21978 research outputs found
Sort by
Essays in Environmental and Health Economics
Get PDF(1) Chapter 1 - this chapter investigates the effects of prenatal exposure to nitrate contamination in drinking water on birth outcomes. While nitrate contamination is a persistent issue in U.S. water systems, little is known about its health impacts, especially at levels below current regulatory limits. I construct a novel dataset linking monthly nitrate measurements in California’s community water systems with individual birth records. Using a panel fixed-effects model, I estimate that second-trimester exposure to nitrate concentra- tions below regulatory limits increases the probability of preterm birth and low birth weight by 15% and 17%, respectively. The findings suggest that current regulations may not adequately protect vulnerable populations and that lowering nitrate limits could substantially improve birth outcomes.
(2) Chapter 2 - While prior literature has documented efficiency gains from privatization, its impact on quality remains underexplored—particularly in sectors with direct public health implications. Using hand-collected data on municipal systems sold to private companies and employing a propensity-weighted difference-in-differences approach, I find that privatization leads to 1.4 fewer Safe Drinking Water Act violations, a 20% decrease in an index of regulated contaminant concentrations, and a 30% decrease in an index for contaminants that pose an immediate threat to human health. These findings indicate that privatization leads to an overall improvement in drinking water quality and back-of-the-envelope estimates suggest economically meaningful benefits to public health, averaging at least $12.6 million per state in the sample.
(3) Chapter 3 - In 2020, the International Maritime Organization instated a fuel sulfur content limit for all global waters in an effort to reduce shipping-related air pollution. This paper examines the impact of this more stringent regulation on coastal air pollution by leveraging differential regulatory exposure along the eastern and western coasts of the United Kingdom. Using a differences-in-differences approach with air quality monitor data, I find little evidence that this regulation significantly reduced coastal PM2.5 concentrations, though several factors may limit the ability to detect an effect.</p
Yes, And-ing the Climate Crisis: Emergent Facilitation Strategies for Improvisational Activities About Climate Change
Get PDFThis dissertation is a combination narrative analysis and performance autoethnography of effective facilitation strategies that emerged from improv activities aimed toward enhancing positive engagement on climate change. The activities were part of a workshop series I facilitated in three different contexts: the University of Iowa, the University of Colorado Boulder, and an arts-science summer school program hosted by SHiFT-COST (Social Sciences and Humanities for Transformation and Climate Resilience, European Cooperation in Science and Technology) in Berlin, Germany. Utilizing a multi-methodological approach, this dissertation seeks to combine multiple areas of study including improvisation, emergent strategy, embodied cognition, queer theory, and climate communications. The purpose is to elucidate new ways of thinking and doing that promote sharing energy, collective imagining, and communal collaboration. It argues that facilitating improvised activities promotes enhanced embodied engagement on issues of climate change. It is designed so that others will be able to utilize the strategies in order to accomplish similar goals.
</p
Low Thrust Trajectory Optimization in the Saturn-Titan System
Get PDFThe growing interest in missions to the moons of Jupiter and Saturn motivates trajectory optimization within these systems. This thesis focuses on the computation of propellant mass optimal trajectories for spacecraft with low thrust propulsive systems in the Saturn-Titan system. The Circular Restricted Three Body Problem (CR3BP) is used to describe the dynamical environment. The equations are modified to incorporate low thrust contributions. The method to generate an initial guess trajectory is defined. Trajectory optimization using primer vector theory is used along with multiple shooting to generate the required transfers. Necessary conditions are derived for the chosen dynamical system. Propellant mass optimal transfers to vertical and halo target orbits are computed to achieve coverage of the polar regions of Titan. The resultant halo orbit transfer is then utilized to generate transfers to a nearby NRHO orbit for closer approaches to the south polar region.</p
Approximately Truthful Forecasting Competitions
Get PDFForecasting competitions have emerged as a promising tool for identifying skilled forecasters, eliciting truthful beliefs, and improving decision-making across domains such as public health, fi- nance, and geopolitics. These competitions typically rely on proper scoring rules to reward accurate predictions, with the assumption that such rules incentivize honest reporting. However, when used in competitive settings where participants are ranked and rewarded based on relative performance, proper scoring rules alone fail to preserve truthfulness. Participants strategically misreport their beliefs to improve their chances of winning, compromising forecast accuracy.
This thesis formally investigates the incentive properties of forecasting competitions and proposes new mechanisms that better align individual incentives with truthful reporting. We first establish fundamental lower bounds on the number of events required to identify high-quality forecasters. We then analyze the standard proper score mechanism, showing it can incentivize forecasters to extremize or hedge their reports depending on their beliefs about the quality of their competitors. Next, we study the Event Lotteries Forecast mechanism, which guarantees truthfulness but show it necessarily requires a large number of events to be effective, which is far from our previously established lower bound.
Instead, we propose using FTRL as an alternative competition mechanism. We show that FTRL is approximately truthful: it incentives forecasters to report something very close to their beliefs. Furthermore, we show it is efficient and achieves the lower bound. We then show that FTRL is robust to correlation, a guarantee that no previous mechanism is able to give. In particular, we introduce the block correlation measure and show it measures the type of correlation we need to account for in the competition setting where standard correlation measures fail. We also derive a concentration bound for block-correlated random variables which may be of further interest. Finally, we show the guarantees of FTRL extend to an entirely different setting: online learning from strategic forecasters. Leveraging our approximate truthfulness results, we give the first no- regret guarantee for non-myopic strategic forecasters in this setting.</p
Construction of the Abstract Other in Online Dating Profiles
Get PDFThis paper serves as an exploration into the concept of the abstract other and the ways in which language users construct the identity of an individual who does not necessarily exist in their online dating profiles. While most interaction includes a negotiation of the identity of concrete others (other speakers or a person who is jointly known by speakers), a few specific interaction types include the prescription of identity onto an imagined other. Online dating profiles are an example of this, as users describe their ideal partner, an imagined figure, thus attributing characteristics to a person that does not exist. Through qualitative and distributional analysis, this paper finds that in addition to listing traits of their ideal partners, users are able to obscure the individualism of the abstract other by using the first-person plural, simultaneously constructing the abstract other and the self in imagined joint activities. Users can also create the abstract other by constructing a different, but related characteristic, leveraging the relatedness between the stated trait and the desired trait. Furthermore, the sets of tools that a dating platform provides to users in order to construct identities have an impact on user identity construction techniques, offering varying affordances and contingencies. Finally, the existence of a profile on a specific platform is itself a part of the identity building process as it is a tool leveraged by users to communicate alignment with the norms of that site.</p
Modeling Trust and Trust Dynamics From Physiological Signals for Operational Supervisory Control During Human-Autonomy Teaming
Get PDFHuman-autonomy teaming is an increasingly integral component of operational environments, including crewed and remotely operated space missions, military settings, and public safety. The performance of such teams relies on proper trust in the autonomous system, thus creating an urgent need to capture the dynamic nature of trust and devise objective, non-disruptive means of precisely modeling trust. This thesis describes the use of bio-signals and embedded measures to create a model capable of inferring and predicting trust. Data (2304 observations) was collected via human subject testing (n = 12, 6M/6F) during which participants interacted with a simulated autonomous system in an operationally relevant, human-on-the-loop, remote monitoring task. Participants reported their subjective trust via visual analog scales and gold-standard surveys. Electrocardiogram, respiration, electrodermal activity, electroencephalogram, functional near-infrared spectroscopy, eye-tracking, were collected during each trial and additional task-based measures were extracted from participant actions. Operator background information was collected prior to the experiment. Features were extracted from all the collected data streams and down selected using Least Absolute Shrinkage and Selection Operator (LASSO). Then, ordinary least squares regression was used to fit the model, and predictive capabilities were assessed on unseen data using leave one trial out per session cross-validation. Across 10 cross-validation splits, the model has a mean Q2 of 0.64 (σ = 0.05). The model has an adjusted R2 of 0.83. These values indicate a high predictive and descriptive fit. Further, the model, while static, captures dynamic changes in trust. Qualitative results show model predictions temporally fluctuating in the same manner as “ground truth” reported trust values. Each sensor stream provided features that were included as trust predictors (40% of neurophysiological features, 39% of psychophysiological features, 56% of operator background information, and 69% of embedded measures were retained). Out of 680 total available features, 279 were retained as predictor variables. The model advances the field of human-autonomy teaming as it captures rapid changes in trust during an operationally relevant task in which multiple facets of trust were altered. It improves upon current models with its large suite of sensors and focus on prospective predictive fit, while many other models in the literature only report descriptive fit performance.</p
Robust Aerocapture Guidance With High-Fidelity Modeling of Aerothermodynamic Uncertainty
Get PDFAerocapture is a mass-efficient orbital insertion method that utilizes atmospheric drag to decelerate a spacecraft, making it an enabling technology for Ice Giant exploration. However, the severe hypersonic environments encountered during aerocapture create significant challenges due to the intrinsic coupling between a spacecraft’s trajectory and its aerothermodynamic environment. Traditional spacecraft design approaches treat these subsystems separately, relying on iterative methods that neglect their direct interaction. This dissertation takes a step toward a more integrated treatment of hypersonic systems by developing robust aerocapture guidance frameworks that directly integrate aerothermodynamic uncertainty information obtained using high-fidelity computational fluid dynamics (CFD). First, a detailed uncertainty quantification (UQ) and global sensitivity analysis (GSA) is conducted at key points along a Neptune aerocapture trajectory, accounting for variations in chemical kinetic parameters and the freestream composition. This analysis yields highfidelity uncertainty bounds for convective and radiative heating distributions, as well as aerodynamic coefficients. Next, the information obtained from the UQ/GSA study is distilled into a reducedorder model using Gaussian process regression (GPR), which enables the efficient propagation of aerothermodynamic uncertainty along an entire trajectory. Finally, two aerocapture guidance algorithms, both based in the principles of sequential convex programming (SCP), are developed. The first is the convex predictor-corrector aerocapture guidance (CPAG) algorithm, which repeatedly solves a deterministic, constrained optimal control problem. CPAG eliminates the need for discrete lateral logic to control out-of-plane error and seamlessly integrates system-level constraints, such as heat load and aerodynamic loading, into its corrective logic. The second takes a stochastic optimal control approach to the aerocapture guidance problem, in which aerothermodynamic, aerodynamic, and atmospheric uncertainty are explicitly modeled within the optimization problem. This approach simultaneously optimizes a feedback control law with uncertainty in mind. Both guidance algorithms are compared against the state of the art and are shown to robustly guide aerocapture vehicles to their target orbit in the presence of uncertainty. By directly integrating aerothermodynamic modeling with trajectory optimization and guidance, this dissertation helps bridge the gap between traditionally separate subsystems in hypersonic vehicle and mission design</p
Quacking the Code: Environmental Justice, Methane, and the Duck Curve
Get PDFPublicly available data fills servers, waiting to be used. It comes in many shapes and sizes: from census data describing who lives where, to emissions reported from every power plant in the country. Much of the data is collected by regulation and reporting requirements and after collection and publication, is largely ignored. In this thesis we used terabytes of publicly available data to progress knowledge about environmental justice in the city of Denver, about how we can improve methane retrievals from satellites, and about how emissions from combined cycle natural gas plants are changed in a world dominated by solar energy.
First, we delve into the chemistry of air pollution in Denver, Colorado. Leveraging publicly available data from the US census, the TROPOMI satellite-based instrument, the national emissions inventory, state highway traffic counts, and highway emissions from NOAA colleagues, we explore who in the metroplex is most affected by air pollution. Additionally, we reach across fields to sociology to begin understanding the strengths and weaknesses of environmental justice study in various fields.
Secondly, we continue the first study with a follow-up focused much more heavily on the sociology and history of the Denver Metroplex and the racial and ethnic groups that live there. History demonstrates why certain groups live where they do in the modern day. This section proves the strength of multidisciplinary study in providing informed solutions to environmental injustices in Denver.
Third, we expand our view from Denver to the Denver-Julesburg basin where oil and gas exploration and drilling overlap considerably with agricultural practices. We show remote sensing of methane from oil and gas production in this region may be 2 affected by agricultural land-uses changing the land surface cover on a seasonal basis, and we developed a correction factor for this using a complex machine-learning model.
Finally, we leave Colorado and instead focus on the whole United States, where the climate crisis has spurred investment and installation of gigawatts of solar and wind energy around the country, and switching natural gas power plants to more efficient combined cycle technology. A quirk in the combined cycle technology, however, may lead to increased air pollutant emissions, like NOx, in urban areas during peak ozone production times.
This thesis advances our understanding of environmental injustices in Denver, Colorado, providing an example for other interdisciplinary studies. This thesis also provides a correction factor for methane measurement and advances our understanding of potential future problems in our energy infrastructure. This thesis also demonstrates the power of harnessing publicly available data, or simply data that has already been collected, but remains unused, for making advances in science that can be used for the good of all people.</p
Procedurally Generated Whole-Body Artificial Skins for Robot Applications
Get PDFBuilding on the foundation of the GenTact Toolbox pipeline, this thesis presents an indepth expansion of its computational design methods for generating whole-body tactile skins for robotic applications. Motivated by the increasing need for robots to safely and effectively interact with unstructured environments, this work addresses the challenge of creating tactile sensors that are precisely tailored to complex robot geometries and customizable for specific operational needs. This work refines key procedural generation techniques—including enhanced smoothing operations and optimized sensor distribution algorithms—to create form-fitting, context-aware sensor arrays. By integrating task-driven simulation with multi-material 3D printing, the pipeline offers greater control over design parameters such as sensor density, wiring, and surface coverage. Six unique capacitive sensing skins were characterized and deployed on a Franka Research 3 robot arm performing a human-robot interaction scenario to validate our approach. The GenTact Toolbox pipeline presents a shift from ”one-size-fits-all” artificial skins towards context-driven and highly adaptable skins that can be customized for individual robots and applications.
The project website is available at https://hiro-group.ronc.one/gentacttoolbox.</p
On the Subjective Valuation of Effort: Computational Insights into Motor Learning and Control
Get PDFEvery movement is a reflection of a person’s subjective values, manifested as a willingness to expend energy to achieve a desired outcome. Economically speaking, a person should maximize the utility of their movements by acquiring as much reward as early as possible with the least amount of effort. However, effort alone paints an incomplete picture of the costs associated with movement. There are other factors at play, such as error, time, and risk, that can modulate how much and when effort is expended. Here, I explore how the subjective valuation of effort relative to other costs can influence movement decisions.
The first study investigates the effect of age on preference for error over effort during motor learning. Older adults performed an arm-reaching adaptation task with greater error than younger adults, which is traditionally interpreted as a deficit in learning. However, I show that a difference in subjective cost of effort can manifest these same larger errors while still learning as much as younger adults.
The second study investigates the interaction of time and effort on preference. The traditional view of effort is that we should be agnostic to when periods of high effort are performed—only total energy matters so long as the goal is accomplished. Using an isometric arm-pushing task, I show that subjects preferred earlier high physical effort to later and that the strength of their preference is also reflected in how they indicated their preference.
The third study pilot tests how a virtual reality system could be used to investigate the effect of reward on walking speed. The fourth and final study builds and extends this pilot study to investigate how immediate reward, a history of reward, and baseline effort costs influence walking speed. I find that subjects walked faster not only for higher immediate reward but also when they experienced a history of higher reward. In addition, subjects walked slower when effort costs were higher.
Collectively, these studies deepen our understanding of how subjective effort cost interacts with error, time, and reward, and how these interactions manifest in our movement decisions.</p
Data sources:
Access Repository Dashboard
Do you manage Open Research Online? Become a CORE Member to access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard! 👇