Forests provide a number of ecosystem services which sustain and enrich the wildlife, human societies, and the environment. However, many disturbances threaten forest ecosystems, making it necessary to monitor their health for optimal management and conservation. Although there are many indicators of forest health, changes in biogeochemical cycling, loss of species diversity, and invasive plants are particularly useful due to their vulnerability to the effects of climate change and intensive agricultural land use. Thus, this work evaluates the use of imaging spectroscopy to monitor forest nutrient status, species diversity, and plant invasions in the Mid-Atlantic region. The research is divided into four separate studies, each of which evaluated a unique application for imaging spectroscopy data at a different scale within the forest. The first two studies examined loblolly pine nutrient status at the leaf and canopy scales, respectively. The first study determined that loblolly pine foliar macronutrient concentrations can be successfully modeled across the Southeastern US (R2=0.39-0.74). Following on these results, the second study focused on the relationship between physical characteristics, reflectance, and nutrients. Reflectance values and W scattering coefficients produced successful nitrogen models across loblolly pine plots at the canopy scale. Regression models showed similar explanatory power for nitrogen, although W scattering coefficients were significantly correlated with nitrogen at multiple wavelengths and reflectance variables were not. However, the direction of some of the correlations with W and the unusually high directional area scattering factor values indicate a need for further experimentation. The third study found that several imaging spectroscopy algorithms were moderately successful in identifying wavyleaf basketgrass invasions in mixed deciduous forests (overall accuracy=0.35-0.78; kappa=0.41-0.53). Lastly, the fourth study used a novel imaging spectroscopy/lidar fusion to identify canopy gaps and measure species diversity of understory vegetation. The lidar algorithm identified 29 of 34 canopy gaps, and regression models explained 49 percent of the variance in gap species diversity. In conclusion, imaging spectroscopy can be used to evaluate ecosystem health through forest nutrient status, nitrogen models, species diversity estimates, and identification of invasive plant species. / Ph. D.

As invasive species are so prominent, the influence of wildfire intensity on fuel loading, invasives, species richness, diversity, and evenness were studied at Shenandoah National Park. Most National Parks identify invasive species as the biggest threat to their goal of maintaining native ecosystems. Eight study sites were stratified into three burn classes (high intensity, low intensity, and control), and three transects were randomly located so that nested plots and fuel transects were measured at a distance of 50 ft (15 m), 150 ft (45 m), and 250 ft (75 m) from a road or trail. Field sampling was conducted between May 15, 2004 and June 30, 2004. A subsample of these plots were used to determine specific gravity and quadratic mean diameter for each size class of fuel and to determine the bulk density of the duff and litter layers. High intensity wildfires initially reduced species diversity and evenness in the tree and herbaceous strata, but after 14 years tree species diversity and evenness returned to levels found in unburned areas, while herbaceous strata diversity was not associated with time since burn. Low intensity wildfires resulted in the greatest impacts in the shrub stratum. Presence of invasive species was associated with more even and diverse vegetation in all strata, perhaps because invasive species were relatively sparse. Fuel loadings were reduced initially by high intensity wildfires, but quickly returned to the same level as unburned areas. Although these initial findings indicate that invasive species will not persist after wildfire disturbance, continual monitoring by National Parks would be prudent. / Master of Science

Eight study sites were established in mid-elevation, south aspect, mixed-oak forests in the Ridge and Valley and Allegheny Mountain physiographic provinces of Southwestern Virginia and West Virginia to address questions concerning the variability in species composition, richness, and relative abundance of vascular plant species in those communities. All forest strata were sampled using a nested plot design. Variability in species richness and species composition was found to be high. Total species richness values ranged from 84 to 273, and Sorrenson's Coefficient of Similarity index values indicated that approximately 46, 38, and 51 percent of the species in the overstory, mid-story, and herb stratum were the same among sites, respectively. However, despite differences in composition and richness, K-S tests revealed significant differences in the distribution of ranked relative abundance only in the mid-story at two sites. Differences did occur in the relative abundance of twelve growth form categories. While tree seedlings and perennial herbs dominated, on average, woody vines and fern species represented substantial coverage on sites in the Allegheny Mountains. Correlations among forest strata were weak. The greatest amount of variation in species richness was attributiable to the standard deviation of a forest site quality index (FSQI), which was thought to represent the variation in microtopography across each site. The lack of correlation and high variability in plant species richness and composition, despite similarities in topographic characteristics, reinforce the inherent weaknesses involved with using the chronosequence approach to studying ecological responses in the Southern Appalachian mixed-oak region. Future remeasurement and long term monitoring of these study sites, following the implementation of silvicultural manipulations, will provide the information needed to make inference on the effects of forest management practices on Southern Appalachian mixed-oak forests. / Master of Science

Bird populations are changing at unprecedented rates in response to human-induced changes to the global environment, and these rates of change are expected to accelerate over the coming decades. Changes in the levels of sulphur dioxide (SO2) in the atmosphere through emissions from power stations pose a potential threat to bird populations. However, avian response to SO2 pollution is poorly understood. Exploring the relationship between avian diversity and SO2 exposure levels will help in determining species sensitive to air pollution. This study seeks to understand the interactions between avian diversity and SO2 concentration levels around Matimba power station so as to have more insight on the level of avian vulnerability to air pollution. Matimba is an important site in South Africa as a second coal fired power station, Medupi, is currently being constructed with additional stations also a possibility. This study represents an important baseline assessment of the avian population status before the additional pollution burden is realised from Medupi. Ten min repeated point counts were conducted at three sample sites with varying distances from Matimba and Medupi power stations. These counts were used to calculate bird species density and diversity. Cloud-free Landsat 8 imagery acquired on 7 January, 2014 was used to derive habitat structure and productivity variables. Elevation variables were derived using a DEM (Digital Elevation Model) obtained from NASA Global Data Explorer. The AERMOD dispersion model was used to characterise spatio-temporal variations in ambient SO2 concentrations around Matimba power station. Multiple regression analysis was then used to ascertain which of these variables (SO2, habitat structure, productivity and terrain) contribute most to the observed variation in bird species density and diversity around Matimba and Medupi power stations. SO2 polluted air did not have an influence on bird species density and diversity at the community level. At species level two species (Batis molitor and Streptopelia senegalensis) exhibited some measure of negative response to SO2 air pollution. However, after further investigation using multiple regression analysis it was revealed that habitat structure had more influence on the density of these two species compared with ambient SO2 concentrations. Bird species density and diversity varied significantly among the sample sites but were not related to the distance to the source of the SO2 air pollution. Evidence obtained from this study revealed that continuous monitoring of the interactions between SO2 polluted air and bird populations is recommended for a more comprehensive understanding of avian susceptibility towards SO2 air pollution and this will also facilitate in the selection of sensitive and relevant species for future ecology studies at other coal-fired power stations. Furthermore, it is expected that SO2 concentrations will significantly increase with the commissioning of Medupi power station thus further necessitating the need for continuous monitoring of bird species densities around Matimba and Medupi power stations. / MSc (Geography and Environmental Management), North-West University, Potchefstroom Campus, 2015

Bird populations are changing at unprecedented rates in response to human-induced changes to the global environment, and these rates of change are expected to accelerate over the coming decades. Changes in the levels of sulphur dioxide (SO2) in the atmosphere through emissions from power stations pose a potential threat to bird populations. However, avian response to SO2 pollution is poorly understood. Exploring the relationship between avian diversity and SO2 exposure levels will help in determining species sensitive to air pollution. This study seeks to understand the interactions between avian diversity and SO2 concentration levels around Matimba power station so as to have more insight on the level of avian vulnerability to air pollution. Matimba is an important site in South Africa as a second coal fired power station, Medupi, is currently being constructed with additional stations also a possibility. This study represents an important baseline assessment of the avian population status before the additional pollution burden is realised from Medupi. Ten min repeated point counts were conducted at three sample sites with varying distances from Matimba and Medupi power stations. These counts were used to calculate bird species density and diversity. Cloud-free Landsat 8 imagery acquired on 7 January, 2014 was used to derive habitat structure and productivity variables. Elevation variables were derived using a DEM (Digital Elevation Model) obtained from NASA Global Data Explorer. The AERMOD dispersion model was used to characterise spatio-temporal variations in ambient SO2 concentrations around Matimba power station. Multiple regression analysis was then used to ascertain which of these variables (SO2, habitat structure, productivity and terrain) contribute most to the observed variation in bird species density and diversity around Matimba and Medupi power stations. SO2 polluted air did not have an influence on bird species density and diversity at the community level. At species level two species (Batis molitor and Streptopelia senegalensis) exhibited some measure of negative response to SO2 air pollution. However, after further investigation using multiple regression analysis it was revealed that habitat structure had more influence on the density of these two species compared with ambient SO2 concentrations. Bird species density and diversity varied significantly among the sample sites but were not related to the distance to the source of the SO2 air pollution. Evidence obtained from this study revealed that continuous monitoring of the interactions between SO2 polluted air and bird populations is recommended for a more comprehensive understanding of avian susceptibility towards SO2 air pollution and this will also facilitate in the selection of sensitive and relevant species for future ecology studies at other coal-fired power stations. Furthermore, it is expected that SO2 concentrations will significantly increase with the commissioning of Medupi power station thus further necessitating the need for continuous monitoring of bird species densities around Matimba and Medupi power stations. / MSc (Geography and Environmental Management), North-West University, Potchefstroom Campus, 2015

At present, species are faced with a host of human-induced impacts that have already led to a loss of biodiversity and shifts in species’ ranges. Knowledge of the ecological and evolutionary processes of dynamic species’ ranges, that are either shifting or expanding, will be key to understanding the ability of species to persist and adapt during periods of environmental change. The aim of this thesis is to investigate the effects of range expansion on the genetics and evolution of species, and understand the processes that facilitate the spatial spread of populations. Using individual-based simulation models, this work demonstrates that the unique selective environment of range expansion can have a significant effect on spatial genetics and the evolution of life-history traits. It is found that the survival and spread of mutations, of various fitness effects, is influenced substantially by landscape features encountered by the expanding range. The incorporation of more specific genetic architecture shows that a substantially increased frequency of fitness peak shifts may occur on the edge of an expanding range than would arise within a stationary population. Range expansion is shown to select for increased dispersal and reproduction at the expense of competitive ability on the front of the expanding wave which results in an accelerating spread rate. The survival and spread of a population during range expansion is also affected substantially by the specific movement behaviour of individuals. The exact nature of suitable habitat, in terms of the number of suitable habitat patches and level of fragmentation, will also greatly affect the ability of a population to survive and spread during a period of environmental change. These findings are synthesised within a conceptual framework that is proposed for the evolution of populations during range expansion, where a flattening of the fitness landscape leads to adaptive revolutions.

Throughout human history man has utilized the environment to varying degrees, depending on technology and population. These ¨ecosystem services¨ have suffered sustained degradation over the centuries, resulting in large investments having to be made to prevent and reverse further changes to the environment. Few studies have attempted to quantitatively compare how these changes, occurring long before modern environmental monitoring programs started, affected important ecosystem services such as species diversity, water quality, carbon burial and soil stability. The aims of this study were to i) assess whether human impact on ecosystem services have varied over time in perspective of relative change, and ii) to assess the individual (per capita) contributions. I used multiple sediment proxies from a 6 m C¹⁴-dated core collected from lake Ekoln, South-Central Sweden, to reconstruct environmental changes while tracking the population growth in the city of Uppsala during the last ten centuries. Through the use of pollen and diatom assemblages I reconstructed the changing terrestrial and aquatic diversities over time, while sediment accumulation rates and the X-ray fluorescence spectroscopy of the sediment was used to reconstruct soil stability, carbon burial and water quality, respectively. In the latter case, sediment phosphorus concentrations were used as a proxy for freshwater eutrophication while metals (mercury and lead) were used to infer inputs of toxic pollutants. Finally, I normalized (z) all data to create meta-data. The z-values and reconstructed population for Uppsala made it possible to differentiate 5 unique time periods based on anthropogenic induced change, which were not previously visible in the data, and all of which have been linked to the most likely historical causes, including the Black Death. The results show that the most significant anthropogenic impacts in terms of pollution volume occurred in the 1960s, while the period from 1200-1500 AD saw the most significant environmental change in terms per head of capita, most likely caused by the shift from woodland to open landscape through twiddening, a process of burning forest to create agricultural land, prior to 1500 AD. Moreover, rapid recovery is visible after the implementation of environmental policies from the 1970s onwards. / <p>Full surname: Kilpatrick van Houte</p>

The secondary productivity of communities is inherently influenced by the availability and quality of food resources. Movement of organic matter (OM) across landscapes can connect adjacent systems by providing subsidies of carbon and nutrients, implying that alterations of environments from their natural state may affect the productivity of neighboring food webs. The intact terrestrial and marine environments of Fiordland provide a setting to study linkages between the land and the sea. The first general objective of this study was to determine if large but nutritionally poor (nitrogen-poor, carbon-rich) inputs of forest litter support marine secondary production, and to identify pathways for incorporation of this material into upper trophic levels. Pools of marine and terrestrial OM had distinct values of [delta]�⁵N, [delta]��C and [delta]�⁴S, providing high power to estimate the relative use of these sources by the food webs of the fjord-head deltas. Deposit feeding invertebrates (e.g. Echinocardium cordatum, Pectinaria australis) directly assimilated plant detritus. Heterotrophic bacteria on the surface of the sediment assimilated forest litter and provided a potential food source for invertebrates. Chemoautotrophic bacteria fix CO₂ that originates from decomposing forest litter, thus providing an indirect pathway for incorporation of forest litter into the food webs. In the deep basins the strength of the flux of uptake by chemoautotrophic bacteria through the benthic food web into the upper trophic levels was demonstrated by hagfish (Eptatretus cirrhatus) obtaining 38-51% of their nutrition from these bacteria. The ability of a community to utilize discrete sources of OM relies on the presence of specific functional feeding guilds. Marine algae provide a highly nutritive (nitrogen-rich) food source for the fjord communities and fluxes of algae into food webs are facilitated by grazing invertebrates and filter feeders. The second general objective of this study was to determine how the effective loss of filter feeders from inner Doubtful Sound would alter the flux of marine-derived OM to the food webs of the delta communities. The low salinity environment imposed by the hydroelectric power station in Doubtful Sound caused a large reduction in the abundance of the infaunal bivalves Austrovenus stutchburyi and Paphies australis from delta habitats. Clams could tolerate periods of freshwater exposure of [less than or equal to]20 days duration, but the constant freshwater conditions in Doubtful Sound decreased survivorship. In 2004/05 the biomass of these species in inner Doubtful Sound (7.28 tonnes) was 29 times smaller than in Bradshaw Sound (214.12 tonnes). The associated loss of biodeposits (~91 tonnes(DW) yr⁻� in Bradshaw Sound vs. 1 tonne(DW) yr⁻� in inner Doubtful Sound) may have also altered the flux of nutritive OM to the infaunal community. The river delta communities in inner Doubtful Sound appear to have a higher reliance on forest litter than those in Bradshaw Sound, which is apparent as low values of [delta]�⁵N and [delta]��C for estuarine fish (Notolabrus celidotus, Hemerocoetes monopterygius), which act as integrators of the benthic community. This study demonstrates important linkages between terrestrial and coastal marine ecosystems and highlights the role of functional diversity in facilitating fluxes of organic material through food webs.

The study aimed to gain understanding about diversity, patterns and processes in natural temperate grassland sites in the ACT. Thirty natural grassland sites were surveyed for plant composition, soil invertebrate abundance and site attributes. Floristic associations were identified using multivariate analysis. These were compared to the other site attributes to determine related characteristics. Total plant species richness was 191 species, of which 41% were exotics. Forbs comprised 56% of the total species, with 69 native forb species. There were 10 native grass species and 12 sub-shrubs recorded. The number of species found within ten 1 m2 quadrats measured in each site ranged from 23 to 56 species. Three floristic associations were identified through classification analysis using frequency data. The Danthonia association occurred in well drained sites on soils with high clay content and low nutrient levels and a greater percentage of bare ground. The sites were assessed as having been subject to moderate to high disturbance in the past and 46% were grazed and 46% mown. Species richness of both natives and exotics was high (quadrat richness averaged 37 species, with 32.8% exotic cover in spring). The Wet Themeda association occurred in poorly drained sites and had also been subject to moderate to high disturbance; 67% of sites were grazed and 25% mown. Phosphorus levels were higher in these sites and the acidity of the soil was also greater. Quadrat richness averaged 39 species, with 35.5% exotic cover in spring. The Dry Themeda association was well drained. These sites had a lower native and exotic diversity, with a mean quadrat richness of 30 species, and 11.1% exotic species cover in spring, but a high species richness in the sites (mean 65.4 species) and a high litter cover of 39.9%. All have been subject to low disturbance and none were grazed. Phosphorus levels were low (7.9 ppm). Wet Themeda grassland sites were more similar to Danthonia grassland sites in terms of species occurrence and frequency than Dry Themeda grasslands, despite both associations being dominated by Themeda australis. While trends emerged regarding management and disturbance levels in the three associations, these were not major differentiating attributes. Invertebrates were collected from soil samples from the sites and identified to order level. Twenty two orders were found, but only three orders, Acarina, Collembola and Coleoptera, were sufficiently abundant to enable statistical tests to be undertaken. Invertebrate order richness and abundance showed strong relationships with vegetation attributes that measured or reflected vegetation structure, with higher abundance in sites dominated by T. australis and where wet soil colour was darker. Both order richness and total abundance of invertebrates were highest in mown sites and lowest in grazed sites. The invertebrate order richness and abundance did not have significant relationships with the floristic associations. A field experiment was conducted in which above- and below-ground gaps were created using the herbicide, Glyphosate, to remove interspecific competition and comparing responses between the Dry Themeda and Danthonia associations. Litter load and soil disturbance were also manipulated. The subsequent establishment of species was studied over two growing seasons, a period of 18 months. Both native and exotic forb richness and cover increased as a result of gap formation. Exotic grass cover and richness showed a strong increase in the first season, but after 18 months had decreased to levels similar to those prior to the application of treatments. Native grasses had not recovered to pre-treatment levels after 18 months. Native forb richness and cover were not inhibited by litter retention, but exotic forb richness and cover were higher in treatments with litter removal. Soil disturbance to a depth of 20 mm had no significant effect on the recruitment of the species. Recommendations for protection and management of the sites were developed from the study. Conservation recommendations are based on the aim of maintaining or enhancing native species diversity and habitat, while protecting threatened species and their habitat. Management plans should recognise requirements based on the floristic associations, diversity of native species, drainage conditions and retention of invertebrate habitat, including structural requirements. Grazing should be retained as a viable management regime, particularly in sites where small species occur. Future research should aim towards a greater understanding of the effects of management on species diversity, including invertebrates, and include a monitoring component with which to determine if management actions should be modified.

National Forest management in the Pacific Northwest is shifting from a focus on commodity production to ecosystem management, in which the health of the entire forest ecosystem is considered, rather than that of a few key species. Ecosystem management includes retention of some live trees following timber harvest (green-tree retention) to preserve biodiversity, imitating the natural fire regime of large, but patchy fires that leave many live trees. How ecosystem management will affect growth and species composition of future forests is an important question. This study takes a retrospective approach to this question by using past disturbance as an analogue to green-tree retention following timber harvest. Using USDA Forest Service timber inventory plot data from the Cascade Mountains of Oregon and SW Washington, 132 unmanaged stands were identified with a tree cohort of 70-110 years old (regeneration) or a tree cohort of 70-110 years old with an overstory of large trees 200+ years old (remnants). All stands were in the Tsuga heterophylla (Raf.) Sarg. zone (Franklin and Dyrness 1973). Single-aged stands represented clearcuts, while two-aged stands served as analogues to stands harvested with green-tree retention. Regeneration basal area/hectare (ba/ha) declined when remnant-tree densities exceeded about 15 remnant trees/ha (R��=0.51) in a relationship roughly described by a sigmoidal curve. Conceptually removing remnant-tree space occupancy effects decreased remnant-tree density's value as a predictor of regeneration ba/ha by about 50% at management-level remnant densities (���45 remnant trees/ha). Thus, it appears that remnant "effects" were a result of both remnant-tree space occupancy and remnant resource use. Douglas-fir ba/ha in the regeneration also declined when remnant-tree densities exceeded about 15 remnant trees/ha (R��=0.60). Western hemlock ba/ha in the regeneration increased slightly with increasing remnant-tree densities (R��=0.19). Western redcedar ba/ha in the regeneration was apparently not related to remnant-tree density (R��=0.02). The degree of aggregation in remnant trees did not appear to affect regeneration ba/ha, but few stands contained the isolated clumps of remnant trees likely under a management scenario. Neither measured site characteristics nor regeneration density was related to regeneration ba/ha across species. Remnant-tree density was apparently unrelated to tree-species diversity in the regeneration. Total-stand ba/ha remained relatively constant across remnant densities. / Graduation date: 1994