calculated at three scales. Neighborhood scale  =  correlation among average detections of predators and prey at all survey points in the developed and forested portion of the 1 km2 study site; forest patch scale  =  correlation among average detections at all survey points in the forested portion of the study site; within patch scale  =  correlation among detections of songbirds at each point within the forest patch and point-specific value of each predator species' or group's utilization distribution scaled to also include landscape-specific relative abundance of each predator. Correlations large enough to reject the two-sided hypothesis that predator and prey abundances are not correlated are indicated by * (P < 0.05) or ** (P < 0.01). Correlations large enough for near-significant P-values (0.05 < P < 0.10) are indicated by †. Predator-prey associations also varied with the spatial extent of analysis. Associations were strongest at the neighborhood scale and weakest at the forest patch scale. At the within patch scale, six of the twelve songbird species appeared to avoid areas of frequent predator activity (Fig. 6A, 6B, 6C). This was often in contrast to associations at larger scales. Bewick's Wrens, Song Sparrows, and Dark-eyed Juncos were negatively associated with Steller's Jays within forest patches, despite weaker positive associations at larger scales (Fig. 6B). Similarly, Bushtits and American Robins were (weakly) negatively associated with Steller's Jays within forest patches, but positively associated at larger scales (Fig. 6B). Pacific-slope Flycatchers and Winter Wrens were (weakly) negatively associated with sciurids within forest patches although they were positively associated at larger scales (Fig. 6C). Consequences for Nesting Success We determined the fates of 784 nests in the forested portions of 20 sites from 1998 to 2004 (7–81 nests per site, combined over years; Table 2). Most nests were built in shrubs (73%) by American Robins, Swainson's Thrushes, Bushtits, and Black-headed Grosbeaks. We found fewer nests on or near the ground (18%); these were built by Song Sparrows, Spotted Towhees, Dark-eyed Juncos, and Wilson's Warblers. Approximately half of all nests fledged young (410 of 784 nests overall or 52%; Table 2; mean fledging success calculated for 20 sites  =  54% ± 2%). Ground-nesting species were more likely to successfully fledge nestlings than shrub-nesting species (χ21  =  4.6, P < 0.05). The three ground-nesting species with n > 10 nests had similar nest success rates (χ22  =  2.3, P  =  0.31), but success rates were different among the five shrub-nesting species (American Robin and Black-headed Grosbeak had below average success, χ24  =  10.4, P < 0.05; Table 2). Table 2 Overall success of nests (1998–2004) and territories (2002–2004) of 15 monitored species in study sites near Seattle, Washington. ‘Success’ for nests  =  fledged at least one young, for territories  =  Vickery score of 5 or more (at least one fledgling observed during the annual breeding season). Open in new tab Table 2 Overall success of nests (1998–2004) and territories (2002–2004) of 15 monitored species in study sites near Seattle, Washington. ‘Success’ for nests  =  fledged at least one young, for territories  =  Vickery score of 5 or more (at least one fledgling observed during the annual breeding season). Open in new tab At the larger scales, the relative abundance of predators was generally not correlated with nesting success. At the neighborhood scale (within each 1 km2 site; Fig. 7A) and forest patch scale (r  =  −0.19, P  =  0.19, n  =  50) predator abundance was weakly negatively correlated with nesting success. The strongest relationship observed using all nests was a pattern of increasing relative abundance of Townsend's chipmunks in the neighborhood with a decrease in proportions of successful nests (Fig. 7B), but this trend disappeared when two outliers were removed (r  =  −0.27, P > 0.25, n  =  18). The strongest (but not significant) correlation using the subset of shrub nests only was a declining proportion of successful shrub nests with increasing relative abundance of American Crows (r  =  −0.37, P  =  0.24, n  =  12). All other (n  =  14) correlations between success of all nests or just shrub nests with combined or individual predator species abundances were even weaker (all |r| < 0.32, all P > 0.20; an insufficient number of ground nests were found for separate analysis). Figure 7 Open in new tabDownload slide Correlation of average predator abundance at a site with the proportion of successful nests or territories (A), as measured by detections per survey of all predators combined at all points in the 1 km2 neighborhood (for nests, n  =  20; for territories, n  =  55). (B) The abundance of Townsend's chipmunks in the 1 km2 neighborhood was correlated with declining nest success (n  =  20). (C) The abundance of sciurids at forested points was correlated with declining success of ground-nesting birds' territories (n  =  44). Figure 7 Open in new tabDownload slide Correlation of average predator abundance at a site with the proportion of successful nests or territories (A), as measured by detections per survey of all predators combined at all points in the 1 km2 neighborhood (for nests, n  =  20; for territories, n  =  55). (B) The abundance of Townsend's chipmunks in the 1 km2 neighborhood was correlated with declining nest success (n  =  20). (C) The abundance of sciurids at forested points was correlated with declining success of ground-nesting birds' territories (n  =  44). At the within patch scale, kernelled estimates of predator abundance (UD values) near nests that failed tended to be greater than those near nests that fledged young, especially for corvids. The effect size (UD height at failed nests – UD height at successful nests, per site) measured the increased probability of detecting predators at failed versus successful nests. Effect size was positive (i.e., predator use was greater near failed nests relative to successful nests) in 14 of 20 sites. However, effect size was significantly greater than 0 (t-test, Pone-tailed < 0.05) at only three of these sites. The average effect size for combined corvid and sciurid detections at 20 sites was small, but marginally greater than 0 (t19  =  1.6; Pone-tailed  =  0.06; Table 3). The effect size for corvid predators was larger and significantly greater than 0 (t17  =  1.6; Pone-tailed  =  0.05). The mean effect size for sciurids was small (t15  =  1.5; Pone-tailed  =  0.08). These results are for all nests; effect sizes were variable and not significantly different from 0 when the smaller sample of only shrub nests was analyzed (Table 3). Table 3 Mean differences in predator utilization distribution values (effect size) at territories and nests failing versus succeeding to fledge at least one nestling. Effect size is interpreted as the change in probability (here scaled from 0–100) of a predator occurring at a failed territory or nest relative to a successful territory or nest; positive values indicate that predators were more likely to be detected in the area around failed territories or nests than successful territories or nests. Mean effect size ± SE and n (for territories, number of site-years; for nests, number of sites) are reported for combinations of predators and prey with n ≥ 10. One-tailed P-values for mean effect sizes are noted if P < 0.10. A separate analysis for ground nests was not conducted because an insufficient number were found. Open in new tab Table 3 Mean differences in predator utilization distribution values (effect size) at territories and nests failing versus succeeding to fledge at least one nestling. Effect size is interpreted as the change in probability (here scaled from 0–100) of a predator occurring at a failed territory or nest relative to a successful territory or nest; positive values indicate that predators were more likely to be detected in the area around failed territories or nests than successful territories or nests. Mean effect size ± SE and n (for territories, number of site-years; for nests, number of sites) are reported for combinations of predators and prey with n ≥ 10. One-tailed P-values for mean effect sizes are noted if P < 0.10. A separate analysis for ground nests was not conducted because an insufficient number were found. Open in new tab Consequences for Songbird Annual Productivity We determined the annual fates of 2119 territories in the forested portions of 28 sites from 2002 to 2004 (Table 2). Seventy annual assessments of sites were adequate for relating predator UDs to productivity (6–108 territories per site per year). Most territories monitored (59%) were of ground-nesting species, including Song Sparrows, Spotted Towhees, Dark-eyed Juncos, and Wilson's Warblers. Other territories commonly found were of open-cup nesting Swainson's Thrushes and Pacific-slope Flycatchers, and the enclosed or cavity-nesting Winter Wrens and Bewick's Wrens. Approximately half of all territories fledged at least one brood (1036 of 2119 territories overall or 49%; Table 2; mean territory success calculated for 70 annual assessments  =  54% ± 2%), the same overall success rate as nests (χ21  =  2.7, P > 0.10). Territory success rates were different among the three nesting guilds (χ22  =  128.6, P < 0.001); ground-nesting species had higher territory productivity than shrub-nesting or cavity-nesting species. However, within all three guilds there were differences in territory success rates among species (ground-nesters, χ23  =  142.8, P < 0.001; shrub-nesters, χ24  =  12.5, P < 0.05; cavity-nesters, χ21  =  16.7, P < 0.001; Table 2). The relative abundance of predators detected during point counts at the 1 km2 neighborhood scale was not generally correlated with annual productivity of territories (Fig. 7A); similar results were obtained using predator abundance from just the forested portion of each study site (forest patch scale; r  =  0.08, n  =  19, P  =  0.76). The strongest relationship was the negative correlation between sciurid abundance in the forest and the proportion of territories of ground-nesting birds that fledged at least one brood (r  =  −0.35, P  =  0.02, n  =  44; Fig. 7C). No independent correlations between other predators and ground-nester productivity or between any predator and shrub-nester productivity were significant (n  =  14, all |r| < 0.40, all P > 0.15). Predator loads were similar at territories that fledged and did not fledge young. Predator UD heights were significantly (P < 0.05) higher at failed territories compared to successful territories for four site-year comparisons, and were higher (as expected), but not significantly so, in 35 other site-year comparisons. In contrast, there were 32 cases in which the average predator load at failed territories was lower than at successful territories, in two cases significantly so. Averaged across sites by years, effect size for all predators combined was small, and not significantly greater than 0 (t70  =  0.2; Pone-tailed  =  0.41; Table 3). However, there tended to be differences in effect sizes for various types of predators (f4,235  =  2.3, P  =  0.06). Corvid effect sizes were more positive than sciurid effect sizes (LSD post-hoc comparison, P  =  0.01). Steller's Jays had the largest positive mean effect size (t16  =  1.6, P  =  0.07), and no other effect sizes approached significance (Table 3). In addition, small effect sizes and large variation resulted in no difference between mean effect size for all territories compared to all nests for corvids and sciurids combined (t27  =  −1.4, P  =  0.18) or for corvids (t18  =  −1.4, P  =  0.17), although the difference was significant for sciurids (t20  =  −2.4, P < 0.05). As an additional comparison of our incidental and corvid mapping data, we tested for the effect of mapping method on the relationship of predator load (as indicated by the UD height at territory locations) to territory failure or success. At only one of 12 sites tested was there a significant interaction effect (f1,17  =  6.3, P  =  0.02), indicating that the method of mapping rarely influenced the results presented above. Discussion In the urbanizing area around Seattle, Washington, the diversity of nest predators and their fundamentally different habitat requirements (Donnelly and Marzluff 2006) forced nesting songbirds to co-occur with some nest predators even if they avoided others. Swainson's Thrushes, American Robins, Spotted Towhees, and Black-headed Grosbeaks, for example, were most abundant in forested portions of neighborhoods and specific forest patches where American Crows and sciurids were uncommon, but where Steller's Jays were abundant. Pacific-slope Flycatchers, Winter Wrens, and Wilson's Warblers also were rare where crows were abundant, but most common where sciurids were abundant. On the other hand, Bushtits and Dark-eyed Juncos were most abundant in the open, developed portions of neighborhoods and forest edges where American Crows were also very abundant. In contrast, Black-throated Gray Warblers and Song Sparrows, while generally forest birds in our area, appeared to avoid most locations where corvids and sciurids were abundant. Predator use of an area at the time of breeding territory selection could fail to indicate predation pressure later in the year. This did not appear to be the case in our study area, as crows and jays began nest-building before most other songbirds, and sciurids were active throughout the year. Rather, the songbirds we studied rarely were able to avoid specific habitat patches with more predators. Conversely, the predator species we observed may have chosen to select habitat with abundant prey (songbirds). Corvids, in particular, have the cognitive abilities to learn over time where there are reliable populations of songbirds and adjust their space use accordingly (Marzluff and Angell 2005). Our study could not distinguish the mechanisms of the positive associations observed between songbirds and their predators, but the possibility of habitat selection by predators in response to prey, in addition to potential avoidance of predators by songbirds, should be considered. Some fine-scale adjustments by songbirds to avoid areas of concentrated nest predator use within forest patches were evident. Pacific-slope Flycatchers, Bewick's Wrens, Winter Wrens, Black-throated Gray Warblers, Song Sparrows, and Dark-eyed Juncos appeared to strongly avoid areas within forest patches that were rich in a single predator species, or all predators combined (within patch scale). Bewick's Wrens, Song Sparrows, Dark-eyed Juncos, and, to a lesser extent, Bushtits and American Robins, did this despite co-occurring with nest predators at larger scales. Morton (2005) observed a similar phenomenon in Ovenbirds, which avoided interior forest with high chipmunk abundance. In our study, such fine-scale adjustments by songbirds occurred most often in response to Steller's Jays. These abundant nest predators appear to prey on nests incidentally to their other daily activities (Vigallon and Marzluff 2005a), so greater jay abundance may in fact translate into greater predation pressure. This would explain why songbirds breeding in areas used frequently by jays or crows tended to fledge young less often than songbirds nesting away from these predators. Beyond the costs of nesting in areas with abundant Steller's Jays, the reproductive success of most songbirds did not suffer great consequences from living with nest predators. Roughly half of all nests and territories we observed fledged young. There were even unexpected positive associations between sciurid use and successful territories overall, as well as between overall predator use and successful territories at some specific sites. In these cases, greater food availability in high-quality territories could be reflected in both territory productivity and local predator utilization (Block and Brennan 1993, Martin and Joron 2003). Some of our study species may be able to maintain relatively high nest success by adjusting breeding habitat selection (forest patch scale) or breeding territory or nest site selection (within patch scale) to avoid areas with more predators (e.g., Bewick's Wrens, Winter Wrens, and Song Sparrows). Other species seem to be positively correlated with predators at different scales, but still maintain high nest success (e.g., Bushtits, Spotted Towhees, and Dark-eyed Juncos), perhaps by effectively concealing nests. Of course, there is no reason that species could not employ both strategies, as perhaps Dark-eyed Juncos do (co-occur with all predators, but avoid areas within forest patches with more Steller's Jays). These more successful species are all permanent residents, which could make them better equipped to respond to predation pressure, and are all ground- or cavity-nesters, which could make them less vulnerable to nest predation overall (Donnelly and Marzluff 2004a, Blewett and Marzluff 2005). Species may also be able to compensate for early nest failure by making multiple nesting attempts per year. Such compensation is indicated by the lower association between predator habitat use and annual territory success than between predator use and individual nest success. Species with lower territory and nest success were mostly shrub- and open-cup nesters, and Neotropical migrants. By studying the success of nests and the annual productivity of some territorial species, we were able to conclude that the reproductive success of Black-throated Gray Warblers, Wilson's Warblers, and Black-headed Grosbeaks was low, with nest predation being one likely cause. Our study was not designed to test whether the forest patches in our study sites were being selected (compared to other available habitat), so it is difficult to say whether or not they are functioning as ‘ecological traps.’ Further research could focus on causes of low productivity and whether these species' assessments of habitat quality lead them into an ecological trap. Additional research could also help elucidate the relative utility of nest monitoring versus annual territory monitoring. Our results suggest that both methods are necessary to fully understand reproductive success of a diverse assemblage of songbirds—some species' nests are found in sufficient quantities, and other species are reliably observed within their territories. But for a few species for which large numbers of nests can be found and where breeding activity can be monitored within many territories, both methods appear equivalent and either method alone should be sufficient (Bonifait et al. 2006). Our inability to closely link predator use of nesting territories with nesting success likely comes in part from the fact that we did not map the occurrence of all known nest predators, particularly nocturnal predators. Snakes, opossums, raccoons, and mice were not mapped, yet we observed garter snakes and opossums preying on video-monitored nests and inferred mouse predation from signs at nests (chewed brain cases of nestlings). Accipiters were also observed preying on nests, but our accuracy at mapping these secretive predators was likely below that for more vocal and visible corvids and sciurids. In our study area, the abundance of rats is negatively correlated with artificial nest survival time (Donnelly 2002). In general, more complete quantification of the nest predator community should improve the ability to relate nesting success to nest predator occurrence (Marzluff and Restani 1999). However, in our case, better quantification of all relevant predators may not dramatically change our findings, because approximately half of songbird pairs and nests in our study were successful. High nesting success by most species in the presence of abundant nest predators is surprising, especially in urbanizing ecosystems where large numbers of native and introduced predators are common (Nilon et al. 1995, Soulé et al. 1998, Crooks and Soulé 1999). Successful nesting in western Washington may occur for several reasons: (1) American Crows, while abundant in developed portions of neighborhoods, are rare in our coniferous forests where human subsidies are uncommon (Marzluff, McGowan et al. 2001, Neatherlin and Marzluff 2004; JCW and JMM, unpubl. data); (2) Domestic cats, which are common predators in many urban settings (Fitzgerald and Turner 2000, Liberg et al. 2000), are rare in our forests, perhaps because of the cool, wet climate and abundant coyotes that regularly eat them (Quinn 1995); (3) Steller's Jays are important, but relatively rare, nest predators in suburban forests because they are sensitive to human disturbance and nest predation (Vigallon and Marzluff 2005b); (4) The disturbance of native vegetation in forest fragments is not extreme. Exotics are just becoming established in most of our study sites and human activity is mostly limited to hiking and biking on trails; (5) Feeders are common adjacent to forest patches and may effectively supplement nesting songbirds; (6) Extensive and nearby wildlands may bolster suburban songbird populations. It is possible that breeding adult survival is low in our forests (in part because of abundant predators), but immigration from less disturbed wildlands quickly restocks lost breeders; and (7) Most of the forests we studied have only been in developed neighborhoods for a few decades. Thus, the high nesting success and species diversity (Marzluff 2005) we have documented may change in the coming centuries of increasingly intensive human appropriation of natural areas. If nesting success declines and functional connections to distant reserves break, the low-density suburban populations of many songbirds will quickly crash. Predator assemblages will likely change as development continues and forest is removed. Our correlations suggest that percent forest cover at the neighborhood scale is positively associated with American Crow abundance and negatively associated with Douglas' squirrel and Common Raven abundance. As the ranges of these species expand or contract, we would expect a parallel response in the productivity of species sensitive to these predators. As forest cover continues to decline, a loss of Douglas' squirrels and Common Ravens may reduce predation pressure for forest-nesting songbirds, but an increase in American Crows may increase predation of songbirds nesting in developed areas and along forest edges. Steller's Jays are likely to be important drivers of future nest predation, as they were during our study, and their reliance on forest edges and openings (Marzluff et al. 2004) rather than subdivisions (Vigallon and Marzluff 2005b) suggests that future predator assemblages in urban forest remnants may be impoverished to the benefit of nesting songbirds. Acknowledgments We thank Roarke Donnelly, Tina Blewett, Cara Ianni, and many urban songbird and crow project field assistants for their help collecting data. Many private landowners graciously gave us their permission to conduct research on their properties. Jeff Hepinstall from the University of Washington's Urban Ecology Research Lab provided assistance with land cover classifications and forest cover calculations. This research was supported by the University of Washington (Tools for Transformation Fund), the National Science Foundation (grants DEB-9875041, BCS 0120024, and IGERT-0114351), and the University of Washingtons College of Forest Resources, particularly its Rachel Woods Endowed Graduate Program and the Denman Sustainable Resource Sciences Professorship. Literature Cited Alberti , M. , E. Botsford , and A. Cohen . 2001 . Quantifying the urban gradient: linking urban planning and ecology. 89 – 115 . In Marzluff , J. M. , R. Bowman , and R. Donnelly , editors. eds . Avian ecology and conservation in an urbanizing world Kluwer Academic Publishers . Boston, MA . Blewett , C. M. , and J. M. Marzluff . 2005 . Effects of urban sprawl on snags and the abundance and productivity of cavity-nesting birds. Condor 107 : 678 – 693 . Google Scholar Crossref Search ADS WorldCat Block , W. M. , and L. A. Brennan . 1993 . The habitat concept in ornithology: theory and applications. Current Ornithology 11 : 35 – 91 . Google Scholar Crossref Search ADS WorldCat Bonifait , S. , M-A. Villard , and D. Paulin . 2006 . An index of reproductive activity provides an accurate estimate of the reproductive success of Palm Warblers. Journal of Field Ornithology 77 : 302 – 309 . Google Scholar Crossref Search ADS WorldCat Borgmann , K. L. , and A. D. Rodewald . 2004 . Nest predation in an urbanizing landscape: the role of exotic shrubs. Ecological Applications 14 : 1757 – 1765 . Google Scholar Crossref Search ADS WorldCat Bradley , J. E. , and J. M. Marzluff . 2003 . Rodents as nest predators: influences on predatory behavior and consequences to nesting birds. Auk 120 : 1180 – 1187 . Google Scholar Crossref Search ADS WorldCat Christoferson , L. L. , and M. L. Morrison . 2001 . Integrating methods to determine breeding and nesting status of three western songbirds. Wildlife Society Bulletin 29 : 688 – 696 . WorldCat Cody , M. L. 1985 . An introduction to habitat selection in birds. 4 – 56 . In Cody , M. L. , editor. ed . Habitat selection in birds Academic Press . San Diego, CA . Crooks , K. R. , and M. E. Soulé . 1999 . Mesopredator release and avifaunal extinctions in a fragmented system. Nature 400 : 563 – 566 . Google Scholar Crossref Search ADS WorldCat Curio , E. 1976 . The ethology of predation Springer-Verlag . New York . Donnelly , R. 2002 . Design of habitat reserves and settlements for bird conservation in the Seattle metropolitan area. Ph.D. dissertation, . University of Washington . Seattle, WA . Donnelly , R. , and J. M. Marzluff . 2004a . Importance of reserve size and landscape context to urban bird conservation. Conservation Biology 18 : 733 – 745 . Google Scholar Crossref Search ADS WorldCat Donnelly , R. , and J. M. Marzluff . 2004b . Designing research to advance the management of birds in urbanizing areas. 114 – 122 . In Shaw , W. W. , L. K. Harris , and L. Vandruff , editors. eds . Proceedings of the 4th international symposium on urban wildlife conservation. May 1–5, 1999 University of Arizona Press . Tucson, AZ . Donnelly , R. , and J. M. Marzluff . 2006 . Relative importance of habitat quantity, structure, and spatial pattern to birds in urbanizing environments. Urban Ecosystems 9 : 99 – 117 . Google Scholar Crossref Search ADS WorldCat Dwernychuk , L. W. , and D. A. Boag . 1972 . Ducks nesting in association with gulls – an ecological trap. Canadian Journal of Zoology 50 : 559 – 563 . Google Scholar Crossref Search ADS WorldCat Fitzgerald , B. M. , and D. C. Turner . 2000 . Hunting behaviour of domestic cats and their impact on prey populations. 151 – 175 . In Turner , D. C. , and P. Bateson , editors. eds . The domestic cat Cambridge University Press . Cambridge, UK . Forstmeier , W. , and I. Weiss . 2004 . Adaptive plasticity in nest-site selection in response to changing predation risk. Oikos 104 : 487 – 499 . Google Scholar Crossref Search ADS WorldCat Franklin , J. F. , and C. T. Dyrness . 1988 . Natural vegetation of Oregon and Washington. 2nd ed Oregon State University Press . Corvallis, OR . Gates , J. E. , and L. W. Gysel . 1978 . Avian nest dispersion and fledging success in field-forest ecotones. Ecology 59 : 871 – 883 . Google Scholar Crossref Search ADS WorldCat Hansen , A. J. , R. L. Knight , J. M. Marzluff , S. Powell , K. Brown , P. H. Gude , and K. Jones . 2005 . Effects of exurban development on biodiversity: patterns, mechanisms, and research needs. Ecological Applications 15 : 1893 – 1905 . Google Scholar Crossref Search ADS WorldCat Hooge , P. N. , and B. Eichenlaub . 1997 . ANIMAL MOVEMENT extension to ArcView, version 1.1 Alaska Science Center Biological Science Office, U.S. Geological Survey . Anchorage, AK . Hutto , R. L. 1985 . Habitat selection by nonbreeding, migratory land birds. 455 – 476 . In Cody , M. L. , editor. ed . Habitat selection in birds Academic Press . San Diego, CA . International Bird Census Committee 1970 . An international standard for a mapping method in bird census work recommended by the International Bird Census Committee. Audubon Field Notes 24 : 722 – 726 . WorldCat Kernohan , B. J. , R. A. Gitzen , and J. J. Millspaugh . 2001 . Analysis of animal space use and movements. 125 – 166 . In Millspaugh , J. J. , and J. M. Marzluff , editors. eds . Radio tracking and animal populations Academic Press . San Diego, CA . Knick , S. T. , and J. T. Rotenberry . 2000 . Ghosts of habitats past: contributions of landscape change to current habitats used by shrubland birds. Ecology 81 : 220 – 227 . Google Scholar Crossref Search ADS WorldCat Liberg , O. , M. Sandell , D. Pontier , and E. Natoli . 2000 . Density, spatial organisation and reproductive tactics in the domestic cat and other felids. 119 – 147 . In Turner , D. C. , and P. Bateson , editors. eds . The domestic cat Cambridge University Press . Cambridge, UK . Luginbuhl , J. M. , J. M. Marzluff , J. E. Bradley , M. G. Raphael , and D. E. Varland . 2001 . Corvid survey techniques and the relationship between corvid relative abundance and nest predation. Journal of Field Ornithology 72 : 556 – 572 . Google Scholar Crossref Search ADS WorldCat Martin , J-L. , and M. Joron . 2003 . Nest predation in forest birds: influence of predator type and predator's habitat quality. Oikos 102 : 641 – 653 . Google Scholar Crossref Search ADS WorldCat Martin , T. E. 1993 . Nest predation and nest sites. BioScience 43 : 523 – 532 . Google Scholar Crossref Search ADS WorldCat Martin , T. E. 1998 . Are microhabitat preferences of coexisting species under selection and adaptive?. Ecology 79 : 656 – 670 . Google Scholar Crossref Search ADS WorldCat Martin , T. E. , P. R. Martin , C. R. Olson , B. J. Heidinger , and J. J. Fontaine . 2000 . Parental care and clutch sizes in North and South American birds. Science 287 : 1482 – 1485 . Google Scholar Crossref Search ADS PubMed WorldCat Marzluff , J. M. 1985 . Behavior at a Pinyon Jay nest in response to predation. Condor 87 : 559 – 561 . Google Scholar Crossref Search ADS WorldCat Marzluff , J. M. 2005 . Island biogeography for an urbanizing world: how extinction and colonization may determine biological diversity in human-dominated landscapes. Urban Ecosystems 8 : 157 – 177 . Google Scholar Crossref Search ADS WorldCat Marzluff , J. M. , and T. Angell . 2005 . In the company of crows and ravens Yale University Press . New Haven, CT . Marzluff , J. M. , R. Bowman , and R. Donnelly . 2001 . A historical perspective on urban bird research: trends, terms, and approaches. 1 – 17 . In Marzluff , J. M. , R. Bowman , and R. Donnelly , editors. eds . Avian ecology and conservation in an urbanizing world Kluwer Academic Publishers . Boston, MA . Marzluff , J. M. , K. J. McGowan , R. Donnelly , and R. L. Knight . 2001 . Causes and consequences of expanding American Crow populations. 331 – 363 . In Marzluff , J. M. , R. Bowman , and R. Donnelly , editors. eds . Avian ecology and conservation in an urbanizing world Kluwer Academic Publishers . Boston, MA . Marzluff , J. M. , J. J. Millspaugh , P. Hurvitz , and M. S. Handcock . 2004 . Relating resources to a probabilistic measure of space use: forest fragments and Steller's Jays. Ecology 85 : 1411 – 1427 . Google Scholar Crossref Search ADS WorldCat Marzluff , J. M. , and M. Restani . 1999 . The effects of forest fragmentation on avian nest predation. 155 – 169 . In Rochelle , J. A. , L. A. Lehmann , and J. Wisniewski , editors. eds . Forest wildlife and fragmentation: management and implications Brill . Leiden, The Netherlands . Millspaugh , J. J. , R. M. Nielson , L. McDonald , J. M. Marzluff , R. A. Gitzen , C. D. Rittenhouse , M. W. Hubbard , and S. L. Sherriff . 2006 . Analysis of resource selection using utilization distributions. Journal of Wildlife Management 70 : 384 – 395 . Google Scholar Crossref Search ADS WorldCat Morton , E. S. 2005 . Predation and variation in breeding habitat use in the Ovenbird, with special reference to breeding habitat selection in northwestern Pennsylvania. Wilson Bulletin 117 : 327 – 335 . Google Scholar Crossref Search ADS WorldCat Neatherlin , E. A. , and J. M. Marzluff . 2004 . Responses of American Crow populations to campgrounds in remote native forest landscapes. Journal of Wildlife Management 8 : 708 – 718 . Google Scholar Crossref Search ADS WorldCat Nilon , C. H. , C. N. Long , and W. C. Zipperer . 1995 . Effects of wildland development on forest bird communities. Landscape and Urban Planning 32 : 81 – 92 . Google Scholar Crossref Search ADS WorldCat Peluc , S. I. 2006 . The role of ecological factors and microclimate shaping nest site selection and reproductive behaviors in song birds. Ph.D. dissertation, . University of California . Riverside, CA . Pietz , P. J. , and D. A. Granfors . 2005 . Parental nest defense on videotape: more reality than “myth.”. Auk 122 : 701 – 705 . Google Scholar Crossref Search ADS WorldCat Quinn , T. 1995 . Using public sighting information to investigate coyote use of urban habitat. Journal of Wildlife Management 59 : 238 – 245 . Google Scholar Crossref Search ADS WorldCat Ralph , C. J. , G. R. Geupel , P. Pyle , T. E. Martin , and D. F. DeSante . 1993 . Handbook of field methods for monitoring landbirds. USDA Forest Service General Technical Report PSW-GTR-144. Raphael , M. G. , D. E. Mack , J. M. Marzluff , and J. M. Luginbuhl . 2002 . Effects of forest fragmentation on populations of the Marbled Murrelet. Studies in Avian Biology 221 – 235 . Ricklefs , R. E. 1969 . An analysis of nesting mortality in birds. Smithsonian Contributions to Zoology 9 : 1 – 48 . Google Scholar Crossref Search ADS WorldCat Robinson , L. , J. P. Newell , and J. M. Marzluff . 2005 . Twenty-five years of sprawl in the Seattle region: growth management responses and implications for conservation. Landscape and Urban Planning 71 : 51 – 72 . Google Scholar Crossref Search ADS WorldCat Rohila , C. M. 2002 . Urbanization in the greater Seattle, Washington area: impacts on vegetation, snags, and cavity-nesting birds University of Washington . Seattle, WA . Rotenberry , J. T. , and J. A. Wiens . 1989 . Reproductive biology of shrubsteppe passerine birds: geographical and temporal variation in clutch size, brood size, and fledging success. Condor 91 : 1 – 14 . Google Scholar Crossref Search ADS WorldCat Schlaepfer , M. A. , M. C. Runge , and P. W. Sherman . 2002 . Ecological and evolutionary traps. Trends in Ecology & Evolution 17 : 474 – 480 . Google Scholar Crossref Search ADS WorldCat Schmidt , K. A. , and C. J. Whelan . 1999 . Effects of exotic Lonicera and Rhamnus on songbird nest predation. Conservation Biology 13 : 1502 – 1506 . Google Scholar Crossref Search ADS WorldCat Sieving , K. E. , and M. F. Willson . 1998 . Nest predation and avian species diversity in northwestern forest understory. Ecology 79 : 2391 – 2402 . Google Scholar Crossref Search ADS WorldCat Silverman , B. W. 1986 . Density estimation for statistics and data analysis Chapman and Hall . London . Soulé , M. E. , D. T. Bolger , A. C. Alberts , J. Wright , M. Sorice , and S. Hill . 1998 . Reconstructed dynamics of rapid extinctions of chaparral-requiring birds in urban habitat islands. Conservation Biology 2 : 75 – 92 . Google Scholar Crossref Search ADS WorldCat SPSS 2001 . Statistical package for social sciences. Version 10.1.3 SPSS, Inc . Chicago . van Winkle , W. 1975 . Comparison of several probabilistic home-range models. Journal of Wildlife Management 39 : 118 – 123 . Google Scholar Crossref Search ADS WorldCat Vickery , P. D. , M. L. Hunter Jr , and J. V. Wells . 1992 . Use of a new reproductive index to evaluate relationship between habitat quality and breeding success. Auk 109 : 697 – 705 . Google Scholar Crossref Search ADS WorldCat Vigallon , S. M. , and J. M. Marzluff . 2005a . Is nest predation by Steller's Jays (Cyanocitta stelleri) incidental or the result of a specialized search strategy?. Auk 122 : 36 – 49 . Google Scholar Crossref Search ADS WorldCat Vigallon , S. M. , and J. M. Marzluff . 2005b . Abundance, nest sites, and nesting success of Steller's Jays along a gradient of urbanization in western Washington. Northwest Science 79 : 22 – 27 . WorldCat White , G. C. , and R. A. Garrott . 1990 . Analysis of wildlife radio-tracking data Academic Press . San Diego, CA . Worton , B. J. 1989 . Kernel methods for estimating the utilization distribution in home-range studies. Ecology 70 : 164 – 168 . Google Scholar Crossref Search ADS WorldCat Zar , J. H. 1996 . Biostatistical analysis. 3rd ed Prentice Hall . Upper Saddle River, NJ . Author notes E-mail: corvid@u.washington.edu © The Cooper Ornithological Society 2007
TI - Consequences of Habitat Utilization by Nest Predators and Breeding Songbirds Across Multiple Scales in an Urbanizing LandscapeConsecuencias De La Utilización De Hábitat Por Parte De Depredadores De Nidos Y De Aves Canoras Reproductivas A Través De Múltiples Escalas En Un Paisaje UrbanizadoHabitat Utilization by Nest Predators
JO - Condor: Ornithological Applications
DO - 10.1093/condor/109.3.516
DA - 2007-08-01
UR - https://www.deepdyve.com/lp/oxford-university-press/consequences-of-habitat-utilization-by-nest-predators-and-breeding-AVCpbAGKvs
SP - 516
VL - 109
IS - 3
DP - DeepDyve
ER -