Validating time-distance helioseismic inversions for non-separable subsurface profiles of an average supergranule

TL;DR

This study validates helioseismic inversions for non-separable supergranule profiles using synthetic data, demonstrating the potential to reliably estimate their depth despite increased model complexity.

Contribution

It introduces a validation approach for helioseismic inversions with non-separable profiles, moving beyond previous assumptions of separability.

Findings

Successfully recovered the peak depth of supergranule models

Demonstrated inversion robustness with increasing parameters

Identified the importance of measurement noise levels

Abstract

Supergranules are divergent 30-Mm sized cellular flows observed everywhere at the solar photosphere. Their place in the hierarchy of convective structures and their origin remain poorly understood (Rincon et al., 2018). Estimating supergranular depth is of particular interest since this may help point to the underlying physics. However, their subsurface velocity profiles have proven difficult to ascertain. Birch et al. (2006) had suggested that helioseismic inferences would benefit from an ensemble average over multiple realizations of supergranules due to the reduction in realization noise. Bhattacharya et al. (2017) used synthetic forward-modelled seismic wave travel times and demonstrated the potential of helioseismic inversions at recovering the flow profile of an average supergranule that is separable in the horizontal and vertical directions, although the premise of this calculation has since been challenged by Ferret (2019). In this work we avoid this assumption and carry out a validation test of helioseismic travel-time inversions starting from plausible synthetic non-separable profiles of an average supergranule. We compute seismic wave travel times and sensitivity kernels by simulating wave propagation through this background. We find that, while the ability to recover the exact profile degrades based on the number of parameters involved, we are nevertheless able to recover the peak depth of our models in a few iterations where the measurements are presumably above the noise cutoff. This represents an important step towards unraveling the physics behind supergranules, as we start appreciating the parameters that we may reliably infer from a time-distance helioseismic inversion.