ABSTRACTS

Ruffed Grouse Ecology and Conservation in Changing Appalachian Forests

Darroch M. Whitaker, Dean F. Stauffer, Patrick K. Devers, Department of Fisheries & Wildlife Sciences, Virginia Tech; Gary W. Norman, Virginia Department Game & Inland Fisheries

Abstract: Hunters and wildlife managers have been concerned about declining ruffed grouse populations throughout the Appalachians. Suspected causes included chestnut decline, late season hunting, and reductions in the extent of early successional forest habitat. Due to the scarcity of information on grouse population regulation in the region, the Appalachian Cooperative Grouse Research Project (ACGRP) was initiated in 1996. Since that time ACGRP cooperators at 12 study sites in 8 states have collected data on survival, reproduction, nutrition, movements and habitat use from ~ 3,000 radio-collared grouse. Preliminary findings suggest that, although hunter-kills may be partially additive to natural mortality, harvests were relatively small and poor recruitment was a far more important factor limiting grouse populations. Hen reproductive success was highly variable between sites and years, a phenomenon resulting largely from a their dependence on unpredictable fall acorn crops. Acorns provide high quality food during winter and, when available, allow hens to enter breeding season in good condition. Improved hen condition results in higher recruitment via increased nesting and renesting rates and higher chick survival. Also, habitat studies suggested that early-successional forest stands on mesic soils, which are typically found in hollows and riparian zones, afford high quality brood-rearing habitat. Given these findings we suggest that grouse management in the region should focus on silvicutural techniques that promote and stabilize hard mast crops and create early-successional habitats on mesic soils. Where grouse populations are declining restriction of hunting may also be warranted. Finally, future prospects for Appalachian grouse populations are discussed. Decreased timber harvests, continued fire suppression, ongoing oak decline, and potential changes in patterns of oak mast production resulting from climate change may all have adverse long-term consequences for Appalachian grouse populations.

Oak Regeneration After Five Silvicultural Treatments in the Southern Appalachians

Jean H. Lorber and Thomas R. Fox, Department of Forestry, Virginia Tech

Abstract: In the eastern United States, oak (Quercus) regeneration is often sparse or poorly competitive after timber harvests, leading to decreased oak stocking in future stands. This trend is more apparent on higher-quality sites where oaks seedlings compete with several faster growing species including red maple (Acer rubrum), sweet birch (Betula lenta) and yellow poplar (Liriodendron tulipifera). It is believed that manipulating stand density using different silvicultural harvests can promote oak regeneration.

This project was a co-operative effort among the U.S. Forest Service, Virginia Tech and MeadWestvaco. There were four study sites altogether; three were located in the Washington/Jefferson National Forest in southwest Virginia and one site was located in the MeadWestvaco Experimental Research Forest in Randolph county, West Virginia. At all four locations, this study looked at effects of 5 types of silvicultural treatments (clearcut, commercial clearcut, group selection, shelterwood, and leave-tree) on regeneration composition and dominance. Additionally, water, light and soil nutrient data were collected from each treatment in order explain the processes by which certain species became dominant. The data was collected four years after harvest.

Our results showed that the dominance of oak regeneration four years after treatment was dependent on site quality but not dependent on silvicultural treatment. The MeadWestvaco site in WVA was the highest quality site (SIoak=80ft.) and had almost no oak regeneration in every treatment. The Newcastle site was the lowest quality site (SIoak=60ft.) and had a sufficient oak presence in every treatment. Even at the MeadWestvaco site, establishment of oak seedlings was not a limiting factor. It was competition from stems of other tree species that was the main deterrent to oak dominance.

This trend of decreasing oak dominance on higher quality sites is consistent with the literature. However, the fact that silvicultural treatments could not help establish oak on the high-quality site is surprising. Even in a shelterwood system, which theoretically favors the intermediately shade-tolerant oak, other tree species are already out competing and killing oak stems. These results reinforce the belief that a one-time overstory treatment is not enough to ensure continued oak dominance in a regenerating stand. Rather, the canopy and understory of a stand should be treated 1-2 decades before a final harvest. This longer-term process will allow the development of large oak regeneration, which will be the source of dominant oak stems once the overstory is finally removed.

Soil nutrient abundance, particularly available nitrogen, is an important indicator for site quality. These 4 sites showed a wide but predictable variability in soil nitrogen levels. The MeadWestvaco site had the highest site index and soil nitrogen levels; concentrations were 4-5 times higher than any other site. Consequently, this site supported the community of tree species, such as black cherry and tulip poplar, that has demanding soil requirements. Soil moisture levels were relatively uniform, with only the lowest quality site (Newcastle) being significantly different from the rest. The three sites that were similar in soil moisture were very different in species composition, indicating that soil moisture levels were only a secondary factor in determining overall site quality.

Efficacy of Localized Management for Reducing Deer Herbivory of Forest Regeneration Areas

B.F. Miller, T.A. Campbell, B.R. Laseter, D.A. Osborn, K.V. Miller, Warnell School of Forest Resources, University of Georgia; and W.M. Ford, USDA Forest Service, Northeastern Research Station

Abstract: White-tailed deer are keystone herbivores that can dramatically affect forest structure and composition. Because traditional management strategies (i.e. sport hunting) are not solving the damage problem in many areas, localized management has been proposed as a possible solution. Localized management involves the "surgical" removal of a social group(s) in an area deemed sensitive to browsing pressure. However, this technique has only been tested in a highly philopatric, low-density, and un-hunted deer herd in New York. In that study, deer did not recolonize areas previously occupied for > 2 years after the removal of an entire social unit. Herein, we describe our experimental investigation of localized management in a high-density and philopatric white-tailed deer herd, and discuss our findings. The study was conducted on the MeadWestvaco Wildlife and Ecosystem Research Forest located in Randolph County, West Virginia. Since February of 1999, 175 female white-tailed deer have been outfitted with radio-collars to identify social groups and estimate range overlap among social groups. During January and February of 2002, 51 deer were removed within a 1.2 km² area, encompassing 2 forest regeneration areas of northern hardwood forest type (both less than 3 years and approximately 14 ha in size). Currently, only one of the remaining 99 radio-collared females is occupying the removal area. Track counts conducted in December 2002 and January 2003 continue to show that the removal area has a lower deer density than surrounding areas. We will collect telemetric, observational and track count data, as well as browse surveys for > 2 years. If localized management can reduce damage in regeneration areas to acceptable levels over the long-term (ca. 4-5 years), forest and wildlife managers could inexpensively ensure regeneration success, sustain forest structure and composition and maintain forest biodiversity.

Localized Management of White-Tailed Deer within Forest Regeneration Areas of the Central Appalachians

T.A. Campbell, B.R. Laseter, D.A. Osborn, and K.V. Miller, Warnell School of Forest Resources, University of Georgia; W.M. Ford USDA Forest Service, Northeastern Research Station

Abstract: Localized management proposes that a semi-permanent (£ 10 year) elimination of white-tailed deer can be achieved through the "surgical" removal of a deer social group(s). However, this hypothesis has only been tested in a highly philopatric, low-density and un-hunted deer herd in New York. In that study, deer did not re-colonize areas previously occupied for ³ 2 years after removal of an entire social unit. These findings are contrary to traditional ideas of deer dispersion patterns in the eastern United States that suggest white-tailed deer will rapidly seek, find and colonize areas of reduced social pressure, particularly when population densities are high. No scientific examination of localized management exists beyond the study in New York. Herein, we describe our experimental investigation of localized management in a high-density and philopatric white-tailed deer herd and discuss our findings to date. The study was conducted on the MeadWestvaco Wildlife and Ecosystem Research Forest located in Randolph County, West Virginia. Since February of 1999, 175 female white-tailed deer have been outfitted with radio-collars to identify social groups and estimate range overlap among social groups. During January and February of 2002, all deer (n = 51) were removed within a 1.2 km² area, encompassing 2 forest regeneration areas of northern hardwood forest type (both less than 3 years and approximately 14 ha in size). Presently, none of the extant radio-collared deer (n = 120) have dispersed into the localized removal area. We will continue to monitor (for ³ 3 years) deer movements and browsing pressure to determine the effectiveness of localized management at reducing deer damage to forest regeneration in the central Appalachians.

Social Groups Of Female White-Tailed Deer: How Much Seasonal Movement Occurs In The Central Appalachians?

Benjamin R. Laseter, Tyler A. Campbell, David A. Osborn, Karl V. Miller, Warnell School of Forest Resources, University of Georgia; and W. Mark Ford USDA Forest Service, Northeastern Research Station

Abstract: Across their range, white-tailed deer form matrilineal groups, composed of older females and their female descendants. In migratory herds, researchers have used summer and winter range commonalities to delineate these matrilineal groups. In the central Appalachians of West Virginia, deer are nonmigratory and seasonal ranges are less discrete than in migratory herds. We hypothesized that group members (as defined by degree of home range overlap) would show more affinity for one another during winter versus summer months. Furthermore, we expected decreases in home range size and number of deer per group in summer. To test these hypotheses, we examined a total of 11,847 radio-telemetry locations (5,176 winter and 6,671 summer) collected from May 1999 to June 2001 for 63 female deer. Average home range size differed between winter and summer months (1.4km² versus 1.1km², respectively) but was not highly significant (P = 0.09). Additionally, average female group size (number of deer/group) was smaller in winter than summer (3.9 versus 5.2, respectively) but was not highly significant (P = 0.07). Our results suggest that in the absence of discrete seasonal ranges, average home range size and group size might not adequately reflect seasonal social group dynamics in the central Appalachians.

Mammalian Predator Response to Intensive Forest Management in the Central Appalachians

Sheldon Owen and John Edwards, West Virginia University; Mark Ford, USDA Forest Service

Abstract: In an attempt to more completely understand predator ecology and the role of predators in an internally fragmented, intensively managed forest, we are monitoring raccoon (Procyon lotor) den site selection and seasonal movements on the MeadWestvaco Wildlife and Ecosystem Research Forest (MWERF). The long-term Appalachian Landscape Ecology Project in progress on the MWERF provides a framework for the study of local and landscape-level factors that influence predator movements and habitat use. We chose to study raccoons as the primary predator species because of their ease of capture, known nest-predating habits, and their link to the Allegheny woodrat/raccoon roundworm (Baylisascaris procyonis) parasite cycle. Also, because of their ability to respond to habitat change at both a micro- and macro-landscape scale, raccoons were chosen for more intensive movement, habitat use, and den site selection research using radio telemetry.

West Virginia is in the middle of a rabies epizootic and the disease has spread to most of the state; however, the majority of positive rabies cases are east of the Allegheny Mountains. The westward spread of raccoon rabies has been much slower than the spread to the northeast, probably in response to lower raccoon populations in the Allegheny Mountains. Although raccoons are the primary host of raccoon rabies the disease can "spill over" into other susceptible species. Identification of possible rabies vectors and their relative densities is key for predator and forest management. Serologic data combined with habitat movements may lend inference to target areas for oral rabies vaccination bait application thus reducing the number of baits and cost. Also, further rabies surveillance can be conducted along the leading edge of rabies spread and in high elevation areas.

Currently 45 raccoons (36 adults, 9 juveniles) have been captured in live traps during 1200 trap nights. Nontarget captures include striped skunks, opossums, red squirrels, and eastern cottontails. We attached radio-collars to 33 adult raccoons. We were able to determine homerange and habitat use (<50 foraging locations each) on 21 individuals (10 females, 11 males). From the 33 radio-collared individuals we found 158 den/diurnal resting sites. Diurnal resting sites consisted of cavities, log piles, rock dens, slash piles, and tree limbs. We have tested 22 individuals (7 females, 15males) for raccoon roundworm by fecal floatation examination. We are also working with USDA Wildlife Services, WV Department of Health and Human Resources, and WV DNR to investigate the epidemiology of raccoon rabies across West Virginia.

Causes of Failed Oak Regeneration

Rachel J. Collins and Walter P. Carson, Department of Biological Sciences, University of Pittsburgh

Abstract: Oak forests are not regenerating and appear to be converting to forests dominated by species of low wildlife and timber value. The causes of failed oak regeneration remain poorly understood and controversial. To address these issues, we have established large-scale experiments that are designed to test four key hypotheses proposed to explain failed oak regeneration and other recent changes in forest species composition in eastern deciduous forests. These include 1. The fire hypothesis: Periodic fires promote oak and hickory regeneration. 2. The herbivory hypothesis: Browsing by high populations of deer prevents oak regeneration and reduces diversity. 3. The seed predation hypothesis: Seed predation by vertebrates prevents oak recruitment and reduces diversity. Finally, 4. The gap hypothesis: Oaks, and other species, require large treefall gaps to regenerate. These hypotheses are not mutually exclusive; thus, we are testing them by using a factorial design where we are experimentally manipulating fire, browsing by deer, vertebrate seed predators, and canopy gap formation. These experiments are fully operational at two sites in West Virginia. We are censusing 128, 20m x 20m plots and permanently tagged and measured approximately 38,000 seedlings, saplings, and adults of all tree species. Our sampling method is specifically designed to tease apart the relative role of the four factors described above on survival of seedlings present at the beginning of the study and on the establishment of new recruits. This will help us understand the relative importance of propagule input and seedling survival. Further, this approach will allow us to rigorously quantify the relative influence of the four factors on size specific mortality and subsequent growth and recruitment among species. We believe we can discover the causes of failed oak regeneration and elucidate the underlying mechanisms responsible for associated shifts in understory tree species abundance that are occurring in many eastern deciduous forests.

One of the unique aspects of this work is that we have the identical experiment set up at two study sites which represent two forest types (MeadWestvaco and Fernow Sites). Therefore, although each study site is self-sufficient, they are highly integrated. For that reason, below we update the accomplishments of both sites.

We recensued the Fernow site including all marked trees greater than 20 cm tall. We also identified, marked and measured all new individuals. This data set contains more than 25,000 individuals. I (Collins) recensused all small and newly emergent seedlings (1-20 cm tall). These small seedlings are in 5, 1 X 1 m plots that are imbedded in each 20 X 20 m plots. In June, Alex Royo and I preformed a census of the herbaceous plant communities in all research plots at both sites (MeadWestvaco and Fernow). These data will give us the unique opportunity to investigate the effects of the treatments (fire, deer herbivory and canopy gaps) at a larger scale and with greater replication than has been done previously on forest herbaceous communities.

Seed Rain Component: Another accomplishment we made this summer collecting the seed rain from the traps. Between the two sites (MeadWestvaco and Fernow) there are over 500 traps and thanks to a tremendous amount of work by Doug Owens and others at the Fernow all the seeds were collected from all traps three times. We are now in the process of identifying and enumerating the seeds. By placing four 0.25m² seed traps in each of the research plots, we will get an excellent estimation of the seed rain in the plot. By monitoring seed rain for all species every year, we will gain insight into the effects that variable seed production has on recruitment.

Ecology and Natural History Investigation of the Endangered Virginia Northern Flying Squirrel

Jennifer M. Menzel and W. Mark Ford, USDA Forest Service, Northeastern Research Station; John W. Edwards and M. Alex Menzel, West Virginia University; Steven Stephenson, Fairmont State College; Jane L. Rodrigue, USDA Forest Service, Northeastern Research Station

Abstract: Little is known about the ecology and habitat requirements of the endangered Virginia northern flying squirrel (Glaucomys sabrinus fuscus) within its distribution in the central Appalachians of Virginia and West Virginia. To gain insight into the habitat requirements of this subspecies, we analyzed nest trees and habitat use of Virginia northern flying squirrels radio tracked during the summers of 2000 and 2001, along with various micro- and macro-habitat characteristics of known occupied and unoccupied areas in 2001 and 2002 on the Monongahela National Forest and the MeadWestvaco Ecosystem Research Forest. Results of the 59 nest trees analyzed indicate that the squirrels change nest trees frequently. Sixty-nine percent of the nests were in cavities and 31% were drey nests. Yellow birch (Betula alleghaniensis), black birch (Betula lenta) and American beech (Fagus grandifolia) were selected for nest trees more than expected. A large portion of nest trees were in larger and taller trees than trees in the surrounding stand. There also were a significant number of trees located next to or near skidder and hiking trails. Results from the habitat use analysis show that the Virginia northern flying squirrels preferred red spruce (Picea rubens) and mixed hardwood-spruce over areas of pure hardwoods and open space at multiple spatial scales. Analysis of habitat characteristics at occupied and unoccupied sites also revealed the importance of increased percent component of red spruce, Norway spruce (Picea abies), eastern hemlock (Tsuga canadensis), or balsam fir (Abies balsamea) to the presence of Virginia northern flying squirrel. The information gained in these analyses is being used to build a model to predict high quality Virginia northern flying squirrel habitat utilizing existing GIS layers and high-resolution satellite imagery. These models will be used to prioritize high quality habitats for restoration and protection in future analyses and should prove helpful in ongoing and future Habitat Conservation Plan efforts in the central Appalachians.

Watershed and SMZ Characterization in the Allegheny Plateau of West Virginia and the Piedmont of Virginia

E.P. Sharp, A.M. Walker, and W.M. Aust, Department of Forestry, Virginia Tech; C.A. Dolloff, USDA Forest Service, Southern Research Station; P.D. Keyser, Forestry Division, MeadWestvaco Corporation

Abstract: Streamside Management Zones (SMZs) are vitally important for the health and management of forested watersheds. SMZs function to protect water quality through nutrient uptake, erosion control, providing channel stability, and thermal buffering. In addition, wildlife corridors are maintained through the use of SMZs. Continuing research of how water quality protection is critical in designing effective SMZs. Forest managers would benefit from this knowledge in choosing the appropriate harvest level and width most suitable for SMZs in their area.

The objective of this study was to conduct a thorough pre-harvest characterization of 18 watersheds in the Peidmont Province of Virginia. Each watershed had a treatment randomly assigned. Our treatments consisted of a 100 ft wide SMZ with a 50% residual harvest level; a 100 ft SMZ with no harvest as the control; a 50 ft SMZ with a 50 % residual harvest level; a 50 ft SMZ with no harvest; or a 15 ft wide uncut stringer.

All data measurements were conducted in established transects at the head, middle, and lower portion of each watershed. Field measurements are broken down into two main areas: in-stream measurements and SMZ measurements. In-stream measurements include examination of large woody debris, sedimentation, Rosgen stream classification, canopy cover, streamwater temperature, litter inputs, stream macroinvertebrates, and streamwater chemistry. The SMZ measurements include overstory vegetation characterization (species, age, size), understory vegetation (ground cover, species composition), erosion, canopy cover, litter layer, and large woody debris. Sediments rods and the USLE were implemented in each watershed to estimate current erosion rates pre-harvest.

The collection of baseline data is finally complete and all of the data is currently being analyzed. Baseline monitoring began in early spring of 2002 and was just completed in early January 2003. During the spring of 2003 harvesting will begin in the study sites. Monitoring will continue until 2010.

Watershed and SMZ Characterization in the Allegheny Plateau of West in Relation to Various Streamside Management Zone Widths and Harvesting Intensities

E.P. Sharp, W.M. Aust, Department of Forestry, Virginia Tech; C.A. Dolloff, USDA Forest Service, Southern Research Station; P.D. Keyser, Forestry Division, MeadWestvaco Corporation

Abstract: Streamside Management Zones (SMZs) are vitally important for the health of forested watersheds under management. SMZs function to protect water quality through nutrient uptake, erosion control, providing channel stability, and thermal buffering. In addition, wildlife corridors are maintained through the use of SMZs.

The implementation of SMZs and SMZ characteristics is variable among different states. For example, most states require different widths of SMZs and varying harvest levels in SMZs. More research towards understanding how SMZs protect water quality is vital is designing SMZs most effectively. Forest managers would benefit from this increased knowledge by understanding the width and harvest levels most suitable for SMZs.

The objectives of this study can be broken down into three main areas. First, we want to characterize watersheds, SMZs, and water quality in first order streams in our study area. Secondly, we want to examine how various SMZ widths and harvest levels affect stream water quality in our study area. Lastly, we want to compare how the Allegheny Plateau Province of West Virginia compares to the Piedmont of Virginia in relation to watershed and SMZ characterization. This study is also simultaneously being replicated in the Piedmont of Virginia. All treatments, data collections, and data analysis are similar.

Our treatments will consist of the following: a 100-ft SMZ width and 50% residual harvest level; a 50-ft SMZ width and 50% residual harvest level; a 50 ft wide uncut SMZ; a 15-ft-wide uncut stringer along streams; and the control will be a 100-ft wide uncut SMZ. The "before and after" effects of harvest on erosion and water quality will be examined.

Field measurements can be broken down into two main areas: in-stream measurements and watershed measurements. In-stream measurements include examination of large woody debris, sedimentation, Rosgen stream classification, Manning streamflow calculation, pool/riffle ratio, canopy cover, streamwater temperature, litter inputs, stream macroinvertebrates, and streamwater chemistry. Watershed measurements include overstory vegetation characterization (species, age, size), understory vegetation (ground cover, species composition), erosion, canopy cover, litter layer, and large woody debris.

During the past year, the bulk of the field measurements were completed, and on-going measurements continue to be collected. Data analysis is in the initial stages and is a continuing process. During the spring of 2002, initial treatment installation and harvest will begin. Monitoring will continue until 2010.

Raptor Species Assemblage, Relative Abundance, and Responses to Intensive Forest Management in the Central Appalachian Mountains

Rebecca D. M. Smith, Tim Dellinger, and Petra Bohall Wood, West Virginia Cooperative Fish and Wildlife Research Unit BRD/USGS and Division of Forestry, West Virginia University; W. Mark Ford, USDA Forest Service, Northeast Research Station; John W. Edwards, Division of Forestry, West Virginia University

Abstract: Avian predators occupy an ecological niche created by a combination of habitat modification due to forest management, a general increase in fragmentation and edge, and recent removal of large predators. In the heavily forested central and southern Appalachian Mountains, habitat relationships for avian predator abundance and diversity have been poorly documented. Direct responses of these predators to intensive habitat modification have rarely been studied. Our study area, MeadWestvaco Wildlife and Ecosystem Research Forest, is an active, industrial forest where on-going research projects of potential prey species have documented considerable mammalian and avian predation effects on various mammal and bird species. The area was divided into six 526-ha compartments representing 2 replications of each of 3 timber harvest rotations (20-, 40-, and 80-year) to study the effects of these harvest rotations on various ecological components. Objectives for this project were to:

1) Determine species assemblage, relative abundance, and basic ecology of breeding and migratory raptors,

2) Provide an understanding of how raptors respond to fragmentation of various harvest intensities, and

3) Determine habitat use patterns of breeding raptors.

We conducted monthly broadcast surveys for diurnal raptors from May-Nov 2000, Mar-Oct 2001, and Apr-Oct 2002 (n = 22). Surveys were conducted at 48 stations established systematically throughout the 6 compartments using standard protocol from the literature with a Great-horned Owl (Bubo virginianus) vocalization. Overall, we detected 9 species, although occurrence of some species varied seasonally. Incidental sightings confirmed 3 additional raptor species not detected on surveys. The surveys were divided into 2 time periods: a Great-horned owl vocalization period and a silent listening period. More raptor responses were detected during the silent listening period. Seventy-four percent of all raptor responses were a type of vocal response (e.g., vocalization only, fly and call, or perch and call). The most common species detected were Red-shouldered (Buteo lineatus) and Broad-winged (B. platypterus) Hawks.

Using analysis of variance and an alpha level of 0.05, the mean number of responses did not differ among the 3 harvest rotations during 2000-2002 for all raptors, edge-dwelling species, forest-dwelling species, or any individual species. During 2001, however, Red-shouldered Hawk responses differed among the 3 harvest rotations (F = 11.2, P = 0.0406) with the highest number of responses in the 20-year rotation.

Nests of Northern Goshawks (Accipter gentilis; n =1), Cooper's Hawks (A. cooperi; n = 1), Red-tailed Hawks (B. jamaicensis; n =6), Broad-winged Hawks (n = 4), and Red-shouldered Hawks (n = 7) were located and monitored; all but three of these nests successfully fledged at least one young over the three breeding seasons.

Songbird Abundance and Nesting Success on Landscapes of Differing Harvesting Intensities on an Industrial Forest

Rachel Dellinger and Petra Bohall Wood, West Virginia Cooperative Fish and Wildlife Research Unit, BRD/USGS and Division of Forestry, West Virginia University; P. D. Keyser, Forestry Division, MeadWestvaco Corporation

Abstract: While many studies have indicated that songbirds can sustain viable populations in areas managed for timber production, what has not been determined are the specific, long-term responses of birds to various intensities of harvesting. Given that logging on public and private lands in West Virginia will continue in the future, it is important to document the effects of these activities on songbird populations. The specific objectives of the present study are to:

1. determine if forest-interior species show negative demographic responses to harvesting intensity and identify the level at which this occurs;

2. determine if early successional species show positive demographic responses to harvesting intensity and identify the level at which this occurs;

3. quantify songbird abundance, richness, and nesting success in landscapes under three different levels of disturbance.

To make it possible to study landscape-level effects, in 1998 the research forest was divided into 2 blocks containing 3 compartments each 526 ha in size. Within the blocks, each of the 3 compartments was randomly assigned a level of disturbance intensity approximating a harvesting rotation of 20, 40, or 80 years. In 2002, an additional block of 3 compartments on the Panther Run Tract was added to the study as a 3^rd replicate. To assess songbird abundance, 18 permanent inventory points were selected per compartment as sites for point counts, and counts were conducted during the breeding season using standard point count protocols. To evaluate nesting success, 2 20 ha nest searching plots were established per compartment, in which full-time technicians found and monitored songbird nests.

The 2 species most commonly detected during point counts during 2001-2002 were the Red-eyed Vireo (Vireo olivaceus) and the Black-throated Blue Warbler (Dendroica caerulescens). Forest interior birds were the most abundant habitat guild throughout the study sites in 2001, followed by interior-edge and lastly edge species (P less than 0.0001). The relative abundance of forest interior, interior-edge, and edge guilds remained the same in 2002 as in 2001 (P less tan 0.0001). In 2001, forest interior species were less abundant in the 80-year rotation compartments than in the other two treatments, and interior-edge species had significantly different abundances among all three treatments with highest abundance in the 40-year rotation and lowest in the 80-year (P less than 0.05). The abundance of edge species was similar among treatments. In 2002, no differences were detected among treatments within any of the guilds.

In 2001, 69 nests of 22 species were found and monitored. Survival rates for all species combined in 2001 were highest in the 40-year rotation compartments and lowest in the 20-year rotation (P less than 0.0001). To increase the sample size on a per-species basis, search efforts in 2002 were limited to 11 target species that represent the range of habitat, nesting, and migratory guilds present on the study sites. Nest survival in 2002 was highest in the 40-year rotation compartments and lowest in the 80-year rotation (P less than 0.00001).

Amphibian Species Diversity, Relative Abundance, and Habitat Use in a Central Appalachian Industrial Forest

Lori Williams and Carola Haas, Department of Fisheries & Wildlife Sciences, Virginia Tech; P. D. Keyser, Forestry Division, MeadWestvaco Corporation

Abstract: In 2002 our goal was to provide a second year of baseline data for the amphibian component of a long-term landscape ecology project on the MeadWestvaco Wildlife and Ecosystem Research Forest (WERF) in Adolph, West Virginia. This ongoing project was designed to monitor the effects of large-scale silviculture on biodiversity and ecosystem health by manipulating the distribution of forest age-classes within a landscape through replicated experiments. The current study will identify habitat features related to amphibian diversity, abundance, and distribution. In the summer of 2002 we used 4 sampling techniques in our research. We sampled 219 sites (10 m x 10 m plots) for amphibians on 9 forest compartments, 6 on the WERF and 3 on adjacent MeadWestvaco property. We used stepwise multiple linear regression to examine relationships between habitat characteristics and amphibian abundance at the forest stand level. At a 0.10 significance level, 4 habitat characteristics were significant in a model to predict abundance: index of % rock cover, leaf litter biomass (g), soil pH, and an index of % canopy cover (R-square=0.2331, p less than 0.0001). We used nighttime area constrained searches to capture surface-active salamanders in 9 reference cuts (28-acre clearcuts), 3 postharvest and 6 preharvest. Paired t-tests revealed significant declines in abundance from preharvest to postharvest on the 3 compartments clearcut in 2001 (p=0.0015, std dev=10.344, std err=2.986). We used coverboards to provide preharvest and postharvest data on the distribution and abundance of salamanders within reference cuts adjacent to mature forest. Similar to results of the nighttime sampling, paired t-tests showed significant declines in abundance from preharvest to postharvest on the 3 compartments clearcut in 2001 (p=0.0011, std dev=3.217, std err=0.657). To study amphibian movement in riparian habitat, we placed a series of drift fences, funnel traps, and pitfall traps at intervals of 15 m, 30 m, 60 m, and 100 m away from a perennial stream, an intermittent stream, a permanent pond, and a seasonal pond. Combining results from all sites, we captured 1712 amphibians, 153 of which were recaptures. Recaptured individuals were found at all distances sampled. 12 salamander species and 5 frog/toad species were captured.

Brook Trout Movement Ecology and Habitat Use in Low Order Streams in the Central Appalachians

Marisa Logan and Kyle Hartman, Division of Forestry, West Virginia University

Abstract: Brook Trout (Salvelinus fontinalis) are the only salmonid native to the Appalachians. They are a resilient species, populating small, low productivity streams. Although such streams have historically been abundant in the Appalachians, their populations have been threatened due to habitat degradations such as logging and acidification of their waters. Stream fish's movement depends on several factors, including water quality, spawning, food availability and suitable habitat. The purpose of this study is to determine microhabitat selection of brook trout in varying habitat complexity and how habitat complexity influences movement. This was accomplished by inserting radio telemetry tags within the fish and tracking them daily and a few for 24 hours. The study site was Birch Fork, a tributary of the Middle Fork River, Randolph County West Virginia, and the property owned by Mead-Westvaco Inc. The fish were originally dispersed in sections with varying habitat complexity. Each time they were located, habitat measurements were taken. These results were compared with habitat availability for the entire stream. Each fish's home range and cumulative range was also calculated. Within the first week of tracking, some fish moved up to a kilometer in one day. Others showed resiliency by swimming upstream, through large, seemingly impossible, debris jams.

Addition of Large Woody Debris to Enhance Stream Habitat and Brook Trout Populations

John A. Sweka and Kyle J. Hartman, Division of Forestry, West Virginia University

Abstract: Large woody debris (LWD) is an important component to stream habitat. It increases organic matter storage; sorts and stores fine sediment; and provides cover and refuge areas for fish. Past timber harvest activities removed much of the source of LWD and current riparian forests lack the age to contribute LWD in amounts equivalent to old growth systems. The purpose of this study was to determine if the artificial addition of LWD to stream systems can increase critical stream habitat for brook trout (Salvelinus fontinalis) and potentially enhance their populations. Such additions may prove to be a useful mitigation tool to the timber industry in the future.

Eight study streams within the Middle Fork River watershed, Randolph Co. WV, were selected for the study. This study began in the fall of 1999 and during the first year, initial stream habitat and brook trout populations were assessed. LWD was added in August 2000 at a density of approximately 5 trees per 100-m of stream. Four of the study streams had LWD added to three 300-m reaches of the stream, while the remaining four had LWD added to only one 300-m reach, so as to compare different scales of habitat manipulation on the brook trout population response.

The addition of LWD has had some positive effect on stream habitat. The addition of LWD increased the volume of LWD within manipulated areas by an average of more than 120%. It appears that the greatest effect of the added LWD occurred during the first year post-manipulation (2000 - 2001). Functionality (i.e. the proportion forming pools, storing organic matter, etc.) did not change between 2001 and 2002. A total of 20 pools were formed from added LWD during the first year post-manipulation, but only three additional pools were formed during the following year. The percent of wetted area comprised of pool habitat increased significantly by the summer of 2002 (paired t-test: t15=-1.89, p (one-tail) = 0.04) in manipulated reaches.

Adult brook trout density was equivalent between the scales of habitat manipulation at the beginning of the study (ANOVA: F1,15 = 0.19, p = 0.67). Density varied over the two years post-manipulation (Repeated measures ANOVA: F4.24 = 18.31, p less than 0.01), but there was no difference between the scales of habitat manipulation (F1.6 = 1.88, p = 0.22). Adult brook trout density in the fall of 2002 did not differ significantly from the initial density in the fall of 1999, indicating factors other than LWD and pool area are currently governing brook trout populations. The increase in the number of pools and percentage of pool habitat may become important during years of drought when refuge areas are most critical. Over the course of this study, stream flow has been moderate with no extremes (i.e. droughts or floods).

Although brook trout populations have not responded to LWD additions, these habitat manipulations have had a positive influence on stream habitat and this inexpensive technique may be used a mitigation tool in the future.

Stand Dynamics and Restoration Implications of a Second-Growth Red Spruce Forest in the Central Appalachians

Thomas M. Schuler, W. Mark Ford, USDA Forest Service, Northeastern Research Station; Rachel Collins, Department of Biological Sciences, University of Pittsburgh

Abstract: Trees and coarse woody debris were inventoried in an isolated montane red spruce forest located in Randolph County, West Virginia. Stand characteristics were quantified and compared to structural characteristics of old-growth forests of similar composition. Based on dendroecological analysis, stand initiation occurred in the early 1920's when watershed-wide timber harvesting was occurring, live tree basal area greater than or equal to 10 cm diameter at breast height (DBH) (44.5 m²~ha^-1), snag density greater than or equal to 10 cm DBH (256 ha^-1), and total volume of fallen logs (CWD) (86 m³~ha^-1) did not differ significantly from old-growth attributes. However, snag basal area greater than or equal to 10 cm DBH (4.6 m²~ha^-1), height of dominant and co-dominant spruce (24.7 m), and maximum red spruce DBH (48.0 cm) were significantly less than would be expected in old-growth forests of similar type. Red spruce made up over 40% of the dominant crown class trees and was often an emergent above the main canopy. However, the stand was largely eastern hemlock in terms of relative importance (51.7%) with red spruce (18%) and red maple (16.7%) representing about equal lesser fractions. Stocking was 147% of the B-line basal area, which was evident in the overall slow growth rate for both spruce and hemlock. Current conditions would support a commercial thinning which would accelerate growth rates of individual trees, increase CWD, could alter species composition to favor the more desirable spruce, and decrease the time required to attain additional structural and functional characteristics of old-growth montane red spruce forests.

Population Reaction-Diffusion Dynamics on Spatially Structured Forest Landscapes

Craig Loehle and Ben Wigley, National Council for Air and Stream Improvement; Patrick Keyser, Forestry Division, MeadWestvaco Corporation

Abstract: Gaining a complete understanding of the development of patch dynamics in a managed forest context requires information on the functional nature and extent of movement barriers. In a truly fragmented system, "non-habitat" areas are viewed as hostile and presumed to act as barriers to normal movements and dispersal across the landscape. This fragmentation ultimately would lead to isolation of populations and the development of a meta-population system. In an internally fragmented forest context, these non-habitat areas (young forests for late successional species, and old forests for early successional species) may be far less hostile than would be the case in an agricultural or urbanized context. Furthermore, the dynamic nature of young forests may make the "non-habitat" areas only ephemeral barriers. Thus metapopulation dynamics may not develop.

We propose to study several species to determine the degree to which the fragmenting effects of the patches on a managed forest serve as barriers. Studies examining small mammals, amphibians, and possibly birds and plants are being designed. These will likely be based on a manipulative study design that will involve the creation of a set of four replicated harvest units that will include isolated "islands" of late successional forest surrounded by an early successional barrier of harvested timber. Connecting corridors will be included as a treatment on some of the harvest units to determine the degree that these contribute to movement across the barrier.

We anticipate initiating field work during 2004 with pre-treatment sampling. Treatments would be applied in 2004-5 with post-treatment sampling in 2005 and 2006. Field data will be used to parameterize several metapopulation models that rely on BIDE data (birth, immigration, death, emigration) to determine if the assumptions of metapopulation theory apply in heavily a forested context. This is an important consideration from a conservation of biodiversity and forest sustainability standpoint.