Computing nonlinear force free coronal magnetic fields

TL;DR

This paper reviews methods for computing nonlinear force-free coronal magnetic fields, compares two numerical approaches, and discusses their performance and challenges, especially in noisy boundary data scenarios.

Contribution

It introduces and compares an optimization-based method with an MHD relaxation method for nonlinear force-free field extrapolation, highlighting their performance and noise sensitivity.

Findings

Optimization method's convergence is affected by boundary noise.

Comparison shows differences in computational efficiency and accuracy.

Adding noise to boundary conditions impacts the reliability of the extrapolation.

Abstract

Knowledge of the structure of the coronal magnetic field is important for our understanding of many solar activity phenomena, e.g. flares and CMEs. However, the direct measurement of coronal magnetic fields is not possible with present methods, and therefore the coronal field has to be extrapolated from photospheric measurements. Due to the low plasma beta the coronal magnetic field can usually be assumed to be approximately force free, with electric currents flowing along the magnetic field lines. There are both observational and theoretical reasons which suggest that at least prior to an eruption the coronal magnetic field is in a nonlinear force free state. Unfortunately the computation of nonlinear force free fields is way more difficult than potential or linear force free fields and analytic solutions are not generally available. We discuss several methods which have been proposed to compute nonlinear force free fields and focus particularly on an optimization method which has been suggested recently. We compare the numerical performance of a newly developed numerical code based on the optimization method with the performance of another code based on an MHD relaxation method if both codes are applied to the reconstruction of a semi-analytic nonlinear force-free solution. The optimization method has also been tested for cases where we add random noise to the perfect boundary conditions of the analytic solution, in this way mimicking the more realistic case where the boundary conditions are given by vector magnetogram data. We find that the convergence properties of the optimization method are affected by adding noise to the boundary data and we discuss possibilities to overcome this difficulty.