MSU Libraries Digital Repository (Michigan State University)
Not a member yet
106711 research outputs found
Sort by
ABUNDANCE, DISTRIBUTION, AND PERSISTENCE OF BIRDS AND MAMMALS UNDER ENVIRONMENTAL CHANGE
No full textThesis (Ph.D.)--Michigan State University. Integrative Biology - Doctor of Philosophy, 2025More than one-fifth of vertebrate species are threatened with extinction. Despite global efforts to conserve biodiversity, species abundance and diversity continues to decline across taxonomic groups and geographic regions. Achieving conservation goals relies on accurate biodiversity monitoring and a deep understanding of how natural and anthropogenic factors influence individual species and communities as a unit. A key theme of my research is advancing the application of quantitative methods to determine the status and persistence of wildlife by maximizing the use of available data. To that end, my dissertation work aims to build knowledge that will help sustain biodiversity in a rapidly changing world. In chapter one, I develop a modified hierarchical distance-sampling model to estimate the abundance of hard-to-detect chimpanzees and elephants in a dense tropical forest within the Albertine Rift ecoregion in east-central Africa. The Albertine-Rift is one of the most biodiverse places in the world, supporting more than half of Africa\u2019s bird species and 40% of mammals. In situations where visibility is limited, indirect measures of species (e.g., nests, dung) serve as proxies for counts of individuals. Current approaches to estimate population abundance using indirect sign data do not adequately account for variations in sign production and spatial patterns of animal density. My model reveals a significant decline in chimpanzees, and an increase in elephants between 2007 and 2021 within the region. My modelling approach produces more precise estimates of covariate effects on animal density by maximizing the use of all the available data to account for long-term and recent variations in abundance. In chapter two, I apply a community distance sampling model to evaluate niche overlap of ecologically similar bird species within a montane tropical forest in the Albertine Rift, where elevation strongly correlates with environmental conditions (e.g., climate, forest type). Although hundreds of bird species live within the region, the underlying mechanisms that facilitate coexistence of many competing species are poorly understood. My model shows that bird species coexistence is determined primarily by abiotic factors (i.e., the environmental elevation gradient) and secondarily by biotic factors (i.e., within-habitat segregation across horizontal space and vertical forest strata). Quantifying niche overlap indices along multiple dimensions provides deep insights into community structuring and a foundation to predict species distributions in response to ongoing environmental change. In chapter three, I assess the impacts of anthropogenic threats across bird and mammal functional groups using global data collected by the International Union for Conservation of Nature. Anthropogenic threats vary in how they affect individual species, taxonomic groups, and biodiversity as a whole. My analysis identified the anthropogenic threats with the highest impacts on bird and mammal functional groups and the functional groups that are most threatened overall. Across different anthropogenic threats, I found that the most vulnerable functional groups were vertivores, aquatic predators, frugivores, and herbivores for birds, and vertivores, aquatic predators, and frugivores/nectivores/granivores for mammals. My results reveal that anthropogenic threats vary in how they affect functional diversity, offering valuable baseline information for how best to target conservation actions for vulnerable species groups. The work in my dissertation offers valuable insights on how to monitor, estimate, and predict wildlife population distributions in response to natural environmental variation and anthropogenic threats. Tracking biodiversity trends over time and space is critical to teasing apart the interacting factors that drive species abundance patterns and determining the consequences of changing environmental conditions. The models and methods in my research are transferable to other systems and taxonomic groups, offering a path forward to improved conservation planning.Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).Includes bibliographical references
PRE-ANALYTICAL SAMPLE PROCESS MODIFICATIONS TO DECREASE TIME TO DETECTION OF SALMONELLA SER. NEWPORT AND LISTERIA MONOCYTOGENES FROM DIVERSE FOOD MATRICES
Get PDFThesis (Ph.D.)--Michigan State University. Comparative Medicine and Integrative Biology - Doctor of Philosophy, 2025Foodborne illnesses continue to negatively affect public health. Current strategies to detect and prevent illnesses rely on prolonged enrichment protocols of 24-48 hours. While rapid methods are constantly being developed, these methods do not consider preanalytical sample processing, which is a critical first step for reliable and reproducible results. Additionally, the time to detection for an assay does not include the preparation and enrichment steps that must be completed to arrive at optimal pathogen numbers to enable detection. To address this gap, the author evaluated and refined the use of a proprietary magnetic nanoparticle functionalized with chitosan (F#1 MNPs) as a preanalytical sample processing tool to capture and concentrate foodborne pathogens from complex food matrices. Two foodborne pathogens, Listeria monocytogenes (gram-positive) and Salmonella ser. Newport (gram-negative) were used to evaluate the F#1 MNPs in strawberries, romaine lettuce, and cotto salami, representing diverse food matrices. These pathogens were chosen for this proof-of-concept study based on their significant public health impact. Chitosan electrostatically binds to the cell-surface structure of bacteria. Therefore, it is hypothesized that the F#1 MNPs also bind to the exterior of pathogens. However, the exact binding mechanism remains unknown. Due to this, all testing used cold-stressed pathogens to simulate their physiological state after food processing. First, statistical design of experiments (DOE) was used to optimize protocols for extracting 64 3 CFU/g of bacterial contamination in diverse matrices with only minor protocol adjustments. This study highlights the potential to standardize protocols and the ability to rapidly adjust them based on regulatory requirements for different pathogens and food matrices. Next, using the same strains and food matrices, the effect of the F#1 MNPs on pathogen enrichment was evaluated. Modifications reduced broth enrichment times to 4-12 hours without inhibiting target pathogen growth on selective agars, expediting the overall time to single-colony isolation. This is especially important for regulatory enforcement that still relies on the isolation of pathogens for downstream testing and outbreak surveillance and investigation. Finally, the use of shotgun metagenomics revealed potential applications beyond bacterial pathogens. The F#1 MNPs can also capture non-pathogenic bacteria, viruses, and fungi, which may have applications such as environmental bioindicators. This further shows the versatility of the F#1 MNPs as a preanalytical sample processing tool in a wide range of detection pipelines, such as multi-organism detection with multiplex assays, pathogen-agnostic testing, and identifying pathogens in emerging food vehicles. By streamlining pathogen extraction and concentration, F#1 MNPs offer significant potential to improve surveillance, outbreak detection and prevention, and overall food safety.Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).Includes bibliographical references
Novel Computational Approaches for Nuclear Fission Theory
Get PDFThesis (Ph.D.)--Michigan State University. Physics - Doctor of Philosophy, 2025Nuclear fission is important for energy production, medicinal applications, nonprolifer-ation efforts, and nucleosynthesis studies. The rapid neutron capture process (r process) contributes to observed abundances of medium- and heavy-mass nuclei, and requires fission data for hundreds of nuclei, most of which lie outside of experimental reach. Therefore, predictive fission models with quantified uncertainties are required. In this thesis, spontaneous fission is described as tunneling through an effective barrier defined using a set collective coordinates, called the potential energy surface (PES), which is computed using nuclear density functional theory (DFT). The half-life is then determined by the tunneling pathway, and the primary fragment yields are approximately determined by its endpoint. Computing uncertainties for these quantities in a Bayesian framework requires tens- to hundreds-of-thousands of calculations, making it computationally infeasible. This problem is exacerbated by the large number of nuclei that participate in the r process. This thesis is divided in two parts. The first discusses the formalism necessary for com- puting spontaneous fission observables. Improvements to the tunneling pathway calculation are presented, and the improved methodology is applied to nuclei with competing fission modes. The second part discusses two strategies for approximating these observables across the r process region of the chart, with quantified uncertainties. The first strategy uses neu- ral networks to emulate the PES. The second uses dimensionality reduction techniques to propagate statistical uncertainties from the energy density functional posterior parameter distribution.Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).Includes bibliographical references
Structure and Motion from Depth and Correspondence Models
Get PDFThesis (Ph.D.)--Michigan State University. Computer Science - Doctor of Philosophy, 2025Recovering structure and motion from videos is a well-studied comprehensive 3D vision task that involves (1) image calibration, (2) two-view pose initialization, and (3) multi-view Structure-from-Motion (SfM). Prior arts are optimization-based methods built over sparse image correspondence inputs. This thesis develops systematic approaches to enhance classic solutions with deep learning models. We introduce EdgeDepth and PMatch for dense monocular depthmaps and dense binocular correspondence map estimations. Since classic approaches typically rely on sparse and accurate inputs, they are less suitable for the dense yet high-variance predictions from dense depth and correspondence models. As a solution, we propose to optimize through the robust inlier-counting-based scoring function, which is widely applied in RANdom SAmpling Consensus (RANSAC). Our system is structured as follows: (1) For image calibration, we introduce WildCamera. The system utilizes a RANSAC algorithm applied to a dense incidence field regressed by a deep model. It calibrates in-the-wild monocular images without checkerboard. (2) In two-view pose estimation, we introduce LightedDepth.It estimates the optimal pose by aligning the depth map with the correspondence map, maximizing the projective inliers. (3) The strategy is extended to a Hough Transform in RSfM for multi-view SfM over a local to frame system. (4) We generalize the RSfM discrete inlier counting scoring function to a smoothed scoring function via marginalizing thresholds for general SfM task. To this end, we formulate a comprehensive system that recovers structure and motion from two-view / local multi-view / large-scale multi-view images with dense monocular depthmap and binocular correspondence maps. Compared to prior arts, our methods show comprehensive improvement on two-view, small-scale, and large-scale multi-view systems.Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).Includes bibliographical references
IDENTIFICATION, ASSESSMENT AND HISTOPATHOLOGY OF PATHOGENIC FUSARIUM OXYSPORUM SPECIES COMPLEX ON SUGAR BEET
Get PDFThesis (M.S.)--Michigan State University. Plant Pathology - Master of Science, 2025Fusarium yellows of sugar beet is caused by fungi in the F. oxysporum species complex (FOSC) and impacts the sugar beet industry from the field to post-harvest storage. The objectives of this work were to 1) identify pathogenic FOSC isolates collected in Michigan on sugar beet and assess their virulence levels and 2) to examine the infection process and to lay the groundwork for identification of potential histopathological differences between F. commune, a member of the FOSC and other FOSC members. The first set of objectives were addressed by isolating, identifying, and assessing virulence of Michigan FOSC isolates in a greenhouse assay using foliar and root severity ratings for Fusarium yellows. In comparison to the controls, of the 35 isolates screened in the greenhouse, 5.7% were classified as moderate virulence, and 60% were classified as low virulence, and 34.3% were non-pathogenic. The moderately virulent isolates will be of benefit to subsequent experiments and resistance screening trials targeted to manage Fusarium yellows in Michigan. For the second set of objectives, mature sugar beets were inoculated with an F. commune isolate and plants were collected every three days post inoculation through 18 days. The bottom half of the roots were fixed and stained for conventional and confocal microscopy. It was observed that F. commune initially colonized the root surface of the tap root and feeder roots. While penetration did occur in the feeder roots, the colonization of the vasculature of the feeder roots did not progress to the main tap root. On the tap root, penetration and colonization of the root interior occurred around the root tip. Hyphae subsequently grew into and traveled via the xylem in cambial rings and the stele from the root tip up to at least as far as the root groove. The knowledge acquired over the course of these experiments will help sugar beet breeding programs and growers make informed decisions on managing Fusarium yellows of sugar beet.Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).Includes bibliographical references
Machine Learning-based Coarse-Grained models for molecular systems
No full textThesis (Ph.D.)--Michigan State University. Computational Mathematics, Science and Engineering - Doctor of Philosophy, 2025Multiscale modeling poses a formidable challenge in computational mathematics, particularly in integrating microscale interactions into meso- or macro-scale constitutive relations. While reduced-order models allow the simulation of extensive systems, their analytical formulations are generally unclosed. Take coarse-grained molecular dynamics as an example, the Mori-Zwangzig formulism decomposes the dynamics into deterministic, memory, and stochastic terms, but the explicit forms for these three terms are unknown. In this thesis, we present a series of studies to solve these problems. Firstly, the main challenge for the deterministic term comes from the high dimensionality and the presence of energy barriers of the free energy surface (FES). We propose a consensus sampling-based approach that reformulates the FES construction as a minimax problem. This framework simultaneously optimizes the function representation of the FES and the training set used to learn it. In particular, the maximization step establishes a stochastic interacting particle system to achieve the adaptive sampling of the max-residue regime by modulating the exploitation of the Laplace approximation of the current loss function and the exploration of the uncharted phase space; the minimization step updates the FES approximation with the new training set. By iteratively solving the minimax problem, the present method essentially achieves an adversarial learning of the FESs with unified tasks for both phase space exploration and posterior error-enhanced sampling. Besides, memory interactions are also important for predicting the collective transport and diffusion processes. To construct this, we introduce a machine-learning-based coarse-grained molecular dynamics model that captures the dissipative many-body contribution. The neural network representation is carefully designed to preserve the physical symmetries and the thermo-consistency. Unlike the common empirical reduced models, the present model is constructed based on the Mori-Zwanzig formalism and naturally inherits the heterogeneous state-dependent memory term rather than matching the mean-field metrics such as the velocity autocorrelation function. Finally, when applied to non-equilibrium systems, models based on the Mori-Zwanzig formalism face inherent challenges. A key issue lies in the Zwanzig projection, which relies on the marginal distribution of the system. We present a data-driven approach for constructing reduced models that retain certain generalization abilities for non-equilibrium processes. Unlike the conventional CG models based on pre-selected CG variables (e.g., the center of mass), the present CG model seeks a set of auxiliary CG variables based on the time-lagged independent component analysis to minimize the entropy contribution of the unresolved variables. This ensures the distribution of the unresolved variables under a broad range of non-equilibrium conditions approaches the one under equilibrium.Through numerical validation, we demonstrate that our model can accurately predict viscoelastic behavior in various non-equilibrium flow regimes.Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).Includes bibliographical references
Assessing Preference for Choice-Making within Activity Schedules among Children with Autism Spectrum Disorder
Get PDFThesis (M.A.)--Michigan State University. Applied Behavior Analysis - Master of Arts, 2025Activity schedules are commonly used in early intensive behavioral intervention to increase student independence in completing a sequence of activities. A study conducted by Deel et al. (2021) taught participants to assemble activity schedules where participants selected the order of activities (choice) and where the sequence of activities was already determined (no-choice). Deel et al. (2021) then evaluated preference for choice-making opportunities embedded in activity schedules. The purpose of the current study was to systematically replicate Deel et al. (2021) and evaluate participant preference for choice or no-choice activity schedules in concurrent operant assessments. All participants learned to assemble and independently complete choice and no-choice activity schedules. Results of the concurrent operant assessments were idiosyncratic across participants. Practical implications and methodological considerations for future research are discussed.Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).Includes bibliographical references
Link Prediction Revisited : From Evaluation Pitfalls to Language Model Synergies
Get PDFThesis (Ph.D.)--Michigan State University. Computer Science - Doctor of Philosophy, 2025Artificial intelligence (AI) and machine learning (ML) have significantly impacted many aspects of daily life, with numerous methods involving structural graph data. Graphs, which model relationships between different entities, are widely used to represent real-world data, including social networks, transportation systems, chemical molecules, power and communication networks, and user-item interactions in recommendation systems. A fundamental task in graph data analysis is link prediction, which predicts connections between entities and is essential for understanding their relationships. For example, link prediction allows us to determine whether two individuals are friends in a social network or if a user will purchase an item in a recommendation system. Consequently, link prediction is crucial for advancing graph-based applications in real-world ML applications. However, several challenges impede progress in this area. Specifically, 1) existing evaluation settings are not unified or rigorous, leading to inconsistent and sometimes suboptimal results, and 2) graph nodes are frequently associated with textual attributes containing rich semantic information, and this data has become increasingly abundant. Language models excel at processing textual data to capture semantic insights. However, effectively integrating textual information with graph data to enhance real-world applications remains under-explored. In light of these challenges, this dissertation seeks to advance link prediction from two main perspectives: 1)identifying evaluation pitfalls across various graph types to inspire more advanced methods for link prediction, and 2) leveraging language models in synergy with link prediction techniques to enhance a range of real-world applications. From the first perspective, we investigate evaluation pitfalls in both uni-relational and multi-relational graphs. Based on these findings, we propose methods to mitigate the identified issues or develop more effective and efficient models. From the second perspective, we explore the use of language models to improve link prediction in recommendation systems, and conversely, apply link prediction techniques to enhance language models in query understanding tasks. By working on these two perspectives, this dissertation not only contributes to the development of more robust link prediction methods but also facilitates their application in practical, real-world scenarios.Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).Includes bibliographical references
Factors impacting pest damage on Pinus strobus L. : Implications for management
Get PDFThesis (Ph.D.)--Michigan State University. Forestry - Doctor of Philosophy, 2025Pinus strobus L. is an iconic forest species native to eastern North America. Managing for high-quality P. strobus timber is complicated by Pissodes strobi Peck, which colonizes and kills terminal leaders, causing bole defects. Pinus strobus is also impacted by Caliciopsis canker. This emerging disease, caused by the native fungal pathogen Caliciopsis pinea Peck, causes shallow bole and branch cankers, dieback, and sapling mortality. Caliciopsis canker disease was detected in Michigan in 2016, but the distribution and severity of the disease in the state is not well known. Management strategies for P. strobi often reduce P. strobus growth. To evaluate trade-offs between growth and P. strobi caused defects, long-term P. strobus stands were established under four different regeneration tactics: monoculture, shelterwood, even-aged mixed-species stand with varying density, and even-aged mixed-species stand with varying hardwood competition (Chapter 1). Trees in these stands were assessed 17 to 19 years after they were planted. Mean annual growth rate of P. strobus was eight times greater in the monoculture than in the shelterwood, however incident rate of bole defects was 64% in the monoculture and 6% in the shelterwood. The even-aged mixed-species stands yielded markedly different results with P. strobus reaching the overstory at the variable density stand but suffering up to 90% mortality at the varying hardwood competition stand. A survey of 66 P. strobus stands across northern Michigan found defects in P. strobus boles in 85% of stands and 11% of surveyed P. strobus (Chapter 2). Pinus strobus crown class and size class were predictive of bole defects but stand density factors (i.e., basal area, trees per hectare, stems per hectare) did not affect the likelihood of defects. Caliciopsis canker disease was widespread throughout the surveyed area. Signs of Caliciopsis canker disease were found in 47% of stands and 7% of surveyed P. strobus. The disease was most frequently found on P. strobus saplings, suggesting Caliciopsis canker has greater impact on regeneration. Lower P. strobus basal area was associated with a higher likelihood of Caliciopsis canker signs and symptoms, likely because lower basal area was more conducive to regeneration. During the stand surveys, bark samples with Caliciopsis spp. ascocarps were collected from P. strobus stems and branches (Chapter 3). From these, 37 isolates of Caliciopsis spp. were isolated, 35 of which were identified as C. pinea by sequencing the ITSrDNA gene region. The other two isolates were a potentially undescribed species of Caliciopsis, C. sp. 1. Pathogenicity tests were conducted on P. strobus excised branches and live seedlings. On excised branches, both species produced cankers larger in area and deeper than the control. In live seedlings, C. pinea produced cankers larger in area than the control while C. sp. 1 produced deeper but not larger cankers than the control. Caliciopsis species were re-isolated from canker margins following both pathogenicity trials. This study confirmed C. pinea as a pathogen of P. strobus and found that it was more virulent than C. sp. 1. In live seedlings, C. sp. 1. was more effective in colonizing down into the sapwood rather than the living cambium suggesting it is a weak pathogen or saprophyte, rather than a primary disease agent. These studies provide new insight into two native damage agents of P. strobus. Quantitative data on trade-offs between growth, competition, and P. strobi damage could help managers weigh different silviculture options (Chapter 1). Surveys found both damage agents are widespread throughout northern Michigan and provide a baseline for future efforts to track damage over time (Chapter 2). Pathogenicity tests showed Caliciopsis pinea is the primary cause of Caliciopsis canker disease, but at least one other species may contribute to the pathogen load (Chapter 3). Future studies are needed to elucidate the epidemiology of this emerging disease and to better inform management decisions.Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).Includes bibliographical references
A NUMERICAL APPROACH FOR EVALUATING FIRE RESISTANCE OF FRP-STRENGTHENED PRESTRESSED CONCRETE BEAMS
Get PDFThesis (M.S.)--Michigan State University. Civil Engineering - Master of Science, 2025In recent years, the demand for repair and strengthening of concrete structures has been increasing due to aging infrastructure, increased service loads, changes in functional use, design deficiencies, updates in building codes, and environmental deterioration. To address these challenges, Fiber-Reinforced Polymer (FRP) composites have emerged as a popular and efficient method for strengthening concrete structures. Due to their light weight, high strength, durability against corrosion, and ease of installation, FRP composites are considered an efficient and reliable option for strengthening both reinforced and prestressed concrete elements. However, despite these advantages, FRP systems are thermally sensitive. At elevated temperatures, the polymer matrix and adhesive bonds deteriorate rapidly, leading to a significant reduction in strength, stiffness, and bond capacity. This vulnerability limits the performance of FRP-strengthened members under fire conditions. As a result, the application of FRP systems in prestressed concrete beams remains limited due to the lack of comprehensive design guidelines and experimental data related to fire exposure.To address this limitation, a rational approach is proposed for evaluating the fire resistance of FRP-strengthened prestressed concrete (PC) beams. This approach expands on conventional fire design principles for PC beams, while incorporating the effects of FRP reinforcement and fire insulation into strength calculations under fire exposure. Simplified equations are utilized to evaluate the cross-sectional temperature distribution in fire exposed FRP-strengthened PC beams, considering both uninsulated and insulated scenarios. These cross-sectional temperature profiles are then utilized to evaluate the reduction in strength of concrete, prestressing steel, and FRP based on their temperature-dependent mechanical properties. The moment capacity of the FRP-strengthened PC beams is determined at various fire exposure durations by applying force equilibrium and strain compatibility principles. The proposed approach is validated by comparing analysis results with available fire test data from FRP-strengthened reinforced concrete (RC) beams. The results show that, without supplementary fire insulation, FRP-strengthened PC beams experience a significant reduction in moment capacity early into fire exposure and can experience failure in 75 min. In contrast, with adequate fire insulation, these beams retain a substantial portion of their load-bearing capacity for up to 3 hours during fire exposure.Description based on online resource. Title from PDF t.p. (Michigan State University Fedora Repository, viewed ).Includes bibliographical references