Using data collected from 2 archaeological surveys of the Upper Cumberland Plateau (UCP), Pogue Creek Gorge and East Obey, a site location model was developed for prehistoric rock shelter occupation in the region. Further, the UCP model was used to explore factors related to differential site selection of rock shelters. Different from traditional approaches such as those that use (aspatial) logistic regression, the UCP model was developed using spatial logistic regression. However, models were also generated using other regression-based approaches in an effort to demonstrate the need for a spatial approach to archaeological site location modeling. Based on the UCP model, proximity to the vegetation zones of Southern Red Oak and Hickory were the most influential factors in prehistoric site selection of rock shelters on the UCP.

Evacuation and sheltering behaviors were modeled for a hypothetical release of uranium hexafluoride (UF6) from Nuclear Fuel Services (NFS) in Erwin, Tennessee. NFS down-blends weapons grade Cold War-era nuclear fuel material and processes highly-enriched uranium occasionally using UF6 onsite. Risk associated with a chemical release to the surrounding residential population was assessed by running 2 scenarios involving an airborne release of UF6 to compare evacuation and sheltering in place actions as effective survival strategies. Risk is minimal and evacuation is recommended for people within a 2-mile radius of the release point. Shelter in place actions are recommended for all critical facilities that have the potential to be affected by a chemical release plume. Oak Ridge National Laboratory’s Radiological Assessment System for Consequence Analysis and Capacity-Aware Shortest Path Evacuation Routing in conjunction with a geographic information system proved to be valuable technological tools in determining evacuation routing and exposure zones.

Abdominal aortic aneurysms are the third leading cause of sudden death in men 60 years and over. AAAneurysm Outreach provides free screenings to residents of Louisiana and beyond. Service areas were calculated for each AAAneurysm Outreach screening event location and stroke center. Data provided by the 2010 U.S. Census, the American Community Survey, and the Behavioral Risk Factor Surveillance System were used to describe demographics of the underserved populations and to identify high-risk areas for targeted interventions. Twenty-five percent of age-eligible Louisianans reside outside optimal drive-time-to-screening-event zones but within spatially clustered areas of increased prevalence rates. The maximum drive-time-to-treatment zones excluded 1,395 residents 65 years and over from timely access to emergency medical care. Results revealed limitations in the geographic breadth of the screening program and small disparities in accessibility to emergency healthcare.

The well-known symbol for conservation, the giant panda bear, and the original red panda have been forced into remote habitats due to anthropogenic disturbance, making ecological study difficult. Therefore the first known species distribution model was created to predict the most likely areas of occurrence within the known range of these elusive animals. These models were then projected onto North America and evaluated against existing breeding programs. Additionally, the close proximity of the Gray Fossil Site and the discovery of the most complete fossil red panda specimens in the world allowed ecomorphological comparisons between the modern red panda, Ailurus fulgens and the fossil red panda, Pristinailurus bristoli. Spatial data and morphometric results from this study will aid conservation work and breeding programs globally.

Increasing concentrations of atmospheric CO2 are in dynamic equilibrium with the oceans. The absorption of CO2 by seawater causes a decrease in seawater pH and calcite saturation state (SS). This process, termed ocean acidification, exerts deleterious effects on marine calcifiers. Studies of symbiont-bearing large benthic foraminifera (LBF) have reported a generally unfavorable response to increased concentrations of carbon dioxide ([CO2]). Experiments and analyses were undertaken to examine the effect of increased [CO2] on the growth rate, ultrastructure, stable isotopes of carbon and oxygen, as well as Mg/Ca of the high-Mg miliolid Archaias angulatus and the low-Mg rotalid Amphistegina gibbosa. A CO2-injection culture study was performed at pH 8.0, 7.8 and 7.6, corresponding to CO2 concentrations of approximately 400 ppm, 800 ppm, and 1,300 ppm. After 2, 4, or 6 weeks of treatment, bags containing groups of approximately 20 previously-imaged live specimens were removed and prepared for the aforementioned analyses. Archaias angulatus responded to increased [CO2] by reducing test growth rate at 1,300 ppm CO2 (pH 7.6) by 50% (p < 0.01, r2 = 36%), increasing its pore area (F(2,3477) = 103.37, p West Florida shelf LBF carbonate production attributed to LBF was estimated by combining interpolations of SS calcite at three treatment levels, corresponding to pH 8.1 (400 ppm CO2), pH 7.8 (800 ppm CO2), and pH 7.6 (1,300 ppm CO2), with a map of the carbonate fraction of seafloor sediment. Growth rates for 10 species were estimated in a meta-analysis of culture studies; these rates were used to model the response of miliolids and rotalids to increased [CO2]. In the model, rotalids responded to higher CO2 concentrations by reducing their average adult size by 20% at 800 ppm CO2 and 40% at 1,300 ppm CO2. Miliolids responded by reducing their average adult size by 40% at 800 ppm CO2 and 75% at 1,300 ppm CO2. Modeled LBF carbonate production for the west Florida shelf is 7 Mt at 400 ppm, 4.8 Mt at 800 ppm, and 2.5 Mt at 1,300 ppm. In a high CO2 world, low-Mg rotalids exhibit modest reductions in test growth rates and carbonate production, whereas high-Mg miliolids exhibit major reductions in test growth rates and carbonate production.

The 52,830-acre Mammoth Cave National Park, located in the karst region of south-central Kentucky, was formally established in July of 1941, culminating nearly three decades of park creation that displaced several thousand residents of the region. This thesis sampled residents using the 1920 manuscript census for the United States Census of Population and Housing and tracked their migration destinations using the 1930 and 1940 manuscript censuses. Migration patterns for the entire sample, as well as by race and homeownership status, were identified through mapping. Out-migrants generally chose locations north, west, and east of the proposed park area, noticeably neglecting the Deep South. Statistical analyses proved significant differences between proportions of Black out-migrants and White out-migrants moving to urban areas, as well as those of homeowners and renters who were not successfully tracked during analysis. The research underlines unintended consequences of the forced out-migration from the proposed Mammoth Cave National Park and several factors that contributed to it. In the process, the thesis fills a gap in research on Mammoth Cave National Park and sheds light on an important aspect of Kentucky’s history.

Modern Geoweb-enabled PPGIS methodologies incorporate interactive map applications as the main driver for public engagement and data collection. However, little research explores exactly how the public interact with these applications to produce spatial data, a fact that contributes to criticisms of final data quality. Usability evaluation offers a solution for developing better PPGIS data production systems by identifying problems in the application interface for which the public engage. Drawing on a case study example of a PPGIS application developed to collect socio-spatial data from members of a random public, this paper addresses usability in a three-stage approach. First, controlled experimentation methods capture performance, preference, and data production metrics. Second, visual and statistical analysis of the captured usability data identify problems in the interface. Results indicate that users learned, became efficient, and were generally satisfied with the application, but also committed errors that may have affected data quality. Third, a solution-oriented critique of the application interface suggests new design options to mitigate future problems in similar applications. The paper ends by providing a conceptual framework for usability as it relates to PPGIS data production and incorporates it into an informed discussion on data quality and future research needs for maintaining the viability of PPGIS projects.

In northwest Alaska, archaeologists hypothesize that environmental variability was a major factor in both growing coastal population density, with large aggregated villages and large houses, between 1000 and 500 years ago (ya), and subsequent decreasing population density between 500 ya and the contact era. After 500 ya people are thought to have dispersed to smaller settlements with smaller house sizes in coastal areas, and perhaps, upriver. This settlement pattern was identified through research at four site locations over 30 years ago. The changing geographic distribution of sites, associated settlement size, and house size has not been examined in detail. A more careful examination of changing northwest Alaskan settlement patterns is needed before larger questions about socio-economic organization can be addressed. I use Geographic Information Systems (GIS) to evaluate the evidence for a geographic redistribution of Arctic peoples during the Late Holocene. The author constructed a database of settlement location and site attribute information, specifically the number of houses within each settlement and the size (m2). Data were collected from a dataset of Western Arctic National Parklands (WEAR), the Alaska Heritage Resource Survey (AHRS) database of archaeological sites in Alaska, 409 unpublished site reports and field notes curated by the National Park Service (NPS) and Bureau of Indian Affairs (BIA), and the results of recent fieldwork in northwest Alaska. A total of 486 settlements were identified within the northwest Alaska with 128 settlements having temporal and site attribute data. The author incorporated settlement size data into a GIS database and then carried out global, Moran's I, local Moran's I, and local Getis-Ord spatial analyses to test whether settlement redistribution occurred and if key settlement locations shifted after 500 ya. The site attribute data (number of houses and average size of houses) are used to test the additional aspects of the proposed settlement pattern change after 500 ya. A total of 83 settlements with 465 houses are used to test if the average size of settlements and average house size changed after 500 ya. The results of the spatial analyses indicate no statistically significant patterns in the spatial distribution of settlements. Site attribute analysis shows no statistical difference in the average number of houses per village or the average size of houses before or after 500 ya. The results of this work build our understanding of regional settlement patterns during the late Holocene. By testing settlement pattern change, i.e. settlement distribution, settlement size, and house size, future research into settlement pattern change can begin to evaluate likely causes for the observed changes. My method, specifically the use of GIS as a method for testing settlement pattern change, can be applied to other regions and temporal scales.

This thesis is a case study of a small agricultural community located along the eastern edge of Everglades National Park, The purpose of this study was to document the way land use decisions have been made and how these decisions have affected this community and the Everglades ecosystem. This research demonstrated that decisions made by the involved agencies have negatively affected both the community of Pariah, Florida and the Everglades ecosystem. Research methods included extensive document research, participant observation and formal and informal interviews. It appears that public concern over “saving the Everglades” has been used to provide a legitimating framework for the achievement of a plurality of personal goals and unstated agency agendas that have little or nothing to do with either the Everglades or the environment in general.

Initial soil surveys are incomplete for large tracts of public land in the western USA. Digital soil mapping offers a quantitative approach as an alternative to traditional soil mapping. I sought to predict soil classes across an arid to semiarid watershed of western Utah by applying random forests (RF) and using environmental covariates derived from Landsat 7 Enhanced Thematic Mapper Plus (ETM+) and digital elevation models (DEM). Random forests are similar to classification and regression trees (CART). However, RF is doubly random. Many (e.g., 500) weak trees are grown (trained) independently because each tree is trained with a new randomly selected bootstrap sample, and a random subset of variables is used to split each node. To train and validate the RF trees, 561 soil descriptions were made in the field. An additional 111 points were added by case-based reasoning using aerial photo interpretation. As RF makes classification decisions from the mode of many independently grown trees, model uncertainty can be derived. The overall out of the bag (OOB) error was lower without weighting of classes; weighting increased the overall OOB error and the resulting output did not reflect soil-landscape relationships observed in the field. The final RF model had an OOB error of 55.2% and predicted soils on landforms consistent with soil-landscape relationships. The OOB error for individual classes typically decreased with increasing class size. In addition to the final classification, I determined the second and third most likely classification, model confidence, and the hypothetical extent of individual classes. Pixels that had high possibility of belonging to multiple soil classes were aggregated using a minimum confidence value based on limiting soil features, which is an effective and objective method of determining membership in soil map unit associations and complexes mapped at the 1:24,000 scale. Variables derived from both DEM and Landsat 7 ETM+ sources were important for predicting soil classes based on Gini and standard measures of variable importance and OOB errors from groves grown with exclusively DEM- or Landsat-derived data. Random forests was a powerful predictor of soil classes and produced outputs that facilitated further understanding of soil-landscape relationships.