Comparison between 2.5D and 3D simulations of coronal mass ejections

Abstract

Context. The shocks and magnetic clouds related to Coronal Mass Ejections (CMEs) in the solar corona and interplanetary space (IP) play an important role in the study of space weather. In order to study the evolution of these IP shocks, numerical simulations of a simplified CME model were performed. Aims. In an earlier study, the effect of the background wind on the evolution of interplanetary shock waves was investigated, where the computations were carried out under the assumption of axial symmetry. The assumption of axial symmetry might be a good approach for the solar corona under conditions of solar minimum, but for the study of CMEs this assumption is definitely no longer valid as CMEs possess clearly a fully three dimensional (3D) structure. From this perspective, the previous simulations were repeated, but now in a three dimensional set-up in order to point out the differences between the 2.5D and 3D simulations and to check the quality and reliability of the 2.5D simulations. Methods. The computations were performed in the framework of ideal magnetohydrodynamics (MHD) and to advance the ideal MHD equations in time a parallel finite volume code with explicit upwind solver was used. The shock waves are generated in a similar way in both the 3D and 2.5D simulations, namely by a simple density-blob model. The 3D and 2.5D simulations are all performed with the same numerical methods and on comparable grids, such that the differences between the simulations are purely due to the dimensionality of the problem, and/or the initial parameters for the CME generation. Results. Three different axisymmetric simulations of CME propagation are compared with the fully three dimensional computation. The 2.5D simulations differ from each other in the parameters used for CME initiation. In a first simulation, the same initial parameters as for the 3D case were taken, in a second simulation the initial amount of mass in the 2.5D and 3D CME was the same, and in a third simulation they had a comparable amount of momentum. It turned out that the latter one compared best with the 3D results. Conclusions. As 2.5D computations are computationally much cheaper than 3D computations, we conclude that the 2.5D simulations of the CME evolution are a good first approach and resemble well the 3D result, provided that the appropriate initiation parameters are chosen.