Planetary host stars: Evaluating uncertainties in ultra-cool model atmospheres

TL;DR

This study compares different stellar atmosphere models for M-dwarfs to quantify systematic uncertainties in stellar parameters, highlighting model discrepancies across various parameters and their impact on synthetic photometry.

Contribution

It provides a comprehensive comparison of four major model atmosphere families for M-dwarfs, assessing their differences and uncertainties in stellar parameters and synthetic colours.

Findings

Model results for higher log(g) deviate less between models.

Synthetic IR colours differ by no more than 0.15 dex across models.

Discrepancies decrease with increasing T_eff.

Abstract

M-dwarfs are emerging in the literature as promising targets for detecting low-mass, Earth-like planets. An important step in this process is to determine the stellar parameters of the M-dwarf host star as accurately as possible. Different well-tested stellar model atmosphere simulations from different groups are widely applied to undertake this task. This paper provides a comparison of different model atmosphere families to allow a better estimate of systematic errors on host-star stellar parameter introduced by the use of one specific model atmosphere family only. We present a comparison of the ATLAS9, MARCS, Phoenix and Drift-Phoenix model atmosphere families including the M-dwarf parameter space (T$_{\rm eff}=2500$K$\,...\,$4000K, log(g)=3.0$\,...\,$5.0, [M/H]=$-2.5\,...\,0.5$). We examine the differences in the (T$_{\rm gas}$, p$_{\rm gas}$)-structures, in synthetic photometric fluxes and in colour indices. Model atmospheres results for higher log(g) deviate considerably less between different models families than those for lower log(g) for all T$_{\rm eff}=2500$K$\,...\,$4000K examined. We compiled the broad-band synthetic photometric fluxes for all available model atmospheres (incl. M-dwarfs and brown dwarfs) for the UKIRT WFCAM ZYJHK, 2MASS JHKs and Johnson UBVRI filters, and calculated related colour indices. Synthetic colours in the IR wavelengths diverge by no more than 0.15 dex amongst all model families. For all spectral bands considered, model discrepancies in colour diminish for higher T$_{\rm eff}$ atmosphere simulations. We notice differences in synthetic colours between all model families and observed example data (incl. Kepler 42 and GJ1214).