Table of contents

PREFACE

The Center for Space Plasma and Aeronomic Research (CSPAR) at the University of Alabama in Huntsville (UAH) and Maison de la Simulation, a joint laboratory at the French Alternative Energies and Atomic Energy Commission (CEA), National Center for Scientific Research (CNRS), University of Paris-Sud, and University of Versailles, organized the 12th International Conference on Numerical Modeling of Space Plasma Flows (ASTRONUM-2017) on June 26—30, 2017 in Saint-Malo, France.

The Program Committee consisted of Tahar Amari (CNRS Ecole Polytechnique, France), Edouard Audit (CEA/CNRS Maison de la Simulation, Gif-sur-Yvette, France, co-chair), Amitava Bhattacharjee (Princeton University, USA), Phillip Colella (Lawrence Berkeley National Laboratory, USA), Anthony Mezzacappa (University of Tennessee, Knoxville, USA), Ewald Müller (Max-Planck-Institute for Astrophysics, Garching, Germany), Nikolai Pogorelov (University of Alabama in Huntsville/CSPAR, USA, chair), Kazunari Shibata (Kyoto University, Japan), James Stone (Princeton University, USA), Jon Linker (Predictive Science, Inc., USA), and Gary P. Zank (University of Alabama in Huntsville, USA).

All papers published in this volume of Journal of Physics: Conference Series have been peer reviewed through processes administered by the proceedings Editors. Reviews were conducted by expert referees to the professional and scientific standards expected of a proceedings journal published by IOP Publishing.

There is a discrepancy between reconnection rates observed in a) various plasma environments, b) plasma PIC simulations and c) resistive fluid simulations. Careful observations of solar flares show reconnection rates between 0.001 — 0.1υ_A (see, e.g., [1]), plasma PIC and Hall-MHD simulations tend to show faster reconnection rate of around 0.1υ_A and pure MHD resistive simulations tend to converge to 0.015υ_A. The common explanation of the difference between b and c [2] is that plasma reconnection is inherently different from resistive MHD reconnection due to the different mechanisms of how individual field lines break and reconnect. This would imply a universal rate of 0.1υ_A in electron-proton plasma. In this paper, I report three-dimensional Hall-MHD simulations with resolution up to 2304 × 4608². Reconnection is mediated by self-excited turbulence in the current layer and is indeed faster than MHD reconnection. However, the reconnection rate goes down as the layer width increases. For the average layer width between two and ten ion skin depths (d_i), we recover "classic" value of 0.1υ_A, however, the trend of the rate to decrease may bring it to the fluid value for widths of around 170d_i, although reaching this width is not possible with current numerical resources. In any case, based on these simulations, now we are certain that the 0.1υ_A value has been favored before due to the limited range of scales available in numerical simulations.

We investigate the spectral properties of plasma turbulence from fluid to sub-ion scales by means of high-resolution three-dimensional (3D) numerical simulations performed with the hybrid particle- in-cell (HPIC) code CAMELIA. We produce extended turbulent spectra with well-defined power laws for the magnetic, ion bulk velocity, density, and electric fluctuations. The present results are in good agreement with previous two-dimensional (2D) HPIC simulations, especially in the kinetic range of scales, and reproduce several features observed in solar wind spectra. By providing scaling tests on many different architectures and convergence studies, we prove CAMELIA to represent a very efficient, accurate and reliable tool for investigating the development of the turbulent cascade in the solar wind, being able to cover simultaneously several decades in wavenumber, also in 3D.

Recent high-resolution 2D radiation-hydrodynamic numerical simulations of the formation and evolution of hot bubbles around evolved stars are described. The simulations take into account the evolution of the stellar parameters such as ionizing photon rate, wind velocity and mass-loss rate for a range of initial stellar masses. For low-mass stars, a planetary nebula with a lifetime of a few thousand years forms around the central hot star, while for massive stars the result is a Wolf-Rayet nebula, which has a lifetime of tens of thousands of years. In both cases, instabilities in the fast wind-slow wind interaction zone produce clumps and filaments in the swept-up shell of nebular material. Turbulent mixing and thermal conduction at the corrugated interface can produce quantities of intermediate temperature and density gas between the hot, shocked wind bubble, and the swept-up photoionized nebular material, which can emit in soft, diffuse X-rays. Sampling of the resultant theoretical spectra helps to make meaningful comparisons with recent observations of planetary nebulae.

Numerical simulations of binary star formation suffer from serious mismatch with observations. Equal mass binaries are abundant in numerical simulations while also unequal mass binaries are commonly observed. The discrepancy should be due to errors in the numerical simulations. In this paper we discuss the evaluation of the Coriolis force as a source of errors in the numerical simulations. Simulations of an accreting young binary is often performed in the frame co-rotating with the binary. We demonstrate that the specific angular momentum changes spuriously at a shock front, if it is evaluated either solely with the density and velocity at the cell center. We show that the spurious change is erased out if a half of it is evaluated from the numerical flux on the cell surface. We name this method of evaluating the Coriolis force HH type since a half is evaluated from the numerical mass flux and the other half is from the momentum density in the cell. We prove that simulations conserve the momentum measured in the rest frame only when HH type Coriolis force is adopted. The numerical error is serious around shock waves since the difference between the numerical mass flux and momentum density is large there. The shock waves drive gas accretion through angular momentum transfer and should be simulated accurately.

We use two-dimensional hydrodynamic simulations to study the formation of nuclear rings and nuclear spirals and the associated mass inflow rates at the centers of barred galaxies. We find nuclear rings form by the centrifugal barrier that the inflowing gas cannot overcome. The size of nuclear rings depend on various galaxy properties such as the bar strength, the bar pattern speed, and the bulge central density: they are smaller in galaxies with a stronger or slower bar, and with a more centrally concentrated bulge. Even a very weak bar potential can induce nuclear spirals that eventually develop into shocks. In galaxies with high shear, nuclear spirals are tightly wound and the shocks are inclined, forming a circumnuclear disk. On the other hand, galaxies with low shear produce loosely wound spirals and perpendicular shocks, without forming a circumnuclear disk. The mass inflow rates driven by the nuclear spiral shocks are enough to account for the observed level of AGN activities in Seyfert galaxies.

A massive and luminous perturber moving across an opaque gas is subjected to a force different from the gravitational friction that it would experience if it were cold. The heat released by the perturber diffuses in the surrounding gas, where it gives rise to a low density region behind the perturber that exerts a force (that we call heating force) in the direction of motion, thus opposed to the standard dynamical friction. We present numerical simulations with nested meshes that confirm the analytical expression of the heating force in the limits of a low and high Mach number, respectively, and we present simulations that show that the dynamical friction exerted on a cold perturber in a gas with thermal diffusion is markedly different from that in an adiabatic gas. We then present numerical simulations of low-mass protoplanets embedded in opaque, viscous discs, that show that when these bodies have a sufficiently large luminosity their eccentricity and inclination can be excited to values comparable to the aspect ratio of the disc. We finally present numerical experiments with very high resolution that try to resolve the flow within the Bondi sphere, in an attempt to study the dependence of the heating force as a function of the ratio of the diffusive to acoustic times across the Bondi radius.

Recent developments in non-ideal magnetohydrodynamic simulations of protoplanetary disks suggest that instead of being traditional turbulent (viscous) accretion disks, they have a largely laminar flow with accretion driven by large-scale wind torques. These disks are possibly threaded by Hall-effect generated large-scale horizontal magnetic fields. We have examined the dynamics of the corotation region of a low mass planet embedded in such a disk and the evolution of the associated migration torque. These disks lack strong turbulence and associated turbulent diffusion, and the presence of a magnetic field and radial gas flow presents a situation outside the applicability of previous corotation torque theory. We summarize the analytical analysis of the corotation torque, give details on the numerical methods used, and in particular the relative merits of different numerical schemes for the inviscid problem.

With the forthcoming VLBI images of Sgr A* and M87, simulations of accretion flows onto black holes acquire a special importance to aid with the interpretation of the observations and to test the predictions of different accretion scenarios, including those coming from alternative theories of gravity. The Black Hole Accretion Code (BHAC ) is a new multidimensional general-relativistic magnetohydrondynamics (GRMHD) module for the MPI-AMRVAC framework. It exploits its adaptive mesh refinement techniques (AMR) to solve the equations of ideal magnetohydrodynamics in arbitrary curved spacetimes with a significant speedup and saving in computational cost. In a previous work, this was shown using a Generalized Lagrange Multiplier (GLM) to enforce the solenoidal constraint of the magnetic field. While GLM is fully compatible with MPI-AMRVAC's AMR infrastructure, we found that simulations were sensible to the divergence control technique employed, resulting in an improved behavior for those using Constrained Transport (CT). However, cell-centered CT is incompatible with AMR, and several modifications were required to make AMR compatible with staggered CT. We present here preliminary results of these new additions, which achieved machine precision fulfillment of the solenoidal constraint and a significant speedup in a problem close to the intended scientific application.

The Sun is located inside a low density cloud known as the Local Interstellar Cloud (LIC), which is part of a group of nearby clouds known as the Complex of Local Interstellar Clouds (CLIC). All of these clouds are contained within the hot Local Bubble, which contains gas at ∼ 10⁶ K and appears to have been created by multiple supernova explosions. We present new results from our ongoing project on simulating the origins and evolution of the Local Interstellar Medium. We aim to model the origins of the CLIC, especially the way that the clouds managed to survive shock passage and reach their current warm (T ∼ 7000 K) and low density (n ∼ 0.2 cm⁻³) state. We find that the magnetic field is important for maintaining the Local Bubble at a high thermal pressure and helps the clouds rebound more effectively after being shocked. Thermal conduction is necessary to make the temperature in the bubble as uniform as observed. We also find that to reach its current state, the Local Bubble requires multiple supernova explosions. The state of the CLIC and the Local Bubble provide important clues as to how supernova feedback operates in the ISM and it interaction with the different phases of the ISM.

The interstellar medium in our galaxy is a complex system which involves multiphase gas in a wide range of temperatures and densities. Accompanied by physical processes such as cooling, gravity and magnetic fields it is known that the supernovae blast waves can reshuffle the ISM by generating a galactic fountain. In our study, we mainly focus on estimating the fuel rate of the fountain which is defined as the supply rate of material from the gas falling back into the galactic disk. We take into account the effect of cooling by combining the use of non-equilibrium ionization (NEI) and collisional ionization equilibrium (CIE). Dealing with cooling in this manner, we aim to reduce computation time by using a prebuilt CIE cooling table at applicable physical conditions instead of constantly considering time dependent ion fractions for NEI cooling.

In order to model the magnetic field amplification and particle acceleration that takes place in astrophysical shocks, we need a code that can efficiently model the large-scale structure of the shock, while still taking the kinetic aspect of supra-thermal particles into account. Starting from the proven MPI-AMRVAC magnetohydrodynamics code we have created a code that combines the kinetic treatment of the Particle-in-Cell (PIC) method for suprathermal particles with the large-scale efficiency of grid-based hydrodynamics (MHD) to model the thermal plasma, including the use of adaptive mesh refinement. Using this code we simulate astrophysical shocks, varying both the Mach-number and the angle between the magnetic field and the shock to test our code against existing results and study both the evolution of the shock and the behaviour of supra-thermal particles. We find that the combined PIC-MHD method can accurately recover the results that were previously obtained with pure PIC codes. Furthermore, the efficiency of the code allows us to explore the available parameter space to a larger degree than has been done in previous work. Our results suggest that efficient particle acceleration can take place in near-oblique shocks were the magnetic field makes a large angle with the direction of the flow.

We explored the generalised Kippenhan-Schluter model non-isothermal stationary states for solar prominences where we made some improvements. The most important is to use recent observed values of radiative losses to build the piecewise heat balance equation. We explicity found several stationary states without and with magnetic shear solutions for the system of equations in the beta plasma range between 0 and 1,6. Finally, we computed magnetic streamlines field for the different shears. Joining all these elements we obtained a catalogue to study the stability of the prominence candidates.

White dwarfs are compact objects with masses comparable to our Sun, but a radius similar to our Earth. They are the final evolutionary stage for about 95% of all stars in the Galaxy, i.e., for all stars that have a final mass less than the Chandrasekhar mass (about 1.4 times the solar mass), the upper mass limit for which hydrostatic equilibrium can be maintained by the degeneracy pressure of electrons at very high densities. The outermost shell of most white dwarfs contains a convective layer. Even if the latter is very thin (≲ 10 km), it is important for mixing properties, observed radiation, and pulsational stability of the whole object. During a long phase white dwarfs have effective temperatures T_eff of about 10000K ∼ 14000K, since the time scale to reach such temperatures by cooling is already ≈ 10⁹ years. Here, we focus on DA (hydrogen-rich) white dwarfs with T_eff ≈ 12000K. This is at the transition from shallow to deep convection zones. Due to very high gravitational acceleration (∼ 10⁶ g at the surface) the material is overturned about five times per second over the distance of a few kilometers. Numerical simulations of such objects have to be done for a compressible flow and feature highly turbulent granules at the surface, which are qualitatively comparable to the convection cells observed at the surface of the Sun. For this study we compare three white dwarf surface simulations with realistic microphysical properties and full 3D radiative transport. The simulations differ in effective temperature, namely, T_eff = 11800K, 12100K, and 12400K. A statistical analysis of the convective processes as function of T_eff is presented.

We consider the effect of the solar/stellar wind collimation towards the solar/stellar wind rotation axis in the heliosheath between the termination shock and the heliopause/astropause. The collimation is due to the magnetic force produced by the toroidal component of the solar/stellar magnetic field. The collimation leads to formation of a two-jet structure and change of topology of the heliopause. Tube-like shape of the heliopause/astropause is formed instead of the commonly accepted sheet-like shape.

Three different situations are explored in this paper: (1) the Sun/star is at the rest with respect to local interstellar medium (LISM), (2) the Sun/star moves with respect to fully ionized LISM, (3) the Sun/star moves with respect to partially ionized LISM. 3D non-dissipative MHD model results have shown that the tube-like structure is formed in the first two cases. The thickness of the heliosheath depends on the model parameters strongly. The case of the partially ionized LISM is the most realistic one for the heliosphere. In this case the collimation towards the solar poles is also observed in the results of 3D kinetic-MHD modeling. However, the tube-like structure of the heliopause is not seen in the numerical results. We argue that this is due to charge exchange between proton and H atom components that results in displacement of the stagnation points in further downwind of the Sun. Note also that dissipative effects (e.g. reconnection at the current sheet in the non-stationary heliosphere) or instabilities could destroy the effects of the solar wind collimation.

We present an innovative numerical method that solves for the multi-fluid plasma equations, including the transport, frictional, and chemical reactions terms, coupled to full Maxwell's equations. The numerical method features a scheme for the electromagnetic field with a proper scaling for the numerical dissipation, a scheme that solves flows at all speeds regimes (from subsonic to supersonic), and implicit time integration to tackle the stiffness of the system. Verification of the numerical scheme is also presented in a wide variety of plasma conditions.

We present early results from a study addressing the question of how one treats the propagation of incertitude, that is, epistemic uncertainty, in input parameters in astrophysical simulations. As an example, we look at the propagation of incertitude in control parameters for stellar winds in MESA stellar evolution simulations. We apply two methods of incertitude propagation, the Cauchy Deviates method and the Quadratic Response Surface method, to quantify the output uncertainty in the final white dwarf mass given a range of values for wind parameters. The methodology we apply is applicable to the problem of propagating input incertitudes through any simulation code treated as a "black box," i.e. a code for which the algorithmic details are either inaccessible or prohibitively complicated. We have made the tools developed for this study freely available to the community.

We present the first results on the construction of an exascale hyperbolic PDE engine (ExaHyPE), a code for the next generation of supercomputers with the objective to evolve dynamical spacetimes of black holes, neutron stars and binaries. We solve a novel first order formulation of Einstein field equations in the conformal and constraint damping Z4 formulation (CCZ4) coupled to ideal general relativistic magnetohydrodynamics (GRMHD), using divergence-cleaning. We adopt a novel communication-avoiding one-step ADER-DG scheme with an a-posteriori subcell finite volume limiter on adaptive spacetrees. Despite being only at its first stages, the code passes a number of tests in special and general relativity.

We describe gas dynamic and plasma effects for transient plasma jets under different conditions. Our Unified Flow Solver (UFS) is evaluated for simulations of mixed continuum-rarefied flows, transient gas expansion dynamics and collisionless plasma expansion. We identify further advances that are required for hybrid fluid-kinetic simulations of laser-ablation and EEE-induced plasma jets.

We present simulations of magnetic reconnection with a newly developed coupled MHD-PIC code. In this work a global magnetohydrodynamic (MHD) simulation receives kinetic feedback within an embedded region that is modeled by a kinetic particle-in-cell (PIC) code. The PIC code receives initial and boundary conditions from the MHD simulation, while the MHD solution is updated with the PIC state. We briefly describe this coupling mechanism. This method is suitable for simulating magnetic reconnection problems, as we show with the example of reconnection in the coalescence of magnetic islands. We compare the MHD, Hall-MHD, fully PIC and coupled MHD-PIC simulations of the magnetic island coalescence. We find that the kinetic simulations are very different from the MHD and Hall-MHD results, while the coupled MHD-PIC simulations can remedy this discrepancy while saving computing time. The diffusion region is well resolved in the kinetic simulations, which is also captured by the coupled MHD-PIC model. The coupled simulation also reproduces the kinetic Hall magnetic fields correctly. We calculate the reconnection rates and find differences between the MHD and kinetic results. We find that the coupled MHD-PIC code can reasonably reproduce the kinetic reconnection rate when a larger PIC feedback region is used, while still saving significant computing time.

Magnetic reconnection is believed to be the driver of many explosive phenomena in Astrophysics, from solar to gamma-ray flares in magnetars and in the Crab nebula. However, reconnection rates from classic MHD models are far too slow to explain such observations. Recently, it was realized that when a current sheet gets sufficiently thin, the reconnection rate of the tearing instability becomes "ideal", in the sense that the current sheet destabilizes on the "macroscopic" Alfvenic timescales, regardless of the Lundquist number of the plasma. Here we present 2D compressible MHD simulations in the classical, Hall, and relativistic regimes. In particular, the onset of secondary tearing instabilities is investigated within Hall-MHD for the first time. In the frame of relativistic MHD, we summarize the main results from Del Zanna et al. [1]: the relativistic tearing instability is found to be extremely fast, with reconnection rates of the order of the inverse of the light crossing time, as required to explain the high-energy explosive phenomena.

We describe a simple and effective algorithm for solving Poisson's equation in the context of self-gravity within the DISPATCH astrophysical fluid framework. The algorithm leverages the fact that DISPATCH stores multiple time slices and uses asynchronous time-stepping to produce a scheme that does not require any explicit global communication or sub-cycling, only the normal, local communication between patches and the iterative solution to Poisson's equation. We demonstrate that the implementation is suitable for both collections of patches of a single resolution and for hierarchies of adaptively resolved patches. Benchmarks are presented that demonstrate the accuracy, effectiveness and efficiency of the scheme.

We performed numerical MHD simulation of the laboratory experiment for creating plasma jets on a NEODIM laser installation. In this experiment, the plasma ejection is formed as a result of the action of a powerful laser on the target. To describe these processes and simulate plasma flow, we chose a numerical method, boundary and initial conditions. We investigated the picture of the flow and compared it with the experiment. We found the distribution of the density of matter at various distances from the target and at different time, and investigated the possible structures of matter on the surface of the detector.

Since molecules exist in the interstellar cloud and they affect the hydrodynamic evolution through their formation and destruction, physical states of the actual interstellar cloud are different from calculations of conventional hydrodynamics simulations. Furthermore, in the case of running star-forming simulations, conventional hydrodynamics models are not enough to explain molecular lines emitted from clouds such as those detected from the ALMA observatory. In order to simulate the chemical evolution of a hydrodynamic cloud, building an efficient chemical network that contains relevant chemical reactions is crucial for cutting down the computation cost. A key factor for generating an efficient chemical network is to avoid using an abnormally small network that contains only a few reactions because using too small a network does not simulate the effects of molecules accurately. There already exist a few chemical hydrodynamics simulation codes, which provide pre-built reduced networks. Although those prebuilt networks make the simulations simple and light, they cannot be used universally for various clouds under diverse initial chemical compositions and environmental conditions. Therefore, it is necessary to build a more flexible network according to the conditions of the model. In this study, we propose to make an automatic network reduction module which builds an optimized closed network corresponding to the specific simulation conditions. As a preliminary result, we test our module with simple primordial test clouds by comparing our results with those obtained with the full network. In the future study, we will validate our reduction module for primordial cloud models and expand its usage to various physicochemical cases such as AGB stars.

We describe the AMReX suite of astrophysics codes and their application to modeling problems in stellar astrophysics. Maestro is tuned to efficiently model subsonic convective flows while Castro models the highly compressible flows associated with stellar explosions. Both are built on the block-structured adaptive mesh refinement library AMReX. Together, these codes enable a thorough investigation of stellar phenomena, including Type Ia supernovae and X-ray bursts. We describe these science applications and the approach we are taking to make these codes performant on current and future many-core and GPU-based architectures.