Mesoscopic Boxes and Subgrid Plasma-Microphysics Prescriptions for Global MHD Modeling of Astrophysical Systems*

Recent years have seen remarkable progress in our understanding of large-scale dynamics of high-energy astrophysical systems, such as accreting black holes and neutron-star magnetospheres, thanks largely to global 3D relativistic magnetohydrodynamics (MHD) simulations. However, our confidence in the applicability of these numerical models to collisionless plasmas in many real systems is hampered by the inability of single-fluid MHD to predict observable radiative signatures directly. This motivates the development of subgrid prescriptions for the partitioning of the dissipated magnetic and bulk-flow kinetic energy among various kinetic-level channels (e.g., nonthermal particle acceleration and electron heating fraction).  The recent advent of such prescriptions, based on gyrokinetic and first-principles particle-in-cell (PIC) kinetic simulations, has had a major impact on plasma astrophysics of black holes. However, these subgrid models are usually formulated in terms of single-point local plasma parameters such as plasma-beta or magnetization sigma. I will argue that such approach is too simplistic, and more sophisticated procedures are needed, involving an analysis of MHD-level plasma and field quantities as well as their spatial variation in quasi-local mesoscopic regions. I will outline a broad plan for implementing this strategy for the cases of magnetic reconnection, Kelvin-Helmholtz instability, and plasma turbulence.