• Non ci sono risultati.

Mineralogical analysis of the Ac-H-6 Haulani quadrangle of the dwarf planet Ceres

N/A
N/A
Protected

Academic year: 2021

Condividi "Mineralogical analysis of the Ac-H-6 Haulani quadrangle of the dwarf planet Ceres"

Copied!
19
0
0

Testo completo

(1)

Mineralogical analysis of the Ac-H-6 Haulani quadrangle of the dwarf planet Ceres

Title

TOSI, Federico; CARROZZO, FILIPPO GIACOMO; ZAMBON, Francesca;

CIARNIELLO, Mauro; FRIGERI, ALESSANDRO; et al.

Authors

10.1016/j.icarus.2017.08.012

DOI

http://hdl.handle.net/20.500.12386/29405

Handle

ICARUS

Journal

318

Number

(2)

ContentslistsavailableatScienceDirect

Icarus

journalhomepage:www.elsevier.com/locate/icarus

Mineralogical

analysis

of

the

Ac-H-6

Haulani

quadrangle

of

the

dwarf

planet

Ceres

F.

Tosi

a,∗

,

F.G.

Carrozzo

a

,

F.

Zambon

a

,

M.

Ciarniello

a

,

A.

Frigeri

a

,

J.-Ph.

Combe

b

,

M.C.

De

Sanctis

a

,

M.

Hoffmann

c

,

A.

Longobardo

a

,

A.

Nathues

c

,

A.

Raponi

a

,

G.

Thangjam

c

,

E.

Ammannito

d

,

K.

Krohn

e

,

L.A.

McFadden

f

,

E.

Palomba

a

,

C.M.

Pieters

g

,

K.

Stephan

e

,

C.A.

Raymond

h

,

C.T.

Russell

i

,

the

Dawn

Science

Team

a INAF-IAPS Istituto di Astrofisica e Planetologia Spaziali, Via del Fosso del Cavaliere, 100, I-00133 Rome, Italy b Bear Fight Institute, 22, Fiddler’s Road, P.O. Box 667, Winthrop, WA 98862, USA

c Max Planck Institute for Solar System Research, Justus-von-Liebig-Weg 3, d-37077 Göttingen, Germany d Italian Space Agency (ASI), Via del Politecnico snc, I-00133 Rome, Italy

e Institute of Planetary Research, German Aerospace Center (DLR), Rutherfordstrasse 2, d-12489 Berlin, Germany f NASA/Goddard Space Flight Center, Greenbelt, 8800 Greenbelt Road, MD 20771, USA

g Department of Earth, Environmental and Planetary Sciences, Brown University, 324 Brook Street, Providence, RI 02912, USA h NASA/Jet Propulsion Laboratory and California Institute of Technology, 4800 Oak Grove Drive, Pasadena, CA 91109, USA

i Institute of Geophysics and Planetary Physics, University of California at Los Angeles, 3845 Slichter Hall, 603 Charles E. Young Drive, East, Los Angeles, CA

90095-1567, USA

a

r

t

i

c

l

e

i

n

f

o

Article history: Received 29 April 2017 Revised 4 August 2017 Accepted 9 August 2017 Available online xxx Keywords: Asteroids surfaces Asteroid Ceres Spectrophotometry Infrared observations

a

b

s

t

r

a

c

t

Ac-H-6‘Haulani’isoneoffivequadranglesthatcovertheequatorial regionofthedwarfplanet Ceres. Thisquadrangleisnotableforthebroad,spectrallydistinctejectathatoriginatefromthecraterHaulani, whichgivesthenametothequadrangle. Theseejecta exhibitoneofthemostnegative(‘bluest’) visi-bletonearinfraredspectralslopeobservedacrosstheentirebodyandhavedistinctcolorpropertiesas seeninmultispectralcompositeimages.BesidesHaulani,hereweinvestigateabroaderareathatincludes othersurfacefeaturesofinterest, withanemphasis onmineralogyas inferredfromdataobtainedby Dawn’sVisibleInfraRedmappingspectrometer(VIR),combinedwithmultispectralimageproductsfrom theDawnFramingCamera(FC)soastoenableaclearcorrelationwithspecificgeologicfeatures.

Ouranalysis showsthatcraterHaulanistandsoutcomparedtoothersurfacefeaturesofthe quad-rangle.Albedomapsobtainedinthenearinfraredrangeat1.2μmand1.9μmrevealthatthefloorand ejectaofHaulaniareindeedapatchworkofbrightanddarkmaterialunits.Visibletonear-infrared spec-tralslopesdisplaynegativevaluesincraterHaulani’sfloorandejecta,highlightingbluish,younger ter-rains.Diagnosticspectralfeaturescenteredat∼2.7μmand∼3.1μmrespectivelyindicateasubstantial de-creaseintheabundancesofmagnesium-bearingphyllosilicatesandammoniatedphyllosilicatesincrater Haulani’sfloorandbrightejecta.Similar,butlessprominent,spectralbehaviorisobservedinother ge-ologicfeaturesofthisquadrangle,whilethegeneraltrendinquadrangleAc-H-6forthesetwomineral speciesistoincreasefromthenorthwesttothesoutheast.However,itisworthnotingthatthe corre-lationbetweenthe∼2.7μmand∼3.1μmspectralparametersisgenerallystrongintheHaulanicrater’s area,butmuchweakerelsewhere,whichindicatesavariabledegreeofmixingbetweenthesetwomajor mineralphasesinmovingawayfromthecrater.Finally,theregionofcraterHaulanidisplaysadistinct thermalsignatureand alocalenhancementincalciumand possiblysodiumcarbonateminerals,which ishardlyfoundintherestofthequadrangleandislikelytheresultofintensehydrothermalprocesses followingtheimpactevent.TheseevidencesalltogetherconfirmtheyoungageofcraterHaulani,asthey havenotbeenerasedormadeelusivebyspaceweatheringprocesses.

© 2017ElsevierInc.Allrightsreserved.

Corresponding author.

E-mail address: federico.tosi@iaps.inaf.it (F. Tosi).

1. Introduction

At about940kmin meandiameter, the dwarf planetCeres is thelargestandmostmassiveobjectinthemainasteroid belt.Its

http://dx.doi.org/10.1016/j.icarus.2017.08.012

0019-1035/© 2017 Elsevier Inc. All rights reserved.

(3)

lowbulkdensity,2162 ± 8kg/m3 (Russelletal., 2016;Park etal.,

2016),isclosertothatoficysatellitesthanofrockybodies,and re-vealsthatits interiormustcontainasubstantialamountofwater. Morphologicalanalysis ofimpact craters suggeststhat this water islikelyconcentratedinthemantleratherthanintheoutercrust. Exposedwaterice,whichonCeresisthermodynamicallyunstable atlow latitudes (Titus, 2015), has been identified only in a few places,favoredby thelocaltopography(Combeetal., 2016;Platz etal., 2016;Nathues etal., 2017; Combe etal., 2017a, thisissue;

Stephanetal.,2017b,thisissue).

The Haulani quadrangle(Ac-H-6)is one out offive that cover the equatorial region of Ceres (Lon 0°−72°E, Lat 22°S to 22°N,

Roatschetal., 2016b). Itis namedafterHaulani (Lon 10.77°E, Lat 5.8°N), the fourth largest impact crater (diameter 34km) in this quadrangle. Other impact craters with an official designation by theInternational Astronomical Union (IAU)and diameters inthe range20–50kmare:Anura(37km),Kondos(44km),Piuku(31km),

Shakaema(47km),andTahu(25km).

Athoroughgeologicmappingofthisregion(Krohnetal.,2017) established that cratered terrain (crt) is the oldest and broadest unit in quadrangle Ac-H-6 (and more generally on Ceres), dis-playing a high crater density, an intermediate albedo (average value0.03) inFC clearfilterimages,andabrownish colorin en-hancedFC color compositeimages.The second broadest geologic type is crater material, which surroundswell-defined craters ex-cept Haulani,and can be further divided intothree units: crater material (c), bright crater material (cb), and dark crater mate-rial (cd). Crater Haulani is a distinct geologic unit in itself, as it appears to be one of the freshest and youngest surface fea-tures on Ceres,being made up by a complexpatchwork of sub-units of varying extent. It exhibits interior smooth plains with flowfeatures originatingfroma hummocky elongated mountain-ous ridge in the center, ponding toward mass-wasting deposits ofthe rim. Several flow features are observed running fromthe crater rim outward to the surrounding area, covering the pre-existing surface (Krohn et al., 2016). The smooth crater floor is laced by pits. Some pit crater chains in the northwestern part are oriented parallel to the rim (Krohn et al., 2016; Sizemore et al., 2017). Most notable geologic units mapped within crater Haulaniare: brightlobatematerial(lb),knobbybrightlobate ma-terial(lkb), smooth lobate material(ls),andsmooth bright lobate material(lsb).Haulani’s brightejecta(crb) iswidespreadover the cereansurface,preferentiallytothewest,extendinginto quadran-gleAc-H-10 ‘Rongo’(Platzetal., 2017; Zambon etal., 2017b,this issue).

CraterHaulaniwasformedonthetransitionbetweenacentral plateauin theeast anda topographic low in thewest (Krohnet al.,2016;Krohnetal.,2017).AdjacenttothewestofHaulanithere is one named domical mountain: Dalien Tholus (length: ∼22km, width:∼14km).WangalaTholus,locatedinthesouthwest,ismuch larger(length: ∼50km) and iscut almost equallyby the bound-arybetweentheHaulaniandSintanaquadrangles(Schulzecketal., 2017).

Magnesium-andammonium-bearingphyllosilicates,carbonates anda spectrally featurelessdarkendmember yetto beidentified, arethe major surfaceminerals on Ceres (De Sanctiset al., 2015) andare ubiquitousat 1-kmspatialresolution (Ammannito etal., 2016).Ammonia,accreted eitherasorganicmatterorasice, may have reacted withphyllosilicates on Ceres during differentiation, leadingtowidespreadexistenceofammoniatedphyllosilicates(De Sanctisetal.,2015). Variationsinthe strengthoftheir diagnostic absorptionfeatures maybe spatially correlated andindicate con-siderablevariabilityintherelativeabundanceofthephyllosilicates, although their composition is fairly uniform. These data, along withthe distinctive spectral propertiesof Ceres relative to other asteroidsandcarbonaceousmeteorites,indicatethatthe

phyllosili-cateswereformedendogenouslybyagloballywidespreadand ex-tensivealterationprocess(Ammannitoetal.,2016).

Besidesthesetwomineralphases,whichdominatethesurface ofCeres atglobalscale,in high-albedoareasthe relative propor-tions of the components changes, with a clear increase in car-bonate content with respect to the dark material. Cerealia Fac-ula, i.e. the bright unit in the middle of the 92-km crater Oc-cator (in quadrangleAc-H-9 ‘Occator’),is especially rich in anhy-droussodiumcarbonate,anditappearstorepresentthemost con-centrated known occurrence of km-scale carbonates beyond the Earth(DeSanctisetal.,2016).ThenorthernflanksofAhunaMons, the highest mountain on Ceres (located in quadrangle Ac-H-10 ‘Rongo’),alsodisplayalargeramountofMg-Cacarbonatesand Na-carbonatescomparedtosurroundingregions(Zambonetal.,2017a; Longobardoetal.,2017,thisissue).

Waterice-richunitswerediscoveredonlyinabouttenspecific places atthesurface. Thevery firstdetectionoccurredinthe ge-ologicallyyoung 10-kmOxo crater, wherethe relatively high lat-itude(∼42°N)andmorphologyprotectsmuchof thesurfacearea fromdirectsolarilluminationformostofthecerean day(Combe etal., 2016).Otheroccurrencesofwaterice-richunitsmixedwith low-albedo components were later discovered at latitudes pole-ward of ∼30° in fresh craters near rim shadows (Combe et al., 2017a, this issue; Stephan et al., 2017b, this issue) and at some permanently shadowed craters (Platz et al., 2016). Furthermore, aliphatic organic matter was found to be mainly localized on a broadregionof∼1000km2closetothe∼50-kmErnutetcrater(De

Sanctisetal.,2017a).However,sincewatericeandorganicsarenot widespreadontheCeressurfacebutonlyoccurinspecificregions of interest outside the Haulani quadrangle, our analysis will use a subset ofspectral parameters that are meaningful for thisand otherquadrangles,consistentwiththesameapproachadoptedfor otherpapersfoundinthisspecialissue.

Here we present a detailed mineralogic study of the Ac-H-6 Haulani quadrangle,similar to what waspreviously done forthe

Av-10OppiaquadrangleonVesta (Tosietal., 2015b).Mostofour mineralogicalanalysisrelies onthelatest calibratednearinfrared spectra acquired by the Visible and InfraRedmapping spectrom-eter (VIR)onboard Dawn.The combination ofVIR data with im-ageryacquiredbytheFramingCamera(FC),showingahigher spa-tialresolution, enablesaccurate correlation ofspectral unitswith geologic units and topography. In addition, we also take advan-tage of the thermal infrared range measured by VIR to derive a spatially-resolvedtemperaturemap ofAc-H-6.Thisunprecedented mineralogic mapping allows for a broader understanding of the processesthatgaverisetothecurrentlyobservedcomposition,and a bettercorrelation withprocessesthat are found inother quad-ranglesonCeres.

2. Data

Launched on27 September 2007,the Dawn spacecraftclosely exploredthegiantasteroidVestafrom16July2011to5September 2012.Aftera2.5-yearstransferperiod,Dawnenteredorbitaround Cereson6March2015,followingapolartrajectory(Russelletal., 2016).

The data used in this paper have been obtained by the two remote sensinginstruments onboard Dawn: VIR andFC. The VIR instrument (De Sanctis et al., 2011) is an imaging spectrometer with an overall spectral range from 0.25 to 5.1μm, designed to determine the mineral composition of surface materials in their geologic context, and to derive spatially-resolved thermal maps of thesurface ofthe object onits dayside in locationswith sur-facetemperatures above∼180K (Tosietal., 2014). TheFC (Sierks etal.,2011)hasabroadbandclearfilterandsevennarrow-band fil-ters(0.4–1.0μm)thatobtainednear-globalcoverageofthesurface

(4)

Table 1

Summary of the main observation circumstances for the different Dawn mission phases at Ceres. Different columns are labeled according to the mission phase name or acronym. Different rows respectively report: start date, stop date, mean altitude of the spacecraft above the surface (km), VIR pixel resolution (km), FC pixel resolution (km), and range of phase angle values (degrees).

Survey HAMO LAMO XMO1 XMO2 XMO3 XMO4

Start date 5 Jun 2015 18 Aug 2015 16 Dec 2015 19 Jun 2016 10 Oct 2016 21 Jan 2017 29 Apr 2017 Stop date 30 Jun 2015 22 Oct 2015 18 Jun 2016 2 Sep 2016 4 Nov 2016 17 Feb 2017 24 May 2017

Mean altitude (km) 4413 1512 378 378 1522 7738 19,457

VIR pixel resolution (km) 1.10 0.38 0.095 0.095 0.38 1.93 4.86 FC pixel resolution (km) 0.41 0.14 0.035 0.035 0.14 0.73 1.82 Phase angle range (deg) 12.4–84.6 23.0–57.5 35.9–80.7 35.9–80.7 58.9–80.0 75.5–84.1 0.0–6.5

revealing detailedalbedo andcolor variations. FC hasan angular (pixel)resolution∼2.67timesbetterthanVIR.FCimageshavealso beenusedto deriveanaccurate shapemodelofthedwarfplanet (Preuskeretal.,2016).

TheorbitalmissionatCereswasdividedinanumberofphases: Survey, High-AltitudeMapping Orbit (HAMO), Low-Altitude Map-pingOrbit(LAMO),ExtendedLow-AltitudeMappingOrbit(XMO1), ExtendedJulingOrbit(XMO2),ExtendedGrandOrbit (XMO3),and Ceres Opposition(XMO4). Each phase hada different length and was carried out from a different altitude,with spatialresolution depending on the altitude over the mean surface. Table 1 sum-marizes the main observational circumstances for these mission phases. In general, the altitude decreased from Survey to LAMO, andthenincreasedagainfromXMO1toXMO4.

ThedifferentstagesoftheDawnmissionatCeresoccurred un-der changing illumination conditions. In a polar orbit, the solar phase anglevariescontinuously,butthevalue measuredat equa-torial sub-spacecraft points’ latitudes increasedoverall from Sur-vey to XMO3, along with a variablecoverage ofthe surface. The mostfavorablecircumstances intermsofcoverage,spatial resolu-tionandilluminationconditionsforspectroscopywereachievedin theSurveyphase,eventhoughthespatialresolutionwasrelatively low comparedtothesubsequentHAMO,LAMOandXMO1stages. Themostlimitedspatialcoverageandlowestsignal-to-noiseratio (SNR) were obtained inLAMO andXMO1, due to shortexposure times,large phase anglesgenerallyinducing long shadowsinthe observedscene,andsmearing.ThismakesVIRdataacquiredinthis periodlessoptimalforuseandinterpretationonaregionasbroad asanentirequadrangleofCeres.

3. Toolsandtechniques

3.1. Spectralparameters

While theaverage reflectancespectrum ofCereslacks distinc-tive features in the spectral range 1.0–2.5μm, the thermally cor-rected wavelength region 2.6–4.2μm as explored by VIR in the early Survey orbit phase, which lasted from 5 to 30 June 2015 and obtainednearly global hyperspectral coverage at ∼1.1km/px, displays a broad asymmetric feature, characteristic of H2

O/OH-bearingmaterials.Withinthisbroadabsorptionareseveraldistinct absorption bands, centered respectively at 2.72–2.73μm, 3.05– 3.1μm,3.3–3.5μmand3.95μm(DeSanctisetal.,2015).The2.72– 2.73μmfeatureisdistinctiveforOH-bearingminerals.This promi-nent band isbest fittedbya mixtureincluding Mg-bearing phyl-losilicates such as Mg-serpentine (antigorite) (Ammannito et al., 2016). NH4-bearing minerals, such as NH4-serpentines and NH4

-smectites, show clear absorptions at 3.05–3.1μm and at 3.3μm (Bishop et al., 2002; Ehlmann et al., 2017), while the 3.95-μm absorption band has been definitively attributed to carbonates (Rivkinetal.,2006;DeSanctisetal.,2015).Theselithologies,plus a spectrally bland dark material that is actually the most abun-dantspectral endmember, areubiquitousat∼1kmspatial resolu-tion (Ammannito et al., 2016; Carrozzo et al., 2017a, this issue).

Fig. 1. Average reflectance spectrum of quadrangle Ac-H-6 Haulani in the range 1–4 μm, as measured by VIR in the Survey mission phase (spatial resolution ∼1.1 km/px) under favorable illumination conditions. The spectrum was corrected for the thermal emission (following the approach described in Raponi et al., 2017 , this issue) and for the illumination and observation geometry by means of the Hapke model ( Ciarniello et al., 2017 ). The solid line represents the mean spectrum while the dashed lines mark the 1-sigma variability. Gray boxes highlight the spec- tral counterparts (position and width) of three order sorting filters placed on top of the VIR infrared focal plane, which induce systematic artifacts in VIR spectra.

The identificationof thedark endmember ischallenging because its spectrum is essentially featureless, except for a tentative ab-sorption band centered at 1μm, which in principle may be at-tributed to iron (Fe). While magnetite (Fe3O4) provides a good

spectral match, alarge amount ofFe isinconsistent withGRaND measurements (Prettymanetal., 2017). Sincecarbonaceous chon-driteis themeteoriticanalogueclosest to theaveragereflectance spectrumof Ceres(ChapmanandSalisbury, 1973), thisdark end-memberisverylikelycarbon-bearingmaterial.

Because the 2.7-and 3.1-μm absorbersare observed in all of the geologic units, spanninga broad range of ages and with no correlation between the age andthese bands’ depths, the corre-spondingmineralphasesare believedto beofendogenousrather thanexogenousnature,whichmeansthattheoutercrustofCeres isintrinsicallyenrichedinthesespecies.

InFig.1weshowan averagereflectancespectrumofthe A-H-6Haulaniquadrangleintherange1–4μm,asmeasuredbyVIRin theSurvey mission phase,where themain spectral signatures so fardescribedcanbeidentified.

Spectralindicesusedinthisworkare:reflectance(measuredat givenwavelengths),spectralslopes,bandcenter,andbanddepth.A spectralslopeismeasuredbyinterpolatingthespectralprofile be-tweentwogivenwavelengthswithastraightline,andcomputing itsangular coefficient.In boththe visibleandnear-infrared spec-tralranges,spectralslopesareindicativeofthecomposition,grain size and age of the surface regolith, therefore highlighting both thechemico-physical state andspaceweatheringprocesses. Band centers(BC)andbanddepths (BD)aretypicalspectralparameters

(5)

usedbothinlaboratoryspectraandinremotesensedspectroscopic datasets to understand the mineralogical composition of a plan-etary surfaceandthe physical characteristicsof the regolith.The bandcenteristhelocationofthereflectanceminimuminsidethe bandafterthespectral continuumremoval, whilethebanddepth isdefined,afterClarkandRoush(1984),as:1-Rb/Rc,whereRband Rc arethe reflectance oftheband andthe spectral continuum at

theBC.

The analysis ofthe band centerfor themain diagnostic spec-tralsignaturesseen intheinfraredspectrum ofCeresshowsthat itdoesnotchangesignificantlyforMg-andNH4-bearing

phyllosil-icates, while it can change for carbonateminerals, where an in-creasein the band center value from 3.95 to 4.01μm is indica-tive ofa smooth transitionbetween Mg-or(Mg,Ca)-rich carbon-atesandNa-richcarbonates(e.g.,Carrozzoetal.,2017a,thisissue;

Palombaetal.,2017b,thisissue).

Banddepths areindicativeofamineral’sabundance,the pres-enceofopaquemineralsorotherphases,andmayalsodependon thegrainsizedistribution(e.g.,Adams,1974;Clark,1999).In gen-eral,thepresenceofopaquemineralstendstoreduceoreven sup-pressan absorption band, dependingon thethicknessof the ab-sorbinglayer.Unlikeband centers,banddepthvaluesareaffected by illumination and observation geometry, hence a photometric correctionisrequiredtoallowa properevaluationofthis param-eter.Inthisregard,Ciarniello etal.(2017) appliedHapke’s model tothe VIR dataset obtainedat Ceres inthe early mission stages, i.e.Survey andpre-Survey approachphasesyieldinga spatial res-olution>1km/px,toperformaphotometriccorrectiontostandard observationgeometry(solarincidenceanglei= 30°,emission an-glee=0°,solarphase angle

α

= 30°)atVIS-IRwavelength.This resultedinalbedoandcolormapsofthesurface.Thesame proce-durewaslaterappliedtoHAMOdata,inordertoalsoallowproper estimation of band depth values for those datasets, which were usedtoproducethespectralmapsdescribedinthenextsection.

3.2.Spectralmaps

This work is largely based on VIR-derived spectral maps of quadrangle Ac-H-6 Haulani, showing a higher level of de-tail compared to global spectral maps released previously (Ammannitoetal.,2016),andonFC-derivedmapsofthesame re-gion.Allmapsuse thelatestproducts available forthesetwo in-strumentsatthetimeofwriting.

Frigeri et al.(2017, this issue) provide details abouthow VIR spectralparameters were derived, filteredandmapped.In partic-ular, VIR spectral indices were computed from individual hyper-spectral images acquired in the Survey and HAMO phases, and thenmosaickedusingaveragedoverlappingvaluesona gridwith a fixed resolution of 140m/pixel (Frigeri et al., 2017, this issue). Thisresolution matches the average resolution of FC images ob-tained in the HAMO phase, to enable a direct comparison be-tween VIR- and FC-derived products. The mineralogical maps of the entire quadrangle, on which our analysis dwells, are based on the band depths of the 2.72–2.73μm (hereafter 2.7-μm) and 3.05–3.1-μm(hereafter3.1-μm)spectralfeatures. Thecomposition andabundanceofcarbonatemineralsisevaluated fromthegobal mapsproducedby Carrozzoet al.(2017a, thisissue),which have been cut out for quadrangle Ac-H-6. In this case, the carbonate composition is mapped by means of the center of the diagnos-ticband at ∼4μm,while its abundance is estimatedby comput-ingthedepth ofthesameband.Stripesandspurious values, dis-connectedfromthegeologiccontext,were firstremovedby using theproceduredescribedinCarrozzoetal.(2016),theneach spec-truminunitsof calibratedradiancefactor (I/F) wascorrected for thermalemission byremoving aPlanckfunction producedbythe bestfitwiththemeasuredspectra(Raponi etal.,2017,thisissue),

andfinally thephotometric correctionmentioned intheprevious section(Ciarnielloetal.,2017)wasappliedtothermallycorrected spectra.While few residual artifacts maystill show up afterthis post-processing,furtheranalysiswilleventuallyexcludethose val-uesthatarenotspatiallycoherentorareclearlythecounterpartof spuriousinstrumentalartifacts.

FCcolorfilterdatawereusedtocreateenhancedcolormosaics anda map of spectral slope in thevisible to near-infraredrange 0.555–0.829μm. The ‘Clementine’ color composite map (Pieters et al., 1994) is of interest to evaluate localspectral variationsat highspatialresolution,whichmaytracevariationsincomposition. ComparingClementinecolordataandVIRspectralmapsisindeed apowerfulwayto highlightpossiblecorrelationsbetween miner-alogyandspectrally distinctstructuresseenincolorimagerywith higherspatialresolution.Ontheotherhand,whiletheClementine colorcodeproved tobe optimalforlunar andVestadata,itmay notnecessarilybethebestpresentationtotracecompositional dif-ferencesinother rocky bodiesofthe SolarSystem. Therefore,we exploredtheHaulaniregionusingasecond,‘enhanced’color com-positemap,whichismoresuitedtoCeres,toinvestigatethe exis-tenceofspectral texturesthatwouldotherwiseremain elusiveor hidden.

The presence of dark and bright material units can be eval-uated by means of albedo or reflectance maps. In our case, we usedthefollowingreflectanceinformation:1)aclearfilterFC mo-saic obtainedduringthe HAMOphase (∼140m/pixel), photomet-rically correctedby applying a phaseangle-dependent brightness function (Roatsch et al., 2016b), and 2) two reflectance mosaics derived from VIR data at1.2μm and1.9μm witha resolution of 140m/pixel,photometricallycorrectedusingthemethoddescribed byCiarnielloetal.(2017)andbrieflyaddressedinSection3.1.Since the 1.2-and1.9-μm wavelengthsin VIRdata are unaffected both byknownorexpectedcereanspectralsignaturesaswellasbyVIR instrumental artifacts,they turnout to be optimalto samplethe spectralcontinuuminaregiondominatedbysolarreflection.Their useinconjunctionwiththe2.7-μmandthe 3.1-μmBDmaps ob-tainedatthesamespatialscaleiskeytohighlightingpotential cor-relations arising betweenreflectance and abundance of Mg-and NH4-bearingphyllosilicates.

BecausetheoverallmineralogyofCeresisdominatedbythese twospecies,ouranalysisusesadensityscatterplotofbanddepths at2.7-μmvs.3.1-μm,andamapofmeasured/modeledbanddepth at2.7μm,tohighlightthoseregionsthatsubstantiallydeviatefrom an ideallylinear dependencebetweenthesetwo spectral indices. Inaddition,weuseafour-colormaptomergethe2.7-μmandthe 3.1-μm BD maps into a single composite map, where colors are indicative of the simultaneous abundance of magnesium-bearing phyllosilicatesandammoniatedphyllosilicates.

3.3. Othermaps:geologicmap,topographicmap,temperaturemap

The availabilityof thegeologicmap by Krohnetal.(2017) al-lowsusto associatethecompositionofthemain surfaceunitsof the Haulani quadrangle to their relative geologic timescales and thereforeachieveaclearpictureaboutthesequenceofeventsthat led to the mineralogythat is observed on Ceres today. A digital terrain model (DTM) ofthe Haulani quadrangle,derived from FC clear-filter data acquired in HAMO with a resolution of 60 pix-els/deg or ∼137m/pixel (Preusker et al., 2016), is also used to cross-checkcorrelationsbetweencompositionandtopography.

Thelong-wavelengthrangeofVIRsensitivity,4.5–5.1μm,is ex-ploitedtosystematicallyretrievesurfacetemperatures.Thesedata were usedboth toproduceglobalmapsandto investigatein de-tailthebehavioroflocal-scalefeatures(Tosietal.,2014).The mo-tivation forourinvestigationis thesearch forthermalanomalies, i.e.regions whoseresponseto insolationdiffers fromtheaverage

(6)

behavior ofthe quadrangleand of the dwarf planet as a whole. The surfaceofCeresisexpectedto havealow thermalinertia at globalandbroadlyregionalscale(Titus,2015;Hayneand Aharon-son,2015),butdeparturesfromthisbehaviormaybefoundatthe localscale,e.g., duetovariationsinregoliththickness, density,or thermal conductivity.For thisreason, we also includeda map of surfacetemperatureoftheAc-H-6quadranglederivedfromVIR in-frareddata.

Inallmaps,includingthespectralmapsdescribedinthe previ-ous sub-section,longitudes aregivenintheCerescoordinate sys-tem(RaymondandRoatsch,2015).Asmallcrater(∼0.4km diame-ter)namedKaitwaschosenbytheDawnTeamtodefinetheprime meridian (Roatsch etal., 2016a).The location of thissmallcrater is within the envelopeof the broad feature identified inHubble SpaceTelescope(HST)datatowhichtheprevioussystem(Archinal etal.,2011)wasanchored.

4. GeneralmineralogyofAc-H-6

4.1. Reflectance,colorsandtopographyfromFCdata

First,wesummarizethemainfeaturesthathavebeendescribed in the Haulani quadrangle based on FC data, providing a high-resolution context for subsequent analysis and interpretation of mineralogydata.Inthisregard,werelyonacomparisonbetween sixmaps,showninFig.2:a)areflectancemap;b)thegeologicmap produced by Krohn etal. (2017); c) a Clementine color compos-itemap;d)an‘enhanced’colorcompositemap,alternativetothe Clementine color map; e)a map ofthe spectral slope computed between0.555and0.829μm;andf)aHAMO-basedDTM.

From the FC-derived clear filter reflectance mosaic (Fig. 2a), whose reflectance values range from 0.01 to 0.06,one can infer theexistenceofbrightanddarkmaterials,whichstandoutas hav-ingsubstantiallyhigherorlowerreflectancesthanthequadrangle’s average value of 0.03. In Ac-H-6, most of these dark and bright materials areassociated withcraterHaulani,whichoverall isone ofthe brightest features onCeres,withits brightinteriorandits extensive ejecta withfar-ranging crater rays (Krohn et al., 2017). ThehighalbedooftheHaulaniareawasfirstrecognizedinHubble SpaceTelescope (HST)images(Lietal., 2006) andthenrelocated inDawnFCApproachandSurveyimages(Lietal.,2016).

Geologic mapping of quadrangle Ac-H-6 (Fig. 2b, fromKrohn etal.,2017)establishedthat,amongbrightunitsfoundinHaulani, knobby bright lobate material (lkb) extends around the northern and eastern crater rim, and smooth bright lobate material (lsb) onlyoccursassmallerbranchesinthesouth-southeastand north-northeast,coveringthesurfacewithstreakyflow-likematerial. Par-ticularlyprominentisabrightlobatematerial(lb),extending west-ward.This materialshowsarelatively smooth surfacewithmany flow fronts.Craterfloorhummockybrightmaterial (cfhb) also oc-cursonlyinHaulanicraterandshowsahighalbedo.Finally,bright crater ray material (crb) deposits occur as bright streaks or ra-dial halos around the Haulani impact crater and cover the ad-jacent terrain with a somewhat thin layer. This unit exhibits a highalbedocomparedtothesurroundingterraininFCclearfilter images.

Amongdarkunits,craterfloormaterialdark(cfd)hasasimilar morphologyandalbedoasthecfunit,butintheenhancedFCcolor imagesthisunitappearsdarkbluetoblack.Thecfdunitwasonly found in Shakaema crater. The crater floor material smooth dark (cfsd) inHaulani contains manypits. Crater dark material (cd) is only found at Kondos crater andalso covers the crater rim and floor,formingacontinuoussurface.

The Clementine composite map of quadrangle Ac-H-6 (red: 0.75/0.44μm, green: 0.75/0.92μm, blue: 0.44/0.75μm) (Fig. 2c)

shows that most of the surface is reddish, while crater Haulani standsout to be the mostprominentmultispectral unit, display-ingan overalllight bluecolor,withbrightcrater raymaterial ap-pearingcyan,darkmaterialappearingbluish/violet,andthe bright-estunitsinthecrater’sfloordisplayingagreenishcolor.However, thesoutheastern innerwall ofHaulani,andsmallerpatches close to the northern rim, which may represent redeposited slumping material from the southern topographic heights (Stephan et al., 2017a), retain thered colortypical of therestof thequadrangle. BesidesHaulani, whichisby far themoststriking feature inthis quadrangle,thereareotherexamplesofbluishmaterialunits high-lightedby theClementine colorcode:1) a smallunnamed crater located at 13.7°S/11.2°E, roughly in the middle of Anura crater’s floor, whichexcavated fresh material frombelow, 2) a small un-namedcratercenteredat20.7°N/7.6°E,i.e.closetotheborderland betweenAc-H-6andAc-H-2quadrangles,and3)theejectaoftwo small unnamed craters, centered respectively at 9.9°S/21.0°E and 17.0°N/62.3°E.

To complementthe analysisof the Haulani quadrangle based on multispectral FC data, we have used a second color scheme, which is more suited to Ceres (Fig. 2d). In this ‘enhanced’ color composite mosaic (red: 0.96μm; green: 0.75μm; blue: 0.44μm), mostofthesurfaceinquadrangleAc-H-6(andmoregenerallyon Ceres)displaysabrownishcolor,whilecraterHaulaniremainsthe most prominentmultispectral unit, with ejecta material display-ing a typicalbluish color. Comparedto theClementine presenta-tion, this color scheme shows many more color shades and en-hances differences in albedo, with bright materials concentrated particularly in the Haulani crater’s floor, apprearing cyan, while dark material units found in the Shakaema and Kondos craters, and to a lesser extent in other craters of smaller size including Haulaniitself, appear dark blue to violet. However, subtle differ-encescanbenoticedinthesoutheasterninnerwall ofHaulanias wellasothersmallerpatcheslocated justbelowitsnorthernrim, whenusing thiscolorscheme compared tothe Clementine color scheme.

Themapofvisualspectralslope,calculatedbetween0.555and 0.829μm(Fig.2e),revealsthatthemajorityofthequadranglehas a greenish and whitish color, indicating a null (flat) or slightly positive slope value. Crater Haulani stands out for its blue/dark blue/redcolorscorrespondingtonegativevaluesofspectral slope, especially in its floor andejecta (Stephan etal., 2017a). In mov-ing away from Haulani,the value of the spectral slope gradually decreasestobackgroundlevels.Extremenegativevalues(downto −0.03)arerecordedinthemiddleandwesternpartofthecrater’s floor,and inthe ejectacloserto the rim, while thesoutheastern inner wall, aswell assome minorlocalized patches in the east-ern,northernandnorth-easterninner wallsclosetotherim, dis-playapositivespectralslope.Thenegativevaluesofspectralslope recordedintheareaofcraterHaulaniareextremenotonlywithin thisquadrangle,butmoregenerallyontheentiresurfaceofCeres, other veryblue surfacefeaturesbeing OxoandAhunaMons, fol-lowedbyOccatorandKupalo(Schröderetal.,2017).

Across quadrangle Ac-H-6, the 0.555–0.829μm spectral slope overall increases in moving from west to east, while other ex-ceptions represented by negative values of spectral slope (albeit lessprominent than in crater Haulani) are found in a small un-named crater at 13.7°S/11.2°E, in the floor of crater Anura, and inthe ejectaofcrater Anura,aswell asin other threesmall un-namedcraters,centeredrespectivelyat:20.7°N/7.6°E,9.9°S/21.0°E, and17.0°N/62.3°E.Agedeterminationsindicatethatthebluish ma-terial is mainly associated with the youngest impact craters on Ceres(Nathuesetal.,2016;Schmedemannetal.,2016).

Fig.2fisaportionoftheDTMofCeresfortheAc-H-6 quadran-gle,wherethecolorpalettereflectsthelocaltopography(Preusker etal., 2016). The maximum topographicrelief that occurs inthis

(7)

Fig. 2. a: Reflectance map of Ac-H-6 Haulani, obtained through photometric correction of FC clear-filter data acquired in the HAMO mission phase. The map scale is ∼140 m/pixel. Surface features that have received an official designation by the IAU are marked in yellow. b: Geologic map of Ac-H-6 Haulani, from Krohn et al. (2017) . The broad brownish geologic unit is mapped as ‘cratered terrain’ ( crt ). Yellow patches of ‘crater material’ ( c ) are found in the vicinity of the main impact craters. c: RGB composite of the Ac-H-6 Haulani quadrangle made from FC color ratios: R : 0.75/0.44 μm, G : 0.75/0.92 μm, B : 0.44/0.75 μm (‘Clementine’ color composite). Color ratios may enhance differences in material and composition. d: RGB composite of the Ac-H-6 Haulani quadrangle made from FC colors: R : 0.96 μm; G : 0.75 μm; B : 0.44 μm (‘enhanced’ color composite). Compared to the Clementine presentation, this color scheme enhances differences in albedo. e: RGB composite of the Ac-H-6 Haulani quadrangle showing spectral slope calculated between 0.555 and 0.829 μm. f: Portion of the digital terrain model (DTM) derived from FC images acquired in HAMO, at a scale of 60 pixels/degree. The color palette follows height in km with respect to the reference ellipsoid (i.e., the curvature of the body is removed), with the highest elevations in brown and the lowest terrains in magenta/white. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

(8)

Fig. 3. a: Reflectance map of Ac-H-6 Haulani, obtained through photometric correction of VIR data acquired at 1.2 μm in the Survey and HAMO mission phases, interpolated on a grid with a fixed resolution of ∼140 m/pixel, photometrically corrected by using the method described by Ciarniello et al. (2017) , i.e. a Hapke photometric correction to standard observation geometry ( i = 30 °, e = 0 °, α= 30 °). b: Reflectance map of Ac-H-6 Haulani, obtained through photometric correction of VIR data acquired at 1.9 μm in the Survey and HAMO mission phases, interpolated on a grid with a fixed resolution of 140 m/pixel. The same photometric correction of ( a ) was applied. In both maps, disconnected lines within a given image are due to the high instantaneous speed of the ground footprints (i.e., projections of the spectrometer’s slit) in the HAMO phase.

quadrangle is ∼9.5km, with the deepest part (down to ∼5.6km belowthe referenceellopsoid)corresponding toa crater floor lo-cated at22.0°N/17.3°Einthe borderlandbetweenquadrangle Ac-H-6HaulaniandquandrangleAc-H-2‘Coniraya’,andtothefloorof crater Kondosclose tothe southern edge, andthe highest eleva-tions (up to ∼3.8kmabove the referenceellipsoid) concentrated in the latitude range 15°S to 20°N and in the longitude range 20–60°E. The available datasets permit investigation of possible correlations betweenvariationsinreflectance, colorand mineral-ogy, and the presence of hills and depressions. The topographic map also serves asa monitoring tool,allowing oneto determine whetheranobservedvariationinspectralindicatorsisreliableand notsimplyduetothecombinationoflocaltopographyand instan-taneoussolarillumination(seeSection3.1).

4.2.1.2-μmand1.9-μmreflectances

Reflectancemapsobtainedinthenearinfraredrangeat1.2μm and 1.9μm, with the method briefly addressed in Section 3.1

(Fig. 3), are consistent with the reflectance map obtained by FC inthevisual rangeandconfirmthatthefloorandejectaofcrater Haulani are a patchwork of inherently bright and dark material unitsalsoatnear-infraredwavelengths.

In particular, the highest reflectance values (∼0.04 at both 1.2μmand1.9μm)are recordedinthecentral andsouthern floor ofthecrateraswell asintheejectathatstretchtothesouthand the western inner wall, while relatively dark materials at these wavelengths (reflectance ∼0.03) are found in the northern inner wall, in the ejecta that extend towards the east and southwest

(9)

Fig. 4. Two spectral slopes, calculated in the near infrared range 1.163–1.891 μm ( a ) and 1.891–2.250 μm ( b ) on the basis of VIR infrared data. These spectral slopes are essentially consistent with the spectral slope obtained from FC data in the visual range 0.555–0.829 μm at higher spatial resolution ( Fig. 2 e): i.e., it is confirmed that crater Haulani generally corresponds to a neat reduction in spectral slopes, with negative values found in the middle of its floor, the western inner wall and the ejecta closer to the rim. In both maps, disconnected lines within a given image are due to the high instantaneous speed of the ground footprints (i.e., projections of the spectrometer’s slit) in the HAMO phase. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

directions in the vicinity of the crater’s rim, and in a dis-tinct patch in the ejecta that extends to the west, a short dis-tance fromthe rim. On the large asteroid Vesta, which was the previous target of the Dawn mission, impact craters often ex-pose layers of both dark and bright material. In that case, the bright layers correspond to fresh pyroxenes processed by rel-atively little space weathering (McCord et al., 2012; Zambon et al., 2016). Conversely, bright spots on Ceres are thought to be the result of phenomena like cryovolcanism (Russell et al., 2016) and post-impact hydrothermal activity (Bowling et al., 2016).

Therestofthequadrangleshowsasubstantialuniformityinthe infraredreflectancevalues,whichisrevealedbythenarrowrange ofvaluesinthegraycolorbars.

4.3. Near-infraredspectralslopes

The two spectral slopes computed inthe nearinfrared range: 1.163–1.891μmand1.891–2.250μm(Fig.4)areconsistentwiththe spectralslopeinthevisualrange0.555–0.829μmathigherspatial resolution(Fig.2e):i.e.,itisconfirmedthat craterHaulani gener-allycorresponds toasubstantial reductioninspectral slopes’ val-ues,withnegativevaluesfoundinthemiddleofitsfloor,the west-erninnerwallandtheejectaclosertotherim,whilemasswasting materialfound inthesoutheastern innerwall showspositive val-uesof spectral slopesimilar to thosemeasured in terrainsfound intheeasternsideofthequadrangle(Stephanetal.,2017a).Other verysmallvaluesofspectralslope,thoughnotnecessarilynegative inthenearinfraredrange,arefoundinthenortheasternejectaof

(10)

Fig. 5. Histograms of 2.7-μm band depth (BD27) and 3.1-μm band depth (BD31) values measured across quadrangle Ac-H-6 (after photometric correction), displayed in dark gray and light gray colors, respectively. A Gaussian fit is superimposed in the light blue color. The most recurrent values are in the range 0.16–0.24 (average value 0.201) for BD27 and 0.04–0.12 (average value 0.083) for BD31. (For interpre- tation of the references to color in this figure legend, the reader is referred to the web version of this article.)

a small unnamed crater located near the northern border ofthe Ac-H-6quadrangle,centeredat20.7°N/7.6°E.

A different behavior in the spectral slopes between the vi-sual range highlighted in the high-resolution FC mosaic(Fig. 2e) andVIRmosaicshighlightedinFig.4(i.e.,localtrendsofspectral slopesinthevisualrangewhichseemtobeneutralinthenear in-frared)ismostlikelyindicativeoftheageofthosematerials, and thereforeofspaceweatheringprocesses.

Thebluematerialbecomesapparentinmorphologicallysmooth areas,withflowfeaturesappearingmorphologicallysmoothatthe highestresolution images(35m per pixel) obtainedinthe LAMO missionphase (Krohnetal., 2016;Stephan etal., 2017a). Physical alterationsuchasamorphizationandaggregationofsmall-grained phyllosilicate-rich material andeven heating up andmelting the impacted surface material to moderate temperatures is expected incaseoflow-velocityimpactevents,whichcouldexplainthe ob-served bluespectral slope (Stephan etal., 2017a). As small scale roughnessisthoughttodominatethescatteringbehavior(Shepard andHelfenstein,2007),thebright,blueejectaofcratersthatshow evidenceforflows,likeHaulani,mayinfactbephysicallysmooth. Highalbedo,bluecolor,andphysicalsmoothnessall appeartobe indicatorsofyouth(Schröderetal.,2017).

4.4. 2.7-μmand3.1-μmbanddepths

Fig. 5 showshistograms related to the frequencyof the band depths at 2.7μm and 3.1μm (hereafter BD27 and BD31) values measuredacrosstheHaulaniquadrangle,sampledwitha0.01bin width.

Consistent with the average reflectance spectrum of the sur-face ofCeresasmeasuredby VIR,wherethe absorptioncentered at2.72–2.73μmis broaderthan theabsorption centered at3.05– 3.1μm(seeFig.1),BD27isgreaterthanBD31,withmostrecurrent valuesin therange0.16–0.24 (average value 0.201) forBD27 and 0.04–0.12(average value 0.083)forBD31. Thesetwo distributions havea narrowwidth, withextremely similarmaximumvaluesof frequency andcomparable shapeshardly modeled by a Gaussian fit.Thisevidencedemostratesthegoodnessofphotometric correc-tion,andtestifiesahomogeneousdistributionoftheabundanceof the featurelessspectral endmember across theentire quadrangle. Forcomparison,inothernearbyquadranglesofCerestheobserved statistics mayshow aslightly differentdistribution (e.g.,Zambon etal.,2017b,thisissue).

This aspect is also important to associate the values ofBD27 andBD31that weobserveto relativeabundances ofMg-richand NH4-richphyllosilicates,respectively.AsanticipatedinSection3.1,

thebanddepthvalue,whichisindicativeofamineral’sabundance, may also be affected by the presence of opaque phases and by the grain size. In order to quantify abundance values, a spectral unmixingmodel hasto be appliedto theobserved spectral data, whichisbeyondthescopeofthiswork,whileitisaddressed else-wherefor a narrower region within quadrangle Ac-H-6,covering onlycraterHaulani(Tosietal.,2017,underreview).However,from thespectralunmixingmodelingcarriedoutpreviouslytoestablish theglobalaveragemineralogyof Ceres,itturns outthat best fits withrespecttotheaveragereflectancespectrumofCeresas mea-suredby VIRareobtainedby usingantigoritewithagrainsizeof 90μm,Mg-carbonateswithagrainsizeof10μm,andafeatureless darkcomponentwithagrainsizeof80μm(thegrainsizeforNH4

-richphyllosilicatescouldnotbedeterminedunequivocally)(Fig.4a andExtended Data Table 2 in DeSanctis etal., 2015). So, under thereasonableassumption thatthesegrain size valuesare recur-rentonabroadlyregionalscaleincludingquadrangleAc-H-6,our BD27 andBD31values maybe associatedto mineralabundances ratherthantovariationsinthegrainsizedistribution.

Fig.6 revealsa map ofthe BD27 andBD31 values across the Haulaniquadrangle.First,acomparisonwiththetopographicmap shownin Fig. 2f shows that the band depth in mostcases does not followthe trendofthe topography, suggestingthat our pho-tometric correction properly removes the effects of topographic slope(see Section 3.1).The moststriking resultisthat low BD27 and BD31 values are associated with crater Haulani’s floor and ejecta(mostnotablythoseextendinginthenorth–southdirection), which appearbluish in FC-derived mapsin both the Clementine andenhancedpresentations,exceptinthesoutheasterninnerwall ofthecrater.Thisevidenceindicatesanoverallnetdecreaseinthe abundanceofMg-andNH4-bearingphyllosilicatesintheregionof

craterHaulani,comparedtothesurroundingterrainsandgeologic units.Other smaller patches showing a similar decreasein these two BDvalues are locatedin the floorsof the craters Shakaema, AnuraandKondos. Inparticular,inthecaseofKondosadecrease oftheBD31valueisobservedwestofthecrater,notfollowedbya proportionaldecreaseintheBD27value.Conversely,the southeast-ern regionofthe quadrangle,corresponding tothe cratered (old-est) terrain unit, is characterized by higher values of BD27 and BD31. Fig. 7 shows a direct comparison between the reflectance spectraof theseregions andthe average reflectance spectrum of quadrangleAc-H-6asawhole.

High values of BD27 not matched by equally high values of BD31 are found in a small unnamed crater centered at about 9.9°S/40.9°E, while the unnamed crater centered at 20.7°N/7.6°E showsan opposite behavior, with its northern ejectadepleted in Mg-bearingphyllosilicates, butits floorshowingan enhancement inNH4-bearingphyllosilicatescomparedtosurroundingterrains.

Insummary, inquadrangleAc-H-6 there isa generaltrend in values BD27 and BD31, and therefore in the abundances of Mg-richandNH4-rich phyllosilicates,to increaseby moving fromthe

northwesttothe southeast.However, thistrend isstrongly influ-enced by crater Haulani, and to a lesser extent by other larger craterssuchasKondos,AnuraandShakaemaaswellassmaller un-named craters. Theirformation significantly alteredan otherwise smoothtransitionintheabundancesofthesetwomineralspecies acrossthisregion.

4.5.BD27versusBD31

We next evaluate the possible correlation between the BD27 andBD31values.Inthiscase,ahighdegreeofcorrelationis indica-tive of a strong (intimate) mixture between the OH-bearing and

(11)

Fig. 6. Distribution of 2.7-μm band depth (BD27) ( a ) and 3.1-μm band depth (BD31) ( b ) values across the Haulani quadrangle, derived from VIR data. These band depths are sensitive to the abundance of magnesium-bearing phyllosilicates and ammoniated phyllosilicates, respectively, and to the presence of opaque contaminants. The overall distribution spans from lower depth (dark blue to blue) to greater depth (red to dark red). The floor and ejecta associated with the spectrally distinct material of crater Haulani have shallower BD27 and BD31 values. The Haulani crater itself displays spectral variations in its interior: BD values are shallower on the floor and stronger in the southeastern and northern walls. Small values are also found elsewhere throughout the quadrangle: other smaller patches showing a similar (but not necessarily parallel) decrease in these two BD values are located in the floors of the craters Shakaema, Anura and Kondos. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

NH4-bearingphyllosilicates,whilemoderate-to-lowcorrelation

val-ueshinttoarealmixtures.Fig.8aisadensityscatterplotofBD27 asafunctionBD31.Becausethereisonebigclusterofpoints,with mostdataconcentratingclosetothemostrecurrentvalues(0.201 forBD27 and0.083forBD31,respectively), alinearmodelproves inadequatetodescribe thetrendoftheratioofMg-rich/NH4-rich

phyllosilicates:asrevealedbythechi-square,thedistributionis es-sentiallynon-linear,andtheaveragecoefficientofdeterminationis moderate-to-low, meaning that a substantial correlation between thesetwo variablesis not found anywhere throughoutthe quad-rangle, but only for specific geologic features found at the local scale.

BymappingthevalueoftheratiobetweenthemeasuredBD27 and the BD27/BD31 linear approximation (Fig. 8b), one can see

that pointswith BD27 values significantly lower than the linear modeldisplayedinFig.8a,i.e.withalowervalueoftheratio Mg-rich/NH4-rich phyllosilicates, generally correspond to the area of

craterHaulani(floor,rimandejecta,especiallythoseextendingin thenorthandsouthdirections)andthenortheasterncornerofthe quadrangle.Thisindicates that theMg-depletion extendsfurther from Haulani than the NH4-depletion. Conversely, values

signifi-cantlyhigherthanthelinearmodel,i.e.withahighervalueofthe ratioMg-rich/NH4-richphyllosilicates,concentrateinthesouthern

andsouthwestern partofthequadrangle,particularlyinWangala Tholus (notable exceptionsbeing the floors ofcraters Shakaema, Anura,KondosandPiuku),andinthemiddle-northernquadrant.

Fig.9isdividedintwo panels.Fig.9arepresentsa2D scatter-plotofBD31vs.BD27fortheentiresurfaceofCeres.Fig.9bshows

(12)

Fig. 7. a: Reflectance spectra of craters Haulani (blue), Anura (cyan), Kondos (green), Shakaema (orange), and of a 5 ° by 5 ° region (Latitude 17–22 °S, Longitude 60 °E-65 °E) in the southeastern region of quadrangle Ac-H-6, as observed by VIR in the HAMO mission phase (pixel resolution ∼0.38 km), compared to the average re- flectance spectrum of the entire quadrangle (black). An arbitrary offset is applied for clarity. Gray boxes highlight the spectral counterparts (position and width) of three order sorting filters placed on top of the VIR infrared focal plane, which in- duce systematic artifacts in VIR data. b: Same spectra of panel ( a ), normalized at 2.67 μm, a wavelength which is unaffected by both known spectral features and in- strumental artifacts. Haulani stands out to have a blue spectral slope and shallower band depths at 2.7 and 3.1 μm compared to other larger and older craters and to cratered terrain in the southeast. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

the same 2D scatterplotof Fig. 9a superimposed on a four-color palette, wherethe color codeis designedspecifically to take ad-vantage ofthe band depth information simultaneously: here col-orsaresuchthatblueindicatesthatthetwoabsorptionbandsare bothstrong,orangemeansthatthetwoabsorptionbandsareboth weak,greenindicatesastrong2.7-μmabsorptionbandandaweak 3.1-μm absorption band, and magenta indicates a strong 3.1-μm absorptionbandandaweak2.7-μmabsorptionband.Thisis moti-vatedbycombiningtwomapsinone,withoutthecolorambiguity thatresultfromclassicalthreecolorcomposites(RGB).

Fig.9provides aconvenientwaytomap theseabundances si-multaneously onthesurfaceofquadrangleAc-H-6.Inthisregard,

Fig. 10 displays two color maps. The first map (Fig. 10a) uses a four-color, two-dimensional table, where all VIR pixels on the map are represented by a color (not black). This is very similar torepresentingoneparameterwithaone-dimensioncolorscheme (aramp), wherethe parametervaluesare representedbetweena minimumandamaximum,andarescaledtothenumberofcolor levels.Inthesecond map(Fig.10b), onlythepurest compositions are represented by colored pixels, while all the others are dark. Thisisjustifiedbyassumingthatthesurfacecompositionwiththe

highest population likely results from mixtures, and therefore it doesnot represent pure chemistry or mineralogy.Pure composi-tionsareusefultoidentifytheoriginsofsurfacematerialsand in-terprettheprocessesinvolvedintheirformation.

ThetwomapsinFig.10overallconfirmasubstantialdepletion inbothMg-richandNH4-richphyllosilicatesinthefloorandejecta

of crater Haulani (except in its southeastern and northern inner walls),whiletherestofthe quadrangleshowsamineralogy sim-ilar to theglobal averagemineralogy ofCeres.Indeed, variability intherelativeabundances ofthesetwomain mineralphasescan alsobefoundlocally,i.e.inthefloorsandejectaofothercratersas discussedbefore,butto a muchlesserextent comparedto crater Haulani.

4.6.4.0-μmbandcenterandbanddepth

Whileglobalmapsofcarbonates,detected onthebasis ofVIR data on Ceres, are addressed elsewhere (Carrozzo et al., 2017a, thisissue), a significant result obtainedfor the Ac-H-6 quadran-gle(Fig.11)isthat thesignatureofcarbonatesisusually centered at3.95μm,i.e.indicativeofMg-or(Mg,Ca)-richcarbonateslikein mostofCeres’surface, exceptin specificareas foundatthe local scale.

In crater Haulani’s western and northwestern inner walls, in the eastern inner wall and in some ejecta that propagate to-wardsthewestandnorthwest,thebandcentermovesto3.97μm, whichisdiagnostic ofCa-carbonatesoralternativelya mixtureof (Mg,Na)-richcarbonateendmembersinthemineralogyofthose ar-eas (Palomba et al., 2017b, this issue). In the southern rim and theejectadeparting fromheretothesouth,thebandcentermay moveupto3.99μm,whichispossiblydiagnosticofNa-carbonates inveryspecificsites.

Thisevidenceisnotobservedintherestofthequadrangle, ex-cept fora distinct patchinthefloorofthe 44-kmcraterKondos, locatedclosetothesouthernedgeofAc-H-6,andintheejctaofa small,freshunnamedcraterlocatedat20.7°N/7.6°E,whichinstead islocatedclosetothenorthernedgeofAc-H-6.Boththesesites re-veala localenhancement intheabundance ofCa-richcarbonates comparedtosurroundingterrains.

4.7.Surfacetemperature

InthecaseoftheHaulaniquadrangle,wehaveseparately con-sideredsurfacetemperature dataretrieved fromVIR infrared ob-servations obtained only in the Survey phase, in order to main-tain uniformity in both the spatialresolution andthe local solar timeorLST(themajorityoftheobservationsacquiredinthis mis-sionphasetookplaceinthecereanmorningbetween10hand12h LST).Forthisreason,therearemajorgapsincoverage,asrevealed byFig.12.Thesurfacetemperaturemapisaccompaniedbyamap ofthe solarincidence angle,so astoenable adirect comparison withtheinstantaneoussolarillumination.

VIR-derived temperatureimages reveal that the only substan-tialthermalsignaturedetectable intheAc-H-6 quadrangle corre-spondstocraterHaulani’smountainousregioninthemiddleofits floor,which is∼8K coolerthan surroundingterrainsobserved at thesamelocalsolartimeandundersimilarsolarillumination,and – toalesserextent– partofitsbrightrimandejecta,hintingthat thealbedooffreshmaterialexcavatedbytheimpact,togetherwith ahighercompactnessofthismaterialinthecrater’speak,maybe responsiblefortheobservedbehavior. Conversely,thermalshades intherestofthe quadrangle,atleastto thespatialscale investi-gatedbyVIRinthisdataset,canbegenerallyexplainedbysolar il-luminationcombinedwiththelocaltopographyatthetimeofthe observation,which suggests that the thickness of the regolith in thisregion,regardlessofalbedodifferences(thatareanywayonly

(13)

Fig. 8. a: Density scatterplot of 2.7-μm band depth (BD27) as a function of 3.1-μm band depth (BD31) computed for the quadrangle Ac-H-6. A linear fit model (white dashed line) is superimposed on the distribution (values of the chi-square ( χ2 ) and the coefficient of determination (i.e., the squared correlation coefficient R 2 ) are also provided).

Most data concentrate around the average values. The cluster shows a large spread and a moderate-to-low correlation in the linear fit model. b: Ratio between the measured 2.7-μm band depth and the BD27/BD31 linear fit model. This map shows where the surface material departs from an ideally linear correlation between the abundance of the two major mineral species. Dark blue to blue colors mark locations that have shallower BD27 than the linear model, while red to dark red shades mark locations that have larger BD27 than the linear model. If BD27 was perfectly correlated with BD31, this map would appear uniformly green. In the map, disconnected lines within a given image are due to the high instantaneous speed of the ground footprints (i.e., projections of the spectrometer’s slit) in the HAMO phase. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

relevantforcraterHaulani)isgenerallylarge enoughtomaskany inhomogeneityintermsofdensity(e.g.,outcropsofcompact ma-terial,pebblesorverycoarsegrainsize)and/orinthethermal con-ductivity(Tosietal.,2014).

CraterHaulani’sthermalsignaturewasfirstnoticedinVIRdata acquiredinFebruary2015 duringtheApproachphase,which cov-ereda substantial fractionof the dayside of Ceres under a solar phaseangleof44.5° Atcoarse spatialresolution (11.4km/px),the region of crater Haulani alreadyshowed a distinct thermal con-trast,being∼5Kcoolerthantherestofthesurfaceasobservedat 11h LST(Tosietal., 2015a). The spatialscaleofthose datacould safelyruleoutshadingeffectsduetolocaltopography.Ceresis ex-pected to havea low thermal inertia, albeit greater than Vesta’s thermalinertia at globaland broadly regionalscale (Titus, 2015; HayneandAharonson, 2015). The maximumdiurnal temperature

isastrongerconstraintthanothertemperaturepointssampledon the dayside, since this maximum shifts from local noon to the afternoon as the thermal inertia of the explored area increases. Therefore,theslowerresponsetoinstantaneoussolarillumination, measuredbyVIRinthelatecereanmorning,couldbeindicativeof asubstantialdeparture(i.e.,increase)inthermalinertiacompared totheglobalaverage,whichisconsistentwiththehigher compact-nessofHaulanicrater’scentralmoundand,toalesserextent,ofits surroundingfeatures.

5. Discussionandconclusions

Ac-H-6 ‘Haulani’ is one of the most interesting quandrangles in the Ceres equatorial belt, and on this dwarf planet in gen-eral.Inthisworkwehaveundertakenathoroughinvestigationof

(14)

Fig. 9. a: 2D scatterplot of 2.7-μm band depth (BD27) as a function of 3.1-μm band depth (BD31) computed for the entire surface of Ceres. b: Same 2D scatterplot of ( a ) superimposed on a four-color scheme that also provides a legend for Fig. 10 . Materials depleted in both Mg-rich phyllosilicates and NH 4 -rich phyllosilicates appear in yellow–

orange colors, while materials enriched in both Mg-rich phyllosilicates and NH 4 -rich phyllosilicates are displayed in blue. The green color marks materials with enhanced

abundance of Mg-rich phyllosilicates and depletion of NH 4 -rich phyllosilicates, while the purple color highlights an opposite situation, i.e. materials depleted in Mg-rich

phyllosilicates and a relatively high abundance of NH 4 -rich phyllosilicates. (For interpretation of the references to color in this figure legend, the reader is referred to the

web version of this article.)

the mineralogyof theHaulani quadrangleas awhole, using var-ioustechniquesdesignedtocomplementandexpandtheanalysis performedinpreviouswork,providingquantitativeresultsthatcan be directlycompared withthoseobtainedforother cerean quad-rangles,which are discussedin other papers ofthisspecialissue by meansofsimilartechniques(Carrozzoetal., 2017b,thisissue;

Combeetal.,2017b,thisissue;DeSanctisetal.,2017b;Longobardo etal., 2017, thisissue;Palomba etal., 2017a;Raponi et al.,2017, thisissue;Singhetal.,2017,thisissue;Stephanetal.,2017b,this issue;Zambonetal.,2017b,thisissue).

The two widespread mineral species discussed in this article, i.e.magnesium-bearingphyllosilicates andammoniated phyllosili-cates,respectivelyidentifiedbydiagnosticspectralsignatures cen-teredat∼2.7and∼3.1μm,showageneralenrichmentinMg-rich phyllosilicates fromnorthwesttosoutheast,whichcorresponds to the oldest geologicunit(cratered terrain).Thedistribution ofthe featurelessspectralendmember,whichmayaffectthebanddepths, isalsooveralluniform,andnoobviouscorrelationbetween miner-alogyandtopographycombinedwiththeinstantaneoussolar illu-minationisobservedinmostoftheHaulaniquadrangle.

In a broader context, these findings are overall consistent with the general mineralogy observed in nearby quadrangles of Ceres. In quadrangle Ac-H-10 ‘Rongo’, located west of quadran-gle Haulani, Mg-bearing phyllosilicates are depleted in the east-ernmostregion includingcraterRongo, anda broaderarea of de-pletion is observed inthe southwestern quadrant, whilethe dis-tribution ofNH4-bearing phyllosilicates is relatively uniform,

ex-cept intheeasternmostregion,which alsoshowsa netdepletion (Zambon et al., 2017b, this issue). Such depletion in both these two main mineralspecies isconsistent withwhat we observein Ac-H-6 west of crater Haulani, and represents a smooth transi-tion between these two adjacent quadrangles. In quadrangle Ac-H-11 ‘Sintana’, located southof quadrangleHaulani, the mineral-ogyisconsistentwithwhatweobserveinthesoutheasternsideof Ac-H-6,withmagnesium-bearingphyllosilicates and ammoniated-phyllosilicates’band intensitythatkeepsincreasingbygoing from lowsouthern latitudestothesouthernmostregions (DeSanctiset al., 2017b,thisissue).QuadrangleAc-H-2‘Coniraya’,locatednorth ofquadrangleHaulani,includesseveralfeaturesofinterest,suchas thewaterice-richcraterOxo(Combeetal.,2016),theorganics-rich

craterErnutet(DeSanctisetal.,2017a),andthepittedterrains-rich crater Ikapati (Sizemore etal., 2017). Here an overall increase in theband depthof magnesium-bearing phyllosilicates is observed from east to west, particularly in the soutwestern corner of the quadrangle,not necessarily correlated with an equal decrease in the band depth diagnostic of ammoniated phyllosilicates (Raponi et al., 2017, this issue). Finally, quadrangle Ac-H-7 ‘Kerwan’, lo-catedeastofquadrangleHaulani,overallshowsenrichmentin Mg-phyllosilicatesinthesouthernregioncomparedtothenorthern re-gion, anddepletion of NH4-rich phyllosilicates in the

northwest-ernregion.Again,thespectralindicesofthesetwomajormineral speciesdonotdisplayasubstantial correlationacrossmostofthe quadrangle,except inspecific regions ofinterest observed at the localscale(Palombaetal.,2017a,thisissue).

From a mineralogical standpoint, the mostprominent surface featurein quadrangleAc-H-6 is by farthe 34-km crater Haulani, one of the youngest geologic features on Ceres, which in turn ismade up by a complex system ofgeologic sub-units with dif-ferent morphology and albedo (Krohn et al., 2017). The floor and the ejecta of crater Haulani display one of the most neg-ative (‘bluest’) spectral slopes at visible to near infrared wave-lengthsacrossthe entiresurfaceof Ceres.Thisblueslope is gen-erallyassociatedwithbrightmaterialunits,whichisindicativeof a younger age and therefore less processed material. The bright andbluest materials observed in and around crater Haulani cor-respondtoasubstantialdecreaseinboththebanddepthsof spec-tral features diagnostic of Mg-rich and NH4-rich phyllosilicates.

In the ejecta, this depletion is more prominent in the north– south direction and at relatively short ranges from the crater, withthedepletion of Mg-richphyllosilicates that extends farther away. The correlation observed between these two spectral in-dices in Haulani’s ejecta is increasingly weaker in moving away fromthe crater. One physical justificationmight be that the en-ergy required to release the hydroxide ion OH− from a phyl-losilicate is less than the energy required to release the ammo-nium ionNH4+, dueto the differentstrengths ofthesechemical

bonds.ThebluishejectaofHaulanialsoextendwestward,crossing theneighboringquadrangleAc-H-10Rongo.Inthatcase,Haulani’s ejecta still show depletion in both Mg-rich and NH4-rich

(15)

Fig. 10. a: Four-color representation of the two spectral parameters: 2.7-μm band depth (BD27) and 3.1-μm band depth (BD31). In this map, all VIR pixels are repre- sented by a color according to the scheme defined in Fig. 9 b: pairs of values (BD27, BD31) are scaled to (X, Y) coordinates of the color scheme. b: Same kind of map as ( a ), but in this case only the purest compositions are represented by colored pixels, while all the others are dark. This is obtained by masking the pixels that host the most frequent values in the pair of parameters. In both maps, disconnected lines within a given image are due to the high instantaneous speed of the ground foot- prints (i.e., projections of the spectrometer’s slit) in the HAMO phase.(For interpre- tation of the references to color in this figure legend, the reader is referred to the web version of this article.)

weakcorrelationatlarge distancefromthecrater (Zambonetal., 2017b,thisissue),consistentwithourfindings within quadrangle Ac-H-6.

Whileneitherwatericenorpeculiarhydratedmineralsare ob-servedtodayintheregionofcraterHaulani,considerable variabil-ity is recorded in its interior. Notable examples are mass wast-ingmaterialfoundinthesoutheastern innerwallandsome mass wasting patches located in the northern inner wall close to the rim,whichshowapositivespectralslopeandanabundanceofthe main mineralspecies comparableto the average valuesrecorded farfromthecrater.

Elsewhere inAc-H-6,negativespectral slopes in thevisibleto near infrared range up to 2.25μm are recorded in specific geo-logicfeatures observed atthe localscale, albeit withsmaller ab-solute values compared to crater Haulani. Notable examples are the smallcrater at13.7°S/11.2°E, in the middle of crater Anura’s floor,andtheejectaoftwounnamedsmallcratersrespectively lo-catedat20.7°N/7.6°Eand 10.0°S/21.0°E. Other local-scalefeatures display negativespectral slopes in the visual rangebut not nec-essarilyinthenearinfrared rangeup to2.25μm.Similarly, a de-creaseintheabundancesofMg-richandNH4-richphyllosilicatesis

foundalsoinotherimpactcratersofvaryingsize(includinggreater thanHaulani’ssize).However,theseoccurrencesdonotaffectlarge areas,butratherpartoftheirfloorsandejecta.

By examining the relationship existing betweenthe values of theband depths at2.7and3.1μmin quadrangleAc-H-6,no par-ticulartrendishighlighted,butamuchstrongercorrelationis ob-servedandmappedoverthoseareasofcraterHaulanithatare sub-stantially depleted in both the two major mineral species, com-paredtootherfeaturesdiscussedinthispaperwhereasimilar de-pletionisobserved.

SincemineralogyonCeresatglobalandbroadlyregionalscales is dominated by magnesium- and ammonium-bearing phyllosili-cates,carbonates,andaspectrallyfeaturelessdarkcomponent(De Sanctisetal.,2015;Ammannitoetal.,2016;Carrozzoetal.,2017a, thisissue),onepossibleexplanationisthattheimpactthatcreated Haulani could haveaffected a portion of Ceres originally rich in hydrated andammoniated material,penetratingthe crustenough toexcavatebrighterandlessprocessedmaterial andcausing inti-matemixingbetweenthesemineralphases.Thisisconsistentwith the evidence that part of crater Haulani’s floor is pitted terrain, whichonCereslikelyformviatherapidvolatilizationofmolecular H2O(possiblywaterice)entrainedinshallowsubsurfacematerials

(Sizemoreetal.,2017).

The work of Schmedemann et al. (2016) and Krohn et al. (2016) determined that the blue-sloped material is the youngest materialonCeres’ssurfaceandinterpretedtheassociationofflow featureswithimpactcratersasindicationforhighlyfluidized ma-terial/impactmelt(Stephanetal.,2017a).Becausetheshallow sub-surface of Ceres may harbor water ice (Prettyman et al., 2017;

Schmidtetal.,2017),theejectaofafreshcraterlikeHaulanimight have been initially a liquid water/phyllosilicate mixture, perhaps asa consequenceofimpact meltingofsubsurface waterice, that turned blue after desiccation (Schröder et al., 2017). The physi-cal smoothness ofthe regolith is consistent withthe initial fluid condition.Alternatively,spectral bluingwouldbe causedby ultra-fine(∼50–100nminsize)grainsthatstronglysticktogether form-ing larger agglomerates (tensof micrometerin size)and/or by a higher abundance ofamorphous material (Stephan etal., 2017a). In both ofthesescenarios, spaceweathering, interms ofdiurnal temperaturevariationsandmicrometeoriticimpacts forminga re-golithlayeroffine-grainedphyllosilicatedust,maybe responsible forthefadingofthebluecolorovertime(Stephanetal.,2017a).

While nowadays the diagnostic absorptions of magnesium-bearing phyllosilicates and ammoniated phyllosilicates are much shallower in the floor of the crater, the impact properties have been able to preserve to some extent the original mineralogyin otherspecificareas,suchaspartsoftheinnerwalls.Ontheother hand, Ca- and possibly Na-rich carbonates found in part of the walls, rim,floorandejectaofcrater Haulani,whichare discussed separately both ina broader context (Carrozzoet al., 2017a,this issue;Palomba etal., 2017b,thisissue) andspecificallyforcrater Haluaniitself(Tosietal.,2017,underreview),arealsoverylikelya byproductoftheimpacteventandarerelatedtobrightermaterial units,similartowhathasbeenobservedinotherspecificplaceson Ceres,mostnotablyincrater Occator(De Sanctisetal., 2016).An enrichmentincarbonates,especiallywhenassociatedwithalower local content inOH-rich andNH4-rich phyllosilicates, is one

evi-dence insupport of intense hydrothermal processestriggered by impactevents,similartothoseobservedinother specificsiteson Ceres(DeSanctis etal., 2016;Combe etal., 2017a, thisissue).In thisregard, itisworth notingthatelsewhereinquadrangle Ac-H-6other impactcraters largerthanHaulanidonotshow thesame contrastintermsofmineralogicalindices(albedo,reducedoreven negativespectral slopes,depletion inMg-rich andNH4-rich

phyl-losilicates, presence ofCa- and Na-carbonates),in supportof the hypothesisthatinitialconditionsofintermsofshallowsubsurface

(16)

Fig. 11. Distribution of 4.0-μm band’s center ( a ) and band depth ( b ) values across quadrangle Ac-H-6 Haulani, derived from VIR data. These parameters are respectively sensitive to the composition and abundance of carbonate minerals. The overall distribution spans from Mg- or (Mg,Ca)-rich carbonates (violet to blue) to Na-rich carbonates (red) ( a ), and from lower depth (violet to blue) to greater depth (red) ( b ). Local enhancement of Ca-rich and possibly Na-rich carbonates is found locally in the region of crater Haulani, most notably in its western inner wall and ejecta closest to the southern rim. Other evidences of Ca-rich carbonates are found in part of Kondos crater’s floor and in the ejecta of a fresh unnamed crater located at 20.7 °N/7.6 °E, even though with smaller concentrations compared to crater Haulani. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

composition andthecharacteristicsoftheimpactevent(e.g., spe-cificenergy)mayprovecrucialintriggeringtheseprocesses,which wouldotherwisebewidespreadonthesurface.Withinquadrangle Ac-H-6,onlythe44-kmcraterKondos,whichdugdeeperintothe crust, shows some mineralogical similaritywithrespect to crater Haulani, even though this evidence is found in a much smaller area,namelypartofitsfloor.Thissuggeststhatcraters olderthan Haulani have experienced relaxationand mass wasting phenom-ena that somehowmixed pre- and post-impact composition, ul-timately reducing the extension and contrast of the carbonate-richareasandachieving,overtime,anintermediate abundanceof magnesium-richphyllosilicatesandammoniatedphyllosilicates.

CraterHaulanishowsadistinctthermalsignature,beingcooler thanthesurroundingregionsobservedbyVIRundersimilar illumi-nationconditions andduringmaximumdiurnalinsolation,which

suggestsa higher thermalinertia, i.e.a local slowerresponse to instantaneousinsolationcomparedtotheaveragesurfaceofCeres. Thisthermalcontrastdoesnotariseinthepittedpartofthefloor (Sizemoreetal.,2017),butratherinthecentralmoundandinthe ejectaclosertothe rim,matchingbrightmaterialwiththebluest spectralslope.Inthisregard,bluishmaterialobservedinother sur-facefeaturesonCeresdoesnotdisplayanequallyprominent ther-mal contrast. One explanation bounding mineralogical and ther-malevidencesisthat partofthematerialthat meltedduringthe impact event may have remained relatively compact up to now, whichwouldexplainthehigherthermalinertiaobservedinthose specificsites.

Althoughthevarietyofcolorsandspectralinformationthatthe DawnmissionwasabletogetintheHaulaniquadrangleis some-whatpeculiarcomparedtomanyotherplacesonCeres,thepicture

Riferimenti

Documenti correlati

Their main conclusions were: (1) confirmation of Ar- gand’s Adria as the Africa promontory, as ancho- red in extensive paleomagnetic data from Adria as reviewed by

OSPZ6. Palynomorphs recovered from Chal i Sheh, Darreh Yas, Kuh e Faraghan and Zard Kuh sections and from core-samples from Well-2, 3, 5, 6 and 8 al- lows the recognition

From the Upper Triassic to the Middle Jurassic the following units can be distinguished (Dozet & Strohmenger 2000; Miler & Pavšič 2008; Dozet 2009): 1)

Our investigation led us to conclude that the optimal region to sample for geochemical and isotope analyses is the middle part of the mid-section of the shell, avoiding the

Cette Dignité, conclut Vico, nous montre que «philosophes et philologues sont restés à mi-chemin de la vérité ; les premiers pour n’avoir pas assis leur argumentation sur la

The prototype consists of three function blocks: local oscillator controlled in voltage (VCO), a latch block that receives the command from the outside to save the measurement

One of the main challenges of the experiment will be the design of detector to veto the photons at small polar angle (SAC)2. The SAC must intercept photons from K L decays that

– The diffuse emission γ-ray spectrum measured by Fermi LAT and Milagro in the inner Galactic plane ( |b| < 2 ◦ , 30 ◦ < l < 65 ◦ ) is compared with the predictions of