Iterative removal of sources to model the turbulent electromotive force

TL;DR

This paper introduces IROS, a novel iterative method inspired by radio astronomy algorithms, to accurately compute dynamo tensors from MHD simulations, improving robustness and noise handling over previous techniques.

Contribution

The paper presents IROS, a new generalization of the CLEAN algorithm, for modeling turbulent electromotive force in dynamo theory from MHD simulation data.

Findings

IROS reliably recovers dynamo coefficients from simulations.

It effectively incorporates priors during inversion.

It performs well even with high noise levels in synthetic data.

Abstract

We describe a novel method to compute the components of dynamo tensors from direct magnetohydrodynamic (MHD) simulations. Our method relies upon an extension and generalisation of the standard H\"ogbom CLEAN algorithm widely used in radio astronomy to systematically remove the impact of the strongest beams onto the corresponding image. This generalisation, called the Iterative Removal of Sources (IROS) method, has been adopted here to model the turbulent electromotive force (EMF) in terms of the mean magnetic fields and currents. Analogous to the CLEAN algorithm, IROS treats the time series of the mean magnetic field and current as beams that convolve with the dynamo coefficients which are treated as (clean) images to produce the EMF time series (the dirty image). We apply this method to MHD simulations of galactic dynamos, to which we have previously employed other methods of computing dynamo coefficients such as the test-field method, the regression method, as well as local and non-local versions of the singular value decomposition (SVD) method. We show that our new method reliably recovers the dynamo coefficients from the MHD simulations. It also allows priors on the dynamo coefficients to be incorporated easily during the inversion, unlike in earlier methods. Moreover, using synthetic data, we demonstrate that it may serve as a viable post-processing tool in determining the dynamo coefficients, even when the power of additive noise to the EMF is twice as much the actual EMF.