ICARUS, a new inner heliospheric model with a flexible grid

¹ Centre for mathematical Plasma Astrophysics, KU Leuven, Celestijnenlaan 200B, 3001 Leuven, Belgium
e-mail: christine.verbeke@oma.be; cgjmverbeke@gmail.com
² Solar–Terrestrial Centre of Excellence – SIDC, Royal Observatory of Belgium, 1180 Brussels, Belgium
³ Institute of Physics, University of Maria Curie-Skłodowska, Pl. M. Curie-Skłodowska 5, 20–031 Lublin, Poland

Received: 6 August 2021
Accepted: 5 January 2022

Abstract

Context. Simulating the propagation and predicting the arrival time of coronal mass ejections (CMEs) in the inner heliosphere with a full three-dimensional (3D) magnetohydrodynamic (MHD) propagation model requires a significant amount of computational time. For CME forecasting purposes, multiple runs may be required for different reasons such as ensemble modeling (uncertainty on input parameters) and error propagation. Moreover, higher resolution runs may be necessary, which also requires more CPU time, for example for the prediction of solar energetic particle acceleration and transport or in the framework of more in-depth studies about CME erosion and/or deformation during its evolution.

Aims. In this paper we present ICARUS, a new inner heliospheric model for the simulation of a steady background solar wind and the propagation and evolution of superposed CMEs. This novel model has been implemented within the MPI-AMRVAC framework which enables the use of stretched grids and solution adaptive mesh refinement (AMR). The usefulness and efficiency (speed-up) of these advanced features are explored. In particular, we model a typical solar wind with ICARUS and then launch a simple cone CME and follow its evolution. We focus on the effect of radial grid stretching and two specific methods or criteria to trigger solution AMR on this typical simulation run.

Methods. For the solar background wind simulation run, we limited the mesh refinement to the area(s) of interest, in this case a co-rotating interaction region (CIR). For the CME evolution run, on the other hand, we apply AMR where the CME is located by the use of a tracing function. As such, the grid is coarsened again after the CME has passed.

Results. The implemented AMR is flexible and only refines the mesh in a particular sector of the computational domain, for example around the Earth or a single CIR, and/or for a particular feature such as CIR or CME shocks. Radial grid stretching alone yields speed-ups of up to 4 and more, depending on the resolution. Combined with solution adaptive mesh refinement, the speed-ups can be much larger depending on the complexity of the simulation (e.g., number of CIRs in the background wind, number of CMEs) and on the chosen AMR criteria, thresholds and the number of refinement levels.

Conclusions. The ICARUS model implemented in the MPI-AMRVAC framework is a new inner heliospheric 3D MHD model that uses grid stretching as well as AMR techniques. The flexibility in the grid and its resolution allows an optimization of the computational time required for CME propagation simulations for both scientific and forecasting purposes.