82837 research outputs found
Sort by
Centrifuge Modeling and Numerical Simulations of Axial Behavior of Helical Piles and Groups in Sand
No full textHelical piles are an innovative type of deep foundation offering advantages over conventional piles. However, their design remains limited by manufacturer guidelines and simplified applications. Key knowledge gaps exist in axial load transfer and soil failure mechanisms of multi-helix piles under axial tension, compression, and repeated axial cycles in sand. Moreover, despite its growing use in battered group configurations, information on the vertical load and bending moment transfer behavior of batter helical piles are unknown. Centrifuge modeling and numerical simulations provide a cost-effective approach to investigating these complex behaviors as an alternative to costly field testing. This research was conducted in sequential phases to examine the axial load transfers within helical piles and groups in sand. The study commenced with nine centrifuge tests conducted at 20-g accelerations on straight-shaft, single-helix, and double-helix piles with inter-helix spacing ratios of 1.5, 2.5 and 3.5 under both axial compression and tension. A helical pile provided about 2.5 times greater capacity than a straight-shaft pile in compression or tension. The shaft resistance was highest near the pile tip and between helices, with the lower helix of double-helix piles carrying up to 2.9 times more load in compression and 1.7 times more in tension than the upper straight-shaft portions of the pile. The end bearing stress for the lower helix was about 2.5 times the upper helix in double helix piles. The study then examined the effects of repeated one-way incremental axial cycles on helical piles by conducting five centrifuge tests at 20-g accelerations in sand. The results revealed that an increase in axial cyclic capacity occurred with increasing cycle amplitude. Repeated cycles resulted in a gradual reduction in capacity, with transition and softening zones developing along the depth. Post-cyclic monotonic capacity was significantly lower than pre-cyclic capacities.
In the third phase, centrifuge tests were conducted at 15-g accelerations using straight-shaft and double-helix piles on pile groups subjected to vertical loading with batter angles of 0°, 10°, and 20°, and spacing ratios of 2 and 3 times the diameter of the helix. Battered double-helix piles supported up to 4.5 times more load than straight-shaft piles. However, larger batter angles and pile spacing reduced capacity by up to 33% and 57%, respectively. The measured bending moments were greater along the diagonal direction of the piles than in normal to the diagonal.
To complement the experiments, three-dimensional (3D) coupled Eulerian-Lagrangian finite element modeling was used to simulate true helix geometry under axial compression in sand. Models of straight-shaft, single-, and double-helix piles with inter-helix spacing ratios of 1.5 and 2.5 were validated against the centrifuge data. The separation of the loads between helices and shafts provides higher helix loads in single-helix and lower helix loads in double-helix piles. Multi-helix piles failed in an individual bearing mode regardless of varied inter-helix spacings.
Finally, 3D standard Lagrangian finite element (FE) analyses were conducted on battered straight-shaft and helical pile groups in sand subjected to vertical loading. Validated FE models confirmed that battered double-helix piles could resist up to five times more load than straight-shaft piles. Interestingly, the upper helix carried 30% more load than the lower under battering, reversing the trend observed in vertical loading. Though the individual bearing mode remained dominant, battering introduced uneven stress around the helices. The research highlights the significance of centrifuge and numerical modeling techniques in understanding helical pile-sand interactions, the internal load transfer mechanism and axial pile failures under various axial loading scenarios, and different pile configurations. The results provide practical design charts that consider pile types, pile spacings, loading conditions, and pile batter angles
Improving EMS Response Times through Threshold-Based Ambulance Repositioning: A Simulation and Genetic Algorithm Framework
No full textTimely ambulance response in life-threatening emergencies can mean the difference between life and death. The effectiveness of Emergency Medical Services (EMS) is commonly measured by response time, defined as the interval between receiving a call and the arrival of an ambulance. Optimizing deployment is therefore critical to ensure timely responses and adequate geographic coverage, particularly during periods of resource scarcity. This thesis presents a simulation–optimization framework that improves EMS performance by integrating dynamic ambulance repositioning strategies with operational thresholds—predefined ambulance availability levels that trigger action.
We develop a discrete-event simulation model that captures spatial and temporal demand, realistic travel times, and operational constraints. The framework evaluates both reactive and selective repositioning strategies. To navigate the large space of possible repositioning configurations, we combine simulation with a Genetic Algorithm that efficiently searches for high-quality repositioning tables—rules that specify where ambulances should be relocated based on availability and demand.
Results show that Genetic Algorithm–generated repositioning tables substantially improve EMS performance. Median response times decrease, lost calls are reduced, and the duration of periods without an available ambulance shortens. These benefits are most pronounced under high-load conditions. Threshold selection proves critical: thresholds set too high lead to excessive ambulance movements and crew fatigue, while those set too low delay intervention and reduce coverage. The optimal threshold range identified here balances responsiveness with operational sustainability.
This research provides practical, data-driven guidance for EMS planners seeking to enhance resilience during capacity shortages. The proposed framework is broadly applicable to other urban EMS systems and demonstrates how simulation, evolutionary optimization, and real-world constraints can be combined to improve both system performance and crew well-being
The Militarization of Urban Public Housing in Postwar Sri Lanka
No full textThis dissertation, situated at the intersection of urban planning, housing, and governance scholarship, examines the discourse of urban public housing in postwar Sri Lanka. Through three stand-alone, distinct, yet interconnected academic papers, the research investigates how the postwar Sri Lankan government strategically leveraged military capacities in Colombo, the country’s capital , for its world-class, slum-free city-making initiative, consequently displacing low-income, underserved urban communities. The research reveals the political and elite motives that significantly shaped the redefinition of the country’s urban policy during this period, clarifying the political objectives of consolidating power to establish an authoritarian government with military support, as well as the subsequent impacts of these trends on the public housing approaches of underserved communities. By employing three interconnected discourse analyses—scholarly and policy discourse, elite and institutional discourse, and the lived experiences of communities—the research provides the rationales and implications of this elite-driven, military-centric urban development approach, highlighting the significant practical and ethical consequences of this strategy on the public housing expectations of low-income, underserved communities
Integrating Taxonomic and Trait-Based Approaches to Evaluate Beta Diversity of Freshwater Invertebrates as Bioindicators of Environmental Change in Western Canada
No full textLakes and streams are among the ecosystems most impacted by recent global change. Aquatic invertebrate communities have long been used as indicators (i.e., bioindicators) of anthropogenic and natural environmental changes. Morphologically defined species have traditionally been used to assess aquatic invertebrates as bioindicators. Later (i.e., 1970s), a species trait-based approach has been recommended in which features of species that define their ecological roles are used to translate taxonomic changes into potential impacts on ecosystem function. Major knowledge gaps exist about how complementary versus redundant these two approaches are across different types of aquatic communities, ecosystems, and environmental changes. My thesis research combines biomonitoring approaches, determining taxonomic and functional turnover (i.e., beta diversity) of mountain zooplankton and stream macroinvertebrate communities to gain insights into ecological factors and potential consequences for ecosystem function. Multivariate data analyses ranging from indirect to direct gradient analyses quantify and illustrate temporal and spatial beta diversities related to environmental change. My analyses of zooplankton communities in naturally fishless alpine lakes stocked with sportfish show that a shift in trait selection from initial tolerance of predation (e.g., body size) to subsequent potential for recolonization (e.g., asexual reproduction) explains their contrasting responses to fish introductions and later removal over several decades. At a broader landscape scale, my analyses indicate that climatic and sportfish variables mainly explain the spatial beta-diversities of zooplankton communities across 85 mountain lakes. However, these drivers are not closely related to the temporal beta-diversities observed within the lakes over the past 60 years. In contrast, spatial beta-diversity of stream macroinvertebrate communities and their traits across tributaries spanning the North Saskatchewan River watershed within Alberta best indicated shifts in human land uses against the backdrop of a natural biogeographical gradient. In conclusion, the high degree of redundancy observed between the taxonomic and trait-based beta diversities of each of these communities allowed for confident interpretation of the results, providing ecological insights into how they were indicative of environmental changes along both spatial and temporal scales. Future research should use such comparative approaches to focus, where possible, on testing the validity of assumptions and models commonly used in bioindicator investigations
Exploring the Survival Strategies of Aerobic Methanotrophs in Oxygen-Limited Conditions; an Interdisciplinary Approach
No full textAnthropogenic pollution is a critical threat to life on Earth. Methane is a powerful
greenhouse gas, emitted directly to the atmosphere as a polluting byproduct of many
industries. It is also emitted from natural sources, and it is produced by microbes whose
growth is promoted by other anthropogenic pollutants, such as nitrogen fertilizers and the
growing accumulation of human and domestic animal waste. Luckily, nature contains a
group of opportunistic, counterbalancing microorganisms capable of consuming methane
and converting it into biomass or value-added products. These so-called methanotrophs are
the organisms which are featured in this dissertation, and in particular, a sub-group of
aerobic methanotrophs that survive in environments where oxygen is transiently available.
One model organism representing this group is a methanotroph in the class
Gammaproteobacteria, Methylomonas denitrificans FJG1. This bacterium is an obligate
aerobe, requiring oxygen to survive, yet it has recently been discovered to simultaneously
consume methane and nitrate in a combined metabolic pathway which allows the bacterium
to survive for extended durations without access to oxygen. This unique ability also
corresponds with a profound phenotypic change from a cream colour to bright pink when
oxygen is depleted.
The first aim of this thesis was to investigate the molecular responses of M.
denitrificans FJG1 to oxygen limitation and its regulation of the methane and denitrification
pathways. Searching a combined set of transcriptome and proteome data measuring the
gene expression and protein production of M. denitrificans FJG1 collected before, during,
and after oxygen depletion, a strong candidate protein for oxygen binding and delivery was
identified. This gene was found to be highly upregulated in response to decreasing oxygen
availability, and the corresponding protein was produced in very high abundance. The gene,
annotated as bacteriohemerythrin, is a homologue, or copy, of a gene for a pink-coloured
iioxygen-transport protein found in the blood of annelid worms that lack the usual
hemoglobin employed by most animals. In Chapter 2, comparative genomics between M.
denitrificans FJG1, other methanotrophic microorganisms, and other non-methanotrophic
bacteria revealed ten homologues in the genome of M. denitrificans FJG1 alone. One,
designated bhr-00 was specific to methanotrophs in the Methylococcales order. The
predicted structure of this Bhr protein is similar to a previously characterized Bhr-Bath
protein shown to bind and deliver O2 to support methane oxidation in the methanotroph,
Methylococcus capsulatus Bath.
In Chapter 3, this discovery of the methanotroph-specific Bhr-00 protein was
employed as a biomarker, in correlation with the most often used pmoA biomarker, to
identify aerobic methanotroph activity in the anoxic regions of a Canadian Boreal lake (Lake
227) using metatranscriptome data. Aerobic methanotrophs in the order Methylococcales
have been identified in bacterial communities of many anoxic environments, yet the
metabolism that supports their presence and abundance has not yet been solved,
particularly in how they access requisite O2 to support methane oxidation. Both bhr-00 and
pmoA transcripts were found in methanotroph metagenome assembled genomes (MAGs)
reconstructed from Lake 227 and the metatranscriptome indicated they are upregulated in a
manner similar to M. denitrificans FJG1 under oxygen limitation. In addition, genes for gas
vesicles and extracellular electron transport were upregulated in some of the methanotroph
MAGs indicating they have evolved distinctive methods for utilizing bhr in conjunction with
motility and alternative terminal electron acceptors in Lake 227.
In Chapter 4, a gene regulatory network (GRN) was developed based on differential
gene expression of M. denitrificans FJG1 grown under oxygen limitation and two different
nitrogen sources across a six point time course, through the development of a methodology
incorporating unsupervised machine learning (ML) algorithms. This methodology was able to
capture and interpret complex regulatory patterns contained in the gene expression values
iiiof 12 individual sampling points by combining two different ML algorithms (ARACNE and
GENIE3) and retaining only those network connections agreed upon by both algorithms. A
comprehensive workflow was developed to strategically maximize confidence in these
models to interpret the links between genes and their regulators when comparing across
growth conditions. Chapter 5 presents an overall conclusion and future research directions
for further understanding of methanotrophs in anoxic ecosystems and their genomic
regulation as they transition from one physicochemical context to another.
Taken together, the work presented in this dissertation identified a key
methanotroph-specific protein, Bhr-00, that promotes survival of Methylococcales bacteria
in ecosystems with limited oxygen availability. This information connected laboratory
observations from a model methanotroph strain with the activity of related methanotrophs
in a natural lake ecosystem and showed potential mechanisms that allow methanotrophs in
the Methylococcales order to thrive in anoxic zones. Last, a new method for interpreting
gene regulatory networks, called ProGRN, was developed with the potential for increasing
scientific return on transcriptome data. The resulting GRN revealed regulatory connections
between methane oxidizing genes and bhr-00 by way of transcription factors sigma 70 and
sigma 24, shedding new light on the regulatory system of methanotrophs when faced with
oxygen depletion
Applications of Machine Learning to Cheminformatics and Biomedical Informatics
No full textMachine learning (ML) offers transformative capabilities for navigating the complexity of chemical and biomedical data. In this thesis, we present two novel systems called RT-Pred and RADOR, that apply ML to address key challenges in cheminformatics and biomedical informatics, respectively. RT-Pred is a GNN-based web platform we developed to predict liquid chromatography retention times (RT) of chemical compounds under diverse chromatographic methods (CMs), including both reversed-phase liquid chromatography (RPLC) and hydrophilic interaction liquid chromatography (HILIC). The system accepts SMILES strings as input and estimates RTs based on learned patterns across molecular structure and chromatographic conditions. It also incorporates a void/retained/eluted classifier to identify compounds with atypical elution behavior and allows users to train custom models using their own experimental RT data. The RT-Pred system was benchmarked across >44 CMs, achieving an average R² of 0.95, with lower mean absolute errors than existing RT predictors. It also features a searchable database of millions of precomputed RTs and is freely available at www.rtpred.ca. In parallel, we built RADOR (Rare Disease Oracle), a biomedical question-answering system designed to extract and organize unstructured rare disease knowledge from literature. RADOR combines large language models (LLMs) with a curated Neo4j-based knowledge graph (Rare Disease Knowledge Graph - RDKG) containing over 400,000 semantically sorted, source-linked triples that span dozens of rare diseases, genes, symptoms, treatments, epidemiological data and more. To build RDKG, we developed a high-recall triple extraction pipeline (TEP) that performs sentence simplification, named entity recognition (NER), entity typing, and relation normalization via LLMs. This TEP achieves an average recall of 0.97 across ten rare diseases (development set) and 0.96 across a validation set of 3 rare diseases. RADOR employs an agentic or agent-based retrieval-augmented generation (RAG) framework that autonomously interprets user queries, performs multi-hop traversal over the RDKG, and retrieves relevant information in real time from PubMed and other web sources to synthesize citation-backed responses. It also supports multilingual input, synonym expansion, and user feedback integration, allowing it to continuously adapt and improve over time. Furthermore, RADOR dynamically generates external resource links to structured biomedical databases such as HMDB, AlphaFold, and PathBank, enriching its responses with context-specific molecular, structural, and pathway level annotations. Unlike generic LLMs, RADOR grounds every answer to peer-reviewed literature, offering literature or web-based citations and transparent domain-aware outputs. These features, combined with its autonomous reasoning, real-time retrieval, explainability, and structured citation, make RADOR a true oracle for rare diseases, capable of delivering comprehensive and verifiable biomedical insights. RADOR is publicly accessible at www.rador.ca.
Together, these two systems illustrate how tailored ML approaches can address longstanding challenges in both compound identification and rare disease information retrieval. By combining algorithmic precision with deep domain expertise, this work demonstrates the potential to create practical, publicly accessible tools that simplify access to high-quality chemical and clinical knowledge
Avian responses to forest disturbances in the Canadian Rockies
No full textDisturbance regimes in the Canadian Rockies have changed substantially over the past century. Historically frequent, low intensity, patchy fires have been replaced by increasingly large, severe, and stand-replacing events, while the mountain pine beetle (MPB) outbreak has affected millions of hectares of forest since 1999. These shifts—driven by settler colonial fire suppression and anthropogenic climate change—have altered forest structure and composition, with uncertain consequences for wildlife. In particular, it remains unclear how bird communities—widely used indicators of ecosystem change— are responding to novel combinations of disturbance type, severity, extent, and recovery trajectories. My thesis investigates how birds respond to two broad-scale forest disturbances in the Canadian Rockies: high-severity wildfire and post-MPB outbreak management. Using data from autonomous recording units (ARUs), I analyzed avian species richness, community composition, and species-specific responses to disturbance types across different temporal and spatial scales.
In Chapter 2, I assessed bird community responses to the 2017 Kenow Wildfire in Waterton Lakes National Park, using ARU data from 337 locations collected between 2018 and 2022. Elevation was the dominant predictor of avian community structure, but wildfire modified these patterns by amplifying species losses at mid-elevations, where bird communities may be more sensitive to disturbance. At lower elevations, fire effects were less pronounced—possibly due to generalist species or faster vegetation recovery—while high-elevation communities appeared relatively unaffected, potentially due to sparser pre-fire vegetation conditions. Species-level models showed varied responses, with mid-elevation old-growth forest specialists exhibiting the strongest declines.
In Chapter 3, I examined bird communities across post-MPB-attacked forests in two regions that experienced temporally distinct outbreaks: a more recent outbreak in Jasper National Park (Jasper) and an earlier, historic outbreak in the Southern Rockies. Using ARU data collected in 2023 and 2024, I compared bird communities across three post-MPB-attack disturbance types: burned, harvested, and left-standing (n = 293). In Jasper, bird communities differed most clearly by disturbance type, whereas in the Southern Rockies, communities converged over time. Indicator species analysis revealed stronger and more distinct species associations in Jasper, while communities in the Southern Rockies showed possible legacy effects of wildfire. Together, these findings highlight how avian responses to disturbance are shaped by cumulative disturbance, elevation, and time, but also by the spatial configuration and scale of disturbance. Across both chapters, bird communities responded most strongly in landscapes where disturbance was spatially distinct, or applied in patchy mosaics. This research contributes to a growing body of work emphasizing the importance of heterogeneity and scale in shaping post-disturbance recovery, and supports forest management approaches that incorporate spatial and temporal variation to sustain biodiversity in dynamic mountain systems
data, materials and article for Spalding, T.L., Gagné, C.L., & Taikh, A. (in press). Early access effects in English compound and pseudo-compound words. Memory & Cognition.
No full textMaterials, data, data analysis code, and draft of manuscript for Spalding, T.L., Gagné, C.L., & Taikh, A. (in press). Early access effects in English compound and pseudo-compound words. Memory & Cognition.
Abstract
Although there is substantial experimental evidence that the morphemic constituents of compound words (e.g., snowball) are activated during compound word access, it is unclear exactly how the presence of constituents impacts word access. A series of experiments using a masked repetition primed lexical decision task investigates the role played by the morphology of compounds in word access. Semantically transparent compound words show consistent advantages relative to their frequency- and length- matched non-compound controls, but opaque compound words do not. For both kinds of compounds, the effect of repetition priming is the same for the compounds and their controls at short prime durations (50 and 100 ms). However, at long prime durations (300 ms), the compounds show more priming than their controls. In short, the compound advantage appears to be independent of the facilitation provided by short duration primes, but affected by long duration primes, and it appears to depend on the semantic transparency of the compound. Pseudo-compound words, such as carpet, provide an interesting comparison to compounds, because the language system cannot, a priori, determine whether they are compounds or not. Pseudo-compounds appear to be more difficult to process than their controls, and at long prime durations they show less priming than their controls. These results suggest that the compound advantage in processing arises relatively late in processing and is sensitive to the match between the semantics/morphology of a constructed compound interpretation and the required whole word.
Keywords: Compound words, Pseudo-compound words, Masked Priming, Morphology, Semantic Transparenc
An Interpretive Description Exploring How Nursing Faculty Teach Pain Assessment and Management in the Clinical Setting
No full textBackground
Nurses require competencies in pain assessment and management to provide safe and compassionate care, yet gaps persist in how effectively undergraduate education prepares students for this aspect of practice. There is limited evidence on how clinical nursing faculty teach pain in practice settings or the factors influencing their teaching.
Purpose
The purpose of my study was to explore how nursing faculty teach pain assessment and management in clinical settings. I examined faculty knowledge, attitudes, and teaching strategies, as well as the influence of clinical and curricular contexts on teaching and learning. I also explored how these elements influenced faculty’s approach to supporting student learning and ensuring that students are prepared to manage pain effectively in practice.
Methods
I used an interpretive description design with mixed methods. Twenty-three faculty from undergraduate nursing programs in western Canada completed the Knowledge and Attitudes Survey Regarding Pain (KASRP). Descriptive statistics identified strengths and gaps in knowledge and attitudes. Fourteen faculty participated in individual interviews and four in a focus group. Transcripts were analyzed using constant comparative analysis. Findings
I identified three themes from my analysis: (1) Internal factors influencing teaching included faculty knowledge, confidence, and prior experiences, which shaped how they engaged students in learning. Faculty had strong foundational knowledge but identified gaps in opioid use, dependency, and cancer pain management. The faculty’s confidence in their teaching role or comfort in the clinical environment influenced whether they focused more on tasks or facilitated reflective, person-centred learning. (2) Contextual factors shaping teaching and learning included unit culture, staff attitudes, and student beliefs. Supportive staff created positive role modelling opportunities, while workload pressures, stigma, and uncertainty about opioid use sometimes limited student learning. Faculty’s knowledge of where pain curriculum was taught in the program influenced clinical teaching, as unclear expectations and inconsistent communication between theory and clinical courses left some faculty unsure of what content to emphasize. (3) Preserving the ideal in practice-based learning described how faculty promoted person-centred care and professional identity development related to managing pain. Strategies included using real patient encounters, reflection, debriefing, storytelling, and advocacy to help students integrate theoretical concepts, develop empathy, and strengthen clinical reasoning.
Implications
My findings highlight opportunities to strengthen pain education in nursing and improve the delivery of pain care. Clinical faculty play a vital role in shaping students’ knowledge, attitudes, and advocacy skills through patient encounters, reflective dialogue, and role modelling of ethical practice. Nursing programs can enhance education by investing in faculty development, clarifying curriculum expectations, and addressing the powerful influence of unit culture on student learning. Supportive clinical environments enable faculty to reinforce compassionate, evidence-informed pain care, while negative norms and stigma create barriers to both learning and patient outcomes. Practical strategies include designating pain education champions, creating accessible teaching resources, and promoting interdisciplinary collaboration to foster consistent, person-centred practices. Future research should evaluate how these strategies, along with initiatives to strengthen unit culture, influence student competencies, faculty teaching practices, and patient outcomes across healthcare settings.
Conclusion
Faculty play a central role in preparing nursing students to assess and manage pain. Their knowledge, confidence, and teaching approaches, combined with the curriculum structure and clinical context, influence how students learn and apply concepts related to pain. Strengthening faculty preparation, clarifying curriculum expectations, and fostering supportive clinical environments can create consistent learning opportunities and help students deliver ethical, person-centred pain care. Continued research evaluating clinical teaching strategies and faculty development will further advance pain education in nursing
Effects of Northern Latitudes on Bat Nightly Activity Patterns in Western Canada
No full textNorth American insectivorous bats provide critical ecosystem services, contributing to natural pest control, agricultural productivity, and human health. These nocturnal mammals typically exhibit variable nightly activity patterns that align with prey availability, though activity timing and structure are also influenced by ambient light conditions. At higher latitudes, extended twilight periods result in delayed sunsets and early sunrises, while day-night transitions occur more rapidly at lower latitudes. Previous research on bat responses to light in northern regions has largely focused on single locations, typically examining how emergence timing shifts across seasons or in response to changing light conditions. This landscape-scale study examines bat populations across a broad latitudinal gradient (49-61°N) in the northwestern portion of their ranges in Western Canada. We aimed to determine how latitude shapes bat activity patterns by examining differences in activity curve shapes, modality, and peak timing across this region, where natural light regimes vary significantly during summer months.
We hypothesized that (1) bat activity curves would be more similar at latitudes closer together than those farther apart, (2) bats at higher latitudes would exhibit unimodal activity patterns due to shorter nights while lower-latitude bats would show bimodal patterns, and (3) peak activity timing would shift with latitude, with northern bats showing earlier peaks due to higher light tolerance and southern bats showing later peaks. We acoustically sampled 127 sites across Alberta and Yukon during May–August of 2021–2022, deploying autonomous recording units to capture bat echolocation calls. We used Wasserstein distance calculations to quantify dissimilarity between latitudinal activity curves, Hartigan's dip test and linear regression to assess changes in activity pattern modality with latitude, and modeled peak activity timing using standardized nightly activity data and kernel density estimates.
Results did not support our initial hypotheses. Wasserstein distances between latitudinal bands showed no systematic relationship to geographic distance, with some adjacent latitudes showing high dissimilarity while widely separated bands exhibited low dissimilarity. Only 24% of our sites exhibited statistically significant multimodal activity patterns, with multimodality predominantly occurring at latitudes below 56°N. However, unimodal patterns were observed across all latitudes, and latitude was not found to be significant in explaining multimodality patterns. Peak activity timing revealed a complex cubic relationship with latitude when all sites were included, but when analysis was restricted to Alberta sites to control for longitudinal confounding, peak activity occurred progressively later at higher latitudes, contrary to our prediction of earlier peaks.
These findings suggest that while factors such as landscape characteristics, species composition, and proximity to roosts are influential in shaping bat activity patterns, latitude, and by extension atmospheric light levels, still demonstrates a measurable impact on the timing of peak activity. The later peak activity at higher latitudes indicates that despite shorter nights, bats delay activity until ambient light levels fall sufficiently, prioritizing darkness over extended foraging time. This research contributes to understanding how bats adapt their temporal activity patterns across large geographic scales and provides baseline knowledge for investigating environmental drivers of bat behavior. As climate change drives range expansions into northern latitudes and artificial light pollution increases globally, understanding natural variation in bat responses to light conditions becomes increasingly important for developing geographically targeted conservation strategies and predicting species responses to environmental change