Crowded-Field Astrometry with the Space Interferometry Mission - I. Estimating the Single-Measurement Astrometric Bias Arising from Confusion

TL;DR

This paper analyzes the impact of confusion bias from nearby stars on the astrometric measurements of the Space Interferometry Mission, providing a model and tools to estimate and mitigate this effect.

Contribution

It introduces a simple phasor model for confusion bias, along with a software tool to assess its impact on SIM observations, and discusses strategies to reduce this bias.

Findings

Less than 1% of targets are vulnerable to confusion bias.

Confusion may be significant if star density exceeds 0.4 star/arcsec^2.

The model provides a good understanding and estimation of confusion bias effects.

Abstract

The accuracy of position measurements on stellar targets with the future Space Interferometry Mission (SIM) will be limited not only by photon noise and by the properties of the instrument (design, stability, etc.) and the overall measurement program (observing strategy, reduction methods, etc.), but also by the presence of other "confusing" stars in the field of view (FOV). We use a simple "phasor" model as an aid to understanding the main effects of this "confusion bias" in single observations with SIM. This analytic model has been implemented numerically in a computer code and applied to a selection of typical SIM target fields drawn from some of the Key Projects already accepted for the Mission. We expect that less than 1% of all SIM targets will be vulnerable to confusion bias; we show that for the present SIM design, confusion may be a concern if the surface density of field stars exceeds 0.4 star/arcsec^2. We have developed a software tool as an aid to ascertaining the possible presence of confusion bias in single observations of any arbitrary field. Some a priori knowledge of the locations and spectral energy distributions of the few brightest stars in the FOV is helpful in establishing the possible presence of confusion bias, but the information is in general not likely to be available with sufficient accuracy to permit its removal. We discuss several ways of reducing the likelihood of confusion bias in crowded fields. Finally, several limitations of the present semi-analytic approach are reviewed, and their effects on the present results are estimated. The simple model presented here provides a good physical understanding of how confusion arises in a single SIM observation, and has sufficient precision to establish the likelihood of a bias in most cases.