Southeastern Naturalist D.R. Brown and T. Weinkam 2014 104 Vol. 13, Special Issue 6 Predicting Bird Community Changes to Invasion of Hemlock Woolly Adelgid in Kentucky David R. Brown1,* and Todd Weinkam1 Abstract - Adelges tsugae (Hemlock Woolly Adelgid [HWA]) is rapidly spreading across Kentucky, yet the potential effects on bird communities of the region are not known. We surveyed birds in Tsuga canadensis (Eastern Hemlock) and other forest types across the Appalachian mountain region of Kentucky during the early stages of infestation when HWA was absent from most sites. Based on associations between birds and forest types, we identified 4 bird species likely to be negatively affected by Eastern Hemlock decline. For one of these species, Setophaga virens (Black-throated Green Warbler), we used landscape metrics of forest composition to make predictions about the density of Eastern Hemlock necessary for persistence. This type of information can help managers establish control strategies for HWA. Introduction Non-native invasive species have the potential to dramatically disrupt ecosystem dynamics at rapid rates. The effects of such disturbances can be extensive and result in a cascade of biological responses (Crooks 2002, Vitousek et al. 1997). The loss of a foundation species, such as Tsuga canadensis (L.) Carrière (Eastern Hemlock, hereafter Hemlock), can have dramatic effects on the composition of communities at different trophic levels (Ellison et al. 2005). In this study, we examine the bird communities of intact Hemlock forests in Kentucky, and predict changes to these communities in response to rapid Hemlock decline caused by an invasive pest, Adelges tsugae Annand (Hemlock Woolly Adelgid, hereafter HWA). Hemlock is a shade-tolerant conifer with a range that extends across much of eastern North America, including eastern Kentucky. In the southern Appalachians, most Hemlock occurs along stream drainages (Narayanaraj et al. 2010), but in Kentucky stands also occasionally occur on the north- and east-facing slopes of ridges. Where abundant, Hemlock plays a foundation role by stabilizing various ecosystem dynamics and influencing composition of plant and animal communities (Ellison et al. 2005). Hemlocks influence forests’ environmental and ecological conditions including local air and hydrologic temperatures, soil pH, and the physical structure of the forest (Ellison et al. 2005, Snyder et al. 2002), which influence the biologic community, forest, and associated streams. HWA is an exotic invasive insect responsible for widespread mortality of Hemlock across eastern North America. HWA feeds on plant nutrients in leaf bases, causing decline in tree health and eventually mortality. It appears to disperse during 1Department of Biological Sciences, Eastern Kentucky University, Richmond, KY 40475. *Corresponding author - david.brown@eku.edu. Manuscript Editor: Jason Davis Forest Impacts and Ecosystem Effects of the Hemlock Woolly Adelgid in the Eastern US 2014 Southeastern Naturalist 13(Special Issue 6):104–116 Southeastern Naturalist 105 D.R. Brown and T. Weinkam 2014 Vol. 13, Special Issue 6 the first instar stage via wind, humans, birds, and other animals (McClure 1990). In recent decades, HWA has spread over most of the native range of Hemlock resulting in changes to forest structure and ecosystem functions (Ellison et al. 2005, Nuckolls et al. 2009, Stadler et al. 2005). The range of HWA has recently expanded to include parts of Kentucky, and the insect is predicted to spread throughout the range of Hemlock in Kentucky (Clark 2010). In the southern Appalachians, Hemlock mortality exceeded 85% seven years after invasion (Ford et al. 2012). Forest simulation models suggest that Hemlock in Kentucky could be completely extirpated within 20 years of HWA invasion, leading to altered forest structure with increased canopy gaps, expansion of Rhododendron spp. (rhododendron), and succession to forests dominated by deciduous hardwood trees (Spaulding and Rieske 2010). The loss of Hemlock and its replacement by other species is also likely to have effects on animal communities. Several studies have highlighted strong positive associations of bird species, including Empidonax virescens (Acadian Flycatcher), Vireo solitarius (Blueheaded Vireo), Setophaga virens (Black-throated Green Warbler), Setophaga fusca Statius Muller (Blackburnian Warbler), and Catharus guttatus Pallas (Hermit Thrush) with Hemlock forests in other regions of the eastern United States (Ross et al. 2004, Tingley et al. 2002). For at least one species, Acadian Flycatcher, breeding density was negatively correlated with HWA-induced Hemlock defoliation, indicating that this species may experience decline due to HWA (Allen et al. 2009). A number of bird species have been associated with positive responses to Hemlock decline due to HWA including Contopus virens (Eastern Wood-Pewee), Molothrus ater Boddaert (Brown-headed Cowbird), Baeolophus bicolor L. (Tufted Titmouse), Sitta carolinensis (White-breasted Nuthatch), Vireo olivaceus (Red-eyed Vireo), Setophaga citrina Boddaert (Hooded Warbler), and several woodpecker species (Tingley et al. 2002). As HWA spreads through Kentucky, land managers are using a combination of chemical and biological control methods to protect individual and forest stands. Thus far, the biological methods, primarily predaceous beetle releases, have been localized and their efficacy largely unkown. Chemical treatment for HWA has been the predominant management approach, but land managers do not have the resources to treat every tree. Unfortunately, there are few guidelines and almost no research on the number, density, spatial extent, and age-classes of trees that need to be preserved to maintain biodiversity and ecosystem functions. During this critical period of early invasion in Kentucky, land managers need accurate information on how bird communities will respond to HWA infestation and how to best manage forests to protect at-risk species. Large-scale coordinated efforts to treat Hemlock for HWA are underway, but these efforts lack specific guidance on the extent and density of Hemlock required to preserve the biological communities that associate with these forests. Our objectives were to describe the bird communities in Kentucky’s Hemlock forests and compare the overall bird community composition to other forest types within the Appalachian mountain region of Kentucky. We also identified Southeastern Naturalist D.R. Brown and T. Weinkam 2014 106 Vol. 13, Special Issue 6 bird species likely to benefit or decline in response to changes in the forest community composition and structure as HWA invades Kentucky. Finally, we used landscape metrics of forest community composition to make predictions about the density of Hemlock necessary for persistence of one at-risk species, Black-throated Green Warbler. Field Site Description and Methods This study took place within Hemlock forests of the Appalachian Mountain region of southeast Kentucky (Fig. 1). This area is in the Coal Fields physiographic region of the Cumberland Plateau, and includes the Level III Ecoregions of the southwestern Appalachians, the central Appalachians, and the western Allegheny Plateau (Woods et al. 2002). We used bird-survey data collected in 2009 from 271 locations including Hemlock forests (n = 123) and other forest types (n = 148). At each location, we assessed the forest community and stand structure using standard forestry techniques (James and Shugart 1970). Protocols differed slightly at sites on US National Forest (n = 16). For all sites, we used a general classification based on the dominant overstory tree species. We categorized sites as Hemlock if it was among the three most dominant overstory tree species within a 50-m radius of the bird-survey location; otherwise, we considered sites non-Hemlock forest. Non-Hemlock habitat was typically deciduous forest with mixed-mesophytic or Quercus spp.(oak)-Carya spp. (hickory) Figure 1. Study locations were clustered on public lands in the Appalachian Mountain region of Kentucky. Solid circles are study locations with Hemlock present. Open circles are forested locations with Hemlock absent within 50-m radius of the site location. Southeastern Naturalist 107 D.R. Brown and T. Weinkam 2014 Vol. 13, Special Issue 6 communities (Braun 1950). Many survey locations were clustered on public lands and were established as part of other land-bird monitoring programs. Birds were surveyed by biologists with special training in bird identification, including personnel from Eastern Kentucky University, Richmond, KY; Kentucky Department of Fish and Wildlife Resources (KDFWR), Frankfort, KY; and the US Forest Service, Daniel Boone National Forest, KY. Bird surveys were conducted as unlimited-distance point counts for 10 minutes. At a subset of 75 sites, we surveyed for an additional 5 minutes using playback of song and calls of Limnothlypis swainsonii Audubon (Swainson’s Warbler) to increase detections of this secretive species. We used non-metric multidimensional scaling (NMS) to illustrate the overall difference in bird communities in Hemlock stands compared to other forest types (Legendre and Legendre 2012). NMS is robust to matrices with many zeroes, which are typical of animal community data, because it uses rank-distances between sampling locations (Clarke 1993, McCune and Grace 2002). This analysis was conducted in Program R using Package Vegan (R Development Core Team 2012). We used Bray-Curtis dissimilarity to construct the distance matrix and limited the analysis to a two-axis solution. To describe the strength of Hemlock habitat associations for individual bird species and to develop predictions for which species will be affected by HWA, we used logistic regression to model presence/absence of bird species based on presence/ absence of Hemlock within 50 m of each survey location using the same surveys (n = 271) described above. To control for variation attributable to the clustered distribution of survey locations, we statistically blocked our data based on the local watershed at the 12-digit hydrologic unit code (HUC-12) scale (n = 24 watersheds). We used the HUC-12 scale because it better reflected the clustered pattern of our sampling design than the next available landscape data at smaller (HUC-14) and larger (HUC-10) scales. We applied this analysis to the 23 bird species that we detected at 50 or more locations. We report the results using Wald’s statistic, P values, and Odd’s ratios. We developed a model of the Hemlock density at the sub-watershed scale that best predicted the occurrence of one at-risk bird species, the Black-throated Green Warbler. Of all bird species within the range of Hemlock, that is the one cited most frequently as being positively associated with Hemlock (Becker et al. 2008, Howe and Mossman 1995, Keller 2004, Tingley et al. 2002). Hemlock density was estimated from a Hemlock distribution map developed specifically for eastern Kentucky using a decision-tree classification based on topographic and climatic environmental variables, and Landsat 7 ETM+ satellite imagery at a resolution of 30 x 30 m (Clark 2010). The Hemlock distribution map covered 25,790 km2 of eastern Kentucky, but did not provide complete coverage of our study area, so we reduced our sample for this analysis by 10 bird-sampling sites. For each bird-survey site, we used ArcGIS v. 9.2 (Environmental Systems Research Institute 2005) to create landscape-sampling polygons based on the sub-watershed described by the USGS 14-digit hydrologic unit code (HUC-14) (n = 70 sub-watersheds with mean size ± SD = 15.0 ± 15.8 km2). Southeastern Naturalist D.R. Brown and T. Weinkam 2014 108 Vol. 13, Special Issue 6 Sub-watersheds included 1–12 bird-sampling locations (mean ± SD, n = 3.8 ± 3.2). We used ArcGIS Spatial Analyst extension to calculate the total number of cells dominated by Hemlock forest within each sub-watershed and then calculated the density of Hemlock habitat in each sub-watershed as the area of Hemlock divided by the area of each HUC-14 sub-watershed. We used the HUC-14 subwatershed scale because this is the smallest watershed-scale data available and we assumed it better reflected the scale at which land managers would make forestry decisions related to management for HWA and its impacts. We calculated the average abundance of Black-throated Green Warblers among all point-count sampling locations within each sub-watershed. The statistical model included the average bird abundance per sub-watershed as the response and a single predictor variable, the total density of Hemlock within the HUC-14 sub-watershed. We used a generalized linear model with a Tweedie distribution and a log link. The Tweedie distribution was appropriate because we used the average bird abundance which had frequent zeros but was otherwise continuously and normally distributed. We used the model-estimated coefficients and standard errors and the average sub-watershed size of sampled sites (15.73 km2) to estimate the minimum density of Hemlock within sub-watershed landscape patches necessary to support the presence of at least one Black-throated Green Warbler per sample. We used SPSS v 19 to conduct the logistic regressions and generalized linear model. Results Based on the spatial separation of bird communities in Hemlock compared to deciduous forest sites, the NMS ordination analysis (stress = 0.18) showed differences in the composition of bird communities between Hemlock and deciduous forest habitats in the Appalachian mountain region of Kentucky (Fig. 2). Birds are highly mobile, and many of the 75 species included in the analysis are habitat generalists, so despite the moderate degree of overlap between communities, the overall pattern suggests bird community structure differs between Hemlock and deciduous forest. Logistic regression indicated 4 of 25 bird species with high detection frequency were positively associated with Hemlock: Acadian Flycatcher, Blue-headed Vireo, Black-throated Green Warbler, and Mniotilta varia (Black-andwhite Warbler) (Table 1). Black-throated Green Warblers were more than 9 times more likely to be detected in Hemlock than deciduous forest types, and Blue-headed Vireos were almost 6 times more likely to be found in Hemlock forests than in deciduous forests. We found ten species to be negatively associated with Hemlock Table 1). The statistical blocking on clustered study areas was significant for some species, suggesting that this approach was warranted and helped to strengthen our results of the regional patterns of bird associations with Hemlock forest. Swainson’s Warblers were detected at only 2 of 271 locations using standard point-count methodology, and at 19 of 75 locations where supplemental playback of song and call was used following standard point counts. Region-wide, Hemlock forests composed 5% of the landscape. Among the 70 HUC-14 sub-watersheds sampled, the average proportion of area classified Southeastern Naturalist 109 D.R. Brown and T. Weinkam 2014 Vol. 13, Special Issue 6 as Hemlock forest was 7% (± 6% SD). For the generalized linear modeling, the average abundance of Black-throated Green Warblers showed a strong positive exponential relationship with Hemlock density at the HUC-14 sub-watershed scale (Wald’s Χ2 = 7.35, n = 70, P = 0.007, Log link coefficients: intercept = -0.484, Hemlock = 5.63; Fig. 3). The minimum area of Hemlock based on the averagesized sub-watershed necessary for the presence of an average of one Black-throated Green Warbler per sample location is 1.35 km2 (95% CI: 0.06–9.35 km2). Discussion This study describes the bird communities of Hemlock forests as having distinct structure and different species associations than bird assemblages in the dominant deciduous forest of the Appalachian mountain region of Kentucky. Three of the 4 species we found to be positively associated with Hemlock—Acadian Flycatcher, Figure 2. Graphical analysis illustrating difference in bird communities in sites with Hemlock forest (solid circles) compared to deciduous forest sites (open circles) in the Appalachian mountain region of Kentucky. Each dot represents a field site. Axis 1 and 2 represent summary values of abundance for all bird species detected. Southeastern Naturalist D.R. Brown and T. Weinkam 2014 110 Vol. 13, Special Issue 6 Blue-headed Vireo, and Black-throated Green Warbler—have been similarly found to associate with Hemlock in other states (Allen et al. 2009, Becker et al. 2008, Keller 2004, Ross et al. 2004, Tingley et al. 2002). Thus, our results are consistent with other Table 1. Association of bird species with Hemlock forest in the Appalachian mountain region of Kentucky. Detections indicate the number of point-count locations where a species was observed. Positive slopes (b) indicate species that were more likely detected in Hemlock forest, whereas negative slopes indicate species more frequently detected in forest without Hemlock. Odd’s ratios indicate the likelihood that a species was detected in Hemlock forest compared to other forest types. Odds Species Detections Wald’s b P-value ratio Positive associations Empidonax virescens (Vieillot) (Acadian Flycatcher) 124 17.68 1.06 less than 0.001 2.90 Vireo solitarius (Wilson) (Blue-headed Vireo) 74 32.13 1.77 less than 0.001 5.88 Setophaga virens (Gmelin) (Black-throated Green 130 62.10 2.25 less than 0.001 9.52 Warbler) Mniotilta varia (L.) (Black-and-white Warbler) 96 18.19 1.14 less than 0.001 3.12 Negative associations Zenaida macroura (L.) (Mourning Dove) 51 6.29 -0.84 0.012 0.42 Coccyzus americanus (L.) (Yellow-billed Cuckoo) 58 5.67 -0.77 0.017 0.46 Melanerpes carolinus L. (Red-bellied Woodpecker) 67 19.00 -1.46 less than 0.001 0.23 Contopus virens (L.) (Eastern Wood-Peewee) 61 26.21 -2.07 less than 0.001 0.12 Vireo olivaceus (L.) (Red-eyed Vireo) 188 4.36 -0.55 0.037 0.57 Cyanocitta cristata (L.) (Blue Jay) 64 4.93 -0.66 0.026 0.51 Sitta carolinensis Latham (White-breasted Nuthatch) 65 13.44 -1.17 less than 0.001 0.30 Hylocichla mustelina (Gmelin) (Wood Thrush) 115 12.15 -0.91 less than 0.001 0.40 Piranga olivacea (Gmelin) (Scarlet Tanager) 132 14.47 -0.96 less than 0.001 0.38 Passerina cyanea (L.) (Indigo Bunting) 50 6.40 -0.92 0.011 0.39 Figure 3. Average s u b - w a t e r s h e d - scale abundance per sampling location of Blackthroated Green Warblers in relation to the relative density of Hemlock within subwatersheds as delimited by 14-digit Hydrologic Unit Codes. The solid line represents the exponential best-fit model and dashed lines are 95% confidence intervals. Southeastern Naturalist 111 D.R. Brown and T. Weinkam 2014 Vol. 13, Special Issue 6 research from across the range of Hemlock and provide state-specific information on bird–Hemlock associations. We also found Black-and-white Warbler to be positively associated with Hemlock, but to a lesser degree than the Blue-headed Vireos and Black-throated Green Warblers. No other studies have identified Black-and-white Warblers as Hemlock associates, and one study found this species to be negatively associated with Hemlock (Keller 2004). Most other bird species previously described as positively associated with Hemlock but not found to be significantly associated with Hemlock in this study, including Setophaga caerulescens Gmelin (Blackthroated Blue Warbler) and Sitta canadensis L. (Red-breasted Nuthatch), occur in only isolated locations within our study region, and their rarity in the study region may mask a possible association. Swainson’s Warbler, a species of special concern in Kentucky, may associate with Hemlock, so it could also be at further risk from the spread of HWA. This species preferentially nests in habitats with dense shrub layers (Bassett-Touchell and Stouffer 2006, Graves 2002), including in young Hemlock in the Appalachian mountains of South Carolina (Lanham and Miller 2006). Swainson’s Warblers are notoriously difficult to detect using traditional survey methods, and our detection rate was less than 1%, yet we detected Swainson’s Warblers at more than 25% of Hemlock sites where we used playback vocalizations of the bird’s song and call. Thus, this species may also be associated with Hemlock habitat in Kentucky, but additional targeted research is needed. It is possible that the positively associated species are actually selecting habitat based on environmental conditions or landscape cues other than the presence of Hemlock. For instance, most Hemlock in Kentucky occurs in stream bottoms. The bird species we report to be positively associated with Hemlock might be selecting habitat based on proximity to streams or other topographic features such as slope, aspect, or elevation. If such characteristics are more important to habitat selection than forest composition and structure, then bird communities may change little in response to the decline of Hemlock. Similarly, the species we describe as being associated with Hemlock also occur in other mesic habitats. If habitat selection is based more on environmental conditions than stand structure, then we may see negligible change in the bird communities as Hemlock declines; however, because Hemlock shows keystone ecosystem functions, the loss of Hemlock may shift local conditions to be more xeric and thus lead to indirect effects on bird communities. Our study was not designed to test environmental factors other than the presence or density of Hemlock, and such research might provide managers with additional valuable knowledge. However, regardless of the proximal causes for the habitat associations, these birds may benefit from the presence of Hemlock in subtle ways that influence survival and reproductive success. For instance, Swainson’s Warblers in the Appalachian mountains of South Carolina preferentially nest in young Hemlock, possibly because leaf litter accumulations in the branches serve as nest decoys that reduce nest predation (Lahham and Miller 2006). Additional research focused on the demographic response, home-range use, and foraging behavior of individual species, such as the Black-throated Green Warbler, would allow more Southeastern Naturalist D.R. Brown and T. Weinkam 2014 112 Vol. 13, Special Issue 6 specific predictions about how the bird community and populations may change due to HWA and other future invasions. Our analysis of the relationship between landscape density of Hemlock and the abundance of Black-throated Green Warblers suggests that relatively high densities of Hemlock within watersheds need to be protected to maintain current densities of this bird. Black-throated Green Warblers often use transitional areas between coniferous and deciduous forests, but they associate with some type of coniferous habitat in most of their range (Collins 1983). In eastern Kentucky, Picea spp. (spruce) and Abies spp. (fir) are absent, and Pinus spp. (pine) forms a minor component of the forest community. Thus, Hemlock represents the primary coniferous species available to Black-throated Green Warblers in Kentucky. As Hemlock distribution shrinks, Black-throated Green Warbler populations may be buffered by resources available in nearby deciduous forests, but the gradual decline of this bird species in Kentucky seems likely without HWA management success. Further research will be necessary to determine the spatial arrangements and age-class structure of Hemlock within these areas that will best support Black-throated Green Warbler and other species. In contrast to our findings regarding likely declines in certain bird populations, our results also enable us to predict that several species will benefit from the decline of Hemlock and its replacement by deciduous forest communities. Of the species we describe as negatively associated with living Hemlock, Eastern Wood-peewee and Red-eyed Vireo were similarly found to associate with non-Hemlock forest types elsewhere in the range of Hemlock (Becker et al. 2008, Howe and Mossman 1995, Keller 2004, Tingley et al. 2002). In Connecticut, these species, among others such as woodpeckers, appeared to benefit from the presence of dead Hemlocks (Tingley et al. 2002). Widespread Hemlock mortality has yet to occur in Kentucky, so we may see increases in these species in future years. Based on our findings, we infer that bird communities of Hemlock forests in Kentucky will likely change in predictable patterns as Hemlocks die and are replaced by other tree species. Several bird species appear to be at risk of losing preferred habitat as Hemlock disappears from eastern Kentucky. Species that are the most closely associated with Hemlock (i.e., Acadian Flycatcher, Blue-headed Vireo, and Black-throated Green Warbler) will likely experience the most dramatic declines in abundance due to HWA, and it is possible that these species will disappear entirely from some areas. As Hemlocks are replaced and forests transition to other communities such as oak- or Acer spp. (maple)-dominated forests, some bird species will likely benefit by the addition of these breeding and foraging areas. This study focused on breeding birds and does not address the impact of Hemlock loss on birds during migration or the winter season. Although winter birds in Kentucky forests tend to be generalists in their habitat use, the loss of cover as Hemlocks decline, at least in areas lacking rhododendron or other year-round cover, could lead to local changes in the winter bird community. The KDFWR, the state agency responsible for monitoring biodiversity in Kentucky, has closely followed the spread of HWA. In anticipation of the effects of HWA, KDFWR recently listed the Black-throated Green Warbler as a species of Southeastern Naturalist 113 D.R. Brown and T. Weinkam 2014 Vol. 13, Special Issue 6 greatest conservation concern under the state Wildlife Action Plan (KY DFWR 2013). This proactive approach to management suggests that managers appreciate the risk of HWA. We recommend that resource managers continue to monitor bird populations and forest ecology, and maintain flexibility to adjust management approaches as research advances. Vose et al. (2013) present a framework for HWA management decisions in the southern Appalachians that could be modified and adopted for use in Kentucky. There are at least two potential biases that could have influenced our results. First, almost all of our 70 study sites were located on protected public lands, 2 of which included old-growth forest, which may differ in composition and age-class structure compared to stands on private lands in the region. However, our results should be broadly applicable because of the large spatial scale of the sampling design and the fact that the relative density of Hemlock in our study sites was similar to the regional average, and to the average within public lands in the region (5%). Another potential bias is that HWA had already invaded the most southern sites at the start of this study. Although there was some evidence of infestation at these locations, it was at an early stage with little or no apparent effects on health of trees and no mortality (T. Weinkam, pers. observ.) Thus, the effect of HWA on birds was likely minimal. Because of Hemlock’s foundational role and the patterns of association described here and in other locations, we predict that a shift in bird communities will occur because of HWA. However, in areas of active HWA control, it may be possible to mitigate changes to existing bird communities. Management for HWA typically involves the release of beetles as HWA predators or insecticide treatments of individual trees. For biological control efforts, rearing of predatory beetles is now occurring at several institutions and millions of beetles have been released across the Hemlock’s range (Vose et al. 2013). In some cases, predator populations have become established and appear to be associated with increased health of Hemlock (Mausel et al. 2010, Onken and Reardon 2010). For chemical HWA control, soil soaking and soil injections of insecticide appear to be the most effective and efficient short-term methods and can provide protection for several years (Cowles 2009, Cowles et al. 2006). Recent research on the primary pesticide used for such treatments, imidacloprid, suggests that it has minimal negative effects on nearby stream invertebrates (Churchel et al. 2011) and soil microarthropods (Knoepp et al. 2012). Other research suggests that 1–2 years following insecticide treatment, HWA abundance does not differ from untreated sites (Falcone and DeWald 2010). Research also suggests non-target effects of insecticide treatment on the abundance of Lepidoptera, an important food for insectivorous birds during critical brood-rearing periods (Dilling et al. 2009, Falcone and DeWald 2010). Furthermore, Falcone and Dewald also found that the combined density of 3 insectivorous migrant bird species, Blue-headed Vireo, Black-throated Green Warbler, and Black-throated Blue Warbler, did not differ between imidacloprid-treated and untreated sites (Falcone and DeWald 2010). However, they did find a higher density of these bird species Southeastern Naturalist D.R. Brown and T. Weinkam 2014 114 Vol. 13, Special Issue 6 in areas with relatively healthy Hemlock cover as measured by foliage mass. This finding suggests a tradeoff between food availability and other factors such that treated Hemlock has reduced food availability for birds, but the increased foliar mass provides benefits such as cover from predators or for nest concealment. Biological control efforts may be successful, but forest management strategies will need to consider tradeoffs including the economic cost of treating Hemlock, alternative strategies of treatment, and the potential for complex indirect ecological effects of control (Vose et al. 2013). In Kentucky, approximately 100,000 trees have been treated with insecticide to date, with some areas receiving follow-up applications (Alice Mandt, Kentucky Division of Forestry, Frankfort, KY, pers. comm.). The baseline data we collected for our study provide opportunities to study the impacts of HWA on birds and the efficacy of control efforts in Kentucky. Our findings provide a starting point for long-term monitoring of bird communities in the Hemlock forests of Kentucky and testable predictions of how species abundance will change with altered forest conditions. Our results also provide land managers with specific guidance on the density of Hemlock necessary to prevent local decline and extirpation of one at-risk bird species, the Black-throated Green Warbler. In combination with other recent and ongoing studies of HWA management approaches, this work helps form a basis for informed species-specific decisions. Acknowledgments We thank all of the biologists that conducted the bird surveys. We thank the Kentucky State Nature Preserves Commission, Kentucky State Parks, Cumberland Gap National Historic Park, Kentucky Department of Fish and Wildlife Resources, US Forest Service, Eastern Kentucky University Division of Natural Areas, and private landowners for permission to conduct this research. Songlin Fei provided the Hemlock map for eastern Kentucky. Support was provided by KY EPSCoR. We also appreciate the hard physical work of all the people, including many volunteers, conducting HWA-control work in Kentucky. This is contribution No. 43 of Lilley Cornett Woods Appalachian Ecological Research Station, Eastern Kentucky University, Richmond, KY. Literature Cited Allen, M.C., J. Sheehan, Jr., T.L. Master, and R.S. Mulvihill. 2009. Responses of Acadian Flycatchers (Empidonax virescens) to Hemlock Woolly Adelgid (Adelges tsugae) infestation in Appalachian riparian forests. Auk 126:543–553. Bassett-Touchell, C.A., and P.C. Stouffer. 2006. Habitat selection by Swainson’s Warblers breeding in Loblolly Pine plantations in southeastern Louisiana. Journal of Wildlife Management 70:1013–1019. Becker, D.A., M.C. Brittingham, and C.B. Goguen. 2008. Effects of Hemlock Woolly Adelgid on breeding birds at Fort Indiantown Gap, Pennsylvania. Northeastern Naturalist 15:227–240. Braun, E. 1950. Deciduous Forests of Eastern North America. Hafner Press, New York, NY. Southeastern Naturalist 115 D.R. Brown and T. Weinkam 2014 Vol. 13, Special Issue 6 Churchel, M.A., J.L. Hanula, C.W. Berisford, J.M. Vose, and M.J. Dalusky. 2011. Impact of imidacloprid for control of Hemlock Woolly Adelgid on nearby aquatic macroinvertebrate assemblages. Southern Journal of Applied Forestry 35:26–32. Clark, J. 2010. Distribution of Eastern Hemlock, Tsuga canadensis, in eastern Kentucky and the susceptibillity to invasion by Hemlock Woolly Adelgid, Adelges tsugae. M.Sc. Thesis. University of Kentucky, Lexington, KY. 161 pp. Clarke, K.R. 1993. Non-parametric multivariate analyses of changes in community structure. Journal of Australian Ecology 18:117–143. Collins, S.L. 1983. Geographic variation in habitat structure of the Black-throated Green Warbler (Dendroica virens). Auk 100:382–389. Cowles, R.S. 2009. Optimizing dosage and preventing leaching of imidcloprid for management of Hemlock Woolly Adelgid in forests. Forest Ecology and Management 257:1026–1033. Cowles, R.S., M. Montgomery, and C.A.S.-J. Cheah. 2006. Activity and residues of imidacloprid applied to soil and tree truncks to control Hemlock Woolly Adelgid (Hemiptera: Adelgidae) in Forests. Journal of Economic Entomology 99:1258–1 267. Crooks, J. 2002. Characterizing ecosystem-level consequences of biological invasions: The role of ecosystem engineers. Oikos 91:153–166. Dilling, C., P. Lambdin, J. Grant, and J.R. Rhea. 2009. Community response of insects associated with Eastern Hemlock to imidacloprid and horticultural oil treatments. Environmental Entomology 38:53–66. Ellison, A.M., M.S. Bank, B.D. Clinton, E.A. Colburn, K. Elliott, C.R. Ford, D.R. Foster, B.D. Kloeppel, J.D. Knoepp, G.M. Lovett, J. Mohan, D.A. Orwig, N.L. Rodenhouse, W.V. Sobczak, K.A. Stinson, J.K. Stone, C.M. Swan, J. Thompson, B. Von Holle, and J.R. Webster. 2005. Loss of foundation species: Consequences for the structure and dynamics of forested ecosystems. Frontiers in Ecology and the Env ironment 3:479–486. Environmental Systems Research Institute. 2005. ArcGIS version 9.2. Redlands, CA. Falcone, J.F., and L.E. DeWald. 2010. Comparisons of arthropod and avian assemblages in insecticide-treated and untreated Eastern Hemlock (Tsuga canadensis (L.) Carr.) stands in Great Smoky Mountains National Park, USA. Forest Ecology and Management 260:856–863. Ford, C.R., K. Elliott, B.D. Clinton, B.D. Kloeppel, and J.M. Vose. 2012. Forest dynamics following Eastern Hemlock mortality in the southern Appalachians. Oikos 121:523–536. Graves, G.R. 2002. Habitat characteristics in the core breeding range of the Swainson's Warbler. Wilson Bulletin 114:210–220. Howe, R.W., and M. Mossman. 1995. The significance of Hemlock for breeding birds in the western Great Lakes region. Pp. 125–139, In G. Mroz and A.J. Martin (Eds.). Hemlock Ecology and Management. Department of Forestry, University of Wisconsin-Madison, Madison, WS. 200 pp. James, F.C., and H.H. Shugart. 1970. A quantitative method of habitat description. Audobon Field Notes 24(6):727–736. Keller, D.A. 2004. Association between Eastern Hemlock (Tsuga canadensis) and avian occurrence and nest success in the southern Appalachians. M.Sc. Thesis. University of Tennessee, Knoxville, TN. 99 pp. Kentucky Department of Fish and Wildlife Resources (KY DFWR). 2013. Kentucky’s wildlife action plan. Frankfort, KY. Available online at http://fw.ky.gov/WAP/Pages/ Default.aspx. Knoepp, J.D., J.M. Vose, J.L. Michael, and B.C. Reynolds. 2012. Imidacloprid movement in soils and impacts on soil microarthropods in southern Appalachian Eastern Hemlock stands. Journal of Environmental Quality 41:469–478. Southeastern Naturalist D.R. Brown and T. Weinkam 2014 116 Vol. 13, Special Issue 6 Lanham, J.D., and S.M. Miller. 2006. Monotypic nest site selection by Swainson’s Warbler in the mountains of South Carolina. Southeastern Naturalist 5:2 89–294. Legendre, P., and L. Legendre. 2012. Numerical Ecology. Elsevier, Amsterdam, Netherlands. 990 pp. Mausel, D.L., S.M. Salom, L.T. Kok, and C. Davis. 2010. Establishment of the Hemlock Woolly Adelgid predator, Laricobius nigrinus (Coleoptera: Derodontidae) in the eastern United States. Environmental Entomology 39:440–448. McClure, M.S. 1990. Role of wind, birds, deer, and humans in the dispersal of Hemlock Woolly Adelgid (Homoptera, Adelgidae). Environmental Entomology 19:36–43. McCune, B., and J.B. Grace. 2002. Analysis of ecological communities. MjM Software Design, Glenden Beach, OR. 300 pp. Narayanaraj, G., P.V. Bolstad, K.J. Elliott, and J.M. Vose. 2010. Terrain and landform influence on Tsuga canadensis distribution in the southern Appalachian mountains. Castenea 75:1–18. Nuckolls, A.E., N. Wurzburger, C.R. Ford, R.L. Hendrick, J.M. Vose, and B.D. Kloeppel. 2009. Hemlock declines rapidly with Hemlock Woolly Adelgid infestation: Impacts on the carbon cycle of southern Appalachian forests. Ecosystems 12:179–190. Onken, B., and R. Reardon. 2010. Fifth symposium on Hemlock Woolly Adelgid in the eastern United States. US Forest Service, Asheville, NC. R Development Core Team. 2012. R: A language and environment for statistical computing. Vienna, Austria. Available online at: http://www.R-project.org/. Accessed 5 January 2012. Ross, R.M., L.A. Redell, R.M. Bennett, and J.A. Young. 2004. Mesohabitat use of threatened hemlock forests by breeding birds of the Delaware River basin in northeastern United States. Natural Areas Journal 24:307–315. Snyder, C.D., J.A. Young, D.P. Lemarie, and D.R. Smith. 2002. Influence of Eastern Hemlock (Tsuga canadensis) forests on aquatic invertebrate assemblages in headwater streams. Canadian Journal of Fisheries and Aquatic Sciences 59:262–275. Spaulding, H.L., and L.K. Rieske. 2010. The aftermath of an invasion: Structure and composition of central Appalachian hemlock forests following establishment of the Hemlock Woolly Adelgid, Adelges tsugae. Biological Invasions 12:3135–3143. Stadler, B., T. Muller, D. Orwig, and R. Cobb. 2005. Hemlock Woolly Adelgid in New England forests: Canopy impacts transforming ecosystem processes and landscapes. Ecosystems 8:233–247. Tingley, M.W., D.A. Orwig, R. Field, and G. Motzkin. 2002. Avian response to removal of a forest dominant: Consequences of Hemlock Woolly Adelgid infestations. Journal of Biogeography 29:1505–1516. Vitousek, P.M., C.M. D’Antonio, L.L. Loope, M. Rejmanek, and R. Westbrooks. 1997. Introduced species: A significant component of human-caused global change. New Zealand Journal of Ecology 21:1–16. Vose, J.M., D.N. Wear, A.E. Mayfield, and C.D. Nelson. 2013. Hemlock Woolly Adelgid in the southern Appalachians: Control strategies, ecological impacts, and potential management responses. Forest Ecology and Management 291:209–219. Woods, A.J., J.M. Omernik, W.H. Martin, G.J. Pond, W,M, Andrews, S.M. Call, J.A. Comstock, and D.D. Taylor. 2002. Ecoregions of Kentucky (color poster with map, descriptive text, summary tables, and photographs). Map scale of 1:1,000,000. US Geological Survey, Reston, VA.