Plasma physics problems in space are often too complex for purely analytic approaches, and too rich in kinetic physics for simple fluid models. Our group occupies a productive niche: we develop and deploy codes that capture kinetic, fluid, and wave physics at the scales where the interesting phenomena actually occur. We use four main classes of simulation, described below, along with strong analytic theory — treating computation and theory as inseparable and equally valued.
Our codes are run on national high-performance computing (HPC) facilities. Graduate students in the group typically spend their first year becoming comfortable with the codes and HPC environment, and spend subsequent years extending the codes and using them to solve new problems. Many of our simulation codes were built from scratch within the group, giving students genuine expertise in both plasma physics and scientific software.
PIC codes represent plasma electrons and ions as individual computational particles, moving under self-consistent electromagnetic fields. At each time step: (1) particle charges and currents are deposited onto a spatial grid; (2) Poisson’s equation (electrostatic) or Maxwell’s equations (electromagnetic) are solved on the grid; (3) particles are pushed with the Lorentz force. This first-principles approach captures wave–particle interactions, resonant heating, Landau damping, and nonlinear saturation mechanisms invisible to fluid models. Our 2D and 3D electrostatic PIC codes, developed in-house over three decades, have produced some of the highest-resolution kinetic simulations of ionospheric turbulence ever achieved. We also deploy electromagnetic PIC for chromospheric and auroral problems.
In many problems, ions can be treated as fluid particles while electrons require a kinetic treatment — or vice versa. Our hybrid codes exploit this separation: typically, ions are represented as macroparticles while electrons are treated as a massless, charge-neutralizing fluid (or vice versa). This approach dramatically reduces computation cost relative to full PIC while retaining the kinetic physics that matters most. Hybrid simulations in our group have been particularly valuable for studying coupled Farley–Buneman / gradient-drift instabilities in the equatorial E region (Young, Oppenheim & Dimant 2017), where the separation of scales between ions and electrons can be exploited.
For large-scale problems and multi-species plasmas — such as the partially ionized solar chromosphere containing electrons, protons, neutral hydrogen, helium, and heavier species — we use multi-fluid codes that track each species with its own fluid equations. These codes sacrifice single-particle kinetics in exchange for the ability to simulate large spatial domains and long time scales at reasonable computational cost. Our multi-fluid chromospheric codes (Evans et al. 2023, 2025) were the first to simulate the thermal Farley–Buneman instability in the solar chromosphere over physically realistic domain sizes, capturing the turbulent heating rate and its dependence on solar conditions.
Finite-difference time-domain (FDTD) methods solve Maxwell’s equations directly on a spatial grid, advancing electric and magnetic fields forward in time. Our group applies electromagnetic FDTD codes to model how radio waves propagate, scatter, and refract through structured ionospheric plasma — including the turbulent irregularities produced by the instabilities we simulate with PIC. The open-source radio wave propagation code released by Green, Longley, Oppenheim & Young (2024, Frontiers in Astronomy and Space Sciences) implements this approach and is now used by groups worldwide.
Many of our most important computational results have been made possible by the rigorous analytic theory developed by Senior Research Scientist Yakov Dimant. His contributions include the unified fluid theory of E×B instabilities in collisional magnetized plasmas, the kinetic theory of meteor plasma formation, the theory of anomalous conductivities from electrojet turbulence, and the thermal Farley–Buneman instability theory in the solar chromosphere. These theoretical results guide simulation design, explain simulation outputs, and in many cases have been published as standalone analytic results in Physics of Plasmas, JGR, and Annales Geophysicae.
Our group holds competitive allocations on national HPC systems through NSF ACCESS (formerly XSEDE) and uses BU’s on-campus Shared Computing Cluster (SCC) for development and moderate-scale production runs.
Large allocation for production-scale PIC and multi-fluid simulations. Our kinetic electrojet simulations routinely use thousands of cores for multi-day runs.
On-campus cluster for development, parameter surveys, testing, and course computation. Enables rapid iteration before committing to national facility runs.
We are porting our innermost PIC loop kernels to CUDA/HIP for GPU-accelerated nodes, leveraging the high arithmetic throughput of modern accelerated architectures.
If you want to write codes that run on thousands of cores, understand plasma physics from first principles, and publish results that advance both solar and space physics, this group is for you. No prior HPC experience required — we provide training from the ground up.
See the BU Astronomy PhD program at bu.edu/astronomy/graduate or contact meerso@bu.edu.