PhD research at UC Riverside

My area of research at UCR is galactic formation and evolution using numerical simulations. This is a very rich topic that I have been interested in since my first astrophysics course as an undergraduate.  My advisor is Gabriela Canalizo.

In order to simulate a galaxy, many phenomena must be included because galaxies are much more than simply gravitationally-bound systems of stars.  For instance, a good simulation must include a consistent simulation of the gas and dust in the galaxy so that star formation can be properly taken into account.  Stellar population evolution must be included because different types of stars interact with the interstellar medium differently at various stages in their evolution.  The interaction between AGN and the host galaxies must be included because AGN perturb the galactic dust and gas and thus affect star formation.  Features of the galactic environment must also be included because gravitational interactions and mergers between galaxies are important in the evolution process.  Finally, cosmological considerations must be included in a good model.

The primary scientific value of galactic simulation is not its predictive power, but rather the inverse–the postdictive utility.  First we make observations of galaxies and then we create simulations which try to reproduce the observations.  Once we correctly reproduce the observations, we can say with a some certainty that the assumptions included in the simulation were at least partially correct.  We then seek alternative assumptions which lead to the same observations.  The ultimate goal is to find a unique set of assumptions which consistently reproduce all observations.  In this way, theories of gravitation on large scales, dark matter, and dark energy can be tested. Insight can also be gained into the conditions in the very early universe.

At the moment, I am particularly interested in learning how different theories of gravitation impact the simulations.  All of the models I am aware of use Newtonian gravitation, which is a causality-violating theory because it assumes that gravitational effects propagate infinitely fast.  Galaxies are sufficiently large and stars move sufficiently fast that the causality violation may become an issue.  Theories of gravitation which are consistent with causality contain extra effects which are absent in Newtonian gravity.  I find this fact somewhat troubling and thus I am interested in finding an existing analysis of this issue. If I cannot find a convincing analysis of the issue, I will investigate it myself.  Before tackling that problem though, I am gaining proficiency with the current state of the art simulation code, GADGET-2.  My doctoral research will focus on more mundane, but important, details of high-resolution simulations using a version of GADGET coupled with visualization codes and observational astronomy tools.

Master’s research at George Mason University

During my time at GMU, I worked in the Space Weather group. My advisor was Robert Weigel. About half of my time was spent learning to do research and acquiring the background knowledge required for the project that I would eventually work on.  I learned UNIX/Linux system administration, shell scripting, Apache web server configuration, , basic fluid dynamics, plasma physics, and MHD.  I also took Rainald Löhner’s computational fluid dynamics course where I learned about unstructured mesh generation, methods of solution for PDEs on unstructured meshes, and techniques of high-performance computing.

The focus of my research was inner-magnetosphere modeling.  The original intent was to acquire the source code to the Rice Convection Model (RCM), which is the standard code for modeling the conditions in the inner magnetosphere.  The plan was to modify the code such that it would be able handle a more realistic (less symmetric) geometric configuration so that the seasonal dependence of the model could be tested.  My advisor and I were unable to acquire a copy of the RCM.  It eventually became clear that the Authors at Rice University were not anxious to share, so I had to write the entire code myself.  I read the published descriptions of the RCM algorithm and I wrote a prototype in MATLAB.  I was able to complete the main structure of the program before I graduated. My description of the project, OpenConvection, can be found here. It is now being used mainly for educational purposes.

Some plots from OpenConvection