The Journal of the American Association J AAVSO Volume 40 Number 2 2012 of Variable Star Observers 49 Bay State Road Cambridge, MA 02138 U. S. A. ε Aurigae Special Edition Alsointhisissue... • BVRI photometry of SN 2011fe in M101 • The 1909 outburst of RT Ser • Photometry and spectroscopy of P Cygni • A W UMa system with complete eclipses • MP Gem—an EB with a very long period? Complete table of contents inside... Historic first: 1.6-m wavelength image of ε Aur, 2009 Nov. 2, as initially processed by John Monnier, based on four telescope beam combination data acquired by MIRC at the CHARA Array and showing the shadow of the disk crossing the face of ε Aur. see page 618 Image courtesy of John Monnier, Univ. Michigan
524
Embed
ε Aurigae Special Editionsouthscienceandgerman.weebly.com/uploads/8/4/6/9/8469541/jaavso... · GEOS RR Lyrae Survey: Blazhko Period Measurement of Three RRab Stars ... Pierre de
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
The Journal of the American Association
JAAVSO Volume 40Number 2
2012
of Variable Star Observers
49 Bay State RoadCambridge, MA 02138
U. S. A.
ε Aurigae Special Edition
Also in this issue...• BVRI photometry of SN 2011fe in M101
• The 1909 outburst of RT Ser
• Photometry and spectroscopy of P Cygni
• A W UMa system with complete eclipses
• MP Gem—an EB with a very long period?Complete table of contents inside...
Historic first: 1.6-m wavelength image of ε Aur, 2009 Nov. 2, as initially processed by John Monnier, based on four telescope beam combination data acquired by MIRC at the CHARA Array and showing the shadow of the disk crossing the face of ε Aur. see page 618
Image courtesy of John Monnier, Univ. Michigan
Editorial BoardGeoffrey C. ClaytonLouisiana State UniversityBaton Rouge, Louisiana
Edward F. GuinanVillanova UniversityVillanova, Pennsylvania
Pamela KilmartinUniversity of CanterburyChristchurch, New Zealand
Laszlo KissKonkoly ObservatoryBudapest, Hungary
Paula SzkodyUniversity of WashingtonSeattle, Washington
The Journal of the American Association of Variable Star Observers
The Council of the American Association of Variable Star Observers2011–2012
Director Arne A. Henden President Mario E. Motta Past President Jaime R. García 1st Vice President Jennifer Sokoloski Secretary Gary Walker Treasurer Gary W. Billings (term ended May 2012) Treasurer Timothy Hager
Councilors
Editor John R. Percy University of TorontoToronto, Ontario, Canada
Associate EditorElizabeth O. Waagen
Assistant Editor Matthew R. Templeton
Production EditorMichael Saladyga
Matthew R. TempletonAAVSO
Douglas L. WelchMcMaster UniversityHamilton, Ontario, Canada
David B. WilliamsWhitestown, Indiana
Thomas R. WilliamsHouston, Texas
Lee Anne WillsonIowa State UniversityAmes, Iowa
ISSN 0271-9053
Edward F. GuinanRoger S. KolmanChryssa KouveliotouArlo U. Landolt
John MartinDonn R. StarkeyRobert J. StineDavid G. Turner
JAAVSOThe Journal of
The American Associationof Variable Star Observers
49 Bay State RoadCambridge, MA 02138
U. S. A.
Volume 40 Number 2
2012
ISSN 0271-9053
ε Aurigae Special Edition
The Journal of the American Association of Variable Star Observers is a refereed scientific journal published by the American Association of Variable Star Observers, 49 Bay State Road, Cambridge, Massachusetts 02138, USA. The Journal is made available to all AAVSO members and subscribers.
In order to speed the dissemination of scientific results, selected papers that have been refereed and accepted for publication in the Journal will be posted on the internet at the eJAAVSO website as soon as they have been typeset and edited. These electronic representations of the JAAVSO articles are automatically indexed and included in the NASA Astrophysics Data System (ADS). eJAAVSO papers may be referenced as J. Amer. Assoc. Var. Star Obs., in press, until they appear in the concatonated electronic issue of JAAVSO. The Journal cannot supply reprints of papers.
Page Charges
Unsolicited papers by non-Members will be assessed a charge of $15 per published page.
Instructions for Submissions
The Journal welcomes papers from all persons concerned with the study of variable stars and topics specifically related to variability. All manuscripts should be written in a style designed to provide clear expositions of the topic. Contributors are strongly encouraged to submit digitized text in ms word, latex+postscript, or plain-text format. Manuscripts may be mailed electronically to [email protected] or submitted by postal mail to JAAVSO, 49 Bay State Road, Cambridge, MA 02138, USA.
Manuscripts must be submitted according to the following guidelines, or they will be returned to the author for correction: Manuscripts must be: 1) original, unpublished material; 2) written in English; 3) accompanied by an abstract of no more than 100 words. 4) not more than 2,500–3,000 words in length (10–12 pages double-spaced).
Figures for publication must: 1) be camera-ready or in a high-contrast, high-resolution, standard digitized image format; 2) have all coordinates labeled with division marks on all four sides;
3) be accompanied by a caption that clearly explains all symbols and significance, so that the reader can understand the figure without reference to the text.
Maximum published figure space is 4.5” by 7”. When submitting original figures, be sure to allow for reduction in size by making all symbols and letters sufficiently large.
Photographs and halftone images will be considered for publication if they directly illustrate the text. Tables should be: 1) provided separate from the main body of the text; 2) numbered sequentially and referred to by Arabic number in the text, e.g., Table 1.
References: 1) References should relate directly to the text.
2) References should be keyed into the text with the author’s last name and the year of publication, e.g., (Smith 1974; Jones 1974) or Smith (1974) and Jones (1974).
3) In the case of three or more joint authors, the text reference should be written as follows: (Smith et al. 1976).
4) All references must be listed at the end of the text in alphabetical order by the author’s last name and the year of publication, according to the following format:
Brown, J., and Green, E. B. 1974, Astrophys. J., 200, 765. Thomas, K. 1982, Phys. Report, 33, 96. 5) Abbreviations used in references should be based on recent issues of the Journal or the listing provided
at the beginning of Astronomy and Astrophysics Abstracts (Springer-Verlag).
Miscellaneous:1) Equations should be written on a separate line and given a sequential Arabic number in parentheses
near the right-hand margin. Equations should be referred to in the text as, e.g., equation (1).2) Magnitude will be assumed to be visual unless otherwise specified.3) Manuscripts may be submitted to referees for review without obligation of publication.
Journal of the American Association of Variable Star ObserversVolume 40, Number 2, 2012
Table of Contents continued on following pages
e Aurigae Special Edition
Highlighting ε Aurigae and Citizen Sky John R. Percy 609
ε Aurigae Papers
The Origins and Future of the Citizen Sky Project Aaron Price, Rebecca Turner, Robert E. Stencel, Brian K. Kloppenborg, Arne A. Henden 614
ε Aurigae—an Overview of the 2009–2011 Eclipse Campaign Results Robert E. Stencel 618
The International ε Aurigae Campaign 2009 Photometry Report Jeffrey L. Hopkins 633
An Analysis of the Long-term Photometric Behavior of ε Aurigae Brian K. Kloppenborg, Jeffrey L. Hopkins, Robert E. Stencel 647
V-band Light Curve Analysis of ε Aurigae During the 2009–2011 Eclipse Thomas Karlsson 668
Report From the ε Aurigae Campaign in Greece Grigoris Maravelias, Emmanuel Kardasis, Iakovos-Marios Strikis, Byron Georgalas, Maria Koutoulaki 679
Photoelectric Photometry of ε Aurigae During the 2009–2011 Eclipse Season Frank J. Melillo 695
Small Telescope Infrared Photometry of the ε Aurigae Eclipse Thomas P. Rutherford 704
UV-Blue (CCD) and Historic (Photographic) Spectra of ε Aurigae—Summary R. Elizabeth Griffin, Robert E. Stencel 714
Ha Spectral Monitoring of ε Aurigae 2009–2011 Eclipse Benjamin Mauclaire, Christian Buil, Thierry Garrel, Robin Leadbeater, Alain Lopez 718
High Cadence Measurement of Neutral Sodium and Potassium Absorption During the 2009–2011 Eclipse of ε Aurigae Robin Leadbeater, Christian Buil, Thierry Garrel, Stanley A. Gorodenski, Torsten Hansen, Lothar Schanne, Robert E. Stencel, Berthold Stober, Olivier Thizy 729
Spectroscopic Results From Blue Hills Observatory of the 2009–2011 Eclipse of ε Aurigae Stanley A. Gorodenski 743
Eclipse Spectropolarimetry of the ε Aurigae System Kathleen M. Geise, Robert E. Stencel, Nadine Manset, David Harrington, Jeffrey Kuhn 767
Table of Contents continued on next page
Polarimetry of ε Aurigae, From November 2009 to January 2012 Gary M. Cole 787
Modeling the Disk in the ε Aurigae System: a Brief Review With Proposed Numerical Solutions Richard L. Pearson, III, Robert E. Stencel 802
A Demonstration of Accurate Wide-field V-band Photometry Using a Consumer-grade DSLR Camera Brian K. Kloppenborg, Roger Pieri, Heinz-Bernd Eggenstein, Grigoris Maravelias, Tom Pearson 815
Stellar Photometry With DSLR: Benchmark of Two Color Correction Techniques Toward Johnson's VJ and Tycho VT Roger Pieri 834
Algorithms + Observations = VStar David Benn 852
An Artist’s Note on Art in Science Nico Camargo 867
JAAVSO Volume 40, Number 2—other papers received
BVRI Photometry of SN 2011fe in M101 Michael W. Richmond, Horace A. Smith 872
New Light Curve for the 1909 Outburst of RT Serpentis Grant Luberda, Wayne Osborn 887
International Observing Campaign: Photometry and Spectroscopy of P Cygni Ernst Pollmann, Thilo Bauer 894
The Pulsation Period of the Hot Hydrogen-Deficient Star MV Sagittarii John R. Percy, Rong Fu 900
GEOS RR Lyrae Survey: Blazhko Period Measurement of Three RRab Stars— CX Lyrae, NU Aurigae, and VY Coronae Borealis Pierre de Ponthière, Jean–François Le Borgne, F. Fumagalli, Franz–Josef Hambsch, Tom Krajci, J–M. Llapasset, Kenneth Menzies, Marco Nobile, Richard Sabo 904
Recent Maxima of 55 Short Period Pulsating Stars Gerard Samolyk 923
The Ross Variable Stars Revisited. II Wayne Osborn, O. Frank Mills 929
GSC 4552-1643: a W UMa System With Complete Eclipses Dirk Terrell, John Gross 941
VSX J071108.7+695227: a Newly Discovered Short-period Eclipsing Binary Mario Damasso, Davide Cenadelli, Paolo Calcidese, Luca Borsato, Valentina Granata, Valerio Nascimbeni 945
Variability Type Determination and High Precision Ephemeris for NSVS 7606408 Riccardo Furgoni 955
Is MP Geminorum an Eclipsing Binary With a Very Long Period? Dietmar Böhme 973
Recent Minima of 150 Eclipsing Binary Stars Gerard Samolyk 975
The Variable Stars South Eclipsing Binary Database Tom Richards 983
A Note on the Variability of V538 Cassiopeiae Gustav Holmberg 986
A Practical Approach to Transforming Magnitudes onto a Standard Photometric System David Boyd 990
The AAVSO 2011 Demographic and Background Survey Aaron Price, Kevin B. Paxson 1010
The Citation of Manuscripts Which Have Appeared in JAAVSO Arlo U. Landolt 1032
Abstracts of Papers and Posters Presented at the Joint Meeting of the Society for Astronomical Sciences and the American Association of Variable Star Observers (AAVSO 101st Spring Meeting), Held in Big Bear Lake, California, May 22–24, 2012
Fast Spectrometer Construction and Testing John Menke 1037
Observations Using a Bespoke Medium Resolution Fast Spectrograph John Menke 1037
Enhancing the Educational Astronomical Experience of Non-Science Majors With the Use of an iPad and Telescope Robert M. Gill, Michael J. Burin 1037
The Rotational Period of the Sun Using the Doppler Shift of the Ha Spectral Line Robert M. Gill 1038
A Single Beam Polarimeter (Poster) Jerry D. Horne 1038
Index to Volume 40 1039
Percy, JAAVSO Volume 40, 2012 609
Highlighting ε Aurigae and Citizen Sky
John R. PercyDepartment of Astronomy and Astrophysics, University of Toronto, Toronto. ON M5S 3H4, Canada; [email protected]
WelcometothisspecialissueoftheJournal of the American Association of Variable Star Observers,featuringpapersoneAurigaeandtheCitizenSkyproject (www.citizensky.org). As Price et al. explain in the opening paper,Citizen Sky grew out of International Year of Astronomy 2009, brilliantlypiggybackingontheinternationalcampaigntoobservethe2009–2011eclipseofabrightbutmysteriousstar.Itincorporatedthe“citizenastronomy”philosophyon which the AAVSO is founded—skilled volunteers participating in andadvancing astronomical research. Itwas a complexmultidisciplinaryprojectwhichrequiredcarefulorganization,facilitatedbyagenerousgrantfromtheNationalScienceFoundation.Thepapersinthisissuerepresenttheculminationoftheastronomicalresearchaspectsoftheprojectbut,asPriceet al.explain,thereweremanyotherpositiveoutcomesoftheprojectwhich,wehope,willcontinuetobearfruitinthefuture.Inparticular,itprovidesanexcellentmodelforhowtoorganizeandmanageacomplexproject,andevaluatetheresults—somethingwhichisrarelydone. I’ve been actively involved in astronomical research for exactly halfacentury,soIhaveexperienced threeeclipsesofeAurigae(if I include the1955–1957onewhichIreadaboutasagraduatestudent).Iwaseducatedinanastronomydepartment(attheUniversityofToronto)whichhadspecialexpertiseinvariableandbinarystars—ofwhicheAurigaeisaprominentexample.Ihaveaspecificinterestintheintrinsicvariabilityofthesupergiantcomponentofthesystemso,unlikemanyobservers,Idon’thavetowaitfortheeclipseseverytwenty-sevenyears.Whatfascinatesmeaboutthepapersinthisissue,andinCitizenSkyingeneral, ishowtheyillustratesomanydiverseandintriguingfacetsofvariablestarastronomy. Firstandforemost:eAurigaeisagenuineastronomicalmystery.Itconsistsof a pulsating blue supergiant and a secondary component, in a 9,896-day(27.09-year)orbit.Thesecondarycomponentissurroundedbyadonut-shapeddiskofgasanddust,whicheclipsestheprimarycomponentforalmostexactlytwoyearseachorbit.Itisstillnotclearwhethertheprimaryisanormalmassivesupergiant,inwhichcasethesecondaryismostlikelyapairofstarsincloseorbit(the“high-massmodel”),orwhethertheprimaryisalow-masssupergiant,and thesecondary isasinglestar (the“low-massmodel”).Wehope that thepapersinthisspecialissuewillhelptoresolvethismystery.ForanexcellentoverviewofeAurigaepriortothe2009–2011eclipse,seeTempleton(2008). Thisisfrontierastronomy.Disksofgasanddustarewidelystudied,becausetheyareubiquitous in theuniverse.Theyare foundaroundformingstars, in
Percy, JAAVSO Volume 40, 2012610
cataclysmicbinarysystems,andaroundsupermassiveblackholesatthenucleiof galaxies. Binary and multiple star evolution is poorly understood, eventhoughbinaryandmultiplestarsarethenorm,nottheexception.Itisatopicwhichisrelevanttosomeofthemostexcitingtopicsinastrophysicstoday.TheoriginofTypeIasupernovae,whicharethemostimportant“measuringsticks”inmoderncosmology,isoneexample. eAurigae has an apparentV magnitude of about 3.0—not much fainterthanthestarsintheBigDipper—soit isvisibleevenincitiessuchasmine.Beginningskywatcherstendtothinkthatallstarsarethesameso,atstarparties,Iliketotellthemabouttheparticularcharacteristicsofindividualbrightstars,especially bizarre stars such as eAurigae. Beginning skywatchers can thenbecomebeginningobservers;eAurigaeiseasytomeasurewiththeunaidedeye(thoughbinocularshelp). Indeed, thiswasthewholepointofCitizenSky; itbringsastronomyandastronomicalresearchtoanyone,ofanyage,anywhere. Asothershavepointedout,eacheclipseofeAurigaebringsanewgenerationofastronomers,andanewgenerationof technology.The1982–1984eclipsecoincidedwiththeblossomingofamateurphotoelectricphotometry(PEP).The2009–2011eclipsewasobservedwithCCDandDSLRcameras,aswellasPEPandvisualtechniques. Itwasalsoobservedwithspectroscopy,atechniquewhichisincreasinglyavailabletoamateurs,andwithpolarimetry—afrontiertechniqueforsuitably-equippedadvancedamateurs.Veryfewprofessionalobservatoriesarecarryingout long-termspectroscopicandpolarimetricobservations, so this isanareainwhichamateurscouldtakeuptheslackinthefuture.Thiseclipsewasalsoobservedwithanopticalinterferometer—GeorgiaStateUniversity’sCHARAarray—so,forthefirsttime,wecan“see”thismysterioussystem. Finally, this eclipse campaign benefitted from modern information andcommunication technology, including email, the Internet, and social media.These made the transfer, display, and storage of information efficient andeffective.Theyalsohelpedtoconnecttheinternationalnetworkofamateurandprofessionalobserversasneverbefore. Youwillnoticethattheauthorsofthepapersinthisissuecomefrommanycountries.Theobservationswhichamateursandprofessionalshavecontributedduring this eclipse campaign come from even more countries, so everycontributorispartofaninternationalcommunityofvariablestarobservers. Furthermore,thearchivaldataoneAurigaestretchbackforover160years.Thedata in theAAVSOInternationalDatabasecover the last seveneclipses(1847–1849,1874–1876,1901–1903,1928–1930,1955–1957,1982–1984,and2009–2011),includingobservationsbytheeminentFriedrichArgelanderofthefirstofthese.ThevariabilityofthestarwasfirstsuspectedbyJohannFritchin1821,duringthe1820–1822eclipse.The1847–1849eclipsewasobservedsystematically byArgelander and others. Hans Ludendorff (1904) suggestedthatthestarwasaneclipsingvariable.By1937,ithadengagedthemindsof
Percy, JAAVSO Volume 40, 2012 611
some of the foremost astronomers of the time: Gerard Kuiper, Otto Struve,andBengtStrömgren(1937)statedthat“thephotometricandthespectroscopicdataoneAurseemtoleadtocontradictionsunparalleledinthestudyofothereclipsing systems”. By the 1955–1957 eclipse, the situation was even morecritical, as stellar astrophysics had matured greatly. Various models wereproposed,includingthepossibility(sinceruledout)thatthesystemcontainedablackhole.Today’sobserversarethusconnected,acrosstime,withyesterday’sobservers and interpreters. The AAVSO centennial, the centennial history(WilliamsandSaladyga2011),andthehistoricalpapersinthecentennialissueofJAAVSO(volume40,number1)weretimelyvehiclesforhelpingtomakethesehistoricalconnections. CitizenSky,andthisspecialissue,wouldnothavebeenpossiblewithoutthesupportofmanyorganizationsandpeople:theNationalScienceFoundationforgrantsupport(DRL-0840188);theAAVSOanditsstaffwhoprovidedin-kindsupport; theCitizenSky team:AaronPrice,RebeccaTurner,RobertE.Stencel,BrianKloppenborg,andArneA.Henden;partnerorganizationsandindividualswhocontributedtopartsoftheprojectsuchastheworkshopsandtutorials;theauthorsofthepapersinthisissue;andalltheobserversandothervolunteerswhocontributedtothediversescientificandeducationalaspectsoftheproject.Special thanks,asalways, to the staffof thisJournal:AssociateEditorElizabethWaagen,AssistantEditorMatthewTempleton,andProductionEditorMichaelSaladyga.Wehopethatyouenjoythisissue!
References
Kuiper,G.,Struve,O.,andStrömgren,B.1937,Astrophys. J.,86,570.Ludendorff,H.1904,Astron. Nachr.,164,81.Templeton,M.2008,http://www.aavso.org/vsotsepsaurWilliams,T.R.,andSaladyga,M.2011,Advancing Variable Star Astronomy:
The Centennial History of the American Association of Variable Star Observers,CambridgeUniv.Press,Cambridge.
JAAVSO Volume 40, 2012612
JAAVSO Volume 40, 2012 613
ε Aurigae Papers
Price etal., JAAVSO Volume 40, 2012614
The Origins and Future of the Citizen Sky Project
Aaron PriceRebecca TurnerAAVSO Headquarters, 49 Bay State Road, Cambridge, MA 02138; address email correspondence to R. Turner, [email protected]
Robert E. StencelBrian K. KloppenborgUniversity of Denver, Department of Physics and Astronomy, 2112 E. Wesley Avenue, Denver, CO 80208; [email protected]; [email protected]
Arne A. HendenAAVSO Headquarters, 49 Bay State Road, Cambridge, MA 02138; [email protected]
Received June 12, 2012
CitizenSkyaroseasaproductofthe2009InternationalYearofAstronomy(IYA)project.AaronPricewasappointedtotheU.S.IYAProgramCommitteewhenitwasfirstconstitutedin2006.Afteraboutayearofplanning,thecommitteeestablishedaWorkingGroupforResearchExperiencesforStudents,Teachers,andCitizenScientistswithAaronas thechair.Thegoalwas touse the IYAinitiativetopromotecitizenscienceprojectstothegeneralpublic.ThegroupmetforthefirsttimeattheAstronomicalSocietyofthePacificmeetinginChicago—2007.At the meeting they decided to focus their efforts around one projectinsteadofpromotingaslateofprojects.Manyspecific ideaswerediscussedduringamorningmeeting,butnothingcaughttheexcitementoftheentiregroup. Intheafternoon,AaronhadahallwaydiscussionwithRickFienberg,theneditor-in-chief of Sky & Telescope magazine. Rick mentioned the upcomingeAurigaeeclipseasapossibletopic.Hehadpreviouslyco-presentedapaperonthetopicwithRobertStencel(“Dr.Bob”)atthe2006meetingoftheInternationalAstronomicalUnion.Itseemedlikeanidealproject.Andastheworkinggroupresearcheditmore,itbecamemoreandmoreenticing.eAurisbrightenoughtobeseenfromthecity,itseclipsehasanamplitudethatcanbedetectedwiththenakedeye,ithappensveryrarely,itisdifficultforprofessionalstomonitor(dueto limited large telescope time)and, aboveall, itwas still anenigma—evenafteroveracenturyofresearchandspeculation. U.S.IYAfundingwasaconstantissue.TheProgramCommitteeonlyhadenoughfundstosupportasmallstaffandafewprojects.TheWorkingGroupknewthatanyprojecttheycameupwithwouldbetargetedatanarrowaudienceandsowouldnotreceivethesamelevelofsupportfromtheProgramCommitteethat a more generalized topic—such as the Galileoscope—received. So theWorkingGroupcameupwithtwoplans.Onewasaimedatwhattheywould
Price etal., JAAVSO Volume 40, 2012 615
dowithzerofunding—whateachmemberofthegroupcouldcontributegratisand/orcandoaspartoftheirregularjob.FortheAAVSO,thismeantissuinganAlert Noticeonthetopic,makingcharts,andrunningaregularcampaign,butnotmuchmore.Theyalsocameupwithaprogramtoimplementiftheycouldraise substantial levelsof support.After these twoplanswere inplace, theylookedforfunding. The National Science Foundation (NSF) seemed like a natural fit.Theyfund citizen science, have a nice history with theAAVSO, and the workinggrouphadsomeexperiencewritingproposalsforthem.However,strongNSFproposalsarenoteasytoassemble.IttookalmostsixmonthstowritetheCitizenSkyproposal.Aaronwroteadraft,circulateditamongsttheWorkingGroup,andthentheyheldconferencecallstodiscussrevisions.Theprocesswentverysmoothly,buttookalotoftime. AAVSO Director Arne Henden was chosen as the project’s principalinvestigator(PI)becauseAarondidnotyethaveaPh.D.(hewasingraduateschoolat thetime).Allof theco-PIscontributedtotheproposal.InadditiontoAaron, they were Lucy Fortson, Jordan Raddick, and Bob Stencel. RyanWyatt also contributed to the proposal and was a member of the advisoryboard along with Chris DuPree, Suzanne Jacoby, Hee-Sun Lee, and DavidAnderson.TheproposalitselfwasoriginallycalledSTARS—ScienceThroughtheAstronomicalResearchofStars.ItwasafterfundingthatitwaschangedtoCitizenSky,anamefirstproposedbyRebeccaTurner. RobertStencelservedasscientificadvisor fromtheverybeginning.Theoriginal idea to involve web cams and DSLR cameras began with him. Heprovidedmuchneededscientificadviceandalsoawelcomesenseofhumorwhenthingsgottightanddeadlinesloomed.Healsorecruitedagraduatestudentofhis,BrianKloppenborg,intotheproject. By the time we submitted the proposal to the NSF’s Informal ScienceEducation (ISE) program, which funds their citizen science projects, thebudget had increased to about $796,000 for a three-year project. It was aninterdisciplinary project with funding for modeling software, workshops, ahigh-endplanetariumshow,journalpublication,ScienceOlympiadmaterials,andmuchmore.ThescientificgoaloftheprojectwastocollectdatatohelpresearchersuncoverwhatiscausingtheeAureclipse.Theeducationalgoalwastoinvolvethepublicinallaspectsofascientificproject.Insteadoffocusingsolelyondatacollection,theprojectaimedtotrainparticipantstoanalyzedata,pose their own research questions, and write up results for a peer-reviewedjournal.Thishadnotbeendonebeforeinalargescalecitizenscienceproject,and theAAVSO was in a unique position to do so thanks to our history ofhavinganengagedmembership.ThegoalwithCitizenSkywas toput it alltogetherinonepackageandmakeitaccessibletonoviceamateurastronomerswiththebeliefthatparticipantswillgetabetterunderstandingofthescientificmethodiftheyaremoreengagedwithitateverystepoftheway.
Price etal., JAAVSO Volume 40, 2012616
TheproposalwassubmittedinJune2008.InNovember2008wereceivedthewordthatithadbeenfavorablyreviewedbytheNSF.OverthenextfewmonthsweansweredquestionsandrespondedtosuggestionsfromtheNSFtostrengthentheproject.Inmid-April2009wefinallyreceivedofficialnotificationthattheprojectwouldbefullyfunded.However,wewerenewtotheNSFsothereweresubstantialadministrative tasks tocompletebefore fundingcouldbegin.Thisadditionaldelaycausedaproblembecause,asweallknow,theskywaitsfornoone.Theeclipsewaspredictedtobegininthelatterhalfoftheyearandweneededpre-eclipseobservations.Intheoriginalproposalweplannedtohavealmostayeartobuildtheproject,materials,andsoon,beforetheeclipsebegan.Nowwehadtodoayear’sworkinafewmonths. Rebecca Turner was assigned the position of Project Manager, meaningshe was responsible for overall implementation of the project. Dr. Bob wasthescientificadvisor.BrianKloppenborgwashiredasastaffmembertoassistRebecca.Thefirstauthorwasfundedtoconductanevaluationoftheproject,whicheventuallymorphedintohisdissertationatTuftsUniversity.TheMorrisonPlanetariumattheCaliforniaAcademyofSciences,actingunderRyanWyatt,was contracted to produce an eight-minute planetarium film for the project.TheAdlerPlanetariumandAstronomyMuseum,actingunderMarkSubbaRao,washiredtodevelopaninteractiveeclipsingbinarymodelingsoftware.JordanRaddickatJohnsHopkinswashiredtoconsultonwebsitedevelopment,whichwasmostlydonebyKateDavis,theAAVSOwebmasteratthetime. Thefirstyearof theprojectwasdedicated tobuilding infrastructureanddatacollection.ThewebsitewasofficiallylaunchedinJune2009.Itconsistedof simplified versions of WebObs, the light curve generator, and the QuickLookfile,tutorialsonobservingandreadinglightcurves,onlineforums,andablogmaintainedbythestaffandDr.Bob.ThefirstCitizenSkyworkshopwasheldinSeptember2009attheAdlerPlanetariuminChicago.Thefocusoftheworkshopwasondatacollection(visual,photoelectric,andspectra)andonthesciencebehindthesystem. Thesecondyearoftheprojectwasdedicatedtobuildingteamsandtrainingindataanalysis.Thenon-observingaspectofCitizenSkywasdesignedaroundtheconceptof teamsofparticipantsworkingtogether.Itwasinspiredbythesuccess of the chart and comparison star database teams that were active intheAAVSO in thepreviousdecade.Asecondworkshopwasheld inAugust2010attheCaliforniaAcademyofSciencesinSanFrancisco.Thefocusofthatworkshopwasondataanalysisandwritingscientificpapers. Thethirdyearoftheprojectwasdedicatedtoobservingtheendoftheeclipseandwritingpapersfor theJournal of the AAVSO (JAAVSO).Thiswasa lesspubliclyactiveperiodastheteamsworkedontheirpapers,mostofwhicharepublishedinthiseditionofJAAVSO, abouteAur,CitizenSky,andrelatedtopics. SowhatishappeningnextwithCitizenSky?Oneofthelessonswelearnedin thisproject is that teamscanindeeddogreatwork.Evidenceof that is in
Price etal., JAAVSO Volume 40, 2012 617
thisJAAVSO issue.However, some teamsdidnotachieve theirgoals.Therearemanyreasonsforthis,butinourviewtherewasoneoveralllessontolearn:project staff cannot appoint leaders of teams; theyhave togroworganicallyfromamongtheparticipants.Themostsuccessfulteamsweremostlytheonesthathadself-nominatedleaders.Theteamswhichhadstaff-invitedleaderswerelesssuccessful,withafewnotableexceptions. AttheendofthethirdyearoftheCitizenSkyProjecttheAAVSOreceivedasupplementalNSFgranttofundathirdworkshop.ThisworkshopwillfocusongeneratingamanualforDSLRPhotometry.Applicationswillbeacceptedinlate2012andearly2013.ItwilltakeplaceMarch22–24,2013,atAAVSOHeadquarters in Cambridge, Massachusetts. The program will comprise amixof1)talksbyexperiencedDSLRphotometristsand2)breakoutsessionsduringwhichsmallgroupsofparticipants(eachwithadesignatedleader)willdeveloppreassignedsectionsofthemanual.Thesegroupswillbedeterminedwellinadvanceandwillbebasedonparticipantinterest,experience,andskills.NopriorDSLRexperienceisrequired,but therewillbesomepre-workshopreadingandpreparationrequired.Theworkshopwillproduceaneasytouse,introductory manual that will help the AAVSO support observers who areinterestedingivingDSLRphotometryatry. WiththeconclusionoftheeAureclipseandnewsoftheDSLRworkshop,theAAVSOhasgivenCitizenSkyanewmission:tobethenewhomeoftheAAVSO’sbrightvariablestaractivities. Initially, theCitizenSkyareaof theAAVSOwebsitewillhosttheAAVSOBinocularObservingProgramandtheDSLR Photometry Program.Additional bright object programs/projects willbe added to the Citizen Sky pages as they develop. Important Citizen SkymaterialswillbemovedfromitsexistingwebsitetotheCitizenSkysectionoftheAAVSOwebsite.TheoriginalCitizenSkywebsitewillbefrozenin2013butwillstayonlineasanarchive. Asforstaff,theyhavealsobuiltonwhattheyhavelearnedaspartofthisproject.Dr.BrianKloppenborg,whoreceivedhisPh.D.atUniv.ofDenverinspring2012,isnowavisitingscientistattheMaxPlanckInstituteforRadioAstronomy in Bonn, Germany, working on interferometry. He continues toinvolveamateurastronomersinhisresearch.Dr.AaronPriceisnowManagerofResearchandEvaluationattheMuseumofScienceandIndustryinChicago,whereheisapplyingthesocialscienceresearchskillshelearnedthroughhisdissertationonCitizenSky’simpactonthescientificbeliefsofitsparticipants.Rebecca Turner has been promoted to Operations Director at the AAVSO,whereshe isapplyingmanyof theprojectandpersonnelmanagement skillsshe learnedasProjectManagerof theCitizenSkyproject.Dr.ArneHendencontinues asDirectorof theAAVSO, andDr.RobertStencel continues as aprofessorattheUniversityofDenver,andisnowplanningobservationsforthenexteclipseofeAurin2036!
Stencel, JAAVSO Volume 40, 2012618
e Aurigae—an Overview of the 2009–2011 Eclipse Campaign Results
Robert E. StencelUniversity of Denver, Department of Physics and Astronomy, 2112 E. Wesley Avenue, Denver, CO 80208;[email protected]
Received March 19, 2012; revised May 30, 2012; accepted June 19, 2012
Abstract Evidenceisprovidedfromthearrayofobservationsamassedduringthe2009–2011eclipse,thatdefinestheenigmaticbinaryeAurigaeascomprisedof an unstable F0-1 Iab star in orbit around a comparable mass upper mainsequencestar(orstars)enshroudedinadiskresultingfromFstarmassloss.Inthispicture,theFstarmaybeundergoingrapidevolutionarychanges,andtherecent67-dayprimaryquasi-periodmaymakeitsuitableforasteroseismicstudies.Thehiddenstar(s)mayhavegainedmassfromtheFstar,andthediskitselfprovidesopportunitiesforstudyofaccretion,dustevolution,anddynamics.
1. Introduction
Duringthe20thcentury,thebrightstareAurigaeconfoundedastronomersbecause the visible member of this single-lined spectroscopic binary starappearedtobeamassive,F-typesupergiantstar,withanequally-massivebutinvisible companion (Guinan and deWarf 2002). Unimpeachable evidencefor the eclipsing object, a disk in transit, was obtained with interferometricimagingduringthe2009–2010eclipse(Kloppenborget al.2010),buildingontheinfrareddetectionofsamebyBackmanet al.(1984).Hoardet al.(2010),and previous authors, argued that the F star is in a volatile, reduced mass,post-AsymptoticGiantBranch(AGB)evoutionaryphase,withthecompanionbeingadisk-enshroudedB-typemainsequencestar—possiblynowthemoremassiveobjectinthesystem.Resolutionofthemassesandevolutionarystateofstarsinthissystemisaprimarymotivatorfortherecenteclipsecampaign,resultsofwhicharesummarizedinthispaper.ThisarticleispartofagroupofarticlescollectedintheJournal of the AAVSO,intendedtodocumentresultsofaninternationalefforttocollecthighqualityobservationsoftheeAursystemduring its 2009–2011 eclipse. Those reports provide details for each facet,butherewesummarizesomeoftheimportantfindings,inrelationtofindingsrecognizedasaresultofthestudiesofpreviouseclipses.
whereRJD=JD–2,400,000andtheuncertainty in timings isat leastone totwoweeks.Thenotionofcontacttimeshistoricallyreferstotangentcrossingsbetweenobjectsthatappearcircularinprojection,butwithaellipsoidaldiskshapeinvolved,thenormalmeaningsofsecondandthirdcontactsarechanged.Also,thesetimingsarecomplicatedbypersistent~0.1magnitudelightvariationsonaquasi-periodof67days(Kim2008).Spectroscopicevidencesuggeststhatdiskmaterialencroachedonthelineofsightmonthspriortophotometricfirstcontact(forexample,inKI7699A—seeLeadbeateret al.,thisissue),orevenyearsprior(Ha,seeChadimaet al.2011).SpectroscopicfourthcontactwasrecentlyannouncedbyobserverThierryGarrell,whoreportedexcessblue-shiftedHaandNaD-lineabsorptiondisappearedcircaRJD55950(2012January25). Theeclipsingobject isa large,550Kflattenedstructure,basedonrecentinfrared photometry and imaging. Although speculated to be a “swarm ofmeteors”byHansLudendorffearlyinthe20thcentury,thefirstphotometricevidencefortheeclipsingobjectwasprovidedbyMitchell(1964)withnine-colorphotometry,whoclaimeda500Kexcess,withaprojectedlinearsizeof50AU!Later,Backmanet al. (1984)observedeAurenteringeclipseandconfirmedthe result, reporting a 500K blackbody, with an apparent size of 8×10–16
ster-radians.At thereferencedistanceof650pc, this translates toanareaof14AUsquared,orequivalenttoarectangularshape1AUtalland14AUlong. The eclipsing object is disk shaped as seen in the near-IR. Thanks toremarkableprogressininterferometricimagingoverthepastdecade,itwaspossible in autumn 2009, during ingress, to detect the shadow of the diskcrossing the faceof theF star (Figure1),using theMIRCbeamcombinerat the Center for HighAngular ResolutionAstronomy (CHARA)Array oftelescopesatopMt.Wilson,California(Kloppenborget al.2010).Moreaboutthisbelow. Theneutralpotassiumlineat7699Åreappearedandshowedvelocityshiftsthatassociateitwiththediskanditsrotation.OriginallyreportedbyLambertandSawyer(1986),extensivemonitoringbyRobinLeadbeater(reportedinthisissue) showed a repeat of the phenomenon, demonstrating disk rotation andshowingstepwisechangesintheaddedequivalentwidthoftheline,suggestiveof disk substructure (Leadbeater and Stencel 2010). Subsequent study ofspectrarevealsanumberoflineswiththisbehavior(seeLeadbeateret al.2012;Schanneet al.2012;GriffinandStencel2012).
Stencel, JAAVSO Volume 40, 2012620
TheonlymoleculedetectedineAur,sofar,istransientcarbonmonoxide.ThiswasoriginallyreportedbyHinkleandSimon(1987),andshownbyStencelet al.(2011)toreappearagainaftermid-eclipse,usingbothmoderateresolutionIRTF+SpeXandhighdispersionGeminiNorthNear-IRSpectrometer(GNIRS)spectra. The data do not clearly confirm the low isotopic 12C/13C ratio ~10(comparedtothesolarvalue,89).TheoverlapofaPfundserieslineofhydrogenatop the 13CO bandhead demands post-eclipse observations where the COcontributionisremoved,allowingdisentanglementoftherelativecontribution.SeeFigures2aand2b.Alessextreme12C/13CratioreducestheevidencefortheFstarbeingalowermass,post-AGBstar. For the first time, transient HeI 10830Å absorption was detected, andit strengthened around mid-eclipse (Stencel et al. 2011), using NASA’sIRTF+SpeXinstrument;seeFigure3.Thislinearisesfroma19.8eVlevelandsuggestsahightemperaturecentralregioninthedisk.ThisisconsistentwithaccretionontoaB0-B5centralstar(Pequetteet al.2011). InfraredmonitoringconfirmedthepredictionbyTakeuti(1986)thatthesideofthediskfacingtheFstarisheatedto1100K(Hoardet al.2012),incontrasttothe550KsidefacingawayfromtheFstar.Amongtheimplicationsofthisarethatthebinaryseparationmightbeevaluated,independentoftheuncertaindistance,andthisdegreeofheatingisrelatedtothematerialpropertiesofthedustinthedisk. Aseriesofthreein-eclipseobservationsofthethefar-ultravioletspectrumofeAurwereobtainedwiththeCosmicOriginsSpectrograph(COS)onHubbleSpaceTelescope (Howell et al. 2011), and it was found that the continuumoutputissomewhateclipsed.Thefar-UVemissionlinesshowevidenceforaPCygni-likeoutflow(Figure4).Bothfactspoint tosurfaceactivityof theFstar,althoughcontributionsfromthediskcentralregionscannotberuledout. Attemptstoobservesolidstatespectralfeatures(ices,PAHs,or10-micronsilicates)failedtodetectany.InstrumentsinvolvedincludedSpeXandBASSatNASA’sIRTF,plusMIRACatMountHopkins(seeStencelet al.2011).Thebroadbandinfraredexcess,combinedwithlackofspectralfeatures,arguesthatthedustiscomprisedoflargeparticles(greaterthan~1micronsize).Kopal(1954)concluded similarlywhenstudying the similarityofeclipsedepthatmanyopticalwavelengths—theeclipse is“gray”due to largeparticlesizes,muchgreater than thewavelengths involved.Asnotedabove,eclipsedepthdeparturefromgrayness isseenatwavelengths longer thanseveralmicronsandintheultraviolet,andthesemayprovideusefulconstraintsonparticlesizeandtype. Photometric monitoring shows persistence of ~0.1 magnitude variationswithaquasi-periodof~67days(Kim2008),bothinandoutofeclipse(Figure5,andseeHopkins(2012)).ThesehavebeenassociatedwithFstaroscillationsandperhapswindevents(seeGriffinandStencel2012).
Stencel, JAAVSO Volume 40, 2012 621
3. System parameters
Among the generally agreed system parameters are the eclipse period,9890±2days,and themassfunction in thissingle linespectroscopicbinary,f(M) =2.51±0.12 (Stefaniket al. 2010).System inclination seems securelyestablishedatverycloseto90degrees,basedoninterferometricimagesshowingthediskeclipsingtheprimarystar(Kloppenborg2010).Giventheseparameters,possiblesolutionsforthemassratio(q)includethe“highmass”case,q~1.1(sumofmasses,M+m~25M
Ä,andsemi-majoraxis,a~25AU)andthe“low
mass”case,q~0.5(M+m~9MÄ
,a~18AU),whereMistheFstarmass,andmreferstothemassofthesecondaryobject–assumedtobeastarinsidethedisk,andwhereaisthesemi-majoraxisoftheorbit.Astrometricsolutionsforbothextremeshavebeenpublished(Strand1959;vandeKamp1978).Workisunderwaytoincludeinterferometricresultsasanewconstraintoncombinedastrometricandphotometricsolutions(Kloppenborg2012),whichfavors thehighermasssolution. What constraints do we have on the mass (M) of the primary star?Classically, the spectral type of the primary star has been classified as anF0Iastar(7500K).Yellowsupergiantstarshavecatalogedmassesof12M
Ä
andabsolutemagnitude,MV=–6.6 (see, for example,StraizysandKuriliene1981). The historic q=1 solution for this single lined spectroscopic binaryimmediately called into question how a high mass companion could remaininvisibleoutsideoftheeclipsesitcauses(StruveandElvey1930).Usingtheapparentmagnitude,V=3.05andthereddening,AV=0.3,theimplieddistanceis740pc(about20%larger thanthatusedbyKloppenborg(2010)basedonthefirstHipparcosparallax).Fromthis,theimpliedluminosityis3.7×104L
Ä,
andtheimpliedradiusis115RÄ
—largerthanmostCepheids!At740pc,theangular diameterwould be 1.5 milli-arcsec (hereafter, mas), at least 0.5massmallerthanrecentinterferometricdeterminations(2.27mas,K-band,Stencelet al.2008).This requires thestar toeitherhaveunusual limbdarkening,orperhapsnotbeasdistantas740pc.Formally,a2.1masV-banduniformdiskdiameterindicatesadistanceof650pc. AninterestingconstraintonbinaryseparationcanbededucedfromrecentthermalIRdata,whichshowsthattheportionofthedisknearesttheFstarrisesto~1100K(Stencelet al.2011;Hoardet al.2012).Thisequilibriumtemperaturefordustparticles,withlowalbedo(0.3tozero),impliesaseparationfromthe7500KF0starof9to12AU.Foradiskradiusof4to5AU(seeKloppenborg2010),theimpliedbinaryseparationis13to17AU,whichismoreconsistentwiththeq=0.5solution.Thesumofmassesinthiscaserangesfrom3to7M
In terms of resolving the ambiguity in distance and masses, twodevelopmentsdeservemention.DissertationworkbyBrianKloppenborg,attheUniversityofDenverat the timeof thiswriting,combinedastrometric,spectroscopic, and interferometric constraints on the orbit absolutedimensions. This resulted in a solution that places the system at 737±67pc, with a separation of 25AU and masses of M=13 and m=11M
Ä.The
interferometricallyconstraineddiskdiameterof7.31±0.66AUreferstotheoptically thickest portions and hence a minimum disk size, relative to thelargersizesimpliedbyspectroscopicconstraints. Whatotherconstraintsarepossibleonthemass(m)ofthesecondarystar?Diskrotationcurvesmaybeobtainedfromtheopticallythinneutralpotassiumlinesat7699Å(LambertandSawyer1986;Leadbeateret al.2012),seenonlyduringeclipsephases,andarguablyarisingfromextremeouterportionsofthedisk.Theirmeasureddiskrotationspeedis~35±–5km/sec.Rotationcurvesfromopticallythickermetalliclines(forexample,TiII4028Å)indicatethediskrotationspeedis42±–2km/sec(Saitoet al.1987).Then,withadiskradiusvalue,wecandetermineaKeplerianrotationperiodatthesespeeds.InterferometricimageswerefittedbyKloppenborg(2010)witha3.8AUellipticalsemi-majoraxis, assuming a 625pc distance. Said rotation speeds (35,42km/sec) haveimplied circumferential rotation rates of 3.25 and 2.70 years, respectively,implying a central mass of 5.2 to 7.5 M
Another constraint on the mass of the secondary can be inferred fromobserved disk dust and gas scale heights. As shown by VEGA+CHARAobservations (Mourardet al.2012), theeclipse inHa is total incomparisonto the partial eclipse seen in the near-IR with the Michigan InfraRed beamCombiner(MIRC)atCHARA—thatis,thehydrogengasextendsatleastafullFstardiameteraboveandbelowthediskplane.Forstrictlythermaldispersionofmaterial,thescaleheight(H)todiskradius(RD)ratioequalsthesoundspeed-to-orbitalvelocityratio,whichis:
[H/RD] = [(γkT / m)/(Gm/RD)]1/2 (1)
(seeLissaueret al. 1996), where γ is the ratio of specific heats (5/3 for ideal gases),mistheatomicmass,andmisthecentralstarmass.TablesandFiguresinLissaueret al.illustratethattheobservedthicknessisseveraltimesthescaleheight.TheobservedratiooftheminimumVEGA-observedthicknessofgas
ofdynamicalstirring,thesemassesareanupperlimit,asthedisk’shydrogenatmosphere could be thicker. Higher gas temperatures and/or a thicker diskscaleheightbothpoint to a lower secondarymass. If theoptically thindiskradiusislargerthanimpliedbythephotometriceclipseduration,theresultantmasscouldbelarger. Arethereanyadditionaldistance-independentdiscriminators?OnemightbetheEddingtonlimitonluminosity,whichisclassicallycomputedtobe:
. Other distance-independent means ofestablishingbinarystarparametersareneeded.
3.1.Natureofthedisk Kloppenborget al.(2010)estimatedthatthedisk-fittedsemi-majoraxisis6.10mas,meaningthefullmajoraxisis12.2mas,or~5.8timesFstardiameter(2.1mas).Giventhereported0.62masE-Wmotionpermonthduringingress(plusaN-Scomponent0.34mas/month,foranetmotionof0.72mas/month),anda2.1masuniformdiskdiameterstar,itshouldtake~2.9monthsforagivenpointinthedisktomoveacrosstheFstardisk,assuminguniformmotion,and~18monthstohavetheentirediskmovepast,whichisclosetotheobservedfirsttothirdcontacteclipseduration.Theellipsemodeldevisedduringingresswas poorly constrained, dictated mainly by the expected length of eclipse.Afterwards,weobtainedthecompletelightcurve(Figure5)andthefirst tofourthcontact lengthindaysisoforderRJD5800–5060=740days,or~24months,and largerstill inspectroscopic terms,suggesting thedisk is largerthanoriginallyestimatedand/or therelativevelocitiesarevarying(slowing,post-periastron). Evidenceexiststhatthediskisstructuredandasymmetric.Therehasbeenoutstanding spectroscopic monitoring of Ha by Mauclaire et al. (2012) andofKI7699ÅbyLeadbeateret al.(2012)andbluespectroscopicfeaturesbyGriffinandStencel(2012),reportedinthisissueandelsewhere,indicatingan
Stencel, JAAVSO Volume 40, 2012624
extended,asymmetricspectroscopicsignatureofthedisk.TheHaequivalentwidthchangesversustimesuggestacompactringmaycontributetolatephasesofingress(RJD55150–55250),whilethediskmainbodyisseenduringtotality(withapossiblemid-eclipsedeclinecircaRJD55450,perhapsduetoionizationrelatedtotheappearanceofHeI10830Å(Stencelet al.2011)).Amorediffusefollowingringassociateswithegress,butlastingpastnominalfourthcontact(RJD55800)withexcessHaabsorptionstillpresentinearly2012,agood150dayspastfourthcontact.Similarly,theneutralpotassiummonitoringhasbeeninterpretedtoshowstepwisechangesinequivalentwidthrelatedtointernalringstructure(LeadbeaterandStencel2010).ThesewerelabeledasringsAthroughF.RingsDandEmightcontributetothelateingressHabehavior,whileringsD-A associate with egress phenomena.As reported elsewhere, however, therelationshipbetweenthedustandgascomponentsisonlyjustbeingrevealedbyspectro-interferometryobtainedwiththeVEGAinstrumentattheCHARAArray(Mourardet al.2012). AdditionalsuggestionsofanionizingphotoninducedshockintheingresssideofthediskhasbeenmadebySaitoet al.(1987).Similarly,amattertransferstreamimpactingtheegresssideofthediskhadbeenadvancedbyStruveandcolleaguesearlyon,andsomethingtothiseffectwasseenagaininblueregionspectra reported by Griffin and Stencel (2012). Finally, the out-of-eclipsephotometricvariationsareconsistentwithsurfacephenomena-drivenmasslossepisodesfromtheFstarthatcontributetoastellarwind,asymmetricallyshapedtoward thecenterofmassand ionizedbyUV light scatteredaway from thesourceinternaltothedisk(Figure6).
3.2.Dustscattering Thecompositionof thedisk isof interestbecausedeterminingdust sizewould provide evidence for the age of the object and the degree to whichplanetessimal formationmay have occurred in the disk. Kopal (1954) notedthewavelength-independenceofeclipsedepthsandconcludedthatlargegrainsmustbepresentwithmulti-micronsizes,muchgreaterthanopticalwavelengths.However,observationstodatehavenotdetectedanytypicalspectroscopicbandsdue tosolids,suchas ice(3microns),PAHs,or10-micronsilicates(Stencelet al.2011).Anestimatefordiskmasscanbemade,basedonthefarIR/sub-mmfluxesreportedbyHoardet al.(2012):
M(dust)=Fnl2d2/(2kTdustkn) (3)
whereFn(250microns)=57mJy=3×10–22W/cm2/micron.Using650pcasthe distance, a dust temperature of 550K and a mass absorption coefficient,kn=3cm2/gm (Juraet al. 2001),wededuce a dustmassof 1.2×1031 gmor
nearly6Jupitermasses.Thisseemstoolargeifthegastodustratiois~100,because the disk mass would be dynamically significant in the system.
Stencel, JAAVSO Volume 40, 2012 625
Disk volume (Kloppenborg et al. 2010) and deduced densities (Hinkle andSimon1987)argueforalowerdustmass,roughlyanEarth-mass.Dustopacitybasedonsub-mmdiskstudiessuggeststhatmassabsorptioncoefficientsarenotwelldetermined.FurtherworkisneededtobettercharacterizethematerialintheeAurigaedisk(seePearsonandStencel2012). Another avenue for exploring the disk content involves polarimetry.BroadbandpolarimetricstudieshavebeenreportedbyKempet al.(1986)andbyCole (2012).These showanoutofeclipsebaselinevalue inphotometricVbandof2.0percent,presumed tobe interstellar,butgrowingandvaryingduringeclipseto2.4percent.ThesubstructureofV-bandpolarizationduringeclipsedoesnotsimplycorrelatewithphotometricchanges.Kempet al.(1986)correctlydeducedthepositionangleoftheeclipsingdiskdecadespriortotheinterferometricimagingofit.Recently,linepolarizationmonitoringhasbecomepossible,greatlyincreasinginsightintodiskandstarphenomena,andgreatlycomplicatingtheanalysis(seeGeiseet al.2012). Onemore lineofevidence tobeexplored involvesadifferencebetweeneclipsebehavioramongredwavelengthlineslikeHaandKI7699Å,andbluewavelengthlineslikeFeI3920Åandothers.Theformershowequivalentwidthvariation due to the eclipse, but remain substantially enhanced around mid-eclipse,whilethebluerlineshaveequivalentwidthexcessthatnearlyvanishesatmid-eclipseandthenreturnsstronglyduringlateeclipse.Thisistosaythatthelinedepthchangesarelessintheblueregionscomparedtotheredregions,suggestiveofparticlesizesdominatedbythegreater-than-micronscale.
4. Archives
Table1listsobservationsthatthisauthorproposedandconducted,manyof which will appear in public data archives maintained by the institutionsinvolved.TheAAVSOprovidesacomprehensivephotometryarchive.Thereareanumberofspectroscopicdatasourcesthatmightnotbefullyreflectedinthisreport,butwellworththemention.InadditiontothecarefuldigitizationbyElizabethGriffinofMt.WilsonandDAOphotographicspectra(plusnewdigitalobservations),additionalmodernhighdispersiondigitalspectrawereobtainedregularly during eclipse by Hideyuki Izumiura at Okayama Observatory, byWilliamKetzebackandcollaboratorsatApachePointObservatory(seeBarentineet al. 2012),byNadineManset andcollaborators atCFHT (seeGeiseet al.2012),byJohnMartinattheUniversityofIllinois,Springfield,Observatory,byNancyMorrisonattheUniversityofToledoRitterObservatory,byUlisseMunari and collaborators atAsiago Observatory, by Klaus Strassmeier andcollaborators (Astronomical InstituteatPotsdam)STELLA telescopeat IAC(seeSchanneet al.2012),andseveralothers.Ideally,thesedataarchivescanbemaintainedbytheiroriginatorsandmadeavailabletointerestedresearchers.Thisauthorwouldbehappytotrytohelpcoordinateanalysisefforts.
Stencel, JAAVSO Volume 40, 2012626
5. Conclusions
ClarificationofthenatureoftheeclipsingobjectineAurwillstandasoneofthemajorachievementsofthiseclipsecycle.WhiletheeclipsehasbroughtattentiontothestudyofeAuranditsmysteries,wenowhaveenteredthelonginter-eclipseperiod,prior to the startof thenexteclipsecirca2036.Currentorbitalelements(Stefaniket al.2010)predictkeytimesthatshouldbeofinteresttoobservers:theFstarreachesmaximumblueshiftcircaautumn2017(anditreachedmaximumredshiftduringautumn2006).EclipseofthediskbytheFstar is anticipatedduring2020and this shouldbedetectable in the infrared,insofarastheobservationalcapabilityexiststhen.Thereismeritinphotometricmonitoringofout-of-eclipsebehavior,althoughthecharacteristicbehavior(67-dayquasi-periodplusovertonesandevolution)seemsestablished(forexample,Kim(2008)andKloppenborg(2012)).TheAAVSO’sphotometricdataarchive(www.aavso.org) provides an excellent resource. Nonetheless, with robotictelescopesanddedicatedpersons,theslowchangesinthissystemcanyetbefollowedduringthisnewestorbitalcycle.
6. Acknowledgements
Therearemany,manypeopletothankforhelpinthisoveralleffort,makingcoverageoftherecenteclipsethemostcompleteinhumanhistory.WithoutJeffHopkins’andLouBoyd’sdedicationtolongtermphotometry,wewouldhaveamuchsparserdataset.MythankstoHalMcAlisterforacceptingmyproposalstousetheCHARAArrayforinterferometricstudyoftheeclipse,andtoJohnMonnierforaccess to theMIRCinstrument thatmade the imagingpossible.ManyfacetsoftheworkpresentedherewerefacilitatedbyBrianKloppenborgandhisamazingcomputerskills.IamgratefultoAaronPrice,RebeccaTurner,andArne Henden atAAVSO for their efforts with the Citizen Sky observercampaignthatenrichedthedataarchiveswiththousandsofnewmeasures,allcarefully archived. Additional spectroscopic monitoring was contributed byRobinLeadbeater,ThierryGarrel,ChristianBuil,ElizabethGriffin,andmanyothers. Key collaborators on HST COS and Spitzer/IRAC data acquisitionincludedSteveHowellandDonaldHoard,withexcellentadvicereceivedfromCOSteammembersStephaneBeland,JimGreen,StevenPenton,andothers.Collaborators obtaining and reducing Gemini North GNIRS data were TomGeballeandRachelMason.Helpfulcommentsonthispaperfromananonymousrefereeareappreciated.MyparticipationwassupportedinpartbythebequestofWilliamHerschelWombleinsupportofAstronomyattheUniversityofDenver,forwhichIamextremelygrateful(astheprojectwouldnothavebeenpossibleotherwise),alongwithanNSF-RAPIDgrant,AST10-16678,totheUniversityofDenver,organizedbytheprescientprogramdirectorDonaldTerndrup,whodeservespublicthanksforrecognizingthesingularopportunityaffordedbyeAur.
Hinkle,K.,andSimon,T.1987,Astrophys. J.,315,296.Hoard,D.,Howell,S.,andStencel,R.2010,Astrophys. J.,714,549.Hoard,D.,Ladjal,D.,Stencel,R.,andHowell,S.2012,Astrophys. J., Lett. Ed.,
748,L28.Hopkins,J.,2012,J. Amer. Assoc. Var. Star Obs.,40,633.Howell, S., Hoard, D., and Stencel, R. 2011, Bull. Amer. Astron. Soc., 43,
257.07.Jura,M.,Webb,R.A.,andKahane,C.2001,Astrophys. J., Lett. Ed.,550,L71.Kemp,J.C.,Henson,G.D.,Kraus,D.,Beardsley,I.,Carroll,L.,Ake,T.,Simon,
T.,andCollins,G.1986,Astrophys. J., Lett. Ed.,300,L11.Kim,H.2008,J. Astron. Space Sci.,25,1.Kloppenborg,B.2012,Ph.D.thesis,UniversityofDenver.Kloppenborg,B.,et al.2010,Nature,464,870.Kopal,Z.1954,Observatory,74,14.Lambert,D.,andSawyer,S.1986,Publ. Astron. Soc. Pacific,98,389.Leadbeater,R.,andStencel,R.2010,http://arxiv.org/abs/1003.3617.Leadbeater,R.,et al.2012,J. Amer. Assoc. Var. Star Obs.,submitted.Lissauer,J.,Wolk,S.,Griffith,C.,andBackman,D.1996,Astrophys. J.,465,371.Mauclaire,B.,Buil,C.,Garrel,T.,Leadbeater,R.,andLopez,A.2012,J. Amer.
Assoc. Var. Star Obs.,40,718.Mitchell,R.1964,Astrophys. J.,140,1607.Mourard,D.,et al.2012,Astron. Astrophys.,544A,91.Pearson,R.,andStencel,R.2012,J. Amer. Assoc. Var. Star Obs.,40,802.Pequette,N.,Stencel,R.,andWhitney,B.2011,Bull. Amer. Astron. Soc.,43,225.05.Saito,M.,Kawabata,S.,Saijo,K.,andSato,H.1987,Publ. Astron. Soc. Japan,
J. Amer. Assoc. Var. Star Obs.,submitted.Stefanik, R., Torres, G., Lovegrove, J., Pera, V., Latham, D., Zajac, J., and
Mazeh,T.2010,Astron. J.,139,1254.
Stencel, JAAVSO Volume 40, 2012628
Stencel,R.E.,Creech-Eakman,M.,Hart,A.,Hopkins,J.L.,Kloppenborg,B.K.,andMais,D.E.2008,Astrophys. J., Lett. Ed.,689,L137.
Stencel,R.,et al.2011,Astron. J.,142,174.Straizys,V.,andKuriliene,G.1981,Astrophys. Space Sci.,80,353.Strand,K.1959,Astron. J.,64,346.Struve,O.,andElvey,C.1930,Astrophys. J.,71,136.Takeuti,M.1986,Astrophys. Space Sci.,121,127.vandeKamp,P.1978,Astron. J.,83,975.
Table1.SelectednewobservationsofeAurduringeclipse,2009–2011,cont. RJD1 Calendar Date Telescope2 Mode
Figure1.Historicfirst:the1.6-micronwavelengthimageofeAur2009November2asinitiallyprocessedbyJohnMonnier,basedonfourtelescopebeamcombinationdataacquiredbyMIRCattheCHARAArray,andshowingtheshadowofthediskcrossingthefaceofeAur.Thescalesareinmilli-arcsecondunits.Image courtesy of John Monnier, Univ. Michigan.
Stencel, JAAVSO Volume 40, 2012630
Figure2a.AportionoftheGeminiNorthGNIRSspectrumofeAurspanningthe12CO(2-0)bandheadat4360cm-1region(2.293microns)showingspectrallinesof12CO(2-0up)and12CO(2-0down), illustratingtheGNIRSresolutioncapable of separating these contributions. The lines show a –25 km/secsystematicblueshift,characteristicofthediskatthisepoch.
Figure 6. Schematic model of eAur that incorporates F star wind focusingand ionizationeffectsdue toscatteringofUVphotonsoriginating inside thedarkdisk.Thismodelaccountsformanyofthenewlyobservedspectroscopicand radial velocity details reported here and in related papers, and providespredictionsforadvancedhigh-resolutionimaginginthefuture.
Hopkins, JAAVSO Volume 40, 2012 633
The International e Aurigae Campaign 2009 Photometry Report
Jeffrey L. HopkinsHopkins Phoenix Observatory, 7812 West Clayton Drive, Phoenix, AZ 85033; [email protected]
Received February 23, 2012; revised April 25, 2012, and June 29, 2012; accepted July 3, 2012
Abstract An International Campaign and Web site were started in Mayof 2006 for the 2009–2011 eclipse of the mysterious star systemeAurigae.Photometric and spectroscopic observations of the eclipse were coordinatedandreported.Theeclipsestarted in thesummerof2009and lasteduntil thespringof2011.Duringthecampaigntwenty-fournewsletterswerepublishedon theweb site andmadeavailable free as .pdf files to read anddownload.Twenty-sixobserversfromfourteendifferentcountriessubmittedphotometricdata in theUBVRIbands.Over3,600high-qualityphotometricobservationsweresubmittedwithnearly2,000observationsinjusttheVband.ThispaperdiscussestheCampaignandreporttheresults.
1. Introduction
Priortotheeclipse,datafrompreviouseclipsesandfrombetweeneclipseswereconsolidated in the formofabook (seeHopkinsandStencel2008).Apaperwasgiven at theSociety forAstronomicalSciences2010Symposiumthatdiscussedtheingressofthelatesteclipse(Hopkinset al.2010).Mostofthephotometricdataforthe2009–2011eclipsewereobtainedintheVband.Thelargestphotometricchangeswereintheintheshorterwavelengths,however.TheUband,whichwasonlyobservedwiththePMT-basedsystems,providedthelargestchanges. Some people thought that during the late spring and early summer thestarsystemwentbehindtheSunandwasnotobservable.Thisisnottrue.ThedeclinationofthestarsystemishighabovetheplaneofthesolarsystemanditnevergoesbehindtheSun.Theproblemisthatatlowerlatitudesitgetsveryclose to thehorizonduring thedarkhoursand thus the lightpasses throughvery high air mass. Extinction becomes a very significant problem. Thoseobserversathigherlatitudescouldobservethesystematlowerairmasses,butasonegoesfarthernorththenumberofdarkhoursdecreasestothepointofthemidnightSun.Someof thehigher-latitudeobserversdidheroicworkduringthesepoorobserving times.Even then thedataaresomewhatconfusingandnoisy.Themid-eclipsebrighteningperiodwasatoneofthesepoortimes.Somedataindicatethebrighteningandsomeindicatenobrightening.Also,midway
Hopkins, JAAVSO Volume 40, 2012634
duringegresswasapoor timeandsomething interestinghappened then too.Thereseemstobeaknee in theegressphotometricplots.Perhapsmoreandbetterdataduringthesetimescanbeobtainedduringthenexteclipsein2036. A Campaign was formed during the 1982–1984 eclipse and thirteennewsletters published. The current Campaign’s web site has some of thesenewslettersavailableonline.Forthe2009–2011eclipsethereweretwenty-fournewsletters published by the Hopkins Phoenix Observatory.All of these areavailableas .pdfsontheCampaignwebsitealongwithagreatdealofotherpertinentdataoneAur.Inadditiontothe2009Campaign’swebsiteathttp://www.hposoft.com/Campaign09.html, a Campaign Yahoo forum was startedandcanbeaccessedat:http://tech.groups.yahoo.com/EpsilonAurigae/
2. Observations summary
AsofDecember26,2011,wehadnearly3,700totalobservationsreportedduring the eclipse, with the visual band having by far the most at nearly2,000observations.Twenty-sixobservers fromaround theworldcontributedobservationaldata.Table1isasummaryoftheobserverscontributions.
3. Data quality
eAurisbrightenoughtobevisuallyseeneasilyeveninmostlightpollutedareas.Onecanevennoticethedimmingoftheeclipsevisually.Whiletheuseofvisualobservationsforplottingchangesinthebrightnessofstarsworkswellwithsomestars,theeclipseofeAurwasnotagoodprojectforvisualwork.Tobeofvaluemagnitudesestimatesmustbeofaresolutionof0.05magnitudeor better. This is an order of magnitude less than what even experiencedvisualobserverscanproduceunderthebestofconditions.ForthisreasontheInternationalCampaigndidnotuseanyvisualmagnitudeestimates. Photometricdatasubmittedtothecampaignhadaveragestandarddeviationsforthreeormoredatapointsofmagnitude0.01.Manyobservationsapproachedastandarddeviationofmagnitude0.001.Thestandarddeviationisusedasanindicationofthequalityofthedatabyrepresentingadataspreadofthreeormore data points.The photometric plots do not have error bars because thestandard deviations, which would be used for the error bars, are too smallinrelation to theplotscale.Mostsubmitteddatahavebeentransformedandcorrectedfornightlyextinction.TheHopkinsPhoenixObservatory(HPO)usedahigh-precisionUBV1P21-basedphotoncountingsystem.AllHPOdataweredeadtimecorrected,transformationcoefficientsdeterminedforeachband,andnightlyextinctioncoefficientsdeterminedforeachband.DatareductionwasdoneusingtheHardieequations.MagnitudesweredeterminedforbothlAurandeAurandthendifferentialcalculationsdoneandnormalizedtothestandardvaluesforlAur’smagnitudes.Nightlysessionsconsistedofthreecomparison
Hopkins, JAAVSO Volume 40, 2012 635
starmeasurementsineachband,threeadjacentskymeasurementsineachband,followedbythesameroutinefortheprogramstar.Therewerethreesetsofthesemeasurementsmadewiththeprogramstarbracketedbythecomparisonstar.Thedatawerealltransformedandextinctioncorrected.Threemagnitudesweredeterminedforeachband,astandarddeviationcalculated,andthemagnitudeaverageforeachbandreportedtofourplaces. TherehasbeensomeconcernabouttheUandBmagnitudesfromtheHopkinsPhoenixObservatory.First,therearenostandardUandBmagnitudesforeAur.Thestarsystemvariessignificantlyandrandomlyout-of-eclipse.ThevariationisgreatestintheUandBbands.Ifoneweretosuggestout-of-eclipsemagnitudes then theaverageofseveralseasonsofout-of-eclipsemagnitudescouldbe specified,but awarningwouldbeneeded indicating that these areaverages.Fromdata takenout-of-eclipsebetweenMarch1986andFebruary2008,thefollowingaremaximum,minimum,andaveragemagnitudesfortheUBVbands:
Oneofthegreatthingsaboutacampaignlikethisisthatonegetsachancetocomparetheirphotometricdatatakenatapproximatelythesametimeagainstothers.Photometriclightcurvesforthecampaignwereupdatedandpublishedonthewebaroundonceamonth.Observerscouldidentifytheirdataandseehowwelltheycomparedtoothers.Noisydataareobvious.Iftheobserverwasinterested, he or she could investigate the differences and improvements intechniqueanddatareduction.
Therewerefourtypesofinstrumentationusedtoacquiredataforthe2009Campaign:DigitalSingleLensReflex(DSLR)cameras,CCDcameras,PinDiode photometers (SSP-3 and SSP-4), and photon counting photometers(PMT-based). The photon counting provided highest quality UBV datafollowedbytheSSP-3forBVdata.ObtainingCCDdatawasmoredifficultas thebrightnesscausedproblems.CCDprovidedBVRIdata.Thelastandnewest equipment for photometry were the DSLR cameras. The DSLRprovidedV-banddatabyusingthegreenchanneloftheRGBimages.WiththeexceptionofacoupleDSLRobserversthedatasubmittedwereverynoisy.However,DSLRcamerasofferanexcellentintroductiontophotometry.OnceanobserverisconfidentwithdoingphotometrywithaDSLR,anentrylevelmonochromeCCDcamerawithBVRIphotometric filterswouldbeagreatwaytostartdoingprofessionalphotometry.
6. Results
6.1.HopkinsPhoenixObservatorydata The Hopkins Phoenix Observatory data were taken with a high-precision UBV photon counting system.All data were transformed, deadtimecorrected,andnightlyextinctiondeterminedandcorrected.Detailsofthe photometric work at the Hopkins Phoenix Observatory is reported inHopkinset al.(2007). When out-of-eclipse the star system has presented tantalizing data(Figure 1).The light is not constant, but varies at a pseudo-periodic rate inall thephotometricbands.Theperiod isnot stable,andvariesunpredictablybetween50and70days.Inadditiontotheperiodvariationstheamplitudesvaryunpredictably.Periodanalysiswasdoneusingperansoperiodanalysissoftware(Vanmunster2007),butnoperiodcouldbedetermined.DetailsofthisworkarereportedinHopkinsandStencel(2007)andHopkinset al.(2008). While not photometry, in an attempt to shed light on the out-of-eclipsevariations high-resolution out-of-eclipse spectroscopy of the star system’sHa region was done at Hopkins Phoenix. The main Ha absorption line isbracketedbyemissionlines(horns)thatgoupanddown,sometimestogetherandsometimescompletelyindependently(seeFigure2).Theysometimesreacha large peak and at other times disappear completely. This is known as the“HydrogenAlphaHornDance.”Noconnectionwas foundbetween theout-of eclipse Ha spectroscopic variations and the out-of-eclipse photometricvariations.Theemissionhornsdidseemtodecreaseandevengoawayforawhile during the eclipse, however. Details of the spectroscopic observationsatHopkinsPhoenixObservatoryarereportedinHopkinsandStencel(2009a,2009b)andHopkins(2012).
Observation techniques varied among observers. Details of each observerare presented on the 2009 Campaign’s web site at http://www.hposoft.com/Campaign09.html.
Therehasbeensomeconcernexpressedbyarmchairphotometristsaboutthecontact times.Thefollowingmayhelpunderstandthecomplexityofthissystemandthemethodologyusedtodeterminethecontactpoints. TheeAurstarsystemisverydifferentfrommosteclipsingbinarysystems.Fortheanalysisofthecontactpointstheclassicalmethodwasused.Inadditiontothenon-classicaleclipsingbody,complicatingtheanalysisarethepseudo-periodicout-of-eclipse(OOE)variations(seeFigure8). Figure9showstheprocedurefordeterminingthecurrenteclipsecontactpoints.Theaverageingressandegressslopeswereusedtofindtheintersectionwiththeaveragetotalityandout-of-eclipsemagnitudes. An archive of UBVRIJH photometric data is available to the public at http://www.hposoft.com/EAur09/Photometry.html and http://www.hposoft.com/EAur09/Data/UBVRIJHData.html. A summary of the data is given in Table 2.
8. Predictions for the 2036–2038 eclipse
While I am unlikely to be around for the next eclipse in 2036 it is stillinterestingtoofferphotometricpredictionsaboutit.OnlytheV-bandcontactpointsareincluded.Thereisstillcontroversyaboutthesecondandthirdcontactpoints.Thesedateswerecalculatedbyadding9,898daystothefirstcontact,mid-eclipse,andfourthcontactofthe2009–2011contacttimes.
EacheclipseofeAurseesanewbreedofequipmentandobservers.Whilemorewaslearnedfromthislatesteclipse,thestarsystemisnotgivingupitssecretseasily.Itseemstobetauntingus.Duringsomeofthemostinterestingtimes,thesystemwasextremelydifficulttoobserve.Theeclipsemaybeoverforanothertwenty-sevenyears,butthestarsystemstillpresentssomeinterestingchallenges. Monitoring and understanding the out-of-eclipse variationswillbeamajorobjective.This is followedby thestrangeHahorndance.Acontinuedfollowingofthestarsystemwithbothphotometricandspectroscopicobservationsissuggested.Thismayprovidesomeadditionalinsightsintothesemysteries.Oneofthenicethingsaboutobservingout-of-eclipseisthetimesofhighairmasscanusuallybeavoided.Observingcanbedoneduringfavorabletimes,suchasintheearlyfall,duringwinter,andinearlyspring.Asindicatedearlier, the most active regions for photometry are the shorter wavelengths.Uband is a especially important band in which to make observations. Thesystemalsooffers excellent spectroscopic learning for thehydrogenBalmerlinesaswellasthesodiumDlines.
10. Acknowledgements
I would like to thank all the observers who contributed data to theCampaign. Inparticular Iwant to thankDr.RobertStencel (“Dr.Bob”)andBrianKloppenborgfor theirhelp,encouragement,and thehonorofworkingwiththembothwiththeCampaignandespeciallywiththeobservingsessionsatCHARAonMountWilsoninCalifornia,andtheMMTonMountHopkinsinsouthernArizona.Thosearetreasuredexperiencesandmemories.
References
Hopkins, J. L. 2012, Small Telescope Astronomical Spectroscopy, HopkinsPhoenixObservatory,Phoenix,AZ.
Hopkins, J. L., Schanne, L., and Stencel R. E. 2008, in The Society for Astronomical Sciences 27th Annual Symposium on Telescope Science,SocietyforAstronomicalSciences,RanchoCucamonga,CA,67.
Hopkins,J.L.,andStencel,R.E.2007,inThe Society for Astronomical Sciences 26th Annual Symposium on Telescope Science,Society forAstronomicalSciences,RanchoCucamonga,CA,37.
Hopkins, JAAVSO Volume 40, 2012 639
Hopkins, J., and Stencel, R. E. 2008, Epsilon Aurigae: A Mysterious Star System,HopkinsPhoenixObservatory,Phoenix,AZ.
Hopkins,J.L.,andStencel,R.E.2009a,eAurigaeHydrogen-aEmissionLineVariations:TheHornDance,J. Amer. Assoc. Variable Star Obs.,37,213.
Hopkins,J.L.,andStencelR.E.2009b,inTheSociety for Astronomical Sciences 28th Annual Symposium on Telescope Science,Society forAstronomicalSciences,RanchoCucamonga,CA,157.
Hopkins,J.L.,et al.2010,inThe Society for Astronomical Sciences 29th Annual Symposium on Telescope Science, Society for Astronomical Sciences,RanchoCucamonga,CA,13.
Table1.2009eAurcampaignphotometricobservercount,cont.1 Observers (AAVSO observer initials are given in parentheses): CH (HEN), Colin Henshaw, Tabuk, Saudi ArabiaCO (OSC), Steve Orlando, Custer Observatory, East Northport, NYCQJ (CQJ), John Centala, Eastern IowaDES (LDS), Des Loughney, Edinburgh, Scotland, UKEAO (SIAK), Iakovos Marios Strikis, Elizabeth Observatory of Athens, Haldrf (Athens), GreeceEGO, Charles Hofferber, East Greenwood Observatory, East Grand Forks, MNEUO, Serdar Evren, Ege University Observatory, Izmir, TurkeyFJM (MFR), Frank J. Melillo, Holtsville, NYGHO (MXL), Richard Miles, Golden Hill Observatory, Dorset, EnglandGO (CLZ), Laurent Corp, Garden Observatory, Rodez, FranceGS (SAH), Gerard Samolyk, Greenfield, WIGVO (MBE), Brian E. McCandless, Grand View Observatory, Elkton, MDHPO (HPO), Jeff Hopkins, Hopkins Phoenix Observatory, Phoenix, AZJBO (BPJ), Paul J. Beckmann, Jim Beckmann Observatory, Mendota Heights, MN JESO, Dr. Mukund Kurtadikar, Jalna Education Society Observatory, Maharashtra, IndiaKO (LHG), Hans-Goran Lindberg, Kaerrbo Observatory, Skultuna, SwedenLO (GSN), Snaevarr Gudmundsson, Lindarberg Observatory, Hafnarfjordur, IcelandMSO, Arvind Paranjpye, MVS IUCAA Observatory, Ganeshkhind Pune, IndiaNKO, Nils Karlsen, Nils Karlsen Observatory, Umea, SwedenNPO, Gary Frey, North Pines Observatory, Mayer, AZRES (SVR), Dr. Robert E. Stencel, University of Denver, Denver, CORLO (HHU), Hubert Hautecler, Roosbeek Lake Observatory, Boutersem Brabant, BelgiumSGGO (CTIO), Tiziano Colombo, S. Giovanni Gatano al Observatory, Pisa, ItalyTP, Tom Pearson, Virginia Beach, VAVO (KTHA), Thomas Karlsson, Varberg Observatory, Varberg, SwedenWWC (CDK), Donald Collins, Warren Wilson College, Ashville, NC2 Equipment key: CCD, CCD Camera and telescope; DSLR, Digital Single Lens Reflex Camera, unguided; SSP-3, PIN Diode photometer with telescope; PMT, Photomultiplier Tube, photon counting with telescope.
H bandOOEMag. 1.605Mag. D0.645Mag.1stContact RJD=55,057 ±15days2ndContact RJD=55,202 ±15daysIngress 145days ±15daysMid-Eclipse RJD=55,395 ±15.5daysTotality AverageMag. 2.050Mag. AverageDepth 0.645Mag. Duration 388days ±15.5days3rdContact RJD=55,590 ±14days4thContact RJD=55,733 ±14daysEgress 143days ±14daysEclipse Duration 676days ±09.5days AverageDepth 0.645Mag. Period —1 The predicted times were calculated by adding the previous eclipse times to the previous determined periods. 2 There were no predictions for the longer wavelengths in the V band.
An Analysis of the Long-term Photometric Behavior of e Aurigae
Brian K. KloppenborgDepartment of Physics and Astronomy, University of Denver, 2112 East Wesley Avenue, Denver, CO 80208; [email protected]
Jeffrey L. HopkinsHopkins Phoenix Observatory, 7812 West Clayton Drive, Phoenix, AZ 85033; [email protected]
Robert E. StencelUniversity of Denver, Department of Physics and Astronomy, 2112 E. Wesley Avenue, Denver, CO 80208;[email protected]
Received June 1, 2012; revised October 4, 2012; accepted October 4, 2012
Abstract
Thelureofa50%reductioninlighthasbroughtamultitudeofobserversandresearcherstoeAureverytwenty-sevenyears,butfewhavepaidattentiontothesystemoutsideofeclipse.Asearlyasthelate1800s,itwasclearthatthesystemundergoessomeformofquasi-periodicvariationoutsideoftotality,butfewconsideredthiseffectintheirresearchuntilthemid-1950s.Inthisworkwefocusexclusivelyontheout-of-eclipse(OOE)variationsseeninthissystem.Wehavedigitizedtwenty-sevensourcesofhistoricphotometryfromeighty-onedifferentobservers.Twoofthesesourcesprovidetwenty-sevenyearsofinter-eclipseUBVphotometrywhichwehaveanalyzedusingmodernperiodfindingtechniques.WehavediscoveredtheF-starvariationsaremulti-periodicwithatleasttwoperiodsthatevolveintimeatDP ≈ –1.5 day/year. These periods are detectedwhentheymanifestasnear-sinusoidalvariationsat3,200-dayintervals.We discuss our work in an evolutionary context by comparing the behaviorfoundineAurwithbona-fidesupergiantandpost-AGBstarsofsimilarspectraltype.Baseduponourqualitativecomparison,wefindthephotometricbehaviorof the F-star in the eAur system is more indicative of supergiant behavior.Thereforethestarismorelikelytobea“traditionalsupergiant’’thanapost-AGBobject.Weencouragecontinuedphotometricmonitoringofthissystemtotestourpredictions.
1. Introduction
eAurigaeisa27.1-yearsinglelineeclipsingspectroscopicbinarythathaslong been wrapped in an enigma. Ever since a dimming of the system was
Kloppenborgetal., JAAVSO Volume 40, 2012648
discoveredbyFritsch(1824),anditsperiodicityestablished(Ludendorff1903),astronomershavespeculatedaboutthecauseofthisvariation.Baseduponradialvelocitymeasurements,eAurwasclassifiedasaspectroscopicbinary(Vogel1903),whichhinted that thedimmingmightbeaneclipse.LaterapplicationofHenryNorrisRussel’sbinarystartheory(Russell1912a,1912b)cametoastrikingrevelation:thecompaniontotheF-typesupergiantwasnearlyequalinmass,yetspectroscopicallyinvisible(Shapley1915). Overthenextdecadeseveraltheorieswereadvancedtoexplainthisstartlingconclusion (for example, Ludendorff 1912; Kuiper et al. 1937; Schoenbergand Jung1938;Kopal1954); however, itwasHuang (1965)whoproposedthat the eclipse was caused by a disk of opaque material that enshroudedthe secondary component.Although Huang’s analytic model replicated theeclipselightcurve,thedisktheoryremainedunprovenuntilBackman(1985)detectedan infraredexcess that corresponded toa500Kblackbody source.Recently,thedisktheorywasvindicatedbyinterferometricobservationsoftheeclipse.TheseimagesshowtheF-starispartiallyobscuredbyadiskofopaquematerial(Kloppenborget al.2010,2011)whichisresponsibleforthedimmingobservedphotometrically. Althoughmostresearchonthissystemconcentratedontheeclipseitself,afewworkshavelookedatthesystemoutsideofeclipse.Sincetheearly1900s,ithasbeenknownthattheF-starexhibits0.1-magnitudevariationsinV-bandoutsideofeclipse.Indeed,theearliestdiscussionoftheseout-of-eclipse(OOE)variationswerebyShapley(1915,p.20).HecommentsonvariationsinvisualphotometrywithanamplitudeofDVis=0.3magnitude.Becauseobservationalerrorswerenotfullycharacterized,Shapleytreatedtheseresultswithcaution.Later, Güssow (1928) spotted aDVis = 0.15-magnitude variation outside ofeclipse that was corroborated by two photoelectric photometers (Shapley1928),therebyconfirmingthepresenceoftheOOEvariations.Shapley(1928)concludedthatthesevariationsarosefroma~355-dayquasi-periodicvariation;however,theexactperiodwaspoorlyconstrainedbythesedata. Afterthe1983–1985eclipse,Kempet al.(1986)proposedthata~100-dayperiodmayexistinpolarimetrydata.Later,Henson(1989)showedtherewaslittletonowavelengthdependenceinthevariations,implyingthatthesourceofpolarizationisThompsonscatteringfromfreeelectrons.Inhisdissertation,HensonfoundintervalswheretherewerevariationsinStokesQ,butlittletonothinginStokesU.ThiswasinterpretedtobecausedbytheF-starhavingtwomajoraxesforpolarization,inclinedatanangleof45degreeswithrespecttoeachother.Likemanyoftheotherstudies,avisualinspectionofHenson’sdatashowedthatsomelongtrendsmayindeedexist,butnothingwasstrictlyperiodic.From the post-eclipse polarimetry, Henson concluded that the photometricand polarimetric variations might be caused by non-radial pulsation in low-order ℓ=1,2m=±1modes.ThisnotionissupportedbytherecentautomatedclassificationofeAurasanaCygvariable(Dubathet al.2011a).
Kloppenborgetal., JAAVSO Volume 40, 2012 649
After the1985eclipse,manyauthors sought todetermineperiodsof theOOEvariation.Usingdatafromthefirstfiveyearsaftertheeclipse,Nhaet al.(1993)foundoccasionalstablevariationalpatternswouldsetin(inparticulararoundJD2447085–2447163)withDU=0.27,DB=0.17,andDV=0.08,andacharacteristicperiodof95.5days.Later,HopkinsandStencel(2008)analyzedtheirinter-eclipseV-bandphotometricdatausingtheperansosoftwarepackage(Husar2006).They found twodominantpeaks in theFourierpower spectrawith65-and90-dayperiods. PerhapsthemostcomprehensiveperiodanalysiseffortwasmadebyKim(2008).TheyusedtheCLEANestFouriertransformalgorithm(Foster1995)andtheWeightedWaveletZ-transform(WWZ;Foster1996)onnearly160yearsofphotometryofeAur.UsingthesetwoalgorithmstheyidentifiedseveralperiodswhichledthemtoconcludeeAurmaybeadoubleormulti-periodicpulsator. Hereweextendtheworkofourpredecessorsbyanalyzingtwenty-seven-years worth of inter-eclipse UBV photometry. In the following sections wediscuss our sources of data, our analysis methods, and the results. We thenconcludewithadiscussionofourresultsinastellarevolutioncontext.
2. Data sources
2.1.Historicphotometry e Aur has a rich history of photometric observations. We conducted acomprehensiveliteraturereviewandfoundtwenty-sevensourcesofphotometryfromeighty-onedifferentobservers.WedigitizedallofthesedataandintendtosubmitthemtotheAAVSOInternationalDatabaseorVizieR(whencopyrightallows)afterthepublicationofthisarticle.Afulldiscussionofthedatasources,assumed uncertainties, and digitization methods is in Kloppenborg (2012);however,wehavesummarizedtheimportantaspectsofthesedatainTable1.
2.2.Phoenix-10 The Phoenix-10 Automated Photoelectric telescope, designed by LouisBoyd,obtainedatotalof1,570U-,1,581B-,and1,595V-bandobservationsofeAurbetween1983and2005.Thesystemconsistedofa1P21photomultipliermountedona10-inchf /4Newtoniantelescope.AdetaileddescriptionofthissetupcanbefoundinBoydet al.(1984b).ThetelescopewasoriginallylocatedindowntownPhoenix,Arizona,untilitwasmovedtoMountHopkinsduringthe summer of 1986. The system was then moved to Washington Camp inPatagonia,Arizona,in1996whereitoperateduntil2005. TheearliestdataoneAurwereobtainedin1983Novemberandcoveredmostofthe1983–1984eclipse(Boydet al.1984a).ThesedatacoverintervalsJD2445646–2455699(Breger1982,file131),JD2445701–2445785(Breger1985, file 136), and JD 2445792–2445972 (Breger 1988, file 137) (1983November 3–1984 September 29). Although the photometer collected data
Kloppenborgetal., JAAVSO Volume 40, 2012650
between1984Septemberand1987September,theoriginaldatawerelostduetoahardwarefailure(Boyd2010).Asfaraswehavebeenabletodetermine,these data were not published in subsequent issues of the IAU Archives of Unpublished Observations of Variable Stars. In addition to these data, ourpublication includes unpublished data from theAPT-10 which starts on JD2447066(1987September27)andendson2453457(2005March27). The standard observing sequence for e Aur was KSCVCVCVCSK(K=Check,S=Sky,C=Comparison,V=Variable)iteratingthroughtheUBVfilters(seeBoydet al.1984b,Table1)with10-secondintegrations.Informationonthetarget,check,andcomparisonstarsaresummarizedinTable2.Thesedifferentialmeasurementswerecorrectedforextinctionandtransformedintothe standard UBV system. The automated reduction pipeline discarded anyobservationswithaninternalstandarderrorof±20milli-magnitudesorgreater.Typicalexternalerrorsare±0.011,±0.014,±0.023magnitudeinV,B,andUfilters,respectively,withmeaninternalerrorsof±0.005,±0.005,and±0.009(StrassmeierandHall1988).Thestabilityofthesystemhasbeensatisfactoryondecade-longtimescales(HallandHenry1992;Hallet al.1986).
2.3.HopkinsUBV Co-authorJeffreyHopkinscollected811U,815B,and993VdifferentialmagnitudesattheHopkinsPhoenixObservatory(HPO)inPhoenix,Arizona.Thedataconsistoftwolargeblocks:thefirstbeganon1982September09andendedon1988December23,andthesecondseriesstartedon2003December04andendedon2011April25.AtthistimethephotometricprogramatHPOended. TheHPOsetupconsistedofa1P21photomultipliermountedtoan8-inchCelestron C-8 telescope with standard Johnson UBV filters. Observationsof eAur were conducted in CSVSCSVSCSVSCS format, each composedof three10-second integrations inoneof the three filters.Nightlyextinctioncoefficientsweredeterminedandapplied.Colorcorrectionwasdeterminedonamonthlybasis.Afterthe1980observingseason,lAurigaewasusedasthesolecomparisonstar.TheassumedmagnitudesforlAurigaewereV=4.71,(B–V)=0.63,and(U–B)=0.12.AsubsetofthesedatahavebeendiscussedinChadimaet al.(2011).
2.4.AAVSOBrightStarMonitor Startingon2009October16,eAurwasplacedontheAmericanAssociationofVariableStarObservers’BrightStarMonitor(BSM)observingprogram.TheBSMconsistedof aTakahashiFS-60CBwith a field flattener; 60-mm f /6.2telescope and a SBIG ST-8XME camera with Johnson/Cousins BV Rc Ic andclearfilters.DatawerereducedbyArneHendenatAAVSOheadquarterstothestandardphotometricsystem.Allcolorandextinctioncorrectionswereappliedbefore data were submitted to theAAVSO International Database. OngoingobservationsfromthisinstrumentcanbefoundintheAAVSOdatabaseundertheobservercode“HQA.’’
Kloppenborgetal., JAAVSO Volume 40, 2012 651
2.5.Solarmassejectionimager Forthesakeofcompleteness,wealsomentionourworkondatafromtheSolar Mass Ejection Imager (SMEI; Simnett et al. 2003).Although SMEI’sprimarymission is tomap the large-scalevariations inheliospheric electrondensitiesbyobservingThompson-scatteredsunlight,italsocollectedprecisionphotometryon~20,000starswithV<8.Througheach102-minuteorbit,mostregions in the skywere coveredby a dozenormore frames throughoneofSMEI’sthreebaffled,unfilteredCCDcameras. The instrument was designed for 0.1% photometry and, when properphotometric extraction is performed, this precision was realized on starsbrighterthanfifthmagnitude.Onaveragetheuncertaintyonfainterstarswasproportionallyworsebytheratioofthestar’sbrightnesstofourthmagnitude,meaningfainteighthmagnitudestarsstillhavebetter-than-groundphotometryprecision.TheinstrumentwasoperationalfromlaunchuntilSeptember2011whenitwasdeactivatedduetobudgetconstraints.Ourworkwiththesedatawillbediscussedinafuturepublication.
3. Analysis
The sources of photometry listed above were very inhomogeneous;consisting of multiple filters, reduction methods, observatories, instruments,and even reference star magnitudes. We created a script that finds overlapsbetweentwodatasetsofthesamefilter,binsthedata,andthencalculatesthecoefficientsrequiredtoscale/offsetthedatatothesamequasi-systemusingaweightedleast-squarestechniqueusingthefollowingequation:
Ai = a + b Bj + c tj (1)
HereAiistheithentryinthereferencephotometryset,Bjisthejthentryinthecomparisonphotometrydatasetoccurringat time tj,a isazero-pointoffset,b accounts for non-Pogson magnitudes (present in only our earliest visualphotometricdata),andccorrectsforatime-dependentdriftofthecomparisonphotometryset. Inallof thefilteredphotometry,onlyawasrequired. In thevisual data, a and b were required.As our work here is concerned with thevariations,ratherthanabsolutelycalibratedphotometry,weselectedHopkinsU,BSMB,andBSMVasreferences.Offsets(a)betweenvariousphotometricsourcesaresummarizedinTable3. AftertheoffsetsweredeterminedwepassedthedatathroughtheWindowsimplementation of WWZ, winwwz. Here we have used data from the rangeJD2446000–2455000in10-daystepswithflow=0.004,fhigh=0.05,andDf=0.0001.OurWWZdecayconstantwas0.0125.
Kloppenborgetal., JAAVSO Volume 40, 2012652
4. Results
In Figure 1 we plot the twenty-seven-years worth of inter-eclipse UBVphotometry from theAPT-10andHPOobservatories.TheVdatahavebeenplotted unaltered, but U have been offset by –1.3 magnitude and B by –0.8magnitude.Internalphotometricuncertainties(~10milli-magnitudes)areaboutthe thickness of the lines. Each observing season consists of approximately200daysofdatawithbi-nightlysamplingfollowedby165-daygapswherethestarwasnotvisibleatnight.AllofthedatawerecorrectedforextinctionandtransformedtothestandardUBVphotometricsystem.Avisualinspectionofthedatarevealsnosingleconsistentperiodispresent. Historically,therehavebeenseveralreportedinstancesofshort-term(thatis,few-daylong)eventswhichwesuspectareflares.AlboandSorgsepp(1974)reportedaDU=0.2,DB=0.1,andDV=0.06brighteningthatlastedfivedaysaroundJD2439968 (1968April21).Similarly,NhaandLee (1983)notedarapid (few hour) 0.4 magnitude rise in the blue filter and 0.2 magnitude intheyellowfilteronJD2445356(1983January21).Wehavenoticeda two-nightflareinourdatasetstartingonJD2446736(1986November01).Thisevent resulted inaDU=0.2,DB=0.1, andDV=0.7photometric increase,strangelyoppositeofhistoricrecords.ContinuedUBVphotometryappearstobeaefficientwayofdetectingtheseevents. InFigure2weplotthecurrenteclipselightcurveinUBVRIJHfiltersfromtheabovesourcesandtheAAVSOdatabase.Wenotethattheeclipseappearsto be slightly wavelength-dependent, particularly from mid-eclipse to thirdcontact.ThiscanbeseenfromthedownwardslopeoftheU-banddataandflattrendinH-band.WesuggestthisisduetoadditionalsmallparticlescomingintothelineofsightfromasublimationzoneontheF-starfacingsideofthedisk. InFigures3through5weplottheWWZresultsfortheU,B,andVfilters,respectively.ThecolorinthesefiguresistheWWZoutputwithhighervalueindicatingstrongerpresenceofthatparticularperiod.BecausetheAPT-10wasnotoperationalduringtheintervalJD2449500–JD245000,weblockedouttheWWZresultinthisregion. Frominspectionofthesefigures,itisclearthatnosingleperiodcanaccuratelydescribe thevariations seen ineAur. InTable4we listpeakswhoseWWZcoefficientwasgreater than100. In theU-banddata the twomostdominantperiodsarecenteredat(8977,102.7)and(12259,87.9)(henceforththe“uppertrack’’).Thesepeakswerespacedapartby3,282dayswithDP=–14.8days.WhentheWWZoutputwashigh,theradialvelocity(seeStefaniket al.2010fordata)andphotometricchangeswereinphase.Thecombinationofthesetwoeffectsleadustobelievetheseeventsshouldberegardedassignificant. OperatingontheassumptionthatthesepeaksprovidedaglimpseofperiodevolutionintheF-star,weintentionallysoughtvariationsfollowingaparallelevolutionary path. Three peaks located at (7166, 90), (10492, 82.7), and
Kloppenborgetal., JAAVSO Volume 40, 2012 653
(13744,68.9)(hereafterthe“lowertrack’’)followedasimilarevolution.Likethe“uppertrack,’’thesepeakswereseparatedbynearly3,300days. PeaksattheselocationsalsoappearedintheB-andV-banddata,althoughwithmuchlowersignificance.TheWWZvalueinwinwwzwasdeterminedbyac2-likemetricthatdoesnotconsidertheuncertaintiesinthedata.Thereforethe additional scatter seen in adjacent B or V photometry, although withinuncertainties,resultsinalowerWWZvalue.Thehigheramplitudesandgreaternight-to-nightconsistencycauses theUdata tohave largerWWZvalues,onaverage,thantheBandVdata. IntheB-bandWWZ,weseeafewpeaksinthe80-to100-dayrangeappearingfromtimetotime,althoughtheyareclearlynotstable.Likewise,intheV-bandWWZthereisasingledominantpeakof(6100,90),butotherwiselittlehintofastablevariationalpattern.WedonotsuggesttheWWZ<100resultsshouldbegivenmuchconsideration;however,intherangeJD24450000–24455000therewereseveralcommensurateperiodsthatappearedtoevolvedownwardatarateofseveraldays/year. InTable5weusedtheaboveobservationsasaguideandpredicteddateswhenstablepulsationalpatternsshoulddevelopand theirperiods.Wenotethat the118-dayperiodaroundJD2445695(near thirdcontactof the1984eclipse)didnotmanifest;however,neartheendofthe2009–2011eclipseasawtooth-likepatternwitha61-to76-dayperioddeveloped(seeFigure2).Thisistantalizinglyclosetothe~73-dayperiodwepredictedwoulddevelopatthistime. To test our extrapolation, we attempted a WWZ analysis on the visualdata.Theonly set that spannedanentire inter-eclipse intervalwascollectedbyPlassman(Güssow1936).Datawithin the interval2422000–2428000(seeFigure6)clearlyshowthepresenceoftheOOEvariationswithcharacteristicperiodsof330–370days.Thisvaluewasnearly100dayslongerthanwhatwepredictforthistimeinterval,implyingourextrapolationshouldnotberegardedashighlypredictiveuntilfurtherperiodcharacterizationisconducted.
5. Discussion and conclusion
Wehavecreatedthefirstlong-termUBVphotometricrecordofeAurusingthe data from the APT-10 and HPO observatories. These data show stablevariationalpatternsdevelopedon3,200-daytimescales.IntheU-bandWWZoutput,wehaveidentifiedtwoparalleltracksofstablevariationsthatevolveatarateofDP=–1.6day/yearandDP=–1.2day/yearfor the“upper’’and“lower’’ track, respectively. Extrapolating these results, we have identifieddates at which we anticipate stable variational patterns will manifest andpredictedtheperiodsthattheyshouldhave.OurextrapolationtoJD2455541predictsa73-dayperiodshouldhavedeveloped;itistantalizinglyclosetothe61- to76-dayperiod thatwasobservedduring the second-halfof the2009–
Kloppenborgetal., JAAVSO Volume 40, 2012654
2011eclipse.Ourinterpretationbelowisbasedonthis“two-track’’notionandlikelyunderestimatesthetruecomplexityinthissystem.InTable4weprovideallpeakswithaWWZ>100fortheUBVphotometryinhopesthattheywillbeusefultofutureresearchers. At the time of this writing, a consistent asteroseismic interpretation forevolvedsupergiant-classstarsdoesnotexistduetotheuncertaintiesunderlyingthetheoreticalcalculationsofmixingtheoryandradiationpressure(Aertset al.2010,ch.2).Therefore,wecannotprovidearigorous,quantitativeinterpretationof the periods which we have observed. Instead, we interpret the observedperiodsqualitativelybycomparing themwithobserved supergiant andpost-AGBbehavior.Regrettably,fewcomparativestudiesofF-typesupergiant/post-AGBstarsexist,especiallymulti-decadesurveys.Thereforeourinterpretationsareinherentlybiased.Wehaveattemptedtodiscussthesebiasesthroughlyandindicatewhereourstudycouldbenefitfromfutureresearch. It had been known for some time that stars near the F0Ia spectral andluminosity class show low-level variations with 0.015–0.025 magnitudeamplitudes in theV-band (Maeder 1980).An investigation by van Leeuwenet al.(1998)oftwenty-foursuper-andhyper-giantstarsfromtheLMC,SMC,andtheMilkyWayusingHIPPARCOSphotometryshowedthatalloftheseBto late-Gstarsexhibitedphotometricchanges thatwerenot strictlyperiodic.Indeed, many of these “periods’’ would be better described as “quasi-’’or “pseudo-periods.’’Across this region of the HR diagram, stars tended toshowvariationson10-to100-daytimescales. Indeed,inthisrespecteAurcouldeasilyberegardedasarecently-evolvedsupergiant.SeveralstarsinthevanLeeuwenet al.(1998)samplewereaclosematch for eAur:At a slightly higher temperature, HD 269541 (HIP 25448,A8:Ia+, in the LMC) hasDVT = 0.1-magnitude variations with several shortperiods in the8- to 40-day interval, and two longer periods at 146 and182days. The slightly cooler HD 269697 (HIP 25892, F5Ia, in the LMC) hadtwoequallysignificantperiodsat48and84dayswithphotometricvariationsbetween0.01and0.05magnitudeinDVT.HD74180(HIP42570,F2Ia,intheMilkyWay)wastheclosestmatchtoeAurinthevanLeeuwenet al.(1998)sample.Thisstarshowedquasi-periodsat53,80,and160dayswithvariationsof0.06magnitude.ItisalsoworthmentioningthatanautomatedphotometricclassificationschemeconsideredeAurtobeanaCygvariable(Dubathet al.2011),aclassofluminoussupergiantsundergoingnon-radialpulsation. Post-AGBstarsmakeupaveryheterogeneousclassofobjects,thereforeit is difficult to discuss their properties, let alone discuss any reasons formembership (or lack thereof) for one particular star. We have searched the“ToruncatalogueofGalacticpost-AGBandrelatedobjects’’(Szczerbaet al.2007)forsystemsofsimilarspectraltypeandidenticalluminositiestoeAur.Thebest-matchisARPup(F0Iab,HIP39376)whichfeaturesmulti-periodic(RVb)pulsationwithDV=0.5andtimescalesof76.4±4-and1,250±300-
Kloppenborgetal., JAAVSO Volume 40, 2012 655
dayperiods(Kisset al.2007).Althoughthetimescalesmatch,thevariationalpattern (that is, highly predictable, stable) does not match what is seen ineAur.LikewisethepatternseeninthecoolerV340Ser(HD158616,F8)hassimilar timescales (87.7dand131d,Arkhipovaet al.2011),but isobviouslymulti-periodicandeasilypredicted.Tocompleteourviewofvariationsseenaround the F0Ia class, long-term photometric studies of the post-AGB starsIRAS10197-5750(A2Iab:e,2MASSJ10213385-5805476),IRAS16206-5956(A3Iab:e, 2MASS J16250261-6003323), IRAS 06530-0213 (F0Iab, 2MASSJ06553181-0217283),HD101584(F0Iabpe,HIP56992),andHD187885(F2/F3Iab,IRAS19500-1709)wouldbebeneficial. Aggregate statisticsofpost-AGBstars imply systemsof similar spectraltypes to e Aur have significantly shorter periods than their supergiantcounterparts.Forexample,Hrivnaket al.(2010)studiedaseriesofC-richpost-AGBstarsandfoundastrongcorrelationbetweentheeffectivetemperatureandperiod.Therelationshippredictsthathigher-temperaturepost-AGBswillhaveshorterperiodsfollowingalineartrend:DP/DTeff=–0.047dayK1.TheirFigure18suggestedeAurshouldexhibitvariationswitha~40-daytimescale,afactorof1.7 less thanwhatwehaveobserved.TheirworkonO-richstarsappearstobeforthcoming(seeShawet al.2011).Inasampleoffivepost-AGBstarsArkhipovaet al.(2011)foundasimilartrend.TheirFigure8predictsperiodsof~65days,afactorof1.25to1.5shorterthanwhatwehaveobserved.Amajorityof theirprogramstarsalsoweremultiperiodic,with ratiosofP1/P2~1.03 to1.09,whereaseAurshowsahigherratioof1.24to1.27. UntilthispointwehavecomparedthevariationalpatternsineAuragainstsinglestars.Asnotedabove,thestablevariationpatternsdevelopat3,200-dayintervals, which is nearly 1/3 of the 27.1-year (9,890-day) orbital period. Itwould appear the companion is influencing the pulsational properties of theF-star.As theorbit is eccentric (e ~0.227ore ~0.249–0.256,Stefaniket al.2010,Chadimaet al. 2010, respectively)onemight anticipate tidal flows tobe induced in the F-star’s tenuously-bound atmosphere (log g ~1, Sadakaneet al.2010)duringperiastronpassage;however,dissipationtimescaleswouldcertainly be less than the nine-year interval seen between stable variationalpatterns(seeMorenoet al.2011,andreferencesthereinforadiscussionofthetheoreticalframework).InsteadwespeculatethatgravitationalforcingduetoorbitalmotionisexcitingnaturalresonantfrequenciesintheF-star(thistheoryisshowntobepossibleinmainsequenceobjects—seeGoldreichandNicholson1989;Rocca1989;WitteandSavonije1999a,1999b;Zahn1975,1977).Thisconjecturepredictsthattheexcitationsshouldrepeatatthesameorbitalphasesandisthereforetestablebycontinuedphotometricmonitoring.Thenextdateswhen sucheventsmighthappenare JD~2457000 (2014December) and JD~2457000 (2019 December).The development of a consistent asteroseismictheoryforsupergiantsmayprovideanearliertestofourhypothesis. Given this information and the photometric behavior discussed above,
Kloppenborgetal., JAAVSO Volume 40, 2012656
we consider it unlikely that the F-star is a post-AGB object and conclude,on a qualitative basis, that the F-star is more likely a traditional supergiant.Implicationsfortheevolutionarystateandphysicalpropertiesofthediskwillbediscussedinforthcomingpublications.
6. Acknowledgements
ParticipantsfromtheUniversityofDenveraregratefulforthebequestofWilliamHershelWombleinsupportofastronomyattheUniversityofDenver.They acknowledge support from National Science Foundation through ISEgrantDRL-0840188(CitizenSky)totheAmericanAssociationofVariableStarObserversandASTgrant10-16678totheUniversityofDenver.Weacknowledgewith thanks the variable star observations from the AAVSO InternationalDatabasecontributedbyobserversworldwideandusedinthisresearch.Thisresearch has made use of NASA’sAstrophysics Data System BibliographicServicesandtheSIMBADdatabase,operatedatCDS,Strasbourg,France.
Var. Stars,No.2371,1.Kemp,J.C.,Henson,G.D.,Kraus,D.J.,Beardsley,I.S.,Carroll,L.C.,Ake,T.
B.,Simon,T.,andCollins,G.W.1986,Astrophys. J., Lett. Ed.,300,11.Kim,H.2008,J. Astron. Space Sci.,25,1.Kiss,L.L.,Derekas,A.,Szabó,G.M.,Bedding,T.R.,andSzabados,L.2007,
Table4.PeakperiodsobservedintheUBVWWZtransformsroughlygroupedbydate.Dateshavebeenroundedtothenearest10.PeriodsandWWZoutputareroundedtointegervalues.MJD=JD–2440000. U B V MJD Period WWZ MJD Period WWZ MJD Period WWZ
Table 5. Predicted dates of stable pulsation features and their periods baseduponextrapolationoftrendsdiscussedinsection4. Upper Track Lower Track JD Period Observed? JD Period Observed?
Table4.PeakperiodsobservedintheUBVWWZtransformsroughlygroupedbydate.Dateshavebeenroundedtothenearest10.PeriodsandWWZoutputareroundedtointegervalues.MJD=JD–2440000,cont. U B V MJD Period WWZ MJD Period WWZ MJD Period WWZ
Figure 2. 2009–2011 eclipse of e Aur in UBVRIJH filters, JD 2454800–2456000(2008November29–2012March13), asmeasuredbyHopkins (UBV), theAAVSOBSM(BVRI),andAAVSOobserversBrianMcCandlessandThomasRutherford(JHdata).TheV-banddata areplotted asobserved, all other filtershavebeenoffset byanarbitraryamountfordisplaypurposes.Theeclipsemayberepresentedbyalineardecreaseinbrightnessof~0.7magnitude,followedbyaflatminimumandthenasharprisebacktoout-of-eclipsebrightness.Theout-of-eclipsevariationsaresuperimposedonthisprofileandresultin60-to100-daycycleswithcharacteristicamplitudesof~0.1magnitudeinU,decreasinginamplitudetowardslongerwavelengths.Noticeduringthesecondhalfof theeclipse theU-band lightcurveslopesdownward,whereas theH-bandhasanupwardslope.Thisatteststothefactthattheeclipsehadwavelength-dependentextinction.
Figure3.U-bandWWZperiod analysisofeAur (c=1.25E–2).The color indicatespowerassociatedwithagivenperiodataparticulartime,whereasthewhitedottracesthe dominant period at a particular time.The region JD2449500–2450000has beenblockedoutduetoalackofdata.Wecautionthereaderthatperiodswithin<200daysofthisregionmaynotbetrustworthy.Seesection4foradiscussionoftheperiodspresentandourinterpretations.
Figure6.VisualphotometryofeAurbyPlassman(inGüssow1936)showingtheintervalJD 2422000–2428000 (1919 February 10–1935 July 16) including the 1927 eclipse.Uncertainties,notplotted,are±1intheleastsignificantdigit(±0.01magnitude).TheOOEvariationsareclearlypresent.Typicalpeak-to-peak timesaredifficult to judge,exceptrightbeforethe1927eclipse,where330-to370-dayperiodsappeartobepresent.
Karlsson, JAAVSO Volume 40, 2012668
V-band Light Curve Analysis of e Aurigae During the 2009–2011 Eclipse
Thomas KarlssonAlmers väg 19, 432 51 Varberg, Sweden; [email protected]
Received August 19, 2011; revised October 28, 2011, November 9, 2011; accepted November 14, 2011
Abstract Timings for the eclipse contactpoints andmid-eclipse, lengthofingress and egress, average magnitude during eclipse, and timings for out-of-eclipsevariationshavebeendeterminedintheVbandforthelongperiodeclipsingbinaryeAurigaeduringthe2009–2011eclipse.ThishasbeendonewithdatafromtheInternationalEpsilonAurigaeCampaign2009andAAVSO.Comparisonwithdatafrompreviouseclipseshasalsobeenmade.
1. Introduction
Theextremelylong-periodeclipsingbinaryeAurigae(period=~27.1years)is stillnot fullyunderstood. Ithasbeenstudiedbymanygroups indifferentwavelengths photometrically, spectroscopically, and interferometrically. The~2-yeareclipsethatoccurred2009–2011presentedanopportunitytoconstrainsomeparametersforthesystem.Asthesystemisbright(V=2.9–3.8)andthedurationoftheeclipseislong,it isasuitabletargetforamateurastronomerswhocancommitalongperiodoftimeforthistypeofproject.Aninternationalcampaign with participants of both advanced amateur and professionalastronomerswasestablished(seethecampaignwebsiteathttp://www.hposoft.com/Campaign09.htmlforfurtherdetails.)PhotometricdatafromthiscampaigntogetherwithtwocontributorsfromtheAAVSO,whohavecoveredthewholeeclipse,arethebasisforthisanalysis.ThecampaignproduceddataintheU,B,V,Rc,Ic,J,andHbandsbutthisanalysisonlycoversthemoststudiedVband.
2. Method
Theobserversinthecampaignusedadiversityofequipmentandreductionmethods(Table1),fromphotometersmountedontelescopes,toCCD-cameraswithtelescopes,orcamera-lensesandstandarddigitalsingle-lensreflexcamerasonatripod.Thecampaignhaspublishedrecommendedreductionmethodsandcomparison stars to use. But the diversity in equipment and individuality inphotometric software and methods used have introduced some differencesamongtheobservers.Tomaketheobservationscomparabletoeachotherthefollowingmethodswereused. Priortotheanalysisallobservationsweredividedintogroupsoffour-day
Karlsson, JAAVSO Volume 40, 2012 669
periods.IneachgroupthemeanofmagnitudeandJDwerecalculated.Eachobserver’sindividualobservationswerethensubtractedfromthecorrespondingmeans,andthestandarddeviationofeachobserver’sdifferenceswascalculated.Observerswithtoomuchspreadintheirdata(SD>0.04)werethenexcludedaltogether,andsomeoutlying(>0.08fromthemean)individualpointsfromtheremainingobserverswerealsoremoved.Newfour-daymeansanddifferenceswerethencalculatedfortheremainingobservers.Fromthesenewdifferencesan offset for each observer was calculated, showing the difference for eachobserver’sdatasetand themeanforalldata.Thisoffsetwas thensubtractedfromeachobserver’sdataandnewone-andfour-daymeanswerecalculated.Thesecorrectedmeanswerethenusedinthefurtheranalysis.TheindividualobservationswiththeoffsetappliedareseeninFigure1,the4-daymeansareinFigure2andthe1-daymeansinFigure3. The offset is based on the assumption that each observer used a similarreductiontechniqueduringtheeclipse,butthechoiceofreductiontechniquesandcomparisonstarsdiffersamongtheobservers,sothateachdatasetistoahigherdegreeconsistentwithitselfthanwiththeotherdatasets.Theideaisthatbyusingthismethodofreconcilingthedatamorefinedetailsinthelightcurveshouldbeseen.
3. Results
3.1.Eclipsetimingsandmagnitude Thefourcontactpointsforatotaleclipsebetweentwosphericalbodiesaredefinedasbeginningofeclipse,beginningoftotality,endoftotality,andendofeclipse.InthecaseofeAurigaethemainstarispartiallyeclipsedbywhatissupposedtobeadustydiscseenalmostedgeon(Huang1965;Kloppenborget al.2011),givingrisetoanelongatedandellipticaleclipsingbodyfromourview.Contact2and3havethereforeforthissystemanambiguousdefinitionandnotthesamephysicalmeaningasforaclassicaleclipsingbinary.Forthissystemtheycouldbeinterpretedas thepointswhere theleadingedgeof theelongateddischascrossedthewholefaceofthemainstarandwherethetrailingedgeofthediscbeginstoleavethefaceofthemainstar. Thecontactpoints(Table2)wereestimatedusingalineartrendlineappliedtotheingress/egressfromFigure2andweredecidedbywherethelinecrossedthe out-of-eclipse mean magnitude of 3.035 (Hopkins 2011) and the meanduringtheeclipseof3.728.Somefurtherworkmaybedonetoproduceamoreprecisemodelofthecurveandthemeansbeforeandafterthecontactpointstoobtainmoreaccuratetimesforthecontacts.Atcontact2thecurveisespeciallysmooth,whichmakes thecontactpointhardtodefine.Contact3couldhaveoccurredaboutaweek later than the trend linesuggestsbecauseof theverysteepbeginningoftheegress.Figures4and5showthegraphsusedtoestablishthecontactpoints.
Karlsson, JAAVSO Volume 40, 2012670
The totality phase after mid-eclipse is 0.027 magnitude brighter thanbeforemid-eclipse.Thebrightestpartofthetotalityisduringmid-eclipse,andthedimmest just beforemid-eclipse.Onaverage, thedimmestpart is in thebeginningandendoftotalityandthebrightestinbetween:
3.2.Out-of-eclipse(OOE)variations Besides the eclipse, the system shows a smaller variation of 0.1 to 0.2magnitudewithanirregularperiodof~2months(see,forexample,HopkinsandStencel2006forarecentstudybeforestartofthe2009eclipse).Theout-of-eclipsevariationsduringtotalitywerecalculatedbyapplyingafourth-orderpolynomialfittothedatapoints24and27daysaroundeachmaximaandminimausingthe1-dayaverages fromFigure3.ThemeantimesandmagnitudesfromthetwosetsareshowninTable3.Theerrorforthespecifieddatesisestimatedtobeonaveragewithin±2daysandthemagnitudewithin±0.01. Amplitude iscalculatedas thedifferencebetween themaximumand themeanofthetwoadjacentminima.Thefirstminimumcouldbeaffectedbytheingress.Theobservationsmade3–4weeksbeforeandaftersolarconjunction
3.3.Ingressandegresscharacteristics DuringingresstherearehintsoftwoOOEvariationswithmaximaaboutJD2455080(2009September05)andJD2455160(2009November24),whichcanbeseeninFigure4inthepartsoftheingresscurvethatlieabovethelineartrend.AnotherOOEvariationjustattheendofingresswithamaximumaboutJD2455215(2010January18)isalsoprobableandcanbeseenasapartwithlowslopeattheendofingress. The egress started with a high rate of change in magnitude, during JD2455620to2455655(2011February27–2011April03).Thechangewas0.0089magnitude/day, the highest that was seen during the whole eclipse. FurtheranalysismustbedonetotellifarisingOOEvariationinteractedtogeneratethishighpace.AsthelastOOEvariationjustbeforeegresswasstrangelyshortand lowit ishard todecideby just lookingat the lightcurve ifanewOOEvariationoccurredduring thisperiod.DuringJD2455655 to2455670 (2011April 03–2011April 18) the slopewas lower, at 0.0051magnitude/day, andthentherewasastrangestandstillorslightdecreaseofbrightnessforabout15daysuntilJD2455685(2011May03).Afterthattheegresswentonatasteadyrateof0.0055magnitude/day,whichisaboutthesameasthemeanduringthewholeegress.Thefluctuationsthatcanbeseenat theveryendofegressareprobablycausedbythedifficultobservationconditionsduringthattimenearsolarconjunction.ThebigchangeofslopearoundJD2455655and the laterstandstillseemtoobigtobecausedbyanyOOEvariation.
3.4Comparisonwithpreviouseclipses InFigure6thereisacombinedlightcurvewiththe4-dayaveragedatafromFigure2togetherwithtwoprominentdatasetsfromtheprevioustwoeclipses,observationsbyGunnarLarsson-Leander1956–1957(Larsson-Leander1959)and Stig Ingvarsson 1982–1984 (Schmidke 1985). The elements from theGeneral Catalogue of Variable Stars(GCVS4;Kholopovet al.1984),epochJD2435629andperiod9,892days,wereusedtophasethedata.Table4showsdatafrompreviouseclipsestogetherwiththedatafromthispaper. Periodanalysesweremadewiththelightcurveandperiodanalysissoftwareperanso(Vanmunster2007)andtheANOVA(analysisofvariance)method.Theperiodscalculated(Table5)are2–5dayslongerthantheperiodfromGCVS4.
Karlsson, JAAVSO Volume 40, 2012672
The trend of decreasing duration of eclipse and egress and increasingduration of totality that was seen during the three or four previous eclipsesisbrokenbythelasteclipse.Infact,the2009–2011eclipseresemblesthatof1955–1957morethanthatof1982–1984,withthemoresimilarlengthofthedifferentphases,thedeepminimumjustbeforemid-eclipse,andthekneehalfwayduringtheegress.In1984thekneewasnotvisibleuntilthesystemhadreachedmagnitude3.30–3.25andtheobservationseasonwasalmostover.Thelackofobservationsattheendofthe1984eclipseisprobablythecauseoftheveryshortegressstated,calculatedfromtheslopeseenatthefirstphaseofegress. Differencesarethedeepminimum,downto3.85,thatwasseenattheveryendoftotalityin1956butnotseeninthetwofollowingeclipses.ThefrequentOOEvariationsduringtotalityseenduring2010seemtonotoccurtothesameextenteither in1956or1983.Maybeitcanbepartlyexplainedbythemoredetailedobservationsdoneduringthelasteclipse. Thepronouncedmid-eclipsebrighteningthatwasevidentin1983wasnotseentoanyhigherdegreein2010.Althoughthebrightestpointduringthetotalityof 2010 occurred at mid-eclipse, it was only 0.02–0.04 magnitude brighterthanthetwosubsequentout-of-eclipsevariations.Itshouldalsobestatedthatthemid-eclipseof1984coincidedwiththetoughestobservationconditionsatsolarconjunction,andnoobservationsweremadeatthetimeofmid-eclipse.Ifacarefulcorrectionforextinctionisnotdoneduringthisperiod,onecouldeasilyobtainvaluestoobrightforeasthemostusedcomparisonstars,lAur,hAur,andHR1644,allliesouthofe.ThisappearancewasseenamongseveralobserversduringMay–June2010. InFigure6onecanalsonoticetheplacementofthehumpsfromtheOOE-variationsduringthetotalityphasebetweenthethreeeclipses.Formostparttheyarenotinphasebetweenallthreeeclipses,withtheexceptionofabrighterphaseatmid-eclipse.ItcontradictstheideafromFerluga(1990),thattheOOE-variationsarecausedbyring-likestructureswithCassini-likedivisionsintheobscuringdiscasitpassesthemainstar.Ifsucharingstructureisstableoneshouldexpectthatthehumpswouldbeinphasebetweentheeclipses.
4. Conclusions
Thefollowingconclusionscanbedrawnfromthisstudy: TheOOEvariationsoccurwiththesameamplitudeandperiodicityduringthe eclipse as the period before eclipse and make it much harder to see thefeaturesoftheeclipseitself. Ingressandegresshavedifferentlengths.Inthisstudyegressisabout20daysshorterthaningress.Comparedtopreviouseclipsesthisrelationhasvariedagreatdeal.Thelengthofegress,especially,hasfluctuatedalot. Theegresshasakneehalfwayswheretheslopechangesabruptly,achangethatis toobigtobeexplainedbyOOEvariations.Furtheranalysishastobe
Karlsson, JAAVSO Volume 40, 2012 673
donetoseeifthegeometryoftheeclipsingdisccanexplaintheshapeoftheegresslightcurveorifsomeotherprocessisinvolved. If the eclipsing body is an homogeneous elliptical disc, then in purelygeometricaltermsthebiggestlossoflightbyapartialeclipseshouldoccurhalfways,andthelightcurveduringtotalityshouldbeslightlyconvex.Instead,theaveragelightcurveisslightlyconcaveduringtotality.Thismeansthatanothermechanismmaybeinvolvedtoexplaintheshapeofthecurve,forexample,anopticallythinnercenterofthediscorsomesortofscatteringeffect. There is also a difference in mean magnitude during the first half oftotality compared to the second half that could be a real feature if OOEvariationsareomitted.
Aurigae’sFirstEclipseoftheMillennium,”inThe Society for Astronomical Sciences 27th Annual Symposium on Telescope Science,heldMay20–22,2008atBigBearLake,California,Soc.Astron.Sci.,RanchoCucamonga,CA,67.
Hopkins,J.L.,andStencel,R.E.2006,“SingleChannelUBVandJHBandPhotometryofEpsilonAurigae,”inThe Society for Astronomical Sciences 25th Annual Symposium on Telescope Science,heldMay23–25,2006,atBigBear,California,Soc.Astron.Sci.,RanchoCucamonga,CA,13.
Huang,S.-S.1965,Astrophys. J.,141,976.Kholopov,P.N.,et al.1985,General Catalogue of Variable Stars,4thed.,Moscow.Kloppenborg,B.,et al.2011,posterattheSeattleAASmeeting11Jan2011,
“InterferometricImagesoftheTransitingDiskintheepsilonAurigaeSystem.”Larsson-Leander,G.1959,Arkiv Astron.,2,283.Schmidke, P. C. 1985, “UBV photometry of the 1982–4 eclipse of Epsilon
Grigoris MaraveliasPhysics Department, University of Crete, GR-71003, Heraklion, Crete, Greece; address email correspondence to G. Maravelias at [email protected]
Emmanuel (Manos) KardasisDepartment of Electronics Engineering, T.E.I. of Pireaus, GR-12244, Egaleo, Greece
Maria KoutoulakiPhysics Department, University of Crete, GR-71003, Heraklion, Crete, Greece
Received May 30, 2012; revised August 27, 2012; accepted September 5, 2012
Abstract WereporttheresultsoftheGreekcampaigntoobservethe2009–2011eclipseofeAurigae.WepresenttheactivitiesorganizedbytheHellenicAmateurAstronomyAssociation(HAAA)inordertopublicizetheeventandtoprovidethenecessaryinformationandtoolstobothfirst-timeandexperiencedobservers.Althoughvisualobservationswere thecoremethod,weproposedand experimented with various techniques. In total, data from 21 observerswere acquired combining different techniques: 302 visual, 95 CCD, and 11DSLRobservations,and5low-resolutionspectra. We were able to construct the light curve of the eclipse and extractsome interesting results, in agreement with previous studies. The system’sV-magnitude drops from ~3.0 to ~3.8 in 131±21 days.The ingress date isestimatedaround theMJD55087±15d (August12,2009)and thesystem isexitingeclipseaftertheMJD55797±15d(August23,2011).Weestimatethedurationofthe2009–2011eclipsetobe710±21days.Aratherpossibletrendformid-eclipsebrighteningexistsonly in theCCD/DSLRdata,which showoscillationsof0.07magnitudeamplitude.
1. Introduction
Thebright(~3V-magnitude)binarysystemeAurigae(eAur,R.A.05h01m58.1s,Dec.+43°49'23.9'',J2000)hasbeenalongmysterysinceitsdiscoveryin early 1800s. The binary consists of a F0 Iab star and a cool, mysterious
Maraveliasetal., JAAVSO Volume 40, 2012680
companionwhicheclipsesthesupergiantevery27.1yearsforalmost2years.Althoughthesystemhasbeenusedextensivelyasatestbedformanytheoriesandobservationalmethods(Carrollet al.1991;GuinananddeWarf2002,andreferencestherein)thedataobtainedsofar,whichextendalmosttwocenturies,havenotbeenadequatetofullyexplainthissystem.Evenafterthelasteclipseof 1982–1984, when an international campaign was launched, the mass andluminosityuncertaintiesremainedstrong,prohibitingthefinalsolution(Stencel1985).Butallthesedataprovideimportantconstraintsallowingtwopossiblescenarios:(i)thehigh-massmodel,wheretheF0starisconsideredayoungstarofmass~15M
interprettheobservations,and,moreover,newdataraisenewquestions(Hoardet al.2010;Kloppenborget al.2010). In2009–2010thesystemwouldundergoanothereclipse,offeringagreatopportunitytoacquiremoreinformation.Thus,anotherinternationalcampaignhasbeenlaunchedinordertocoordinate,collect,andprovideallthenecessarymaterialtoperformscientificusefulobservationsfrombothprofessionalandamateur astronomers: the International e Aur Campaign 2009–2011 (http://www.hposoft.com/Campaign09.html) organized by Jeff Hopkins and RobertStencel,andtheCitizenSkyproject(http://www.citizensky.org)organizedbytheAmericanAssociation of Variable Star Observers (AAVSO; http://www.aavso.org). The contribution of amateur astronomers is considered valuablesincethesystemisbrightenoughtobeobservedbytheirequipment(evenwithunaidedeye). This fact motivated us, the Hellenic Amateur Astronomy Association(HAAA;http://www.hellas-astro.gr),topublicizethiscampaigninGreece.InHAAAourmaingoalistopromotetheamateurastronomyperformedusingthenecessarymethodologyinordertoobtainscientificallyvaluableobservationsand contributing to pro-am collaborations. Thus, the contribution to such aprojectwasconsideredasauniqueopportunitytoparticipateinandtopromoteittotheGreekamateurcommunity. ThisworkreportstheresultsofthiscampaigninGreece,includingtheaimsandtheactivitiesorganized,insection2,adescriptionofthemethodsusedandtheobservationscollected,insection3,andasmalldiscussionontheresults,insection4.
of theAAVSO/Citizen Sky staff.The main goals of this campaign were to:(i) inform the Greek community about the rarity and the importance of theeAur’s eclipse, (ii)provide thenecessarymaterial forobservations forbothexperienced and first-time observers, (iii) collect all observations by Greekobservers,(iv)forwardthesetotheAAVSOdatabase,(iv)constructthelightcurveoftheeclipsealongwithanyinterestingresults.
2.2.Activities Inordertobetterpromotethecampaign,wehavecreatedthededicatedwebpage “The observational program of e Aur” (currently available inGreek athttp://www.hellas-astro.gr/article.php?id=765&topic=variables&subtopic=&lang=el) inwhichwepublished relatedmaterial.This includedanalyticalguidesonhowtoperformvisualobservations,maps,tips,newsandupdatedinformationonthesystem,andafrequentlyupdatedplotofallobservationscollected. Thispagehasalreadybeenupdatedthirteentimesanditwillcontinueasnewinformationandresultswillbecomeavailable.UpdatesweremailedtotheobserverswhohavealreadysubmittedobservationsandwerepublicizedatthetwomainforaofamateurastronomersinGreece,Astrovox(underthethread“epsilonAur”,http://www.astrovox.gr/forum/viewtopic.php?t=10578&highlight=\%C7\%ED\%DF\%EF\%F7\%EF\%F5) and AstroForum (under the thread“epsilon Aur”, http://astroforum.gr/forum/viewtopic.php?t=3862&highlight=\%C7\%ED\%DF\%EF\%F7\%EF\%F5), with almost 5,000 reads each. Atthesame time, therewasalsoa threadactive in the forum(under the thread“Greekon-goingvisualobservations”,http://www.citizensky.org/forum/greek-going-visual-observations)oftheCitizenSkyproject,whichwasinformingtheinternationalcampaignabouttheprogressinGreece. Aformalcallforobservationswasmadeduringatalkatthe6thPanhellenicConference on Amateur Astronomy held in Alexandroupolis, Greece, onSeptember26,2009(Maravelias2009),withmorethan500participants.Thetalk was also accompanied by a practical mini-workshop during the night,in which volunteers tried the visual observing technique on the system. Inaddition,twooftheregularmeetingsoftheHAAAwerededicatedtoeAur,regardingtheprogressoftheobservationsandoftheeclipsealongwithnewresultsobtained.Numerousinformaldiscussionstookplaceatthemeetingsandviathemailinglist. eAur was presented in all talks and workshops related to variable starsthattookplaceaftertheinitializationofthecampaigninMarch2009.Wetookadvantageofeveryeventtopublicizetheneedforobservationsandthecampaignitself(thatis,thePanhellenicAstropartiesof2009and2010,atAnavra-FthiotidaandMt.Parnonas,respectively,andthe“SunandVariableStarsMeeting”attheUniversityofAthensin2010).Especiallyinthecasesofworkshopsweofferedthe opportunity for hands-on experience of visual and digital observations.
Maraveliasetal., JAAVSO Volume 40, 2012682
AlthoughwedidnotcirculateanypressreleasesabouttheeAureclipse,there was one article, to the best of our knowledge, published in a nationalcirculation newspaper dedicated to this event (“Vima Science,” January 17,2010).TheauthorreferredtobothhistoricalandnewdataaboutthesystemanddidnotfailtorefertotheCitizenSkyprojectandtheGreekcampaign.
3. Observations
The campaign in Greece was heavily based on visual photometricobservations. However, we still experiment with other kind of observations,includingDSLRphotometry(GMwasamemberoftheDSLRDocumentationand Reduction team of the CitizenSky project (http://www.citizensky.org/teams/dslr-documentation-and-reduction,responsibletodevelopobservationalmethods and tutorials), CCD photometry (IMS was an official member ofthe International Epsilon Aur Campaign 2009 (http://www.hposoft.com/Campaign09.html)), and low-resolution spectroscopy.The observations willbe web-archived and made available through theAAVSO ftp site at ftp://ftp.aavso.org/public/datasets/gmaraj402.txt. We present each method andobservationsobtainedinthefollowingsections.
3.1.Visual Visualobservationsareallobservationsthatareobtainedusingtheeyeasaphotometer,eitherunaidedorthroughanopticalsystem(thatis,binocularsoratelescope).SinceeAurisabrightobjectthemajorityoftheobservationsweremadeusingunaidedeyeandbinoculars. The method used for visual observations is simply based on thecomparisonofthetargetstar(eAur)withtworeference(comparison)stars,usuallyhAurwith3.2V-magnitude(invisualobservationsallmagnitudesareroundedtothefirstdecimalplace),zAurwith3.8V-magnitude(actuallyit isaneclipsingbinaryof~3.75V-magnitudewithadropof~0.1duringeclipse,every2.7yearswitheclipseduration~40days,butitisconsideredas non-variable during the major part of the eAur eclipse except for theperiodNovember–December2009whenitwasineclipse—duringthattimeitwasavoidedasacomparisonstar),or58Perwith4.3V-magnitude.Bycomparingourtargetwithafainterandabrighterstar,wewereabletoplaceitbetweenthetwomagnitudes.Usuallytheerrorofthesemeasurementsisnot given, and an accuracyof a few tenthsofmagnitude is assumedwiththe most experienced observers going down to 0.1, although a subject ofcontroversy(Priceet al.2006).Insomecaseswewereabletogodownto0.05magnitudeerror,partlyduetothesmalldifferencebetweenzAurand58Per(wheneAurwasin-between)andpartly to thegrowingexperienceandconfidencewiththefield. Byapplying thismethodwewere able to collect 302observations fromtwenty-onepersons,presentedinTable1.
Maraveliasetal., JAAVSO Volume 40, 2012 683
3.2.CCD TheCCDobservationswereperformedbyusinganATik16ICmonochromecamera,equippedwithaZenit55mmf /2.8lens,andaJohnsonVphotometricfilter with the whole setup mounted on an equatorial mount. The reductionprocedure followed is the standardone forCCDobservations: (i) create themaster-dark, themaster-bias,and themaster-flat images(usually twenty-fiveimageswerecombinedforeachmaster image), (ii)subtractmaster-darkandmaster-biasfromeachscienceimage,(iii)divideeachofthesebythemaster-flat,(iv)alignandstackimagesofeachset(usuallytensetsofthirtyscienceimageseach),(v)performphotometryusingtheMaxImDLsoftware. The comparison stars used were h Aur and z Aur. Since the latter isan eclipsing binary the V-magnitude to use was provided each week byJeffHopkins (coordinator of the InternationaleAurCampaign2009).ThereductionandthephotometrywereperformedaccordingtotheguidelinesoftheInternationalCampaign. Usingthistechniquewecollected95CCDobservations,allacquiredbyasingleobserver.DetailsarepresentedinTable2.
3.3.DSLR DSLRphotometryreferstotheuseofanormalDigitalSingle-LensReflexcamera (DSLR) or any digital photography camera which: (i) can produceimages in a RAW data format, (ii) can focus semi-manually, (iii) is able tomanually select a shutter speed/exposure timeof several seconds, (iv) has awideenoughfield-of-viewtogetavariablestarandacomparisonstarintheimage.Inordertoobtaintheimagesneeded,thecameraisusuallymountedonasimpletripodwithatypicallensof50–90mmandexposuresofsomesecondstocapturethebrightstars(Kloppenborget al.2012,inthisvolume). ThedatareductionofDSLRobservationsfollowsthatoftheCCD.Inourcasethough,biasandflatfieldswerenotavailable,soaslightlydifferentprocesswasfollowed:(i)themaster-darkwascreatedasnormal(fromdarkimages),(ii)usingthreescienceimageswecreatedthemaster-flat(mediancombineoftheimages),(ii)subtractthemaster-darkfromscienceimages,(iii)divideeachof these by the master-flat, (iv) align and stack images of each set (usuallyaroundeightimages),(v)separatestackedimagestotheirRGBBayerpattern,andkeeponlytheGreenchannel(closermatchtotheVfilterpassband),(vi)performphotometry. The DSLR Documentation and Reduction team of the Citizen Skyproject has developed standard guides (http://www.citizensky.org/content/dslr-documentation-and-reduction)forsomewidelyusedsoftwarepackageswithin theamateurastronomercommunity(iris, aip4win, maximdl) inordertopresenteasywaystoreducedataandperformphotometry.Inthisworkweusediris(freesoftware)forimagereductionandphotometry,alongwiththespreadsheetsprovidedforthispurposeattheCitizenSkysite.WeusedhAur
3.4.Spectra3.4.1.80-mmrefractor As low resolution spectroscopy is available to amateur astronomers, asample was obtained with a Sky Watcher 80-mm Apochromatic refractorequipped with an ATiK 16 IC camera (640×480 px) and a Baader BlazeGratingSpectroscope(207lines/mmgratingwithadispersionof1267Å/mmandwavelengthcoverage~3800–6800Å). The spectraextractionwasperformed through the rspec software.Usingthe standard libraries, anA7Vspectral type starprofilewas selected for theidentification of the Balmer lines as well as for the wavelength calibration.Fromthecalibratedspectrum,theBalmerlineswereremoved,leavingonlyafeaturelessspectrumcomposedofthecontinuumemissionofthestarandtheinstrumentalresponseofthesystem.Bydividingthisresultbytheinstrumentalresponseweobtainedthefinalnormalizedspectrum. Usingthismethodwecollectedfourspectra,obtainedbyasingleobserver(Table3).OneofthesespectraispresentedinFigure1.
3.4.2.1.3-mreflector We were also able to use the slit spectrograph mounted on the 1.3-mtelescope at Skinakas Observatory (http://skinakas.physics.uoc.gr/en/) toacquireonespectrumofeAurduringitseclipse,on30September,2010.Thetelescopeisequippedwithaslitspectrograph(1302lines/mmwithadispersionof70.44Å/mmandwavelengthcoverage~5210–7280Å)anda2000×800ISASITeCCD. Thedatareductionperformedinirafincluded:(i)biassubtraction,(ii)flat-fielding,tocorrectpixel-to-pixelvariationsacrossthechip,and(iii)wavelengthcalibration.Thefinalcorrectionwouldbe toremove thecontinuumwith thehelpofastandardstar.Sincenostandardwasobservedthatnightandnogoodone was available we decided to leave the spectrum uncorrected.Althoughpresent,thecontinuumdoesnotprohibitusfromidentifyingbasicfeaturesinthespectrum. Onlyonespectrumwasobtainedwiththissetup(Table3)andispresentedinFigure2.
Maraveliasetal., JAAVSO Volume 40, 2012 685
4. Discussion
4.1.Statistics ThemainaimofthecampaigninGreecewastocollectalltheobservationsmadebyGreekobserversduringthe2009–2011eclipse(Figure3).Atotaloftwenty-oneindividualsmanagedtoobtain302visualobservations.Outofthetwenty-oneobservers,onlythreewerenotengagedwiththecampaignrunbytheHAAAandtheirobservationswereacquiredthroughtheAAVSOInternationalDatabase.However,themajorityoftheobservers(~86%)wereparticipantsofthiscampaign(thatis,theysubmittedtheirobservationsdirectlytoHAAA)andprovidedalmostall theobservations(93%).It is interesting topointout thatthirteenparticipants(~72%)werefirst-timersinvisualobservationsofvariablestars,butonlythreeofthemobservedmorethanacoupleoftimes.Althoughwewereexpectingalargercontribution,wehopethatthenewobserverswillcontinueobservingothervariablestars.Intotalthevastmajorityofthedata(255observations,almost84%of the total sample)camefromonly fourpersons,alreadyexperiencedobservers. Outoftheeighteenparticipantsofthecampaign,sixwerenotmembersoftheHAAA,whichmeansthatactuallytherewasanumberofpeopleoutsidetheHAAAinterestedinthecampaign.Inaddition,sevenoftheparticipantswerealreadyAAVSOobservers(thatis,theyhadanAAVSOobservercode). Althoughthecampaignwasheavilybiasedtowardsthevisualobservations(74% of the whole sample), there had been systematic work with digitalobservations(DSLR/CCD).Onlythreepeopleobservedusingdigitalsystems(twoDSLRusersandoneusingboth).Oneofour team(IMS)usedaCCDcameraaspartoftheInternationaleAurCampaign2009,producingthemajorityofthedigitalresults(almost90%ofalldigitalobservations). Spectroscopicobservationswerealsoattemptedbytwoobservers(ISMandGM),butmainlyastests.ExperienceandtimeavailabilitywaslimitedtofullyexploitthepowerfultoolofspectroscopyforeAur,buttheknowledgegainedcouldbeusedinfutureprojects.
deviation(s)were calculated. In thecaseswhereonlyoneobservationwasobtained,anerrorof0.1magnitudewasassumed (corresponding to thebestcase scenario—experiencedobservers).Thenall thedatawhichwerewithinthe3×srangewerekept(indicatedasthe“reduceddata”inFigure4).Weusedthesedatatoobtainthemedianvalueanditscorrespondingerrorforeachbin(representedasthe“median”lineinthesamefigure).Byvisualinspectionofthefinalresultsomeinterestingfeaturesoftheeclipsecanbeidentified. TheV-magnitudeofthesystembeforeenteringeclipsewas~3.0.Thesmalldropof0.1magnitudeaftertheMJD55050(August6,2009)cannotidentifythebeginningoftheeclipse(duetothe0.1-magnitudeoscillationsofthesystem(Carroll1991;Hopkinset al.2008).Onlylater,withinMJD55077–55097,wehaveaclearerindicationoftheingress,whichcouldbeplacedaroundtheMJD55087(August12,2009),withanerrorequivalenttothebinsize(thatis,±15days).Thedateiswithinthepredictedrangeofdates(Hopkinset al.2008). OnlyafterMJD55218(February2,2010),eAurseemstohavereacheditsfaintstate(totality)atmagnitude~3.8,losingalmost0.8magnitudein131±15days,inagreementwiththevaluesof137daysforthe1982–1984eclipseand135daysforthe1955–1957eclipse(Carrollet al.1991).TherewasasmalltrendofbrighteningafterMJD55261(March5,2010),whichcouldtemptustocreditittothemid-eclipsebrightening.However,sincetheerrorsarelarge,thebrighteningisnotstatisticallysignificant.Moreover,duringthisperiodeAurwasgettingloweronthehorizonandafterpassingbehindtheSun(June2010),itwasagainlowonthehorizon,whentheobservationswereresumed.Thispositiondefinitelyaffectedtheobservationsduetotheairmass. AftertheMJD55376(June28,2010)wenoticeascatterofvaluesaroundmagnitude ~3.7–3.75. There can be no estimation when the system passedfromthethirdcontact(startofengress)duetothedatascatter.OnlyafterMJD55797(August23,2011)wecanacceptthateAurwastotallyoutoftheeclipsewithaV-magnitude~3.1.Usingthepreviouslyestimateddateofingress(MJD55087)wecalculatethedurationofthe2009–2011eclipsetobe710±21days,whichisactuallywithin2–3sfromthepreviouseclipsedurations,647daysin1982–1984and670daysin1955–1957(althoughthereisatrendfordecreasingduration(Carrollet al.1991).
4.3.CCDandDSLRobservations We present the digital data (DSLR and CCD observations) in Figure 5.AllobservationswereobtainedwithinMJD55128–55545(October23,2009–December14,2010),wheneAurwasalreadyineclipse,withthemajorityofthedataobtainedduringtotality.Thus,therearenoadditionaldatatoallowforthedeterminationofingressorengress.Nevertheless,weobservemodulationsof~0.07magnitude,inagreementwithpreviousresults(Hopkinset al.2008;Carrollet al.1991).Thefaintestvalue,withinerrors,thateAurreachedwasmagnitude 3.789 ± 0.003. Moreover, the oscillations displayed in Figure5,
4.4.Spectra As there have been only a few tests with spectroscopy, we present thebestspectraobtainedinFigures1(spcA)and2(spcB).Theobservationswereobtained during the eclipse of eAur and, as such, its spectrum would be acompositeofthemainstarandthedisk.Thus,itisoutofthescopeofthisworktopresentaclassificationoranyspectralresultsregardingwiththenatureoftheobjects,butrathertopresentasampleoftheobservationsperformedandthelinesidentified. Thereisanoverlapofthetwospectraintherange5200–6700Å,wherethemostprominentfeaturesaretheNaIDoubletll5890,5896lines(characteristicfeatureofF-toM-typestars(Monteset al.1999))andtheHal6563line.Bothoftheselinesarevariableduringtheeclipse,revealingpropertiesforboththediskandtheprimarystar(Barsonyet al.1986;Chadimaet al.2011).OutsidethisregioninspcAallBalmerlinesareevidentwithsomeadditionalfeaturesaroundll4040, 4480, and 5050 but we were unable to resolve which linestheyare(thoughthel4480linecouldbetheMgIIlineatl4481Å).However,inspcBwewereabletoidentifynumerousmetalliclines.ThemostabundantmetalisironwithlinessuchasFeIIll5235,5274,5316,5363,6148,6238,and6247 lines and FeIll5226,5325,5657, and 5780 lines.Also present are theNiIIIl5534lineandtheSiIIll6347,6371lines(Chadimaet al.2011).Inthiscase,theNaIDoubletisalsonicelyresolvedtoitstwoseparateabsorptionlines(ll5890,5896,seeinsetgraphinFigure2).
5. Conclusions
The current work is a report of the results obtained from the Greekcampaigndedicatedtotheobservationofthe2009–2011eclipseofeAur.WehavebeensuccessfulininformingtheGreekamateurastronomicalcommunityabout the eclipse and its importance. Furthermore, we publicized the eventand the appropriate material for both experienced and first-time observers,byusing internet resources (dedicatedwebpage, threads inwell-knownfora)andtalks/workshopsatmajorastronomicalevents.Wemanagedtocollect413
Maraveliasetal., JAAVSO Volume 40, 2012688
observations (302visualestimates,95CCD,and11DSLRmeasurements,5low-resolutionspectra)from21Greekindividuals,whichhavebeensubmittedtotheAAVSOInternationalDatabase. Wewereable toconstruct the lightcurveof theeclipseand,evenundersome limitations of the data, to extract some interesting results. By visualexaminationofthelightcurvewenoticedthesystem’sV-magnitudedroppedfrom~3.0to~3.8,in131±21days,inagreementwithCarrollet al.(1991).WeestimatedtheingressdatearoundtheMJD55087±15days(August12,2009),withinthepredictedrangeofdates,andtheexitoftheeclipseaftertheMJD55797±15days(August23,2011).Thedurationofthe2009–2011eclipsewasfoundtobe710±21days,withintheerrormarginsofpreviouseclipses(Carrollet al.1991).Althoughwecannotconfirmthemid-eclipsebrighteningbythevisualobservations,theCCD/DSLRdatapresentedaratherpossibleindication.Moreover, 0.07-magnitude oscillations were present in the CCD/DSLR datainagreementwithpreviousobservations(Hopkinset al.2008;Kim2008).Inaddition,wepresentedourfirstattemptsatspectroscopicobservationsofeAur,whichresultedintheidentificationoftheNaIDoubletll5890,5896lines,theSiIIll6347,6371lines,andnumerousFeIandFeIIlines(Barsonyet al.1986;Chadimaet al.2011).
6. Acknowledgements
The authors are grateful to the observers: Athanasios Douvris, NikosFlemotomos,DimitrisGkionis,PanagiotisKottaridis,KrikisManolis,DimitrisManousos, Eleni Maraki, Serafim Ntovolos, Kostas Dimitris Panourakis,Paschos, Giorgos Stefanopoulos, Vasilis Takoudis,Andromachi Tsouloucha,Lefteris Vakalopoulos, George Vithoulkas, Giorgos Voutiras, and OrfeasVoutiras. TheauthorswouldliketothanktheAAVSOanditsCitizenSkyproject,especiallyAaronPriceandBrianKloppenborg.GMwouldliketoacknowledgethehelpofPabloReigregardingspectroscopyobtainedatSkinakasObservatoryand thevaluablediscussionswithThodorisBitsakisandPaoloBonfini. IMSacknowledgesthehelpofJeffHopkinsandRobinLeadbeater. Also, we acknowledge the supporting communities of the free/opensoftwares iris, python/matplotlib, and iraf, which have been used for thiswork.ThisresearchhasmadeuseofNASA’sAstrophysicsDataSystem.
References
Barsony,M.,Lutz,B.L.,andMould,J.R.1986,Publ. Astron. Soc. Pacific,98,637.Carroll, S. M., Guinan, E. F., McCook, G. P., and Donahue, R. A. 1991,
Guinan,E.F.,andDeWarf,E.2002,inExotic Stars as Challenges to Evolution,eds. C.A. Tout and W. Van Hamme,ASP Conf. Ser. 279,Astron. Soc.Pacific,SanFrancisco,121.
Hoard,D.W.,Howell,S.B.,andStencel,R.E.2010,Astrophys. J.,714,549.Hopkins, J. L., Schanne, L., and Stencel, R. E. 2008, in The Society for
Astronomical Sciences 27th Annual Symposium on Telescope Science, Held May 20-22, 2008 at Big Bear Lake, CA,SocietyforAstronomicalSciences,RanchoCucamonga,CA,67.
Kim,H.2008,J. Astron. Space Sci.,25,1.Kloppenborg,B.,Pieri,R.,Eggenstein,H.-B.,Maravelias,G.,andPearson,T.
2012,J. Amer. Assoc. Var. Star Obs.,40,815.Kloppenborg,B.,et al.2010,Nature,464,870.Maravelias, G. 2009, In Proceedings of the 6th Panhellenic Conference on
Amateur Astronomy, Alexandroupolis, Greece, 25–27 September 2009,143(inGreek).
Name Number of Participant1 HAAA AAVSO Observations Member2 Code3
Douvris,Athanasios 1 yes no DXA Flemotomos,Nikos 1 yes no — Georgalas,Byronas 1 yes yes — Gkionis,Dimitris 5 yes yes — Kardasis,Manos 44 yes yes KMO Kottaridis,Panagiotis 5 yes yes KPAA Krikis,Manolis 1 yes no — Manousos,Dimitris 4 yes no MUQ Maraki,Eleni 1 yes no — Maravelias,Grigoris 64 yes yes MGK Ntovolos,Serafim 1 yes yes — Panourakis,Kostas 1 no no PKO Paschos,Dimitris 1 no no PDIA Stefanopoulos,Giorgos 41 yes yes STF Strikis,Iakovos-Marios 106 yes yes SIAK Takoudis,Vasilis 1 yes yes — Tsouloucha,Andromachi 2 yes no — Vakalopoulos,Lefteris 1 yes yes — Vithoulkas,George 19 no no VGK Voutiras,Giorgos 1 yes yes — Voutiras,Orfeas 1 yes yes —
Totals(persons/observations) 21 302 18 12 10
1 Defined as an observer who submitted his/her observations directly to the HAAA (in “no” cases the data were retrieved from the AAVSO International Database). 2 Member of the Hellenic Amateur Astronomy Association (HAAA). 3 Observer who submits his/her observations to the AAVSO has a unique observer code—this is given when applicable.
Maraveliasetal., JAAVSO Volume 40, 2012 691
Tabl
e2.
DSL
Ra
ndC
CD
obs
erva
tions
obt
aine
dby
Gre
eko
bser
vers
dur
ing
the
2009
-201
1ec
lipse
ofe
Aur
.
Nam
e Fi
lter*
Cam
era
Lens
f-r
atio
Ex
posu
re(s
) N
o.
O
bs.
Gki
onis
,Dim
itris
G
reen
C
anon
300
D
Nik
on2
00m
m
2.8
51
Gki
onis
,Dim
itris
G
reen
C
anon
300
D
Tam
ron
300m
m
5.6
51
Kar
dasi
s,M
anos
G
reen
C
anon
300
D
Can
on1
7–85
mm
(@8
5mm
)5.
610
7
Stri
kis,
Iako
vos-
Mar
ios
Gre
en
Can
onE
OS
5DM
rkII
Ta
mro
n24
–105
mm
(@1
05m
m)
4.5
51
Stri
kis,
Iako
vos-
Mar
ios
Gre
en
Can
onE
OS
350D
Reb
elX
TI
Tam
ron
24–1
05m
m(@
105
mm
)4.
55
1S
triki
s,Ia
kovo
s-M
ario
sV
AT
iK1
6IC
mon
ochr
ome
Zeni
t55m
m
2.8
2–5
95* G
reen
cor
resp
onds
to th
e G
reen
cha
nnel
of t
he B
ayer
pat
tern
(RG
B) fo
r D
SLR
cam
eras
, V c
orre
spon
ds to
the
John
son
V ph
otom
etri
c fil
ter,
used
fo
r C
CD
pho
tom
etry
.
Tabl
e3.
Log
ofs
pect
rosc
opic
obs
erva
tions
obt
aine
dby
Gre
eko
bser
vers
dur
ing
the
2009
–201
1ec
lipse
ofe
Aur
.
Nam
e D
ate
Tele
scop
e C
amer
a G
ratin
g W
avel
engt
h D
ispe
rsio
n
(line
s/m
m}
(Å)
(Å/m
m)
Stri
kis,
Iako
vos-
Mar
ios
2Ja
n20
10
ED8
0mm
AT
iK1
6IC
(640
×48
0)
BB
G* (
207)
38
00–6
800
1267
Stri
kis,
Iako
vos-
Mar
ios
30Ja
n20
10
ED8
0mm
AT
iK1
6IC
(640
×48
0)
BB
G(2
07)
3800
–680
012
67S
triki
s,Ia
kovo
s-M
ario
s1
Feb
2010
ED
80m
m
ATiK
16
IC(6
40×
480)
B
BG
(207
)38
00–6
800
1267
Stri
kis,
Iako
vos-
Mar
ios
27A
ug2
010
ED8
0mm
AT
iK1
6IC
(640
×48
0)
BB
G(2
07)
3800
–680
012
67M
arav
elia
s,G
rigor
is
30S
ep2
010
Skin
akas
’1.3
m
ISA
SIT
e(2
000×
800)
13
02
5210
–728
070
.44
* Baa
der B
laze
gra
ting.
Maraveliasetal., JAAVSO Volume 40, 2012692
Figure1.Low-resolutionspectrumofeAurobtainedonAugust27,2010.ASkyWatcher80-mmApochromaticrefractorwasusedequippedwithanATiK16 IC (mono) and a Baader Blaze grating (see section 3.4.1). CharacteristicBalmerlinesareshownalongwiththeNaIDoubletlines.SpectrumobtainedbyIMS,reducedbyRobinLeadbeater.
Figure2.Low-resolutionspectrumofeAurobtainedonSeptember30,2010.Skinakas’ 1.3-m telescope was used, equipped with an ISA SITe and a slitspectrograph(seesection3.4.2).TheHaandNaIDoubletlinesareprominentalongwithaseriesofFeIandFeIIlines,andSiIIlines.SpectrumobtainedbyGMwiththehelpofPabloReig,andreducedbyMK.
Maraveliasetal., JAAVSO Volume 40, 2012 693
Figu
re3
.The
200
9–20
11e
clip
seo
fe
Aur
as
deriv
edf
rom
obs
erva
tions
(vi
sual
isin
dica
ted
with
ad
ot,C
CD
isin
dica
ted
with
a
cros
s,D
SLR
isin
dica
ted
with
atr
iang
le)
obta
ined
by
Gre
eko
bser
vers
,mem
bers
or
noto
fth
eH
elle
nic
Am
ateu
rAst
rono
my
Ass
ocia
tion
(the
listo
fcon
tribu
tors
isp
rese
nted
inT
able
s1a
nd2
).
Maraveliasetal., JAAVSO Volume 40, 2012694
Figure5.Plotofthedigital(DSLRidentifiedastriangles,andCCDidentifiedas crosses) observationsduring the2009–2011 eclipseofeAur (see section4.3fordetails).TheDSLRpointjustbeforeApril1st,2010,isapoorresultprobablyduetothepresenceofthinclouds.
Photoelectric Photometry of e Aurigae During the 2009–2011 Eclipse Season
Frank J. MelilloHoltsville Observatory, 14 Glen-Hollow Dr, E-16, Holtsville, NY 11742; [email protected]
Received February 13, 2012; revised March 1, 2012; accepted March 1, 2012
Abstract A total of 100 V-band photometric observations were made atHoltsvilleObservatoryforthe2009–2011eclipseofeAurigae.Alightcurvehasbeenplottedusingdatafromtheseobservationswhichcoversthephasesbefore, during, and after the eclipse. The light curve shows precise timingduringthefirst,second,third,andfourthcontacts,whichmarkimportantpartsoftheeclipse.ThemagnitudesanddurationoftheeclipseinthephotometricVbandarediscussed.
1. Introduction
e Aurigae is different from other binary star systems. It has been amysteriousstartoussince1821whenitwasfirstnoticedthatitwasavariable,butitwasnotwellknownuntiltheearly1900s.Thisstaristhelongestperiodeclipsingbinaryeverstudied.What’ssomysteriousabouteAuristhatwedonotknowmuchaboutitseclipsingobject.Eventhoughtheeclipsingobjecthasbeenwellidentifiedasexisting,itisstillunclearwhatitismadeof.Withits27.1-yearperiod,it isasloweclipselastingatotalof1.75years.Regardlessofwhatequipmentyouusetostudythisbinarysystem,thereisnoclearclueabout theeclipsingobject.Usingphotometry, spectroscopy,polarimetry,andinterferometry,wearebeginningtogetanindicationofitsnature.Duringthetimeoftheeclipse,eAurdimsabout0.8magnitudeandthenreturnstoanormalbrightness. What isknownabout thisbinarystar system is that theeclipsingbody’sorbitisinourlineofsight.Every27.1years,eAurundergoesaneclipsewhichmeanstheunknownsecondaryobjectpassesinfrontofastarasseenfromtheEarth,causingthelighttodim(Hopkinset al. 2006).TheeAursystemissostrangethatitisunique.Withsuchararebinarystarsystem,itisdifficulttostudytheeclipseevery27.1years.SomethingisorbitingeAur,butwhatisit? eAurisknowntobeaF0-typestaranditisahighlyluminoussupergiantatVmagnitude3.00.TheeclipsingobjectorbitingthisF0-typestardoesn’texhibitanyspectrumofitsownbutitismostlikelytobeahiddenB-typestar.Thediskmustbedarkbuthavetransparentregions.Butduringtheeclipse,stillitisthickenoughtoobscuresomeofthelightfromtheF0star.Thevisuallightcandimabout0.8magnitudeeventhoughthespectrumdoesnotchange.Also,theF0
Melillo, JAAVSO Volume 40, 2012696
staritselfisavariableanditisperhapsasemiregulartype.Itisapulsatingstaroutsideoftheeclipsewithaperiodof60to65days.Theamplitudecanchangeasmuchas0.15magnitudeinadditiontotheeclipse. eAurhasbeenstudiedsubstantiallyevery27.1years.The1982–1984eclipsewas observed by many amateur astronomers, by ground-based professionalobservatories, and from space. No matter what data were collected, still itwasacompletelymysteriousstar.Thereissomuchinterestinthisbinarythatall observers are trying to squeeze out as much detail as possible about thesecondary.Whatwereallyseeisonlythelightoftheprimarystar. ThelongawaitedeAur2009–2011eclipseisfinallyover.Priorto2009,many organizations and observing campaigns were formed (Hopkins et al. 2008).Thisauthorisparticipatinginoneorganization,whichisrunbyJeffreyHopkinsofPhoenix,Arizona.During the2009–2011eclipse,manyamateurastronomerswereequippedwiththemostadvancedtechnology,especiallyinspectroscopy, which plays an important role for eAur where the secondarycan be detected at certain wavelengths. Also, the Center for High AngularResolutionAstronomy(CHARA)operatesaninterferometeronMountWilsoninCalifornia(Kloppenborget al.2010).Withup tosix telescopescombinedtogether in infrared light, they have successfully imaged eAur. They wereable to detect an elongated object partially obscure the primary disk.The eAureclipseof1982–1984wassuccessfulbut leftusmanyquestions for the2009–2011season.Hopefullymanyamateurandprofessionalastronomerscanshednewlightonthesecondarythistime.ThisauthorcarriedoutphotoelectricphotometryinVlightduringthe2009–2011eclipse,anditisdescribedhere.
2. Method
PhotoelectricphotometrywastheonlymethodthisauthorusedtomonitortheentireeclipseofeAur.TheobservationsprovidedanexcellentcoverageinVband.ItisaJohnson-typefilterwithapeakspectralresponseat5400nm.The readings were taken photometrically using a SSP3 OPTEC photocountingphotometercoupledonaCelestron8-inchf /10telescope.lAurwasthecomparisonstaratVmagnitude4.71.Itislocatedjustfourdegreeseast-southeastofeAuranditwasanexcellentchoicetocompareitsbrightnessandtominimizetheatmosphericcoefficientduringthe2009–2011eclipseseason. Foreachset,fourreadingsweretakenattenseconds’integrationtimeforeAur,skyreadings,andthenthecomparisonstar,lAur.Thisauthorsometimesmonitored three to four sets, dependingon theweather condition and time.Once the observing night was over, the readings could be calculated. ThephotometricreadingsofeAur, thesky,andlAurwereaveraged.Then, thesky was subtracted from both stars’ readings. That left the ratio brightnessbetweeneAurandlAur.Oncetheratiowascalculated,eAur’sbrightnesscouldbedeterminedwiththeknownmagnitudeoflAur.Also,theStandard
Melillo, JAAVSO Volume 40, 2012 697
Deviation(SD)wascalculatedtoanalyzehowmucherrorwasinthereadings.Mostofthetime,theerroramountwasfrom0.012magnitudetoasmuchas0.0427whenallthereadingswerecalculated.Thealtitudewasalsocalculatedto determine the air mass during the time of the observations. During theobservingrun, thehigher thestars, thebetterchanceofgettingtheaccuratereadingsfromthephotometer. Many factors had to be also considered: seeing conditions, the winds,periodicclockdriveerror,polaralignment,andthestabilityofthephotometer,whichcanallaffectthereadings.Duringmostnights,theseeingconditionwasaboveaveragewithnowinds,andtheperiodicclockdriveerrorwasnoticeableattimes.Beingthataportableobservatorywasused,thepolaraxiswasalignedascloseaspossibletothecelestialnorthpole.Therefore,inspiteofthesmalldrift,thestarstillstayedinsidethereticlecircleduringthelengthoftime.FortheSSP-3OPTECphotometer itself, theunitwas turnedonat leastanhourbeforethestartofthefirstcounts.This“warmup”routinestayedonuntilthephotometer dark count was stable enough for accurate readings. In fact, thecoldertheoutsideairtemperaturewas,themoretimethephotometerneededtowarmup.Theunitwasrunningona9-voltbatterytoavoidthepowercordtangle-upduringthenight’srun.Inthephotoelectricphotometrymethod,theaccuracy can be little as 0.01 magnitude.This author’s readings were closeenoughtogeneratetheshapeofthelightcurve(seeFigure1andTable1).
3. Observations
eAurislocatedintheconstellationAurigaeasathirdmagnitudeobjectandpassesnearlyoverheadmostevenings,asseenfromtheHoltsvilleObservatory.AllphotoelectricphotometryresultsweretakenfromHoltsvilleObservatory,locatedunderamoderatelylight-pollutedskyfiftymileseastofNewYorkCity.eAuriseasilyvisibletothenaked-eyeanditisoneofthethreestarsformingtheasterism“TheKids,”nearaAur(Capella). ThisauthorwasgearingupforthefirsteclipseoftheMillennuim.Priortothestartoftheeclipse,thefirstcontactwaspredictedtooccurinAugust2009(Hopkins et al. 2009).The observations commenced on December 3, 2008,during the 2008–2009 observing season to develop a baseline for the lightcurve.Thephotometricreadingsweretakennightly,weatherpermitting,untilthestarreachedaconjunctionwiththesuninJune2009.Withinthisperiod,eAurwasshowingaslightvariationaveragingVmagnitude3.00out-of-eclipse(OOE).Thevariationhadnothingtodowiththeeclipsingobject.eAuritselfisavariableandperhapsasemiregulartypewithamplitudeof0.10magnitude.ThelastreadingsoftheobservingseasoninVbandshowednoindicationoftheupcomingeclipse. After the solar conjunction, the observation resumed for the 2009–2010observingseason.Thefirst readingwasonAugust14,2009, in the
Melillo, JAAVSO Volume 40, 2012698
morningsky.Even though theexact startof theeclipsewaspredicted tobeonAugust6,2009,everything lookednormal.After severalweeks, itgotveryinteresting.Sincethestartofthenewobservingseason,eAurhaddimmedconsiderably.Thisauthordidn’tknowwhetherornotthiswasduetoeAur’svariationitselforthestartoftheeclipse.AseAurcontinuedtodim, thepartialeclipsehadactuallybegun,but itwasn’tconfirmeduntilaftermid-Septemberwhenthemagnitudedroppedbelow3.15V. TheexcitementbuiltupaseAur’sbrightnesscontinuedtofall.ThisauthortookphotometricreadingsinVbandonanaverageoftwonightsperweek.Atcertainpointsofthedecliningphase,eAur’sbrightnessreductionsloweddownabit.Thiswasduetothevariationofthestaritself.Itformedasomewhatwavypatternasitcontinuedtodim. Atthispartofthedecliningstage,theauthorwasgettingreadyforthesecondcontact.Itwasn’tuntiltheendofDecember2009whenthebrightnessdeclinesloweddownalotat3.70V.ItstillshowedaslightdimminguntillateFebruary2010when the lightcurve reached rock-bottomat3.78V—but thiswasnotnecessarilythesecondcontact.WhileeAuritselfvariedwithin0.10magnitudethewholetime,thismighthaveinterferedwiththeactualtimeofthesecondcontact,whichshouldbethebeginningofthetotality.ThesecondcontactwaspredictedonDecember19,2009,anditlookedlikeitwasrightonschedule,buttheOOEvariabilitycausedthelightcurvetodimevenfurther.Afterthesecondcontact,eAurshowedasmallvariationevenintotality.Thelightcurvecouldbeeasilyidentifiedasasemiregulartypewitha65-dayperiod. Hereistheinterestingpart:eAurisknowntoshowasurgeofbrightnessasitnearstomid-totality,whichmaycauseperhapsasmallgapinthemiddleoftheeclipsingbodytoallowtheF0startoshinethrough.Thismysteryhasnot been explained yet, but hopefully it can be solved this time. The mid-totalitywasscheduledtohappenbyearlyAugust2010.ThisauthortriedtotakereadingsasmuchaspossiblewhileeAurwasapproachingsolarconjunction,but,unfortunately,thereadingsweretakenatlowaltitudeinthenorthwestaftersunset.The accuracy of the photometric measurement was not as great, butanyattempttotakereadingswasworthwhile.Accordingtothelasttwonightsin May 2010, this author may have caught a surge of brightness to 3.55V.Thephotometricresultsmaynotbeenprovenyetbecausetheauthordidnotknowwhetherthiswascausedbyatmosphericturbulencefromthelowaltitude,eAur’svariationitself,ortheF0starshiningthroughthegap.eAurwasnotinagoodpositiontotakefurtherreadingsanditwentthroughthesolarconjunctioninJune2010. The2010–2011observingseasonbeganwheneAurwasjustpastthemid-eclipsestage.TheobservationsresumedonAugust13,2010,earlyinthemorningsky.Therewasnosurgeofbrightnessatmid-eclipsethatwasexpectedjustshortlybeforethesolarconjunction.Atthistime,theexpectedsurgeofbrightnessmayormaynothaveoccurredatexactlymid-eclipse.Thisisyettobedetermined.
Melillo, JAAVSO Volume 40, 2012 699
ThebrightnessofeAurvariedwithin0.10magnitudethroughoutthewholesecondhalf of the eclipse,with themagnitude averaging at 3.71V.Even attotality,eAurclearlydemonstrateda65-dayvariationunrelatedtotheeclipsingbody.Thestarwasmonitoredonanaverageofonceperweek,andtheauthorpreparedforthethirdcontact,whichwaspredictedforMarch19,2011.Duringthat month, the star’s brightness became a little strange. According to theirregular65-daypattern,eAurwas expected toget a littlebrighter. Instead,itwentoff trackabit.Asthis time,wedon’tknowif thiswasrelatedtotheeclipsingbodyortheresultofthestar’sirregular-typebehavior.AfterMarch19,2011,thebrightnessbegantorisequiterapidly,whichindicatedthethirdcontacthadpassed.Thetotaleclipsewasoverandthiswasthebeginningofthepartialphase.Ithadrisenatafasterpacethanduringthedecliningstage.The fourthcontactwas scheduled forMay19,2011.Earlier thatmonth, thebrightnessleveledoffat3.25V.Again,thiscouldbeeAur’svariationitself,butatthesametime,thephotometricreadingsweregettingdifficulttoobtainduetotheloweraltitudeinthenorthwest.Thefourthcontact,whichmarkstheendofthepartialphase,wasprobablynotcaught.InlateMay2011,eAurwasapproachingsolarconjunctionandthereforenomorereadingsweretaken. More observations of e Aur began in September 2011, after solarconjunction. Even though the eclipse ended after the fourth contact aroundlateMayorJune,thisauthorcontinuedtomonitorthisstarjusttoconfirmthattheeclipsewasover.Thestarreturnedtoanormalbrightnessatanaverageof3.00V,justasbeforetheeclipse.AftermonitoringtheentireeclipseofeAur,thephotometricobservationsceasedaftertwoyearsandninemonths.
4. Conclusion
OnlytheVbandatapeakspectralresponse5400nmwasusedtodeterminetheaccurateshapeofthelightcurve.Theresultmaybesimilartothe1982–1984 eclipse but a closer look might detect some differences, especially inspectroscopy. However, the mid-eclipse brightening is still a mystery and itwasn’twellobservedduetothetimeofthesolarconjunction.Itisstillprematuretodrawanyconclusionwhetherthemid-eclipsebrightnesstookplaceinthiscycle. Thecontactpointsneedfurtheranalysisinordertodrawanyconclusion.It isquitedifficult toobtain thedataof thecontact timejustbyoneperson.Therearegapsbetweentheobservationsandprobablytheactualcontacttimemaybemissedevenbyoneday.Buttheauthor’slightcurvewasconstructed(see Figure 1) and we see two different types of variations.WhileeAur isa semiregular typevariable star, the65-day cycle0.10magnitude is clearlyseeninthelightcurve.Secondly,theeclipsingobjectofa27.1-yearperiodisobvious—seen at 0.70 magnitude amplitude.With 100 data points, one canshowthattheeclipsetookplacebutnotclearlymarkthecontactpoints.The
Melillo, JAAVSO Volume 40, 2012700
actual timeof thecontactpointsmaybeconfusedbyeAur’s65-daycycle,andthe0.10-magnitudevariationisenoughtoburytheactualcontacttimes.Using only the author’s data, it would be premature to draw more specificconclusionsaboutthecontacttimes.Datafromotherobserversmayfillinthegapsandthereforethecontacttimesmaybedetermined.Also,theresultsmaybecomparedwitheclipsesofthepasttoseewhetherthereisanysignificantdifference. While the photometric data gathered by participants is muchmoresophisticatedthanforthepasteclipses,thetrulynewtypeofdataisthespectroscopy.With today’sadvanced technology, thiscollectivephotometricand spectroscopic data set is the best ever obtained. The eAur 2009–2011eclipse isbehindusand the resultswillbestudiedformanyyears tocome.Still,therewillbemorequestionsthananswers.Hopefully,wewillhavemoreanswersbythenexteclipsein2036–2038.
Hopkins J. L., et al. 2008, “Gearing Up for e Aur’s First Eclipse of theMillennium,”Proceedingsforthe27thAnnualConferenceoftheSocietyforAstronomicalSciences.
Small Telescope Infrared Photometry of the e Aurigae Eclipse
Thomas P. Rutherford201 Clear Branch Circle, Blountville, TN 37617; [email protected]
Received May 15, 2012; revised June 18, 2012; accepted June 18, 2012
Abstract Near-infraredphotometryofeAurigae,intheH-andJ-bands,wasundertaken during the 2009–2011 eclipse using telescopes of moderate size(8-inchand14-inchdiameter).InstrumentsofthissizesuccessfullycollectedscientificdataintheH-andJ-bands.ObservationsweremadefromthecampusofEastTennesseeStateUniversity(ETSU),JohnsonCity,Tennessee,thecampusofKingCollege,Bristol,Tennessee,andfromtheauthor’shome.Signal/Noiseratiosofapproximately45wereobtainedduring timesofmaximumeclipse.HigherS/Nratioscouldhavebeenobtainedbyextendingthelengthoftimeontarget.S/Nratiosofalmost100wereobtainedoutsideofeclipse.Theinfraredlightcurvesproducedcloselyparallelthelightcurveinthevisualrange(V),beingabout0.5magnitudebrighterinHand0.7magnitudebrighterinJ.Theeclipsewaseasilydetectedandfollowedthroughoutitsduration.TherateofingresswasshallowerthantherateofegressinboththeH-andJ-bands.Thebackgroundvariationsoftheprimarystarwerereadilydetected.1. Introduction
ThevariablestareAurhasbeenofinteresttoastronomersforalmosttwocenturiesduetoitslongperiod(27.1years)andthelongduration(approximatelytwoyears)ofitseclipses.Withsuchalongeclipse,theoccultingobjectcannotbe another star, but must be some other type of object.The exact nature oftheeclipsingobjecthasbeenamystery,althoughmoreinformationbecomesavailableeach timeaneclipseoccurs.This isdue tobothan increase in thenumberofobserversandalsocollectionofdatainadditionalwavelengthsnotutilizedinpreviouseclipses.Thecurrentmodeloftheoccultingobjectisthatofalarge,cooldiskwhichcontainsacentralB-typestar(Stencel2011). Infraredphotometryhas,untilrecently,beentheprovinceofprofessionalastronomers,usinglargetelescopes,sophisticatedinfrareddetectors,andhigh-altitudeobserving sites.UKIRT,SOFIA, IRTF, andSpitzer are a fewof theprofessionalinstrumentsthatcometomind.Itisnowpossiblefortheaveragevariablestarobserver,usingacommontelescopeofmodestsize,toworkinthenear-infraredpartoftheelectromagneticspectrumandtoproducehighquality,scientificallyusefulresults.TherecenteclipseofeAurprovidedanexcellentopportunitytodemonstratethesecapabilities;duringpreviouseclipses,infraredstudieswere restricted toprofessionalobservatories since amateurobserversdidnothavethecapabilty.
Rutherford, JAAVSO Volume 40, 2012 705
Therearenumerousadvantages(andsomedisadvantages)fortheobserverwho works in the near-infrared when compared to visual-band photometry(Templeton2011a):
2) Atmospheric extinction,while present, is lowcompared to thevisualwavelengths.Astheeclipseneareditsend,somedatawerecollectedattheveryhighairmassof5;thishighairmassdidcauseanincreaseinnoise,buttheS/Nratiowasstilladequate.
4) There are few infrared observers. While for some stars, visual-bandobservers (V in particular) will find their data lost in the crowd and socontributelittle,thisisnotthecaseintheinfrared.Forthedurationofthisproject(November2008–present)therewereonlytwoobserverssubmittingH-andJ-banddatatotheAAVSO.
The equipment used during the eclipse consisted of an Optec SSP-4photoelectricphotometerwithamanualfilterslideholdingH-andJ-bandfilters.TheH-andJ-bandfilterband-passesareclosesttotheMaunaKeaObservatory(MKO)andCaltech/Tololo(CIT)systems(Henden2002).Thegallium-arsenide(GaAs)detectorintheSSP-4isonemillimeteracross(about100arcsecondsatthefocalplaneofan8-inchSCToperatingatf /10).Aversionofthephotometerisavailablewithasmaller,0.3-mmdetector,butthesmallerdetectorgiveslesssatisfactoryresults(Hopkins2006a). TheSSP-4photometeristheresultofacollaborativeprogrambetweentheAAVSOandOptec,Inc.(West2007).Itcontainsaneyepieceforcenteringthetargetandcomparisonstars,aflipmirrorforsendingtheimagetothedetectoraftercentering,andafilterslidecontainingtheH-andJ-bandfilters;thefilterslideandflipmirrorareoperatedmanually. The software which Optec provided for use with the photometer wasutilizedformostoftheobservations,althoughanalternatesoftwarepackage,
Rutherford, JAAVSO Volume 40, 2012706
written by Brian Kloppenborg at the University of Denver, was used onoccasion (Kloppenborg 2011). This software gave greater control over thephotometer,but theOptecsoftwaremadeabettermatchwith theparticularspreadsheetthatwasinuse.BothsoftwarepackagesranwithoutissueonanolderlaptopcomputerrunningWindowsXPPro.Thiscomputerwasusedforbothtelescopeandphotometercontrol.Thecomputerhad1GBofRAMandasingle-coreCeleronprocessor;ithadnoproblemscontrollingbothinstrumentsatthesametime. Threedifferenttelescopeswereutilizedduringtheeclipse:an8-inchMeadeLX200,aCelestronC14,andaMeade8-inchLX90.ThetwoMeadeSCTswerefork-mountedonequatorialwedges,whiletheC14wasmountedonaGerman-equatorial mount (GEM).All of the telescopes proved adequate to the task,althoughtheC14,duetoitslargeraperture,gavethebestresultsintheshortestperiodoftime.
3. Observations
Whenpreparationsbeganfortheupcomingeclipseduringthesummerof2008, an 8-inch Meade LX200 located at ETSU’s Powell Observatory wasutilized.Itisreportedthata10-inchtelescopeistheminimumusefulsizeforinfraredphotometryofeAurduetothefaintnessofthetraditionalcomparisonstar,lAur(Lucaset al.2006).lAurisfaintintheinfrared(magh=3.33,magj=3.62),makingitdifficulttousewithasmallertelescope.Thesolutionwastouseabrightercomparisonstar—theAAVSOrecommendationsfortelescopesunder 0.25 meter (10 inches) are to use Capella as the comparison star andbTau(elNath)asthecheckstar(Templeton2011b). ThetwoMeadetelescopeswerecontrolledwithMeade’sautostarsoftware.Althoughothercontrolsoftwarewastested,theMeadesoftwareprovedtobevery capable and worked well. The C14 was controlled using starry night pro 5,whichalsoworkedwell.Thegoalsinitiallywere:1)collectpre-eclipsedataoneAur,2)takemeasurementsofstandardstarstobeusedforcalibratingthephotometer/telescopecombination,and3)measureothervariablestarsontheAAVSO’sIRphotometrylist. The firstobservationsofeAuroccurredonNovember2,2008.The starwaskeptunderobservation(fromtheETSUobservatory)fromthatdateuntilitwaslosttothesuninlateMay2009.eAurwasre-acquiredonthemorningofJuly24,2009,andatthatpoint,itbecameobviousthattherewouldsoonbeaproblem.Theauthor’s“dayjob”isthatofahighschoolchemistryteacher—onceschoolbeganafterthesummerbreak,itwouldnotbepossibletomaketheforty-five-minutedrivetoETSUearlyinthemorning,getsetup,collectthedata,geteverythingputaway,andthengettoschoolintimeforthestartofclasses. Asearchforasuitabletelescopetousefromtheauthor’shomewasstartedand a usedMeade8-inchLX90 (UHTCcoatings)with an equatorialwedge
Rutherford, JAAVSO Volume 40, 2012 707
wasfound.Itwaspurchasedandsetuponaconcretepierattheauthor’shome.The LX90 was not as accurate as the LX200, but its tracking and pointingabilities were more than adequate for use with the 1-mm detector of thephotometer.TheonsetoftheeclipseinAugust2009waseasilydetected. Astheeclipsedeepened,itbecameapparentthatalargertelescopewouldgivebetterresults,asthebrightnessofeAurwasdecreasing.AdiscussionwithAAVSO DirectorArne Henden, while he was on a speaking engagement atETSU,ledtotheadvicethatlongertimeontargetwasonesolution(Henden2009).Thisworked,butthetimeinvolvedincreasedaswell. Atthispoint,about1.5hoursofcollectingdatawererequiredforthetwodatapoints(oneeachinHandJ).Theinfraredskyisnottransparentoverlongperiodsof time(Henden2002),especiallyintheSoutheasternUnitedStates.A larger telescope would allow shortened observation times. King College,in Bristol, Tennessee, offered the use of their Celestron C14 located at thecollege’sBurkeObservatory.Thecollege’sphysicsdepartmentalloweduseofthetelescopewhenitwasnotbeingotherwiseutilized. The C14 made a large difference in the S/N ratio. It was used for theremainder of that observing season, although a move back to the ETSUtelescopewasrequiredinlateMay(treesbecameanissue).Duringthelastyearoftheeclipse,theC14begantobeutilizedmoreoftenbyKingCollegeandsoobservationsmovedbacktotheLX90andtheLX200.Currently,theLX90istheprimaryinstrumentusedforpost-eclipsemonitoring.
4. Data collection
Priortoanight’sobservations,thephotometerwaspoweredupandhookedto the laptop computer. The photometer drew from its own external powersupply,butwascontrolledbysoftwareonthelaptopthroughaUSBcable.Thephotometer’stemperaturecontrolwassetto–40°Celsius(belowambient). Whenfirstturnedon,thephotometertypicallyproducedhighdarkcounts—thesedecreasedas the instrumentstabilized(30–60minutesafterpower-up).Thedarkcountswouldnormallydropslowlyduringtherun,butnotsofarthattheybecameaproblem.While thephotometerwasstabilizing, the telescopewassetup,polar-aligned,andplacedundercontrolofthelaptop(theC14waspermanently mounted, the other two telescopes were not). Prior to startingthedatarun,thephotometer’sdarkcountswerecheckedandthegaincontroladjustedtobringthemintotherangerecommendedbyHopkins(2006b).Thephotometerwasthensettoagainof100,witha10-secondexposure. The comparison star,aAur (Capella),wasmeasuredwithbothHand Jfilters.The telescopewas thenoff-set slightlyanda readingon the skywasmade,onceagaininbothfilters;thepatternusedwas“compH,compJ,skyJ,skyH.”Thetelescopewouldthenbepointedatthevariablestar(eAur)andthesamepatternrepeated“variableH,variableJ,skyJ,skyH.”Thenbackto
Rutherford, JAAVSO Volume 40, 2012708
aAur where the pattern was repeated once again.The last measurement ofthenightwasofthecheckstar(bTau)sinceithadlowestpriority(SchmidkeandHopkins1990).Thisprocedureproducedonedatapoint ineachfilter.Aminimumofthreesuchobservationswererequiredsothatmeaningfulstatisticscouldbecalculated;more than threeobservationswerenecessarydue to thesmallsizeofthetelescopesused. Thephotometer software,providedbyOptec,allowed thephotometer totake one exposure, three sequential exposures, or four sequential exposures.Whentheprojectfirstbegan,threeexposuresoftensecondseachweretaken,followedbythreeexposuresontheskyandthenbacktothecomparisonstarwherethreemoreexposuresweremade(accordingtotheabovedescription).Thesethreeobservationswerethenaveraged.Thiswasrepeatedatleastthreetimes(orfiveorseven,andsoon.).Eachgroupofthreereadingswasaverageddown toone readingand then the three (or fiveor seven, and soon) singlereadingswereaveraged togivea singlenumber for the star’sbrightness forthatnight’sobservations.Thiswasthenrepeatedusingtheotherfilter.Thetimeinvolvedforthiswassignificant,especiallyinthebeginning. Thismethodworkedwell,exceptwhentheskywasrapidlychangingorforobservationsmadeathighairmass—theskywouldsimplychangetooquicklybetweenonemeasurmentandthenext.ItwasrealizedthataveragingthedatatwicedidnothelpintermsofS/Nandaddedagreatdealoftimetotheoperation.Astheeclipseneareditsend,“singleshot”readingsbegantobetaken—comp,sky, var, sky, comp, sky—one reading each, not grouped into threes. Thisproduceddatathatweregenerallybetter,ifthestarswereunderchangingskyconditions,thanthe“groupofthree”methodmentionedabove.Becauseofthespeed at which the readings could be taken, data could be collected at highairmassandstillnotbetoonoisy. Originallythispatternwasrepeatedthreetimes,butitwasfoundthatthetotalcountswerelow,(exceptwhenusingtheC14),soitwasincreasedtofivesets,thentoseven,andfinallytoninesets.Ninerepetitionswerenotpractical,soeventuallysevensetsweresettledupon—thisgavethebesttrade-offbetweentimeinvolvedandS/Nratio.
5. Data reduction
AMicrosoftexcelspreadsheetwasdeveloped,usingaformatthatallowedtheobservertosimplycutandpastthedatafromthephotometer’snativeoutputfiledirectlyintothespreadsheet. The data output of the photometer contained columns of data,representingsuchthingsasdate,time,filtersused,gainsettings,lengthofexposure,andnumberswhichrepresentthephotonsdetectedduringtherun.Itwas inanexcel-readable format.Severalquantitiesmustbecalculatedfromthisinformation:
Rutherford, JAAVSO Volume 40, 2012 709
1) Thetelescope’slocation—thisinformation,alongwiththesiderealtime,wasneeded tocalculate the localhourangleand from that, atmosphericextinction,
Theaveragedphotometer readingsmentionedpreviously represented thenumber of photons that arrived from the target star during each exposure.Severaladditionalcorrectionstotheserawcountswerenecessaryinordertoget thestar’s truemagnitudeineachfilter.Asummaryoftheproceduresarelistedbelow.ThefullproceduresandequationsusedmaybefoundinHendenandKaitchuck(1982),Hall(1988),andSchmidkeandHopkins(1990):
3) Apply an extinction correction to the differential magnitude of thetargetstar.Thisextinctioncorrectionisnecessaryduetothedifferenceinairmassesbetweenthecompandtargetstars.InthecaseoftheaAur-eAurpair,thedifferencewasslightwhilethestarswererisingintheeast(latesummer),butincreasedastheytraveledwestward,reachingamaximuminlatespring.
An examination of Figure 2 and Figure 3 shows that the eclipse beganaroundJD2455046 inHandJD2455057 in J (3August2009 inHand14August2009inJ).Thereisalargeamountofscatterinthepre-eclipseJbanddatapointsandthismightexplainthevariationinthestartingdatesbetweenthetwofilters.TheeclipseappearedtoendonJD2455707inboththeHandJbands(26May2011).
7. Conclusions
Near-infrared photometry in the H and J bands can be successfullyundertaken using a telescope of modest size, such as an 8-inch Schmidt-Cassegrain. Care should be taken in the choice of comparison, target, andcheckstarsinordertokeepsignal/noiseratiosashighaspossible.TheeAureclipsedidfallwithinthereachofthistypeofsetup,butatitsdeepest,thedropintheS/Nratiowasveryapparent.Targetstarsforsuchasetupshouldbebrighterthanaboutmagnitude2inthechosenfilters,ifpossible.Fainterstarscanbemonitored,butrequirelongdatacollectiontimesinordertokeepthesignalatacceptablelevels.
Rutherford, JAAVSO Volume 40, 2012 711
8. Acknowledgements
Theauthorwouldliketothankthefollowingindividualsandinstitutionsfortheirguidanceandassistanceinthisproject:Dr.GaryHenson,ETSU;Dr.Beverly Smith, ETSU; Dr. Ray Bloomer, King College; Dr. Edward Burke,KingCollege(deceased);ETSUDepartmentofPhysicsandAstronomy;andtheAmericanAssociationofVariableStarObservers. Specialthankstomywifeandfamilyforallthelonghoursandcomplaintsabouttheclouds.
References
Hall,D.S.,andGenet,R.M.1988,Photoelectric Photometry of Variable Stars,Willmann-Bell,Richmond,VA.
Lucas, G. A., Hopkins, J. L., and Stencel, R. E. 2006, in The Society for Astronomical Sciences 25th Annual Symposium on Telescope Science,SocietyforAstronomicalSciences,RanchoCucamonga,CA,25.
Schmidke, P. C., and Hopkins, J. L. 1990, Workbook for Astronomical Photoelectric Photometry,HPODesktopPublishing,Phoenix,AZ.
UV-Blue (CCD) and Historic (Photographic) Spectra of e Aurigae—Summary
R. Elizabeth GriffinDominion Astrophysical Observatory, 5071 West Saanich Road, Victoria, BC V9E 2E7, Canada; [email protected]
Robert E. StencelDepartment of Physics and Astronomy, University of Denver, 2112 East Wesley Avenue, Denver, CO 80208; [email protected]
Received March 19, 2012; accepted March 27, 2012
Abstract Whiletherearenumerous“newspectroscopicstudies”ofeAurigaereportedinthisspecialeditionofJAAVSO,theonesummarizedhereisbelievedtobeuniqueontwocounts:itconcentratesontheblueandnear-UVspectralregions, and it incorporates historical spectra from the previous eclipses of1983and1956.Themoredatathatcanbecollated,acrossallwavelengthandtimebase-lines,themoreconclusivethefinalmodelofthisbafflingobjectislikelytobe.Amorelengthypaperthatincludesillustrationsofthespectraisbeingpreparedforpublicationelsewhere.Thisshortcontributionsummarizestheeffort thathas so fargone intodata acquisitionandpreparation, and theprincipalresultsthatarenowemerging.
1. Data: recent CCD spectra
Wehaveobservedandanalysed~150newCCDspectraofeAurigaeatblue and near-UV wavelengths, recorded during the 2010 eclipse with thecoudéspectrographoftheDominionAstrophysicalObservatory(DAO)1.2-mtelescope.Thewavelengthspanobservableinoneexposureisabout145Å,sowetargettedspectralregionsofspecificinterest,andmonitoredjust thoseasopportunitypermitted.Thegreatmajoritywascentrednearl3950Åsoastoinclude theCaiiHandK linesand thenearbystrongground-stateFei lines;somewhatlessfrequentmonitoringwascentredonHdandspannedthestronglow-excitationFeilinesbetweenl4045–4071Å,andacomparablenumberofspectrawerecenteredneartheMgiidoubletatl4481Å,whereausefulmixtureoflow-andhigh-excitationlinesoccurs.Someobservationswerealsomadeintheredspectralregion,nearHa.Allthespectrawerereducedintheworldco-ordinatesystembyasemi-automaticpipeline,andwerethenextractedinstepsof0.01Åandlinearizedtoanapparentcontinuumheightof100%.IforiginalS/Nratioswereratherlow,sequentialspectrawereco-added.
Griffin and Stencel, JAAVSO Volume 40, 2012 715
2. Data: historic photographic spectra
Wedigitizedover130historicspectraofeAurfromMountWilson(datingback to the1930s) and from theDAO(dating from1971)using theDAO’sPDSmicrophotometer, resurrectingandadoptingprocedures thathave stoodthe test of time. Spectrophotometric calibrations were determined by meansofspecialexposurestoalightsourcethroughaperturesofknowngeometricalcharacteristics.Thosespectrawerealsoextractedinregularstepsof0.01Åandlinearizedtothelocalstellarcontinuum.ManyhadveryacceptableS/Nratios,thoughnonequiteashighascanbeachievedwithamodernCCDdetectoronabrightstar.Nevertheless,theseheritagedatacontributeuniquelytothisstudy,byrevealingwhichofthemanyline-profilechangeshaverepeatedatidenticalphasesofpasteclipses,therebyfurnishingimportantconstraintsforamodelofthestructureandformationoftheoccultingdisk.TheycouldthenbemergedwiththerecentCCDonessoastoprovideamorecompletedataset,whichweexaminedbyrangingeachspectralregionwithphase,withinthewavelengthspansdefinedbytheCCDspectra.
3. Properties of the disk
We adopted the parameters of the radial-velocity orbit of Stefanik et al.(2010)(P=9896.0±1.6days,K=13.84±0.23kms–1,T=HJD2434723±80),andalignedthespectrainthevelocityrest-frameoftheFstar.Thezerophaseofthatorbitsolutioncorrespondstoperiastron,sophotometriceclipsebeganclosetophase0.056andendedclosetophase0.130.Mid-eclipse(secondaryconjunction)wasatphase0.091. Manyauthors—e.g.,Struve(1956),Wright(1958),Hack(1962),LambertandSawyer(1986),andFerlugaandMangiacapra(1991)—havedrawnattentiontocuriousline-profilechanges,particularlyduringearlyingressandlateegress,oftenreferringtothemas“linedoubling.” (a)Theprincipalline-profilechangesaretheextraabsorptioncomponentsthatbecomesuperimposedonground-stateandlow-excitationlines.Thenewfeaturesarered-shiftedduringingressandblue-shiftedduringegress,andarebestseeninlinesofFeionaccountofthelatter’sstrengthandnumber.Thelinesthemselvesdonotactually“split,”sincethetwodistinctcomponentsoriginateinquitedifferentregionsofthesystem. (b)ComparisonsoftheFeiprofilesatdifferenttimesfurnishlimitstothephases (and hence on the geometry) when that “doubling” commenced andceased. It is apparent that the occulting disk has a “tail” which trails muchmoreextensivelythandoesanymaterialatitsleadingedge.Ourspectraalsodemonstrated the existence of more rarefied material in the extremes of the“tail,”notunlikethatinacool-starchromosphere. (c)Comparisonsofspectraofthesamephasesbutdifferentorbitalcyclesindicatethatthestructureofthediskisstable,atleastonatime-scaleofacentury
At all orbital phases the spectra of the system reveal more subtle line-profilechanges:theabsorptionlineswhichpresumablyoriginateintheF-starphotosphere(sincetheyshownophase-relatedvelocitychanges)canweaken,orbroaden,orstrengthenalittle.Dividingallspectraofagivenregionbyonerecordedfarfromeclipseaccentuatesthechanges,andshowsbroademissionfeaturesthatgrowandfadeatallphasesontime-scalesofdaysorweeks,withassociated narrowing (or deepening) of the K-line wings resembling a rise(orfall)inTeff.Fortherecenteclipse,therewassomeevidenceofcorrelationbetweenthegrowthofemissionandtheCepheid-likepulsationsof~67days’period (Kim 2008), but the recorded photometry had not been sufficientlyplentifulinthepasttoinvestigatesuchcorrespondencesinearlieryears,noraretheavailablespectraadequateforathoroughinvestigation:aquarter-phaseisonly17days,andspectramustcorrespondtonullormaximum—arequirementthatisunlikelytobefulfilledgiventheeclecticnatureofthearchivaldataforasystemofsuchlongperiod.
5. The spectrum of the disk
The absorption of the disk itself can be isolated by dividing eclipse-phase spectra by one of the system far from eclipse.The procedure cannotbe fully precise or absolute because of the small intrinsic variations in thesystem’sspectrumreferredtoabove;thediskspectrumalsocontributestothesystem’soneatallphases,thoughthosecontributionsaresmallcomparedtothe substantial absorptionwhich theedgesof thediskcreateduringeclipsephases.Theprime features thatwe thus isolateare the red-shifted (ingress)andblue-shifted (egress) absorption lines; theyvanishduring thephasesofcentral eclipse. Their respective velocity displacements are surprisinglyconstant,andwhilebothareconsiderablysharptheblue-shiftedonesappearbroaderatphaseswhichareactuallybeyondtheendofphotometriceclipse.Itisthereforeonlyattheleadingandtrailingedgesofthediskthatopticallythinmaterialissufficientlyconcentratedtogiverisetodetectableabsorptionfeatures.However,thepresenceofvariableemissionalsointhesystemrequiresmodellingbeforeaquantitativeassessmentofthepropertiesofthatabsorbingmaterialcanbemade. Qualitatively,thediskfeaturesareverysimilartotheabsorptionlineswhichariseinacool-starchromosphereandwhichcanbeisolatedduringingressor
Griffin and Stencel, JAAVSO Volume 40, 2012 717
egressphasesofan“atmospheric”eclipseinazAursystem.ThelinesineAurare ground-state or low-excitation transitions of easily-ionized elements likeTiiandFei,whereastherearenocorrespondingfeaturesinMgiil4481Å,forexample.Aswithastellarchromosphere,thetotalabsenceoflinewings,evenforhydrogenlines,indicatesasurfacegravity,logg,<1. Acomparisonbetweentheingressandegressfeaturesofthediskshowsthatthediskisnotsymmetrical:itsleadingedgeismorecompressedthanitstrailingedge.Althoughthevelocityshiftsarelargelyconstant,duringtheextremeendof theegressphases(phase~0.125onward) theblue-shift invelocitybeginstodecrease.Aroundthesametimethesharpnessofthefeaturesisparticularlyaccentuated.ThoseobservationssuggestthatweareseeingmaterialwhichisflowingalongaveryconfinedpathfromtheFstartothedisk.
Ha Spectral Monitoring of e Aurigae 2009–2011 Eclipse
Benjamin MauclaireObservatoire du Val de l’Arc, route de Peynier, 13530 Trets, France; [email protected]
Christian BuilCastanet Tolosan Observatory, 6 place Clémence Isaure, 31320 Castanet Tolosan, France
Thierry GarrelObservatoire de Juvignac, 19, avenue du Hameau du Golf, 34990 Juvignac, France
Robin LeadbeaterThree Hills Observatory, The Birches, CA7 1JF, England
Alain Lopez529 rue Buffon, 06110 Le Canet, France
Received February 22, 2012; revised May 30 2012; accepted May 30, 2012
Abstract Wepresent and analyzeeAurigaedata concerning the evolutionoftheHalineontheoccasionofthe2009Internationalobservationcampaignlaunchedtocovertheeclipseofthisobject.WevisuallyinspectthedynamicalspectrumconstructedfromthedataandanalyzetheevolutionwithtimeoftheEW(EquivalentWidth)andoftheradialvelocity.Thespectroscopicdatarevealmanydetailswhichconfirm thecomplexityof theeAur system.Theobjectis far from being understood. In particular, according to our measurements,theeclipsedurationhasbeenunderestimated.Acompleteanalysisofdetailsrevealed by our data would require much time and effort. Observers areencouragedtocontinuemonitoringtheHalineoutofeclipseinthehopethatitwillprovidefurtherimportantinformation.
1. Introduction
eAurigaeisoneofthemostintriguingeclipsingstarsystemswhichhaspuzzledastronomersfornearly200years.Themaineclipsingperiodiscloseto27.1yearsandthefirstspectroscopicsurveyswereundertakenduringthe1929and1956eclipses.Alargecampaignwasalsoorganizedfor1982–1984.Forareviewofliteraturepriortothe2009–2011eclipse,seeGuinanandDeWarf(2008).Therearealsonumerouspapersbeingpreparedasaresultofthe2009–2011eclipse.Despite the concentrated efforts, someaspectsofeAurremainamystery.
Mauclaireet al., JAAVSO Volume 40, 2012 719
eAurisclassifiedasanA8IabstarinanAlgoltypeeclipsingbinarysystem(simbaddatabase2011).TheprevailingmodelisofanF-typestarwithahotclumpyHydrogendiscandanobjectofunknownnaturewhichproducesaneclipse phenomenon lasting almost 2 years every 27.1 years.There may beamid-eclipsebrighteningbut solarproximitymakes thephotometry suspectat those times. Recently, it has been suggested that the eclipsing object is a550Kdustydiskseenedgeon,heatedonthesidefacingtheFstarto1100K.ThisdiskmaycontainaB5Vstar(Hopkinset al.2011)thatcouldcontributeto emission wings surrounding Ha line. Light curves feature 0.1 magnitudevariationsbothinsideandoutsideeclipse.VariationshavealsobeenobservedintheEquivalentWidth(EW)andRadialvelocityofspectrallinesoutsideeclipse.ThesevariationsmightbeFstaroscillationsandwind. Duringthe1982–1984eclipse,thisstarwasstudiedbyamateurobserversusing multiband photometric methods. The e Aur system was not clearlydescribeddespitealltheacquireddata. Twenty-seven years later, an international campaign was organized tomanagebothspectroscopicandphotometricobservationsbyamateurobserverswiththeaimofproducingdatawithimprovedtimeresolutioncomparedwiththat achieved during previous eclipses. In this article, we present amateurspectroscopicHalinemonitoringfromFebruary12,2008,toNovember12,2011.
2. Observations
In 2008, Jeff Hopkins (http://www.hposoft.com/Campaign09.html)organizedtheinternationalobservationcompaignofthe2009eAureclipse.Weacquiredmorethan250highresolutionspectraoftheHalinecoveringthethreeyearsaroundeclipse.Theseshowsignificantvariabilitythroughoutthisperiod.TheeffectoftheeclipseisclearlyseeninthislinefromtheendofApril2010toendofApril2011.ThespectrausedforthisstudywererecordedbyfiveobserversinEurope. Most observations were made using LHIRES3 spectrographs (LHIRES3andeShellareproductsfromShelyakInstrument,Grenoble,France:http://www.shelyak.com). C. Buil used an eShell spectrograph that covers wavelengthsfrom4500Å to7000Å.Telescopediameterswerebetween0.2mand0.3m.Spectralresolutionisabove10,000andmostofthetimearound15,000.Meanexposuretimewas2,000s.AllsetupsarereportedinTable1.
3. Reduction and analysis method
TherawobservationsareavailablefromRobinLeadbeater’seAursurveywebpage(http://www.threehillsobservatory.co.uk/).Thespectrawerereducedusingstandardprocedurestoproducecalibratedandnormalizedlineprofiles.Most of the reduction was done using SpcAudace (http://bmauclaire.free.
Mauclaireet al., JAAVSO Volume 40, 2012720
fr/spcaudace/) pipelines.The reduction steps were: preprocessing, geometriccorrections, and registration. Then line profiles were extracted with skybackground subtraction. Wavelength calibration was done using calibrationlampspectrabeforeandaftereachacquisition series tominimise theeffectsofcalibrationdrifts.The instrumental responsewas then removed.Anoffsetwas computed using telluric lines to achieve a final wavelength calibrationRMS uncertainty of 0.03Å. Finally a heliocentric velocity correction wasapplied depending on the observation date. All wavelengths l are giveninÅ(Ångström). Equivalent Width (EW) measurements were computed betweenl6550 and l6577 using linear integration and an extracted continuumobtained from a fit to the local star continuum (about EW’s computation:http://bmauclaire.free.fr/astronomie/spectro/experiences/ew/ewconvention/).The Chalabaev algorithm (Chalabaev 1983) was used to estimate theuncertainty,whichismostlydependentonthesignal-to-noiseratio,andappearsto overestimate the uncertainty compared with the actual scatter observedaroundthelongtermtrend. Radial velocity is computed in two steps because the Ha line isasymmetric:
• TheGaussianflankof the linewasreproducedon theoppositesideofthesymmetryaxis(theoreticalwavelengthoftheline)andshiftedtofittheline’soppositeflank;
A dynamical spectrum was computed using 177 spectra corrected toheliocentricvelocityandcroppedtol6550–l6575.Alinearinterpolationwasusedtoproduceanimagewitha1-daysamplinginterval.Suchinterpolationdoesn’tintroducebiasforouranalysisasthepurposeofFigure6istoshowglobal behavior of the eclipse spread over several hundred days. Dates arelogged in MJD (that is, JD–2400000). All computations were performedin SpcAudace. The monitoring covers a period of 720 days. Most of theinformationgeneratedbyourmonitoringof theHa line iscontainedin thisdynamicalspectrum. Analyzing this complex image turns out to be cumbersome, however.Thisisthereasonwhywehavesimplifiedtheanalysisbyconcentratingonthe evolutionwith timeof theEW that canbe compared toVmagnitude,andoftheradialvelocitythatcanbeusedtostudyeclipsingphenomena.Ofcoursewehave tokeep inmind thatEWloses itsphysicalmeaningwhenappliedtocomplexlineprofileslines,asitisthecaseforeAur,arelikelytoresultfromacombinationofseveralsources.Butbeforeanalyzingthesequantities, letus firstexamine thespectral lineprofilesatdates thatshowimportanttransitions.
Mauclaireet al., JAAVSO Volume 40, 2012 721
4. Behavior of the wings
Outside eclipse, the Ha line profile comprises a central absorption coreflankedbyemissionfeaturesontheredandbluewings(seeFigure1).ThesefeaturesarehighlyvariableasdescribedbyGolovin(2007). IntheregionoftheHalineareabsorptionlinesidentifiedastelluriclinesatl6543.91,l6547.71,l6548.32,l6552.63,l6557.17,l6568.81,l6572.09,l6574.85,andl6586.68. Aswecan see inFigures2and6, fromMJD55250onwards therewasadditionalabsorptioninthecorewhichbroadenedrapidly,engulfingfirsttheredemissionandbyMJD55340alsotheblueemissioncomponent.NotethisisincontrasttotheKIl7699lineabsorptionwhichstarteddecreasinginintensityduringthisphase(Leadbeateret al.2011). During ingress and into totality (see spectra atMJD55390.44andMJD55496.42) through the mid-eclipse point, the absorption core deepened andbroadenedslightlyontheredside.Theadditionalabsorptionmovedtotheblueand,atMJD55520,theredemissionfeaturereappeared. Atthebeginningofthedecreasingphase(seespectrumatMJD55627.46),the Ha line became narrower, with an emission component at the red side.Then,fromMJD55853.42,theblueedgeemissioncomponentreturnedasjustbeforeeclipse.Afterthemaineclipsephase(seespectrumatMJD55878.39),theblueandredwingswerebothpresentbutsmall,startingtoresemblethepre-eclipseprofile(Figure1). Attheendofthesurveyperiodtherestillappearstobeanexcessabsorptiononthebluesideofthecentralabsorptionregioncomparedwithtypicalpre-eclipsespectra,possiblyduetothecontinuedpresenceoftheeclipsingdisc.However,the inherent variability of this at all phases makes the statement uncertain. Today’sunderstanding(Stencel2011)isthattheFstarissemi-stableandcapableofproducingvariabilityinlinesinandoutofeclipse.Thediskisonlymodifyingtheopticalspectrumduringitspassage.
5. EW evolution with time
EWmeasurementswerecomputedbetweenl6550andl6577.Althoughthesignal-to-noiseratiovariesbetweenobservationsandincludestelluriclineswhichimpactonEW,theeffectmostofthetimeisrathersmall.ThedataqualityallowsareliableestimationoftheEW.Asmentionedearlier,thesinglequantityEWisagrosssimplificationofthecomplexnatureoftheline. However, this quantity is the integral of the distribution of luminosityversus wavelength. It can thus be compared to similar integrals such as theV magnitude.As shown in Figures 3 and 4, equivalent width (EW) and Vmagnitude(Vmag.)areanti-correlated.GiventhatEW>0forabsorptionlines,theeclipsingobjectoccultstheFstarHydrogendiskasfirstminimuminEW
Mauclaireet al., JAAVSO Volume 40, 2012722
andinVmag.evolutionarebothclosetoMJD55250,andassecondminimumand V magnitude evolution are also both close to MJD 55630. These datesdefinetotalityinnerlimits(seeTable2). Whilesecondandthirdcontacts(andmid-eclipse)timesarewell-definedbyourEWasafunctionofDatetrend(Figure3),thedefinitionoffirstandfourthcontacttimesarelessobvioushereanddonotcorrespondtophotometriccontacts:thesedates(secondandthirdcontacts)arelikelytobelinkedwiththedensestendsoftheeclipsingobject. Duringtheeclipsephaseandoutsideittoo,therearemanysmallvariationsinEWandVmag.ThissuggeststhattheoccultingobjectandFstarHydrogendiskmaybeclumpy.TheFstarmayhavealsoanintrinsicpulsatingactivity(Kempet al.1986;Stencel2011)thatproducessuchvariations. Ha EW has irregular variations like small steps during its increasinganddecreasingphases.SimilarbehaviorhasbeenobservedontheKIl7699asborptionline.Ithasbeeninterpretedasanindicationofstructures(possiblyring-like)withinthedisc(LeadbeaterandStencel2010).Continuedobservationduringegressmayhelptoclarifythis.
6. Radial velocity evolution with time
Figures 2 and 6 show how shapes are shifted in radial velocity.During the beginning phase of the eclipse (see spectra at MJD55390.44 and MJD 55496.42), the absorption line became red shifted(+14.79±1.37km/s). During the end phase of the eclipse (see spectrumat MJD 55627.46), the Ha line first returned to the position seen atMJD 55390.44 and then the absorption line still remained blue shifted(–31.59±1.54km/s at MJD 55853.42). See Table 3 for measurements atkeydates. Anemissioncomponent(Figure5)appearedinthecoreoftheHalinecloseto the restwavelength fromMJD55150onwardas theabsorption increasedinthisregion.Thisbecamemoreclearlydefinedasthesurroundingfluxleveldroppedfurtherandmovedacrosstheregionfromredtoblue.Theshapeoftheemissioncomponentisrevealedastheabsorptionregionbroadenedandsweptacrossitthroughmideclipse.Itisclearthattheconstantemissioncomponentisonlyrevealedasthesurroundingfluxleveldrops. Duringoursurvey,theemissioncomponentmeasuredbygaussianfittingremainedcenteredon6562.71±0.03Å,which, in termsof radial velocity,amountsto–4.97±1.54km/s.ThisdoesnotaccountfortheeAursystemicradialvelocityestimatedat–2.26±0.15km/s(Stefaniket al.2010),leadingtoacorrectionof+0.049Å(that is,+2.26km/s).NotethatFigure6isnotwellenoughresolvedtoseesuchasmallshift.Computationsweredoneonthelineprofiles.
Mauclaireet al., JAAVSO Volume 40, 2012 723
7. In quest of new models
ThelightvariationsofeAurinandoutofeclipsehavebeentheobjectofmanystudies. During1983eclipseKempet al.(1986)analyzedpolarizationdata.TheysuggestedthattheFstarisanon-radialpulsatorandthatitssurroundingdiskistiltedwithrespecttotheorbit. In 1991, Ferluga and Mangiacapra (1991) suggested that to explain theshape of the light curve, the disk is not a continuous aggregate of dust, butinsteadaseriesofringswithaCassini-likedivision.Thismodelwasmoreorlessvalidatedbyobservations. Duringthiseclipse,awidevarietyofobservationshavebeenundertaken:infrared, ultraviolet, interferometry, photometry, and high resolution spectralmonitoring.Thus,aconsiderableamountofinformationisnowavailable(seeStencel2010;Hopkinset al.2011foranoverview).Itnowremainstodevelopamodelthatfitsallthedataathand.Undoubtedly,thehighresolutionspectralmonitoringdatawillbeveryimportantforconstrainingthesesmodels.
8. Conclusions
Our Ha monitoring ofeAur shows that, contrary to what was forecast,the effects of the eclipse extended beyond December 2011. Post-eclipseobservationsareneeded.R.Stencelwelcomesanyoutsideeclipsespectroscopiccontributions to thecampaignover thecomingmonthsandyears,especiallythosecoveringtheNaDlines. Wehaveobservedsimilaritiesanddiscrepanciesbetween theEWandVmagnitude evolution with time. The discrepancies remain to be explained,but that isbeyond thescopeof thisarticle.Wealsowereable todefinekeydates in theeclipsingphenomenon.However,much remains tobeanalyzed.ObviouslytheHamonitoringbringsalotofinformationwhichshouldplacemanyconstraintsonthemodelsconceivedbyscientistsabouteAur. Amateur spectroscopists are now able to monitor bright targets with aspectralresolutionofabout15,000.Suitablyequippedamateursconstituteateamwith long termmonitoringcapacitywhich iswidelydistributedovertheplanet. Inanycase,wehopethatthisinformationwillhelpscientiststosolvethemysterieshiddenbehindthisfascinatingobject.
9. Acknowledgements
This campaign would not have been possible without Jeff Hopkins’motivation and constant efforts to encourage amateur observations. Theauthors acknowledge Robert Stencel for his help and encouragements
Mauclaireet al., JAAVSO Volume 40, 2012724
thoughoutthiscampaign.Theyalsokindlythanktherefereeforconstructivecommentsandveryhelpfulwork. Allofthedatapresentedinthispaperwereobtainedfromamateurobservers.The authors also acknowledge Eric Barbotin, Stéphane Charbonnel, ValérieDesnoux, Stanley Gorodenski, Keith Graham,Torsten Hansen, James Edlin,BrianE.McCandless,ÉricSarrazin,LotharSchanne, JoséRibeiro,FrançoisTeyssier,OlivierThizy,andJohnStrachanforalltheworktheycarriedoutatotherwavelengthsandspectralresolutionsonthismysteriousstar.
Figure 3. Plot showing evolution with time of Equivalent Width computedbetweenl6550andl6577.NotethetwominimaatMJD55250±2and55630±2correspondingtosecondandthirdcontactsdates.
Mauclaireet al., JAAVSO Volume 40, 2012 727
Figure 4. Light curve of e Aur using data from the AAVSO InternationalDatabaseshowingVmagnitudeevolutionwithintime.LightcurvecourtesyofAAVSO.
Figure5.SpectrumofeAur showingemissioncomponent atbottomofHaabsorption line. Observations of 12.939/7/2010, R. Leadbeater, TN 0.25m,Lhires32400g/mm,0.107Å/pixel.
Mauclaireet al., JAAVSO Volume 40, 2012728
Figure 6. Dynamical spectrum of Ha line from JD 2455120.605 to JD2455840.48. Interpolationbetween spectrawasused toget a smooth image.SinusoidallinesatbothHalineedgesaretelluriclineslyinginlineprofilesthatarecorrectedfromheliocentricvelocity.Theblackdashesalongtherightaxisshowtheactualobservationdates.
Leadbeateretal., JAAVSO Volume 40, 2012 729
High Cadence Measurement of Neutral Sodium and Potassium Absorption During the 2009–2011 Eclipse of ε Aurigae
Robin LeadbeaterThree Hills Observatory, The Birches, CA7 1JF, UK; address email correspondence to R. Leadbeater at [email protected]
Christian BuilCastanet Tolosan Observatory, 6 Place Clemence Isaure, 31320 Castanet Tolosan, France
Thierry GarrelObservatoire de Foncaude, Juvignac, France
Stanley A. Gorodenski9440 E. Newtown Avenue, Dewey, AZ 86327
ofstructurewithin theeclipsingobject.TheFstar is totallyobscuredby theeclipsingobjectattheNaDwavelengthduringeclipse.TheradialvelocityoftheFstarandthemeanandmaximumradialvelocityoftheeclipsingmaterialinfrontoftheFstaratanygiventimehavebeenisolatedandtrackedthroughouttheeclipse.Thequasi-periodicvariationsseenintheFstarradialvelocity(RV)outsideeclipsecontinuedduringtheeclipse.Itishopedthattheseresultscanbeusedtoconstrainproposedmodelsofthesystemanditscomponents.
1. Background
eAurigae is a naked eye eclipsing binary system with a period of 27.1yearsandaprimaryeclipseofabouttwoyears.Despitehavingbeenstudiedforalmosttwocenturies,ourunderstandingoftheexactnatureofthesystem,theprimarystar,anditslargelyunseencompanion,isstillincomplete. Ateacheclipse,anewgenerationofastronomersequippedwiththelatesttechnologytacklestheproblem.ArecentmultiwavelengthstudyofthesystemproposedthatthespectralclassFprimarystarismostlikelyanevolvedpost-AGBstarandthataB5Vclassstarisembeddedinthecool(~1000K)opaquematerialwhichcausestheeclipse(Hoardet al.2010,hereafterreferredtoastheHHSmodel). Interferometricmeasurementsmadeduring the2009–2011eclipsehaveestablishedthattheeclipsingobjectisanelongatedopaquecigarshapedobject,mostlikelyadiscseenalmostedgeon,whichcoversthesouthernhalfofthestarduringeclipse(Kloppenborget al.2010). Advances in sensor technology and the availability of affordable highresolutionspectrographsallowedanetworkofadvancedamateursusingsmallaperturetelescopestocontributeduringthe2009–2011eclipse.Theobjectivewas to study theevolutionof theoptical spectrumof the system throughouttheeclipsewithimprovedtimeresolutioncomparedwiththatachievedduringprevious eclipses. Over 800 spectra were collected and these are availableonline(http://www.threehillsobservatory.co.uk/epsaur_spectra.htm).
TheobservationsaresummarizedinTable1.TheK7699lineobservationswere made by Leadbeater and Schanne at a resolution of 0.35Å usingLhires IIILittrowspectrographs,modified to reach thiswavelength.Variousspectrographdesignswithresolutionsfrom0.35–0.65ÅwereusedfortheNaDlineobservations.TheMg4481linemeasurementsweremadewiththeeShelechellespectrographsofBuil,Thizy,andGarrelattypically0.5Åresolution.
2.1.Choiceoflines Outside eclipse, the eAur spectrum shows an interstellar K7699 line(Welty and Hobbs 2001). There is no detectable contribution from the Fstar spectrum. This was confirmed by examining spectra outside eclipse(Lambert and Sawyer 1986;Welty and Hobbs 2001) recorded at differentphaseswhichshowednovariationinRVabovethelevelofthemeasurementuncertainty.IfastellarcomponentwaspresentitwouldbedetectableduetothechangeinRVofthatcomponent.Afterremovaloftheconstantinterstellarcomponent,theremainingK7699lineuniquelydescribestheabsorptionduetotheeclipsingcomponent.Spectraofthislinehadalsobeentakenduringthepreviouseclipse,thoughatlessfrequentintervals(LambertandSawyer1986). TheNaDlinesalsoshowstrongadditionalabsorptionduringeclipse(Barsony et al. 1986); however, the analysis of these lines is morecomplexduetothepresenceofcomponentsfromboththeFstarandtheinterstellarmedium. TheMg4481linearisesfromtheF-starphotosphereandisabsentfromthecooleclipsingobjectspectrumdue to thehighexcitation levelof theformer(FerlugaandHack1985),soactsasareferenceforchangesintheFstarRV.
2.2.Reductionofspectra The spectra were initially individually reduced by each observer (dark,flat, geometric corrections, cosmetics removal, background subtraction, andbinning).ExceptforStober,whosespectrawerenormalisedusingafittothecontinuum, the spectra were also corrected for instrument response using astandardstar.Noatmosphericextinctioncorrectionwasapplied.Fluxisrelativetothecontinuum.ThesoftwareusedbyeachobserverislistedinTable1. ThespectrawerethenfurtherreducedandlineparametersmeasuredbyRLusingvisual spec software.Thepre-reducedspectra foreach lineof interestwerefirstnormalizedusingafittothelocalcontinuum.Tomaximizetheglobalprecision of the NaD and K7699 line wavelength calibration, the originalcalibrationswerecheckedusingtelluriclinesvisibleinthespectraandsmalloffsetsappliedasrequired.TherearenotelluricsintheMg4481lineregionsotheobservers’originalThArlampcalibrationswereusedforthisline.Theestimatedresidualuncertaintyinwavelengthis0.02Å.Thetelluriclineswerethen removed, by dividing by hot line-free star spectra taken with the same
For this paper, eclipse start and end dates of JD 2455070 and 2455800have been adopted based onV-band photometric data submitted toAAVSO(Mauclaireet al.2012).TheparametersusedforcalculatingthepropertiesofthesystemarelistedinTable2.
3.1.KI7699Åline Figure1 shows theevolutionof theK7699 line throughout theeclipse.Each row represents an interval of 2 days. The dates of the spectroscopicmeasurementsaremarkedonthetimeaxis.Intermediaterowsbetweenmeasuredspectraareinterpolated.Notethattheinterstellarcomponent(measuredfrompre-eclipsespectra)hasbeenremovedsotheplotshowsjustthecontributionfromtheeclipsingobject.Figures2and3showtypicalprofilesfortheK7699linewiththeinterstellarcomponentpresentandremoved,respectively. Figure4showsthevariationinthestrength(equivalentwidth,EW)oftheK7699lineincludingtheinterstellarcomponent,withmeasurementsfromtheprevious eclipse (Lambert and Sawyer 1986) superimposed for comparison.Figure 5 shows the same EWs for the K 7699 line with the interstellarcomponent removed. The estimated uncertainty in EW is 15mÅ, based onrepeat measurements and comparison with coincident observations made byotherobserversduringeclipseingress(Ketzeback2009). TheRVtrendfortheK7699lineafterremovaloftheinterstellarcomponentisshowninFigure6.ThemaximumvelocitycomponentinthematerialinfrontoftheFstaratanygiventimeduringingress(lineprofilerededge)andegress(lineprofileblueedge)isalsoplotted.
3.2.NaDlines Figure7showstheevolutionoftheNaDlinesthroughouttheeclipse.Eachrow represents an intervalof2days.Toproducea consistent setof spectra,higherresolutionspectrawerefilteredtogiveacommonresolutionof0.65Å.The dates of the spectroscopic measurements are marked on the time axis.Intermediaterowsbetweenmeasuredspectraareinterpolated.Figure8showstypical Na D2-line profiles throughout the eclipse, while Figure 9 shows aselection of full NaD-line profiles at a higher resolution of 0.35Å obtainedduringthesecondhalfoftheeclipse.ThetotalEWoftheNaDline(sumofD1andD2)asafunctionoftimebeforeandduringtheeclipseisplottedinFigure10. TheRVsoftheNaDlines(meanofD1andD2)andoftheMg4481lineasa functionof timebeforeandduring theeclipseareplotted inFigure11.
Leadbeateretal., JAAVSO Volume 40, 2012 733
Alsoshownisa linearfit to theMg4481linedataand,forcomparison, theradialcomponentoftheorbitalvelocityoftheFstar(Stefaniket al.2010).
4. Discussion
4.1.Theextentoftheeclipsingobject Interferometric imaginghas shown thedrop inbrightnessduringeclipsetobeduetoanelongatedopaqueobject(athinrotatingdiscseenalmostedgeon)whichcoversthesouthernhalfoftheFstar(Kloppenborget al.2010).Thechangesweseeinthespectrumduringeclipsearisefromabsorptionwithinagaseousregionsurroundingthisopaqueobject.Increasedabsorptionwasfirstdetectedin theK7699line95daysbeforephotometricfirstcontactandwasstilldetectableattheendofthesurvey,215daysafterfourthcontact.Adoptingascaleof3.8AUfortheradiusofthediscaspertheHHSmodel,weestimatethatthisgaseousregionextendsbeyondtheouteredgesoftheopaqueregionby1.2AUontheingresssideandatleast2.6AUontheegressside. OutsideeclipsetheNaDlines,acombinationofcontributionsfromtheFstarandtheinterstellarmedium,arenotfullysaturated.Duringeclipse,however,the lines saturate and the residual flux in the coreof the lines falls close tozero.This is seenparticularlyclearly in thehigher resolution spectra showninFigure9takenduringthesecondhalfoftheeclipse,inwhichtheeclipsingdisccomponentoftheline,blue-shiftedduringthishalfoftheeclipse,showsadistinctlyflattenedbottomatjust0.03ofthefluxrelativetothecontinuum.GiventhattheopaqueregioneclipsesonlythesouthernhalfoftheFstar(asseenintheinterferometricimages)thissuggeststhattheNa-absorbingregionextendsatleasttheF-starradius(0.6AU)abovetheopaqueregion,completelycoveringtheFstar. Asmallpeakisvisible,~10km/sbluewardsoftherestwavelength,inthehigh resolution NaD absorption lines during the second half of the eclipse(Figure9).Thiswasalsoobservedduringthepreviouseclipse(Barsonyet al.1986).Itislikelythatthisisduetoapartialseparationoftheeclipsingdisc/F-star components (bothblue-shiftedduring thishalfof theeclipse) and theinterstellarcomponent.Thesameeffect isseen in theK7699 linewhere theseparationofthetwocomponentsisclear(Figure2). The EW of the Na D lines at the end of the campaign period was stillsignificantly higher than pre-eclipse values (p << 0.001, mean 2.20Å range2.18–2.24,8valuesRJD55954–56005asopposedtomean1.79Årange1.67–1.86,6valuesRJD54455–54779);however,thiscouldbedueatleastinparttotherelativeseparationofthecomponents.(Becausethelineissaturated,theEWwilldependontheoverlapbetweenthecomponentswhich,inturn,dependsontherelativeRV).Extendedpost-eclipsemonitoringcouldhelpclarifythis. TheEWofboththeK7699andNaDlinesdroppedaroundmid-eclipse.InthecaseofthesaturatedNaDlinethiscouldbecaused,forexample,bya
Leadbeateretal., JAAVSO Volume 40, 2012734
narrowingofthealreadysaturatedlineprofileratherthananetreductionintheamountofabsorbingmaterial.ThiswouldnotbethecasefortheunsaturatedK7699line,however,andsuggeststhateitherthereislessabsorbingmaterialintheseinnerregionsofthediscorthattheconditionsclosertothecentralstardonotallowtheparticulartransition(duetoradiationfromthecentralstar,forexample).The intervalbetween the leadingand trailingEWmaxima for theK7699lineis265days,whichcorrespondstoaregion2.1AUindiameter. ThereissignificantasymmetryintheK7699lineexcessabsorptionbetweentheleadingandtrailingregionsofthedisc.Thetrailingmaximumis70%higherthantheleadingmaximum(780mÅasopposedto460mÅ,seeFigure5).Asalreadymentioned, the tailof theabsorptionextendssignificantly furtherontheegressside.Thelateralextentofthemainregionofabsorption,however,issimilarforbothhalves(200daysfrom30%tomaximumabsorptionduringingresscomparedwith220daysduringegress). TheminimumfluxintheK7699lineprofileremainedsignificantlyabovezeroduringtheeclipse(theminimumlevelwas0.18relativetothecontinuum)so,providedthatthematerialproducingthislineextendsabovethedisctothesameextentasthatfortheNaDline,coveringtheFstarcompletely,wecanconcludethatwhenwelookattheK7699lineweareseeingthroughthefullthicknessofthematerialand,therefore,thelineprofileincludescontributionsfromalldepthswithintheabsorbingregioninfrontoftheFstaratthetime.
4.2.Comparisonwiththepreviouseclipse ThereisgoodoverallagreementbetweenthetotalK7699lineEWtrendduring this eclipse andmeasurementsmadebyLambert andSawyeron thislineduringthepreviouseclipse,offsetby9,896days.ThelargestdifferencesoccuratRJD55340and55615butLambertandSawyerplottedasmoothcurvethroughtheirmoresparsedataandattributedtheresidualscattertomeasurementerror,soitisnotclearifthesedifferencesaresignificant.
4.3.Structurewithinthedisc The trendof excessK7699absorptionduring eclipse (Figure5)didnotprogresssmoothlyduringingressandegressbutproceededinaseriesofsteps.Thestepsduringingresshavebeeninterpretedasanindicationofstructurewithinthediscmaterial(LeadbeaterandStencel2010).Thereisnoobvioussymmetryinthesefeaturesbetweeningressandegress,asmightbeexpectedforasimplesystemofconcentriccircularrings,thoughmorecomplexstructuressuchasanellipticalsystemorspiraldensitywavesproducedbythetidalinfluenceoftheFstar,asseeninothercircumstellardiscs(forexample,Mutoet al.2012),arenotruledout.(NotethatfeaturesarealsoseenintheNaDlinetotalEWtrendbutthesedonotcorrespondwiththestepsseenintheK7699EWandmaybecausedbyinteractionsbetweenthevariouscomponentswhichmakeuptheline.Theseinteractionsarediscussedinsection4.5.)
Leadbeateretal., JAAVSO Volume 40, 2012 735
4.4.Thediscrotation ThetrendoftheK7699lineRVafterremovaloftheinterstellarcomponent(Figure 6) is smooth throughout the eclipse with little scatter (except at thestartandendofeclipsewhenthelineisweak)andnoobviousshorttimescalefeatures.Giventheknownquasi-periodicvariabilityseenintheRVoftheF-starlines(Stefaniket al.2010),thisisadditionalconfirmationthattheK7699lineisnotsignificantlycontaminatedbyanycontributionfromtheFstar.Sincewe are seeing through the full depth of the disc at this wavelength, the RVvaluesplottedhereareameasureofthemeanvelocityalongourlineofsightofallthevectorsoftheorbitingcomponentsoftheabsorbingmaterialinfrontoftheFstarataparticularphaseoftheeclipse.Whileitmayprovepossibletomodelsuchaparametergivenadetaileddescriptionof thepropertiesanddistributionofthematerialsurroundingthedisc,aperhapssimplerparameter,themaximumvelocitycomponentintheline,hasalsobeenplottedinFigure6.This is somewhat more difficult to measure as it involves estimating thewavelengthof theedgeof thelinewhereitmeets thecontinuum,andthis isreflected in the increasedscatter. Ithas thepotential,however, tobeused tocalculateanorbitalvelocitycurveforthediscsince,formaterialinKeplerianorbits,themaximumradialvelocityatagiventimewillbeduetotheinnermostmaterialinfrontoftheFstaratthattimeviewedtangentiallyand,hence,givesa direct measurement of the orbital speed of that material.This has alreadybeenattemptedusingthedatafromthefirsthalfoftheeclipse(LeadbeaterandStencel2010)wherea figureof5.3 solarmasseswasestimated for thedisccomponent.However,thisvalueissensitivetotheRVadoptedfortheeclipsingcomponentasawholeduetoitsorbitalmotion,whichisnotknowncurrently.We speculate that this orbital motion is, at least in part, responsible for theasymmetryseenintheRVcurvefromingresstoegress,(highervaluesofRVseenonegressandadecliningRVinthelaterpartoftheeclipsecomparedwiththelevelRVduringingress);however,afullorbitalsolutionforthesystemwillbeneededtoclarifythis.
4.5.TheradialvelocityoftheNaDandMg4481lines TheNaDlinesareacombinationofcomponentsfromtheeclipsingdisc,the interstellar medium, and the F star. The interstellar component will beconstantasfortheK7699linebuttheFstarRVwillhaveanorbitalcomponentandquasi-periodicvariations(Stefaniket al.2010).Thenetresult(Figure11),althoughbroadlyshowingthesameswingfromredtoblueduringeclipse,isquite different in detail from that of the disc component extracted from theK7699line. NotenoughisknowncurrentlyabouttheinterstellarorF-starcomponentsto allow the net effect of the eclipsing disc to be isolated in the NaD line.This might be possible in the future using more data obtained outside ofeclipse. (Throughout the orbit, the NaD line F-star component will move
Leadbeateretal., JAAVSO Volume 40, 2012736
independentlyoftheinterstellarcomponentsoitshouldbepossibletoisolatethetwocomponents.)TheNaDlinealsobecomessaturatedduringtheeclipsesothelineprofileisnotasimplelinearcombinationofthethreecomponents.Nonetheless,wecanmakesomequalitativeobservations. TheMg4481lineisproducedbytheFstaronly(FerlugaandHack1985)sotheRVofthislinecanbeusedasareferencefortheF-starNaDcomponentRVduringeclipse.TheMg4481linedataclearlyshowthatthequasi-periodicvariationsinFstarRVseenoutsideeclipsecontinuedduringtheeclipse.NoteinFigure11thattheNaD-lineRVfollowsthesevariationsintheintervalRJD55080–55150.Thedisccomponentthenstartstodominate,rapidlyincreasingthe RV to approximately +16 km/s before reversing through mid eclipseto approximately –23 km/s. There is again some correlation between Na DandMg4481RV in the intervalRJD55570–55700, but thedisc absorptionstronglysaturatestheNaDlinesduringthisperiodsothesensitivitytotheF-starvariationsisexpectedtobelow.ByaboutRJD55800theRVhadreturnedclosetotheFstarRVasmeasuredbytheMg4481line.
5. Further work
Additionalspectratakenoutsideeclipseareneededtodeterminetheextentoftheabsorptionduetotheeclipsingdiscontheegresssideandtoseparatetheinterstellar,F-star,andeclipsingdisccomponentsintheNaDlines. Theechelle spectrausedhere tostudy theNaDandMg4481 linesalsocovermanyother linesknown to showchangesduringeclipse (FerlugaandHack 1985).A similar analysis of these lines may reveal more informationaboutthestructureandconditionswithintheeclipsingdisc.
6. Acknowledgements
WearegratefultoJeffHopkinsfororganizingtheInternationalCampaignfor the observation of eAur during this eclipse and to all who submittedobservations. We thank the eAur spectral monitoring team atApache PointObservatory(W.Ketzeback,J.Barentine,et al.)forallowingusaccesstotheirK7699linedataduringingress.WeacknowledgewiththanksthevariablestarobservationsfromtheAAVSOInternationalDatabasecontributedbyobserversworldwide and used in this research. R.E.S. is grateful for the bequest ofWilliamHerschelWombletotheUniversityofDenverinsupportofastronomy,andforsupportunderNationalScienceFoundationgrant#AST1016678totheUniversityofDenver.
Leadbeateretal., JAAVSO Volume 40, 2012 737
References
Barsony,M.,Mould,J.R.,andLutz,B.L.1986,Publ. Astron. Soc. Pacific,98,637.Ferluga,S.,andHack,M.1985,Astron. Astrophys.,144,395.Hoard,D.W.,Howell,S.B.,andStencel,R.E.2010,Astrophys. J.,714,549.Ketzeback,B.2009,privatecommunication,ApachePointObservatory.Kloppenborg,B.,et al.2010,Nature,464,870.Lambert,D.,andSawyer,S.1986,Publ. Astron. Soc. Pacific,98,389.Leadbeater,R.,andStencel,R.2010,arXiv:1003.3617v2[astro-ph.SR].Mauclaire,B.,et al.2012,J. Amer. Assoc. Var. Star Obs.,40,718.Muto,T.,et al.2012,Astrophys. J., Lett. Ed.,748,L22.Parthasarathy,M.,andFrueh,M.L.1986,Astrophys. Space Sci.,123,31.Stefanik,R.P.,Torres,G.,Lovegrove,J.,Pera,V.E.,Latham,D.W.,Zajac,J.,
Figure7.PlotshowingtheevolutionoftheNaDlines(darkercolorssignifyincreased absorption). Tick marks on the right-hand y-axis indicate actualmeasurements. Intermediate rows are interpolated.The gaps are the periodsaroundsolarconjunction.
Figure 9.A selection of NaD-line profiles covering the second half of theeclipseat0.35Åresolution.
Leadbeateretal., JAAVSO Volume 40, 2012742
Figure 11. Radial velocity of the NaD lines (mean of both lines) and theMg4481line.AlinearfittotheMg4481linedataandtheF-starRV(basedonStefaniket al.2010)arealsoshown.Allvelocitiesareheliocentric.
Figure 10. NaD-line equivalent width (sum of both line components) as afunctionoftimebeforeandduringeclipse.
Gorodenski, JAAVSO Volume 40, 2012 743
Spectroscopic Results From Blue Hills Observatory of the 2009–2011 Eclipse of e Aurigae
Stanley A. Gorodenski9440 East Newtown Avenue, Dewey, AZ 86327; [email protected]
Received February 7, 2012; revised February 17, 2012; accepted February 28, 2012
Abstract The purpose of this paper is to report spectroscopic results ofeAurigaeduringthe2009–2011eclipse.SpectraofthesodiumDlinesandanabsorptionlineoccurringatapproximately5853ÅweretakenfromFebruary13,2010,toOctober10,2011,withanLHIRESIIIspectrographanda16-inchMeadetelescopeatBlueHillsObservatoryinDewey,Arizona.Equivalentwidthandradialvelocitydatasupportthepresenceofavoidorringstructurewithintheeclipsingdisk,andtheysupportacentraldiskclearingaroundanunseenprimarycentralobject.Theresultsalsoindicatethediskdoesnotendatfourthcontact but continues for a significant distance.Analysis of radial velocitiesdemonstratedtheprofileofthe5853ÅlinehasadiskcomponentinadditiontotheprimaryF0starcomponent.Asplitlineatthislocationwasobserved.Fromtheequivalentwidthprofileofthe5853Ålinethedurationofthesplitlineeventwasestimatedtobe101days.Otherlesserresultsarepresentedanddiscussed.
1. Introduction
eAurigaeisaneclipsingbinarysystemwithaperiodof27.1yearsandaneclipsethatlastsalmosttwoyears.TheprimarystarisanF0staralthoughrecentworkindicatesitmaybeahighlyevolvedobject(Leadbeater2011).TheeclipseisthoughttobecausedbyadiskofdustrotatingaroundanunseencentralobjectwhichisinorbitaroundtheprimaryF0star.Thelasteclipseandcampaignwas1982–1984.Forthe2009–2011eclipseaninternationalcampaignwasorganizedbyDr.RobertE.StencelandJeffreyL.Hopkins.Thereisamajordifferencebetweenthiscampaignandthepreviousone.Alotoftheamateurcontributiontothepreviouscampaignwasintheformofelectronicphotometry.SincethenarangeofhighperformanceCCDcamerashavebecomeavailabletotheamateur,aswellasrelativelyinexpensivehighresolutionspectrographssuchasSBIG’sSGS spectrograph and Shelyak’s LHIRES III and eShel spectrographs. TheauthorownsanLHIRESIIIspectrographwitharesolutiongreaterthan18,000.At the encouragement of Jeff Hopkins, the author worked on the sodium Dabsorptionlinesforthiscampaign. This paper consists primarily of observations and analyses with somediscussionandinterpretation.AlthoughthesodiumDabsorptionlinesfromthediskareconfoundedwithabsorptionlinesfromtheprimaryF0star,noattemptwasmadetosubtracttheF0contributionfromthespectra.Doingsorequiresan
Gorodenski, JAAVSO Volume 40, 2012744
out-of-eclipseestimateoftheF0absorptionlines.Thiscanbedone,butusingitrequiresonetoassumetheF0radialvelocitiesandequivalentwidthsdonotchangeduringtheeclipse.Theauthorprefersnottomakethisassumptionandsoanout-of-eclipsecomponentwasnotsubtracted.Becausesuchasubtractionwould have involved a constant component in any event, it is felt one canstill obtain meaningful results and interpretations. In addition to the sodiumD lines, this paper also includes a line located at approximately 5853Å.Anevaluationofanout-of-eclipsespectrumtakeninJuly2008byOlivierThizyhasnotbeensuccessful in identifyingthis line.Onepossibility isBariumII.(Theprofileis20080727-031309-epsAur-5x300s_P_1C_FULL.fitandisfoundattheeAurspectraldatabasemaintainedbyRobinLeadbeaterathttp://www.threehillsobservatory.co.uk/astro/epsaur_campaign/epsaur_campaign_spectra_table.htm.)Inlieuofadefiniteidentification,itwillbereferredtoasthe5853Ålineinthispaper.Itexhibitsanunexpectedlinesplitwhichwillbediscussed. Thesodiumdoubletconsistsoftwolines,D2 (λ = 5889.95Å) and D1 (λ = 5895.924Å).Figure1isatypicalspectrum,normalized,ofthesodiumDlinesregion, about 175 Ångstroms wide, taken on February 26, 2010. Figure2contains profile examples to illustrate the evolution of the D lines. It alsoillustratestheresolutioncapabilitiesoftheLHIRESIIIspectrograph. The data contained inTables 1, 2, 3, and 4, are available for downloadfrom theAAVSO ftp site at ftp://ftp.aavso.org/public/datasets/jgoros402a.txt,jgoros402b.txt,jgoros402c.txt,andjgoros402d.txt.
2. Instrumentation and methods
Theworkwasconductedattheauthor’spersonalobservatory,BlueHillsObservatory(http://users.commspeed.net/stanlep/homepagens.html),locatedinDewey,Arizona.TheequipmentconsistedofaMeade16-inchLX200Rtelescope(vintage2006),theLHIRESIIIspectrograph,anST8-XMEcameraforimagingthe spectra, and a Meade DSI I camera for guiding.The spectrograph’s slitwidthwassetat22-microns,whichenabledthespectrographtoobtainhigherresolutionspectrathanpossiblewiththewiderwidth,usuallyaround30to35microns,normallyusedbyownersoftheLHIRES.Variousguidingsoftwarewas tried and tested.The major ones used were Meade’s envisage and phd.ccdsoftcapturedtheimages,iris(version5.57)wasusedtoreducethespectra,and vspec and audela were used to calibrate and do additional processing.Greatcarewastakentoobtainverygoodcalibrationsusingtelluriclines.Aftercalibrationthetelluriclineswereremoved(usingvspec)andthespectrawerethenheliocentriccorrected.Afterallthisprocessing,thereducedspectrawereanalyzedusingspss,astatisticalsoftwarepackage.Thegraphsandtablesinthispaperwereproducedwithspss. Table1givesthedatesthespectraweretaken.Becauseoftheshortintegrationtime—onlytenminutes—manytimesmorethanonespectrumpernightwere
Gorodenski, JAAVSO Volume 40, 2012 745
obtained.The result was a total of seventy spectra over the time period theauthorparticipatedinthecampaign.Thedatapointsoftheseadditionalspectrapernightcanbeseeninthegraphsinthispaper.
3. Equivalent width
Equivalent width (EW) was estimated using the method described byGorodenski (2011).With this method a continuum is estimated with a leastsquarespolynomialregressionmodel.Thebeginningandendpointsofalinefor thepurposeofcomputinganEWaredefinedwhere the line is judgedtocome in contact with, or cross through, the estimated continuum.At timesduring theevolutionof theNaD lines, themidpointbetween theD2andD1linesdropsbelowtheestimatedcontinuum.WhenthishappenedthecrossoverpointsforestimatingtheequivalentwidthsweretakentobemidwaybetweentheDlines. Table1hastheequivalentwidths(inÅngstroms)and95%upperandlowerconfidencelimitsfortheNaDlinesandthe5853Åline.Figures3and4areplotsoftheequivalentwidthsoftheNaD2andNaD1lines,respectively. TheestimatedcontactdatesoftheeclipseinBbandfromHopkins(2012)are:
Theapplicabledates(thelastthree)areshowninallthegraphsinthispaper. Thediskinclinationhasbeenreportedtobenearlyedgeon(HopkinsandStencel2008)andatmost90°±2°(Kloppenborg2012).As thediskmovesacross(eclipses)thestar,theamountofdiskmaterialcontributingtoaspectrumshould be at a minimum at mid-eclipse, gradually increase to a maximumat third contact, and after third contact it should decline. One would expectequivalentwidthstoconformtothesechangingamountsofdiskmaterial.TheprofileoftheequivalentwidthsinFigures3and4generallyagreeswiththeseexpectations.Theequivalentwidthsareatalowpoint(althoughnotthelowestaswillbeshortlydiscussed)atmid-eclipse,andtheygraduallyincreaseaftermid-eclipseuntil they reachahighpointat thirdcontact.After thirdcontacttheyexhibitadeclineasexpected.However, thedeclineappears tocontinuewellafterfourthcontact(150daysbeyondfourthcontact)whentheprimaryF0starshouldbefarclearofthedisk.Therelativelyconstantslopeofthedeclinefor150daysafter the star reached fourthcontact indicates thediskmaterialdoesnotabruptlyendatthispoint,butgraduallydisappears. There are a few irregularities in the graphs that appear to be real, thatis, theyarenotasamplingerrororaproblemwith thedata.One isadipat
Gorodenski, JAAVSO Volume 40, 2012746
approximatelyJanuary16,2011.Theauthorinterpretsthisasavariationintheamountofdiskmaterial,meaningthediskmaterialisnotdistributeduniformlyacrossthediskandmaybepatchyorclumpy.ThisagreeswithLeadbeateret al.(2011)whointerpretstructurewithinthediskasbeingresponsiblefortherateofincreaseofEWduringingress. Theotheranomaly,amajorone,isatthebeginningoftheseries.Theseries(profile)ofequivalentwidthsstartatasignificantlylowerlevel thanatmid-eclipse.There isadefinite increasing trend thatstarts fromthebeginningofthe series and peaks on March 28, 2010. (Based on the author’s experiencewithregressionanalysis,astatisticallysignificantpolynomialregressionmodelwouldfitthispartoftheseries.Itshouldbeobvioustothereaderthatthistrendisnotrandomerror.)ThisanomalyoverthesametimeperiodisalsoseenintheradialvelocitygraphsofD2andD1inFigures6and7,respectively,inthesectiontofollowonradialvelocities. TheEWandradialvelocitydatatogetherareevidenceofaringstructuretothedisk(LeadbeaterandStencel2010;Seebodeet al.2011),oralargevoidinthedisk(theauthor’shypothesis).Ifaringgaporvoidexists,therewouldbelessmaterialcontributingtoalinestrengthand,hence,theequivalentwidthwouldbesmallerinvalue.AsthediskmovesfromfirstcontacttothepointwheretheprimaryF0starjustclearstheinnerboundaryofthedisk,radialvelocitiesshouldincrease, not drop, according to Kepler’s Law of Planetary Motion.A largevoidorringgapexplainsthedropinradialvelocitiesforthefollowingreason.Becausetherewouldbenodiskmaterialinthisspacetherewouldbenodiskmaterialmovinginaradialdirectionawayfromtheobserver(theradialvelocityvaluesarepositiveandsothemovementwouldbeawayfromtheobserver).Thecontributiontotheobservedradialvelocitiesthencanonlycomefromthepartsofthediskoneithersideofthevoid.Whattheobserverseesistheradialcomponent(theothercomponentisthepropermotioncomponent)ofthediskasitrotates.Thisradialcomponenthasasmallervelocityvaluethanmaterialmoving directly away from the observer would have at the location of thehypothesizedvoidorringgap.AsthevoidorgapmovesacrosstheprimaryF0star,radialvelocitiesstartincreasingasdustmaterialmovesintothelineofsight. Figure5isagraphofthe5853Ålineequivalentwidths.ItcanbeseenthattheEWprofile ismuchmorevariable.However,a secondorderpolynomialregressionmodel,showninthegraph,wasstatisticallysignificantatthe0.001level,indicatingthe5853ÅlineprofilegenerallyfollowsthesametrendastheNaDlines.(AsecondorderpolynomialwasstatisticallysignificantforbothDlinesandinthesamedirectionasthe5853Åpolynomialbutisnotdisplayedinthegraphsbecauseofaconcernoftoomuchgraphclutter.)Thelargedropthatbottomsoutbetweenthirdandfourthcontactcanbeexplainedbyasplitlinewhichwillbediscussedlater.
Gorodenski, JAAVSO Volume 40, 2012 747
4. Radial velocity and the 5853Å split line
An estimate of radial velocity requires an estimate of the center of aspectralline.Fourdifferentkindsofestimatesweremadewiththeassistanceofvspec.Theyfallintotwomajorcategories:EWversusVisual.EWmeansthat thebeginning and endpoints of a line for line center estimationwerethesamebeginningandendpoints thatwereused forestimating the line’sequivalentwidth(usingthemethoddescribedinGorodenski2011).BecauseofthefrequentoccurrenceofalongwingonthebluesideoftheD2line,itwasfeltthismethodmightnotgivegoodlinecenterestimates.Therefore,aseparatesetofestimateswasmadebasedonavisual“guess”ofthebeginningand end points of the lines in vspec. An attempt was made to minimizeinclusionofasymmetriclongwings.AlthoughthisVisualestimateisdifficulttodefine,anattemptwasmadetobeasconsistentfromspectrumtospectrumasmuchaspossible. Withineachofthesetwocategories,thatis,EWandVisual,twokindsoflineestimateswereperformed in vspec.Onewasabarycenter lineestimate,whichwillbecalled“Barycenter,”producedbycheckingaboxinvspeccalled“LineCenter.”Theotherisa“Gaussianfit”estimate.However,vspecusestheline location of the barycenter of the Gaussian fitted line as the line center,thatis,thelinecenterestimateisnottheestimatedparameteroftheGaussianfunction.However,itwillstillbecalledthe“Gaussian”estimate. ThelinecenterestimatesofeachofthesodiumDlinesconsistedofthefourtypesjustdescribed.However,the5853ÅlineismuchweakerthantheDlines.Becauseofthisitwasfeltgettinga“Visual”estimateforthe5853Ålinewouldhavebeenverydifficult,maybeimpossibletoconsistentlyestimate.Asaresult,thebeginningandendpointsforestimatingalinewerethebeginningandendpoints thatwereused to estimate the equivalentwidths.Consequently, therewereonlytwoestimates(incontrasttothefourfortheDlines),theBarycenterandtheGaussianestimates. Adecisionhadtobemadeastowhichofthefourestimatesforsodiumandthetwoforthe5853Ålinearethebestforcomputingradialvelocities.Todothisastandarddeviationwascomputedforeachofthelineestimationmethodsforeachelementline.Table2containsthesestandarddeviations.Forthe5853ÅlinetheBarycenterestimateshavealowerstandarddeviationthantheGaussianestimates.Hence,theBarycenterestimateswereusedforcomputingthe5853Åradialvelocities.ForbothsodiumDlinesthe“Visual”estimateshavesmallerstandarddeviationsthanthe“EW”estimates.Within“Visual,”theBarycenterestimateshavesmallerstandarddeviations.Hence,fortheDlinestheVisual-Barycenterestimateswereusedtocomputeradialvelocities.Table3containsthe line center estimates for all three lines. Table 3 also contains a vspecestimateofthelinecentersofthe5853Asplitline,thesplitlinementionedintheintroductionofthispaper.
Gorodenski, JAAVSO Volume 40, 2012748
Table4containstheradialVelocities(km/sec)forNaD2,NaD1,5853Å,andthe5853Åsplitline.Byconvention,apositivevalueindicatesaredDopplershift, that is, theobject ismovingawayfromtheobserver.AnegativevaluemeansablueDopplershift,oranobjectmovingtowardtheobserver. Figures 6 and 7 are plots of the radial velocities for Na D2 and Na D1,respectively, with a statistically significant (0.001) 2nd order polynomialregressionlineforeach.Thereareanumberofthingstonotice.First,atmid-eclipse it is expected that the radial velocity should be equal to zero. Thehorizontalsolidlineisthezeroradialvelocityline.TheestimatedpointswheretheinterpolatedNaD2andNaD1linescrossthehorizontalzeroradialvelocitylineareAugust8,2010,andAugust4,2010,respectively. However,thesearenotmid-eclipseestimatesbecausetheeAursystemismovingtowardEarthatabout2.5km/sec,±0.9km/sec.Consequently,whatmightbearedshiftradialvelocityof2.5km/secwitheAurastheframeofreferencewillappearasazerokm/secradialvelocityfromEarth,thatis,withEarthastheframeofreference.Therefore,2.5km/sechastobeaddedtotheradialvelocitiesinthispapertoconverttotheeAurframeofreference.Whenthisisdone,theauthor’sdatasuggestthemid-eclipsedatesforNaD2andNaD1areAugust18,2010,andAugust16,2010,respectively.BecausethereisnoquantumexplanationthatcouldallowtheDlinetohavedifferentmid-eclipsedates, these dates can be taken as independent estimates of the actual mid-eclipsedate.Averagingthetwogivesanestimatedmid-eclipsedateofAugust17,2010. TheincreasingradialvelocitiesfromthebeginningoftheseriestoMarch28,2010,havealreadybeenexplainedasbeingduetoapossibleringgap,orapossiblevoidinthedisk. TheradialvelocitycurvefromaboutMay6,2010,toaboutNovember1,2010,inbothFigures6and7supportsthehypothesisthatthediskhasacentralclearareaaroundanunseencentralobject.Inotherwords,thediskhasanouterand inner edge, or boundary. Based on this hypothesis, one would expect adecreasingradialvelocityfromMarch28,2010(assumingthispartofthecurveisatorneartheinnerboundaryontheingresssideofthedisk)tomid-eclipseandthenanincreasingonefrommid-eclipsetoNovember1,2010(assumingthis part of the curve is at or near the inner boundary on the egress side ofthedisk)andthisiswhatisobserved.Theexplanationforthisexpectationisthesameastheonegiventoexplainthedropinradialvelocitiesintheabovesectiononequivalentwidth.Essentially,withoutcentraldust, theobserverisseeingthediskonbothsidesofthevoid.Asaresult,whatisbeingobservedistheradialcomponent(theothercomponentbeingthepropermotioncomponent)totherotationalmotionofthedisk.Asthemid-eclipseisapproachedthisradialcomponentbecomessmaller,eventuallygoingtozero.Thereverseoccursontheothersideofmid-eclipse.Thisexplainstheshapeanddirectionofthelinebetweenthesedates.
Gorodenski, JAAVSO Volume 40, 2012 749
AscanbeseeninFigures6and7,theradialvelocitiesstillhavenotreachedthezeroradialvelocitylevelafter4thcontact.ThelastdatapointinthegraphisOctober10,2011.Thisis150dayspastfourthcontact.BythistimetheprimaryF0starshouldbewellclearofthedisk.Yet,radialvelocitiesonthisdateareover8km/sec(seeTable4).ThisisadditionalevidenceinsupportoftheEWdatathatthediskcontinueswellbeyondfourthcontact. Figure8 is theradialvelocityplot for the5853Åline. It isof interest todeterminewhetherthislineisundergoingthesameradialvelocitychangeastheNaDlines.Ifitis,thenthiswouldbeevidencethereisadiskcomponenttothe5853Åabsorptionline.ThesecondorderpolynomialsinFigures6and7fortheDlinesarestatisticallysignificant(0.001).Althoughthe5853ÅradialvelocitiesexhibitmorevariabilitythanthoseoftheNaDlines,astatisticallysignificant(0.001)secondorderpolynomialmodelalsofitsthedataandisinthesamedirectionastheDlinespolynomials.Thisisevidencethatthereisadiskcomponenttothe5853Åabsorptionline.However,thisisnotcertain.ItcouldallstillbefromtheprimaryF0star. There isonebigdifference in the5853Ålineprofilecompared to theDlines.StartingjustafterJanuary27,2011,the5853Åradialvelocitiesundergoa largeblueshiftdropandreachahighof28.866km/seconApril13,2011(spectrumnumber1inTable4).TheNaDlinesdonotexhibitsuchadrop.ThelargestblueshiftfortheD2andD1lineswere23.842km/secand23.843km/sec,respectively,onJanuary17,2011(spectrumnumber1inTable4). Thelargedropinthe5853Åradialvelocitiesiscausedbytheconfoundingeffectofthe5853Åsplitline.Figure9isagraphofthesplitlineradialvelocitieswith a superimposed statistically significant (0.001) first order polynomialregressionline.ItcanbeseenthesplitlineradialvelocitieshadalreadystarteddroppingonJanuary17,2011,andreachedahighblueshiftof27.954km/seconApril26,2011(spectrumnumber1inTable4).Thesedatesdonotexactlymatch those for the5853Å linebut areclose.Thereare two reasons for theinexactmatch:thesplitlinedataisveryincomplete,andthelinecenterscouldonlyberoughlyestimatedinvspec.Figure10showstheevolutionofthesplitlineandillustratessomeofthedifficultiesofgettinglinecenterestimates.Theprofile on December 25, 2010, was included to demonstrate the latter. Thisprofileexhibitswhatmightbecalledaplateau,notawelldefinedlinesuchastheoneonMarch28,2010. Suchaplateaumakesitimpossibletoestimateasplitlinecenter.Inaddition,therewereundoubtedly5853ÅspectrathatcontainedtheeffectsofasplitlineonEWandradialvelocity,buttheeffectwastoosmalltobenoticeableinalineprofile. Thestartandenddatesofthesplitlinephenomenoncannotbedeterminedfrom the split line data itself, but it appears a reasonable estimate might beobtained fromFigure5, thegraphof the5853Åequivalentwidths.AvisualinspectionofthegraphplacesthestartandendofthesplitlineatJanuary27,
2. Theequivalentwidthdata for theD lines support thehypothesis thatsomeofthevariationinEWaretheresultofanon-homogeneousdisk,thatis,adiskthatispatchyorclumpy.
4. Theestimatedmid-eclipsedatebasedonthesodiumDlinesisAugust17, 2010. This estimate does not differ substantially from the projectedmid-eclipsedate(madepriortomid-eclipse)ofAugust4,2010.
8. Theradialvelocitydatasupportstheexistenceofacentralclearingaround theunseenprimaryobjectof thedisk, that is, thediskhasaninnerboundary.
9. The 5853Å absorption profile contains a split line. The estimateddurationofthesplitlineeventwasestimatedtobe101days.
6. Acknowledgements
Iwould like to thankJeffHopkins forencouragingme to join theeAurCampaign, and for encouraging me to concentrate my spectroscopy on thesodiumDlinesregion.IwouldalsoliketothankJeffHopkinsforsuggestingthatIpurchasetheMeade16-inchLX200Rtelescopewaybackintheyear2005.
Gorodenski,S.2011,Soc. Astron. Sci. Newsl.,9(no.2),5.Hopkins, J. L. 2012, eAur Campaign website (http://www.hposoft.com/
EAur09/Starinfo.html).Hopkins, J.L.,andStencel,R.E.2008,Epsilon Aurigae: a Mysterious Star
System,HopkinsPhoenixObservatory,Phoenix,Arizona.Kim,H.2008,J. Astron. Space Sci.,25,1.Kloppenborg,B.2012,privatecommunication.Leadbeater,R.2011,“TheInternationalEpsilonAurigaeCampaign2009–2011.
A Description of the Campaign and early Results,” arXiv:1101.1435v1[astro-ph.SR),67.
Leadbeater, R., and Stencel, R. E. 2010, “Structure of the Disc of EpsilonAurigae:SpectroscopicObservationsofneutralPotassiumduringEclipseIngress,”arXiv:1003.3617v2[astro-ph.SR].
Figure 5.The 5853Å line equivalent widths (in Ångstroms) with upper andlower 95% confidence limits, and a superimposed second order polynomialregressionline.
March 28, 2010 September 24, 2010 December 7, 2010
December 25, 2010 January 17, 2011 January 27, 2011
February 12, 2011 March 5, 2011 March 10, 2011
March 23, 2011 April 26, 2011 May 8, 2011
May 31, 2011 September 18, 2011
Geiseetal., JAAVSO Volume 40, 2012 767
Eclipse Spectropolarimetry of the ε Aurigae System
Kathleen M. GeiseRobert E. StencelUniversity of Denver, Department of Physics and Astronomy, 2112 E. Wesley Avenue, Denver, CO 80208; address email correspondence to [email protected]
David HarringtonJeffrey KuhnInstitute for Astronomy, University of Hawaii, 2680 Woodlawn Drive, Honolulu, HI 96822
Received April 13, 2012; revised August 27, 2012, October 17, 2012; accepted November 4, 2012
Abstract Therecenteclipseoftheenigmaticbinarystarsystem,eAurigae,offered a special opportunity to explore the role of spectropolarimetryin discovery of unknown facets of the objects involved. Here we presentspectropolarimetricresultsforH-alpha,H-beta,CaI(422.6nm),andKI(769.9nm)basedonmorethan50epochsofhighdispersionspectraobtainedwiththeESPaDOnSinstrumentatCFHTduring2006–2012.
1. Introduction
The target, eAur, is a single line spectroscopic binary that features anopaquedisk,surroundingahiddencompanion,thatcausesalengthyeclipseevery 27 years—for a reading list, see, for example, Stencel et al. (2011).Instrumentationadvancesofthepastdecadehaveenabledaremarkablesetofnewspectropolarimetricdatatobeobtainedduringthe2008–2011eclipse.TheESPaDOnSinstrument(Donati2003)attheCanada-France-HawaiiTelescopeobtainedmorethan50epochsoffullStokespolarimetryfrom3800Åto10000Å.Prior efforts have revealed broadband polarization changes during eclipses,successfully predicting some disk characteristics, as well as demonstratingpost-eclipse variability (Kemp et al. 1986; Cole 2012). SpectropolarimetricobservationsmaycontributetoourunderstandingofthesystembyrevealingthenatureanddistributionofgaseousmaterialintheF-staratmosphereandintheoccultingdisk.Thispaperprovidesadescriptionofdataanalyzedtodate.Our preliminary results indicate that the increased polarization observed inbroadbandduringeclipseisalsopresentinmanyspectrographiclines.
Geiseetal., JAAVSO Volume 40, 2012768
2. Observations
Thedatawereobtainedusing theEPSaDOnS instrument at theCanada-France-Hawaii telescope (CFHT). ESPaDOnS is a cross-dispersed échellespectropolarimeterdesignedtoobtainacompleteopticalspectruminasingleexposure,witharesolvingpowerofabout70,000.TheeAurdatausedinthisreportwereobtainedwithabout106countsperspectralbin.Thenormalizedintensity (normalized to 1) corresponds to an average uncertainty of about1×10–3,forasignal-to-noise(S/N)ofc.1000.AllfourStokesparametersweretakenforeachobservation(seeTable1foralogofobservations). WeretrievedESPaDOnSobservationsofeAurfromtheCanadianAstronomyData Centre (CADC), CFHT Science DataArchive (http://www1.cadc-ccda.hia-iha.nrc-cnrc.gc.ca/cadc/).FiftyepochsofspectropolarimetricobservationsofeAur span the time from pre-eclipse (beginning February 2006) throughtheeclipsephasesandincludepost-eclipseobservations(forexample,January2012).The data were automatically reduced with Upena, CFHT’s reductionpipelineforESPaDOnS.Upenauseslibre_esprit,whichisaproprietarydatareductionsoftwaretool(Donatiet al.1997).
2.1.ContributionsofpolarimetrytothestudyofeAur Polarizationsignaturesoccurwhensymmetriesarebroken.Possiblesourcesofpolarizationare rotationdeformities, asphericalwinds, tidaldistortions inbinary systems, the Chandrasekhar effect during eclipse of binary systems,photosphericinhomogeneitiesincludingradiationinhomogeneities,andmatterstreamsoraccretion.Polarizationsignaturesduetoscatteringdependuponthenatureanddistributionofthescatterers:electrons,atomicspecies,and/ordustgrains.Broadbandpolarizationsignatures(forexample,seeCole2012;Hensonet al.2012;Kempet al.1986;Coyne1972)mayincludecontributionsfromboththecontinuumandlinesthatmayinvolvedifferentscatteringagents. The spectropolarimetric observations included in this paper revealanisotropies in gaseous atomic species intrinsic to the F star, as well asgeometric and/or scattering effects of the disk during eclipse. One possiblegeometricsourceofpolarizationinlinesduringeclipseistheChandrasekhareffectassociatedwithlimbpolarizationof theFstar.Onepossiblesourceofpolarizationout-of-eclipse isanisotropyassociatedwithstellarpulsation, forexample.Ourgoal is to identify the sourceor sourcesofpolarization in thelinesbothin-anout-of-eclipserevealedbytheseexcellentESPaDOnShigh-resolutionobservations.ESPaDOnSobservationscannotbeused tomeasurecontinuumpolarization(http://www.cfht.hawaii.edu/Instruments/Spectroscopy/Espadons/).
WhereI/Icdenotes thenormalizedintensity.Percentq (%q),percentu (%u).andpercentv (%v),whereq,u,andvarethenormalizedStokesparameters,werecalculatedinamannersimilartoEquation(2). Assessing the true errors in polarimetric measurements is crucial toestablishingthephysicsofthesource.WeassumethattheStokesparametersQandUarenormallydistributedabouttheirtruevalues,butithasbeenshownthat linear polarization follows a Rice distribution (for example, Clarke andStewart1986).TheRiceprobabilitydistributionisgivenby
(2)%p= ×100P /Ic
I /Ic
Here p0 is the underlying, true, polarization, I0 is the zeroth-order modifiedBessel function, and the underlying Gaussian noise has variance σ2. In thelimitofhighsignal-to-noise(S/N),theRicedistributionapproachesthenormaldistribution, with a mean that approaches p0 and a standard deviation thatapproaches σ (Vaillancourt 2006). The Rice mean is given by
R( p|p0,s)=ps2
exp –p2+p02
2s2 I0
pp0s2
⎤⎦
⎡⎣
⎛⎝
⎛⎝ (3)
HereL½(x) is a Laguerre polynomial of order ½, μ is the mean, and σ is the standarddeviation.TheRicevarianceisgivenby
(4)mR = √⎛⎛p
2 sL½–m2
2s2⎛⎛
Here L2½(x) is a generalized Laguerre polynomial Ln
(α)(x) with n = ½ andα = 2. The Rice standard deviation may be found by taking the square root ofthevariance.
(5)12
ps2 L2½⎛⎛–m2
2s2⎛⎛
⎛
⎛⎥
⎠
⎛⎥ ⎛
⎛⎥
⎠
⎛⎥
Geiseetal., JAAVSO Volume 40, 2012770
Bydefinition,theparameterpisapositivedefinitequantity.TheindividualpolarizationvaluescalculatedfromEquation(1)willalwaysbepositiveandnon-zerobecausetheindividualvaluesofQandUwillgenerallybenon-zero.AtlargeS/N( p /σ ≥ 4) the maximum likelihood and most probable estimators forthelinearpolarization(forexample,SimmonsandStewart1985)convergeto
Thebarredvariablesdenotethemean. Linear polarization position angle, Θ, may be computed from Stokes qanduasfollows(Bagnuloet al.2009).Positionangleswillrangefrom0to180degrees.
(7)–p= ( –q 2+–u 2)√
TheESPaDOnSdataarenoisieratshorterwavelengths(thatis,towardtheblue)thanlongerwavelengthsbecausethedetectorislesssensitiveintheblue.Webinnedthelinearpolarizationdatawithawavelengthbinof0.015nmusingtheerror-weightedmean(forexample,Taylor1997)toboostsignal-to-noiseatshorterwavelengthsandfordeepabsorptionlines.Wedeterminedthroughtrialanderror that thiswas the largestbinsize thatdidnotseriouslydegrade theresolutionoftheline.Insomecases,wecombinedseveralepochs(usingtheerror-weightedmean)tofurtherreducenoisepriortowavelengthbinning,usingthecriterionthatthelineprofiledidnotchangebetweenbinnedepochs. Wecomputed themeanforbothqanduusing thebinneddataand thencomputedthemeanlinearpolarizationusingEquation(7).TheRicemeanand
(8)12
⎛⎛ ⎛⎛Q= tan–1 uq +Q0
Q0=
0ifq>0andu ≥ 0180ifq>0andu<0
90ifq<0⎨⎧
⎩⎨⎧
⎩
Q=45ifq=0andu>0135ifq=0andu<0⎨
⎧
⎩⎨⎧
⎩
(9)
Geiseetal., JAAVSO Volume 40, 2012 771
variancewerecomputedforeachpusingEquations(4)and(5),adoptingthemean polarization for μ in those equations. Finally, we bias-corrected the linear polarization using Equation (6), adopting the Rice standard deviation for σ in thatequation.Weadoptedtheerrorbarsgiveninthepipelinereductionastheuncertainty in both normalized intensity and normalized Stokes parametersandpropagatedthoseuncertaintiesincalculationsofpercentStokes(see,forexample, error bars in Figure 2). We also propagated the uncertainty (Ricestandarddeviation) in the calculationof%p (see, for example, errorbars inFigure1). For polarizations with large S/N, the confidence regions approach thosegivenbyanormalGaussiandistributioncenteredonthebias-correctedvalueofp, with 2σ corresponding to 95% and 2.6σ corresponding to 99% confidence for p/σ ≥ 4 (Vaillancourt 2006). The maximum likelihood estimator (Equation 6) of theunderlying(“true”)polarizationconvergeswithallotherestimatorswhenp/σ is greater than 4. For p/σ greater than 3 and less than 4, the maximum likelihood estimator may not completely correct for bias; the polarization isconsidered the upper bound.Values of p/σ less than 1.4 correspond to zero polarization (Vaillancourt 2006). We identified spectroscopic regions withsignificantlinearpolarizationbyflaggingpolarizationpeaksforp/σ ≥ 4. We rotated theunbinneddataby27degreesusinga rotationmatrix (forexample,Code andWhitney1995;Bagnuloet al. 2009) to align instrumentnorthwiththerotationaxisofthesystemasdescribedbyKloppenborget al.(2010) before binning by epoch and wavelength. We confirmed that theinvariant,(Q2+U2),wasconservedunderrotation(intensityisunaffectedbyrotationand%pisunaffectedaslongastheinvariantisconserved). WeverifiedthatthenullparametersprovidedbytheESPaDOnSpipelinecontainednosignal,indicatingthatanyinstrumenteffectswereremovedbythedatareduction.WewerealsocarefultochangethesignofStokesUasdirectedbytheESPaDOnSFITSfileheaders.
4. Analysis
4.1.Linearpolarizationtimeseriesandqu-plots Initial analysis focused on the stronger lines in order to assess whethertheeclipseresultedinchangestothepolarization.Asampletime-seriesofH-alphalineprofilesandpolarizationsareshowninFigure1.Asampleplotof%qvs.%u(aqu-plot)isshowninFigure2.Theerrorbarsinthequ-plotarethe1s-propagated(assumedGaussian)averageuncertainties inboth%qand%u.Theanglefromthe+qaxismeasuredcounterclockwisetoafeatureinthequ-plot corresponds to twice the position angle (2Θ) as measured east of north andmaybeausefuldiagnosticofthegeometryofthescatteringgivingrisetothepolarization.Wewerecareful to exclude%qand%ucontributions fromneighboringlinesintheseplotswhenpossible.
Geiseetal., JAAVSO Volume 40, 2012772
4.2.Hydrogenalpha Wefit thepre-eclipseH-alpha line (restwavelength656.280nm)withaGaussianfunctionandadoptedtheHWHMoftheGaussian(c.25kms–1)asthelinecore.Wefurtherdefinedthewingsasfollows:thebluewing(–125kms–1to–25kms–1)andtheredwing(25kms–1to125kms–1).Regionsoutsideofthesedefinedareasconsistentlymappedto(0,0)inthequ-plots.Thesedefinitionsaremaintainedthroughoutthefollowinganalysis. H-alpha exhibited persistent polarization in the line core in all epochs(seeFigure1).Duringpre-eclipse,thelinecoreaccountedforanearlylinearexcursionof(–%u,–%q)inthequ-plot(greendiamondsymbol,Figure2).Thelinewas largely symmetric,withboth red andblue emissionwings evident.ThelineappearedslightlybroadenedtotheredwhencomparedtotheGaussianfunction.Theblueandredemissionwingsexhibitednopolarizationfeaturesinthisbinnedepoch,orinanypre-eclipseepoch.AtnophasedidtheH-alphalinereachzerointensity. HarringtonandKuhn(2009)notedthestrongpresenceofspectropolarimetricsignaturesinandaroundtheabsorptivecomponentsof theH-alphaemissionlineinHerbigAe/Bestarsthattheycalled“polarization-in-absorption.”Theyalsoidentifiedabroadpolarizationsignatureacrossemissionfeaturesformanyclassical Be stars in their sample. HerbigAe/Be stars are embedded in coldgasanddust,whichmaybeequatoriallyenhanced,whereasclassicalBestarsarerapidrotatorscharacterizedbyionizedequatorialmaterial.Theequatorialmaterialsurroundingthesetwotypesofstarscontributestodistinctlydifferentpolarization signatures.Thepolarization featuresweobserved in pre-eclipseeAur spectra are similar to “polarization-in-absorption”; this suggests thateAurdoesnothaveanequatorialenhancementofionizedmaterial. At mid-eclipse, and for many epochs following mid-eclipse, the lineexhibitedacentralemissionfeature(presumedrecombination,seeStencelet al.2011).%pincreased,consistentwithchangestobroadbandpolarizationreportedbyCole (2012) andKempet al. (1986).Thecorepolarizationpeakappearsnotchedin theseepochs,possibly indicatingadepolarizationassociatedwiththeemissioncore.Atmid-eclipse,theblueandredabsorptionwingsexhibitedbroad %p polarization.There were excursions in the qu-plot for features inthelinecore(–%q;greendiamondsymbol),aswellastheblue(+%q,+%u;bluesquaresymbol)andred(+%q,–%u;redtrianglesymbol)absorptionwings(Figure 2). These qu-plot excursions were not affected by binning and areevidentinseveralobservationsaroundthistime. By lateeclipse, the lineexhibitedabroad,deep,blue-shiftedabsorption.Normalizedintensitydroppedtoabout10%atthedeepestpartoftheline,butsignal-to-noise remained above 700 and the ratio p/σ ranged from 3 to 25 for polarizationgreaterthan0.4%,significantwithintheRicestatistics.Thecentralcore polarization remained strong (>1%) in late eclipse, but the blue wingpolarizationincreased(>0.5%at–100kms–1),whiletheredwingpolarization
Geiseetal., JAAVSO Volume 40, 2012 773
decreased (that is, negligible at 100 km s–1). In the qu-plot, the line coreaccountedforthe–%qfeatures(greendiamonds),thebluewingaccountedforthe+%qfeatures(bluesquares),andtheredwingpolarization(redtriangles)was largely concentratedat (0,0) (Figure2).Thus, thepolarizationbehaviorfollowedthelinebehaviorinthisepoch. TheH-alphalinedidnotreturntoitspre-eclipseformbythetimeofourlastpost-eclipseobservation(January17,2012).Theredemissionwingreappearedat about the pre-eclipse level, but the blue emission wing was masked by abroad(>–150kms–1),shallow(normalizedintensityc.0.9)absorption.Linecorepolarization remainedabove1%and thebluewingpolarization featuredisappeared. The line core polarization maintained a largely (–%q, +%u;greendiamonds)orientationinthequ-plot(seethepost-eclipsebinnedepochpresentedinFigure2). Clearly,thepassageoftherotating,darkdiskinfrontoftheFstarinducespolarizationsignalsawayfromlinecenter.ContinuingobservationsmaybeabletodemonstratewhetherthepersistentlinecorepolarizationtrackstheFstaroradisk-tiedsourcevelocityaroundtheorbit.
4.3.Hydrogenalphalinearpolarizationpositionangle Wecalculated thepositionanglefor(%q,%u)pairsfor theH-alpha lineinmid-eclipse(Figure3).Thedataarerotatedtothestellarframeandarenotbinned. Notice that the linear polarization position angle appears randomlyscattered outside of the line, which is expected. The position angles thatcorrespond to the line core (± 25 km s–1 from rest wavelength) are plottedin green; position angles corresponding to the blue-shifted absorption wing(–125kms–1to-25kms–1)areplottedinblue;andpositionanglescorrespondingtothered-shiftedabsorptionwing(+25kms-1to+125kms-1)areplottedinred.Positionanglesonlyrangefrom0° to180°becauseof thenatureof theStokesparameters;apositionangleof180°isconsistentwith0°.Noticethelinecorepolarizationisoffsetbyabout90°fromthewings.Thismaybeanopacityeffect.Comparethesedatatothemid-eclipsequ-plot(Figure2).
4.4.Hydrogenbeta Unlike the pre-eclipse H-alpha line, the H-beta line (rest wavelength486.135nm) was better fitted by a Lorentzian profile and we adopted theHWHMofthisprofile(c.35kms–1)asthelinecore.Wefurtherdefinedthewingsasfollows:thebluewing(–125kms–1to-35kms–1)andtheredwing(35kms–1to125kms–1)andnotedthatregionsoutsideofthedefinedareasconsistentlymappedto(0,0)inthequ-plots.Thesedefinitionsaremaintainedthroughoutthefollowinganalysis. H-betaalsoexhibitedpersistentpolarizationinthelinecoreinallepochs(see Figure 4). In pre-eclipse, the normalized intensity was at 13%, but thesignal-to-noise remained above 1,000. The ratio p/σ was consistently greater
Geiseetal., JAAVSO Volume 40, 2012774
than4for%pgreaterthan0.22%.The–%qexcursioninthequ-plot(Figure5)corresponds to linecorepolarizationanddiffers inorientation fromH-alphapre-eclipse(seeFigure2).Thelineitselfappearednearlysymmetricandred-shiftedbyaboutthevelocityoftheFstaratthatphase. The H-beta line maintained a deep and broad absorption at mid-eclipsethatdeepened furtherby lateeclipse.%p increased inmid-and lateeclipse.Atmid-eclipse,thenormalizedintensityfelltoabout5%atthedeepestpartoftheline,butsignal-to-noiseremainedgreaterthan500.Thelinearpolarizationwasgreatestatthelinecore,withsmallercontributionsfromtheblueandredabsorption wings. The ratio p/σ remained consistently greater than 4 for %p ≥ 0.2% in this epoch. There were excursions in the qu-plot for features in the linecore(–%q;greendiamond),aswellastheblue(+%q,+%u;bluesquare)andred(+%q,–%u;redtriangle)absorptionwings(Figure5).Thesequ-plotexcursionsaresimilartothoseexhibitedbyH-alphaforthisepoch. By lateeclipse, the lineexhibitedabroad,deep,blue-shiftedabsorption.Thelineisclearlysaturatedinlateeclipse,fallingtojust1.6%intensityatthedeepest point and, although S/N is nearly 200 here, the ratio p/σ is only 1.2; therefore %p = 0. The ratio p/σ ranged from 3 to 11 for %p > 0.4 outside the saturatedregion.Thecentralcorepolarizationappearstohavedecreased;thebluewingpolarization(outsideofthesaturatedregion)increased(>2%at–70kms–1),whiletheredwingpolarizationdecreased(c.0.2%at70kms–1)frommid-eclipselevels.Inthequ-plot,thelinecoreaccountedforthe–%qfeatures(p/σ ≥ 3 at –20 km s–1; p/σ ≥ 4 at –10 km s–1andthroughouttheremainderofthecore),theunsaturatedbluewingaccountedforthe+%qfeatures(Figure5;bluesquares)andtheredwingpolarizationwaslargelyconcentratedat(0,0).The polarization behavior also followed the line behavior in this epoch andqualitativelyresemblestheH-alphapolarization. The H-beta line did not return to its pre-eclipse form by the time ofour last post-eclipse observation (January 17, 2012). The line was deeper,broader,andblue-shiftedbyabout–20kms–1withastrong(c.1.5%),narrowpolarizationpeakcenteredonthisvelocity.Thelinewasdeep,butunsaturated,withnormalizedintensitynearly8%andsignal-to-noiseabove500.Theratiop/σ remained above 4 for %p greater than 0.5%. The line core polarization maintainedalargely–%qorientationinthequ-plot(seethepost-eclipsebinnedepochpresentedinFig.5).
4.5.HydrogengammaandHydrogendelta In pre-eclipse, H-gamma (rest wavelength 434.047 nm) appeared toexhibitlow-levelpolarizationfeatures,buttherewereinsufficientdatapointscorresponding to p/σ ≥ 4. H-delta (rest wavelength 410.008 nm) in pre-eclipse showed no polarization features. By mid-eclipse, both lines becamesaturated, making analysis of line core polarization impossible. However,both lines exhibited absorption wing polarization meeting the p/σ ≥ 4 criteria.
Geiseetal., JAAVSO Volume 40, 2012 775
These features exhibited excursions in the qu-plot of +%u for blue-shiftedabsorptionand–%uforred-shiftedabsorptionforbothlines,whichisconsistentwith both H-alpha and H-beta polarization during this epoch. The linesremainedsaturated in lateeclipseand therewerenosignificantpolarizationfeaturespost-eclipseforeitherline.
4.6.Potassium(769.896nm) TheKIline(restwavelength769.896nm)describedhereistheweakerlineofadoubletthatarisesfromthegroundstate.Theout-of-eclipselineisthoughttohaveaninterstellarorigin(WeltyandHobbs,2001).TheKI769.896nmlineexhibitednopolarizationsignaturesuntilaftermid-eclipse.WeidentifiednoF-starcontribution;varyingabsorptionandpolarizationfeaturesdescribedbelowmaybeattributedtothedisk. Inpre-eclipse,thelineprofileremainedconstant;thestellarradialvelocitywas red-shifted with respect to the line rest wavelength (see Figure 6); andnolinearpolarizationfeatureswereobserved(seeexamples,Figures6and7).AGaussianfittothelineprofilereturnedaHWHMof0.02nm. Afterfirstcontact,thelineexhibitedared-shifted(about20kms–1)featureinitiallyaboutthesamedepthasthelinecore(normalizedintensityabout0.7)thatdeepenedtoanormalizedintensityofabout0.3.Therewerenosignificant(p/σ ≥ 3) polarization features evident until after second contact. The weak (c.0.3%),narrowlinearpolarizationfeaturethatappearedaftersecondcontactwascenteredonthered-shiftedcomponentoftheline. Bymid-eclipsethelineappeareddeeperandbroaderthanduringpre-eclipseepochs.Abroad,weak(<0.2%)linearpolarizationfeatureappearedwhichwascentered on the line (see Figure 6). In Figure 6, the ratio p/σ is greater than 3 for polarizationabove0.1%andisgreaterthan4forpolarizationabove0.15%;thepolarizationissignificantbyRicestatistics.Thestellarradialvelocitywasnotshiftedwithrespecttothelinerestwavelengthatmid-eclipse. The line developed a broad, blue-shifted component after mid-eclipse.Polarizationremainedlow(below0.2%)untillateeclipse,whenthelinewasverybroad.WefitaGaussianfunctiontothelateeclipselineprofile;thefityieldedaHWHMof0.05nm,morethantwicetheHWHMmeasuredinpre-eclipse. The Gaussian centroid corresponded to a shift of –20 km s–1 fromtherestwavelength.Alarger(c.0.5%)polarizationfeaturewascenteredonthe blue-shifted component of the line in this epoch (see Figure 6), whichcorresponded toa (–%q,+%u;bluesquares) loop in thequ-plot (Figure7).The star’s radial velocity became blue-shifted with respect to the line restwavelengthaftermid-eclipse. Thelineretainedablue-shiftedcomponentafter4thcontactthatdecreasedinbreadth anddepthover time.Therewerenopolarization features evidentafter4thcontact.Thelinehadnearlyreturnedtoitspre-eclipseformbyourlastobservation(January17,2012).
Geiseetal., JAAVSO Volume 40, 2012776
We examined the stronger line of the K I doublet, (rest wavelength766.490nm) and discovered that it exhibited line profile and polarizationvariations similar to the weaker line.At late eclipse, the line exhibited abroad,blue-shiftedabsorptioncomponent.Alinearpolarizationfeaturewascentered on the blue-shifted component and the qu-plot also exhibited a(–%q,+%u)loop.
4.7.Calcium(422.673nm) TheCaIline(restwavelength422.673nm)describedherearisesfromthegroundstateandshowsapersistentcorepolarizationsignatureinpre-,mid-andlateeclipsephases(Figure8).Thelateeclipsepolarizationappeareddramaticallygreater than the pre-eclipse phase.After first contact, the Ca I line exhibitsvariationssimilartotheKIline,withanadditionalabsorptioncomponentatabout+20kms–1fromthelinerestwavelengthbeforemid-eclipseandablue-shiftedadditionalabsorptioncomponentdevelopingatabout–20kms–1aftermid-eclipse. Wefitthepre-eclipseCaIlinewithaGaussianfunctionandadoptedtheHWHMoftheGaussian(36kms–1)asthelinecore.TheGaussiancentroidwasdisplacedfromrestby0.08nm(24kms–1),butthecalculatedradialvelocityofthestarwasonly8kms–1forthisepoch.WeconfirmedtheGaussianHWHMusing data for two later epochs (20081216, 20080213) when the Gaussiancentroidandstellarradialvelocityweremorecloselyaligned.Wedefinedthewingsasfollows:thebluewing(–125kms–1to–36kms–1)andtheredwing(36kms–1to125kms–1).Regionsoutsideofthesedefinedareasconsistentlymappedto(0,0) in thequ-plots.Thesedefinitionsaremaintainedthroughoutthefollowinganalysis. Thelineappearedasymmetricinthepre-eclipseepoch(Figure8)witharedabsorptioncomponentextendingbeyondthelinecore,atabout50kms–1.Alarge(c.1%)polarizationpeakwasnearlycenteredonthestarandasmaller(0.4%)linearpolarizationpeakwasassociatedwiththered-shiftedcomponent. The presence of two peaks in %p for this pre-eclipse epoch may beattributedtooneofthefollowing:(1)thelineisopticallythickatthelinecore(aswithH-alpha),(2)morethanoneasymmetricregioncontributestotheCaI422.6nmpolarization,or(3)thelinecontainsablendofspecieswithvaryingdegreesofpolarization.Twospecies,FeI(422.743nm,3.3eV,doublet)andTiII(422.733nm,1.13eV,multiplet33),arecandidatesforpossibleblendedspecies corresponding to the velocity offset of c. 50 km s–1. The possibleblended feature persisted at about the same polarization strength throughoutthetimeseries.ThecorrespondingFeIdoublet(422.545nm)toourcandidatelineexhibitednopolarizationinanyepoch.ThetwoothermembersoftheTiIImultiplet(421.818nmand420.592nm)didnotshowpolarizationfeatures.Hack(1959)identifiedaTiIIlineofcomparableenergytoourcandidateTiIIline(TiII,454.5nm,1.13eV,multiplet30)asasolelyFstarline.Thefeature
Geiseetal., JAAVSO Volume 40, 2012 777
atc.50kms–1maybeaTiIIcomponentattributabletotheFstar. The qu-plot (Figure 9) describes two dominant position angles for thelinein thispre-eclipseepoch; theexcursionof(–%q,+%u;greendiamonds)correspondstothelinecoreandtheexcursion(+%q;redtriangles)correspondstotheredwing.Thepositionanglescorrespondingtotheexcursionsinthequ-plotdifferbyabout90degrees. TheCaI422.6nmlineincreasedin%pbymid-eclipse(peakpolarization>1% centered at the line rest wavelength. The polarization increased frommid-throughthelateeclipseandbroadenedonthebluewardsideastheblueabsorptioncomponentappearedinthespectra. Thequ-plot(Figure9)correspondingtomid-eclipsemaybecomplicatedbythepresenceofapossibleblendedline.Thelinecorecorrespondsto(–%q,+%u;greendiamonds),theexcursionof(+%q,+%u;redtriangles)correspondstotheredabsorptionwing(apossibleblend)andtheblue-shiftedabsorptioncorrespondsto(–%q,–%u;bluesquares).Asimilarscenariocorrespondstothelateeclipsequ-plot,withthenotableexceptionthatthedegreeofpolarizationhasobviouslyincreasedforthelinecenterat(–%q,+%u;greendiamonds). The%pdecreasedafter4thcontactandthelinereturnedtoitspre-eclipseshape.Thelinearpolarizationexhibitedtwopeaksin%p,onecenteredontheFstarandtheotheroffsetfromtheFstarvelocitybyabout+50kms–1,centeredonthepresumedTiIIfeature.Thequ-plot(Figure9)describestwodominantposition angles for the line in post eclipse; the excursion of (–%q; greendiamonds)correspondstothelinecoreandtheexcursion(+%q;redtriangles)correspondstotheredwing. WeexaminedotherCaI lines todiscover if thepolarizationbehaviorofthegroundstatetransitionwasconsistentwithotherenergytransitionsofthisatomicspecies.AnotherCaIline(restwavelength430.774nm)arisesfromahigherenergylevel(1.9eV)andexhibitedsimilarlineprofilechangesduringeclipse, however, there were insufficient data points corresponding to p/σ ≥ 4 to comparepolarizationchangesofthislinewithchangesobservedinthegroundstateCaIline.WeobservednolinearpolarizationsignaturesinotherCaIlinesofcomparableenergytransitionlevelstotheCaIlineat430.774nm. ThepolarizationbehaviorofthislinesuggeststhatCaI(422.6nm)tracespolarizationassociatedwiththeFstaritself,aswellasdiskeffects.Kim(2008)noted a 67-day out-of-eclipse light variation. We speculate that the F starpolarizationfeaturesweobservedmightbeassociatedwithupwellingandlargescalebrightconvectiveregions.
Thepresenceofpersistentpolarization inspectral linessuchasH-alpha,H-beta, and the Ca I 422.6 nm line out-of-eclipse suggests that asymmetrypersistsintheFstarforextendedperiods.H-alphapolarizationwasidentifiedinESPaDOnSobservationsofeAurdatedFebruary7and8,2006(HarringtonandKuhn2009),atphase0.925,morethanthreeyearsbeforetherecenteclipseand about two years before periastron. Those observations, as well as theH-alphaobservationspresentedhere,indicatedthattheblueandredemissionwingsarenotpolarized,suggestingpossiblesymmetryintheemittingregion(attheseshiftedvelocities/temperatures),orinsufficientopticaldepthtogeneratedetectablepolarizationfromscatteringintheregion.WealsoobservedthattheH-alphalinedoesnotsaturate,unlikeH-betaandothersintheBalmerseries.This suggests that a broad emission exists, contributing additional H-alphaphotonstothelinecore. WeobservedthatH-alphaandH-betaexhibiteddifferentpositionanglesinpre-eclipse.TheH-betaabsorptionmay indicate thepresenceofequatoriallyalignedhydrogengas—theopticallythickcomponentscatteringat90degreesperpendicular to the optically thin component. The H-alpha polarizationpositionanglesmayincludeacomponentfrominterstellarpolarization,ortheymayindicateamorecomplexdistributionofhydrogengasat thoseenergies.Chadimaet al.(2011)demonstratedthattheatmosphereofthediskstartstobeprojectedagainsttheFstarasearlyasthreeyearsbeforethebeginningofthephotometriceclipse;ourobservationsseemtocorroboratetheirfindings. H-alpha and H-beta showed similar mid- and late eclipse behavior inqu-space.The linecoresandwingsseem toagreeabout the rangeofanglesinvolved,suggestingthatthedominantfeaturesarisefromthesameorientationinthesky(gaseousmaterialaboveandbelowthedisk).Anoffsetof90degreesbetween linecoreandabsorptionwings is consistentwith theeffectopacitymayhaveonscattering. The polarization behavior of the K I 769.9 nm line confirmed that thislinehasnoF-starcomponent;thelinemaybeconsideredabellwetherforlowexcitationlinesaffectedbythediskduringeclipse.Limbpolarizationcannotbethesolecontributortopolarizationsignaturesduringeclipse(forexample,Kempet al.1986)becausethislinecannothavealimbpolarizationcomponent.Thedisk exhibited significantly strongerpolarization features in late eclipsethan in any other eclipse epoch. Many observers have noted the asymmetryintheline(forexample,Leadbeateret al.2012).Theincreaseinpolarizationmay indicate that thedensityof scatteringmaterial increased in lateeclipse.PearsonandStencel(2012)notethe“dawn”faceofthediskmayrotateintothe
Geiseetal., JAAVSO Volume 40, 2012 779
line-of-sightduringlateeclipseepochs.Thewarmed,presumablysublimated,materialcouldcontributetotheincreasedresonantscatteringintheline.Theposition angles of scattering in late eclipse deviate from strict equatorial orpolaralignment;modelingisrequiredtoreplicatethelate-eclipseStokes%qand%ubehavior. The pre-eclipse polarization features of the Ca I 422.6 nm line showedcontributions from the F star as well as from the eclipse. The increase inpolarizationinlateeclipseisconsistentwiththeKIlinebehaviorandmayalsosuggestanincreaseinthedensityofscatteringmaterial. Thespectralandlinearpolarizationfeaturespresentedhereareasampleofthefeaturespresentinthedataset.Onlyafewepochshavebeenpresentedforbrevity.WefoundsignificantchangestolinearpolarizationinlinessuchasH-alpha,H-beta,CaI422.6nm,andKI796.6nmpresentedhere.
6. Conclusions and next steps
The analysis present here is awork inprogress and is not a finalword.There are many spectral features whose linear polarization characteristicsremaintobedescribed.WeareoptimisticaboutthepotentialinthesedatatohelpcharacterizepolarizationfeaturesthatmaybeattributedtotheFstaritself,aswellasthepolarizationthatarisesfromtheeclipsingdisk.Ournextstepsinasubsequentpaperwillincludeidentificationandclassificationofspectralfeatures that exhibit polarization, an analysis of the position angle of linearpolarizationfeatures(tofullydescribe the linearpolarizationvector),andananalysisofscatteringbehaviorwhentheFstarisnotuniformlyeclipsed.7. Acknowledgements
TheauthorsaregratefulforsupportofthisworkinpartfromabequestinsupportofastronomyfromtheestateofWilliamHerschelWomble.BasedonobservationsobtainedattheCanada-France-HawaiiTelescope(CFHT),whichisoperatedbytheNationalResearchCouncilofCanada,theInstitutNationaldesSciencesdel’UniversoftheCentreNationaldelaRechercheScientiqueofFrance,andtheUniversityofHawaii.WewouldliketothankRobertoCasini,Elizabeth Griffin, Philip Judge, Brian Kloppenborg, Bruce Lites, and JanStenfloforhelpfuldiscussions. The authors are very grateful to the referee, John Landstreet, for manyhelpfulsuggestionsandimprovementstothework.
to Death,AIPConf.Proc.1429,Amer.Inst.Physics,Melville,NY,140.Kemp,J.C.,Henson,G.D.,Kraus,D.,Beardsley,I.,Carroll,L.,Ake,T.,Simon,
T.,andCollins,G.1986,Astrophys. J., Lett. Ed.,300,11.Kim,H.2008,J. Astron. Space Sci.,25,1.Kloppenborg,B.,et al.2010,Nature,464,870.Leadbeater,R.,et al.2012,J. Amer. Assoc. Var. Star Obs.,40,729.Pearson,R.L.,andStencel,R.E.2012,J. Amer. Assoc. Var. Star Obs.,40,802.Simmons,J.F.,andStewart,B.G.1985,Astron. Astrophys.,142,100.Stefanik,R.P.,Torres,G.,Lovegrove,J.,Pera,V.E.,Latham,D.W.,Zajac,J.,
andMazeh,T.2010,Astron. J.,139,1254.Stencel,R.,et al.2011 Astron. J.,142,174.Taylor,J.1997,Introduction to Error Analysis, the Study of Uncertainties in
Physical Measurements, 2nd Edition, University Science Books, NewYork.
Figure 3. Position angle calculated for (%q, %u) pairs for the H-alpha line in mid-eclipse.Thedataare rotated to thestellar frameandunbinned.Notice that the linearpolarization position angle appears randomly scattered outside of the line, which isexpected.Thepositionangles thatcorrespondto the linecore(±25kms–1 fromrestwavelength) are plotted in green; position angles corresponding to the blue-shiftedabsorptionwing(–125kms–1to–25kms–1)areplottedasbluestar-shapes;andpositionanglescorrespondingtothered-shiftedabsorptionwing(+25kms–1to+125kms–1)areplottedasredstar-shapes.Positionanglesonlyrangefrom0° to180°becauseof thenatureoftheStokesparameters;apositionangleof180°isconsistentwith0°.Noticethelinecorepolarizationisoffsetbyabout90°fromthewings.Thismaybeanopacityeffect.Comparethesedatatothemid-eclipseQU-plot(Figure2)andmid-eclipselineprofile(Figure1).
Geiseetal., JAAVSO Volume 40, 2012784
Figure4.TimeseriesofH-beta(486.135nm)lineand%pprofilesforpre-,mid-,andlate-eclipseepochs.Velocityiscenteredonthelinerestwavelength.TheF-starradialvelocityisindicatedbyastarsymbol.Polarizationdataareepoch-(seetext)andwavelengthbinned(binsize0.015nm).Averageerrorsin%pareshown.Thelineisclearlysaturatedinlateeclipse,fallingtojust1.6%intensityatthedeepestpointand,althoughS/N is nearly 200 here, the ratio p/σ is only1.2;therefore%p=0.Theratiop/σ ranges from 3 to 11 for %p > 0.4 outsidethesaturatedregion.
Polarimetry of e Aurigae, From November 2009 to January 2012
Gary M. ColeStarphysics Observatory, 14280 W. Windriver Lane, Reno, NV 89511; [email protected]; www.starphysics.com
Received March 20, 2012; revised May 17, 2012; accepted June 25, 2012
Abstract During the 2010–2012 eclipse of eAurigae, the author obtainedlinearpolarizationmeasurementsduring200nightsofobservationoverthreeobservingseasons.Theseobservationsbeganbeforesecondcontactandhaveextended some six months into the post-eclipse period. Measurements weremade in V, B, and R photometric bands. The polarization of eAurigae wasobservedtovarybynearly0.6%peaktovalleyduringthisperiodincyclesofvaryingduration.Thesevariationsresemble,ataqualitativelevel,thoseseenbyKempandHensonduringthe1984eclipseegress.Inparticulartheyshowevidenceoflocalpolarizationactivityextendingwellpast4thcontact.
1. Introduction
eAurigae is an F-type supergiant star approximately 2,000 light yearsdistant.Normallyseenatmagnitude3,itundergoesan18-month-longeclipseevery27years.Thesecondaryobjectactsasanopaquedisk,partiallycoveringtheprimary,whichdimsthesystem’slightbyapproximatelyonemagnitude.Whilethediskisassumedtocontainahotembeddedstar,thespectrumofthesecondaryhasnotbeenobserved. At the suggestion of Dr. Robert Stencel (University of Denver) in May2009,theauthorbeganaseriesofbroadbandpolarimetricobservationsofeAurduring its recenteclipse.Thepurposewas toextendmeasurementsobtainedduringthe1984eclipsebyDr.JimKempandcolleagues(Kempet al.1986)andfollowedupbyhisstudent,Dr.GaryHenson(Henson1989). Those observations had revealed significant variations in polarizationduring and after the eclipse. Dr. Stencel suggested that similar observationsduringtherecenteclipseshouldbeusefulforcomparison. Observationsreportedhereinweremadeoverthecourseofthreeobservingseasons. Season 1 from November 2009 to February 2010, season 2 fromSeptember2010toMay2011,andseason3fromJuly2011intoJanuary2012.Ongoinginstrumentdevelopmentledtosignificantimprovementsinthequalityofmeasurementsoverthecourseofthiswork.
Cole, JAAVSO Volume 40, 2012788
2. Instrumentation
There are no commercial sources for small telescope astronomicalpolarimeters.Myinstrumentshavebeenpurpose-builtforthisprojectfollowingthe general concepts used on the Hawaii 88-inch telescope as described inMasieroet al.(2007). An imaging telescope becomes a dual beam polarimeter with theintroduction of a rotating half wave retarder and a calcite beamsplitter.Thewaveplatematerial used is an achromaticpolymerof the same specificationasintheHawaiiinstrument.Thebeamsplitterisa13-mmcalciteSavartplatefabricatedbyHalboOptics,whichprovidesaseparationof1mmattheCCDfocal plate.This is approximately 135arcseconds at the plate scale used ontheC8telescopeand55arcsecondsontheC14.Thisismorethanenoughtocleanlyseparatetheslightlydefocusedandastigmaticstarimages.Theinitialwaveplate rotator was constructed with a USB-controlled servo motor andplasticgearing. TheSavartplateandafocalreducerweremountedintothenoseofanSBIGST6camera.Therotator,BVRfilterselector,andcamerawereattachedtoaC8opticaltubeandtheentireassemblywasmountedontoanexistingautomatedC14telescope. Thisarrangementwasusedforthefirstseasonofobserving.Forthesecondseason, the camera was replaced with a SBIG ST402 camera. This greatlyreduced thermal noise and reduced download time. The rotator, which hadfrequentbreakdownsandpoorgearing,remainedinuseuntilFebruaryof2010whenitwasreplacedwithahighprecisionrotatorengineeredbyOptec,Inc. Forthefinalseasonofthiswork,theinstrumentwasreconfiguredtoworkin the C14 optical system.The waveplate rotator was mounted between thefocuserandtheexistingfour-wayinstrumentselector.AdoublefoldorthogonalmirrorassemblyrelaysthebeamintotheSavart+cameraassembly. These ongoing changes have yielded significant improvements in dataqualityoverthecourseofthisproject.Inthefirstseasonofmeasurements,theuncertainty, ∆p, in the degree of linear polarization was ~ ∆p ± 0.1%; in the second season this was improved to ~ ∆p ± 0.05%, and further refined to ~ ∆p ±0.03%inthethird. Note: The Johnson BVR photometric filters are used in series witha 400-700nm luminance filter so as to match the effective range of theachromaticwaveplate.ThisresultsinaslightlyreddenedBandaslightlytruncatedRbandpass.
Instrumentalpolarizationisdetectedbyobservingzeropolarizationstandardstars.Severalhundredmeasurementsofsuchstarsindicatethatsuchinstrumentaleffects, if present, are less than 0.03% and show no preferred orientation. Theangleofpolarization,q, isdefinedby IAUconventionsasananglefromtheNorth-Southline,towardsEast.Themeasuredangle,therefore,mustbeadjustedsoastomaintainthisconvention.Thefirstlevelofadjustmentistosetthe“zeropoint”ofthewaveplatetoreplicatetheconvention.Thiswasdoneusinganexternally-mountedsheetpolarizerorientedalongthedeclinationaxis.Asecondstepwastorefinethefiduciaryanglebyusingmeasurementsofstandardstars. Inpractice, theauthorhas found itverydifficult tomaintain theangularprecision of the instrument over the course of these observations. Frequentrotatorbreakdowns,mechanicalrotation,andreconfigurationshavedisturbedtheeffectivealignmentandhencethereportedangles.Frequentobservationsof HD 21291, whose stability was affirmed from HPOL measurements (U.Wisconsin2012),wereusedforangularcorrection. The modulation efficiency was checked by comparing large sets ofinstrumentalresultswithcatalogvaluesforstarsofknownhighpolarization.Theresultsobtainedmatchexpectedvalueswithintheexperimentaluncertainties.Nocorrectionsformodulationefficiencyhavebeenmade.
A floating square aperture that is aligned to the centroid of each star issizedtocapture97.5%ofthetotalsignalonthefirstframe.Thisaperturesizeisthenusedforallimagesinthedataset.Alargerconcentricapertureisusedto estimate thebackground signal.Thedualbeammethod is self-calibratinganddoesnotrequireflatfielding,butthisissubjecttotherequirementthatthestarimagesarerecordedontothesamepixellocationswithineachdatapair.Tomaintain this requirement,anydatasets thathaveexcessivemotionbetweenframesareexcludedfromcalculations. Once the intensity values have been extracted from all data sets of theobservation, thedegreeofpolarizationandanglearecomputedaccording tothemethodsofTinbergen(1996).Thesamearecalculatedforeachindividualdatasetforcomparison.TheerrorthatcanbeascribedtophotonstatisticsiscalculatedaccordingtothemethodofSerkowski(1974). The polarization value is adjusted for zero point bias according to themethoddescribedbyClarke(2010).Thevalueofthisadjustmentistypicallylessthan0.01%forourobservationsandinsignificantintermsoftheuncertaintyassociatedwitheachmeasurement;theS/Nratioishighandthevalueofpislargerelativetothemeasurementuncertainties.6. Target observations
In Table 1 the result for 200 nights of observation are presented. Eachmeasurement consists of multiple sets of four images taken at waveplatepositionsof0,22.5,45,and67.5degrees.Onmanynights inwhichseveralmeasurementswereobtained,thesehavebeenaveraged.Theerrorsestimatedfor each polarization measurement are stated in the adjoining columns.Theanglesareshownintherightmostthreecolumns.Theerrorsassociatedwiththeanglesarediscussedinthenextsection.
7. Sources of uncertainty
ObservationspriortoMarchof2011weresubjecttoimprecisewaveplatepositioning.Theoriginal rotatingdeviceusedplasticgears to translateservomotion.Aseachcogofthemaingearcorrespondstonearly3degrees,asmallrangeofrandomvariationisinevitable.Therewasalsoaninfrequentproblemof“hopping.”According to theanalysisdonebyRamaprakashet al. (1998),anyuncertaintyofangularpositioninggeneratesapolarizationuncertaintyofsimilarmagnitude.Henceifa±1degreeerroractuallyoccurredfromtimetotime,itwouldadd0.05%totheuncertaintyofanindividualmeasurement.Thissourceofthisproblemwasreducedbyatleastafactorof25whentheOptecRotatorwasinstalledlateinseason2.
Cole, JAAVSO Volume 40, 2012 791
Observations from this site encounter significant turbulence, bad seeing,windbuffeting,andimperfectguiding.Theeffectsofseeingaretomovethecenterofilluminationandtochangethedistributionoflightwithineachstarimage.Thesefactorsinducesmallrandomerrorsintotheresultsthataredifficulttoquantify.Anexaminationofthereportedvariationsinclosely-spacedtimesequentialmeasurementssuggest that thedatacarryerrorabout50%greaterthanpredictedsimplyfromthephotonflux. For the data presented in this paper, the errors reported in the table andchartsforeachmeasurementaredoublethosecalculatedbaseduponthephotonstatisticsoftheactualimages.Theauthorbelievesthisprovidesaconservativeestimateofthetrueuncertaintyintotalpolarization. Theuncertaintyinreportedanglesofpolarizationismuchworsethatthatofthemagnitudeofpolarizationbecause1)itisderivedfrommeasurementsofstandardstarswhichhaveatleastequivalentuncertainty,and2)becausetherewerefrequentmechanicaladjustmentsmadeduringthecourseofobservations.Theauthorbelievesthattheanglesreportedcontainbothrandomandsystematicerrorsthatmayreachthelevelof±1degree.
8. Discussion
The polarization measurements (%) in the BVR photometric bands areplottedasatimeseriesinFigure1.Thefirstseasonofobservationscoversthetimeapproachingthesecondcontactduringtheeclipse.Thepolarizationvalues(p)arehighandarematchedagainintheearlypartofthesecondhalfofthesecondseasonleadinguptotheendoftheeclipse.Thethirdobservingseasonisbeyond theeclipsephaseandshowsstrongpolarizationactivitywith fallsandrecoveriescoveringseveraltensofdayswithamplitudesmuchthesameasduringtheeclipse.Particularlyontheprotractedfallofsome70daysatthebeginningofthesecondseason,soonaftermid-eclipse,thecolordependenceofthepolarizationclearlyshowsinFigure1,withthevaluesofpfallingfromBthroughVtoR. Figure2displayspandthepositionangle,q,fortheVbandoverthethreeseasons.Thereappears tobeacorrelationbetween the threepeaksofpandtroughsinqduringthesecondseason,but this isnotquiteasapparentpost-eclipse.Thisbehavior indicates that thevariable intrinsicpolarization isatasignificantangletothatofthelargeinterstellarcomponent. Several researchers, including the teamat thePineBluffObservatoryoftheUniversityofWisconsin,havemadesynoptic,out-of-eclipseobservationsofthisbinarysystem.During1990to1996,theirmeasurementsrevealedp(V)valuesvaryingfrom1.96%to2.06%withq(V)~144degrees.Thisweassumetobetheinterstellarcomponent. Figure3displaysthenormalizedStokesparametersqanduasafunctionof time. Direct comparison of the polarimetric behavior of this eclipse with
Cole, JAAVSO Volume 40, 2012792
thatreportedbyKempet al.(1986)ofthepreviousonecannotbemadeastheearlierauthorsprovidenormalizedStokesparameter time-lineplotswith theq,uvaluesexpressedinaco-ordinateframeat5degreesrelativetothestandardequatorialsystem.However,astheangulardifferencebetweenframesisverysmall,strongqualitativecomparisonscanbemade. Inparticular,thebehaviorofbothqanduduringtheegressphaseisextremelysimilar with small oscillations in q and a smooth rise in u in both data setsfollowedbyarapiddecline.Themagnitudeandfrequencyofvariationinbothdatasetsshowconsiderablesimilaritybutnotdirectrepetitionexceptategress. A standard method for investigation of polarimetric data is to plot thenormalizedStokesparametersq,uasCartesianco-ordinates.ThishasbeendoneinFigure4foralltheV-banddata.Althoughthescatteroftheplottedpointsisverylarge,simpleinspectionshowsthatthereisatrend.Thisbecomesmorereadily apparent when the data are grouped into smaller temporal segmentsas has been done in Figure 5.While each segment has a different origin intheq,u-plane,theslopesduringthesefourselectedtimeintervalsareclosetobeingparallel.ThisisshownmoreclearlyinFigure6wherethefoursegmentsareoverlaid.Translatingthesegradientsintocelestialangles,thepolarizationalmovementappearsalongpositionanglesbetween7and22degrees.Giventheleveloferrorintherawdata,thesevaluesareclosetothat(5degrees)takenbyKempet al.(1986)fromtheastrometricdataofvandeKamp(1978)asbeingrelatedtotheorbitalplanedirectionprojectedontothesky.TheuncertaintyofthisangleisnotgivenbyKempet al.(1986)anditisdifficulttoascertainthebestvaluefromthesketchprovidedbyvandeKamp(1978).Asthetrendanglesareclosetotheestimateoftheorbitalprojectionangle,short-termpolarizationchanges, chiefly affecting theqvalues, canbe surmised asoriginating fromscatteringbyclumpymaterialintheorbitalplane.Thedriftsmaybeaffectedbyintroductionofnewclumpsandtheirdecay,orbychangesintheirdistancefromtheilluminatingstar.Astowhytheoriginofthetrendlinesmovesaboutto produce the more blurred picture that Figure 5 presents, with significantchangesintheuparameter,amoresophisticatedmodelisrequired. Finally,thereappearstobecorrelationhoweverbetweenseveraloftheUbandphotometricmaximaoftherecenteclipse(eAurcampaign2011)andthepolarimetricmaximainthesedata.Inparticular,atJD2455460,atJD2455545,andategressthevaluesofUandpseemtomovetogether.
9. Future work
In future seasons it is the author’s intention to monitor the system inthe same frequentmannerasduring theeclipse toconfirm thenatureof thevariabilityandsearchforshort-termperiodicities.AsimultaneousprogramofUandBphotometrywillbeaddedtosearchforanydirectcorrelationbetweentheopticalandpolarimetricvariability.
Modeling the Disk in the ε Aurigae System: a Brief Review With Proposed Numerical Solutions
Richard L. Pearson, IIIRobert E. StencelUniversity of Denver, Department of Physics and Astronomy, 2112 E. Wesley Avenue, Denver, CO 80208; address email correspondence to [email protected]
Received April 19, 2012; revised July 19, 2012; accepted August 7, 2012
Abstract Parameters associated with the opaque disk in e Aurigae areexplored in thecontextofcircumstellarandproto-planetarydisk theory.Theobservedblackbodytemperaturesofthedisk,at550and1150K,areprimarilydiscussed. Brief reviews of previous work are included that describe andattempttoexplainthistemperaturegradient.HeatingfromonlythecentralBstarprovidesabasaltemperatureofabout250K.Anaccretionrate(fromthedisktotheBstar)of10–7M
eAurigaeisasingle-linespectroscopicbinarythatfeaturesanopaquediskaroundacompanionthatcauseslengthyeclipsesevery27years(forareadinglist,seeCarrollet al.1991;Lissaueret al.1996;Stencelet al.2011).Figure1illustrates the configuration. Interferometric imagingproves the existenceofa disk andprovides somepreliminarydimensions for it (Kloppenborget al.2010), based on a highly uncertain Hipparcos distance of 625pc (Perrymanet al. 1997). The eclipse of 2010 provided a wealth of new data, from far-ultraviolettofar-infraredandsub-mmwavelengths(Hoardet al.2010;Stencelet al.2011;Hoardet al.2012).Forthiswork,wearespecificallyinterestedintheobservedsurfacetemperaturesofthediskasstatedinHoardet al.(2012)andprovidedhereinTable1. ThereisanunresolvedconcerninregardstopinpointingtheevolutionarystateofeAur.Twopartsofthisconcerndealwiththestellarmassesandthedistancetothesystem.First,themassesareunknown,butthebrightprimarystarresemblesanF0supergiantthatmaycontainasmuchas15–25M
2002).Littletonoopticalspectrumexistsforthedisk-shroudedcompanionandthusitcannotbeclassifieddirectly.Spectroscopicinformationprovidesamassfunction value of 2.53 (Stefanik et al. 2010) and admits distance-dependentmassratiosof0.5–1.1.Eclipsedataandultravioletfluxessuggestthehiddencompanion could be a ≈ 6 M
ÄB5Vstar(Hoardet al.2010),with thebright
staranapproximately3MÄ
hyper-luminouspost-AGBstarinarapidstateofevolution(LambertandSawyer1986;Takeuti1986a;Saitoet al.1987;ShefferandLambert1999). Second, the actual distance to eAur is not well defined. The system isat the limitofavalidHipparcosparallacticdistance.Therealsoseemstobeintrinsic variability in the F star, which provides a varying star photocenterfor the parallactic measurements (Kloppenborg et al. 2011). Of course, anaccuratedistancemeasurementrefinestheconstraintsonthesystem’sphysicalparameters (forexample, stellarmasses, star-to-disk separations, andsoon).Someof the typicalparameters,using thehigh-errorHipparcosdistance,arelistedinTable1. Previous analytical and numerical modeling techniques have focused onresolvingthequestionablestateofeAur,byunderstandingthedisktoconstraincertainparametersoftheentiresystem.Theseincludediskthicknesslimitations(Lissauer et al. 1996), disk temperature studies (Takeuti 1986b; Hoardet al. 2010; Takeuti 2011), and spectral energy distribution (SED) matching(MuthumariappanandParthasarathy2012),forexample.ManyofthesemodelsusedtheHipparcosdistancetodefineparametersand/orusedonlyspecificpartsofthesysteminthemodeling(thatis,consideringonlythediskandBstar). In order to support the plethora of observations, modeling techniquesmustfurtherexplorethecompletenatureofthisdisk.Theobservedminimumand maximum disk temperatures (Tnoon = 1150 ± 50 K and Tmidnight = 550 ±50K)provideanavenueofinvestigation(Hoardet al.2012).Onecanlooktoreproducethesetemperaturesinanalyticalandnumericalstudies,byexaminingfoureAurdisk-heatingscenarios:
Themotivationof thispaper is todescribe theneed tobuildacompletemodelof theeAur systembyusingnumericalmodeling techniquesand theobservedtemperatures.Fromthis,adistancecanbedetermined(asdescribedin section 2.3.2), thereby clarifying the evolutionary state. Sections 2.1 and2.2 establish possible disk basal temperatures and clarify that the observedtemperature gradient must be achieved by an external source. Section 2.3
A very useful tool in constraining certain parameters of the system arethe disk temperatures. By matching observationally constructed SEDs withblackbodytemperaturecurves,twotemperatureshavebeenestablished(Hoardet al.2012).HalfofthediskfacingoppositetheFstaristhe“midnight’’sideandfoundtobe550±50K.ThesidefacingtheFstar,“noon,”isfoundtohaveatemperatureof1150±50K.Decipheringhowandwhythediskexhibitstheobservedtemperaturegradientisimportantinunderstandingtheactualstateofthesystem. Weexploreheatingeffectsonthediskbythreedifferentsources:heatingfromthecentralstar,accretionalheating,andtheexteriorstellarsourceareallinvestigatedbelow.Briefreviewsofpriorworkareincludedaswell.
2.1.Centralstarinput We first consider a discussion concerning the effect of the central starradiationonthesurroundingdisk.Takeuti(1986b)analyticallysolvedforthetemperatureandscaleheightofthediskatspecifiedradii,basedonacentralBstarof4M
Ä,3R
Ä,andTeff=15000K.Blackbodyequilibriumtemperatures
of355and263Kwere foundatdisk radiiof2 and3AU, respectively.Nonumericalradiativetransferanalysiswasperformed.Thesignificanceofthesevaluesisdiscussedbelow. Morerecently,MuthumariappanandParthasarathy(2012,hereafterM&P)investigatedthecomposition,dustparticlesize,outerradiustemperature,andmass of the disk in eAur, using a two-dimensional, photon-tracking MonteCarlocode.TheirSEDresultswerebasedonenergyinputonlyfromaninternalB5V star. The system was assumed to be at the Hipparcos distance. Theycreatedmodelswithdustcompositionsofamorphouscarbon,ISMdistribution(60%silcatesand40%carbonates),andamorphoussilicates.Also,particlesizedistributions corresponding to minima--maxima ranges of 0.05–0.2mm and10–100mmwereapplied.AKuruczfluxmodelofa7700K,F0Iaepost-AGBstarwascombinedwiththeoutputSEDs.Then,comparisonsweremadewiththeobservationallydeterminedSED. TheirSEDfittingresultedina5×10–3M
Ädiskcomposedofcarbonates,
withgrain sizes10–100mm, andanouterdisk temperatureof252KatRout=3.8AU.Theothermodelsresult inouterdisk temperaturesof261–293K.These are, obviously, lower than the either of the observed Tmidnight or Tnoon.Takeuti (1986b)determinedaverycomparable temperature,asstatedabove.ThoughM&Pprovidenodiscussion concerning this, these findings indicate
2.2.Accretioninput Wenowexplorethedisktemperature,specificallyTmidnight,basedsolelyonaccretionalheatingfromthedisktothecentralstar.Armitage(2010,specificallysection3.3therein)describesasetofgeneraldiskequationsdescribingaccreting,Keplerian disk systems (for additional discussion, see Lin and Papaloizou1985).Turbulentmotionisusedastheprimarysourceofangularmomentumtransportwithinthedisk;itisportrayedasaturbulentviscosityinEquation1ofTable2.TheShakura-Sunyaevaparameter(ShakuraandSunyaev1973)isusedtodefinethedisk’sviscosity. The equations displayed in Table 2 have been adopted from Armitage(2010) and applied to the eAur disk.A previous iteration of this analyticalcalculationisfoundinTakeuti(1986b),whousesaformofEquations5,9,and10inTable2(alongwithacentralBstarasdescribedinthefirstparagraphofsection2.1) tocalculate temperaturesandscaleheightsatvariousdiskradii.Updatedtemperaturesandscaleheights,aswellasadditionalparameterssuchasdiskmassanddensity,arepresentedinTable3. If the central star mass (Mstar), accretion rate (M ), and disk radius (rdisk)are known, the equations inTable 2 depend only upon the mass absorptioncoefficient (k), the mean molecular weight (m), and the viscosity parameter(a).Therefore,bydefiningthesesixterms,afullsetofparametersdescribingthesystemcanbeoutput.Theopacitytemperaturedependence(k=k0T )andconstant(k=5×10–3cm2g–1)wereselectedfromLinandPapaloizou(1985)andPollacket al.(1985).Amolecularweightof1.5wasselected,whichislowenoughtoadmitthatthediskhasahighgas-to-dustratio,butshowsitisnotcompletelymadeofhydrogengas(m=1).Hartmannet al.(1998)determineda=0.01forTTauristardisksandthathasbeenadoptedhere.AccretionrateswereobtainedfromPequetteet al.(2011a),whoconcludesthathavingaM ≠ 0 was apossiblewaytoreproducecomparablemodelSEDstotheobserved.Pequetteet al.(2011a)statethatonlyahighaccretionrateprovidesanappropriateUV-to-IRratio.Fortheanalyticcalculationshere,twotypicalaccretiondiskratesof10–6and10–7M
observedthicknessofthediskanditsradius.Lissaueret al.(1996)showsplotsofz/Rd,whichisequivalenttoN h/RoutwhereNisanintegernumberofscaleheightsinadistancez.TheanalyticalcalculationsinTable3showthatforthehighestaccretionrate,noneoftheN h/Routvaluesfallnear0.08.However,ratiosfromthe10–6and10–7accretionratesarecomparable.The10–6rateallowsforN=1orN=2,dependingonthecentralstarmass.The10–7rateseemstoprefereitherN=2orN=3,whichalsocoincideswiththeLissaueret al.(1996)models. Ifweassumeadustsublimationtemperatureof1500K,thentheTcvaluesforthehighestaccretionratearetoolarge,especiallyifparticlesaresupposedtobecoalescingintoprotoplanetaryobjects.However,asseeninFigure2,theonlyM tooutputaTdisknear theobservedmidnight temperature is the1.5×10–5 M
Ä/yr rate. If accretional heating was solely responsible for providing
the observed 550 K “midnight’’ temperature, then the predicted disk massfromthehighaccretionratemustbediscussed.Thediskmass fromTable3is41×10–3M
2.3.Companionstarinput ThetwoprevioussectionshaveillustratedtheimportanceofincludingtheF0star'sradiationinthemodelingtoexplainthedisk’sazimuthaltemperaturegradient.Thoughanexaminationofseparateheatingcomponentscanprovidespecificconstraints(thatis,abasaltemperature),thenextstepistoincludeallofthecomponentsofthesystem.Abriefreviewofpreviousworkrelatingtothedisktemperatureisdescribedtofurtherjustifytheneedfornewnumericalcomputations. The results of these new calculations will help provide anindependentmeasurementofeAur’sdistance.
2.3.1.Previousanalyticalstudies As previously mentioned, Takeuti (1986b) calculated disk temperaturesandscaleheightsfromtwoseparateradiationsources.Puttingthemtogether,Takeuti (1986b) postulates a possible disk configuration: a cool, thin disk
Pearson and Stencel, JAAVSO Volume 40, 2012 807
near “midnight” and subject only to the central star’s radiation; a crescent-shaped,opticallythickregiondominatedbytheFstar’sradiation;andasmall,optically thinregion,directlyopposite theFstarnear thedisk’sedge.Usingthisconfiguration,asteady-statediskassumption,andalimitingTmidnight=500K,anaccretionrateof2×10–11M
Ä/yrisfound.Thisaccretionratewouldnot
producetheobservedUVexcess(Hoardet al.2010).Takeuti(1986b)outlinesthat further ultraviolet and infrared observations are needed to resolve thequestionsconcerningthesystem. Twenty-five years later, Takeuti (2011) performed some analyticalcalculationsthatexploredtheouter-edgedisktemperaturevariationalongthediskplane,asheatedbyF0andB5Vstars.TheHipparcosdistanceof625pcwasusedtogeneratelinearseparationsinthesystem.Itisnotedthatsincethe“noon”temperatureistheaveragetemperatureoftheF0-facingsideofthedisk—similartotheaverage“midnight”temperatureoverthehalfofthediskfacingawayfromtheF0star—themaximumtemperatureat“noon”willbeslightlygreater than theaveragevalue,Tnoon (seeFigures1and2,aswellasTable2inTakeuti2011).Exploratorycalculationsconsistingofphysicallydescriptiveheatcapacitiesprovidedvariabletemperaturesaroundthedisk.However,theimplications(composition,particlesize,andsoon)oftheheatcapacitieschosenwerenotconsidered.Amaximumtemperatureof1200Kprovidedanupperlimitforthecalculations.Temperaturesatthe“midnight”positionwerefoundtobe250,500,and750K.Scaleheightswerealsocalculated toprovideanadditionalphysicalcorrelationwithprescribedmodels.Accretionalheatingwasignoredduring thiscalculation.TheworkofTakeuti (2011)provideshelpfulconstraintsthatcanbeusedinnumericalmodelcalculations.
2.3.2.Proposednumericalanalysis The processing power and speed of current computers allow numericalsimulationstoeffectivelydemonstratephysicalconfigurationsofastronomicalsystemsquickly.ForeAur,usinga three-dimensional (3D),photon-trackingMonteCarlocodewillpermitaninspectionofdisktemperaturesaccordingtoradiationfrombothstellarcomponents,aswellasaccretionalheating.However,the unknown (or rather, the known but highly uncertain) distance createssignificantproblemsinanalyzingtheradiationeffectsonthedisk.Therefore,adistance-—independentofanyobservationsotherthantemperature—canbedetermined, based on the reproduction of the known temperatures, Tnoon andTmidnight,onthedisk. Figure3displaystheprocessbywhichadistancecanbedetermined.First,adistance (dModel) is selected.Then,dModel isused toconvert allof thewell-determinedangularseparationsintolinearmeasurements.TheseparametersaretheninputintotheMonteCarlocode.Next,theMonteCarlocodeoutputsadusttemperaturefile(describedas“DiskTemperatures”inthefigure).ThepointsalongtheouteredgeofthediskfacingtheFstar,willbeaveraged— —Model
noonT< >
Pearson and Stencel, JAAVSO Volume 40, 2012808
and compared with . If the modeled temperatures do not equal theobserved,anotherdistanceischosenandtheprocessiscontinued.However,ifthetemperaturesareequivalent,thenthesystemcanbedefinedbythechosendistanceandadditionalparametersarecalculated. Therewillbeaninherentrangeofdistancesthatarecapableofmatchingtheaveraged“noon”temperature,duetotheobservedtemperaturehavinganassociatederrorof±50K.However,thedistancerangesshouldbesufficientlysmalltoprovideconclusiveresults.Oncethedistancehasbeenestablished(withacertainamountofuncertainty),furthersystemsparameterscanbefinalized:themassfunction,themassratio,thetwostellarmasses,andthestellarradii.Knowingthese,theevolutionarystateofthesystemcanbeclarified. Anotherfeaturethatcanbeexploredisthedisk’scooling(andheating)rate.However,thenumericalmodelingonlyprovidesatemperaturesnapshotofthe“stable”system.Norotationofthediskisaccountedfor,andhence,thedustdistributed in thediskhasnoprior incidentheator radiation.Therefore, theoutputtemperaturesonlyaccountforthesnapshotofheatingthedustreceivesatthespecificmomentbeingexamined;atemperaturedistributionwillstillbepresentalongtheouterridgeofthedisk(forexample,Tnoon>Tmidnight),duetoshieldingeffectsoftheflareddiskandtheincreaseddistancefromtheFstarradiation.Atemperatureprofileofthediskcanbeconstructed(seeFigure4foranexample)andexamined. Comparisonsofthesnapshottemperatureprofilecanbecomparedagainstaprofileofarotatingdisk.Ifadiskrotationisassumed,thedustparticlesdirectlyinlinewiththecompanionFstar,at“noon,”willheatuptosometemperature,Tnoon.Asthedisk’sdustrotates,theamountofradiationfromtheFstardecreases;this allows thedust tobegin its coolingprocess.Thedustbegins toheatupagainwhenitagainreceivestheFstarradiation(seeTakeuti2011).Therateatwhichtheheatingandcoolingoccursishighlydependentonthecompositionof the dust. Once the maximum and minimum temperatures are acquired inthenumericalprocessoutlinedabove,temperature-timeprofilescanbecreatedfollowing Takeuti (2011). Various disk compositions will result in varioustemperature-time profiles. The shape and slope of profile can be comparedwiththesnapshottemperatureprofilestoassessthecompatibilityofthetwo. Thereareafewotherinputparametersthatwillneedtobefullyexplored.For instance, theaccretion rate.Since it isanunknown,numerous iterationswithvariousvaluesofM willneedtobecomparedateachchosendistance.Limitations have been placed on M previously, but since this is the firstcomplete3Dnumericalanalysisofthewholesystem,afullrangeofaccretionalrateswillbeexplored.Anotherrequiredinputisthedensity,composition,andparticlesizedistributionofthedisk.Theseaspectshavealargeimpactonhowthediskmaterialinteractswiththevariousheatingsourcesandhowitcools.Eachwillbeexploredwiththetechniquedescribedabove.Acompletesetofsolutionsareforthcoming.
2.3.3.Asimpleexampleofa3Dnumericalcalculation Todemonstrate theprimary results from this 3Dmethod, a hypotheticalexample is described here. Following the prescribed outline in Figure 3, anarbitrary distance of 830 pc was chosen. This places a user-defined F0 star(7750K,150R
Ä,3M
Ä)about30AUfromadisk-envelopedB5Vstar(15000K,
4RÄ
,6MÄ
)inradmc-3d.Asilicatediskwasmodeledwithinnerandouterradiiof1and4AU.Only10000photonpacketswerelaunchedduringtheradiativetransfercalculation,creatingaverystatisticallypoorsetofsolutions;however,thisissufficientforthisconjecturedcalculation. Theaveragedisk-planetemperatureatRout=4AUoftheF0-facingdiskwasabout1500K.ThistemperatureishigherthanthevaluespredictedbyTakeuti(2011)and ,butstillrelativelyreasonable.Theareaofthediskatthe“noon”positionmaxedatabout2500K,whichissignificantlyhigherthantheTakeuti(2011)estimationofabout1600K.Anaverage“midnight”temperatureofabout900Kwasfound,whichisalsolargerthanthe .Acompleteanalysis, as described previously, is not presently provided. A graphicalrepresentationisshowninFigure4. Itisnotedthatthisbriefexampleusesonlyasilicatediskandprovidesnoaccretionalheating.Afullanalysiswillexaminethediskcompositionandwillincludeanaccretionalrate.Still,thedisregardforaccretionalheatingmaybeaplausibleassumptionbasedontheB5Vtemperature:iftheoutwardradiationpressurefromtheB5Vstarislargeenough,itwouldseverelylimitthemassflowfromthedisktotheB5Vstar. A final note is made here concerning the and the coldesttemperatureof theradmc-3dmodel.Abasal temperatureofabout250Kwasfound in sections 2.1 and 2.2. The work by Takeuti (2011) place this basaltemperaturenearthe“dusk”positiononthedisk(refertoFigure1).Therefore,onecanexpect theobservedTmidnight tobelargerthanat“dusk.”However, inthishypotheticalcalculation,anaveragebasaltemperaturewasfoundatabout500K.Adjustments to the distance, separation, composition, transmissivity,and/orotherinputparametersofthesystemcanbemadetofindreasonableandcomparable“dusk,”“noon,”and“midnight”temperatures. Thoughjustahypotheticalnumericalcalculation,thisexampleshowsthatthetemperatureresultswillbeabletoprovidedistanceanddiskcompositionpredictions.Itoutlinestheusefulnessofincorporatingbothstarsinanumericalsimulationandtheneedtoexploretheeffectsofaccretionalheating.
ObservednoonT< >
ObservedmidnightT< >
ObservedmidnightT< >
Pearson and Stencel, JAAVSO Volume 40, 2012810
3. Conclusion
WehavedemonstratedtheneedforfutureanalyticalandnumericalmodelsofeAur to include the radiationeffectsof thecompanionstar.Additionally,wehaveproposedawaytodeterminethedistancetothesystembyusingtheobservedtemperatures.Abriefreviewofpreviousdisktemperaturemodelingwas provided to give context to the proposed numerical solutions. Furtherclarification regarding theevolutionary stateof theeAur systemmay resultfromthedescribedanalysis.
4. Acknowledgements
Weappreciate the expert adviceprovidedbyBrianKloppenborgwhileconstructing this article.This work was supported in part by the bequestofWilliamHerschelWomble in support of astronomyat theUniversity ofDenver,forwhichwearegrateful.
References
Armitage, P. J. 2010, Astrophysics of Planet Formation, ed. P. J.Armitage,CambridgeUniv.Press,Cambridge.
Carroll, S. M., Guinan, E. F., McCook, G. P., and Donahue, R. A. 1991,Astrophys. J.,367,278.
Guinan, E. F., and Dewarf, L. E. 2002, in Exotic Stars as Challenges to Evolution,ed.C.A.ToutandW.vanHamme,ASPConf.Ser.,279,Astron.Soc.Pacific,SanFrancisco,121.
K.,andMais,D.E.2008,Astrophys. J., Lett. Ed.,689,L137.Stencel,R.E.,et al.2011,Astron. J.,142,174.Takeuti,M.1986a,Astrophys. Space Sci.,120,1.Takeuti,M.1986b,Astrophys. Space Sci.,121,127.Takeuti,M.2011,Publ. Astron. Soc. Japan,63,325.
Table1.AdoptedeAursystemparametersusingd=625pc. Parameter Value Comments*
Figure 4.A plot describing the temperature distribution from a hypotheticalnumerical modeling reconstruction of e Aur, from radmc-3d. The average“noon’’ and “midnight’’ temperatures were calculated by averaging thetemperatureslocatedattheouteredgeofthedisk,intheassociatedareasofthedisk.Theminimumandmaximumtemperaturesassociatedwiththenumericalmodelingareshown,aswellasahypotheticaltemperaturegradientrepresentedbythesolidblackline.Again, thelineisnotafit toactualdata: it issimplyagraphical representationofapossible temperaturedistributionon thedisk,outlinedbytheminimumandmaximumtemperatures.Noaccretionalheatingordiskrotationwasused.
Figure 3. A schematic showing the process of solving for the distance toeAur.Notethatthereareafewotherinputparametersbesidesjustthebinaryseparation.These includedustcomposition,dustmolecularweight,anddustdensity distributions. See text for further information regarding analyticalstudiesandnumericalmodeling,specificallywithradmc-3d.
Kloppenborgetal., JAAVSO Volume 40, 2012 815
A Demonstration of Accurate Wide-field V-band Photometry Using a Consumer-grade DSLR Camera
Brian K. KloppenborgDepartment of Physics and Astronomy, University of Denver, 2112 East Wesley Avenue, Denver, CO 80208; [email protected]
Roger Pieri37 C rue Charles Dumont, 21000, Dijon, France; [email protected]
Grigoris MaraveliasPhysics Department, University of Crete, GR-71003 Heraklion, Crete, Greece; [email protected]
Tom Pearson1525 Beachview Drive, Virginia Beach, VA 23464; [email protected]
Received May 8, 2012; revised July 3, 2012; accepted July 3, 2012
Abstract TheauthorsexaminedthesuitabilityofusingaDigitalSingleLensReflex(DSLR)cameraforstellarphotometryand, inparticular, investigatedwidefieldexposuresmadewithminimalequipmentforanalysisofbrightvariablestars.Amagnitude-limitedsampleofstarswasevaluatedexhibitingawiderangeof(B–V)colorstakenfromfourfieldsbetweenCygnusandDraco.ExperimentscomparinggreenchannelDSLRphotometrywithVTphotometryoftheTycho2catalogue showed very good agreement. Encouraged by the results of thesecomparisons,amethodforperformingcolor-basedtransformationstothemorewidelyusedJohnsonVfilterbandwasdevelopedandtested.ThismethodissimilartothatrecommendedforTycho2VTdata.TheexperimentalevaluationoftheproposedmethodledtorecommendationsconcerningthefeasibilityofhighprecisionDSLRphotometryforcertaintypesofvariablestarprojects.Mostimportantly,wehavedemonstratedthatDSLRcamerascanbeusedasaccurate,wide field photometers with only a minimal investment of funds and time.
thecenterwavelengthsofthesefiltersdonotmatchthewavelengthsofstandardJohnson filters. Furthermore, features that improve image quality like built-insoftwarenoisereductionalsodistortthetruephotometricsignatureofstarsof interest. Fortunately, camera manufacturers have given consumers accessto the RAW pixel data, which is often free of any on-camera processing ornoisereduction. ExamplesofphotometrytestingofDSLRsincludeworkbyHoot(2007).HetestedaCanonEOS350DasastellarphotometerbyimagingaseriesofLandoltfieldstars.UsingtheLandoltV,B,andRfilters,hecomputedoffsetsandcolorandextinctiontransformations.Hootfoundsignificant(~0.13mag.)uncertaintiesinthecolorcorrectioncoefficientsthatwerefarinexcessoftheinstrumentalRMS(~0.003mag.)values.Fromthisheconcludedtheremustbesomeoutlyingsystematiceffectcausingthelowqualityoffit.Furthermorehefoundthat“nosingleexposuresettakenwiththisDSLRfamilyofcameracanaccuratelyspanmorethan2.5magnitudes.’’Outsideofthisnarrowwindow,theerrorsincreaserapidly,thereforedecreasingtheutilityofusingDSLRcamerasforfieldswithwidemagnituderanges. SubsequenttopublicationofHoot’sarticle,DSLRcameraswereputtothetestonvariousastronomicaltargets.Littlefield(2010)andGuyonandMartinache(2011)haveshownDSLRcamerasarecapableof10milli-magnitudeorbetterphotometry that is suitable for tracking transiting exoplanets. Fiacconi andTinelli (2009)haveshownDSLRcamerascan trackpulsatingvariablestars,likeXXCyg.Allthreeofthesestudiesmadeuseofdifferentialphotometrythatdoesnotrequireaprecisetransformationtoastandardphotometricsystem.ArecentstudybyPataet al.(2010)showsthatsuchtransformationsareindeedpossible,providedthatthespectralpropertiesoftheobservedstarsareknown. AspartoftheAmericanAssociationofVariableStarObservers’(AAVSO)CitizenSkyProject,severalparticipantselectedtouseDSLRcamerastotrackthe2009–2011eclipseofeAurigae(Stencel2008;GuinanandDewarf2002).The success of this method (Kloppenborg and Pearson 2011; Kloppenborget al.2011)inspiredourworkwhichconfirmsthatDSLRcamerascanindeedbeusedas accurate, photometrically calibrated,wide-fieldphotometersoverawiderangeofcolorsandnearly fivemagnitudesofbrightnesswithaverymodestinvestmentoftimeandequipment.Whattheylackinflexibility,DSLRcamerasmakeupforinpriceandportability.
ofitsdynamicrangebycomparingsensorresponsetoaseriesofilluminatedtargets.Targetbrightnesslevelsweremeasuredusingadedicatedphotometer.AtISO100,the450Dappearstobelinear(within±0.5%)upto14,300ADU,wherethecamerasharplydepartsfromthelineartrend(fullwellsaturation).TheADCclippingleveloccurredat~15,800ADU,ratherthantheexpected16,384fora14-bitcamera.ThecalibrationfactoratISO100hasbeenmeasuredto2.27e-/ADU.Thereforethe1electronto1ADUcalibrationsettingresidessomewherebetweenISO200andISO400.Abovethissetting,thedynamicrangeofthecameraisreduced.WeusedstandardNikkor200-mmand85-mmlenseswhosefieldsofvieware6.36×4.24degreesand15×10degrees,respectively. Wemeasured the spectral responseofourcamera’sRed (R),Green (G),andBlue(B)Bayerarrayfilters.Figure1aisaplotoftheCanon450DGfiltercomparedtotheJohnson’sVfilterdefinition(MaízApellániz2006).Wefoundthe G channel is shifted blueward by 12nm relative to Johnson’sV-filter.Atransformationequationis,therefore,requiredtoadaptDSLRmeasurementstothisphotometricstandard.Likewise,wefounda2-nmblueshiftbetweenTychoVTandDSLRG(seeFigure1b),butthisshiftisnearlynegligible.Thisresultisquiteinterestingasitshowsthereshouldbelittletonocorrectionbetweenthe450DGchannelandTychoVT.Thesepropertieswillbeverifiedanddiscussedingreaterdetailbelow. Ofnotearesomeusefulfeaturesofthiscameratoaphotometrist.Thecamerahas a 10× magnified live-view display that is very useful when defocusing.Unfortunately,wefounditisoftennotpossibletoframethefieldofinterestbyusingtheliveviewat1×zoom.Insteadweusedtheopticalviewfinderwitharightangleadapteradded.Additionally,weusedareddotfinderattimes.Tobestsimulatethefacilitiesavailabletootherobservers,wemountedourcameraon a small, undriven, equatorial mount to accelerate framing of the chosenfields.Asimpletripodcanalsobeused.
2.2.Choiceofstarfields Our star fields were selected to optimize testing the limits of DSLRphotometryandtherelatedpost-processingsteps.InchoosingourtestfieldweintentionallyexcludedregionsintheplaneoftheMilkyWaytoavoidblendingtargetstarswithpotentiallyunseen,butdetectable,backgroundstars.OurfourfieldsarefoundbetweentheconstellationsCygnusandDracoboundedbyR.A.19h48mto18h05m and Dec. +48˚ 07' to +54˚ 28'. These fields total 15.85 × 6.36 degreesor101squaredegreesandhave~283starsbetweenV=3.7and8.75with(BT–VT)valuesof–0.2to2(seeFigures2aand2b).Ofthese,VizierandVOPlotreport76starsareatriskofblendingwithinthephotometricapertureand27starsaresuspectedvariables(seeTable1).MoststarsinthefieldareAFGandK spectral types,with a fewB- andM-type stars (seeFigure2c).BecauseofthelimitednumberofM-stars,wewerenotabletochecktheknowntransformationissuesinvolvingtheseobjects(Perrymanet al.1997).Wealso
Kloppenborgetal., JAAVSO Volume 40, 2012818
obtainedimagesfromasecond,largerregionusingthe85-mmlens.ThisareaextendedbetweenR.A.19h48mto17h10m and Dec. +44˚ to +59˚ in three fields. ThesedatahavesignificantfielddistortionsneartheedgeoftheFOVthatmustbedealtwithdelicately.Wewilldiscussthesedataandourmethodofreductioninafuturepublication.FortheremainderofourdiscussionweshallrefertoourobservedfieldsasCD3S,CD4S,CD5S,andCD6SassummarizedinTable2.
2.3.Dataacquisition OurobservationsbeganonJuly14,2011,andendedonSeptember28,2011.All imaging was done from the same location: +47˚ 19' N, +05˚ 01' E at 250 maltitude.Alldataweretakeningroupsof10to15images(hereaftera“series’’)withtheCanon450DandNikkor200-mmlensatf /4,ISO100with12.3-secondexposure times, except a few of the CD3S field which uses only 8-secondexposures.Thenumberofimagesperserieswerechosentoreducethenoisefromatmosphericscintillationtoafewmilli-magnitudes.Allimagesweredefocusedslightly(seeFigure3)sothatthestellarimagecoversroughlya10-to15-pixeldiameter(plustrailingduringtheexposure).Mostoftheimageswereacquiredwithin15to30degreesofthezenithwheredifferentialairmassisnegligible.FourseriesforCD5SandCD6Swere takenat40 to45degreesofzenithatabout1.5airmasses.Fromtheseimages,wefoundthatthebrighteststarsinoursample(V~3.7)haveaSNR>>1,000andstarsat7thmagnitudehaveSNR~200. Because we used a fixed focal length lens, it was possible to use a flatimagerecreatedeveryfewmonths.Theflatisobtainedbyimagingadiffuselyilluminated,non-glossy,fine-grainedwhitesurface(forexample, thebackofhighqualityphotopaper).A1%cross-surfaceuniformityofourflatfieldingsourcehasbeenverifiedusingadedicatedphotometer.
3. TheCanon’ssystematicoffsetof1024ADU(possiblyincludedindark)is first subtracted from the raw.The resultant image is then flat fielded.Eventhoughtheimagelooksveryuniform,wehavenoticedasmallnon-uniformity in the the outer perimeter of our images, and therefore weexclude a 50-pixel-wide border from photometric analysis. We do notapplyadarkimageaswehavefoundtheseshortexposuresatISO100haveminimaldarkcurrentnoise.Anyresidualcross-imagegradientswilleasilybedetectedinstage2,manifestingintheextinctioncoefficient.
4. Nextwecreateatemporaryluminance(R+G+B)imagetodetectstars.Theimageissampledatseveralpoints todeterminethebackgroundandisfitusingapolynomial.Thisfunctionisthensubtractedfromtheimage,removinganyremainingsystematicbackground.Theresultingluminanceis again measured and pixels residing above 3-sigma (noise) above thedarklevelareselectedascandidateobjects.Theareaaroundthebrightestpixelsareanalyzedtodetermineiftheyarepartofastar.Ifconfirmed,thefootprintof thestar ismeasuredand itsgeometriccentroidcalculated toensurepropercenteringduringaperturephotometry.
5. Because of diurnal motion, our star images are trailed.Therefore weperformaperturephotometryusingrectanglesratherthanannuli.Ourinnerapertureis21×13pixelsandtheouteris51×51pixels.Thebackgroundlevelforeachstar,determinedfromtheouteraperture,issubtractedfromtheforegroundapertureandtheRGBintensitiesofthestarareextracted.
3.1.1.Stellarblending Onedownsideofwide-fieldDSLRphotometryisthatrelativelylowangularresolutionofthecameracanblendtwonearbystarstogether.Uponexaminingour instrumental output table from the steps above, we found a significantnumberofoutliersfromthecalibrationtrendweexpected.Someofthesestarswereundoubtedlyvariables,butinmostcasestheywereaffectedbyblendingwithfaintbackgroundstarsthatfellwithintherectangularaperture. Tosolvethisissuewewroteasmallpieceofsoftwarethatusesdatafrom
Kloppenborgetal., JAAVSO Volume 40, 2012820
the Tycho 2 catalog down to VT=13.All stars that fall within the apertureare selected and their approximate flux is computed. If the background starcontributionexceeds0.012magnitude,weflagitinourdatatables.IfthePointSpreadFunction(PSF)responseof thecamerawerebettercharacterized,wespeculateitwouldbepossibletoremovethebackgroundstarcontribution,butthiswasbeyondthescopeofthispaper.Anystarsthatwereaffectedbyblendingwerenotusedinouranalysis.
3.1.2.Variablestars Inasimilarmannertoblendedstars,wehavealsoflaggedvariablestars.Wehavecollectedaggregatestatisticsonvariability,spectraltypes,andluminosityfrom the Tycho 2 (Høg et al. 2000a, 2000b), Hipparcos Input (Turon et al.1993),Hipparcos Main(Perrymanet al.1997),andtheTycho 2 Spectral Types(Wright2003)catalogues.Ofourtargets,twenty-fourwereflaggedasvariablestars from the input catalogs;however, sixof these stars showedno signofvariationwithinourmeasuredaccuracy.
3.1.3.Finalstarselectionandoutputquality These observations have provided 201 stars from the four fields fromVTmagnitude3.8to8.8.Ofthese,67havebeendeselected(18variables,52blended),yieldingatotalof134starsforfurtherprocessing.AggregatestatisticsofthesestarsareshowninFigures2a,2b,and2c.
3.2.Stage2 ThesecondstageofdatareductionmirrorsthetechniquesemployedintheCitizenSkyIntermediateReductionSpreadsheet.ThismethodtransformstheDSLRinstrumentalGmagnitude(denotedusingnhereafter)intothestandardphotometricsystem(denotedusingcapitallettersVandB)usingthestandardtransformationcoefficientmethod(comparetoHendenandKaitchuck(1982)andreferencestherein).Thismethodessentiallyfitstheobservedinstrumentalmagnitudes to a 3D surface todetermine the transformation coefficient (e),extinction coefficient (k' ), and zero-point offset (z
n).The remainder of this
section reviews this method by outlining the mathematics required to findtheseparameters.
3.2.1.Determiningeandzn
For differential photometry in which airmass may be neglected, thetransformation coefficient (e) and zero point offset (z
point offset of the camera.The subscript i denotes the ith calibration star intheimage.BecauseofthewayinwhichCMOSsensorsaremanufactured,weassume,tofirstorder,thattheresponseofeachpixelinthecameraisnearlyidentical.Therefore, ifproperbackgroundand flat subtractionmethodshavebeenappliedeandz
3.2.2.Airmasscorrections Theabovemethodofcalibratingisgoodforimagesofsmallangularextent(that is, those with < 3˚ FOVs) at zenith angles less than 34 degrees. Beyond thispoint,thedifferentialairmassacrossthefieldcancontributesignificantlytotheerror.FirstorderairmasscorrectionsmaybeappliedtoDSLRimagesusingthefollowingequation(HendenandKaitchuck1982):
(V – n)i=–k'nXi+e(B–V)i+z
n (3)
where thenewly introducedvariable,k'n, is theextinctioncoefficientandXi
is the airmass. This equation has the same functional form as a geometricplaneinthreedimensions:z=Ax+By+C.Ifweassumethattheinstrumentalmagnitude,ni,dependsonlyonthetermsontherightsideoftheaboveequation,thenwemaysolve theaboveexpressionfor thecoefficients (–k'
It isnotnecessary towriteacomputercode to solve theseequations,asmany spreadsheetprogramsandprogramming languages alreadyhavebuilt-
Kloppenborgetal., JAAVSO Volume 40, 2012822
in routines forsuchapurpose.Forexample,excel/openoffice calchave the“linest’’ function which we have employed in the Intermediate ReductionSpreadsheetontheCitizenSkywebsite.Ifyouwishtowriteyourownreductioncode,Python’s“scipy.optimize.leastsq’’functioncanbeusedforthistask. After thecoefficientsaredetermined, themagnitudeof the jth star in thefieldofviewmaybedeterminedbyrearrangingEquation6:
Vj = nj+–k'nXj +e(B–V)j+z
n (6)
Note that these equations require that the color of the target stars mustbeknowna priori!ThisplacesaDSLRcameraatasignificantdisadvantagebecause even though the Blue and Red channels are measured, they do notcorrespond to any standard photometric filters. Furthermore, the spectralresponseoftheRedandBluepixelsoftenareasymmetricwithmodestredandblueleakscomparedtostandardfilters.ThisdisadvantagecanbemitigatedbyusingacatalogthatcloselyrespondstoDSLRG,therebyresultinginanear-zerovalueforeandmitigatingthecolorcontributiontofinalV-bandoutput.
3.3.Verification InadditiontotheAPL-basedpipelinedescribedabove,oneofus(H.B.E.)createdanalternativereductionmethodthatusesaip4win(BerryandBurnell2005)tostackimages,sourceextractor(BertinandArnouts1996)toautomaticallyfind and perform aperture photometry on stars, and scamp (Bertin 2006) toperformastrometricstarassociation.Theoutputisthenprocessedinthesamemannerasstep2describedaboveusingascriptwehavewritteninR(Matloff2011).This second pipeline produced results identical (within uncertainties)tothemethoddescribedabove.PleasecontactH.B.E.ifyouareinterestedinvirtualmachineimageofthefreelyredistributablecomponentsofthispipeline.
4.1.1.Tycho For our first calibration catalog we used the Tycho 2 Catalogue (Høget al. 2000a, 2000b). As discussed above, the difference in transmissionbetweenVTandDSLR-Gfilters isminimal. InFigure5aandFigure5bweadjustedDSLR-GtoVTusingonlyazeropointoffset,z
n.Theresiduals(that
is, transformed–catalog) appear normally distributed about zero and shownosystematictrendsasafunctionofcolor,confirmingthesuspicionthatthedifferenceintransmissionbetweentheVTandRGB-GfilterscanberegardedasminimalfortheCanon450D.
4.1.2.ASCC Unlike Tycho VT, the Johnson V measurements in the ASCC catalog(Kharchenko2001)shouldexhibitamodestcolor transformationcoefficient.Foralmostallofourtargetstars,ASCCcontainsVTmagnitudestransformedtoVJ . In Figure 6a we plot the instrumental magnitudeninst as a function oftheASCC2.5JohnsonV.Todemonstratethecolorcorrelation,thedifferencebetweeninstrumentalandcatalogmagnitudeisplottedasafunctionof(B–V)inFigure4b.NotethatunlikeouranalysisfortheVTphotometry,andunliketheanalysisinHoot(2007),wefindasignificantcorrelationbetweentheresidualsandcolor. NextweappliedthefullcolorcalibrationtothedatatoyieldFigures5cand5d.Asidefromasmallexcursionbetween0.4<(B–V)<0.5(whichcontainsonly stars with V > 8 with very poor SNR), we find the data are normallydistributed. Likewise, residuals as a function of magnitude appear normallydistributedwithinanenvelopethatis∝1/SNR.Combined,theseimplythatthetransformationequationsforDSLRGtoVJarevalidacrossourentirerangeofcolorsandmagnitudesinoursample.
4.1.3.SIMBAD InFigures5e,5f,and6cweessentiallyrepeattheASCCexperimentusingaheterogeneousreferencecatalogassembledfromSIMBADqueries.Boththemagnitude and color residual diagrams show a loss of precision of ~ 0.005mag.,with a few stars shiftingby asmuchas0.02mag.Althoughcertainlywithin the statisticaluncertaintiesquoted in thecatalog, thesedeviationsareoftenoutsideoftheinternalinstrumentaluncertainty,implyingthedeviationsareduetoerrorsinthecatalogmagnitudes.Wecautionthereaderthatusingaggregatecatalogs,likeblindqueriesfromSIMBAD,mayresultindegradedprecision.ThereforewesuggestthatDSLRphotometrybeperformedusingastandardreferencecataloglikeASCCor,preferably,TychoVT.
5.1.Onthenumberofreferencestars As discussed above, one may calibrate data using the simple method(Equation 2) using only two reference stars and the airmass-correctedversion(Equation6)withonlythreestars.However,whenusingtheselowerlimitsasaguide, the readermustbeaware thatan incorrect identification,bad instrumental magnitude, or incorrectly referenced catalog value willsignificantlydegrade (ifnot invalidate) the results.Asa ruleof thumb,werecommendsixtoninereferencestarsthatbracketthetargetobject(s)incolor,airmass,andmagnitudesothatthevaluesofk',e,andz
nmaybeinterpolated
ratherthanextrapolated. In many cases, satisfying all of these requirements is not possible. Thereadermaybetemptedtoincludefainterreferencestars,butwiththatcomeslargerstatisticaluncertainty.Inourwork,wefoundstarswithaSNR>100areoftenacceptableandstronglycautionagainstusinganystarwithaSNR<100asacalibrator.
5.2.Airmasscorrections Most of our images were taken fairly close to the zenith (with typicalairmasses not exceeding 1.2), therefore the differential extinction across theimagewasnegligible.TwoserieswiththegreatestairmassinfieldCD6Sdidshowasignificantcorrelationofresidualsasafunctionofairmass(compareequation 6). Including airmass correction in the (planar) fit did improve theresiduals significantly when compared to the (linear) color-corrected fit.In general, airmass corrections should be applied whenever the differentialairmassacrosstheentireFOVmultipliedbytheextinctioncoefficientexceedthedesiredlevelofaccuracy.
Beforeweconsiderthesethreecasesinfurtherdetail,wewishtodescribetwomethodsbywhichblendsmightbeidentified.Inalargeensembleofstars,blendswillonlyhavealimitedeffectontheresultantphotometry.Blendedstarscanbedetectedinatleasttwoways.Givenpositionsfromanastrometriccatalogand the formal resolution of the imaging setup, stars should be consideredblended if the photometric extraction apertures overlap. Software that doesthis is available from author H.B.E. by request. If the reader wishes to useanempiricalmethodfordeterminingblends,blendedstarsmaybeidentifiedbyfindingstarsthatskewthephotometryerrorhistogram.Typicallyavisualinspectionoftheplotthatcomparesmeasuredmagnitudestocatalogdatawillhave some obvious outliers. These are most frequently caused by blending,variability,orothersourcesoferror(forexample,smallclouds). Thethreesourcesofblendingdeservefurtherdiscussion.Themethodbywhichthephotometristchooses todecreaseblendingdependsonthescienceobjectivetheywishtoachieve.Inourwork,alossof25%ofoursamplewasinconsequentialaswestillobtainedphotometryon150starswithonlyten10-secondexposures.Ofthefifty-twostarsthatwerelostduetoblending,almostall of them were due to the limited resolution of the optics.We could havesimplyincreasedtheangularresolutionofoursetupbyzooming,butthenwewouldrequirealargernumberofexposures. Inthecasewherestarsareblendedduetodefocusingortrailing,themethodofresolvingtheblendbecomesmoredifficult.Inthecaseofdefocus-inducedblending, focusing the image is the obvious solution; however, defocusingensures an accurate measure of the star’s light. In the high photon regime,one shoulddecrease the exposure lengthwhileproportionally increasing thenumberofexposures.Whenanalyzingthedata,theimagesshouldbealignedand stacked. Likewise, for bright stars that become blended due to trailing,the exposure length can be decreased while the number of exposures isproportionallyincreased. Thisadvicewilllikelynotextendtothephoton-limited,faintstarregime.Indeed, theauthorsareunawareofanystudy that theoreticallydiscusses thetrade-offsbetweenchangingtheintrinsicresolution,focus,andtraillengthwhileprovidingexperimentalverificationofanypublishedclaims.Untilsuchaworkiscompleted,wesuggestthereadercarefullyconsiderthesciencetheywishtoachieveandchooseasetupbestsuitedforthejob.Theprocedurewedescribehereinisideallysuitedforwide-fieldbright-starphotometricmonitoring,butclearlynotforobservingfaintstarswithlittlephotometricvariation.
6. Acknowledgements
Theauthorswould like to thank theAAVSO’sCitizenSkyproject.B.K.acknowledges support from NSF grant DRL-0840188. B.K. thanks Dr.Doug Welch for discussions that piqued his interest in DSLR photometry.
Kloppenborgetal., JAAVSO Volume 40, 2012826
The authors would also like to thankArne Henden for answering questionsconcerningphotometriccalibrationdatabases.Wearegratefultotheorganizersof theCitizenSkyworkshopsIandIIduringwhichwefirstdiscussed theseefforts.ThisresearchhasmadeuseoftheSIMBADdatabase,operatedatCDS,Strasbourg, France and several online collaborative tools like theAAVSO’sIRCchannel,GoogleDocs,andPastebin.
References
Berry, R., and Burnell, J. 2005, The Handbook of Astronomical Image Processing,Willmann-Bell,Richmond,VA.
Bertin,E.2006,inAstronomical Data Analysis Software and Systems XV,eds.C.Gabriel,C.Arviset,D.Ponz,andE.Solano, ASPConf.Ser.,351,Astron.Soc.Pacific,SanFrancisco,112.
Bertin,E.,andArnouts,S.1996,Astron. Astrophys., Suppl. Ser.,117,393.Fiacconi,D.,andTinelli,L.2009,Open Eur. J. Var. Stars,114,1.Guinan, E. F., and Dewarf, L. E. 2002, in Exotic Stars as Challenges to
Guyon, O., and Martinache, F. 2011, in Techniques and Instrumentation for Detection of Exoplanets V, ed. S. Shaklan, Proc. SPIE, 8151, 815111C,AmericanInst.Physics,Melville,NY.
Henden, A. A., and Kaitchuck, R. H. 1982, Astronomical Photometry, VanNostrandReinholdCo.,NewYork.
Høg,E.,et al.2000a,Astron. Astrophys.,355,L27.Høg,E.,et al.2000b,Astron. Astrophys.,363,385.Hoot, J. E. 2007, in The Society for Astronomical Sciences 26th Annual
Symposium on Telescope Science, Soc. Astron. Sciences, RanchoCucamonga,CA.,67.
Bull. Amer. Astron. Soc.,43,158.12.Littlefield,C.2010,J. Amer. Assoc. Var. Star Obs.,38,212.MaízApellániz,J.2006,Astron. J.,131,1184.Matloff,N.2011,The Art of R Programming: A Tour of Statistical Software
Páta, P., Fliegel, K., Klíma, M., Blažek, M., Řeřábek, M. 2010, in Applications of Digital Image Processing XXXIII,ed.A.G.Tescher,Proc.SPIE,Amer.Inst.Physics,Melville,NY,77982H.
Paxson,K.B.2010,J. Amer. Assoc. Var. Star Obs.,38,202.Perryman,M.A.C.,EuropeanSpaceAgencySpaceScienceDepartment,and
theHipparcosScienceTeam1997,The Hipparcos and Tycho Catalogues,
Kloppenborgetal., JAAVSO Volume 40, 2012 827
ESA SP-1200 (VizieR On-line Data Catalog: I/239), ESA PublicationsDivision,Noordwiijk,TheNetherlands.
Table 1. The 27 variable stars detected in our survey.1
Tycho VT Series Mean Typical SD Series Min. Series Max. No. Tycho Identification (mag). (mag.) (mag.) (mag.) (mag.) Sigmas Classif.
3528-2121-1 8.077 7.979 0.015 7.954 8.011 2 DA3529-1447-1 7.648 7.633 0.010 7.600 7.660 3 U3533-2577-1 5.219 5.195 0.003 5.182 5.209 5 1U3534-302-1 7.450 7.290 0.010 7.284 7.300 1 1D-3536-1939-1 7.359 7.379 0.006 7.356 7.420 7 M3536-2022-1 8.371 8.400 0.023 8.335 8.436 3 C3538-2150-1 8.436 8.415 0.015 8.351 8.483 53539-137-1 7.759 7.783 0.008 7.757 7.810 3 CU3539-1700-1 6.831 6.797 0.004 6.702 6.884 24 1U*0.513539-2623-1 8.313 8.389 0.015 8.345 8.461 5 3R3548-2346-1 7.227 7.234 0.005 7.225 7.246 2 U*2.793550-579-1 8.339 8.310 0.029 8.276 8.367 2 U*1.253551-1744-1 7.459 7.136 0.005 7.103 7.158 7 P3552-1543-1 8.438 8.464 0.014 8.427 8.513 33552-394-1 8.000 8.078 0.011 7.999 8.223 13 P3553-999-1 8.260 8.266 0.012 8.249 8.282 1 U3554-100-1 7.753 7.743 0.007 7.699 7.780 6 1U3554-1071-1 6.014 6.071 0.003 6.039 6.109 13 5U3555-686-1 7.559 7.551 0.008 7.525 7.582 4 U3564-1126-1 8.121 8.129 0.010 8.103 8.173 4 U3564-3159-1 6.231 6.215 0.003 6.151 6.256 21 5U3569-331-1 8.117 8.118 0.012 8.061 8.161 5 1U3908-1123-1 7.652 7.626 0.009 7.620 7.631 1 U3918-1829-1 5.867 5.932 0.003 5.921 5.943 4 3U3920-1660-1 8.451 8.401 0.014 8.367 8.431 2 D3920-1971-1 3.884 3.873 0.002 3.861 3.880 6 C53934-27-1 7.405 7.413 0.007 7.374 7.436 6 U1 Most stars have min./max. values that are 2+ times the typical nightly standard deviation. Of particular interest are TYC 3536-2022-1 which was labeled as “stable” in the Tycho 2 input and main catalogs. Stars TYC 3538-2150-1 and TYC 3552-1543-1 have no variability designation. All three stars have no variability designation in SIMBAD. Tycho classifications are: S=Standard, C=Stable in input/main catalog, U, P, M, R = confirmed variables. Numbers in Tycho classifications indicate variation type, see Perryman etal. (1997), and Wright etal. (2003) for designations.
Kloppenborgetal., JAAVSO Volume 40, 2012828
Table2.Summaryoftheobservedfieldsandexposureinformation. Field* Center Near TYC No. No. No. No. Air-mass R.A. Dec. Series Images Days Stars h m ˚ ' V<8.8
CD3S“A” 1935 +5116 3568-2325-1 21 245 9 56 1.0–1.14CD4S“B” 1909 +5123 3554-275-1 6 76 3 40 1.02–1.2CD5S“C” 1843 +5130 3539-1697-1 6 76 3 48 1.05–1.34CD6S“D” 1815 +5130 3537-1538-1 7 85 4 57 1.13–1.53*Fields are 15.9 × 6.4 degrees. The instrumental magnitudes for each of the 40 series of observations are available from R.P. by request.
Figure2.GeneralfieldpropertiesoftheCygnus-Draco15.85×6.36deg.fields,used in our experiment, and statistics for selected stars (after rejection ofblendedandvariablestars):(a)magnituderanges,Tycho2;(b)(BT–VT)colorranges;(c)Approximatespectraltypes.
(a)
(b)
(c)
Figure 3. A typical RAW star imagefromourdataset.Asmentionedabove,we use a non-tracking camera mount,hencethestarimageshowssignificanttrailing. The Red Green Blue natureof the camera’s Bayer array is clarlyvisible. The 21×13 pixel aperture isshownforreference.
Figure 5 (d, e, f). Residuals of the simple calibration method as a functionof color, magnitude, and catalog: (d) ASCC color residuals; (e) SIMBADmagnitude residuals; (f) SIMBAD color residuals.Aside from an increasedspread at fainter magnitudes (due to a lower SNR), there appears to be noremaining systematic residuals after application of the calibration procedurediscussedabove.Thestatisticaldistributionsofresidualsareshowntotheleftof they axes.All residuals followaGaussiandistributionwithavery smalloffset(~1mmag).Thedistributionistypifiedbyas=22mmaguncertaintywhichimprovesbynearlyafactoroftwoforstarswithV<8).ThisatteststothequalityofthecatalogandcapabilitiesoftheDSLRcamera.WebelievetheSIMBADresultsareskewedduetoinvalidcolordeterminationsinherentinainhomogeneouscatalogsystem.
Stellar Photometry With DSLR: Benchmark of Two Color Correction Techniques Toward Johnson’s VJ and Tycho VT
Roger Pieri37 C rue Charles Dumont, Dijon, 21000, France; [email protected]
Received April 24, 2012; revised June 18, 2012; accepted August 17 2012
Abstract DSLRsarenowroutinelyusedformeasuringtheVmagnitudeofstarsthroughtheirGchanneloutput.ThisrequiresatransformationtoJohnsonV,usingacorrectionbasedonthecatalogue(B–V)colorindicesofthestars.Thispaperreviewstheresponsesof theinvolvedpassbandsandproposesanalternate solutionusinga synthetic filtermadebycombining the threeRGBDSLRchannels.Theassessmentofthetwotechniquesthroughexperimentationbeinguncertain,wehave chosen touse a computer simulation instead.ThissimulationcombinesthemeasuredspectralresponsesoftheDSLRchannels,theatmosphericreddening,andstarspectrafromthePickleslibrary.
1. Introduction
The recent ε Aurigae campaign (thanks to AAVSO’s Citizen Sky, hereafter “CS”)hasbeenagreatopportunitytoexperimentwiththeDSLR's(DigitalSingleLensReflex)photometrycapabilities.Wehavehadtoobserveunderverydifferentconditions,sometimesnearzenith,butalsoatverylowelevationandhighairmass,duringtheconjunctionperiodthatwasanimportantphaseoftheeclipse.ThetypicalCSDSLRobserverworkswithastandardDSLRequipped with a regular lens and mounted on a photo tripod.The author'slens is a 200 mm, f /4, but most people use shorter focal length, typically70mm, f /2.8.Accuratephotometryisachievabledowntomagnitude6withthisconfiguration.Lessaccuratedatacanbegathereddowntomagnitude8.Ideally,anobservationmadeeveryfewdays(whenweatherpermitted)wasneededtoobtaingoodcoverageofthephenomenonovertwoyears.Foranamateur this ispossibleonlynearhome,mostof the timeunderurbanskyconditions,withtypically15to20minutesofobservation.Thisalsomeansunder an extinction much worse than any professional observatory wouldencounter. Not only was neutral extinction a problem, but also reddeningathighairmasswasmakingthingsmorecomplex(unusualforprofessionalobservers).Aspecificobservingmethodanddataprocessing techniquehasbeendevisedbytheCitizenSky’sDSLRteamtofitallsuchconditions.Itisknown at CS as the “intermediate spreadsheet” method (hereafter “CSIS”)andcanbefoundontheCSwebsite(AAVSOCitizenSky2009). AsimplermethodisalsousedbyCSpeoplewhenobservationsaredonebelowairmass1.5.Itisknownasthe“beginnerspreadsheet.”Itinvolvesonly
Initially,theauthorusedthetoolsandpublicationsprovidedbyC.Builonhiswebsite(Buil2012),hisbook(Buil1991),andthetutorialsfromCitizenSky(AAVSOCitizenSky2009),butsoondevelopedhisownsoftwarepipelinetomaketheoperationsmoreproductiveandtoexperimentwithvarioustechniquesincludingasyntheticJohnsonVfilter. Thegoalwastobetterextractusefuldatafromobservationsunderthepoorε Aur conjunction conditions. Another aim was to push the DSLR to perform at itsbestinbrightstarphotometry. TheDSLRitselfhasG-andB-colorfilterswhichdonotmatchthestandardCousins-JohnsonBandVpassbands(hereafterBJandVJ).TheDSLRRchannelistoofarfromtheCousinsRctobefaithfullytransformed.AspecifictransformationandextinctioncorrectiontechniquehasbeendevisedtotakecareofboththeGpassbandandtheCSamateurcontext(HendenandKaitchuck1982;KloppenborgandPearson2011).ThisCSIStechniqueisbasedonthecatalogue(B–V)andV-magnitudeofastar“ensemble”thatdeterminestherelatedcorrectioncoefficients. In the CSIS the differential neutral extinction is determined through amappingofthemagnitudedeviationsofstarsofalargeensembledistributedin the DSLR field-of-view (FOV). To do this, the fully “color corrected”instrumentalmagnitudeofthestarsisthencomparedtothecataloguemagnitudeof the stars. The resulting magnitude differences that are a function of thedifferentialairmassdeterminetheneutralextinctiongradientacrosstheFOV. Statistics have been extracted from the ε Aur observations and critical cases have been identified in this CSIS technique. The single-equationsystem(Equation5)usedbytheCSIShassomedrawbacksdependingonthedistributionofstars.ToovercomethatissuetheauthorsoonconsideredacolorcorrectiontechniqueindependentofthecataloguedatausinganRGBsyntheticfilterequivalenttoVJorVT(hereafter“VSF”).Itisimplementedbycombiningthe three RGB channel signals of the DSLR. The first experiments showedinterestingpossibilitiesbutitwasobviousthatanaccurateassessmentofthesetechniqueswasdifficulttoachievethroughobservations.Thenitwasdecidedtouseacomputersimulationforthatstudy. SeveraloptionsofprocessingoftheneutralpartoftheextinctionintheVSFtechniqueremainunderstudy.Theauthorplanstoaddressitinafuturepaperbasedonthebestoption. Bytheway,thepresentpaperaddressesonlythecoloraspectofboththeCSISandtheauthor’sVSFtechnique.ForsuchstudyonlytheDSLRresponseandtheslopeof theatmospheretransmissioncurveareusedinadifferentialphotometryschemewhereallstarsareatthesameairmass.
Pieri, JAAVSO Volume 40, 2012836
Carrying out such a simulation needed a library of stellar spectra. ThePickleslibrary(Pickles1998)hasbeenfoundtomeettheneedwithits131starsfromOtoM,atallluminositiesandseverallevelsofabundance.Otherinspiringsources have been the many papers from Bessell: in “UBVRI passbands”(Bessell1990)heanalyzed the issuesofvarious filter implementationsusedunderseveralphotometricsystems.Thepaperincludesasyntheticmagnitudecalculationusing theVilnius spectra library.Theapproach isvery similar towhatisreportedinthispaper.Hislastpaperhasaninterestingappendixpointingoutsomeissuesofsuchbroadpassbands(BessellandMurphy2011). Thepaperfrom(TokunagaandVacca2005),thephotometryintroductionof(Romanishin2006),andastudyofthecalibrationofTycho-2andJohnsonsystems(Apellániz2006)havealsobeenveryinformative. AnotherimportantsourcefortheDSLRapplicationistheHipparcosmission(Perrymanet al.1997).LookingatthevarioussystempassbandsrevealedthattheTychoVTpassbandiscomparabletotheDSLRgreenresponse.InadditiontheTycho-2catalogue(Høget al.2000)catalogueisauniform,veryaccurate,and reasonablyprecise reference.Thisopenedanewpossibility thatwillbeanalyzedhereafterandcomparedtoJohnsonVJusingboththeCSISandtheVSFtechniques. AtthispointtheauthorwouldliketodiscusstheaccuracyofthecataloguescomparedtowhatcanbeobtainedthroughDSLRphotometrytechniques.Suchtechniques,basedonmappingofmagnitudedifferences,areverysensitivetothecatalogueerrors.Fortunately,DSLRshavealargeFOVthatallowsonetoworkwithalargeensembleofstars(50to100starsaccessibleina6×4degreefield).ThisaveragesbothcatalogueerrorsandtheimperfectcolordistributionintheFOV.Ata1000signal-to-noiseratiotheDSLRprovidesresultsreproduciblewithinafewmillimagnitudes.Mostcatalogueshavelargeruncertainties,andfromcataloguetocataloguesuchuncertaintiesareoftenmuchlarger;thiscouldaffect the determination of both the color transformation coefficient and thedifferentialneutralextinctionwiththeCSIS(asitisbasedonamappingofitfromanensembleofstars).
3. Color Passband: DSLR’s RGB versus Johnson’s VJ and Tycho-2 VT
TheDSLRR,G,Bchannel responses are shown inFigure1.Theyhavebeenobtainedfromaclearday,sunlightobservationusingaslitandgratingmountedinfrontoftheDSLRlens,thencorrectedfromthestandardASTMsunspectrumat1.5airmass(ASTMInternational2008)andtheRGBbalanceofthegrating.ItistobenotedthatDSLRcurvesaretheresponsesofthefilterstack(R,G,B,IR,andUVcut)plusthesensorandthelens.ThelinearRAWoutput of the DSLR has been used; that excludes any processing from theDSLR—inparticular,anygammaapplicationoranytransformationtosRGBorothercolorprofile.TheDSLRwasastandardCanon450DwithaCMOS
Pieri, JAAVSO Volume 40, 2012 837
sensor,14bits,equippedwithaNikkor85mmlens.ItsIRandotherfilterstackhasnotbeenmodified. IntheUBVRI(andother)photometricsystems(Bessell1990),magnitudesare calculated from the photon-count from photometric chains of severalpassbands.Figure2showstheJohnsonV(hereafter“VJ”)passbandanditsrelationshipwiththeDSLRGchannel.TheversionusedasreferenceinthisstudyistheonerecommendedinthePickles(Pickles1998)spectralibrary.ThisversionisverysimilartotheonerecommendedbytheCDS.Theyareusedinanumberofpublications,butotherversionsalsoexist(BessellandMurphy2011). Figure3showstherelationshipbetweentheTycho-2VTpassbanddefinition(Apellániz2006)andtheDSLRGchannelresponsecurve. OtherDSLRcolorchannelsresponsescouldbeseenatthewebsiteofC.Buil(Buil2012).Theyareverysimilar.Apparently,therearenolargedifferencesbetweenrecentDSLRorDSC(digitalstillcameras;largedifferencesmaybefoundinoldercameras). ThecolorfilterarrayoftheDSLRisoftheBayertype(Bayer1975)withasquareRG/GBarrangement.ThatmeanstheGsignalisobtainedfromtwosub-pixelsinsteadofonlyoneforRandB.Inaddition,asshowninFigure1,thetransmissionoftheredfilterismuchlowerthanGandB.AsaresulttheSNRoftheGchannelismuchhigherthantheBand,inparticular,theRchannel.TheRAWresultingcolorbalanceismoreorlessCIE(CommissionInternationaledel'Eclairage)typical(Greenet al.2002).Inmostcommonusage(sRGBandJpegoutput)theRGBlevelsarerealignedto1,1,1.Thentherepresentationisnon-linear(0.45gamma)codedon8bitsandtosomecolorprofilelikesRGB.ThisstudyusesonlythelinearRAWoutputofR,G=(G1+G2)/2andBdata(saidADUs,proportional tophoton-count,hereafterallphoto-countdataaredenotedas“G”,italicizeduppercase,andmagnitudesas“Gm”). The instrumentalmagnitude,Gm, isusuallycalculatedwith theequation(Equation3)fromtheDSLRG-channeloutput,G,thentransformedtoVJGtm usingEquation4.IfwecomparetheGresponsetotheVJresponse(Figure2),thereisalargecommonsurfacebetweenthemboth,butalsosignificantdifferences. Thebluesideroll-offisthelargestsignificantdifference.ThecentroidoftheGresponseisat530nmandtheVJoneat542nm.Buttheblueroll-offoftheGresponseisshifted26nmtothebluecomparedtothesteeproll-offoftheVJcurve.Inadditionthereisablueleakatthefootofthecurvebetween450and400nm. Ontheredsidethedifferenceislessimportantandopposite:missingredinsteadoftoomuchblue.TheredsideGroll-offisshiftedtothebluebyabout5nm.Figure3showstheTychoVTpassbandwhichismuchmoresimilartotheGchannel.Thecentroidspositionsaredifferentbyonly2nm.Theresultingcolorcorrectionbetweenbothshouldbeminimal.WewillalsostudythatcaseasitoffersarealopportunityforDSLRphotometry.
Pieri, JAAVSO Volume 40, 2012838
4. Atmosphere transmission
The bandwidths of all DSLR-involved passbands are not small. Besidestheimpactofthemeanextinctionatthecentroid,thepassband’sinternalcolorbalance is affected by the atmosphere transmission variation as a functionof the wavelength (essentially the slope of this function). The result is theproduct of both: the camera response S(l) and the atmosphere transmissionT(l).Thismore-or-lessmovestheresultingcentroidtothered.Theatmosphereaffects all passbands including the references, Johnson’sVJ, andTychoVT.The atmosphere attenuates the short wavelengths more than the long ones.This is shown in Figure 5, extracted from the Moon’s solar irradiance table(Desvignes 1991).That effect is weighted by the length of the light path intheatmosphereknownas theairmassnumber(AM),after theBeer-Lambert-Bouguerequation:
T(l,AM)=e–e(l)AM (1)
WheretheairmassAMisafunctionofthestarzenithaldistancez(orelevationh=90–z),ande(l)istheextinctioncoefficient.IntheDesvignestablee=0.23atl=550nm.ItcouldalsobeextractedfromthesolarirradianceatAM0andAM1.5 from theASTMstandard (ASTMInternational2008).The solarenergystandardsbetterfitoururbanamateurskyconditionsthanthedatafromprofessionalobservatories.
Where M is the airmass in a planar homogeneous atmosphere model (validdownto10degreeselevation),andAMthecorrectedairmassaftertheHardie’sapproximation (valid down to 5 degrees elevation) (Ilovaisky 2006). Theseequationsarenotused in thesimulationwhere the inputof the transmissiontableisdirectlyAM. Inthefollowingsimulationonlythechromaticpartoftheextinctionisusedandstudied(thetransmissionslope).TheneutralpartofitanditsdifferentialeffectintheDSLRFOViseliminatedbythedifferentialphotometryandthefactallstarsaresetatthesameairmassinthissimulation.
5. CSIS color transformation from the DSLR Gm, instrumental magnitudes, to Gtm, VJ, or VT magnitudes
The CSIS color transformation (Henden 2000) is a linear interpolationbetweenthecolorindicesofthestars,thetargetoneandatleasttworeferencestars(an“ensemble”beingbetter).Thecolorindexisthemagnitudedifferencebetween two passbands like (B3 – V). (Pickles 1998; B3 denotes a specific
Pieri, JAAVSO Volume 40, 2012 839
Johnson B filter adopted for the Pickles spectra. It is not far from the filterrecommendedbytheCDS.)Themagnitudedifferencefromathirdpassbandislinearlyinterpolatedfromtheknowncolorindices.Theinterpolationcoefficient“kD”isexpectedtobespecifictoagivenDSLR.Asimilartermshallalsoaccountforthereddeningduetothechromaticpartoftheatmosphericextinction“kK”.Itshallbereadjustedforeachstarfieldasafunctionofitsdistanceangletothezenith.Within40degreesofthezenithitaccountsforabout2millimagnitudesonlywithintheusual±0.7(B–V)range,thenitcouldbeignored.Atairmass1.5 ~ 4 the impact shall be taken into account (3 ~ 30 millimagnitudes).
Gm=m0–2.5log(G/G0) (3)Gtmi=Gmi+zp+kD(B3 – V )i+kK(B3 – V)i+kNXi (4)
GmistheinstrumentalmagnitudefromtheGandG0photon-countissuedfromthegreenchanneloftheDSLRrespectivelyforatargetstarandacomparisonstar.(B3 – V )isthecataloguecolorindexofthestar(i).Gtmisthetransformedmagnitudeof the star. zp is the zero-point, themagnitude constant from theinstrument.InagivenobservationkDandkKcan’tbeseparatedwhensolvingtheequationsystem(Equation5)andshallbereplacedbykC=kD+kK.IntheDSLRFOV,Xiisthedifferentialairmassofthestar(i)fromthereferencestarorensemble. InthesimulationreportedinthispaperG0isthephoton-countofanA0Vstarofmagnitudem0=0and(B3 – V )~0resultinginzp~0.ThenallstarsoftheFOVbeingsetatthesameairmass,allXi=0andkNarenotapplicable. Fromactualobservations, thecoefficientsforagivenDSLRandagivenFOV are determined from the catalogue values (B3 – V ) and Vjm of severalreferencestarsandtheirmeasuredGmvalues(orhereinafterBT – VTandVTminthecaseofTycho-2).
Vjmi=Gmi+zp+kc(B3 – V )i+kNXi (5)
ApplyingtherelationEquation5totheobservationofmorethanthreestars(i)formsanoverdeterminedequationsystemthatissolvedusingtheleastsquareerror algorithm. This method is used to determine zp, kC, and kN from thisensembleofstars.OnlykCisusedinthereportedsimulation. Characterizing the DSLR kD through observations instead of throughsimulation is tricky.Suchobservationsshouldbemadeunderverygoodskyconditions,nearthezenith,usingwell-documented,stablestarsofamatching(B – V)range,athighSNR(>300),andrepeatedseveraltimes.TheextinctionshouldbeindependentlycheckedusingtheBouguer’slinemethod(Ilovaisky2006). If the resulting extinction parameters are far from standard and/orirregular as a function of the star’s elevation, the sky conditions should beconsideredinadequateforacalibration.
Pieri, JAAVSO Volume 40, 2012840
6. Definition of a DSLR “VSF” synthetic filter delivering a VJ- or VT-like magnitude
ItisnotrecommendedtoapplyEquation8todataobtainedfromobservations.The various errors due to the observational conditions and the catalogueuncertaintiescanhaveasignificantimpactonthe(a,b)coefficients.SucherrorscanunbalancethecorrectionsfromtheBandRchannels.Thiscouldresultinagoodcorrectionforagivenstarsetbutonethatisnotoptimalacrosstheoverall(B–V)range.It’sbettertousethecoefficientsdeterminedfromtheRGBpass-bands. Theway(a,b)workscanbeunderstoodascontrollingthecentroidandthebandwidthoftheVSF.When(a,b)varyinoppositedirectionstheycontrolthebandwidth,otherwisetheyshiftthecentroidoftheVSF. This G,R,B combination results in a synthetic filter (VSF) which has aresponsesimilartoVJ(orVT)inthevisibledomain.TheVSFpassbandshapeisnotexactlythesameasVJbutrespondssimilarlytothespectrumcontinuum(Figure18).TheVSFresponsetospecificfeaturesofthespectra,likeBalmer’slinesofbluestarsormolecularbandsoftheredones,coulddiffer(Figure4).A first check has been made using the black body spectrum at various startemperatureswithexcellentresults(withinonemillimagnitude). TheproposedVSFtechniquehasalsobeenextensivelytestedonthestarfieldoftheeAurcampaignwithexcellentresults.Undergoodskyconditionsthestandarddeviationwithinanensembleoffivestars(B–Vfrom–0.18to1.22)
Pieri, JAAVSO Volume 40, 2012 841
andtheircataloguevalueswasoftenaslowasacoupleofmillimagnitudes.Butthatfieldhadnospectrum-criticalstar.Next,therearefewreferencestarsofvarioustypeshavingVJmagnitudesknownwithhighaccuracyinasinglefieldof5~15degrees(DSLRFOV).Ithasbeenjudgedeasier,moreaccurate,andinformativetoassessthistechniquethroughacomputersimulationusingthecoherentPicklesspectralibraryandthemeasuredRGBresponsesoftheDSLR. IntheCSIStransformationthechromaticpartoftheextinctionisincludedinkC,andweseethatitshallbesetbeforethedeterminationofthedifferentialneutralextinctionintheDSLRFOV. ThiscanbeachievedwiththeVSFtechniquebyadaptingthecolorbalanceof thesyntheticfilter.This is justmakingthe(a,b)coefficientsadaptivetoameasureoftheatmosphericreddening(Equation9).TheBc/GcratiooftheB,GoutputsoftheDSLRisanexcellentmeasureofit.Itisverystableduringanobservationandlooksmuchmorereliablethananysinglechanneloutput.
a=a0+a1Bc/Gc b=b0+b1 Bc/Gc (9)
TheBc/GcratiousedisthatofareferencestarintheFOVorofanensemblecenteredonthe(B–V)range.ThisatmosphericreddeningcorrectionshouldnotbeconfusedwiththemainfiltersynthesisoperatedbyEquation6.Itproducesonly a limited centroid shift of the synthetic filter that compensates for thereddening,andafirstorderlinkto(a,b)hasbeenfoundaccurateenoughfortheCanon450D.AsecondordercouldbeneededforsomeolderDSLRshavingalargeblueleakintotheGchannel.
7. Pickles spectra library
ThePickleslibraryincludes131spectraofstarswith(B3 – V )from–0.38to1.816andO5toM10spectraltypes.AllluminosityclassesItoVarerepresented.Thespectracomprisewavelengths1150Åto25000Åataresolutionofabout500(Dl/l).Thelibraryhasbeenconstructedbycombiningthedatafromsixteenother libraries with a dedicated combination and verification methodology.DetailsmaybefoundinPickles(1998). ForthepurposeoftheDSLRsimulationtherangefrom350nmto750nmhas been extracted from the original data (Figure 4). The library providescalibrated spectra in energy density per Ångstrom; then the flux has to beconvertedtophoton-count.Theoriginalnormalizationat5556ÅofthespectrahasbeenreadjustedfornormalizingtheoutputoftheVJ(orVT)passbandtozeromagnitudeatzeroairmassforallstars.ThismakestheassumptiontheVJmagnitudesaredefinedontopoftheatmosphereaftertheJohnsondefinition(Bessell1990). MoststarsofMspectraltypeshowlargecolortransformationerrors.Thisisduetotheirspectrumhavinglargeabsorptionmolecularbandsresultingintheir(B3 – V )colorsnolongerreflectingtheblack-bodyfluxandpredictingthe
Pieri, JAAVSO Volume 40, 2012842
correctfluxcorrectiontoVJ.ForMstars,amagnitudecorrectionbasedonthe(V–Rc)colorismoreappropriate(RcisnotaccessibletoDSLRexceptifIRcutandIRdyefiltersareremoved).Mstarshavebeenprocessedseparatelyand110O-to-Kstarsremaininthemainanalysis. The ninth star of the library ofA0V spectral type has been used as thecomparisonstar(G0photon-count)inthemagnitudecomputations.ItsVJ(VT)magnitudehasbeensettozerobutthecolorindexdeterminedbyitsspectrumis0.015.Thisinducessomefewmillimagnitudesglobalshiftoftheendresult.
8. Computer simulation of the DSLR outputs and color processing
responseS(l).TheresultsafterintegrationarethenewVjorVTandR, G, B photon-counts. The lastphase is tocompute theendresultsofboth techniquesfromthephoton-countsusingEquations3–4and6–7.TheresultsarethefinalmagnitudesVjm(thereference,orVTm),Gtm,Gcm(thecorrectedresults). Hereweshouldbringtomindthatallstarsareatthesameairmassunderevaluation.Nodifferentialneutral extinction is involved in the result.Theseresults are the deviations due to the combined camera passbands and theatmospheric reddening, depending on the star’s color. They are graphicallyshownbelow. InafirstpassthesoftwareisalsousedtocomputeonlyR,G,BandcalibratethecoefficientskC,a,andbusingEquations5and8. Allthatcomputingchain,uptoEquation3orEquation7,usesthenotionoftransmissionandphoton-countinsteadofthelogarithmicprocessingusual
9.1.Airmass1instrumentalresultsunderbothVJandVTsystems Johnson System—Figure 6 shows the respective instrumental magnitudedeviationsatairmass1oftheJohnsonVJpassband,Vjm,andtheDSLRgreenchannel,Gm.Vjmwasnormalizedto0atairmass0.Thedeviationsaremoreorlessalinearfunctionof(B3–V ).Thepeak-to-peakdeviationforGmisabout260mmgacrossthecolorindexrangeof2.2magnitudes.Theslopeisroughly0.118.ThisGm(B3–V )curveiswhatweshallusetotransformtoJohnsonVJ. Thereisasetofoutliersabove(B3–V )=1.1magnitude.ThosearetheMstars,whicharebluerthantheKstars,althoughtheyarecooler,duetolargemolecularabsorptionbandsintheirspectra.TheireffectivetemperatureismuchlowerthanexpectedfromtheircolorindexdefinedbytheBJandVJpassbands.Inthefollowingtheyareexcludedfromthemainanalysis.Thisissuehasbeendenoted in the Hipparcos and Tycho report (Perryman et al. 1997) and theauthorsrecommendnottotransformtheMstarmagnitudes(likeTycho,wecanonlyworkwithBandV,astheDSLRRistoofarfromRc). TheVjmJohnsonoutputitselfshowsasmalldeviationatairmass1duetothe star color—about10millimagnitudesacross the range.This is smallbutwouldneedalinearcorrectionwhenprecisephotometryisrequired. Tycho System—Figure7 shows the comparativedeviationsof theTychoVT passband VTm and the DSLR green channel Gm, without correction, atairmass1.VTmisnormalizedto0atairmass0.TheVTmdeviationatairmass1issomewhatlarger(19millimagnitudes)thanVjm.ThisisduetothelargerandbluerVTpassband.Theresponseof theDSLRgreenchannelshowsamuchsmaller deviation than under the Johnson system (23 millimagnitudes peak-to-peakinsteadof260).Thisconfirmstheanalysis(section3)madefromtheresponsecurvesandtheircentroids.TheoutliersaretheMstarsandafewKstarsasseenintheVJplot(Figure6),butwithasmallerscatter. This confirmation is important, showing the Tycho-2 catalogue shouldbe the reference for DSLR photometry. This catalogue is accurate in themagnituderangeaccessiblewithouta telescopeandprovideswell-quantifieduncertainties.
9.3.Johnson’sVJ—airmass1and4—CSISandVSFtechniques Figures10to13showthedeviationsofGtmandGcmforthe110starsoftypeO-to-K. CSIS—TheCSIStransformationresult,Gtm(Figures10and12)hasbeenoptimizedwith respectively kC=0.135 and0.098 for airmass 1 and4.This0.135valueisingoodagreementwiththeonetypicallyused,atzenith,foroureAurobservations. Thedeviationpatternissimilaratbothairmassesandformstwolinesfroma maximum around 5 millimagnitudes at (B3–V ) = 0.33 and a range of 20millimagnitudes.Usingslightlydifferentcoefficientsforthecolorindexranges–0.4~0.33and0.33~1.8(twodifferentslopes)wouldimproveit.Thisshouldreducetheerrorrangetoabout8millimagnitudesexceptafewoutliers.ThoseoutliersareK-typestarsthatshowsomeproblemssimilartotheMtypes.OnthebluesidethereareoneB3IandoneF5Ioutlierwithoutclearreasontodeviate,althoughtheadditionalvioletfluxinsupergiantsissuspectedtobethecause. VSF—TheVSFresult,Gcm(Figures11and13)hasbeenoptimizedwith,respectively,(a=0.284,b=0.224)and(a=0.149,b=0.244)forairmass1and4.Thisisjusttheexperimentalvalueforb,andasomewhathighervaluefora.Buttheredcorrectioncoefficienthasalargevariabilitydependingontheexperimentalconditionsduetoitsweakeffect.aequaltoorabove0.284isnotuncommon. The airmass 1 deviation pattern differs from Gtm with a well containedsection above (B3–V ) = 0.3 within 7 millimagnitudes and no K outlier.Atairmass 1 the section below 0.3 shows more dispersed results (within 15millimagnitudes)oftheOBAstars. Thepatternatairmass4isreducedtoa12millimagnituderangebuthasthesameshapeasairmass1.BotharebettercontainedthanGtm,theresultoftheCSIStransformation.
9.4.TychoVT—airmass1and4—CSISandVSFtechniques Figures14to17showthedeviationsofGtmandGcmforthe110starsoftypeO-to-K. CSIS—TheCSIStransformationresult,Gtm(Figures14and15),hasbeenoptimizedwith,respectively,kC=0.009and–0.019forairmass1and4.This0.009iswellinagreementwiththevaluetypicallyused,atzenith,inanothermassivesurveyexperimentbasedonTycho-2. Thepatternofthedeviationsatairmass1ismoreorlesssimilartothatof VJ but with a reduced amplitude of 10 millimagnitudes excluding the
Theaccuracyofthesimulationresultsdependsuponthreeitems:thespectralibrary,thepassbandresponses,andtheatmosphereextinctionmodel.Thelastisjustastandardmodel,andtheobservationsclearlyshowlargedeviationsfromsuchamodel.TheBc / Gcratiodependencytotheairmassisusuallystableandtheneutralextinctionmuchmorevariable.Thisistheresultofthevariabilityofthehighaerosolcontentofoururbanskies.SuchaneutralcomponenthasnoimpactandanyslopechangeofT (l,Am)isacolorbalancedifferencethatwouldjustaffectthecorrectioncoefficients,nottheenderrorlevelshownbythesimulation(thechromaticanddifferentialneutralcorrectionsareindependentintheVSFtechnique). ThenextpointistheDSLRresponsecurveaccuracy.ThepassbandshavebeentestedbycomparingthesimulatedRGBoutputswiththephysicalRGBoutputs of theDSLRandhavebeen found tobewithin a coupleof percentandcorrected.Theresponsecurveshavebeenmeasuredusingthreedifferentreference light sources and found similar. The differences impacted thecorrectioncoefficientbylessthan2%. The spectral features: lines, bands, but also some small differences inlargebandsofthecontinuum,areexpectedtobethereasonforthestar-to-stardeviationsseen in thesimulation results.ThenwehaveastandarddeviationfromPickles(hereafterSD)ateachwavelengththatranksfrom5%to0.5%.Butwedon’tknowthecorrelationfromonewavelengthtothenextthatcouldformsuchsmall“bumps”ofthecontinuum(afewpercent)overabandwidthlarge enough to interact with the roll-off of the DSLR passbands. In thehypothesis,wherewepropagatethoseSDswithoutanycorrelation,wewillgetaworseSDof3millimagnitudesattheendwhichseemssomewhatoptimisticwithregardstothesimulationresults.Then,ifwetaketheworstcasepossiblecorrelation,wewillgetadeviationof50millimagnitudeswhichisobviously
Pieri, JAAVSO Volume 40, 2012846
fartoolargeandunlikely.Thisaspectwouldneedadeeperstudyofthelibraryconstruction,whichcombinesavariablenumberoforiginalspectradependingthewavelengthandthestarspectraltype.Therearecasesbasedonafewdataonly.Anyhow,thePickleslibraryisanexcellentresourceandthebestofthosetestedbytheauthorforthiswork. The VSF technique has been extensively used during the ε Aur campaign withgoodresults (reportscanbe foundunderPROCatCitizenSkyand theAAVSO). It permits us to achieve a fully VJ color response of the DSLRindependentlyfromthedifferentialneutralextinctioncorrection(andothererrorsources).ThisavoidspossiblecontaminationfromotherfactorsthatexistswiththeCSIStransformation.ThisCSISissuedependsonthestarcolordistributionandtheirpositioninthefieldofview.TheVSFtechniqueisalsoindependentofthecataloguesafteritscalibrationisset. The VSF technique has been very effective when we had to observe ε Aur at low elevation during the conjunction. It is also very helpful in cases ofcolorchangeofavariable.ThishasbeenrecentlythecaseofzAur,forwhichinterestingresultshavebeengathered. AfurtherimprovementoftheDSLRphotometrywillbetousetheTycho-2 catalogue as a primary reduction reference. The DSLR instrumental colorcorrection isvery lowunder thisphotometric system;only thehighairmasschromaticeffectissignificant.Thisshallfurtherimprovetheaccuracyofbothcolor and neutral extinction corrections thanks to the better uniformity andaccuracy of this catalogue. If a Johnson’s Vjm is needed the end result canbeconvertedusingtheTychorecommendedmethods(Perrymanet al.1997).Amoredetailed tableof corrections as a functionofBT–VT is provided in(Bessell2000).
11. Conclusion
Itispossibletoachieveacolorcorrectionoftheinstrumentalmagnitudesof a DSLR to Vj or VT, within ±5 millimagnitudes, with both the CSIStransformationandtheVSFtechnique.TheCSIStransformationwouldneedsomerefinement(secondordercoefficient,ortwoslopes)toachieveitacrossa wide B–V range under the Johnson system.The color and airmass of thereferencestarsusedintheCSISshouldbewell-balancedandbracketthetargetstarsinallaspects. M-starmagnitudesarenotwell-transformablefromtheDSLRpassbands,withpossibleerrorsaslargeas60millimagnitudes.SomelateKstarscouldalso deviate by 20 millimagnitudes after the CSIS transformation. Theluminosityandtheabundanceofthestarsseemnottobesignificantfactorsofdeviation. The VSF technique provides the best accuracy, and is independent ofthecatalogue(B–V),ofthedifferentialneutralextinctionandothervarious
Pieri, JAAVSO Volume 40, 2012 847
errorsources.ItshowsnospecificerrorfortheKstarsandgenerallyworksbetterwith red stars.Using it under theTycho-2 systemprovides a furtherimprovement, reducing the color related errors and also providing a moreuniform reference for determining the differential neutral extinction in thefieldofviewoftheDSLR.
12. Acknowledgements
TheauthorwouldliketothanktheCitizenSkystafffortheirinitiativeandsupport during the ε Aur campaign, Brian Kloppenborg for his support and involvementinDSLRphotometry,andtherefereeMikeBessellforhisusefulcommentsonthepaper. This work has made use of the SIMBAD database, operated at CDS,Strasbourg,France.
Perryman,M.A.C.,EuropeanSpaceAgencySpaceScienceDepartment,andtheHipparcosSciencTeam.1997,The Hipparcos and Tycho Catalogues,ESA SP-1200 (VizieR On-line Data Catalog: 1/239), ESA PublicationsDivision,Noordwiijk,TheNetherlands.
Pickles,A.J.1998,Publ. Astron. Soc. Pacific,110,863.Romanishin, W. 2006, An Introduction to Astronomical Photometry Using
AM Gc mean Bc /Gc mean VJ/kc VJ/a VJ/b VT/kc VT/a VT/b
0 1027 0.805 0.145 0.340 0.217 0.020 –0.064 0.089 1 800 0.750 0.135 0.284 0.224 0.009 –0.085 0.073 2 624 0.701 0.125 0.235 0.231 –0.002 –0.103 0.058 3 488 0.656 0.110 0.190 0.238 –0.012 –0.121 0.045 4 382 0.616 0.098 0.149 0.244 –0.019 –0.136 0.030*Table 1 shows the various coefficients used in the simulation at various airmasses. “Gc mean” and “Bc /Gc mean” are the mean output intensities of the DSLR channels for the O-to-K spectra “ensemble”(orofaF6Vstar).Theextinctioncoefficientusedisε=0.23at550nm.TheBc /Gc ratio is used to measure the atmospheric reddening as a function of the airmass, AM. Gc mean results from the atmosphere transmission for the G channel after the Moon’s model (Desvignes 1991). “kc”, “a”, and “b” are the correction coefficients defined in sections 3 and 4, applied to both Johnson and Tycho standards. The values at AM0 are the correction coefficients of the DSLR G channel passband alone, normally invariant.
Pieri, JAAVSO Volume 40, 2012 849
Figure 4. Pickles library (Pickles1998), samples of M3V andA0V starsof the131 spectranormalizedat555.6nm. Spectra are in energy density perwavelength(nm).
Figure 5. Transmission factor of theatmosphere at airmasses 1, 2, and4, computed from the solar energydensity (Desvignes 1991). Theextinctioncoefficientis0.23at550nm.
Figure 2. Photonic response of theJohnson VJ passband compared to theDSLRgreen(G).
Figure3.PhotonicresponseoftheTychoVT passband compared to the DSLRgreen(G).
Wavelength (nm)
S(l
) A.U
.
Wavelength (nm)
S(l
) A.U
.
Wavelength (nm)
F (la
mbd
a)
Wavelength (nm)
Tran
smis
sion
Fac
tor
Figure 6. Johnson’sVjm and DSLR Gmmagnitude deviation, without correction,atairmass1under theJohnson’ssystem.Vjm is normalized at airmass 0.The Gmerrorsare11timesthoseofFigure7.
Figure 7. Tycho VTm and DSLR Gmmagnitude deviation, without correction,atairmass1undertheTychosystem.VTmisnormalizedatairmass0.TheoutliersareMstarsandacoupleofKstars.
Pieri, JAAVSO Volume 40, 2012850
Figure 8. e_Vjm magnitude deviation ofthe VJ passband at airmass 0, 1, and 4,under the VJ photometric system. Theresultsshownarethoseofthe110O-to-Kspectraofthelibrary.
Figure9.e_VTmmagnitudedeviationofthe VT passband at airmass 0, 1, and 4,under the VT photometric system. Theresultsshownarethoseofthe110O-to-Kspectraofthelibrary.
Figure 15. e_Gcm deviation of the VSFfilteroftheDSLRagainstTychoVTmatairmass1.
Figure 14. e_Gtm deviation of the CSIStransformationof theGmDSLRchanneltoTychoVTmatairmass1.
Figure 16. e_Gtm deviation of theCSIS transformation of the Gm DSLRchannel to Tycho VTm at airmass 4.OutliersarefewKstars.
Figure 17. e_Gcm deviation of the VSFfilteroftheDSLRagainstTychoVTmat
Figure 18. Response of the synthetic filter for a correction toVJ at airmass 0.Thiscorrespond to the correction of the DSLR alone. The DSLR R and B outputs arecombinedtoGweightedwiththecoefficients“a”and“–b”fromTable1(a=0.34;b=0.217).Theresultingcentroidsofthetwopassbandskeepwithin1nminthe–0.4~1.8(B3–V )range.Thenegativesectionintheblueprovidesacompensationoftheresidualexcessoftheblueroll-offoftheVSF.Itworkslikeacomplementarytransformation.
Benn, JAAVSO Volume 40, 2012852
Algorithms + Observations = VStar
David Benn73 Second Avenue, Klemzig, South Australia; [email protected]
Received May 15, 2012; accepted June 25, 2012
Abstract vstar is a multi-platform, free, open source application forvisualizing and analyzing time-series data. It is primarily intended for usewithvariablestarobservations,permittinglightcurvesandphaseplots tobecreated,viewedintabularform,andfiltered.Periodsearchandmodelcreationaresupported.Wavelet-basedtime-frequencyanalysispermitschangeinperiodovertimetobeinvestigated.DatacanbeloadedfromtheAAVSOInternationalDatabaseorfilesofvariousformats.vstar’sfeaturesetcanbeexpandedviaplug-ins,forexample,toreadKeplermissiondata.Thisarticleexploresvstar’sbeginningsfromaconversationwithArneHendenin2008toitsdevelopmentsince2009inthecontextoftheAAVSO’sCitizenSkyProject.Scienceexamplesareprovidedandanticipatedfuturedirectionsareoutlined.
1. Introduction
AconversationwithAAVSODirectorArneHendonat the23rdNationalAustralian Convention ofAmateurAstronomers (NACAA ) in 2008 plantedtheseedforasuccessortoGrantFoster’sdos-basedvstarprogram(Figure1),initiallycreatedforusewiththeAAVSO’sHands-on Astrophysicseducationalmaterial, later renamed Variable Star Astronomy. The motivation for theauthorwassimple: theopportunity todevelopaneasy touse,powerfuldatavisualizationandanalysistoolthatamateurandprofessionalastronomerswouldwanttouse. CorrespondenceoverthenextyearculminatedinAAVSOstaff(includingAaronPrice,ArneHenden,andSaraBeck)comingupwithaninitialspecificationforanewJava-basedmulti-platform(windows, mac os x, linux, opensolaris)implementationofvstar.RegularcommunicationwithSaraBeckcommencedinMay2009.Sincethen,vstardevelopmenthasconsumedmostoftheauthor’ssparetime,familyandothercommitmentspermitting,andbroughtagroupofpassionatepeopletogether.
2. Feature overview
Fundamentally, vstar’s purpose is to permit time series data (ostensiblyvariablestarobservations)fromavarietyofsourcestobeloadedandanalyzed.The initial specification called for data to be loaded from the AAVSOInternational Database (AID), files conforming to the AAVSO Download
Benn, JAAVSO Volume 40, 2012 853
File format, and a simple subset of the latter for personal observations (JD,magnitude,andoptionally:error,observercode,andvalidationflag). Apartfromlightcurveandphaseplotcreation,theloadeddatasetcanbeviewed and searched in tabular form.Selectionof anobservationon a lightcurveorphaseplotissynchronizedwithobservationtablerowselection,andthereverseisalsotrue.ListsoffavoriteAIDobjectscanbecreated.Discrepantobservationscanbeexcludedand/orreportedtoAAVSOHeadquarters. Observationbandscanbedisplayedorhiddenthroughaplotcontroldialog,whichalsoaffectswhatisseenbydefaultintheobservationtable.Simplefilterscanbedefinedtoyieldanewseriesforanalysisandtheobservationtablecanbe searchedusing regular expressions.SeeFigures2 and3 for examplesofdifferent views. Plots can be zoomed and panned.The usual print and savefunctionsareprovided. Binnedmeanscanbecreatedfortherawlightcurveorphaseddata.Errorbarsformeansandobservationscanoptionallybedisplayed.Aninformationwindow shows a breakdown of the loaded dataset by band and a “signalsignificance”statistic in theformofANOVAfor thebandthat is thecurrentsourceforbinnedmeans. Theotherbroadcategoryoffunctionalityisanalysis.Thefirstsub-categoryisperiodanalysis.Invstar,thisisanimplementationoftheDateCompensatedDiscreteFourierTransform(DCDFT)algorithm(Ferraz-Mello1981),yieldingapowerspectrumandtableof“top-hits”foraspecifiedseries, frequencyorperiodrange,andresolution. From within the DCDFT result window, a phase plot can be created. Inaddition,oneormoreperiodseachwithoneormoreharmonicscanbeselectedto create a model.A model’s Fourier coefficients can be viewed along withrelativeamplitudesandphases.Multipleperiodscanoptionallyberefinedviathe CLEANest (Foster 1995) algorithm.When a model is created, it is alsosubtracted from the series on which the DCDFT was performed to yield asecond,“residuals,”series.DCDFTcan thenbeapplied to theseresiduals tolookforfurthersignals(periods),aprocessoftencalled“pre-whitening.”SeeFigure8forasampleDCDFTpowerspectrumandphaseplotresultingfromatop-hitselection. Thesecondmajoranalysiscapabilityistime-frequencyanalysisintheformoftheWeightedWaveletZ-Transform(WWZ;Foster1996).Theuserspecifiesaseries,periodorfrequencyrange,andresolution,andananalysisofchangeinperiodovertimeistheresult.Thiscanbeviewedasa2Dgraph,acontourplot,arotatable3Dgraph,orintabularform.Periodscanbeselectedforphaseplotcreation.Figure4showsperiodchangeforTUMibetween1913and2009.HerethecolorrepresentstheWWZstatistic,thestrengthofaperiodicity,ataparticulartime.ThisexampleisdiscussedinFoster(2010). Anotherkindofmodelthatcanbecreatedisapolynomialfit,alongwiththecorresponding residuals series (forexample, for finding theminimumor
Benn, JAAVSO Volume 40, 2012854
maximummagnitudeinacycleofaMiradataset).Figure5showsapolynomialfit of degree7 foroCet in the JD range2451460 to2451560alongwith a5-daybinnedmeansseries.Thecross-hairsareonthehighestpointofthecurve,atamagnitudeofaround3.315andaJDofaround2451492.8.ThisexampleisbasedupononefoundinFoster(2010). vstar’sin-builtfeaturesetcanbeenhancedbycreatingplug-ins(injava).Thiswill be elaboratedon in another section.Thebeginningsof a scriptingcapabilityexist,permittingcertainoperationstobeautomated.
3. Early development
The original specification led to an initial round of questions. A keydecisionearlyonwas tomovethecontentfromaworddocument toaWiki(initially hosted byAAVSO).This facilitated a dialogue between the authorandtheAAVSOthroughSaraBeck,resultingintheWikibeingannotatedwithquestions and answers. By the end of May 2009, a set of requirements wascreatedthatdeterminedwhatwouldbedevelopedduringphaseone. ItwasdecidedthattheprojectwouldbehostedonSourceForge.AAVSOstaffmemberRichardKinnehelpedestablishthisandarguedthatvstarshouldbelicensedundertheAfferoGNUPublicLicense,aweb-deployment-friendlyversionofthenormalGNUPublicLicense.FeedbackonearlyuserinterfaceprototypeswassoughtfromAAVSOstaffmembers. Itwasrewardingtoreachthepointatwhichvstarcouldbeusedtoloadadataset(fordCep)andcreateaphaseplot.Althoughasimplefeature,itwasatthispointthattheauthorbegantoglimpsehowpowerfulatoollikethiscouldbe.
4. Citizen Sky team
LeadinguptothefirstCitizenSkyworkshopin2009,thevstarSoftwareDevelopment team became the first Citizen Sky project, facilitated byRebeccaTurner. In July 2009, Michael Umbricht, an astronomer at Brown University’sLadd Observatory, contacted the author through Citizen Sky to say that hewanted tohelpon theproject as a tester.After about threemonthsof initialdevelopment,Michael’sdomainknowledgeandearlyfeedbackontheproto-vstarwasvaluable.Leadinguptothefirstworkshopandforquiteafewmonthsthereafter,Michaelplayedakeyroleintesting,promotingvstar,andtriallingitinaclassroomsetting. Soonafterthefirstworkshop(aroundSeptember2009)AdamWeberjoined.He and the author had worthwhile design and implementation discussions.Adamcontributedanumberofbugfixesandhelpedtoimprovethelookandfeelofvstar,especiallyundermac os x. NicoCamargo,ayoungartistlivinginChicago,attendedthefirstCitizen
Benn, JAAVSO Volume 40, 2012 855
Skyworkshopandwecorrespondedafterward.Hehasplayedanimportantroleinimprovingvstar’sappearancebycreatingtoolbarbuttonanddesktopicons.Hiswillingnesstohelp,oftenatshortnotice,hasbeenmuchappreciated. Overthelifetimeoftheprojectnumerouspeoplehaveexpressedinterestin contributing. Following through was not always possible due to othercommitmentsorbecausethetimeandeffortrequiredtolearnenoughabouttheexistingcodebasetomakeacontributionwasprohibitive.Manywhodidnotdirectlycontribute todevelopmentor testoften stillprovidedsuggestionsorencouragement.TheCitizenSkyvstardevelopmentteamforumcapturestheteam’sinteractions.Testinganddocumentationarenotespeciallypopularanditwasnotgenerallyeasytointerestpeopleintheseactivities.Astheteamsizegrew,communicationoverheadsrose,withlesstimeavailabletotheauthorforsoftwaredevelopment,comparedwiththefreneticpaceofthefirstfewmonths.Overheadsreducedastheteamstabilized. Themostimportantaspectoftheteamwastheconfluenceofdiverseskills,knowledge,andexperience.Asleaddeveloper,theauthorcoulddefertootherswithgreaterdomainexpertiseorartisticskill.CommunicationmediasuchasWiki, instantmessaging,email,andoccasionalcalls largelycompensatedforthedistance across thePacific separating the author frommost of the team.Questions left with Sara and other team members would often be answeredduringtheauthor’snight.AnAASposterpapercoverstheteamaspectinmoredetail(Hendenet al.2010).
5.2.NACAA2010workshop TheNationalAustralianConventionofAmateurAstronomers(NACAA)isheldeverytwoyearsandhasalreadybeenmentionedinrelationtoitsroleingettingvstarstartedin2008.Twoyearslatertheauthorranafull-dayworkshopatthe24thNACAA.Version2.0wasreleasedandannouncedontheAAVSOwebsiteinconjunctionwiththatevent. Feedback from Australian and New Zealand amateurs and members ofVariableStarsSouth (VSS) led to severalnewSourceForge tracker items. Italsoreinforcedtotheauthorcertainuserinterfacechangesthatwouldimprovetheenduserexperience,primarilybyincreasingtheamountofrealestateforplotsandtablesinthemainwindowandmovingsecondaryfunctionstodialogs(seeFigures6and7).
Benn, JAAVSO Volume 40, 2012856
5.3.CitizenSky2 At the secondCitizenSkyworkshop,heldat theCaliforniaAcademyofSciences, San Francisco, in September 2010, the author gave a presentationshowing how vstar could be used to carry out period analysis, minimaidentification with a polynomial fit, and an overview of other capabilitiesaddedsince2009.Again,theresultingfeedbackwasimportantforimprovingthetool. Leadinguptothisworkshop,AaronPricecreatedthe“5StarTutorial”fordataanalysis,acompaniontotheCitizenSky“10StarTutorial,”forobservationofvariablestars.The5StarTutorialshowedhowtousevstartocreateabinnedmeansseries,carryoutperiodsearchwithDCDFT,andcreatephaseplots. Heinz Bernd-Eggenstein was present at the workshop and showed theauthorabuginwhichvstarmisbehavedwithnumericinputinthepresenceofaGermanlocalebeingsetonthehostoperatingsystem.Fixingthisseeminglysimple problem took some weeks after the workshop. Additional localeimprovements are on the roadmap, such as providing language specific (forexample,German,Spanish,andPortuguese)versionsofvstar.
5.4.NACAA2012update Four years after the initial conversation with Arne Henden, a talk anddemonstrationofprogresssincethe2010fulldayworkshopwasgivenatthe2ndVariableStarsSouthColloquiumheldinconjunctionwiththe25thNACAAinBrisbane. TimespentduringbreakswithVSSmembers,inparticularMarkBlackford,DavidMoriarty,AlanPlummer,andTomRichards,wasbeneficialaswetalkedabouttheiruseofvstar.Forexample,theneedforanASASdatasourceplug-inwasexpressed.
6. From fortran to java
Evenbeforedevelopmentofvstarhadbegun,MatthewTempletonpointedtheauthortoapaperabouttheDCDFTalgorithm(Ferraz-Mello1981).GrantFosterhadpreviouslywrittenanimplementationofDCDFTinbasicwhichwassubsequentlyconvertedtoafortranversionbyMatthewaspartofAAVSO’sTSconsole-basedprogram. When the time came to implement DCDFT for vstar, there were a fewchoices:
• Convert the fortran ts code to java. The author experimented withfortrantojavatranslators,butatthetime,nonewasfoundtobewithoutimportantbugs.
Benn, JAAVSO Volume 40, 2012 857
• Convert the fortran TS code to c and call from vstar as “nativemethods.”
• Convertthefortran tscodetocandthentojava.
fortran toc translatorsexist, inparticularf2c.Thiswasmadeuseof inordertobeabletosource-leveldebugthetscodeincformwhenpossiblebugsinthefortrancodewerefound.Thecodegeneratedbysuchtranslatorstendstobecrypticandhavedependenciesuponspeciallibraries.Callingccodethroughjava’s native interface mechanism would have required natively compiledlibrariesforeachoperatingsystem,somewhatantitheticaltovstar’sgenerallyplatform-independentimplementation. Otherthanthis,therearesufficientincompatibilitiesbetweentheconsoleandtextmenudrivennatureanddatastructuresofthetsprogramcomparedwithvstar’sinternalsanduserinterface,thatintheend,amixtureofthefirstandsecondoptionswasemployed. First,theappropriateliteraturewasreadtogetanunderstandingofDCDFT(Ferraz-Mello 1981). Next, a perl script was written to perform a partialtranslationof thets fortrancode into java.Next,each lineofemitted javacodewas inspected for logicalequivalencewith the fortran code.Thecorealgorithmswere translatedandextracted in thiswayandexposed to the restof vstar through the appropriate menu items and dialog boxes. For testingpurposes, ts was treated as a reference implementation for DCDFT. Unittestswerewrittenfor the java implementationof thealgorithmandchecked(automatically,aftersomeinitialmanualchecking)againsttheresultsgeneratedbyts for thesame input.Thesamestrategywasused to implementand test wwzinvstar. There is a difference in emphasis between implementing and testingalgorithmsfromapublicationsuchasMeeus(1991)—whichwasusedforJDcalculationsinvstar—anddoingsofromtheliterature.ThepurposeofabooksuchasMeeusistounambiguouslydescribeanalgorithmandprovideatleastminimaltestcases.Apaperthatdescribesanalgorithmdoesnothavequitethesame obligation. Hence the approach of using “battled-tested” fortran as areferenceimplementationandsemi-automatedtranslationasopposedtowritingfromscratchfromtheliterature,whilemakinguseofittobolsterunderstandingofthealgorithm.Intheend,thekeybenefitisthatthepowerfulfunctionalityofthefortran tsand wwzcodehasbeenmadeavailableinvstar,andalongtheway,somebugswereuncoveredintheoriginalfortrancode.
Perhapsnottoosurprisingisthatobservationsourceplug-inshavesofarbeenthemostcommonlywrittenorrequested,sincethismakesitpossibletoloaddatasetsfrommorediversesourcesthanthosedefinedbytheAAVSO. Todate,observationsourceplug-inshavebeenwrittenforKeplermissionpublicdatareleaseFITSfilesandSuperWASPsurveyFITSfiles.BothofthesewerewrittenincollaborationwithDougWelch.OtheruserssuchasKenMogulandAlanPlummerpromptedthedevelopmentofanAAVSOsimpleandextendeduploadfileformatobservationsourceplug-in.Thisformatcanbehand-craftedor generated by an application such as vphot for use in uploading multipleobservationstotheAAVSOInternationalDatabase.Someplug-ins,ifcommonlyusedlikethisone,mayeventuallybeaddedintothecoreofvstar’scodebase. Anexampleofthesecondcategorywasaplug-inwrittenbySaraBecktoshowwhichobservationsina loadeddatasetweremadebyobservers in,forexample,Ireland. Asfarastheauthorisaware,noperiodanalysisplug-inshavebeenwrittenyet,buttwocandidatesareAoV(Schwarzenberg-Czerny1989)andFastChi-squared(Palmer2009).Internally,DCDFTandWWZaretreatedasplug-ins,implementingtherequiredinterfaces,theonlydifferencebeingthattheyarenotdynamicallyloadedwhenvstarstarts. Anexampleof the“arbitraryoperation”plug-in typewascreatedby theauthor when AAVSO member Mike Simonsen expressed a need to selectdatapoints on a light curve plot to determine mean time between selectedobservations (along with mean magnitude).An example of the last plug-incategoryisaJDtocalendardatecalculator.Thevstarplug-inlibrary(http://www.aavso.org/vstar-plugin-library) and the vstar webpage (http://www.aavso.org/v-star-overview)canbeconsultedformoreinformation.
8. Deployment
java webstart™isthedeploymentmechanismusedwhenlaunchingvstarfromtheAAVSOwebpageandremainstheeasiestwaytobeginusingvstar.webstart™imposes“securityconstraints”similar to that imposedbya javaapplet.Ittooksometimetomakevstarworkproperlywithintheseconstraints,especiallyinthepresenceofdynamicallyloadedplug-ins.
Benn, JAAVSO Volume 40, 2012 859
For each formal release, there were one or more testing releases madeavailable to the Citizen Sky team through webstart™. Each formal releasehasbeenaccompaniedbyadownloadablearchive(fromSourceForge)thatcanbeunzippedandrunasanormallocalapplication.Morerecently,operating-systemspecificlauncherprograms/scriptswereaddedtoimprovethelocalrun-timeexperience,inparticularbymakingstart-upmemoryallocationequivalenttothewebstart™deploymentandbyaddingadesktopiconfordos/windowsandmac os x.
9.1.Cepheididentification Inanexampleofprofessional-amateurcollaboration,amateurastronomerKen Mogul has partnered with Doug Welch, a professional astronomer atMcMasterUniversity,onalong-termprojectusingtherobotictelescopenetworkAAVSOnettoobtainhigh-qualityphotometrictimeseriesofType2CepheidsinSloangrizfilters.Ambiguouslyclassifiedvariables(CEP,CEP:)werealsoincludedtoensurethattheirclassificationscouldbeimproved(Welch2012).vphot is used to process images and vstar is used to help eliminate targetsthatdon’tmeettheprojectcriteria.InKen’swords:“Ihavealsobeenabletodataminethedatafornewvariablesoutsidethegoaloftheproject.vstarhasenabledmetomakeintelligentsuggestionstotheprofessionalonwhattolookatandwhattodismisstherebysavingtheprofessionaltimetofocusonthebigpicture”(Mogul2012). By using vphot and vstar together, Ken was easily and quickly able tofirstreducethedataforeveryotherstar,apartfromtheoriginaltarget,inthefield-of-view being studied, to look for possible variability. Then vstar’sDCDFT and phase plot features (see Figure 8) enabled him to immediatelyphasethedataintoaclassification-revealinglightcurve,suchasintheexampleinFigure8ofstar2MASSJ03145502+5618172(Mogul2012).KenwasalsoabletoreclassifyGVAurasaClassicalCepheidasshowninFigure9. Again, inKen’swords:“Eventuallywithenoughdata,vstarwillenablemetobeatthecenteroftheactionconceptually,withoutmyhavingtospendagreatefforttolearnandunderstandthingslikeIRAF.Thesetoolsareashiningexampleofhowtomakecitizenscienceaviableforcegoingforward.AAVSOnet,vstarandvphothaveenabledme,withnoequipment…togofromobservingtoorganizingtoanalyzinganddrawingusefulconclusions…withonlyacomputerandInternetconnection…apossibilitywhichwasalmostunimaginableevenadecadeago,excepttotheveryfarsighted”(Mogul2012).
Benn, JAAVSO Volume 40, 2012860
9.2.Keplerdatamining Using vstar’s Kepler data source plug-in and DCDFT and phase plots,KevinAltonhasexploredeclipsingvariablestarsystemsobservedbytheKeplerspacecraft,lookingforchangesinminima/maximaacrosscycles.AccordingtoKevin,thisisapreludetopotentiallymappingmagneticactivitycyclesand/orstarspotmigrationinselectedcontactbinaries(Alton2012).
9.3.LongtermeAurigaelightcurve BrianKloppenborghasusedvstar’sWWZfeaturetolookforchangingormultipleperiods ineAur’s long-termphotometricarchive lightcurve.Briancommented that nothing has been found so far, as others before him havediscovered (Kloppenborg 2012). Brian’s use ofWWZ exposed a bug in thefortran and java implementation (in thepresenceof significantgaps in thedata).Thisisnowontheinevitablelistofthingstofix.
9.4.Lightcurvesforillustrationinarticles Articles containing light curves saved from vstar have appeared inAustralian Sky & TelescopebyVSSmemberAlanPlummerandbyothersinVSSnewslettersandelsewhere.
9.6.Periodsearchandphaseplots Aspartoflearningaboutvariablestarphotometryandanalysisinhisworkon the SPADES project (Richards 2012), VSS member David Moriarty hascarried out period search with vstar on photometric data obtained for thecontacteclipsingbinarysystemTWCruandcomparedittoperiodandepochvaluesfoundinGCVS,Dvorak,andKrakowrepositories.
• Memory footprint reduction.Forexample,eachobservationdata-pointis represented by an object that consumes more memory than it should.Addressingthiswillpermitlargerdatasetstobeloaded.
• Betterdocumentation, in the formof ausermanual asopposed to thecurrentminimalHelpmenuitemandoccasionalarticles.
• More analysis features, for example: time of minimum/maximum, forexample,foreclipsingbinaryepochdetermination;alternativestoDCDFT,suchasAoVandFastChi-Squared.Thesecanalsobedevelopedbyothersasplug-ins;O–Canalysis;Lowessfit,forexample,forminimadetermination(Foster2010).
• AAVSO download files can be generated as lines of comma-, tab-, orspace-separatedvalues.Unlessvaluesarequoted,ambiguityispossible(forexample,commasincommentsfields).Thishasbeenflaggedasanissueandwillbepursuedsincesomefilescontainlinesthatcannotbeunambiguouslyparsedbyvstar(oranytool),somustbeomittedatloadtimeandreportedaserroneoustotheuser.
• Increasing the power of models, for example: addition of arbitraryterms,suchasforobserverbias(aspermittedbyTS);makingiteasiertoaccumulateperiodsthroughouttheprocessofpre-whiteningforsubsequentmodelcreation;mergingtwoormoreexistingmodels;modelcreationfromWWZ.
• Althoughplug-inscanbecreatedtoperformarbitraryfilteringoperations,the current in-built filtering capability is limited to the equivalent of anexpression language over a subset of an observation’s fields, supportingconjunction (logicalAND), relational operations (>, <, =, <=, >=), andnegation. This could be made more powerful by permitting expressionscontainingdisjunctions(logicalOR)andallowingallobservationfieldstobeused.
• The scripting API should grow to permit more operations to beautomated.
TheauthorwouldliketothankArneHendenandAaronPriceforgettingitallstartedandforongoingencouragement;andSaraBeckformentoring,beingan AAVSO liason, and providing regular support and enthusiasm. RebeccaTurnerhasbeensupportiveasCitizenSkyprojectmanager.DocKinneprovidedsystemadministrationsupport,helpedestablish theSourceForgeprojectsite,andsortoutlicenseissues.MatthewTempletonpatientlyansweredquestionsaboutalgorithms.Allprovidedencouragement. MichaelUmbricht,AdamWeber,NicoCamargo,HeinzBernd-Eggenstein,Jaime Garcia, Doug Welch, Ken Mogul, and Mike Simonsen—vstar teammembersandusers—providedfeedback,assistance,testing,andencouragement.ThanksalsotoGrantFosterforwritinghislightcurveanalysisbook(Foster2010), introducingme to theR statistical analysis language, and for helpfultechnicaldiscussions.Thesepeoplehavegivengenerouslyof their timeandtalentsinresponsetomyquestions. Members of VSS have also provided encouragement, advice for futuredirections,andbetterstill,havemadeuseofitintheirresearch(forexample,MarkBlackfordandDavidMoriarty). Alittlemorethantwoyearsafterthevstarprojectstarted,IreceivedemailfromArneHendentosaythatIwastherecipientofthe2011AAVSODirector’sAward.Considering theprevious recipientsof thisaward, Iamhonoredandhumbledtohavereceivedthis.Membersofthevstarteam,andothers,helpedtomakethispossible. MysupportivewifeKarenandchildrenNicholasandHeatherhavebecomevery tolerantof theirhusband’s/father’snocturnalnature,especiallyover thelastfewyears.Nicholasalsohelpedwiththedos/windowslauncher. The author greatly appreciates the opportunity to develop vstar, toparticipate in the Citizen Sky andAAVSO communities, and to attend bothCitizenSkyworkshops. Thetitleforthispaperwasinspiredbya1976bookbycomputerscientistNiklausWirth—Algorithms + Data Structures = Programs.
Nico Camargo4233 N. Hermitage Avenue, #3A, Chicago, IL 60613; [email protected]
Received May 14, 2012; revised October 22, 2012
Abstract Idecidedtowriteaboutartinsciencebytellingpersonalanecdotessurroundingmyinvolvementinastronomy.Theseportraywhatdroveme(withacareerintheFineArts) totryscientificillustrationthroughinvolvementintheAmericanAssociationofVariableStarObservers’CitizenSkyproject toobservetheeclipseofeAurigae.TheseaccountsdefinewhatIbelievetobeatthecoreofbeingascientificillustrator:theimportanceofmaintainingaccuracyand factual detail without compromising the compelling visuals that evokecuriosity.This,asmostpeoplerealize,isanimportantfactorinbeingabletosuccessfullyengagethepublicinscience—particularlyinastronomy.However,myintentionintellingtheseanecdotesgoesdeeperthanstatingtheimportanceof disseminating scientific knowledge through imagery—for which ampleexamplesandliteraturealreadyexist.Instead,I’mreallyaftercontrastingtheroleofascienceillustratorversusthatofanartist.Indoingso,IwillunderscorewhatIbelieveillustratingphenomenainsciencereallyisallabout.ThisnoteisnotintendedasascholarlytreatisebutratherpersonalreflectionsrelativetomyinvolvementinCitizenSky.
funweekendoflecturesandconversationswithpassionate,brightindividualstrying to set up anunprecedented campaignof observationsduring thenext640–730daysofeclipse. My enthusiasm for astronomy was entrenched in curiosity about howtheuniverseworks.VeryearlyrecollectionsofCarlSagan’sCosmosandtheSTS Challenger disaster on television are embedded in my memory to thisdayandhadimpactedthewayIviewscience.Particularlyspaceexplorationintrigued me—from the standpoint of human achievement and engineeringadvancement—becauseitwasacommontopicatthedinnertablegrowingup.Thatfamousimageofanastronautwearingajet-pack,floatingaloneinspaceaboveEarth,wasinmygrandfather’shomeoffice,andalate1980stelevisionmini-seriesabouttheApollomissionscementedmyutmostadmiration,wonder,andthirstforlearningaboutspaceandtheuniverse.Whatallofthesememorieshave in common is awe-inspiring imagery.Whether on television or in stillpictures, these images, aswell asmanyothers (certainly some fromsciencefictionmoviesaswell)wouldstirmycuriositytolearnaboutwhatisbeyondourhome. BackinChicagoattheAdlerPlanetarium,Iwassoakinginalltheinformationavailable about eAur: light curves from past observation campaigns, crudeillustrations,dataanalyses;that,andmuchmoreinformation,providedseveralpossibilitiestoillustratewhatwaspresumedtobeaneclipsingbinarysystemthatwasjusttoofartobecapturedinphotographicformat—notevenwithourmost powerful observatories.What eAur was remained a mystery awaitingresolution,andtome,thismeantanopportunitytoinspireotherswithmyownspace images—ones that would tell what would otherwise be a difficult-to-comprehendcelestialconundrum.Imetthepodcast’svoice,RebeccaTurner,andtherestofCitizenSkycrewthatweekendandfromthatpointon,Iwascommittedtoprovidingthemwithillustrationsthatwouldhelptosupporttheirmission.
2. Nature has better imagination
In 2010 Citizen Sky met again, that time in San Francisco, to discussprogress on eAur’s eclipse and analyses of new data.After lectures, onceagain,IengagedinconversationswithheadastronomersandexpertsoneAurtofindoutwhatthenewleadinghypothesesworthillustratingwere.Iwantedtodepictexactlywhat theyenvisioned,ormoreappropriately,what thenewinformationwastellingthem.About2,000light-yearsaway,theonlywaytobringthiseclipsingbinarysystemtolifeisbyimaginingandillustratingit,andeverynewpieceofthepuzzlemeantthatwecouldhaveamoreconcreteideaof the system’s objects. Evidence of the system’s objects’ approximate size,positioning,molecularcomposition,andshapealreadycreatedanexactmodelofhoweAurcouldlook,however,therewerenewdetailsthatdismissedearlierhypotheses.
Camargo, JAAVSO Volume 40, 2012 869
When I began illustrating e Aur, I looked at earlier illustrations of thesystem,aswellasonesofothersimilarsystems.Onehadablackholewithspewing jets of energy coming from the middle of this companion disk—acompelling image that was negated by data confirming that there were noblackholesineAur.(That’snottosaythatimpressivejetsofenergyspewingfromblackholesdon’texist,orhaven’tbeenseen—theyhave.)AnothereAurillustrationhaditsvisiblestar(whatcametoberecognizedasalargeF-typestar)pullingmassfromtheobscurecompanionthroughtheinnerLagrangianpointandcreatingadiskofmaterialifitsown—thatideawasalsorejectedaftermoresoliddatacametolight.Again,traffickingmaterialbetweentwocelestialobjectsthattrespasstheirRocheLobedoesexist.Mypointisthatatthetime,itwasanybody’sguesswhatwasgoingonineAur,butasmoreknowledgewasacquired,theseguesseswereappropriatelymarginalized.Iwassupposedtoillustratethenew“guesses,”anditwasmydutytomakeanattractiveimagewhilekeepinginmindthatanyextrapolationofmyownwouldwronglymaketheillustrationbecomesomethingotherthaneAurigae—justliketheonespriortomine,justlikemyfirstones. Today, my favorite eclipsing binary system remains much of a mysterydespitehard-earnednewdata.Itsdark,humongousdisc-likecomponentcouldbe a planetary system in its infancy, or one in limbo that will never formplanets due to gravitational pulls from its inside (where a small B star wasfound)andtheF-starcompanion.eAurcouldendupbeingseveralsurprisingthings(butnot toomany).However, theseplausiblesurprises, I’msure,willbefascinating.Naturehasawayofrevealingherself,inthemost“out-there”ofways—sheoutdoesevensciencefictionflicks.Fromplanetswithfourstars,to planetary formation itself, we should just illustrate natural phenomena ashonestlyaspossible,sincethoseimagesarecertainlyas(ifnotmore)creativeandimaginativethananyartistcouldimagine.
3. Artwork created in support of Citizen Sky and the AAVSO
• illustration commemorating the receipt of the AAVSO’s 20 millionthvariablestarobservation• October page in the American Astronomical Society’s 2011 calendar,commemoratingtheAAVSO’scentennial•potentialdesignsfortheAAVSOlogo
4. Not apples and oranges, more like apples and onions
Asanemergingvisualartist,Iunderwentanintentionalshiftfromnebulousabstractconceptualismtotheconstrictionsofimperativepredestinarianisminpicture-making.Thatshifteffectivelyconvertedmeintoanillustratorforthetimebeing.Thisdualitymustbeaddressedbecause,whilethejuxtapositionofthesetwodifferentfieldshasprovidedanabundanceofintellectandcreativityinmyart,onemustunderstandtheextentsandlimitationsthatthesetwofieldsimpose on each other. To say that a scientific illustrator took some artisticliberties when rendering a natural phenomenon implies misdirection—suchas fabrication of facts, exaggeration, and image entitlement, to name a few.Conversely,whenanartistrendersideasasvisualexplanationsofthemselves,withoutcontextambiguity,orabstractionofanykind, thisartistbecomesanillustrator. There are certain things in the universe that can’t be directly translatedinto visual form, so we use symbols, approximations, and the imagination.But,scientificillustratorsfillinthisgap,andtheymustderivetheirworkfromresearchandscientifichypotheses.Ascientificillustratorhastheobligationtoexplainascientificsubjectconcisely,thoroughly,andaccurately,whetherthemediumisanimationorillustration,andindoingso,heorshemustexcitetheimaginationwithoutmisleadingtheviewer. Beauty and aesthetics are very important aspects of scientific imagery.Undoubtedly, science illustrators must make their subject matters visuallyattractive and unforgettable. That being said, scientific imagery should alsobereserved,humble,andhaveananonymousquality.Itshouldbeeconomicand focused solely on its depictions, without compromising on its scientificaccuracy. Scientific illustrators are not making art; they are not making apersonal statement; they are not trying to convey an idea indirectly or perambiguitiesandinsinuations,noraretheyspeakingaboutthemselves,abstractconcepts (like loveor life),or society. Instead, theyareprovidingaprecise,hyper-realistic representation(asbestas theycan)of thephysicalworldandits natural occurrences. Scientific illustrators have formal and conceptualconstrictionsandareobligedtocapturethesubjectmatterexplicitly.Theartistryinscientificimageryliesinhoweffectiveandwell-executedthesubjectmatteris,inordertomakelearningfaster,easier,andmoreenjoyable. MyhopeisforothercreativeindividualstofollowthepointsIlayoutinthisnote,andtorememberthattheessenceofscientificillustrationliesinthecomprehensionofknowledge.
JAAVSO Volume 40, 2012 871
JAAVSO Volume 40, Number 2—other papers received
Richmond and Smith, JAAVSO Volume 40, 2012872
BVRI Photometry of SN 2011fe in M101
Michael W. RichmondPhysics Department, Rochester Institute of Technology, 84 Lomb Memorial Drive, Rochester, NY, 14623; [email protected]
Horace A. SmithDepartment of Physics and Astronomy, Michigan State University, East Lansing, MI 48824; [email protected]
Received March 19, 2012; revised April 4, 2012; accepted April 11, 2012
Abstract We present BVRI photometry of Supernova 2011fe in M101,starting2.9daysaftertheexplosionandending179dayslater.Thelightcurvesandcolorevolution show thatSN2011febelongs to the“normal’’ subsetoftypeIasupernovae,withadeclineparameterDm15(B)=1.21±0.03mag.Aftercorrectingforthesmallamountofextinctioninthelineofsight,andadoptinga distance modulus of (m – M ) = 29.10 mag to M101, we derive absolutemagnitudes MB = –19.21, MV = –19.19, MR = –19.18, and MI = –18.94.WecomparethevoluminousrecordofvisualmeasurementsofthiseventtoourCCDphotometryandfindevidenceforasystematicdifferencewhichdependsoncolor.
1. Introduction
Supernova(SN)2011feinthegalaxyM101(NGC5457)wasdiscoveredbythePalomarTransientFactory(Lawet al.2009;Rauet al.2009)inimagestakenonUT2012Aug24andannouncedlaterthatday(Nugentet al.2011a).AstheclosestandbrightesttypeIaSNsinceSN1972E(Kirshneret al.1973),and moreover as one which appears to suffer relatively little interstellarextinction,thiseventshouldprovideawealthofinformationonthenatureofthermonuclearsupernovae. WepresentherephotometryofSN2011feintheBVRIpassbandsobtainedattwosites,startingonedayafterthediscoveryandcontinuingforaspanof179days.Section2describesourobservationalprocedures,ourreductionoftherawimages,andthemethodsweusedtoextractinstrumentalmagnitudes.Insection3,weexplainhowtheinstrumentalquantitiesweretransformedtothestandardJohnson-Cousinsmagnitudescale.Weillustrate the lightcurvesandcolorcurvesofSN2011feinsection4,commentbrieflyontheirproperties,anddiscussextinctionalongthelineofsight.Insection5,weexaminetherichhistoryofdistancemeasurementstoM101inordertochoosearepresentativevaluewithwhichwethencomputeabsolutemagnitudes.UsingaverylargesetofvisualmeasurementsfromtheAAVSO,wecomparethevisualandCCDV-bandobservationsinsection6.Wepresentourconclusionsinsection7.
Richmond and Smith, JAAVSO Volume 40, 2012 873
2. Observations
This paper contains measurements made at the RIT Observatory, nearRochester, New York, and the Michigan State University (MSU) CampusObservatory, near East Lansing, Michigan. We will describe below theacquisitionandreductionoftheimagesintoinstrumentalmagnitudesfromeachsiteinturn. TheRITObservatoryislocatedonthecampusoftheRochesterInstituteofTechnology,at longitude77:39:53West, latitude+43:04:33North,andanelevationof168metersabovesealevel.ThecitylightsofRochestermakethenortheasternskyespeciallybright,whichattimesaffectedourmeasurementsofSN2011fe.WeusedaMeadeLX200f /1030-cmtelescopeandSBIGST-8E camera, which features a Kodak KAF1600 CCD chip and astronomicalfiltersmade to theBessellprescription;with3×3binning, theplate scale is1.85secondsperpixel.TomeasureSN2011fe,wetookaseriesof60-secondunguidedexposuresthrougheachfilter;thenumberofimagesperfilterrangedfrom10,atearlytimes,to15or20atlatetimes.Wetypicallydiscardedafewimages in each seriesdue to trailing.Weacquireddark and flatfield imageseachnight,switchingfromtwilightskyflatstodomeflatsinlateOctober.ThefilterwheeloftenfailedtoreturntoitsproperlocationintheR-band,so,whennecessary,weshiftedtheR-bandflatsbyasmallamountinonedimensioninordertomatchtheR-bandtargetimages.Wecombined10darkimageseachnighttocreateamasterdarkframe,and10flatfieldimagesineachfiltertocreateamasterflatfieldframe.Afterapplyingthemasterdarkandflatfieldimagesintheusualmanner,weexaminedeachcleanedtargetimagebyeye.WediscardedtrailedandblurryimagesandmeasuredtheFWHMofthoseremaining. TheXVista (TreffersandRichmond1989) routines starsandphotwereused to find stars and to extract their instrumental magnitudes, respectively,usingasyntheticaperturewithradiusslightlylargerthantheFWHM(whichwastypically4"to5").AsFigure1shows,SN2011feliesinaregionrelativelyfreeoflightfromM101(seealsoSupplementaryFigure1ofLiet al.2011).Asacheckthatsimpleaperturephotometrywouldyieldaccurateresults,weexaminedhigh-resolutionHSTimagesofthearea,usingACSWFCdataintheF814Wfilteroriginally takenaspartofproposalGO-9490 (PI:Kuntz).Thebrightesttwosourceswithina5"radiusofthepositionoftheSN,R.A.=14h03m05.733s, Dec, = +54˚ 16' 25.18" (J2000) (Li et al. 2011),haveapparentmagnitudesofmI~_21.8andmI~_22.2.Thus,evenwhentheSNisatitsfaintest,inourfinalI-bandmeasurements,itismorethanonehundredtimesbrighterthannearbystarswhichmightcontaminateourmeasurements. Between August and November 2011, we measured instrumentalmagnitudes from each exposure and applied inhomogeneous ensemblephotometry (Honeycutt 1992) to determine a mean value in each passband.StartinginDecember2011,theSNgrewsofaintintheI-bandthatwecombined
Richmond and Smith, JAAVSO Volume 40, 2012874
thegoodimagesforeachpassbandusingapixel-by-pixelmedianprocedure,yieldingasingleimagewithlowernoiselevels.Wethenextractedinstrumentalmagnitudesfromthisimageinthemannerdescribedabove.Inordertoverifythatthischangeinproceduredidnotcauseanysystematicshiftintheresults,we also measured magnitudes from the individual exposures, reduced themusingensemblephotometry,andcomparedtheresultstothosemeasuredfromthe median-combined images.As Figure 2 shows, there were no significantsystematicdifferences. The Michigan State University Campus Observatory lies on the MSUcampus,atlongitude05:37:56West,latitude+42:42:23North,andanelevationof 273 meters above sea level. The f /8 60-cm Boller and Chivens reflectorfocuses light on anApogeeAlta U47 camera and its e2V CCD47-10 back-illuminated CCD, yielding a plate scale of 0.56 arcsecond per pixel. FilterscloselyapproximatetheBessellprescription.Exposuretimesrangedbetween30and180seconds.Weacquireddark,bias,andtwilightskyflatfieldframesonmostnights.Onafewnights,highcloudspreventedthetakingoftwilightskyflatfieldexposures,soweusedflatfieldsfromtheprecedingorfollowingnights.TheI-bandimagesshowconsiderablefringingwhichcannotalwaysberemovedperfectly.Weextractedinstrumentalmagnitudesforallstarsusingasyntheticapertureofradius5.4seconds.
3. Photometric calibration
In order to transform our instrumental measurements into magnitudesin the standard Johnson-Cousins BVRI system, we used a set of localcomparisonstars.TheAAVSOkindlysuppliedmeasurementsforstarsinthefieldofM101(Henden2012)basedondatafromtheK35telescopeatSonoitaResearchObservatory(Simonsen2011).WelistthesemagnitudesinTable1;notethattheyareslightlydifferentfromthevaluesintheAAVSO’son-linesequenceswhichappeared in late2011.Figure1shows the locationof thethreecomparisonstars. TheAAVSOcalibrationdataincludedmanyotherstarsintheregionnearM101.Inordertocheckforsystematicerrors,wecomparedtheAAVSOdatatophotoelectricBVmeasurementsinSandageandTammann(1974).ForthefivestarslistedasA,B,C,D,andGinSandageandTammann(1974),whichrange 12.01 <V < 16.22, we find mean differences of –0.013 ± 0.038 maginB-band,and–0.009±0.022maginV-band.WeconcludethattheAAVSOcalibrationsetsuffersfromnosystematicerror inBorVat thelevelof twopercent. Unfortunately, we could not find any independent measurements tochecktheRandIpassbandsinasimilarmanner. InordertoconverttheRITmeasurementstotheJohnson-Cousinssystem,we analyzed images of the standard field PG1633+009 (Landolt 1992) todeterminethecoefficientsinthetransformationequations
Richmond and Smith, JAAVSO Volume 40, 2012 875
B=b+0.238(043)*(b–v)+ZB (1)
V=v–0.077(010)*(v – r)+ZV (2)
R=r–0.082(038)*(r – i)+ZR (3)
I=i+0.014(013)*(r – i)+ZI (4)
Intheequationsabove,lower-casesymbolsrepresentinstrumentalmagnitudes,upper-case symbols Johnson-Cousins magnitudes, terms in parentheses theuncertaintiesineachcoefficient,andZthezeropointineachband.WeusedstarsA,B,andGtodeterminethezeropointforeachimage(exceptinafewcasesforwhichGfelloutsidetheimage).Table2listsourcalibratedmeasurementsofSN2011femadeatRIT.ThefirstcolumnshowsthemeanJulianDateofalltheexposurestakenduringeachnight.Inmostcases,thespanbetweenthefirstandlastexposureswaslessthan0.04day,butonafewnights,cloudsinterruptedthesequenceofobservations.ContactthefirstauthorforadatasetprovidingtheJulianDateofeachmeasurementindividually. The uncertainties listed in Table 2 incorporate the uncertainties ininstrumentalmagnitudesandintheoffsettoshifttheinstrumentalvaluestothestandardscale,addedinquadrature.Asacheckontheirsize,wechosearegionofthelightcurve,875<JD–2455000<930,inwhichthemagnitudeappearedtobea linear functionof time.Wefitastraight line to themeasurements ineachpassband,weightingeachpointbasedon itsuncertainty; theresultsareshown inTable3.The reducedc2values,between0.9and1.6, indicate thatouruncertaintiesaccuratelyreflectthescatterfromonenighttothenext.Thedecline rate is smallest in the blue, but it is still, at roughly 130 days afterexplosion,significantlyfasterthanthe0.0098mag/dayproducedbythedecayof56Co. TheMSUdataweretransformedinasimilarway,usingonlystarsAandB.ThetransformationequationsforMSUwere
B=b+0.25(0.03)*(b – v)+ZB (5)
V=v–0.08(0.02)*(b – v)+ZV (6)
I=i+0.03(0.02)*(v – i)+ZI (7)
Intheequationsabove,lower-casesymbolsrepresentinstrumentalmagnitudes,upper-case symbols Johnson-Cousins magnitudes, terms in parentheses theuncertaintiesineachcoefficient,andZthezeropointineachband. Table4listsourcalibratedmeasurementsofSN2011femadeatMSU.DuetothelargerapertureoftheMSUtelescope,exposuretimeswereshortenough
WeadopttheexplosiondateofJD2455796.687±0.014deducedbyNugentet al.(2011b)inthefollowingdiscussion.Figure3showsourlightcurvesofSN2011fe,whichstart2.9daysaftertheexplosionand1.1daysafterNugentet al.(2011a)announcedtheirdiscovery. In order to determine the time and magnitude at peak brightness, wefit polynomials of order 2 and 3 to the light curves near maximum in eachpassband,weightingthefitsbytheuncertaintiesineachmeasurement.Welisttheresults inTable5, includingthevaluesfor thesecondarymaximuminI-band.Weagainuselow-orderpolynomialfitstomeasurethedeclineintheB-band15daysafterthepeak,findingD15(B)=1.21±0.03.Thisvalueissimilartothatofthe“normal’’SNeIa1980N(Hamuyet al.1991),1989B(Wellset al.1994),1994D(Richmondet al.(1995),and2003du(Stanishevet al.2007).ThelocationofthesecondarypeakinI-band,26.6±0.5daysafterand0.45±0.03magbelowtheprimarypeak,alsoliesclosetothevaluesforthoseotherSNe. AlthoughthereisasyetlittlepublishedanalysisofthespectraofSN2011fe,(Nugent et al. 2011b) state that the optical spectrum on UT 2011 Aug 25resembles that of the SN 1994D; on the other hand, (Marion 2011) reportsthat a near-infrared spectrum on UT 2011Aug 26 resembles that of SNe IawithfastdeclineratesandDm15(B)>1.3.Wemustwaitfordetailedanalysisofspectraofthiseventasitevolvestoandpastmaximumlightforasecurespectralclassification,butthisverypreliminaryinformationmaysupport thephotometricevidencethatSN2011fefallsintothenormalsubsetoftypeIaSNe. WeturnnowtotheevolutionofSN2011feincolor.Inordertocompareits colors easily to those of other supernovae, we must remove the effectsof extinction due to gas and dust within the Milky Way and within M101.Fortunately,thereappearstobelittleinterveningmaterial.Schlegelet al.(1998)useinfraredmapsofdustintheMilkyWaytoestimateE(B–V)=0.009inthedirectionofM101.Patatet al.(2011)acquiredhigh-resolutionspectroscopyofSN2011feandidentifiedanumberofnarrowNaID2absorptionfeatures;theyuseradialvelocitiestoassignsometotheMilkyWayandsometoM101.Theyconvertthetotalequivalentwidthofallcomponents,85mÅ,toareddeningofE(B–V)= 0.025±0.003usingtherelationshipgiveninMunariandZwitter(1997).Note,however,thatthistotalequivalentwidthisconsiderablysmallerthanthatofallbutasinglestarinthesampleusedbyMunariandZwitter(1997),sowehavedecidedtodoublethequoteduncertainty.AdoptingtheconversionsfromreddeningtoextinctiongiveninSchlegelet al.(1998),wecomputetheextinctiontowardSN2011fetobeAB=0.11±0.03,AV=0.08±0.02,AR=0.07±0.02,andAI=0.05±0.01.
Richmond and Smith, JAAVSO Volume 40, 2012 877
Afterremovingthisextinctionfromourmeasurements,weshowthecolorevolutionofSN2011feinFigures4through6.TheshapeandextremevaluesofthesecolorsaresimilartothoseofthenormalTypeIaSNe1994Dand2003du.InFigure4,wehavedrawna line torepresent therelationship(Lira(1995);Phillipset al.(1999))forasetoffourtypeIaSNewhichsufferedlittleornoextinction.The(B−V)locusofSN2011feliesslightly(0.05to0.10mag)totheredsideofthisline,especiallynearthetimeofmaximum(B−V)color.GivenourestimatesoftheextinctiontoSN2011fe,thissmalldifferenceisunlikelytobeduetoourunderestimationofthereddening.
5. Absolute magnitudes
InordertocomputethepeakabsolutemagnitudesofSN2011fe,wemustremove the effects of extinction and apply the appropriate correction for itsdistance.Theprevioussectiondiscussestheextinctiontothisevent,andwenowexaminethedistancetoM101.SincethefirstidentificationofCepheidsinthisgalaxy26yearsago(Cooket al.1986),astronomershaveacquiredeverdeeperand larger collectionsofmeasurements.Shappee andStanek (2011)providea listof recentefforts,whichsuggests thatCepheid-basedmeasurementsareconvergingonarelativedistancemodulus(m – M)=10.63magbetweentheLMCandM101.Ifweadoptadistancemodulusof(m – M)LMC=18.50magtotheLMC,thisimpliesadistancemodulus(m – M)M101=29.13toM101.Thisissimilartooneofthetworesultsbasedontheluminosityofthetipoftheredgiantbranch,(m – M)M101=29.05±0.06(rand)±0.12(sys)mag(ShappeeandStanek 2011), though considerably less than the other, (m – M)M101 = 29.42± 0.11 mag (Sakai et al. 2004). We therefore adopt a value of (m – M)M101= 29.10 ± 0.15 mag to convert our apparent to absolute magnitudes. Notethattheuncertaintyinthisdistancemodulusisourroughaverage,basedonacombinationoftherandomandsystematicerrorsquotedbyotherauthorsandthescatterbetweentheirvalues.ThisuncertaintyinthedistancetoM101willdominatetheuncertaintiesinallabsolutemagnitudescomputedbelow. Using this distance modulus, and the extinction derived earlier for eachband,wecanconverttheapparentmagnitudesatmaximumlightintoabsolutemagnitudes.WelistthesevaluesinTable6. Phillips (1993) foundaconnectionbetween theabsolutemagnitudeof atypeIaSNandtherateatwhichitdeclinesaftermaximum:quickly-decliningevents are intrinsically less luminous. Further investigation (Hamuy et al.1996; Riess et al. 1996; Perlmutter et al. 1997) confirmed this relationshipand spawnedseveraldifferentmethods toquantify it.Weadopt theDm15(B)method,whichcharacterizesaneventbythechangeinitsB-bandluminosityinthe15daysaftermaximumlight.ThelightcurveofSN2011feyieldsDm15(B)= 1.21 ± 0.03 mag, placing it in the middle of the range of values for SNeIa.Prietoet al.(2006)computelinearrelationshipsbetweentheDm15(B)and
Richmond and Smith, JAAVSO Volume 40, 2012878
peak absolute magnitudes for a large sample of SNe. If we insert our valueofDm15(B)intotheequationsfromtheirTable3forhostgalaxieswithsmallreddening,wederivetheabsolutemagnitudesshownintherightmostcolumnofTable6.TheexcellentagreementwiththeobservedvaluessuggeststhatourchoiceofdistancemodulustoM101maybeagoodone.
6. Comparison with visual measurements
PerhapsbecauseitwasthebrightestSNIatoappearintheskysince1972,SN 2011fe was observed intensively by many astronomers. The AAVSOreceivedover900visualmeasurementsoftheeventwithinsixmonthsoftheexplosion.SinceitwasobservedsowellwithbothhumaneyesandCCDs,thisstarprovidesanidealopportunitytocomparethetwodetectorsquantitatively. WeacquiredvisualmeasurementsmadebyalargesetofobserversfromtheAAVSO;notethatthesehavenotyetbeenvalidated.Weremovedasmallnumber of obvious outliers, leaving 880 measurements over the range 799<JD–2455000 < 984. For each of our CCD V-band measurements, weestimatedasimultaneousvisualmagnitudebyfittinganunweightedlow-orderpolynomialtothevisualmeasurementswithinNdays;duetothedecreasingfrequencyofvisualmeasurementsandthelesssharplychanginglightcurveatlatetimes,weincreasedNfrom5daysto8daysatJD2455840andagainto30daysatJD2455865.WethencomputedthedifferencebetweenthepolynomialandtheV-bandmeasurement.Figure7showsourresults:thereisacleartrendforthevisualmeasurementstoberelativelyfainterwhentheobjectisred.Ifwemakeanunweightedlinearfittoallthedifferences,wefind
(visual–V)2011fe=–0.09+0.19(04)*(B – V) (8)
wherethenumberinparenthesesrepresentstheuncertaintyinthecoefficient. WeknowoftwoothercasesinwhichvisualandothermeasurementsoftypeIaSNearecompared.PierceandJacoby(1995)retrievedphotographicfilmsofSN1937C,whichwereoriginallydescribedinBaadeandZwicky(1938),re-measured themwithaphotodensitometer,andcalibrated theresults to theJohnsonV-bandusingasetoflocalstandards.TheycomparedtheirresultstothevisualmeasurementsofSN1937CmadebyBeyer(1939)andfound
(visual–V)1937C=–0.63+0.53*(B – V) (9)
WeplotthisrelationshipinFigure7usingadottedline.JacobyandPierce(1996)discussedthedifferencesbetweenvisualmeasurementsofSN1991TfromtheAAVSO to CCD V-band measurements made by Phillips et al. (1992). WehaveextractedthemeasurementsofPhillipset al.(1992)fromtheirFigure2andcomparedthemtothevisualmeasurements,usingthemedianofallvisual
Wefindtheslopetobethemoreinterestingquantityintheserelationships,sincetheconstantoffsettermmaydependonthechoiceofcomparisonstarsforvisualobservers.Althoughatfirstblushtheslopesappeartobequitedifferent,ifoneexaminesFigure7carefully,onewillseethatthetrendisquitesimilarforallthreeSNeifonerestrictsthecolorrangeto(B – V)>0.5.Themaindifferencebetween these three events, then, lies in the measurements made when theSNe were relatively blue. Could that difference be real? We note that SNe1991T (definitely) and 1937C (probably) were events with slowly declininglightcurvesandhigher thanaverage luminosities,whileSN2011fedeclinedat an average rate and, for our assumed distance to M101, was of averageluminosity.AsPhillipset al.(1992)describes,thespectrumofSN1991TwasmostdifferentfromthatofordinarySNeIaatearlytimes,beforeandduringitsmaximumluminosity;itisalsoattheseearlytimesthatSNeshinewithbluelight.CouldthecombinationofphotometrybythehumaneyeandphotometrybyCCDreallydistinguishordinaryandsuperluminousSNeIaatearlytimes?Theevidenceisfartooweakatthistimetosupportsuchaconclusion,butwelookforwardtotestingtheideawithfutureevents. Stanton(1999)undertookamoregeneralstudy,comparingthemeasurementsofasetofroughlytwentystarsnearSSCygmadebymanyvisualobserverstotheJohnsonVasafunctionof(B – V).Hefoundarelationship
OurmulticolorphotometrysuggeststhatSN2011fewasa“normal’’typeIaSN,withadeclineparameterDm15(B)=1.21±0.03mag.AftercorrectingforextinctionandadoptingadistancemodulustoM101of(m – M)=29.10mag,wefindabsolutemagnitudesofMB=–19.21,MV=–19.19,MR=–19.18,andMI=–18.94,whichprovidefurtherevidencethatthiseventwas“normal’’initsopticalproperties.Assuch,itshouldserveasanexemplaroftheSNewhichcan act as standardizable candles for cosmological studies.Acomparisonofthe visual and CCD V-band measurements of SN 2011fe reveals systematicdifferencesasafunctionofcolorwhicharesimilartothosefoundforothertypeIaSNeandforstarsingeneral.
Richmond and Smith, JAAVSO Volume 40, 2012880
8. Acknowledgements
WeacknowledgewiththanksthevariablestarobservationsfromtheAAVSOInternational Database contributed by observers worldwide and used in thisresearch.WethankArneHendenandthestaffatAAVSOformakingspecialeffortstoprovideasequenceofcomparisonstarsnearM101,andforhelpingtocoordinateeffortstostudythisparticularvariablestar.MWRisgratefulforthecontinuedsupportoftheRITObservatorybyRITanditsCollegeofScience.WithoutthePalomarTransientFactory,theastronomicalcommunitywouldnothave received such early notice of this explosion.We thank the anonymousrefereeforhiscomments.
References
Baade,W.,andZwicky,F.1938,Astrophys. J.,88,411.Beyer,M.1939,Astron. Nachr.,268,341.Cook,K.H.,Aaronson,M.,andIllingworth,G.1986,Astrophys. J., Lett. Ed.,
2011a,Astron. Telegram,No.3581,1.Nugent,P.E.,et al.2011b,Nature,480,344.Patat,F.,et al.2011,arXiv,1112.0247.Perlmutter,S.,et al.1997,Astrophys. J.,483,565.Pierce,M.J.,andJacoby,G.H.1995,Astron. J.,110,2885.Phillips,M.M.1993,Astrophys. J., Lett. Ed.,413,L105.Phillips, M. M., Wells, L.A., Suntzeff, N. B., Hamuy, M., Leibundgut, B.,
New Light Curve for the 1909 Outburst of RT Serpentis
Grant LuberdaWayne OsbornYerkes Observatory, 373 W. Geneva St., Williams Bay, WI 53191; email correspondence should be sent to [email protected]
Received October 6, 2011; revised October 24, 2011; accepted October 24, 2011
Abstract Anew light curve for the1909outburst of theunusual novaRTSerpentis hasbeenderived.Publishedobservationshavebeen compiled andnew brightness measures determined from archival photographic plates.ModernphotometryhasbeenusedtoplacetheobservationsapproximatelyontheUBVsystem.Theoutbursthadanoverallincreaseofatleast5magnitudesandreachedmaximumin1917MaywithB=11.0andV=10.1
1. Introduction
TheunusualvariableRTSerpentis(=AN7.1917=NovaSer1909;J2000coordinates=R.A.17h39m51.98s,Dec.–11°56'39.0")wasdiscoveredbyWolfin1917(Wolf1917)andindependentlydiscoveredbyBarnardtwoyearslater(Barnard1919a).Asearchofarchivalphotographsshowedtheobjectwasnotvisibleonanyplatespriorto1909butafteroutburstremainednearmaximumbrightnessforseveralyears(Bailey1919,Wolf1919,Shapley1919)afterwhichitbeganaslowdecline(Shapley1927).RTSerisnowclassifiedasoneoftheraresymbioticnovaeobjects(Nussbaumer1992). OurinterestinthisobjectbeganwhenwecameacrossaplateintheYerkescollectionwiththeenvelopenotation“Remarkablevariable.”SomesleuthingrevealedthephotographwasoneofBarnard’sobservationsofRTSerduringoutburst.Oursearchfora lightcurveof theevent revealedonly theoldonepublished by Payne-Gaposchkin and Gaposchkin (1938). It seemed obviousthatderivinganewlightcurvemorecloselytiedtothetraditionalUBVsystemwouldbeworthwhile.
2. Observational Data
WehavecollectedavailableobservationsofthemagnitudeofRTSerpentiswithin twenty-five years of its outburst in 1909, i.e., data prior to 1935.The observations are of two types: magnitudes determined from images onphotographicplatesandvisualestimates.
Barnard(1919a,1919b),Bailey(1919,1921),andShapley(1919,1923,1927).Thepublisheddatahavebeensupplementedinthreeways.First,wehavemadeeyeestimatesofthebrightnessofRTSeronforty-twoplatesfoundintheYerkesObservatorycollection.Second,wehaveusedthedigitalcopiesoftheplatesof the University of Heidelberg’s Bruce Telescope that are available online(http://www.lsw.uni-heidelberg.de/projects/scanproject/) both to make eyeestimatesandtodeterminemagnitudesthroughaperturephotometry.Third,wehavedeterminedroughmagnitudesandepochsforthe“unpublishedHarvardobservations”plottedonthePayne-GaposchkinandGaposchkin(1938)lightcurve.
2.2VisualObservations VisualobservationsofRTSeraround the timeof itsoutbursthavebeenpublished by Mundler (1919), Barnard (1919b), Graff (1919, 1921, 1922,1927), and Lacchini (1921, 1929, 1933). We also downloaded the 121 datapointsforRTSerintheAAVSOInternationalDatabase(AAVSO2011)thatarewithinourtimewindow.AreviewoftheAAVSOobservationsshowedthatallbutthreearevisualestimatesbyLacchini.Theyincludehisthirteenpublishedmagnitudeestimatesbutwithmoreaccurateepochs.Forsomereason,however,his first eight observations have AAVSO magnitudes between 0.2 and 0.3magnitude systematically fainter thanhispublishedvalues for those epochs,withthepublisheddatabeingmoreconsistentwithhissubsequentbrightnessestimates.Thereisalsoonecaseofanepochwithaten-daydifference.Wehaveadoptedthepublisheddatainthediscrepantcases.
3. Reductions
We endeavored to place the diverse observations approximately on theUBVsystem.Specifically,pseudo-Bandpseudo-Vmagnitudeswerederivedfor the photographic observations and the visual measures, respectively.WebeganbydeterminingBandVmagnitudesforacomparisonstarsequencethatincludedthestarsthathadbeenusedbytheearlierinvestigators.TheadoptedBandVdataaregiveninTable1alongwiththeidentificationofthestarsandthesourceoftheBVphotometry.TychocataloguedataweretransformedtotheUBVsystemusingtherelationsofBessell(2000);AAVSOdataarefromCCDphotometry on the UBVRcIc system. Given the uncertainties in the RT Sermagnitudes,thecomparisonsequencevaluesareusuallygivenonlytoatenthofamagnitude.
whichliststheplatenumber,theJuliandateofmid-exposure,andthepseudo-Bmagnitude.Thefirstdigitsoftheplatenumberindicatetheapertureofthecameraemployed,withtwoormoreexposuresoftenbeingtakensimultaneouslywithdifferentlenses. Wolf (1919) published magnitudes from Heidelberg plates taken withtelescopes of three different apertures. Some plates—those taken with the41-cm (16-inch) Bruce Telescope—have been digitized and are availableonline.Wedownloaded thedigitized imagesofplatesshowingRTSerandboth performed aperture photometry and made eye-estimates using ourcomparisonsequence;ourresultsaregiveninTable3.Thesedatashowedthatacorrectionof+2.0magisneededtoconvertWolf’spublishedmagnitudestoapproximateBones,andthisfactorwasusedtotransformthepublisheddatafortheundigitizedplates. The observations from the Harvard plates were more difficult to adjust.Magnitudesaregiveninsixdifferentreferences(Bailey1919,1921;Shapley1919,1923,1927;Payne-GaposchkinandGaposchkin1938).Severalgivethemagnitudefromthecritical1909July9plate,andwecomparedthedifferentvalues to each other and to an eye-estimate we were able to make for thisplateusingourcomparisonsequence.OuranalysisindicatesthatallHarvardobservations other than those of Shapley (1919) are on the same system(referred toas the“HarvardSystem”).Anempiricallyderivedadjustmentof+0.6 magnitude was used to transform the Harvard system to our pseudo-Bsystem;thecorrectionusedforthe1919Shapleydatawas+0.1magnitude.
3.2VisualObservations Ingeneral,thevisualobserverslistedthecomparisonstarsusedandtheiradoptedmagnitudes.Wecomparedthelistedcomparisonstarmagnitudeswiththestars’Vonestoobtaincorrectionsfortransformingthevisualobservationsapproximately to theV system.The adopted corrections were +1.0 mag forBarnard’sobservations,+0.9magforGraff’s,and+1.1magforLacchini’sandfor theAAVSOdata.Mundler’sobservationsweremadewithaphotometer;thecorrectionaveraged0.0magbutdependedslightlyonwhichcomparisonstarsheusedforagivenobservation.Weestimateourzeropointshiftsmaybeuncertainbyuptothreetenthsofamagnitude.
4. Results
OurderivedBandVlightcurvesareshowninFigure1.TheBlightcurveincludesdatafromplatesofthefieldtakenbeforetheoutburst.Noevidenceoftheprecursorwasseenonplatesextendingbackto1891(Bailey1919),withthedeepestonesreachingpast16thmagnitude.Therisetomaximumtookplacebetween1908June,whentheBmagnitudewaslessthan16.4,and1910March,when the B magnitude was approximately 12. Unfortunately, few plates are
Luberda and Osborn, JAAVSO Volume 40, 2012890
availablethatwereexposedduringthetimeofrise.ThereisaHarvardplateof1909July9thatdefinitelyshowsRTSer.Oneofus(WO)hadtheopportunitytobrieflyexaminethisplateandthebrightnesswasestimatedatabout14.9usingour comparison sequence; the value from the corrected Harvard observers’estimatesput itat14.5.WealsofoundtwopossibleimagesofRTSerat theplatelimitonthedigitizedHeidelbergplatesof1909June9;theseindicateamagnitudeofroughly14.5. From1910to1917, thenova’smagnitudegradually increasedaboutonemagnitude, reaching a maximum of about B = 11.0 around 1917 May (JD2421350).ThiswaswhentheobjectwasdiscoveredbyWolfandbegantobefollowedbythevisualobservers.ThemaximuminVwas10.1,indicatingaB-Vofapproximately0.9. Aftermaximum, thebrightnessbeganadecline,witharateofabout0.4magnitude/yearforfouryearsafterwhich theratedecreasedto less than0.1magnitude/year.By1928both theBandtheVmagnitudeshaddecreasedtoabout 13.0.Thus, the B–V in the 1930s was approximately zero, indicatingthenova’scolorshiftedtowardtheblueasitsbrightnessdecreased.Thislikelyreflects thegradual spectraldevelopmentof strongemission lines,whichby1928dominatethespectrum(AdamsandJoy1928);theselineswouldenhancethe photographic brightness relative to thevisual.Other novae, suchTPyx,havealsoshownsimilarcolorevolution(Schaefer2010).Wecaution,however,thatboththevisualandphotographicestimatesforthelateryearsareuncertain.Inparticular,thephotographicdataafterJD2425100arefromtheunpublishedobservationsplottedonPayne-Gaposchkin’sgraphwhilethevisualobserversnodoubtreliedoncomparisonstarsforwhichonlyphotographicmagnitudeswereavailableforthesefainterestimates.Lastly,wenotewedonotfindstrongevidenceofthebrightnessfluctuationspreviouslyreported.
Figure 1. The newlight curve of the 1909outburst of RT Ser.Open triangles are Vmagnitudes,filledcirclesare B magnitudes, andthe dashes show the Bmagnitude plate limitwhen the star was toofainttobedetected.
Pollmann and Bauer, JAAVSO Volume 40, 2012894
International Observing Campaign: Photometry and Spectroscopy of P Cygni
Ernst PollmannEmil-Nolde-Str. 12, 51375 Leverkusen, Germany; [email protected]
Thilo BauerHeimerzheimer Str. 2, 53332 Bornheim, Germany
Received February 3, 2012; revised March 19, 2012; accepted March 19, 2012
Abstract In this combined campaign on the Luminous Blue Variablestar PCygni we are trying for the first time by way of contemporaneousmeasurementsofphotometricVbrightnessandHaequivalentwidth(EW)torealizea longtermmonitoringof the intrinsicHa-line flux.Thephotometricobservers of AAVSO and BAV (Germany) and a spectroscopic observergroup(Japan,France,Spain,Germany)startedobservingforthiscampaigninNovember2008attherequestofBerndHanischandoneofus(EP)inorderto continue former investigations whose results were based on multi-dailyaveragingofVandEW.AdditionaldatafromliteratureenableustorepresentthequantitativebehavioroftheHa-linefluxforthetimespanAugust2005toDecember2011,whichbehaviorreflectsvariabilitiesinmass-lossrate,stellarwinddensity,andionizationstructure.
1. Introduction
The international observing campaign, Photometry and Spectroscopy ofPCygni, begun in 2008 by Bernd Hanisch and one of us (EP) (Templeton2008a,2008b),isacooperativeprojectoftheAmericanAssociationofVariableStar Observers (AAVSO), Active-Spectroscopy-in-Astronomy (ASPA), andBundesdeutsche Arbeitsgemeinschaft für Veränderliche Sterne (BAV). Onegoal of the campaign is the monitoring of the behavior of the Hα-line equivalent width(EW)andthecontemporaneouschangesoftheV-bandmagnitudeofPCyg.Anothergoalistogatherfurtherinformationabouttheintrinsicfluxofthisspectralline. Asbackgrounditshouldbementionedthatalotofdifferentinvestigationsduringthelastdecadeshavebeencarriedouttoclarifythecausesofbrightnessandemissionstrengthvariationandpossiblelinksbetweenthem,theextensionsof the Hα-emitting wind, and line structures in different spectral areas. One of the earliest interesting investigations of the Hα emission was performed by Scuderiet al.(1992)inordertodeterminepropertiesofthemass-loss. Fiveyearslater,Najarroet al.(1997)carriedoutadetailedparameterstudyofthelinestrength,lineshape,andenergydistributionfortheHandHeIspectrum,
Pollmann and Bauer, JAAVSO Volume 40, 2012 895
in order to understand the nature of P Cygni and its wind. de Jager (2001)describedphotosphericmodelstoexplainoutwardmotionsintheatmospherein context of luminosity and brightness variations. The investigation of thelong-term spectral and quasi-simultaneous photometric behavior of P CygniofMarkovaet al.(2001a,2001b)wasthemainimpetustostartourcampaign. The (for us) important question of the quantitative evaluation of the Hα emissionand its reference to the radialdistributionof theemitting regionsaroundPCygwasinvestigatedinacomprehensiveinterferometricstudybyBalan et al. (2010). Relative to our investigation, it seems to give certainparallelsinthecurrentstudyofRichardsonet al.(2011).Theyareshowinginlong-termstudiesthecorrelationbehaviorofthecontinuumfluxand(notcontemporaneous)photometricV-data, and theyconclude that theyvary indifferentmannerincontextoflong-termandshort-termtimescales(thenon-contemporaneous nature of their data is one of the substantial differencesfromcampaign). InourcampaignitisassumedthatthevariabilityoftheEWiscausedbyvariationsofthecontinuumfluxandnotbyvariationsofthelineflux,whichwouldindicatevariationsinthestellarwinddensity.Therefore,thevariabilityofthecontinuumfluxshallbeourprimaryconcernwhenthepropertiesofthestellarwindsandrateofmasslossarestudied. Photometricandspectroscopicchanges inPCygniareshowntobeanti-correlatedonshort-andlong-termscales.Weobservedatotalchangeof35Åinthe equivalent width (EW) of the Hα line and of ~ 0.25 magnitude in the V-band brightness.OurobservationsextendfromJD2454671(23July2008)throughJD2455880(14November2011).
2. Results
Figures1and2illustrateourobservations.Figure1comparesofthebehaviorof the V magnitude (top) and the Hα-EW (below) during our campaign. Figure 2 is a plot of the Hα-EW (black points) as a function of photometric V magnitude (opencircles)ofPCygfromMarkovaet al.(2001b). AscanbeseeninFigure1,whentheEWdecreases,thecontemporaneousstellarbrightnessincreasesandviceversa.Sofar,ourownresultsinFigure1agree well with the results of Markova et al. (2001b), which are shown inFigure2forcomparison.Strictanti-correlationisexpectedifthevariationofthe continuum flux is independent from variations of the EW. If the Hα line fluxisconstantovertime,anincreaseofthecontinuumbrightnesswillyieldasmallerlinefluxfromthemeasuredEWandviceversa. To find out if and how the flux obtained from the spectral line profilesvaries, the EW measurements are corrected for the effect mentioned in theprevioussection.Fromthedefinitionsof
Pollmann and Bauer, JAAVSO Volume 40, 2012896
(1)
and the relation between stellar magnitudes and continuum flux variationsF2/F1=10–0.4(m2–m1),itfollowsthatthelinefluxisF=CEW/10(0.4Vphot). HereCisaconstantfactor.Inpractice,wecorrectEWwithasimpledivisionby10(0.4*Vphot).Thederivedquantityisthennotthelinefluxinphysicalunits,butaquantityproportionaltothephysicallineflux,correctedforcontinuumvariations. It is important to consider the absolute flux of the line becauseitsvariationsarecausedbytheeffectsofmassloss,stellarwinddensity,andchangesoftheionizationstateofchemicalelementsinit. Inthecurrentcampaignwehavealreadyobtained122nearlysimultaneousmeasurements of the EW and the flux in the V-band (Figure 3). Strictlyapplied, the continuum flux at 6563 Å should be used. But here ΔV is a good approximationsincethecolorindicesofPCygnidonotvarygreatly(Markovaet al.2001b,p.903). Figure3attemptstodisplayifandtowhatextenttheintrinsiclineflux(acontinuum-corrected EW) depends onV-magnitude. From a statistical pointofviewonecansaythatthelow0.25correlationcoefficient(whichshouldbezeroafter thecontinuumcorrection),withconsiderationof themeasurementuncertainties, suggests the conclusion that the Hα line flux is independent of V-magnitude.Withconsiderationofstandarddeviationandpossiblyotherkindsof errors, the temporal variation of the line flux of Hα in the plot in Figure 4 will representtheresultofvariationsinthemasslossrate,stellarwinddensity,andchangesoftheionization.The122EWandcontemporaneousV-measurementsof the current campaign are, of course, froma statistical point of view, stillnot sufficient to make firm statements regarding the simultaneous temporalbehaviorofVandtheintrinsiclineflux.Inordertoachievethisaim,furthermultiyear,simultaneousspectroscopicandphotometricmeasurementswillbecontinuedinthiscampaign.Maybethenwewillhaveaopportunitytoreporthereagainaboutthestateoftheresults.3. Acknowledgements
Wearegrateful toDr.DietrichBaade(ESO-München),Dr.OtmarStahl(Landes, Sternwarte Heidelberg), and Prof. Dr. Edward Geyer (formerlyDirector Observatorium Hoher List, University Bonn) for their criticalcommentswhich led to essential improvementsof thiswork.Wealsooffermany thanks to all the participants of the campaign for their worthwhilecontributionsandmeasurements. Thefollowingobserverstookpartattheproject:AAVSO(V-magnitude)—AdrianOrmsby,RobertE.Crumrine, JimFox,KateHutton,NickStoikidis,DavidB.Williams,E.G.Williams,CharlesL.Calia,ThomasL.Peairs,Jeffery
lo – l
l l
o
dlEW=∫
Pollmann and Bauer, JAAVSO Volume 40, 2012 897
G.Horne,MikeDurkin,DesmondLoughney,WolfgangVollmann;Spectroscopy(Hα-EW)—Mitsugu Fuji, Benjamin Mauclaire, Joan Guarro, Bernd Hanisch, Ernst Pollmann, Thierry Garrel, Valerie Desnaux, Olivier Thizy, Jean NoelTerry,ChristianBuil,StephaneCharbonnel,PierreDubreul,AlainLopez(fromBalanet al.(2010)). TheEW,V(phot),andlinefluxdataareavailableatthefollowingwebsite:http://astrospectroscopy.de/Data_PCyg_Campaign/Data%20table%20of%20the%20campaign.txt References
Figure1.Photometricandspectroscopicobservations:V-magnitude(top)andthe Hα-EW (bottom) of P Cyg during the campaign (including data of Balan et al.2010).
Figure 2. Plot of the Hα-EW (black points) as a function of photometric V-magnitude(opencircles)ofPCyg(fromMarkovaet al.2001b).
Pollmann and Bauer, JAAVSO Volume 40, 2012 899
Figure 4. Intrinsic flux of the Hα line of P Cyg, 22 August (JD 2453605) to 15 December2011(JD2455911).
The Pulsation Period of the Hot Hydrogen-Deficient Star MV Sagittarii
John R. PercyRong FuDepartment of Astronomy and Astrophysics, University of Toronto, Toronto. ON, M5S 3H4, Canada; [email protected]
Received January 20, 2012; revised April 30, 2012; accepted April 30, 2012
Abstract MV Sgr is a hot, hydrogen-deficient star which has undergoneRCrB fadings. We have used self-correlation analysis and Fourier analysisofCCDV-bandphotometryintheAAVSOInternationalDatabasetoidentifyaperiodof8.0daysin thisstar; theamplitudeisabout0.03magnitude.Thevariabilityismostlikelyduetopulsation.
1. Introduction
Hydrogen-deficient stars (Clayton 1996; Werner and Rauch 2008) are ararebutverydiverse(Jeffery2008a)groupofobjects,inadvancedandunusualstagesofevolution. In thispaper,weareconcernedwithhydrogen-deficientstars which can undergo the R CrB phenomenon—unpredictable fadings,followedbyslowreturn tomaximumbrightness.Mostof theapproximatelyfiftysuchstarsinourgalaxyarecool—“classical”RCrBstars—butafewhotmembersofthegrouphavebeendiscovered. Therearetwoproposedmechanismsforproducing“classical”RCrBstars:themergerofaheliumwhitedwarfandacarbon-oxygenwhitedwarf,andafinalheliumshellflash(Ibenet al.1996;SaioandJeffery2002).Observationalevidenceseemstofavortheformermechanism,thoughafewRCrBstarsmaybeproducedbythelattermechanism(Clayton2011).Thehothydrogen-deficientstarsdonotnecessarilyarisefromthesamemechanism(s)asthecoolones,andmayinfacthavediverseevolutionaryhistories(DeMarcoet al.2002). ManyRCrBstarsarealsopulsatingvariables,andthepulsationmaybepartly responsible for the mass loss that leads to the fadings. Pulsating hothydrogen-deficient variable stars have been classified as PV Tel stars, butJeffery(2008b)arguesthatthisclassificationshouldbereplacedbythreenewclasses,basedonthepulsationperiodandmodeinthestar. MVSgr(AAVSO1838–21,HV4168,V~13.35)wasdiscoveredtobeanRCrBstarin1928byMissIdaWoods(Hoffleit1959),andhasbeenstudiedbyvarioustechniquessincethen(DeMarcoet al.2002).ItsatmosphericpropertiesareTeff=16,000±500Kandlogg=2.48±0.30,andpulsationshadnotbeenfoundinMVSgrasof2008(Jeffery2008b). SinceobservationsofthisstarhavebeenmadebyAAVSOobservers,oneof
VisualandCCDV-banddatawere takenfromtheAmericanAssociationofVariableStarObserversInternationalDatabase(AID;Henden2012).Therewereatotalof2,315visualobservations,and138CCDV-bandobservations.Theformerweremadeby33differentobservers;thelatterweremadebyG.DiScala,M.Simonsen,J.Temprano,andD.Wells. ThemostnumerousV-bandobservationswereintheseasonJD2455644–2455852. Self-correlation analysis (Percy and Mohammed 2004) of theseshowedaclearperiodofabout8.0days,withafullamplitudeofabout0.03magnitude,andatleasteightrepeatingminima,indicatingcoherentvariability(Figure1).Self-correlationanaysisofthewholeV-banddatasetshowedaperiodof8.0±0.1days,withasimilaramplitude.Themeanerroroftheobservations,asdeterminedfromtheinterceptontheverticalaxis,is0.03magnitude.Self-correlationanalysisofthevisualdatadidnotshowadetectablesignal,whichisprobablyduetothesmallamplitudeandthemuchhighernoiselevel. TheanalysiswasrepeatedwithFourieranalysis,usingtheperiod04software(Lenz and Breger 2005). For the V-band data, in the season JD 2455644–2455852,and in thewholedataset, thehighestpeakswereat frequenciesof0.128cycle/day(period7.8days)andand0.122cycle/day(period8.2days),respectively;thelatterspectrumisshowninFigure2.Ineachcase,thehighestpeak was only slightly higher than the next-highest peak. It did, however,agreewiththeperiodfoundfromself-correlationanalysis.Forthevisualdata,thehighestpeakwasatafrequencyof0.204cycle/day(period4.9days),butthiswasonlyslightlyhigherthanthenoiselevel,andmaynotbesignificant,especiallyconsideringthesmallamplitudeandthemuchhighernoiselevelinthevisualdata.
3. Discussion and conclusions
MV Sgr displays a period of 8.0 days, which we assume to be due topulsation.Thesignal isnotstrong,but it isquiteclear intheself-correlationdiagram,andisconsistentwiththeresultsoftheFourieranalysis.ThisenablesustoplacethestarinJeffery’s(2008b)PVTelIsub-class. MV Sgr was not known to be pulsating (Saio and Jeffery 1988; Jeffery2008b).Wenowhaveonemorepieceofusefulinformationaboutthisstar.Also,thediscoveryofpulsationprovidesfurthersupportforthepossibleconnectionbetweenpulsationandtheRCrBphenomenoninhydrogen-deficientstars.
ASPConf.Ser.391,Astron.Soc.Pacific,SanFrancisco,3.Jeffery,C.S.2008b,Inf. Bull. Var. Stars,No.5817,1.Lenz,P.,andBreger,M.2005,Commun. Astroseismology,146,53.Percy,J.R.,andMohammed,F.2004,J. Amer. Assoc. Var. Star Obs.,32,9.Saio,H.,andJeffery,C.S.1988,Astrophys. J.,328,714.Saio,H.,andJeffery,C.S.2002,Mon. Not. Roy. Astron. Soc.,333,121.Werner,K.,andRauch,T.,eds.2008,Hydrogen-Deficient Stars,ASPConf.Ser.
391,Astron.Soc.Pacific,SanFrancisco.
Percy and Fu, JAAVSO Volume 40, 2012 903
Figure 2. Fourier spectrum for CCD V-band observations of MV Sgr. Thehighestpeakisatafrequencycorrespondingtoaperiodof8.0days.
Figure1.Self-correlationdiagramforCCDV-bandobservationsofMVSgrduring the season JD 2455644–2455852 when the observations are mostnumerous. There are repeating minima at multiples of 8.0 days: 8, 16, 24(weak),32,40,48,56,64...days,andmaximaat4,12,20,28(weak),36,44,52... indicating coherent variability.These minima and maxima are marked.Theminima(includingatDt=0),donotextendbelow0.025becausethisistheaverageerroroftheobservations.TheminimaandmaximadieoutatlargeDtbecauseofthescarcityofDmagnitudeswiththeseDtvalues.
de Ponthièreet al., JAAVSO Volume 40, 2012904
GEOS RR Lyrae Survey: Blazhko Period Measurement of Three RRab Stars—CX Lyrae, NU Aurigae, and VY Coronae Borealis
Pierre de Ponthière15 Rue Pré Mathy, Lesve, Profondeville 5170, Belgium
Jean–François Le Borgne14, Avenue Edouard Belin, F–31400 Toulouse, France
F. FumagalliCalina Observatory, Carona, Switzerland
Franz–Josef Hambsch12 Oude Bleken, Mol, 2400, Belgium
Tom KrajciP. O. Box 1351, Cloudcroft, NM 88317
J.-M. Llapasset 66 Cours de Lassus, F–66000 Perpignan, France
Kenneth Menzies318A Potter Road, Framingham, MA 01701
Marco Nobilevia Cantonale 53, 6942 Savosa, Switzerland
Richard Sabo2336 Trailcrest Drive, Bozeman, MT 59718
Received December 14, 2011; revised December 28, 2011; accepted January 9, 2012
Abstract We present the results of collaborative observations of three RRLyraestars(CXLyr,NUAur,andVYCrB)whichhaveastrongBlazhkoeffect.ThisworkhasbeeninitiatedandperformedintheframeworkoftheGEOSRRLyrSurvey(GroupeEuropéend’ObservationsStellaires).Fromthemeasuredlightcurves,wehavedeterminedthetimesandthemagnitudesatmaximum.Thetimesofmaximahavebeencomparedtoephemeridestoobtainthe(O–C)valuesandfromaperiodanalysisofthese(O–C)values,theBlazhkoperiodisderived.TheBlazhkoperiodsofNUAur(114.8days)andVYCrB(32.3days)arereportedhereforthefirsttimeandamoreaccurateperiodforCXLyr(68.3days)hasbeenobtained.ThethreestarsaresubjecttostrongBlazhkoeffect,butthiseffecthasdifferentcharacteristicsforeachofthem.Whenwecomparethevariationsofmagnitudeatmaximumandvariationsof(O–C)valueswith
de Ponthièreet al., JAAVSO Volume 40, 2012 905
respect to theBlazhkophase, thesevariationsare inphase, inopposition,oreveninquadrature.
1. Introduction
ThemainobjectiveoftheGEOSRRLyrSurveyistofollowthevariationsofperiodandBlazhkoeffectofbrightandwell-studiedRRLyraestars.Thesevariationsare followed in the long termwithTAROTrobotic telescopes(LeBorgneet al.2007andPorettiet al.2008).ThesecondobjectiveofthesurveyistheobservationofBlazhkoeffectofunder-studiedRRLyraestars.Theresultspresentedhereareinkeepingwiththisobjective. The RR Lyrae stars of Bailey type ab (RRab) are pulsating stars with aperiodbetween0.4and0.7day.SomeRRabstarsexhibitaphaseandamplitudemodulation. This phenomenon, known for a century, is called the Blazhkoeffect.Itisrecognizedthatthiseffectisstillnotwellunderstood.RRabstarsexhibiting the Blazhko effect appear to show a variety of characteristics.Recent continuous, high precision photometry from the Kepler satellitedocuments a period doubling for some RR Lyrae stars (Szabó et al. 2010).Withourground–basedsmallaperturetelescopesandtheirlimitedphotometricaccuracy,weattempttodeterminetheBlazhkoperiodofneglectedRRabstars.MonitoringduringseveralyearsisneededtodeterminetheBlazhkoperiodandtocharacterizetheBlazhkobehavior.Wehaveanalyzedthevariationsofthemagnitudeatmaximumand(O–C)valuewithrespecttotheBlazhkophaseforthreedifferentstars(CXLyr,NUAur,andVYCrB). Afterdarkandflatfieldcorrectionswiththemaxim dlsoftware(DiffractionLimited 2004), aperture photometry was performed using either aip4win(BerryandBurnell2001)orlesvephotometry(dePonthière2010),acustomsoftware which also evaluates the SNR and estimates magnitude errors. Nocolorcorrectionshavebeenappliedtothemeasuredmagnitudes.Thetimesofmaximaofthelightcurveshavebeenevaluatedwiththesamecustomsoftwarefitting the light curve with a smoothing spline function (Reinsch 1967).WehaveusedtheANOVAalgorithmofperanso(Vanmunster2007)toderivetheBlazhkoperiodfromthetimesofmaxima.
2. CX Lyr
ThestarCXLyrisclassifiedintheGeneral Catalogue of Variable Stars (GCVS;Samuset al.2011)asanRRabvariablestarwithaperiodof0.61664495day. CX Lyr observations during the second half of 2008 (JD 2454637 to2454783)havebeenpreviouslyreportedbydePonthièreet al.(2009).Duringanewobservationcampaignfrom2009to2011(JD2455041to2455807),weobtainedforty–onenewmaxima. Thecomparison starsusedby theauthorsaregiven inTable1.The star
de Ponthièreet al., JAAVSO Volume 40, 2012906
coordinatesandmagnitudesinBandVbandswereobtainedfromtheNOMADcatalogue (Zacharias et al. 2011). C1 was used as magnitude reference andtheothers as check stars.The choiceof different comparison stars creates amagnitudeoffsetduetotheircolordifferences.Thisoffsethasbeenevaluatedbycomparingthemagnitudesofacommoncheckstarandtakenintoaccount.Table2providesthelistof thesenewobservationsandFigure1showsthe(O–C)values.Forthesakeofcompleteness,observationsobtainedbyG.Maintz(Huebscheret al.2008,2010)andolderGEOSobservationsareincludedinthetableastheyareusedinthepresentanalysis. A linear regression of all available (O–C) values has provided a newpulsationperiodof0.616758day.The(O–C)valueshavebeenre–evaluatedwiththisnewpulsationperiod.Thenewelementsare:
The Blazhko period was determined by a period analysis of the (O–C)valueswiththeANOVAalgorithm.Themostsignificantperiodis68.3±0.4days(5.34c/y).TheperiodogrampresentedinFigure2indicatesotherpeaksat56.6days(6.45c/y),84.1days(4.34c/y),and113.3days(3.22c/y)whichareone-yearsamplingaliases. Thereisalsoanotherpeakat136days,thatis,twicethemostsignificantperiod.Data from theyear 2010 (JD2455300 to2455500) indicate that thesuccessive Blazhko cycles are not identical (Figure 1). The variations ofsuccessive cycles create spectral response at a multiple of the fundamentalperiod.An (O–C) folded light curve at 136days,would show twomaxima.A similar period analysis of the magnitude at maximum with the ANOVAalgorithmhasprovidedsimilarconclusions. Thefolded(O–C)andmagnitudeatmaximumcurvesversustheBlazhkophasearegiveninFigure3aand3b.Itcanbeseenthatthesetwocurvesarenearlyinphase,withtheminimareachedatthesameBlazhkophase.
3. NU Aur
The star NUAur is classified in the GCVS (Samus et al. 2011) as anRRab variable star with a period of 0.53941672 day and a Blazhko periodof179days.Duringa firstobservationcampaign,betweenDecember2006andFebruary2007(JD2454081to2454135),theeighteenobtainedmaximaclearlyshowedastrongBlazhkoeffectbutdidnotallowadeterminationoftheBlazhkoperiod.Theobservationof seventy–fivemaxima resulted froma second series of observations between December 2008 and March 2011
de Ponthièreet al., JAAVSO Volume 40, 2012 907
(JD2454752to2454640).ThecomparisonstarsaredocumentedinTable3.Star coordinates and B and V magnitudes are those found in theAAVSO’sComparisonStarDatabase(VSD).Thetimesofmaximumand(O–C)valuesaregiveninTable4andFigure4.TheobservationsofG.MaintzhavealreadybeenpublishedbyHuebscheret al.(2009)andthoseofK.MenziespublishedbySamolyk(2011). A linear regression on the (O–C) values has provided the followingelements:
Thefirst twopeaks(114.4±1.7and170.1±2.6days)areanaliaspair.Onefrequencyisthealiasatonecycleperyearoftheother.Thethirdperiod(227.1)daysisapproximatelydoublethefirstone(114.4).Theperiodof170.1daysisclosetothevaluereportedintheGCVS(175days).Thesealiasesareartifactsarisingfromgapsbetweennormal6–monthobservingseasons.WiththeSpectralWindowtoolinperanso,wehavetriedtodeterminewhichpeaksare artifacts of the seasonal sampling. This algorithm calculates the patterncausedbythestructureofgapsintheobservations.TheoutputoftheSpectralWindowisgiveninFigure5b,whereitcanbeseenthattheartifactpeaksarebroad.Thelistofprominentpeaksis:
we can not eliminate the second possible period of 170.1 ± 2.6 days. Moreobservationsareneededtoremovethisambiguity. Usingtheadoptedperiodof114.4days,thefolded(O–C)andmagnitudeatmaximumcurvesversustheBlazhkophasearegiveninFigure6aand6b.ItcanbeseenthatthesetwocurvesarenotinphaseaswasthecaseforCXLyr.ForNUAurstarthetwocurvesareinquadrature.
4. VY CrB
ThestarVYCrBisclassifiedintheGCVS(Samuset al.2011)asanRRabvariablestar.VYCrBisalsodesignatedasGSC2576–0980(SpaceTelescopeScienceInstitute2001).ItwasidentifiedasanRRabstaronphotographicplatesbyAntipin(1996).VYCrBishereinidentifiedasAntipin’sVar23withaperiodof0.462957day. We observed two maxima of VY CrB in April 2007 (JD 2454215 and2454216) and forty–nine maxima betweenApril 2010 andAugust 2011 (JD2455302to2455784).TheselectedcomparisonstarsaregiveninTable5.StarcoordinatesandBandVmagnitudeareobtainedfromtheNOMADcatalogue.Thetimesofmaximumand(O–C)valuesaregiveninTable6andFigure7.ThistablealsoincludesapreviousobservationobtainedbyA.Paschke(AgererandHuebscher2002). A linear regression of the (O–C) values has provided the followingelements:
Itisinterestingtoplotthemagnitudeatmaximumversusthe(O–C)values.If thesequantitieswerevarying in time as sinusoids andwere inphase, theresultinggraphwouldbeastraightlineinthefirstandthirdquadrants.Iftheywere in phase opposition, the graph would be a straight line in the second
de Ponthièreet al., JAAVSO Volume 40, 2012 909
andfourthquadrantsand if theywere inquadrature thegraphwouldexhibita circle.Theperiodicalvariationsofmagnitudeatmaximumand the (O–C)valuesarenotsinusoidal,butthecorrespondingparametricrepresentationwillneverthelessprovideusefulinformation. Thesegraphs for the threestarsaregiven inFigure10.For theCXLyr,the points are scattered along two segments forming a right angle but thegeneraltrendisaslopeat45degreesindicatingthat(O–C)andmagnitudeatmaximumare inphaseasshowninFigures3aand3b.Thepointsalong theverticalsegmentcorrespondtotheBlazhkophasesbetween0.0and0.5andtheotherpointsalongthehorizontalsegmentcorrespondtothesecondpartoftheBlazhkoperiod. InthediagramofNUAur,thepointswithamagnitudefainterthan12.9aregroupedonacircle,themagnitudeatmaximumand(O–C)valuesareinphasequadratureasshowninFigures6aand6b.Thegroupofpointswithamagnitudefainter than 12.9 are created by the non–repetitive behavior from Blazhkocycles. The full data set for NUAur covers more than ten Blazhko cycles. IntheVYCrBgraph,thepointsarescatteredalongacurvewithaslopeofabout135degrees.Themagnitudeatmaximumand(O–C)curvesareinphaseopposition. ForCXLyrandVYCrB,the(O–C)errorsarelargerwhenthemagnitudesatmaximumareat theirgreatestvalue.This ispartiallyduetoa lowerSNRbutmainlybecausethelightcurveatmaximumisflatter,whichleadstoalessprecisemaximummeasurement.
6. Conclusions
ThisstudyindicatesthatregularobservationsoverseveralseasonsoryearsbyamateurscanleadtothecharacterizationoftheBlazhkoeffectofRRLyrstars:thisisoneofthemainobjectivesoftheprofessional–amateurprogram“GEOSRRLyrSurvey.”Theseresultsshouldencourageamateurs to join inmeasurementcampaigns. The measurement of RR Lyrae stars having a strong Blazhko effecthighlightsthefactthatthiseffectisnotstandardfromonestartoanother,assatellite-basedobservations(CoRotandKepler)haveshown.Eachstarhasaparticularbehavioranditmaynotrepeatexactlyfromonecycletoanother.AcompleteastrophysicalmodelofBlazhkoeffectforRRabstarsshouldbeabletoexplainthesebehaviordifferences.
7. Acknowledgements
This research made use of the GEOS RR Lyr database http://rr–ly.ast.obs–mip.fr,hostedbyInstitutdeRechercheenAstrophysiqueetPlanétologie,Toulouse,France,andoftheSIMBADdatabase,operatedatCDS,Strasbourg,
de Ponthièreet al., JAAVSO Volume 40, 2012910
France. The authors acknowledge AAVSO Director Arne Henden and theAAVSO for the use of theAAVSONet telescopes at Sonoita (Arizona) andCloudcroft (NewMexico).Theywould also like to thankG.Maintz andA.PaschkefortheircontributionstotheGEOSdatabase. The authors wish to recognize their affiliation with the followingorganizationsorinstitutions:PierredePonthière—AAVSO,GroupeEuropéend’ObservationsStellaires,France(GEOS);Jean–FrançoisLeBorgne—GEOS,UniversitédeToulouse,France;F.Fumagalli—GEOS;Franz–JosefHambsch—AAVSO,BundesdeutscheArbeitsgemeinschaft fürVeränderlicheSternee.V.,Germany(BAV),GEOS,VerenigingVoorSterrenkunde,Belgium(VVS);TomKrajci—AAVSO;KennethMenzies—AAVSO,GEOS;MarcoNobile—GEOS;RichardSabo—AAVSO,GEOS.
References
Agerer, F., and Huebscher, J. 2002, Inf. Bull. Var. Stars, No. 5296, 1 (BAV MitteilungenNo.152).
Antipin,S.V.1996,Inf. Bull. Var. Stars,No.4343,1.Berry, R., and Burnell, J. 2001, The Handbook of Astronomical Image
Gerard SamolykP.O. Box 20677, Greenfield, WI 53220; [email protected]
Received January 6, 2012; accepted January 6, 2012
Abstract Thispapercontainstimesofmaximafor55shortperiodpulsatingstars (primarily RR Lyrae and d Scuti stars). This represents the CCDobservations receivedby theAAVSOShortPeriodPulsator (SPP)section in2011alongwithsomeearlierdata.
1. Recent observations
The accompanying list contains times of maxima calculated from CCDobservationsmadebyparticipantsintheAAVSO’sShortPeriodPulsator(SPP)Section.Thislistwillbeweb-archivedandmadeavailablethroughtheAAVSOftp site at ftp:ftp.aavso.org/public/datasets/jsamog402.txt.Theseobservationswere reduced by the writer using the peranso program (Vanmunster 2007).ColumnFindicatesthefilterused.Theerrorestimateisincluded. RRLyrstarsinthislist,alongwithdatafromearlierAAVSOpublications,are included in the GEOS database at: http://rr-lyr.ast.obs-mip.fr/dbrr/dbrr-V1.0_0.php.ThisdatabasedoesnotincludedScutistars. ThelinearelementsintheGeneral Catalogue of Variable Stars(Kholopovet al.1985)wereused tocompute theO–Cvaluesformoststars.Fora fewexceptionswhere theGCVSelementsaremissingorare insignificanterror,lightelementsfromanothersourceareused:VYCrB(Antipin1996),AHLeo(Schmidtet al.1995),VYLMi(HendenandVidal-Sainz1997),andGWUMa(Hintzet al.2001).
References
Antipin,S.V.1996,Inf. Bull Var. Stars,No.4343,1.Henden,A.A.,andVidal-Sainz,J.1997,Inf. Bull Var. Stars,No.4535,1.Hintz,E.G.,Bush,T.C.,andRose,M.B.2005,Astron. J.,130,2876.Kholopov,P.N.,et al.1985,General Catalogue of Variable Stars,4thed.,Moscow.Schmidt,E.G.,Chab,J.R.,andReiswig,D.E.1995,Astron. J.,109,1239.Vanmunster,T.2007,peransoperiodanalysissoftware,http://www.peranso.com
Samolyk, JAAVSO Volume 40, 2012924
SWAnd 55786.7504 85137 –0.3869 V R.Sabo 0.0009SWAnd 55793.8288 85153 –0.3850 V R.Sabo 0.0014SWAnd 55815.9392 85203 –0.3886 V R.Sabo 0.0014SWAnd 55827.8846 85230 –0.3847 V R.Sabo 0.0023SWAnd 55844.6873 85268 –0.3887 V K.Menzies 0.0006SWAnd 55852.6494 85286 –0.3876 V R.Sabo 0.0015SWAnd 55874.7608 85336 –0.3902 V R.Sabo 0.0012XXAnd 55796.8764 23120 –0.4746 V R.Sabo 0.0020XXAnd 55833.7370 23171 –0.4741 V R.Sabo 0.0015ZZAnd 55891.6583 56017 0.0266 V K.Menzies 0.0009ATAnd 55808.8146 21827 –0.0041 V R.Sabo 0.0025ATAnd 55832.8716 21866 –0.0068 V R.Sabo 0.0020ATAnd 55847.6745 21890 –0.0099 V R.Sabo 0.0026ATAnd 55868.6556 21924 –0.0039 V R.Sabo 0.0026ATAnd 55889.6287 21958 –0.0059 V N.Simmons 0.0020DMAnd 55857.7959 31949 0.0667 V K.Menzies 0.0018SWAqr 55777.9087 66798 –0.0022 V R.Sabo 0.0007SWAqr 55790.7716 66826 0.0002 V R.Sabo 0.0009SWAqr 55806.8453 66861 –0.0017 V R.Sabo 0.0010SWAqr 55836.7000 66926 –0.0017 V R.Sabo 0.0010AAAqr 55824.7823 57607 –0.1347 V R.Sabo 0.0015TZAur 55844.8578 91766 0.0127 V K.Menzies 0.0006TZAur 55844.8597 91766 0.0146 V G.Samolyk 0.0011TZAur 55869.9259 91830 0.0136 V R.Sabo 0.0009BHAur 55596.4980 28163 –0.0010 V K.Menzies 0.0012BHAur 55835.9501 28688 0.0039 V R.Sabo 0.0014BHAur 55872.8920 28769 0.0025 V K.Menzies 0.0009BHAur 55919.8685 28872 0.0018 V R.Sabo 0.0012NUAur 55575.6370 30616 0.2937 V K.Menzies 0.0020NUAur 55576.7099 30618 0.2878 V K.Menzies 0.0023NUAur 55589.6434 30642 0.2753 V K.Menzies 0.0031NUAur 55857.7213 31139 0.2631 V K.Menzies 0.0013RSBoo 55683.7390 36872 0.0069 V K.Menzies 0.0015RSBoo 55762.6015 37081 0.0055 V K.Menzies 0.0006STBoo 54957.6753 57491 0.0754 V R.Poklar 0.0018STBoo 55748.6354 58762 0.1041 V K.Menzies 0.0015STBoo 55759.8333 58780 0.1007 V R.Sabo 0.0013STBoo 55791.5561 58831 0.0867 V K.Menzies 0.0010
Table continued on following pages
Table1.RecenttimesofmaximaofstarsintheAAVSORRLyraeprogram. JD(max) Star Hel. Cycle O–C F Observer Error 2400000 +
Samolyk, JAAVSO Volume 40, 2012 925
SZBoo 55624.6954 53526 0.0118 V K.Menzies 0.0020TVBoo 55666.7388 99364 0.0756 V R.Poklar 0.0014TVBoo 55743.6272 99610 0.0744 V K.Menzies 0.0021TWBoo 55746.7935 54212 –0.0665 V R.Sabo 0.0008UUBoo 55647.6959 42815 0.2347 V K.Menzies 0.0014UYCam 55905.759 76170 –0.095 V G.Samolyk 0.004RWCnc 55596.5712 29314 –0.3343 V K.Menzies 0.0024RWCnc 55906.8342 29881 –0.3331 V G.Samolyk 0.0015TTCnc 55615.7041 27813 0.1189 V R.Poklar 0.0012TTCnc 55914.8750 28344 0.0982 V K.Menzies 0.0012RRCet 55826.8102 40948 0.0099 V R.Sabo 0.0020RRCet 55852.8042 40995 0.0116 V R.Sabo 0.0013RRCet 55913.6401 41105 0.0144 V R.Sabo 0.0020RUCet 55897.6386 27412 0.1202 V R.Sabo 0.0037RZCet 55875.7351 43025 –0.1840 V R.Sabo 0.0025RZCet 55896.6656 43066 –0.1885 V R.Poklar 0.0021TUCom 55628.6952 57760 0.1216 V K.Menzies 0.0023TUCom 55648.5661 57803 0.1347 V K.Menzies 0.0018TUCom 55719.6807 57957 0.1307 V K.Menzies 0.0019VYCrB 55760.8200 30207 –0.1321 V R.Sabo 0.0019XXCyg 55798.6319 84108 0.0025 V K.Menzies 0.0004XZCyg 55740.7980 24895 –2.1385 V N.Simmons 0.0011DMCyg 55772.7961 31416 0.0683 V R.Sabo 0.0010DMCyg 55785.8154 31447 0.0720 V R.Sabo 0.0015DMCyg 55797.5700 31475 0.0705 V K.Menzies 0.0008DMCyg 55803.8663 31490 0.0689 V R.Sabo 0.0011DMCyg 55809.7499 31504 0.0745 V R.Sabo 0.0012DMCyg 55825.6998 31542 0.0697 V R.Sabo 0.0012DMCyg 55835.7749 31566 0.0681 V R.Sabo 0.0011DMCyg 55852.5725 31606 0.0713 V K.Menzies 0.0009DMCyg 55862.6500 31630 0.0722 V R.Sabo 0.0013DMCyg 55875.6641 31661 0.0706 V R.Sabo 0.0009RWDra 55754.8872 36976 0.1972 V R.Sabo 0.0008RWDra 55767.7599 37005 0.2253 V R.Sabo 0.0006RWDra 55814.7015 37111 0.2177 V R.Sabo 0.0006RWDra 55825.7456 37136 0.1889 V R.Sabo 0.0009RWDra 55829.7266 37145 0.1836 V R.Sabo 0.0013RWDra 55833.7180 37154 0.1888 V R.Sabo 0.0013
Table continued on following pages
Table1.RecenttimesofmaximaofstarsintheAAVSORRLyraeprogram,cont. JD(max) Star Hel. Cycle O–C F Observer Error 2400000 +
Samolyk, JAAVSO Volume 40, 2012926
RWDra 55865.6079 37226 0.1887 V R.Sabo 0.0013RWDra 55873.5769 37244 0.1852 V R.Sabo 0.0009XZDra 55733.7791 28973 –0.1425 V K.Menzies 0.0013XZDra 55794.7806 29101 –0.1326 V K.Menzies 0.0007XZDra 55847.6739 29212 –0.1305 V R.Sabo 0.0017XZDra 55874.8360 29269 –0.1287 V R.Sabo 0.0024RRGem 55563.3943 35757 –0.4458 V M.Martinengo 0.0009RRGem 55575.7103 35788 –0.4465 V R.Poklar 0.0008RRGem 55584.4513 35810 –0.4463 V M.Martinengo 0.0008RRGem 55586.4336 35815 –0.4505 V A.Marchini 0.0005RRGem 55596.7671 35841 –0.4471 V G.Samolyk 0.0008RRGem 55603.5197 35858 –0.4488 V K.Menzies 0.0009RRGem 55603.5198 35858 –0.4487 V N.Simmons 0.0009RRGem 55833.9453 36438 –0.4633 V R.Sabo 0.0010RRGem 55862.9464 36511 –0.4659 V R.Sabo 0.0010RRGem 55868.9045 36526 –0.4675 V R.Sabo 0.0009RRGem 55901.8832 36609 –0.4656 V R.Sabo 0.0010RRGem 55915.7883 36644 –0.4663 V R.Sabo 0.0012GQGem 55578.7200 44258 –0.1941 V K.Menzies 0.0019GQGem 55605.5376 44305 –0.1932 V K.Menzies 0.0023TWHer 55741.7979 85577 –0.0142 V R.Sabo 0.0007TWHer 55761.7789 85627 –0.0132 V R.Sabo 0.0006TWHer 55775.7646 85662 –0.0135 V R.Sabo 0.0006TWHer 55797.7417 85717 –0.0144 V R.Sabo 0.0010TWHer 55823.7159 85782 –0.0142 V R.Sabo 0.0006TWHer 55872.4671 85904 –0.0142 V K.Menzies 0.0005VXHer 55767.7440 74702 0.0009 V R.Sabo 0.0010VXHer 55798.7055 74770 –0.0030 V R.Sabo 0.0009VZHer 55751.6914 43065 0.0728 V N.Simmons 0.0010ARHer 55757.8655 30434 –1.3137 V R.Sabo 0.0023ARHer 55798.7326 30521 –1.3390 V R.Sabo 0.0010ARHer 55806.7177 30538 –1.3444 V R.Sabo 0.0012DLHer 55785.7309 29710 0.0390 V R.Sabo 0.0016DYHer 55745.7557 150078 –0.0270 V R.Sabo 0.0007DYHer 55761.8064 150186 –0.0285 V R.Sabo 0.0011DYHer 55777.7105 150293 –0.0279 V R.Sabo 0.0006LSHer 55786.6014 120367 0.0228 V K.Menzies 0.0020SZHya 55611.7238 27795 –0.2801 V G.Samolyk 0.0021
Table continued on following pages
Table1.RecenttimesofmaximaofstarsintheAAVSORRLyraeprogram,cont. JD(max) Star Hel. Cycle O–C F Observer Error 2400000 +
Samolyk, JAAVSO Volume 40, 2012 927
SZHya 55624.6736 27819 –0.2241 V R.Poklar 0.0010SZHya 55639.6686 27847 –0.2718 V G.Samolyk 0.0026SZHya 55653.6876 27873 –0.2211 V R.Poklar 0.0012UUHya 55649.7164 30868 0.0036 V R.Poklar 0.0011DHHya 55648.7256 50046 0.0777 V R.Poklar 0.0009RRLeo 55604.6817 27209 0.1104 V R.Poklar 0.0009RRLeo 55666.6594 27346 0.1102 V K.Menzies 0.0013RRLeo 55905.9802 27875 0.1150 V G.Samolyk 0.0014TVLeo 55635.7908 27635 0.1192 V G.Samolyk 0.0014TVLeo 55664.7188 27678 0.1145 V R.Poklar 0.0018WWLeo 55567.8683 34313 0.0446 V K.Menzies 0.0027WWLeo 55648.6415 34447 0.0365 V G.Samolyk 0.0018WWLeo 55648.6480 34447 0.0430 V K.Menzies 0.0039AALeo 54196.5951 24388 –0.0751 V G.Lubcke 0.0013AHLeo 55576.8477 14856 0.0293 V K.Menzies 0.0018VYLMi 55565.8731 9652 0.0049 V K.Menzies 0.0011VYLMi 55603.7593 9724 0.0086 V K.Menzies 0.0024SZLyn 55838.8355 146965 0.0227 V G.Samolyk 0.0014SZLyn 55875.8397 147272 0.0227 V R.Sabo 0.0008SZLyn 55890.9069 147397 0.0231 V G.Samolyk 0.0007SZLyn 55921.7637 147653 0.0229 V N.Simmons 0.0008SZLyn 55921.8832 147654 0.0219 V N.Simmons 0.0010RZLyr 55747.6734 28488 –0.0232 V G.Samolyk 0.0013RZLyr 55756.8719 28506 –0.0271 V R.Sabo 0.0012RZLyr 55770.6657 28533 –0.0368 V K.Menzies 0.0019RZLyr 55771.6878 28535 –0.0372 V R.Sabo 0.0007RZLyr 55777.8244 28547 –0.0355 V R.Sabo 0.0009RZLyr 55793.6703 28578 –0.0381 V R.Sabo 0.0012RZLyr 55816.6874 28623 –0.0270 V R.Sabo 0.0017RZLyr 55837.6507 28664 –0.0246 V R.Sabo 0.0008RZLyr 55862.7012 28713 –0.0250 V R.Sabo 0.0017CXLyr 55745.7667 36772 1.1336 V R.Sabo 0.0025AVPeg 55744.8970 30623 0.1366 V R.Sabo 0.0015AVPeg 55760.9014 30664 0.1356 V R.Sabo 0.0011AVPeg 55771.8337 30692 0.1374 V R.Sabo 0.0011AVPeg 55792.9147 30746 0.1382 V R.Sabo 0.0010AVPeg 55798.7691 30761 0.1370 V K.Menzies 0.0009AVPeg 55807.7492 30784 0.1384 V R.Sabo 0.0010
Table continued on next page
Table1.RecenttimesofmaximaofstarsintheAAVSORRLyraeprogram,cont. JD(max) Star Hel. Cycle O–C F Observer Error 2400000 +
Samolyk, JAAVSO Volume 40, 2012928
AVPeg 55823.7558 30825 0.1397 V R.Sabo 0.0013AVPeg 55832.7326 30848 0.1379 V R.Sabo 0.0010AVPeg 55864.7437 30930 0.1382 V R.Sabo 0.0008AVPeg 55868.6454 30940 0.1362 V K.Menzies 0.0012AVPeg 55872.5512 30950 0.1382 V K.Menzies 0.0006BHPeg 53952.8716 22758 –0.1271 V G.Samolyk 0.0027BHPeg 54372.7303 23413 –0.1188 V G.Samolyk 0.0023BHPeg 55759.8356 25577 –0.1224 V R.Sabo 0.0017BHPeg 55825.8398 25680 –0.1404 V R.Sabo 0.0017BHPeg 55834.8125 25694 –0.1416 V R.Sabo 0.0024RVUMa 54219.6391 19536 0.1079 V G.Lubcke 0.0013RVUMa 55607.9182 22502 0.1211 V G.Samolyk 0.0017RVUMa 55634.6024 22559 0.1259 V N.Simmons 0.0011RVUMa 55634.6039 22559 0.1274 V G.Samolyk 0.0015RVUMa 55763.7830 22835 0.1219 V R.Sabo 0.0013RVUMa 55807.7821 22929 0.1234 V R.Sabo 0.0024AEUMa 55584.5518 232282 0.0002 V G.Samolyk 0.0006AEUMa 55584.6348 232283 –0.0028 V G.Samolyk 0.0007AEUMa 55584.7285 232284 0.0049 V G.Samolyk 0.0011AEUMa 55584.8102 232285 0.0006 V G.Samolyk 0.0006AEUMa 55584.8936 232286 –0.0020 V G.Samolyk 0.0005AEUMa 55584.9814 232287 –0.0003 V G.Samolyk 0.0008AEUMa 55890.7752 235842 0.0029 V G.Samolyk 0.0004AEUMa 55890.8557 235843 –0.0026 V G.Samolyk 0.0007AEUMa 55890.9446 235844 0.0003 V G.Samolyk 0.0007AEUMa 55907.7213 236039 0.0037 V G.Samolyk 0.0009AEUMa 55907.8063 236040 0.0026 V G.Samolyk 0.0006AEUMa 55907.8872 236041 –0.0025 V G.Samolyk 0.0007GWUMa 55921.7701 19301 0.0021 V G.Samolyk 0.0013GWUMa 55921.9671 19302 –0.0041 V G.Samolyk 0.0011
Table1.RecenttimesofmaximaofstarsintheAAVSORRLyraeprogram,cont. JD(max) Star Hel. Cycle O–C F Observer Error 2400000 +
Osborn and Mills, JAAVSO Volume 40, 2012 929
The Ross Variable Stars Revisited. II
Wayne OsbornO. Frank MillsYerkes Observatory, 373 W. Geneva Street, Williams Bay, WI 53191; email correspondence should be sent to [email protected]
Received May 23, 2012; accepted June 15, 2012
Abstract Better magnitudes and epochs have been determined for 190 ofthe379confirmedandsuspectedvariablestarsdiscoveredbyRossfrom1925to1931.Accuratepositionshavebeendeterminedforthoseobjectsforwhichunambiguousidentificationshadbeenlacking.TheseincludeanumberofcasesforwhichRoss’spublishedcoordinateshavelargeerrors.
1. Introduction
This is the second of two papers giving identifications and improvedmagnitudesandepochsforthesuspectedvariablestarsdiscoveredbyF.RossofYerkesObservatoryintheperiod1925and1931.PaperI(OsbornandMills2011)discussedthosestarsinRoss’sfirst,second,seventh,eighth,ninth,andtenthlistsofvariables(Ross1925,1926a,1928b,1929,1930,1931).Thispapergivesourresultsfortheremainingstars—thoseinRoss’sthird,fourth,fifth,andsixthlists(Ross1926b,1927a,1927b,1928a). TherationaleforthisworkalongwithourmethodsweredescribedinPaperIandwillnotberepeatedhere.Wewillmerelysaythatwehavere-examinedtheoriginalphotographicplatesusedbyRossforhisdiscoveriestounambiguouslyidentifytheobjectsanddeterminebettermagnitudesandepochsofobservationthanthosepublishedbyRoss.TheonlychangeinourprocedurewastoincludedatafromtheThird U.S. Naval Observatory CCD Astrograph Catalog(UCAC3;Zachariaset al.2010)indeterminingourapproximateBmagnitudes.
2. Results
Our resultsarepresented inTable1.Thecolumnsgive, respectively, theRoss star number, the corresponding variable star name (and, when needed,anotheridentificationbelow),theJulianDates(actuallyJD–2400000.0)andBmagnitudesforthetwocomparedplates.Theepochsaregeocentric,thatis,theyhavenot been converted toheliocentricones.Following each table arenotesformanyofthestars;thesegivesuchinformationaserrorsdetectedinRoss’spapersandcommentsontheidentification. Ofthe190starsconsideredinthispaper,159arenamedvariables,twenty-twoaresuspectedvariableslistedintheNSVcatalogue(Samuset al.2009),
Osborn and Mills, JAAVSO Volume 40, 2012930
sevenwereobservationsofminorplanets,oneisamissingBDstar,andoneavariableBLLacobject(BaumentandCudworth1981).OftheNSVobjects,wedonotseethevariationseenbyRossforR118usingthesameplatematerial,andthesupposedvariationsofR119andR164maybeduetoplateflaws. The case of R148 is interesting, this being a listing in the Bonner Durchmusterung catalogue (Argelander 1903), BD+34°531, for which Rossfoundthereisnocorrespondingstar(irrespectiveofthefactthatRoss’s1875coordinateswereincorrectlyprecessedfromtheBD’s1855position).Hisnotecardforthisstarhasthecomment“KustnerwritesthiswasintheskycertainlyonOct21,1856.”TousitseemspossiblethatBD+34°531resultedfromsomeerrorbythecompilersoftheBD.Thestarwiththesamenumberinthe+33°zone,BD+33°531,liesatthesamerightascensionandjustslightlysouth.ThefollowingcomparestheBDcatalogueentriesforthetwostars:
Thatbothstarsliesoclosetothezoneboundary,therevisioninthepositionforthemissingstar,andthesimilarityinnumber,magnitude,andcoordinatesforthetwostarssuggestconfusionwiththedatamayhaveledtotwocatalogueentriesfromtheobservationsofjustoneobject. Withthispaper,alltheRossvariableshavebeenreviewedandwecannowexpandupon Marsden’s (2007) comments concerning theRoss observationsthatwereofminorplanets.Forthefifteensuchcases,theaveragedifferencebetween the actual time of exposure of the plate and Marsden’s computedtimewas0.0dwithanRMSdispersionof0.15d,strongconfirmationthatourJulianDateshavebeencalculatedcorrectly.Theaveragedifferencebetweenourpseudo-BmagnitudesandtheexpectedVmagnitudesfortheminorplanetsis0.9mag,or0.75ifonediscrepantvalueisomitted.Notonlyisthisingoodagreement with the average (B–V) of around 0.8 for minor planets (Zellneret al.1975),itsupportsboththatRoss’sestimateswerevisualmagnitudesandthatourestimatesareapproximatelyontheBsystem. AsmentionedinPaperI,Ross’snotecardscontaincommentsforanumberofthestars,includingsomevisualobservationsbyhimorhiscolleagueswiththe Yerkes 40-inch refractor. This information is presented in an appendix,TableA,soitwillnotbelost.
3. Corrections to paper I
Soon after the publication of Paper I, a reader brought to our attention
Osborn and Mills, JAAVSO Volume 40, 2012 931
severalerrors.Mostinvolvethetenstarswehadnotedas“notlistedinNSVcatalog.” We had used SIMBAD to check for cross identifications betweentheRossvariablesandNSVobjects,butitwaspointedoutthatanumberofNSVobjectshadyettobeincorporatedintotheSIMBADdatabase.Usingtheon-line version of the General Catalogue of Variable Stars (http://www.sai.msu.su/gcvs/cgi-bin/search.htm)alltenobjectsturnedouttoindeedhaveNSVnumbers.TheseidentificationsandafewothercorrectionstoPaperIaregivenTable2.
4. Acknowledgements
This research made use of the SIMBAD database, operated at CDS,Strasbourg,France,andtheDigitizedSkySurveys(DSS)producedattheSpaceTelescopeScienceInstituteunderU.S.GovernmentgrantNAGW-2166.TheDSS images are fromphotographicdataobtainedusing theOschinSchmidtTelescope on Palomar Mountain and the UK Schmidt Telescope throughfundingprovidedbyTheNationalGeographicSociety, theNationalScienceFoundation,theSloanFoundation,theSamuelOschinFoundation,theEastmanKodakCorporation,andtheUKScienceandEngineeringResearchCouncil.
References
Argelander,F.W.A.1903,Bonner Durchmusterung des Nordlichen Himmels,2nd corrected ed. (BD Catalogue), A. Marcus and E. Weber’s Verlag,Bonn.
Baumert,J.H.,andCudworth,K.1981,Inf. Bull. Var. Stars,No.2039,1.Bedient,J.R.2003,Inf. Bull. Var. Stars,No.5478,1.Bidelman,W.P.,andCudworth,K.1981,Inf. Bull. Var. Stars,No.2055,1.Bidelman,W.P.,andvanAltena,W.F.1972,Inf. Bull. Var. Stars,No.744,2.Harwood,M.1960,Ann. Sterrewacht Leiden,21,387.Marsden,B.G.2007,Perem. Zvezdy,27,3.Osborn,W.,andMills,O.F.2011,J. Amer. Assoc. Var. Star Obs.,39,186.Ross,F.E.1925,Astron. J.,36,99(firstlistofvariables).Ross,F.E.1926a,Astron. J.,36,122(secondlistofvariables).Ross,F.E.1926b,Astron. J.,36,167(thirdlistofvariables).Ross,F.E.1926c,Astron. J.,36,172(thirdlistofpropermotionstars).Ross,F.E.1927a,Astron. J.,37,91(fourthlistofvariables).Ross,F.E.1927b,Astron. J.,37,155(fifthlistofvariables).Ross,F.E.1928a,Astron. J.,38,99(sixthlistofvariables).Ross,F.E.1928b,Astron. J.,38,144(seventhlistofvariables).Ross,F.E.1929,Astron. J.,39,140(eighthlistofvariables).Ross,F.E.1930,Astron. J.,40,34(ninthlistofvariables).Ross,F.E.1931,Astron. J.,41,88(tenthlistofvariables).
Osborn and Mills, JAAVSO Volume 40, 2012932
Samus,N.N.,et al.2012,General Catalogue of Variable Stars,publishedon-lineathttp://www.sai.msu.su/gcvs/cgi-bin/search.htm.
R114: Marsden (2007) showed this suspected variable was not a minor planet observation.
R115: Marsden (2007) showed this suspected variable was not a minor planet observation.
R118: Probably not variable. Ross indicated a variation of two magnitudes between his two epochs, but we do not see a brightness change for the star he marked as the variable on his finding chart, nor for any nearby star, on these same plates.
R119: Probably not variable. The brightening detected by Ross on the 1909 plate taken with the 10-inch camera seems to result from a plate flaw—a dark smudge that does not look like a star image. No brightening was seen on the 6-inch plate. Ross (1926c) discussed plate flaws found on his plates in his paper that gave the proper motion stars detected in the R119 field.
R127: Ross published his fainter magnitude as 15 but his note card has 15(?). We do not believe the variable was seen but that Ross detected the close companions to the variable.
R130:Ross’publisheddeclinationhasa28'errorbuthisfindingchartandourre-examinationofthe plates confirm this identification.
R136: Bidelman and Cudworth (1981), Bedient (2003), and Marsden (2007) showed object seen in 1925 was a minor planet.
R144: Marsden (2007) showed this suspected variable was not a minor planet observation.
R148: This special case is discussed in the paper.
R160: Marsden (2007) showed this suspected variable was not a minor planet observation.
R164: The apparent variability is probably due to a plate flaw. The star image is brighter on the 1915 10-inch plate, but not on the 6-inch plate taken simultaneously.
R166: Marsden (2007) showed this suspected variable was not a minor planet observation.
R168: Marsden (2007) showed object seen in 1905 was a minor planet.
R184: Ross’ published magnitudes for the two epochs are reversed.
R190: Marsden (2007) showed object seen in 1909 was a minor planet.
R194: Only Ross’ 10-inch plate could be located so we could not confirm the variability on the second plate set.
R200: Baumert and Cudworth (1981) identified this as the BL Lac object Ohio I 090.4.
R204: Marsden (2007) showed object seen in 1927 was a minor planet.
R205: Date of second observation was 1927 Feb 1, not Feb 2 as published by Ross.
Table continued on next page
Table1.IdentificationsandimproveddataforRossVariables105–294,cont. Ross Variable / First JD B Second JD B Other identification 2400000+ 2400000+
Osborn and Mills, JAAVSO Volume 40, 2012 939
R206: Bedient (2003) and Marsden (2007) showed object seen in 1909 was a minor planet.
R207: Ross inverted the identifications of variable and comparison star on his finding chart.
R218: Marsden (2007) showed this object was not a minor planet observation. We found Ross’ published coordinates are in error. His finding chart and our re-examination of the plates confirm this identification.
R225: Marsden (2007) showed this suspected variable was not a minor planet observation.
R228: Ross’ published magnitudes for the two epochs are reversed.
R230: Marsden (2007) showed object seen in 1915 was a minor planet.
R231: Marsden (2007) showed object seen in 1915 was a minor planet.
R239: This star is very close to QT Sct, but a careful comparison of the finding chart for that object (Harwood 1960) with the field suggests R239 is not the same object as QT Sct.
R243:Ross’publisheddeclinationhasan11'errorbutre-examinationoftheplatesbybothBidelmanand van Altena (1972) and ourselves as well as the Ross finding chart confirm this identification.
R244: Ross’ published declination has a 2.3° error but his finding chart and our re-examination of the plates confirm this identification.
R249: Marsden (2007) showed this suspected variable was not a minor planet observation.
R252:Ross’publisheddeclinationhasa30'errorbuthisfindingchartandourre-examinationofthe plates confirm this identification.
R262:Ross’publishedpositionhasa21'errorbuthisfindingchartandourre-examinationoftheplates confirm this identification.
R271: Marsden (2007) showed this suspected variable was not a minor planet observation.
R275: The variable may be the star slightly north of our adopted identification.
R279: Marsden (2007) showed this suspected variable was not a minor planet observation.
R284: Ross’ published right ascension has a 21' error but his finding chart confirms thisidentification. The 1927 Oct 22 plates could not be located and the listed magnitude for this date is based on Ross’ estimate.
R286: The 1927 Oct 22 plates could not be located but Ross’ finding chart confirms this identification. The listed magnitude for this date is based on Ross’ estimate.
R287: The 1927 Oct 22 plates could not be located but Ross’ finding chart confirms this identification. The listed magnitude for this date is based on Ross’ estimate. The variability was confirmed on a 1922 Nov. 26 plate (JD 2,423,385.535) which gave a magnitude of 12.6.
R288: The 1927 Oct 22 plates could not be located but Ross’ finding chart confirms this identification. The listed magnitude for this date is based on Ross’ estimate.
R290: The 1927 Oct 22 plates could not be located but Ross’ finding chart confirms this identification. The listed magnitude for this date is based on Ross’ estimate.
R291: The 1927 Oct 22 plates could not be located but Ross’ finding chart confirms this identification. The listed magnitude for this date is based on Ross’ estimate.
Ross’snotecards forhisearlierdiscoveriescontainentries showing thatobservationsofsomeof thesuspectedvariablesweremadewith the40-inchrefractorbyhimorcollaborators.Thefollowingtablereproducesthesenotesverbatim, omitting the common note that the object might be an asteroid.A“v”referstothevariablewhileletters(a,b,m,n)refertocomparisonstarsthatareindicatedonthefindingcharts.Forthosecaseswhereavisualmagnitudeestimatewasmade,wefollowthereproducedcommentswiththeapproximateJulianDateoftheobservationandapproximateVmagnitude,derivedbyusingmodernmagnitudevaluesforthecomparisonstars.
TableA.Ross’snotesandderivedVmagnitudeestimates. Ross Comments on Ross’s note cards reproduced verbatim JD m(V)
GSC 4552-1643: a W UMa System With Complete Eclipses
Dirk TerrellDepartment of Space Studies, Southwest Research Institute, 1050 Walnut Street, Suite 300, Boulder, CO 80302; [email protected]
John GrossSonoita Research Observatory, 1442 E. Roger Road, Tucson, AZ 85719; [email protected]
Received October 6, 2011; revised October 17, 2011; accepted November 17, 2011
Abstract ObservationsandanalysisofGSC4552-1642arepresentedandthesystem is shown to be aW UMa system with complete eclipses.The total/annularnatureoftheeclipsesresultsinaphotometricmassratiothatshouldbereliablebutthelightcurvesdohaveappreciableamountsofthirdlight.
1. Background
GSC4552-1643islistedasstar828321intheGettel,et al.(2006;hereafterGGM)catalogueofovercontactbinarystarswithanorbitalperiodofabout0.27day.VisualinspectionofthelightcurvegivenbyGGMshowedthatitmighthavecompleteeclipses.Giventhegreatlystrengthenedanalysisthatcompleteeclipsescanprovide,webeganobservingthesystemwithJohnsonBandVandCousinsIfiltersattheSonoitaResearchObservatoryinSonoita,Arizona.Weused the20-inch telescopeequippedwithaSantaBarbara InstrumentGroupResearch STL 6303 CCD camera. Calibration (bias, dark, flat) and aperturephotometry were done with iraf. Observations were made on six nights inMarchandApril2011withatotalof283observationsintheBfilter,459inV, and436inI.GSC4552-0002wasusedas thecomparisonstarandGSC1399-0059wasthecheckstar.Theimageswerereducedusingiraftobiasanddark subtract the frames before flatfielding, and the calibration frames werecreatedbymedian-combiningatleasttwentyrawimages.Aperturephotometrywas performed on the images using the iraf phot package. Data from theAAVSO PhotometricAll-SkySurvey (APASS) (Hendenet al. 2010) for thestarsareshowninTable1.Thecomparisonminuscheckstarvaluesshowednovariabilitygreaterthan0.01magnitudeoverthecourseoftheobservingrun.
2. Analysis
Weanalyzedourobservationswiththemostrecentversionofthephoebeprogram (Prša and Zwitter 2005) which is based on the Wilson-Devinneyprogram(wd;WilsonandDevinney1971;Wilson1979).TheJ–Hvaluefrom
Terrell and Gross, JAAVSO Volume 40, 2012942
2MASS(Skrutskieet al.2006)is0.40±0.03,whichcorrespondstoaneffectivetemperatureofabout5600K(KenyonandHartmann1995).TheB–VvaluefromAPASSis0.78±0.08,correspondingtoatemperatureofaround5400K.SincetheJ–HandB–Vvaluesareingoodagreement,wesetthemeaneffectivetemperatureofstar1(thestareclipsedatprimaryminimum)equalto5500Kforthelightcurvesolution.Theadjustedparametersweretheorbitalinclination(i), themeaneffectivetemperatureofthesecondary(T2), themassratio(q),themodifiedsurfacepotentialofthecommonenvelope(–1),theheliocentricJuliandateoftheprimaryminimum(HJD0 ),theorbitalperiod(P),thebandpassluminosityoftheprimary(L1),andthirdlight(l3).Unadjustedparameterssuchasthegravitydarkeningexponents(g1,g2)andbolometricalbedos(A1,A2)wereset to their theoreticallyexpectedvalues for convectiveenvelopes (0.32and0.5,respectively).Thesquarerootlimbdarkeninglaw,usingtheVanHamme(1993)coefficients(x1,x2,y1,y2),gavethebestfitstothelightcurves.Weusedwd mode 3, appropriate for overcontact binaries of this type. The derivedparametervaluesareshowninTable2.Thethirdlightvaluesaretheratioofthethirdlightvaluetothetotalsystemlightatquadrature.Thefitshowsthattheslightlydeepereclipse(by0.01magnitudeinV)isthepartialone,technicallymakingthisanA-typesystem,butthelightcurveclearlyshowsasymmetriesandlargescatter,sowecannotmakethatclassificationwithhighconfidence.Figure1showsthefitstotheobservations. In W UMa systems like this, there is an astrophysical interest in thetemperaturedifferenceofthetwocomponents.Theprimarystar’stemperatureisestimatedfromtheobservedB–VandJ–Hcolorswithanapproximateerrorof200Kfromtheobservationalerrorsofthecolors.ThelightcurvesolutiongivestheerrorinT2,giventheassumedT1fromtheobservedcolors.ThusthetrueuncertaintyinT2issimilartotheuncertaintyinT1.However,theuncertaintyinthetemperaturedifference,T2–T1,issimilartotheformalsolutionerrorinT2sinceT1isassumedtobefixed.Toillustratethis,were-didthelightcurveanalysiswithT1=5700KandfoundT2=5858K,foratemperaturedifferenceof158K,comparedtoavalueof152KforthesolutioninTable1.So,whiletheindividualtemperaturesareuncertainatthelevelof200K,theuncertaintyintheirdifferenceismuchsmaller,oforder10–20K.
3. Conclusions
Because of its total/annular eclipses and overcontact configuration, wemight expect the photometric mass ratio to be well-determined (Terrell andWilson2005).However,TerrellandWilson(2005)didnotconsidertheeffectofthirdlightonthemassratiodetermination.Thissystemshowsappreciablethirdlight,withthefluxfromthethirdbodycomparabletothatofthesecondarystaratquadrature.SimulationslikethoseofTerrellandWilson(2005)areunderwaytodeterminehow third lightmight affect their findings, althoughwedonot
Terrell and Gross, JAAVSO Volume 40, 2012 943
References
Gettel,S.J.,Geske,M.T.,andMcKay,T.A.2006,Astron. J.,131,621.Henden,A.A.,Terrell,D.,Welch,D.,andSmith,T.C.2010,Bull. American
expectittohavealargeeffectonthemassratiodetermination.AsdiscussedbyTerrellandWilson(2005),thephotometricmassratioforatotallyeclipsingovercontactsystemisaccuratebecauseofthehighaccuracytowhichtheratioof the stellar radii canbedetermined.Since third light affects thedepthsofbotheclipses,ratherthanoneortheotherstrongly,wedonotexpecttheratiooftheradiitobestronglyaffected,andthephotometricmassratioshouldbereasonably accurate.The luminosity ratio in this system is, however, not soextremeas topreclude the ability tomeasure a spectroscopicmass ratio, sothissystemcouldhaveitsabsolutedimensionsdeterminedaccuratelybecauseof thecompleteeclipses,even if thephotometricmass ratioshouldprove tobeinaccurateduetothirdlightissues.Aspectroscopicstudyofthissystemissorelyneeded.
Table1.Colorsofthevariable,comparison,andcheckstars. Star B–V
i 88.9±0.6° T2 5652±10K q 0.327±0.004 Ω1 2.435±0.008 HJD0 2452500.187±0.002 P 0.2698999±0.0000001day L1/(L1+L2)B 0.687±0.010 L1/(L1+L2)V 0.697±0.009 L1/(L1+L2)I 0.705±0.009 l3B 0.163±0.008 l3V 0.185±0.007 l3I 0.212±0.008 A1,A2 0.5(fixed) g1,g2 0.32(fixed) x1,y1(Bfilter) 0.77,0.09(fixed) x1,y1(Vfilter) 0.46,0.35(fixed) x1,y1(Ifilter) 0.19,0.48(fixed) x2,y2(Bfilter) 0.72,0.14(fixed) x2,y2(Vfilter) 0.43,0.37(fixed) x2,y2(Ifilter) 0.18,0.48(fixed)* The quoted errors are the formal errors from the differential corrections solution. Since the primary star’ s temperature is estimated from the observed B–V and J–H colors with an approximate error of 200 K; the secondary star’ s uncertainty in temperature is similar. The formal error is more indicative of the precision to which the solution can fit the ratio of the stellar temperatures, rather than the absolute temperatures.
Figure1.BVI lightcurvesofGSC4552-1643and the fits fromtheWilson-DevinneysolutiongiveninTable2.
Damassoetal., JAAVSO Volume 40, 2012 945
VSX J071108.7+695227: a Newly Discovered Short-period Eclipsing Binary
Mario DamassoAstronomical Observatory of the Autonomous Region of the Aosta Valley, fraz. Lignan 39, 11020 Nus (Aosta), Italy; INAF associated; [email protected] and [email protected]
and
Dept. of Physics and Astronomy, University of Padova, vicolo dell’Osservatorio 3, I-35122 Padova, Italy
Davide CenadelliPaolo CalcideseAstronomical Observatory of the Autonomous Region of the Aosta Valley, fraz. Lignan 39, 11020 Nus (Aosta), Italy; INAF associated
Luca BorsatoDept. of Physics and Astronomy, University of Padova, vicolo dell’Osservatorio 3, I-35122 Padova, Italy; INAF-OAPd associated
Valentina GranataDept. of Physics and Astronomy, University of Padova, vicolo dell’Osservatorio 3, I-35122 Padova, Italy
Valerio NascimbeniDept. of Physics and Astronomy, University of Padova, vicolo dell’Osservatorio 3, I-35122 Padova, Italy; INAF-OAPd associated
Received March 28, 2012; revised April 13, 2012; accepted April 16, 2012
Abstract WereportthediscoveryofanEWvariable,VSXJ071108.7+695227,withashortorbitalperiodof~0.238day.Thisperiodisveryclosetothelowerlimitof~0.22daythathasbeenfoundforEWsystems.Herewepresentanddiscuss photometric and spectroscopic data of the variable, collected at theAstronomicalObservatoryoftheAutonomousRegionoftheAostaValleyandattheAsiagoAstrophysicalObservatory.ThelightcurvesshowsomeasymmetriesandthespectrasuggestadK4classificationforthetwocomponents.Itcouldbeinterestingtocarryoutfurtherobservationsofthissystematdifferentepochsbecausesuchsystemsfrequentlyshowvariationsinperiodandinthefeaturesofthelightcurve.
Damassoetal., JAAVSO Volume 40, 2012946
1. Introduction
We report the discovery of VSX J071108.7+695227 (= 2MASSJ07110876+6952276), a short-periodeclipsingbinary system.Weclassify it as anEWtypevariable.Thisvariable, located in theconstellationCamelopardalis(R.A. 07h 11m 08.76s; Dec. +69° 52' 27.6"; epoch J2000.0), was discoveredthankstophotometricobservationsperformedattheAstronomicalObservatoryof theAutonomousRegionof theAostaValley (OAVdA),andat theAsiagoObservatory.Thesystemhasanestimatedorbitalperiodof~0.238day. The variable deserves further investigation mostly because the assessedperiod is very short and pretty close to the sharp cut-off at the lower limitof~0.22dayfoundbyNortonet al. (2011) for thesebinaries.Onlyaminorpercentage of EW variables have periods shorter than ours. For example,Rucinski (2006) found thatonlya few tensoutof the4,638EWsystemsoftheASASsamplehaveaperiodshorterthanorsimilartoours,Weldrakeet al.(2007)foundonlytwooutoffifty-eightshortperiodeclipsingbinariestohaveP<0.238day,andMilleret al.(2010)abouttenoutof533.Todate,thelargestsampleofEWsystemswithaperiodshorterthanoursismadeupofthefifty-threeEWbinarieswithP<0.2314daylistedbyNortonet al.(2011).Moreover,suchclosesystemsshowseveralfeaturesthatchangeovertime,astheorbitalperiodandtheshapeofthelightcurves,aswediscussbelow. Allthisconsidered,wepursuedaphotometricandspectroscopicanalysisofourvariableaimedatbettercharacterizingitsphysicalproperties.WedeterminedB,V,andRlightcurves,weassignedthetwocomponentstoaspectraltype,welookedforpossibleactivityindicatorsinthespectralikeemissionlines,andwesoughtasymmetriesinthelightcurves.
2. Instrumentation and methodology
Sincetheendof2008,attheOAVdAisunderwaytheimplementationofan extensiveobservational campaignaimedat finding small-sizedextrasolarplanetsaroundMdwarfsusingthephotometrictransitmethod(Damassoet al.2010;Giacobbeet al.2012).ThefirstobservationsofVSXJ071108.7+695227wereperfomedattheendofDecember2011,duringthecommissioningtestsofthenewtelescopearraywhichwillbeusedforafive-yearsurveyofhundredsofMdwarfsinthesolarneighborhood.Thearrayiscomposedoffouridentical40-cm f /8.4 Ritchey-Chrétien telescopes, each on a 10 MICRON QCI2000mountandequippedwithaFLIProLine1001ECCDcamera. After the discovery, we photometrically followed up on the binarysystemwiththemaintelescopepresentinOAVdA,an81-cmf/7.9Ritchey-Chrétiencoupledtoaback-illuminatedCCDcameraFLIProLinePL3041-BBandstandardBVRIfilters.Thesensoris2048px×2048px,withapixelarea of 15 × 15 μm2.The system has a FoV of 16.5×16.5 arcmin2 with a
2.1.Photometricdata Thescientificframestakenduringthediscoverynightandthefollow-upobservations were reduced and analyzed with the software package teepee(Transiting ExoplanEts PipElinE) developed by some of the authors. Adetaileddescriptionofthebasicfunctionsoftheteepeepipelineisprovidedin Damasso et al. (2010). In short, the final result of the data processingis thegenerationof thedifferential light curvesof all the stars in theCCDfieldwhichwereautomaticallyrecognizedineveryframeoftheseries.Theensemblephotometryisobtainedtestinguptotwelveapertures.Thebestsetofcomparisonstarsisautomaticallydeterminedforapre-selectedtargetfromalistofthe100brightestobjectsfoundinthefield,excludingtheonestooclosetotheCCDborders.TheapertureandsetofreferencestarsselectedattheendofthedataprocessingaretheoneswhichgivethesmallestRMSfortheentirelightcurveoftheobjectofinterest. Figure2 (a,b, c) shows thenormalizeddifferential lightcurvesofVSXJ071108.7+695227obtainedinB,V,andRfiltersonJanuary16,2012,duringthefollow-upobservationswiththe81-cmtelescope.Thedifferentialmagnitudesarecalculatedasthedifferencebetweentheaverageinstrumentalmagnitudesofthecomparisonstarsandtheinstrumentalmagnitudesofthevariablestar.ThecomparisonstarsautomaticallyselectedbytheteepeepipelineforeachfilterarelistedinTable2.ThelightcurvesclearlyunveiltheEWvariabilitytypefortheobjectandtheobservationscoveralmost1.75timestheorbitalperiodofthetwocomponentsofthesystem,whichweestimatedtobe~0.238dayusingtheFastChi-SquaredalgorithmdescribedinPalmer(2009)andfreelyavailableat http://public.lanl.gov/palmer/fastchi.html. Figure 3 shows the normalizedV-band lightcurve foldedaccording to theorbitalperiod.Thezeroepoch isHJD(0)=2455943.438. We estimate a mean V magnitude for the object through the relationsdescribed inPavlov(2009),whichuse theJ–Kcolor indexfromthe2MASScatalogue(Skrutskieet al.2006)andtheUfandUamagnitudesfromtheUCAC-3catalogue(USNO2012):
whereUSNO(B)=14.8andUSNO(R)=13.9,weobtainedB=15.0andR=13.9. Moreover, thestaralsoappears in theAPASScatalogue(AAVSO2012),which provides other estimates for B and V magnitudes: B=15.485±0.368;V=14.468±0.289.
2.2.Spectroscopicdata Tobettercharacterizethevariableandlookforpossibleactivityindicatorslike emission lines, on January 26, 2012, we took nine consecutive spectrawiththe182-cmCopernicotelescopeattheAsiagoObservatory(http://www.pd.astro.it/asiago/), using the AFOSC CCD camera. Two different opticalconfigurations were used (grism3 and grism8) with different dispersion andspectralcoverage.ThemainfeaturesoftheseconfigurationsaresummarizedinTable1.Wetookfivespectrausingthegrism3andfourwiththegrism8. InFigure3weoverplotted(solidanddashed)verticallinescorrespondingtotheorbitalphasesatwhichtheninespectraofthesystemwerecollected.Itcanbeseenthatthefirstones(withbothconfigurations)weretakenatmaximumbrightness.TheyareshowninFigure4.Theotherspectradonotshowsizeabledifferencesascomparedtothese,atleastatthelowdispersionaccessibletous.Itcanbenoticed,however,thattheyweretakenatanotverydifferentphase.3. Discussion
At theepochofourobservationsthelightcurvesof thevariableshowedsomeclearasymmetries,asexpectedfromsuchshort-periodsystems.ThemostsignificantisanevidentO’ConnellEffectasthemagnitudeofthetwomaximaisdifferent (DavidgeandMilone1984).Theconventionalway toassess thesizeofthiseffectistomeasurethepeakmagnitudeaftertheprimary(deepest)minimumsubtractedfromthepeakmagnitudeafterthesecondaryminimum.From Figure 2 we derive that ΔB ~ ΔV ~ ΔR~+0.05.Furthermore,acloserinspectionoftheBcurverevealsasmall“shoulder”alongtherisetowardsboththesecondarymaxima.Thisispossiblyduetotheocurrenceofstarspots. Infact,itiswell-knownthatthatEWvariablesareusuallyheavilyspottedsystems,astheoreticallypredictedbyBinnendijk(1970).Hewasthefirsttopropose that dark spots exist on the surfaces of the components to explainthe asymmetry and variability of the light curves at different epochs. Thishypothesis was later demonstrated by Doppler imaging (see for instanceHendryandMochnacki2000).OneoftheauthorsofthepresentpaperalreadydealtwiththisissueanddetectedthepresenceofstarspotsinlightcurvestakenatdifferentepochsofoneEWvariablewithanorbitalperiodof~0.355day(Damassoet al.2011).
Damassoetal., JAAVSO Volume 40, 2012 949
Starspotsareapossibleexplanationof theO’ConnellEffectaswell,butthereismuchuncertaintyinthisrespect(WilseyandBeaky2009). Itwouldbeinterestingtocollectnewdataofthissystematdifferentepochstoputinevidencepossibleevolutionsthatourdata,takenduringasinglenight,cannotreveal.Inparticular,short-periodsystemslikeourscanshowvariationsinperiodorintheshapeofthelightcurves,likeaninversionoftheprimaryandsecondarymaximaleadingtoanO’ConnellEffectwithoppositesign(seeforexamplethecaseofV523CastreatedinZhangandZhang2004). Letusnowturntothediscussionofthespectroscopicdata.ThespectrainFigure4displaythetypicalshapeofKstars,withoutanyevidentsuperpositionof different spectral types.This suggests that both components have similarspectraandsurfacetemperatures.Thisisinaccordancewiththefactthattheminimainthelightcurvesarealmostequalindeepness. WUMabinariesareexpectedtobelocatedwithinorinproximityofthemainsequencestars(Biliret al.2005).Thiscanbeconfirmedbyourspectra:theprominenceoftheMgHbandaround5200Åandthetooth-shapedMgHfeature around 4770Å strongly suggest that the stars are dwarfs (Gray andCorbally2009,p.262). Toobtainamorepreciseclassification,weresortedtothelineshighlightedinFigure4(c):theCaIlineat6162Å,theBaII-FeI-CaIblendat6497Å,andtheHα line at 6563 Å. The latter directly correlates with temperature, while the first twoinverselycorrelatewithtemperature,sothattheCaIorBaII-FeI-CaIblendto Hα ratio is a useful temperature indicator. Althoughinlate-typestarsitisoftendifficulttospotoutanactualcontinuumtoreferto,wecalculatedtheequivalentwidthsofthesethreelineswhichareonlyindicative.Theyare:
Wλ (CaI) = 1.7ÅWλ (blend) = 0.8ÅWλ (Hα) = 0.4Å
WedecidedtocompareourspectratotheonestabulatedbyAllenandStrom(1995)becausetheypossessaspectralcoverageandresolutionquitesimilartoours. The CaI to Hα ratio suggest a dK4 classification. Afurtherclueastohowtoclassifythetwostarscanbegivenbytheslightdifferenceindepthofthetwominima.Suchdifferenceisnearlyequalto0.05magnitude,thatis,~4.7%influx.CallingT1,2thesurfacetemperaturesofthetwocomponentswecanroughlyinferthat(T1/T2)
4 ~1.047 → T1/T2~1.012,thatis,thecomponent1ishotterthanthecomponent2by1.2%.Attemperaturesbetween 4000 and 5000K this means a difference of ~50K. Unfortunately,thesurfacetemperatureoflowermainsequencestarsisnoteasytocalculateandwithina samespectral type therecanbeavarianceofmore than100K(seeforexampleCasagrandeet al.2006).Hence, thecalculated temperature
Damassoetal., JAAVSO Volume 40, 2012950
differencecannotbesafelyassumedastotestifyadifferenceofspectraltypebetweenthetwocomponents.Intheend,werestuponadK4classificationforbothcomponents.Inaddition,thisconclusionisconsistentwithaclassificationbasedonthecolorindices:accordingtoCox(2000)adK4starisexpectedtohaveJ–H=0.58andH–K=0.11,andourvariablehasJ–H=0.583andH–K=0.125. Furthermore,wecantrytodeducethemassesofthetwostars.Ifwecallatheoverallsemi-majoraxisofthesystem(thatis,thesumofthetwosemi-majoraxeswithrespecttothecenterofmass)expressedinAUs,Ttheperiodexpressed inyears,andM1,2 themassesof the twocomponentsexpressed insolarunits,wehave:
If we consider the Hα line at 6563Å we have that Δλ ~ 8 Å, that is, the widthofthelineshouldextendaboutfrom6554.5to6571.5Å.Ifweconsiderthe CaI line at 6162 we have Δλ 8Å, that is, the width of the line should extend aboutfrom6154to6170Å.Thesevaluesareconsistentwithourspectra,butbecauseofthelowdispersionwecanattainwebelievewecannotdrawanysureconclusioninthisrespect. Wefinallylookedforpossibleactivityindicatorsinthespectralikeemissionlines,butwefoundnoneatthelevelofresolutionaccessibletous.
Miller,V. R., Albrow, M. D., Afonso, C., and Henning, T. 2010, Astron. Astrophys.,519A,12.
Monet,D.et al.1998,USNO-A V2.0 Catalog of Astrometric Standards, U.S.NavalObservatory,Flagstaff,AZ.
Norton,A.J.,et al.2011,Astron. Astrophys.,528A,90.Palmer,D.M.2009,Astrophys. J.,695,496.Pavlov, H. 2009, “Deriving a V magnitude From UCAC3” (http://www.
hristopavlov.net/Articles/index.html).Rucinski,S.M.2006,Mon. Not. Roy. Astron. Soc.,368,1319.Skrutskie,M.F.,et al.2006,The Two Micron All Sky Survey, Astron. J., 131,
astrometry/optical-IR-prod/ucac).Weldrake,D.T.F.,Sackett,P.D.,andBridges,T.J.2007,Astron. J.,133,1447.Wilsey,N.J.,andBeaky,M.M.2009,inThe Society for Astronomical Sciences
28th Annual Symposium on Telescope Science (May 19–21, 2009), Soc.Astron.Sci.,RanchoCucamonga,CA,107.
Table2.ComparisonstarsusedfordifferentialphotometryinB,V,andRbands. Star R.A. Dec. Band h m s ° ' ''
UCAC-3320-037770 070919.041 +695506.66 B V UCAC-3320-037772 070921.093 +695600.27 B UCAC-3320-037776 070924.095 +695659.02 B V R UCAC-3320-037779 070928.970 +695143.34 B R UCAC-3320-037783 070932.398 +695057.44 B UCAC-3320-037784 070932.649 +695842.06 B UCAC-3320-037787 070934.315 +694949.11 B V R UCAC-3320-037794 070942.534 +695059.13 B UCAC-3320-037795 070943.173 +695442.21 B UCAC-3320-037798 070945.939 +695515.41 B UCAC-3320-037799 070947.551 +695226.48 B UCAC-3320-037820 071009.948 +695517.80 B UCAC-3320-037821 071011.189 +695045.90 B UCAC-3320-037823 071014.455 +695034.28 R UCAC-3320-037827 071020.912 +694940.04 B V R UCAC-3320-037828 071021.806 +695230.72 B UCAC-3320-037831 071023.959 +695124.78 B R UCAC-3320-037835 071031.401 +695352.17 B UCAC-3320-037849 071038.339 +695718.61 B UCAC-3320-037857 071042.666 +694914.61 B UCAC-3320-037860 071044.576 +695152.55 B UCAC-3320-037862 071047.011 +695135.23 B UCAC-3320-037867 071054.909 +695217.17 B V R UCAC-3320-037871 071103.636 +695231.88 B UCAC-3320-037876 071107.800 +695127.49 B V UCAC-3320-037881 071110.879 +695129.52 B V UCAC-3320-037885 071114.572 +695607.65 B V R UCAC-3320-037896 071127.544 +694808.74 B UCAC-3320-037901 071132.558 +695203.82 B V R 2MASS07110649+6947296 071106.492 +694729.61 B 2MASS07094405+6952413 070944.051 +695241.40 B
Table1.Opticalconfigurationsusedtoacquirethespectraofthevariable. Configuration Spectral coverage Mean dispersion Resolution
Variability Type Determination and High Precision Ephemeris for NSVS 7606408
Riccardo FurgoniKeyhole Observatory, Via Fossamana 86, S. Giorgio di Mantova (MN), Italy; [email protected] Astrofili Mantovani—Gorgo Astronomical Observatory MPC 434, S. Benedetto Po (MN), Italy
Received May 11, 2012; revised June 4, 2012; accepted June 14, 2012
Abstract Aphotometric campaignanalysisof the starNSVS7606408hasbeenconductedinordertodeterminethetypeofvariabilityandhighaccuracyephemeris.Bycombiningtheobtaineddatawithotherdatasetsavailable,itwastriedtoimprovethedeterminationoftheperiod,highlighting,however,apossibleminimalchangeoftheperiodovertheyears.AtMay2,2012,theephemeriscalculated for this variable star is HJDmin = (2456015.44998±1.1×10–4)+E×(0.35482573±3×10–8)+E2×(2.4×10–10±1×10–11)
1. Introduction
In recent years many wide field photometric surveys have led to thediscoveryofalargenumberofvariablestarsofdifferenttypes.Theseinclude:NSVS (Northern SkyVariability Survey—an extensive variability survey ofthe sky north of Dec. −38˚ with daily time-sampling and a one-year baseline, by the Los Alamos National Laboratory; SuperWASP (Wide Angle Searchfor Planets)—UK’s leading extra-solar planet detection program comprisingaconsortiumofeightacademicinstitutions,includingCambridgeUniversity,theInstitutodeAstrofisicadeCanarias,theIsaacNewtonGroupoftelescopes,Keele University, Leicester University, the Open University, Queen’sUniversityBelfast, andSt.Andrew’sUniversity;ASAS (AllSkyAutomatedSurvey)—alow-costprojectdedicatedtoconstantphotometricmonitoringofthewholeavailable sky,which isapproximately107 starsbrighter than14thmagnitude—the project’s ultimate goal is detection and investigation of anykindofphotometricvariability.Despitethis,therehasneverbeenmadeaprecisestudyofmanyofthemwiththeaimofdeterminingtheclassofvariabilityandobtainingphaseplotswithareasonablylowscattering.Mostoftheinstrumentsdedicatedtolarge-scaleresearchuseshortfocallengthtelescopeswhichgiveextremelycrowdedimagesofstarfields,makingthephotometricmeasurementsdifficultandsometimesinaccurate.Thedatabasesgeneratedbythesesurveys,however,areaparticularlyvaluablesourceofdatainobtainingconfirmations
The data were obtained with a Celestron C8 Starbright, a Schmidt-Cassegrain optical configuration with aperture of 203mm and centralobstructionof34%. Thetelescopewaspositionedatcoordinates:45°12'32''N10°50'20''E(WGS84),inaruralareawithlowtomediumlightpollution.ThetelescopewasequippedwithafocalreducerBaaderPlanetariumAlanGeeIIabletobringthefocallengthfrom2030mmto1396mm.Thefocalratiowasreducedtof /6.38fromtheoriginalf /10. Thepointingwasmaintainedwith aSynthaNEQ6mountwith softwaresynscan3.27guidedthroughaBaaderVarioFindertelescopeequippedwithaBarlowlenscapableofbringingthefocallengthofthesystemto636mmandfocalratiooff /10.5. The guide camera has been a Magzero MZ-5 with Micron MT9M001monochromesensorequippedwithanarrayof1280×1024pixels.Thesizeofthe pixels is 5.2 μm × 5.2 μm for a resulting sampling of 1.68 arcsec/pixel. The CCD camera has been a SBIG ST8300m with monochrome sensorKodakKAF8300equippedwithanarrayof3352×2532pixels.Thepixelsareprovidedwithmicrolensesforimprovingthequantumefficiency,andthesizeof the pixels is 5.4 μm × 5.4 μm for a resulting sampling of 0.80 arcsec/pixel. Thecamerahasaresolutionof16bitswithagainof0.37e-/ADUandafull-well capacity of 25,500 electrons. The dark current is 0.02e-/pixel/sec at atemperatureof–15°C.Thetypicalreadnoiseis9.3e-RMS. Thecameraisequippedwith1,000×antiblooming:afterexhaustivetestingithasbeenverifiedthatthezoneoflinearresponseisbetween1,000and20,000ADU,althoughatupto60,000ADUthelossoflinearityislessthan5%.TheCCD is equipped with a single-stage Peltier cell ΔT = 35 ± 0.1° C which allows coolingatastationarytemperature.
Theobservationswere carriedoutwithout theuseofphotometric filterstomaximizethesignal-to-noiseratio.ThespectralsensitivityoftheCCD,asshowninFigure1, ismaximumatawavelengthof540nm,makingthedatamorecompatiblewithamagnitudeCV(ClearFilter—ZeroPointV). TheobservationswereconductedovereightnightsaspresentedinTable1. TheCCDcontrolprogramwasSoftwareBisque’sccd softv5.Once theimageswereobtained,calibrationframesweretakenforatotalof30darkof55secat–5°C,80darkflatof2secat–5°C,and100flatof2secat–5°C.Thedarkflatsanddarksweretakenonlythefirstobservationsessionandusedforallothersessions.TheflatswereperformedforeachsessionasthepositionoftheCCDcameracouldbevariedslightly,aswellasthefocuspoint. Thecalibrationframeswerecombinedwiththemethodofthemedianandthe masterframes obtained were then used for the correction of the imagestaken.AllimageswerethenalignedandanastrometricreductionwasmadetoimplementtheastrometricalcoordinatesystemWCSintheFITSheader.Theseoperationswereconductedentirelythroughtheuseofsoftwaremaximdl v5.18madebyDiffractionLimited.
4. The measurement of NSVS 7606408
ThestartobemeasuredwasobservedwithintheNorthernSkyVariabilitySurvey(NSVS;Wozniaket al.2004),whichisaphotometricsurveyforstarswithanopticalmagnitudebetween8and15.5.The researchwasconductedduringthefirstgenerationoftheRoboticOpticalTransientSearchExperiment(ROTSE-I) using a robotic system composed by four photographic lenseswithoutphotometric filtersconnected toCCDs.The researchwasconductedbyLosAlamosNationalLaboratory(NewMexico)inordertocovertheentirenorthernhemisphereofthesky.Inayearofworkbetween1999and2000anaverageof150photometricmeasurementsfor14millionstarswasmade. At the end of this research an automatically extended data analysis wasmadethatledtothediscoveryofvariablestellarsources.NSVS7606408hasbeen recognized as a source with confirmed variability (Shaw et al. 2009).Identificationdataforthisstarareasfollows:
Thechoiceofmakingobservationswithouttheuseofphotometricfiltersrequiredaproperstudyofthefieldinordertoidentifycomparisonstarswithcolor similar to the variable under study, and with adequate signal-to-noiseratio,toobtainasaccuratedifferentialphotometryaspossible. Thechoicewasmadebyanalyzing themeasures relating tophotometric2MASS(Skrutskieet al.2006)bandsJ,H,andKasmentionedintheCarlsberg Meridian Catalog 14 (CMC14; Copenhagen Univ. Obs. et al. 2006). In theabsenceofprecisemeasuresofthemagnitudeinthestandardVJohnson-Cousins,forstarsusedintheanalysisitwasdecidedtoderivetheirVmagnitudefromthecatalogCMC14r'magnitudeasdescribedinDymockandMiles(2009).Thiswayensuresagoodreliabilityandafinalerrortheoreticallynotmorethan0.05magnitudeinV.Theconversionformulaisasfollows:
Thepossiblespectral typehasbeenderivedfromthe2MASSJ–Hvalueinstead, as shown in CMC14. The method is described in Stead and Hoare(2002)withtheresultsasshowninTable2. Withinthestarfieldanalyzed,onlytwostarshaveanadequatesignal-to-noise ratioandcolor similar to thevariableunder study.TheirpositionwithrespecttothevariableisshowninTable3andFigure3. ThroughdifferentialphotometrytheaveragemagnitudeofNSVS7606408is13.572±0.006CVdeterminedbyplacingthecomparisonstar’smagnitudeat13.325CV.WhereasthevariabilityofthestarisbyfarsuperiortothegapthatthisvaluehaswiththeVmagnitudederivedfromCMC14(=13.641),itcanbeconsideredanacceptableresult.Themeanmagnitudeerrorforallthesessionsis0.006CV.Forthelongestobservingsessionthelightcurvesofthe
Furgoni, JAAVSO Volume 40, 2012 959
variablestarandthecheckstarinrelationtothecomparisonstarareshowninFigure4. The check star shows an almost complete absence of significant trends:thisfactshowsthatthecolorindexofstarsusedisgoodenoughfordifferentialphotometrycarriedoutinunfilteredconditions.
5. Data analysis
Before proceeding further in the analysis, the time of the light curvesobtained was heliocentrically corrected (HJD) in order to ensure a perfectcompatibilityofthedatawithobservationscarriedoutevenataconsiderabledistanceintime.Fromhereon,theobtaineddatasetwillbereferredtoasFRICinthiswork,(FRICbeingtheauthor’sAAVSOobservercode). Thedeterminationoftheperiodwascalculatedusingthesoftwareperiod04(Lenz and Breger 2005), using a Discrete Fourier Transform (DFT). Theaveragezero-point(averagemagnitudeoftheobject)wassubtractedfromthedatasettopreventtheappearanceofartifactscenteredatafrequency0.0oftheperiodogram.Allthedatapointswereweightedinrelationtotheamountofthemagnitudeerror.Theperiodogramwascalculatedwithafrequencyrangefrom0to25cd–1(cyclesperday)andstep0.00118537583cd–1.TheresultisshowninFigure5. The calculated frequency is 5.636659(89) cd–1 (SNR 12.9) with anamplitudeof0.164±0.001CV.Theperiod thusappears tobe0.1774100(28)day with a time of minimum 2456008.35376(17) HJD. The calculation oftheuncertaintieswascarriedoutwithperiod04usingthemethoddescribedinBregeret al.(1999).TheresultingphasediagramisshowninFigure6. In theVSXcatalogue (Watsonet al. 2012)maintainedby theAmericanAssociationofVariableStarsObservers(AAVSO),thestarislistedasapossibleeclipsingvariableWUMa-typeorasapulsatingvariabledSct.Byobservingthisphasediagramonecanseethatthecurveshapemuchmoreresemblestherotationofaneclipsingvariableratherthanapulsatingone.Ingeneral,itisnoteasytodistinguishbetweenEWeclipsingbinariesandpulsatingstarssuchasdSctorRRCsincemanyofthemhavesymmetriclightcurves,andsometimescomparable periods. In any case it is possible to distinguish them using thefollowingcriteria:
Based on these considerations and the more plausible classification of thisvariableasaWUMasystem,IproposeanewperiodforNSVS7606408equalto P=0.3548200(28) day and HJDmin=2456015.45016(17). The new phasediagramisshowninFigure7.Thelowdispersionphasediagramallowsustorecognizethefollowingparameters:
• Primaryminimumplacedatphase0isdeeperthantheminimumplacedat phase 0.5. The difference in intensity is estimated on the order of0.028±0.005magnitude(theresultisobtainedbycalculatingthedifferencebetweentheaveragevalueofthedataplacesbetweenphase0.95andphase0.05andthoselocatedbetweenphase0.45andphase0.55).
• The maximum placed at phase 0.25 is slightly less bright than themaximumplacedatphase0.75.Thedifferenceinintensityisestimatedon the order of 0.013±0.005 magnitude (the result is obtained bycalculatingthedifferencebetweentheaveragevalueofthedataplacesbetweenphase0.2andphase0.3and those locatedbetweenphase0.7andthephase0.8).
Asaseriesofphotometricmeasurementsareaccurateandprovidedwithlowscatter, thetemporalextensionof themisakeyrequirementtoallowanaccuratedeterminationof theperiod.The longer their extension, in fact, themorenoticeablewillbetheeffectthataminimumerrorinthedeterminationoftheperiodwilllendtothephasediagramobtained.Eventhephenomenonofaliasingtendtodecreaseifinthepresenceoftemporalcoveragesufficientlylongandprovidedwithadequateresolution.Myobservationsgiveaphaseplotwithreduceddispersion;untilnow,nophasediagramhasbeenpublishedforNSVS 7606408, except that associated with NSVS data, which however, isplaguedbyaverynoticeablescattering. Itwasthereforedecidedtoevaluatealltheexistingphotometricobservationsforthisstarabletoprovideareasonableamountofdatatoimprovetheperiodspecifiedabove.
Furgoni, JAAVSO Volume 40, 2012 961
Thefollowingsurveysweretakenintoconsideration:
• NSVS ThestarislistedasNSVS7606408.Thereare233observationsmadefromApril5,1999,toMarch30,2000,withameanmagnitudeof13.72±0.07(theoriginaldataare275pointsbutarereducedto233aftertherejectionontheflagNGoodPoints).Thedatawereobtainedthroughthe SkyDot service operated by LosAlamos National Laboratory (SkyDatabase for Objects in Time-Domain. Data at http://skydot.lanl.gov/nsvs/star.php?num=7606408&mask=32004). I have made a heliocentriccorrectionofthedata.
• SuperWASP The star is listed as 1SWASP J114856.59+260230.1.Thereare5,916observationsmade fromMay2,2004, toMay17,2008withameanmagnitudeof13.73±0.04(Tamflux2).Thedata,providedbythePublicSWASPArchiveintheformofDataFits,wereextractedusingfvsoftware,onlyforthevaluesTmid,Tamflux2,andTamflux2_Err.(Tmid=HJDMeanTimeofExposurefromJDReferencewhich is2453005.5;Tamflux2=Originalflux(Tamuz)correctedonthebasisofadecorrelationtechnique highlighted by Collier Cameron et al. (2006). This flux isprovidedinmicrovegasanditsconversiontomagnitudesisgivenbytheformulamag.=15–2.5log(Flux);Tamflux2_err=errorcalculatedonfluxTamflux2.) The SuperWASP data are already heliocentrically correctedandevenfor theaverage timeofexposure.ThewideavailabilityofdatarelatedtothisdatasetledmetomakeaselectiononthosethatshowedtheTamflux2_Errlessthan0.05magnitude.Thenumberofobservationsusedisthusdownto3,158withatimespanfomMay4,2004,toMay16,2007,andameanmagnitudeof13.706.
• ASAS ThestarislistedasASAS114858+2602.5.OnlydatarelatingtoASAS-3(V-Magnitude)arepresent,nottoASAS-2(I-Magnitude).Thereare139observationsmadefromJanuary2,2003,toMarch2,2009,withameanmagnitudeof13.67±0.03.(Theoriginaldataare279pointsbutarereducedto139afterdiscardingallthosewithflagD(worstdata,probablyuseless)andall thosehavingat leastonemeasure invalid(Mag0;Mag1;Mag2;Mag3;Mag4=29.999.Theaveragemagnitudewascalculatedbyperforming for eachgivendata theaverageof thevalues Mag1,Mag2;Mag3;Mag4correspondingtodifferentaperturesusedforthecalculationofthestellarflux.)ThedatawereprovidedbytheAstronomicalObservatory,UniversityofWarsaw,andarealreadyheliocentricallycorrected.
Table4showstheperioddeterminationsforeachdatasetusingthesoftwareperiod04.(Whenthesamecriteriausedfordeterminingtheperiodanduncertaintiesof FRIC dataset data were also used here, they have been omitted.) TherespectivephasediagramsandFourierspectraareshowninFigures8through13.
Furgoni, JAAVSO Volume 40, 2012962
7. Combining the available data
DuetothedifferentspectralsensitivityoftheCCDsusedinthesesurveys(aswellasthelackortheuseofdifferentfilters),thecombinationofthedatasetscan not be trusted as regards the determination of the amplitude variation.Using theDFT,only theSWASPdataset showsanamplitudeofvariationatallsimilartotheFRICdataset(0.164±0.001)whilealltheotherdatasetsshowslightlydifferentvalues.Moreover,theNSVSandASASdatasetshaveastrongscatteringwhichcouldmakeinaccuratethecombinedFourieranalysis.SinceitisunclearhowthesedifferencesaffecttheDFTofthecombineddata,onlytheFRICandSWASPdatasetswillbeusedincombinationforthiscalculation.Inthiscase,theDFTmaybeusedonlytoobtainapossibleperiodtobeusedasagoodstartingpointforamoredetailedO–Canalysis. InordertocombinethedatasetsitwasfirstnecessarytonormalizethemeanmagnitudeoftheSWASP,NSVS,andASASdatasetstotheaveragevalueoftheFRICdataset.(FortheNSVSandASASdatasetstheoffsethasbeenappliedonlyfortherealizationofthecombinedphasediagramshowninFigure15.)TheappliedoffsetsaregiveninTable5. Theperiodogramobtained,calculatedonthefrequencybetween0and15(cd–1)withstepsof0.00001712275(cd–1)isshowninFigure14. Thechoiceofcombiningthetwodatasetshasledtoasignificantreductioninerrorsrelatingtotheperiod,theamplitudevariationandthedeterminationofthetimeofprimaryminimum. The calculated period is 5.63656978(8) cd–1 (SNR=13.068), whichcorrespondsto2P=0.35482573(3)day,theamplitudeis0.1645±0.0007mag.with HJDmin=2456015.44998(11). The resulting phase diagram for all thedatasetsbasedonthesenewparametersisgiveninFigure15. Visuallywecanseethat theNSVSdataset,whichis theearliest in time,isclearlynotinphasewiththeFRICandSWASP.Tryingtoslightlyvarytheperiod to improve the phase of NSVS we immediately lose the coincidenceof other datasets. Now it would be necessary to use an O–C analysis evenif theNSVS,ASAS,andSWASPdatasetsdidnothavea sampling rate thatallows to verify precisely the times of minimum. Due to the fact that therearenoobservations in the literature regarding the timeofminimum for thisvariablestarwewillproceedasfollows:foreachdatasetwillbetakenastimesofminimumthetimesinwhichthelightintensityislessthanmagnitude13.75.(ThelownumbersofASASdataandtheirhighdispersionrendersthemuselessforthistypeofanalysis.ThisdatasetwasnotusedforthecreationoftheO–Cdiagram.)Obviously,becauseofscattering,eventhiscalculationwillbenoisy,but if a period change is present, theO–Cdiagram should show the typicalshape(Figure16). Comparedtotheprimaryminimumobservedon2456015.44998(11)HJD,the earlier NSVS primary minimum that best represents the time shift with
Furgoni, JAAVSO Volume 40, 2012 963
respecttothecalculatedephemerisisoutofphaseby0.042(1)day.Hereitisimportant to remember that thevertical shapeof thedifferentminima in thegraphisduesolelytotheuncertaintyofthederivedtimeoftimimumrelatedtothedataset’sscattering.Ihaveusedforfittingasecondorderpolynomialandthegoodresultobtainedsuggeststhatwearefacedwithaprogressiveslowingoftheperiod.Now,ifweassumethattheperiodisslowingdownforaprobablemasstransferbetweenthetwostars,isnecessarytocalculatehowmucheachcyclemustslowdowntoaccumulatein13,326cyclesatimeshiftof0.042(1)day. (These are the complete cycles that the star has made since the earlierNSVSprimaryminimumuntilHJDmin=2456015.44998±0.00011HJD.) It isnecessary that theperiod increases4.8×10–10±1×10–11daysateachcycletobeabletoaccumulateatimeshiftof0.042(1)dayin13,326cycles.Asthecalculationisbasedonverysmallquantitiesitisnecessarytoprovideadoublecheck: it ispossible toevaluate the reliabilityof thispredictionbylooking at the data provided by period04 concerning the calculation of theperiodofeachdataset. If the period really increases by this small amount in each cycle it isnecessary that in the 13,326 cycles which separate the two minima it isincreasedby6.396×10–6days.Theperiodcalculatedbyperiod04fortheearlierNSVSdatasetis0.3548184(14)dayandifweaddtothisthevalue6.396×10–6daysweobtain0.3548248(14),whichisveryclosetotheperiodobtainedbyperiod04fortheFRICandSWASPdatacombinedasshowninTable5. Inotherwords,thedifferentlengthoftheperiodfoundbytheDFTinthemeasurementscarriedoutin1999and2000byNSVSincomparisonwiththosecarriedoutmorerecentlyisverysimilartothevariationoftheperiodcalculatedwiththeO–Cdiagram.
8. Conclusions
By combining the available data I have tried to classify the star NSVS7606408,originallyconsideredintheliteratureapossibleWUMa-ordSct-typevariablestarwithaperiodof0.17740547day.Analyzingtheobtaineddataitisbelievedpossibletoidentifythetypeofvariabilityandaprecisecalculationoftheephemeris.Thefinalresultingdataareasfollows:
VizieROn-lineDataCatalog:I/304.Dymock,R.,andMiles,R.2009,J. British Astron. Assoc.,119,149.Lenz,P.,andBreger,M.2005,Commun. Asteroseismology,146,53.Liu,Q.-Y.,andYang,Y.-L.2003.,Chin. J. Astron. Astrophys.,3,142.Shaw,J.S.,et al.2009,“FindingperiodicvariablesintheNSVS”(https://www.
Table 1. Activity log for the data collection concerning the FRIC dataset.TheNExTvalueindicatestheNetExposureTimewhichisthesessiondurationspentinpurephotonscollection. Date UTC Start UTC End NExT1 Useful Airmass Airmass (dd-mm-yyyy) (hh:mm:ss) (hh-mm-ss) (hh-mm-ss) Exposures Min Max
Table 3. Comparison star positions and information with respect to NSVS7606408. Object R. A. (2000) Dec. (2000) 2MASS CMC14 CMC14 Probable UCAC3 UCAC3 J–H r'mag derived spectraltype h m s ° ' " Vmag.
Table4.Perioddeterminationresultsforeachdataset,usingperiod04software. Dataset Checked Step SNR Period Amplitude Range (days)
NSVS 0–25(cd–1) 0.00013895(cd–1) 7.4 0.3548184(14) 0.170±0.005 SWASP 0–25(cd–1) 0.0000452(cd–1) 11.9 0.35482227(13)0.164±0.001 ASAS 5–6(cd–1) 0.0000222(cd–1)* 4.0 0.3548227(14) 0.157±0.027* Several problems afflicting this dataset: high scattering, a few observations available. Despite this, forcing the search for a frequency only between 5 (cd–1) and 6 (cd–1) we obtain a result consistent with other datasets.
Figure 1. Quantum Efficiency of the Kodak KAF8300 monochrome sensor withmicrolens, equipping the CCD SBIG ST8300m used for the data collection of thisstudy.
Figure 4. Differential photometry of NSVS 7606408 and the check star UCAC3233-104578 in relation to the comparison star UCAC3 232-105102.The 26/03/2012sessionisshown,whichisthelongestone.Thecheckstarshowsanalmostcompleteabsence of significant trends: this fact shows that the color index of stars usedis good enough for differential photometry carried out in unfiltered conditions.
Figure14.FourierspectrumoftheFRIC+SWASPdatasets.Anoffsetof–0.158maghas been applied to SWASP to normalize the mean magnitude to that of FRIC.Theperiodogramwascalculatedwithafrequencyrangefrom5to6cd–1(cyclesperday)andstep0.00001712275(cd–1).
Figure 15. Phase plot of FRIC, SWASP, NSVS, and ASAS dataset with period of0.35482573 day.We can see that the NSVS dataset, which is the earliest in time, isclearlynotinphasewiththeFRICandSWASP.
Hoffmeister(1963)discoveredthevariablestarMPGeminthefieldnGemof the “Sonneberger Felderplan” (Figure 1). He wrote: “the star is invisible1944February24/25(2plates),February25/26(7plates),onallotherplatesbright,perhapswithvery lowchanges.Onneighboringdaysnoplatesexist.Perhaps anAlgol-like star with a long period.” I have checked these platesagainandcanconfirmthatthestarwasinvisibleonthosedates. Gessner(1973,whereMPGemislistedasS7958Gemonp.589)observedMPGemonallplatesavailableatSonnebergin1973.Shecouldnotfindanyfurtherminimaandnoted:“thevariableconsistsofcomponents,thesoutherncomponent is blue (Palomar atlas, page 417).” The variable is the southcomponent. Nofurtherobservationsofthestarareknown.Ithereforehaveobservedthestaronall119laterplatesoftheSonnebergskypatrolfromtheyears1981to1994.Icouldnotfindanyfurtherminima.AlightcurveforMPGemisshowninFigure2, basedonmagnitudes estimated from theSonnebergplates.ThecomparisonstarwasUSNO1050-04327215(mpg15.6).
2. Discussion and conclusion
BrightnessmeasurementsinBVRIJHKofthestarexistindifferentcatalogs.ThesemeasurementsshowthatitprobablyisastarofspectraltypeA. Thestar isabout16thmagnitude inVandsocouldbemonitoredeasilywith amateur telescopes. The VSP chart is very good for finding, but mostplanetariumprograms(Guide8.0)showthestaratafaultyposition.ThecorrectpositionisR.A.06h48m33.3s, Dec. +19º 37´ 15˝ (J2000). I would like to call for theobservationofthestar,sothatitispossibletoconfirmtheperiod,whichispresumablyverylong.
Gerard SamolykP.O. Box 20677; Greenfield, WI 53220; [email protected]
Received January 23, 2012; accepted January 23, 2012
Abstract Thispapercontinuesthepublicationoftimesofminimaforeclipsingbinarystars fromobservations reported to theAAVSOEBsection.Timesofminima from observations made from April 2011 through December 2011alongwithsomeunpublishedtimesofminimafromolderdataarepresented.
1. Recent Observations
The accompanying list contains times of minima calculated from recentPEP and CCD observations made by participants in theAAVSO’s eclipsingbinaryprogram.Thislistwillbeweb-archivedandmadeavailablethroughtheAAVSOftpsiteat ftp://ftp.aavso.org/public/datasets/gsam2j402.txt.This list,alongwiththeeclipsingbinarydatafromearlierAAVSOpublications,isalsoincludedintheLichtenkneckerdatabaseadministratedbytheBundesdeutscheArbeitsgemeinschaftfürVeränderlicheSternee.V.(BAV)at:http://www.bav-astro.de/LkDB/index.php?lang=en. These observations were reduced by theobserversorthewriterusingthemethodofKweeandVanWorden(1956).Thestandarderrorisincludedwhenavailable.ColumnFinTable1indicatesthefilterused.A“C”indicatesaclearfilter. ThelinearelementsintheGeneral Catalogue of Variable Stars(GCVS;Kholopovet al.1985)wereusedtocomputetheO–Cvaluesformoststars.ForafewexceptionswheretheGCVSelementsaremissingorareinsignificanterror, light elements fromanother sourceareused:CDCam (BaldwinandSamolyk 2007), CW Cas (Samolyk 1992a), V1115 Cas (Kholopov et al.2011), Z Dra (Danielkiewicz-Krośniak and Kurpińska-Winiarska 1996), DF Hya (Samolyk 1992b), EF Ori (Baldwin and Samolyk 2005), GU Ori(Samolyk1985).LightelementsforV471Cas,BCEri,V728Her,CDLyn,V1128Tau, KM UMa, and MSVir are from up-to-date linear elements ofeclipsing binaries (Kreiner 2012). O–C values listed in this paper can bedirectly compared with values published in the AAVSO Observed Minima Timings of Eclipsing Binariesseries.
Samolyk, JAAVSO Volume 40, 2012976
References
Baldwin,M.E.,andSamolyk,G.2005,Observed Minima Timings of Eclipsing Binaries No. 10,AAVSO,Cambridge,MA.
Baldwin,M.E.,andSamolyk,G.2007,Observed Minima Timings of Eclipsing Binaries No. 12,AAVSO,Cambridge,MA.
Danielkiewicz-Krośniak, and E. Kurpińska-Winiarska, M., eds. 1996, Rocznik Astron. (SAC68),68,1.
Kholopov,P.N.,et al.1985,General Catalogue of Variable Stars,4thed.,Moscow.Kholopov, P. N., et al. 2011, General Catalogue of Variable Stars, Online
Edition(http://www.sai.msu.su/gcvs/gcvs/index.htm).Kreiner, J.M.2012,Up-to-date linear elementsof eclipsingbinaries (http://
www.as.up.krakow.pl/ephem/).Kwee,K.K.,andVanWorden,H.1956,Bull. Astron. Inst. Netherlands,12,327.Samolyk,G.1985, J. Amer. Assoc. Var. Star Obs.,14,12.Samolyk,G.1992a, J. Amer. Assoc. Var. Star Obs.,21,34.Samolyk,G.1992b, J. Amer. Assoc. Var. Star Obs.,21,111.
JD(min) STANDARDSTAR HEL. CYCLEO-C F OBSERVER ERROR 2400000+
Star JD (min) Cycle O–C F Observer Standard HJD 2400000+ (day) Error (day)
The Variable Stars South Eclipsing Binary Database
Tom RichardsDirector, Variable Stars South, Pretty Hill Observatory, P.O. Box 323, Kangaroo Ground, Vic 3097, Australia; [email protected]
Received October 1, 2012; revised October 18, 2012; accepted October 18, 2012
Abstract VariableStarsSouth(VSS)hasthreeactiveprojectsusingelectronicdetectorstostudyeclipsingbinaries,especiallyEAs.InadditiontosupplyingJDobservationaldata to theAAVSO InternationalDatabase,VSSmaintainsa database of observed times of minima (ToM) and linear light elementsderived from the ToMs. This database, located on the VSS website www.variablestarssouth.org, is updated monthly. In addition, the same page linkstotheareasforthethreeprojectswhichmaintainextensiveobservationalandanalyticdata.
1. Introduction
Variable Stars South (VSS) is an online organization of astronomersinterestedinstudyingsouthernvariablestars.ItisaresearchsectionoftheRoyalAstronomicalSocietyofNewZealand.AllinformationaboutVSSincludingitscollecteddatacanbefoundonitswebsite,www.variablestarssouth.org.VSSprimarilyactsasahosttoprojectsorganizedbyindividualsorteams,andthreesuchare:
Although these projects have very different goals, they all collect andanalyzeelectronic(DSLRandCCD)timeseriesdataoneclipsesofbinarystarsystems.Informationabouteachproject,includingsciencecases,observationalrequirements,andguidesforobserversandanalysts,aswellasobservationaldata and analyses, can be found under the Research Projects menu on theVSSwebsite.Collaborationintheseprojects isopentoanyastronomerwithappropriateequipment.Theeclipsedatafromthethreeprojectsarecollectedintoasingledownloadablefile, theVSSEclipsingBinaryDatabase.JDdatafromtheseobservationsarealsosuppliedbyindividualobserverstotheAAVSOInternational Database—a requirement on all VSS observational projects.Thereisnofixednortherlylimittothetargetsinanyoftheprojects—themainrequirement is to be well observable from temperate southern latitudes.AtpresentthemostnortherlytargetisLTHerinSPADES,at+09°57'52''.
• Excel Observation Files of every observation set obtained—all to afixedformat,assubmittedbytheobserver.
• AnExcelResultsFilemaintainedby the analyst responsible for thattargetsystem,containing:allToMsmeasuredfromtheObservationFiledata; sets of linear light elements (LEs) published by others (such asGCVS);calculationsofO–CoftheToMsagainstapublishedLEset(suchasGCVS);andwheresufficientmeasuredToMsexist,alinearestimateoftheLEsfromtheToMs.
System—Nameofthebinarysystem.Formatisconstellationabbreviation,thenGCVSidentificationwhereavailable,with3-digitV...identificationsexpandedwithaleading"0"(forexample,AraV0536).WheretheGCVSidentification is not available the constellation abbreviation is retainedthen another catalogue identifier is used, whose provenance should beobvious.The table is sorted lexically on this column, so, for example,"Sco"entriesprecede"Sgr",and"GruRU"precedes"GruW".
Error—Measured uncertainty in HJD_min, to one or two significantfigures.
Min V mag—MeasuredmagnitudeofminimuminJohnsonVband.
Min B mag—MeasuredmagnitudeofminimuminJohnsonBband.
Min R Mag—MeasuredmagnitudeofminimuminCousinsRband.
Notes—Anyusefulinformationontheprecedingdata.
Next come four columns recording measured linear light elements. In rowswhere these occur, they are derived from linear regressions on the minimameasurements recorded in that row and the preceding rows for that system.Thusitispossibletocountthenumberofdatapointsintheregression;andforagivensystemlaterrowswillhaveregressionsbasedonmoredatapoints.
E0—The HJD zero epoch for the elements, to the same number ofsignificantdecimalplacesastheerror(nextcolumn).Thiswillbeclosetooneofthemeasuredminimainthisorprecedingrows.
E0 error—The measured uncertainty in E0, to one or two significantfigures.
The VSS EB database is a downloadable file in CSV text-only formatcontainingdataonsoutherneclipsingbinariesobtainedinthreeVSSprojects.Thedata,updatedmonthly,consistofmeasuredtimesofminima,lightelements,andspectra.Atpresentitcontainsdataon91systems.
Holmberg, JAAVSO Volume 40, 2012986
A Note on the Variability of V538 Cassiopeiae
Gustav HolmbergKarl XI-gatan 8A, SE-222 20 Lund, Sweden; [email protected]
Received December 28, 2011; revised February 9, 2012; accepted February 13, 2012
Abstract CCD observations of V538 Cas have been made on nine nightsduring three weeks using the AAVSO Bright Star Monitor. No significantvariationswerefound.
1. Discovery
Weber (1958a) discovered variations in the star BD +60˚ 201 (HD 7681, HIP6084,R.A.01h18m07.2s Dec. +61˚ 43' 04" (J2000)) by analyzing plates takenbetween1942 and1958.The starwas found tovarybetween9.0 and9.6 photographic magnitude and Weber, who did not publish a light curve,suspecteditcouldbeaneclipsingvariableoftheAlgoltype.AfurtherstudyofthestarwasmadebyHäussler(1974)whoused135patrolcameraplatestofindthestarvaryingirregularlybetween9.44and10.01photographicmagnitude.Heclassifieditasanirregularvariablestar,typeIsb. Onthebasisofthesestudies,thestarwasdesignatedV538Cas.Theentryin the General Catalogue of Variable Stars (GCVS; Samus et al. 2011) ofthestarhastypeIsbandgivesthemagnituderangeas9.4–10.6photographicmagnitude,andspectral typeK5III.Laterobservationshavere-classified thespectrumasM0III(Henryet al.2000).
2. New observations
V538 Cas is, at 7.7V, a bright star. It is thus a suitable object for theAAVSOBrightStarMonitor (BSM;AAVSO2009).This instrumentaimsatfillinganicheincurrentphotometrictelescopeecologywherenotmanyCCD-equippedtelescopesareoperating:thereareinterestingstarsthataredifficulttoobserveusingmostmoderntelescopesbecausethestarsaretoobrightandrisksaturatingtheCCDdetector.Findingsuitablecomparisonstarsforbrightstarsintelescopeswithsmallfieldsofviewcanalsobeachallenge.Severalof thecurrentphotometric surveys, suchasNSVSandASAS,onlymeasurestarsfainterthanabouteighthmagnitude,givingmuchroomforworkforaninstrumentsuchastheBrightStarMonitor.TheBSMisa6-cmf /6.2refractorwithaSBIGST-8XMECCDwitha fieldofviewof84'×127',operated forAAVSObyTomKrajciattheAstrokolkhozfacilityinNewMexico. 30-second (V)and60-second (B)exposuresofV538Caswereobtained
Holmberg, JAAVSO Volume 40, 2012 987
withtheBSMonninenightsduringthreeweeksinNovemberandDecember2011.Theimageswereanalyzedusingvphot(AAVSO2011)andthebrightnessofthestarmeasuredrelativetoanensembleoffivestars.Theresults(Table1)showverysmallornosignificantvariationsduringthistimeinterval. These observations are in line with other measurements from the post-photographicera.AphotometricstudybyHenryet al.foundslightshort-timescalevariationsontheorderofahundredthofmagnitude(Henryet al.2000).Hipparcos(Perrymanet al.1997)consistentlymeasuredV538Casaround7.75withvariationsontheorderofhundredthsofamagnitude.TASS,TheAmateurSkySurvey (2012),observed the staron fouroccasionsduring threeweeks,findingaconstantbrightness,alsointhevicinityof7.75.Togetherwiththenewobservationsreportedhere,thisleadsthepresentauthortoconcludethatV538Cas,atleasttoday,doesnotshowthetypeofvariationsonceattributedtoit.
3. Discussion
Althoughnottheprimarypurposeofthisshortcommunication,anattemptwillalsobemadetotrytoresolvethediscrepancybetweendatashowingrapidvariationsontheorderof0.6magnitudeormore,anddatashowingnoorverysmallvariationsinthestarknownasV538Cas. Onepossibilityisofcoursethatthisstarhaschangeditsbehaviorsincethemid-20thcentury.AnotheristhatWeber’sinitialguess,thatthisisanEAstar,iscorrectandthatfurthereclipseshavenotbeencoveredintheobservationsmadewiththeBSM,Hipparcos,TASS,andbyHenryet al.Butcantherebeanotherexplanation? Asimilardiscrepancyfound in the literaturemayprovideaclue.Weber,also in1958,publishedhisdiscoveryofaCepheid in theopenclusterNGC7789,varyingbetweenphotographicmagnitude11.2and12.2.Thediscoverywas confirmed by another observer, Romano (Weber 1958b). Cepheids inopen clusters have great astrophysical importance, andWeber’s finding wastherefore followed up. Burbidge and Sandage found no variations at all inthe star in their photographic photometry of the cluster using the 100-inchHooker telescope (Burbidge and Sandage 1958). Furthermore, Starrfield’sphotoelectricmonitoringofthestarwitha24-inchreflectoratLickduringthreehourspernightduringthreenightsfoundnovariationsinthestar.Photographicphotometryonaseriesofplatestakenwiththe20-inchCarnegieastrographatLickgaveasimilarresult:constantbrightness(Starrfield1965).MeasurementsonplatesfromtheHarvardCollegeObservatoryplatearchivesgaveasimilarresult(Janes1977). Thus,wehavetwocases inwhichWeberfoundvariationsofquite largeamplitude,bothofwhichwerefirstconfirmedbyanotherobserverusingsimilartypeofcameras(Weberusedsmall-scalephotographiccameras)thatwasnotconfirmedbylaterobservers.Starrfield,intryingtoresolvethisdiscrepancy,
Holmberg, JAAVSO Volume 40, 2012988
pointedoutthatapossiblesolutioncouldbethesmallplatescaleofWeber’scamera; thesuspectedvariablestarwasthereforeimperfectlyseparatedfromanother star, and seeing variations could produce a spurious impression ofvariability(Starrfield1965).PerhapsthisorsomeotherphotographiceffectcanaccountforthevariationsfoundinV538CasbyWeberandHäussler.
4. Acknowledgements
The author wishes to thank AAVSO Director Arne A. Henden for hisgenerous support.This research has made use of the simbad database (CDS2007),operatedatCDS,Strasbourg,France.
theHipparcosScienceTeam1997,The Hipparcos and Tycho Catalogues,ESA SP-1200, ESA Publications division, Noordwiijk, The Netherlands(Hipparcos data on V538 Cas accessed using http://www.rssd.esa.int/hipparcos_scripts/HIPcatalogueSearch.pl?hipepId=6084).
Samus,N.N.,et al.2011,General Catalogue of Variable Stars,onlineversion(http://www.sai.msu.su/gcvs/gcvs/index.htm), Sternberg Astron. Inst.,Moscow.
Table 1. Bright Star Monitor measurements of V538 Cas in November–December2011. JD Magnitude filter
2455881.85384 9.422 B 2455881.85444 7.696 V 2455882.80720 9.433 B 2455882.80782 7.715 V 2455888.78655 9.422 B 2455888.78714 7.713 V 2455889.68596 9.455 B 2455889.68656 7.699 V 2455892.65948 9.429 B 2455892.66008 7.715 V 2455893.65671 9.425 B 2455893.65730 7.718 V 2455894.67079 9.435 B 2455894.67139 7.702 V 2455895.65212 9.404 B 2455895.65272 7.697 V 2455896.64828 9.399 B 2455896.64888 7.706 V
Boyd, JAAVSO Volume 40, 2012990
A Practical Approach to Transforming Magnitudes onto a Standard Photometric System
David Boyd5 Silver Lane, West Challow, Wantage, OX12 9TX, UK; [email protected]
Received December 30, 2011; revised March 15, 2012; accepted April 5, 2012
Abstract We describe a practical implementation, convenient for amateuruse, of a method of transforming instrumental magnitudes onto a standardphotometricsystemindifferentialCCDphotometry.
1. Introduction
MostamateurastronomersusingaCCDcameratomeasurethebrightnessofanobjectofinterest,beitavariablestar,anasteroid,orsomeotherdistantlight source, employ a procedure called differential photometry. In this themagnitudeof the targetobject is foundbycomparing itsbrightness tootherstarsinthesamefieldofviewwhosemagnitudesareknown.Themeasureofanobject’sbrightness foundby integrating the imageof theobject recordedby the CCD is called its instrumental magnitude. Increasingly amateurs arebeingencouraged toobserveusingphotometric filtersand to transformtheirmeasured instrumental magnitudes onto one of the standard photometricmagnitudesystems. Transforminginstrumentalmagnitudesinthiswaygreatlyincreasestheirusefulness for scientific analysis. They can be combined with similarly-transformed measurements from other observers into a single internally-consistentdataset.ThemostcommonstandardphotometricsysteminusebyamateurstodayistheJohnson-CousinsUBVRIsystemanditisthiswhichweshallusehere,inparticularitsBVRI subset.However,theapproachdescribedisapplicabletoanystandardsystem. There are two main reasons why instrumental magnitudes differ fromstandard magnitudes: atmospheric extinction and a mismatch between theresponseoftheequipmentinuseandthestandardsystem.Byobservingstarswhose magnitudes are very accurately known (“standard” stars), correctionscanbefoundwhichenableinstrumentalmagnitudestobetransformedontothestandardsystem.Overtheyearsseveralformulationsofthesetransformationshavebeendevisedandpublished,usuallywiththeexpectationthatphotometricskieswillbe readilyavailable for theiruse.Wedescribehereavariationonpreviously published methods which amateurs may find better suited toconditionstheyarelikelytoexperience. Firstweexplainhowatmosphericextinctionandinstrumentalcharacteristicsaffectobservationsandbrieflydiscussthescopeofapplicabilityoftheproposed
Boyd, JAAVSO Volume 40, 2012 991
approach.Nextwedescribethetransformationequationsanddiscusschoosingstandard stars. We then explain how to find the transformation parametersandusethemtotransforminstrumentalmagnitudesandcolorindices.Finallywe use measurements of Landolt standard fields to illustrate the benefits oftransforminginstrumentalmagnitudes.
2. Atmospheric extinction and instrumental transformation
Atmosphericextinctionhastwoprincipalcauses:scatteringofincidentlightbyaerosolsincludingdustandwatervapor,andmolecularRayleighscattering.Aerosolscattering is relativelyuniformacross thespectrum,risingslowlyatshorterwavelengths,and is thedominantcauseofextinctionatwavelengthslongerthanabout500nm.Rayleighscatteringincreasesveryrapidlyatshorterwavelengths anddominates at theblue endof the spectrum.The amount ofaerosol scattering can vary widely depending on the aerosol content of theatmosphere,whereasRayleighscatteringremainsrelativelyconstant.Furtherinformationabout atmospheric extinctioncanbe found in (Green1992) and(Stubbset al.2007). Wavelength-independentscatteringisrepresentedbyafirstorderextinctioncoefficientmultipliedbytheairmassinthedirectionofthestar,andwavelength-dependentscatteringbytheproductofasecondorderextinctioncoefficient,theairmass,andthecolorindexofthestar.Sincefirstorderextinctionisprimarilycausedbyaerosolscatteringitcanpotentiallychangerapidlyduringasinglenightorfromnighttonightasthedustormoisturecontentoftheatmospherechanges.SecondorderextinctionismainlyduetoRayleighscatteringandisgenerallyconsideredtobestableovertimeatagivenlocation.BecauseoftherapidriseinRayleighscatteringattheblueendofthespectrum,secondorderextinctionshouldprimarilyaffectmeasurementsmadethroughaBfilterwhilemeasurementsthroughV,R,andIfilterswillbemuchlessaffected. Thepresenceofa thinlayerofcloudorhazereducestheincominglightequally at all wavelengths, as verified experimentally (Honeycutt 1971). Inpracticethisaddsaconstant“gray”termtoatmosphericextinction,whichhastheeffectofchangingtheimagezeropoint.Thincloudisoftendifficulttodetectvisuallybutitspresenceneednotpreventgoodqualityphotometryprovideditisuniformandstable(Hardie1959).Nevertheless,themostreliableresultswillalwaysbeobtainedunderclearskies. InpracticeallBVRIfilterandCCDdetectorcombinationsdifferintheirspectral response from the standard Johnson-Cousins system. This meansthattheywillproduceresultswhichdifferfromthestandardsystemandthisdifferencewillvarywith thecolorof thestar.Since themismatch isusuallysmall,itisnormallyassumedthatitcanbecorrectedbyalineartermcomprisingtheproductofaninstrumentaltransformationcoefficientandthecolorindexofthestar.Providedtheinstrumentalcomponentsdonotchange,anychangein
Boyd, JAAVSO Volume 40, 2012992
the transformation coefficient, for instance due to long term changes in thetransmission properties of the optical components and filters, is likely to besmallandslow. So,whilefirstorderextinctionmaychangefromnighttonight,andevenduringasinglenightasatmosphericconditionschange,anychangesinsecondorderextinctionatthesamelocationandinstrumentaltransformationwiththesameequipmentarelikelytohappenonlyslowlyovertimeasnotedinWelch(1979)andHarriset al.(1981).Itwouldneverthelessbeprudenttoremeasurethematregularintervalstomonitorforanysuchchange.
3. Applicability
The approach to obtaining transformation parameters described here isaimedatthoseusingCCDcamerastoimagetheskythroughsufficientlylongfocal length optical equipment that the field of view is small, typically lessthan one degree across. In photometry it is normally good practice to workaboveanaltitudeofabout30degrees,corresponding toanairmassof2, toavoidtheworsteffectsofatmosphericextinction.Underthesecircumstanceswecanassumethatallstarsbeingmeasuredinthesameimageareatthesamezenithdistanceandhencearemeasuredthroughthesameairmass.Theerrorintroducedbythisassumptionis,intheworstcase,onlyafewthousandthsofamagnitude.Itshouldbenotedhoweverthatundersomecircumstances,forexamplewhenimaginglargefieldsusingDSLRcamerasorimagingclosetothehorizonwhereairmassincreasesrapidlywithreducingaltitude,itmaynotbevalidtoassumethatallstarsintheimagehavethesameairmassandamorerigorousanalysismustbecarriedout.
whereBandVarethestandardB-andV-bandmagnitudesofthestar,bandvaretheinstrumentalmagnitudesofthestarmeasuredinBandVfilters,k'bandk'varetheB-andV-bandfirstorderextinctioncoefficients,k"bbvandk"vbvaretheB-andV-bandsecondorderextinctioncoefficientsforthe(B–V)colorindex,Xistheairmassofthestar,TbbvandTvbvaretheB-andV-bandinstrumentaltransformationcoefficients for the (B–V)color index, andZbandZvare the
Boyd, JAAVSO Volume 40, 2012 993
B-andV-bandimagezeropointswhichareofcoursethesameforallstarsinanimage. Weadopttheconventionthatthefirst,oronly,letterinasubscriptindicatesthefilterpassband,whilethesecondandthirdlettersindicatetherelevantcolorindex. As we noted earlier, rapid atmospheric changes may cause first orderextinction coefficients and image zero points to vary from image to image,whereassecondorderextinctionandinstrumentaltransformationcoefficientsshouldremainconstantorchangeonlyslowlyovertime. The first order extinction, second order extinction, and instrumentaltransformationcoefficientshaveusuallybeendeterminedbyseparateprocesses.ThetwocommonmethodsoffindingtheextinctioncoefficientsaretheBouguerandHardiemethods.TheBouguermethodinvolvesfollowingaclosegroupofstandardstarsastheymovethroughawiderangeofairmassoverthecourseofanight.TheHardiemethodrequirescloselytimedobservationsofpairsofredandbluestandardstarsatwidelyseparatedairmasses.Bothmethodsdependonatmosphericextinctionremainingconstantoverthewiderangesoftimeand/ortheskyinvolvedinmakingthesemeasurements.Instrumentaltransformationcoefficientsarefoundbyobservingacloselyspacedgroupofstarswithawiderangeofcolorsatlowairmass.Descriptionsofthesemethodscanbefoundin,forexampleWelch(1979)andRomanishin(2002). However,aspointedoutin(Harriset al.1981),determiningtheseparametersseparately in thisway is sub-optimalandusually requires iteration toobtainthebestsolution.Theyarguethat it isbetter tomakeuseofallobservationstogether to determine the required parameters. Our approach is a simplifiedversionofthatdescribedbyHarriset al.whichmeetstheneedsofdifferentialCCDphotometryandmaybeeasierforamateurstoimplementinpractice.Ithasthemeritofreachingasolutioninstageswithgraphicalverificationateachstageratherthana“black-box”multilinearleastsquaresapproach.Thismakesiteasiertospotproblemswithindividualstars,internalinconsistencieswithinthedata,orhumanerror. Wecanrearrangeequations(1)and(2)asfollows:
imagezeropointandfirstorderextinctioncorrection,andZ'v=Zv–k'vXisasimilartermfortheV-band. SinceweareworkingwithsmallfieldsandcanassumeallstarsintheimagebeingmeasuredhavethesamevalueofairmassX,thetermsCbbv,Cvbv,Z'b,andZ'varethesameforallstarsintheimage.Ifweplot(B–b)against(B–V)forallstarsinanimage,thegradientwillgiveusCbbvatthevalueofXforthatimage,andsimilarlyplotting(V–v)against(B–V)givesCvbv. Similar equations to (5) and (6) relatemagnitudesmeasuredusingotherfiltersandcolorindices.Forexample:
(R–r)=Crvr(V–R)+Z'r (7)
(I–i)=Civi(V–I)+Z'i (8)
wheredefinitionsof the termsCrvr,Civi,Z'r, andZ'i followbyanalogywiththoseabove. Hence, if we can measure values of Cbbv, Cvbv, Crvr, and Civi (henceforthdescribedcollectivelyas theCparameterswhere,generically,C=T–k''X),eachwiththecorrespondingvalueofX,thenwecandeterminetheinstrumentaltransformationsTandsecondorderextinctioncoefficientsk''.Asweshallsee,thesearetheparametersrequiredtotransformdifferentialCCDphotometry.
5. Sources of standard stellar magnitudes
Inordertomeasuretheseparameters,weneedtoidentifygroupsofstarswhich(a)willfitwithinasmallCCDfieldofview,(b)containasmanystarsaspossiblewithaccuratelyknownmagnitudesineachofthefilterpassbands,and(c)spanaslargearangeofcolorindexaspossible. Traditionally, the gold standard for calibration is the set of equatorialstandardstarsmeasuredovermanyyearsbyArloLandolt,seeLandolt(1992,2009,2011).Thesehavearootmeansquare(rms)Vmagnitudeuncertaintyofabout0.004mag.WhileLandolt standard starshave the advantageofbeingvisiblefromalllatitudes,forobserversfarfromtheterrestrialequatortheydonotriseveryhighintheskyandsotraversearelativelysmallrangeofairmass.Using only those standard stars would necessitate extrapolating to air massvaluesoutsidethisrange,aprocedurewhichwouldinevitablyreduceaccuracy.UsingonlyLandoltstarsthereforelimitstheaccuracyattainablewhentryingtoaccountforairmassdependenteffects.AnotherslightdrawbackofLandoltstarsisthatitissometimesdifficulttoincludemanyofthemwithawiderangeofcolorindexwithinatypicalCCDfield. For this reasonwe investigatedusinggroupsofsecondarystandardstarswhichhavethemselvesbeencalibratedusingLandoltstars,inparticularoneswhichculminatenearthezenithathigherlatitudesandcanthereforebeobservedatornearanairmassof1.SuitablefieldsofstarshavebeenmeasuredintheB,
6. Finding the instrumental transformation and second order extinction coefficients
First we select a suitable field containing stars with known, accurately-measured B, V, R, and I magnitudes covering as wide a range of colors aspossible,suchastheoneshowninFigure1.Weidentifythestarsinthefieldweintendtouseasoursecondarystandards.Wethenwaitforaclearnightwithstableskyconditionswhenthefieldiswellplacedandtakebetweenfiveandtenimagesofthefieldthrougheachfilter.Wetakecaretorecordthebrighteststarswithashighsignaltonoiseaspossibleineachfilterwhileavoidingnon-linearity or saturation of the CCD. This typically takes only a few minutesforeachfilterduringwhichwerequiretheskyconditionstoremainclear.Wecalibratetheimagesusingdarkandflatfieldsandusingaperturephotometrymeasure the instrumental magnitudes b, v, r, and i for each of our standardstarsineachimagetakenwiththecorrespondingfilter.Wethencalculatemeanvaluesandstandarddeviationsofb,v,r,andiforeachstaroveralltheimagestakenwitheachfilter.WealsocalculateameanvalueofXofallthestarsinalltheimagestakenwitheachfilter. Tocheckthatskyconditionshaveindeedremainedstablethroughouteachsetofimages,wecancomparestandarddeviationsoftheinstrumentalmagnitudesofeachstarineachfilterwiththeestimatedmeasurementuncertaintiesoutputbythephotometrysoftware.Iftheformeraresubstantiallylargerthanthelatter,
Boyd, JAAVSO Volume 40, 2012996
it isan indication thatskyconditionswereprobablyunstableand theresultsshouldnotbeusedforcalibrationpurposes. Weknowvaluesof(B–V),(V–R),and(V–I)andcancalculatevaluesof(B–b),(V–v),(R–r),and(I–i)foreachstandardstar.Wecanalsocalculatetheuncertaintieson(B–b),andsoon,foreachstarbyaddingthequoteduncertaintyon its standard magnitude B and the standard deviation of its instrumentalmagnitudebinquadrature. We plot (B–b) vs (B–V), (V–v) vs (B–V), (R–r) vs (V–R), and (I–i) vs(V–I).Assumingthetransformationsweareseekingarelinear,weexpectthedatapointstolieapproximatelyonastraightline.If thelineishorizontal, itindicatesthatourmeasuredinstrumentalmagnitudehasnocolordependency.If,asismorelikely,thelineisatanangletothehorizontalitmeansthatthereisacolordependencyinourobservationswhichweneedtocorrect. Figure2showsmagnitude-colorplotsof(B–b)vs(B–V),(V-v)vs(B–V),(R–r)vs(V–R),and(I–i)vs(V–I)forstarsinthefieldofthevariableEECep.Theequipmentusedwasa0-35mSCT,SXVR-H9CCDcamera,andAstrodondichroicB,V,R,andIfilters. Fromequations(5) to(8) thegradientsofstraight linesfittedto thedatapointsintheseplotsgivevaluesoftheCparametersatthemeanairmassXcalculated for each filter, and we can calculate their uncertainties from thescatterinthedatapointsaboutthefittedlines. Werepeatthisprocessforseveralmorefieldsofsecondarystandardstarssatisfyingour selectioncriteria andcoveringaswidea rangeof airmass aspossible.EachfieldprovidesvaluesoftheCparametersandacorrespondingvalueofXforeachfilter.Thedatadonotallhavetobecollectedonthesamenightsince,asnotedearlier,theseparametersshouldbestableovertimesodatafromseveralnightscanbecombined. SinceforeachoftheCparameters,C=T–k''X,ifwenowplotthevalueoftheCparameterforeachfilteragainstthecorrespondingvalueofXandmakeweightedlinearfits,wecanobtainvaluesfortheinstrumentaltransformation(T)andsecondorderextinction(k'')coefficientsforeachfilterwithestimatesoftheiruncertainties.InthesefitsweweighteachvalueofCwiththeinversesquareofitsuncertainty. Figure3showsplotsoftheCparametersvsXobtainedfromtwelvesetsofobservationsoffourfieldsimagedovertwonights.AlsoshownintheseplotsarethestraightlinesrepresentingthefitsforTandk''. Table1liststhevaluesoftheinstrumentaltransformationandsecondorderextinctioncoefficientsfoundfromtheseanalyses.Rememberourconventionisthatthefirstletterinasubscriptindicatesthefilterpassband,whilethesecondandthirdlettersindicatetherelevantcolorindex. ThesecondorderextinctioncoefficientsfortheV,R,andIfilterpassbandsare small and consistent with zero within experimental uncertainty. This isas expected given that extinction in these filters is primarily due to aerosol
Boyd, JAAVSO Volume 40, 2012 997
scattering which has minimal wavelength dependency. Therefore, consistentwith conventional wisdom, we will assume these have a value of zero.ThesecondorderextinctioncoefficientfortheBfilterissignificantlynon-zeroasexpectedfromtherapidriseinRayleighscatteringatshorterwavelengths. By observing several fields of secondary standard stars covering a widerangeofcolorandairmass,slightdifferencesbetweentheassumed“standard”magnitudes for individual stars and their “true” values are averaged out. Inthiswaywecanoffsettheloweraccuracyinthemagnitudesofoursecondarystandardstarscomparedtoprimarystandards.
7. First order extinction coefficients
Valuesofthefirstorderextinctioncoefficientsk'arenotcalculateddirectlyandarenotneededfordifferentialCCDphotometrywithsmall fieldswhereweassumeXisconstantoverthefield.TheyarecontainedintheimagezeropointsZ'=Z–k'Xineqns(5)to(8).ValuesofZ'canbefoundfrominterceptsofthelinearfitsinFigure2atcolorindex0butingeneralthesewillnotyieldconsistent values of Z and k' since these parameters vary with changes inatmospheric transparency. Nevertheless, if conditions are sufficiently stableduringonenight,wemightexpectZandk'toremainapproximatelyconstant.Inthatcase,ifweplotthevaluesofZ'foreachfilteragainstthecorrespondingvaluesofXfordataobtainedonthatnight,weshouldgetastraightlinewhosegradientgivesthemeanvalueofk'fortheappropriatefilteronthatnight(seeFigure4).Thecorrespondingvaluesofk'arelistedinTable2.Asexpected,k'graduallydiminishesas thewavelength increases, reflecting the reduction inaerosolscatteringatlongerwavelengths.
8. Transforming instrumental magnitudes to standard magnitudes
NowwehavealltheparametersneededtotransforminstrumentalmagnitudesmeasuredindifferentialCCDphotometryontotheBVRIstandardmagnitudesystem. Indoing this it is easier toworkwithequationswhichcontainonlyinstrumentalratherthanstandardmagnitudesontherighthandside.Bysimplemanipulationofeqns(5)to(8)wederivetheequationswerequire.
andthevarious“z”termsontherighthandsideareimagezeropointswhicharethesameforallstarsinanimage. Since,generically,C=T–k''XandthevaluesofTandk''areknown(seeTable1),ifweknowtheairmassXofanimagewecancalculatethevaluesoftheappropriateCparametersandhencetheC'parameters. Supposewewant to find the standardVmagnitudeof avariable star inafieldcontainingseveralcomparisonstarswithknownmagnitudes.WetakeseveralimagesofthefieldthroughBandVfiltersandmeasuretheinstrumentalmagnitudesbandvofthevariableandcomparisonstarsineachimage.KnowingthemeanairmassXofthestarsineachimagewecalculateCbbv=Tbbv–k''bbvXand Cvbv =Tvbv – k''vbvX and hence C'vbv from equation (14) for each image.Usingequation(10)andknowingthestandardandinstrumentalmagnitudesforthecomparisonstars,wecandeterminethezeropointzvforeachimage.Sinceweknowthe instrumentalbandvmagnitudesof thevariable,wecanagainuseequation(10)tocalculateitsstandardVmagnitudeineachimage.Finally,usingtheseindividualmeasurementsoftheVmagnitudeofthevariable,wecancomputeitsmeanandstandarddeviation. AsimilarprocedurewillyieldvaluesforB,R,andIcalculatedfromthemeasuredinstrumentalmagnitudesb,v,r,andiusingeqns(9),(11),and(12).
11. Transforming magnitudes measured for Landolt standard fields
Asanapplicationofthisapproach,BVR-filteredinstrumentalmagnitudeswere measured for stars in three Landolt standard fields and transformedas described above. The rms residuals between the standard and derivedmagnitudesbeforeandaftertransformationareshowninTable3.Theseresultsclearlyshowtheimprovementachievedbytransformingmagnitudesontothestandardsystem.
12. Conclusion
Wehavedescribedanddemonstratedapracticalapproachtofindingandapplying the transformations required tobring instrumentalmagnitudesontoastandardphotometricsystemindifferentialCCDphotometry.Thisincludescorrecting for second order atmospheric extinction where appropriate. It ispossible to use well-measured secondary standard stars in fields spanning awiderangeofdeclinationstoenablefullcoverageoftheairmassrangefromobservingsitesfarfromtheterrestrialequator.Skyconditionsmustremainclearandstableforlongenoughtoobtainshortseriesoffilteredimagesofthefieldsrequiredtocovertherequiredrangeofairmass.Theseimagescanbeobtainedonseveralnightsand theresultscombined.Thisapproachmaybeeasier forsomeobservers,particularlythoseoperatingathighterrestriallatitudes,tousethanotherapproaches.
13. Acknowledgements
IamgratefultoDr.ChrisLloydandDr.RichardMilesforhelpfulsuggestionsinpreparingthispaperandtoBrianWarnerforstimulatingmythoughtsonthissubject. Constructive comments from an anonymous referee have helped toclarifyandimprovethepaper.
Table3.rmsresidualsbetweenstandardandderivedB,V,andRmagnitudesbeforeandaftertransformationforstarsinthreeLandoltstandardfields. Field Air mass rms residuals between standard and derived magnitudes
Presented at the 101st Spring Meeting of the AAVSO, Big Bear Lake, CA, May 22–24, 2012
Received August 23, 2012; revised October 1, 2012; accepted October 2, 2012
Abstract
The author discusses his new remote observatory under pristine skies andthe intensive observations of variable stars he is accomplishing. The starsunderinvestigationaremainlycataclysmicvariables,observedinresponsetoAAVSO,CBA,andVSNETalerts;othertypes,suchasRRLyraestars,werealso observed. Examples are presented of dense observations of differentcataclysmicvariablesaswellasanRRLyraestar.Featuredisthefirstbrightoutburst of SVAri (NovaAri1905) since its discovery, as well as the firstoutburstofUGWZcandidateBWScl.ResultsforVWHyi,anothercataclysmicvariable,willalsobeshown.Furthermore,anintensivelyobservedRRLyraestarwillbehighlighted.
1. Introduction
Itisanamateurastronomer’sdreamtoobserveunderpristinedarkandclearskiesnearlyeverynight likeat thesiteswhere theprofessionalastronomicalobservatoriesarelocated.Suchadreamnormallynevercomestrue.However,modern techniques and infrastructures in most countries make it possiblenowadays to observe from remote sites using off-the-shelf technology. TheauthorinstalledaremoteobservatoryunderthedarkskiesoftheAtacamaDesertclosetothetownofSanPedrodeAtacama,Chile.ThetelescopeishousedatSPACE (San Pedro deAtacama Celestial Exploration (http://www.spaceobs.com/index.html)).Theowner,alsoanamateurastronomer,formerlyworkedattheEuropeanSouthernObservatory(ESO)atthebigtelescopesitesinChile.In2003hestartedSPACE,whichhasbeenextendedtotelescopehostingforthelastcoupleofyears.Icontactedhimin2009anddecidedtoestablishmyobservatoryathisplace. Unfortunately,deliveryofthetelescopetookmuchlongerthananticipated,andonly in July2011was I able to install thedome,mount, and telescope.SinceAugust1,2011,theremoteobservatoryhasbeenproducingdataeveryclearnight.Sofar,inlessthan8.5monthsofoperation,thisamountedtoabout220data-takingnights.Notbad,isitnot?
Hambsch, JAAVSO Volume 40, 20121004
IamcollaboratingontheobservationofcataclysmicvariablestarswithJ.PattersonoftheCenterforBackyardAstrophysics(CBA)andT.KatooftheVSNETemailalerts.SinceIamalsointerestedinobservingRRLyraestars,IcollaboratewiththeGroupeEuropéend’ObservationsStellaires(GEOS).InBelgium,IamamemberoftheWerkgroepVeranderlijkeSterren(WVS)onHighAmplitudeDeltaScuti(HADS)stars.IamalsomemberoftheAAVSOand the German Bundesdeutsche Arbeitsgemeinschaft für VeränderlicheSterne(BAV). During themanyclearnights Igathera lotofdataonmanystars.Sincespace here is limited, I restrict myself to some highlights: the outbursts ofSVAri,BWScl,andVWHyi;andIalsolookatanRRLyrae-typestarwithastrongBlazhkoeffect.
2. Observatory
The remote observatory in Chile houses a 40-cm f /6.8 Optimized Dall-Kirkham(ODK)fromOrionOptics,England.TheCCDcameraisfromFingerLakesInstruments(FLI)andcontainsaKodak16803CCDchipwith4k×4kpixelsof9mmsize.ThefilterwheelisalsofromFLIandcontainsphotometricBVIfiltersfromAstrodon. Figure1showsanimageoftheremotetelescopeinChile.Itishousedinaclamshelldome,makingeasymovementof the telescopepossiblewithouttheneedtofollowwithashutterofanormaldome.Imagesofanight’ssessionare either acquiredwith acp or ccdcommander automation software. Furtheranalysisintermsofdeterminationofthebrightnessofthestarsisdoneusinga program developed by P. de Ponthierre (2010). The data are then finallysubmittedtotheAAVSO.3. SV Arietis (Nova Arietis 1905)
As a first example, I show the results of the campaign on SV Arietis(NovaAri1905).SVAriwasdiscoveredonNovember6,1905,byM.andG.Wolf(1905)inHeidelberg,Germany,ataphotographicmagnitudeof12.0.Itwasreportedthatithadbrightenedfrommagnitude22.1.HimpelandJantsch(1943)reportedapossiblesightinginSeptember1943,atamagnitudeof15.7,butthiswasnotconfirmed.Nobrightnessincreasehasbeenobservedforthisstareversince. Then,on2011August2.788UT,R.StubbingsobservedthefieldofSVAriandsawanobjectatmagnitude15.0.Hesentanalertviathemailinglistcvnet-outbursttoaskforconfirmation.TheoutburstinformationwasalsogivenviatheVSNETmailinglist.ThisleadtotheconfirmationbyG.Masi(2011)andR.Fidrich(2011).IsawthosealertsviaVSNET,andsinceG.Masiimmediatelytookatimeseriesandobservedsuperhumps,Ialsodecidedtogoafterthisstar
Hambsch, JAAVSO Volume 40, 2012 1005
andstarteda timeseriesatAugust2.844UT,just1.34hoursafter theinitialdiscovery.AfirstanalysisofthedataofG.MasibytheKyototeam(Ohshimaet al.2011)showedthatthestarisprobablyanSUUMatypedwarfnovawithasuperhumpperiodofabout1.54hours.Thepresentoutburstseemsalsonotasbrightastheoneduringthediscoveryofmagnitude12.However,anothermailfromT.Kato(2011a)onVSNETreportedthatprobablytheoriginalbrightnessestimatewastoooptimisticasitseemsthatmanybrightnessdeterminationsofM.Wolfwereabout2magnitudestoobright. Figure2showstheobservationsImadeofthisstaroveraperiodofmorethantwoweeks.Duringthisperiodthestardroppedmorethanonemagnitudeinbrightness.Figure2alsoclearlyshowsindividualnightlybrightnessvariationsofabout0.3magnitude. Figure 3 shows the time series observation made during the first night.Clearly a superhump of 0.3 magnitude is visible. The magnitude of thesuperhumpsreducedtoabout0.15magnitudeafteracoupleofdays. BasedontheanalysisbytheKyototeam,therehavebeendistinctstagesintheevolutionofsuperhumpsinSVAri.ThemeanperiodbeforeAugust4was0.05574(18)daywhichsincethenshortenedto0.05519(5)day.Also,laterdatashowedthatSVAriwasgraduallydeclining(~0.05mag./day).ThisseemedveryslowforanordinarySUUMatypedwarfnova.Thus,thisobjectmightbeaWZSge-typedwarfnova.Thisinformationisbasedone-mailexchangesviaVSNET.Towards the end ofAugust 2011, SVAri had dimmed towardsmagnitude18.
4. BW Sculptoris
AnotherexampleofanintensivelyfollowedstarisBWSculptoris,whichwentintoitsfirst-everobservedoutburston2011October21.BWSclisalsoacataclysmicvariablestar.OnthesamedaytheAAVSOSpecial Notice #261(AAVSO 2011a) was published mentioning the outburst of BW Scl. It wasvisually observed by M. Linnolt on October 21.3146 at a magnitude of 9.6(visual).TheoutburstwasconfirmedbyA.Plummeratmagnitude9.4(visual).ThestarhasconflictingclassificationsintheliteratureandisprobablyaWZSge-typedwarfnova.OnOctober25, theAAVSO Alert Notice 449 (AAVSO2011b)wasissuedconcerningthisoutburst. Ihadalreadybeen follwing this star in itspre-outburstphase.However,Imissed theoutburst, as I thought the starwasnotdoingmuch,andceasedobserving it on October 14—just a week before the outburst took place. Ofcourse, I restarted observations immediately after the news was spread andfollowedthestaroveraperiodof2.5months.Figure4showsthedevelopmentofthebrightnessoverthefullobservingperiod. Afterafewdaysnicesuperhumpsofabout0.25magnitudedeveloped,ascanbeseen inFigure5.Afteraweek into theoutburst theearlysuperhump
At the request of professional astronomers from SouthAfrica (P.Woudt(2011)),IbeganobservationsofVWHyijustasthestarwentintosuperoutburst,althoughthatwasasurprisetothepros,asthesuperoutburstwasnotexpectedyet.Myobservationstriggeredsatelliteobservationsofthestar. VWHyiisapopularcataclysmicvariableintheSouthernsky.Manystudieshavebeenperformedon this star; see, for example,AAVSO (2010).Duringquiescencethestarisatmagnitude14.4,Normaloutburstshappenonaverageevery27.3daysandlastabout1.4days.Thesuperoutbursthappensonaverageevery179daysandlastsforabout12.6days.Figure6showsimpressivelythisbehaviorofVWHyiwithonesuperoutburstandtwonormaloutbursts. ThestardevelopsstrongsuperhumpsduringitsoutburstascanbeseeninFigure7.Thevariationofthosehumpsreaches0.5magnitude.
6. Example of an RR Lyr star, V1820 Ori
Thistypeofvariableisnamedaftertheprototype,thevariablestarRRLyraeintheconstellationLyra.RRLyrstarsarepulsatinghorizontalbranchstarswithamassofaboutone-halfofourSun’s.Theirperiodisshort,typicallylessthanoneday. MyinterestinobservingRRLyrstarsis,ontheonehand,duetotheshortperiodofthosestars—withinonenightyoucanseequiteachangeinbrightness.On the other hand, the stars also show some brightness modulations, knownas the Blazhko effect. Back in 1907 S. Blazhko observed this effect for thefirsttimeinthestarRWDra(seeSmith2004).TheBlazhkoeffectisnotwell-understood and needs further observational campaigns. Recently due to theKeplerandCoRoTsatellitemissions,more insight into thisphenomenonhasbeengainedasthesatellitescan,ofcourse,observethestarscontinuously,whichis impossible for Earth-bound observations. Nevertheless, observations fromEartharealsoveryvaluable,ascanbeseeninmanypublicationsonthissubjectintheastronomicalliterature.Figure8showsthephasediagramoftheRRLyrstarV1820Ori,whichhasbeenobservedfromChileoverafullseason(morethan3months).Itisobviousfromthefigurethatthelightcurveisnotregular,the maximum brightness is changing over more than 0.5 magnitude, and themomentofmaximumtimeischanging.Soarathercomplexlightcurveistheresult.Presentlythedataareunderanalysis,andIintendtopublishtheresults.
Hambsch, JAAVSO Volume 40, 2012 1007
7. Conclusion
The remote observatory under pristine skies in the Atacama Desert ofChileopensupgreatpossibilities toobservevariablestars.Intensivefollow-upobservationsovermanydays,weeks,orevenmonthsarepossibledue tothestableweatherconditions.Thegivenexamplesshowimpressivelywhatispossible.Iwelcomecollaborationsinordertocontributetoscientificresearchofcommoninterest.
References
AAVSO. 2010, Variable Star of the Month (http://www.aavso.org/vsots_vwhyi).
AAVSO.2011a,AAVSO Special Notice #261,AAVSO,Cambridge,MA.AAVSO.2011b,AAVSO Alert Notice 449,AAVSO,Cambridge,MA.dePonthierre,P.2010,LesvePhotometrysoftware(http://www.dppobservatory.
Aaron PriceAAVSO Headquarters, 49 Bay State Road, Cambridge, MA 02138;[email protected]
Kevin B. Paxson20219 Eden Pines, Spring, TX 77379; [email protected]
Received February 16, 2012; revised March 13, 2012; accepted March 13, 2012
Abstract In2011,theAAVSOconductedasurveyof615peoplewhoareorwererecentlyactiveintheorganization.Thesurveyincludedquestionsabouttheirdemographicbackgroundandvariablestarinterests.Dataaredescriptivelyanalyzedandcomparedwithpriorsurveys.Resultsshowanorganizationofveryhighlyeducated,largelymaleamateurandprofessionalastronomersdistributedacross108countries.Participants tend tobe loyal,with the average timeofinvolvement in theAAVSO reported as 14 years. Most major demographicfactorshavenotchangedmuchover time.However, theaverageageofnewmembers is increasing.Also, a significant portion of the respondents reportbeingstrictlyactiveinanon-observingcapacity,reflectingthegrowingmissionoftheorganization.Motivationsofparticipantsaremorealignedwithscientificcontributionthanwiththatreportedbyothercitizenscienceprojects.Thismayhelpexplainwhyathirdofallrespondentsareanauthororco-authorofapaperinanastronomical journal.Finally, there is someevidence thatparticipationintheAAVSOhasagreater impactontherespondents’viewof theirrole inastronomycompared to thatexpected through increasingamateurastronomyexperiencealone.
1. Introduction
TheAAVSO is a large,multinational citizen scienceorganizationdatingbackto1911.Theorganizationhasexperiencedsignificantchangeinthepasttwodecades(WilliamsandSaladyga2011),yetourlastsurveyofmembershipwasconductedin1994.Astheorganizationbeginsplanningforthefuture,itwastimetousedatatocharacterizethosewhoareactiveandcontributingtotheAAVSO,inallitsforms.Thesedatacanbecomparedwithcurrentassumptionsandbeliefsoftheorganizationandalsousedasatoolforplanningnewinitiativesanddirection.
2. Prior surveys
TheAAVSOhasconductedanumberofmembershipsurveysinthepast35 years. These surveys included a mix of demographic (such as “age”),
Price and Paxson, JAAVSO Volume 40, 2012 1011
programmatic(suchas“observingtrends”)andoperational(suchas“evaluationofstaff”)items. In 1976, AAVSO Director Janet Mattei, with staff members Linda M.Blizzard and Josefa M. Manella, mailed a survey to members. The surveyincludedsectionsonheadquartersoperations,communications,observations,meetings,andpublications.Approximately200–300responseswerereceivedandare stored in theAAVSOarchives (AAVSO1976).However,noknowntabulationorreportoftheresponsesisknowntoexist. InJanuaryof1980,theAAVSOmailedasecondsurveytomembers.Itwasasmallersurvey,withtenquestionsfocusedondemographicsandobservingactivities.AsummaryofresultswerereportedinWaagen(1980).267surveyswerereturnedasofthewritingofthatreport.Noneofthequestionsonthe1980surveywerealsoonthe1976survey,sothetwosurveyscouldbeseenascomplementary. In1994,WayneM.LowderdesignedasurveyofmembersandobserversinresponsetoarequestbyarecentlyconvenedFuturesStudiesGroup.Thesurveyhad 79 items divided into sections focused on demographics, publications,otherresourcesofinformation,astronomicalactivities,observing,headquartersoperations,meetings,anduseofpersonalcomputers.420surveyswerereturnedandtabulatedbyAAVSOstaff(TanjaFoulds,ShawnaHelleur,DennisMilon,and Barbara Silva). Results were presented to the Futures Studies Group inthe form of an executive summary written by Lowder in September 1994,whichexistsintheAAVSOarchives(Lowder1994).TheFuturesStudyGrouppresented results to theAAVSO Council at theAAVSOAnnual Meeting inOctober1994(Hazen1995). TheAAVSOhasalsorunafewsurveysover thepast fewyearsfocusedonmorespecific topics. In2010, theAAVSOconductedasurveywitheightquestions about AAVSO meeting experiences. This survey was distributedexclusively online through theAAVSO website and received 88 responses.Since 2009, theAAVSO’s Citizen Sky project has asked participants a fewoptionaldemographicquestionswhentheyfirstregisteredfortheCitizenSkywebsite.1,385oftheseresponseswereanalyzedinapaperbyPriceandLee(inpress)andsomeofthosedataareincludedhere.
3. Survey design and methodology
The goal of the 2011 survey was to better characterize the AAVSOmembershipsothatstaffandtheCouncilcanmakebetterdecisionsregardingmembershipactivitiesandfuturedirectionsoftheorganization,suchastestingcurrentassumptionsaboutmembersandalsolookingforunexpectedresultsinthedata.Assuch,thesurveyitemsweredesignedtoreportontherespondents’educationalandprofessionalbackgroundsandtheirexperienceintheAAVSOandastronomyingeneral.
Price and Paxson, JAAVSO Volume 40, 20121012
The survey was designed prior to the known existence of the previoussurveys.Yet the itemson thesurveysareoftenquitesimilar,whichwefeelisatestamenttothevalidityofthechosenitems.Twentyofthosewhowereprivately invited to take the survey tested the first draft. Only technicalchangescameoutofthatpilottest.Thesurveyhasamaximumof27items(seeAppendixA).However,someitemsareconditionedonlytoappearbasedoncertainresponsestoearlieritemsandallitemswereoptional.Sotheresponseratevariesitem-to-item. The survey was placed on the Survey Monkey website so that resultscouldbeautomaticallytabulated.Wepostedthesurvey’sURLtotheAAVSODiscussionGroup(~490subscribers)andontheAAVSOwebsite(417reads).We sent ane-mail to thosewhowerecurrentAAVSOmembersorwhohadmadeanobservationwithinthelastfiveyears(~2,400e-mails).Weidentifiedeightpeoplewhomet thatcriteriabutwhodidnothavee-mailaddressesonrecord.Forthem,weprintedacopyofthesurveyandsentitusingpostalmail.Atotalof691validresponseswerereceivedtotheonlinesurveyandfouroftheprintedsurveyswerereturned. We tabulated the results into an Excel spreadsheet. For the open-endeditems,wecoded them intoa setofcategories that included99%ormoreofthe responses.Whenaparticular itemresponsecould fit intomore thanonecategory,thefirstcategorymentionedintheresponsewasused.Thiswasbasedontheassumptionthatthefirstitemmentionedbytherespondentwasthemostimportant to them. The ranking items were scored on an ordinal scale (thehighest ranking item is assigned a “1” and the rest are ranked accordingly).Wetreateditemsthatwerenotansweredasmissingdata.WemadeourfinalanalysiswiththePASWStatistics18software(formerlySPSSStatistics,nowIBMSPSSStatistics).
3.1.Definitions For the purposes of this survey, we refer to respondents as anyone whoansweredatleastoneitemofthesurvey.Also,wecombinedCharge-CoupledDevice (CCD), Photoelectric Photometry (PEP), and Digital Single-LensReflex(DSLR)technologiesbeneaththeumbrellatermofdigital technologies.Visual observationsincludeanyobservationmadewiththeeye,whichincludesnaked eye, binoculars, and telescopic observations made with an eyepiece.Membership statuswasassignedbasedonAAVSOheadquartersrecordsasofJanuary2012. The profession categories included items were taken from the U.S.Department of Labor categories used in the 2011 U.S. Census (U.S. Dept.ofLabor2011).Thecategoriesofobjectswere taken from thehighest levelcategoriesusedbytheVariableStarIndex(VSX)(AAVSO2011),whicharebasedoncategoriesoriginallydevelopedbytheGeneral Catalogue of Variable Stars(GCVS;Kholopovet al.1985).
Price and Paxson, JAAVSO Volume 40, 2012 1013
4. Results
4.1.Age Themeanageofsurveyrespondentsis53(N=671;Figure1).Thereisnosignificantdifferencebetweenthemeanagesofmenandwomen,norofthemeanagesbetweendigitalandvisualobservers.The1994surveyreportsfrequenciesinstead of means. The mean frequency category was “41–50” (N=420).CitizenSkymembers report ameanageof41 (N=1,385).Sky & Telescopemagazine reports a mean subscriber age of 51 (New Track Media 2010). TheAAVSOmaintainsanarchiveofmembershipapplicationsdatingbackto1911.Almostallapplicationsincludeeithertheapplicant’sageorbirthdate.Werandomlyselected615applicationsandplottedtheirageasofthemomenttheyjoinedtheorganization(Figure2).Therearesomegapsinthedatafromincompleterecords(namely1911–1918,1922–1928,and2004–2008).Overtheentire100-yearperiod, theaverageageofanAAVSOmembershipapplicantwas 37 years old. During 1911–1921, the average new member age was 40years.Newmemberagedroppedto28yearsduringtheyears1967–1977andwas51yearsfortheyears2001–2011. The average age of observers does not vary much among the variousobservingtechniquesincludedinthesurvey(Figure3).TheageofPEPobserversistheonlycategorythatstandsout.Wefoundslightlymorevariationbetweenthe average ages of those engaged in non-observing activities (Figure4). Inparticular, the more “high-tech” activities of programming and data miningtendtohaveslightlyyoungerparticipants.
4.2Yearsactive “Years active” is a variable computed from subtracting the first year arespondent reported to be active in theAAVSO from 2011. It represents anapproximation of how long survey respondents have been active in theorganization.Themeanis14.3years(SD=15,N=598).Thereseemstobeadropoffatyears2and5,afterwhichdropoutratesflatten(Figure5).MembersoftheAAVSOtendtobeinvolvedintheorganizationforsixyearslongerthannon-members,adifferencewhich is statisticallysignificant,F (1,503)=22.0,p<.001.
The1994surveyincludedfrequenciesofthenumberofyearsrespondentshaveobservedvariable stars (notea slightdifferencebetween theirquestionabout “observing” and our question about “being active”) separated into 5-yearbins(Figure6),sowewereunabletocomputeamean.Theonlymajordifferencebetweenthe1994and2011distributionsisadropoffafter30yearsthatappearsinthe1994survey,butnotinthe2011survey.
In order to get more detail from the period around the 1980 survey, werandomlyselected26observersfromtheAAVSOInternationalDatabase(AID)whohadsubmittedobservations in1977(chosen tomatch thesame intervalbetween1994and2011)andpulledtheiroriginalmembershipapplicationtosetadatefortheirjoiningoftheorganization.Wecomputedthedifferencebetweenthatdateand1977as theirAAVSOAge(Figure7).Themeanagewas10.8years.ThemeanAAVSOAgeofmembersinthe2011surveywas16.8years.Thedistributionsaresimilar,butnotthesame.Thedrop-offsseeninthe2011surveyatyears2and5occurinthe1977data,butayearortwolater.Overall,thetrendsareverysimilar.
4.3Membershipstatus Membershipstatuswasdeterminedbylookingupanobservercodeore-mailaddress(whenprovided)intheAAVSOmembershipdatabaseonJanuary18, 2012. At that time, 54% of respondents were official members of theAAVSO. In the 1994 survey, 85% of the respondents were members of theorganization(N=417),accordingtoself-reporteddata.Inthe1994survey,wereceivedobservationsfrom660observers.In2011,wereceivedobservationsfrom1,050observers.Theobserver/membership ratiodifferencemay reflectmore thegrowthofobservers rather than the lossofmembers.TheAAVSOdoesnotkeeparecordofmembershiptotalsperyear. Wealsolookedforarelationshipbetweenmembershipstatusandwhethertherespondentreportstobeanactiveobserverornot.Wefoundnosignificantrelationship. However, for those who were active, there was a significantrelationship between the techniques they used and their membership status,F(390,8)=2.21,p=.03(Figure8).Themostinterestingresultisthat60%oftelescopicCCDobservers(N=143)aremembersand43%oftelescopicvisualobservers(N=138)aremembers.Thisdifferenceisstatisticallysignificant,F(315,1)=7.36,p<.01.
4.4.Gender 92% of respondents identified as male (N=634) and 8% identified asfemale(N=44).Themeanageforwomenis49,whilethemeanageformenis53,howeverthedifferenceisnotstatisticallysignificant(p=.06).Thelowsampleoffemalesmakesitdifficulttolookforrelationshipsbetweengenderandothervariablesinthesurvey.Inthe1994survey,thedistributionwas94%maleand6%female,veryclosetothecurrentratio.Sky & Telescopereportsagenderratioof95%maleand5%femaleintheir2010advertisingratecard(NewTrackMedia2010).TheCitizenSkygenderdistribution is78%male,19%female,and3%unreported(N=1,385).
4.5Country 108differentcountrieswererepresentedinoursurveyresults.About49%ofrespondentswerefromtheUnitedStates.Therestwerewidelydistributedamong theother107 countries (Figure9). 28 countrieswere representedbymembersand46countrieswererepresentedbyactiveobservers.
4.6Formaleducation Almost a quarter of the respondents claimed to have a terminal degree(Ph.D.,M.D.,J.D.,andsoon)intheirfield(Figure10).About76%reportaBachelor’sdegreeorhigher,whichisclosetoSky & Telescope’srateof77%(NewTrackMedia2010).
There is a significant relationship betweenprofessionandobservationtype,F(9,438)=3.44,p<.01. Most of that significance is due to theincreasededucationlevelsofthespectroscopicobservers (N=32) and decreased educationlevelsreportedbythesunspotobservers(N=22)(Figure11).
4.7.Profession Thereporteddistributionofprofessions(N=615)canbebrokendownintotwocategoriesofhighandlow(Figure12).Themostcommonprofessionswereinscience,computerscience,engineering,andeducation.Thosefourcategoriesaccountforabout57%oftherespondents.Therestoftheothercategorieswereroughlyeven,withmanagementandhealthcare leading thegroup.BuildingandGroundsCleaninghadthefewestresponses.
Figure 12. Distribution of professions according to U.S. Department of Labor categories: a.Transportationandmaterial;b.Salesandrelated;c.Protectiveservice;d.Production;e.Personalcare and service; f. Office and administrative support; g. Mathematical and computer; h.Management;i.Life,physical,andsocialscience;j.Legal;k.Installation,maintenanceandrepair;l.Healthcarepractitionersandtechnical;m.Foodpreparationandserving;n.Financial;o.Farming,fishing,andforestry;p.Engineering,architectureandsurveyors;q.Education,training,andlibrary;r.Constructionandextraction;s.Communityandsocialservices;t.Buildingandgroundscleaning;u.Arts,design,entertainment.
The 1994 survey had a similar question, but with fewer professioncategories to choose from.Table2 is a comparisonof results from the twosurveysincategoriesthataresimilar.Ingeneral,thereisnotmuchdifferencethatcannotbeexplainedbydifferencesinthedefinition/labelingofcategoriesbetweenthesurveys.
4.9.Motivation Supporting science and research (“citizen science”) is the most popularreasonrespondentsgaveforbeingactiveintheAAVSO.Aclosesecondisaninterestinvariablestars(Figure14).Around9%ofrespondentsareactiveintheAAVSOduetoprofessionalorgraduatestudentresearch.
Price and Paxson, JAAVSO Volume 40, 2012 1019
4.10.Barrierstoactivity Toinvestigatebarrierstoactivityweaskedanopen-endedquestionwordedas:“IfyouarenotcurrentlyactiveintheAAVSO,whatisthemainreason?”Almost all respondents who previously reported to be inactive (N=192)providedananswertothisitem.Notethevaguedefinitionof“active”inthequestion.Wepurposelyallowedtherespondenttodefineactivityintheirownwaysoas to include thosewhowouldotherwisebeactive innon-observingcontexts.Time(43%)wasbyfar themainreportedreasonrespondentswereinactive (Figure15).Other astronomy interestswas second (14%), followedbyalackofequipment/poorlocation(12%).Muchofpoorlocationcommentswerereportedasproblemswithlightpollution.
Figure 15. Major reasons to be inactive /barrierstoactivity(N=192).
4.11.Referralsources People learn about the AAVSO from a variety of sources (Figure 16).Wordofmouth(25%)ismostcommon,followedcloselybyotherastronomycluborconferences(19%),Sky & Telescopemagazine(19%),andtheInternet(18%).The1994surveyalsoincludedaquestionaboutreferralsources.Itwasstatedas:“HowdidyoufirsthearabouttheAAVSO?”andhad427responses(Figure17).
The biggest difference between the 1994 and 2011 surveys is the 35%to19%dropinreferralshareattributedtoSky & Telescopemagazine.Sky & Telescopehaslongbeenanimportantsourceofbrandingfortheorganization,butithasdroppedinsignificance.ThisislikelyduetotheabilityoftheAAVSOtoreachamateurastronomersdirectlythroughtheInternet.Ifyouaddupthe
Price and Paxson, JAAVSO Volume 40, 20121020
referralshareofthe“Internet”and“Sky & Telescope”inthissurveyitreaches37%, which is very close to the 35% share in the 1994 survey (and in linewithgeneralsocietaltrends).Clubandbookscontinuetohavesimilarsharesbetweenthe1994and2011surveys,at18%–19%and12%–9%respectively.Interestingly,non-Sky & Telescopemagazinesalsohaveasimilarsharebetweensurveys,at7%–8%. Inthe2011survey,wecodedtalksintothe“Wordofmouth”category.Ifyoucombinethe“Other(talks…)”and“Friends”categoriesinthe1994survey,thentheytoohaveasimilarsharewiththe2011“Wordofmouth”categoryat28%–25%. The1980surveyincludedareferralquestionaswell.Itwasstatedas:“SourceofinformationabouttheAAVSO?”.Waagen(1980)dividedtheresponsesinto5categories(Figure18).Ingeneral,theyareconsistentwiththeothersurveys.Themajordifferencebeingthelargesharebooksheldin1980(21%)asopposedto 1994 (12%) and 2011 (9%). Club referrals also dropped between 1980(23%)and1994(18%)whileremainingconsistentfrom1994to2011(19%).
Figure 19.Active and inactive observer rates(N=691).
Price and Paxson, JAAVSO Volume 40, 2012 1021
Wewereinterestedtoseewhethertheincreaseininactiveobserverswasrelated to any of the non-observing activities in the AAVSO. An ANOVA(ANalysisOfVariance)foundasignificantdifferencebetweentheactiveandinactivegroupsbasedonwhethertheyarealsoactiveinanon-observingactivity,F(10,644)=1.9,p<0.05.Figure20isaplotofthemeanobservationactivityvalue(lowernumbermeansmoreactive)groupedbynon-observingactivitieslistedinthissurvey.Itisofnosurprisethatthemanyactiveobserversarethoseinvolvedinthechartprocess,sincechartshaveadirectimpactonobserving.Itisalsointerestingthattheleastactivearethoseinvolvedinfinancialaspectsoftheorganization.Programmerstendtobeactiveobserversaswell.Beyondthat,therestofthecategoriesareroughlyeven.
Figure 20. Activity rates of respondents whoparticipateinnon-observingactivities.
4.13.Faintestobservation We asked observers to report the magnitude of the faintest observationthey“typically”observe(Figure21).Wethenaskedwhetherthatobservationwasmadevisuallyorwith“CCD/DSLR/PEP/OtherDigitalsystem.”Thiswasmore useful than asking if a person was a “visual or CCD” observer sincemanyusebothtechniques.Instead,thisquestiontellsuswhichtechniquetheyused to get their faintest observation, which will almost always be “CCD”
Figure 21. Faintest observations (magnitude)“typically” recorded by respondents. Mean =13.75;StandardDeviation=3.45;N=365.
Price and Paxson, JAAVSO Volume 40, 20121022
for CCD/combined observers and “visual” for visual-only observers (for amoredetaileddiscussionofobservation type see theobservation technologysection). The goal of these questions was to establish magnitude ranges forvisualorCCDcampaigns.Observersweredividedroughlyinhalfbetweenthetwo(Figure22).Themeanfaintestvisualobservationwas12.02andthemeanfaintestCCDobservationwas13.75.Thevisualdistributiondropsoffsharplyaroundmagnitude15–16whiletheCCDdistributionfadesmoregraduallyuntilaroundmagnitude20(Figure23).
Figure 22. How the faintest observation inFigure21ismeasured.
Figure 23. Distribution of the faintest “typical” observation made through CCD or visualmeasurements.
4.15.Observationtechnology We provided respondents with a list of observation technologies andasked them to select which types they “actively use” (N=460, Figure 26).Interestingly,40%ofthosewhoreportedtobetelescopicCCDobserversarealsoactiveasvisualtelescopic(N=54)ornakedeye(N=34)observers.Also,thenumberofthosewhoreportedtobeobserversutilizingspectra(N=32)wassurprising,sincetheorganizationhasnoformalgrouporsectiontostandardizethis activity. Lastly, more than half of all observers use multiple observingtechniques(Figure27).
4.16.Non-observingactivities Byfar,themostpopularnon-observingactivitiesinvolvedpublicoutreach(Figure 28). Specifically, giving talks and writing about theAAVSO. Priorsurveysdidnotincludeitemstoaddressparticipationintheseareas.Wefoundno significant relationships between non-observing activity and observationtechnologyorwithmembershipstatus.
4.17.Objectinterest Whenaskedtorankobjectsbyinterest,wefindaclusteringofobjectsintotwocategories:highlyranked(Pulsating,CV,EB,Novae,Extragalactic)andlesserranked(YSO,Sun,Rotating,NonStellar)(Figure29).Respondentswereonlyallowedtoprovideoneobjectforeachranking.However,theywerenotrequiredtorankeveryobject.Somerespondentsonlyrankedtheobjectstheyweremost interested in (example: theyonly ranked the top5). Inaseparatetabulation,weassignedarankingof9(thelowestrankpossible)tothoseobjectsthathadmissingdata,withthejustificationthattherespondenthadnointerestinthatobject.Theonlymajordifferenceisthatthedropoffbetweenthehighlyrankedgroupandthelesserrankedgroupincreases.33typesofobjectswereincludedinthe“Other”category(Table3).Manyofthesetypesofobjectscouldbeincludedinexistingsurveycategories.
4.18.Meetingattendance Themeannumberofmeetingsattendedbyarespondentwas1.5(N=584;Figure 30). However, that number is significantly skewed because of fourrespondentsofreportedbetween20–60meetings.Theoverallsurveymedianvalue is 0, meaning that the vast majority of respondents had not attendedanAAVSO meeting before. By excluding those who have not attended any
meetings, and the two persons who reported 40 and 60 meetings, then theaveragemeetingattendanceis4.2.Thisnumberreflectsthenumberofmeetingssomeonewouldtypicallyattendiftheyhaveattendedatleastone.
We looked for relationships between meeting attendance and the typeof observations people make and the types of non-observing activities theyparticipatein(Figure31).Fortypesofobservations,wefoundnosignificantrelationships.Fornon-observingactivities,wedidfindasignificantrelationship,F(573,10)=15.4,p<.001.However,mostofthatstatisticalrelationshipisduetothosewhoselected“HQVolunteer.”WebelievethistobeaperplexingvariablebecausethosewhovolunteeratHeadquarterstendtolivenearHeadquarters,thusattendtheannualmeetingquiteoften.InFigure31,wezoomedinontheothercategories.Themaindifferencebetweentheactivitiesthatinvolvemoreinvestmentofinitiative(writing,programming,mentoring,andsoon)thaninactivitiesthataremoreprocedural-based.Finally,membersaremorelikelyto
Amoredetailedinvestigationoftheagequestionisplannedforafuturestudy.Another major difference between surveys is where and how respondentsheardabouttheAAVSO.TheInternetisreplacingmanyofthereferralswhichpreviouslymayhavecomethroughSky & Telescopemagazine.Itisinterestingthat referrals from books and non-Sky & Telescope magazines have notchanged much between the 1994 and 2011 surveys.This suggests the issuemaybe specific toSky & Telescope,perhapsdue to itspreviouslydominantpositionasamajorreferringsourceand/orbecauseSky & Telescope ismorecloselyassociatedwithadvancedamateurastronomersthancasualreaders.ItispossiblethattheInternetisimpactingnewsstandsales(casualreaders)lessthancirculationsales(moreadvancedreaders). In terms of observing methodology, active observers are pretty evenlydividedbetweendigitalandvisualobserving.Sincethe1990s,therehasbeendiscussionaboutcompetitionbetween the two typesofobserving.However,wefoundthat40%oftelescopicCCDobserversarealsovisualobservers.Also,abouthalfofallCCDobserversandaquarterofnewCCDobserversbeganasvisualobservers.Thissuggeststhelinedistinguishingthesetwogroupsismuchfuzzierthanhasbeenadvertised. Therearesomesurprisingresultsaswell.First,AAVSOactivityisrelatedtoagreaterincreaseinself-efficacyinastronomythanonewouldfindthroughincreasedastronomyexperiencealone (asmeasured throughage).That is, itispossiblethatthedemandsofvariablestarresearchhaveagreaterimpactonhowoneviewstheirknowledgeofastronomythantheactivityofthetypicalamateurastronomer.Thiscouldhintatgreaterlearningtakingplaceinactivecitizenscienceprojectswhencomparedtotypicalamateurastronomyprojects.Afuturestudyisplannedtoinvestigatethisresult.AsecondsurprisewasthenumberofrespondentswhoreporttobeactiveintheAAVSO,butalsoreporttobeinactiveobservers.ThisreflectstheincreasedscopeoftheAAVSOoverthepasttwodecades.Whenthepasttwosurveyswereconducted,non-observingactivitieswerenotevenconsideredunlesstheywereinsupportofobserving.Now,manyparticipantsoftheAAVSOarededicatedtoimportantprojectssuchasprogramming,analysisofdata,publicoutreach,andsoon.ObservingisstilltheheartoftheAAVSO,with60%ofthosewhoreportparticipationinatleastonenon-observingactivityalsoreporttobeactiveobservers.Anothersurprisewasthehighlevelofeducationreported.Overhalfofparticipantsreportagraduatedegree,with24%reportingaterminaldegreeintheirfield(Ph.D.,M.D.,J.D.,andsoon).Also,13%identifyasprofessionalastronomers(N=86).Oneofthemost interestingsurprises is thenumberofcountriesrepresentedbyAAVSOmembership(108).Thismayreflectthesignificantworkthatvolunteershaveputintotranslatingourtrainingmaterialsintootherlanguages,butitislikelyabiggerreflectionoftheuniversalappealofvariablestars.Therearesimplysomanytypesofstarsandsomanyopenquestions,thatalmostanyonecanfindaproject or object of interest.Finally, thenumberof respondentswhohave
Price and Paxson, JAAVSO Volume 40, 20121028
authoredorcoauthoredapaper inascientific journalwasquitehigh (36%).ThisisoneofthemajordistinguishingcharacteristicsbetweentheAAVSOandothercitizenscienceorganizations,whotendtofocusonusingparticipantstocontributedataforprofessionalstoanalyzeandpublish. Raddick,et al.(2010)identified12categoriesofmotivationfromGalaxyZooparticipantsthroughanalysisofinterviewsandonlineforumposts.Aswithoursurvey,theircategorieswerereducedfromopen-endeddiscussion(intheircase, interviews were also included). In areas where our categories overlapwiththeirs,wefindsimilaritiesandsomesignificantdifferencesinmotivationrates.Theyreportonly1%oftheirresponsesaremotivatedby“science”while“scienceandresearch”wasthemotivationof35%ofourrespondents,whichwasourhighestcategoryofmotivation.TheGalaxyZooprimarymotivationcategorywas“astronomy”at39%.Ifthatreferstointerestinastronomy,thenitisanalogoustoour“interestinvariablestars”categorythatwascitedby32%ofourrespondents.Ourtwogroupsreportthesamelevelofmotivationintermsofcontributingdatatoagreatercause.13%oftheGalaxyZooparticipantsciteadesireto“contribute”asamotivationoftheirparticipationwhileanidentical13%ofourrespondentsciteadesireto“sharedata.” Thisstudyhasanumberoflimitations.First,itisastudyofactiveorrecentlyactiveparticipantsof theAAVSO.Somedata, suchas the ranked interest intypesofobjects,willbeskewedtowardscurrentoperations(peopleinterestedinexoplanets,forexample,mayhavedroppedoutoftheorganization).Sothisdatashouldnotbeusedasaguideforthefuture,butonlyasasnapshotofthepresent.Second,thecodingoftheopen-endeditemswaslimitedtoonecodeperitem.Soitmayoversimplifytheresultsofthoseitems.Thereisastrikingsimilaritybetweenourresultsandpastsurveysontheseitems,whichsuggestsstrongvalidity.Finally,thisisselfreporteddata.Thusitincludesbiasescausedbyhumannatureanddifferentdefinitionsofterminology.Forexample,somerespondentsreporttobeprofessionalastronomersyetalsoreporttonothaveaPh.D.Wearenotstatingtheyarenotprofessionals,justthatrespondentswillhavedifferencedefinitionsof the term“professional.”Tosome, it requiresa“Ph.D.” while for others it denotes publishing in journals while still othersapplythetermtoanyonecontributingscientificallytoastronomicalresearchatanylevel.
6. Conclusion
ThisisasummaryreportoftheAAVSO2011DemographicSurvey,whichincludedcurrentandrecentAAVSOparticipants.Comparedwithpastsurveysofthistype,itshowsanorganizationthatislargelysimilarindemographics.Respondents were active in a wide variety of observing and non-observingactivitiesandareinterestedinawidevarietyofobjects.TheAAVSOreflectsa “big tent” mentality, with room for everyone interested in variable stars.
Price and Paxson, JAAVSO Volume 40, 2012 1029
There are signs in thedata of some challenges, such as a population that isgrowing older and the presence of a very significant gender gap. However,thesearenotlimitedtotheAAVSOalone.Asadescriptiveanalysis,wemakeno predictions for the future. However, the results can be used, along withother surveysandanalysis, to identify futurepathsandopportunities for theorganization.
References
AAVSO 1976, AAVSO Archives, JAM Adm., Box 3 and 4, Survey ofMembership,AAVSO,Cambridge,MA.
AAVSO2011,VariableStarIndex(http://vsx.aavso.org).Beatty,K.2000,Sky & Telescope,100,3.Hazen,M.L.1995,J. Amer. Assoc. Var. Star Obs.,23,143.Kholopov,P.N.,et al.1985,General Catalogue of Variable Stars,4thed.,Moscow.Lowder,W.M.1994,AAVSOArchives,JAMOrg.,Box3B,FuturesStudy(2),
This is the printed version of the AAVSO 2011 Demographic survey. All items are worded exactly as they appeared online, except for the state and country items, which included drop down lists. In a few areas, screen shots of the online form were used in the printed survey as well.
The Citation of Manuscripts Which Have Appeared in JAAVSO
Arlo U. LandoltDepartment of Physics and Astronomy, Louisiana State University, Baton Rouge, LA 70803; [email protected]
Received May 3, 2012; revised May 14, 2012; accepted May 15, 2012
Abstract A study is presented of the use by the astronomical communityof themanuscriptspublished inThe Journal of the American Association of Variable Star Observers(JAAVSO).
1. Introduction
The Journal of the American Association of Variable Star Observers (JAAVSO) was established in 1972 as “a place where professional and non-professional astronomers can publish papers on research of interest to theobserver”(Mayall1972).Mayallwentontowritethat“thenewjournalwouldincludeAbstractsofpaperspresentedatAAVSOmeetings,variousCommitteereports,MinutesoftheMeetingsoftheAAVSO’sCouncil,Observer’sTotals,andBookReviews.”Additionalinsightsconcerningtheformationanddevelopmentof JAAVSO are discussed in theAAVSO’s centennial history (Williams andSaladyga2011).ThecontentofJAAVSOhasevolvedoverrecentyears,nowcontaininglessofthebusinessandmoreofthescholarshipoftheorganization. ThecurrentauthorsandreadersofJAAVSOareamixtureofamateurandprofessionalastronomers.Theseindividualshaveanabidinginterestincelestialphenomenawhichvaryinbrightnessandcolor.ThekindsofdatapresentedinJAAVSOmanuscriptshavechangedovertime.Historically,thedataprimarilywere visual or photographic.At present, data may be visual, photoelectric,or CCD-based. Both the precision and accuracy of the published data haveimprovedwithtime.SufficeittosaythattheauthorsofJAAVSOpapersareavariedandtalentedpopulation. Onthewhole,andcertainlyhistorically,JAAVSOpapersdwelledmoreondata acquisition and data presentation. Data interpretation is not pursued totheextentonefindsintheprofessionaljournals.ItisimportantthattherebeajournalsuchasJAAVSOforitprovidesanopportunitytoabroadercommunitytowriteandtopublishusefulandqualityresearch.
2. Method
TheAstrophysicsDataSystem(ADS)isheadquarteredattheSmithsonianAstrophysical Observatory at Harvard. It is funded by NASA and contains
Landolt, JAAVSO Volume 40, 2012 1033
bibliographic databases, is a digital library, thereby providing access toastronomical informationto theworldvia theWorldWideWeb.Oncein theADS,andintheJournal/Volume/Pagesection,aparticularjournal’scodecanbeentered;forJAAVSO,thatcodeis“JAVSO.”Next,thedesiredjournalvolumenumbercanbeentered.ThisstepleadstoQueryResultswhichlistthepapersintheirorderofappearanceinthespecifiedvolume.TheexistenceoftheletterC indicates that thearticlewascited.Clickingon theCbringsup thecitingarticles,whosenumbersthenmaybecounted. IttranspiresthattheADSdoesnotindexallarticlesinagivenvolumeinits article, or paper, count.Two volumes, numbers one and nineteen (whichappearedin1972and1990,respectively),werecarefullysearchedinaneffortto ascertain which titles were not included by theADS. In volume one, theTreasurer’sReport,ObserverTotals,andCommitteeReportswerenotindexedasarticles.Similarly,involumenineteen,theIntroductiontothevolume,thelistingofthepapersatFirstEuropeanMeetingoftheAAVSO,theMinutesoftheAnnualMeeting,theDirector’sannualreport,theCommitteeReports,andtheTreasurer’sreportwerenotidentifiedbytheADS.Allthesewrittenitemshavethecharacteristicthat theyarenotastronomicalresearcharticles,hencea likely reason for omission from theADS index count.Another reason forinclusionoromissionis that thedecisiontoincludeornot includedependedonthedecisionmaker,differentfromtimetotime.Thisstudywaslimitedtopaperscontainingastronomicaldata. This exercise reviewed the thirty-nine volumes of JAAVSO which werepublishedintheinterval1972through2010.Thereview,thecountingofthecitations, of the authored articles identified by theADS, was completed onOctober,28,2011. Afilewascreatedwhosefunctionwas tocontainacountof thenumberoftimeseachpaperinagivenvolumewascited.Thisfilecontainedamatrixconstructed with the volumes listed in the row, and the possible number ofcitations, from zero to n, listed in the column. The matrix was filled in bytabulating the citations present, or not, if there were none.As an example,volume one contained thirty-one articles, as defined by theADS, of whichtwentypapersneverwerecited,fivepaperswerecitedonce,fourpaperswerecitedtwice,onepaperwascitedthreetimes,andonepaperwascitedfourtimes,foratotaloftwentycitationsforthevolume.
3. Results
SomeaspectsofthisinvestigationhavebeentabulatedinTable1.TheTablecontainsthefollowinginformation:thevolumenumberinthefirstcolumn,theyearofpublicationinthesecondcolumn,thenumberofpaperswhichappearedinthevolumeincolumnthree,andthetotalnumberofcitationsreceivedbythepapers in thatvolume incolumn four.Column fivecontains thenumber
Landolt, JAAVSO Volume 40, 20121034
of citations due to JAAVSO authors for a given volume, or, another way ofdescribingit,columnfivegivesthenumberofcitationsincolumnfourwhicharebyJAAVSOauthors.Thelastcolumnprovidesthepercentageofthecitationswhich were due to JAAVSO authors. This column is of interest because itindicateswhetheronlyindividualswhopublishinJAAVSOarecitingJAAVSOpapers. The goal of this table is to compare the number of citations by theastronomicalcommunityasafunctionofvolume,withthenumberofcitationsbyJAAVSOauthors. The information tabulated in Table 1 indicates that 1,545 papers werepublishedinJAAVSOoveritsfirstthirty-nineyears,arateof39.6paperspervolume.Those1,545papershavebeencited1,296times,or,0.84citationsperpaper. Ofthetotal1,296citations,464,or35.8%,werecitationsbyotherJAAVSOauthors.And,64.2%,abouttwo-thirds,ofallcitationstoJAAVSOpaperscamefromthegreaterastronomicalcommunity.Atotalof1,296citationsinthirty-ninevolumesleadsto33.2citationspervolume. The last column in Table 1 provides the percentage of JAAVSO authorcitationsforeachvolume.Ontheaverage,37%ofthecitationstopapersinagivenvolumearebyJAAVSOauthors. Eighteenpapersinthesethirty-ninevolumeswerecitedtenormoretimes.Thepaperwiththemostcitations,nineteen,wasbyPercy,et al.,1985,JAAVSO,vol.14,p.1.Appendix1liststheeighteenpapersinchronologicalorder,andwhichwerecitedatleasttentimes,throughOctober28,2011.Thenumberinbrackets foreach listedpaperprovides thecitationcount for thatpaper.Thepurposeinlistingthemostcitedpapersistoshowthebreadthoftheprojectsfound useful by the user community. It should be emphasized that a smallnumberofcitationsdoesnotmeanthatapaperhasnovalue.Indeed,suchapaper’scontentmaybecrucialtoaciter’sownresearch. Insummary,onecansaythatthearticles,thepapers,whichappearintheissuesofJAAVSO,arereadby,andarecitedby,researchersacrosstheamateurandprofessionalcommunity.
References
Mayall,M.M.1972,J. Amer. Assoc. Var. Star Obs.,1,1.Williams,T.R.,andSaladyga,M.2011,Advancing Variable Star Astronomy:
The Centennial History of the American Association of Variable Star Observers,CambridgeUniv.Press,Cambridge.
Landolt, JAAVSO Volume 40, 2012 1035
Vol. Year Number Total Number Percent of Papers Number of JAAVSO JAAVSO of Cites Cites Cites
Appendix 1Papers published in JAAVSO having the most citations. The papers are listed in chronological order. The number in brackets for each listed paper provides the citation count for that paper.
Abstracts of Papers and Posters Presented at the Joint Meeting of the Society for Astronomical Sciences and the American Association of Variable Star Observers (AAVSO 101st Spring Meeting), Held in Big Bear Lake, California, May 22–24, 2012
Fast Spectrometer Construction and Testing (Abstract)
John Menke22500 Old Hundred Rd, Barnesville, MD 20838; [email protected]
Abstract This paper describes the construction and operation of a medium resolution spectrometer used in the visual wavelength range. It is homebuilt, but has built in guiding and calibration, is fully remote-operable, and operates at a resolution R = 3000. It features a fast f /3.5 system, which allows it to be used with a fast telescope (18-inch f /3.5) with no Barlow or other optical matching devices.
Observations Using a Bespoke Medium Resolution Fast Spectrograph (Abstract)
John Menke22500 Old Hundred Rd, Barnesville, MD 20838; [email protected]
Abstract Designing and building a medium resolution (R = 3000) spectrograph was the relatively easy part. The really challenging part is learning how to use it: learning the characteristics of the spectrograph, choosing the right kind of astronomical problems, learning the best methods of taking data, and figuring out how to analyze the results. I have used several observing projects to “commission” this system, including measuring the Doppler shifts in several WUMa type stars. I will briefly describe the spectrograph but discuss in more detail the early experiences of using it.
Enhancing the Educational Astronomical Experience of Non-Science Majors With the Use of an iPad and Telescope (Abstract)
Robert M. GillMichael J. BurinCalifornia State University, San Marcos, Physics Department, 333 S.Twin Oaks Valley Rd., San Marcos, CA 92096; [email protected], [email protected]
Abstracts, JAAVSO Volume 40, 20121038
Abstract General Education (GE) classes are designed to broaden the understanding of all college and university students in areas outside their major interest. However, most GE classes are lecture type and do not facilitate hands-on experimental or observational activities related to the specific subject matter. Utilizing several astronomy application programs (apps), currently available for the iPad and iPhone, in conjunction with a small inexpensive telescope allows students unique hands-on experiences to explore and observe astronomical objects and concepts independently outside of class. These activities enhance the student’s overall GE experience in a unique way not possible prior to the development of this technology.
The Rotational Period of the Sun Using the Doppler Shift of the Hα Spectral Line (Abstract)
Robert M. GillCalifornia State University, San Marcos, Physics Department, 333 S.Twin Oaks Valley Rd., San Marcos, CA 92096; [email protected]
Abstract The fact that the sun rotates is obvious by observing the daily motion of sunspots. The overall sunspot movement to the west is a result of this solar rotation. However, solar rotation can also be determined by observing the solar spectrum at the solar limbs. The absorption lines in the spectrum will display a Doppler shift since the east limb is coming toward the observer and the west limb is moving away. The velocity of the limb, relative to the observer, can be determined from these spectral line shifts. Knowing the solar radius, the rotational period can be calculated.
A Single Beam Polarimeter (Poster abstract)
Jerry D. Horne3055 Lynview Drive, San Jose, CA 95148; [email protected]
Abstract As astronomical polarimetry is an emerging field of study for amateur astronomy, the background, theory, and instrumentation of astronomical polarimetry is reviewed. Additionaly, the design and construction of a simple single beam polarimeter is presented, together with the results of its initial calibration.