Spectral properties of near-Earth and Mars-crossing asteroids using Sloan photometry B.Carrya,b,E.Solanoc,d,S.Eggla,F.E.DeMeoe,f aIMCCE,ObservatoiredeParis,PSLResearchUniversity,CNRS,SorbonneUni-versites,UPMCUnivParis06,Univ.Lille bEuropeanSpaceAstronomyCentre,ESA,P.O.Box78,28691VillanuevadelaCan˜ada,Madrid,Spain cCentrodeAstrobiologia(INTA-CSIC),DepartamentodeAstrofisica.P.O.Box78,E-28691VillanuevadelaCan˜ada,Madrid,Spain dSpanishVirtualObservatory eDepartmentofEarth,AtmosphericandPlanetarySciences,MIT,77MassachusettsAvenue,Cambridge,MA,02139,USA fHarvard-SmithsonianCenterforAstrophysics,60GardenStreet,MS-16,Cambridge,MA,02138,USA 6 1 Abstract 0 2Thenatureandoriginoftheasteroidsorbitinginnear-Earthspace,includingthoseonapotentiallyhazardoustrajectory,isofboth nscientific interestandpracticalimportance. We aimhereatdeterminingthetaxonomyofalargesampleofnear-EarthandMars- acrosser asteroids and analyze the distribution of these classes with orbit. We use this distribution to identify the source regions J of near-Earthobjects and to study the strength of planetaryencountersto refresh asteroid surfaces. We measure the photometry 9ofthese asteroidsoverfourfiltersatvisible wavelengthsonimagestakenbytheSloanDigitalSkySurvey(SDSS).Thesecolors are used to classify the asteroids into a taxonomy consistent with the widely used Bus-DeMeo taxonomy (DeMeo et al., Icarus ] P202,2009)basedonvisibleandnear-infraredspectroscopy. Wereporthereonthetaxonomicclassificationof206near-Earthand E776 Mars-crosser asteroids determinedfrom SDSS photometry,representingan increase of 40% and 663% of knowntaxonomy .classificationsinthesepopulations. UsingthesourceregionmapperbyGreenstreetetal. (Icarus,217,2012),wecompareforthe h pfirst time the taxonomic distribution among near-Earth and main-beltasteroids of similar diameters. Both distributions agree at -the few percentlevelforthe innerpartof the Main Belt andwe confirmthis regionas a main sourceof near-Earthobjects. The o effectofplanetaryencountersonasteroidsurfacesarealsostudiedbydevelopingasimplemodelofforcesactingonasurfacegrain r tduringplanetaryencounter,whichprovidesthe minimumdistance at whicha close approachshouldoccurto trigger resurfacing s aevents.Byintegratingnumericallytheorbitofthe519S-typeand46Q-typeasteroidsinoursamplebackintimefor500,000years [andmonitoringtheirencounterdistancewithVenus,Earth,Mars,andJupiter,weseektounderstandtheconditionsforresurfacing events. ThepopulationofQ-typeisfoundtopresentstatisticallymoreencounterswithVenusandtheEarththanS-types,although 1 vbothS-andQ-typespresentthesameamountofencounterswithMars. 7 8Keywords: Near-Earthobjects,Asteroids,composition,Photometry 0 2 01. Introduction bits cross that of the telluric planets, form a transient popu- . 1 lation. Their typical lifetime is of a few million years only 0 Asteroids are the leftovers of the building blocks that ac- (Bottkeetal.2002;Morbidellietal.2002)beforebeingejected 6cretedtoformtheplanetsintheearlySolarSystem. Theyare from the Solar System, falling into the Sun, or impacting a 1alsotheprogenitorsoftheconstantinfluxofmeteoritesfalling planet. Thesepopulationsarethereforeconstantlyreplenished : von the planets, including the Earth. Apart from the tiny sam- byasteroidsfromthemainasteroidbelt,thelargestreservoirof iple of rock from asteroid (25143) Itokawa brought back by X knownsmallbodies,betweenMarsandJupiter. the Hayabusa spacecraft (Nakamuraetal. 2011), these mete- The resonances between the orbits of asteroids and that of r aoritesrepresentoursolepossibilitytostudyindetailsthecom- Jupiterhavebeenlongthought(Wetherill1979;Wisdom1983) position of asteroids. Identifying their source regions is cru- toprovidethekickineccentricitynecessarytoplaceasteroids cial to determine the physical conditions and abundances in on planet-crossing orbits. It was later found that the secular elements that reigned in the protoplanetarynebula around the resonance ν , delimiting the inner edge of the main belt, and 6 youngSun(see,e.g.,McSweenetal.2006). Fromtheanalysis the 3:1 mean-motionresonance (MMR) with Jupiter, separat- of abolidetrajectory,itis possibletoreconstructits heliocen- ing the inner from the middle belt, were the most effective, tricorbitandtofinditsparentbody(e.g.,Gounelleetal.2006), compared to the 5:2 resonance, for instance, which tends to butsuchdeterminationshavebeenlimitedtoafewobjectsonly ejectasteroidsfromthesolarsystem(seeMorbidellietal.2002, (Rudawskaetal.2012). for a review). The major role playedby the ν resonancewas 6 Amongthedifferentdynamicalclassesofasteroids,thenear- confirmed by the comparison between the reflectance spectra EarthandMars-crosserasteroids(NEAsandMCs), whoseor- ofthemostcommonmeteorites,theordinarychondrites(OCs, 80%ofallmeteoritefalls),thedominantclassinthenear-Earth space, the S-type asteroids (about65% of the observedpopu- Emailaddress:[email protected](B.Carry) PreprintsubmittedtoIcarus January12,2016 lation,Binzeletal.2004),andthedynamicalfamilyofS-types 2. VisiblephotometryfortheSloanDigitalSkySurvey asteroidslinkedwith(8)Floraintheinnerbelt(Vernazzaetal. 2.1. TheSloanDigitalSkySurvey 2008). The NEAs also represent ideal targets for space explo- TheSloanDigitalSkySurvey(SDSS)isa wide-fieldimag- ration owing to their close distance from Earth. This prox- ingsurveydedicatedtoobservinggalaxiesandquasarsatdiffer- imity is quantified by the energy required to set a space- entwavelengths. From1998to2009,thesurveycoveredover craft on a rendezvous trajectory and is often expressed as ∆v 14,500 square degrees in 5 filters: u′, g′, r′, i′, z′ (centered (in km/s), the required change in speed. This is the reason on 355.1, 468.6, 616.5, 748.1 and 893.1 nm), with estimated why the first mission to an asteroid targeted the Amor (433) limitingmagnitudeof22.0,22.2,22.2,21.3,and20.5for95% Eros(Veverkaetal.2000),whyallthetargetsofsample-return completeness(Ivezic´etal.2001). missions were selected among NEAs: (25143) Itokawa for JAXA Hayabusa (Fujiwaraetal. 2006), (101955) Bennu for 2.2. TheMovingObjectCatalog NASAOSIRIS-REx(Origins-SpectralInterpretation-Resource In the course of the survey, 471,569 moving objects were Identification-Security-RegolithExplorer,Laurettaetal.2011), identifiedintheimagesandlisted inthe MovingObjectCata- (162173)RyuguforJAXAHayabusa2(Yanoetal.2010),and logue (SDSS MOC, currently in its 4th release, including ob- (175706) 1996FG3 and (341843) 2008EV5 for the former servationsthroughMarch 2007). Amongthese, 220,101were ESAM3/M4candidateMarcoPolo-R(Baruccietal.2012)and successfullylinkedto104,449uniqueobjectscorrespondingto ARM (AsteroidRedirectMission,Abelletal. 2015), andwhy known asteroids (Ivezic´etal. 2001). The remaining 251,468 the recentpropositionfor a demonstrationprojectof an aster- moving objects listed in the MOC corresponded to unknown oid deflection by ESA, AIDA (Asteroid Impact & Deflection asteroidsatthetimeoftherelease(August2008). Assessment, Murdochetal. 2012), targets the NEA (65803) First, we keep objects assigned a number or a provisional Didymos. This latter point, the protection from asteroid haz- designation only, i.e., those for which we can retrieve the or- ard,iscertainlythemostfamousaspectoftheasteroidresearch bital elements. Among these, we select the near-Earth and knowntothegeneralpublic,andhastriggeredmanyinitiatives Mars-crossers asteroids according to the limits on their semi- leading to breakthroughsin NEA discovery and characteriza- majoraxis,perihelion,andaphelionlistedinTable1,resulting tion of their surface and physical properties (see, e.g., Binzel in2071observationsof1315uniqueobjects. We thenremove 2000; Stokesetal. 2000; Ostroetal. 2002; Binzeletal. 2004; observations that are deemed unreliable: with any apparent Jedickeetal. 2007; Mainzeretal. 2011a; Muelleretal. 2011, magnitudesgreaterthanthelimitingmagnitudesreportedabove amongothers). (Section2.1),oranyphotometricuncertaintygreaterthan0.05. InbothattemptingtolinkNEAsandMCstransientpopula- These constraintsremovea largeportionof the dataset(about tionswith theirsourceregionsandmeteoritesanddesigninga 75%), primarily due to the larger typical error for the z′ fil- protection strategy, the study of their composition is key. In- ter. Whilethereisonlyasmallsubsetofthesampleremaining, deed, dynamicalstudiesallowstodeterminerelativeprobabil- weareassuredofthequalityofthedata(seeDeMeoandCarry ities of the origin of asteroids belonging to those populations 2013,foradditionalinformationonthedefinitionofphotomet- (e.g., Bottkeetal. 2002; Greenstreetetal. 2012). These links riccuts). Additionally,forhighererrors,theambiguityamong are however not sufficient, and must be ascertained by com- taxonomicclassespossibleforanobjectbecomessolargethat positionalsimilarities(Vernazzaetal.2008;Binzeletal.2015; theclassification(Section3)becomesessentiallymeaningless. Reddyetal.2015). Moreover,differentcompositionsyielddif- Inthisselectionprocess,wekept588observationsof353indi- ferent densities and internal structure/cohesion (Carry 2012), vidualasteroidsfromtheSDSSMOC4,aslistedinTable1. andanasteroidonaimpacttrajectorywithEarthofagivensize will requirea differentenergyto bedeflectedordestroyedac- 2.3. IdentifyingunknownobjectsintheMOC4 cordingtoitsnature(JutziandMichel2014). Asmentionedabove,morethanhalfoftheMOC4entrieshad Here,weaimatclassifyingalargenumberofnear-Earthand notbeenlinkedwithknownasteroids.Atthetimeoftherelease Mars-crosserasteroidsintobroadcompositionalgroupsbyus- (August 2008), about 460,000 asteroids had been discovered ing imaging archivaldata. We presentin Section 2 the proce- and 350,000were numbered (i.e., had well-constrained orbits dureweusedtoretrievethephotometryatvisiblewavelengths allowing easy cross-matching with SDSS detected sources). fromthepubliclyavailableimagesoftheSloanDigitalSkySur- Thecurrentnumberofdiscoveredasteroidshasnowrisenabove vey (SDSS). We describe in Section 3 how we use the SDSS 700,000,withmorethan370,000numberedobjects. Wethere- photometrytoclassifytheobjectsintothecommonly-usedBus- foreusetheimprovedcurrentknowledgeontheasteroidpopu- DeMeo taxonomyofasteroids(DeMeoetal. 2009), following lationtolinkunknownMOCsourcestoknownobjects. the work by DeMeoandCarry (2013). We present the results We use the Virtual Observatory (VO) SkyBoT cone-search oftheclassificationinSection4beforediscussingtheirimpli- service (Berthieretal. 2006), hostedat IMCCE1, for that pur- cationsfor sourceregionsin Section5 and forsurface rejuve- pose. SkyBoT pre-computes weekly the ephemeris of all nationprocessesinSection6. known Solar System objects for the period 1889-2060, and 1http://vo.imcce.fr/webservices/skybot/ 2 Class a(au) q(au) Q(au) MOC4 SVO-MOC SVO SVO griz gri min. max. min. max. min. Atens – 1.0 – Q♂ 0.983 – – 10 1 Apollos 1.0 – – 1.017 – 14 18 82 70 Amors – – 1.017 1.3 – 29 73 111 40 Mars-crosser – – – Q♂ – 310 383 622 567 Total – – – – – 353 474 825 678 Table 1: Definition ofthe dynamical classes ofnear-Earth andMars-crosser asteroids usedinpresent study, basedonthesemi-major axis (a), perihelion (q), andaphelion(Q)oftheirorbit. AlltheobjectshaveaperihelioninwardofMarsaphelion(Q♂)at1.666au. ThenumberofobjectslistedintheSDSSMOC4 (Section2.2),identifiedinSDSSMOC4usingSkyBoT(SVO-MOC,Section2.3),andbytheSVONEAproject(Section2.4)arealsolisted. SeeFig.1forthe distributionoftheseobjectsintheorbitalelementspace. 80 keepobjectswithahighprobabilitytobelinkedwiththeMOC Main-belt Near-Earth sources. Atens To furtherrestrictthe numberof false-positiveassociations, 60 Apollos Amors we comparetheposition,apparentmotion,andapparentmag- on () Mars-crosser nitude of the MOC sources to that predicted from ephemeris atio 40 provided by SkyBoT, based on the database of orbital ele- n ments AstOrb2. We consider successful association of SDSS cli n sources with SkyBoT entry if the positions are closer than I 20 30′′,theapparentV-Johnsonmagnitudesdonotdifferbymore than0.5, andthe apparentmotionsare co-linear(differencein dαcos(δ)/dt and dδ/dt of less than 3′′/h). However, neither 0 SkyBoTnorMOC4provideestimatesontheuncertaintyinthe 0.8 q=1.0 apparent velocity. The only information is the uncertainty in q=1.3 thevelocitycomponentsparallelandperpendiculartotheSDSS 0.6 scanning direction. The mean value of this error (both in the y q=1.67 parallelandin the perpendiculardirection)is of 1′′/h. We are cit takingthisvalueasonestandarddeviationtosetthecutabove. ntri 0.4 Q Of the 251,468 unidentified MOC sources, SkyBoT provides ce = c 1 known asteroids within 30 arcseconds for 68,497 (27%), cor- E .0 responding to 41,055 unique asteroids. We trim this value to 0.2 57,646(36,730asteroids)forwhichtheassociationcanbecon- sideredcertain.Thevastmajorityofthesenow-identifiedaster- 0.0 oidshaveorbitswithinthemainbelt(35,404,correspondingto 0.5 1.0 1.5 2.0 2.5 3.0 3.5 96%),butsomeareNEAs(48,0.1%),orMCs(73,0.2%).Their Semi-major axis (au) respectivenumbersare reportedin Table 1. The complete list ofMOCentriesassociatedtoknownasteroids(277,747entries Figure1:DistributionoftheNEAsandMars-crossersstudiedhereasafunction associatedto141,388asteroids)isfreelyaccessible3. of their osculating elements (semi-major axis, eccentricity, and inclination). Theblackdotsrepresentthefirst10,000main-beltasteroids,thegreydotsthe 532NEAswithspectralclassificationfromtheliterature(seeSection3.2),and 2.4. TheSVONear-EarthAsteroidsRecoveryProgram theblackcircles,redsquares,bluetriangle,andgreenstarstheAtens,Apollos, Inaddition,wesearchtheimagesoftheSDSSforNEAsand Amors,andMars-crossersweclassifyfromSDSS. MCs that were either not identified as moving objects by the automaticSDSSpipeline,rejectedbytheMOCdataselection4, orimagedafterthelatestcompilationoftheSDSSMOC4(i.e., stores their heliocentricpositions with a time step of 10 days, observedafterMarch2007). Indeed,onlymovingobjectswith allowingfastcomputationofpositionsatanytime. Thecone- an apparent motion between 0.05 and 0.050 deg/day were in- searchtoolallowstorequestthelistofknownobjectswithina cludedintheMOC,leavingasignificantfractionofNEAsun- fieldofviewatanygivenepochasseenfromEarthintypically cataloged(Solanoetal.2013). less than 10s. We send 251,468 requests to SkyBoT, corre- We use the resources of the citizen-science project “Near- sponding to the 251,468 unknownobjects in the MOC4, cen- EarthAsteroidsRecoveryProgram”oftheSpanishVirtualOb- teredontheMOC4object’scoordinates,atthereportedepoch servatory (SVO) which was originally designed for this very ofobservation,withinacircularfieldofviewof30arcseconds. Although many asteroids among the 700,000knownhave po- 2http://asteroid.lowell.edu/ sition uncertainty larger than this value (as derived from their 3http://svo2.cab.inta-csic.es/vocats/svomoc orbital parameter uncertainty), this cut ensures that we only 4http://www.astro.washington.edu/users/ivezic/sdssmoc/sdssmoc.html 3 purpose: to identify and measure the astrometry of NEAs in g r archival imaging data (Solanoetal. 2013). For each Aten, 22 N = 71 N = 531 griz griz Amor, Apollo, and Mars-crosser listed by the Minor Planet N = 57 N = 469 Center5 (MPC),itsephemerisarecomputedovertheperiodof gri gri 20 operationoftheSDSSimagingsurvey(1998to2009)andcom- paredtothefootprintsoftheimagesofthesurvey. Theimages -0.25 0.25 -0.25 0.25 18 possibly containingan objectbrighter than the SDSS limiting 40 40 magnitude (V=22) are then proposed to the public for iden- e tification through a web portal6. Since the beginning of the ud 16 20 20 projectin2011,over2,500astrometrymeasurementsofabout gnit a 600NEAsnotidentifiedintheMOChavebeenreportedtothe SS m 22 i z MPC(seeSolanoetal.2013,fordetailsontheproject). D N = 399 N = 93 S griz griz To compute the photometry of the NEAs measured by the N = 413 gri users we first searched in the photometric catalog of the 8th 20 SDSS Data Release7. If no photometry associated with the -0.25 0.25 -0.25 0.25 NEA was found, we ran SExtractor on the correspondingim- 18 ages and calibrated the SExtractor magnitudes by comparing 40 40 themwiththeSDSSmagnitudesofthesourcesidentifiedinthe 16 20 20 image. Owingtothemorestringentlimitingmagnitudeinthez′ fil- 16 18 2 0 22 16 18 2 0 22 ter,manyasteroidsareidentifiedoverthreebands(g′r′i′)only. SVO magnitude We also report these objects here, although deriving a taxo- nomicclassificationisofcourselessaccurate.Overall,wecol- Figure2:Comparisonoftheapparentmagnitudemeasuredbyoursystem(SVO lect1194fourbands(g′r′i′z′)photometrymeasurementsof825 magnitude)withthatreportedintheSDSSMOC4(SDSSmagnitude)forthe uniqueasteroidsand976three-bands(g′r′i′)photometrymea- fourg′,r′,i′,andz′filters.Foreach,thenumberofcommonobjectsisreported, forbothfour-bands (black circles) andthree-bands (greysquares) sets. The surements of 678 distinct asteroids (Table 1). We present in insertedhistogramsshowthedistributionofthemagnitudedifference(MOC- Fig.2acomparisonofourmeasurementswiththemagnitudes SVO) reportedintheSDSSMOC4forthecommonasteroidsinboth sets,showingexcellentagreement(valuesagreewithastandard deviationof0.05mag). unityinfilter g′. Second,wecomputetheslopeofthecontin- uum over the g′, r′, and i′ filters (hereaftergri-slope), and the z′−i′ color(hereafterzi-color),representingthebanddepthof 3. Taxonomicclassification apotential1µmband,becausetheyarethemostcharacteristic spectraldistinguishersinallmajortaxonomies(beginningwith TheSDSSphotometryhasbeenusedtoclassifyasteroidsac- Chapmanetal. 1975). The classification into the taxonomyis cordingtotheircolorsbymanyauthors(e.g.,Ivezic´etal.2002; thenbasedonthesetwoparameters. Nesvorny´ etal. 2005; Parkeretal. 2008; Carvanoetal. 2010). As a results of the limited spectral resolution and range One key advantage of the survey was the almost simultane- of SDSS photometry, we group together certain classes into ousacquisitionofallfilters(5minintotal),hencelimitingthe broadercomplexes(seecorrespondencesinTable2).Foraster- impact of geometry-relatedlightcurveon the apparentmagni- oidswithmultipleobservationsthatfallundermultipleclassifi- tude. Here we follow the work by DeMeoandCarry (2013, cations,weusethetree-likeselectiontoassignafinalclass(see 2014) in whichthe class definitionsareset to be as consistent DeMeoandCarry(2013)fordetailsandDeMeoetal.(2014a) aspossiblewithpreviousspectraltaxonomiesbasedonhigher foranexampleofaspectroscopicconfirmationcampaignofthe spectral resolution and larger wavelength coverage data sets, SDSS classification used here). We successfully classify 982 specifically Bus and Bus-DeMeo taxonomies (BusandBinzel asteroidsfromthesampleof1015near-EarthandMars-crosser 2002;DeMeoetal.2009). Wepresentconciselytheclassifica- asteroidswithfour-bandsphotometry(i.e.,97%ofthesample). tionschemebelowandrefertoDeMeoandCarry(2013)fora For objects with three-bands photometry only, we set their completedescription. z′ magnitude to the limiting magnitude of 20.5 (Ivezic´etal. 2001) as an upper limit for their brightness. We then classify 3.1. FromSDSStoBus-DeMeotaxonomy theseasteroidsusingtheschemepresentedabove. Becausethe magnitudeof20.5inz′isanupperlimit,theactualzi-colorfor First, we convertthe photometryinto reflectance (using so- these asteroids may be overestimated. The classification can lar colors from Holmbergetal. 2006) and normalize them to thereforebe degenerated,all the classes with similar gri-slope andlowerzi-colorbeingpossible. Weassigntentativeclassifi- cationto254asteroidsfromthesampleof678near-Earthand 5http://minorplanetcenter.org/ 6http://www.laeff.cab.inta-csic.es/projects/near/main/ Mars-crosserasteroidswiththree-bandsphotometry(i.e.,37% 7http://cdsarc.u-strasbg.fr/viz-bin/Cat?II/306 ofthesample). 4 SDSS&Literature Bus-DeMeo Diameter: 10.0 1.0 0.1 (km) 110044 A,AR,AS A B B 110033 C,C:,Cb,Cg,Ch C ersers D,DT D crosscross 110022 K,K: K ars-ars- MM L,Ld L 110011 Q,Q/R,R,RQ Q 11 O,Q/R/S,R,RS S MPC S,S:,S(I→V),Sa,Sk,Sl,Sq,Sq:,Sr S oidsoids 110033 Spectra-only V,V: V erer Common spectra/SDSS E,M,P,X,X:,XT,Xc,Xe,Xk X h Asth Ast 110022 SDSS U,C(u),R(u),S(u),ST,STD,QX U artart EE ar-ar- 110011 ee Table2: Correspondencebetweentheclassesfromdifferenttaxonomies(e.g., NN TholenandBarucci1989;BusandBinzel2002)foundintheliterature,usedto 11 classifytheSDSSphotometry(inbold),andtheirequivalence inthereduced 1100 1155 2200 2255 versionofthetaxonomybyDeMeoetal.(2009)adoptedhere, followingthe AAbbssoolluuttee mmaaggnniittuuddee ((HH)) work by DeMeoandCarry (2013). We strive to preserve the most extreme classes (like A,B)andweconvert thetentative classification (e.g., RS)into Figure3:Distributionoftheknowntaxonomyfornear-EarthandMars-crosser theirbroader,safer,complex(S).Wealsoconsiderasunknown(U)anydataof asteroids as a function of their absolute magnitude H. The white histogram insufficientquality. correspondstothetotalnumberofdiscoveries(listedinAstOrb),thegreytothe taxonomicclassificationfoundintheliterature,thebluetotheoverlapbetween literatureandpresentstudy,andtheredthefour-bandsset. Anapproximative Inallcases,wemarkobjectswithpeculiarspectralbehavior conversion todiameter is alsoreported, using D = 1329× pv−0.5 ×10−0.2H, with the historicalnotation“U” (forunclassified), anddiscard withthegeometricalbedo pvtakenas0.2(themajority,≈60%,oftheobjects classifiedherebeingS-types). them fromthe analysis. Thereare 33and424asteroidsin the four-bands and three-bands photometry samples respectively for which we cannot assign a class. These figures highlight photometry.The254asteroidswiththree-bandsphotometryare theambiguityraisedbythelackofinformationonthepresence listedseparatelyinTableB.2,becausetheirtaxonomicclassifi- orabsenceofanabsorptionbandaroundonemicron,towhich cationislessrobust. Inmanycases,theupperlimitof20.5for thez′filterissensitive. their z′ magnitude provides a weak constraint on their taxon- omy,andclasseswithhighzi-color(mainlyV-types)aremore 3.2. Gatheringclassificationsfrompaststudies easilyidentified.Thissamplebasedonthree-bandsphotometry only is therefore biased, but it can be used as a guideline for Many different authors have reported on the taxonomic selectingtargetsforspectroscopicfollow-ups. We concentrate classification of NEAs. We gather here the results of belowonthesamplebasedonfour-bandsphotometry. Dandyetal. (2003), Binzeletal. (2004), Lazzarinetal. The206NEAspresentedherehaveabsolutemagnitudesbe- (2005), deLeo´netal. (2006), deLeo´netal. (2010), tween12and23.Oursamplefullyoverlapswiththesizerange ThomasandBinzel (2010), Popescuetal. (2011), Ye (2011), of the 523 NEAs characterized by visible/near-infrared spec- Reddyetal. (2011), Polishooketal. (2012), Sanchezetal. troscopypublishedtodateandrepresentsanincreaseof≈40% (2013), and DeMeoetal. (2014b). These authors used differ- ofthecurrentsamplesize(Fig.3). Asignificantfraction(46%) ent taxonomic schemes to classify their observations, using of the NEA population with H<16 (about 2km diameter for either broad-bandfilter photometryor spectroscopy, at visible an albedoof0.20)hasa taxonomicclassification. Forsmaller wavelengths only or also in the near infrared. We therefore diameters,thefractiondropsquicklyto10%andless. Thesub- transposetheclassesofthesedifferentschemes(Tholen1984; kilometerpopulationof NEAs is thereforestill poorlycatego- TholenandBarucci 1989; BusandBinzel 2002; DeMeoetal. rized. 2009) into the single, consistent, set of 10 classes we already Theabsolutemagnitudeofthe776MCsreportedhereranges use for the SDSS data. Here also, we attribute the historical from11to19. Oursamplerepresentsthefirstclassificationof “U” designation for objects with apparently contradictory sub-kilometricMars-crossers,andasixfoldincreasetothesam- classifications (e.g., QX or STD in Ye 2011). These patho- pleof117MCsfromspectroscopy(Fig.3).SimilarlytoNEAs, logical cases represent15% of the objects with multiple class about40%oftheMCpopulationwithH<14(about5kmdiam- determinations. Intotal,wegather1022classificationsfor648 eterforanalbedoof0.20)nowhasataxonomicclassification, objectslistedintheliterature. andthefractiondropsquicklyto10%forsmallerdiameters. 4.1. Taxonomyandorbitalclasses 4. Results We present in Fig. 4 how the different classes distribute We list in Table B.1 the photometry and the taxonomy of among the orbital populations. As already reported by all982near-EarthandMars-crossersasteroidswithfour-bands Binzeletal. (2004), the broad S- (including Q-types), C-, 5 X-complexes, and V-types dominate the NEA population, objects must be injected in the NEA space to explain the cur- the minor classes (A, B, D, K, L) accounting for a few rent observed population. We use our sample of 982 NEAs percentsonly,similarlytowhatisfoundintheinnermainbelt andMCs to identifypossiblesourceregions. For that, we use (DeMeoandCarry 2013, 2014). We find that the S-complex the source region mapper8 by Greenstreetetal. (2012), built encompassestwiceasmanyobjectsastheC-andX-complexes, on the result of numericalsimulationsof the orbitalevolution compared to the threefold difference reported by Binzeletal. of test particules in the five regions defined by Bottkeetal. (2004). (2002): ν secularresonance,3:1mean-motionresonancewith 6 ThedistributionoftaxonomicclassesamongMCsissimilar Jupiter,Mars-crossers(MC),OuterBelt(OB),andJupiterFam- to that of Apollos and Amors (with only 8 Atens, our sample ilyComet(JFC). suffers from low-number statistics). Our findings of V-types For each object, we compute its probability P to originate i accounting for roughly 5% of all MCs may therefore seem from the ith source region. We then normalize all P for each i puzzling considering the lack of V-types among the ≈100 sourceregion. ThesumofthenormalizedP overagiventax- i classified MCs highlighted by Binzeletal. (2004). It is, onomic class therefore represents the fraction of objects (by however,anobservationselectioneffect.Indeed,Mars-crossers number) of this class in the source region (Table 4). Uncer- are inner-main belt (IMB) asteroids which eccentricity has tainties are computed from the source region mapper uncer- been increased by numerous weak mean-motion resonances tainties, quadraticallyadded with the margin of error (at 95% (“chaotic diffusion”, see MorbidelliandNesvorny´ 1999). The confidencelevel)to accountforthe sample size. We can then IMBhostingthelargestreservoirofV-typesinthesolarsystem compare these predicted fractions to the observed distribution (Vestoids,BinzelandXu1993),V-typeswereexpectedinMC oftaxonomicclassesforeachsourceregion. space. ThevastmajorityofNEAswithtaxonomicclassificationsus- The size distribution of known Vestoids however peaks taindiametersoflessthan5km(absolutemagnitudeabove14). at H=16 (about 1.5km diameter), and the largest members Because thedistributionoftaxonomicclassesin the mainbelt have an absolute magnitudeof 14 (we use the list of Vestoids varieswithdiameter(DeMeoandCarry2014),weneedtocom- fromNesvorny2012). All the V-typesidentifiedhere havean parethepredictedfractionofTable4withobjectswithH>14 absolutemagnitudeabove14,andsodoes(31415)1999AK23 in each source region. For the first time, thanks to the large (H=14.4), the first V-type among MCs reported recently by dataset of asteroidsprovidedby the Sloan Digital Sky Survey Ribeiroetal.(2014).Only33MCshadbeencharacterizedwith (SDSS), suchinformationis availableforthe innerpartof the thisabsolutemagnitudeorhighertodate,andthepreviouslack mainbelt(DeMeoandCarry2013,2014)andisreportedinTa- ofreportofV-typesamongMCsisconsistentwithourfindings. ble4(3to5kmdiameterrange). The comparison of the ν and 3:1 source regions (delim- 6 iting the inner belt) with the observations in Fig. 5 shows a 4.2. Low-δvasspacemissiontargets very good match of the distributions (correlation coefficient of 0.99, maximum difference of 5%). This validates the Withinthe206NEAserendipitouslyobservedbytheSDSS, dynamical path from the inner belt to the near-Earth space we identify 36 potential targets for space missions based on as described in the model by Greenstreetetal. (2012). Al- their accessibility. We select all the NEA with a δv below though the inner belt is widely accepted as a major source 6.5km/s. As a matter of comparison, the requiredδv to reach forNEAsandmeteorites(e.g.,Bottkeetal.2002;Binzeletal. the Moon and Mars are of 6.0 and 6.3km/s (e.g., Abelletal. 2004;Vernazzaetal.2008;Binzeletal.2015),therelativecon- 2012). We listinTable3thebasiccharacteristicsofthesepo- tributionofthedifferentsourceregionsparticularlywithrespect tential targets, together with the targets already, or planned to toasteroidsizeisstillamatterofdebate. Forthefirsttime,we be,visitedbyspacecraft. compareherepopulationsofthesamesizerange. Among the list of low-δv objects, we find a large majority Unfortunately,thereisnosimilardatasetfortheothersource ofS-types,followingtheirdominanceinthesamplepresented regions (MC, OB, and JFC) to be compared with our predic- here of about60%. We, however,note the presenceof poten- tion.Wecanstillnotetheoveralltrendofincreasingfractionof tialD-,L-,andK-types. Todate,ofthe24taxonomicclasses, C/D/X-typesin OBandJFC comparedwiththe innerregions, onlyC-(Mathilde,Ceres),S-(Ida,Eros,Gaspra,Itokawa,and as expected. The fraction of K-types peaks in the outer belt, Toutatis),Xe-(Steins),Xk-(Lutetia),andV-types(Vesta)have place of the Eos family, also as expected, although a steady beenvisitedbyspacecraft.ThesepotentialD-,L-,andK-types distributionofK-typesacrosssourceregionsisalsoconsistent targetsmayrepresentgoodopportunitiesforexploration. Data withinuncertainties.Wealsonoteastrongcorrelationbetween in the visible can only suggest the presence of an absorption MCandsmallinner-beltasteroidpopulations(correlationcoef- bandat1µm,andnear-infrareddataisrequiredtoconfirmthese ficient0.97),whichisnotrelatedtotheoriginofNEAs,butof potentialclassifications. MC themselves, via chaotic diffusion from the inner belt (see MorbidelliandNesvorny´ 1999;Micheletal.2000). 5. Sourceregions ApeculiarfeatureofTable4isthehighfractionofS-typein The population of NEAs being eroded on short timescale 8UpdatedbyS.Greenstreetfromtheoriginalmappertoincludetheproba- (<10My)byplanetarycollisionsanddynamicalejections,new bilityofthesourceregionsoftheMCsthemselves. 6 all source regions. Based on presentset of NEAs and the dy- namicalmodelbyGreenstreetetal.(2012),oneasteroidbelow 5kmdiameteroutoffourshouldbeanS-typeintheouterbelt and the same applies to Jupiter family comets. Although the census of composition in this size range (3–5km) is far from being completein these regions, there is a biastoward detect- 4 50 Aten: 8 ing S-typesdue to their high albedo(≈0.20)comparedto that oftheC/P/D-typesfoundthere(around0.05,seeMainzeretal. 40 Ngri 3 2011b; DeMeoandCarry 2013, for albedo averages over tax- N griz 30 onomic classes) S-types are minor contributors to the outer belt for diameters above 5km (DeMeoandCarry 2014) and 20 searchesforS-typematerialhavebeenunsuccessfulamongCy- 1 beles, Hildas, andTrojans(seeEmeryandBrown2003,2004; 10 Emeryetal.2011;Fornasieretal.2004,2007;YangandJewitt 0 2007, 2011; Roigetal. 2008; Gil-HuttonandBrunini 2008; A B C D K L Q S V X Marssetetal.2014). However,only4%ofthesample(42ob- jects) are predicted to originate from the OB and JFC source 50 Apollo: 106 regions. The results for these regionsis, thus, based on small 40 Ngri numberstatistics. ThelargefractionofS-typesinthesesource N regions suggests nevertheless that dynamical models may re- 30 32 griz 30 quirefurtherrefinements. 20 18 ) % 11 ( 10 n 4 4 5 o 1 1 i 0 t la A B C D K L Q S V X u p po 50 Amor: 116 51 he 40 Ngri t f N o griz n 30 o i ct 20 21 a r 15 F 10 9 5 5 3 3 3 1 0 A B C D K L Q S V X 434 50 MC: 873 40 Ngri N griz 30 20 160 116 10 23 24 22 31 26 34 0 3 A B C D K L Q S V X Taxonomic classes Figure4:Distributionofthetaxonomicclassesforeachdynamicalgroup:Aten, Apollo,Amor,andMars-crosser.Thenumberofobjectwithfour-bandphotom- etryineachclassisreported.Emptyandfilledbarsstandsforthethreeandfour bandphotometrysamplesrespectively. 7 60 Inner Main Belt 3-5 km NEAs from ν 6 NEAs from 3:1 MMR %) Fraction of the population ( 2400 0 A B C D K L S V X Figure5: Relativefractions(bynumber)oftaxonomicclassesforasteroidsin theinnerpartofthemain-belt(2.0–2.5au)withdiameterbetween3and5km (computed fromDeMeoandCarry2013,2014)compared withthepredicted fractionsoriginatingfromν6and3:1resonances(seeSec.5) 8 Designation Type Class δv(km/s) H s(%/100nm) z-i a(au) e i(o) 2004EU22 X Apollo 4.420 23.00 0.78 0.044 1.175 0.162 5.3 1996XB27 D Amor 4.750 22.00 0.85 0.123 1.189 0.058 2.5 2000TL1 C Apollo 4.870 22.00 -0.09 0.002 1.338 0.300 3.6 2001QC34 Q Apollo 4.970 20.00 0.69 -0.228 1.128 0.187 6.2 1999FN19 S Apollo 5.020 22.00 0.96 -0.135 1.646 0.391 2.3 2000SL10 Q Apollo 5.080 21.00 0.86 -0.189 1.372 0.339 1.5 1994CN2 S Apollo 5.150 16.00 1.12 -0.073 1.573 0.395 1.4 2002LJ3 S Amor 5.280 18.00 1.07 -0.258 1.462 0.275 7.6 2004UR C Apollo 5.320 22.00 0.13 -0.025 1.559 0.406 2.4 2006UP S Amor 5.350 23.00 1.43 -0.074 1.586 0.301 2.3 1994CC S Apollo 5.370 17.00 0.99 -0.210 1.638 0.417 4.7 2010WY8 K Amor 5.670 21.00 0.92 -0.042 1.385 0.136 6.0 2002XP40 S Amor 5.720 19.00 1.63 -0.092 1.645 0.296 3.8 1993QA D Apollo 5.740 18.00 1.02 0.174 1.476 0.315 12.6 2004RK9 C Amor 5.760 21.00 0.14 -0.050 1.837 0.426 6.2 2001FC7 C Amor 5.780 18.00 0.16 -0.034 1.436 0.115 2.6 1977VA C Amor 5.940 19.20 0.34 -0.010 1.866 0.394 3.0 2001WL15 C Amor 6.000 18.00 0.34 -0.179 1.989 0.475 6.9 2000XK44 L Amor 6.080 18.00 1.24 0.020 1.724 0.385 11.2 2003BH V Apollo 6.090 20.00 1.79 -0.273 1.456 0.356 13.1 2000NG11 X Amor 6.130 17.00 0.27 0.089 1.881 0.368 0.8 2000RW37 C Apollo 6.150 20.00 0.45 -0.166 1.248 0.250 13.8 2001FD90 V Amor 6.200 19.00 0.75 -0.419 2.046 0.478 7.3 2002PG80 S Amor 6.210 18.00 1.09 -0.225 2.013 0.438 4.4 2004VB S Apollo 6.260 20.00 1.04 -0.200 1.458 0.409 10.9 1993DQ1 S Amor 6.270 16.00 1.20 -0.207 2.036 0.493 10.0 2000YG4 Q Amor 6.300 20.00 0.61 -0.155 2.211 0.503 2.6 2004KD1 C Amor 6.330 17.00 0.13 -0.110 1.720 0.331 10.1 2004RS25 C Amor 6.410 20.00 0.39 -0.057 2.128 0.479 6.7 2004QZ2 S Amor 6.470 18.00 -5.74 -0.227 2.260 0.495 1.0 2001FY S Amor 6.530 18.00 1.36 -0.132 1.886 0.327 4.7 2009OC S Amor 6.540 20.00 0.85 -0.153 2.137 0.446 4.6 2004XM35 S Amor 6.560 19.00 1.13 -0.111 1.837 0.301 5.4 2005QG88 K Apollo 6.560 20.00 0.99 -0.056 1.728 0.493 11.3 1999KX4 V Amor 6.580 16.00 1.60 -0.432 1.457 0.293 16.6 2002TY57 S Amor 6.600 19.00 0.86 -0.154 1.922 0.327 3.5 Itokawa S Apollo 4.632 19.20 1.324 0.280 1.6 Bennu C Apollo 5.087 20.81 1.126 0.204 6.0 Ryugu B Apollo 4.646 19.17 1.189 0.190 5.9 Didymos X Apollo 5.098 17.94 1.644 0.384 3.4 Table3:ListofNEAswithalowδv,hencepotentialtargetsforspacemissions.ForeachNEA,thetaxonomicclassdeterminedhereisreported,togetherwiththe dynamicalclass,theabsolutemagnitudeandtheorbitalelements.Thetargetsoftheasteroid-deflectionmissionAIDA(Didymos)andofthereturn-samplemissions Hayabusa(Itokawa),OSIRIS-REx(Bennu),andHayabusa2(Ryugu)arealsoincludedforcomparison. 9 Class Sourceregions IMB 3−5km MC ν MMR OB JFC (#) (%) 6 3:1 A 0.4±5.9 0.5±6.3 0.3±5.5 0.0±3.0 0.0± 3.0 9 0.4 B 2.4±5.8 1.7±5.4 2.9±6.1 6.1±7.5 0.4± 4.3 34 1.5 C 15.7±5.2 20.5±5.4 22.7±5.6 38.6±6.6 47.3±10.6 589 25.7 D 2.5±5.7 2.3±5.6 2.5±5.8 5.8±7.2 9.7± 8.5 20 0.9 K 2.8±5.9 2.5±5.8 2.4±5.7 4.8±6.9 1.8± 7.3 57 2.5 L 3.8±5.9 3.4±5.7 4.4±6.1 1.9±5.1 0.4± 4.2 73 3.2 Q 3.6±5.6 4.3±5.8 6.3±6.4 3.8±5.8 2.4± 5.5 0 0.0 S 52.8±4.6 44.5±4.6 42.0±4.8 23.8±4.8 23.4± 6.9 1145 50.0 V 13.0±5.5 15.6±5.7 12.5±5.5 7.0±5.0 3.0± 4.7 247 10.8 X 3.2±5.5 4.7±6.0 4.0±5.8 8.1±7.2 11.7±10.3 117 5.1 Table4: Relativefractionofeachtaxonomicclass(bynumberofobjects)fromeachsourceregiondefinedbyBottkeetal.(2002)andGreenstreetetal.(2012), comparedwiththedistributionoftaxonomictypesamong3to5kminnerMainBelt(IMB)asteroids(whereQ-typesaremergedwithS-types,seeDeMeoandCarry 2013). 10