文档下载 文档下载导航
设为首页 | 加入收藏
搜索 请输入内容:  
 导航当前位置: 文档下载 > 所有分类 > Tests for Selection Effects in GRB High-Energy Correlations
侵权投诉

Tests for Selection Effects in GRB High-Energy Correlations

Several correlations among observables derived from modelling the high-energy properties of GRBs have been reported in the literature. We show that well-known examples of these have common features indicative of an origin in selection effects. In particula

GRBCorrelations

3

portantforSwiftduetoarcsecondX-raylocalizations.Westressthatacorrelationisnotintrinsicsimplybe-causetriggerthresholdingcanberuledoutasin uenc-ingthesurveys(see,e.g.,Amatietal.2002).Stud-iesareunderwaytobetterunderstandthetruncationsandtopotentiallyinferwhetheranyintrinsice ectsarediscernibleintheradically-truncatedpre-Swiftsurveys(e.g.,Ghirlandaetal.2008).

3.THEANATOMYOFACORRELATION

Pre-Swift,afterglow-erasatellitesdonothavesensitiv-itiesthatvarybyordersofmagnitude(see,e.g.,Band03);we ndthatdiagnosingthereliabilityofacorrela-tionplottedonalog-logscaleforthesedataalonecan-note ectivelybedonebylookingatvariationsinthenormalizationsalone.Wecaninsteadexaminehowtheregressionslopeandscattervarywithassumptionsonthenatureofthecorrelation.Inthissectionweoutlineacartoon-ish,butinsightful,mathematicalmodel.

IfweassumethatallGRB tparametershavethesameerror,theerrorscanbeignored.ConsiderobserverframequantitiesX=log(Sbol)andY=log(Efor′example.Thesetranslatetosourceframequantitiespk,obs),X=X+bZ=log(Eiso)andY′=Y+aZ=log(Eexample.Here,Z=log(1+z)anda=1,b≈pk),for3(accurateto20%for0.5<z<5,representingthebulkofknownGRBredshifts).

Ifweassumeanoriginintheobserverframe,crosstermsinvolvingtheobservablesandzcanbeignored,andtheexpectedsource-frameregressionm= slopeis:

m Z2

srcobs+ab X2

.(1)

Here, . denotesanaverageaftersubtractingawaythemeaninXorZ,andm1givestheobsistheslopeintheobserverframe.Equationslopeoftheapparentsourceframerelationgiventheobserveddata,whereasdeter-minationofthetruesourceframerelationrequiresde-tailedknowledgeoftheGRBratedensityandluminos-ityfunctiontoimputethemissingdata.Althoughre-centprogresshasbeenmadeonmeasuringthese(e.g.,Kistleretal.2007),wedonotattemptheretorecon-structmGRBobs[mstudiessrc,true].Moreover,forcomparisonwithmostwhichsimplyassumemsrcEquation,true=m(e.g.,Firmanietal.2005),weonlyrequire1.srcRelativelyindependentofwhichsurveyweuse, Z2 / X2 ≈0.1,andmsrc≈mobs.Thisisthetypi-calexpectedbehaviorforGRBcorrelationswhereXorYaretypicallymuchmorebroadlydistributedthanZ.However,arti cialcorrelationsinthecasem(e.g.,considerWillingaleetal.2007),obsmayalsooccurdependingonthealgebraicarrangementofobservablesgoingintoXandY.

ThescatterinYaboutthebest- tregressionlineis

σ2src= (Y msrcX)2 +(a msrcb)2 Z2 .

(2)

CovariancetermsbetweenXandZorbetweenYandZ—whichgenerallyacttodecreasethescatterinthesourceframeascomparedtotheobserverframe—aredroppedbecauseweassumethecorrelationisnon-intrinsic.Tobeexplicit,rearrangementlikeEpk∝E0.5

ofanequation

,andthisisoshowsthattheobservablesshouldvarywithzimpliescrosstermsinEquation2.The

crosstermsgotozeroifthecorrelationisdueto(z-independent)observationale ects.

The rstterminEquation2isboundfromthebottombyσ2obs;equalityisobtainedformsrc=mobs.Thesec-ondterminEquation2impliesanincreaseinσ2

overσ2

src

obsunless,coincidentally,a=msrcb.Interestingly,this“coincidence”issatis ednearlyorexactlyformanyoftheknownGRBcorrelations.

Asanexample,considertheGhirlandacorrelation:Epk∝E0γ.7et,whereEγ∝Eisoθ2

al.(2004)

ergyreleasecorrectedforbeamingintoajetjetistheen-ofangleθcaseofauniformdensitymediumsurrounding

jet.ForthetheGRB(Sarietal.1999),θ2

3/4T3/4thetimeinthejet∝E 1/4observeriso(1+z) framewherejet,whereTjetisthejettinge ectsareexpectedtobecomeapparent.Sub-stitutingintoequations1and2,itfollowsthatX=3/4log(SbolTjet),Y=log(Epk,obs),a=1andb≈3/2.Hence,a≈msrcb,asalsodemonstratedgraphicallyinFigure2A.TheGhirlandaetal.(2004)correlationis,therefore,e ectivelyindependentofthemeasuredphoto-metricredshifts.ForthedatainFigure2A,σσinsigni cantobs=0.17dexandσsrc=0.16dex(i.e.,0.9decreasefromanF-test).

Asaninterestingside-notewhichfurthersuggestsapaucityofintrinsicphysics,thissamebalancingischaracteristicofsimilarrelationsfoundassumingin-steadawind-strati edmedium(Navaetal.2006)orevenallowingθ(Liang&ZhangjettovaryarbitrarilywithT2005).ThelatterrelationhasjetandEthesmall-isoestσsrc,whichleadsustosuspectthattheonlyreasonthataverylargenumberofcorrelationsdonotaboundintheliteratureisthatwetypicallyrestricttothosethancanbeinterpretedphysically,or,shortofthis,werequiretheobservablestoevolvewithredshiftinaphysicallyplausiblefashion(e.g.,Tjet,src=Ttheseconstraints,itispossiblejet/(1+z)).With-outtoinventhighly-signi cantbutabsurdrelations(e.g.,redshiftedSwifttriggernumberversusredshiftedburstduration;Figure2B)whichmakeamockeryofallcorrelationstudies.

4.MINIMALREQUIREMENTSFORNEWCORRELATIONS

WeshowaboveforGRBsobservedwithmultiplesatel-litesthatfourexamplecorrelationsreportedintheGRBliteraturecanbeexplainedintermsofselectione ects.Contrarily,thereisnoconvincing,non-a-posteriorex-planationforthecorrelationsinthesourceframe.ThisstronglysuggeststhatthesehavenothingtodowiththeintrinsicphysicsofGRBs.Ourresultscanbetestedfur-therusingpre-Swiftdataonceweextendthosecatalogsdowntothedetectionthreshold uxlevelsorestablishthedominanttruncationmechanisms.

Independentoftheargumentthatthecorrelationsareduetotruncatione ects,acommonz-independenceiseitherveryoddordamning.Aswediscuss,thisisonecharacteristicofatight,apparentsource-framecorrela-tionwhicharisespurelyduetoselectione ects.Itcannotbeaconsequenceofintrinsicphysics,becausetheGRBcannotpossiblyknowhowthedistancetotheobservershouldvarywithz.Wecautionthatthebalancingof1+ztermsimpliesthecorrelationscannotbeusedtoin-ferz.Moreover,usingtheobservablesandspectroscopicredshiftstoconstructaHubblediagramandtestconcor-dancecosmology(e.g.,Schaefer2007)iscircular,because

上一页第3页

你可能喜欢

热门文档

相关文档

站点地图 | 文档上传 | 侵权投诉 | 手机版
本站所有资源均来自互联网,本站只负责收集和整理,均不承担任何法律责任,如有侵权等其它行为请联系我们.
文档下载 Copyright 2013 doc.guandang.net All Rights Reserved.  -  email