J/ApJ/798/122 Abundances from SEGUE Stellar Parameters Pipeline (Miller+, 2015)
A machine-learning method to infer fundamental stellar parameters from
photometric light curves.
Miller A.A., Bloom J.S., Richards J.W., Lee Y.S., Starr D.L., Butler N.R.,
Tokarz S., Smith N., Eisner J.A.
<Astrophys. J., 798, 122 (2015)>
=2015ApJ...798..122M 2015ApJ...798..122M
ADC_Keywords: Stars, variable ; Abundances, [Fe/H] ; Effective temperatures ;
Surveys
Keywords: methods: data analysis; methods: statistical; stars: general;
stars: statistics; stars: variables: general; surveys
Abstract:
A fundamental challenge for wide-field imaging surveys is obtaining
follow-up spectroscopic observations: there are >109 photometrically
cataloged sources, yet modern spectroscopic surveys are limited to
∼fewx106 targets. As we approach the Large Synoptic Survey Telescope
era, new algorithmic solutions are required to cope with the data
deluge. Here we report the development of a machine-learning framework
capable of inferring fundamental stellar parameters (Teff, logg, and
[Fe/H]) using photometric-brightness variations and color alone. A
training set is constructed from a systematic spectroscopic survey of
variables with Hectospec/Multi-Mirror Telescope. In sum, the training
set includes ∼9000 spectra, for which stellar parameters are measured
using the SEGUE Stellar Parameters Pipeline (SSPP). We employed the
random forest algorithm to perform a non-parametric regression that
predicts Teff, logg, and [Fe/H] from photometric time-domain
observations. Our final optimized model produces a cross-validated rms
error (RMSE) of 165K, 0.39dex, and 0.33dex for Teff, logg, and
[Fe/H], respectively. Examining the subset of sources for which the
SSPP measurements are most reliable, the RMSE reduces to 125K,
0.37dex, and 0.27dex, respectively, comparable to what is achievable
via low-resolution spectroscopy. For variable stars this represents a
~12%-20% improvement in RMSE relative to models trained with
single-epoch photometric colors. As an application of our method, we
estimate stellar parameters for ∼54000 known variables. We argue that
this method may convert photometric time-domain surveys into
pseudo-spectrographic engines, enabling the construction of extremely
detailed maps of the Milky Way, its structure, and history.
Description:
To facilitate the construction of a large training set, we conducted a
systematic, spectroscopic survey of variable sources in Stripe 82, an
equatorial, ∼315deg2 field that was repeatedly imaged by the SDSS.
Spectroscopic observations were carried out, during 2011 June and
July, using the multi-fiber Hectospec instrument on the MMT. In sum,
we obtained 5914 Hectospec spectra of 5825 unique sources.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
table3.dat 77 5914 SEGUE Stellar Parameters Pipeline (SSPP) stellar
parameters for Hectospec targets
table5.dat 45 53781 Final model predictions for the University of
Washington variable source catalog (UWVSC) sources
--------------------------------------------------------------------------------
See also:
VII/260 : The SDSS-DR7 quasar catalog (Schneider+, 2010)
VII/233 : The 2MASS Extended sources (IPAC/UMass, 2003-2006)
J/AJ/145/13 : Metal-poor stars from SDSS/SEGUE. I. Abundances (Aoki+, 2013)
J/ApJS/203/32 : Machine-learned ASAS Classif. Cat. (MACC) (Richards+, 2012)
J/AJ/142/112 : KIC photometric calibration (Brown+, 2011)
J/MNRAS/414/2602 : Automated classification of HIP variables (Dubath+, 2011)
J/AJ/141/90 : SEGUE stellar parameter pipeline. V. (Lee+, 2011)
J/ApJ/728/26 : QSO selection based on photom. variability (Macleod+, 2011)
J/A+A/494/739 : Automatic classification of OGLE variables (Sarro+, 2009)
J/AJ/134/973 : SDSS Stripe 82 star catalogs (Ivezic+, 2007)
J/AJ/138/466 : NSVS variables automated classification (Hoffman+, 2009)
J/AJ/136/2070 : SEGUE stellar parameter pipeline. III. (Allende Prieto+, 2008)
J/AJ/136/2050 : SEGUE stellar parameter pipeline. II. (Lee+, 2008)
J/AJ/134/973 : SDSS Stripe 82 Variable Source Catalog (Ivezic+, 2007)
J/A+A/475/1159 : Supervised classif. of variable stars (Debosscher+, 2007)
J/PASP/110/863 : A Stellar Spectral Flux Library: 1150 - 25000 A (Pickles 1998)
Byte-by-byte Description of file: table3.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 7 I7 --- ID Source identifier in the UWVSC (G1)
9- 12 A4 --- P [1-9. 10] Target priority for Hectospec
observations (1)
14- 18 A5 --- Flags Analysis flags returned by the SSPP (2)
20- 26 F7.1 K Teff [4006/8643]?=-9999 Effective temperature
28- 34 F7.1 K e_Teff ?=-9999 Uncertainty in Teff
36- 37 I2 --- o_Teff [0/10] Number of methods used by SSPP
to derive Teff
39- 46 F8.2 [cm/s2] log(g) [0.7/5]?=-9999 Log surface gravity
48- 55 F8.2 [cm/s2] e_log(g) [0/1.6]?=-9999 Uncertainty in log(g)
57 I1 --- o_log(g) [0/7] Number of methods used by SSPP to
derive log(g)
59- 66 F8.2 [Sun] [Fe/H] [-3.4/0.7]?=-9999 Metallicity
68- 75 F8.2 [Sun] e_[Fe/H] [0/1.1]?=-9999 Uncertainty in [Fe/H]
77 I1 --- o_[Fe/H] [0/9] Number of methods used by SSPP to
derive [Fe/H]
--------------------------------------------------------------------------------
Note (1): Each target was assigned a relative priority based on its brightness
and total number of observations (see section 2.2.2).
Criteria for the priorities as follows:
1 = quasar-like targets, likely periodic variables, or bright sources with
either mean observed u, g, or r≤15mag, or i≤14.8mag, or z<14.7mag.
2 = Targets outside the main stellar locus in color-color space
(see Figure 1) or those with ugriz≤16mag.
2.5 = Targets are either extremely bright, ugriz≤14.6mag, or quasar-like,
or likely periodic variables, or high-amplitude variables.
Only those can have multiple entries in the catalog.
3 = Targets have been observed ≥30 times in the r band and have
ugriz≤17mag.
4 = Targets with ugriz≤17mag.
5 = Targets have been observed ≥30 times in the r band and have
ugriz≤18mag.
6 = Targets with ugriz≤18mag.
7 = Targets have been observed ≥30 times in the r band and have
ugriz≤18.5mag.
8 = Targets with ugriz≤18.5mag.
9 = Targets have been observed ≥30 times in the r band and have
ugriz≤19mag.
10 = Targets with ugriz≤19mag.
Note (2): Flags of SEGUE Stellar Parameters Pipeline, see
Lee et al. 2008AJ....136.2022L 2008AJ....136.2022L, table 7:
----------------------------------------------------------------------------
Position Flag Description Category Param
----------------------------------------------------------------------------
First
n Appears normal --- Yes
D Likely white dwarf Critical No
d Likely sdO or sdB Critical No
H Hot star with Teff>10000K Critical No
h Helium line detected, possibly very hot star Critical No
l Likely late type solar abundance star Cautionary Yes
E Emission lines in spectrum Critical No
S Sky spectrum Critical No
V No radial-velocity information Critical No
N Very noisy spectrum Cautionary Yes
Second
n Appears normal --- Yes
C The photometric g-r color may be incorrect Cautionary Yes
Third
n Appears normal --- Yes
B Unexpected Hα strength predicted Cautionary Yes
from Hδ
Fourth
n Appears normal --- Yes
G Strong G-band feature Cautionary Yes
g Mild G-band feature Cautionary Yes
Fifth
n Appears normal --- Yes
P Parameters reported for 5.0≤S/N<10.0 Cautionary Yes
N No parameters Critical No
----------------------------------------------------------------------------
The nominal condition for the five letter flag combination is "nnnnn", which
indicates that the SSPP is satisfied that a given stellar spectrum (and its
reported g-r colors and S/N) has passed all of the tests that have been
performed, and the stellar parameters should be considered well determined.
--------------------------------------------------------------------------------
Byte-by-byte Description of file: table5.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 7 I7 --- ID Source identifier in the UWVSC (G1)
9- 18 F10.6 deg RAdeg Right Ascension in decimal degrees (J2000)
20- 28 F9.6 deg DEdeg Declination in decimal degrees (J2000)
30- 34 F5.0 K Teff [4116/8422] Effective temperature
36- 39 F4.2 [cm/s2] log(g) [2.3/4.6] Log surface gravity
41- 45 F5.2 [Sun] [Fe/H] [-2.3/-0.01] Metallicity
--------------------------------------------------------------------------------
Global note:
Note (G1): As a byproduct of a search for standard stars in Stripe 82,
researchers at the University of Washington constructed a publicly
available variable source catalog (UWVSC; Ivezic et al.
2007, J/AJ/134/973; Sesar et al. 2007AJ....134.2236S 2007AJ....134.2236S;
http://www.astro.washington.edu/users/ivezic/sdss/catalogs/S82variables.html).
--------------------------------------------------------------------------------
History:
From electronic version of the journal
(End) Greg Schwarz [AAS], Emmanuelle Perret [CDS] 18-May-2015