21090 research outputs found
Sort by
Development of Facilitated Transport Membranes with Metal-Chelating and Hydrogel-Like Properties for Efficient Olefin/Paraffin Separation
No full textOlefins are essential feedstocks for producing a wide range of polymers and chemicals, and their separation from associated paraffins is crucial to obtaining high-purity olefins (>99.5%). However, due to their similar molecular weights and volatilities, olefin/paraffin separation remains a significant challenge. Distillation, the conventional method for this process, is both capital- and energy-intensive. Therefore, the development of alternative technologies that can be effective, sustainable, and energy-efficient for olefin/paraffin separation is a focus. Facilitated transport membranes (FTMs) offer a promising alternative by employing metal ion carriers, typically Ag⁺, to chemically facilitate olefin transport. This work developed a series of FTMs using polymeric matrices with both metal-chelating and hydrogel-like properties, achieving high olefin permeability, selectivity, and stability for olefin/paraffin separation.
The first study focused on FTMs based on chitosan and silver nitrate. A large amount of silver was incorporated into the membrane by simply immersing a pre-formed chitosan membrane into an aqueous silver nitrate solution. Through sorption and diffusion, Ag⁺ ions and water were effectively loaded into the membrane, facilitated by the abundant amine groups in chitosan and their chelating interactions with silver ions. The membrane's high-water uptake created an ideal microenvironment for olefin-silver complexation, as well as the migration of both the complexes and silver ions. This study highlights the crucial role of both Ag⁺ and water loading in achieving optimal facilitated olefin transport. To further enhance water retention and membrane performance, a modification using dilute citric acid treatment was proposed. This approach contributes to preserving a small fraction of protonated amine groups, thereby improving the membrane’s water retention capacity and ultimately enhancing its separation efficiency.
Next, a poly(vinyl alcohol) (PVA)/poly(vinyl amine) (PVAm)-based membrane was identified as a promising candidate for facilitated olefin transport. These two linear polymers were interpenetrated into a water-insoluble chitosan framework, forming an interpenetrating network (IPN) that achieves membrane insolubility without conventional crosslinking. This approach maximizes the availability of abundant amine and hydroxyl groups within the IPN, enhancing chelating interactions with Ag⁺ and enabling higher Ag⁺ loading compared to crosslinked membranes. The IPN structure allows for a water uptake of up to 2.32 g/g-polymer while maintaining sufficient mechanical strength for facilitated olefin transport, which involves the migration of complexes and silver ions within the membrane.
To further optimize the membrane structure, a PVA/PVAm composite FTM with a gradient structure was developed using vapor-solid interfacial crosslinking. Beneath the highly crosslinked outer surface, abundant hydroxyl and primary amine groups were retained to facilitate chelation-based Ag⁺ loading, while the enhanced polymer chain mobility in the interior provided additional free volume for Ag⁺ and water loading. The ultrathin crosslinked surface, formed through interfacial crosslinking, acted as an effective barrier to paraffin molecules while maintaining permeability for olefins, leading to high perm-selectivity in olefin/paraffin separation.
Another potential membrane matrix, derived from natural waste cocoons, was investigated for olefin/paraffin separation. Pristine fibroin FTMs exhibited limited performance due to low silver salt and water uptake, attributed to their rigid structure and low swelling capacity. Moreover, their brittleness and mechanical instability prevented them from withstanding the stress induced by Ag⁺ bonding at high concentrations. To address these limitations, fibroin and sericin were blended with chitosan. Both membranes demonstrated increased Ag⁺ and water loading capacities, better film-forming properties, and enhanced olefin permeability and olefin/paraffin selectivity. The β-sheet structure of fibroin in the chitosan/fibroin-Ag⁺ membrane provided greater structural rigidity, reducing paraffin permeability and mitigating competitive effects during mixed-gas permeation, ultimately leading to higher olefin/paraffin selectivity.
To further investigate the gas permeation behavior and transport mechanisms in the water-swollen FTMs, the permeability, solubility, and diffusivity of olefins and paraffins were analyzed in four different FTMs. The study demonstrated that the permeability of the FTMs was significantly influenced by Ag⁺ and water content. Sorption tests showed that olefin solubility increased with pressure, deviating from Henry's law due to the combined effects of Ag⁺ complexation and gas condensability. Diffusivity calculations indicated that paraffins generally had higher diffusivity than olefins, primarily governed by molecular size
Development of Microwave-Microfluidic Sensors for Microplastic Detection in Environmental Samples
No full textMicroplastics (MPs), plastic particles smaller than 5~mm, are emerging as a significant environmental threat due to their widespread presence in ecosystems and potential health impacts. They originate from both primary sources, such as microbeads in personal care products, and secondary sources, like the degradation of larger plastics. MPs can accumulate in aquatic life, pose risks to food chains, and carry toxic pollutants. Despite their environmental significance, detecting MPs in natural settings is challenging due to complex particle characteristics and the limitations of current detection methods.
Several well-established methods have been developed for detecting and monitoring MPs in aqueous samples. Fourier-transform infrared and Raman spectroscopy are among the most widely used techniques due to their unique ability to identify chemical compositions at the molecular level. However, these methods generally require bulky, expensive equipment and skilled personnel. Additionally, they are offline techniques that involve time-consuming and labor-intensive sampling processes. As a result, there is growing demand for affordable and user-friendly MP sensing techniques suitable for on-site applications.
Electrical sensing methods-including resonance microwave spectroscopy, dielectric spectroscopy, high-frequency impedance spectroscopy, and electrical impedance spectroscopy-offer unique advantages for on-site detection of MPs due to their compact detection systems and scalability for multi-location testing. Each method interacts differently with the electrical properties of the material, offering diverse capabilities for MP detection.
Although most electrical sensing methods share similar working principles, resonance microwave spectroscopy stands out as a promising solution due to its broader frequency range (typically 0.1-100 GHz), enabling more versatile and precise detection of various particle types. Microwave sensing differentiates materials based on their permittivity, making it highly sensitive for detecting MPs, which typically exhibit permittivity values (∼2-3.5) distinct from their natural surrounding materials, such as water (∼80), wet sediments (∼10-30), and blood (∼50-60). Furthermore, microwave sensors can be integrated with planar technologies, such as printed circuit boards (PCBs) and microstrip antennas, to create compact, lightweight, durable, and cost-effective systems, offering a practical solution for continuous measurements in real-world applications.
This thesis presents the development of microwave-microfluidic sensors for detecting and characterizing MPs in aqueous environmental samples, offering a scalable and cost-effective solution for real-time monitoring. It starts with an exploratory study capable of only concentration monitoring and progresses to an enhanced sensing platform capable of monitoring both size and concentration. Then, continuous flow is added to the sensing platform to enable single-particle monitoring, which leads to MP size, type, and concentration characterization. In the final stage, the application of the microwave-microfluidic sensor extends from environmental to biomedical contexts.
The thesis begins by exploring the integration of microwave sensing with microfluidic platforms for detecting MPs in water. Experimental investigations were conducted using polyethylene microspheres of two different sizes (20 μm and 70 μm). The results indicate that the resonance frequency shift depends on particle size, concentration, and temperature. While experimental trends largely align with numerical simulations, the observed shifts were less pronounced than predicted, and the detection limits were higher than MP concentrations typically encountered in freshwater environments. These findings highlight the need for improved sensitivity and expanded applicability.
Building on this, a sensitivity-enhanced microwave sensing platform was introduced using coupled planar microwave resonators to characterize both the size and concentration of MPs in real time. The design incorporates an interdigital capacitor (IDC) structure with a traditional split-ring resonator (SRR) to enhance sensitivity. A disposable sample holder enables multiplex testing without cross-contamination, making the system field-deployable. The sensor was optimized through simulation and validated experimentally with MPs of three sizes (20 μm, 70 μm, and 275 μm) at various concentrations (100k, 1000k, and 10,000k particles/L). The results confirmed the sensor's ability to monitor particle size and concentration accurately. However, since experiments were conducted with fixed sample volumes in the microliter scale, continuous flow integration was needed to improve statistical robustness.
The next project introduces an innovative AI-powered microwave-microfluidic platform that enables comprehensive analysis of MPs. The system analyzes particle size, concentration, and type using a K-nearest neighbors (KNN) algorithm trained on raw sensor data. Environmental samples are prefiltered into specific size ranges, and single-particle detection enables precise quantification of MP concentration. This approach offers a scalable solution for real-time monitoring of MP contamination across a wide size range (20-300 μm).
The final project demonstrates the potential of the microwave-microfluidic sensor in biomedical diagnostics. The platform was adapted to detect biomarkers in biological samples, focusing on monitoring amylase levels in postoperative peritoneal drainage fluid as an indicator of anastomotic leakage. This work highlights the broader utility of the sensor system for cost-effective, non-invasive real-time monitoring in both environmental and clinical settings
Evaluation of Information Access Systems in the Generative Era
No full textThe rapid advancement of information access technologies, including neural retrieval models and generative information-seeking systems, has outpaced traditional evaluation methodologies, exposing fundamental gaps in assessing their effectiveness. Existing evaluation frameworks struggle to adapt, particularly in the presence of sparse relevance labels, limiting their ability to fairly and comprehensively compare retrieval and generation-based systems. The emergence of large language models (LLMs) further complicates evaluation, as they challenge conventional assessment paradigms while offering new opportunities for automated evaluation. To address these issues, it is crucial to first identify flaws in current evaluation methodologies and then develop more robust, efficient, and adaptable assessment strategies.
This thesis begins by demonstrating that evaluation based on sparse labeling introduces substantial biases and inconsistencies in system rankings, often failing to recognize genuine improvements in retrieval effectiveness. We show that in traditional IR benchmarks,stronger models may retrieve highly relevant but unjudged documents, leading to underestimation of their performance. To mitigate this, we propose an alternative evaluation approach based on distribution of retrieved results and labeled data using Fréchet Distance. This method not only improves robustness in the presence of sparse labels but also facilitates direct comparison between retrieval-based and generative models on a common evaluation scale.
We then investigate how LLMs can be leveraged to evaluate IR systems, distinguishing between their use for evaluating retrieval-based methods and generative IR systems. A key focus of this work is the role of LLMs in automated relevance judgments. We systematically compare different LLM-based relevance assessment methodologies, highlighting the lack of standardization in evaluating these approaches. To address this gap, we propose a structured framework that evaluates relevance judgment methods based on their alignment with human labels and their impact on system rankings. Furthermore, we examine the effect of prompt formulation on LLM-based evaluation, demonstrating how prompt variations can significantly influence the consistency and reliability of assessment outcomes.
Finally, we extend our study beyond retrieval-based evaluation to assessing generated content across multiple applications. We explore retrieval-assisted methods for evaluating generative textual content, IR-inspired approaches for assessing text-to-image generation models, and a broader framework for evaluating LLM-powered applications. These contributions lay the foundation for a new generation of evaluation methodologies that keep pace with evolving information access technologies, ensuring that improvements in retrieval and generative AI systems can be accurately and meaningfully assessed
Pulsed Laser micro welding of Si to Cu for die attachment
No full textA die attach material joins a semiconductor die (typically Si) to a metallic substrate (typically
Cu) for mechanical support and heat dissipation, in microelectronic circuits. In aerospace
applications, electrical vehicles (EV), there is increasing demand for die attach technologies
that can withstand operating temperatures in excess of 200 0C and sustain high mechanical
stresses. Use of lead-based solders is prohibited as per regulations. Implementation of other
methods of die attachment such as soldering, eutectic bonding, sintering and epoxy bonding,
encounter challenges related to die attachment failure on multiple fronts. This thesis studies the
feasibility of high power short pulse laser spot micro-welding process for die attach alternative.
Practical concern to reduce energy requirements leads to the enhancement in absorptivity of
Cu substrate by application of black marking on its surface exposed to laser. This also reduces
the possibility of thermal damage to the joint. The detailed laser parameters, specifically laser
peak power and pulse duration, required for damage-free joining, are discussed in Chapter 3,
which highlights the effectiveness of the black marking and Ag interlayer in reducing the peak
power for joining. Laser spot welds with high peak power and short pulse duration prove to be
ideal for micro-joining. The corresponding temperature characteristics and the microstructure
evolution have been discussed in Chapter 4. While maximum temperatures attained indicate
the potential for phase change, cooling rates can influence the mechanical properties of the
joint.
Some trials were conducted to evaluate the shear load for failure of the bonds, as specified by
MIL-STD-833, for various spot-welding patterns and reported in Chapter 5. This test
successfully proves that the utilization of a laser for spot welding is feasible for obtaining a
joint with satisfactory mechanical load-bearing capacity
High-Temperature Metamorphic Reactions from the Macro-Scale to the Micro-Scale
No full textMetamorphic reactions are the basis of metamorphic petrology and the lens through which we interpret metamorphic rocks and processes. They serve both as tectonic indicators, revealing the pressure--temperature history of a rock, and as tectonic drivers, responsible for the production of fluids and melts that are critical to many geological processes in the crust. At high temperatures, two types of reactions that have emphasized importance are melting reactions that have implications for large-scale crustal reworking and reactions that grow accessory minerals that we use to date petrological processes.
Zircon is the most common geochronometer, but its behaviour at high temperatures is poorly understood. Zircon-forming reactions were investigated in granulite-facies meta-granitoids in the Grenville province to better understand how zircon grows during metamorphism at high temperatures. Zircon growth occurred during retrogression as a result of melt crystallization and titanomagnetite breakdown. With this information, the dates of metamorphic zircon that were measured were interpreted as cooling dates, and provided additional context that suggests that the major orogenic phase of Grenville Orogen may have begun tens of millions of years earlier than previously thought.
Zircon was also used as a proxy to investigate the kinetics of trace elements in intergranular melt during melt crystallization in a migmatite. Key trace elements including Hf, U, Th, Y, and heavy rare earth elements were analyzed in multiple metamorphic zircon rims to compare relative concentration of zircon that grew coevally in the same thin section. The significant differences observed in concentration of these elements across zircon grains suggests rates of diffusion of these key trace elements are slower than zircon growth in migmatites. Zircon growth probably occurred as a result of size-dependent interface-controlled growth, implying that Zr diffusion was relatively fast in the melt.
On the macro-scale, evidence of regionally extensive H2O-fluxed melting reactions have been observed in multiple distinct tectonic environments across the globe, yet there is no generic tectonic model that explains regional-scale H2O-fluxed melting in the crust. Regional scale H2O-fluxed melting was studied in the Muskoka domain of the Grenville province. In the Muskoka domain, H2O-fluxed melting dominated throughout the region and until now, the source and mechanism of the H2O transport into the Muskoka domain has been unclear. Multiple examples of pegmatites with amphibole and leucosome-rich reaction selvages were found throughout the domain that show how H2O may have been transported into and through the Muskoka domain. Using a two-stage melting model, it was shown that melt generated at depth through hydrate-breakdown melting contains enough H2O to readily melt the rocks in the Muskoka domain through diffusive H2O-fluxed melting, with no fluid exsolution required.
Metamorphic reactions are used to understand regional tectonics, but there are significant gaps in our understanding of these reactions on both the micro-scale and the macro-scale. The geochronological tools that are used to unravel metamorphism are based on micro-scale processes that are still poorly understood. Simultaneously, our understanding of macro-scale tectonic processes involving H2O transport in the crust, which influence our interpretation of metamorphic rocks, is limited with H2O-fluxed melting. This thesis addresses our limits of understanding and shows how understanding metamorphic reactions allows us to better understand regional tectonics
Fine-Grained Visual Entity Linking through Promptable Segmentation: Applications in Medical Imaging
No full textImage analysis in domains that produce large amounts of complex visual data, like medicine, is challenging due to time and labour-constraints on domain experts. Visual entity linking (VEL) is a preliminary image processing task which links regions of interest (RoIs) to known entities in structured knowledge bases (KBs), thereby using knowledge to scaffold image understanding. We study a targeted VEL problem in which a specific user-highlighted RoI within the image is used to query a textual KB for information about the RoI, which can support downstream tasks such as similar case retrieval and question answering. For example, a doctor reviewing an MRI scan may wish to obtain images with similar presentations of a medically relevant RoI, such as a brain tumor, for comparison. By linking this RoI to its corresponding KB document, search of an imaging database with VEL-guided automatically-generated tags can be performed in a knowledge-aware manner based on exact or semantically similar entity tag matching.
Cross-modal embedding models like CLIP present straightforward solutions through the dual encoding of KB entries and either whole images or cropped RoIs, which can then be matched by a vector similarity search between these respective learned representations. However, using the whole image as the query may retrieve KB entries related to other aspects of the image besides the RoI; at the same time, using the RoI alone as the query ignores context, which is critical for recognizing and linking complex entities such as those found in medical images. To address these shortcomings, this thesis proposes VELCRO—visual entity linking with contrastive RoI alignment—which adapts an image segmentation model to VEL using contrastive learning by aligning the contextual embeddings produced by its decoder with the KB. This strategy preserves the information contained in the surrounding image while focusing KB alignment specifically on the RoI. To accomplish this, VELCRO performs segmentation and contrastive alignment in one end-to-end model via a novel loss function that combines the two objectives. Experimental results on medical VEL show that VELCRO achieves an overall linking accuracy of 95.2% compared to 83.9% for baseline approaches
Extensions of the Tutte Polynomial and Results on the Interlace Polynomial
No full textIn graph theory, graph polynomials are an important tool to encode information from a graph. The Tutte polynomial, first introduced in 1947, is one of the most important graph polynomials due to its universality. Here, we present three classic definitions of the Tutte polynomial via a deletion-contraction recursion, via rank and nullity, and via activities. We will touch on the significance of this polynomial to the field of mathematics to motivate an extension to signed graphs. Extending the polynomial to retain deletion contraction and inactivity information, we introduce an extended Tutte polynomial to allow for the construction of a Tutte like polynomial on signed graphs. Using the extended information, we examine the monomials of these polynomials as grid walks. Using grid walking and the extended Tutte polynomial, we investigate the relationship between the Tutte polynomial of a graph and that of its bipartite representation. This is done with a view toward the construction of a Tutte like polynomial for oriented hypergraphs.
While many graph polynomials are directly related to the Tutte polynomial, there are also a wide variety of polynomials related in special cases only. One such polynomial is the Martin polynomial and, related to it, the interlace polynomial. Here, we discuss how these two polynomials are related and how results on the Martin polynomial can be extended to the interlace polynomial. The Martin invariant, a specific evaluation of the Martin polynomial, obeys the symmetries of the Feynman period. The Feynman period of a graph is useful in quantum field theory, but difficult to compute and thus there is interest in finding graph invariants that have the same symmetries. It was established that the interlace polynomial on interlace graphs was equal to the Martin polynomial on the associated 4-regular graph. While only graphs that do not contain a set of forbidden vertex minors are interlace graphs, the interlace polynomial is defined over all graphs. We discuss how this provides a way to try and extend the notion of Feynman symmetries via the interlace polynomial and some specific classes of graphs with formulas. Additionally, the interlace polynomial is only equal to the Martin polynomial for interlace graphs of 4-regular graphs, but the Martin polynomial is defined for 2k-regular graphs. Thus, we work toward creating an interlace-like polynomial for graphs derived from 2k-regular cases of the Martin polynomial
Modeling and Experimental Studies of Electrolyte for Zinc Battery Systems
No full textZinc-based batteries have emerged as promising alternatives to lithium-ion systems due to their low cost, inherent safety, and sustainability. However, challenges in electrolyte stability, dendrite formation, and limited lifetime have constrained their practical deployment. Addressing these barriers requires both experimental innovation and data-driven approaches for electrolyte and materials design.
This thesis aims to advance zinc battery technologies through four interconnected directions: (i) development of novel electrolyte systems, (ii) optimization of zinc deposition and cycling performance, (iii) machine learning methods for early-stage lifetime prediction, and (iv) accelerated discovery of functional ionic liquids.
Molten salt-derived Zn(TFSI)2 electrolytes were investigated for zinc–oxygen batteries across a wide temperature range. Electrochemical, spectroscopic, and microscopy analyses revealed the structural evolution of zinc oxide nanosheets during cycling and highlighted that Zn(TFSI)2·8H2O suppresses parasitic reactions more effectively than Zn(TFSI)2·18H2O, enabling improved reversibility and energy storage potential. Complementary studies on hydrated ZnCl2 electrolytes demonstrated how temperature and hydration level influence zinc nucleation, morphology, and cycling stability. ZnCl2·10H2O achieved 99.2% coulombic efficiency over 50 cycles, while higher operating temperatures increased discharge capacity from 17 mAh g-1 at -10 °C to 72 mAh g-1 at 40 °C.
Artificial intelligence approaches were developed to classify and predict battery lifetime from early cycling data. Machine-learned models achieved up to 96% accuracy after only two cycles and 98% with additional data, while human-selected electrochemical features showed strong generalizability across chemistries. Deep learning methods reached 99.5% accuracy with extended cycling data but proved less transferable to systems with distinct degradation profiles. In parallel, convolutional and generative adversarial neural networks were applied to accelerate the discovery of ionic liquids for zinc batteries. These models improved property prediction and successfully generated new candidate electrolytes with enhanced performance at room temperature.
Overall, this work provides an integrated experimental–computational framework for electrolyte optimization, electrochemical performance improvement, and AI-driven materials discovery. The findings pave the way toward more durable, efficient, and scalable zinc-based energy storage technologies
From Spin Vorticity Models to Spin Liquids on the Octochlore Lattice
No full textNearest-neighbour spin ice has been central to the study of frustrated magnetism for nearly three decades, providing a framework that reveals emergent gauge fields and monopole excitations within geometrically frustrated spins on the pyrochlore lattice. The geometry of corner-sharing tetrahedra admits only a single symmetry-equivalent nearest-neighbour bond, strongly constraining the range of allowed interactions.
Recently, a new frustrated lattice of corner sharing octahedra, dubbed the octochlore lattice, has emerged as a promising platform for novel spin liquid phases. Unlike the pyrochlore, the octahedra permit distinct intra-octahedral interactions, greatly expanding the variety of realizable models. Building on the work of Szabó et al., where the spin-ice analogue was studied in a restricted region of parameter space, this thesis pursues two complementary directions. First, we investigate the spin vorticity model, in which the monopole excitations of spin ice are replaced with string-like excitations analogous to closed current loops. Second, we identify all the long-range ordered phases at the second nearest-neighbour level, fully elucidating the intra-octahedra model of Szabó et al. through an irreducible representation analysis. In doing so, we discover a novel classical U(1) analog to the celebrated X-cube model of fracton topological order.
Overall, this work demonstrates that the octochlore lattice of corner-sharing octahedra constitutes a next-generation platform for three dimensional frustrated magnetism, uniquely capable of hosting exotic spin liquid phases with potential realizations in rare-earth based antiperovskites and potassium rare-earth fluorides
Anuran Habitat Associations and Minimums: Identification, Application, and Implications
No full textAnurans (frogs and toads) are model species for habitat conservation in disturbed areas due to their reliance on a range of land covers and sensitivity to fragmentation. As a group of species that are reliant on a landscape permeability to access different habitat throughout the year, their presence across the landscape can provide valuable information about minimum habitat amounts and distributions necessary to maintain habitat connectivity for a range of species. In Ontario, historical landscape disturbance and the availability of a long-term anuran monitoring database provides a study area suitable for exploring critical landscape characteristics and thresholds related to anuran species presence. Like many jurisdictions in North America, natural heritage planning in Ontario has adopted minimum habitat disturbance and buffer recommendations but low anuran occupancy in disturbed areas suggests they are insufficient to ensure long-term species presence. Requirements related to habitat connectivity in particular lack detail regarding the amount, type, placement, and permeability of land covers across the landscape. To address these fundamental gaps, this thesis identified ecological thresholds amongst several habitat connectivity and landscape composition metrics designed to model anuran habitat usage in Southern Ontario. It also discusses the land use implications of implementing biologically informed and ecologically meaningful minimum habitat protections for maintaining anuran habitat connectivity.
Data for anuran occupancy was derived from acoustic survey data in the Birds Canada amphibian monitoring database. Due to regular species turnover and challenges with acoustic detection, the number of monitoring years required to account for all species present in a breeding wetland is uncertain. Using a sample of 66 wetlands in Southern Ontario with at least eight years of acoustic monitoring data and four identified anuran species, I constructed species accumulation curves and determined that a minimum of three years of standardized acoustic monitoring data area is required to capture a complete picture of anuran species composition. With this result, 290 wetlands across Southern Ontario with anuran acoustic data of six commonly occurring species were determined to be suitable for further analysis.
In addition to several commonly used landscape composition metrics used in amphibian habitat modelling (e.g., proportion of forest/wetland, urban, and vegetated land cover), several other metrics related specifically to anuran habitat connectivity were designed and compared to anuran presence. Using Generalized Additive Models (GAMs), each metrics’ explanatory strength of anuran species presence was compared. I detail the construction and performance of one particular metric that explicitly models the isolated impact of vehicle impact on road crossing survival; accounting for the exponential reductions in crossing survival with increasing traffic intensity in a way that is not captured by least cost, cost distance, or circuit theory modelling approaches. This metric was compared against an unaltered cost distance tool, with three maximum movement distances and three minimum survival cut-offs of 50%, 25%, and 5%. The top-performing road crossing survival metrics explained up to 22% deviance, performed best for the spring peeper and gray treefrog, and performed similarly to the top-performing unaltered resistance models. A model that accurately captures the dynamics of road mortality for amphibians and which can be applied to other taxa is important for effective conservation and land use planning.
I follow this with an expanded detailing of the construction of additional habitat connectivity metrics modelling juvenile anuran dispersal and overwintering habitat connectivity, exploring all possible combinations of explanatory metrics to identify additive habitat influences. These metrics were designed to address gaps in connectivity modelling methods which have not accounted for anuran movement behaviours. The juvenile dispersal metric divides the landscape around a source wetland into discrete wedges to simulate the auto-correlated and laterally restricted movement behaviours exhibited by juvenile dispersing anurans and outputs a measure of connectivity that reflects the proportion of directions that have reachable breeding habitat. The overwintering connectivity metric took a more traditional approach, using typical resistance modelling to output a metric quantifying the accessibility of overwintering forest habitat surrounding a breeding wetland. Both metrics used anuran occupancy as a response variable. The explanatory strength of landscape composition and habitat connectivity metric varied substantially between the six individual species, indicating notable differences in life history processes, landscape interactions, and species monitoring. Landscape composition metrics consistently explained the highest amount of variance in single-variable models regardless of species, explaining as much as 37% of deviance in anuran species occurrence using proportion of forest and wetland cover within 3000m. Habitat connectivity metrics related to road crossing survival, juvenile dispersal, and overwintering habitat performed best for the spring peeper and gray treefrog, explaining as much as 26% of variance in occupancy for spring peepers using the overwintering habitat connectivity metric. Various areas for improvement and discussions of the implications of these results are included. The performance of landscape composition and connectivity metrics at different maximum movement distances varied, indicating species-specific differences in the effects of habitat composition, distribution, and permeability across difference spatial scales. In multi-variable GAMs, the top performing models were again for the spring peeper and gray treefrog, with up to 48% of variance in gray treefrog occupancy explained. Habitat connectivity metrics were often included in the top-performing models, suggesting that they contribute additional finer-detail significant explanatory strength to anuran population distributions and with further refinement can be valuable tools for landscape planning.
Critical thresholds in landscape composition and habitat connectivity were identified across all six species and spatial scales using a segmented regression approach. Significant thresholds were most commonly identified in the landscape composition metrics and strongest with the natural cover composition. For the spring peeper and gray treefrog which had the strongest threshold-type relationships, thresholds at ~40-60% natural cover, ~5-35% of forest and wetland cover, and ~12-55% of urban cover were identified, varying depending on radial distance. Visual breakpoint estimation differed from statistical breakpoint estimation occasionally, emphasizing the importance of critical interpretations of statistical results. Threshold presence was also moderately consistent with non-linear relationships identified in the GAMs, suggesting that species-landscape relationships require critical examination when determining biological recommendations. The identification of thresholds amongst the various habitat and connectivity metrics refines our understanding of minimum biological requirements for anurans and provides additional evidence for science-based conservation efforts.
In the concluding of this thesis, I discuss the financial efficiencies related to infrastructure, health, and planning, as well as societal improvements to quality of life for implementing biologically informed minimum habitat protections for anurans and other species reliant on landscape permeability. Identifying and planning around existing wildlife movement corridors and broader habitat thresholds associated with landscape composition and permeability can encourage planning bodies to produce a denser, more efficient, and more equitable landscape for both humans and the wildlife populations living alongside them