9
$\begingroup$

For just hobbyist purposes (3D starmap) - I'm trying to infer temperature from EDR3 data to make a rough assessment of stellar type.

I tried applying figure 2 from this doc: https://arxiv.org/pdf/1008.0815.pdf

$ \log(T_{\mathit{eff}}) = 3.999−0.654(C_{XP})+0.709(C_{XP})^2−0.316(C_{XP})^3 $

Where $C_{XP} = G_{BP} - G_{RP}$ (from Gaia TAP over Astroquery)

However, it seems to only work for G type stars when comparing for known $T_{\mathit{eff}}$ for the same stars identified in SIMBAD. So with that formula I get 5619 from SIMBAD but 5630 for the G type star.

However, for M stars, with temperature in SIMBAD of 3111, I'm getting values less than 1.

Is there another formula to use for dimmer (and brighter) stars? Or perhaps someone has done preprocessing that is available via TAP somewhere else?

Similarly, I'm curious if there are any groups out there that have identified possible binary/multi-star systems.

https://gitlab.com/godotuniverse/galaxygen/-/blob/main/run_data.py

Thanks for any pointers!

$\endgroup$
0

2 Answers 2

3
$\begingroup$

The paper you got your relation from states it is only valid for C_XP<1.5. i.e. For K-type stars and hotter. If you extrapolate a non-linear polynomial fit outside its limits of application then don't be surprised if it returns nonsense values.

You can then see (Fig 9) that the conversion from colour to temperature for redder colours and cooler stars becomes highly sensitive to metallicity, so unfortunately there is no one-to-one mapping between Bp-Rp colour and temperature for those stars.

You might have more luck using G-Rp, which should be less sensitive to metallicity.

$\endgroup$
3
$\begingroup$

The GAIA database includes teff_gspphot as an existing parameter. The methods used to calculate temperatures are complex, see for example:

https://gea.esac.esa.int/archive/documentation/GDR2/Data_analysis/chap_cu8par/sec_cu8par_process/ssec_cu8par_process_priamteff.html

"For our temperature estimation, we refrain from using a simplistic polynomial model, which would make too restrictive a-priori assumptions about the mathematical form of the colour-temperature relation. Instead, we use GBP−G and G−GRP as features to estimate Teff using extremely randomised trees (Geurts et al. 2006, hereafter ExtraTrees). This machine-learning algorithm comes up with a non-parametric model for the colour-temperature relation, which is far more general than a model of polynomial class."

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .