Using bootstrap to assess uncertainties of VLBI results I. The method and image-based errors

TL;DR

This paper introduces a bootstrap-based method to accurately estimate uncertainties and significance in VLBI observations, improving over traditional error estimates by accounting for correlated noise and array heterogeneity.

Contribution

The paper presents a novel bootstrap approach for uncertainty estimation in VLBI imaging, addressing correlated noise and array heterogeneity issues.

Findings

Bootstrap errors align better with nominal coverage.

Method effectively handles heterogeneous arrays like Space VLBI.

Can incorporate instrumental calibration factors.

Abstract

Very Long Baseline Interferometric (VLBI) observations of quasar jets enable one to measure many theoretically expected effects. Estimating the significance of observational findings is complicated by the correlated noise in the image plane. A reliable and well justified approach to estimate the uncertainties of VLBI results is needed as well as significance testing criteria. We propose to use bootstrap for both tasks. Using simulations we find that bootstrap-based errors for the full intensity, rotation measure, and spectral index maps have coverage closer to the nominal values than conventionally obtained errors. The proposed method naturally takes into account heterogeneous interferometric arrays (such as Space VLBI) and can be easily extended to account for instrumental calibration factors.