|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|||||||||
|
|
||||||||||
|
|
|
|||||||||
|
|
|
|
|
|
|
|
|
|
||
|
|
|
|
|
|||||||
|
|
||||||||||
|
LEETOWN SCIENCE CENTER STUDY PLAN NUMBER: 2055 TITLE: Modeling Stand Vulnerability and Biological Impacts of the Hemlock WoolyAdelgid | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
Principal Investigators:
CONTENTS: LSC "Umbrella" Study PlanAppendix A: AEL component study plan Appendix B: RDL component study plan Appendix C: SAFL component study plan In the last two decades, substantial declines in eastern hemlock (Tsuga canadensis) have been observed throughout its range, resulting in widespread concern (Lapin, 1995; Evans, 1995; Hemlock Wooly Adelgid Working Group, 1995). Hemlock defoliation has largely been attributed to the hemlock woolly adelgid (Adelges tsugae), an exotic aphid-like insect that is native to Japan. Resource managers expect the adelgid to continue to spread and consequently the entire hemlock forest community may be threatened. The impact of removal of this important climax forest species on the ecology of Appalachian forests is poorly understood, but has the potential for significant disturbance to biotic communities by changing energy inputs, micro-climatic environments, and physical habitat structure. The purpose of this research is to provide managers with an empirical model of stand vulnerability that can be used to assess the potential risk of hemlock decline due to hemlock wooly adelgid. Secondly, we propose to determine how avian and aquatic communities associated with hemlock stands are most likely to be impacted Stand Vulnerability Modeling (Phase I). In Shenandoah National Park (SHEN), hemlock defoliation and tree mortality have advanced at such a rapid pace that near complete mortality of some hemlock stands occurred within 3-4 years, while other stands have been less severely impacted. In other areas, such as Delaware Water Gap National Recreation Area (DEWA), the adelgid is present in large numbers but no significant mortality of hemlocks has been observed. The patchy nature of hemlock decline suggests that landscape-level processes may affect hemlock mortality either by regulating the dispersal potential of hemlock woolly adelgid, or by affecting the sensitivity of the trees themselves. For example, hemlock woolly adelgid can disperse by wind or by attaching to migrating animals including birds, deer, and man (McClure 1990). Consequently, prevailing winds, number of and distance to roads, proximity to migrating bird habitats, and stand composition, edge, and topography may contribute to hemlock woolly adelgid dispersal and ultimately to the extent of mortality. Likewise, sub-optimal habitat characteristics such as low soil moisture may be related to landscape variables such as aspect and elevation, which may act synergistically with hemlock woolly adelgid infestation to induce mortality. The relationship between hemlock stand mortality and landscape features needs to be fully examined. This information could be used to develop empirical models designed to predict stand vulnerability from landscape information. National Park managers could use this information to design pro-active strategies in areas where hemlock woolly adelgid has not spread such as Great Smoky Mountains National Park (GSMNP), or where infestation is recent and significant mortality has yet to be evidenced, such as at DEWA. The goal of phase 1 of this research will be to construct an empirical model of hemlock stand vulnerability at site, landscape, and regional scales for SHEN (see study plan, below). Initial investigations will focus on this area due to the presence of HWA and the differential pattern of hemlock infestation and mortality observed in these parks over the past 5-7 years. Results from this modeling effort will be integrated with phase 2 efforts in GSMNP, and DEWA, to predict potential impacts on biotic communities from loss of the eastern hemlock forest component. Potential Ecological Effects (Phase II) Hemlock stands may have co-evolved with other flora and fauna in stable, co-dependent associations. Preliminary data (Ross, unpublished), for example, show exclusive use of hemlock bench and ravine habitats (as opposed to equivalent hardwood sites) by 3 species of wood warbler and vireo (solitary vireo, black-throated green warbler, and blackburnian warbler), mixed use by 2 species (red-eyed vireo and ovenbird), and exclusive use of hardwood benches and ravines by 1 species (American redstart) of those surveyed. As stated, the loss of eastern hemlock has unknown consequences and may involve changes in energy inputs, microclimate, and physical habitat structure for several biotic communities. Hence, there is an urgent need to characterize the contribution of hemlock forests to avian biodiversity in large, forested landscapes. Following the initiation of the stand vulnerability modeling, we will characterize breeding bird populations in matched hardwood and hemlock stands of DEWA and, later, in GSMNP. Much of this data already exists and sampling schemes have already been developed for both areas. We will stratify and compare specific habitat types (bench, ravine, and mid-slope) within hardwood and hemlock stands for bird usage. From those data, models will be developed and coupled with stand vulnerability models to predict changes in avian communities at both areas. In addition to avian communities, loss of hemlock may have tremendous impacts on the aquatic ecosystems with which they are usually associated. Perhaps the most sensitive indicator of changes in aquatic environments are aquatic invertebrates. Invertebrate sampling currently is being conducted at both SHEN and GSMNP. We will supplement this monitoring effort with additional surveys in hemlock communities if necessary. OBJECTIVES: The goal of this research is to develop a hierarchical model to predict hemlock woolly adelgid (HWA) infestation and mortality patterns on eastern hemlock and associate those patterns with biotic and abiotic causative or associative factors. Specifically, we will:
We see three management implications arising from this research:
HYPOTHESIS TO BE TESTED: The null hypothesis of this research is that the level of eastern hemlock decline due to hemlock woolly adelgid is not related to site, landscape, or regional scale environmental factors. Rejection of the null hypothesis would allow spatially explicit models of hemlock decline to be built that may assist in monotoring and management of HWA dispersal and damage. A second null hypothesis is that avian and aquatic communities do not change after adelgid infestation and hemlock mortality. Additional predictive models will be built to assess the potential for biotic community changes due to loss of the eastern hemlock forest component in eastern National Park ecosystems. PROCEDURES: This research is being carried out by three LSC component labs (Aquatic Ecology Laboratory, Research and Development Laboratory, and Southern Appalachian Field Laboratory). The
research is being carried out in two phases: (1) landscape modeling and (2) potential ecosystem
effects. AEL will have lead responsibility for phase 1 with assistance from SAFL. SAFL and
RDL will have lead responsibility for phase 2 with assistance from AEL. Specific research plans
for each phase can be found in component specific study plans (see below: DATA: Data for this project will be collected and managed by the individual components and housed at AEL, SAFL, and RDL (see individual component study plans for details). GIS database will be developed jointly among the three component laboratories. All data bases will be documented according to the federal metadata documentation standard and made available for distribution at the conclusion of the project (except for licensed satellite imagery and other restricted distribution data). SCHEDULE: FY98: Coordination between AEL, RDL, SAFL on study components and plans, initiation of AEL component (phase 1), low-level initiation of SAFL component (phase 1). FY99: Continuation of AEL component (phase 1), initiation of RDL component (phase 2), continuation of low-level SAFL component (phase 1). FY00: Continuation of RDL, component, initation of full SAFL component (phase 2), completion of AEL field and lab components. FY01: Completion of RDL field component, continuation of SAFL component. FY02: Coordination/integration of model results RDL, SAFL, and AEL. Publication of results. DURATION OF STUDY: Begin: June 1998 End: June 2002 SAFETY: Foul weather gear appropriate for the season and temperature in temperate woodlands will be required in the field. First aid kits will be maintained in field vehicles. ANIMAL WELFARE:Not Applicable KEY STAFF: John Young, Aquatic Ecology Laboratory, Leetown, WV Craig Snyder, Aquatic Ecology Laboratory, Leetown, WV Dave Morton, Aquatic Ecology Laboratory, Leetown, WV Joe Clark, Southern Appalachian Field Laboratory, Knoxville, TN Frank van Manen, Southern Appalachian Field Laboratory, Knoxville, TN Bob Ross, Research and Development Laboratory, Wellsboro, PA ESTIMATED COSTS: FTE: AEL:
RDL:
SAFL:
EXPECTED PRODUCTS: Computer model(s) linking landscape attributes to hemlock woolly adelgid dispersal, hemlock health, defoliation risk, and biotic impacts. Paper(s) published in peer-reviewed journals such as Conservation Biology, Ecosystems, Landscape Ecology, Ecological Applications, Remote Sensing of the Environment, or Photogrammetric Engineering and Remote Sensing. Intended audiences are the conservation community, technical specialists, and resource managers. LITERATURE CITED: Evans, R. A. 1995. Hemlock ravines at Delaware Water Gap National Recreation Area: A highly valued, distinctive, and threatened ecosystems. Proceedings of the Delaware Water Gap 30th anniversary Symposium, November 1995. Pocono Environmental Education Center, Bushkill, Pennsylvania. HWA Working Group. 1995. Hemlock Woolly Adelgid Strategic Plan. 12 p. Lapin, B. 1995. The impact of hemlock woolly adelgid on resources in the lower Connecticut River Valley. Northeastern Center for Forest Health Research, Hamden CT. 45 p. McClure, M.S. 1990. Role of wind, birds and mammals in the dispersal of hemlock woolly adelgid, Adelges tsugae, Annad (Homoptera: Adelgidae). Environmental Entomology 19: 36-43. APPENDIX A STUDY PLAN AMENDMENT LEETOWN SCIENCE CENTER AQUATIC ECOLOGY LABORATORY TITLE: Modeling Stand Vulnerability of Eastern Hemlock to Hemlock Woolly Adelgid BACKGROUND AND JUSTIFICATION: This study will investigate site, landscape, and regional factors that may influence hemlock woolly adelgid (HWA) induced mortality in eastern hemlock or HWA dispersal. Spatially- explicit computer models will be constructed to examine these multi-scale influences in an attempt to determine the mechanisms leading to hemlock infestation, defoliation, and decline. Landscape-level data analyzed in a geographic information system (GIS) will provide the means to determine the range of environmental conditions that may govern the distribution and susceptability of hemlock forests. Digital landscape data (e.g. maps) have been used to characterize the manner in which environmental factors such as slope, topographic position, soil type, and moisture regimes combine to form distinct eco-physiographic environments on the landscape (Bailey, et al. 1993, Band 1989, Coughlan and Running 1989). Digital representations of terrain (eg. digital elevation models) are useful for quantifying landscape characteristics such as topographic position, soil moisture, solar radiance, and a host of other environmental parameters (Jenson 1991, Martz and Garbrecht 1993, Skidmore 1990, Carter 1988, Jenson and Dominique 1988). Specific morphological features such as ridges, gullies, coves, and saddles can be determined from elevation models to assess habitat conditions over large areas (Skidmore 1990). Vegetation occurrence can be assessed using remote sensing techniques (Campbell 1987, Quattrocchi and Pelletier 1991) and modeled by combining factors such as slope position, solar radiance, soil moisture, and microclimate (Band et al. 1993, Twery, et. al. 1991, Davis and Goetz 1990). Distance modeling and spatial overlay and regression techniques can be used to evaluate the relationships among environmental variables in a GIS (Burrough 1986, Tomlin 1990). Elevation and other topographic factors are major influences in the distribution of biotic communities in mountainous environments. On a local scale, higher elevations generally correspond to cooler ambient temperatures (e.g. 4 C cooler mean annual temperature for every 1000 meter rise in elevation in GSMNP). Over a larger area, temperature is influenced by latitude and longitude along with elevation. Hopkins (1938) described this relationship as it relates to phenology. Hopkins (1938) stated that 122 meter increase in elevation, 1 degree north in latitude, and 5 degrees west in longitude are all equivalent to a 4 day delay in phenology. An extension of this concept of climatically similar areas can be made to include topographic factors and used to assess "ecological equivalency". Ecologically equivalent areas are areas of similar growing conditions that can be mapped for predicting phenology and species community distributions. This may be a better option than using elevation alone over a broad area, when describing biological distributions. Klopfer and McCombs (1997) mapped ecological equivalency as relative phenology across Virginia to analyze forest community distributions. McCombs (1997) also used this climatic relationship to analyze the distribution of Northern dusky flying squirrels (Glaucomys sabrinus). The combination of relative phenology (based on elevation, latitude, and longitude) with other factors such as slope, aspect, solar radiation, land form (e.g. convexity), precipitation, geology, soil, and slope position could result in a powerful map of ecological equivalency for microclimate (scale would depend on resolution of base data), although exact calibration techniques need to be researched. Using such a map, the distribution of eastern hemlock could be assessed in relation to ecologically equivalent growing environments across its range in the Eastern U.S. Hemlock stands defoliated by hemlock woolly adelgid could be mapped and compared to hemlock stands not impacted by the hemlock woolly adelgid in ecologically equivalent areas. This could reveal patterns of HWA infestation related to landscape condition, and help to determine the scale at which landscape influences are most important (ie. microsite, landscape, region). If such a relationship can be identified, areas that are ecologically equivalent to current zones of infestation could be delineated as potential sites of future defoliation. This would be an important tool in the monitoring and management of HWA dispersal. OBJECTIVES: The objectives of this research are to:
HYPOTHESIS TO BE TESTED: The null hypothesis of this research is that the level of hemlock stand decline from hemlock wooly adelgid is not related to site, landscape, or regional level factors of climate, stand condition and configuration, or terrain influences. Rejection of the null hypothesis, if supported by data analysis, would allow spatially explicit models of hemlock decline to be built that may assist in monotoring and management of HWA dispersal and damage. PROCEDURES: Initial investigations will focus on SHEN due to the differential pattern of hemlock infestation and heavy mortality observed in the park over the past 5-7 years. The Aquatic Ecology Laboratory (USGS-BRD) has already constructed topographic characterization models of SHEN and DEWA as part of a research program aimed at assessing potential biodiversity losses due to hemlock wooly adelgid (Young et al., 1998). Several terrain variables (elevation, slope, aspect, terrain shape, solar radiance) were shown to influence hemlock distribution in relation to non-hemlock forests. This model will be enhanced using new data and extended to incorporate additional environmental variables and potential vectors for adelgid dispersal (e.g., prevailing winds, road density, bird usage of hemlock stands, deer browsing, etc.). The resulting terrain-based model will be used to stratify hemlock stands into ecological equivalency classes for sampling and comparison and to assess stand vulnerability due to site-, landscape-, and regional-level factors (figure 1). To more accurately assess the location and severity of hemlock decline, we will re-map hemlock stands in SHEN using multi-year satellite imagery. Landsat Thematic Mapper (TM) images will be purchased for 3-year intervals starting in 1986 and ending at the most currently available image (e.g., 1984, '88., '92, '94, '97). Due to cloud cover and other confounding factors, the dates may be a year before or after the target date. Two full Landsat scenes (path16, row 33 and path 16, row 34) are required to cover the SHEN study area. The broad extents of the scenes will allow examination of hemlock stands outside of the SHEN boundary, including significant portions of Jefferson/George Washington National Forest. The first step will be to identify hemlock stands on the earliest images (i.e., 1986). Ground data from intact hemlock stands, knowledge of local experts, information from SNP's Long Term Ecological Monitoring plots, and other existing maps and datasets will be used as training and accuracy assessment data. Once an accurate map of hemlock stands is created for 1984, these same areas can be re-assessed with imagery from successive years. Previous studies (Royle and Lathrop 1997) have found that a TM4/TM3 band ratio is effective in evaluating eastern hemlock forest health. This "vegetation index" will be used to analyze the condition of eastern hemlock trees in relation to known infestation and defoliation by HWA. Maps of damage will be created for each year along with an overall map of defoliation rate. The most current map will be assessed using recent field data. The intermediate maps will be assessed for accuracy from color infrared aerial photography when possible. Techniques to better map current hemlock damage, including the use of high-resolution (eg. 1-5 meter) imagery, will be explored. Site level factors affecting eastern hemlock tree growth and survival will be collected from transect surveys of hemlock stands stratified into ecological equivalence classes using a landscape factor-based, randomized block sampling design. A sub-sample of hemlock stands will be selected within each sampling block and assessed for site factors. Specifically, measures of soil moisture, temperature, and salinity will be collected using a digital soil probe. Soil pH will be measured in the field using a soil test kit. Measures of solar insolation, air temperature, and canopy closure will be collected using digital light and temperature meters and a densitometer. Transects will be oriented perpendicular to stream channels to assess gradients in slope, soil moisture, and light availability. Regional climatic patterns will be mapped using data from the National Climatic Data Center, and National Park Service weather stations and will include information on prevailing wind speed and direction, precipitation, air temperature, and air quality. Other regional patterns of topography, soils, geology, and vegetation will be derived from existing databases. Statistical modeling will be conducted using a combination of logistic regression and other multi-variate techniques to assess the relationship between hemlock decline and biotic and abiotic environmental factors at multiple scales. Site, landscape, and regional factors will be analyzed for associations with both HWA dispersal mechanisms and hemlock stand condition indicators (figure 1). MANOVA tests will be used to test differences between site-level environmental factors among stands of similar ecological equivalency. We will also explore the use of Bayseian statistical methods to construct spatially explicit probability models of stand vulnerability based on landscape condition factors. DATA: Data for this project consist of existing GIS coverages derived from US Geological Survey Digital Elevation Models, US Geological Survey Digital Line Graphs, and Park Service vegetation data layers at map scales of 1:24,000. Maps of hemlock condition and decline will be derived from Landsat Thematic Mapper satellite imagery (map scale approx. 1:100,000), aerial photography (map scale approx.1:20,000-1:40,000), and NPS databases. Satellite imagery for 1994 was provided through a cooperative agreement with the Virginia GAP Analysis project. Satellite imagery for other years will be purchased through the USGS Earth Resources Observation System archives. Additional GIS data layers will be incorporated as needed from existing USGS, NPS, NOAA, EPA, and other government data sources. All data will be housed in the Aquatic Ecology Laboraory's Geographic Information Systemfacility. At the end of the study, all data used in the project will be made available to the public with accompanying metadata documentation, except where licenses restrict distribution (i.e. satellite imagery). SCHEDULE: FY98: Coordinate with RDL and SAFL and NPS on components of study design and field sampling protocols, acquire imagery, begin image mapping of hemlock defoliation, research model parameters. FY99: Finish draft of image maps, field surveys to check mapping accuracy, collect site specific environmental data, begin model construction. FY00: Data analysis and modeling FY01: Publication FY02: Coordination/integration of model results with RDL, SAFL, NPS DURATION OF STUDY: Begin: June 1998 End: June 2002 SAFETY: Foul weather gear appropriate for the season and temperature in temperate woodlands will be required in the field. First aid kits will be maintained in field vehicles. ANIMAL WELFARE: Not Applicable COOPERATORS/PARTNERS: Tom Blount, Shenandoah National Park, Luray, VA James Akerson, Shenandoah National Park, Luray, VA Joe Clark, Southern Appalachian Field Laboratory, Knoxville, TN Frank van Manen, Southern Appalachian Field Laboratory, Knoxville, TN Bob Ross, Research and Development Laboratory, Wellsboro, PA KEY STAFF: John Young, Fish and Wildlife Biologist (GIS Specialist), Aquatic Ecology Laboratory, Leetown, WV Craig Snyder, Ecologist, Aquatic Ecology Laboratory, Leetown, WV Dave Morton, Biologist (Remote Sensing), Aquatic Ecology Laboratory, Leetown, WV EXPECTED PRODUCTS: Computer model(s) linking landscape attributes to hemlock wooly adelgid dispersal, hemlock health, defoliation risk, and biotic impacts. Paper(s) published in peer-reviewed journals such as Conservation Biology, Ecosystems, Landscape Ecology, Ecological Applications, Remote Sensing of the Environment, or Photogrammetric Engineering and Remote Sensing. Intended audiences are the conservation community, technical specialists, and resource managers. LITERATURE CITED: Bailey, R.G., M.E. Jensen, D. Cleland, P.S. Bourgeron. 1993. Design and use of ecological mapping units. In: M.E. Jensen and P.S. Bourgeron, eds. Eastside Forest Ecosystem Health Assessment, Volume II: Ecosytem Management: Principles and Applications. Portland, Oregon: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. Band, L.E. Automating topographic and ecounit extraction from mountainous forested watersheds. 1989. AI Applications. Vol. 3, No.4. pp. 1-11. Band, L.E., P. Patterson, R. Nemani, and S.W. Running. 1993. Forest ecosystem processes at the watershed scale: incorporating hillslope hydrology. Agricultural and Forest Meteorology. Vol. 63. pp. 93-126. Burrough, P.A. 1986. Principles of Geographic Information Systems for Land Resources Assessment: Monographs on Soil and Resources Survey No. 12. Oxford University Press. Oxford, England.194pp. Carter, J.R. 1988. Digital representations of topographic surfaces. Photogrammetric Engineering and Remote Sensing. Vol. 54, No. 11. pp. 1577-1580. Campbell, J.B. 1987. Introduction to Remote Sensing. Guilford Press. New York. 551 pp. Coughlan, J.C. and S.W. Running. 1989. An expert system to aggregate biophysical attributes of a forested landscape within a geographic information system. AI Applications. Vol. 3, No. 4. pp. 35-43. Davis, F.W. and S. Goetz. 1990. Modeling vegetation pattern using digital terrain data. Landscape Ecology. Vol.4, No. 1. pp 69-80. Hopkins, A.D. 1938. Bioclimatics, a science of life and climate relations. U.S. Dept. Agr., Misc. Publ. 280. Washington, D.C. 188pp. HWA Working Group. 1995. Hemlock Woolly Adelgid Strategic Plan. 12 p. Jenson, S. K. 1991. Applications of hydrologic information automatically extracted from digital elevation models. Hydrological Processes. Vol 5. pp. 31-44. Jenson, S.K. and J.O. Dominique. 1988. Extracting topographic structure from digital elevation data for geographic information system analysis. Photogrammetric Engineering and Remote Sensing. Vol. 54. No. 11. pp. 1593-1600. Klopfer, S.D., and J.W. McCombs II. 1997. Using Hopkins Bioclimatic Law to classify ecologically similar areas. GIS and Remote Sensing Poster Session. Fourth Annual TWS Conference, Snowmass, CO. September 1997. Lapin, B. 1995. The impact of hemlock woolly adelgid on resources in the lower Connecticut River Valley. Northeastern Center for Forest Health Research, Hamden CT. 45 p. Martz, L.W. and J. Garbrecht. 1993. Automated extraction of drainage network and watershed data from digital elevation models. Water Resources Bulletin. Vol. 29, No. 6. pp. 901-908. McClure, M.S. 1990. Role of wind, birds and mammals in the dispersal of hemlock woolly adelgid, Adelges tsugae, Annad (Homoptera: Adelgidae). Environmental Entomology 19: 36-43. McCombs, J.W., II. 1997. Geographic information system topographic factor maps for wildlife management. M.S. Thesis. VPI and SU, Blacksburg, VA. 141pp. Quattrochi, D.A. and R. E. Pelletier. 1991. Remote Sensing for analysis of landscapes: An Introduction. In Quantitative Methods in Landscape Ecology, M.G. Turner and R.H. Gardner, eds. Springer-Verlag. New York. pp. 51-76. Royle, D.D. and R.G. Lathrop. 1997. Monitoring hemlock forest health in New Jersey using Landsat TM data and change detection techniques. Forest Science, Vol. 43, No. 3. Pp. 327-335. Skidmore, A.K. 1990. Terrain position as mapped from a gridded digital elevation model. Int. J. Geographical Information Systems. Vol. 4, No. 1. pp. 33-49. Tomlin, C.D. 1990. Geographic Information Systems and Cartographic Modeling. Prentice Hall, Englewood Cliffs, NJ. 249 pp. Twery, M.J., G.A. Elmes, and C.B. Yuill. 1991. Scientific exploration with an intelligent GIS: predicting species composition from topography. AI Applications. Vol. 5, No. 2. pp. 45-53. Young, J.A., C.D. Snyder, D.R. Smith, and D.P. Lemarie. 1998 (In prep). A landscape-based sampling design to assess biodiversity losses from eastern hemlock decline. APPENDIX B. STUDY PLAN AMENDMENT TITLE: Avian biodiversity and obligate species associations in eastern hemlock forest habitats BACKGROUND AND JUSTIFICATION: The eastern hemlock (Tsuga canadensis) thrives in cool, moist, hillside and ravine environments throughout the eastern United States and Canada. Hemlock stands are valued for their riparian forest habitat, their commercial use (timber and horticulture), and as desirable recreational resources on public lands. Viewed as dominant overstory species in the late-successional state of many forests, hemlock stands may have co-evolved with other flora and fauna in stable, co-dependent associations. Preliminary data, for example, show exclusive use of hemlock bench and ravine habitats (as opposed to equivalent hardwood sites) by three species of wood warbler and vireo (solitary vireo, black-throated green warbler, and blackburnian warbler), mixed use by two species (red-eyed vireo and ovenbird), and exclusive use of hardwood benches and ravines by one species (American redstart) of those surveyed (Table 1). The extent to which some species depend on hemlock is unknown. Within the past two decades, substantial declines in eastern hemlock have been observed throughout the eastern United States, resulting in considerable federal and state concern. Widespread hemlock defoliation has largely been attributed to the hemlock woolly adelgid (Adelges tsugae). In a recent biocontrol workshop, 32 federal, 4 state, 6 private, and 11 university experts ranked the hemlock woolly adelgid as the most serious exotic pest in the east but also as the one with the most potential for biocontrol. Loss of this important forest species has unknown consequences for the ecology of Appalachian forests but may involve changes in energy inputs, microclimate, and physical habitat structure for several biotic communities. Hence there is an urgent need to characterize the contribution of hemlock forests to avian biodiversity in large forested landscapes and to determine how avian communities associated with hemlock stands are most likely to be impacted. Hemlock-specific biodiversity and unique species associations in hemlock ecosystems are not well known. In many eastern-slope forests hemlock stands are fully invaded by the HWA and trees are in various stages of decline. Should restoration of these ecosystems be needed in future years, knowledge of former biodiversity is essential. In addition to species lists and richness comparisons, this work will likely generate significant differences in species density estimates, unique community associations, and macrohabitat requirements. Classification of species present by trophic level and feeding guilds will also enable us to determine how avian communities are structured in hemlock versus hardwood forests. This research initiative relates directly to the responsibility of the Biological Resources Division (BRD) of the USGS to "assist resource and land managers, particularly in the Department of Interior (DOI), by providing them with sound biological information and with assistance in applying the information to their needs." Specific goals of BRD addressed by this work are (1) characterize natural processes and identify factors that influence the quality or quantity of the Nation's biological resources at all levels of biological organization and (2) facilitate sound management of the Nation's biological resources by collaborating with partners in all phases of our work. National Park Service (NPS) land managers at the Delaware Water Gap National Recreation Area (DEWA), Great Smoky Mountains National Park, and Shenandoah National Park (SHEN) have requested this work as a priority need under "Population trends and habitats of neotropical migrant birds in national parks (Bureau Information Needs, BIN, number 5). The primary program element is exotic species with secondary program elements of ecosystems and status and trends. In addition, one of the stated goals of the Northeast Regional Work Group (NEWG) of Partners in Flight (PIF), dating to March 1994, is to "research and monitor effects of large-scale declines in forest health due to agents such as woolly adelgid, gypsy moth, and acid rain." Representatives from USEPA, USFS, USFWS, and several state fish and wildlife agencies are participating in and coordinating these efforts. The Pennsylvania PIF's priority research goal is to "identify, prioritize, and conduct research on population and community dynamics, habitat requirements, and songbird habitat assemblages." Present knowledge of avian community relationships with hemlock forests includes studies from New Jersey and Wisconsin/Michigan. Benzinger (1994a, b) reported three species (northern goshawk, blue-headed vireo, and black-throated green warbler) as found almost exclusively, five (red-shouldered hawk, barred owl, Acadian flycatcher, winter wren, and hermit thrush) usually, and four (Cooper's hawk, red-breasted nuthatch, magnolia warbler, and Blackburnian warbler) often in eastern hemlock of New Jersey. Howe and Mossman (1995) similarly showed five species (red-breasted nuthatch, winter wren, blue-headed vireo, black-throated green warbler, and Blackburnian warbler) to be significantly associated with hemlock forests in Wisconsin and upper Michigan. Four additional species (brown creeper, hermit thrush, northern parula, and yellow-rumped warbler) were considered to be "characteristic hemlock birds". Assemblages changed or densities of these birds decreased where forests were managed. Preliminary observations within DEWA showed three of the above species (blue-headed vireo, black-throated green warbler, and Blackburnian warbler) to be relatively abundant in hemlock ravines (Evans et al. 1996). Further preliminary observations showed a dozen territorial species in hemlock stands in early July, three of which (blue-headed vireo, black-throated green warbler, and Blackburnian warbler) were discussed above and not found in matched hardwood stands (Table 1). Relative abundances and densities during the breeding season are still not known for the forest habitats within the park. OBJECTIVES: The principal objective of this research is to characterize avian community composition in hemlock forests of eastern parks and to compare the structure and function of these communities in hemlock habitats (bench, slope, and ravine) to matched hardwood habitats. The ultimate objective is to couple empirical models of stand vulnerability (due to infestation by the hemlock woolly adelgid) with known structural and functional attributes of hemlock avian communities to predict how and where avian communities are most likely to change. HYPOTHESIS TO BE TESTED: Hemlock avian communities are characterized by a unique species composition (compared to matched hardwood) with structural and functional guilds that vary among habitat types: bench, slope, and ravine. Table 1. Comparison of the Number of Study Sites where Potentially Hemlock-Associated Bird Species were Heard or Sang During Aquatic Biodiversity Sampling, 7-11 July 1997. In Parenthesis are the Number of Sites in each Catagory Sampled.
PROCEDURES: Point-count techniques (Ralph et al. 1993, 1995) will be used to quantify breeding bird populations in matched hardwood and hemlock stands of the Delaware Water Gap National Recreational Area. Use of an existing study design (Ross et al. 1996, Smith et al. 1996) with established sites will allow us to stratify and compare specific habitat types (bench, ravine, and mid-slope) within hardwood and hemlock stands for the first time and to maximize the usefulness of data through relation to existing aquatic and terrestrial databases. Breeding bird field work will be initiated in June 1998 with pilot studies to determine variability estimates and detection magnitude for each species present. These data should allow us to determine the number of count points needed for each category of forest and habitat type. Ralph et al. (1995) recommend a minimum of 30 points, 250 m apart, to account for variability in the absence of pilot studies. Our established sampling design of 28 sites and 6 forest habitat types would require 6 points per site in such a case. Count point locations will then be determined ( 250 m apart) and maps of count circles developed in the field from May to July 1999. Winter bird surveys will be conducted from December 1999 to February 2000. Owls will also be surveyed following methods described by Mahan et al. (1997). Taped vocalizations of five species (northern saw-whet, eastern screech, barred, long-eared, and great-horned) will be played back at the same point locations in late winter-early spring 2000 (March to April). Following a 5-min equilibration period after reaching the site, the calls of each species in turn will be played back in four bouts of 15 sec followed by a 45 sec period of silence. Vocal or visual responses will be recorded. Breeding bird data acquisition will begin in late May 2000 and extend to early July if necessary, with 5-minute point counts between sunrise and 10 a.m. daily. Individuals within 50 m of the count center will be distinguished from all others detected, for abundance estimates (to assure uniform detection rates among species; Ralph et al. 1995). Modelling work will be done by ecologists at AEL and SAFL and comparisons made to southern Appalachian hemlock stands, which may exhibit different species associations. DATA: Critical Data Primary data are the replicated counts for each bird species detected in each forest and habitat type. These data will comprise an index of abundance for each species comparable across treatment categories. Analysis Statistical comparisons will be made using general linear modelling (SAS 1987) and multivariate techniques to evaluate the extent to which hemlock-dominated forests and forest habitats support distinct avian communities. Acceptance/Rejection Criteria The level for rejection of null hypotheses will be p 0.05. Location of Data Data, records, correspondence, and related study documents will be stored in the office of the Ecologist or General Biologist and at station research files of RDL-Wellsboro. The Ecologist and Laboratory Director are responsible for stored data, which will consist of both hard copies and electronic disk format. At the conclusion of the study, all data and documents will be stored in the station archives for a period of 5 years. SCHEDULE: FY98: Coordinate with AEL and SAFL on major components of study; pilot studies and subsequent power analyses FY99: Point location determinations; mapping FY00: Winterbird surveys; owl surveys; breeding bird surveys FY01: Data analysis; modelling FY02: Publication DURATION OF STUDY: Begin: June 1998 End: June 2002 SAFETY: Fowl weather gear appropriate for the season in temperate woodlands will be required in the field. First aid kits will be maintained in field vehicles. ANIMAL WELFARE: Not applicable. COOPERATORS/PARTNERS: Elizabeth A. Johnson, Delaware Water Gap National Recreation Area (NPS) Richard A. Evans, Delaware Water Gap National Recreation Area (NPS) Roles: Provide point count data for power analyses, housing, volunteer/intern field assistance, staff assistance Carol Schell, Great Smoky Mountains National Park Role: Provide data (bird community) KEY STAFF: Robert M. Ross, Ecologist, Research and Development Laboratory - Wellsboro Randy M. Bennett, General Biologist, Research and Development Laboratory - Wellsboro EXPECTED PRODUCTS: A publishable paper in peer-reviewed journals such as Conservation Biology, Ecological Applications, or Auk is expected. Intended audience is the conservation community and land or natural resource managers, especially NPS. LITERATURE CITED: Benzinger, J. 1994a. Hemlock decline and breeding birds - I: hemlock ecology. Rec. N. J. Birds 20(1):2-12. Benzinger, J. 1994b. Hemlock decline and breeding birds - II: effects of habitat change. Rec. N. J. Birds 20(2):34-51. Evans, R. A., E. Johnson, J. Schreiner, A. Ambler, J. Battles, N. Cleavitt, T. Fahey, J. Sciascia, and E. Pehek. 1996. Potential impacts of hemlock woolly adelgid (Adelges tsugae) on eastern hemlock (Tsuga canadensis) ecosystems. Pages 42-57 in S. M. Salom, T. C. Tigner, and R. C. Reardon (eds.), Proc. 1st Hemlock woolly Adelgid Review, Charlottesville, Virginia, October 12, 1995. U.S.D.A. Forest Service, Morgantown, WV. Howe, R. W., and M. Mossman. 1995. The significance of hemlock for breeding birds in the western Great Lakes region. Pages 125-139 in G. Mroz and J. Martin (eds.), Hemlock Ecology and Management. Proc. Regional Conf. on Ecol. and Mgmt. of E. Hemlock, September 27-28, 1995, Iron Mountain, Michigan. Dept. of Forestry, Univ. of Wisconsin, Madison, WI. Mahan, C., K. Sullivan, K. C. Kim, R. Yahner, and M. Abrams. 1997. Sampling protocols and procedures: biodiversity profile assessment of eastern hemlock forests. Center for Biodiversity Research, The Pennsylvania State University, University Park. Ralph, C. J., G. R. Geupel, P. Pyle, T. E. Martin, and D. F. DeSante. 1993. Handbook of field methods for monitoring landbirds. Gen. Tech. Rep. PSW-GTR-144. Albany, CA: USDA, Forest Service, Pacific Southwest Research Station. 41p. Ralph, C. J., J. R. Sauer, and S. Droege. 1995. Monitoring bird populations by point counts. Gen. Tech. Rep. PSW-GTR-149, USDA, Forest Service, Pacific Southwest Research Station, Albany, CA. 18p. Ross, R. M., C. Snyder, D. Smith, J. Young. 1996. Aquatic Biodiversity in Eastern Hemlock Forests. NBS Study Plan, Leetown Science Center, Kearneysville, WV. 19p. SAS (Statistical Analysis System). 1987. SAS/STAT guide for personal computers. Version 6. SAS Institute, Cary, NC. Smith, D. A., C. Snyder, and J. A. Young. 1996. Sampling design methodology - hemlock biodiversity research program. Progress Report, Aquatic Ecology Laboratory, USGS, Biological Resources Division, Leetown, WV. APPENDIX C STUDY PLAN AMENDMENT LEETOWN SCIENCE CENTER SOUTHERN APPALACHIAN FIELD LABORATORY TITLE: Potential Avian Community Impacts of the Hemlock Woolly Adelgid in Great Smokey Mountains National Park BACKGROUND AND JUSTIFICATION: The potential decline of hemlock in the southern Appalachians could lead to changes in associated biotic communities due to changes in stand composition and structure, microclimate, and energy cycles. The association between hemlock forests and avian communities is largely undocumented. Therefore, there is a need to determine the association between avian species and hemlock forests and to develop tools to predict how species compositions may change. Because the hemlock woolly adelgid has not yet reached the southern Appalachian region, such predictive tools could be timely for proactive management. Focus on the impact of hemlock decline on breeding birds in hemlock stands is warranted for a number of reasons. Some of the species that typically occur in hemlock stands have experienced population declines over the past 30 years, based on Breeding Bird Survey data (Sauer et al. 1997). In addition, hemlock communities may support important source populations for sustaining regional avian communities and represent critical sites of high avian diversity. Because birds are sensitive to environmental perturbations at a number of levels, birds also may be good indicators of broad-based ecological changes linked to hemlock decline. Despite the considerable literature base available on birds, relatively little research has been done to characterize avian communities in hemlock stands. We are aware of no existing avian-habitat models that would allow for prediction of how hemlock decline will affect birds. The research proposed here would fill this critical need. Following the initiation of the stand vulnerability modeling by AEL, we will characterize breeding bird populations in matched hardwood and hemlock stands based on data from the Cherokee National Forest in Tennessee. Much of these data already exist and sampling schemes have already been developed. From these and additional field data, we will develop habitat models to predict the composition of bird communities and, coupled with stand vulnerability models, to predict changes in avian communities for various projected scenarios of hemlock decline. We will develop separate models for the southern Appalachian and mid-Atlantic regions and validate the models with independent data. OBJECTIVES: The objectives of this research are to: (1) develop GIS-based habitat models of bird species in the southern Appalachian and mid-Atlantic regions; and (2) apply and test these models to their respective regions to predict potential changes in bird communities in response to projected scenarios of hemlock decline. HYPOTHESIS TO BE TESTED: The composition of avian communities in hemlock forests are different from other forest types and can be predicted by use of landscape-scale GIS variables. PROCEDURES: Models of bird-habitat relationships will be developed and tested for all bird species for which sufficient data exist for hemlock and hardwood stands. We will use data from the 200 point-count surveys conducted on the Tellico District of the Cherokee National Forest during a seven-year period (1992-98). Using the SAA databases and other available databases, models for all species with sufficient data will be developed based on two modeling approaches: logistic regression and Mahalanobis distance functions. We will use a jackknife procedure to test model accuracy. The results of these tests will be compared between modeling approaches to determine which modeling approach is the most accurate and provides the best fit across all of the individual species models. Models developed from the Delaware Water Gap will be tested with bird data from the Cherokee National Forest to test the ability to apply models region-wide. Similarly, models developed from the Cherokee National Forest will be tested with bird data from the Delaware Water Gap. Bird data from both regions may ultimately be pooled to develop regional models that optimize predictive accuracy. DATA: Avian data will be collected based on a series of point counts placed in hemlock and hardwood stands, with one point placed in the interior of each stand surveyed. Each point will be censused for a 10-minute period, with all point counts occurring between sunrise and four hours after sunrise. Data will be divided based on the first three minutes of each point count, the next two minutes and the entire 10-minute period to allow for comparison of the results with breeding bird survey data and other national monitoring initiatives. Bird locations will be assigned to distance classes from the point, including 0-50 m and >50 m. Each point will be visited once during the breeding season (May 15- July 1). Geospatial data for this project consist of existing GIS coverages derived from the Southern Appalachian Assessment (SAA) database (Hermann 1996) and the Mid-Atlantic atlas (Jones et al. 1998). Additional GIS data layers will be incorporated as needed from existing USGS, NPS, NOAA, EPA, and other government data sources. All data will be stored on SAFL computers. At the end of the study, all data used in the project will be made available to the public with accompanying metadata documentation, except where licenses restrict distribution (i.e., satellite imagery). SCHEDULE: FY98: Coordinate with AEL, RDL, and NPS on components of study, design field sampling protocols, research model parameters, establish field sampling protocols. FY99: Coordinate with AEL, RDL, and NPS on preliminary results of hemlock vulnerability modeling. FY00: Field data collection. FY01: Data analysis and modeling. FY02: Coordination/integration of model results with AEL, RDL, NPS. DURATION OF STUDY: Begin: June 1998 End: June 2003 SAFETY: Foul weather gear appropriate for the season and temperature in temperate woodlands will be required in the field. First aid kits will be maintained in field vehicles. ANIMAL WELFARE: Not Applicable COOPERATORS/PARTNERS: John Young, Aquatic Ecology Laboratory, Leetown, WV Bob Ross, Research and Development Laboratory, Wellsboro, PA KEY STAFF: Frank van Manen, Senior Computer Systems Specialist, University of Tennessee/Southern Appalachian Field Laboratory, Knoxville, TN David A. Buehler, Associate Professor, Department of Forestry, Wildlife and Fisheries, University of Tennessee, Knoxville, TN Joseph D. Clark, Laboratory Director, Southern Appalachian Field Laboratory, Knoxville, TN EXPECTED PRODUCTS: GIS-based models linking model predictions of changes in hemlock communities due to hemlock woolly adelgid infestation to changes in avian communities. Papers published in peer-reviewed journals such as Journal of Wildlife Management, Wildlife Society Bulletin, Wilson Bulletin, and Conservation Biology. Intended audiences are the conservation community, technical specialists, and resource managers. LITERATURE CITED: Hermann, K.A., ed. 1996. The southern Appalachian Assessment GIS data base CD ROM set. Southern Appalachian Man and Biosphere Program, Norris, Tenn. Jones, K.Bruce, Riitters, Kurt H., Wickham, James D., Tankersley Jr., Roger D., O'Neill, Robert V., Chaloud, Deborah J., Smith, Elizabeth R., and Anne C. Neale.1997. An Ecological Assessment of the United States Mid-Atlantic Region: A Landscape Atlas. EPA/600/R-97/130, U.S. Environmental Protection Agency, Office of Research and Development, Washington DC. McClure, M.S. 1990. Role of wind, birds and mammals in the dispersal of hemlock woolly adelgid, Adelges tsugae, Annad (Homoptera: Adelgidae). Environmental Entomology 19: 36-43. Sauer, J. R., J. E. Hines, G. Gough, I. Thomas, and B. G. Peterjohn. 1997. The North American Breeding Bird Survey Results and Analysis. Version 96.4. Patuxent Wildlife Research Center, Laurel, MD. U.S. Department of the Interior || U.S. Geological Survey 11700 Leetown Road, Kearneysville, WV 25430, USA URL: http://www.lsc.usgs.gov Maintainer: lsc_webmaster@usgs.gov Last Modified: September 11, 2002 dwn Privacy Policy and Disclaimers || FOIA || Accessibility | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||