Skip to Content

2018-2019 Blue Waters Graduate Fellows

There were 81 applicants for this year's fellowship competition. Of those, 60 were male, 18 female, and 3 who declined to specify their sex. There were 7 candidates that indicated they were Hispanic, 2 American Indian of Alaska natives, 9 Asian, 1 African American, 1 native Hawaiian or Pacific Islander, 64 white.

Below is a summary of the project for the 10 finalists.

Elaad Applebaum

Elaad Applebaum

Rutgers University

Though we can model realistic dwarf galaxies, differing star formation recipes yield degenerate results, leaving the specific mechanism of star formation unconstrained. We propose breaking this degeneracy by turning to the regime of ultra-faint dwarf (UFD) galaxies, which we have simulated, for the first time, in a cosmological context. Unlike classical dwarfs, different star formation models lead to different numbers, distributions, and formation densities of UFDs. We will therefore tie detailed predictions from these models to upcoming observations. Using cosmological simulations to constrain the detailed physics of star formation is an entirely novel approach, opening a new avenue for studying galaxy evolution and star formation.

Katelyn Barber

Katelyn Barber

University of North Dakota

Convectively-induced turbulence is an aviation hazard that is a forecasting challenge because operational weather models are too coarse to resolve turbulence. Indices are commonly used to aid pilots in avoiding turbulence. Over tropical oceans, turbulence prediction systems rely heavily upon nowcasting and indices not designed for tropical convection. In this study, a large number of numerical simulations of tropical oceanic and mid-latitude continental convection will be performed to produce statistically robust results. This study will address the current shortcomings in turbulence prediction systems and will determine the limitations of mid-latitude continental indices currently used for turbulence prediction over tropical oceanic regions.

Shanna Chu

Shanna Chu

Stanford University

I will improve current understanding of intermediate-depth earthquakes, which occur in a regime where normal earthquake physics models break down. Current methods of studying earthquake source properties use over-simplified physical models due to computational cost. Using Blue Waters' resources, I will more accurately compute and bound earthquake source properties using hierarchical Bayesian models on large datasets from tectonically diverse regions, running a Hamiltonian Monte Carlo algorithm. I will use simulations from the waveqlab3d wave propagation code to validate the physics of models used in data-fitting. By better estimating source properties, I hope to constrain the energetics and mechanisms of intermediate-depth earthquakes.

Robert Cieri III

Robert Cieri III

University of Utah

Vertebrate lungs are incredibly complex and diverse structures. Variation in airflow patterns in these lungs remains unexplored because the lungs are too complex for flow to be measured directly. High performance computing and computational fluid dynamics (CFD) are novel tools for investigating these structures. Using CFD, I plan to compare lung airflow patterns among monitor lizards, a diverse group of reptiles with especially diverse lungs. With Blue Waters support, I would expand my thesis to open up a new line of inquiry in biology, investigating how lung airflow patterns evolved with habitat, body size, and metabolic rate. This research could help explain the evolution of unidirectional airflow in lungs and open these structures up for biomimetic applications.

Andrew Emerick

Andrew Emerick

Columbia University

Developing codes that can take advantage of current and future generations of computing environments is essential to advancing our understanding of the Universe. Using the cosmological hydrodynamics code Enzo, I developed a revolutionary new approach capable of following the interactions between individual stars and entire galaxies. However, the size of these galaxies is restricted by current scaling limitations. With my Blue Waters Fellowship I will port physics modules from my work to the the Enzo-P/Cello project, which has demonstrated efficient scaling up to 256,000 cores on Blue Waters. I will test how well our current findings scale to more massive galaxies and will open the work for use by the broader astrophysics community.

Alexander Gurvich

Alexander Gurvich

Northwestern University

Massive outflows of gas from galaxies known as galactic winds are crucial in regulating galactic star formation rates. These outflows are observed to consist of multiple phases, including both hot (~million K) and cold, molecular, gas. Previous simulations either lacked the resolution or the chemistry treatment necessary to capture these phases. This limits the ability of previous work to produce powerful galactic winds and also accurately predict observations of their chemical diagnostics. We propose to develop and employ a revolutionary GPU-accelerated time-dependent chemistry code to enable us to run a suite of high resolution simulations that will, for the first time, capture the full range of phases of galactic winds.

Kara Marsac

Kara Marsac

Colorado School of Mines

Over-extraction of water for hydraulic fracturing is a top risk to the United States' water resources. Within the Permian Basin, an oil and gas basin in Texas, 87% of wells are in areas of high or extreme water stress, and by 2020 it is estimated that 13 billion gallons of water will be used for oil and gas development here. A proposed alternative to freshwater for hydraulic fracturing is saline groundwater, however there is concern that mineral precipitation could reduce permeability. To test the feasibility of this alternative water source, we propose using advanced, numerical reactive transport simulation to understand mineral precipitation reactions that result from natural brines mixing with formation waters in the Permian Basin.

Ruijie Darius Teo

Ruijie Darius Teo

Duke University

Iron-sulfur clusters exist in a variety of proteins, including DNA-binding proteins like DNA primase and mutY DNA glycosylase. Due to their redox activities, these clusters participate in charge transfer. However, as tyrosine-to-cysteine mutations like Y345C in primase and Y179C in mutY are up-regulated in cancer, the aim of this project is to investigate how this type of mutation impacts protein-DNA charge transfer, DNA repair, and redox signaling. Molecular dynamics simulations and electronic structure calculations will be performed to fulfill this aim. A robust ab initio method will also be developed to calculate iron-sulfur cluster electronic couplings, a crucial parameter required to define charge transfer rates.

Walter Torres

Walter Torres

Duke University

Circulation around islands affects the transport of nutrients, contaminants, and sediment as well as larval retention and connectivity. Recent field observations of currents around a small island show a persistent counter-clockwise flow, however little is known about the mechanism(s) by which the circulation pattern is established and maintained and if it is a general feature of other island systems. Herein, I propose to use a combination of idealized and realistic computational fluid dynamics experiments validated by field measurements to identify the role of four candidate mechanisms that may explain this circulation: radial buoyancy gradients, alongshore radiation stresses, pressure gradients from mesoscale eddies, and upstream island wake turbulence features.

Samuel Whitman

Samuel Whitman

University of Colorado Boulder

In this project, enhanced gas turbine stability, efficiency, and emissions will be enabled through study of turbulence-flame interactions in the presence of intense coherent vortex structures. High-resolution direct numerical simulations of bluff body stabilized turbulent flames will be performed on the Blue Waters machine using a massively parallel next-generation reacting flow code that includes adaptive mesh refinement and embedded boundaries. Energy content and modal structures in the resulting data will be analyzed to address the modeling challenges that limit stability and efficiency in current designs. The results will aid in the development of future turbines and reduced order models to support a more renewable energy grid.