Student Theses - James M. Dyer

Eastern hemlock (Tsuga canadensis) is an important foundation species, altering its environment and supporting a variety of organisms not generally found in deciduous forests. The hemlock forests of the Hocking Hills region in southeast Ohio are particularly unique, occurring in isolated pockets in ravines and on steep slopes on the western edge of the range of the species. However, this unique ecosystem is under threat from hemlock woolly adelgid (HWA, Adelges tsugae), an invasive insect pest which entered the state in 2012.

The goal of this study is to characterize the hemlock stands of the Hocking Hills just prior to HWA infestation and describe how these forests have changed in the past decade in terms of growth and mortality. An additional goal is to record non-hemlock species growing alongside these hemlock stands which could become dominant following HWA-induced mortality, as well as any invasive plants present which could pose a threat to native diversity following such a disturbance. In order to accomplish these goals, thirty 20 x 40 m plots established roughly a decade ago were resampled. Within these plots, the diameter of tagged trees was measured and their health and canopy position assessed, and saplings counted as well. Transects measuring 10 x 50 m were established on the upper slopes or ridges above each plot to record non-hemlock species that have the potential to seed into the hemlock stands below. The existence of any invasive plants was also recorded in both plots and transects.

It was found that hemlock, for the most part not yet impacted by HWA, was still the dominant species in all plots, with an average importance value of 47.78. Storm damage and competitive thinning rather than HWA infestation appeared to account for the majority of mortality. Growth and mortality rates among hemlocks (on average, 1.18% annual growth and 13.24% total mortality) were both comparable to those seen in deciduous species on the plots. A mixture of hardwood species such as tulip poplar (Liriodendron tulipifera), chestnut oak (Quercus montana), white oak (Quercus alba), sweet birch (Betula lenta) and red maple (Acer rubrum) were common components of hemlock stands. Therefore, these trees can be expected to be among the first species to dominate the canopy following future hemlock mortality. Chestnut oak dominates the deciduous stands above the hemlock plots, with red maple, white oak, and red oak (Quercus rubra) also commonly recorded. Each of these species could contribute to the composition of post-HWA forests by dispersing to the slopes and valleys below. Invasive plant species were found at one-third of sampling sites, with Japanese stiltgrass (Microstegium vimineum) being by far the most numerous. This is worrying, as the expansion of invasive species following hemlock mortality could potentially inhibit the regeneration of native species. In short, a variety of deciduous species appear poised to take advantage of increased resource availability in the absence of hemlock, but unfortunately the same can also be said for invasive plants. The results of this study can inform forest management and conservation efforts in the isolated and unique hemlock stands of southeast Ohio.

This project acquired high-resolution, RGB images of a deciduous forest in Athens, Ohio from an unmanned aerial vehicle over the course of a year. With UAVs, greater temporal and spatial scales exist today than what were available in the recent past with the full capabilities of these images yet to be fully assessed. The goal of the project was to follow the different phenology patterns of various deciduous trees to accurately map the project site while formulating a reproducible methodology for future work. Acquiring mosaics for 29 dates in 2017, a supervised Maximum Likelihood Classification scheme was employed to classify various species using 15 of the mosaics. Simultaneously, spectral profiles were built for 16 common species at the site using 25 of the mosaics. A community level map displaying three main groups was produced for the study site along with a phenology calendar which highlights key dates for identifying various species. The results of this project show that many species have a high amount of overlap in their spectral signals following phenology, making differentiation of individual species difficult when employing a pixel-based approach alone.

This research examined 10 similar surface coal mines in southeast Ohio to assess the factors broadly influencing the cover and composition of the successional vegetation mosaic. Transects were surveyed on opposing north- and south-facing aspects within each mine to comprehensively characterize the vegetative community. Statistical analysis strongly indicated that measures of vegetation cover were almost exclusively associated with edaphic properties such as soil pH and electrical conductivity. In clear contrast, species richness and the composition of the vegetative community were influenced primarily by physical landform features and the size and shape of mine extents. While some features of surface mines may act as barriers to the dispersal of native species, the element of distance is the primary limitation to natural succession. Areas furthest from the functional edge seem to be more readily occupied by non-native species once conditions improve, representing a missed opportunity for native establishment and may act as a barrier to future establishment of natives. These results may begin to simplify issues for understanding the long-term recovery of mine lands, in terms of the goals of reclamation and restoration. Furthermore, it highlights the need for incorporating more ecologically-based objectives and practices into landscape planning and management.

Restoration of American chestnut (Castanea dentata (Marsh.) Borkh.) is currently underway in eastern North American forests. American chestnut and eastern hemlock (Tsuga canadensis (L.) Carr.) trees historically co-occurred in these forests. Today, hemlock-dominated forests are in decline due to hemlock wooly adelgid (Adelges tsugae Annand) infestation, and as such, may serve as appropriate habitat for chestnut reestablishment. To investigate this notion, I evaluated the performance of American chestnut seedlings planted under healthy eastern hemlock-dominated canopies. Two process-oriented greenhouse experiments were also performed to study the response of American chestnut to drought stress and to test the competitive performance of chestnut against red maple (Acer rubrum (L.)), the most abundant hardwood found in the understory of regional hemlock-dominated forests. After two growing seasons, mean chestnut seedling survival in the field experiment was 6.6%. Seedling survival was significantly higher among trees that lacked a protective tree tube, suggesting low light levels played a major role in seedling mortality. In the drought stress experiment, American chestnut exhibited significantly higher mortality under a severe drought treatment compared to the control group. The moderate stress treatment also responded poorly to drought; however, results did not differ significantly with the control group. In the competition experiment, American chestnut grew significantly taller than red maple. The results of this study caution against the underplanting of bare-root chestnut seedlings in low light conditions or where moisture may be limiting to establishing seedlings. The results also validate the classification of American chestnut as an aggressive understory competitor, when environmental conditions are appropriate.

Between July and October 2005, 42 sites across Ohio Brush Creek watershed were surveyed to assess the spatial distribution of native freshwater mussels (Unionidae). Freshwater mussel shells were recorded at 28 out of 42 sites representing 14 native species. A total of thirteen species were recorded at 19 sites as living or fresh dead. Associations between the presence, diversity, and abundance of freshwater mussels and coarse-scale variables (drainage area, stream gradient, and percent land cover) and fine-scale variables (200 meter stream-reach habitat features based on Ohio EPA’s Qualitative Habitat Evaluation Index (QHEI)) were explored using correlation and chi-square analysis. The presence, diversity, and abundance of mussel shells were associated with both coarse and fine scale variables. Drainage area and stream reaches with excellent channel development, high amounts of habitat cover, maximum water depths > 1 meter, and riffle depths > 5 cm were all associated with the presence, diversity, and abundance of mussels. Stream gradient was also associated with mussel shell presence and diversity, however was not associated with shell abundance due to the high abundance of fat mucket shells in upper reaches of the watershed. Sites with the highest diversity and abundance occurred along the mainstems of Ohio Brush Creek and the West Fork. Thirty-seven native mussel species have been recorded in the watershed. Unfortunately over 40% of these species are listed as either endangered, threatened, or of special concern. Sedimentation due to agricultural runoff and deforestation of riparian corridors has been identified as the primary threat to freshwater mussels in Ohio Brush Creek watershed. It is imperative to collect data that can be explored to find spatial and temporal patterns that exist amongst the mussel community in Ohio Brush Creek watershed. This data can also be used to help guide stream habitat restoration and native mussel re-establishment projects in the watershed.

This study explores the environmental control of two key elements of water balance on fire behavior at two sites in southeastern Ohio, Arch Rock and Tar Hollow. The two elements of water balance are potential evapotranspiration (PET) and actual evapotranspiration (AET). Fire behavior is expressed in this study as fire radiative energy (FRE), which is estimated from time-sequence fire radiative power by remote sensing of infrared radiation during fires. Previous studies have shown that both fuel moisture and fuel load are primary driving forces of fire behavior at a coarse scale. Therefore, in this study, PET is used to suggest fuel moisture in spring and AET is used as a surrogate for productivity during the growing season in order to explain FRE at a fine scale. Results show that a water balance approach helped to explain FRE using categorical AET and PET. At the pixel level, the water balance approach was less successful because the variables did not relate reasonably with space and because of other fine-scale local effects (hydrology, decomposition rates) that were not captured by PET and AET. Aspect strongly influences solar radiation and PET. Thus, aspect was used additionally to assess its indirect effects on FRE. Overall, FRE responded well to aspect at both study sites.

The eastern hemlock (Tsuga canadensis) forests that inhabit the cool moist ravines of the Hocking Hills in southeastern Ohio are in the projected path of the hemlock woolly adelgid (Adelges tsugae). This introduced pest has destroyed entire eastern hemlock stands, including mature old-growth forests, and drastically altered ecosystems. A summary of available knowledge about the ecological significance of eastern hemlock as well as considerations for managing pest populations of the hemlock woolly adelgid is presented. Geographic Information System (GIS) techniques combining high-resolution aerial imagery and ancillary information such as topographic position and soil were applied to map the spatial extent of Eastern Hemlock on the landscape. Baseline information about the structure and composition of local Eastern Hemlock forests, as well as the dynamics of the stands and growth trends are established through a dendroecological analysis of core samples and plot inventories. Recognizing the need to understand the spread of HWA and the possible significant role of migratory birds in dispersal, a survey of birds associated within the hemlock stands via call recordings was conducted. The objective of this research is to add to the scientific knowledge of local eastern hemlock forests and to promote proactive pest management strategies and improved monitoring for early detection.

For twenty years the Ohio EPA has characterized stream health using fish and macroinvertebrate community assemblages, as well as by local habitat within the stream. Since the adjoining landscape may also influence stream health, this research explores a variety of easily attainable GIS land use/cover, population, and forest connectivity data to find which best correlate with stream health. Four different metrics (percent native fish, darters/sculpins, intolerant, and sensitive species) were evaluated against forty-eight different independent variables using multiple regression. Three variables were important for predicting stream health: density of abandoned mine openings in the upstream catchment, forest connectivity in a 66 ha area surrounding the sample point, and percent wetlands (from NLCD) within the HUC-14 subwatershed. Percent forest cover (GIRAS) in the upstream riparian buffer was also important for intolerant species. Despite known macro- and micro-controls on fish communities, landscape variables were able to explain up to half the variance in these fish metrics.

The hemlock wooly adelgid (Adelges tsugae) is a non-native, invasive pest killing eastern hemlock (Tsuga canadensis) trees in the eastern

United States

. The purpose of this research is to evaluate wind and people as dispersal mechanisms and the factors limiting spread of the adelgid in Great Smoky Mountains National Park. The results from binary discriminant analysis indicate that people have the biggest influence on spreading HWA, wind was not a significant factor of dispersal, and cold temperatures can limit the range to which it will spread. This information can be applied in the fight against the adelgid because it can help to define a search area with a higher probability of infestation and to define monitoring and treatment standards for these areas.

The colonization of exotic species in forested areas is a growing concern to the biological diversity of our forests. It is known that exotic species are more common in highly fragmented, disturbed landscapes (e.g. forested landscape with transmission line corridors present) with higher levels of light, temperature and photosynthetically active radiation. This study concentrates on the colonization of exotic species along the transmission line corridor/forest edge, and their ability to spread into the forest interior. In Athens County, Ohio, a landscape has been created in which exotic species can potentially use transmission line corridors as a conduit to move along the corridor as well as from the corridor to the forest. Edge or interior position, aspect (cool or warm) and topographic position (ridge or valley) are environmental factors that may affect the ability of exotic species to move along the corridor/forest edge, then spread into the forest interior by providing environmental conditions favorable to the exotic species.

Overall, 90 pairs (one ridge or one valley) of sites were sampled. At each of these 180 sites, plots were established on warm/cool aspects, and at forest edge/interior, resulting in 720 sample sites. At 2 and 30 meters inside the forest on each side of the corridor, nested circle quadrats of 1-m² for herbs and 5-m² for shrubs and trees were constructed and percent cover per species was estimated using the Modified Daubenmire Scale. A Principal Coordinates Analysis (PCO) was performed, and PCO scores then served as input to a MANOVA to determine if there is a significant difference in exotic species abundance from edge or interior position, warm or cool aspect and ridge or valley topographic position (interactive effects of the three environmental variables were also analyzed). MANOVA results indicate that edge or interior position and an interactive effect of edge or interior position and aspect significantly influence the abundance of exotic species. Although not statistically significant, topographic position may also play a role in the abundance of exotic species. Potential exotic species colonization and displacement of native interior species may increase in a forested landscape that is dissected by transmission line corridors. Eradication and prevention of further exotic species colonization may be the best solution to control exotic species from dominating local forests.

This research evaluates the effectiveness of two protected areas of the northeastern Ecuadorian Amazon (Cuyabeno Fauna Production Reserve and Yasuní National Park) in protecting wildlife populations in the long-term. Because, currently, there are no methods to determine the efficiency of protected areas with reference only to ecosystem characteristics, five large mammal species were chosen as indicator species (Myrmecophaga tridactyla, Ateles belzebuth, Alouatta seniculus, Panthera onca, and Tapirus terrestris). The methodology combined remote sensing, geographic information systems, and population viability analysis in order to estimate the current carrying capacity of the protected areas and the extinction probability of the indicator species during the next one hundred years. The results suggest that individually Cuyabeno and Yasuní cannot be considered efficient because, under current management conditions, they do not protect minimum viable populations. Persistence probability is higher if migration and gene flow are allowed between the protected areas. Finally, a reserve network joining Cuyabeno and Yasuní with a biological corridor was designed to increase the conservation potential of both areas.

The study describes land use land cover (LULC) change in the Gwynns Falls Watershed of Baltimore County, Maryland during the 20^th century. Procedures include (1) mapping the extent of LULC change by integrating air photos from 1938, 1957, 1971 and 1999, with a digitized forestry map from 1915 into a GIS, (2) performing landscape pattern analysis on forest cover for each year, and (3) overlaying LULC maps in the GIS to determine change over time. LULC categories mapped were Urban-High Tree Density, Urban-Low Tree Density/High Permeability, Urban-Low Permeability, Agriculture, Old Field & Pasture, Forest, Water, and Transitional. Minimum mapping unit is 1 ha.

The Gwynns Falls Watershed transformed from an area covered by relatively equal proportions of Agriculture, Old Field/Pasture, Forest, and Urban-Low Permeability surfaces, to an area dominated by Urban-Low Permeability surfaces during the 20^th century. From 1938-1957, 46% of the landscape in the common study area changed, 37% change in LULC occurred 1957-1971, and 35% of LULC changed 1971-1999. Transitions in agriculture allowed the expansion of forest cover onto old fields, similar to the pattern seen elsewhere in the eastern deciduous forest. Increasing development reduced and dispersed the Forest patches, potentially affecting species persistence and dispersal.

Species located in habitat islands tend to show similar isolation effects to those found on true islands. The isolation often causes a reduction in fitness usually directly related to decreased genetic variation. The study species, Clethrionomys gapperi, Gapper's red- backed vole, is a boreal species restricted to mountaintops in the southern part of its range. Populations of C. gapperi from eastern North America were examined for evidence of this reduction in variation through both genetic and morphological means. Tissue from live- trapped voles from New Hampshire, West Virginia, and Tennessee was analyzed for heterozygousity using microsatellite primer pairs. Bilateral characters on skulls from four museum collections were measured and then used in a comparison of fluctuating asymmetry levels among different groups. The DNA analysis reveals that populations found in the south had a lower number of banding patterns in the southern populations, both within an individual population and overall. Morphological analysis of skulls for evidence of fluctuating asymmetry reveals higher levels of fluctuating asymmetry for more southern populations. As high levels of fluctuating asymmetry are often related to low levels of genetic variability, these results are in agreement with the genetic analysis that the southern populations had lower levels of genetic variability. The southern mountaintops are therefore acting as habitat islands, isolated from each other and the more northern "mainland", and have the expected reduction in genetic variation in the isolated populations.

The wolf is identified by the federal government as endangered in the 48 contiguous states, excluding Minnesota where it is listed as threatened; North Dakota's Turtle Mountain pack is listed endangered. Over the past two decades, a relationship has been identified in the Great Lakes Region between wolf presence/absence, road density and human population density. Road density, in particular, has been described as a "yardstick" to measure habitat suitability, resulting in the use of this variable in recovery efforts. Road density and human population density together serve as surrogates for human access to the species. Access, resulting in poaching, increases the difficulty in preserving this keystone species

The objective of this thesis is to examine fine-grained habitat suitability throughout the state of North Dakota. The anthropogenic factors examined include human population density and road density. This thesis reclassifies the state based on parameters from North Rolette Unorganized, the minor civil division (MCD) in the Turtle Mountains in which wolves are denning. In addition, this thesis examines the importance of prey biomass on potential wolf numbers and the significance of spatial scale in defining habitat.

Results suggest the following: (1) large areas of habitat exist within North Dakota in which human population density and road density are low; (2) models of habitat suitability for wolves are formation specific; differences exist as to threshold values from the GLR (forest type formation) and North Dakota (prairie formation); (3) differences between the MCD and county spatial scale exist with regards to wolf habitat suitability in North Dakota; (4) no predictive model can express wolf habitat adequately as long as humans continue to view this endangered species as destructive. Based on the variables examined, the state of North Dakota can support a wolf population. As indicated by numerous authors, wolves will not successfully reestablish their former range so long as people destroy them.

On May 31, 1985, a series of tornadoes ripped through the Allegheny National Forest, in northwestern Pennsylvania. One of the tornadoes, an F4, hit Baker Island, an Allegheny River Wilderness Island, and the surrounding upland. This thesis investigates the woody species regeneration of this riparian site as it compares with the eastern and western upland sites, eleven years after the devastating natural event.

From aerial photographs, three distinctive areas were chosen for sampling on Baker Island and both sides of the Allegheny River: (1) an undisturbed zone, (2) a disturbed zone, and (3) a transitional zone (a designated "edge" area in the upland sites). Island sampling used belt transects consisting of contiguous 20x50m (1000 m² quadrats, while the upland sites were sampled with a single 20x50m quadrat in each area, for all trees (h ≥10cm dbh) and large saplings (h ≥ 2m, <10cm dbh). The small saplings and seedlings were sampled inside a nested 5x10m (50m²) quadrat and a 2x5m (10m²) quadrat, respectively, on the upland sites and within a 1x1 m (1m²) quadrat on Baker Island. Species identification and counts were taken at all sites and dbh measurements were recorded for trees and large saplings. Ordination, TWINSPAN, and cluster analysis helped determine if the regeneration processes were similar between each area.

Results indicate that succession is proceeding on the upland sites; early successional species dominate the eastern and western uplands, such as sassafras and quaking aspen. Beech has a significant presence in the understory on the upland. In contrast, Baker Island has no regeneration occurring, with the exception of a few patches of choke cherry and American elm. The main reason suggested for the limited recovery of woody species on the island is reed canary grass. It dominates the understory everywhere on the island, suppressing any seedling growth that might establish.

Basal area and density decreased in all areas, which is expected after a tornado of this intensity. Within the disturbed areas there was no selectivity in which trees were damaged. Everything in the path of the tornado was either snapped or uprooted. A species-specific response did occur on Baker Island as the damaged lessened, however. Based on existing evidence, silver maple suffered more snapping, while sycamore was more likely to uproot.

Riparian environments are very unique and have not been studied as deeply as upland forest areas have been. This study is significant in helping to bridge the gap in the literature in how a riparian system responds to a natural disaster of this intensity. It also adds to the body of knowledge already established concerning upland forest recovery from a devastating tornado.

With increasing population and urbanization, the need for understanding resource users and the environment in which they reside is necessary for proper management of urban fisheries. For urban residents to enjoy nature within city limits, managers must plan for aquatic amenities such as fishing and other aquatic activities.

Understanding the needs of the fishery user ultimately will help managers to increase the quality of the resource and life in the city. The purpose of this study was to identify the key resource concerns and attitudes of anglers of a fishery located in Winnipeg, Manitoba,

Canada

. The spatial variations within the city and analysis of data in terms of accessibility, use, and perception was the focus of the study.
The user base was sampled by conducting 269 personal interviews from May to August over three distinct areas of the city (Red River south of the forks, Red River north of the forks, and Assiniboine River). The results revealed that there were two primary spatial variations in perceptions of Winnipeg anglers. Firstly, water quality concerns were highest in the North area of the city and secondly, safety concerns were greatest in the Assiniboine area.

The current managers of the resource will use the results to assess urban fishing strategies within Winnipeg. Urban fishing studies are important to managers and planners because with the knowledge of the user base and the resource, the fishery can be enhanced, promoted, and protected.

Graduate and Honors Theses Abstracts