An extensive library of synthetic spectra covering the far red, RAVE and GAIA wavelength ranges

0.77 MB·English

by  Tomaz Zwitter

Astronomy&Astrophysicsmanuscriptno.synthe˙asaccepted February2,2008
(DOI:willbeinsertedbyhandlater)
An extensive library of synthetic spectra covering the far red,
RAVE and GAIA wavelength ranges
Tomazˇ Zwitter1,FiorellaCastelli2,3,andUlisseMunari4,5
1 UniversityofLjubljana,DepartmentofPhysics,Jadranska19,1000Ljubljana,Slovenia
4 2 CNR-IstitutodiAstrofisicaSpazialeeFisicaCosmica,ViadelFossodelCavaliere,00133,Roma,Italy
0
3 INAF-OsservatorioAstronomicodiTrieste,ViaG.B.Tiepolo11,34131,Trieste,Italy
0
4 OsservatorioAstronomicodiPadova,SedediAsiago,36012Asiago(VI),Italy
2
5 CISAS,CentroInterdipartimentaleStudiedAttivita`Spazialidell’Universita`diPadova,Italy

n
a
Receiveddate..............;accepteddate................
J

0
Abstract. A library of 183588 synthetic spectra based on Kurucz’s ATLAS9 models is presented for the far red spectral
2
interval (7653 – 8747 Å). It is characterized by 3500 K ≤ T ≤ 47500 K, 0.0 ≤ logg ≤ 5.0, −3.0 ≤ [M/H] ≤ +0.5,
eff
2 0 ≤ Vrot ≤ 500kms−1,ξ = 2kms−1.Thewholegridofspectraiscalculatedforresolvingpowers8500,11500and20000.
v Asectionofthegridisalsocomputedfor[α/Fe]=+0.4andformicroturbulentvelocities0and4kms−1.Thelibrarycovers
9 thewavelength rangesand resolutions ofthetwoambitious spectroscopic surveys bytheground experiment RAVEandthe
6 spacemissionGAIA.Cross-sectionsacrossthemulti-dimensionaldata-cubeareusedtoillustratethebehaviourofthestrongest
0 spectrallines.Interpretationofrealdatawillhavetoincludeinterpolationtogridsubsteps.Wepresentasimpleestimateofthe
1 accuracyofsuchaprocedure.
0
4
Keywords.Astronomicaldatabases:spectroscopic–Stars:fundamentalparameters–Surveys:GAIA–Surveys:RAVE
0
/
h
p 1. Introduction disentangle their age and metallicity. Composite systems in-
-
o cludenormalellipticalgalaxies(Molla&Garcia-Vargas2000;
Observationsinthefarredinterval(7650–8750Å)arebecom-
r Saglia et al. 2002; Cenarro et al. 2003), dwarf ellipticals
t ingincreasinglyusedfordeterminationofbasicstellarparam-
s (Michielsen, De Rijcke & Dejonghe 2003) and active galax-
a eters. The advantagesof this intervalare its low sensitivity to
ies (van Groningen 1993; Diaz, Terlevich & Terlevich 1989;
:
v reddening,highphotonbudgetforlate–typestars,goodsensi- Nelson & Whittle 1999; Schinnerer, Eckart & Tacconi 2001;
i tivity of modern CCD detectors, and above all richness of its
X Marquez et al. 2003). Cenarro et al. (2002) studied the be-
spectralfeatures.
haviourofCaIIline-strengthindicesintermsofeffectivetem-
r
a The far red interval includes several important spectral perature, surface gravity, and metallicity. Their analysis was
lines (Munari 1999): (a) the K I infrared resonant doublet
basedonasetof706realspectra.Amuchlargersyntheticspec-
(7664.907 Å, 7698.979 Å), (b) the Na I non-resonant dou-
trallibrarypresentedinthispapercanbeusedtocomplement
blet (8183.256Å, 8194.821Å), (c) the O I triplet (8446.247,
theseresultsonCaIIandotherlinesinthereddomain.
8446.359,8446.758Å),(d)theFeImultiplets60,401andoth-
erswithexcitationpotentialsbetween2.17and5.00eV,(e)the Two of the largest forthcoming spectral surveys are cen-
headofthePaschenseriesofhydrogen,and,mostimportantly, teredonthefarredspectralinterval.GAIAisthe cornerstone
(f )theCaIInon-resonantinfraredtriplet(8498.018,8542,089, 6 mission of ESA, approved for a launch around 2010. It is
8662.140Å).Thelatterispresentinallspectraltypesbetween aimedatprovidingmicro-arsecastrometryand∼10-bandpho-
B8 and M. A furtheradvantageis that the intervalof 8400to tometryfor∼ 109 starsbrighterthanV = 20.Brightertargets
8750Åisnearlyfreefromtelluricabsorptions(Munari1999). (V <17.5)willalsobeobservedbyanon-boardspectroscopic
TheCaIItripletisapowerfuldiagnostictool:thestrength instrumentoperatinginthe8480–8747Åwavelengthrangeat
of Ca II lines depends on metallicity, but it is essentially in- aresolvingpowerR ≡ λ/∆λ = 11500(Katz2003).Theother
sensitiveto ageofstellar population(Garcia-Vargas,Molla& survey,RadialVelocityExperiment(RAVE),hasjuststartedin
Bressan 1998; Schiavon, Barbuy & Bruzual 2000; Vazdekis April (Steinmetz 2003). This is an internationalcollaboration
et al. 2003).These results have been used to study behaviour whichusestheUK-SchmidttelescopeattheAngloAustralian
of Ca II in composite stellar populations and to attempt to Observatory(AAO)equippedwithafiber-opticspectrographto
obtainspectrainasimilarwavelengthdomain(8410–8750Å)
Sendoffprintrequeststo:T.Zwitter and at a resolving power R ∼ 8500. The goal is to observe
2 TomazˇZwitteretal.:Anextensivelibraryofsyntheticspectracoveringthefarred,RAVEandGAIAwavelengthranges
Fig.1. Graph of calculated spectra without alpha enhancement and with a micro-turbulent velocity of 2 km s−1. Metallicity is coded with
differentsymbolswhichareplottedwithsmallhorizontaloffsetsforclarity.Eachspectrumiscalculatedfor11(T < 7000K)or14(T ≥
eff eff
7000K)differentrotationalvelocitiesandatthreedifferentresolvingpowers(seetext).
35 million stars brighter than V = 16 at declinationssuitable withknownvaluesofphysicalparameterscanyieldmuchmore
forAAO.Theprimarymotivationforbothsurveysistoobtain than radial velocity: in fact the effective temperature, gravity,
stellar radialvelocities to be used in studies of Galactic kine- metallicity(withcertainelementabundances),rotationalveloc-
matics.Butacross-correlationwithalibraryofstellarspectra ity,andthepresenceofdifferentkindsofpeculiaritiescanallbe
TomazˇZwitteretal.:Anextensivelibraryofsyntheticspectracoveringthefarred,RAVEandGAIAwavelengthranges 3
determined.ThereforeanysoftwaredevelopmentfortheGAIA Table1.Propertiesofmodelsatdifferentmetallicities.Thefirstthree
missionaswellasanalysisofthefirstRAVEdataneedstouse columns give values of metallicity, α-enhancement coefficient and
anextensivelibraryofstellarspectrainthefarreddomain.The microturbulent velocity (ξ). The last two columns state if the ”no-
same holds for individual spectroscopic studies in this wave- overshooting” approximation totheconvectiontreatmentandifnew
opacitydistributionfunctionswereused.
length range which are becomingever more numerousowing
tothepreparationoftheGAIAmission(seee.g.Munarietal.
[M/H] [α/Fe] ξ[kms−1] NOVER ODFNEW
2001,Zwitteretal.2003,Marreseetal.2004).
+0.5 0.0 2 yes no
+0.2 0.0 2 no no
0.0 0.0 0 yes no
0.0 0.0 2 yes no
0.0 0.0 4 yes no
–0.2 0.0 2 no no
–0.5 0.0 2 yes no
–0.5 0.4 2 yes yes
–1.0 0.0 2 yes no
–1.0 0.4 2 yes yes
–1.5 0.0 2 yes no
–1.5 0.4 2 yes yes
–2.0 0.0 2 yes no
–2.0 0.4 2 yes no
–2.5 0.0 2 yes no
–3.0 0.0 2 no no
–4.0 0.0 2 no no
Fig.2. As Fig. 1, but for models with alpha-enhancement ([α/Fe]= in a uniformmanner.So a paralleleffortwaslaunchedtocal-
+0.4).Allmodelswerecalculatedforthemetallicities[M/H]=–0.5, culatea gridofsyntheticspectrausingATLAS9modelsfrom
–1.0,–1.5,and–2.0. Kurucz.Thefirsttwopapers(Munari&Castelli2000,Castelli
& Munari2001)exploredthe gridin the temperature-gravity-
metallicity space for non-rotating stars, assuming a micro-
turbulent velocity of 2 km s−1. Altogether 952 spectra were
presented.
These early workswere aimed to coarsely butrapidly ex-
ploreanessentiallyunknownwavelengthrangetoassist early
planningand instrumentdesign for GAIA. This phase is now
over,andthecommunityrequirementsarenowmovingtoward
data and analysisreductionpipelines.A muchmoreextended
and complete grid is now essential, and to provide one is the
aim of this paper. The grid made available with this paper
shouldmeetthecommunityrequirementsin testingalgorithm
coding for some time to come. The one to be used to ana-
lyzetheactualGAIAdatawhentheywillbefinallyassembled
around 2015–2018will be computed only in the next decade
takingfulladvantagesofthecontinuousadvacementsinthein-
putphysics,codingand atomic constantsthat will be reached
by that time. The present grid is also timely presented to as-
sist with analysis of RAVE spectra that are already routinely
obtainedatAAO.
Fig.3.AsFig.1,butformodelswithmicro-turbulentvelocitiesof0 The present grid extends earlier calculations by adding
and4kms−1.Allmodelshavesolarabundances. more spectra, inclusion of stellar rotation and presentation
of results at different spectral resolving powers. Furthermore
Systematic observationsof MKK standard stars in the far a limited number of spectra corresponding to enhanced α-
red spectral domain were presented by Munari & Tomasella element abundances and different values of micro-turbulent
(1999) and by Marrese, Boschi & Munari (2003). It is dif- velocity are presented. The database consists of more than
ficult however to obtain high–signal–to–noise–ratio observa- 183500 spectra and is freely available in electronic form via
tions to map all relevant combinations of stellar parameters CDSaswellasviatheESAwebserver.
4 TomazˇZwitteretal.:Anextensivelibraryofsyntheticspectracoveringthefarred,RAVEandGAIAwavelengthranges
2. Computationofsyntheticspectra gravity–temperature plane except for hot low-gravity mod-
els which are notradiatively stable. Low–temperaturespectra
We computed synthetic spectra for the 7650-8750 A inter-
(T < 5000K) werecomputedfora sparser setof metallici-
eff
val for almost all ATLAS9 grids of model atmospheres with
tiesduetolargerequirementsofcomputingtime.Thesespectra
different metallicities and micro-turbulentvelocities available
willbeaddedonlinewhencompleted.Fig.2correspondstoα–
at the Kurucz web-site (http://kurucz.harvard.edu).We gener-
enhanced cases and Fig. 3 to those with a different value of
ally used the NOVER models which differ from the previous
microturbulentvelocity.Notethateachofthesymbolsactually
Kurucz (1993a) models for the convection in that they were
correspondsto11(T < 7000K)or14(T ≥ 7000K)spec-
eff eff
recomputed by Castelli with the overshooting option for the
trawithdifferentvaluesofrotationalvelocity(seeTable1)and
convectionswitched off (Castelli et al. 1997). When NOVER
tothreedifferentresolvingpowers.
grids were not available, we used the Kurucz (1993a) mod-
Allspectraareavailableasasciifilesgroupedintodifferent
els. Because new ATLAS9 models based on updated Opacity
directoriesaccordingtotheirresolvingpowerandtemperature.
DistributionFunctions(ODFNEWmodels)havebeenrecently
ThefilenamesareinastandardformatidentifiedinTable3.So
computed (Castelli & Kurucz 2003), we used them for a few
f765875v010r20000m05t05250g45k2nover.asccorrespondsto
metallicities with T ≥ 5000 K and for all the spectra with
eff a flux calibrated spectrum between 7650 and 8750 Å, with
Teff ≤4750K.Infact,themaindifferencebetweentheNOVER V = 10 km s−1, λ/∆λ = 20000, [M/H] = −0.5, T =
rot eff
models and the ODFNEW models is the use of new Opacity 5250K,logg=4.5,ξ =2kms−1,andnoα–enhancement.
DistributionFunctions(ODFs)computedwithTiO linesfrom
The calculated grid is by far too large to present all of
Schwenke(1998)insteadoffromKurucz,andincludingH O
2 its properties here, so we explore only sample cross-sections
lines, not considered at all in the previous ODFs. Also the
across the grid. Figure 4 is a greyscale presentation of the
ODFNEW models computed up to now are available at the
spectrawhichwerenormalizedtoenhancelinevisibility.Each
Kurucz web-site (ODFNEW grids). Details of the adopted
panelshowsvariationalongoneparameteraxis, startingfrom
modelsaregiveninTable1.
aspectrumofanon-rotatingK0Vtypestar.Notethatallspec-
The synthetic spectra were computed with the SYNTHE
trawerecalculatedinawiderwavelengthdomain,butonlythe
code of Kurucz (1993b) at a resolving power of 500000.
8400–8750Årangeisplottedforclarity.
The source of atomic data was Kurucz & Bell (1995), the
The temperature panel of Fig. 4 clearly shows the impor-
source of molecular data, except TiO, was Kurucz (1993c),
tance of sharp Ca II lines for any radial velocity study. The
while the source of TiO data was Kurucz (1999a), who sup-
panel is a textbookexample of the expected behaviourof the
plied Schwenke’s (1998) computations in SYNTHE format.
Paschen lines and metallic lines. The metallicity panel illus-
Di-atomicmoleculesusedinmodelcomputationarediscussed
trates that the Ca II lines remain strong even at the lowest
in Kurucz (1992) and Kurucz (1993d). The source for H O
2 metallicities and the gravity panelshows their presence in all
dataisPartridge&Schwenke(1997),asdistributedbyKurucz
luminosityclasses.Therotationalvelocityandresolvingpower
(1999b). The solar abundances are from Anders & Grevesse
panelsshowhowthelinesgetsmearedathighrotationalveloc-
(1989), except for the spectra based on the ODFNEW mod-
itiesorifobservingatlowresolvingpowers.
els. In this case the solar abundances are from Grevesse &
The steps in the calculated grid are relatively small, but
Sauval(1998).Eachspectrumcomputedforagivenmodelat-
the coverage is not continuous. As an example, the step in
mospherewas then broadenedfor severalvalues of rotational temperature is 250 K (for T ≤ 10000 K). This is larger
eff
velocityV andforthreeresolvingpowers:R=8500,11500,
rot than the baselined accuracy of temperature determination for
and20000byassumingaGaussianinstrumentalprofile.These
bothGAIAandRAVEsurveys.Sothegridwillhavetobein-
resolutionswere chosenbecause they correspondto the base-
terpolated to smaller steps. Figure 5 illustrates the errors in-
lined values for the RAVE survey, the GAIA mission and a
troduced by a simple linear interpolation. At a certain grid
typicalCassegrain-fedEchellespectrograph,respectively.The
point i with the parameter values p we compare the true
i
resultingfinalsyntheticspectrawereresampledto2pixelsper synthetic spectrum S(p) with the spectrum S′ obtained from
i
resolutionelement(R=8500,11500)or2.5pixelsperresolu-
a linear combination of spectra at neighbouring grid points:
tionelement(R=20000). S′ = fi−1S(pi−1)+ fi+1S(pi+1). The weights fi−1 and fi+1 are
optimizedsothatR[S(pi)−S′]2dλisminimal.Thedifference
between the interpolatedvalues of parameters p′ and the true
3. Gridofsyntheticspectra
ones p canthenbeexpressedinunitsofagridstep:
i
Rofansgyensthaentidc sstpeepcstrfaoraraellgsiveevneninbaTsaicblepa2ra.mWeeterasdoopftthaecgormid- ∆≡ p′−pi = fi+1(pi+1−pi)+ fi−1(pi−1−pi) (1)
monconventionofquotingmetallicityandenhancementofα– pi+1−pi−1 (fi+1+ fi−1)(pi+1−pi−1)
elementsin logarithmicunits with respectto the solar values. Figure5showsthatlinearinterpolationisaccurateto ∼< 10%
Thegravityisinlogarithmiccgsunits.Detailsofallcalculated ofthe gridstep.Note thatthisisthe worstcase scenario,cor-
parametercombinationsare given in Figures1–3.Spectra are respondingtoareconstructionofthespectrumatthemiddleof
placed in gravity–temperature planes, with metallicity coded thegridinterval.Linearinterpolationwouldbemoreaccurate
byasymboltype.Figure1coversthemostnumerousspectra, for spectra lying closer to one of the grid points. The results
i.e. the oneswith no α–enhancementand with microturbulent couldbeimprovedfurtherbyemployingnon-linearinterpola-
velocity of 2 km s−1. The computed spectra cover the whole tionschemes.Onemayconcludethatlinearinterpolationitself
TomazˇZwitteretal.:Anextensivelibraryofsyntheticspectracoveringthefarred,RAVEandGAIAwavelengthranges 5
Fig.4. Cross-sections through the multidimensional [Temperature, Metallicity, Gravity (cm s−2), Rotation, Resolving power] data-cube of
computedspectra.Parametervaluescorrespondtoanon-rotatingK0Vtypestar(T = 5250K,[M/H] = 0.0,logg = 4.5,V = 0kms−1,
eff rot
ξ =2kms−1,λ/∆λ=20000),exceptfortheparameterwhichisallowedtovaryonaparticularplot.Allspectrawerenormalizedbyasingle
cubicsplinefittotheirupperenvelope.Thenumberofgreyscalehues(identifiedatthetopofthefigure)isintentionallykeptsmall,soasto
allowreadingthedepthofindividualspectrallinesfromthegraph.Ionsgivingrisetothemostprominentspectrallinesareidentifiedatthe
top.Syntheticspectraspanawiderwavelengthrangeandincludevariationofadditionalparameters(α-enhancement,differentmicro-turbulent
velocities)notshownhereforclarity.
does not introduce errors exceeding 25 K in temperature (for ratio complicates their determination (Bailer-Jones 2003, see
T < 10000K),0.05dexin [M/H]orlogg and1 kms−1 in also Fig. 1in Zwitter2002).Also,spectra ofrealstars donot
eff
V . Note thatothererrorsaremoreimportant:degeneracyof correspond exactly to the synthetic spectra due to their pecu-
rot
parameter values fitting spectra with a limited signal to noise liarities (e.g. emission lines, varied abundances of individual
6 TomazˇZwitteretal.:Anextensivelibraryofsyntheticspectracoveringthefarred,RAVEandGAIAwavelengthranges
Table2.Generalrangesofparametersofthespectra.Notallcombinationswerecalculated(fordetailsseeFigs.1–3).
Parameter Range SteporValues
Temperature 3500...47500K step250K(T ≤10000K),500or1000K(T >10000K)
eff eff
Metallicity +0.5...–3.0 step0.5&valuesof+0.2and–0.2
Gravity 0.0...5.0 step0.5
α–enhancement 0.0...0.4 (0.0,0.4)
Micro-turbulence 0...4kms−1 (0,2,4kms−1)
Rotationvelocity 0...100kms−1(T <7000K) (0,2,5,10,15,20,30,40,50,75,100kms−1)
eff
0...500kms−1(T ≥7000K) (0,10,20,30,40,50,75,100,150,200,250,300,400,500kms−1)
eff
Resolvingpower 8500...20000 (8500,11500,20000)
Wavelength 7653...8747Å 2pixelsperresolutionelement(2.5pixelsatR=20000)
elements, non-LTE effects, and non-static atmospheric struc- Table3.Filenamesofindividual spectra.Meaningofcorresponding
ture). charactersisgiven.
character meaning
1 f:fluxedspectrum(ergs−1cm−2Åsteradian−1)
n:normalizedtothecontinuumflux
2–4 startingwavelength(nm)
5–7 endingwavelength(nm)
9–11 rotationalvelocityV (kms−1)
rot
13–17 resolvingpowerR
18 p:[M/H]≥0.0
m:[M/H]<0.0
19–20 10×ABS([M/H])
21 t:noα–enhancement
a:[α/Fe]=0.4
22–26 effectivetemperatureT (K)
eff
28–29 10×logg
31 micro-turbulentvelocityξ(kms−1)
4. DiscussionandConclusions
Thegridof syntheticspectrain the 7653–8747Åintervalpre-
sented here is the most extensive so far. Altogether the grid
contains 183588 ascii files. They span three spectral resolu-
tions correspondingto a typical Echelle spectrographand the
spectrographs of the RAVE and GAIA surveys. This should
facilitate comparisonof the results obtainedwith differentin-
struments. Spectra with other resolving powers (R < 20000)
canbeeasilycomputedfromthegrid.
The synthetic spectra we have computed can be used as
templatesforthe determinationofradialvelocity,theprimary
goaloftheRAVEandGAIAsurveys.Theyalsopermittheuser
to derive the primary parametersof stellar atmospheres: tem-
perature, metallicity, gravity and rotational velocity. The grid
includessubsetsofspectracomputedfor+0.4enhancedabun-
dances of the α elements and for different values of the mi-
Fig.5. Errors introduced by linear interpolation of the grid. Each
graph explores variation of one parameter, with the values of other croturbulentvelocity.Itwas shownthatthegridcan beeasily
parameters being fixed at: T = 5250 K, [M/H]=0.0, logg = 4.5, interpolatedtoapproximatespectrawithparametervaluesbe-
eff
V =0kms−1,R=20000.Theordinateisthedifferencebetweenthe tweenthestepsinthegrid.Theerrorsofparameterscomputed
rot
value obtained from linear grid interpolation of the given parameter with a simple linear interpolation do not exceed 10% of the
andthetruevalue,expressedinunitsofthegridstep.Notethatlinear gridstep.Thecomputedspectradependontheadoptedvalues
interpolationyieldsvaluesaccurateto ∼< 10%ofthegridstep. ofindividualelementabundances.Inparticularthesolarspec-
trum computedwith Anders & Grevesse’s (1989)solar abun-
dancesforalltheelements(Fig.6)featurestoostrongOIand
TomazˇZwitteretal.:Anextensivelibraryofsyntheticspectracoveringthefarred,RAVEandGAIAwavelengthranges 7
Fig.6.Comparisonofcomputedandobservedspectra:(a)observedsolarfluxatlasresampledtoR=20000(Kuruczetal.1984);(b)computed
spectrum(T =5750K,[M/H]=0.0,logg= 4.5,V =2kms−1,ξ =2kms−1,R=20000)(c)sameas(b)butfor[M/H]=-0.5.Allspectra
eff rot
areplottedinnormalizedflux.Thetopmosttwospectrahavebeenverticallyoffsetforclarity.
Fe I lines. This is consistent with a recently suggested lower edges financial support from the Slovenian Ministry for Education,
oxygensolarabundancelog(N(O)/N(H))=−3.31dex(Prieto, Science and Sports, CNRS and the Royal Society as well as warm
Lambert & Asplund 2001) and with a lower Fe I abundance hospitality of GAIA groups at the Observatoire de Meudon and the
log(N(Fe)/N(H)) = −4.5dex(BellotRubio& Borrero2002) MullardSpaceSciencesLaboratorywherepartofthisworkhasbeen
completed. U.M. acknowledges financial support from the Italian
as obtained from two and three dimensional hydrodynamical
Space Agency contract ASI-I-R-117-01 and the Italian Ministry of
modelatmospheres.Thiscomplementsclassicalmethodsofdi-
EducationCOFIN2001grant. WethankR.Sordoforassistingwith
agnosticlineratiosforthefarredspectralintervaldiscussedby
partsofthegridcomputation.
Munari(2002).
Properties of Kurucz ATLAS model atmospheres in the
contextoftheredspectralintervalandtheGAIAmissionhave
alreadybeenreviewedbyNesvaciletal.(2003).Plez(2003)es- References
timatedthepotentialoftheMARCSmodels.Hauschildtetal.
AndersE.,GrevesseN.1989,Geochim.Cosmochim.Acta,53,197
(2003)presentedthecomputationsofnon-stationaryandnon- Bailer-Jones C.A.L. 2003, in GAIA Spectrsocopy, Science and
sphericallysymmetricatmosphereswithsignificantamountsof Technology(U.Munari,ed.),ASPConf.Ser.,298,199
dust.FinallyThe´venin,Bijaoui&Katz(2003)reviewedthede- BellotRubioL.R.,BorreroJ.M.2002,A&A,391,331
termination of chemicalabundancesfrom GAIA spectra. The CastelliF.,GrattonR.G.,KuruczR.L.1997,A&A,318,841
commonconclusionis thatthe observationalpotentialsof the CastelliF.,MunariU.2001,A&A,366,1003
CastelliF.,KuruczR.L.2003, inModellingof StellarAtmospheres,
GAIA mission ask for improvements in the computation of
IAUSymp.210(N.E.Piskunovetal.,eds.),inpress
synthetic stellar spectra. The RAVE collaboration is starting
CenarroA.J.,GorgasJ.,CardielN.,VazdekisA.,PeletierR.F.2002,
to obtain a very large number of high quality spectra in the
MNRAS,329,863
far red spectral interval. They guarantee that this wavelength
CenarroA.J.,GorgasJ.,VazdekisA.,CardielN.,PeletierR.F.2003,
rangeisrapidlyrisinginimportance.Wearecurrentlyextend-
MNRAS,339,L12
ing the grid presented here to a larger wavelength interval. A DiazA.I.,TerlevichE.,TerlevichR.1989,inActiveGalacticNuclei
continuouscoveragefrom250nmto1050nmshouldfacilitate (D.E.Osterbrock&MillerJ.S.,eds.),560
acomparisonofspectraobtainedinotherwavelengthdomains Garcia-VargasM.L.,Molla´M.,BressanA.1998,A&AS,130,513
to hugeobservationalsets comingfromthe RAVE andGAIA GrevesseN.,SauvalA.J.1998,SpaceSci.Rev.,85,161
surveys. vanGroningenE.1993,A&A,272,25
Hauschildt P.H., Allard F., Baron E., Aufdenberg J., Schweitzer
Acknowledgements. We thank the anonymous referee for very A. 2003, in GAIA Spectrsocopy, Science and Technology (U.
relevant comments on the first version of the paper. T.Z. acknowl- Munari,ed.),ASPConf.Ser.,298,179
8 TomazˇZwitteretal.:Anextensivelibraryofsyntheticspectracoveringthefarred,RAVEandGAIAwavelengthranges
Katz D. 2003, in GAIA Spectrsocopy, Science and Technology (U.
Munari,ed.),ASPConf.Ser.,298,65
KuruczR.L.,FurenlidI.BraultJ.,TestermanL.,1984.NationalSolar
Observatory,Sunspot,NewMexico,SolarFluxAtlasfrom296to
1300nm
KuruczR.L.1992,RvMexAA,23,45
KuruczR.L.1993a,ATLAS9StellarAtmospheresProgramsand2km
s−1grid,CD-ROMNo.13
KuruczR.L.1993b,SYNTHESpectrumSynthesisProgramsandLine
Data,CD-ROMNo.18
KuruczR.L.1993c,Diatomicmoleculardataforopacitycalculations,
CD-ROMNo.15
Kurucz R.L. 1993d, in Molecules in the Stellar Environment, IAU
Coll.146(ed.U.G.Jorgensen),282
KuruczR.L.,BellB.1995,AtomicLineList,CD-ROM,No.23
KuruczR.L.1999a,TiOLineList,CD-ROMNo.24
Kurucz R.L. 1999b, H2O Linelist from Partridge and Schwenke
(1997),noIDs,CD-ROMNo.26
MarreseP.M.,BoschiF.,MunariU.2003,A&A,406,995
MarreseP.M.,MunariU.,SivieroA.,MiloneE.F.,ZwitterT.,Tomov
T.,BoschiF.,BoecheC.2004,A&A,413,635
Marquez I., Masegosa J., Durret F.,Gonzalez Delgado R.M.,Moles
M.,MazaJ.,PerezE.,RothM.2003,A&A,409,459
MichielsenD.,DeRijckeS.,DejongheH.2003,ApJ,597,L21
Molla´M.,Garcia-VargasM.L.2000,A&A,359,18
MunariU.1999,Balt.A.8,73
MunariU.,TomasellaL.1999,137,521
MunariU.,CastelliF.2000,A&AS,141,141
Munari U., Tomov T., Zwitter T., Milone E.F., Kallrath J., Marrese
P.M.,BoschiF.,PrsˇaA.,TomasellaL.,MoroD.2001,A&A,378,
477
MunariU.2002,inGAIA:AEuropeanSpaceProject(O.Bienayme
andC.Turon,eds.),EASPubl.Ser.,2,39
NelsonC.,WhittleM.1999,Adv.SpaceRes.,23,891
Nesvacil N., Stu¨tz C., Weiss W.W. 2003, in GAIA Spectrsocopy,
Science and Technology (U. Munari, ed.), ASPConf. Ser.,298,
173
PartridgeH.,SchwenkeD.W.1997,J.Chem.Phys.,106,4619
Plez B. 2003, in GAIA Spectrsocopy, Science and Technology (U.
Munari,ed.),ASPConf.Ser.,298,189
PrietoC.A.,LambertD.L.,AsplundM.2001,ApJ,556,L63
Saglia R.P., Maraston C., Thomas D., Bender R., Colless M. 2002,
ApJ,579,L13
SchiavonR.P.,BarbuyB.,BruzualG.2000,ApJ,532,453
SchinnererE.,EckartA.,TacconiL.J.2001,ApJ,549,254
Schwenke D.W. 1998, Chemistry and Physics of Molecules and
Grains in Space. Faraday Discussions No. 109. The Faraday
DivisionoftheRoyalSoc.ofChem.,London,.321
SteinmetzM.2003,inGAIASpectrsocopy,ScienceandTechnology
(U.Munari,ed.),ASPConf.Ser.,298,381
The´veninF.,BijaouiA.,KatzD.2003,inGAIASpectrsocopy,Science
andTechnology(U.Munari,ed.),ASPConf.Ser.,298,291
VazdekisA.,CenarroA.J.,GorgasJ.,CardielN.,PeletierR.F.2003,
MNRAS340,1317
ZwitterT.,2002,A&A,386,748
ZwitterT.,MunariU.,MarreseP.M.,PrsˇaA.,MiloneE.F.,BoschiF.,
TomovT.,SivieroA.2003,A&A,404,333