We have two types of benchmark that we consider in WP125500 — "benchmark data sets" and "benchmark targets". These categories are not mutually exclusive, e.g., measurements of an eclipsing binary may be used as benchmark data set, and the same star can also be a benchmark target.

Benchmark data sets

Benchmark data sets will be used to test and improve methods and models that will be used to deliver PLATO data products. For example ...

• Kepler photometry for bright solar-like stars plus high resolution spectroscopy can be analysed using the PLATO tools under development for asteroseismology and spectral analysis. The radius of the star from asteroseismology plus the effective temperature from the spectral analysis can be combined to estimate the distances to these stars. These results can then be compared to direct distance measurements from the Gaia mission.
• The radii of nearby solar-type stars from asteroseismology using photometry from the TESS mission can be tested against the radius measured directly using the Gaia parallax and the stars' angular diameters measured using interferometry.
• Precise and accurate mass, radius and effective temperature measurements for normal stars in eclipsing binary systems can be used to calibrate free parameters in stellar models, e.g., Valle et al. 2017A&A...600A..41V.
• Masses of stars in binaries derived from asterosesimology using TESS photometry can be tested against the masses measured directly using the spectroscopic orbit and astrometry.

Benchmark targets

Benchmark targets are stars that can be observed by the PLATO mission where independent measurements of the star's properties are available. Benchmark targets will be to make an "end-to-end" assessment of the accuracy and precision of PLATO data products, particularly the mass, radius and age estimates from asteroseismology.

The quality of the measurements available to calibrate and assess the various PLATO data products varies depending on the type of benchmark star observed. The following classification will be used to help in the prioritisation of tasks related to establishing a set of benchmark targets.

Level B1

Direct, accurate and precise measurements of a star's fundamental properties with negligible model dependence, e.g., masses of stars in double-lined (SB2) spectroscopic binaries.

Level B2

Properties inferred using stellar models or other theoretical input where the level of systematic errors due to the model dependence is well understood, e.g., ages of stars in clusters; mass, radius and age estimates for stars from asteroseismology.

Level B3

Stellar properties inferred from empirical relations, e.g., ages from gyrochronology.

Note that a benchmark target may be classified differently according to the stellar property under consideration, e.g., the radius of star from its measured angular diameter and parallax is Level B1, the mass of this star estimated from isochrones is Level B2 , and the age of this star from its rotation rate is Level B3.

A priority for WP125500 is to establish a set of Level B1 benchmark targets to calibrate and assess PLATO data products, supplemented by a set of Level B2 benchmark targets for stellar properties where it is not possible or practical to establish Level 1 benchmark targets. Level B3 benchmark targets can be used in statistical studies to assess the reliability of PLATO data products, but should be avoided as calibrators because they may introduce systematic errors that are very hard to quantify, e.g., is the age scale from gyrochronology of "single" stars applicable to all planet host stars?

Establishing benchmarks for PLATO will require the development of new and improved techniques that can be tested against benchmark data sets, e.g., radii of Kepler targets estimated using methods tested against directly measured angular diameters for similar nearby stars would promote the asteroseismic radii from benchmark level B2 to B1.