No description available.

This thesis examines vegetation community change at two temporal scales in the Piedmont of North Carolina. Using long-term plots in the Duke Forest, I examine decadal-scale changes in community composition of the forest understory and shed light on the potential drivers of that change. Using historical data from colonial survey records, I study presettlement forest communities of the Piedmont and attempt to reconstruct Piedmont forests as they may have been in the time before European arrival. The pattern of successional change in southeastern United States Piedmont forests has been assumed from chronosequence studies over the last half century. However, these assumptions for forest understory herb-layer populations and communities have not been tested using long term data sets. Using permanently marked plots in the Duke Forest (Durham, NC, USA) re-censused after a 23 year time step, species richness and community changes at 25m2 and 1000m2 scales are examined. I look at changes across life forms and examine these changes in relation to measured stand and environmental factors. Although total species richness stayed relatively constant through the 23 year step, herb richness declined with a concomitant increase in woody richness. Plot composition change was remarkably consistent and this change was not correlated to any measured stand or environmental factors. These community-level changes are consistent with previously reported changes in the understories of hardwood dominated stands in the Duke Forest, suggesting that landscape scale drivers may be more important than within-stand successional processes in patterning herbaceous communities at this time. Combined with growing evidence from other studies, this work indicates that forests in the temperate region may be experiencing changes different from those predicted by successional chronosequence studies. It indicates that one of the primary drivers of this change is the explosive growth of deer populations in the last two decades. Witness trees recorded in historical surveys have been used to reconstruct presettlement vegetation in many parts of North America, leading to a better understanding of vegetation patterns before the effects of Europeans. For some parts of North America, Government Land Office records make the process of reconstructing vegetation patterns easier - thus more is known about these areas. Because of the unique and unplanned nature of settlement in the southeastern U.S., less is known about the presettlement vegetation in this area of the country. Using a reconstructed cadastral map of a section of the North Carolina Piedmont, I was able to plot the positions of trees on the historical landscape. These data were then used to understand and reconstruct the composition of presettlement forests. Although the vegetation of some areas of the Piedmont is similar to what was expected, I find significant differences with the expected presettlement composition. In particular, pine species were common in some areas and rare in others, indicating that different disturbance regimes were active on the landscape. / Dissertation

A significant question currently facing environmental managers is how to accurately and efficiently quantify forest diversity and resources. Numerous studies have demonstrated the use of modern spatial analytical tools , such as geographical information systems (GIS), remote sensing devices, and statistical models for predicting the distribution of dominant vegetation cover types. This study examines the ability of generalized additive models (GAMs) to delineate structural diversity in forested ecosystems (specifically the Uinta Mountain Range in Utah) using GIS tools and satellite spectral data, and analyzes the effect of including different forms of satellite data in model construction (i.e., Landsat thematic mapper (TM), advanced very high resolution radiometer (AVHRR), and the GAP Analysis TM-classified map). Based on the assumption that vegetation composition, as well as structural diversity, is a function of environmental gradients, temperature, precipitation, elevation, aspect, slope, and geology were included as independent environmental variables. Probability surface maps were generated for presence of forest , presence of lodgepole pine, basal area of forest trees, percent cover of shrubs, and density of snags. The maps were validated using an independent set of field data collected from the Evanston Ranger District within the Uinta Mountain Range . In general, the models tended to underpredict at large numbers and overpredict at locations that were sampled as having no forest cover. The models predicting the presence of forest and lodgepole pine were 88% and 80% accurate, respectively, within the Evanston Ranger District and an average of 62% of the predictions of basal area, shrub cover , and snag density fell within an approximate 15% deviation from the field validation values . The addition of TM spectral data and the GAP Analysis TM-classified data were found to contribute significantly to the models' predictions, with some contribution from AVHRR data. The methods used in this study provide a systematic approach for delineating structural features within forest habitats, thus offering an efficient spatial tool for making management decisions.

Temperature isoclines are shifting with global warming. To survive, species with thermal niches must shift their geographical ranges to stay within the bounds of their suitable habitat, or acclimate to a new environment. Mathematical models that study range shifts are called moving-habitat models. The literature is rich and includes modelling with reaction-diffusion equations. Much of this literature represents space by the real line, with a handful studying 2-dimensional domains that are unbounded in at least one direction. The suitable habitat is represented by the set over which the demographics (reaction term) has a positive net growth rate. In some cases, this is a bounded set, in others, it is not. The unsuitable habitat is represented by the set over which the net growth rate is negative. The environmental shift is captured by an imposed shift of the suitable habitat. Individuals respond to their environment via their movement behaviour and many display habitat-dependent dispersal rates and a habitat bias. Such behaviour corresponds to a jump in density across the interface of suitable and unsuitable habitat. The questions motivating moving-habitat models are: when can a species track its shifting habitat and what is the impact of an environmental shift on a persisting species. Without closed form solutions, researchers rely on numerical methods to study the latter, and depending on the movement of the interface, the former may require numerical tools as well. We construct and analyse two numerical methods, a finite difference (FD) scheme and a finite element (FE) method in 1- and 2-dimensional space, respectively. The FD scheme can capture arbitrary movement of the boundary, and the FE method rather general shapes for the suitable habitat. The difficulty arises in capturing the jump across a shifting interface. We construct a reference frame in which the interfaces are fixed in time. We capture the jump in density with a clever placing of the nodes in the FD scheme, and through a Lagrange multiplier in the FE method. With biological applications, we demonstrate the power of our solvers in advancing research for moving-habitat models.

Coal mining has had profound impacts in the Appalachian region, initiating a need to understand the implications of traditional and current reclamation practices on wildlife. I evaluated wildlife use of reclaimed sites of varying ages and cover types in southwestern Virginia. I compared reclaimed sites to another form of anthropogenic disturbance (clearcut) and relatively undisturbed mature forest. Birds were surveyed during early mornings throughout the breeding season in 2007 and 2008 using the point count method. Amphibians were surveyed using artificial cover, constrained-time night searches, and auditory pond surveys. Microhabitat data were collected at each sampling point and were combined with landscape-level GIS information to relate habitat characteristics and wildlife patterns. I observed 80 bird species using reclaimed areas, clearcuts, and mature forest. Pre-regulation sites (prior to the Surface Mining Control and Reclamation Act of 1977) supported the highest number of species overall. Cluster analysis identified 4 bird associations based on habitat characteristics. I developed site-specific, landscape-level, and mixed-scale logistic regression models to identify habitat characteristics that best predicted the presence of 27 species. For 18 species, mixed-scale models performed best, suggesting the importance of a multi-scale approach to habitat analysis. Salamanders were generally not detected on reclaimed areas, possibly due to the lack of soil moisture, leaf litter, and woody debris on young sites. Frogs were present in all water bodies surveyed, suggesting the importance of managing ponds and wetlands on reclaimed sites. Identifying and focusing on important habitat characteristics will help managers enhance post-mining land for wildlife. / Master of Science

This research aims to develop a process for eliciting expert knowledge and incorporating this knowledge as prior distributions for a Bayesian logistic regression model. This work was motivated by the need for less data reliant methods of modelling species habitat distributions. A comprehensive review of the research from both cognitive psychology and the statistical literature provided specific recommendations for the creation of an elicitation scheme. These were incorporated into the design of a Bayesian logistic regression model and accompanying elicitation scheme. This model and scheme were then implemented as interactive, graphical software called ELICITOR created within the BlackBox Component Pascal environment. This software was specifically written to be compatible with existing Bayesian analysis software, winBUGS as an odd-on component. The model, elicitation scheme and software were evaluated through five case studies of various fauna and flora species. For two of these there were sufficient data for a comparison of expert and data-driven models. The case studies confirmed that expert knowledge can be quantified and formally incorporated into a logistic regression model. Finally, they provide a basis for a thorough discussion of the model, scheme and software extensions and lead to recommendations for elicitation research.

The greater sage-grouse (Centrocercus urophasianus; sage-grouse) is a species of conservation concern in Utah and range-wide due to declines in populations and threats to sagebrush habitat on which they depend. To effectively conserve the species, detailed site-specific knowledge of ecology and distribution is needed. To expand knowledge of local populations within the West Box Elder Sage Grouse Management Area (SGMA) and gain insights into the effectiveness of vegetation treatments intended to benefit sagegrouse, I radio marked and tracked 123 (68 female, 55 male) sage-grouse and conducted sage-grouse pellet surveys on 19 conifer removal projects. Widespread habitat restoration measures designed to benefit sage-grouse have highlighted the need for prioritization tools to optimize placement of sage-grouse habitat projects. I generated seasonal habitat models to predict sage-grouse habitat use within the West Box Elder SGMA using a suite of vegetation and topographical predictors and known sage-grouse locations. Model fit was good with brood, early summer, late summer, lekking (early spring), and non-breeding models reporting an AUC of >0.90; nest and winter models reported an AUC of 0.87 and 0.85, respectively. A vegetation disturbance history was built for the study area from 1985 to 2013; however, the vegetation disturbances mapped were not a strong predictor of sage-grouse seasonal habitat-use. To evaluate effectiveness of conifer reduction treatments I used fecal pellet and in concert with radio-telemetry data. Increased sage-grouse use of conifer treatments was positively associated with sage-grouse presence in adjacent habitats (P = 0.018), percent shrub cover (P = 0.039), and mesic environments within 1000 m of treatments (P = 0.048). Sage-grouse use of conifer treatments was negatively associated with conifer canopy cover (P = 0.048) within 1000 m of treatments. To investigate sample bias related to individual bird behavior or capture trauma I monitored 204 radio-marked sage-grouse within the West Box Elder and Rich-Morgan- Summit SGMAs in Utah between January 2012 and March 2013. Sage-grouse that flushed one or more times prior to capture had higher brood (P = 0.014) and annual survival (P = 0.027) than those that did not. Sage-grouse that experienced more capture trauma had decreased annual survival probabilities (P = 0.04).

Les prédateurs marins supérieurs, ici les cétacés et les oiseaux marins, doivent développer des stratégies optimales d’utilisation des ressources et des habitats. Notre objectif était d’explorer leurs habitats en fonction de leurs coûts de vie. Nous avons postulé que les prédateurs coûteux étaient contraints d’occuper les habitats de meilleure qualité alors que les prédateurs plus économes pouvaient occuper les habitats de qualité moindre. Nous nous sommes basés sur des guildes de cétacés et oiseaux définies selon leurs coûts de vie et les observations de survols aériens dans trois régions tropicales (l’Atlantique Ouest tropical, le Sud Ouest de l’Océan Indien et la Polynésie française). Nous avons construit des modèles additifs généralisés à partir de variables physiographiques (ex : profondeur), océanographiques (ex : activité tourbillonnaire) et biologiques (ex : chlorophylle et micronecton) pour décrire la qualité des habitats. Nous avons d’abord modélisé les habitats des cétacés et oiseaux à l’échelle régionale. Les cétacés coûteux occupaient les habitats de meilleure qualité alors que les cétacés plus économes occupaient aussi les habitats de qualité moindre. La distribution des oiseaux reflétait principalement celle des colonies et leur dépendance à la qualité de l’habitat semblait moins claire. Nous avons ensuite mis en évidence des propriétés génériques de distribution des cétacés et fourni des prédictions circumtropicales. Cette thèse a donné un nouvel aperçu des stratégies d’utilisation des habitats des prédateurs supérieurs à la lumière de leurs coûts de vie. Ces prédictions spatiales ont des implications majeures pour la gestion de ces espèces et de leurs écosystèmes. / Marine top predators, here cetaceans and seabirds, must develop optimal strategies of resource and habitat utilization. The main goal of this dissertation was to investigate cetacean and seabird strategies of habitat utilization in relation to their energetic costs of living. We hypothesized that predators with high costs of living should be constrained to high quality habitats, whereas less active predators could cope with habitats of lesser quality. We studied the habitats of cetacean and seabird guilds defined according to their likely costs of living. We relied on sightings collected from aerial surveys in three tropical regions (the western tropical Atlantic, the Southwest Indian Ocean and French Polynesia). We built generalized additive models based on a range of physiographic (e.g. depth), oceanographic (e.g. mesoscale activity) and biological variables (e.g. chlorophyll concentration and micronekton) to describe the quality of pelagic habitats. We first modeled cetacean and seabird habitats at the regional scale. Energetically costly cetaceans appeared to be constrained to the highest quality habitats, whereas less active cetaceans exploited habitats of lesser quality. Seabird distributions primarily reflected colony locations and their dependences on habitat quality were less clear. We then highlighted generic properties of cetacean distributions and provided predictions at the circumtropical scale. This dissertation gave new insights on top predator strategies of habitats utilization in light of their costs of living. These spatial predictions have significant implications for the management of these species and of their pelagic ecosystems.

Butterfly weed (Asclepias tuberosa) is a primary larval food source for the monarch butterfly (Danaus plexippus). Planting more butterfly weed may stimulate declining monarch populations. To that end, a habitat suitability map was created for Mississippi in GIS using soil pH, soil texture, and land cover. Herbarium data were derived from the Southeast Regional Network of Expertise and Collections (SERNEC) database. Environmental data were from the USDA National Resource Conservation Service geospatial data gateway. Frequency analysis was used to assign scores to environmental variables of SERNEC occurrences using a suitability index. Global positioning systems (GPS) locations of butterfly weed were collected to validate the model. The most suitable model with 78.9% of GPS points in medium to high suitability was a weighted sum overlay with land cover 50%, soil pH 25%, and soil texture 25%. The suitability map may enable conservationists to identify suitable sites for butterfly weed in Mississippi.

In the past century, hydropower development in Sweden has been facilitated by the lack of appropriate environmental legislation. This exploitation has resulted in the current poor ecological status of most Swedish water bodies affected by hydropower, and in the need for new policies that reflect the 2000 European Water Framework Directive. Thus, the Swedish Agency for Marine and Water management (SwAM), the Swedish Energy Agency (SEA) and Svenska Kraftnät have elaborated a new national plan for the revision of almost all hydropower licenses in the next 20 years, and Vattenkraftens Miljöfond has carried out a pilot project on Ljungan river, to test different working methods for regional collaboration, and to find effective solutions for achieving the greatest possible benefit for the aquatic environment, while ensuring an efficient national hydropower supply. This thesis focuses on the fish habitat rehabilitation of a portion of Ljungan river highly affected by hydropower practices. By combining a 1D/2D hydraulic model, calibrated and validated, with a habitat model that defines the hydraulic preferences of the target fish species, it was possible to run different simulations and to quantify the optimal habitat obtainable in each scenario, following the PHABSIM methodology. The outputs of this analysis were produced in the form of Composite Suitability Index (CSI) maps and Weighted Usable Area (WUA)­discharge curves. The former resulted to be a very useful tool for analyzing the spatial distribution of suitable habitats for certain ecological processes, and for planning in­stream habitat improvement measures, while the latter have been efficiently used to identify and quantify the maximum habitat improvement achievable with the different mitigation scenarios. This approach helped to firstly assess the habitat improvement that would result from removing the weirs located along the river, leading to the conclusion that the removal would be effective only if combined with other morphological changes of the river channel. Additionally, the coupled eco­hydraulic model has been useful for the design of ecological flow scenarios able to achieve different levels of habitat improvement, including scenarios aimed at minimizing energy production losses. The use of the PHABSIM methodology resulted to be very suitable for the scale and scope of the project. However, it also required to limit the iii ecological drivers considered in the analysis and to make some simplistic assumptions about target species behavior, which must always be underlined and acknowledged when using the results in decision­making processes. / Under det senaste århundradet har vattenkraftsutbyggnaden i Sverige underlättats av bristen på lämplig miljölagstiftning. Detta utnyttjande har lett till att de flesta svenska vattenförekomster har en dålig ekologisk status och att det behövs en ny politik som återspeglar det europeiska ramdirektivet för vatten från 2000. Således har Havs­ och vattenmyndigheten, Energimyndigheten och Svenska Kraftnät utarbetat en ny nationell plan för översyn av nästan alla vattenkraftslicenser under de kommande 20 åren, och Vattenkraftens Miljöfond har genomfört ett pilotprojekt i Ljungan älv för att testa olika arbetsmetoder för regionalt samarbete och hitta effektiva lösningar för att uppnå största möjliga nytta för vattenmiljön samtidigt som man säkerställer en effektiv nationell vattenkraftförsörjning. Denna avhandling fokuserar på återställandet av fiskhabitatet i en del av Ljungan älven som är starkt påverkad av vattenkraftsutövning. Genom att kombinera en 1D/2D hydraulisk modell, kalibrerad och validerad, med en habitatmodell som definierar målfiskarternas hydrauliska preferenser var det möjligt att göra olika simuleringar och kvantifiera den optimala livsmiljö som kan erhållas i varje scenario, enligt PHABSIM­metodiken. Detta tillvägagångssätt har använts för att först och främst bedöma den förbättring av livsmiljön som skulle bli resultatet av att ta bort de dammar som finns längs älvsträckan, vilket ledde till slutsatsen att avlägsnandet skulle vara effektivt endast om det kombineras med andra morfologiska förändringar av flodfåran. Dessutom har den kopplade ekohydrauliska modellen varit användbar för att utform ekologiska flödesscenarier som kan uppnå olika nivåer av habitatförbättring, inklusive scenarier som syftar till att minimera förlusterna i energiproduktionen.