VVV survey near-infrared colour catalogue of known variable stars
Fábio R. Herpich, Carlos E. Ferreira Lopes, Roberto K. Saito, Dante Minniti, Alessandro Ederoclite, Thiago S. Ferreira, Marcio Catelan
AAstronomy & Astrophysics manuscript no. 34356corr_revised © ESO 2021February 8, 2021
VVV survey near-infrared colour catalogue of known variablestars (cid:63)(cid:63)(cid:63)
F. R. Herpich , C. E. Ferreira Lopes , R. K. Saito , D. Minniti , , A. Ederoclite , T. S. Ferreira , M. Catelan , Universidade de São Paulo, IAG, Rua do Matão 1226, Cidade Universitária, São Paulo 05508-900, Brazil National Institute For Space Research (INPE / MCTI), Av. dos Astronautas, 1758 - São José dos Campos - SP, 12227-010, Brazil Departamento de Física, Universidade Federal de Santa Catarina, Trindade 88040-900, Florianópolis, SC, Brazil Departamento de Fisica, Facultad de Ciencias Exactas, Universidad Andres Bello, Av. Fernandez Concha 700, Las Condes, Santi-ago, Chile Vatican Observatory, V00120 Vatican City State, Italy Instituto de Astrofísica, Facultad de Física, Pontificia Universidad Católica de Chile, Av. Vicuña Mackenna 4860, 7820436 Macul,Santiago, Chile Millennium Institute of Astrophysics, Santiago, ChileReceived month day, year; accepted month day, year
ABSTRACT
Context.
The Vista Variables in the Via Lactea (VVV) near-infrared variability survey explores some of the most complex regions ofthe Milky Way bulge and disk in terms of high extinction and high crowding.
Aims.
We add a new wavelength dimension to the optical information available at the American Association of Variable Star ObserversInternational Variable Star Index (VSX-AAVSO) catalogue to test the VVV survey near-infrared photometry to better characterisethese objects.
Methods.
We cross-matched the VVV and the VSX-AAVSO catalogues along with Gaia Data Release 2 photometry and parallax.
Results.
We present a catalogue that includes accurate individual coordinates, near-infrared magnitudes (
ZY JHK s), extinctions A Ks ,and distances based on Gaia parallaxes. We also show the near-infrared CMDs and spatial distributions for the di ff erent VSX typesof variable stars, including important distance indicators, such as RR Lyrae, Cepheids, and Miras. By analysing the photometric flagsin our catalogue, we found that about 20% of the stars with measured and verified variability are flagged as non-stellar sources, evenwhen they are outside of the saturation and / or noise regimes. Additionally, we pair-matched our sample with the VIVA catalogue andfound that more than half of our sources are missing from the VVV variability list, mostly due to observations with low signal-to-noiseratio or photometric problems with a low percentage due to failures in the selection process. Conclusions.
Our results suggest that the current knowledge of the variability in the Galaxy is biased to nearby stars with lowextinction. The present catalogue also provides the groundwork for characterising the results of future large variability surveys suchas the Vera C. Rubin Observatory Legacy Survey of Space and Time in the highly crowded and reddened regions of the Galacticplane, as well as follow-up campaigns for characterising specific types of variables. The analysis of the incorrectly flagged stars canbe used to improve the photometric classification of the VVV data, allowing us to expand the amount of data considered useful forscience purposes. In addition, we provide a list of stars that are missed by the VIVA procedures for which the observations are goodand which were missed due to some failure in the VIVA selection process.
Key words.
Galaxy: disk – Galaxy: bulge – Galaxy: stellar content – Stars: variables: general
1. Introduction
The variable stars in the Milky Way are almost countless. Sincethe discovery of the first of these objects, the number of vari-able stars has increased consistently. Pigott & Englefield (1786)provided one of the first catalogues of stellar variability, con-taining a total of 12 objects known as being variable at thattime. Many other catalogues have been released since then, es-pecially within the past two decades, when large digital sur-vey telescopes became operational, scanning the sky night afternight, such as the Infrared Astronomical Satellite (IRAS, Neuge-bauer et al. 1984), the Two Micron All-Sky Survey (2MASS, (cid:63)
Based on observations taken within the ESO Public Survey, Pro-gramme IDs 179.B-2002. (cid:63)(cid:63)
Data used in this work is fully and only available in electronic format the CDS via anonymous ftp to cdsarc.u-strasbg.fr (130.79.128.5) orvia http: // cdsweb.u-strasbg.fr / cgi-bin / qcat?J / A + A / Kleinmann et al. 1994), the Sloan Digital Sky Survey (SDSS,York et al. 2000), the Catalina Sky Surveys (Drake et al. 2014,2017), the Panoramic Survey Telescope and Rapid ResponseSystem (Pan-STARRS, Kaiser et al. 2002), the UKIRT InfraredDeep Sky Survey (UKIDSS, Lawrence et al. 2007), the AKARIFar-infrared All-Sky Survey (AKARI, Ishihara et al. 2010), theWide-field Infrared Survey Explorer (WISE, Wright et al. 2010;Chen et al. 2018a,b), and the Gaia spectroscopic survey (Gaia,Gilmore et al. 2012; Gaia Collaboration et al. 2016), as well theongoing All-Sky Automated Survey for Supernovae (ASAS-SN;Jayasinghe et al. 2018). In addition, the large microlensing sur-veys, such as the Massive Astrophysical Compact Halo Objects(MACHO; Alcock et al. 1993), the Optical Gravitational Lens-ing Experiment (OGLE; Udalski et al. 1993), the Experiencepour la Recherche d’Objets Sombres (EROS; Aubourg et al.1993), the Microlensing Observations in Astrophysics (MOA;Bond et al. 2001), the Disk Unseen Objects (DUO; Alard et al.
Article number, page 1 of 14 a r X i v : . [ a s t r o - ph . S R ] F e b & A proofs: manuscript no. 34356corr_revised , Ivezic et al. 2008), which is expectedto discover millions of variables. Another important aspect toconsider is the contrast between near-infrared and optical vari-ability. The searches performed at di ff erent wavelengths resultin di ff erent relative numbers, and this must be considered whenthe number of variable stars that will be present in the LSSTcatalogues is predicted (e.g. Pietrukowicz et al. 2012).However, most of the current projects searching for vari-ables operate with optical telescopes, thus avoiding the inner-most Milk Way plane, where high extinction and crowding limitthe depth at optical wavelengths. Therefore an infrared surveyis more suitable for the mission of observing deeper into theGalaxy plane, as does the VISTA Variables in The Vía Láctea(VVV ), which is an European Southern Observatory (ESO )public survey that mapped the bulge ( − . ◦ (cid:46) l (cid:46) + . ◦ and − . ◦ (cid:46) b (cid:46) + . ◦ ) and the inner southern part of the disk(294 . ◦ (cid:46) l (cid:46) + . ◦ and − . ◦ (cid:46) b (cid:46) + . ◦ ) of our Galaxyusing five near-infrared bands ( Z , Y , J , H and K s ; Minniti et al.2010) plus a variability campaign in K s band with typically 100epochs per field in the period 2010-2017. The VVV Survey wascompleted in 2017 when its extension, the VVV eXtended Sur-vey (Minniti 2018), started observing to increase the observedarea of VVV from 562 to 1700 sq. deg., filling the gaps betweenthe VVV and the VISTA Hemisphere Survey (VHS McMahonet al. 2013).Near-infrared surveys such as the VVV are very e ffi cient atlow Galactic latitudes close to the plane, where existing opti-cal variability surveys are usually blinded by the absorption inthe interstellar medium. For instance, VVV data were used tofind RR Lyrae stars within 100 arcmin from the Galactic Cen-tre (Contreras Ramos et al. 2018). In the same way that Baade’sWindow was important for optical studies of the Galactic bulge,the near-infrared surveys also profit from the study of the win-dows that have recently been found in the Milky Way plane (e.g.Dante’s Window; Minniti et al. 2018).Focusing on the subject of variables stars, the InternationalVariable Star Index (VSX), provided by the American Associ-ation of Variable Star Observers (AAVSO) , compiles a largenumber of known variable objects into a single database contain-ing 1 432 563 objects as of February 1, 2020 . This work com-piles information from various catalogues such as those men-tioned before, together with a brief analysis of missed sourcesin one of the more complete variability catalogues for the galac-tic bulge existing to date, the VISTA Variables in the Vía Lácteainfrared variability catalogue (VIVA, Ferreira Lopes et al. 2020).Here we present a catalogue with near-infrared colours forthe VSX sources based on VVV data. This work provides use-ful information such as the colours in the ZY JHK s bands, theextinction in the near-infrared ( A K s ), and the distances based on http: // adsabs.harvard.edu / abs / https://vvvsurvey.org/ We downloaded the present VSX catalogue from the AAVSOdatabase in February 2020. As this catalogue grows constantly, it is ex-pected to be even larger at the time of publishing. the parallaxes of Gaia Data Release 2 (Gaia DR2; Gaia Collab-oration et al. 2018, 2019). Near-infrared colour-magnitude dia-grams (CMDs) and the surface density distribution for the di ff er-ent types of variable stars include important distance indicatorssuch as RR Lyrae, Cepheids and Miras. Section 2 presents thedata we used and some statistics of our sample, in Section 3 weclassify the variable stars in our catalogue in context with dis-tance and period ( P ), in Section 4 we discuss the sources thatwere missed in the construction of the VIVA catalogue, and inSection 5 we summarise our results.
2. Data
Many catalogues of variable stars are available electronically,such as the General Catalogue of Variable Stars (GCVS, Samus’et al. 2017), the All-Sky Automated Survey (ASAS, Pojmanskiet al. 2005), and the Catalog and Atlas of Cataclysmic Variables(Downes et al. 2001). As part of this sample of catalogues, ob-jects are discovered by important variability surveys of the innerMilky Way such as OGLE (e.g. Udalski et al. 2015), MOA (Al-cock et al. 1997) , and even VVV. The VSX catalogue containsnames, positions, period, the VSX type , and astronomical in-formation, such as the constellation they belong to, and the pass-band that was used to measure the variability, which is mostlyobserved in the optical. While the VSX catalogue contains data for the whole sky, theVVV survey covered a total of 562 sq. deg., which is about 1.4 %of the celestial sphere. Although it covers only a small fractionof the sky, the VVV concentrated its observations on the mostcrowded regions of the southern sky, which are the Milk Waybulge and southern plane. The projected cone of view covers ∼
30 % of the Milk Way stars, thus providing one of the mostcomplete catalogues of the inner Milk Way: almost a billionsources have been detected (Alonso-García et al. 2018).We matched the VSX catalogue, containing 1 432 563sources, with the VVV catalogues for the 348 individual tilescovering the bulge and disk portions, which contain 428 260 599sources. This resulted in 701 256 objects that match to within 1arcsec. We refer to them as the V SX sample. Table 1 shows thefirst few stars of the resulting match (the full table is available inthe complementary data) . The full description of all columnsappearing in the table is given in Section 2.4. VVV data in-clude the coordinates (equatorial and Galactic) and the standardsingle-epoch aperture photometry for the VVV Data Release 4(VVV-DR4) provided by the Cambridge Astronomical SurveyUnit (CASU ) in the five VISTA filters ( ZY JHK s), with photo-metric flags indicating the likely morphological type based onthe aperture curve of growth (Saito et al. 2012). The K s-bandvalues are mean magnitudes calculated from the multi-epoch ob-servations (typically 50-100), while the ZY JH -band magnitudesare averages of a few (typically 2-4) observations taken at ran-dom phases for the periodic variables. The CASU reduction pro-cess produces a flag to identify the quality of the source, where -1means best quality photometry of stellar objects, -2 means bor-derline stellar objects, 0 means noise, + In Table 1 some VSX and Gaia (see Section 2.3) columns were sup-pressed to better fit the page. http://casu.ast.cam.ac.uk/ Article number, page 2 of 14erpich et al.: VVV survey near-infrared colour catalogue of known variable stars objects, -7 indicates sources containing bad pixels, and -9 meanssaturated sources (Saito et al. 2012). Using these flags to selectunsaturated, noisy, and extended sources in the JK s bands, wecollected a sample of 281 536 variables stars that can be reliablystudied and whose unsaturated light curves are available throughthe VISTA Science Archive (VSA) (we call these CVVS, for“constrained VISTA variable sources”).While this paper focuses on the near-infrared colours, VVV K s-bands light curves are publicly available and can be retrievedthrough the VSA by querying the VVV DR4 synoptic source ta-ble . A search for new variable stars in the VVV light curveswas performed using di ff erent techniques (e.g. Ferreira Lopes& Cross 2016a, 2017a; Ferreira Lopes et al. 2018) and resultedin the discovery of a large number of previously unknown vari-able stars. These discoveries are beyond the scope of this paper,and the resulting catalogue is presented in Ferreira Lopes et al.(2020). Complementing the near-infrared colours from the VVV, we alsopresent the total extinction in the K s band ( A K s ) in our catalogue,integrated along the entire line of sight for each source and pro-vided by the VVV extinction maps. In all cases, the values arethe mean A K s over an area of 10 ×
10 arcmin around the tar-get position and based on the Cardelli et al. (1989) extinctionlaw. For the bulge area, the total extinction A Ks was taken di-rectly from the Bulge Extinction And Metallicity (BEAM) Cal-culator (Gonzalez et al. 2012) , and for the disk area, the A Ks values were calculated from the EJK map presented in Minnitiet al. (2018), assuming the Cardelli et al. (1989) law. Relativeextinction for the other VISTA filters were provided by Catelanet al. (2011) and are listed in Table 2 for the optical V band aswell.Fig. 1 shows the surface density distribution along with theGalactic coordinates and the CMD for the disk and bulge forthe V SX sample, colour-coded by the total extinction (inte-grated along the entire line of sight for each object) calculatedfor the VVV data A K s . A K s values for all objects are avail-able electronically as supplementary material. A K s varies from A K s < .
01 mag in the outer bulge up to A K s ∼ Recently, another huge database has been publicly released. It isthe second data release (DR2) of the optical Gaia mission, whichcontains positional data for over one billion sources at G , BP , and RP bands, as well as precise parallaxes and proper motion mea-surements (Gaia Collaboration et al. 2018). The dataset allowedus to match the VVV + VSX and Gaia-DR2 catalogues, selectingthe matches with the smaller separation within 1 arcsec radius.A total of 590 824 pairs were found within 1 arcsec, for whichparallaxes are available in Gaia DR2 (we did not distinguish be-tween positive and negative parallaxes because of the biases thatthis procedure would introduce in the sample, e.g. Bailer-Joneset al. 2018, BJ18 hereafter), and we can estimate the distanceto them assuming a naive value of d = /ω . Distances basedon Gaia parallaxes were also retrieved from BJ18, obtained on http://horus.roe.ac.uk/vsa/index.html http://mill.astro.puc.cl/BEAM/calculator.php the basis of a three-dimensional model of the Galaxy instead ofthose obtained by simply inverting the parallax. From the to-tal initial variables in our sample, we found 207 439 pairs forwhich BJ18 distances are available. A comprehensive descrip-tion of distance estimates based on Gaia parallaxes is presentedin Luri et al. (2018). The full description of all columns fromGaia that appear in the catalogue is given in Section 2.4.Fig. 2 shows these resulting sources coloured by the A K s asadopted for Fig. 1, but for the distances calculated from the Gaiaparallax. Selecting only VVV sources whose J and K s magni-tudes have flags between -9 and 0 (see Section 2.1 for a moredetailed overview of the meaning of the flags) and with the dis-tance measured by BJ18, we obtain a sample containing 113 786objects. The data we release contain all VVV + VSX matchedvariables, regardless of whether they are good according to the JK s flag criteria. The discrepancy between the measured dis-tance directly from the parallax and that from taken from BJ18 isquite evident. We can fairly assume that this correction is highlyimportant because many of the Galactic sources in our samplewould be located much farther away than they should when onlyconsidering the parallax, i.e. Galactic sources have distances big-ger than the size of the Milk Way. Fig. 2 also shows the projecteddistribution of stars along with the height z of the Galaxy as afunction of the geometrical distance y . It is interesting to notethat the stars appear to avoid the Galactic centre (around b = ◦ ),as is also visible in the spatial distribution of Fig. 1. This is anatural and known phenomenon and is the consequence of thehigh interstellar extinction along the line of sight of the disk andbulge. As Table 1 does not show many of the columns in our catalogue,in this section we describe all the columns from VVV that weremodified or created by us that are available electronically as sup-plementary material (we do not describe those of the alreadypublic catalogues such as VSX and Gaia).The following columns are directly obtained from the VVVcatalogue: – RA_(VVV) and
DEC_(VVV):
J2000.0 coordinates from theVVV catalogue (in degrees). – L and B: The Galactic coordinates l and b as obtained for theVVV (in degrees). – MAG_ i: Magnitude of the sources measured by VVV, where i = { Z , Y , J , H , K s } are the respective bands of the VVV sur-vey. – ERR_ i: Errors of the magnitudes of VVV to the respectivebands i = { Z , Y , J , H , K s } . – F_ i: The photometric flags for the bands where i = { Z , Y , J , H , K s } .The following columns are obtained from the information ofVVV catalogue or from the complete match VVV + VSX + Gaia: – A Ks : The total extinction in the K s band. – distance: Distances calculated using the Gaia parallax( d = /ω , in kpc). – dist_BJ: Distances from BJ18 (estimated distance ‘r_est’,in kpc). – (x,y,z)_BJ: Distances based on dist_BJ . – CandidateType:
Variables reclassified in Section 3.
Article number, page 3 of 14 & A proofs: manuscript no. 34356corr_revised
Table 1. V SX catalogue. In this sample table we suppressed a few columns from the original AAVSO catalogue (AUID, Period, Type, and Mag.),VVV (errors and flags), and from Gaia (parallax). RA / DEC coordinates are J2000 from VVV. Blank entries indicate unavailable data.
VSX RA DEC L B
Z Y J H K s A Ks Dist_BJName (deg) (deg) (deg) (deg) (mag) (mag) (mag) (mag) (mag) (mag) (kpc)ASAS J123010-6323.5 187.54 -63.39 -59.45 -0.62 10.93 10.74 10.58 10.47 10.34 0.65 ...GDS_J1229509-635129 187.46 -63.86 -59.44 -1.09 12.72 11.84 10.40 10.25 9.01 0.78 ...NSV 5667 187.73 -62.77 -59.41 0.01 13.35 13.15 12.82 12.96 12.54 1.08 0.790GDS_J1230188-633823 187.58 -63.64 -59.41 -0.86 11.14 10.69 10.12 9.72 9.37 0.67 1.623GDS_J1230367-632806 187.65 -63.47 -59.39 -0.69 13.50 13.21 12.83 12.38 12.13 0.65 1.696GDS_J1231003-631727 187.75 -63.29 -59.36 -0.51 12.93 10.79 9.07 8.59 7.62 0.72 5.226GDS_J1230537-633727 187.72 -63.62 -59.35 -0.84 10.83 9.16 8.38 9.29 8.39 0.67 1.849NSV 5671 187.77 -63.38 -59.35 -0.60 12.56 12.33 12.05 11.80 11.62 0.65 0.607NSV 5675 187.88 -62.93 -59.33 -0.15 12.65 12.34 12.24 12.72 12.23 0.91 2.128GDS_J1231035-633759 187.76 -63.63 -59.33 -0.85 11.85 10.24 9.04 8.70 ... 0.88 3.162GDS_J1231255-633150 187.86 -63.53 -59.30 -0.75 12.62 12.33 11.88 11.57 11.36 0.72 0.846GDS_J1231445-631137 187.94 -63.19 -59.29 -0.41 12.07 11.88 11.49 11.37 11.07 0.84 0.792GDS_J1232008-624848 188.00 -62.81 -59.28 -0.03 12.34 12.20 ... ... ... 0.96 ...NSV 5684 187.97 -63.21 -59.27 -0.42 13.40 13.17 12.85 12.56 12.38 0.84 0.839GDS_J1232131-624557 188.05 -62.77 -59.26 0.02 12.29 12.16 12.80 ... ... 0.96 1.932NSV 5689 188.03 -63.07 -59.25 -0.28 13.27 13.06 12.75 12.74 12.45 0.84 1.316ASAS J123144-6324.3 187.93 -63.41 -59.27 -0.62 10.09 9.89 ... ... ... 0.72 1.215GDS_J1231514-633314 187.96 -63.55 -59.25 -0.77 11.92 10.31 8.98 ... 7.58 0.88 4.274NSV 5683 187.96 -63.83 -59.23 -1.04 12.75 12.31 11.89 11.60 11.30 1.03 2.327GDS_J1232312-625756 188.13 -62.97 -59.22 -0.17 12.36 11.79 12.03 13.92 11.76 0.91 3.955GDS_J1232281-630555 188.12 -63.10 -59.21 -0.31 13.56 11.66 10.14 12.43 8.57 0.84 2.840NSV 5692 188.14 -63.23 -59.19 -0.44 13.33 13.15 12.89 12.67 12.60 0.79 1.319NSV 5696 188.20 -63.10 -59.18 -0.30 13.62 13.47 13.17 13.10 12.88 0.84 1.431GDS_J1232513-630146 188.21 -63.03 -59.17 -0.23 13.94 13.70 13.59 13.46 13.10 0.91 0.821NSV 5710 188.33 -63.02 -59.12 -0.22 12.72 12.56 12.27 12.81 12.30 0.91 0.932NSV 5701 188.24 -63.81 -59.11 -1.01 17.79 16.16 14.28 12.35 11.55 0.82 3.727ASAS J123326-6256.2 188.36 -62.94 -59.11 -0.14 12.90 12.40 13.77 ... ... 0.91 1.426NSV 5699 188.24 -63.85 -59.10 -1.05 11.08 10.84 10.45 10.07 9.87 0.82 ...NSV 5713 188.37 -63.26 -59.08 -0.46 13.51 13.29 12.99 13.20 12.96 0.71 0.971VW Cru 188.33 -63.51 -59.09 -0.71 13.32 13.00 13.74 ... ... 0.72 1.235CM Cru 188.48 -62.83 -59.07 -0.03 14.36 13.49 12.52 12.66 ... 0.93 3.656GDS_J1233486-630008 188.45 -63.00 -59.07 -0.20 12.34 ... 12.74 ... ... 0.91 2.442GDS_J1233485-630858 188.45 -63.15 -59.06 -0.35 12.27 12.08 ... ... ... 0.72 9.390GDS_J1233514-631706 188.46 -63.29 -59.04 -0.48 12.29 11.93 12.05 ... ... 0.71 6.043NSV 5716 188.43 -63.82 -59.02 -1.01 12.00 11.57 11.20 10.99 10.79 0.82 1.587GDS_J1234027-632932 188.51 -63.49 -59.01 -0.69 12.03 11.90 12.33 ... ... 0.72 1.715GDS_J1233526-635121 188.47 -63.86 -59.00 -1.05 11.78 11.26 10.71 10.19 11.41 0.68 0.544NSV 5719 188.46 -63.93 -59.00 -1.13 12.96 12.64 12.37 12.08 11.83 0.51 0.901GDS_J1234165-631955 188.57 -63.33 -58.99 -0.52 12.35 11.85 11.91 ... ... 0.69 6.543GDS_J1234253-630811 188.61 -63.14 -58.99 -0.33 12.34 11.92 12.13 ... ... 0.83 5.700
Table 2.
Relative extinctions for the
ZY JHK s VISTA filters, and for theoptical V band. Adapted from Catelan et al. (2011).
Relative Valueextinction A V / A Ks A Z / A Ks A Y / A Ks A J / A Ks A H / A Ks ff erent interpretations in terms of stellarpopulation and Galactic structure. For instance, a smaller num- ber of objects is seen in the innermost bulge area. This so-calledzone of avoidance is also present in the distribution of the VVVNovae catalogue (Saito et al. 2013) and is caused by the high ex-tinction that a ff ects the observations in the optical wavelengths,as seen in most variability surveys. This absence of objects inthe innermost regions is even more evident in the distributionof long-period variable candidates from Gaia (Mowlavi et al.2018), which is also based on optical observations. Observableobjects in these regions are mainly from the foreground disk, ascan be seen in the y , z projected distances (bottom panels of Fig.2). On the other hand, closer objects are probably saturated inthe VVV observations and thus are not included in our sample.The highest density of variable sources is seen in the interme-diate bulge region ( − ◦ < b − ◦ ) of Fig. 3 and is mostly causedby RR Lyrae that were detected in variability surveys such asOGLE and VVV. Especially for OGLE, the footprint is easily Article number, page 4 of 14erpich et al.: VVV survey near-infrared colour catalogue of known variable stars
Fig. 1.
Top: Distribution for all 281 536 stars in the match VVV + VSX. The data points are colour-coded by the total extinction A K s , integratedalong the entire line of sight for each object. Values vary from A K s < .
01 mag to A K s ∼ A K s parameter. The arrow shows the reddening vector for the diagram. seen across the bulge (see figure 7 in Soszy´nski et al. 2011). RRLyrae are metal-poor population II stars that are detected in largenumbers in the bulge region, compared to a small fraction thatis present throughout the Galactic disk, which should be domi-nated by metal-rich population I stars (e.g. Dékány et al. 2018;Iorio & Belokurov 2020). This section is dedicated to gathering the information availablefor our VVV + VSX + Gaia catalogue. From the 701 256 variablesin the match VSX + VVV, we selected all objects with flags lowerthan 0 and greater than -9 ( i.e. unsaturated point sources). Wehave the following numbers: 474 105 in Z band, 420 775 in Y ,366 282 in J , 353 220 in H , and 346 614 in K s . There are 195 580sources with good photometric quality in all five bands.From the 590 824 objects with Gaia parallaxes, 206 624 havenegative measures, with a minimum value of -443.735. These parallaxes are unreliable and hence were not used for any anal-ysis involving distances, although the stars are not excludedfrom the table. The highest parallax value is for an M-typevariable star in the bulge direction: 282.189 arcsec. For thepositive values of the parallax, the mean value is 0.468 arcsecwith a standard deviation ( σ ) of 0.990. The mean correspond-ing distance from the naive Gaia’s parallax is 28.326 kpc with σ = .
582 kpc. For the BJ18 distances, the average is d BJ18 = .
246 kpc with σ = .
625 kpc.
During the definition of the sample for the analysis of the VIVAcatalogue, we realised that a considerable number of stars haveflags 0 (noise) and 1 (non-stellar objects), but most of them arecompatible and classified as point sources in VSX. We anal-ysed the conditions of these measurements and briefly investi-gated whether they are misclassifications or spurious data. Fig. 4
Article number, page 5 of 14 & A proofs: manuscript no. 34356corr_revised
Fig. 2.
Distances calculated directly from the Gaia parallax (left) and distances from BJ18 (right). Upper panels: ( x , y ) projection. The blacksemi-circle identifies the 15 kpc radius from the Galactic centre as a symbolic limit for the disk of the Galaxy. Lower panels: ( y , z ) projection.The red cross identifies the geometrical position adopted to the centre of the Galaxy assuming d = K s band. The sample is constrained in the VVV J and K s magnitudes and for those stars with distances measured using BJ18. shows the histograms of the stars (as identified in VSX) col-lected for these two flags. Out of 701 256, 2393 (0.34%) thatwere flagged as noise ( F _ K s =
0) and are shown in the greenhistogram in Fig. 4, about half of them are fainter than the de-tection limit (the flag for the K s band is typically 17.5, althoughnot constant, depending on the extinction of the region beingconsidered) . The right side of Fig. 4 shows the number of vari-ables corresponding to each defined VSX class. For this analysis, all objects assigned as suspected variables in VSX (for which acolon is added to the class) were added to the group “Other”.While flag 0 contains just a small number of objects, negligi-ble for most purposes, flag 1 has a much more significant numberof stars. We found that 133 808 (19%) stars in our main sampleare classified as non-stellar objects. Fig. 4 shows the histogramand the distribution per variable class for these objects, and al-though part of the sample lies beyond (or very close to) the de-tection limit, these sources are better behaved, and most of themare within the saturation-detection limits. Moreover, most of the Article number, page 6 of 14erpich et al.: VVV survey near-infrared colour catalogue of known variable stars
Fig. 3.
All 112 932 variables of our sample distributed through the five main classes we defined with the Galactic coordinates l and b . Symbols arecolour-coded following the legend in the figure. Fig. 4.
Green: Histogram (left) for the sources with flag 0 (noise) for the K s band and their distributions in the VSX classes (right side). Magenta:Same, but for flag 1 (non-stellar objects) for the K s band. sources ( ∼ F _ K s = ZYJH and find the same conclusions with similar numbers. Thisinformation can be further used to constrain their definitions andmaybe improve the accuracy rate of these flags for future re-leases of the extended VVV survey (VVVX ).
3. Variables through distance and period
As the number of di ff erent types of variables is high and thus pre-vents a clearer view in the diagrams, the similar variable types were merged following the VSX denomination for simplicity.We separated them into the five main groups following the def-initions of the VSX for Eruptive, Rotating, Cataclysmic, Pulsat-ing and Other, which contains all objects that are not includedin the previous groups. The last group alone does not followthe VSX denomination strictly and might contain a mix of starswith unknown variability type and / or exotic sources such as ac-tive galactic nuclei, gama-ray bursts, quasars, and microlensingevents. We did not add the composite types from VSX to thesegroup (e.g. Algol systems, in which the low-mass component isclose to its inner Roche lobe, EA / SD ), for instance.We here then present some results using the period fromVSX and other properties in terms of the five groups we just de-fined. For instance, we only used the K s magnitude and selected Article number, page 7 of 14 & A proofs: manuscript no. 34356corr_revised the good sources accordingly using only the flags in this bandto constrain our sample. As a result, we have a higher numberof variables, with a total of 130 855 objects, of which we show129 938 in Fig. 5. The missing objects are stars that we wereunable to classify into any of the five main classes defined here.Fig. 5 shows the histograms normalised to 1 at their max-imum for the defined classes of variables as a function of thedistance calculated using the BJ18 distances. The Eruptive starshave two peaks at 1.3 kpc and 2.1 kpc, while the Other groupexhibits a broad distribution from ∼ ∼ ff erent typesof variables, this is an expected e ff ect, possibly composed of amix of Pulsating and Rotating stars. This assumption is beingexplored by other works.Because exploring all the VSX classes, which are fundamen-tally related to our five main groups, is beyond of the scope ofthis work, we concentrated on one large group, the Pulsatingstars, and three particular VSX types from the group Other. Theyare the VSX types MISC (stars whose classification was not pre-cise enough with the automatic methods that were applied andthey can be either red variables or irregular types stars), S (vari-ables with rapid light changes that have not been studied so far),and VAR (unclassified variable stars).The result of this selection are the 167 892 stars (here weintroduce another cut to the sample because of the A K s mea-surements, which cause 507 stars to be dropped from consid-eration) shown in Fig. 6 for the extinction-corrected K s magni-tude ( K s c ) as a function of the BJ18 distance and colour-codedby the extinction A K s . We can identify five main groups: fourhorizontal groups ( K s c > K s c ∼ . K s c ∼ .
5, and K s c <
12) for all distance ranges, and the fifth group, whichspans from 9 < K s c (cid:46)
14 but is concentrated at smaller distances( d < K s magnitude for all variables shown in Fig. 6, exceptfor those that do not have a measured period in the VSX cat-alogue for a total of 14 727 objects. The 2055 objects that aremissing in this sample relative to those shown by Fig. 6 cor-respond to the stars with undetermined periods, and Fig. 6 andFig. 7 show that they are mostly objects of high extinction thatare mainly located at the centre of Galactic plane (see Fig. 1).Fig. 7 shows two dominant groups, most composed of RR Lyrae(shorter periods) and Cepheid stars (longer periods). We studythe VSX classes individually in this projection.To extend the study of Pulsating stars and the VSX variablestypes, we plot them in a diagram of magnitude K s (the appar-ent and the absolute) as a function of the period, but identifyingthe variable types that are considered (see Appendix B for thefull names and descriptions of each type in Fig. 8). The resultsare shown in Fig. 8 along with the contours representing the 1 σ boundary for the more populous groups. We easily identify twomain distinct regions populated by RR Lyrae on one side and thesemi-regular variables on the other. In between these two maingroups lie the W Virginis variables (CWA and CWB), which oc-cupy the upper portion of the diagram and several unstudied vari-ables that occupy the lower portion (e.g. the unstudied variableswith rapid light changes, S, form a well-defined group in this re-gion along with other non-classified variables of the types MISC,PULS, and VAR).The apparent magnitude of the types shown in Fig. 8 appearsto be very well-defined in di ff erent regions of the diagram, theybecome more inhomogeneous when distributed over the periodversus absolute magnitude plane, at least on the magnitude axis. Table 3.
Candidates to VSX variable types for the unclassified classesMISC, PULS, and VAR using the period vs. K s magnitude space fromFig. 8. Type DSCT RRC RRAB S SRMISC 17 22 27 122 371VAR 5 1 6 23 681 24We can still draw vertical lines in both panels of Fig. 8 usingthe contours as parameters for the divisor line. The first line isdrawn at log P = − .
65 and separates the δ Scuti type of vari-ables (DSCT and HADS) from the RR Lyrae with symmetriclight curves (RRC). The second line is located at log P = − . P =
0. Finally, the fourth line is drawn atlog P = .
24 and separates the semi-regular variables (SR) fromthose with shorter periods.Based on these definitions, we can now identify the unclas-sified stars from our sample into the most populous VSX types,at least as a first approximation. We can then classify the typescalled MISC, PULS, and VAR depending upon which side of thelines they occupy. We did not classify type S when they occupya region of the diagram that is not populated by any other well-defined VSX type in our sample (at 0 < log d < .
24 and justbelow where the CWA, CWB, and SRS are located). They mightbe W Virginis or semi-regular pulsating variables or even moreprobably, they are variables other than pulsating stars and areto be classified accurately through other observations. We con-sidered type S as a variability type of stars and included MISC,PULS, and VAR located at 0 < log d < .
24 as S types. Ourclassification is loosely based on about two parameters: periodand K s magnitude, and the term can be misread as a determina-tion when in fact we intend to provide candidates that need to bestudied in more detail to confirm their final type. Applying thedefinitions above (for the entire table) with good K s photometryand well-determined periods, we account for a total of 337 559stars that are available for a reclassification, we find the numberof candidates listed in Table 3. No PULS were considered. Theseresults are included in the supplementary material as a new col-umn in the table called CandidateType . We call attention to thestars of VAR type, however, which are by definition variablesof undefined type. This nomenclature is also used for variablecandidates, which makes the period found in the VSX databasenot completely reliable and implies that the strange gap aroundlog P = .
24 might be an artefact due to incorrect estimatesof the period (we thank the anonymous referee for pointing thisout). As a consequence, we must be careful when these stars areused for science. On the other hand, we consider them importantobjects to be targeted by variability surveys and / or follow-ups,therefore we kept them in the analysis of Fig. 8.
4. VIVA comparison
The VIVA catalogue (Ferreira Lopes et al. 2020) is based on avariability analysis of the VVV-DR4 Data Release that is thesame as we used here. It was compiled following a series ofrecommendations in the New Insight into Time Series Analysis(NITSA - Ferreira Lopes & Cross 2016b, 2017b; Ferreira Lopeset al. 2018). These criteria provide a good prescription for select-
Article number, page 8 of 14erpich et al.: VVV survey near-infrared colour catalogue of known variable stars
Fig. 5.
Histograms for the VSX classes of variables normalized to 1 relative to the distance calculated as in the Section 2.3.
Fig. 6.
Distance of the group defined as Pulsating stars along with theVSX types MISC, S, UNC, and VAR from the group Other as a functionof the extinction-corrected magnitude K s colour-coded by extinction A K s . ing variable stars. We caution that the VIVA catalogue only con-sidered stars with more than ten non-flagged observations (num-ber of good observations, or NG > ppErBits ) smaller than 256 and stack-flag ( flag ) equal to0. The first is related to the archive curation procedures and thesecond to potential matching problems. These thresholds wereadopted to reduce problems regarding the integrity of the detec- The full description of the detection flag can be found at http://horus.roe.ac.uk/vsa/ppErrBits.html . Fig. 7.
Extinction-corrected K s magnitude as a function of the periodfrom the VSX catalogue, colour-coded by A K s . tion. On the other hand, all VVV sources having at least one ob-servation and at least one counterpart with VSX catalogue withinan 1-arcsec radius are considered in V SX.By matching with V SX within a radius of 1 arcsec, wefound that 330838 single-detection point sources are not presentin the VIVA catalogue. Three main reasons might be the ori-gin of these di ff erences: (a) our sample contains sources withfewer than ten non-flagged observations, (b) the signals reportedin other wavelengths are not clear in the IR, and (c) the proce-dures used by Ferreira Lopes et al. (2020) fail. To understandthese di ff erences, we analysed the stars whose periods are re- Article number, page 9 of 14 & A proofs: manuscript no. 34356corr_revised
Fig. 8.
Diagram for the period as a function of the extinction-corrected K s (top panel) and absolute M K s (bottom panel) magnitudes. The VSXvariable types are highlighted. VSX types with fewer than ten stars were removed because they are not statistically relevant. The group calledLPV (long-period variables) was removed because the measurement of its period in the VSX catalogue has low precision. Contours represent the68th percentile (the contour colour di ff ers from the point colour) of the types with more than 100 points and are represented as RRAB by theyellow contour, RRC in dark green, PULS in cyan, CWB in gray, MISC in blue, SR by the purple line, and VAR by the golden line. The verticaldashed purple lines represent the separation limits between DSCT and RRC (log P = − . P = − . P = P = . N in the bottom panel is the total number of variables shown. ported in the literature, using the signal-to-noise ratio (S / N) es-timated with harmonic fits to select our sample (e.g. FerreiraLopes et al. 2015a,b).We selected a sample of 10 793 stars considering two crite-ria: S / N > / N > NG higher than 100. Following the VSX clas-sification, this sample is composed mainly of eclipsing binaries ( ∼ . ∼ . ∼ . ∼ . Article number, page 10 of 14erpich et al.: VVV survey near-infrared colour catalogue of known variable stars riods might be inaccurate (Carmo et al. 2020), as well as somethat appeared to have a very low amplitude signal.On the other hand, for the selected objects displaying goodsignals (for which three examples are shown in Fig. 9), we ob-served that ∼
98% of them belong to the VVV bulge fields.These regions are more crowded and dustier than those locatedin the disk area, and as a consequence, the number of missedsources is expected to be higher. Additionally, noting that thepeak of the magnitude distribution of our sample is at K s ∼ / N (see middle panel of Fig. 9). These correspond to less than0 .
5% of the variable stars with more than ten non-flagged obser-vations. We also observed that some flagged observations maybe useful for identifying LPVs (see the top diagrams in Fig. 9).Only the LPVs appear to show variability at the flagged measure-ments. On the other hand, the flagged observations might also in-dicate the zero-point error (see the middle panels in Fig. 9). Wecaution that these observations must be used carefully becausethey can also increase the noise of the signal.In summary, our results agree with those of Ferreira Lopeset al. (2020), that is, the VIVA catalogue includes about 99%of all variable stars found in the VVV Data Release 4 that havegood signal. The missed sources correspond to faint stars mainlywith a low S / N. Additional observations from VVVX can helpsolve this matter. Moreover, Hajdu et al. (2020) identified twoindependent types of bias in the photometric zero-points in VVVdata that can also improve the data quality.
5. Conclusions
We presented a near-infrared catalogue that includes accurate in-dividual coordinates, magnitudes, and extinctions as well as dis-tances based on Gaia parallaxes for variable stars presented inthe VSX catalogue and also in the VVV near-infrared colourcatalogue. Our CMDs, colour-colour diagrams, and surface den-sity distributions for the di ff erent types allow us to give a globalcharacterization of Galactic variables stars, including importantdistance indicators such as RR Lyrae, Cepheids, abd Miras.Our results suggest that the current knowledge of variabil-ity in the Galaxy is biased to nearby low extinct stars and thattherefore the larger part of the Galaxy disk and bulge is still un-explored by stellar variability studies.We provide a set of several thousand candidates to di ff erentVSX types of variables using the period space to constrain theregion they occupy in the period versus magnitude diagram. Thisresult can be used to target specific samples of stars to determinethe characteristics through follow-up campaigns.Future studies of the Milky Way variable stars will be rev-olutionised by the LSST survey in the next decade, for whichthe discovery of hundreds of thousands of variables is expected.Thus, the present catalogue also provides the groundwork tocharacterise the results of future projects. Another important as-pect to consider is the near-infrared versus optical variability.The searches performed at di ff erent wavelengths result in di ff er-ent relative numbers of variables, and this must be consideredwhen the number of Rubin Observatory findings in this field ispredicted (e.g. Pietrukowicz et al. 2012).Near-infrared surveys such as the VVV are e ffi cient at lowGalactic latitudes close to the plane, where optical surveys areusually blinded by the absorption in the interstellar medium. Inthe same way that Baade’s Window was important for opticalstudies of the Galactic bulge, the near-infrared surveys also profitfrom the study of the windows that have recently been found inthe Milky Way plane (Minniti et al. 2018). By analysing the photometric flags from the VVV catalogue,we identified a misclassification regarding noise and mainly non-stellar objects. These stars are misclassified even inside the mag-nitude range that allows a good verification of the data and typi-cally corresponded to ∼
20% of our sample for all VVV bands.Additionally, we identified a considerable number of vari-ables inside our sample that were missed during the selection ofthe VIVA catalogue. They are mostly sources with low signal-to-noise ratio, stars with photometric problems, and for about1% of them, the procedure used to identify the variability failed.These sources can help to improve the variability analysis of theVVV survey and to verify the accuracy of the VVV photomet-ric flags. This information is valuable to improve the proceduresand completeness of future releases.
Acknowledgements.
We thank the anonymous referee for the useful suggestionsto improve this paper. We gratefully acknowledge the use of data from the ESOPublic Survey program ID 179.B-2002 taken with the VISTA telescope, and dataproducts from the Cambridge Astronomical Survey Unit (CASU). F. R. H. thanksto Federal University of Santa Catarina for the computational support, and theIAG / USP and FAPESP program 2018 / / Brazil through projects 308968 / / / RCUK’sPCI grant DPI20140066. C.E.F.L. acknowledges a PCI / CNPQ / MCTIC post-doctoral support, MCTIC / FINEP (CT-INFRA grant 0112052700), and the Em-brace Space Weather Program for the computing facilities at INPE. T.F acknowl-edges the financial support from the PIBIC CNPq / Brazil.
References
Alard, C., Guibert, J., Bienayme, O., et al. 1995, The Messenger, 80, 31Alcock, C., Akerlof, C. W., Allsman, R. A., et al. 1993, Nature, 365, 621Alcock, C., Allen, W. H., Allsman, R. A., et al. 1997, ApJ, 491, 436Alonso-García, J., Saito, R. K., Hempel, M., et al. 2018, A&A, 619, A4Aubourg, E., Bareyre, P., Bréhin, S., et al. 1993, Nature, 365, 623Bailer-Jones, C. A. L., Rybizki, J., Fouesneau, M., Mantelet, G., & Andrae, R.2018, AJ, 156, 58Bond, I. A., Abe, F., Dodd, R. J., et al. 2001, MNRAS, 327, 868Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1989, ApJ, 345, 245Carmo, A., Ferreira Lopes, C. E., Papageorgiou, A., et al. 2020, MNRAS, 498,2833Catelan, M., Minniti, D., Lucas, P. W., et al. 2011, in RR Lyrae Stars, Metal-PoorStars, and the Galaxy, ed. A. McWilliam, Vol. 5, 145Chen, X., Wang, S., Deng, L., de Grijs, R., & Yang, M. 2018a, ApJS, 237, 28Chen, X., Wang, S., Deng, L., de Grijs, R., & Yang, M. 2018b, ApJS, 237, 28Contreras Ramos, R., Minniti, D., Gran, F., et al. 2018, ApJ, 863, 79Dékány, I., Hajdu, G., Grebel, E. K., et al. 2018, ApJ, 857, 54Downes, R. A., Webbink, R. F., Shara, M. M., et al. 2001, PASP, 113, 764Drake, A. J., Djorgovski, S. G., Catelan, M., et al. 2017, MNRAS, 469, 3688Drake, A. J., Graham, M. J., Djorgovski, S. G., et al. 2014, ApJS, 213, 9Ferreira Lopes, C. E. & Cross, N. J. G. 2016a, A&A, 586, A36Ferreira Lopes, C. E. & Cross, N. J. G. 2016b, A&A, 586, A36Ferreira Lopes, C. E. & Cross, N. J. G. 2017a, A&A, 604, A121Ferreira Lopes, C. E. & Cross, N. J. G. 2017b, A&A, 604, A121Ferreira Lopes, C. E., Cross, N. J. G., Catelan, M., et al. 2020, MNRAS, 496,1730Ferreira Lopes, C. E., Cross, N. J. G., & Jablonski, F. 2018, MNRAS, 481, 3083Ferreira Lopes, C. E., Dékány, I., Catelan, M., et al. 2015a, A&A, 573, A100Ferreira Lopes, C. E., Neves, V., Leão, I. C., et al. 2015b, A&A, 583, A122Gaia Collaboration, Brown, A. G. A., Vallenari, A., et al. 2018, A&A, 616, A1Gaia Collaboration, Eyer, L., Rimoldini, L., et al. 2019, A&A, 623, A110Gaia Collaboration, Prusti, T., de Bruijne, J. H. J., et al. 2016, A&A, 595, A1Gilmore, G., Randich, S., Asplund, M., et al. 2012, The Messenger, 147, 25Gonzalez, O. A., Rejkuba, M., Zoccali, M., et al. 2012, A&A, 543, A13Hajdu, G., Dékány, I., Catelan, M., & Grebel, E. K. 2020, Experimental Astron-omy, 49, 217Iorio, G. & Belokurov, V. 2020, arXiv e-prints, arXiv:2008.02280Ishihara, D., Onaka, T., Kataza, H., et al. 2010, A&A, 514, A1
Article number, page 11 of 14 & A proofs: manuscript no. 34356corr_revised
Fig. 9.
Time domain (on the left side) and phased light curves (on the right side) for examples of variable stars that are missing in the VIVAcatalogue. The light curves display the VSX nameat the top and the VIVA ID inside the legend box, and the corresponding phase diagram showsthe VSX type (Type), the VVV K s magnitude, the number of good observations NG, and the S / N estimated with harmonic fits at the top, with theVSX period and the ppErrbits flags in the legend box inside the diagram. Blue points represent ppErrbits = ppErrbits (cid:44) Ivezic, Z., Axelrod, T., Brandt, W. N., et al. 2008, Serbian Astronomical Journal,176, 1Jayasinghe, T., Kochanek, C. S., Stanek, K. Z., et al. 2018, MNRAS, 477, 3145Kaiser, N., Aussel, H., Burke, B. E., et al. 2002, in Proc. SPIE, Vol. 4836, Sur-vey and Other Telescope Technologies and Discoveries, ed. J. A. Tyson &S. Wol ff , 154–164Kim, S.-L., Lee, C.-U., Park, B.-G., et al. 2016, Journal of Korean AstronomicalSociety, 49, 37Kleinmann, S. G., Lysaght, M. G., Pughe, W. L., et al. 1994, Ap&SS, 217, 11Lawrence, A., Warren, S. J., Almaini, O., et al. 2007, MNRAS, 379, 1599Luri, X., Brown, A. G. A., Sarro, L. M., et al. 2018, A&A, 616, A9McMahon, R. G., Banerji, M., Gonzalez, E., et al. 2013, The Messenger, 154, 35Minniti, D. 2018, The Vatican Observatory, Castel Gandolfo: 80th AnniversaryCelebration, 51, 63Minniti, D., Lucas, P. W., Emerson, J. P., et al. 2010, New A, 15, 433Minniti, D., Saito, R. K., Gonzalez, O. A., et al. 2018, A&A, 616, A26Mowlavi, N., Lecoeur-Taïbi, I., Lebzelter, T., et al. 2018, A&A, 618, A58Neugebauer, G., Habing, H. J., van Duinen, R., et al. 1984, ApJ, 278, L1Pietrukowicz, P., Minniti, D., Alonso-García, J., & Hempel, M. 2012, A&A, 537,A116Pigott, E. & Englefield, H. C. 1786, Philosophical Transactions of the RoyalSociety of London Series I, 76, 189Pojmanski, G., Pilecki, B., & Szczygiel, D. 2005, Acta Astron., 55, 275Saito, R. K., Hempel, M., Minniti, D., et al. 2012, A&A, 537, A107Saito, R. K., Minniti, D., Angeloni, R., et al. 2013, A&A, 554, A123Samus’, N. N., Kazarovets, E. V., Durlevich, O. V., Kireeva, N. N., & Pas-tukhova, E. N. 2017, Astronomy Reports, 61, 80Soszy´nski, I., Dziembowski, W. A., Udalski, A., et al. 2011, Acta Astron., 61, 1Udalski, A., Szymanski, M., Kaluzny, J., et al. 1993, Acta Astron., 43, 289Udalski, A., Szyma´nski, M. K., & Szyma´nski, G. 2015, Acta Astron., 65, 1Wright, E. L., Eisenhardt, P. R. M., Mainzer, A. K., et al. 2010, AJ, 140, 1868York, D. G., Adelman, J., Anderson, Jr., J. E., et al. 2000, AJ, 120, 1579 Article number, page 13 of 14 & A proofs: manuscript no. 34356corr_revised
Appendix A: Merged classes and what they contain
In this section we list all VSX types that were merged into thefive large classes discussed in Section 3. The classes include thefollowing types: – Eclipsing: E, EA, EB, EP, EW, EC, ED, ESD, AR, D, DM,DS, DW, EL, GS, HW, K, KE, KW, PN, SD, WD. – Rotating: ACV, BY, CTTS / ROT, ELL, FKCOM, HB, LERI,NSIN ELL, PSR, R, ROT, RS, SXARI, SXARI / E, TTS / ROT,WTTS / ROT. – Pulsating: ACEP, ACYG, AHB1, (B), BCEP, BCEPS, BL,BLAP, BXCIR, CEP, CW, CWA, CWB, CWB(B), CWBS,GWLIB, CW-FO, CW-FU, DCEP, DCEP(B), DCEPS,DCEPS(B), DCEP-FO, DCEP-FU, DSCT, DSCTC, DSCTr,DWLYN, GDOR, HADS, HADS(B), L, LB, LC, LPV, M,O, PPN, PULS, PVTEL, PVTELI, PVTELII, PVTELIII,roAm, roAp, RR, RRAB, RRC, RRD, RV, RVA, RVB,SPB, SR, SRA, SRB, SRC, SRD, SRS, SXPHE, SXPHE(B),V361HYA, V1093HER, ZZ, ZZA, ZZB, ZZ / GWLIB, ZZO,ZZLep. – Eruptive: BE, cPNB[e], CTTS, DIP, DPV, DYPer, EXOR,FF, FSCMa, FUOR, GCAS, I, IA, IB, IN, INA, INAT, INB,INS, INSA, INSB, INST, INT, IS, ISA, ISB, RCB, SDOR,TTS, UV, UVN, UXOR, WR, WTTS, YSO, (YY), ZZA / O. – Other: Includes all Cataclismic Variables (AM, CBSS,CBSS / V, CV, DQ, IBWD, N, NA, NB, NC, NL, NL / VY,NR, S, SN, SN I, SN Ia, SN Iax, SN Ib, SN Ic, SN Ic-BL,SN II, SN IIa, SN IIb, SN IId, SN II-L, SN IIn, SN II-P, SN-pec, UG, UGER, UGSS, UGSU, UGWZ, UGZ, UGZ / IW, V,V838MON, VY, ZAND), X-Ray (BHXB, HMXB, IMXB,LMXB, X, XB, XBR, XJ, XN, XP, XPR), and other objects(AGN, APER, BLLAC, CST, GRB, Microlens, MISC, non-cv, NSIN, PER, QSO, S, SIN, Transient, UNC, VAR, VBD).
Appendix B: Summary of VSX classes