J/ApJS/261/33 Phot. metallicity prediction of RR Lyrae stars (Dekany+, 2022)
Photometric metallicity prediction of fundamental-mode RR Lyrae stars in the
Gaia optical and Ks infrared wave bands by deep learning.
Dekany I., Grebel E.K.
<Astrophys. J. Suppl. Ser., 261, 33 (2022)>
=2022ApJS..261...33D 2022ApJS..261...33D
ADC_Keywords: Stars, variable; Abundances, [Fe/H]; Photometry, UBVRI;
Surveys; Models
Keywords: RR Lyrae variable stars ; Metallicity ; Light curves ;
Neural networks
Abstract:
RR Lyrae stars are useful chemical tracers thanks to the empirical
relationship between their heavy-element abundance and the shape of
their light curves. However, the consistent and accurate calibration
of this relation across multiple photometric wave bands has been
lacking. We have devised a new method for the metallicity estimation
of fundamental-mode RR Lyrae stars in the Gaia optical G and
near-infrared VISTA Ks wave bands by deep learning. First, an existing
metallicity prediction method is applied to large photometric data
sets, which are then used to train long short-term memory recurrent
neural networks for the regression of the [Fe/H] to the light curves
in other wave bands. This approach allows an unbiased transfer of our
accurate, spectroscopically calibrated I-band formula to additional
bands at the expense of minimal additional noise. We achieve a low
mean absolute error of 0.1 dex and high R2 regression performance of
0.84 and 0.93 for the Ks and G bands, respectively, measured by
cross-validation. The resulting predictive models are deployed on the
Gaia DR2 and VVV inner bulge RR Lyrae catalogs. We estimate mean
metallicities of -1.35dex for the inner bulge and -1.7dex for the
halo, which are significantly less than the values obtained by earlier
photometric prediction methods. Using our results, we establish a
public catalog of photometric metallicities of over 60,000 Galactic
RR Lyrae stars and provide an all-sky map of the resulting RR Lyrae
metallicity distribution. The software code used for training and
deploying our recurrent neural networks is made publicly available in
the open-source domain.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table1.dat 112 55723 *I-band photometric parameters and metallicities of
RRab stars in the bulge and disk areas of the
OGLE Collection of Variable Stars (OCVS)
table4.dat 71 58652 G-band photometric metallicity estimates and basic
light-curve attributes of Gaia DR2 RRab stars
table5.dat 80 4447 KS-band photometric metallicity estimates and
basic light-curve attributes of the VVV bulge
RRab stars discovered by
Dekany & Grebel (2020, J/ApJ/898/46)
--------------------------------------------------------------------------------
Note on table1.dat: Results are shown for stars that pass the following
criteria: Cphi≥0.8, Atot≤1.2, S/N≥50.
--------------------------------------------------------------------------------
See also:
I/345 : Gaia DR2 (Gaia Collaboration, 2018)
J/AJ/108/1016 : Kinematics of local RR lyrae stars. I. (Layden, 1994)
J/AJ/110/2319 : Abundances in RR Lyr variables (Clementini+ 1995)
J/A+A/312/111 : [Fe/H] from RR Lyrae light curves (Jurcsik+, 1996)
J/ApJS/197/29 : Chemical compositions of 11 RR Lyrae (For+, 2011)
J/ApJ/782/59 : Abundances of 8 RR Lyrae subclass C stars (Govea+, 2014)
J/MNRAS/447/2404 : Equivalent width of 21 RR Lyrae stars (Pancino+, 2015)
J/MNRAS/466/2602 : Blazhko effect in Galactic RR Lyrae (Prudil+, 2017)
J/ApJ/848/68 : Abund. & RVs of stable and Blazhko RRc stars (Sneden+, 2017)
J/ApJ/857/54 : JHKs photometry of VVV RR Lyrae stars (Dekany+, 2018)
J/MNRAS/478/4590 : Light curves of RR Lyrae variables in M31 (Tanakul+, 2018)
J/A+A/622/A60 : Gaia DR2 misclassified RR Lyrae list (Clementini+, 2019)
J/ApJ/898/46 : NIR LCs of RRab stars from the VVV survey (Dekany+, 2020)
J/ApJ/908/20 : Field RR Lyrae as galactic probes. II. (Crestani+, 2021)
J/A+A/657/A123 : OGLE IV & Gaia EDR3 data for RR Lyrae (Oliveira+, 2022)
http://ogledb.astrouw.edu.pl/~ogle/OCVS/ : OCVS home page
Byte-by-byte Description of file: table1.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1 A1 --- Field Field identifier (1)
3- 7 I5 --- ID [1/68197] Object identifier within the Field (1)
9- 10 I2 h RAh Hour of Right Ascension (J2000)
12- 13 I2 min RAm Minute of Right Ascension (J2000)
15- 19 F5.2 s RAs Second of Right Ascension (J2000)
21 A1 --- DE- Sign of the Declination (J2000)
22- 23 I2 deg DEd Degree of Declination (J2000)
25- 26 I2 arcmin DEm Arcminute of Declination (J2000)
28- 31 F4.1 arcsec DEs Arcsecond of Declination (J2000)
33- 38 F6.3 [Sun] [Fe/H] [-8.7/3.2] Metallicity estimate
40- 45 F6.3 mag Imag [10.78/21.3] Mean I band magnitude
47- 51 I5 --- Nep [16/15016] Number of epochs in the light curve
53- 60 F8.6 d Per [0.27/1] Period
62- 66 F5.3 --- Atot [0.016/1.2] Total (peak-to-valley) amplitude
68- 72 F5.3 --- A1 [0.006/0.5] The A1 Fourier parameter
74- 78 F5.3 --- A2 [0.001/0.3] The A2 Fourier parameter
80- 84 F5.3 --- A3 [0/0.2] The A3 Fourier parameter
86- 92 F7.4 --- phi21 [5.16/10.5] The φ21 Fourier parameter
94- 99 F6.4 --- phi31 [2.7/9.1] The φ31 Fourier parameter
101-105 F5.3 --- Cphi [0.8/1] Phase coverage (1-maximum phase lag)
107-112 F6.1 --- S/N [50/8796] Signal-to-Noise
--------------------------------------------------------------------------------
Note (1): Field and ID designations, as:
b = bulge (48153 occurrences); in Simbad
d = disk (7570 occurrences); in Simbad
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table4.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 19 I19 --- Gaia Gaia DR2 source identifier
21- 26 F6.3 [Sun] [Fe/H] [-3.1/-0.06] Mean Metallicity prediction
from a model ensemble
28- 32 F5.3 [Sun] e_[Fe/H] [0.01/0.8] Standard deviation of [Fe/H]
34- 39 F6.3 mag Gmag [9.4/21] Mean Gaia DR2 G band magnitude
41- 43 I3 --- Nep [21/247] Number of epochs in the light curve
45- 52 F8.6 d Per [0.3/0.98] Period
54- 58 F5.3 --- Atot [0.12/1.4] Total (peak-to-valley) amplitude
60- 64 F5.3 --- Cphi [0.85/1] Phase coverage (1-maximum phase lag)
66- 71 F6.1 --- S/N [30/5443] Signal-to-Noise
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table5.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 5 I5 --- [DG2020] [7/13250] Dekany & Grebel 2020, J/ApJ/898/46
identifier
7- 8 I2 h RAh [17/18] Hour of Right Ascension (J2000)
10- 11 I2 min RAm Minute of Right Ascension (J2000)
13- 17 F5.2 s RAs Second of Right Ascension (J2000)
19 A1 --- DE- Sign of the Declination (J2000)
20- 21 I2 deg DEd Degree of Declination (J2000)
23- 24 I2 arcmin DEm Arcminute of Declination (J2000)
26- 29 F4.1 arcsec DEs Arcsecond of Declination (J2000)
31- 36 F6.3 [Sun] [Fe/H] [-2.6/-0.06] Metallicity estimate
from a model ensemble
38- 42 F5.3 [Sun] e_[Fe/H] [0.008/0.5] Standard deviation of [Fe/H]
44- 49 F6.3 mag Ksmag [12/16.3] Mean KS band magnitude
51- 53 I3 --- Nep [33/469] Number of epochs in the light curve
55- 62 F8.6 d Per [0.29/0.98] Period
64- 68 F5.3 --- Atot [0.058/0.7] Total (peak-to-valley) amplitude
70- 74 F5.3 --- Cphi [0.8/0.99] Phase coverage (1-maximum phase lag)
76- 80 F5.1 --- S/N [30.4/272] Signal-to-Noise
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Prepared by [AAS], Emmanuelle Perret [CDS] 29-Sep-2022