Thesis (M.S.)--Oregon State University, 2001. / Typescript (photocopy). Includes bibliographical references. Also available on the World Wide Web.

In this study four different aspects of primary successions on glacier forelands in South-Central Norway are investigated. First, general vegetation and environment trends, which have been observed by other scientists, were studied on two forelands. It is shown that the two selected forelands (one low- and one high-altitude foreland, Nigardsbreen and Storbreen, respectively) differ in the speed of vegetation cover change and in the overall species richness developed over the last 250 years since deglaciation. This is attributed to differences in the severity of the environment between the two forelands. Some support for an early peak in the number of species was found for the Nigardsbreen foreland, whereas the Storbreen foreland shows no significant trend in the development of species richness. Life-forms and soil-depth change on both forelands in the predicted manner. A novel approach for the use of Ellenberg indicator values for the estimation of pH-values from the existent vegetation shows highly significant correlations between predicted pH values and trends measured for similar terrain ages from a study by Messer (1988). Investigation at two different scales indicated that small-scale disturbances (< 100 m) may influence the mosaic of the environment, which affects the results of the chronosequence approach. By using a strict geometrical sampling design more variable conditions entered the analyses than with the stratified random sampling schemes adopted by others, such as Matthews (1979b). Some loss in the power of analysing the data occurred, but the results are still consistent with hypotheses about primary succession on glacier forelands. Second, new aspects of the temporal development and successional change of plant communities were investigated on the same two forelands. There is some evidence of succession from a pioneer community to the local climax community of the surrounding area for both forelands. There is also support for the correlation of timedependent environmental variables with the successional development of plant communities. A new application of Ellenberg indicator values in ordinations was assessed for its value in estimating average environmental conditions derived from the vegetation. As with the previous analyses, investigation of community development at two different scales supported the importance of sampling scale on the analysis of succession data. Third, spatially explicit analyses of distribution patterns of dominant woody species was developed as a means to investigate processes and formulate hypotheses about primary succession on glacier forelands. The data for this part of the study was collected on five glacier forelands, three low-altitude (Nigardsbreen, Austerdalsbreen and Bødalsbreen) and two high-altitude (Storbreen and Høgvaglbreen). Pattern descriptors including perimeter/area index, autocorrelation measures and fractal dimensions, support the hypothesis of differential colonisation of wind and animaldispersed species. Even more convincing is the result that low- and high-altitude forelands show significant differences in the patterns exhibited by wind and animals dispersed species. This result is thought to be related to the differences in environmental severity affecting the establishment of colonising species. Finally, logit modelling of the distributions of dominant woody species provides insights into the sampling effort needed to gather enough data for meaningful analyses. It also leads to the identification of important factors influencing the distribution of those species.

Thesis advisor: Tara Pisani Gareau / Salt marshes provide important ecosystem services, including carbon sequestration. Permanently inundated ponds are prominent features in the marsh landscape, encompassing up to 60% of the total marsh area, but they are rarely considered in biogeochemical assessments. I investigated two ponds in Plum Island Estuary, MA to measure and analyze their metabolism. The ponds varied in size and vegetation cover. Oxygen concentrations and pH values were recorded in 15-minute intervals during the entire study period. The ponds regularly become hypoxic or anoxic during night. This is a problem for the estimation of respiration rates which are based on nighttime measurements. To investigate this potential underestimation, several approaches to estimate respiration were used. First, additional measurements of surface water concentrations of dissolved inorganic carbon were made. A comparison of respiration estimates based on oxygen and DIC changes during tidal isolation revealed a reasonable agreement for the most time but not during periods of high productivity during the day or late at night. At this point, oxygen concentrations are so depleted that a change in concentration – the indicator of respiration – is barely detectable. However, DIC based respiration rates indicate that respiration is occurring under these hypoxic/anoxic conditions. This saturation changes during periods of tidal inundation, when a nighttime peak in oxygen concentrations indicates that the flood water is relatively enriched in oxygen compared to the pond water. On three days, it was tested whether under these conditions the oxygen-based respiration rate was higher than under hypoxic conditions (i.e., during tidal isolation). The rates were indeed higher than those under tidal isolation but still not in the range of DIC-based rates. Overall, metabolic rates differed between the two ponds in magnitude, which is likely caused by different vegetation cover, but may be influenced by size, sampling period, and duration as well. / Thesis (BS) — Boston College, 2018. / Submitted to: Boston College. College of Arts and Sciences. / Discipline: Departmental Honors. / Discipline: Earth and Environmental Sciences.

Pang Ying Wai. / Thesis submitted in: August 2003. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2004. / Includes bibliographical references (leaves 134-144). / Abstracts in English and Chinese. / ABSTRACT --- p.i / ABSTRACT (CHINESE) --- p.ii / ACKNOWLEDGEMENTS --- p.iii / TABLE OF CONTENT --- p.iv / LIST OF TABLES --- p.viii / LIST OF FIGURES --- p.x / Chapter CHAPTER ONE --- INTRODUCTION / Chapter 1.1 --- Conceptual Framework --- p.1 / Chapter 1.2 --- Objectives and Significance of the study --- p.3 / Chapter 1.3 --- Study Area --- p.4 / Chapter 1.4 --- Organization of Thesis --- p.9 / Chapter CHAPTER TWO --- LITERATURE REVIEW / Chapter 2.1 --- Landscape Ecology --- p.10 / Chapter 2.1.1 --- Definition of Landscape Ecology --- p.10 / Chapter 2.1.2 --- Dimension of Landscape Ecology --- p.13 / Chapter 2.2 --- Landscape metrics --- p.17 / Chapter 2.3 --- Application of Remote Sensing in Landscape Ecology --- p.25 / Chapter 2.4 --- Wetland Habitats and Situations in Hong Kong --- p.29 / Chapter 2.5 --- Summary --- p.39 / Chapter CHAPTER THREE --- METHODOLOGY / Chapter 3.1 --- Data Description --- p.40 / Chapter 3.1.1 --- Satellite Data --- p.40 / Chapter 3.1.2 --- Topographic Map Sheets and Digital Maps --- p.45 / Chapter 3.2 --- Satellite Data Preprocessing --- p.45 / Chapter 3.3 --- Landuse and Land Cover Classification --- p.47 / Chapter 3.4 --- Landscape Units Extraction --- p.51 / Chapter 3.5 --- Landscape Metrics Identification and Extraction --- p.54 / Chapter 3.6 --- Disturbance Level Identification and Extraction --- p.60 / Chapter 3.7 --- Inter-Classes and Temporal Comparison of Class-based and Landscape-based metrics --- p.61 / Chapter 3.8 --- Summary --- p.62 / Chapter CHAPTER FOUR --- LAND COVER CLASSIFICATION RESULT / Chapter 4.1 --- Introduction --- p.63 / Chapter 4.2 --- Clustering Result --- p.63 / Chapter 4.3 --- Land cover classification results --- p.66 / Chapter 4.4 --- Accuracy Assessment --- p.83 / Chapter 4.5 --- Implication from land covers change --- p.84 / Chapter 4.5.1 --- Mangrove Changes --- p.84 / Chapter 4.5.2 --- Encroachment of Built-up area --- p.89 / Chapter 4.6 --- Summary --- p.91 / Chapter CHAPTER FIVE --- LANDSCAPE METRIC ANALYSIS / Chapter 5.1 --- Introduction --- p.94 / Chapter 5.2 --- Landscape perspective analysis --- p.94 / Chapter 5.2.1 --- "Area, Density and Edge" --- p.95 / Chapter 5.2.2 --- Fractal Dimension and Shape Indices --- p.98 / Chapter 5.2.3 --- "Contagion, Interspersion and Diversity" --- p.99 / Chapter 5.2.4 --- Disturbance information --- p.101 / Chapter 5.3 --- Class perspective analysis --- p.105 / Chapter 5.3.1 --- Mangrove --- p.105 / Chapter 5.3.1.1 --- "Area, Density and Edge" --- p.105 / Chapter 5.3.1.2 --- Fractal Dimension and Shape Indices --- p.107 / Chapter 5.3.1.3 --- Contagion and connectivity --- p.107 / Chapter 5.3.1.4 --- Disturbance information --- p.109 / Chapter 5.3.2 --- Reed bed --- p.111 / Chapter 5.3.2.1 --- "Area, Density and Edge" --- p.111 / Chapter 5.3.2.2 --- Fractal Dimension and Shape Indices --- p.112 / Chapter 5.3.2.3 --- Contagion and connectivity --- p.113 / Chapter 5.3.2.4 --- Disturbance information --- p.114 / Chapter 5.3.3 --- Fishponds --- p.116 / Chapter 5.3.3.1 --- "Area, Density and Edge" --- p.116 / Chapter 5.3.3.2 --- Fractal Dimension and Shape Indices --- p.117 / Chapter 5.3.3.3 --- Contagion and connectivity --- p.118 / Chapter 5.3.3.4 --- Disturbance information --- p.119 / Chapter 5.4 --- Discussion --- p.121 / Chapter 5.4.1 --- Landscape degradation from landscape perspective --- p.121 / Chapter 5.4.2 --- Implication of landscape metrics on land use planning --- p.123 / Chapter 5.4.3 --- Factors affecting the usage of landscape metrics --- p.124 / Chapter 5.5 --- Summary --- p.126 / Chapter CHAPTER SIX --- CONCLUSION / Chapter 6.1 --- Summary of Findings --- p.129 / Chapter 6.1.1 --- Summary of landscape composition --- p.129 / Chapter 6.1.2 --- Summary of landscape configuration --- p.130 / Chapter 6.2 --- Limitations of the Study --- p.132 / Chapter 6.3 --- Recommendations for Further Studies --- p.133 / REFERENCES --- p.134 / APPENDIX --- p.145

The alpine treeline ecotone is primarily caused by climate at continental and regional scales, but at finer scales exhibits increasing complexity and linkages with abiotic factors. In order to better understand the dynamics and geographic properties of treeline, a three-pronged study was conducted. The study's first objective investigated what factors account for local treeline variation by comparing a climate-based predictive surface with actual treeline elevation. After finding that a mean growing season temperature of 10.27°C was the strongest predictor of treeline at continental scales, statistical analyses attempted to model the differences between predicted and actual treeline elevation based on local climatic and geomorphological controls. Site elevation and latitude were found to correlate most strongly. Roughness also exerted positive influences, as well as precipitation when coupled with certain terrain factors. The second objective explored the distribution of treeline carbon at site scales by analyzing field data gathered in August 2009 in Glacier National Park, Montana. This confirmed the expected relationship of upright trees containing the highest biomass density, tundra the lowest, and krummholz in the middle. A high degree of dead organic matter was discovered, which may have important implications for treeline as a carbon source or sink when considering treeline's slow decomposition rates, and how these will change under climate warming. An additional analysis investigated the potential for relative biomass gain with future treeline advance, based on site and regional differences in slope and environmental lapse rates. The third objective asked what the potential for change in treeline biomass is across the western United States is by using freeze-thaw days as a proposed indicator. This potential importance is based on existing knowledge of the facilitating relationships between solifluction, surface geomorphology, and seedling establishment and survival. Freeze-thaw days were found to be highly variable, but correlated most strongly with elevation, suggesting increased temperature variability at higher elevations. The concluding chapter synthesizes the findings and maps potential biomass gain and freeze-thaw days together to highlight treeline sites and regions with the greatest potential for advance in a warming climate.

Thesis (M.S.)--University of Nevada, Reno, 2005. / "(December 2005)." Includes bibliographical references (leaves 36-42). Online version available on the World Wide Web.

Thesis (M.S.)--Brigham Young University. Dept. of Geography, 2005. / Includes bibliographical references (p. 71-78).

Chronic wasting disease (CWD) has been a known threat to Saskatchewans wild cervid populations for more than a decade. As host movements can affect the spread of a disease across the landscape, disease models and management strategies should incorporate information regarding movement patterns of the host population in question. I used radio telemetry to study mule deer (<i>Odocoileus hemionus</i>) captured between 2006 and 2008 in a CWD-endemic region of southern Saskatchewan. Using location data from 152 individuals, I investigated home range size and patterns of direct and indirect contact (measured using proximity and shared space use) in relation to sex, habitat, and landscape structure. <p>Home ranges (95% fixed kernel) of GPS-collared deer in this study averaged 21.4 km² (n = 94). Male home ranges (mean = 29.5 km², n = 56) were larger than those of females (mean = 16.1 km², n = 38), which could have implications for CWD prevalence differences between sexes. Of the landscape variables tested, topographic ruggedness was inversely related to home range size and Shannons diversity (a measure of both habitat richness and evenness) was positively related to home range size. <p>Potential direct contact events were identified when two deer were located within 25 m of each other at the same point in time. These events occurred more often between February and April, agreeing with the tendency of mule deer to aggregate into large groups during the late winter months, and suggesting that this may be an important time period for disease transmission. Contact also occurred more than expected in cropland, whereas areas of shared use occurred more than expected in grassland, shrub/wood habitat, and rugged terrain. Smaller home ranges and greater degree of shared space use within areas of rough topography may lead to greater risk of environmental contamination with the infectious CWD agent in these areas. In contrast, the relationship between cropland and probability of direct contact may imply greater risk of direct CWD transmission between deer occupying this habitat. <p>These results identify connections between particular landscape factors and risk of CWD transmission and will be used, in combination with results of related studies, to develop a model of CWD spread in Saskatchewan. This will in turn aid management agencies in developing methods to more effectively manage the disease and control its movement outside of affected regions.

Human-dominated landscapes often have habitat loss and fragmentation. These characteristics described at the landscape scale, called landscape elements, influence species diversity and distribution. These landscape elements include such descriptions as the amount of habitat in the landscape and the degree of fragmentation of the habitat. "Optimization of landscape pattern" studies which landscape elements will maximize species diversity and/or distribution. Some general conclusions have emerged from this research. For example, for some bird species the size of the habitat patch in which a species nests has been found to be more important than landscape variables. However, preliminary research suggested that landscape elements such as the matrix are important urban areas. My study addressed this problem by asking: which elements of the landscape are most important for predicting avian species richness and abundance in the Greater Toronto Area? A literature review revealed a number of variables that have been found to influence bird species diversity within a landscape: area of habitat in which the species nests, amount of habitat within the landscape, degree of fragmentation, vegetation characteristics of the habitat patch, and area within the landscape deemed urban in municipal land-use designations (amount of urbanization). From this literature, I formulated four hypotheses describing the most important variables for avian diversity: (1) the area of the habitat patch is most important, (2) only variables describing the habitat patch itself are important, (3) the area of the habitat patch is important, but landscape variables should also be considered and (4) urbanization is most important. These hypotheses were considered competing explanations of bird species diversity at the landscape scale. <br /><br /> A database of breeding bird data and landscape information, in a geographic information system platform, was used to investigate the comparative strength of the competing hypotheses for the Greater Toronto Area. A mathematical expression with a Poisson model format was created to represent each hypothesis. The model selection technique based on Kullback-Leibler information using the Akaike Information Criterion was deemed most appropriate for the comparison of the models. Four separate Poisson model competitions were completed using two habitat types and two response variables: species richness and total abundance. In three of the four competitions, the best model included the habitat area and the amount of urbanization in the landscape. In the forth competition, this model was considered as strong as another model which included habitat area, amount of habitat in the landscape and degree of fragmentation. The results from the model competition support the hypothesis that habitat area is important, but landscape variables must also be considered to explain avian richness and total abundance. It appears that maintaining native bird biodiversity in the Greater Toronto Area should focus on preserving and possibly increasing habitat area and decreasing adjacent urbanization. Exploration of the best model in the forest analysis with the richness response variable found that a 10% increase in habitat area cause approximately a 10% increase in species richness, and a 10% increase in urban area caused approximately a 20% decrease in species richness. Consequently, current natural heritage planning in Ontario should consider urban development as an important negative effect on native birds.

Human-dominated landscapes often have habitat loss and fragmentation. These characteristics described at the landscape scale, called landscape elements, influence species diversity and distribution. These landscape elements include such descriptions as the amount of habitat in the landscape and the degree of fragmentation of the habitat. "Optimization of landscape pattern" studies which landscape elements will maximize species diversity and/or distribution. Some general conclusions have emerged from this research. For example, for some bird species the size of the habitat patch in which a species nests has been found to be more important than landscape variables. However, preliminary research suggested that landscape elements such as the matrix are important urban areas. My study addressed this problem by asking: which elements of the landscape are most important for predicting avian species richness and abundance in the Greater Toronto Area? A literature review revealed a number of variables that have been found to influence bird species diversity within a landscape: area of habitat in which the species nests, amount of habitat within the landscape, degree of fragmentation, vegetation characteristics of the habitat patch, and area within the landscape deemed urban in municipal land-use designations (amount of urbanization). From this literature, I formulated four hypotheses describing the most important variables for avian diversity: (1) the area of the habitat patch is most important, (2) only variables describing the habitat patch itself are important, (3) the area of the habitat patch is important, but landscape variables should also be considered and (4) urbanization is most important. These hypotheses were considered competing explanations of bird species diversity at the landscape scale. <br /><br /> A database of breeding bird data and landscape information, in a geographic information system platform, was used to investigate the comparative strength of the competing hypotheses for the Greater Toronto Area. A mathematical expression with a Poisson model format was created to represent each hypothesis. The model selection technique based on Kullback-Leibler information using the Akaike Information Criterion was deemed most appropriate for the comparison of the models. Four separate Poisson model competitions were completed using two habitat types and two response variables: species richness and total abundance. In three of the four competitions, the best model included the habitat area and the amount of urbanization in the landscape. In the forth competition, this model was considered as strong as another model which included habitat area, amount of habitat in the landscape and degree of fragmentation. The results from the model competition support the hypothesis that habitat area is important, but landscape variables must also be considered to explain avian richness and total abundance. It appears that maintaining native bird biodiversity in the Greater Toronto Area should focus on preserving and possibly increasing habitat area and decreasing adjacent urbanization. Exploration of the best model in the forest analysis with the richness response variable found that a 10% increase in habitat area cause approximately a 10% increase in species richness, and a 10% increase in urban area caused approximately a 20% decrease in species richness. Consequently, current natural heritage planning in Ontario should consider urban development as an important negative effect on native birds.