J/ApJ/954/141 The MADGICS 8621Å DIB cat. from GaiaDR3 RVS (Saydjari+, 2023)
Measuring the 8621Å Diffuse Interstellar Band in Gaia DR3 RVS spectra:
obtaining a clean catalog by marginalizing over stellar types.
Saydjari A.K., Uzsoy A.S. M, Zucker C., Peek J.E.G., Finkbeiner D.P.
<Astrophys. J., 954, 141 (2023)>
=2023ApJ...954..141S 2023ApJ...954..141S
ADC_Keywords: Interstellar medium; Spectra, optical; Equivalent widths;
Space velocities; Reddening; Stars, distances;
Parallaxes, trigonometric
Keywords: Diffuse interstellar bands ; Astronomy data reduction ; Catalogs
Abstract:
Diffuse interstellar bands (DIBs) are broad absorption features
associated with interstellar dust and can serve as chemical and
kinematic tracers. Conventional measurements of DIBs in stellar
spectra are complicated by residuals between observations and best-fit
stellar models. To overcome this, we simultaneously model the spectrum
as a combination of stellar, dust, and residual components, with full
posteriors on the joint distribution of the components. This
decomposition is obtained by modeling each component as a draw from a
high-dimensional Gaussian distribution in the data space (the observed
spectrum)-a method we call "Marginalized Analytic Data-space Gaussian
Inference for Component Separation" (MADGICS). We use a data-driven
prior for the stellar component, which avoids missing stellar features
not well modeled by synthetic spectra. This technique provides
statistically rigorous uncertainties and detection thresholds, which
are required to work in the low signal-to-noise regime that is
commonplace for dusty lines of sight. We reprocess all public Gaia DR3
RVS spectra and present an improved 8621Å DIB catalog, free of
detectable stellar line contamination. We constrain the rest-frame
wavelength to 8623.14±0.087Å (vacuum), find no significant
evidence for DIBs in the Local Bubble from the 1/6th of RVS spectra
that are public, and show unprecedented correlation with kinematic
substructure in Galactic CO maps. We validate the catalog, its
reported uncertainties, and biases using synthetic injection tests. We
believe MADGICS provides a viable path forward for large-scale
spectral line measurements in the presence of complex spectral
contamination.
Description:
The primary data used in this work are the Gaia DR3 RVS spectra, which
were obtained from the Gaia archive (Gaia Col. 2016A&A...595A...1G 2016A&A...595A...1G ;
see I/337, superseded by I/355). These are optical spectra spanning
8460-8700Å.
The primary Gaia analysis of these spectra falls under the General
Stellar Parameteriser-spectroscopy (GSP-Spec) module, which is a
purely spectroscopic analysis to obtain chemophysical stellar
parameters from combined RVS spectra (Recio-Blanco+ 2023A&A...674A..29R 2023A&A...674A..29R).
See Section 2.
File Summary:
--------------------------------------------------------------------------------
FileName Lrecl Records Explanations
--------------------------------------------------------------------------------
ReadMe 80 . This file
madgics.dat 498 7789 *MADGICS 8621Å Diffuse Interstellar Band (DIB)
catalog
--------------------------------------------------------------------------------
Note on madgics.dat: Along with the main catalog file including the MADGICS
("Marginalized Analytic Data-space Gaussian Inference for Component
Separation") method output, we also provide a subset of values in the
Gaia DR3 catalog (see I/355), the Gaia DR3 DIB cat. (2023A&A...674A..40G 2023A&A...674A..40G),
the Bailer-Jones distance catalog (CBJ; see I/352), and the result of
dustmap queries along the lines of sight. Any other value desired from
external catalogs can be obtain by cross-match or using the Gaia
source ID.
--------------------------------------------------------------------------------
See also:
I/337 : Gaia DR1 (Gaia Collaboration, 2016)
V/146 : LAMOST DR1 catalogs (Luo+, 2015)
I/352 : Distances to 1.47 billion stars in Gaia EDR3 (Bailer-Jones+, 2021)
I/355 : Gaia DR3 Part 1. Main source (Gaia Collaboration, 2022)
J/A+A/550/A103 : Model 1D (LHD) and 3D (CO5BOLD) sp. (Allende Prieto+, 2013)
J/ApJ/774/72 : 19 DIBs and 8 other interstellar lines (Kos+, 2013)
J/A+A/550/A108 : DIB in VLT-FLAMES Tarantula Survey (van Loon+, 2013)
J/A+A/585/A12 : Nearby early-type stars spectroscopic survey (Bailey+, 2016)
J/ApJ/885/131 : ∼200 high-mass SFR plx & PMs with VLBI (Reid+, 2019)
J/MNRAS/506/150 : The GALAH+ Survey DR3 (Buder+, 2021)
J/A+A/654/A130 : GSP-spec line list (Contursi+, 2021)
J/A+A/654/A116 : Measured DIB 8620 (Zhao+, 2021)
http://faun.rc.fas.harvard.edu/saydjari/GaiaDIB/ : Online data access
http://zenodo.org/record/8303423 : Associated data to the paper
http://www.cosmos.esa.int/web/gaia : Gaia home page
Byte-by-byte Description of file: madgics.dat
--------------------------------------------------------------------------------
Bytes Format Units Label Explanations
--------------------------------------------------------------------------------
1- 19 I19 --- GaiaDR3 Gaia DR3 source ID (gaiasourceid)
21- 26 I6 --- Seq [529/998803] Index of public RVS spectra,
sorted on Gaia source ID (linear_indx)
28- 36 F9.4 --- sSNR [15/1346.4] SNR of the spectrum (stellar_snr)
38- 52 F15.11 deg RAdeg [0/360] Gaia DR3 right ascension (ICRS)
at Ep=2016.0 (ra)
54- 68 F15.11 deg DEdeg [-85.82/79.36] Gaia DR3 declination (ICRS)
at Ep=2016.0 (dec)
70- 81 F12.8 deg GLON [0/360] Galactic longitude
83- 94 F12.8 deg GLAT [-78/71] Galactic latitude
96-106 F11.7 kpc X [-11.43/11.6]? Galactic cartesian coordinate X
108-118 F11.7 kpc Y [-15.17/10.23]? Galactic cartesian
coordinate Y
120-129 F10.7 kpc Z [-5.6/3.6]? Galactic cartesian coordinate Z
131-139 F9.6 --- U [-1/1] Galactic cartesian X-coordinate
direction (unit direction vector)
141-149 F9.6 --- V [-1/1] Galactic cartesian Y-coordinate
direction (unit direction vector)
151-159 F9.6 --- W [-0.98/0.95] Galactic cartesian Z-coordinate
direction (unit direction vector)
--------------------------------------------------------------------------------
161-167 F7.4 0.1nm EW [-1.1/1] MADGICS DIB equivalent width in Å
169-174 F6.4 0.1nm e_EW [0.0018/0.2] EW uncertainty (ew_err)
176-183 F8.3 0.1nm lambda [8613.6/8633.4] MADGICS DIB center
wavelength (stellar frame) in Å (lam)
185-189 F5.3 0.1nm e_lambda [0.08/1.8] lambda uncertainty (lam_err)
191-195 F5.3 0.1nm sig [0.5/4] MADGICS DIB width in Å
197-201 F5.3 0.1nm e_sig [0.09/2.8] sig uncertainty (sig_err)
203-209 F7.2 km/s VdLSR [-427/383] MADGICS DIB velocity in LSR
(vdustlsr)
211-215 F5.2 km/s e_VdLSR [2.7/62] VdLSR uncertainty (vdustlsr_err)
(1)
217-223 F7.2 km/s VdLSRCO [-427/381] MADGICS DIB velocity in LSR
(of Dame+ 2001ApJ...547..792D 2001ApJ...547..792D CO map)
(vdustlsr_COframe)
225-231 F7.5 [-] logdChi2 [1.6/2.6] MADGICS log of change in χ2
from adding DIB component (log10delchisq)
233-239 F7.4 --- dibSNR [3.79/13.56] MADGICS DIB signal-to-noise
ratio (significance of detection) (dib_snr)
(2)
241-246 F6.4 --- Chi2 [0.005/0.4] Total χ2 per degree of
freedom for MADGICS component model
(totchisqperdof)
--------------------------------------------------------------------------------
248-255 F8.5 mas Plx [-1.1/27.9]? Gaia DR3 parallax of background
star (parallax)
257-263 F7.5 mas e_Plx [0.008/0.8]? Plx uncertainty (parallax_error)
265-271 F7.2 km/s RV [-340/379] Gaia DR3 radial velocity of
background star (solar barycentric)
(vstarbary)
273-277 F5.2 km/s e_RV [0.1/32] RV uncertainty (vstarbary_error)
279-284 F6.4 mag E(BP-RP) [0.0002/4.1]? E(BP-RP), color excess modeled
by Gaia (ebpminrp_gspphot)
286-292 F7.4 mag E(B-V)sfd [0.0048/73.1] Reddening E(B-V) along line of
sight, according to SFD dustmap
(Schlegel+ 1998ApJ...500..525S 1998ApJ...500..525S) (sfd)
294-299 F6.4 mag E(B-V)B19 [0/3.48]? Reddening E(B-V) at location of
background star, according to Bayestar19
(Green+ 2019ApJ...887...93G 2019ApJ...887...93G) 3D dust map
(bayestar19)
301-305 F5.3 0.1nm EWGSP [0.013/0.46]? GSP-spec DIB equivalent width
in Å (dibew_gspspec)
307-314 F8.6 0.1nm e_EWGSP [0/0.24]? EWgsp uncertainty
(dibewgspspecuncertainty)
316-324 F9.6 0.1nm erEWNGSP [-1/0.11]? Uncertainty on DIB equivalent width
value occuring from noise part
(dibewnoisegspspecuncertainty)
326-332 F7.3 0.1nm lamGSP [861.3/862.8]? GSP-spec DIB center
wavelength (stellar frame) in Å
(dibgspspeclambda)
334-341 F8.6 0.1nm e_lamGSP [0.001/0.4]? lamGSP uncertainty
(dibgspspeclambda_uncertainty)
343-348 F6.4 0.1nm p2GSP [0.2/3.5]? GSP-spec DIB width of the
Gaussian profile (p2) in Å (dibp2_gspspec)
350-355 F6.4 0.1nm e_p2GSP [0.01/1.7]? p2GSP uncertainty
(dibp2gspspecuncertainty)
357-363 F7.4 --- p0GSP [-0.074/-0.0092]? GSP-spec DIB Gaussian
amplitude (depth; p0) (dibp0_gspspec)
365-375 I11 --- QFGSP [-2/5]?=-2147483648 GSP-spec DIB pipeline
quality factor, QF (dibqf_gspspec) (3)
377-384 F8.5 [-] logChi2GSP [-4.5/-1.2]? Log of GSP-Spec χ2 for
stellar fit (logchisq_gspspec)
386-426 A41 --- FlGSP GSP-Spec quality flag string chain
(flags_gspspec) (4)
428-437 F10.4 pc rgeoCBJ [35.7/17697]? CBJ median geometric distance
(rmedgeo)
439-448 F10.4 pc b_rgeoCBJ [35.6/15310]? 16th-percentile geometric
distance (rlogeo)
450-459 F10.4 pc B_rgeoCBJ [35.8/21086]? 84th-percentile geometric
distance (rhigeo)
461-470 F10.4 pc rpgeoCBJ [35.7/16376]? CBJ median photo-geometric
distance (rmedphotogeo)
472-481 F10.4 pc b_rpgeoCBJ [35.6/13131]? 16th-percentile
photo-geometric distance (CBJ)
(rlophotogeo)
483-492 F10.4 pc B_rpgeoCBJ [35.8/19078]? 84th-percentile
photo-geometric distance (CBJ)
(rhiphotogeo)
494-498 I5 --- FlCBJ [10000/11133]? CBJ distance inference flags
(flag_cbj) (5)
--------------------------------------------------------------------------------
Note (1): This uncertainty does not include uncertainty in the stellar frame,
which should be added in quadrature to obtain the true uncertainty.
Note (2): This is a continuous quality metric that can be used as a cut to
obtain cleaner catalogs.
Note (3): Quality flags from Gaia Col. 2023A&A...674A..40G 2023A&A...674A..40G (Table 2).
Code as follows:
--------------------------------------------------------------
QF ρ0 ρ2 (Å)
--------------------------------------------------------------
0 <0.15 AND >max(RA,RB) 1.2-3.2
1 <0.15 AND <max(RA,RB) & >RB 1.2-3.2
2 <0.15 AND >max(RA,RB) 0.6-1.2
3 <0.15 AND <max(RA,RB) & >RB 0.6-1.2
4 <0.15 AND <max(RA,RB) & <RB 0.6-1.2
5 <0.15 -
--------------------------------------------------------------
Note (4): Definition of each character in the GSP-Spec quality flag string
chain (flags_gspspec), including the possible values (Col. 3);
see Table 2 in Recio-Blanco+ 2023A&A...674A..29R 2023A&A...674A..29R -
Flag as follows:
-----------------------------------------------------------------------------
Chain character Considered Possible
Number-Name quality aspect values
-----------------------------------------------------------------------------
1 vbroadT vbroad induced bias in Teff 0,1,2,9
2 vbroadG vbroad induced bias in log(g) 0,1,2,9
3 vbroadM vbroad induced bias in [M/H] 0,1,2,9
4 vradT VRad induced bias in Teff 0,1,2,9
5 vradG VRad induced bias in log(g) 0,1,2,9
6 vradM VRad induced bias in [M/H] 0,1,2,9
7 fluxNoise Flux noise induced uncertainties 0,1,2,3,4,5,9
8 extrapol Extrapolation level of the parametrisation 0,1,2,3,4,9
9 negFlux Negative flux wlp 0,1,9
10 nanFlux NaN flux wlp 0,9
11 emission Emission line detected by CU6 0,9
12 nullFluxErr Null uncertainties wlp 0,9
13 KMgiantPar KM-type giant stars 0,1,2
14 NUpLim Nitrogen abundance upper limit 0,1,2,9
15 NUncer Nitrogen abundance uncertainty quality 0,1,2,9
16 MgUpLim Magnesium abundance upper limit 0,1,2,9
17 MgUncer Magnesium abundance uncertainty quality 0,1,2,9
18 SiUpLim Silicon abundance upper limit 0,1,2,9
19 SiUncer Silicon abundance uncertainty quality 0,1,2,9
20 SUpLim Sulphur abundance upper limit 0,1,2,9
21 SUncer Sulphur abundance uncertainty quality 0,1,2,9
22 CaUpLim Calcium abundance upper limit 0,1,2,9
23 CaUncer Calcium abundance uncertainty quality 0,1,2,9
24 TiUpLim Titanium abundance upper limit 0,1,2,9
25 TiUncer Titanium abundance uncertainty quality 0,1,2,9
26 CrUpLim Chromium abundance upper limit 0,1,2,9
27 CrUncer Chromium abundance uncertainty quality 0,1,2,9
28 FeUpLim Neutral iron abundance upper limit 0,1,2,9
29 FeUncer Neutral iron abundance uncertainty quality 0,1,2,9
30 FeIIUpLim Ionised iron abundance upper limit 0,1,2,9
31 FeIIUncer Ionised iron abundance uncertainty quality 0,1,2,9
32 NiUpLim Nickel abundance upper limit 0,1,2,9
33 NiUncer Nickel abundance uncertainty quality 0,1,2,9
34 ZrUpLim Zirconium abundance upper limit 0,1,2,9
35 ZrUncer Zirconium abundance uncertainty quality 0,1,2,9
36 CeUpLim Cerium abundance upper limit 0,1,2,9
37 CeUncer Cerium abundance uncertainty quality 0,1,2,9
38 NdUpLim Neodymium abundance upper limit 0,1,2,9
39 NdUncer Neodymium abundance uncertainty quality 0,1,2,9
40 DeltaCNq Cyanogen differential equivalent width quality 0,9
41 DIBq DIB quality flag 0,1,2,3,4,5,9
-----------------------------------------------------------------------------
Note (5): CBJ (Bailer-Jones+ 2021, Cat. I/352) flags as follows:
The flag field in the catalog is a string of five integer digits ABBCC.
A : Source magnitude compared to the limit used to make the prior
0 = Source has no G-band magnitude
1 = G ≤ Glim
2 = G > Glim
B : Hartigan dip test for unimodality. Left digit geometric, right digit
photogeometric
0 = unimodal hypothesis okay
1 = possible evidence for multimodality
C : QG models used in prior. Left digit bluer model, right digit redder model
0 = null (no model)
1 = one-component Gaussian
2 = two-component Gaussian
3 = smoothing spline
--------------------------------------------------------------------------------
Acknowledgements:
Saydjari Andrew [andrew.saydjari at cfa.harvard.edu]
(End) Emmanuelle Perret [CDS] 14-Sep-2023