In the format provided by the authors and unedited.SupplementaryInformation

Remotedetectionofwidespreadindigenouswaterinlunarpyroclasticdeposits

RalphE.Milliken1&ShuaiLi1*

Affiliations:1DepartmentofEarth,Environmental,andPlanetarySciences,BrownUniversity,Providence,RI02912

*Correspondenceto:njushuaili@gmail.com

Contentsofthisfile:SupplementaryText

SupplementaryFigures1-6SupplementaryTables1-5

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

SUPPLEMENTARY INFORMATIONDOI: 10.1038/NGEO2993

NATURE GEOSCIENCE | www.nature.com/naturegeoscience 1

MappingofwaterinlunarpyroclasticdepositsAssumptionsofGrainSizefortheLunarSurface BasedonourlaboratoryandnumericalsimulationsitisclearthattheESPAT-wt.%H2Otrendisdependentonparticlesize(seeMethods).However,onlyseverallocationsontheMoonhavebeendirectlysampled,andnoneoftheselocationscorrespondtothelarge,regionaldarkmantledeposits(DMDs)discussedhere.Actualmeangrainsizesandranges,includingthoseforglassbeadsatregionalDMDs,thusremainunknownformuchofthelunarsurface.GrainsizeandsortinginreturnedApollosampleshavebeenmeasuredandaresummarizedin(21),whereitisnotedthatthemeangrainsizeoflunarsoilsrangesfrom40-800µmandmostmeanvaluesfallbetween45-100µm.Basedonthelatter,wehaveadoptedanaverageparticlesizeof60-80µmtorepresentthelunarsurface(regolith)inourstudy,arangethatfallswithinthetypicalmeasuredvaluesforlunarsoils.Ithasalsobeenobservedthatglass-richsamples(e.g.,greenglassfromApollo15andorangeglass/blackbeadsfromApollo17)exhibitsmallermeangrainsizesthantypicallunarsoils,closertoameanvalueof~40µm21.Tobeconservativewehaveadoptedthesame~60-80µmgrainsizeforthedarkmantledeposits,thoughadoptingsmallervalues(e.g.,40µm)wouldresultinhigherestimatesofH2Ocontent. ThoughApollopyroclastic-richmaterialsarebettersortedandexhibitsmallermeangrainsizesthantypicallunarsoils21,ithasalsobeenshownthatthefiner(<25µm)sizefractionofsoilsistheonemostsimilartospectraofbulksoilsatvisibleandnear-infraredwavelengths50(thoughallcomponentswillcontributetotheobservedspectralsignaturetosomedegree).Therefore,althoughtypicallunarsoilshavelargermeangrainsizesthanpyroclastic-richsoils,itispossiblethatthespectralbehaviorofthesetwosoiltypesissimilarintermsoftheir‘effective’grainsize.However,thesegrainsizeeffectswereevaluatedonlyforlabspectraforwavelengths<2.6µm,notfortheOH/H2O-bearing3µmwavelengthregion,anditispossiblethatthespectrallydominantsizefractionvarieswithwavelength.Inaddition,watermaybepreferentiallyhostedinagglutinates22,whichtendtobelargerparticlesandmorevolumetricallyabundantincoarsersizefractionsoflunarsoils50.Ifcorrect,thiswouldsuggestthatthewaterabsorptionsobservedinthe3µmregionofM3dataaremorerepresentativeofthelargesizefractionoflunarsoilsandnotthe‘finestfraction’. Theseissueshighlightthecomplexityoflinkingparticlesizeforlunarregolithmaterialstospectralobservations.Forthepurposeofthisstudy,akeyquestioniswhetherornottheproposeddifferencesinwatercontentbetweenpyroclasticdepositsandsurroundingterrains(‘typical’lunarregolith)canbeexplainedsimplybydifferencesinparticlesize.Themostextremecasewouldbetoassumefineparticlesdominatethe3µmspectralregionfortypicalhighland/mareregolithwhereaslargeparticlesdominatethesamewavelengthregionforDMDs.Asanexample,~90%ofM3pixelswithinthe±35°latitudezoneexhibitEPATvaluesbetween0-0.01,whichwouldyieldanestimateof0~60ppmH2Oassuming60-80µmdiameterparticles.Ifitisinsteadassumedthatafinerfractionisthespectrallydominantcarrierofthewatersignaturefortheseregions,thenan

assumedparticlesizeof<45µmwouldyieldanupperlimitof≤180ppmH2Oforsoilsinthislatitudezone.Incontrast,ESPATvaluesforpyroclasticdepositsapproachvaluesof0.04-0.06.Fortheseregionstoexhibitasimilarwatercontentof180ppmthespectrallydominantparticlesizeat~3µmwouldneedtobeontheorderof≥100µm,or≥2.5timesthemeanparticlesizethathasbeenmeasuredforactualglass-richApollosoils.ThisisanunlikelyscenariogiventhatbothDMDsandsurroundingregionsaresubjecttomicrometeoritebombardmentandcomminutionandthatmeasurementsofreturnedsamplesindicateglass-richsoilsexhibitsmallermeangrainsizesthantypicallunarsoils,asdescribedabove.

WealsonotethatthemajorityofM3spectrawithinthe±30°latitudezoneexhibitnoorveryweakabsorptionsnear3µm(e.g.,ESPATvaluescloseto0;seeFig.3b),thustheywouldalwaysyieldH2Ocontentsatorcloseto0ppmregardlessofwhichvaluesareassumedforparticlesize.Incontrast,spectraforpyroclasticdepositsexhibitclearwaterabsorptionsandarethusmorehydratedthansurroundingterrains.ThisisparticularlyclearatlargeDMDssuchasAristarchus,Harbinger,RimaBode(seeFig.1)andatsmallerDMDssuchasthoseshowninFig.2,whereESPATvaluesare0(noabsorptionbandobserved)interrainsthatsurroundthepyroclasticdeposits.Together,theseobservationsindicatethatpyroclasticdepositsarepreferentiallyenrichedinOH/H2Ocomparedwithtypicallunarregolithinthe±35°latitudezone.Thoughcurrentlylacking,grainsizemeasurementsofdirectlysampledDMDmaterialswouldprovidethebestconstraintonestimatesofabsolutewatercontentforthesedeposits.VariationsinestimatedwatercontentasafunctionofassumedparticlesizecanbeevaluatedbyusingtheslopesandtrendspresentedinSupplementaryFigure2.If~100-125µmdiameterparticlesarepreferred,forexample,thentheESPATvaluesforfiguresshownhereshouldbemultipliedby0.34.ApplicationtoM3DataandComparisonwithPreviouslyMappedDMDs Asnotedinthemaintext,themajorityofpyroclasticdeposits(DarkMantleDeposits,DMDs)classifiedas‘verylarge’and‘large’byGaddisetal.18exhibitevidenceofincreasedwatercontentcomparedtosurroundingterrains,asdomanysmallerdeposits,includingseveralthatwererecentlyidentifiedbyGustafsonetal.19.Inthecaseofthelatter,ourresultsconfirmthatsomeofthesedepositsarelikelytoberelatedtovolatile-drivenpyroclasticeruptions.Therelationshipbetweenourwatercontentmapsandpreviouslyreported‘verylarge’pyroclasticdepositsispresentedinSupplementaryFig.3,wherethenumberscorrespondtothedepositslistedinSupplementaryTable5(locationsandareasarefromGaddisetal.18).Aclose-upviewofwaterassociatedwithdarkmantledepositsatOrientaleispresentedinSupplementaryFig.4,andhistogramsofestimatedwatercontentsforselectdepositsarepresentedinSupplementaryFig.5.ThehistogramsdemonstratethatasignificantnumberofM3pixelsforeachdepositareaboveanassumedbackgroundlevelof100ppm(estimatedbackgroundwaterduetosolarwindhasnotbeenremovedintheseplots).Absolutewatercontentestimatesarebasedonanassumptionof60-80µmaveragediameterregolithparticles;histogramswouldshifttolowervaluesassuminglargerparticlesandhighervaluesassumingsmallerparticles.Basedonthisparticlesize,itisclearthatmostpixelswiththesepyroclasticdepositscontain<400ppmOH/H2O(after

subtracting100ppmofbackground).Notallpreviouslymappedpyroclasticdepositsarevisibleattheglobal/regionalscaleofSupplementaryFigs.3-4(seeexamplesofsmallerdepositsinFig.2).SinusAestuum,forexample,doesexhibitasmallregionofenhancedwatercontent,asshowninSupplementaryFig.6,butthisregionismuchsmallerthanpreviouslyreportedextentsofthisdeposit.ThisindicatesthatSinusAestuumislesshydratedthanotherpyroclasticdepositsofsimilararealextent.CompilationofWaterContentsMeasuredinLunarSamples

ThedatapresentedinSupplementaryFig.2cforlunarsamplesisacompilationofpreviouslypublishedvalues51-59.Resultsfromselectpyrolysismeasurementsthatwereperformedonlunarregolith,mineralseparates,agglutinates,androckssampledfromtheApollo11,12,14,15,16,and17landingsitesarepresentedinSupplementaryTable3.Forcomparisonwithpyrolysismeasurements,SIMSresultsforApollo11agglutinate,Apollo15volcanicglasses,Apollo16agglutinate,andApollo17agglutinateandvolcanicglassesarepresentedinSupplementaryTable4.Ithasbeenarguedthatwaterobservedintheseagglutinatesarelargelytheproductofsolarwindprocesses22,whereaswaterobservedinthevolcanicglassesareindicativeofwaterinmagmasourceregions5-6.

Forpyrolysisresults,onlywateramountsmeasuredfromtemperaturesabove

300°Careincludedinordertominimizepotentialeffectsofadsorbedterrestrialwater.However,itmustbeconsideredthatalldatamaysufferfromsomedegreeofterrestrialcontaminationthatcouldintroducelargeuncertaintiesintheδDvalues(whereδD=[(D/H)sample/(D/H)standard-1]×1000),whichisoneofthemostimportantindicesusedtodeterminetheoriginoflunarwater6,22,51-59.Indeed,thelargerangeinδDvaluesfromthepyrolysisapproach,someofwhichoverlapterrestrialvalues,havebeenconsideredanindicationoflikelyterrestrialcontamination51-57.However,recentmeasurementsonsoilgrainsandagglutinatealsoconfirmthatthesematerialscontainwateroriginatingfromsolarwindimplantation22-23.Regardlessoforigin,itcanbeseenthattheaveragewatercontentofexaminedApollosamplesis63.0ppm(SupplementaryTables1-2andSupplementaryFig.2),withmostvalues<100ppm.Therefore,weadoptavalueof100ppmtorepresentlunarsurfacewaterthatislikelyattributabletosolarwindprocesses.Valuesabovethisthresholdatequatoriallatitudesareconsideredtoexhibit‘excess’waterfromothersources.

References50. Pieters, C., Fischer, E., Rode, O. & Basu, A. Optical effects of space weathering: the role

of the finest fraction. Journal of Geophysical Research 98, 20817-20824 (1993). 51. Epstein, S. & Taylor Jr., H. The concentration and isotopic composition of hydrogen,

carbon and silicon in Apollo 11 lunar rocks and minerals. Proceedings of the Apollo 11 Lunar Science Conference, Supplement in Geochemica et Cosmochimica Acta 1, 1085-1096 (1970).

52. Epstein, S. & Taylor Jr., H. O18/O16, Si30/Si28, C13/C12 and D/H studies of Apollo 14

and 15 samples. Proceedings of the Third Lunar Science Conference, Supplement 3 in Geochirnica et Cosmochimica Acta 2, 1429-1454 (1972).

53. Epstein, S. & Taylor Jr., H. The isotopic composition and concentration of water,

hydrogen, and carbon in some Apollo 15 and 16 soils and in the Apollo 17 orange soil. Proceedings of the Fourth Lunar Science Conference, Supplement 4 in Geochimica et Cosmochimica Acta 2, 1559-1575 (1973).

54. Friedman, I., O’neil, J. R., Adami, L. H., Gleason, J. D. & Hardcastle, K. Water,

hydrogen, deuterium, carbon, carbon-13, and oxygen-18 content of selected lunar material. Science 167, 538-540 (1970).

55. Friedman, I., O’Neil, J. R., Gleason, J. D. & Hardcastle, K. The carbon and hydrogen

content and isotopic composition of some Apollo 12 materials. Proceedings of the Second Lunar Science Conference, MIT Press 2, 1407-1415 (1971).

56. Friedman, I., Hardcastle, K. G. & Gleason, J. D. Water and carbon in rusty lunar rock

66095. Science 185, 346-349 (1974). 57. Merlivat, L., Lelu, M., Nief, G. & Roth, E. Deuterium, hydrogen, and water content of

lunar material. Proceedings of the Fifth Lunar Science Conference, Supplement 5 in Geochimica et Cosmochimica Acta 2, 1885-1895 (1974).

58. Greenwood, J. P. et al. Hydrogen isotope ratios in lunar rocks indicate delivery of

cometary water to the Moon. Nature Geoscience 4, 79-82 (2011). 59. Saal, A., Hauri, E., Van Orman, J. A. & Rutherford, M. Hydrogen isotopes in lunar

volcanic glasses and melt inclusions reveal a carbonaceous chondrite heritage. Science 340, 1317-1320 (2013).

SupplementaryFig.1. ESPAT/watercontentmapsderivedfromM3databrokendownbyopticalperiod.Watercontentvaluesarebasedontheassumptionof60-80µmdiameterparticles;estimatedsolarwindbackgroundvalueof100ppmhasnotbeensubtracted.

SupplementaryFig.2.DatausedtoestimateESPAT-wt.%H2Otrendforvolcanicglassesandestimatesoflunarbackgroundwatercontent.(a)Terrestrialanorthosite,terrrestrialMORBglasses,andsyntheticglassesoflunarcompositiondefinelinearESPAT-H2Otrendsforwhichtheslopevarieswithparticlesize(symbols).Similartrendsarepredictedfromforward-modeledreflectancespectrausingmeasuredextinctioncoefficientsforvolcanicglass(dashedlines).(b)zoomed-inportionoutlinedbygrayboxin(a)showingsimilaritybetweenmeasuredandcalculatedvaluesatlowwatercontent;watercontentsaresimilarfordifferentcolorsbecausethesamesampleswereusedtoevaluateparticlesizeeffects.(c)HistogramofwatercontentsfromM3data(graybars)inthe±30°latitudezoneassuming60-80µmdiameterparticlesfromtrendshownin(a).WaterinlunarsamplesdeterminedbypyrolysisandSIMSmeasurementsareshownforcomparison.TypicalvaluesforM3inthiszoneare<100ppm,similartowaterlossesobservedinpyrolysismeasurementsandlikelyattributabletoasolarwindorigin.(d)ExamplelaboratoryreflectancespectraforsyntheticorangeandlunarglassesshowingbroadOH/H2Oabsorptionnear~3µm.WavelengthpositionforESPATcalculation(2.85µm)ismarkedandisnearthemaximumabsorptionpoint.RELABsampleIDsRM-REM-139AandRM-REM-140A.

SupplementaryFig.3. ESPAT/watercontentmapsderivedfromM3data(allopticalperiods)comparedwithdistributionof‘verylarge’pyroclasticdepositslistedinGaddisetal.18.Watercontentvaluesarebasedontheassumptionof60-80µmdiameterparticles;estimatedsolarwindbackgroundvalueof100ppmhasnotbeensubtracted.a)globalviewforlatitudezone±35°.b)zoomed-inviewforaportionofthenearsideoftheMoon.NumberscorrespondtopyroclasticdepositslistedinSupplementaryTable5.Thoughnotvisibleatthescaleofthesemaps,nearlyallofthenumbereddepositsexhibitsomedegreeofenhancedhydrationrelativetosurroundingterrains.

SupplementaryFig.4. DarkmantledepositsinferredaspyroclasticdepositsalongsouthwesternrimofOrientale.a)LROCwide-anglecamera(WAC)imageshowingdarkmaterialsringingapotentialvent(yellowarrow);b)ESPAT/H2Omapoverlainonimageshownin(a).

SupplementaryFig.5. Histogramsofestimatedwatercontentvaluesforselectpyroclasticdeposits.Valuesarebasedontheassumptionof60-80µmdiameterparticles.Allpixelsmayinclude~100ppmofwaterattributabletosolarwindimplantation,butnumerouspixelsarebeyondthisvalueforthesedeposits.Pixelsintheseregionswithvaluesof0(nowaterabsorptionfeature)werenotincludedinthisanalysis.

SupplementaryFig.6. WatercontentofpyroclasticdepositsatSinusAestuum.a)LunarReconnaissanceOrbiterCamerawideangle(WAC)imageshowinglocationsofdarkmantledepositsinferredtobepyroclasticdeposits.b)ESPAT/H2Omap(1km/pixel)overlainonWACimage.Notethatmostofthedarkmantlematerialdoesnotexhibitevidenceforincreasedwatercontent.c)Histogramofwatercontentsformapshownin(b);veryfewpixelsareabovetheestimated~100ppmsolarwindbackgroundatthisspatialscale,indicatingthisdepositcontainslesswaterthanotherlargepyroclasticdeposits.Pixelswithvaluesof0(nowaterabsorptionfeature)werenotincludedinthisanalysis.

SupplementaryTable1. Watercontentsforterrestrialmid-oceanridgebasaltglassescorrespondingtodatashowninSupplementaryFig.2.

SampleName InitialWaterContent(ppm) ReferenceD9A 976.49

[Shimizuetal.,2015]D38A 16227.00D44A 815.44D44B 888.00

SupplementaryTable2. WatercontentsforglassesofsyntheticlunarcompositionusedtoderiveESPAT-wt.%H2OtrendsandcorrespondingtodatashowninSupplementaryFig.2.

SampleName WaterContent(ppm) ReferenceOrangeglass(OGV-5) 390 [Wetzel,2014]Yellowglass(YG-5) 305

SupplementaryTable3. WatercontentsandδDvaluesforApolloregolith,rock,andmineralsamplesmeasuredviapyrolysiscorrespondingtodatashowninSupplementaryFig.2.

Reference SampleName H2O(ppm) δD(‰) Temperature

Merlivatetal.,1974

15600regolith 46.8 -59

>600°

15600regolith 68.4 -9568501regolith 93.6 -11572501regolith 90 -8872501regolith 82.8 -8278501regolith 104.4 -72

15229agglutinate 228.6 -5215229agglutinate 7.2 -242

70215basalt 48.6 -16575035basalt 44.6 -9275035basalt 50.4 -252

EpsteinandTaylor,1973

15021.4regolith 48.6 -252 630°-melting64421.21regolith 109.8 -280 500°-melting65513.1regolith 268.2 -173 500°-melting62221.8regolith 199.8 -88 500°-melting74220.22regolith 46.8 -154 500°-melting

Friedmanetal.,1971 12030.30regolith 7 -300 430°-1025°

Friedmanetal.,1970

10046.21agglutinate 455 -756 300°-950°10046.22agglutinate 372 -576 300°-950°10060.11agglutinate 152 -865 300°-950°

EpsteinandTaylor,1970

10084.2rockandminerals 90.4 -380 500°-1400°

10084.1rockandminerals 208.4 -268 500°-1400°

10061rockandminerals 160.2 -409 500°-1400°

EpsteinandTaylor,1972 14422regolith 48.6 -419 550°-600°

Friedmanetal.,1974

66095.62rock 9 -194 690°-1350°66095.31rock 13.5 -153 690°-1350°

SupplementaryTable4. WatercontentsandδDvaluesforApolloglassandagglutinatesamplesmeasuredviaSIMScorrespondingtodatashowninSupplementaryFig.2.

SampleName H2O(ppm) δD(‰) Reference

10084Agglutinate

470 4555

Liuetal.,2012

27 5413225 -562282 -789246 -843161 -731188 191207 267

70051Agglutinate200 -711209 -844203 -781

64501Agglutinate 79 -558

74220Glass

5.0 446

Saaletal.,2013

3.5 1944.6 2835.5 4124.6 4534.9 2914.4 2353.8 4163.2 2698.7 3926.7 2995.7 3303.7 4063.1 3963.5 307

15426LowTiglass

69.3 47451.9 34740.7 34939.2 45934.8 80432.7 219930.8 225828.1 84327.9 54527.8 224425.1 50222.8 46922.0 121016.8 216016.3 580

15427VeryLowTiglass

33.8 57828.6 180214.4 404512.1 104311.4 83010.8 2099.4 7619.1 10288.7 6167.4 806.9 3716.4 -3786.2 487

6.1 6266.1 25.7 -6165.6 -935.5 -1195.5 -2135.5 -1395.3 3185.3 8375.3 505.1 5234.9 2734.4 -3384.4 -1654.2 22094.1 -7333.6 -651

SupplementaryTable5.Locationsandareasforlunarpyroclasticdepositsclassifiedas‘verylarge’inGaddisetal.18.DepositnumbercorrespondstonumbersshowninSupplementaryFig.3.Deposit# Name Latitude Longitude Area(km2)

1 Aristarchus 26.7 -52.3 490132 Aestuum 6.6 -5.9 103573 RimaBode 11.9 -3.4 66204 SmythiiNW 1.1 84.8 58515 Orientale -30.3 -97.5 53216 Cruger -16.7 -66.5 48287 SulpiciusGallus 21.7 9.4 43228 Vaporum 10 7.9 41299 TaurusLittrow 20.2 30.7 294010 NectarisSE -22.4 40.5 290511 Harbinger 26.6 -43.4 287712 Dopplemayer -28.1 -40.5 262813 SmythiiSW -6 85 253914 Titius -26.7 103.9 215915 Petavius -23.5 61 164516 OppenheimerNW -34.8 -168.2 150017 HumorumSW -26.6 -44.4 147218 Riccioli -2.4 -75.5 108919 Cleomedes 26.6 54.7 108420 MoscovienseSE 24.7 151.2 1022