Standard candles from the Gaia perspective

0.6 MB·English

by  Laurent Eyer

Standard candles from the Gaia perspective
Laurent Eyer • Lovro Palaversa • Nami Mowlavi •
Pierre Dubath • Richard I. Anderson •
Dafydd W. Evans • Thomas Lebzelter •
Vincenzo Ripepi • Laszlo Szabados • Silvio Leccia
2 • Gisella Clementini
1
0
2

n
a
J

3
2
Abstract main will be achieved thanks to unprecedented astro-

The ESA Gaia mission will bring a new era to the metric precision, whole-sky coverage and the combina-
]
R domain of standard candles. Progresses in this do- tion of photometric, spectrophotometric and spectro-
S scopic measurements. The fundamental outcome of
.
h LaurentEyer the mission will be the Gaia catalogue produced by
p the Gaia Data Analysis and Processing Consortium
Geneva Observatory, University of Geneva, 1290 Sauverny,
-
o Switzerland (DPAC), which will contain a variable source classifi-
r LovroPalaversa cation and specific properties for stars of specific vari-
t
s Geneva Observatory, University of Geneva, 1290 Sauverny, ability types. We review what will be produced for
a
Switzerland Cepheids, RR Lyrae, Long Period Variable stars and
[
NamiMowlavi eclipsing binaries.

1
Geneva Observatory, University of Geneva, 1290 Sauverny,
v
Switzerland Keywords stars:distance; stars: variables; stars: bi-
9
8 ISDC, Geneva Observatory, UniversityofGeneva, 1290Versoix, naries; stars: statistics; cosmology: distance scale;
Switzerland
8 space vehicles; surveys; catalogs
4 PierreDubath
1. ISDC, Geneva Observatory, UniversityofGeneva, 1290Versoix,
0 Switzerland 1 Introduction
2 RichardI.Anderson
1
Geneva Observatory, University of Geneva, 1290 Sauverny,
: Thesubjectofstandardcandlesisafundamentalscien-
v Switzerland
tificcasethatwillgreatlybenefitfromtheGaiamission.
i DafyddW.Evans
X In this respect Gaia is unique and this scientific sub-
InstituteofAstronomy,CambridgeUniversity,MadingleyRoad,
r jectwilltakeadvantagefromallaspectsofthemission.
CambridgeCB30HA,UK
a
The astrometry will obviously provide a major contri-
ThomasLebzelter
bution. However, other aspects of the Gaia measure-
DepartmentofAstronomyUniversityofVienna,Tuerkenschanzs-
trasse17,1180Vienna,Austria ments will contribute to this subject as well. The clas-
VincenzoRipepi sicalstandardcandlesareRRLyraeandCepheidstars.
But Gaia will also offer the possibility to exploit other
INAF - Astronomical Observatory of Capodimonte, Salita
Moiariello16,80131Napoli,Italy classes of variable stars as standardcandles. Examples
LaszloSzabados of“non-classical”standardcandlesincludeLongPeriod
Konkoly Observatory of the Hungarian Academy of Sciences, Variables (LPVs) (Feast et al. 1989; Matsunaga et al.
1121Budapest,Hungary 2009), OGLE Small Amplitude Red Giants (OSARGs;
SilvioLeccia Wray et al. 2004), eclipsing binaries (Paczyn´ski 1997)
INAF - Astronomical Observatory of Capodimonte, Salita and Large Amplitude δScuti stars (McNamara 1997).
Moiariello16,80131Napoli,Italy Certain types of stellar variability occur within spe-
GisellaClementini cificmassandmetallicityrangesandatgivenevolution-
INAF-Osservatorio Astronomico di Bologna, via Ranzani 1, arystages. ThesizeoftheGaiadatasetwillensurethat
40127, Bologna,Italy
all these cases will be covered statistically. Gaia data
2
willallowustoestablishthefundamentalastrophysical number contains estimated dead-times). The launch is
properties of these stars, in particular their luminos- foreseenfor2013. Therewillbeanalertsystemandin-
ity, and thereby establish their usefulness as standard termediate data releases throughout the mission. The
candles. They will also determine their cosmic scatter. final results will be made available by 2020–2021.
Gaia will therefore be able to test the universality of A summary of performances can be found on the
many different standard candles. Gaia webpage (http://www.rssd.esa.int/Gaia) under
The astrometric precision of the Hipparcos satellite Science Performances. The numbers displayed in ta-
has been exploited to its limit in the case of standard bles 1 and 2 are extracted from this webpage (as of
candles. Standard candles like Cepheids or RR Lyrae October 2011).
can be considered “far” for Hipparcos precision. The
statistical properties of distance, parallaxand absolute The scanning law The Gaia scanning law has been
magnitudearecomplex. Feast & Catchpole(1997)con- designedin orderto optimize the astrometricresults of
strainedthezeropointoftheperiod-luminosityrelation the mission. The Gaia satellite has two fields of view
with a subtle and sound method. However, their zero (FOV) of 0.7×0.7 deg2 each. These two viewing direc-
point has been subject to discussions. It is worth men- tions are separated by an angle of 106.5 deg. The two
tioning that the number of Cepheids retained for con- FOV images are superposed on the same focal plane
strainingtheirsolutionwasonly26. Thedistancemod- thatconsistsof106CCDs,totalingnearly1billionpix-
ulus of the Large Magellanic Cloud (LMC), obtained els. As the satellite rotates around its axis with a pe-
with the Hipparcos-derived zero-point is 18.7±0.10, riodof6hours,thestarsaresweepingthroughthefocal
which seems a bit too far compared to other standard plane. The CCDs are read in Time Delay Integration
estimates that are close to 18.5 (cf. Clementini et al. mode. The rotation axis of the satellite is precessing
2003). onaSun-centeredconewithanopeningangleof45de-
DecadesafterthehistoricaldiscoveryoftheCepheid greesandaprecessionperiodof63days. Thisconstant
period-luminosity relation by Henrietta Leavitt, sur- angleoftherotationaxiswithrespecttotheSun’sposi-
veys towards the Magellanic Clouds such as OGLE, tion gives the peculiar dependency of the scanning law
MACHO and EROS, provided a remarkable contribu- oneclipticcoordinates. Theaveragenumberoftransits
tion to the subject of standard candles. Even if the (one passage through the FOV) is about 70 but varies
distances to the Magellanic Clouds are not known pre- between 40 and 250, depending on the sky position.
cisely, they can be considered at a fixed distance with Properties of the scanning law have been presented
onlymoderatedepthandthereforethedifferenceinap- by Eyer & Mignard (2005). Although there have been
parent magnitude can be interpreted as a difference in changesinsatellitedesign,the conclusionsofthe study
luminosity. This is whythe MagellanicCloudshaveof-
remainvalidfor the AstrometricField (AF). The spec-
fered such a high scientific potential. Gaia,as a whole-
tral window of Gaia sampling contains high peaks
sky survey, will observe the Magellanic clouds, as well
at high frequencies, limiting aliasing when compared
as the halo and the plane of the Galaxy (with paral-
to large scale ground-based surveys (see Eyer et al.
laxes). The inter-comparison of different populations
2009). The period recovery of periodic signals is very
ofstandardcandle classeswitha singleinstrumentwill
high even at relatively low signal to noise ratio (see
have an enourmous impact, since, for the first time,
Eyer & Mignard 2005). Handling of semi-regular or ir-
they can be calibrated to a homogeneous reference.
regular variables might be difficult due to the sparsely
sampled Gaia time series. Furthermore, the semi-
regular or irregular variables may contaminate the
2 A quick review of the Gaia mission
samples of periodic objects. Particular cases such as
double-mode Cepheids or Blazhko RR Lyrae stars will
In this section, we present properties of the Gaia mis-
also require special analysis.
sion that are relevant to standard candles.
Gaia is a spacecraft of the European Space Agency
The astrometric performance The mission’s success
(ESA) thatwillbe locatedatthe LagrangianL2point,
depends critically on the astrometric performance
1.5millionkmawayfromEarth. Itwillobserveabout1
achieved. Therefore, performance estimations have
billionobjectswithamagnitudebetweenV ≃6and20
been the subject of constant attention throughout the
mag. Themeasurementsgatherastrometric,photomet-
development of the spacecraft. The latest numbers
ric, spectrophotometric and spectroscopic data. The
available,re-evaluatedduringthe criticaldesignreview
lengthofthemissionis5yearswithapossibleoneyear
in April 2011, are displayed in Table 1.
extension. For a duration of 5 years, the average num-
ber of measurements will be about 70 per object (this
3
Table 1 Astrometric error for end-of-mission parallax as Table2 Radialvelocityend-of-missionerrorasafunction
a function of spectral typeand magnitude of spectral typeand magnitude
B1V G2V M6V Spectral type V Radial velocity error
V −I −0.22 0.75 3.85 B1V 7 1kms−1
Bright stars 6<V <12 6<V <12 8<V <14 12 9kms−1
5-14µas 5-14µas 5-14µas G2V 13 1kms−1
V=15 26µas 24µas 9µas 16.5 13kms−1
V=20 330µas 290µas 100µas K1III-MP 13.5 1kms−1
−1
(metal-poor) 17 13kms
These numbers result from studies which have been
recently quite stable; the current numbers are consis- Gaia performs also low resolution spectrophotome-
tentwiththoseobtainedinthepastfewyears. Globally, try. The Blue Photometer (BP) covers the wavelength
the performanceis within the initial requirementswith range from 330 to 680nm while the Red Photometer
only some minor non-compliances. (RP) covers the range 640–1050nm. The RP has red-
The numbers shown in Table 1 represent estimated enhancedCCDssothatlongerwavelengthsarereached.
errors on parallax at the end of the mission. They The requirements were not formulated in terms of ac-
should be multiplied by 0.8 and 0.5 in order to ob- quiring desired astrophysical quantities, but in terms
tain errors for end-of-mission position proper motion ofphotometric precisioninpseudo-bands. The BPand
(µas/year), respectively. RPspectrahaveeach60samples. Thephotometricpre-
In theory, the performance improves with mission cision of the integrated or “mean” spectra are given in
length L. The parallax and position errors scale as Figure1. However,again,thesenumbersshouldbeseen
L−0.5 while the proper motion error varies as L−1.5. as theoretical limits. The lower limit of the calibration
However,theastrometricsolutionofthefirst18months error is estimated to be at the level of 10 to 30mmag.
islikelytobeaffectedbysystematicerrors. Thegoalof It should be noted that the error estimations from the
the consortium will be to search and correct for these April2011reviewhavelargererrorsforthislowerlimit.
errors through the subsequent iterations in such a way Finally, the Sky-Mapper (SM) CCDs will also produce
as to reach the expected accuracy at the end of the photometricmeasurements. Themeannumberoftran-
mission. As a consequence, the scaling of the different sits in G, BP and RP is estimated to reach 70 over 5
solutions produced during the missionmay not strictly years.
follow the above rules.
ForaCepheidlocatedat12kpcwitha10dayperiod, The Radial Velocity Spectrometer performance The
a relative parallax error of 10% is expected assuming Radial Velocity Spectrometer (RVS) is a near-infrared
no extinction. The same relative error is reached for a instrument with a resolution of 11,500 and spans the
6 kpc Cepheid if the extinction is A = 5 mag. For wavelengthrangefrom847to874nm,whichcoversthe
V
Cepheids, Gaia will cover a significant fraction of the Calcium triplet. The RVS instrument will survey the
Galaxy with a very good precision. whole-skyuptomagnitudeV ∼15or17(dependingon
thespectraltypeofthestars)withend-of-missionerror
The photometric and spectrophotometric performance levels from 1 to 10 kms−1, depending on the spectral
The astrometric field is also producing a white light type and magnitude, cf. Table 2. The number of mea-
(called G-band) magnitude. As there is no filter, the surements in the RVS will be about 40 per object over
bandpass is only limited by the optical properties of the5yearmission. Thenumberoftransitsforthe RVS
the system (reflectivity of the mirror, response of the is reduced, since there are only 4 dedicated CCDs per-
CCDs, etc.). The wavelength coverage is from 330 to pendicular to the scanning direction whereas the SM,
1050nm. The photometric G-band precision should be G, BP, RP instruments have 7.
ofveryhighquality,asitisthesumofthe9CCDmea-
surements over one FOV transit. The transit/epoch
photometry accuracy as a function of the magnitude is 3 Standard Candles
given in Figure 1. The lower limit of the calibration
error is estimated to be at the level of 1 mmag. The Theknowledgeaboutstandardcandleswillbenefitfrom
per-CCD photometry will also be available and will al- allaspects ofthe Gaia mission: its astrometry,its pho-
low detection of variability on very short time scales tometry and spectrophotometry, as well as its spectro-
(tens of seconds). metric radial velocity measurements.
4
45$67"&’()*+&"(668’(6+9*
2-"%./03
,-"%./01
2-"%./01
,-"%./03
!"#$%"&’()*+&
Fig. 1 Per transit photometric error in G band, integrated BP and RP as a function of G magnitude. The sawtooth
structure at the bright magnitudes is a consequence of the gating system which allows observations of bright stars by
limiting theexposed part of theCCD, thusreducingthe integration time
Theastrometrywillallowcalibrationofluminosities Furthermore Gaia will conduct a global survey, col-
of the standard candles thanks to the Gaia parallaxes. lecting photometry and spectrophotometryon a multi-
For a given variability type, there exists an interplay epochbasisthatwillallowthedetectionofnewobjects
between luminosity, distance (distribution within the in each standard candle category, e.g. new Cepheids,
Galaxy) and Gaia precision for the corresponding ap- new RR Lyrae stars and eclipsing binary systems, see
parent magnitudes. However the number of standard Table 3. Due to its diversity the impact of such a
candles of a given type with good and useful astrome- harvest is difficult to forecast. However, the physics
try will increase by one to several orders of magnitude driving the variability and instability, and in particu-
with respect to the present situation. lar how metallicity affects the variability properties of
The Gaia multi-epoch photometry is also advanta- a star will be systematically studied. Both aspects are
geous in the case of standardcandles with large ampli- essential for the calibration of standard candles using
tudes. Forthesecases,lightcurvescanbemodeledand Gaiaastrometry. Inaddition,wemayfindentirelynew
mean luminosities can be defined. Finally, with uni- classes of standard candles within the Gaia data.
form and homogeneous photometric and spectrophoto- Spectrophotometry will also provide estimates of
metric measurements of Gaia, calibrations of period- stellar parameters for the standard candles.
luminosity-color relations can be established. Radial velocity data that will be obtained for the
Another less often mentioned benefit from astrom- mostluminousobjectswillallowcomputationofphysi-
etry is that Gaia will be able to determine the or- calparametersofsinglestarpulsators,usingtheBaade-
bit of astrometric binary stars from their movement Wesselink method and orbital parameters of binary
on the sky. With radial velocity measurements also systems with the Wilson-Devinney-like code (see Sec-
from Gaia, the physical orbit can be determined (e.g. tion 4.3).
Zwahlen et al.2004). Thisisanotherwaytodetermine The detection of new standard candle objects and
distances purely geometrically. characterization of their astrophysical parameters will
5
provide a wealth of data to test, on a statistical level, stars observed in the instability strip do not show the
the universality of standard candle relations such as expected photometric variability. The comparison of
the period-luminosity-metallicity relation. Once cali- certaintypesofvariablestarsindifferentknownstellar
brated to high accuracy with Gaia, standard candles populations,suchasourGalaxy,openorglobularclus-
can be used to extend studies of Galactic and extra- ters, or in the local groupof galaxies,provides a useful
galactic structure beyond the astrometric performance way to learn about those variable stars as a function
capabilities of Gaia. As examples, we mention tracing of different properties such as the metallicity. Open
ofthegalacticbarusingOSARGs(Wray et al.2004),or clusters, for example, are natural laboratories for this,
thepossibilitytoconstrainthethree-dimensionalstruc- as their member stars are assumed to share the same
ture of the Large and Small Magellanic Clouds by us- age, initial chemical abundances, distance and redden-
ing period-luminosity relations of pulsating red giants ing. In this way, the already listed 2100 open clusters
(Lah et al. 2005). Synergies of Gaia with other large in the disk of our Galaxy (Dias et al. 2002), spanning
surveys such as LSST are also easily envisioned. The a large interval in age and Galacto-centric distances,
RR Lyrae stars calibrated with Gaia can be used by have been used as an excellent tool to probe both the
LSST to fully characterize the halo of our Galaxy. chemical and dynamical structure and the evolution
Differentstandardcandlescorrespondtostarsatdif- of the Galactic disk. The variety of variable star con-
ferent stages of evolution. If the evolutionary stages of tentfromone cluster to another,associatedto the spe-
standard candles are known, the formation history of cific variability properties characterizing each phase of
the Galaxy can be traced (see e.g. Clementini 2011). evolution, provides independent measurements for the
Another interesting application of standard can- physical parameters of open clusters. Yet, this con-
dles is in mapping the distribution of the interstellar nection is not well known so far. One difficulty is the
medium. determination of the membership of the cluster. Gaia
Despite many studies devoted to variable stars on will put the question of membership on solid ground
one side and to populations of stars on the other side, (see van Leeuwen 2009 for what has been done with
not many studies have so far been devoted to the com- Hipparcos).
bination of both fields, i.e. the study of variable stars In Table 3 we review the number of discovered ob-
in relation to stellar populations. Here are two exam- jects by different large scale surveys. The estimates
ples of questions showing the need for more studies in by Eyer & Cuypers (2000) are only for the Galaxy or
this domain. A basic question that arises when talking some component of the Galaxy. The large uncertainty
about variablestars in stellar populations concernsthe inthese numbers shows that there are many unknowns
fraction of stars that are expected to vary. The answer inthisdomain. FromTable3,theoptimist’sviewwould
is not obvious, as attested by the difficulty to predict be that Gaia will multiply by nearly 5 the number of
the numberofvariableobjectsexpectedtobe observed Galactic RR Lyraestars,by 10 the number ofGalactic
inaspecificsurvey,seeTable3. Forexample,thenum- Cepheids,bynearly40theGalacticLPVsandbynearly
ber of eclipsing binaries predicted to be detected by 1000the Galactic eclipsing binaries. To this table vari-
Gaia varies from 0.5 million (Dischler & Soederhjelm ables from the LMC and SMC are added. Due to the
2005) to 7 million (Zwitter 2002), which represents a bright limit of Gaia at V∼6 mag, only few Cepheids
factor of ten uncertainty. Moreover, it is not always willbe missed(inthe Hipparcoscataloguewefound 22
possibletopredictwhetherastaratagivenlocationin stars for which the maximum light is brighter than 6).
theHR-diagramwillpulsateornot. Starsareexpected A word on Supernovae: Gaia will have an alert sys-
to pulsate in some specific areas of the HR-diagram, tem that will detect supernovae and other transient
called instability strips, e.g. the “classical instability events. The estimated number of supernovae brighter
strip” is hosting Cepheids, RR Lyrae or δScuti stars. than magnitude 19 is around 6,000. One third of these
The identification of the borders of those instability willbeobservedbeforetheirmaximum. Thissubjectis
strips has been a successful tool to better understand coveredby Gilmore in this volume.
the pulsational mechanism and the parameters driving
the photometric variability at the surface of the stars.
The crucial role of convection and of its coupling with 4 Gaia Data Processing and Analysis
pulsationhas,forexample,beenhighlightedinexplain- Consortium activities on standard candles
ing the red border of the Cepheid instability strip. It
explains why stars located in the HR-diagram between Software development is often underestimated in large
theinstabilitystripandtheredgiantbranchdonotpul- scale projects. For Gaia, however, it was recognized
sate. However, what remains unexplained is why some early-onthatthesoftwaredevelopmentisakeyelement
6
Table 3 Numbers of RR Lyrae stars, Cepheids, Long Period Variable stars, eclipsing binaries, known in the Galaxy,
the LMC and SMC and predicted numbers in the Galaxy for the Gaia mission. For Hipparcos the numbers are taken
from ESA (1997); the numbers from ASAS should be take with care (see Berdnikov et al. 2009, variable types have been
selected with the “only” option) and are taken from the ASAS webpage as of May 2011, the estimates for Gaia are from
Eyer & Cuypers (2000); The “other” line are other publications: Galactic Cepheids (Fernieet al. 1995), Gaia Cepheids
(Windmark et al. 2011), Gaia eclipsing binaries (Dischler & Soederhjelm 2005, Zwitter 2002), OGLE-III SMC RR Lyrae
stars (Soszyn´ski et al. 2010b), OGLE-III LMC RR Lyrae stars (Soszyn´ski et al. 2009a), OGLE-III bulge RR Lyrae stars
(Soszyn´ski et al.2011),OGLE-IIISMCCepheids(Soszyn´ski et al.2010a),OGLE-IIILMCCepheids(Soszyn´ski et al.2008),
EROSLMCLPV(Spano et al.2011),OGLE-IISMCeclipsingbinaries(Wyrzykowskiet al.2004),OGLE-IIILMCeclipsing
binaries (Graczyk et al. 2011)
RRLyrae Cepheid LPV Eclipsing bin.
Known
Hipparcos 186 273 1,238 917
ASAS 1,635 872 2,793 5,911
Other (bulge) 16,839 509
Predicted
Eyer & Cuypers (bulge) 15,000-40,000 2,000-8,000 200’000 3,000,000
(halo) 70,000
Other 9,000 500,000; 7,000,000
LMC 24,906 3,361 37,047 26,121
SMC 2,475 4,630 1,351
for its success. Indeed, the Gaia data processing and terized. A number of attributes are computed to char-
analysis is a tremendous task, due to the large amount acterizethesources. Someofthemreflectglobalstellar
of raw data to be processed (in the order of 100 com- properties, such as mean color or absolute magnitude,
pressed Terabytes in 5 years), but even more so due whereasothersdescribesomeofthelightcurvefeatures.
to the complex and interwined relationships between Anumberofstatisticalparametersarederivedfromthe
astrometry, photometry, spectrophotometry and spec- magnitudedistribution. SinceweknowthatGaiahasa
troscopy. In addition, since the targeted accuracy is relatively good performance on periodic objects, a pe-
higher than anything ever obtained before for so many riod search is carried out and the folded light curves
stars, the processes have to be self-calibrating, going are modeled with Fourier series. Many harmonics are
througha numberofiterations,with eachsetofresults fitted, but only those that are significant according to
providing inputs for the next run. This is most obvi- an F-test are kept. The number of harmonics is also
ous for the astrometric global iterative solution. The limited if there are gaps in the time sampling to avoid
global reference frame of sky positions will be built non-physical large model excursions in regions devoid
gradually, measuring/modeling parallaxes and proper of measurements.
motions, andeliminating deviating points such as mul-
Step 3: Variability classification Once variables are
tiple stars.
identified and characterized, multiple classification
The task of the variability analysis and processing
methods are applied. We decompose this into three
wasgiventotheCoordinationUnit7(CU7),wherethe
subtasks: supervised methods, clustering techniques,
work is decomposed into several steps.
extractors (a specific variability type is selected us-
ing all the astrophysical knowledge). Automated and
Step 1: Variability detection The photometric (CU5)
efficient variable star detection and classification are
and spectroscopic (CU6) groups are in charge of de-
critical components of large-scale surveys. They are
tecting variable objects applying general-purpose algo-
required both to study stellar population properties
rithms, such as some statistical standard tests. The
andtoprovidecandidatesforfurtherdetailedinvestiga-
Special Variability Detection (within CU7) is defined
tion of individual cases. Tests have been performedfor
to implement specific algorithms which take advantage
supervised methods by using cleaned Hipparcos light
of what we know about a particular type of variability.
curves to evaluate the ultimate performance of peri-
All variableobjects arethen storedinto the Variability
odic (Dubath et al. 2011) and non-periodic (Rimoldini
Database.
et al., in preparation) star classification schemes. The
classificationresultassociatesagivenvariablestarwith
Step 2: Variability characterization Oncevariableob-
a membership probability to a given class of variable
ject candidates are identified, their behavior is charac-
stars.
7
Step4: Specific ObjectStudies Inthethreefirsttasks, detectionof BlazhkoRR Lyraeandthe estimate ofthe
data for all objects are processed in a systematic way. Blazhko period (when possible) has been implemented
In Specific Object Studies, specific algorithms are ap- in the CU7 processing chain.
plied to objects as a function of their variability class. ForCepheids,thepipelinewillbeabletodistinguish
For example, the processing required at this stage for between populationI (ClassicalCepheids) andpopula-
the periodic Cepheid stars is different from the one tionIIobjects(BLHer,WVirclasses)thatobeydiffer-
required for the usually rather erratic distant Active ent Period-Luminosity or Period-Luminosity-Color re-
GalacticNuclei. AftertheSpecificObjectStudiesstep, lations.
all available information about the variables has been In this context, the question of binarity is also im-
extracted from the Gaia data and is available in the portant. The presence of a companion leads to smaller
Variability Database. photometric amplitudes and an altered mean luminos-
We willbe able to validate the automatedclassifica- itycomparedtothelightcurvesofsingleCepheids(e.g.
tion,inthisstep,byanalyzingtheobjectsingreaterde- Klagyivik & Szabados2009). Thiswouldleadtoaner-
tailsandstudyingthepropertiesofsub-samples(some- roneous distance estimate from the period-luminosity
times manually). The Specific Object Studies step relation,iftheeffectofthecompanionisnottakeninto
should also re-evaluate the membership probability as- account.
signed by the automated classification. Stellar parameters can be derived with the knowl-
We present what is foreseen for the Specific Object edge of multi-epoch radial velocities, measured by
Studies of standard candles, namely in section 4.1 for Gaia. The radii and the distances of Cepheids and
RR Lyrae stars and Cepheids, in section 4.2 for Long RR Lyrae can be derived with the Baade-Wesselink
Period Variables, in section 4.3 for Eclipsing Binaries (BW) method, with an error of a few % for bright ob-
and in section 4.4 for Supernovae. jects and about 10% at V magnitudes of 13-14. The
resulting quantities can be compared with the val-
Step5: GlobalVariabilityStudies Inthenextstep,the uesobtainedthroughGaiaastrometry,providingstrin-
Global Variability Studies task will investigate larger- gent constraints on the systematic errors affecting the
scale properties of variability, e.g. the period distribu- method. Inparticular,itwillbepossibletofixthevalue
tion for all Cepheids, a color-magnitude (or HR) di- ofthe “projectionfactorp”, aproportionalityconstant
agram with iso-contours of variability amplitude, etc. between pulsational and radial velocities, which is the
Given the large number of objects, special tools are in most important source of error for any version of the
development in order to facilitate the evaluation and BW technique. This will allow to safely use the BW
usage of the database content. method for faint objects outside the Galaxy.
4.2 Long Period Variables
4.1 Cepheids and RR Lyrae stars
Long Period Variables (LPVs) constitute a class of red
A good characterization of Cepheids and RR Lyrae
giant variables classically defined by the Mira stars
starsisessentialifthose objectsaretobe usedasstan-
and semi-regular variables. These variables are known
dard candles. The automatic data processing imple-
to obey several nearly parallel period-luminosity rela-
mented in the CU7 pipeline should first derive the pe-
tionsrelatedto the variouspulsationmodes. Mirasare
riod(s) and the Fourier decomposition parameters of
knowntobe fundamentalpulsators,andvariableswith
thelightcurves,i.e.theamplituderatiosandphasedif-
smalleramplitudesaretypicallyfirstorsecondovertone
ferences (see Leccia et al. 2011). These parameters al-
pulsators. TheirP-Lrelationshavebeenidentifiedfrom
low the identification of the pulsation mode(s), funda-
observations, thanks to large scale surveys initiated by
mental or higher overtones. They will also be useful
MACHO (Wood et al. 1999), OGLE (Soszyn´ski et al.
to estimatethe metalabundance of those objects and
2009b), and EROS (see e.g., Spano et al. 2011). Un-
their intrinsic colors.
fortunately, the modeling of the pulsation of those red
Double-mode RR Lyrae variables will be detected
giantsisdifficultduetothecouplingbetweenpulsation
and the period ratio fundamental/first overtone used
and convection. The agreementbetween predicted and
to estimate the mass of the targets.
observedrelationsisquitegood,buttheimpactofstel-
A significant sample of RR Lyrae (about 30%) are
larparametersonthepulsationpropertiesisstillpoorly
expected to be affected by the Blazhko effect, which
understood. Largescalesurveyswithmulti-epochpho-
is a modulation of the amplitude and/or phase of the
tometryofredgiantsintheMagellanicClouds,thedis-
periodic signal, that can lead to an erroneous determi-
tances to which are known, have been the driving tool
nationofmeanmagnitude andstellarparameters. The
to study the P-L relations of LPVs.
8
Byprovidingthedistancestoalargesetofredgiants help the sub-classification. Initial guesses of some of
in our Galaxy, Gaia will bring substantial additional the parameters may also be helpful for the computa-
dataforthosestudies,wideningthe rangeofstellarpa- tion of the full parameters of the binary system with a
rametersandrelatingthezeropointsoftherelationsto Wilson-Devinney-type code, a task in the hand of Co-
a direct measurement. It will also be possible to study ordination Unit 4 in the Gaia consortium.
various subgroups within the Milky Way and to place
nearbyandwell-studiedLPVs ontoadistinct P-Lrela- 4.4 Supernovae
tion. Thisstudy,however,requirestheknowledgeofthe
bolometric correction for each star, the value of which The light curves of cataclysmic variables and super-
strongly depends on the atmospheric chemistry (O or novae will also be characterized by CU7 and the re-
C-rich, with or without dust). The O or C-rich nature sults will be made available in the Gaia catalogue of
of the star can be assessed from the relative strengths variable stars. Contrary to the alert system of CU5
of the TiO and CN molecular bands around 7780 and whose purpose is to alert the scientific community as
8120 ˚A, which fall within the spectral range of the BP early as possible and hence is derived from basically
spectrometer of Gaia (6500-10000 ˚A). First investiga- calibrated data, the data analyzed by CU7 will consist
tionstoseewhethertheresolutionandsensitivityofthe offully calibrateddatathatcoverthe lightcurveofsu-
spectrometerissufficienttoallowforagooddistinction pernovae over the entire duration available at the time
into O or C-richstarsled to promising results awaiting of catalogue publication.
fine tuning once Gaia is in orbit.
Furthermore, special care has to be taken on a pos-
sible shift ofthe photo-centerofredgiants due to their 5 Conclusions
largeradiiandthepossibilityoflargescaleandvariable
structures ontheir surface,e.g. asthe resultofconvec- Gaia’s scientific impact on standardcandles will be re-
tion. The impact on the astrometry of Betelgeuse in markable. Clearly, there will be some limitations re-
theHipparcosH band,forexample,isestimatedtobe lated to statistical aspects of the analysis: aliasing in
p
oftheorderof3.4mas(Harper et al.2008). Thiseffect theperiodsearchandtheestimationsofthecircumstel-
in Gaia is investigated by the Coordination Unit 4 in lar or interstellar extinction as well as the bolometric
the Gaia Consortium. correction. However, Gaia will calibrate many stan-
dard candles and these will contribute to many differ-
4.3 Eclipsing binaries ent topics. In this article we touched on a few: testing
the universality of standard candles, deriving statisti-
Eclipsingbinariesaretraditionallysub-classifiedasEA, cal properties of different types of standard candles,
EB or EW types. EA, or Algol (β-Persei), types have searching for new standard candles, constraining stel-
the eclipses well defined in their light curves, with the lar evolution through standard candles, improving the
possibility to identify the times of their beginning and knowledge of the formation history of the Galaxy and
their end in the folded light curve. EB, or β-Lyrae, extending Gaia’s results to investigate Galactic struc-
types display a continuous variation of the light curve ture. Some of the problems might be used to our ad-
over an orbital cycle, preventing the identification of vantageandsomestandardcandlescouldbe utilizedto
the eclipsetimes. EW,orWUrsaeMajoris,typeshave crosscheck Gaia’s astrometry or even to establish ex-
similar depths of the primary and secondary eclipses. tinction maps (Windmark et al. 2011).
Pojmanski(2002)suggestedaclassificationbasedon
thephysicalpropertyofthebinary,i.e. detached,semi-
detached or contact binary. These three categories can
theoretically be identified from the a2 and a4 parame-
ters of the Fourier decomposition of the light curves in
cosine series (Rucinski 1993, Pojmanski 2002).
The automated variability processing pipeline for
Gaia will also characterize the geometry of the folded
light curves of eclipsing binaries and estimate the du-
ration of the eclipses and their depths and phases. A
study is underway to explore the orbital parameters
that can be estimated based solely on the geometrical
characterization of the folded light curves in order to
9
References Spano,M.,Mowlavi,N.,Eyer,L.,etal.2011,arXiv:1109.6132
vanLeeuwen, F. 2009, Astron. Astrophys.,497, 209
Berdnikov, L. N., Kniazev, A. Y., Kravtsov, V. V., Pas- Windmark, F., Lindegren, L., & Hobbs, D. 2011, Astron.
tukhova,E.N.,&Turner,D.G.2009,AstronomyLetters, Astrophys., 530, A76
35, 39 Wood, P. R., Alcock, C., Allsman, R. A., et al. 1999,
Clementini, G., Gratton, R., Bragaglia, A., et al. 2003, As- Asymptotic Giant Branch Stars, 191, 151
tron. J., 125, 1309 Wray, J. J., Eyer, L., & Paczyn´ski, B. 2004, Mon. Not. R.
Clementini, G. 2011, EAS Publications Series, 45, 267 Astron. Soc., 349, 1059
Dischler, J., Soederhjelm, S. 2005, The Three-Dimensional Wyrzykowski,L.,Udalski,A.,Kubiak,M.,etal.2004,Acta
Universewith Gaia, 576, 569 Astron., 54, 1
Dias, W. S.,Alessi, B. S.,Moitinho, A.,& L´epine, J.R.D. Zwahlen,N.,North,P.,Debernardi,Y.,etal.2004,Astron.
2002, Astron.Astrophys., 389, 871 Astrophys., 425, L45
Dubath, P., Rimoldini, L., Su¨veges, M., et al. 2011, Mon. Zwitter, T. 2002, Exotic Stars as Challenges to Evolution,
Not. R.Astron. Soc., 414, 2602 eds. C.A. Tout & W.van Hamme, ASPC 279, 31
ESA,1997,TheHipparcosandTychoCatalogues, ESASP-
1200
Eyer,L.,&Cuypers,J.2000,IAUColloq.176: TheImpact
of Large-Scale Surveys on Pulsating Star Research, 203,
71
Eyer, L., & Mignard, F. 2005, Mon. Not. R. Astron. Soc.,
361, 1136
Eyer, L., Mowlavi, N., Varadi, M., et al. 2009, SF2A-2009:
Proceedings oftheAnnualmeetingoftheFrenchSociety
of Astronomy and Astrophysics, 45
Fernie, J. D., Evans, N. R., Beattie, B., & Seager, S. 1995,
Information Bulletin on Variable Stars, 4148, 1
Feast, M. W., Glass, I. S., Whitelock, P. A., & Catchpole,
R.M. 1989, Mon. Not.R. Astron. Soc., 241, 375
Feast, M. W., & Catchpole, R. M. 1997, Mon. Not. R. As-
tron. Soc., 286, L1
Graczyk, D., Soszyn´ski, I., Poleski, R., et al. 2011, Acta
Astron., 61, 103
Harper,G.M.,Brown,A.,&Guinan,E.F.2008,Astron.J.,
135, 1430
Klagyivik,P.,&Szabados,L.2009,Astron.Astrophys.,504,
959
Lah, P., Kiss, L. L., & Bedding, T. R. 2005, Mon. Not. R.
Astron. Soc., 359, L42
Leccia, S., Ripepi, V., Clementini, G., et al. 2011, in: The
Fundamental Cosmic Distance Scale: State of the Art
and the Gaia Perspective, on line proceedings, available
on: http://www.na.astro.it/ESFdistance/
McNamara, D. 1997, Publ.Astron. Soc. Pac., 109, 1221
Matsunaga, N., Kawadu, T., Nishiyama, S., et al. 2009,
Mon. Not. R. Astron.Soc., 399, 1709
Paczyn´ski, B. 1997, The Extragalactic Distance Scale,
edited by M. Livio (Cambridge University Press) 273
Pojmanski, G. 2002, Acta Astron., 52, 397
Rucinski, S.M. 1993, Publ. Astron. Soc. Pac., 105, 1433
Soszyn´ski, I., Poleski, R., Udalski, A., et al. 2008, Acta
Astron., 58, 163
Soszyn´ski, I., Udalski, A., Szyman´ski, M. K., et al. 2009,
Acta Astron., 59, 1
Soszyn´ski, I., Udalski, A., Szyman´ski, M. K., et al. 2009,
Acta Astron., 59, 239
Soszyn´ski, I., Poleski, R., Udalski, A., et al. 2010, Acta
Astron., 60, 17
Soszyn´ski, I., Udalski, A., Szyman´ski, M. K., et al. 2010,
Acta Astron., 60, 165
Soszyn´ski, I.,Dziembowski, W.A.,Udalski,A.,etal.2011,
Acta Astron., 61, 1 ThismanuscriptwaspreparedwiththeAASLATEXmacrosv5.2.