SHAREOK Repository
Not a member yet
49261 research outputs found
Sort by
A Machine Learning Based Multi-model ensemble Approach to reconstruct the historical monthly precipitation over Oklahoma using NOAA's SPEAR dataset
Get PDFGeneral Circulation Models (GCMs) are important tools in simulating and projecting future precipitation at the decadal scale. However, it is inevitable that simulation and projection errors and uncertainty exist in GCMs, hindering their applications for regional water resources planning. Different post-processing tools are available to address the uncertainty issues associated with GCMs and to utilize these tools better for regional water resources planning. For example, a multi-model ensemble (MME) could reduce uncertainties from different GCMs and help reduce the model biases from a single model. In this study, we employed multiple Machine Learning algorithms (MLs) to combine ensemble members from NOAA’s Seamless System for Prediction and EArth System Research (SPEAR) to reconstruct historical monthly precipitation over Oklahoma during a study period (1981-2014). The employed MLs include Random Forest (RF), eXtreme Gradient Boosting (XGBoost), Support Vector Machines (SVM), and Classification And Regression Trees (CART). The performances of the employed MLs are benchmarked with Simple Model Averaging (SMA), Bayesian Model Averaging (BMA), and Reliability Ensemble Averaging (REA). Our result echoes previous studies where the raw precipitation simulation from SPEAR presents significant simulation bias and marginal simulation skills. Different spatial and seasonal patterns of the simulation bias and skill are also observed over our study region. All the employed multi-model averaging techniques have delivered better performances than any single ensemble member from SPEAR. The employed MLs have outperformed SMA, BMA, and REA, which is evident from the reduction of bias and skill improvement. In general, this study highlights future applications of other data-driven techniques in post-processing the multi-model simulation from GCMs
Freedom and Restraint: A Self-Control Perspective on Political Ideology
Get PDFWithin the United States the topic of political ideology is an evolving and complex matter. The goal of this thesis is to examine an understudied area within this topic – the impact that self-control has on political ideology, and how it may influence the formation, expression, and/or evolution of political beliefs. This study deploys a quantitative approach utilizing Multinomial Logistic Regression to examine the relationship between self-control and political ideology.
The research begins by establishing a theoretical framework examining the concept of Self-Control, nested within Social Control Theory. Drawing on existing literature, these theories offer a paradigm through which we can view political ideology, where the erosion of faith in the political system sometimes results in political deviance. While this research does not measure political deviance, the findings have both academic and practical implications. I found that as a main effect, self-control does not significantly impact on political ideology; however, when the moderating effects of sex and race are included, the effects of self-control on political ideology differ across race and class, suggesting self-control as a potential influence on the political views, and therefore potentially the political behavior, electorate. In terms of practical implications, this study can inform public policy, with the goal of mitigating and impacting impulsive decision-making motivated by political attitudes and behaviors, which in turn can aid in the development of targeted interventions to quell political deviance/violence that has been seen at elevated levels within the United States over the past decade
The impact of contextual bias on novice examiners in firearms examinations and fingerprints analyses
Get PDFDespite efforts to maintain objectivity in analysis and interpretations, bias and human error still impact forensic science. Diversity in the perspectives, ideologies, techniques, and experience of forensic scientists generates a lack of standardized systems and methodologies in the analysis of evidence and introduces bias into forensic science. This leads to complications regarding interpretation and can result in subjectivity being introduced into testimony. The literature discusses the importance of comprehending bias as a whole and its impact on various fields in forensic science. The use of subjective judgments can substantially influence the analysis of fingerprints, firearms, pathology, bite marks, and crime scenes. Subjectivity can occur due to previous contextual information being given to an individual prior to examination. Objective analysis is crucial for factual investigations and to prevent wrongful convictions. This study focuses on DNA contextual information and how it impacts the conclusions of novice analysts when analyzing firearms and fingerprints. In this study, twenty-one participants from a forensic science analysis lab course conducted bullet comparisons and eighteen participants from an advanced fingerprint course analyzed fingerprints. A randomly assigned between-subject Fisher's exact test was used to analyze the data. In the examination of both fingerprints and firearms, participants were given contextual information regarding a crime scene, a suspect in custody, and whether DNA found at the scene matched, did not match, or was unknown to be a match to the suspect. Fingerprint novices were then asked to analyze and compare an unknown fingerprint from the crime scene to one of the suspects known prints and firearms novices were asked to compare a bullet collected from the crime scene to a bullet fired from the suspect's weapon. Examiners were also asked to indicate their level of confidence in their conclusions. Results indicated that when novices were not provided contextual DNA information, they formed conclusions with 100% accuracy in both fingerprints and firearms examinations. Statistical analysis indicated that for the fingerprint task, there was no significant difference in the proportions of correct answers for all groups and no significant differences in the median confidence levels of participants in different groups. For the firearms task, there was also no significant difference in the proportions of correct answers for all groups. However, confidence levels for novices in the firearms task were significantly higher when no contextual DNA information was provided. In both fingerprints and firearms, participants had 100% accuracy in the conclusion when analyzing prints or bullets in the control group vs. lower percentages of accuracy in the experimental group in which participants were provided with DNA contextual information
MULTI-SCALE HYBRID DATA-DRIVEN FRAMEWORK FOR ELECTRIC ENERGY FLOW: TRANSIENT ANALYSIS AND RESILIENCY SOLUTION FOR NEXT-GENERATION POWER GRID
Get PDFIllustrating by compelling case studies, this study discusses the challenges to the design of resilient power grid with inverter-based renewable energy resources posed by curses of dimensionality and heterogeneity in the complex multi-scale interactive dynamics by utilizing the network-based data-driven approaches, and further explains the essentiality of having hybrid data analytics with multi-scale stochastic autocovariance analysis to overcome those newly recognized curse of varietal complexity in next-generation power grids. It underscores the necessity for resilient functions beyond traditional reliability, explores essential resiliency aspects of power grids integrated with renewable resources, emphasizing proactive mechanisms to ensure stability during disruptions.
By integrating numerical methods and differential embedding techniques, the framework enables precise modeling of energy flow dynamics across temporal and spatial scales, crucial for accurate transient analysis. Simulation case studies further illustrate its capability in distinguishing nonlinear dynamics induced by IBRs from traditional synchronous generators, showcasing robustness and accuracy even under varying penetration levels of IBRs. This innovative approach not only enhances modeling accuracy and prediction capabilities but also provides insights essential for managing and controlling the complex dynamics of modern power grids
Forecasting Quasi-Linear Convective Systems and Mesovortex Tornado Potential Using the Warn-on-Forecast System (WoFS)
Get PDFAbstract: Quasi-linear convective systems (QLCSs) can produce multiple hazards (e.g., straight-line winds, flash flooding, and mesovortex tornadoes) that pose a significant threat to life and property, and are often difficult to accurately forecast. The NSSL Warn-on-Forecast System (WoFS) is a convection-allowing ensemble system developed to provide short-term, probabilistic forecasting guidance for severe convective events. Examination of WoFS's capability to predict QLCSs has yet to be systematically assessed across a large number of cases for 0-6-hr forecast times. In Part I of this study, the quality of WoFS QLCS forecasts for 50 QLCS days occurring between 2017-2020 is evaluated using object-based verification techniques. First, a storm mode identification and classification algorithm is tuned to identify high-reflectivity, linear convective structures. The algorithm is used to identify convective line objects in WoFS forecasts and Multi-Radar Multi-Sensor system (MRMS) gridded observations. WoFS QLCS objects are matched with MRMS observed objects to generate bulk verification statistics. Results suggest WoFS's QLCS forecasts are skillful with the 3- and 6-hr forecasts having similar probability of detection and false alarm ratio values near 0.59 and 0.34, respectively. The WoFS objects are larger, more intense, and less eccentric than those in MRMS. A novel centerline analysis is performed to evaluate orientation, length, and tortuosity (i.e., curvature) differences, and spatial displacements between observed and predicted convective lines. While no systematic propagation biases are found, WoFS typically has centerlines that are more tortuous and displaced to the northwest of MRMS centerlines, suggesting WoFS may be overforecasting the intensity of the QLCS's rear-inflow jet and northern bookend vortex.
One of the most challenging aspects of QLCS prediction is the ability to accurately forecast mesovortex tornado potential. The current paradigm employed by the National Weather Service for forecasting the likelihood that a QLCS will produce a mesovortex and, potentially, a tornado is known as the Three Ingredients Method (3IM). This technique's first two ingredients are grounded in RKW Theory, which states there should be a balance between the strength of the QLCS's cold pool and the magnitude of line-normal shear to produce an optimal balanced state that favors strong, upright updrafts and enhanced low-level rotation. However, mesovortices have been observed despite sub-optimal line-normal shear within the framework of the 3IM. Additionally, observational studies examining prolific mesovortex- and tornado-producing QLCSs found a component of the low-level environmental shear field that was
oriented line-parallel, which is not included in the 3IM. Recent observational and modeling work suggest line-parallel shear fosters the development of strong, cyclonic mesovortices, thus increasing the overall likelihood of QLCS tornadoes.
Motivated by these recent results, Part II of this study examines three tornadic QLCS case studies from 12 May 2022, 30 March 2022, and 15 December 2021 to investigate the potential of using the upstream kinematic environment to predict QLCS mesovortices in WoFS. Using object-based identification methods and a new technique to isolate the leading line of the WoFS QLCS objects, we are able to calculate the shear components upstream of the QLCS. Results suggest that near-storm shear components can be combined with information about the local line geometry to provide a system-relative forecast of mesovortex tornado potential in WoFS. A new composite parameter, called QTor, characterizes how favorable the environment is for the development of mesovortex tornadoes based on the magnitudes of 0-1-km line-parallel shear, 0-3-km line-normal shear, and the local tortuosity/curvature of the leading line. For each case study, the QTor forecasts are able to predict areas favorable for the development of mesovortex tornadoes that also coincide with the documented tornado reports for each case. However, small orientation differences between the WoFS and MRMS leading lines results in QTor false alarms, which will need to be accounted for and mitigated in the future. The long-term goal of this project is to revise the 3IM and bring it into a more objective framework that can then be implemented into any convection-allowing model
Expert-Guided Machine Learning for Meteorological Predictions Across Spatio-Temporal Scales
Get PDFThis dissertation emphasizes the contribution of expert knowledge in the development and assessment of machine learning (ML) models within the Earth sciences, specifically Meteorology. Despite the common focus on achieving high skill scores, conventional metrics may inadequately capture the nuanced patterns learned by these models. This dissertation underscores the importance of incorporating end-user feedback, demonstrating that with this feedback, tailored yet flexible ML models can effectively learn specific meteorological patterns while remaining applicable to broader contexts.
The first focus of ML development is in identification of above-anvil cirrus plumes (plumes). In satellite imagery, plumes serve as critical indicators of impending severe weather, often appearing 30 minutes before reported events. Their real-time identification is particularly valuable in radar-deficient regions, where they offer insights into the convective environment. However, manually labeling plumes is labor-intensive and requires specialized expertise. To streamline this process, I develop a deep learning (DL) model trained on expert-annotated data to create skillful pixel-level plume classifications using remote sensing data that is available globally. This approach was tested on combinations of spectral data across the contiguous United States, showing above-average object correspondence with human-derived labels.
Another focus of this dissertation is leveraging ML models for severe hail prediction on localized scales. Existing ML models have demonstrated proficiency across the United States during spring and summer but have struggled to capture the nuanced spatio-temporal dynamics of thunderstorm development in local contexts. Addressing this gap, I develop a novel localization technique that prioritizes storm object weighting without imposing substantial additional burdens on model developers. Results indicate that localized weighting of storm objects matches or outperforms existing ML approaches, while improving the physical relevance of the top predictors in the trained ML model.
Lastly, leveraging extensive satellite data archives, this dissertation addresses the challenge of efficiently creating training sets that accurately represent large-scale Earth science datasets. This work explores clustering approaches to capture regional nuances within a vast dataset of remote sensing data, focusing on a straightforward use case with an established baseline: land cover classification. By using surface reflectance bands in a random forest (RF) model, I compare classification outcomes between randomly sampled datasets of varying sizes and datasets created using clustering. The clustering approach produced a training sample that was 200% smaller than the largest sample studied, yet it achieved a 77% increase in F1 score. This suggests that clustering may offer an effective alternative (or addition) to increasing computing power when modeling "Big Data"
Spatiotemporal gap-filling of NASA Deep Blue aerosol optical depth over CONUS using the UNet 3+ architecture
Get PDFDue to sensor and algorithmic constraints, satellite aerosol optical depth (AOD) retrievals are spatially incomplete over clouds, deserts, and other bright surfaces. These gaps in satellite AOD datasets represent a significant challenge in characterizing aerosol distributions at a daily temporal resolution. These challenges are essential to overcome due to aerosol impacts on human health, the economy, weather, and climate. The task of filling these gaps in AOD datasets is known as AOD gap-filling. In this study, we propose using a deep learning (DL) architecture called UNet 3+ to perform this task. The model is trained on Deep Blue (DB) AOD retrievals from Terra, Aqua, and S-NPP, MERRA-2 reanalysis AOD, meteorological and land-use variables from NAM, and HMS smoke polygons. Through spatial evaluations against AERONET and DB AOD, we show that such an approach is feasible over CONUS, even in the semi-arid western U.S. where historically, topography, bright surfaces, and snowpacks have made AOD gap-filling a challenging problem. We created spatiotemporal datasets of daily gap-filled DB AOD from 2012-2022 over CONUS at a 12 x 12 km2 resolution with statistical evaluations of RMSE~0.08 and r~0.84 against collocated AERONET retrievals. This dataset will be a starting point for future aerosol-related studies, such as acute daily PM¬2.5 (particulate matter with an aerodynamic diameter smaller than 2.5 µm) exposure studies. Some potential challenges exist with the suggested approaches and, more generally, in estimating AOD and PM¬2.5. The first is the sampling bias that naturally arises from AOD retrievals, whether from ground- or satellite-based sensors. As no retrievals are performed in cloudy pixels, any model trained on this data is only aware of the dynamics of clear-sky AOD, and we cannot directly validate estimations over cloudy-sky areas (not even with ground truth AERONET). However, because this research aims to use AOD to estimate PM¬2.5 and hydrological effects tend to lower aerosol concentrations (e.g., washout from precipitation), the methods can be justified as providing an upper-bound (acute) estimate for exposure
How Early Childhood Hope Lessons Impact Students and Teachers
Get PDFWhile focusing on instruction and academics, today’s classroom teachers must also meet the individual needs of students who come to the learning environment socially and emotionally unprepared. Positive psychology Hope lessons may be one way to strengthen skills which must be in place before learning can happen. This case study explored the perception of eight pre-kindergarten teachers as they reflected on early childhood Hope Theory lessons involving goals, pathways, and motivation. Through interviews, observations, and artifact analysis, over the course of a seven-month span of time, this work explored whether Hope instruction made a positive difference for students and determined how including these lessons in the school day impacted teachers. The term “impact” for this dissertation is defined by the researcher as the assessment of the perceived effect of Hope lessons on teachers and students following lesson implementation based on data collected from participant interviews, observations, and artifacts. Findings indicated that teaching Hope is possible, even with young students, and that Hope positively impacts social-emotional wellbeing and academic growth. Hope lessons were found to build relationships between students, teachers, parents, and the community at large. Including the concept of Hope within school curriculum was determined to be important. The results of this study inform future practices for all educators, and recommendations are presented for administrators, teachers, social-emotional and academic curriculum decision-makers, teacher-preparation entities, and legislators
