Variability in GRMHD simulations of Sgr A$^*$: Implications for EHT closure phase observations

TL;DR

This study investigates the variability of closure phases in GRMHD simulations of Sgr A* to inform EHT observations, revealing that variability depends on baseline size and accretion flow models, impacting image reconstruction.

Contribution

It provides the first detailed analysis of closure phase variability in GRMHD models of Sgr A*, highlighting implications for observational strategies and image reconstruction methods.

Findings

Closure phase variability increases with baseline size.

MAD models show less variability than SANE models.

Variability impacts image reconstruction techniques.

Abstract

The observable quantities that carry the most information regarding the structures of the images of black holes in the interferometric observations with the Event Horizon Telescope are the closure phases along different baseline triangles. We use long time span, high cadence, GRMHD+radiative transfer models of Sgr A$^*$ to investigate the expected variability of closure phases in such observations. We find that, in general, closure phases along small baseline triangles show little variability, except in the cases when one of the triangle vertices crosses one of a small regions of low visibility amplitude. The closure phase variability increases with the size of the baseline triangle, as larger baselines probe the small-scale structures of the images, which are highly variable. On average, the jet-dominated MAD models show less closure phase variability than the disk-dominated SANE models, even in the large baseline triangles, because the images from the latter are more sensitive to the turbulence in the accretion flow. Our results suggest that image reconstruction techniques need to explicitly take into account the closure phase variability, especially if the quality and quantity of data allow for a detailed characterization of the nature of variability. This also implies that, if image reconstruction techniques that rely on the assumption of a static image are utilized, regions of the $u-v$ space that show a high level of variability will need to be identified and excised.