Master of Science (MS), Bowling Green State University, 2024, Geology

The long-nosed peccary (Mylohyus nasutus) and flat-headed peccary (Platygonus compressus) are two of the most recently extinct members of Family Tayassuidae from North America. The aim of this study was to create ecological niche models (ENMs) for both species during the Heinrich Stadial 1, Bølling-Allerød, and Younger Dryas intervals of the latest Pleistocene across the contiguous United States and parts of northern Mexico to provide insight on their distribution, how it changed over the time intervals, and what environmental (climate and floral frequency) factors affected both species' ranges just prior to their extinction around the Pleistocene/Holocene transition. The Neotoma Paleoecology Database, Paleobiology Database, and published research articles were used to compile peccary occurrence data. Climatic raster data were derived from PaleoClim. Floral data (specifically pollen abundance) was compiled from Neotoma, then ISODATA clustering was used in GIS SAGA to create frequency ratio maps for several dominant floral groups to rasterize floral abundancy maps. Peccary occurrence data and environmental rasters were input into Maxent, which was used to create both jackknife and response curve ENM models. Lastly, the ENMs were combined with Paleoindian archeological data (via p3k14c) to provide insight on human and peccary relationships. M. nasutus was found to have insufficient dated occurrence points to create ENMs for the targeted time slices. P. compressus was found to have had a large potential range across much of the modeled region throughout all three time intervals. P. compressus was very tolerant of vast ranges in temperature and preferred to live in forested habitats, but avoided areas with low precipitation, high precipitation seasonality, forests abundant in oak, or more open grassland/scrubland. P. compressus' large potential range through both cold and warm intervals of the Late Pleistocene suggests that the changing environmenta (open full item for complete abstract)

Committee: Margaret Yacobucci Ph.D. (Committee Chair); Jeff Snyder Ph.D. (Committee Member); Peter Gorsevski Ph.D. (Committee Member) Subjects: Geographic Information Science; Paleontology

Master of Arts, Miami University, 2021, Geography

The Pacific Golden Chanterelle (Cantharellus formosus) is a widely sought-after mushroom most abundant in the forests of Washington and Oregon, USA. This project used the species to investigate how accurately the species distribution could be modeled using natural history (herbarium) as model training data and citizen science (iNaturalist) as validation data. To combat the potential sampling bias towards population centers an effort variable weighting scheme was used to consider observations in harder to reach areas more than those in easier to access areas. Four models were created and run using the natural history data as training points: Random Forests (RF), Maxent, General Linear Model (GLM), and Artificial Neural Network (ANN); the effort variable was only applied to the ANN and GLM models. Out of these four, RF was found to perform the best with an equitable skill score (ETS) 0.987 when tested against the iNaturalist citizen science validation points. Overall, this project provides a good proof of concept and framework for the use of herbarium and citizen science data for use in biogeographical modeling projects in the future.

Committee: Mary Henry (Advisor); Jessica McCarty (Committee Member); Nicholas Money (Committee Member) Subjects: Geography

Master of Science (MS), Bowling Green State University, 2016, Geology

The spatial paleoceanography of the entire Western Interior Seaway (WIS) during the Cenomanian-Turonian Oceanic Anoxic Event has been reconstructed quantitatively for the first time using Geographic Information Systems. Models of foraminiferal occurrences—derived from Dempster-Shafer theory and driven by fuzzy sets of stratigraphic and spatial data—reflect water mass distributions during a brief period of rapid biotic turnover and oceanographic changes in a greenhouse world. Dempster-Shafer theory is a general framework for approximate reasoning based on combining information (evidence) to predict the probability (belief) that any phenomenon may occur. Because of the inherent imprecisions associated with paleontological data (e.g., preservational and sampling biases, missing time, reliance on expert knowledge), especially at fine-scale temporal resolutions, Dempster-Shafer theory is an appropriate technique because it factors uncertainty directly into its models. Locality data for four benthic and one planktic foraminiferal species and lithologic and geochemical data from sites distributed throughout the WIS were compiled from four ammonoid biozones of the Upper Cenomanian and Early Turonian stages. Of the 14 environmental parameters included in the dataset, percent silt, percent total carbonate, and depositional environment (essentially water depth) were associated with foraminiferal occurrences. The inductive Dempster-Shafer belief models for foraminiferal occurrences reveal the positions of northern and southern water masses consistent with the oceanographic gyre circulation pattern that dominated in the seaway during the Cenomanian- Turonian Boundary Event. The water-mixing interface in the southwestern part of the WIS was mostly restricted to the Four Corners region of the US, while the zone of overlap of northern and southern waters encompassed a much larger area along the eastern margin, where southern waters occasionally entered from the (open full item for complete abstract)

Committee: Margaret Yacobucci Dr. (Advisor); Peter Gorsevski Dr. (Committee Member); Andrew Gregory Dr. (Committee Member) Subjects: Earth; Geographic Information Science; Geography; Geology; Marine Geology; Oceanography; Paleoecology; Paleontology; Statistics

Master of Science, The Ohio State University, 2023, Plant Pathology

My work aimed to develop a comprehensive environmental risk mapping framework for beech leaf disease (BLD) using an ensemble approach in species distribution modeling. The recent emergence of BLD poses a significant risk to American beech trees in the northeastern region of the United States, and accurate risk mapping can assist in targeted management strategies, particularly biosurveillance. To generate reliable present and future models, I first modeled the current distributions of American beech in its full U.S. range as well as BLD. Then I explored the impact of climate change on future American beech habitat suitability as well as BLD distribution. The models integrate present and projected climate data to anticipate alterations in suitable habitats and potential shifts in species distribution ranges. The ensemble approach I used combines multiple boosting and bagging approaches with machine learning or regression-based algorithms. These include multiple adaptive regression splines, flexible discriminant analysis, random forest, and boosted regression trees to generate a robust risk map. The findings of this research offer a crucial understanding for managing forests and preserving ecosystems, which encompasses pinpointing vulnerable regions for BLD and evaluating the potential influence of climate change on American beech populations. The ensemble approach employed in species distribution modeling effectively produced highly predictive maps, while the habitat suitability models emphasized the significance of taking climate change into account when devising conservation strategies.

Committee: Pierluigi E. Bonello (Advisor); Desheng Liu (Committee Member); Christopher G. Taylor (Committee Member); Laurence V. Madden (Committee Member) Subjects: Environmental Science; Forestry; Plant Pathology

Doctor of Philosophy, Miami University, 2016, Ecology, Evolution and Environmental Biology

The Tigris and Euphrates Rivers and their tributaries form the arteries of life in the central part of the Middle East, where climate change and anthropogenic disturbance have been evident in recent decades. While the Tigris River has a long history of human use, the conservation status for the majority of its basin is poorly known. In addition, planning for conservation, given limited time, funds, and prior information, has remained a challenge. In my dissertation research, I sampled 53 randomly selected sites in the Kurdistan Region (the KR) of northern Iraq, a poorly studied region of the Upper Tigris and Euphrates freshwater ecoregion, for water quality bioindicators, mayflies (Insecta, Ephemeroptera), stoneflies (Plecoptera), and caddisflies (Trichoptera) (a.k.a. EPT). I identified the mayflies to the finest possible taxonomic level and created the first Iraqi checklist and larval key to nine families, nine subfamilies, 19 genera, and 13 subgenera, and supported it with 117 state-of-the-art scientific illustrations using fresh specimens collected during my study (Chapter 1). I performed an initial species morphological identification for mayflies and stoneflies, then identified them genetically after sequencing the full-length of the mitochondrial cytochrome oxidase subunit 1 (COI) gene (658 base pairs). I introduced Genetic Similarity Blocks (GSBs), a genetic-based analysis which was used along with morphology and other genetic-based analyses to overcome the taxonomic impediment and accelerate species identification. I delineated Operational Taxonomic Units (OTUs) using genetic-based analyses, then matched OTUs to delineate Species-Like Units (SpLUs). I compared and contrasted SpLUs morphologically and found five stonefly and more than 55 mayfly taxa, the majority of them being new records for Iraq, and many of them potentially new to science (Chapter 2). I identified 76 planning units within aquatic ecosystems in the KR and prioritized a subset of them for EP (open full item for complete abstract)

Committee: David Berg (Advisor); Bruce Cochrane (Committee Member); John Morse (Committee Member); Michael Vanni (Committee Member); Thomas Crist (Committee Member); Mary Henry (Committee Member) Subjects: Animal Sciences; Aquatic Sciences; Biology; Conservation; Ecology; Environmental Studies; Wildlife Conservation; Zoology

Master of Arts, Miami University, 2016, Geography

A current assessment of wildebeest (Connochaetes taurinus) distribution throughout the Masai Mara National Reserve (MMNR) and adjoining group ranches has not been investigated for over 15 years. This information is greatly needed to protect populations of wildebeest and their ranges. MaxEnt, a statistical model, was used to evaluate influential factors for wildebeest distribution and predict suitable habitats throughout the northern extent of the Serengeti National Park, the MMNR, and adjoining group ranches. There were thirty five abiotic and biotic variables used to create two distribution models in MaxEnt across the study area for two different time periods. Both models performed well with training AUCs >0.80. Precipitation seasonality, isothermality, and distances to lodges were the greatest contributing variables to wildebeest distribution in the November model. Kauth-Thomas wetness, annual temperature range, and distances to camps were significant factors for wildebeest distribution in the June model. Predictive maps from the June 2010 model revealed higher concentrations of predicted habitat suitability in areas historically impacted by the expansion of mechanized farming practices. It is recommended that the MMNR collaborate with local group ranch conservancies to secure seasonal dispersal sites for wildebeest and impose land use policies in unprotected areas. Results from MaxEnt also revealed that bioclimatic variables and soil and plant moisture are significant contributors to wildebeest distributions. The MMNR should evaluate the potential effects imposed by climate change to wildebeest distributions and populations.

Committee: John Maingi Dr. (Advisor); Amélie Davis Dr. (Committee Member); Thomas Crist Dr. (Committee Member) Subjects: Animal Sciences; Biographies; Ecology; Geographic Information Science; Geography; Statistics

Doctor of Philosophy (PhD), Wright State University, 2015, Environmental Sciences PhD

Humans have caused drastic changes in ecosystems and communities through their modification of the natural landscape. Rare species, often highly specialized, are more impacted by these changes. Pieris virginiensis is a rare butterfly native to eastern North America that is a species of concern due to negative influences from habitat loss and plant invasion. This thesis discusses several threats to P. virginiensis, including habitat loss, climate change, competition, and the cascading effects of a novel European invasive plant, Alliaria petiolata, that attracts oviposition but does not allow for larval survival. First, I examined a local extinction event and attributed it primarily to several seasons of poor weather and extreme climatic events, but with contributions by an increasing deer population and the introduction of A. petiolata. Second, I found that A. petiolata attracts approximately two-thirds of total eggs, but no larvae survive on the novel host. I tested several chemical causes of larval death and identified two potential contributors: sinigrin, which delays growth, and alliarinoside, which reduces survival. I also examined competition between P. virginiensis, its host plants, and novel competitors in the habitats. First, I looked at shared habitat use between P. virginiensis and another, exotic Pierid butterfly P. rapae. Although habitats are occasionally shared, P. rapae is most likely not a large influence on the success or failure of P. virginiensis. Second, I examined the influence of A. petiolata when it competes with two native host plants of P. virginiensis, and found differential effects of each life stage of A. petiolata on the native host plants. Finally, I used a combination of species distribution modeling and genetic sequencing to determine the current and future states of P. virginiensis given the changing climate and other stressors on P. virginiensis populations. Although secure currently, future stressors will most likely cause (open full item for complete abstract)

Committee: Don Cipollini Ph.D. (Advisor); John Stireman Ph.D. (Committee Member); Jeffrey Peters Ph.D. (Committee Member); Thaddeus Tarpey Ph.D. (Committee Member); Francie Chew Ph.D. (Committee Member) Subjects: Biology; Botany; Climate Change; Conservation; Ecology; Environmental Science; Environmental Studies

Master of Arts, Miami University, 2014, Geography

The ability to predict American beech distribution (Fagus grandifolia Ehrh.) from environmental data was tested by using a geographic information system (GIS) in tandem with species distribution models (SDMs). The study was conducted in Butler and Preble counties in Ohio, USA. Topography, soils, and disturbance were approximated through 15 predictor variables with presence/absence and basal area serving as the response variables. Using a generalized linear model (GLM) and a boosted regression tree (BRT) model, curvature, elevation, and tasseled cap greenness were shown to be significant predictors of beech presence. Each of these variables was positively related to beech presence. A linear model using presence only data was not effective in predicting basal area due to a small sample size. This study demonstrates that SDMs can be used successfully to advance our understanding of the relationship between tree species presence and environmental factors. Large sample sizes are needed to successfully model continuous variables.

Committee: Mary Henry PhD (Advisor); David Gorchov PhD (Committee Member); Jerry Green PhD (Committee Member) Subjects: Botany; Ecology; Forestry; Geography