A test of Gaia Data Release 1 parallaxes: implications for the local distance scale

TL;DR

This study compares Gaia DR1 parallaxes with Cepheid-based distances, finding good overall agreement and implications for the local Hubble constant, while highlighting potential systematic issues in early Gaia data.

Contribution

First validation of Gaia DR1 parallaxes against Cepheid distances, providing insights into their accuracy and impact on the local Hubble constant estimation.

Findings

Gaia DR1 parallaxes agree well with Cepheid predictions

Systematic underestimation in bright star parallaxes confirmed by HST data

Implication of a 0.3% decrease in the local Hubble constant estimate

Abstract

We present a comparison of Gaia Data Release 1 (DR1) parallaxes with photometric parallaxes for a sample of 212 Galactic Cepheids at a median distance of 2~kpc, and explore their implications on the distance scale and the local value of the Hubble constant H_0. The Cepheid distances are estimated from a recent calibration of the near-infrared Period-Luminosity P-L relation. The comparison is carried out in parallax space, where the DR1 parallax errors, with a median value of half the median parallax, are expected to be well-behaved. With the exception of one outlier, the DR1 parallaxes are in remarkably good global agreement with the predictions, and the published errors may be conservatively overestimated by about 20%. The parallaxes of 9 Cepheids brighter than G = 6 may be systematically underestimated, trigonometric parallaxes measured with the HST FGS for three of these objects confirm this trend. If interpreted as an independent calibration of the Cepheid luminosities and assumed to be otherwise free of systematic uncertainties, DR1 parallaxes would imply a decrease of 0.3% in the current estimate of the local Hubble constant, well within their statistical uncertainty, and corresponding to a value 2.5 sigma (3.5 sigma if the errors are scaled) higher than the value inferred from Planck CMB data used in conjunction with Lambda-CDM. We also test for a zeropoint error in Gaia parallaxes and find none to a precision of ~20 muas. We caution however that with this early release, the complete systematic properties of the measurements may not be fully understood at the statistical level of the Cepheid sample mean, a level an order of magnitude below the individual uncertainties. The early results from DR1 demonstrate again the enormous impact that the full mission will likely have on fundamental questions in astrophysics and cosmology.