Magnetogram filtering techniques for global coronal modelling

ABSTRACT Global coronal models for space weather forecasting often require filtered-down magnetogram data for their inner boundary prescriptions. Often, the technique of spherical harmonics decomposition is used for this purpose. However, this method leads to large oscillations of the magnetic field intensity around active regions. To avoid these oscillations, in this paper, (i) an adjusted version of spherical harmonics decomposition, (ii) a non-linear diffusion filter, and (iii) a Yaroslavsky filter were implemented in Python and applied to a reference magnetogram with high activity: the HMI (Helioseismic Magnetic Imager) magnetogram for 2022-11-22T01:30:00. First, these filters are analysed separately to show their effects and optimal parameter settings. Afterwards, four simulations with the same mesh size were run with the COolfluid COronal uNstrUcTured (COCONUT) solver to see the impacts of the new boundary prescription on the simulated density, radial velocity, and on the magnetic field topology. Validation was done against electron tomography. The new filters did not present non-physical oscillations, captured the large structures, and removed the small ones. All filters ran in the order of minutes or less of CPU time. Spherical harmonics-based filters were found to be more suitable for coarser, rapid simulations; however, they could lead to inaccurate features. The non-linear diffusion and Yaroslavsky filters were found to be more suitable for a finer grid, resulting in more accurate solution features. Validation against tomography showed a good capture of low-density regions and the equatorial streamer.