..

Research

AGORA Project: Comparison of Galaxy Simulation Codes

Working with Ji-hoon Kim, Santi Roca-Fàbrega, Joel R. Primack, and the AGORA Collaboration

The AGORA Project performs controlled comparisons of galaxy simulations run with different numerical codes but common initial conditions and physical prescriptions. My work examines the formation and evolution of satellite populations, galactic disks, the circumgalactic medium, and galaxies undergoing major mergers. By identifying which results are robust across simulations and which depend on numerical methods, these comparisons help guide the development of more reliable galaxy-formation models.

FOGGIE Project: Ram-Pressure Stripping and Absorption Signatures of Gas from Dwarf Satellites

Working with Anna Wright, Stephanie Tonnesen, and Jingyao Zhu

I use high-resolution FOGGIE cosmological simulations to follow LMC/SMC-like satellite galaxies during their infall into Milky Way-like hosts. We study the fate of stripped gas and its observational signatures.

Star-Cluster Formation with a Hydrodynamics + Direct N-body Code

Working with Yongseok Jo, Eunwoo Chung, and John Wise

We develop and apply hybrid simulations that combine hydrodynamics with direct N-body integration of individual stars. We use these simulations to study dense star-cluster formation, runaway stellar collisions, and the formation of ultra-faint galaxy progenitors.

Galaxy-by-Galaxy Resolution Study and Machine-Learning-Based Correction

I match individual galaxies across simulations with different numerical resolutions using their merger trees, enabling direct galaxy-by-galaxy comparisons rather than comparisons of statistical populations alone. This approach allows me to quantify how numerical resolution affects stellar mass, gas content, and the evolutionary histories of galaxies.

I also develop tree-based machine-learning models that infer high-resolution galaxy properties from lower-resolution simulations. By using information about galaxy structure, environment, neighboring systems, and earlier evolutionary stages, these models both correct resolution-dependent biases and reveal the nonlinear physical connections that govern galaxy formation.