WORKSHOP ON NEW VIEWS OF THE MOON II:

UNDERSTANDING THE MOON THROUGH THE INTEGRATION OF DIVERSE DATASETS

SEPTEMBER 22-24, 1999 FLAGSTAFF, ARIZONA WORKSHOP ON NEW VIEWS OF THE MOON II: UNDERSTANDING THE MOON THROUGH THE INTEGRATION OF DIVERSE DATASETS

Flagstaff,Arizona September 22-24, 1999

Conveners Lisa Gaddis and CharlesK. Shearer

Sponsored by Lunar andPlanetary Institute Museum of NorthernArizona National Aeronautics and Space Administration U.S. Geological Survey, Flagstaff

Lunar and Planetary Institute 3600 Bay Area Boulevard Houston TX 77058-1113

LPI ContributionNo. 980 Compiled in 1999 by LUNAR ANDPLANETARY INSTITUTE

The Institute is operated by the Universities Space Research Association under Contract No. NASW- 4574 with the National Aeronautics and Space Administration.

Material in this volume may be copied without restraint forlibrary, abstract service, education, or personal researchpurposes; however, republication of any paper or portion thereof requires the written permission of the authors as well as the appropriate acknowledgment of this publication.

Abstracts in this volume may be cited as

Author A. B. and Author C. D. (1999) Title of Abstract. In Workshop on New Views of the Moon II: Understanding the Moon Throughout the Integration of Diverse Datasets. LPI Contribution No. 980, Lunar and Planetary Institute, Houston.

This volume is distributed by

ORDER DEPARTMENT Lunar and Planetary Institute 3600 Bay Area Boulevard Houston TX 77058-1113, USA Phone: 281-486-2172 Fax: 281-486-2186 E-mail: [email protected]

Mail order requestors will be invoiced for the cost of shipping and handling. LP] Contribution No. 980 iii

Preface

• This volume contains abstracts that have been accepted for presentation at the Workshop on New Views of the Moon II: Understanding the Moon Through the Integration of Diverse Datasets, September 22-24, 1999, in Flagstaff, Arizona. The workshop conveners are Lisa Gaddis (U.S. Geological Survey, Flagstaff) and Charles K. Shearer (University of New Mexico). Color versions of some of the images contained in this volume are available on the meeting Web site (http://cass.jsc.nasa.gov/meetings/moon99/ pdf/program. pdf). Logistics, administration, and publications support for this meeting were provided by the staffof the Publications and Program Services Department at the Lunar and Planetary Institute.

LP/ Contribution No. 980 V

Contents Prospecting forLunar Resources with Global Geochemical and Multispectral Data C. C. Allen ...... I

Photometric Imaging of the Moon fromthe Robotic Lunar Observatory J. M. Anderson and H. H. Kieffer ...... 2

The Lunar Crustal Thickness fromAnalysis of the Lunar Prospector Gravity and Clementine Topography Datasets S. Asmar, G. Schubert, A. Konopliv, and W. Moore ...... 3

Three Paradigms of Lunar Regolith Evolution A. Basu, D. S. McKay, and S. J. Wentworth...... 4

Lunar Global Petrologic Variations D. B. J. Bussey, P. D. Spudis, and J. J. Gillis ...... 5

Integrationof the Ultraviolet-Visible Spectral Clementine Data and the Gamma-Ray Lunar Prospector Data: Preliminary Results ConcerningFeO, Ti 02, and Th Abundances of the Lunar Surface at Global Scale S. D. Chevrel, P. C. Pinet, G. Barreau, Y. Daydou, G. Richard, S. Maurice, and W. C. Feldman ...... 6

The Distribution of Titanium in Lunar Soils on the Basis of Sensor and In Situ Data Fusion P. E. Clark and L. Evans ...... 7

Digital Elevation Models of the Lunar Surface A. C. Cook and M. S. Robinson ...... 8

Simulating the Formation of Lunar Floor-Fracture Craters Using Elastoviscoplastic Relaxation A. J. Dombard and J. J. Gillis ...... :...... : ...... IO

Sample ReturnMission to the South Pole Aitken Basin M. B. Duke, B. C. Clark, T. Gamber, P. G. Lucey, G. Ryder, and G. J. Taylor ...... 11

Subpixel Detection of Pyroclastic Materials in Clementine Ultraviolet-Visible Data W. H. Farrand and L. R. Gaddis ...... 12

Enhanced Hydrogen Abundances Near Both Lunar Poles W. C. Feldman, S. Maurice, D. J. Lawrence, I. Getenay, R. C. Elphic, B. L. Barraclough, and A. B. Binder...... 14

Laser Argon-40-Argon-39Age Studies of DarAl Gani 262 Lunar Meteorite V. A. Fernandes, R. Burgess, and G. Turner...... 15 vi Workhop on New Views of the Mon I

Progress Toward Characterization of Juvenile Materials in Lunar Pyroclastic Deposits L. R. Gaddis ...... 16 Lateral and Vertical Heterogeneity of Thorium in the Procellarum KREEPTe rrane: As Reflected in the Ejecta Deposits of Post-Imbrium Craters J. J. Gillis and B. L. Jolliff...... 18 The Optical Maturity of Ejecta fromLarge Rayed Craters: Preliminary Results and Implications J. A. Grier,A. S. McEwen, PG. Lucey, M. Milazw, and R. G. Strom ...... 19 The Lunar Imager/SpectroMeter for the SELENE Mission J. Haruyama, H. Otake, T. Matsunaga, and the LISM Working Group ...... 21 The Composition and Origin of Selected Lunar Crater Rays B. R. Hawke, D. T. Blewett, P G. Lucey, C. A. Peterson, J. F. Bell III, B. A. Campbell, and M. S. Robinson ...... 22 Lunar Gruithuisen and Mairan Domes: Rheology and Mode of Emplacement J. W Head and L. Wilson ...... 23 A Multispectral Analysis of the FlamsteedRegion of Oceanus Procellarum D. J. Heather, S. K. Dunkin, P D. Spudis, and D. B. J. Bussey ...... 24 Petrogenesis of Magnesian-Suite Troctolites and Norites P C. Hess and E. M. Parmentier ...... 26 Ages of Oceanus Procellarum Basalts and Other Nearside Mare Basalts H. Riesinger and J. W Head Ill...... 27

High-Resolution Mapping of Lunar Crustal Magnetic Fields: Correlationswith Albedo Markings of the Reiner Gamma Class L. L. Hood, A. Yingst, D. L. Mitchell, R. P Lin, M. Acuna, and A. B. Binder ...... 28 Solar-Wind-ImplantedVolatiles in the Lunar Regolith J. R. Johnson, T. D. Swindle, and P G. Lucey ...... 29

Thorium Enrichment Within the Procellarum KREEPTe rrane: The Record in Surf ace Deposits and Significancefor Thermal Evolution B. L. Jolliff, J. J. Gillis, and L.A. Haskin ...... 31

Space Weathering in the Fine Size Fractions of Lunar Soils: Soil Maturity Effects L. P Keller, S. J. Wentworth, D. S. McKay, L. A. Taylor, C. Pieters, and R. V.Morris ...... 32

The Deepest Lunar SPA Basin and Its Unusual Infilling: Constraints Imposed by Angular Momentum Considerations G. G. Kochemasov ...... 34 LP/ Contribution No. 980 vii

North-Polar Lunar Light Plains: Ages and Compositional Observations U. Koehler,J. W. Head Ill., G. Neukum, and U. Wolf...... 34

Lunar Meteorites and Implications for Compositional Remote Sensing of the Lunar Surface R. L. Korotev ...... :...... 36

Iron Abundances on the Moon as Seen by the Lunar Prospector Gamma-Ray Spectrometer D. J. Lawrence, W. C. Feldman, B. L. Barraclough, R. C. Elphic, S. Maurice, A. B. Binder, and P. G. Lucey ...... 38

DiscriminationBetween Maturity and Composition from Integrated Clementine Ultraviolet-Visibleand Near-InfraredData S. Le Mouelic, Y. Langevin, S. Erard, P. Pinet, Y. Daydou, and S. Chev rel...... 39

Topographic-Compositional Relationships Within South Pole Aitken Basin P. G. Lucey, J. Boltzmann, D. T. Blewett, G. J. Taylor, and B. R. Hawke ...... 41

High-Energy Neutrons fromthe Moon S. Maurice, W. C. Feldman, D. J. Lawrence, R. E. Elphic, 0. Gasnault, C. d'Uston, and P. G. Lucey ...... 42

The Interior of the Moon, Core Formation, and the Lunar Hotspot: What Samples Te ll Us C. R. Neal ...... 43

Asymmetric Evolution of the Moon: A Possible Consequence of Chemical Differentiation E. M. Parmentier, S. Zhong, and P C. Hess ...... 44

The Distribution of Anorthosite on the Lunar Farside C. A. Peterson, B. R. Hawke, P G. Lucey, G. J. Taylor, D. T Blewett, and P D. Spudis ...... 46

The Moon as a Spectral Calibration Standard Enabled by Lunar Samples: The Clementine Example C. M. Pieters ...... 47

Regional Dark Mantle Deposits on the Moon: Rima Bode and Sinus Aestuum Analysis S. Pinori and G. Bellucci ...... 49

Lunar Elemental Abundances fromGamma-Ray andNeutron Measurements R. C. Reedy and D. T. Vaniman ...... 50

Lunar South Pole Topography Derived fromClementine Imagery M. R. Rosiek, R. Kirk, and A. Howington-Kraus ...... 52

Intention and Intension in the Integration of Lunar Data Sets: The Great Instauration G. Ryder ...... 53 viii Workshop on New Views of the Moon ll

Naming Lunar Mare Basalts: Quo Vadimus Redux G. Ryder ...... 55

Origin and Evolution of the Moon: Apol�o 2000 Model H. H. Schmitt ...... 56

Magmatism of the Lunar Highlands and the Early Paleoproterozoic Magmatism of the Earth: Similarities and Distinctions E. V. Sharkovand 0. A. Bogatikov ...... 58

Mare Magmatism of the Moon and Oceanic Magmatism of the Earth: Similarities and Distinctions E. V. Sharkov and 0. A. Bogatikov ...... 59

MareBasaltic Magmatism: A View fromthe Sample Suite With and Without a Remote Sensing Prospective. C. K. Shearer, J. J. Papike, and L.R. Gaddis ...... 59 Structure and Composition of the Lunar Crust P. D. Spudis, D. B. J. Bussey, and B.R. Hawke ...... 61

The Mineralogy of the Youngest Lunar Basalts M. I. Staid and C. M. Pieters ...... 62 The Lunar Atmosphere and Its Intimate Connection to the Lunar Surface: A Review S. A. Stem ...... 64

Apollo 17 Soil Characterization for Reflectance Spectroscopy L. A. Taylor, C. Pieters, A. Patchen, R. V. Morris, L. P. Keller, S. Wentworth, and D. McKay ...... 64

Global Geochemical Variationon the Lunar Surface: A Three Element Approach D.R. Thomsen, D. J. Lawrence, D. T Vaniman, WC. Feldman, R. C. Elphic, B. L. Barraclough, S. Maurice, P. G. Lucey, and A. B. Binder ...... 66

Siderophile Element Systematics and the Moon's Core-Mantle DifferentiationHistory P.H. Warren ...... :...... 68 The Transport of Magma from the Mare Source to the Surface M. A. Wieczorek andR. J. Phillips ...... 69

A View of the Lunar Interior Through Lunar Laser Range Analysis J. G. Williams, D. H. Boggs, J. T.Ratcliff, C. F. Yoder, and J. 0. Dickey ...... 71 A Study of an Unmanned Lunar Mission for the Assay of Volatile Gases fromthe Soil L. J. Wittenberg, I. N Sviatoslavsky, G. L. Kulcinski, and E. A. Mogahed ...... 72 LP/ Contribution No. 980 1

Abstracts

PROSPECTING FOR LUNAR RESOURCES WITH to map Fe abundances across nealy the entire luna surface. These GLOBAL GEOCHEMICAL AND MUL TISPECTRAL data ca supprt identifcation of Fe-rich regions as smal as a fw DAT A. C. C. Allen, Lockheed Martin Space Operations hundred meters across. Clark ad McFadden [8] ad Elphic et al. [9] Corporation, 2400 NASA Road 1, Houston TX 77058, USA fnd good agreement between y-ray/neuton and multispctal Fe ([email protected]). determinations fr most aea on the Moon. 02 - Correlated with Volcanic Glass Beads: Experi­ Introduction: Laboratory expriments have demonstated cor­ ments. The H-reduction experiments cited above [2] also showed relations between the abundances of some luna resources ad spe­ submillimeter volcanic glas beads could be highly desirable fed­ cifc chemical ad mineraogical paaeters of surface materias. stocks fr luna O production. Iron-rich species, represented by Global remote sensing fom the Apollo, Gaileo, Clementne, and glassy (orage) ad crystalline (black) beads, promise particulaly Luna Prospector missions, combined with Eth-baed observa­ high O yields. Apllo 17 volcaic glass sample 74220, composed tions, has now quatifed these parameters. Combining experiments predominatly of orage glas beads with an average dameter of 40 ad remote sensing allows the frst prospcting fr resources across mm, contains 17.8 wt% Fe2•. Reducton of this saple yielded 4.3 the entire luna surface. wt% 0, well above the regression line defned by the expriments on The Moon's rocks ad soil contain resources that could be used 16 luna soils. Sample 74001 is dominated by black crystaline beads, to supprt a future luna base fr reseach or fr launching expdi­ the isochemical equivalent of orage glases. Reduction of 74001 tions into deep space. Luna O is the resource most likely to be yielded 4.7 wt% 0, the highest value fr any lua sample. exploited frst, fr both propulsion ad lif supprt [1]. The ef­ Global Remote Sensing. Extensive aeas of the lunar surface ciency of O extaction fom lunar materas ha been shown to depend covered by volcanic glas beads have been delineated using Eh­ stongly on the materal's chemical and mineraogica compsiton. baed data ad spacecraft orbital photography [IO]. Chemical com­ Some luna materials may be rae and vital enough on Earth to psitions of the deposits have been estmated ad mixing ratos of justif the cost of their ret. The rare isotop JHe, which ca be glassy and crystalline glass beads have been determined [11]. used in a efcient and low-plluting fsion pwer reaction, has Clementine multispcta imagery ha been employed to determine been cited as one such resource [2]. Sola-wind helium, including the precise extent ad estmate the thickness of one widespread 3He, is implated in lunar regolith grans ad ca be released by deposit, that of the Aristachus Plateau [12]. heatng. Helium concentaton has been shown to corelate with soil Ice - Correlated with Hydrogen in Polar Cold Traps: composition. Theory. Luna ice could prove to be a hghly advatageous O source, 02 - Correlated with Iron in Lunar Soil: . Experiments. compaed to O derived fom soil reduction, in ters of process Over 20 digerent processes have been propsed fr O production complexity ad power requirements. Thus, a deposit of ice on the on the Moon [3]. Among the simplest ad bst studied of these Moon is a ptentially imprtant resource. processes is the reduction of oxides in luna minerals ad glass using Permanently shadowed pla craters have been modeled a pos­ H gas. Allen et al. [4] reprted the results ofO exttion experments sible cold taps fr water ice, derved either fom indigenous sources on 16 luna soils. Each saple was reacted in fowing H fr 3 hat or fom comets [ 13]. Any luna ice deposits must occur in extemely 1050°C. Total O yield corelated stongly to each saple's initial restcted areas nea the ples. Only crater interiors that ae pra­ Fe2• abundace. A linea leat-squaes ft of O yield vs. Fe2+ fr 16 nently shadowed fom sunlight ae cold enough to have retained lunar soils yielded a regression line with a slope of 0.19, a intercept volatles fr a signifcat part of the Moon's history. of 0.55 wt% 0 ad an r2 vah,1e of 0.87. Oxygen yield did not Global Remote Sensing. Clementine images ad Eah-based signifcatly corelate with the abundace of any element except Fe. rada have demonstated that craters whose interiors ae never ex­ Thus, 0 yield fom luna soils ca be predicted with nealy 90% posed to sunlight do exist near the luna ples [14, 15]. Water ice ha confdence based solely on their iron abundances. Te potential fr been tentatively identifed in some of these craters using Clementne 0 production at ay locaton on the Moon ca be predicted if the bi-static rada data [ 14]. soil's Fe concentation is known. Luna Prospector mappd the epitheral and fast-neuton fuxes Global Remote Sensing. On a global scale, Fe concentation in across the entre Moon [16]. Low epithermal fluxes are corelaed the nea surface has been estmated fom data reted by a variety with concentations of H, and by extension ice, in the near-surface. of spacecraft. Iron wa one of several elements meaured fom nea­ Such signatures were obsered nea peranently-shadowed craters equatorial orbits during the Apllo 15 and 16 missions, using y-ray at both luna poles. Calculation of the actual amount of water ice in spectomet [5]. These data covered approximately 20% of the luna these cold taps is stongly model-depndent but each pla region surface, with spatial resolutions of -100 km. An improved could contan as much as 3 x 109 t of water ice. y-ray spectrometer ad a neuton spctometer, fown on the Luna 3He - Correlated with Titanium in Lunar Soil: Prospector spacecrat in a pola orbit, provided Fe abundance data Experiments. The concentation of sola wind-implated He in mae fr the entre lunar surface, agan at a spatial resoluton of -100 km regolith increases with the soil's Ti content [17]. This corelaton is [6]. apparently due to prefrential asorption of He by ilmenite gains. A technique fr iron asessment baed on orbital multispect Experiments simulating interaction of the sola wind with terestial imaging ha been developd by Lucey et al. [7]. This method core­ ilmenite resulted in a releae cure very simla to the releae cure lates Fe abundace to a pameter derived fom reflectace values at fom the Apllo 11 regolith saples [18]. Thus, Ti in mae soils is 750 ad 90 r. Te authors use data fom the Clementine spacecrat a predictor of the 3He resource [19]. 2 Workhop on New Views of the Moon l

Global Remote Sensing. Data from the Lunar Prospector y-ray the U.S. Geological Survey Flagstaff Field Station in Arizona. Two andneutron spectrometers have been used to determine regolith Ti separatecamera systems are attached to a single telescope mount and abundances across the entire Moon at a spatial resolution of -100 km boresighted to the same pointing direction. The visible/nearinfrared [ 6]. Clementine multispectraldata have also been employed to quan­ (VNIR) camera uses a 512 x 512 pixel CCD and 23 intermediate- tifyTi at a resolution of a few hundred meters [7]. Good correlation width interference filters forwavelength selection. The shortwave has been determined between Ti abundancesdetermined by the two infrared(SWIR) camera uses a 256 x 256 pixel cooled-HgCdTe methods [9]. infraredarray and nine intermediate-width interference filters.Table 1 Challenges Remaining: The fullintegration of laboratoryex­ andFig. 1 provide information on the instrumental passbands.Sepa- periments has not been completed, andchallenges remain forlunar rate 20-cm-diameter Ritchey-Cretien telescopes areprovided for the resource prospecting, including: (1) integrating the higher precision two cameras. The optics are designed to image the entire Moon ofy-ray chemical analysiswith the higher spatial resolution of mul­ within each camera's field of view, resulting in instrument pixel tispectral chemical analysis; (2) combining chemical and spectral scales of 4 and 8 arcsecpixeJ- 1 (-7.4 and 15 kmpixeJ- 1 for the sub- datasets to locate additional deposits volcanic glass beads, corre­ Earth point on the Moon) for VNIR and SWIR respectively.Detailed sponding to the full range of compositions found in lunar soils; information on the instrumentation can be found in Anderson et al. (3) demonstrating the relationship of enhancedH concentrationsto [3]. ice concentrationsin permanentlyshadowed craters; and ( 4) refining the correlationof solarwind gasconcentrations to soil chemistryand mineralogy. TABLE 1. ROLO filterbandpasses. References: [1] Joosten B. K. andGuerra L. A. (1993) AIAA Space Prog. Tech. Conj Paper 93-4784, Am. Inst. Aeronautics A. M T QE Band Astronautics. [2] Wittenberg L. J. et al. (1986) Fusion Tech., JO, (nm) (nm) (%) (%) 167-178. [3] Taylor L. A. andCarrier W. D. III (1992) in Engineer­ ig, Construction, and Operations in Space I, pp. 752-762, Am. VNIR Soc. Civ. Eng. [4] Allen C. C. et al. (1996) JGR, 101, 26085-26095. 347.3 32.5 62 15 u [5] Davis P. A. Jr. (1980) JGR, 85, 3209-3224. [6] FeldmanW. C. 352.5 31.6 52 15 u et al. (1996) LPS XVII, 355-356. [7] Lucey P. G. et al. (1995) 405.0 16.2 54 15 V Science, 268, 1150-1153. [8]ClarkP. E. andMcFaddenL.A. (1996) 412.7 12.5 52 15 C LPS XVII, 227-228. [9] Elphic R. C. et al. (1996) Science, 281, 415.1 17.8 56 15 V 1493-1496. [10] Gaddis L. R. et al. (1985) Icarus, 61, 461-489. 441.8 9.6 61 15 C [11] Weitz C.M.etal.(1998)JGR, 103, 22725-22759. [12]McEwen 466.5 20.0 86 16 b 475.7 18.4 85 16 A. S. et al. (1994) Science, 266, 1959-1962. [13] ArnoldJ. (1979) b 488.1 7.9 81 17 C JGR, 5659-5668. [14] Nozette S. et al. (1996) 84, Science, 274, 545.0 18.8 91 25 y,V 1495-1498. [15] MargotJ.L. eta!.(1999) Science, 284, 1658-1660. 550.3 8.7 72 25 C [16] Feldman W. C. et al. (1998) Science, 281, 1489-1493. 554.9 18.1 92 25 y,V [17] Cameron E. N. (1988) in Te Second Conference on Lunar 666.7 8.3 87 35 C Bases and Space Activities of the 21st Centur, pp. 89-97, LPL 694.8 16.8 84 36 R [18] Harris-KuhlmanK. R. and Kulcinski G. L. (1998) in Engineer­ 705.5 16.7 87 36 R ing, Construction, and Operations in Space VI, pp. 533-540, Am. 747.1 8.7 72 32 C Soc. Civ. Eng. [19] Taylor L. A. (1990) in Engineering, Construc­ 765.6 16.8 95 28 tion, and Operations in Space II, pp. 68-77, Am. Soc. Civ. Eng. 776.5 16.9 98 28 867.7 13.9 78 16 C 875.3 18.4 92 15 I 885.2 16.0 94 15 I PHOTOMETRIC IMAGING OF THE MOON FROM THE 934.6 17.6 94 7 ROBOTIC LUNAR OBSERVATORY. J. M. Anderson and 944.7 18.8 94 6 H. H. Kieffer, U.S.Geological Survey, Flagstaff AZ 86001,USA. SWIR Introduction: As partof the calibration programfor the NASA 944.3 21.5 65 67 Earth Observing System(part of NASA's Earth Science Enterprise) 1062.2 27.1 52 66 J [l], the U.S. Geological Surveyoperates the Robotic Lunar Obser­ 1246.5 23.3 62 66 J vatory(ROLO). The ROLO project is designed to produce a photo­ 1542.7 48.6 56 67 H metricmodel of the nearsidelunar surface for all phase andlibration 1637.8 23.4 66 67 H angles visible from Flagstaff[2]. Goals for this photometricmodel 1984.8 38.5 54 70 are 2.5% absolute and 1.0% relative uncertainty. Although the model 2131.8 54.7 64 65 K 2256.3 48.2 54 63 is principally intended to produce radiance images of theMoon for K 2390.3 58.2 48 60 use in calibration of Earth-orbiting spacecraft, the ROLO data and K model will also provide importantinformation for studies of thelunar A-effective wavelength forflat spectrum; �I-half-power bandwidth;T-fil­ soil. ter transmissionfor equivalent rectangularbandpass; QE-detector quan­ Instrumentation: An astronomical observatory dedicated to tum efficiency; Band-corresponding standardmagnitude band, if the radiometryof the Moon hasbeen constructedon the campusof applicable, "c" indicates cross-calibrationradio meter band. LP/ Conrribution No. 980 3

400 800 1000 C 1 f=,-:--'>:�--.=c�

�0.8 ...:... ·§0.8 lelf--�-=- - .,0.4 ... f 0.2 >" � o ...... ,_....._.....,_.....__._""-" ..__...... =-o== .," 1000 1500 2000 2500 .ll Wavelength (nm) 0 )," -5 Fig. 1. Robotic Lunar Observatory Filter Response. Te tansmission "' cures fr VNIR (a) and SWIR (b) flters (solid lines) are plotted a a 1996 1999 2000 2001 gnction of wavelength. Also shown are a normalized solar spectum (dashed), nominal atospheric tansmission (dotted), and the quantum -5 0 5 efciency' of the ROLO detectors (dot-dash). (Ts plot is simila to Fig. Sub-Observer Lngitude (degrees} 4 in Kiefer and Anderson [5].) Fig. 2. Robouc Lunar Obseratory Lunar Obseratons. Obserauons of the Mon made with the SWI camera through 1999 June ae shown as thick bars in te phae/librauon space available to ROLO. Phase angle Observations: Routine imaging has been in progress since late is indicated by the slopes of the bars - horizontl ba indicate 0° phase 1995 fr VNIR bd late 1997 fr SWI, bd is expected to continue and vertical bars indicate 90° phase. Positive slopes correspond to through at leat 2002. ROLO obseses the Moon every clear night obseratpns afer Full Moon. Te thin lines connect topoentc librauon between the frst bd last quaer phases of the moon. On such nights, angles fr Aagstaf at midnight fr consecuuve night between First and the Mon is imaged through all 32 flters at half-hour intesals during Last Quarter Mons fom 1998 through 202. Several yeas of conunued the time that the Moon is above the 60° zenith bgle. Obsesatons observauons are necessary to fll in the empty regions of the phase/ libraton space. of stbdard stas to meaure atmospheric extnction bd detector respnsivity drifts occupy the remainder of the nighttime. Measure­ ments of the dak curent and detector bias levels ae made during the metic models of the luna surface that ae psition, phae, and dusk bd dawn periods fr VNI and throughout the night fr SWIR. wavelength dependent. Kiefer and Anderson [5] bd Grbt et al. [6] Flat feld correctons bd absolute radibce calibrations ae provided provide exaples of calibrating spacecra instents with existing through obsesations of a Spctalon plate illuminated by a NIST­ ROLO data. The dataset will also provide a wealth of informbion taceable 100 W FEL lap. Raw data ae converted to ISIS [4] aout the geological propertes of the luna surface. The photometc cubes ad stored on CD-ROM. Detailed infration on the obses­ fnction of many thousbds of points on the Moon will be measured ing procedure is also found in Anderson et al. [3]. fr phae bgles between 2° bd 90° fr wavelengths within 0.35-2.4 Data: As the development of data-reduction softwae fr the µm. Absolute bd noralized color ratios bd their phase-angle ROLO project progresses bd additonal data ae accumulated, the depndence will enale detaled studies of the mneral abundances of raw data ae repeatedly processed into a calibrated fr.Corections the Luna soil. Infrmation on the photometic fncton depndence fr instument respnse chaacteristcs, photon scatterng processes, on phase and wavelength will aid investigations of the surfbe pa­ bd atmospheric extinction ae applied to the raw lunar images to ticulate stucture of the soil. produce exoatmospheric radiance images of the Moon. These images References: [1]Butler J.J. bdJohnsonB. C. (1996) Te Earth ae then transfrmed to a fxed selenographic-grid projecton de­ Obserer, 8 (1 ). [2] Kiefer H. H. bd Wildey R. L. (1996) J. Atmos. signed to accommodate all of the pssible viewing geometes of the Ocean. Technol., 13, 360. [3] Anderson J.M. et al. (1999) Pub. Ast. ROLO telescop [2]. A preliminay discussion of results fr the total Soc. Pacic, 111, 737. [4] Gaddis L. et al. (1997) LPS X1, 387. iradibce of the Moon derved fom ROLO VNIR images acquired [5] KieferH. H. bd AndersonJ. M. (1998)Proc. SPlE, 3498, 325. through Aprl 1998 was published by Kiefer bd Anderson [5]. At [6] Grbt I. F. et al. (1998) Proc. SPIE, 3498, 337. that time, diffcultes in adequately determining the atospheric extinction limited the bcuacy of the derived luna iradibce values. Signifcbt improvements in the reducton softwae have been devel­ TH LUNAR CRUSTAL TICKNESS FROM ANALYSIS oped since that time bd meaurement scatter is expected to be OF THE LUNAR PROSPECTOR GRAVITY AND reduced to s2% fr the data-processing rn planed fr the summer CLEMNTINE TOPOGRAPHY DATASETS. S. Asmar 1 , of 1999. G. Schubert1 , A. Konopliv2, bd W. Moore1, 1Depament of Eh As of June 1999 , ROLO has acquired over 2200 cubes of raw bd Space Sciences, Univeristy of Califra, Los Angeles, Los Moon images with VNI bd over 1200 cubes with SWI. Figue 2 Angeles CA 7095, USA, 2Jet Propulsion Laboratory, Califra shows the phae/libraton coverage fr VNI bd SWI together Institute of Technology, USA. through June 1999.By 2002, ROLO expcts to have acquired roughly 3500 images of the Moon through each VNIR filter bd nearly 300 The Luna Prospctor spacecraft has mapped the gravity feld of images through each SWI flter, or more thb 10,00 absolutely the Moon to a level of resolution never achieved befre, bd a calibrated images of the Moon. These data ae used to create photo- sphercal haonic representation to degree bd order 10 is aval- 4 Workshop on New Views of the Moon 11

able. When combined with the topography datasetproduced by the It is possible that Apollo 12 and15 sites have the thickest regolith Clementine mission, the resulting Bouguer anomaly map is inter­ andthe Apollo 16 site has the thinnest. It is alsopossible that Apollo preted to model the thickness of the lunar crust. Such models are 12 and 15 basalts arecomminuted faster than Apollo 16 highland crucial to understandingthe lunar thermalhistory and the formation rocks andApollo 14 and17 soils are products of mixed parentage. of geological featuressuch as mascon basins, several more of which If a soil becomes continually finer as it matures until agglutination have been newly discovered fromthis dataset. A two-layer planetary catches up, and if comminution is differential-dependent on the model was used to compute the variations of the depth to the lunar physical properties of the constituents, then the composition of the Moho. The thickness values ranged fromnear Ot o 120 km. There is bulk soil has to match the composition of some "fulcrum" grain size significant agreement with previous work using the Clementine fraction,say X. Grain size fractions>X and

raise questions ranging fromthe unquestionable use of IsfFeO as the Ti, normalizedto the chrondritic ratiofor these elements.] Later work universal maturity parameterof lunarsoils to global elemental maps [2] applied a ternarydiagram approach to look at global lunarpetro­ of the Moon fromremote-sensing data [e.g., 12,15]. logic variations. This work used the Fe-(Thffi)0 technique as this had Regolith Differentiation: Global distribution of rock types the most spatial coverage with the available data andalso appeared and global compositional variations on the Moon are interpreted to be adequate at distinguishing between different rock types. In the fromremo tely sensed signals (UV-VIS-IR; y-ray; secondary XRF). ternary diagram, the apexes were assigned the average Fe and(Th/ These signals integrate differentdepths of the regolith. Hence, con­ Ti)c values of ferroananorthosite, mare basalt, and KREEP rocks. vergence of raw data and first-order interpretation cannot be ex­ Each apex was assigned a primary color while the center of the pected without calibration with actual samples.The very surface of triangle was represented by gray. Each point on the lunar surface, the Moonis covered with very finedust [ 16] that contributessignifi­ covered by the Apollo geochemical instruments, wasthen assigned cantly to all remote-sensing signals. Size fractions of the lunar a color depending on where in the ternarytheir composition placed regolith differ in chemical and mineralogic compositions. Many them. The resultant petrologic classification map shows how the processes contributeto regolith differentiation including the loss of petrologicunits varyspatially. The main results fromthis work were volatile elements such as Si [17]. In addition, the distributionof np­ asfollows: (1) The highlandscontain largeareas of relatively pure Fe0 and largergrains ofFe0 affects IRspectra [18]. It appearsthat the ferroan anorthosite; (2) KREEP/Mg suite rocks represent a small credibility of global maps from recent remote-sensing data (para­ percentage of the upper lunarcrust; (3) farside outcrops of KREEP/ digm?) will increase if calibrations are carried out on differentgrain Mg suite rocks are associated with areasof crustalthinning, particu­ size fractionsof the lunarregolith [e.g., 19]. larly on the floorof South Pole Aitken Basin; (4) the average com­ Conclusion: Much of the global understanding of the Moon position of the highlands is richer in Fe than ferroan anorthosite, comes from remotely sensed and sample analyses of the regolith. which supports the magma oceanhypothesis of crystal formation; Recent data, fromTEM's nanometer-scale to kilometer-scale foot­ and(5) regions of the eastern limband farsidehighlands arerelatively prints of orbital and Earth-based observation, call for a review of more maficthan average highlands.These areashave a high density paradigmsin lunarregolith science. of darkhalo craters, supporting the idea that mare volcanism oc­ Acknowledgments: This research was supported by NASA curred in this region before the end of the heavy bombardment. GrantNAG5-4018. Global Coverage: This earlierwork utilizedthe-Apollo gamma References: [l] McKayD.S.et al. (1974) Proc. LSC5th, 887- andX-ray orbital datasets. These data provided limited coverage of 906. [2] MendellW.W. andMcKay D. S. (1975) Te Moon, 13, 285- the lunarsurface (mostly confinedto the equatorial latitudes). The 292. [3] Basu A. (1977) Proc. LSC 8th, 3617-3632. [4] McKay y-ray instrumentcovered approximately 19% of the lunar surface D.S. and BasuA. (1983)Proc. LPSC 14th, Bl93-B199. [5] Heiken while the X-ray only covered 9%. With the Clementine and Lunar G. et al., eds. (1991) Lunar Sourcebook, 736 pp. [6] Pieters C. M. Prospector datasets, we now have global maps of Fe, Ti, andTh et al. (1993) JGR, 98, 20817-20824. [7] LindstromD. J. (1999) LPS [3-6]. Apart fromglobal coverage, anotherimportant advantageof X, Abstract#2014. [8] Housley R. M. et al. (1973) Proc. LSC 4th, the new datasets is higher spatial resolution. The resolution of the 2737-2769. [9] Taylor L.A. and Cirlin E-H. (1985) in ESR Dating Apollo instrumentswas 15 km forthe X-ray and lOOkmfor they-ray and Dosimetr, pp. 19-29, Ionics. [10] Morris R. V. (1980) Proc. [l]. The Fe and Ti maps arederived fromthe full-resolutionClementine LPSC 11th, 1697-1712. {11] Keller L. P. and McKay D.S. (1997) UV-VIS data, i.e., -250 m/pixel. The resolution of the Th data, GCA, 61, 1-11. [12] Lucey P. G. et a!. (1995) Scienc�, 268, 1150- obtained by Lunar Prospector's neutron spectrometer, is currently 1153. [13] Hapke B. (1998) Workhop on the New Views on the -150 km [5], but will be available in the future with a spatial Moon, 34-35, LPI, Houston. [14] Christofferson R. et al. (1996) resolution of 60 km [7]. The other improvement provided by the MAPS, 31, 835-848. [15] Blewett D. T. et al. (1997) JGR, 102, recent lunarmissions is the errorassociated with the data. The errors 16319-16325. [16] Horz F. et al. (1972)Apollo 16 Prelim. Sci. Rpt., associated with the Fe, Ti, andTh values obtained by Apollo were pp. 7-24-7-54. NASA, Washington, DC. [17] Naney M. T. et al. 10-25 wt%. The errorof the Clementine-derivedFe andTi values is (1976) Proc. LSC 7th, 155-184. [18] Keller L. P. et al. (1998) -1%, while the Th data have anerror of -1 ppm. Workhop on the New Views on the Moon, 44-45, LPI, Houston. We intend to investigate the petrologicvariations on the Moon at [19] Taylor L. A. et al. (1999) LPS X, Abstract #1885. a global scale using the new Clementine and Lunar Prospector elemental maps for Fe, Ti, and Th. We shall use the technique described in Davis andSpudis [2]. An initial study has been under­ LUNAR GLOBAL PETROLOGIC VARIATIONS. D. B. J. taken that looks at some regions tha� were covered by the Apollo Bussey 1, P. D. Spudis2, and J. J. Gillis3, 1Code SCI-SO, European geochemistry data.Two mareregions, one in Imbrium andthe other Sapce Agency, EuropeanSpace Research andTechnology Centre, in Procellarum, match well with the results using the Apollo data. 2200 AG Noordwijk, The Netherlands, ([email protected]), The highlandterrain appears problematic. The calibration of the Th 2LunarandPlanetaryInstitute, Houston TX 77058, USA, 3Department data [6] is based on the assumptionof a constantbackground. This of Earth andPlanetary Sciences,Washington University, St. Louis is a valid assumption where Th counts are well above background MO 63130, USA. limits, but as count rates decreasevariations in Th concentrationare more sensitive to background fluctuations.Eventually we will cir­ Introduction: An initialattempt at producing petrologicprov­ cumvent this problem by using the lower-altitude (i.e., higher reso­ ince maps of the lunar highlands combined orbital and sample lution) Prospectordata anda calibration derived fromdeconvolution geochemical data in variationdiagrams [ 1]. Threedifferent variation of the y-ray spectrawith properattention to background variations.

diagramswere produced:Mg* (= 100 Mg/Mg+Fe) vs. [(Th/Ti)0, Al The Th/Ti vs. Fe technique provides more geologic information vs. Mg*/(Thffi)0, andFe vs. (Th/Ti)0• ([Thffi]0 is the ratio of Th to thancan be extractedfrom solely using the elemental data because it 6 Workhop on New Vews of the Moon I

quantitatively determines each pixel's composition in terms of the ment of the validity of these coefficients forevery class of absolute end-member compositions: ferroan anorthosite, mare basalt, and abundance, the LP data have been transformedinto absolute abun­ KREEP rocks. Moreover, the digital map illustrates the extent and dances forthe whole Moon. lateral transition of one rock-type to another, thus allowing the The Th LP data have been converted into abundances(ppm) using sample data to be placed in a regional and global context. Th concentrationsin average soils from the Apollo and Luna sites Initial examination confirmsearlier conclusions [2] that the Mg given in [5]. Values of Th abundances for these samples range suite is not a major contributorto at leastthe upper lunar crust.We between 0.5 and 13 ppm [3,5]. A nonlinear regression has been summarize elsewhere in this volume evidence that Mg-suite rocks applied to the LP data, except forlow count rate values ( correspond­ may be a significantcontributor to the composition of the lower crust ing to Th abundancesbelow l.27 ppm) forwhich a first-orderlinear [8]. Work continues on the problem of the crustal fractionof the Mg regression has been applied. suite, needed to determine the Mg# of the bulk crust, an important Statistical tests demonstratethat the LunarProspector FeO, Ti0 2, constraint on lunarorigin [9]. andTh abundancesestimates we have produced through the regres­ References: (l]DavisP. A. andSpudis P. D. (1985)LPSXXVI. sions are reliable. For the LP maps, the uncertainty of absolute FeO [2] Davis P.A. andSpudis P. D. (1987) LPSXXVII. [3] Lucey P. G. andTi0 2 estimates is on the sameorder as that for CLM, i.e., 2 and et a!. (1998)JGR, 103, 3679. [4) BlewettD. T. eta!. (1997)JGR, 102, 1.5 wt% respectively. For the Th, abundanceestimates, the uncer­ 16319. [5] Lawrence D. J. et al. (1998) Science, 281, 1405-1560. tainty is -1-2 ppm. [6) Gillis et al. (l 999) LPS XX, Abstract #1699. [7] Lawrence et al. Results: At first order, our global FeO and Ti02 abundances (1999) LPS XX, Abstract #2024. [8) Spudis P. D. et al., this volume. maps fromLP are in very good agreement with those obtained from [9] Drake M. J. (1986) Origin of the Moon, LPI, 105 pp. CLM [2]. The Th abundance map is also in good agreement with a recent global Th abundance map by [6], although concentrations have a higher range (0-20 ppm) in our case, and may be slightly overestimated. INTEGRATION OF THE ULTRAVIOLET-VISIBLE However, a more detailed comparison between CLM and LP SPECTRAL CLEMENTINEDATA AND THEGAMMA-RAY abundance maps of FeO and Ti02 reveals regional differences. LUNAR PROSPECTOR DATA: PRELIMINARY RE­ Differences between the two datasets are expected because of instru­ SULTS CONCERNINGFeO, TiO2, AND Th ABUNDANCES mental andobservational causes (spatial resolution, depth of obser­ OF THELUNAR SURFACE AT GLOBAL SCALE. S. D. vation), but also because they are sensitive to different types of ChevreJI, P. C. Pinet1, G. Barreaul, Y. Daydou1, G. Richard1, S. information (mineralogy vs. chemistry).Maps of differencesin abun­ Maurice1, and W. C. Feldman2, 1Unite Mixte de Recherche 5562, dancesbetween LP and CLM data have been produced for both FeO Group forSpace Researchand Geodesy, ObservatoireMidi- Pyrenees, andTi0 2. In relation to the precision of the LP and CLM data, we Toulouse, France ([email protected]),2 Mail Stop D-466, Los consider that differencesbetween the two datasetsexceeding 2 wt% AlamosNational Laboratory, Los AlamosNM 87545, USA. are significant. Concerning FeO contents, differencesup to 4-5 wt% are ob­ The Clementine mission (CLM) produced global multispectral servedin the South Pole Aitken Basin, andgenerally at high latitude ° data that resulted in a map of FeO and Ti02 concentrations of the (up to 70 ) in the southern hemisphere, as well as in the northern lunar surface [1,2]. The recent Lunar Prospector (LP) mission re­ border (Irid�-Gruithuisen domes region) and the southern border turnedthe frrstglobal data forthe distributionof surface abundances (Sinus Aestum) of the Imbrium Basin. Most important differences, of key elements in lunarrocks, using a y-ray spectrometer(GRS) and up to 6-7 wt%, are found in Mare Serenitatis, Crisium, and neutron spectrometer (NS) [3,4]. Integrating CLM mineralogical Fecunditatis. spectralreflectance and LP chemical data is importantto enhance our Concerning Ti02, most important differences (4-5 wt%) be­ view oflunar crustorigin and evolution, lunarvolcanism, andsurface tween LP and CLM abundances are located within Mare processes. Tranquillitatis.Differences (about 2-3 wt%) also occur in the north­ Data Reduction (P): Iron, Ti, andTh having relatively large ernborder of the Imbrium Basin (lr-Gruithuisen domes region) and compositional variation over the lunar surface, as well as strong in the western part of Oceanus Procellarum (from southwest of isolated peaksin the GRS spectra, informationconcerning the distri­ Kepler, up to the Aristarchus Plateau). bution andconcentration of these elements has been derived from Regions forwhich differences in FeO and/orTi0 2 content may maps of corrected( cosmic ray, non symmetricresponse of the instru­ arise fromthe presence of heterogeneous terrains due to geological ment) counting rates only, without converting them into absolute transitions at the scale of observationof LP have been identified on abundances [3). Maps produced contain count rates in equal-area the basis of the spatially degraded FeO andTi0 2 CLM maps. These projection averaged into 5 x 5° latitude/longitude bins, from-90 ° to regions correspondto Mare Crisium, the westernpart of Procellarum, +90° latitude and-l 80° to + 180° longitude. and MareHumorum. In this work, we have used the CLM global FeO and Ti02 The surface heterogeneity effectcannot be invoked in the case of abundances (wt%) maps [2] converted at the LP spatial resolution the South Pole Aitken (SPA) Basin interior andits immediate vicin­ (-150 km/pixel) to produce FeO and Ti02 GRS abundancemaps, ity. However, LP indicates a lower content in FeO thanCLM (8-11 through a linear regression based on the analysis of the scatter wt% vs. l 1-13 wt% respectively).This observation suggests that the distributionof both datasets. The regression coefficientshave been SPA materials may contain a lower proportionof Fe than was previ­ determinedfrom the data taken between -60° and+60 ° latitude to ously thought. This may imply a smaller proportionof mantlemate­ avoid uncertainties in the CLM spectral data due to nonnominal rials excavated during the impact event than predicted in [7] andgive conditions of observation at high latitudes. After a critical assess- some constraintsto the models of largebasin formation,[e.g., 7-9). LP! Contribution No. 980 1

Some regions such as Mare Serenitatis, Tranquillitatis, basalt petogenesis. One of the ealiest Ti maps [3] showed Ti Fecunditatis, and Crisium do present signifcat digerences in FeO variations in neaside maria on the bais of groundbased spct ad TiO2 abundaces, the LP estimates being lower tha the CLM reflectance measurements. A map of Ti derived fom y-ray measure­ ones. This might be ascrbed to a diferent prception of the infuence ments on Apollo 15 and 16 was produced at about the same time [4], of the mae regolith reworking by the LP geochemichal obsesauon, ad wa improved upon considerably by Davis ad coworkers [5], sensitive to the subsurface regolith, ad the CLM refectace spc­ who efectively removed sources of spurious variation fom Fe ad troscopy observation, sensitive to the surface mineralogy. Al or REE (e.g., Th) interference, ad calibrated Ti on the bases of In the case of the Gruithuisen domes/Iridum region, LP maps lading-site soil averages. In recent yeas, spctral refectance mea­ show FeO and TiO2 with higher abundaces (by -30%) tha CLM. surements fom Clementine have been used by Lucey ad coworkers They also display high concentations of Th (46 ppm). In that [6] to produce global Ti distribution maps a well. As we indicated particular cae, the nonagreement between LP and CLM abundances previously, the Lucey and Davis maps agree to first order. Mea­ could be linked to particula typs of materials. More precisely, it while, we are using the concept of sensor data fusion to combine may indicate that Fe and Ti ae not present in the mbor mineral phae measurements fom the AGR (Apllo y-ray) and CSR (Clementine only, i.e., silicates ad oxides respctively (obesed by CLM), but ae Spectral Refectace) techniques with ground tuth fom luna soils also present in other minerals ad/or amorphous phases . The to uulize the diferences between the two maps to understand the Gruithuisen/ Iridium region coresponds to Imbrum ejecta materials distibution of Ti within luna soil components, as we have done with mapped as Fra Mauro deposits [10). Sbples reted fom the Fra Fe [7]. This technique should be verifed ad applied on Luna Mauro fration (Apollo 14) ae impact melts that have undergone Prospector y-ray meaurements of Ti [8], as the calibrated data multiple episodes ofbrecciation. They display a bulk composition of become available within the next couple of yeas. KREEPbasalts adclasts ofbrecciated feldspathic rocks [l O]. KREEP Luna Ti is fund principally in the mineral ilmenite, and is basalts have Th concentauons raging between 3 and 10 ppm [5] asociated with certan compnents of luna soil [9]: crystalline which is in good agreement with the LP obseration. The Gruithuisen ilmenite mneral fagments ad high Ti-bearing glas. All data indi­ region also corespnds to possible extended nonmae volcaism cate that Ti is associated with maia and mafc mineras. In AGR ad units having spectal characteristics close to those of Red Spts units CSR datasets, Ti is hghest on the neaside ad in the maria, parcu­ (i.e., a stong absortion in the UV domain) associated with volcaic laly in southes Serenitatis/norther Traquillitatis ad northwest­ domes [11 ]. At the present time, the tue nature of the Gruithuisen er Procellam. Unlike CSR-derived values, AGR Ti values show dome materials is still unknown. From morphologic evidence, these modest increae (up to 0.7%) on the norther farside. Tale 1 shows materials may result fom silicic or flspathic-rich KREEP basalts the relationship between average Ti values fr typical regions [10). [12). The diffse presence of such a material in the region may Both techniques show a primarily unimodal distibution with a therefre also contribute to te high concentation of Th obsered by shoulder in the high Ti direction and a primary mode at approxi­ LP. Higher concentations of Fe and Ti obsered by LP may result mately 0.2% (0.15% frCSR ad0.25% fr AGR data), buttheAGR fom their presence in the highy brecciated melt materias and in the Ti data have more stucture in the shoulder, including a appaent nonmae volcaic materials (probaly in a amorphous phae). minor mode at 1.2% Ti, representing nonhighad aea. Recalibrang References : [ 1] Malaet E. ad P. G. Lucey (1996) LPS XVI, the two datasets on the basis of matchng the paks and rage on the 797-798. [2] Lucey P. G. et al. (1998) JGR, 103, 3679-3699. histograms would not account fr this additional stucture in the [3]LawrenceD.J.etal. (1998)Science, 281, 1484-1489. [4] Feldman AGR data. Low Ti areas, which occu predominatly on the faside, W.C. et al. (1998) Science, 281, 1489-1493. [ 5) Korotev R. L. (1998), ae not represented well in the luna sample collection, atough JGR, 103, 1691-1701. [6] GillisJ.J.etal. (1999)LPSX. [7] Lucey several meteorites thought to be of luna highlad origin show Ti P. G. et al. (1998) JGR, 103, 3701-3708. [8] Pieters et al. (1997) abundances averaging-0.2% [11). (Erorbas on AGR Ti values ae GRL 24, 1903-1906. [9] Schultz P.H. (1997) LPS XXVI, 1259- -0.5%.) 1260. [10) Spudis et al. (1988) LS XVI, 155-168. [11) Chevrel et Gaa-ray meaurements refect intinsic Ti surface composi­ al. (1999)JGR, in press. [12)HeadH. W. etal. (1978) LPS IX, 488- tion, regardless of the physical or chemical state of Ti [1]. Lucey and 489. coworkers [6] have attempted to produce an equivalent bulk Ti map

THEDISTRIBUTION OF TITANIUM I LUNAR SOILS TABLE 1. Regional AGR and CSR Ti abundnces. ON THEBASIS OF SENSOR AND IN SITU DATA FUSION. Region AGR CSR P. E. Clak 1,2 and L. Evas2,3, 'Catholic University of America, 2Mail Proellam North 2.9 3.0 Code 691, NASA Goddard Spbe Fight Center, Greenbelt MD Poellarum South 2.4 2.5 3 20771, USA ([email protected]), Computer Science Mae Imbrum Est 1.3 1.8 Corration, 65 Hembree Pakway, Suite A, Atlata GA 30076, Mare Imbrium West 2.4 2.6 USA. Mare Traquillitats 3.0 3.5 Mare Serenitats 2.5 3.0 A variety of remote-sensing meaurements have been used to map Mae Crisium 1.3 2.4 the distibution of elements on the Moon as a meas of providing Mare Smythii 2.6 1.5 constnts on the processes fom which its crst ad major teres Farside Northeast 0.9 0.5 originated [1,2). Discussed here is Ti, whch is incororated into Farside Northwest 1.3 0.4 Farside Southwest 0.7 0.6 refactory mineras such a ilmenite during the latter stages of difer­ Farside Southeast 0.5 0.6 entiation, ad is thus a most usefl element fr understading mae 8 Workhop on New Vews of the Moon I

by nonnalizing the spectral feature at 415 nm to remove the effects DIGITAL ELEVATION MODELS OF THE LUNAR of physical variations from soil to soil. The nonnalizationis based on SURFACE. A. C. Cook1 and M. S. Robinson2, 1Center forEarth laboratory measurements of available samples, primarily nearside and Planetary Studies, National Air andSpace Museum, Washington maria,and is undoubtedly optimized forthe proportions in which Ti­ DC 20560-0315, USA ([email protected]), 2Department of bearing components are found in these soils. In Table 2, AGR and Geological Sciences, Northwestern University, EvanstonIL 60208- CSR Ti abundances (with CSR filteredto match resolution and field 2150, USA ([email protected]). of view of AGR data) for landingsites are comparedto abundances of soil components in which Ti would principally be found: crystal­ Introduction: Several digitalelevation models(DEMs) have line ilmenite fragmentsand Ti-bearing glass. In particular,mare soils been produced at a scale of lkm/pixel and covering approximately tend to have a much higher proportion of opaque mineral grains [ 12], one-fifthof the lunarsurface. These were produced mostly by semi­ which appearsto be well correlatedwith Ti abundance, as shown in automatically matching the stereo available between Clementine Table 2. The relationship between Ti abundance and proportion of UV-VIS images, although some localized DEMs have been pro­ Ti-bearing glass ([12], weighted for Ti abundance according to [9]) duced by applying this technique to Apollo Metric stereo pairs,or by is not asdirect. Table2 also shows the relationshipbetween landing digitizing an existing Apollo Metric contour map. The DEMS that site soil average, AGR, and CSR Ti abundances (with CSR filtered result fromClementine UV -VIS images, although of poorer height to match resolution and field of view of AGR data) . AGR Ti values accuracy (±300-600 m for a single matched point) [I] than the show the best agreement with Ti soil averages, CSR Ti values being Clementine laser altimeter point measurements (<±100 m) [2], do somewhat higher in high-Ti soils, and lower in low-Ti soils. These provide considerably higherspatial resolution (e.g., every kilometer results would tend to support the argument that the technique for vs. every tens of kilometers) and allow topography in the polar deriving Ti fromthe CSR measurementsis optimized forone of the regions to be determined. primaryTi-bearing components,most likely opaque mineral grains, Method: Nadir-pointingClementine UV-VIS stereo pairsare and is thus overcompensating and reporting higher Ti values in mare automatically stereo matched using a patch-based matcher [3] and areas, and underreporting Ti in highland areas with much lower fed through a stereo intersection camera model to yield a digital proportions of opaque mineral grains. terrain model (DTM) of longitude, latitude, and height points. The References: [1] Adler I. andTrombkaJ. (1970) Geochemical DTM foreach stereo pair is then replotted and interpolatedto form Exploration ofthe Moon and Planets. [2] Taylor S. R. (1986) Origin map-projected DEM tiles. The DEM tiles can then be fitted to of the Moon (W. K. Hartmann et al., eds.), pp. 125-144, LPI, absolute heightlaser altimeter points, or iteratively to each other, to Houston. [3] Johnson T. et al. (1977) Proc. LSC 8th, 1029-1036. forma DEM mosaic. Uncertainties in UV-VIS camera pointingand (4] ArnoldJ. et al. (1978) LPS IX, 25-26. [5] Davis P. (1980) JGR, the need to accumulate a sufficiently good topographic SIN ratio 85, 3209-3224. [6] Lucey P. et al. (1998) JGR, 102, 3769-3699. necessitates the use of I km pixels forthe UV-VIS derived DEMs. [7] Clark P. and McFadden L. (1999) JGR, in press. [8] FeldmanW. For Apollo Metric stereo, an internal camerageometry correction C. et al. (1998) Science, 1497-1500. [9] Heiken G. et al. (1991) and a full photogrammetric block adjustment must be performed Lunar Sourcebook (G. H. Heiken et al., eds.), Chapters 2 and 5, using ground- control points to derive a DEM. The image scale of CambridgeUniv.,NewYork. [10] Clark P.andEvans L. (1999)LPS Apollo Metric,as well as the stereo angle,allow fora DEM with 100 XXX. [11] Lindstrom M. et al. (1991) GCA, 55, 2999-3007. m pixels anda height accuracy of ±25 m. Apollo Metricimagery had [12] PapikeJ. et al. (1998) Planetar Materials, Chapter 5. previously been used to derive contour maps formuch of the lunar equatorial regions; however, to recover this information in digital form these maps must be digitized. Regions Covered: Most of the mare areas mapped contain noticeable topographic noise (see Fig. I). This results fromthe stereo TABLE2a. Landing site Ti abundances. matcher failingin regions of low texture and contrast. Below is a summary of the regions forwhich DEMs have been generated, and Site %Ti %Ti %Ti %Ti some of the features visible in each (heights are referencedto a 1738- Soi!Av AGR CSR AGR-CSR km radius sphere): ° ° ° ° Ap 12 1.8 1.6 2.4 +0.8 Apollo 15 landing site ( 1 E-7 E, 20 N-30 N). Features include Ap 14 1.0 0.8 0.6 -0.2 Hadley Rille (justvisible) andseveral mountains [4]. Ap 15 0.95 1.0 0.6 -0.4 Apollo 17 landing site (30.0°E-3l.5°E,19.5°N-2l.0°N). This Ap 16 0.3 0.35 0.25 -0.2 DEM was derived by digitizing contours from a map [5]. Ap 17 2.5 3.6 4.2 +0.6 Alpine Valley (2°W-8°E, 47°N-51 °N). The valley andTrouvelot crater arevisible [6]. Topography rangesfrom+ 1 km to -4 km. The TABLE2b. Landing site Ti-bearingcomponents. valley walls are

Kepler Crater (38°W, 8°N). Using this DEM [7] a perspective view of this crater ha been produced with a USGS color-rauo map overlaid. Korolev Basin ( 175°W-160°W, 10°S-20°N). Seventy-fve per­ cent of the Korolev basin is present in this DEM together with a pre­ Nectarid bain the size of Korololev just to the north [8]. The topography in this DEM rdges fom 0 km to 9.6 km. Named craters visible include Mach, McMath, Icas, and Tsdder. Luna 9/13 landing site (70°W-60°W, 0°N-30°N). The topo­ graphc range is fom -0 km to -4.5 km, the latter being on the foor of Hevelius A crater. Mare Crisium (50°E-70°E, 0°N-20°N). This DEM [4] covers the lower half of the bain dd suounding highldds. The topo­ graphc rdge is fom+ 3 km to-5 km. Several protuding foor crater rims ae visible on the bain foor (e.g., Picad, Yerkes, dd Lick). Mare Orientate ( 120°W-75 °W, 40°S-0°N). This aea was cov­ ered well in stereo, dd a regional north-south slope is detectable across the bain [9]. The topography varies fom +6 km on the westes rim to -5 km on the bottom of Maunder crater. Three rings of the basin ae clearly visible in the topgraphy. Mare Serenitatis ( 10°E-40°E, 10 °N-40°N). Highldd aea ae well depicted in the DEM, a ae craters Bessel, Dawes, Le Monier, Menelaus, Plinius, Posidonius, dd Ross. North polar region (60°-90°). An dalysis of this DEM [10,11] revealed the newly discovered Sylvester-Nansen pre-Nectarian im­ pact basin (30-400 km diceter, centered on 45°E, 83°N). The Schwarzchild dd Be'lkovich bains ae clealy visible, as is the northes extent of the Imbrium basin. Several seconda ejecta scour maks ae present in te topgapy and spreb out radially fom Mae Imbrium. The crater Hayn contains a north-south running graben which bisects its cental peaks. Saha crater area (90°E-105°E, 5°S-5°N). Stereo imagery is missing in the center of this region [7]. Schickard (60°W-50°W, 50°S-40°S ). The foor interior crater­ lets and surounding highands ae clearly visible [7]. Sinus lridum (40°W-30°W, 30°N-50°N). See Fig. 1 [4]. South polar region (90°S-60°S). An anaylsis of this DEM [4,7,10-15] has revealed two newly discovered pre-Nectaid im­ pact basins (Schrodinger-Zeeman, 250 km in diceter, centered on 165°W, 81°S, dd 330 km diceter Bailly-Newton, centered on 57°W, 73°S). Several other prominent basins ae visible, including the southes half of the South Pole Aitken (SPA) Bain, Bally, Schrodinger, par of Clavius, and the Amundsen-Gdswindt Bain. The topographic rdge in this DEM is considerably greater than the north pla DEM due to the presence of the SPA topgraphy. Tsiolkovsk ( 124°£, 19°S). In this DEM [16) the westes faks Fig. 1. Digital elevation models of Sinus Iridwn. Black = low (or no data) and white = high. of Tsiolkovsky ae shown in great detail. Tycho (20°W-0°W, 50°S-40°S ). This DEM shows Tycho crater dd its suoundings a well as the northes interior of Maginus Clementine laser btimeter DEM [2], the stereo-derived DEM con­ Crater. tains considerably more spatal detail. The DEMs produced fom Discussion: The best results obtained fom stereo matchng lie Clementine stereo are also complementary to Clementine laser altim­ over highldd regions, where there exists much surfae texture fr eter meaurements, because the latter prfor better over fat mare the stereo matcher to lock onto. Mae regions ae daker and rela­ aeas, but less well in highldd regions. Some of these DEMs are tively fatureless and hence result in noisy DEMs, which largely available on the Intesetat http://www.nam.edu/ceps/reseachcook preclude the detecton of wrnkle rdges. Nevertheless, craters dd top.htl. peas within mare aeas can be resolved and ptfles extacted fom Acknowledgments: We would lie to thank DLR Berlin fr these fatures. One of the drawbacks of the Clementne-derived the computing time fr some of the ealier DEMs and University DEMs is that the stereo coverage at latitudes below ±60° in lattude College London/Laser-Scd for the Gotcha stereo-matching sof­ contain many gaps. However, when compaed to d interpolated ware, originally written by T. Day. 10 Workhop on New Views of the Moon I

References: [1] Cook A. C. et_al (1997) Planet. Space. Sci., 44, fr the rim. Tis form closely approximates the long-wavelength 1135-1148. [2] Smith D. E. et al (1997) JGR, 102, 1591-1611. behavior of complex craters, while ignoring higher-fequency topg­ [3] Day T. et al (1992) Int. Archv. Photgrm. Remot. Sens., 29-B4, raphy, save the rim. This approximation is appropriate because crater 801-808. [4] Cook. A. C. et a (1996) Proc. 24th Vemadsky-Brown relaation is strongly controlled by long-wavelength topography. Micro., Moscow. [5) Robinson et al. (1999) LPS XXX, Abstact Loading is accomplished assuming a unifrm gravity feld (1.62 #1931. [6) Cook A. C. ad Riesinger H. (1996) LP S XXVII,249-250. m s-2) and a unifrm density of 290 kg m-3. The initial stress state [7] Cook A. C. et al. (1996) Annu. Geophys., 14(lll), C803. is set to be hydrostatic, ,with an additonal pressure term to account [8) Konopliv A. (1999) prsonal comunication. [9] Cook A. C. fr any overlying topography. This additional pressure term is ta­ et al. (1997) Annu. Geophys., 15(lll). [10) Cook A. C. et al. (1999) peredexpnentially with depth ( e.g., 7]. While te simulations quickly LPS XXX, Abstact #1154. [11) Cook A. C. et al. (1999) JGR, settle on a prefred stess state, ad while the fnal solution is fairly submitted. [12) Hogman H. et a. (1996) Annu. Geophys. 14(lll). insensitive to the choice of the e-flding depth of the taper, selecting [13) Cook A. C. ad RobinsonM. S. (1998)3rdintl. Conf Expl. and a e-folding depth close to the diaeter of the crater sets the initial Utiliz. ofthe Moon, Moscow, 21. [14] Spudis P. D. et al. (1998) Proc. stess state nea the prefered state. of New Views of the Moon I. [15) Cook A. C. and Robinson M. S. We asume a diually averaged surface temprature of -20 °C, (1998) 3rdIntl. Conf Expl. and Utiliz. of the Moon, Moscow, 21. and allow temperature to increase with depth at a rate of 50 K/km-1• [16) Cook A. C. et al. (1994) BAAS, 26, 1097. Assuming a theral conductivity of 2 W m-1 K-1 [8), this gradient tanslates to a heat fow of lOmW m-2, an extemely high value fr the Moon. Temprature, of course, will not increase without bound. To maximize relaation, we allow our temprature profile to increase SIULATING THE FORATION OF LUNAR FLOOR­ linearly until it reaches the solidus (assumed to be 120 °C) at a depth FRCTURE CRTER USJG ELASTOVISCOPLASTIC of 24.4 km, at which point it is kept constant. The presence of melt RLAXATION. A. J. Dombad ad J. J. Gillis, Depatment of will drop the bulk viscosity; however, we have no rheological contol Eath ad Plaetary Sciences ad McDonnell Center fr Spae fr patial melts. Therefre, we mae no attempt to simulate this Sciences, Washington University, One Brookings Drive, St. Louis situation. MO 63130, USA ([email protected]; [email protected].) Elastoviscoplastic rheological model. In general, geologic ma­ terials ca behave in three main ways: elaticaly, viscously (via Introduction: Luna foor-facture craters fred during the solid-state creep), and brittly (plastcity is a continuum approach to height of mae basat emplacement. Due to a general temporal ad simulate this phenomenon). We combine these three deformation spatia relation with the maria, these craters, numbering some 200, mechaisms in an extended Mawell solid, where the total stain ca may be diagnostic of the theral stucture of the crst durng this be broken down into a simple summation of the elastic, creep, ad time [I). As the nae suggests, these craters exhibit brittle failure, plastic strains. generally limited to the cent foor region. That, ad a shallower In relaation phenomena in general, the system takes advatage depth than fesh lunar craters [l), has led to two main theories as to of any meas possible to eliminate deviatoric stesses by relaxing their fration: laccolith emplacement under the crater [1-3] ad away the topography. Previous aalyses have only modeled the viscous relaation (2,4 ). The implications of each model fr the state viscous respnse. Compaatively, the elastic respnse in our model of the Moon's crst during this time ae quite different (2], so the ca augment the relaxation, to a pint. This efect decreases a the viabilit of each model must be checked. Laccolith emplacement has elastic response becomes stifer; indeed, in the limit of infnite elastic been treated elsewhere [1-3]. However, previous attempts to study Young's modulus (and with no platcity), the solution converges on the relaation of the craters have assumed only a unifrm, Newtonian the purely viscous soluton (5). Igneous rocks comon to the luna viscous response of the nea surface to the topographic driving nea-surace have Young's modulii in the rage of 10-100 GPa [8). frces, ad simply postulated that the factures resulted fom tensile To maximize relaxation, we use a Young's modulus of 10 GPa. stesses associated with foor uplif (2,4). Here, we use a more (There is negligible sensitvity to the other elastc modulus, the sophisticated rheological model (5] that includes not only non­ Poisson's ratio; we use 0.25.) Newtonia viscous behavior (i.e., the viscosity is stess-depndent), For the viscous response, we use a flow law fr steady-state creep but also incorporates elatic behavior and a plastic component to the in thoroughly dried Columbia diabase (9), because the high plagio­ rheology to directly simulate the formaion of the foor factures. The clase (-70 vol%) ad orthopyroxene ( -17 vol%) content is similar to results of our simulations show that while elatoviscoplastic relax­ the compsiton of the luna highlad crst a described by remote ation is ptentially viable fr lager foor-facture craters, it is not sensing [10) ad sample studies [11): noritic aorthosite. This fow viable fr craters with diaeters �60 km, the size of the majority of law is highy non-Newtonia, i.e., the viscosity is highly stress for-facture craters [l]. depndent. That, ad the vaiabilit with temprature, stads in The Moel: We employ the fnite element method, a numerical stong contat to previous examinations of lunar foor-fture cra­ technique well suited fr bounday-vaue problems, via the commer­ ter relaauon [2,4]. cially available MARC softwae package. To test the viability of To model discrete, brittle fultng, we assume "Byerlee' s rule," a topographic relaaton, our goa is to prepae the simulations as to stadard geodynamical techique [e.g., 7). We implement this "rle" mze the amount of relaxaton. with a agle of interal ficton of -40°, ad a higher-than-noral We take advatage of the natural axisymmety of craters, simulat­ cohesion of -3.2 MPa (to approximate the breaking of unfactured ing one radial plane. Inital shapes are based on data fr fesh craters rock). The actual behavior of geologic materials is more complex fom Pike (6). To simplif implementation, a furth order plyno­ than in our rheological model, so the uncertaintes in the platicity do mia is used fr the basin, while a third order inverse gnction is used not represent the state-of-te-art eror. LP/Contribution No. 980 11

100 larger craters, even within only 10 Ma, suggests that relaxationmay have been at leastpartly responsible forthe formationof largerfloor­ fracture craters. However, parameters in our simulations were se• 80 • l0Myr lected to maximize the amount of relaxation.Using more plausible choices (e.g., lower heat flows) will result in less relaxation,poten­ D 800Myr D tially much less.Consequently, we conclude that topographic relax­ 60 0 4.55 Gyr ation was not responsiblefor the formation of lunar floor-fracture 0 craters . D • Acknowledgments: This research is supported by a NASA 40 GSRP Fellowship (AID). � • References: [1] Schultz P. H. (1976) Moon, 15, 241-273. 8 [2] Wichman R. W. and Schultz P.H. (1995) JGR, 100, 21201- 20 • 21218. [3] Wichman R. W. and Schultz P.H. (1995) Icarus, 122, � 193-199. [4]Hall J.L.et al. (1981)JGR, 86, 9537-9552. [5] Dombard • A. J. and McKinnon W. B. (1999) LPSC XXX,Abstract #2006 . 0 .__..._._... _ ___.____.__._ ___.....__.___,__� [6] Pike R. J. (1977) In Impact and Explosion Cratering (Roddy 0 20 40 60 80 100 120 et al.,eds.), Pergamon,489-509. [7] Jaeger J.C. andCook N. G. W. (1979) Fundamentals of Rock Mechanics, Chapman and Hall. Diameter (km) [8] Turcotte D. L. and Schubert G. (1982) Geodynamics, Wiley. [9] Mackwell S. J. et al. (1998) JGR, 103, 975-984. [10] Pieters Fig. 1. State of rlaxaton fr luna crter of various sizes, based on C. M. (1986) Rev. of Geophys., 24, 557-578. [11] Taylor S. R. apparent depth under our model bsumptions. Craters <6 km see only (1975) Lunar Science: A Post-Apollo View, Pergamon. minor rlaxaton.

SAMPLE RETURN MISSION TO THE SOUTH POLE Results: Figure 1 shows the relaxationstate of lunar craters of AITKEN BASIN. M. B. Duke1, B. C. Clark2, T. Gamber2, P. G. various diameters, under our model assumptions. The relaxation Lucey3, G. Ryder1 , and G. J. Taylor3, 1Lunar and Planetary Institute, state is based on comparingthe current apparentdepth with the initial 3600 Bay A,rea Boulevard, Houston TX 77058, USA apparentdepth of the crater. (Apparentcrater depth is the depth of the ([email protected]; [email protected]),2LockheedMartin centralpoint of the crater referenced to some outside, background Astronautics,Denver CO 80201, USA ([email protected]), elevation.) Figure 1 is plotted at 3 different times: 10 Ma (a time 3University of Hawai'i at Manoa, Honolulu HI 96822, USA which bothHall et al. [4] and Wichman and Schultz [3] consider a ([email protected]; gjtay [email protected]). reasonable length of time for relaxation), 800Ma (an upper limit, based on the associationoffloor-fracture craters with maria emplace­ The South Pole Aitken Basin (SPA) is the largest and oldest ment [3]), and4.55 Ga (anabsolute upperlimit formost geodynamic observedfeature on the Moon. Compositional and topographicdata processes in this solarsystem). The simulation of a 120-km diameter fromGalileo, Clementine, andLunar Prospector have demonstrated crater terminated after -3.5 Ga at -75% relaxation because of a that SPA represents a distinctive major lunarterrane, which has not numericalinstability; however, scaling fromFig. 1 indicates a final been sampledeither by sample returnmissions (Apollo,Luna) or by relaxation state of 80-85%. lunarmeteorites. The floorof SPA is characterizedby maficcompo­ As expected, largercraters relaxmore rapidly thansmaller cra­ sitions enrichedin Fe, Ti, and Th in comparisonto its surroundings ters. Moreover, most relaxation occurs early on, particularly for [1,2]. This composition may represent melt rocks from the SPA largercraters, an effectthat attests to the highly nonlinearrheology. event, which would be mixtures of the preexistingcrust and mantle While largercraters canexperience substantial relaxation, thedegree rocks.However, the Fe content is higher than expected,and the large of relaxationof smaller craters canbe quite small.Craters of 40 km Apollo basin, within SPA, exposes deeper material with lower iron diameteror less see relaxationof< 5% after4.55 Ga. For comparison, content. Some of the Fe enrichment may represent mare and all of the craters examined byHall et al. [ 4] were 40 km in diameter cryptomare deposits. No model adequately accounts forall of the or less and saw relaxation of -70%. characteristics of the SPA and disagreements are fundamental. Is The inclusion of plasticity does speed up the relaxationprocess. mantle material exposed or contained asfragments in melt rockand For example, an 80-km crater relaxes to -26% at 4.55 Ga without breccias? If impact melt is present, did the vast sheet differentiate? plasticity, compared to -33% withplasticity (see Fig. 1). The result­ Wasthe initial mantle andcrust compositionally different fromother ing plastic strains, also as expected, are concentrated within the regions of theMoon? Was the impact event somehow peculiar, (e.g., centralfloorregion, with plasticitypenetrating most deeply forlarger a low-velocity impact)? The precise time of formationof the SPA is craters. This plastic zone is, however, shallow (-8 km), and the unknown, being limited only by the initial differentiation of the magnitude of the strainsis somewhat small (-3%). Indeed, a 20.km Moon and the age of the Imbrium event, believed to be 3.9 b.y. crater does not experience any brittle failure on its floor, nor is its The questions raised by the SPA can be addressed only with relaxationsped up by the inclusion of plasticity. detailed sampleanalysis. Analysis of the melt rocks, fragments in Discussion: If it is assumed that one process was responsible breccias, and basalts of SPA can address severalhighly significant for the formation of allfloor-fracture craters, then this relaxation problems forthe Moon and the historyof the solarsystem. The time model clearly fails. However, the moderate degreeof relaxation of of formation ofSPA, based on analysisof melt rocks formedin the 12 Workhop on New Views of the Moon I

event, would put limits on the period of intense bombadment of the addressed by saples fom SPA; choice of landing site to maimize Moon, which ha been infred by some to include a "terminal the probability of addressing the first-order problems; saple size cataclysm." If close to 3.9 Ga, the presumed age of the Imbrium ad the distribution between regolith ad rocklet samples; details of Bain, the SPA date would confr the luna cataclysm. This epi­ sample collection (rage fom lander, depth, avoidance of contai­ sode, if it occued, would have affected all of the plaets of the inner nation fom lader); and environmental control constraints on samples solar system, ad in particula, could have been critical to the history (maimum tempratures, acceptable leak rates on Earth). Te premse of lif on Earth. If the SPA is signifcantly older, a more orderly is that ths mission will have to be kept as simple ad a low-cost as cratering history may be infred. Secondly, melt-rock compositions possible to be selected by the Discovery program. and clasts in melt rocks or breccias may yield evidence of the com­ References: [I) Lucey et al. (1998) JGR, 103, 3701-3708. position of the luna mantle, which could have been penetated by the [2) Lawrence et al. (1998) Science, 281, 1484-1489. [3] Waen et impact or exposed by the rebound process that occured after the al. (1995) Eos Trans. AGU, 77, 33. impact. Tirdly, study of mare ad cryptomare basalts could yield frther constraints on the age of SPA and the theral histor of the crust and matle in that region. The integration of these data may SUBPIL DETECTION OF PYROCLASTIC MATERIALS allow infrences to be made on the nature of the impacting body. I CLE:MENTI ULTRVIOLET-VISIBLE DATA. W. H. Secondary science objectives in saples fom the SPA could Faad1 and L. R. Gaddis2, 1Space Science Institute, 1540 30th include aalysis of the regolith fr the lattudinal effects of sola­ Street #23, Boulder CO 80303, USA ([email protected]), 2U.S. wind irradiation, which should be reduced fom its equatoral values; Geological Survey, Astogeology Progra, 2255 N. Gemin Drive, possible remnat magnetization of ver old basalts; and evidence fr Flagstaf AZ 86001, USA. Imbrium Basin ejecta ad KREEP materials. If a sapling site is chosen close enough to the ples, it is pssible that indirect evidence Introduction: Lunar pyroclastic deposits represent a style of of pola-ice deposits may be fund in the fn of oxidzed or hy­ volcaism different fom that respnsible for the food basalts that fll drated regolith consttuents. the mae basins. As volatile-coated, primitive materials originating A saple retus mission to the Moon may be possible within the deep (-400 km) within the Moon, these products of explosive volca­ constaints of NASA's Discovery Progra. Recent progress in the nic erptions ae aso importat as probes of matle composition [ 1] development of sample ret canisters for Genesis, Stardust, ad ad a a potential resouce [2] fr future settlers. While mby of the Mas Saple Ret missions suggests that a small capsule can be luna pyroclastic depsits ae spatially resticted ad relatively sml reted directly to the gound without a paachute, thus reducing its in size, they ae eaily resolvable at the spatial scae (-10 mpixel) mass ad complexity. Ret of a 1-kg sbple fom the luna surface of the Clementine UV-VIS camera. Recent studies confrm previous would appa to be compatible with a Delta II clas launch fom results indicating that these depsits ae not compositionaly unifrm Earth, or possibly with a piggyback opprtunity on a commercial [3-4], ad suggest that further analyses can help to identif pssible launch to GEO. A total mission price tag on the orderof $10 million genetic relationships among luna pyroclatic dpsits, chaacterize would be a goal. Target date would be late 2002. Saples would be their juvenile components, ad clarif their relatonships to neaby reted to the cuatorial facility at the Johnson Space Center fr maria. description ad allocation fr investigations. Among the juvenile materials fom sampled luna pyroclastic Concentration of mlligra-to gram-sizedrocklets is a very egec­ depsits ae the orange glass ad devitifed black beads fund in tive stategy fr sample studies of the luna regolith. A rake accom­ the Taurs-Littow Valley [5-6) ad the green glass found by plished this type of sampling in the Apollo missions. For the SPA Apollo I5 [I]. Recent studes [3-4] suggest that deposits dominated saple ret mission, either a small rover or a arm on a lader by materials such a these may represent end members in the ob­ would deliver regolith to a sieving mechanism that retains fagments seted compsitional vaiatons among the luna pyroclatic deps­ in the 1-lOmsizerage.Approximately 10%ofthe massofApllo its. Here we present preliminary results of aalyses fcused on the 16 regolith samples, which were fom possibly simila highlad use of the Clementine UV-VIS data fr chaacterizing the composi­ terain, consisted of fagments in the size range. To ret 1 kg of rock tion ad distibution of juvenile pyroclastic materials. Our test case fagments, approximately 5 x 1( cm3 of regolith would have to be fr detailed mapping of a luna pyroclastic deposit is that of the sapled. Waen et al. [3] suggested 7-10 mm as the optimum size Apollo 17 lading site in the Tauus-Littow (TL) Valley (Fig. la, b). fr individual saples, which would require more regolith to be Although black beads domnate the obsered spctal refectace at sieved. this site [5-6), saple data show that the pyroclatic esptonchaged This mission would represent the frst lander mission to the luna chaacter, producing frst orage glas ad then black beads [3]. To faside ad, as such, would require that a comunication link be assess the compositional vaiability of this deposit, especially our established with the Earth. A growing number of assets at the Sun­ ability to distinguish the orage glases, we apply techniques based Earh L-1 libration point may provide access to a viable communica­ on spectal mixture analysis [7] to detect materals at subpixel scales. ton link, avoiding the need fr a comunications orbiter. The The low albedo and subdued absorption fatures of the Taus­ mission need only be designed to last through a single luna day, Littow deposit make this a chalenging task. which could make it relatvely staightfrward; if a rover is chosen a Detectability Analyse: In recent years, several subpixel de­ the implementation fr sapling, it may be possible to keep the rover tection techniques have been developed fr use with terestal air­ alive fr longer. This would be a costbeneft tadeof to be deter­ bore imaging spctometer data [e.g., 8-10). A technique that is mined a pat of the mission aalysis. fnctionally equivalent to spect mixture aalysis, the orthogonal Issues on which the luna sample community should make input subspace projection (O SP) technique [11] is used fr the simulations include: identifcaton of additional scientifc problems that can be presented here. In OSP, a taget spectum is projected onto a sub- LP/ Contribution No. 980 13

60, cd 80, the orcge soil taget was added in abundances of 90, 80, 60, and 40%. The 100-sample set was then reduced via OSP (Fig. 3). For this example, the orange soil was detectable only at the 90 cd 80% aundance levels. It wb fund that the addiuon of higher noise levels (-1 % ), made the orange glbs undetectable even at the 90% level. ijowever, using background materials more representative of the highlands made the orcge soil detectable at lower abundances. Future Work: These results suggest that we should be able to map the distibution of juvenile pyroclastic materias, such b the orange glasses, using the Clementine UV-VIS data cd subpixel caysis techniques such b spectal mixture calysis [7] cd fre­ ground background calysis [10]. Given the low albedo of these materials, a high fll fctor will be required on a pr pixel basis in order to achieve that mapping; however, obseratons made by

endmembers used in simulaton 0.2.----.----r---r---.-----,----,

0.18

0.16

e 0.14

�Q) 0.12 c::: 0.1

400 500 600 700 800 900 1000 Waveleng (nm)

Fig. 2. End members used in detectability simulation. Fig. 1. (a) 750-nmband for Taurus-Littrow Valley. The Sculptured Hills are in the upper right and South Massif is in the lower left. North is toward the top. (b) Albedo-normalized [14] composite of 950-, 750-, and 415-nm bands. Colors are due to both compositional and maturity differences. Bright blue units such as those in the crater cluster (upper detecton simulaton middle) expose fresh regolith of highland composition.Other green and blue-green features mark the locations of highland massifs and fresh 0.8 craters. Soils with substantial pyroclastic components, primarily black beads,are shown i bright red with yellow overtones. (To view this figure 0.6 in color, go to http://cass.jsc.nasa.gov/meetings/moon99/pdfl8053.pdf.) g 0.4

"O 0.2 space that is orhogonal to a set of background specta. In this process, the response fom the background spcta ae nulled and tat 0 of the taget is maximized. For the T site, the specta used fr the -0.2 simulation (Fig. 2) included three laoratory-measured sbple spc­ ta [12] convolved to the fve UV-VIS bcdpasses, cd two spcta -0.4 extacted fom UV-VIS data over the T Valey. The taget spectum -0.6 was the orcge soil sample 74220 fom the Shorty Crater rim. 0 20 40 60 80 100 "Background" specta were fom samples 74221 (a gray soil fund sample number nea the orcge soil) cd 75111 (a dk mae soil). From the UV-VIS data, additonal background spcta were obtained at the mare/high­ Fig. 3. Normalized response from detection simulation. Sample 20 lcd interface and fom the "crater cluster" area in the T Valey. In contained 90% of the orangesoil target, sample 40contained 80%, sample the simulaton, the bbkground spcta were randomly mixed in ebh 60 contained 60%, and sample 80 contained 20%. T higher response of 10 sbples with 0.1 % Gaussian noise added. For samples 20, 40, in samples 20 and 40 indicate detection of the target. 14 Workhop on New Views of the Moon II

others, including orbital obsesations by the Apollo 17 astronauts data to search fr relatively small spatial-scale enhacements in H [14], have indicated that these abundaces are met for several pyro­ abundances at bothlunar poles. Maps were constucted of epitheral­ clastc deposits. The production of such maps will help to constain neutrons corected fr elementa abundace variations by subtract­ the dynamics of pyroclastic erptions on the Moon by providing ing 7% of measured theral-neutron counting rates. Although the inforation on the type, relative quatity, and distribution of juve- spatial resolution of the LPNS at 30-km altitude is -55 km FWHM, ° ° nile volcanic materials [15]. , we binned all the data in 0.5 x 0.5 spatial pixels and then applied References: [l]DelaoJ. W. (1986)Proc. LPSC 16th, in JGR, a 30-km FW Gaussia smooting algorithm. Resultat pla 91, D201-D213. [2] Hawke B. R. et a. (1990) Proc. LPSC 20th, maps of corrected epitheral counting rates ae shown in Fig. 1. 249-258. [3] Weitz C. A. et al. (1998) JGR, 103, 22725-22759. Inspction reveals discrete depressions in counting rates that are [4] Gaddis L. R. et al. (1999) JGR, in press. [5] Adams J. B. et al. suprimpsed on more g�nerally distributed depressions that sur­ (1974) Proc. LSC 5th, 171-186. [6] Pieters C. M. et al. (1974) round both poles. Comparison with the rada-meaured pla topg­ Science, 183, 1191-1194. [7] Adams J. B. et al. (1993) in Remote raphy shows tat the aeas of most depressed epithermal counts rates Geochemical Analysis (C. Pieters and P. Englert, eds.), pp. 145-166. in the south overlie craters that have foors in peranent shade. [8] Farrand W. H. ad Harsayi J. C. (1995) JGR, JOO, 1565-1578. Fuherore, these depressions are neither cylindrically symmetc [9] Faand W. H. ad HasayiJ. C. (1997) Rem. Sens. Environ., 59, about either ple nor do they minimize at the ples. Simila maps of 64-76. [10] Smith M. 0. et al. (1994) IGARSS '94, 2372. fat neutons (not shown here) reveal a single, statistically signifcant [11] Hasayi J: C. and Chag C-1. (1994) TGARS, 32. [12] Adams depression centered on the maximum depression of the epithermal J. B. and McCord T. B. (1971) Proc. LSC 2nd, 2183-2195. neutons at about 88°S, 20°E. Comparison between high-altitude [13] Lucchitta B. K. and Schtt H. H. (1974) Proc. LSC 5th, 223- ad low-atitude epithenal maps reveals a lager depression at low 234. [14]PouchG. W. adCapagnaD.J. (1990)PE &RS, 56, 475- altitudes in the south but the sae magnitude of depression in the 479. [15] Gaddis L. R., this volume. north. Hydrogen enhancements in the south must therefre have spatial scaes comparable to the spatal resolution of LPNS, 55-km FWH, yet consist of smaller clumps more unifrmly distbuted ENHANCED HYDROGEN ABUNANCES NEAR BOTH over the LPNS feld of view in the norh. A quatitative comparison LUNAR POLES. W. C. Feldma1, S. Maurice2, D. J. Lawrence1, between measued epithenal countng rates ad numerical simula­ I. Getenay2, R. C. Elphicl, B. L. Baacloughl, and A.�- Binder3, tions using the measured pola topgraphy [5) yields the fllowing: Los Alaos Natonal Laboratory, Los Alamos NM 87545, USA (1) H abundances within the prently shaded craters nea the ([email protected]), 2Observatoire Midi-Pyrenees, Toulouse, south ple ae equivalent to a water-ice mass faction of 1.5±0.8%, Frace, 3Lunar Reseach Institute, Gilroy CA 95020, USA.

Chemical aalyses of all samples of the Moon reted to Earth show that the luna surface is highly depleted in volatles [l]. Specif­ cally, the H content of luna soils averages only 50 ppm, which ca b explained in terms of surface implantation of sola-wind H. We note that all reted saples come fom near-equatorial latitudes where daytime temperatures ae sufciently high that water is not stable to evaporation, photo dissociation, ionization, ad eventual 49:::J.t loss to space through pickup by the sola wind. However, it has long 00 been postulated that a signifcant faction of water delivered to the Moon by comets, meteoroids, ad interplaneta dust ca be stably er, tappd within the prently shaded foors of pola craters where ' "': ., temperatures ae sufciently low so that sublimation times ca be longer tha several billion yeas [2-3]. .,I(\< - -"'Is- • - Recent results fom aalysis of the high-alttude (100 ± 20 km) prion of the Lunar Prospctor Neuton Spctometer (LPNS) dataset �yithcr-mcl r:eutror.s ot t-..,e sc1.,tr. pole (l.:,iN) [4] have revealed that H abundances nea both luna ples ae 0 enhaced relative to that which exist at equatorial latitudes. Because this average enhacement is not much lager tha the nea-equatorial OOJ.O average of 50 ppm, it is reasonable to ask how much of the pla-H enhancement comes fom the sola wind ad how much comes fom luna impacts by solid interplaetar materials. Perhaps the low temperatures at pla latitudes could reduce loss rates of sola-wind­ implanted H sufciently to account fr the infered digerence be­ tween average plar ad equatoria H abundaces. Although the fregoing suggestion is plausible, neither labora­ tor simulatons on reted soil saples nor numerical simulations of H loss rates fom the radiaton-daaged surfaces of soil grains have been prored to prove its feasibility. We t to address this Fig. 1. Counting rates of (epithermal = 0.07 x thermal) neutrons queston by aalyzing the low-altitude (30 ± 15 km) portion of LPNS polewad of ±75 ° lattude. LP/ Contribution No. 980 15

(2) the enhaced H within these craters must not be buried beneath The basaltic clast shows a very fne intergraula texture with the surface by more tha -IO g cm-2 (5 cm at a density of 2 g cm-3), some lager plagioclase grains. The two clasts are separated by a (3) the H abundance nea both poles averages 100 ppm above that melt-glas matrix that contains plagioclase with the sae chemical known fom nea-equatorial reted soil samples (50 ppm, [1 ]), ad composition as those plagioclases within the fldspathic clast. A ( 4) the total mas content of H poleward of -7 5° is -20 x 106 t, ad de vitrifed mafc-glass spherule of -150 mm diameter wa identifed that pleward of+ 75° is -150 x 106 t. If the enhaced H within all within the matix as well a other small fagments of pssible glass regions of prmaent shade nea both poles is in the fr of water ice, spheres. Some dendritic veins ca be obsered on the outer pa of the use of the aea estimates given by Magot et a. [5] yields estimates section, which formed after breccia lithifcation[l-2]. of 135 (240) x 106 tin the 2250 (400) km2 of shaded aeas polewad Exprimental Methods: Two laser 4Ar-39Ar experiments are of -87.5°, and 62 x 106 tin the 1030 km2 of shaded aeas polewad being carried out: (1) infaed laser-steppd heating of feldspathic of +87.5°. fagments that have been sepaated fom the meteorite; ad (2) We note, however, that neuton obserations by themselves ca­ ultaviolet laer spt fsion (50 µm) of a slice of meteorite (-1.0 x 0.5 not uniquely identf the chemical frm of H in luna regolith. That cm) showing both clast types. Preliminary results fom experiments is, LPNS data caot discrminate solar-wind-implated H fom OH, (1) have been obtained so far. Three fldspathic fagments (DAG1- H20, etc. In the absence of a tempratue map of the luna ples ad 3, each weighing -1 mg) have been heated using between 2030 a knowledge of the retenuvity of H by luna soil grains a a function steps, of increasing laser output power, each of 1 min duraton. of temperature, suffcient information is not yet available to uniquely Argon-40-Argon-39 Age Constrainu: Figure l is a plot of identify water-ice depsits at the luna poles. Nevertheless, the foors 36Ar/40Ar vs. 39Ar/4Ar fr 20 extactions obtaned fom DAG3. of pranently shaded pola craters ae predicted to have tempra­ Overal, the data show increasing 36Ar/ 4 Ar and decreasing 39Ar/ 4 Ar tures suffciently low that water ice should be stable for billions of as heating proceeds. In detail the releae is more complex; pints 1- yeas, yet uncratered suraces nea the pola cap are too high in 4 show a signifcant decrease in 39Ar/4Ar coresponding to a temperature to retain water ice [6]. Tis suggests that at least some increase in appaent age fom 1.3 to 5.1 Ga. This vaiation may be of the enhaced deposits of H identifed by LPNS are in the fr of attibuted to the combined efects of terestrial alteration that result water ice. in loss of radiogenic 4Ar ad implantation of luna 4Ar in grain References: [1] Haskin and Warren (1991) in Lunar boundaies that lead to a increase in apaent age. Sourcebook A User's Guide to the Moon (G. H. Heiken et al., eds.), Points 5-20 defne a linea corelation interpreted to represent p. 357, Cabridge Univ. [2] Watson K. et al. (1961) JGR, 66, 1598. mixing between radiogenic 4Ar, releaed at interediate tempra­ [3] Aold J. (1979) JGR, 84, 5659. [4] Feldma et al. (1998) ture, and tappd Ar that dominates the high temperature release Science, 281, 1496. [5] Magot et al. (1999) Science, 284, 1658. (Fig. 1 ). Te intercept of the corelation on the 39Ar/ 4Ar ais core­ [6] Vasavada A. R. et al. (1998) Icars, submitted. sponds to a age of 1.95 Ga. This age is low compaed with other luna meteortes and is most likely related to a shock event. The trappd component released at high-temperature ha 36Ar/4Ar of -2.8, which is close the value of -2.5 reported previously fr the bulk LASER ARGON-40ARGON-39 AGE STUIS OF DAR meteorite [2]. AL GANI 262 LUNAR MTEORITE. V. A. Ferades, R. The increaing proportion of tapd Ar releaed at high tempra­ Burgess, adG. Tuer, DepaentofEathSciences, University of ture is contary to the usual release patter obtained fr luna soils Manchester, Oxford Road, Manchester Ml3 9PL, UK [5]. Two pssible explaations ae given: (1) Ar loss fom grain ([email protected]) boundaries occured at depth within the luna regolith, while pre­ sering tapd Ar with higher 36Ar/40Ar in the center of grains; or Introducton: The luna meteorite Dar al Gani 262 (DAG 262) wa fund in the Sahaa Desert in Libya on Mach 23, l 997. Ts wa the frst luna meteorite fund in a dsert ad is the thirteenth luna meteorite discovered. DAG 262 is a polymict aorthositic luna highlad breccia [1-2). TABLE 1. Electon-mcroprob aalyses of low-C Poxene and pla- 4 39Ar dating technique has been applied to DAG 262 in The Ar- golase ad of a devitfed glass-spherle fund within te matix. an attempt to determine the crystallization age ad shock events exprienced by this meteorite. Previous studies have indicated that Low-C Proxene Plagolase Dvitfed glas this meteorite may have sufered up to fur shock events [3]. Due to the brecciated nature of the rock ad the likelihood of multple shock SiO 38.35 4.23 35.19 events, a laser-probe technique ha been used to aalyze individual 2 TiO2 0.1173 0.0145 0.0671 components (minerals ad clats) of the meteorite. Alz3 0.0840 35.3244 0.0000 Sample Decription: The sample supplied to us (-1.5 g) shows Crz03 0.0845 0.0000 0.0338 two distinct clast types (1) fldspathic ad (2) baaltc, the latter not FeO 21.42 0.19 36.47 having been previously described [1-]. Plagioclaes show factures MnO 0.3321 0.0000 0.4154 ad undulatory extincton as the result of shock events [1-3]. The MgO 39.38 0.19 27.73 fldspathic clast has small, round low-Ca pyroxenes distibuted CaO 0.1625 19.6293 0.0623 KiO 0.0236 0.0025 within it.Electon-microprobe analyses of feldspa ad pyroxene ae 0.00 Nap 0.0237 0.4024 0.0268 given in Table 1. The bulk composition a deterined by [4] suggests NiO 0.0490 0.0000 0.0000 that the source of this clast is the froa-anorthositic suite. 16 Workhop on New Vews of the Moon I

0.45 ,------, PROGRESS TOWARD CHARACTERIZATION OF 0.40 JUVENILE MATERIALS IN LUNAR PYROCLASTIC DEPOSITS. L. R. Gaddis, Astrogeology Team, U. S. Geological 0.35 ,,7-19 �.,s Survey, 2255 N. Gemini Drive, Flagstaff AZ 86001, USA 0.30 ---!� 1412 ([email protected]). ct 0.25 '�,o 9 �3t,� � 0.20 Introduction: In recent analyses, the 5-bandClementin e UV­ {- ,. ·· .... VIS data have been used to examine the compositions of lunar 0.15 .iec. .•_ 7 pyroclasticdeposits [e.g., 1-3). A primary goal of these analyses has 0.10 ...... , been the characterization of the primaryvolcanic or juvenile compo­ nents of these deposits. The compositions,physical and morphologi­ 0.05 . +-...... cal characteristics, and spatial distributions of juvenile volcanic 0.00 �������������������·., 0.000 0.005 0.010 0.015 0.020 0.025 materials provide informationon the distributionof primary mafic 39Art°Ar materials on the Moon, conditions required fortheir eruption at the surface, and the behavior of lunar volcanic processes over time. Fig. 1. Argon-36/Argon-40 vs. 39Ar/40Ar laser-stepped heating results Using currentanalytical techniques with the new Clementine UV­ for DAG3. The numbers adjacent to the data points indicate the laser VIS global mosaic [4-6), anddata from the GLGM2 geophysical . extraction step. models [7), to supplement ongoing work with Earth-based spectral reflectance analyses and laboratory investigations, my colleagues andI have adopted a three-pronged approach to these issues involv­ ing: (1) compositionalanalyses of lunar pyroclastic deposits; (2) (2) release of trappedAr from a low-K-bearing mineral present in characterization of the relations between effusive and explosive feldspathicfragments. Anomalously high apparent ages (>4.5 Ga) lunarvolcanism; and(3) examination of the global occurrence and obtainedat low temperaturefrom steps 3 and4 indicate that DAG3 distribution of lunar pyroclasticdeposits. This report and related containsimplanted lunaratmosphere 40Ar presumablyacquired near work [8] describe progress towardremote characterization of the the surface of the lunarregolith. This is inconsistent with the expla­ compositions ofjuvenile materials in the pyroclastic deposits located nation (1) that loss occurred at depth. Fine-grained pyroxene is a at Taurus-Littrowand J. Herschel. These studies have implications common component in the matrix of DAG 262 [2], has low K forcharacterization of the relations between the products of effusive content, and will retain Ar to a higher temperature than plagioclase and explosive volcanism on the Moon. makingexplanation (2) the most likely. Background: Analyses of lunarpyroclastic materials, prima­ Release of Ar fromfragment DAG 1 was similar to that of DAG3 rily the juvenile picritic glasses, provide unique information on the giving an indistinguishable intercept age of2.0 Ga. However, DAG2 composition of the mantle [9] and on the nature and origin of contained much lower levels of trappedAr and 19% of the total39 Ar associatedvolatile elements in an otherwisevolatile-depleted envi­ released at high temperature gave an age of 3.05 Ga. This age is ronment [10). Possible fundamental differences between picritic slightly below previously inferredtotal gas ages forlunar meteorites glasses and mare basalts (e.g., lesser fractional crystallization and andmay be related to the crystallizationevent. greater depth of origin forglasses [11]) support theiridenti fication Cosmic Exosure Ages: These are based on plots of 37Ar/3 6 Ar asthe best examplesof primitive materials on the Moon, and attest vs 38Ar/36Ar andgive ages of 185 Ma, 252 Ma and122 Ma forDAG 1, to their importance in characterizing the lunar interior and as a DAG2 andDAG3 respectively.The cosmic-ray exposure ages indi­ startingplace forunderstanding the origin and evolution of basaltic cate the components ofDAG 262 had differentexposure histories, an magmatism on the Moon. Remote-sensing analyses of these deposits expectedresult given that this meteorite is a breccia. have helpedus to identifythe characteristic componentsof some of Chemistry: The K/Ca weight ratio calculated from39 Ar/3 7Ar these deposits [ 12-14 ], to begin to constrain the distribution oflunar of 3.6 x 10-3 is similar to the bulk ratio obtained by [2] of 3.8 x volcanic deposits [15), and to understandthe styles of eruptionand 10-3. A much lower K/Ca of 0.6 x 10-3 obtained from feldsparby emplacement of basalts on the Moon [1,16). To fullyappreciate the electron microprobe is evidence for the impurity of the feldspar role of pyroclastic volcanism on the Moon, we must understand the separates analyzed for the age study. The low K/Ca ratio of the ranges of composition, spatialand temporal distribution, relation­ feldsparcompared to the bulk indicates that there is anadditional K­ ship to effusive volcanic deposits, and modes of occurrence and bearing phase in DAG 262 that remains to be identified. formationof pyroclastic deposits.Until additional samples areavail­ Currentinvestigations are aimed at using the UV laser to deter­ able, remote analyses of lunar pyroclastic deposits are a primary mine 40Ar_39Ar ages fromindividual minerals within DAG 262. meansof fullycharacterizing these deposits. :cknowledgments: We wish to thank I. Franchi(Open Uni­ More than 100 pyroclastic deposits are identified (see http:// versity) for providing the sample of DAG 262. This work was wwwflag.wr.usgs.gov/U SGSFJag/Space/LunPyro) on the Moon. supported by the EuropeanCommission via a TMR fellowship to Their characteristics,summarized here briefly,are described in detail VAF. elsewhere [e.g., 1-3). Lunarpyroclasticdepositsaredarkandsmooth­ References: [1] BischoffA. and Weber D. (1997) MAPS, 32, surfaced, and they are observed in association with sinuous rilles, A13-A14. [2] BischoffA. et al. (1998) MAPS, 33, 921-935. irregulardepressions, or endogenic craters within highlands and/or [3] Nishiizumi K. et al. (1998) LPS XXIX, Abstract #1957. on the floors of old impact craters situated along the margins of many [4] JoliffB.et al. (1999) LPS XIX, Abstract#2000. [SJ Burgess R. mare-filledbasins on the lunarnearside and farside. Lunar pyroclas­ andTurner G. (1997) MAPS, 33, 921-935. tic deposits aredivided into two classes on the basisof size, morphol- LPI Contribution No. 980 17

ogy, and occurrence [e.g., 14]. The -12 largedeposits areof regional highlandand mareunits may be observed for both large andsmall extent (up to several tens of thousands km2), while small deposits deposits [2]. (-90 in number) are more localized,typically only several hundred To characterizejuvenile materials in lunar pyroclasticdeposits, km2. Large deposits were probably emplaced via Strombolian-style we are using the Clementine multispectral data at both UV-VIS and or continuous fire-fountaineruptions, with wide dispersion of well­ NIR wavelengths(-0.4-2.7 µm) to study the large deposit at Taurus­ sorted pyroclasts (16]. A significant component of Fe+2-bearing Littrow(TL) [8] and the small deposit atJ. Herschel crater. Although volcanic glass beads was identifiedin manyof the regional pyroclas­ black beads dominate the observed spectralreflectance at the TL site tic deposits[e.g., 12-14]. AtTaurus-Littrow(-SOiem west of Apollo [12,13], sample data show that the pyroclastic eruption changed 17), black beads (the crystallized equivalent of Fe+2-bearing orange character, producingfirst orange glass andthen black beads (10). To glasses) are the characteristic ingredient of the regional pyroclastic assess the compositional variability of this deposit, especially our deposit[12-13]. Both orange andblack beads were recognized as ability to distinguish and map the orange glasses, we apply tech­ pyroclastic, with variations in coolingtime in a firefountain produc­ niques based on spectral mixture analysis to detect materials at ing quenched, crystallized, and/orcomposite droplets [10-12]. subpixel scales. Preliminary detectability analyses using the Small pyroclastic deposits were formed via Vulcanian-styleor Clementine UV-VIS data andthe Orthogonal Subspace Projection intermittenteruptions [17], with explosive decompression removing technique with targetand background spectra(from both the labora­ a plug of lava or caprock within a conduit and forming a vent tory and the scene) indicate that orangesoil at TL is detectableat the depression. Smallpyroclastic deposits were divided into three classes 90 and80% abundance levels([8] andreferences therein). To evalu­ on the basisof their "l.0-mm" or maficabsorption bands[18]. Most ate our ability to map the distributionof different types of juvenile Group 1 deposits are mixtures of highlands-rich countryrock and pyroclastic materials at individual andmultiple sites, and thus to glassy juvenile material with small amounts of basaltic caprock provide constraints on their eruption mechanisms, we will apply material. Group 2 deposits consist largely of fragmented basaltic these results to TL and to related depositsat Sulpicius Gallus, Rima material,with insignificantamounts of highlandand glassy materi­ Bode, Sinus Aestuum, andMare Vaporum. als.Group 3 depositsare dominated by olivine and orthopyroxene; An additional approach to the characterization of the juvenile the olivine is almost certainlyassociated with juvenile material, and materialsin lunarpyroclastic depositsfocuses on spectral analysisof the orthopyroxene is likely to have been emplaced as a result of the small deposit at J. Herschel Crater. This deposit is somewhat erosion andentrainment of the wall rock (14]. unique in that it is the only currentmember of Group 3 and it appears RecentWork: Although differingorigins have been proposed to have a substantial olivinecomponent [e.g., 19]. At visible to near­ forthe large andsmall pyroclasticdeposits, comparison of Clementine infraredwavelengths, olivine hasa broad, relatively deep absorption UV-VIS spectralreflectance data for these deposits shows a linear band at 1.0 µm, but it has no absorption feature at 2.0 µm. Iron­ trend(Fig. I), suggestingsome compositional relationshipbetween bearing pyroclastic glass,which may be present in additionto or in the two typesof deposits. Weitz et al.[I], in studying several large lieu of olivine, hasbroad absorption bandsat both 1.0 and2.0 µm pyroclastic deposits (Taurus-Littrow, Sulpicius Gallus, Mare [e.g., 14]. To determine whether olivine is indeed the dominant Vaporum, Rima Bode, Sinus Aestuum, Aristarchus Plateau, and juvenile componentat J. Herschel, the Clementine NIR data were Orientale), proposedthat this lineartrend represents a "mixingline" examined forthis area. Although calibrationof these data is currently between those large deposits with -100% crystallized beads and under way andis thus incomplete at present, comparisonswithin the those with -100% glass beads. A Sinus Aestuum dark spot is the NIR dataset can be made.No absorption featureat 2.0 µmis observed crystallized end member,and Aristarchus is the glass end member. at the volcanicvents at J. Herschel. This supports the identification However, the overlap between the small and large deposits and of olivine as the primaryjuvenile component and the absence of between the pyroclastic deposits and nearby mare andhighlands volcanic glasses as major contributors to the spectralsignatures in units (Fig. 1) suggests that (1) the juvenile pyroclasticmaterials are this area.Application of the Lucey-derived compositional(FeO and

not glassesin all cases and(2) substantialspectral contributions from TiO2) and maturity relations[20] shows affinities between the J. Herschel deposit andancient mariain nearbyMare Frigoris.

J.-.0:, .. {'i W.....l1) [1] Weitz et al. (1998) 22725-22759. .�'.>f:i !B acdol(I) References: JGR, __ ., .,.._.,,, X � .... (::I) ,.-,_,,ICb�� (2) Gaddis et al. (1999) in press. [3] Gaddis et al. (1999) r=i...... c.-: • .;....,...!2, JGR, LPS ISl i (�!!.!.t) -··,\p('li<,!11� El! .-o, ---..., X, 1732. [4] Eliasonet al., this volume. [5] McEwen et al.(1998) i:S:lg�::.r, ...... :-:,�•� _.,, __ o. 5 t • __ LPS XIX, 1466-1467. [6] Pieters et al. (1997) online document. Lunar� Dopoobo: t SmoAondlargo [7]Neumannet al. (l996)JGR, 16841-16863; Lemoine et al.(1997)

0.7 fl'l'\1111 .....7 JGR, 16339-16359. [7] Gaddis et al. (1998) LPS XIX, 1807-1808. \ ,.� (8) Farrandand Gaddis, this volume. [9] Delano(1986) Proc. LPSC 0 0.65 Weitzetal.,98 16th, in JGR, 91, D201-D213. [10) Heiken et al. (1974) GCA, 38, 1703-1718. [11) Shearer and Papike (1993) GCA, 57, 4785-4812; Papike et al. (1998)Rev. Mineral., 36, ch. 5. (12) Adams et al. (1974) LSC V, 171-186. [13] Pieters et al. (1973) JGR, 78, 5867-5875. i (14] Gaddis et al. (1985)/carus, 61, 461-488.[15] Head(1974) LSC t V, 207-222. [16] Wilson and Head (1981) JGR, 78, 2971-3001. 0.5 ...._ ..... _. .. (17] Head andWilson (1979)LPS X, 2861-2897. [18] Hawke et al. 1 1.05 ,., ,.,s 1.2 1.25 (1989) LPS XIX, 255-268; Hawke et al. (1990) LPS X, 249-258. ssonso weaker-+- -stroger [19] McCord et al. (1981). [20] Luceyetal. (1998a)JGR, 103, 3670- Fig. 1. 3699; Lucey et al.(1998b) LPS XXIX, Abstract#1356. 18 Workshop on New Views of the Moon II

LATERAL AND VERTICAL HETEROGENEITY OF Model: One interpretation fr the origin of the high-Th ma­ THOR IN mE PROCELLARUM KREP TERANE: terial is that subsurface KREEPy materials have been excavated by AS RFLECTED I TH EJECTA DEPOSITS OF POST­ impact craters [1). The materia excavated may be either volcaic IR CRTERS. J. J. Gillis and B. L. Jollif, Department KREEP (e.g., Apennine Bench Formation), KREEPy impact-melt of Ea ad Plaeta Sciences, Washington University, St. Louis breccia frmed by the Imbrium impact (e.g., Fra Mauro Fora­ MO 63130, USA ([email protected]). tion) [1], or other KREEP-rich crstal material. Determinng whch type of material is responsible fr the elevated Th and its extent is Introduction: The Procellam KREEPTerrae [I ,2J displays importat to understading the premare ad possibly the prebasin the highest concentrations of Th on the Moon [3]. However, loca­ statigraphy of the Imbrium-Procellam Region. tions of elevated Th in this region appea to be radom. As obsered Merging the 5° Th data with the shaded relief map, we obsere in the 5° pr pixel equal-aea Th data [3-J, ad made more evident that the highest Th concentrations are not related to pre-Imbrium in the preliminary 2° data [SJ, Th is enhaced around the craters upper crsta materials. The Apennines, Alpes, and Caucasus Moun­ Arstillus, Aristachus, Kepler, Mara, the Apennine Bench fra­ tans represent the pre-Imbria highlands material ad do not express tion, ad the Fra Mauro region, while noticeably ad uexpctedly concentatons of Th, FeO, ad TiO2 a high a the most Th-rich lower in other locations (e.g., Archimedes, Coprcus, Eratosthenes, materials expsed within the Procellarm KREEP Tere. We ob­ ad Plato). We have exained the compsition of the materials sere that, in general, these masifs contain 10-14 wt % FeO [1 OJ ad present in these regions with the goal of understading the patchy 4-7 ppm Th [2-3]. nature to the distibuton of Th and ultimately to decipher the geo­ Determining whether the Th signal is fom KREEP basalts or logic processes that have concentated the Th. KREEPy impact-melt breccias canot be done with the Clementne At present time, the published resolution of the Lua Prospctor daa bcause the two rock types ae compositionally and mineralogi­ Th y-ray data is low (5° pr pixel [3-4)), but this will soon be cally too simila (e.g., the Th-rch, mafc impact-melt breccia in the suprceded by signifcatly higher-resolution data (2 ° pr pixel [5)). Apollo saple collection ae dominated by a KREEP-basalt like Even at tis improved resolution, however, it is diffcult to resolve component [� 1 )). Mapping the distibution ad sizes of craters and the units that ae the maor source of Th. In a attempt to circumvent whether they display elevated Th concentations or not, should this problem, we employ the higher-resolution Clementine multi­ reveal the depth ad thickness of the KREEP-rch materias, ad spectal data fr those regions mentioned above. We use the UV­ whether they ae ubiquitous (i.e., impact-melt breccia) or more rb­ VIS-derived compositiona inforation and the spctal properties domly distbuted; ths might be tben a an indicator of localized of craters, and their ejecta a dill holes through the mae-basalt KREEP-basalt fows. surface to investigate the thickness ad composition of underlying Discussion and Conclusions: Within the southeates region material. With this infrmation we attempt to piece together the of the Imbrium basin, there ae two Th hot spots. The frst is asoci­ statigraphy and geologic histor of the Imbrium-Proellam re­ ated with the crater Aristillus, ad the latter with the Apennine Bench gion. Foration [12). Adjacent to these two hot spts ae craters with a Methps: We processed the fve-bad multispectral data fom lower Th signature: Archimedes ad Autolycus. We obsere in the the Clementine Mission (415, 750, 90, 950, ad 100 nm [6)) using ejecta of Aristillus, a region of signifcatly lower FeO (1014 wt%) ISIS sofwae ad calibration parameters developd by the USGS, relative to the suounding mae baalt (Table 1; Fig. 1). The crater Flagstaff, Arizona [7,8J. Final image mosaics are in equal-aea sinu­ Autolycus, 50 km to the south, did not excavate simila Iow-FeO soidal projection, ad have a resolution of 250 mpixel. Using the material. We suggest that the Iower-FeO material in the ejecta of method of [9] we produced maps of FeO ad TiO2 compositon. Here Aristillus coresponds to Th-rich material; the FeO content observed we examine the Th, FeO, ad TiO2 compsition ad spectal prop­ in Aristillus ejecta is comparable to that of KREEP basalt or mafc erties of the craters discussed above ad their ejecta, with the goal of impact melt breccia (10-12 wt% FeO). We determine that this Iow­ describing the materials they excavate (Tale 1). FeO, Th-rich material is volcanic KREEP, as opposed to Imbrium impact melt, on the bais that the low-Fe material is exposed more prominently in ejecta in the northes portion of Aristillus. Our asumption is that if the layer underlying Aristillus wa contnuous, TABLE l. The compositionsof theseareas were recorded utilizing a more widespread ad unifr low-Fe signature would be obsered Clementine derived FeO and TiO2 [7] and Th from Prospectory-ray data (2,3]. in the ejecta deposit. Archimedes, 110 km southwest of Aristillus, impacted te north­ Area of Interest Crater Th± 1 FeO± 1 Basalt es porton of the Apennine Bench prior to the erption of KREEP Diameter ppm wt% 1bickness basat [12]. Archimedes rim material is not as enriched in Th a the Apennine Bench, and tere ae diferences between the two in FeO Apennine Bench 7 10-14 0 concentation (Table 1) and in their continuum slop. Archimedes Archimedes 83 6 12-16 0 exhibits a much steepr or "redder" contnuum slop tha te Apnnine Aristillus 55 7 10-14 4.4km Bench. This steep slope suggests the presence of glassy material [13 J. Autolycus 35 6 14-16 >2.8 The glasy material is concentated aound a unnaed crater on the Copernicus 93 6 4-8 <7.4 southes rm of Archimedes (4.5°W, 28.2°N) and along the northes Eratosthenes 58 6 10-14 0 rim of Archimedes. We suggest two pssibilities, or a combination Kepler 32 8 14-18 ,3.2 km of the two, to explain the low-Th signal fom Archimedes: ( 1) The Mairan 40 7 12-15 0 Apennne Bench prior to KREEP basat eruption was lower in Th ( 4- Plato 101 5 9-13 0 7 ppm, e.g., simila to the Apnnine massifs) ad KREEP baalts ae LP! Contribution No. 980 19

The volume oflow-FeO material excavated by the impact was sub­ statial enough to mak the high-FeO signatre ofmare basalt ad the high-Th signal fom a subsurface KREEPy materia. On the other hand, the rim of Plato, like Archimedes, has a steep "reddish" con­ tinuum, indicating that it may be covered by a glasy, perhaps pyroclastic, material. Surounding Plato ae numerous rilles and volcanic vents, indicbing that plains-type volcanism occured in the area ad supporting the pssibility that these glassy materials ae pyroclastic deposits. Therefre, the low-Th compsition of the ma­ terials surounding Plato may be explaned by pyroclatic material covering the original basin rim materal ad concealing its inherent Th concentauon, or it may simply have struck a region oflower-Th content. The smallest crater in this study, Kepler, displays the highest concentation ofTh (Table 1). Kepler and its ejecta have a noticeably higher-FeO concentrauon (14-18 wt% FeO) compaed with other areas ofelevated Th. We note that the unit on which Kepler is situated has a lobate morhology. This obsesation suggests that volcanic fows, containing a high concentation ofTh, reached the suace in this area. These fows may represent KREEP basalt fows like those within the Apennine Bench Formation, or they may represent a higher-FeO variant. We fnd that the composiuonal data fom Aristillus, Autolycus, Archimedes, and Kepler fvor the prsence of KREEP baalt over "upper crsta," proximal ejecta Moreover, the Fra Mauro Forma­ tion represents "deep crustal," more distal Imbrium ejecta [14]. Furtherore, this techique of aalyzing the compsition of crater ejecta not only allows us to estimate the composition of subsurface materias but also yields mae basalt thicknesses. A more comprehen­ sive teatent of this is being prepared ad will result in a high­ resolution map of mae tickess. Acknowlegents: This work wa supprted by NASA grants NAGS-6784 (BU) ad NAGS-4172 (LAH). References: [1] Jollif B. L. et a. (1999) JGR, submitted. [2] Hakin et al. ( 1999) JGR, submitted. [3] Lawrence D. et al. ( 1998) Science, 281. [4] Gillis J. J. et al. (1999) LPS XX, Abstat #1699. [5] Lawrence D. et al. (1999) LPSC XXX, Abstract #2024. [6] Nozette S. et a. (1994) Science, 266. [7] Eliason E. M. (1997) LPS XVI, Abstrat #1198. [8] McEwen A. S. (1996) LPS XVI. [9] Lucey P. G. et al. (1998) JGR, 103. [10] Bussey D. B. J. ad Spudis P. D. (1998) in Workhop on New Views of the Moon, LPI. [11] Korotev R. L. (1999) JGR, submitted. [12] Hackman R. J. (1966) USGS Map 1-463. [13] Wells E. ad Hapke B. (1977) Sci­ Fig. 1. Clemenune-derived FeO fom orbit 168 (center longitude 1 °E). ence, 195. [14] Jolliff et al., this volume. Te cental peak crater is Aristllus, ad to te lower right is Autolycus. Arisullus has exposed a relatvely low-FeO unit and it is likely that tis mterial accounts fr the increased T detected in the Lunar Pospector TH OPTICAL MT OF EJCTA FROM LARGE y-ry sp. Exteme FeO values assoiatd with steep-slope inside crater RAYED CRATERS: PRELIMINARY RSULTS AND rim ae an artfct of albedo variauons on Sun-fcing ad shaded slopes. ILICATIONS. J. A. Grier1, A. S. McEwen1, P. G. Lucey2, M. Milazo1, ad R. G. Stom1, 1University of Arizona, Luna ad Plaetary Laborbor, 1629 E. University Boulevad, Tucson AZ absent in the rim ofArch imedes; or (2) the glasy (possibly pyroclas­ 85721, USA ([email protected]), 2Hawai'i Institute of tic) materal layering the rim ofthe Archimedes, dilutes any high-Th Geophysics adPlaetology, University ofHawai'i at Maoa, 2525 material present with low-Th material. Corea Road, Honolulu HI 96822, USA. The craters Copercus ad Plato display some ofthe lowest Th vaues fr craters within the Procellarum KREEP Terane (Table 1). Introduction: Shoemaker [1-2] agued tat several lines of The impact that produced Copercus was not only lage enough to evidence suggest a signifcat increae in the production of lage excavate through the mare basat but also through a KREEPy layer, craters on the Ea late in geological tme (-20-300 Ma). We have ifpresent, beneath the baalt to expose low-FeO materia in its ejecta. conducted a preliminay test of this hypthesis via a study of lage 20 Workhop on New Views of the Moon I

recent craters on the Moon, and have concluded that the lunarrecord no evidence for a changein the cratering rate since Tycho ( -109 Ma) does not supportthis idea. compared with the cratering rate since Copernicus (-810 Ma). The relatively recent ( <1 Ga) flux ofasteroids and comets form­ There are two additional results from this preliminary study that ing craters on the Earth and Moon may be accurately recorded by are relevantto lunarstratigraphy based on our analyses of Aristillus, lunar craters whose ejecta are optically immature, and therefore Autolycu,s and Lichtenberg Craters. We find optically mature ejecta distinguishable from the overall mature background soils. Lucey around the crater Lichtenberg (20-km diameter). Mare lavas are et al. [3-4] have developeda methodology for extracting anoptical em placed over the bright ejecta and raysofLichtenberg; this has been maturity (OMAn parameterfrom multispectralimages. We have cited as evidence for Copernican-age mare volcanism. Our data generated a preliminary dataset of average radial profiles of OMA T indicate that this superposition does not require anunusually young values aroundlarge craters (

et al. (1999) JGR, in prep. [11) Lucey P. G. et al. (1999) JGR. TABLE I. Terrain cameraspecification. [12) Horz F. (1985) in Lunar Bases and Space Activities of the 21st Mass (radiation unit) <10 kg (withMl) Century (. W. Mendell, ed.), LPL Power consumption 24W Focallength 72.5 mm Fnumber 4 THE LUNAR !MAGER/SPECTROMETER FOR THE Stereo angle ±150 SELENE MISSION. J. Haruyama1, H. Otake1, T. Matsunaga2, Detector type ID CCD (4098 pixels) andthe LISM WorkingGroup, 1 Advanced Mission ResearchCenter, Sensor pixel size 7 x7 µm Spatial resolution !Om National Space Development Agency ofJapan,Sengen 2-1-1 Tsukuba ° lbaraki 305-8505, Japan ([email protected]), Fieldof view 19.3 Swathwidth on ground 35 km(nominal mode) 2Department of Environmental Science and Technology, Tokyo 17.5 km(half mode) Institute of Technology, 4259 Nagatuta Midori-ku, Yokohama 40 km(full mode) Kanagawa 226-8502, Japan. B/H 0.57 SIN >100 Introduction: The Moon is the nearest celestial body to the Bandwidth 450--700 nm Earth.The investigation of the Moon gives clues asto the originand Quantization 10 bit evolution of not only the Moon itself, but also the Earth. ISAS MTF >0.2 @ Nyquist freq. (Institute of Space and Astronautical Science) and NASDA (Na­ Datacompression DCT method tionalSpace Development Agency of Japan)are coopera tively plan­ Pixelexposure tm 1.625 ms3.25 ms 6.5 ms ning to launch a Moon explorer namedSELENE (Selenological and DCTtable 32 patterns Engineering Explorer) in 2003.Phase C study has been started since Dataamount per day 50 Gbit Solarelevation angle in <40° lastApril 1999. SELENE will load 14 mission instrumentsto obtain TC operation scientificdata of the Moon. The scientificdata will be also used for investigatingthe possibilityof futureutilization of the Moon. SELENE will carry out the orbiter mission for one year and subsequently TABLE 2. Multibandimager specification. separate the propulsion module that will be a lander and send the radio signal forthe differentialVLBI fortwo months. The orbiter will Mass (radiationunit) <10 kg(with TC) end its role when the propulsion module is separated. Imagery and Power consumption <17W spectral profiling are most useful techniques for understanding the Focallength 65mm structure and materials of a planetarysurface. From the SELENE Fnumber 3.7 mission, we will obtain data with high spatialand spectralresolution Bandassignment 415,750,900,950,1000 nm(VIS) by LISM (Lunar Imager/SpectroMeter). LISM consists of three 1000,1050,1250,1550 nm(NIR) type sensors: a terrain camera (TC), a multiband imager (Ml), and a Detector 2D CCD(VIS) 2D lnGaAs detector(NIR) spectral profiler(SP). We introduceLISM in this paper. Sensor pixel size 13 x13 µm(VIS) Lunar Imager/Spectrometer Overview: LISM will obtain 40 x 40 µm(NIR) digitized global imaging and spectral profilingdata of the Moon with Spatial resolution 20 m(VIS) high-quality spatial resolution, wavelength resolution, andsignal-to­ 62 m(NIR) noise ratio. Data of LISM will be also combined and analyzedwith Fieldof view 110 the data of other mission instrumentsof SELENE. The LISM group Swath width on ground 19.3 km consists of about 20 researchers, who have responsibility on LISM Quantization 10 bit (VIS) development including science requirement decision, instrumenta­ 12 bit (NIR) tion, software development, and operation. The current status of Band width 20 nm(for 415, 750, 900,950 nm) LISM development is in phaseC, or PM (preflightmodel) designing 30 nm(for 1000, 1050, 1250 nm) 50 (1550 andmaking phase. From the next year, we will startdesigning FM nm nm) SIN >100 (flightmodel). MTF >0.2 @ Nyquist freq. Terrain Camera Overview: The main purpose ofTC is global Datacompression DPCMmethod stereo mapping with high spatial resolutions. To investigate sub­ Compressionrate to< 80% kilometer size morphologyfor all over the Moon in detail, we used Pixelexposure time 2.0, 4.1, 8.2 ms(VIS) images with spatial resolution of 10 m and height resolution of 20- 6.6, 13.2, 26.4 ms(NIR) 30 Dataamount per day 50 Gbit m. TC consists of two slant telescopes, each of which has a linear ° ° CCD detector, andtakes images by pushbroom scanning.The swath Solarelevation angle in 30 -90 of TC is 35 kmwhen thealtitude of SELENE orbiter is 100km, so Ml operation that images taken at serial paths are overlapped for across-track direction. DCT lossy compression method has been adopted to reduce the tremendousdata. The data will be nominally compressed to 30% or less. scopes,each of which has a two-dimensional detector with band-path Multiband Imager Overview: The main purpose of MI is filters. Data of several lines of the detectors, each of which corre­ global multiband mapping of spatial resolutionof 20 m forvisible spond to band assignments, are read out. MI takes images by range, and60 m for near-infrared range. MI consists of two tele- pushbroom scanning. The swath of MI is 20 km when the altitudeof 22 Workhop on New Views of the Moon I

TABLE 3. Spectral profiler specification. others have emphasized the role of secondary craters in producing rays [e.g., 1-3]. Pieters et al. [2] presented the results of a remote­ Mass (radiation unit) <8 kg sensing study of a portion of the ray system north of Copernicus. Power consumption 38 W (VIS, NIRl, 2) They provided evidence that the present brightness of the Copernicus Focal length llOmm Fnumber 4 rays in this sector is due largely to the presence of a component of Detector type Si pin photo diode (VIS) highlandejecta intimately mixed with local marebasalt, andthat an InGaAs (NIRl,2) increasing component of local material is observed in the rays at Sensor pixel size 50 x 500µm (VIS) 50 x 200µm progressively greater radial distances fromthe parent crater. These (NIRl,2) results have been questioned, and the origin of lunar rays is still Spatial resolution 500m uncertain [e.g., 4]. In an effort to better understand the processes Field of view per pixel 500m responsiblefor the formationoflunar rays, we have utilized a variety Numberof band 84 (VIS) ofremote-sensing data to study selected rays associated with Olbers A, 100(NIRl ) Lichtenberg, the Messier Crater Complex, and Tycho [ 6,9]. The data ll2 (NIR2) include near-IR reflectance spectra(0. 6-2.5 µm) and3 .8-cm and 70- Cooler peltiercooler (for NlR2) Quantization 16 bit cm radarmaps. In addition, Clementine UV-VIS images were used Wavelength resolution �nm to produce high-resolution FeO, TiO2, and maturity maps for the Data compression NIA variousrays using the methods presented by Lucey andet al. [ 10, 11]. Pixel exposure time selectable (TBD) Results and Discussion: Messier Crater Complex. Messier Data amount per day 0.7 Gbit (14 km in long dimension) and Messier A (diameter=11 km) are ° ° SIN >2500 @VIS located near2 s, 47 E in MareFecunditatis. Major rays fromthese ° ° Solarelevation angle 0 -90 craters occur to the south and west of the parentcraters. Spectrawere obtained forportions of the rays west andsouth of the crater complex, aswell as for MessierA andnearby mature mare regions. The spectrum of Messier A exhibits an extremely deep (29%) SELENE orbiter is 100km. We adopt a DPCM loss-less compres­ ferrousiron absorption bandcentered at 0.98 µm, and a freshmare sion method to avoid irreparable loss of informationabout subtle composition is indicated. Both the near-IR spectrum and the FeO differenceof reflectance. image clearly demonstrate that Messier A did not penetrate the Spectral Profiler Overview: The main purpose of SP is to Fecunditatis marefill. In addition, Messier A crater exhibits strong obtain detailed continuous spectral data of the Moon in the 500- returnson both the 3.8-cm and 70-cm depolarizedradar images [6-- 2600-nmrange to understanddistribution and chemical composition 8]. The spectrumcollected forthe ray west of Messier A has a 15% of minerals on the surface. SP has one set of optics with three absorption featurecentered at 0.99 µm. The ray has slightly enhanced detectors for three ranges (VIS: 500-1000nm, NIRl: 900-1700 values in the depolarized 3.8-cmradar image, but no enhancement is nm, and NIR2: 1700-2600nm). NIR2is cooled to 220 K by a peltier apparent in the 70-cm dataset. The mature mareunit adjacent to this cooler. The spatial resolution is 500m for across-trackand along­ ray has a spectrumwith a shallower band depth (12%) and a similar trackdirections. Since we require high accuracy to analyzeSP data, band center. The FeO image indicates that the west ray exhibits FeO we execute on-board calibration with using halogen lamps. values similar to adjacent maredeposits. The brightnessof the ray Lunar Imager/SpectrometerOp eration Overview: The spec­ west of Messier A is due to the presence of largeamounts of fresh tral profilerwill be operated in all daytime. The total amountof data mare basalt. Near-IR spectra as well as the FeO and maturity maps of TC andMI is so largethat TC andMI arenot operated at the same indicate that the ray south of the Messier complex is alsodominated time. Still, operations of TC andMI should be carried out in limited by freshmare material. daytime. To identify where SP is observing,however, highly com­ Tycho Ray in Mare Nectaris. A major ray from Tycho Crater pressed TC images or one-bandimages of MI areplanned to be taken crosses much of MareNectaris. The 80-km long ray segment north­ simultaneouslywith SP data east of Rosse Crater (17.9°N, 35°E) has a somewhat conical shape, ranging in width from 8 km near Rosse to 16 km northwest of Bohnenberger Crater. Spectrawere obtained forRosse Crater (diam­ THECOMPOSITION AND ORIGIN OF SELECTED LUNAR eter= 12 km), mature mareunits, and two small areason the Tycho CRATERRAYS. B. R. Hawke1, D. T. Blewett', P. G. Lucey1, ray northeastof Rosse. Both of the spots on the ray are located near C. A. Peterson 1, J. F. Bell 1112, B. A. Campbell3, andM. S. Robinson4, a Tycho secondary crater cluster -1400 km fromthe center of the 1Planetary Geosciences, Hawai'i Institute of Geophysics and parentcrater. The spectrumcollected fora mature marearea east of Planetology, University ofHawai' i, Honolulu HI 96822, USA, 2Center Rosse exhibits an 8.5% absorptionfeature centered at 0.98 µm. Both forRadiophysics and Space Research,Cornell University, Ithaca NY ray spectrahave 11.6% bands centered at-0.99 µm. It appearsthat 14853, USA, 3Center for Earth and Planetary Studies, National Air the ray in the areasfor which spectrawere collected aredominated andSpace Museum, Washington DC 20560, USA, 4Department of by fresh mare debris. These results are in agreement with those Geological Sciences, Northwestern University,Evanston IL 60208, presented by Campbellet al. [6]. These workers noted that the Tycho USA. secondary craters in the cluster areeasily seen in high-resolution 3.0- cm radarimages, anda radar-bright areaextends 10-15 km down­ Introduction: The nature and origin of lunar crater rays has rangeofTycho fromthe approximate centerof the cluster. In addition, long been a source of major controversy.Some lunarscientists have they noted that the radar-bright region exhibited a deeper "1-µm" proposed that rays are dominated by primary crater ejecta, while featurein multispectralratio images andsuggested that fragmental LPI Contribution No. 980 23

material wa emplaced well dowmage of the visible secondaries, [10] Lucey P. et al. (1995) Science, 268, 1150. [11] Lucey P. et al. perhaps by a secondary debris surge. The FeO ad maturity images (1999) JGR, submitted. support this interretation. In summa, analyses of near-IR refectance spectra, multispec­ tra imagery, ad a variety of rada data suggest that Tycho ray in Mae Nectas is dominated by fesh local material excavated ad LUNAR GRUITHUISEN AN MAIRN DOMS: RHEO­ emplaced by secondar craters. While some highlads material fom LOGY AN MODE OF EMLACEMNT. J. W. Head1 ad Tycho is undoubtedly present in the ray, the major fctor that pro­ L. Wilson2, 1Department of Geological Sciences, Brown University, duces the bri�huess of the ray is the immature mae baalt. Providence RI 02912,USA, 2Environmental Studies Division, Olbers A Ray. This Copercan-aged impact crater (diameter= Insttute ofEnvironmental and Naura Sciences, Lancater University, 43 km) located in the highlands on the Moon's westes limb (8.1° N, Lacaster LAl 4YQ, UK. 77.6 °W) exhibits a extensive ray system. Eight nea-IR refectance specta were obtained fr a prominent ray that extends northeast of Abstract: The luna steep-sided Gruithuisen ad Mairan domes Olbers A across Oceaus Procellaum. Three of these spcta were are morphologically ad spectaly distinctive stuctures that appa fr small (3- km in diameter) aea nea the intersection of two simila to terestial fatures chaacterized by viscous magma. We major ray elements -385 km northeat of Olbers A. Three specta use the baic morphologic ad morphometic chaacterstics of the were collected fr a prtion of the ray imediately northeast of domes to estmate their yield stength (-lO Pa), plastc viscosity Seleucus Crater. This ray segment is -550 km fom the rim of the (-109 Pa s), ad efusion rates (-50 m3/s). These values ae simila paent crater. Two specta were obtaned fr difse ray elements in to terestial rhyolites, dacites, and baaltic adesites ad support the the sae genera aea. Specta were also obtaned fr Olbers A Crater hypothesis that these domes ae a unusual vaiation of typical as well as both mature mae and fesh craters nea the ray. highlads ad mae compositions. Typical dike geometies ae pre­ All specta were anayzed ad spectal mxing model studies were dicted to have widths -50 m and lengths -15 km. The mgma rise conducted using the techniques described by Blewett et al. [5]. Three speed implied by this geomet is very low, -7 x 10-5 ms, ad the component mixing studies were perored using mature mae, fesh Reynolds number of the motion is -2 x 10-, implying a completely mae, ad either fesh or mature highlad material as end members. lamina fow regime. The specta obtained fr aea nea the ray intersection ae dominated Introduction: The Gruithuisen ad Maira domes represent by mare material. However, highlad debris is quite abundant (con­ examples of topgaphically, morhologically, ad spectally dis­ tibuting 30-50% of the fux to the spectra). Perhaps these high tinctve stuctures on the Moon [1-2] that appa to be cadidates fr values ae due.to the fct that two rays cross in the aea fr which the very viscous magma In this aalysis, we use the basic morphologic spectra were collected. Surprisingly, even lager amounts (35-55%) ad morphometric chaacteristics of the domes as a bais fr estimat­ of highlad material were fund to be present in portions of the ray ing of their yield stength, plastic viscosity, eruption rates, ad dike northeast of Seleucus. Lesser aounts (2638%) of highlad debris feeder geomet ( e.g., dike width ad length). We begin by asumng were determined to be present in the more diffse segments of the ray. that each dome is a single stcture emplaced on a fat plain ad use Maturity ad FeO maps produced fom Clementine UV - VIS images Blake's [3] treatent of domes modeled as Bingha platics to suggest that the ray is less mature tha the adjacent terain ad that establish the order of magnitude of the yield stength, (t), ad platc it contains a signifcat amount of highlads debris. viscosit (h). Infred values of tae-3 x lO Pa fr the Grithuisen Lichtenberg Crater Rays. Lichtenberg Crater (diaeter= 20 km) domes and 1.0 x 1O Pa fr Maira features. Next we use an empirical is located in Oceaus Procellam on the westes prtion of the luna frula given by Moore ad Ackeran [4] to relate the plastic neaside (31. 8°N, 67. 7°W). This Copoercan-aged impact stucture viscosity, to the yield stength. The most likely value of h fr the displays a relatively high-albedo ejecta blanket ad ray system to the Grithuisen domes is - 1 x 1010 Pa s and fr the Maira domes is north ad northwest. However, Lichtenbertg ejecta is embayed by perhaps within a fctor of two of 5 x 108 Pa s. As a specifc exaple, mae baalt south and southeast of the crater. The FeO map produced we teat the Gruithuisen Gaa dome as a underlying symetric fr the Lichtenberg region indicates that the ejecta ad rays north and dome of radius = 10 km and unknown thickness on which ae northwest of the crater exhibit relatively low FeO abundaces. These suprimpsed three fow lobes; we assume that these merge towad depsits appear to be dominated by low-FeO highlads debris. The the summit, where they have a common thickess. The values (t= 7. 7 maturity image demonstates that these highlads-rich ejecta deps­ x 104 Pa; h = 3.2 x 108 Pa s) ae, as expected, signifcatly less tha its ad rays are fully mature. Hence, the Lichtenberg rays exhibit a the frst approximations, which ignored the detailed morhology of relatively high abedo because of their compsition. These mature the dome. highlands-rich rays appa bright in compason to the ajacent Summar of Reologcal Parameters: Given the resoluton mature mae surfaces. These "compsitional" rays stad in stak of the images used and the consequent diffcultes in mapping ad contast to the imatity rays associated with the Messier Crater meauring fatures, it might be agued that measurements fr t ad Complex. h represent a single Bingham platc magma fr which t = (10 ± 3) References: [1] ShomaerE. (1962) in Physics andAstronomy x 1( Pa and h = ( 6 ± 4) x 108 Pa s. Altesatvely, we could sepaate of the Moon, p. 283. [2] Pieters C. et al. (1985) JGR, 90, 12393. the two goups of domes ad chaacterize the Grithuisen dome [3] Oberbeck V. (1971) Moon, 2, 263. [4] Schultz P. ad Gault D. magma by t= (12± 3) x 10 Pa, h= (10± 5) x 108 Pa s ad the Maira (1985)JGR, 90, 3701. [5] BlewettD. et a. (1995)JGR, 100, 16959. dome magma by the slightly smaller values t = (8 ± 4) x 10 Pa, h = [6] Capbell B. et a. (1992) Proc. LPS, Vol. 22, 259. [7] Zisk S. (5 ± 4) x 10s Pa s. et al. (1974) Moon, JO, 17. [8] Thompson T. (1987) Earth, Moon, Eruption Rte Estimate: We now t to the estimation of Planets, 37, 59. [9] Hawke B. et al. (1996) LPS XXVII, 507. the erption rates of the magmas frming the vaious dome units. To 24 Workshop on New Views of the Moon I

do this we assume that the advance of the flowfront of each flowunit A MULTISPECTRAL ANALYSIS OF THE FLAMSTEED or dome-forming episode was limited by cooling. Pinkerton and REGION OF OCEANUS PROCELLARUM. D. J. Heather', 3 Wilson [5] show that a varietyof types of lava flow cease to move S. K. Dunkin 1, P. D. Spudis2, and D. B. J. Bussey , 1 Department of when the Gratz number, a dimensionless measure of the depth of Physics and Astronomy, University College London, Gower Street, penetration into the flowof the cooled boundarylayer, has decreased London WCIE 6BT, UK ([email protected]),2 Lunar and Planetary fromits initially very high value to a critical value of -300. Blake's Institute, 3600 Bay Area Boulevard, Houston TX, 77058 USA, model of Binghamplastic domes shows the way in which the radius 3EuropeanSpace Research andTechnology Centre,European Space of the dome grows as a function of time. The assumptionthat flow Agency, Noordwijk, The Netherlands. lobes are cooling-limited always leads to effusion-rate estimates of lower bounds. This is because any flow unit assumed to be cooling­ Introduction: According to Pieters et al. (1), the Flamsteed limited may in fact have been volume-limited (i.e., ceased to flow areaof OceanusProcell arumis representative of basalts that have yet simply because the magma supply was exhausted), implying that it to be sampled. They studied the area in detail using telescopic data had the potential to travel further than the observed distance and to identify seven distinct mare flows. This diversity makes the therefore had a higher effusionrate than that deduced. For each of the Flamsteed region an ideal candidate for Clementine multispectral Gruithuisendomes we have either a second smaller dome or a flow studies. The region studied here is far smaller thanthat covered by lobe superimposed on a larger, earlierdome. It is very tempting to [I], but the higher spatialresolution of the Clementine data will allow assume that the second unit in each case is a breakout at the vent us to make a freshinterpretation of the nature of our restricted area caused by the magma supply continuing after the first unit has beforeexpanding to encompass the surrounding regions. The pri­ reached its cooling-limited length. This automatically implies that maryaim of this work is to use Clementine UV-VIS data to analyze the effusion ratesdeduced fromthe large dome geometries( - 48, 24, flowson a smaller scale anddetermine the stratigraphyof the mare, and119 m3/s forGruithuisen Delta, Northwest, and Gamma,respec­ using impact craters as probes to measure the thicknessof mare lavas tively) are the realistic estimates and that the rates foundfrom the wherever possible. smaller units are underestimates. For the Mairan domes we arenot Data Reduction: We used the Clementine UV-VIS data to able to resolve multiple lobe structuresand must regardthe effusion produce a multispectralimage of the Flamsteed areafrom 0.60°N to rates of - 50 m3/s as lower limits on the truerates. 16.06°S and 308.34°E to 317.12°E. The data were processed at a Geometries ofFeeder Dikes: To explore the conditionsunder resolutionof200m/ pixel using the ISIS softwareprogram (available which the magmas erupted, we assume that in each casethe melt through the USGS), and the photometric coefficients tabulated in reached the surface from a source region at the base of the an­ Blewett et al. [2]. In addition to the multispectral image, a "true orthositic crust. In order to keep a dike filled with buoyantmagma color" image, FeO map using the algorithms of [2], and a TiO2 map openfrom this depth, we require a driving pressure of -33 MPa. Any using the algorithms of [3] were generated. In conjunction with a driving pressure smaller than this value wouldallow the lower end of 750-nm Clementine mosaic and Lunar Orbiter photographs, these the diketo close andand so cause it to pinch shut from the magma images formed thedataset used forthis analysis. For more details on source region. Thus the absolute pressure in the source region is at the data-reduction procedure used, please contact the authors. least -490 MPa. The typical width of the dike held openby the above Mapping the Flows: The areastudied here lies in the south­ driving pressure is expected to be within a factor of 2 or 3 of -80 m. easternportion of Oceanus Procellarum,and covers approximately We now calculate the rise speed of magma driven by the source 134,500 km2, extending from the mare-highland boundary (to the pressure through a dike of mean width and horizontal length. Using south) up to and including the Flamsteed P ring. The number of the mean rheological properties for all ofthe domes, t = 10 x 104Pa spectrally distinct flows in the area is striking in the Clementine and h= 6 x 108 Pas, together with aneruption rate of 50 m3/s (typical mosaics, ranging fromhigh-Ti flowsin and around FlamsteedP to of the majority of the values), we findW = 50 mand L= 15 km. The low-Ti flows atthe edges of the mare-highlandboundaries. From a magma rise speedimplied by this geometry is very low, -7 x 1Q-5 preliminary analysis, we have identified at least five flows in the mis,and the Reynolds number of the motion is - 2 x lQ-8,implying multispectral image alone(Fig. 1). Sunshine and Pieters [4] found a completely laminar flow regime. Considering the uncertainties in three distinct flowswithin the Flamsteed Pring using high resolution manyof the parametersinvolved, a dike width of -50 mis certainly CCD images from a groundbased telescope. We find evidence for consistent with the estimate, obtained fromconsideration of the static only two: a younger high-Ti flow (flow A in Fig. 1) overlying an stability of the dike, that the width should be within a factor of 3 of older lower-Ti flow (flow C in Fig. 1). However, we have not yet -80 m. In addition, a fissurelength of -15 km is consistent with the reduced the data forthe most easternpart of the ring,and it is possible Delta dome being elongate with its main vents -12 km apart, andthe that further flow(s) couldbe found in our missing section. fact that the Gamma and Northwest domes together could be re­ The low-Ti flowsat some of the mare highland contacts to the gardedas defining an underlying fissureabout 18 km long. It is less south are exceptionally bright in the 750/415 nm channel of the clearwhether the Mairandomes should be regardedas being fedby multispectral image (flow E in Fig. 1). These areascorrelate with three separate dikes or in some combination by two dikes. intermediate FeO and TiO2 content (the TiO2 map is shown in References: [1] Head J. W. and McCord T. B. (1978)Science, Fig. 2), and seem to be the oldest flows visible on the surface, 199, 1433-1436. [2] Chevrel S. eta!. (1999) JGR, in press. [3] Blake probably extending overa large area beneaththe later flows. S. (1990) in LavaFlows and Domes: Emplacement Mechanisms and The extent of each flow so far identified is shown in Fig. 1. Hazard Implications, IA VCEI Proceedingsin Volcanology,vol. 2. Boundaries were defined according to multispectral and albedo (J. Fink, ed.), pp. 88-128, Springer Verlag, Berlin. [4] Moore H.J. properties. Detailedstudies of the TiO2 map (Fig. 2) and maturity andAckerman J. A. (1989) NASA Tech. Memo TM-4130,387-389. data (taken through observationsof crater densitiesfrom the Orbiter [5] PinkertonH. andWilson L. (1994) Bull. Volcan., 56, 108. frames andan optical maturityimage produced using the algorithm LP/ Contribution No. 980 25

I.OS

e 1.00

�a: 0.9S >, § 0.90 -nejecta _nwaU 0. i 85 -sejccta "2 -swal 'i;.. 0.80 i 0.15

-l, 0.70

.WO 450 S00 S50 600 650 700 7S0 800 850 900 950 1000 1050

Wavelength (nm)

Fig. 3. A 5-point noralized refectance plot of the crater boxed in Fig. 2. The wall and ejecta to the south show a fr stonger mafic absorpuon band than the northes units, which are probably a mix of excavated highland and basaltc material. Te 80-m depth of the crater terefre gives an upper limit fr the mare thickess in the area.

of Lucey et a. [5]), will improve this map. Work is continuing in c attempt to delineate cleaer fow boundaies. Fi. 1. Outline mp of the fows in the Ramsteed aea. H = Highlad Stratigraphy of the Region: The prima aim of this work is material; A = youg, high-Ti fow; B = high-Ti fow, perhaps linked to to determine te tickness of the mae fows as one moves out fom flow A; C = older low-Ti fow; D = interediate-Ti fow; E = low-Ti the highand boundary into Ocecus Pocellarum. First-order indica­ fow, probably the oldest in the area, ver bright in te red (750/415 nm) tions of thickess cc be obtained by searchng fr highland outcrops channel; Ha = Mons Hansteen. (To view this figure in color go to witn the mara The Flamsteed aea shows mcy such outcrops (see http://cass.jsc.nasa.gov/meeungs/moon99/pdf8032.pdf.) Fig. 1), cd the lava must be quite thin close to these. A more asolut idea of basalt thickness can be obtained by calculatng the depths of craters that have dug through the lavas to expse hghland material below. These craters can be identifed fom multspectal images cd 5-pint specta Previous work has sug­ gested thb a cyc color in the multispectal fbe represents hgh­ land materal [ 6-7), cd that yellows cd geens ae feshly excavated baalts [8]. However, we have recently fund that a cyc color cc also result fom a feshly excavated high-Ti basalt. In order to diffrentiate btween the hgh-Ti cd hghland signature, it is neces­ sar to look at the FeO and TiO2 fames and plot 5-point specta to look fr the absorpton at 0.95 mm that is chaacteristc of pyroxenes in the basats. These obserations have shown there to be candidate craters in the Flamsteed region which have excavated highlcd material, c exbple of whch is shown in Fig. 3. The example crater displays a basaltic signature with aclea0.95µm absorption in its south wall and ejecta, whle the absorption in the north wall and ejecta is fa weaker. The norther deposits ae also relatively low in TiO2 cd FeO, cd probably represent a mix ofbaatic cd hghlcd material. The crater is 8 km in diameter, so will have excavated to a depth of -80 m (using the depth:dibeter rato of 1:10 given by Croft [9]); this is therefre c uppr limit to the thickness of the basalts at the crater's norther edge. In additon, there are several aea where craters close together excavate spctally distinct mberias. Tese may indicate boundaries of subsurface mare fows, cd will allow for a more detailed stati­ Fig. 2. A grayscale Clementne TiO2 map of the Flamsteed region, clealy showing the distnct fow uniu of bth high-and low-Ti content graphic picture to be constucted. (lighter a have a high TiOi content compaed to the darker regons). Future Work: We intend to map te lava fow cd crater 5-point specta of te bxed crater, which may have excavated through distibution across the Flamsteed region, using craters to deduce te entre mae sequence, ae shown in Fig. 3. depths to the highland-mare contact where pssible. Flamsteed will 26 Workhop on New Views of the Moon I

then be combined with adjoining areas of Oceanus Procellarum, A. In this model, the parent magmas to the magnesian-suite gradually developing a complete picture of the stratigraphy and troctolite are impact melts formedby meltingthe ferroananorthosite basalt thickness across the basin. This work will form part of a crust andthe early-most cumulates of the magma ocean. The mafic continuing project in which we aim to study maria across the whole cumulates exist below the crust through the agency of cumulate Moon, providing a global perspectiveof lunarvolcanic history. overturnas the original cumulate pile establishes gravitational equi­ Acknowledgments: We would like to acknowledge the con­ librium [ I 0]. A magma ocean withMg* 84-88 is required to produce tinuing supportof the USGS in the reduction of Clementine data. The olivines withMg* 92-95 [3]. Such highMg* values require that the ISIS software (distributedby the USGS) was run on Starlinkfacili­ magma oceanexperienced fractionalrather thanequilibrium crystal­ ties, funded by the Particle Physics andAstronomy Research Council lization even at the earliest periods of crystallization. The high and (PPARC) in the UK. This work wascarried out while SKD was a variable REE content is explained by melting of variable amounts of PPARC ResearchFellow. the KREEP layer that survived the overturn. The mafic cumulate References: [I] Pieters C. M. (1980) JGR, 85, 3913-3938. cannot be solely of olivine, however, since the magnesian norites [2] Blewett et al. (1997) JGR, 102, 16319-16325. [3] Lucey et al. require an orthopyroxene-bearing source. An orthopyroxene-bear­ (1998) JGR, 103, 3679-3699. [4] Sunshine J.M. and Pieters C. M. ing dunite or even anolivine-rich harzburgite is required. This would (1990) LP/Tech. Rpt. No. 91-03, 59-60. [5] Lucey et al.(1998) LPS imply that the magma oceanwas less olivine normative than typical XXIX, Abstract#1356. [6] Gillis andSpudis (1995) LPS XXVI, 459- terrestrialmantle. 460. [7] Bussey and Spudis (1996) LPS XXVI, 183-184 [8] Dunkin The impact model nicely reconciles constraintsI to IV, whereas andHeather(l999)LPSXXX, Abstract#l l80. [9] Croft S. K. (1980) V simply requires multiple impacts. Indeed, since magnesian-suite Proc. LPSC llth, 2347-2378. rocks were collected from Apollo 11, 12, 14, 15, 16, and 17 [11], impact melting in this model was a common and widespread event. If this model is correct,why are the ferroananorthosite and magne­ PETROGENESIS OF MAGNESIAN-SUITETROCTOLITES siansuite plutonic trendsso distinct?Why isn't there a continuum of AND NORITES. P. C. Hess andE. M. Parmentier,Department of compositions? It is also noteworthy that upper mantle samples are Geological Sciences, Brown University, Providence RI 02912, USA. rareto nonexistent in the lunarcollection [I]. In fact, neither ferroan anorthosites or cumulate mantle samples aretypically foundwithin Models for the petrogenesis of magnesian-suite troctolites and the ejecta deposits that include magnesian-suite samples [12]. norites areseverely constrainedby a number of importantgeochemi­ B. In the cumulate remelting model, the highMg* valuesof the cal features. magnesiantroctolite arecontributed also by a source dominated by I.Most troctoliteshave olivines withMg* values that rangefrom early olivine cumulates of the magma ocean. The normative plagio­ 84 to 92 [I]. Rarespinel troctoliteshave extremely primitive olivines clase andorthopyroxene componentsare supplied by sinking parcels with Mg* 95-96 [2]. Hess [3] showed that a peridotite source with of quenched crustof the convecting magma ocean [4] or perhaps by Mg*>91 was required to produce the most magnesian troctolite interaction with partial melts generated by melting the primitive parent liquids known at that time. But these parent liquids were interior of theMoon [13], if indeed such an interior exists. In either stronglyundersaturated with respectto plagioclaseat the low pres­ case, the maficcumulates of olivine, perhaps with small amounts of sures expected of the lunar anorthosite crust. Such liquids must orthopyroxene (see earlier discussion), must dominate the source. crystallize significantquantities of olivine in order to bring plagio­ Pressure-release melting associated with the cumulate overturncre­ clase onto the liquidus; the fractionationof olivine must reduce the ates the parentmagmas forthe magnesian suite. Sufficientmelting is Mg* value of the plagioclase-saturated liquids. It follows that required to leave only orthopyroxene and olivine in the source to troctoliteswith Mg* of 90-92 cannot be produced in this way. The maintain low normativediopside contents. problem is far worse for thespine! troctolite withMg*96. This model satisfiesconstraints I, II, andIV but hasdifficulty with II. The crystallization sequence of magnesian suite magmas is III, the high KREEPcontents of theMg-suite parent liquids. Possibly non-tholeiitic; i.e., the crystallization of low-CaO pyroxene precedes the high KREEP contents are the result of assimilation of high level high-Cao pyroxene [4]. Indeed, troctolitesand norites arenearly free KREEP deposits.However, if KREEP is akin to ferroanbasalt then of discrete CaO-rich pyroxene [I], a featureimplying a melt and a the assimilation of this mass would lower Mg* values. If KREEP is source with low normative diopside, certainly much lower than in quartz normative and granitic, assimilation would compromise the typical terrestrialIherzolite. olivine-normative character of the parent liquids; KREEP-rich m. The parent magmas for magnesiannorites and troctolitehave troctoliteparent liquids would evolve to norite parentliquids. This calculated REE contents typically between I and2x high-K KREEP latter viewpoint is probablythe more acceptable one. [5-6] or about 300-600x chondrite! If this model is correct, then convective overturnand pressure IV. The parent magmasto the most primitive troctoliteare hot and release meltingmust not only begin promptly to create old Magne­ have temperatures that exceed 1270°C at I bar [7]. sian-suite rocks but must also continue for 200-300more m.y. to V. Radiogenic isotope ages for magnesian norites imply that account for the youngest members of theMagnesian-suite [1]. The magma generation waslong lived, contemporary with the formation rapid formation of the cumulate pile is consistent with the 182W­ of ferroananorthosite and extending to( andpossibly beyond)4.2 AE isotopic anomaliesthat require that heterogeneous anddiscrete source [1,8]. regions must be established in the first50 m.y. of the origin of the Three models of petrogenesisare tested against the petrochemical Moon [14]. Perhaps thermal rather then chemcially driven convec­ features listed above:These models include (1) impact melting, (2) tion is needed to explain the originof young Magnesian norites. cumulate remelting, and(3) assimilationof cumulates by late-stage C. Korotev [15] has proposedthat Apollo15 norite andthe entire liquids of the magma ocean. Magnesian-suiteis the product of assimilation of olivine cumulate LP/ Contribution No. 980 27

mantle by the late stage KREEP liquid residual to the magma ocea. Oceanus Procellam in order to include the lagest mae baalt­ This would account fr the KEEP paadox, i.e., the combination of covered aea into our investigation. high concentrations of incompatible elements and more refactor Introduction: It is known that lunar mae basalts cover about major elements as represented, fr example by high Mg* values. This 17% of the luna surface, but represent only 1% of the volume of the model ha a number of prblems, however: The KREEP liquid is luna crust [2]. A signifcat portion of luna mae basalts ae ex­ produced by the cooling ad factional crystallization of the magma posed within Oceaus Procellam, ad for large aeas absolute age ocea. The residual liquid is a nea-solidus phase ad must exist at data ae still lacking. From geologic mapping we know that there are T <1100°C. In order to produce the paent liquids to the magnesia very young basalts aound crater Lichtenberg, but how old are they suite by assimilaton, the KEEP liquid must somehow be reheated really? How do Oceanus Procellam baalts compae to basalts in to temperatures of 1270°C or higher. How this is done is certainly the previously investigated basins? Whitford-Stak ad Head [3] problematic. Secondly, this residual liquid has a evolved major­ defined 21 basalt types in Oceaus Procellam, and these basats element compsiton cont to the refactory major-element cha­ show ages fom 3.8 Ga to 2.4 Ga [4]. Our approach is to use acter of toctolites ad nortes. Moreover, this liquid will be diop­ previously published maps [e.g. 3,5] as a basis to defne homoge­ side-rich(noratively) conta to constaint II. If this magma body neous basalt units on Clementine color-ratio mosaics ad to prform exists beneath the "Great Luna Hot Spt" at -3.9 AE a required to new crater counts for these basalts. This is simila to the approch we produce Apollo 15 KEEP basalts, it is not clea how the gravity took fr our previous study [6,7]. aomalie•s related to the Imbrium mascon (ad others) can be main­ Reults and Conclusions: On the basis of aalysis of easter tained. A soft mantle would rela ad eliminate the gravity aomaly. ad cent neaside mae bains we addressed the fllowing ques­ Finally, if the magma ocea wa in existence at the tme of the tions. Imbrium impact, how did subsequent mare basats readily pass 1. How long was lunar mare volcanism active? Our data show through this zone ad erpt without KEEP assimilation? that luna volcaism was active fr at least 2.0 b.y. According to our Conclusions: Models A ad Bae compaable and viable. Both data, volcaism started at about 4.0 Ga ad ceased at -2.0 Ga. sufer fom the KEEP paradox, however: How ae large masses of However, numerous dark hao craters in the Austae region suggest KEEP incorrated into the toctolite paent magma without com­ the presence of a crptomae. We conclude that volcanism must have promising the highMg* values? Several importat conclusions nev­ been already active, at leat in the Nectaria Period. Schultz ad ertheless derive fom this aalysis: (1) cumulate over must bring Spudis [8] suggested that luna mae volcaism mb have stated a olivine-rich cumulates to the upper matle; (2) the magma ocean ealy as 4.3 Ga. The baalts that were dated in ou previous study were exprenced factional rather tan equilibrium crystallizaton through­ erpted primaly in the lmbria Period fom 3 .4 to 3.8 Ga. We fund out most of its history; ad (3) it is unlikely that lage volumes of only a minor number of basalts that erpted during the Eatosthenia primitive Moon exist- tholeiite baalts do not exist on the Moon! Period. In addition, there wa no evidence fr basalts ofCoperca Alteratively, the moon is depleted in normative diopside, and (4) age in the six investigated impact basins, but such basalts ae exposed olivines with Mg* > 90 require a bulk Moon with Mg* > 84. in the Oceaus Pocellaum [9]. In Oceaus Procellam these basats References: [1] Papike J. J. et al. (1998) Rev. Mineral., 36, embay the Coperca crater Lichtenberg, ad it is thought that these 5189. [2] Snyder G. A. et al. (1999) LPS XXX. [3] Hess P.C. (1994) basalts ae the very youngest basats on the luna surface with a age JGR, 99, 19083-19093. [4] Hess P.C.(1998)LPSXIX. [5] Papike a low a probably <1 Ga [9]. In order to aswer the queston over J.J. et al.( 1996) GCA,60, 3967-3978. [ 6) SheraisJ. W. adMcGee what time spa luna volcanism was active, the absolute ages of the J.J.(1998) GCA, 56, 3809-3823. [7]Morse S. A. (1980)Basaltsand the Lichtenberg basalts have to be known. Therefre we will meaure Phase Diagrams, Springer Verlag, p. 162. [8] Nyquist L. E. ad Shih the crater size-fequency distbuton of these basalts as well as other C-Y. (1992) GCA, 56, 2213-2234. [10) Hess P. C. ad Parentier baalts north of Lichtenberg that were mapped asCopercan basats E. M. (1995) EPSL, 134, 505-514. (11) Waen P. H. (1993) Am. by Wilhelms [9]. Geologic mapping (9) also reveaed that signifcat Mineral., 75, 360-376. (12) Ryder G. et al. (1997) GCA, 61, 1083- porions of Oceaus Proellam ae covered wit Erathosthenia 1105. [13) Sherais J. W. and McGee J. J. (1999) JGR, 104, 5891- basalts. From these maps it seems that Eratosthenian volcaism was 5920. [14] Lee D. C. et al. (1997) Science, 278, 1098-1103. more widespread in Oceanus Procellarm tha in other luna mae [15] Korotev R. L.(1999) LPS XXX. basalt regions besides Imbrium. We will determine absolute ages fr Oceaus Procellam basalts in ordr to investigate variations in the actvity of mae volcanism with time. AGES OF OCEANUS PROCELLARUM BASALTS AN 2. Whatis the spatial distribution ofages? Soderblom et al. [10] OTHR NEARSIE MARE BASALTS. H. Riesinger and reported diferences in age, chemisty, ad magnetsm between J. W. Head II, Depaent of Geological Sciences, Brown University, baalts of easter ad wester basins. According to this work, easter Providence RI 02912, USA ([email protected]) basalts ae generally older, more magnetized, and less rbioactve tha their wester counterpars. Results fom Lunar Prospctor [11] Abstract: In a previous papr [1] we reported on crater size­ show high concentatons of T ad K in and aound the neaside fequency distibuton daa fr 139 spctaly ad morphologically wester maria thus being consistent with Soderblom' s obseratons defined basalt units in six nearside impact basins (Australe, concerng the amount of radioative elements. So far we invest­ Tranquillitatis, Humboldtianum, Humorum, Serenitatis, and gated basalts in six lunar impact basins that ae distbuted over the Imbrium). We also presented results on the relationship between Ti entire luna nearside, therefre alowing us to study large-scale concentaton ad age of a basalt unit, the infuence of crustal thick­ diferences in age ad geochemisty between easter and wester ness on the erpton of a basat onto the luna surface, ad the fux basalts. Baed on this investigaon, we fund tha volcaism was of luna mare volcanism. In this papr we expand our study to actve longer in the wester pats of the neaside tha in the easter 28 Workhop on New Vews of the Moon I

regions. However, this trend is not very pronounced in our data and a consequence, our data indicate that the youngest basaltsare often westernmostparts, i.e., OceanusProcellarum are still missing. Geo­ exposed in or near areas with the relatively thinnest crust. On the logic maps [9] show that large areaswithin Oceanus Procellarum are basis of these observations,we conclude that in areaswith a thinner covered with Eratosthenianbasalts. These Eratosthenian basalts are crust,dikes still could reach the surface even later in lunar history, exposed in the centralparts of Oceanus Procellarum, and Imbrium whereas in other regions the dikes stalled in the crust andcould not basalts arelocated closer to the highland/mare boundaries. We will propagate to the surface. In Oceanus Procellarum we will investigate test for this distribution and provide new age data for the Oceanus the correlationbetween crustalthickness andcorresponding age in Procellarumbasalts. order to test for if the youngest basaltsare exposedover regions with 3. How does the Ti content var with time? For basalts in the six systematically thinner crust and if our previous results also hold for investigated basins,we generally see that TiO2-rich basalts tend to be the basalts in this area. older thanTi O r poorbasalts. Our results areconsistent with the trend References: [1] Hiesinger H. et al. (1999) LPS XX, Abstract found in the sample collection that indicate that radiometric older #1199. [2) Head J. W. III (1976) Rev. Geophys. Space Phys., 14, basalts aregenerally richer in Ti than younger basalts. However, a 265-300.[3] Whitford-StarkJ. L. and Head J. W. III ( 1980) JGR, 85, very important result of our previous investigation is that the TiO2 6579-6609.[4] Boyce J.M. andJohnson D. A. (1978) Proc. LPSC contents andages of the basalts within each investigated basin arenot 9th, 3275-3283. [5] Pieters C. M. (1978) Proc. LPSC 9th, 2825- correlated. Our data indicate that TiOrrich andTiO 2-poor basalts 2849. [6] Riesinger H. et al. (1998) LPS XIX, Abstract #1242. canerupt simultaneously in differentlocations within the basin. In [7] Riesinger H. et al. (1998) LPS XIX, Abstract#1243. (8) Schulz each single basin, the TiO2 concentrations canvary independently P.H. and Spudis P. D. (1983) Nature, 302, 233-236. [9] Wilhelms fromthe ages of the units. This probably suggests a highly heteroge­ D. E. (1987) U.S. Geologi cal Survey Spec. Paper 1348. neous mantleor source region. With the recent study we wish to test [10) Soderblom L. A. et al. (1977) Proc. LSC 8th, 1191-1199. if this observationalso holds forOceanus Procellarum basalts. Spec­ [11) Lawrence D. J. et al. (1998) Science, 281, 1484-1489. tral mapping of basalt types [5] indicates that for large parts of [12] GiffordA. W. andEl-BazF.(1981)Moonand Planets, 24, 391- Oceanus Procellarum younger basalts are more Ti rich than older 398. [13) Zuber M. T. et al. (1994) Science, 266, 1839-1843. basalts, thus somewhat reversing the trend found in the returned (14) YingstR. A. andHeadJ.W. III (1997)JGR, 102, 10909-10931. samples.Because of the immense importance of this observation for (15) Head J. W. III and Wilson L. (1992) GCA, 56, 2155-2175. the development of petrogenetic models, absolute age determina­ tions arenecessary forthese basalts. 4. How much basalt was erupted with time? Once spectrally HIGH-RESOLUTION MAPPING OF LUNAR CRUSTAL homogeneous units are mapped, their areal extent can easily be MAGNETIC FIELDS: CORRELATIONS WITH ALBEDO · measured. In combination with reasonable assumptions [e.g., 12] of MARKINGSOFTHEREINERGAMMA CLASS. L. L.Hood 1, 3 the thicknessof a basalt,we areable to estimate the erupted basalt A. Yingst1, D. L. Mitchell2, R. P. Lin2, M. Acuna , and A. Binder4, volume. In our previous study we assumedthat a single basalt flow 1 Lunarand Planetary Laboratory,University of Arizona, Tucson AZ unit is -10 m thick. For the basaltsin the six investigated basins,we 85721-0092, USA ([email protected]), 2Space Sciences see that the main volcanicactivity occured in the Imbrian Period and Laboratory, University of California-Berkeley,Berkeley CA 94720- was at least larger by a factor of 3.5 between 3.4 and 3.7 b.y. 7450, USA, 3Goddard Space Flight Center, GreenbeltMD 20771, comparedto the periodbetween 2.9 and 3.4 b.y. We conclude that USA, 4LunarResearch Institute, Gilroy CA 95020, USA. mare volcanismwas not equaly active over long periods of time but peaked in the ImbrianPeriod. As mentioned above, there are large During the last eight months of the Lunar Prospector mission areas in Oceanus Procellarum that were previously mapped as (December 1999-July 1999), the spacecraft was placed in a rela­ Eratosthenianin age (9). We will investigate the influence of these tively low-altitude (15-30-km perapsis), near-polar orbit that al­ basalts on the distribution of active mare volcanism over time. Is lowed high-resolution mapping of crustal magnetic fields. We report there a second peakof volcanic activity in the EratosthenianPeriod, here initial studies of the correlation of locally strong magnetic or arethe Eratosthenian basaltsof OceanusProcellarum part of the anomalies with unusual, swirl-like albedo markings of the Reiner decline of volcanic activity observed in the other areas? Gamma class. Based on thiscorrelation, which is knownfrom earlier 5. How does crstal thickess infuence basalt eruptions? Besides studies of Apollo subsatellite magnetometer data [1], it has been high-resolution color imaging data that are used forthe definitionof proposed that the swirls represent regions whose higher albedoshave spectrally homogeneous units, Clementine provided us with esti­ been preservedvia deflectionof the solar-windion bombardmentby mates of the crustal thickness of most of the Moon [13). In our strongcrustal fields (2). This model in turndepends on the hypoth­ previous study, we used this datasetand found that the maximum esis that solar-wind implantedHi s at least one component of the thickness of thecrust where basalts occur is 50-60km. Our results process thatoptically matures exposed silicate surfaces in the inner areconsistent with the results of Yingst and Head [ 14] and imply that solar system [3]. Specifically,it is hypothesized that implantedHacts crustalthickness is a limiting factorfor the eruptionof lunarbasalts as an effectivereducing agent to enhancethe rate of production of onto the surface. We see that ascending magmas preferentially ex­ nanophase metallic Fe particles from preexisting silicates during trudeto the surface where the crustis thin, i.e., the lunarnearside or micrometeoroid impacts. According to the model, the curvilinear the impact basins. On the farside, the crustal thicknessis larger and shapes of these albedo markingsare caused, at least in part, by the therefore the magmas stall and cool in dikes before they reach the geometryof ion deflectionsin a magneticfield [4]. surface. Thickening of the crustwith a related compressional stress The improved resolution and coverage of the Prospector data fieldwithin the crust or sinking of the buoyancytrap over time makes allow more detailed mapping of the fields, especiallyon the lunar it more and more difficultfor dikes to extrudeto the surface (15). As farside. Thispermits a more quantitativetest of whetherall albedo LI Contribution No. 980 29

makings of this clas ae associated with stong local magnetic Procellarm (location: 58.5°W, 7.5°N). The-contour interva is 3 nT felds. Ony if the latter condition is met can the sola-wind defection ad the mean spacecraft altitude is 18 km to within the accuracy hypothesis be valid. allowed by the resolution of the map (30 km or -1°); stong magnetic Field Mapping: The basic procedure fr mapping crstal mag­ aomalies corelate closely with swirl locations. Individual orbit netic felds using Luna Prospector magnetometer data fllows that profles (whose resolution along the orbit tack is compaable to the developed fr analysis of Apllo subsatellite magnetometer data [5]. spacecrat altitude of 18 km) also demonstate a good corelation of The specifc mapping steps ae (1) selection of mission time interas field magnitude with surface albedo. suitable fr mapping crustal felds; these ae limted essentially either In order to investigate the corelation of magnetic felds with the to times when the Moon is in a lobe of the geomagnetc tail or to times location of swirl fatures, we have reexamined available lunar imag­ when the Moon is in the solar wind but the spacecraft is in the luna ery (Luna Orbiter, Apllo, ad Clementne) to identif and map wake; the data ae tansfrmed to a radial, eat, and north coordinate swirl locatons within regions where swirls have previously been system with meaurements given as a fncton of spacecraft latitude, mapped [ 4-5). In these images, swirls were distinguished fom other longitude, and altitude; (2) visual editing of individual orbit seg­ high-albedo features such as crater rays by their curilinea shaps ments sleeted fr minimal exteral feld disturbaces; (3) minimiza­ and increased visibility in frad-scattered light. Digital maps of tion of remaining low-fequency exteral felds fr individual orbit swirls identifed by all available imagery were then superposed on data segments by quadratic detending; ad (4) two-dimensiona maps of the feld magnitude at the spacecraft attude. Baed upn fltering of individual orbit segments to produce a vector feld map aaysis of these composite magnetic/geologic maps, we draw the along the slightly curved surface defned by the spacecraft altitude; preliminary conclusion that swirl fatures ae associated with mag­ maps of the three feld components (radia, east, ad north), the feld netic aomalies revealed by Luna Prospctor. magnitude, ad the spacecraft atitude ae constcted. For data Detailed maps of these swirl features ae curently being con­ obtained at low to mddle lattudes, the horizontal resolution of the structed fr the magneticaly stong regions atpdal to the Imbrum, feld maps is limited by the orbit-tack sepaaton (-30 km at the Serenitatis, ad Crisium Basins. equator). References: [1] Hood L. L. et al. (1979) Science, 204, 53-57. Correlation Analysis: Maps of the feld magnitude have been [2] Hood L. L. ad Schubert G. (1980) Science, 208, 49-51. constucted within limited selenographic regions baed mainly on [3] Lucey P. G. et al. (1998) LS XIX. [4] Hood L. L. ad Willias data acquired in Mach, Aprl, ad May of 1999. This was a time C.R. (1989) Proc. LPSC 19th, 99-113. [5] Hood L. L. et al. (1981) period when the orbit plae wa nearly aligned with the Sun-Mon JGR, 86, 1055-1069. [6] SchultzP. H. and SmkaL.J.(1980)Nature, line so that feld mapping was pssible at times when the Mon wa 284, 22-26. in the sola wind a well as when the Moon was in the geomagnetic tail. Most of te coverage is across the luna faside. However, Fig. 1 shows a example of a feld map produced fom sola-wind SOLAR-WI-ILANTED VOLATIES I TH LUNAR wake data for a region including Reiner Ga on wester Oceaus REGOLITH. J. R. Johnson1, T. D. Swindle2, and P. G. Lucey3, 1U. S. Geological Surey, Branch of Astogeology, Flagstaff AZ 86001, USA ([email protected]), 2Lunarad Plaetay Laoratory, University of Arizona, Tucson AZ, USA, 3Hawai'i Institute of Geophysics ad Planetology, University of Hawa'i, Honolulu HI, USA.

Introduction: The epithermal neuton distribution fom Luna Prospector (LP) and the H abundace ad maturity of retured luna soils ae being used synergistically to estimate global H distribution. The ability to determne luna H abundace using a combination of remote sensing ad laboraory aalyses broadens our understanding of the retentivity of volatiles in the luna regolith and the nature of the sources fom which the volatiles arived (i.e., sola wind, cometar impacts). This is relevat to the seach fr water ice at the lunar ples as well a other potential resources such a H-3. Kowledge of these pssible resouce-rch regions will be importat fr fture mssion planning, instent selection, ad lading-site aalyses. Methoolog: The Level 1 caibrated epither neuton dis­ tibuton maps (initially calibrated to remove major instumental egects by the LP tea) [1-2] have been used fr frst-order calibra­ tion to H abundace. The equaton of Feldma et a. [3], i.e., [epitermal* = epitheral-(0.068 * therma)] was used to mmize the efects of compositona vaatons between mae ad highlads (likely due to major-element concentatons such as Ti, Fe) on the epitheral neuton countng rates. Figue 1 shows a example of a Clementine mosaic at the sae scale a the epiteral* neuton fux Fig. 1. maps. 30 Workhop on New Views of the Moon l

Comparisonof the average H abundances forthe Apollo landing 100.0 site soils with the epithermal* neutroncounts for the landing sites 90.0 from Fig. I results in the rather poor correlation shown in Fig. 2. 80.0 Neglecting the Apollo 11 sample, the sense of the correlation is --·g, 0, 7.0 A14 correct(fewer epithermal* neutrons implies greater H abundance), .2, but the correlationis still mediocre (r=-0.574). This may result from � 60.0 A1S i. 1f using the average values for the landing sites as opposed to the .. 50.0 individual soil samplevalues. 40.0 Results: It canbe seen in the epithermal * map (Fig. I) that the highest counting rates (likely the lowest H abundances) correspond 30.0 weII to regions of fresh,Copernican-age impacts (e.g., Tycho, Hayn, 20.0

Jackson, possibly Stevinus). This implies a dearth of solar-wind­ 10.0 implantedgases, which could be explained by the immaturity of the 0.0 regolith in these regions. But there are also mare regions near the 5.0 510.0 5.0 530.0 550.0 limbs (Orientale, Crisium) with high values that require other expla­ Eera1• nt (cn! nations. Further, while the darker regionsin the polar areashave been Fig. 2. Epithermal* neutron counts vs. average Apollo landing-site H interpreted as water-ice-bearing regions [3), the low-valued high­ abundances. Error bars represent standard deviations of epithermal* · landterrain west of Sinus Iridium requires furtherstudy. neutron counts and average H values among measured soils. Linear fit Overall, the solar-wind fluence model's predictions of highest excludes the Al 1 sample (r = -0.574). volatile abundances on the centralfarside and lowest values on the central nearside are not immediately apparent in the epithermal* H abundancesto refinethe correlation.Once a H map is constructed, map. However, if the lower counting rates (higher H abundances)at it will be used to test a solar-wind fluence model for the Moon ° ° longitudes 75 -I 80 E arereal, it would be consistent with the solar­ developed previously [4,9] that represents relative solar-wind-im­ wind model ( at least forthat portionof the easternlimb andfarside ). plantedelemental abundances (assumingminimal variations in im­ Future Work: Hydrogen contents of 54 Apollo bulk lunar pact history andchemical compositionof the soils, aswell as minimal soils [ 4-6) will be comparedto the epithermal neutron countsfor the saturation of soils with volatiles). Deviations of the H distribution landing sites to determine more precisely an average correlation. fromthe idealized solarfluence model will be investigated by incor­ This will be applied to theLP data to estimateH content on the lunar porating the effects of local magnetic fields [10-14], composition surface. Because exposure ageaffects retained solar-wind-implanted (e.g., Ti, Fe) (15-18], and soil maturity [17]. volatileabundances, incorporationof soil maturity using the 1/feO Further, previous work by the authors has shown that abundances parameter (reduced:total Fe ratio) [7-8) and/or spectrally-derived ofH-3 (a potential lunar resource relevant as afuel for future nuclear maturity data fromlunar soils will also be used in conjunction with fusionpower [19-23]) in the lunar regolith canbe estimated using the fluence model in combination with surface maturity andTi maps constructedusing Clementine multispectral data [24-26]. The mod­ eled solar-windfluence distribution map will be updated based on observations of the epithermal neutron-derivedH maps, and the Clementine and LP Ti-abundance maps will be incorporated with Clementine surface maturity maps to construct revised estimates of H-3 abundance. Conclusions: The calibration of the LP neutronspectrometer data to provide maps of H abundance will be important in under­ standingthe nature of the solar-wind-implantedvolatiles in the lunar regolith, particularly the distribution, migration, and retentivityofH relevant to understanding the possible presence of water ice at the lunar poles.Hydrogen maps will be useful in separatingthe near­ surface geologic history (e.g., solar-wind implantationand microme­ teorite bombardment history related to surface maturity) from the underlying crustal evolution,the knowledge of which is a required component of our total understandingof the planetary history of the Moon.The preliminary observationsmade here regarding correla­ tions of epithermal* neutron counts with geologic features arein­ triguing and will require furtherstudy. Potential improvements to models of solar-windfluence will assistin mapping other volatiles in the regolith, particularlyH-3, which alongwith Hand water ice are some of the most promising potentiallunar resources available on the Fig.1. (a) Clementine mosaic with craters Jackson (J), Tycho (T), Hayn Moon [e.g., 6,20]. (H), and Stevinus (S) marked and overlain with 30° latitude/longitude References: [1] Binder A. B. (1998) Science, 281, 1475-1476. grid; (b) epithermal* neutron distribution map. Simple cylindrical [2] Feldman W. C. et al. (1998) Science, 281, 1489-1493. projection, 0.25°/pixel, with 180° W at left edge. [3) FeldmanW. C. et al. (1998) Science, 281, 1496--1500.[4] Fegley • LP/ Contribution No. 980 31

B. and Swindle T. D. (1993) in Resources of Near-Earth Space, pp. be a unifing concept fr a number of petologic and geochemica 367-426, Univ. Arizona. [5] Des Maas D. J. eta. (1974) Proc. LSC obsesations. 5th, 1811-1822. [6] Carter J. L. (1985) in Symposium on Lunar Remotely-sensed Thorum Distribution in the Procellarum Bases and Space Activities of the 21st Century(W.W. Mendell, ed.), KREEP Terrane: From the initial Luna Prospector y-ray spec­ pp. 571-581, LPL [7] Morris R. V. (1976) Proc. LSC 7th, 315-335. trometer data (-5° resolution [6]) and fom the preliminary low-orbit [8] Morris R. V. (1978) Proc. LPSC 9th, 2287-2297. [9] Swindle data [7], there appeas to be a number of relatively hotter "spts" T.D. etal.(1992)LPSXl, 1395-1396. [10] HoodL.L. ad Huag within the PKT in terms of Th concentration. Some of the hotter spts Z. (1991)JGR, 96, 9837-9846. [11] HoodL. L. eta. (1979) Science, corespond to interediate-sized craters that penetrated volcaic 204, 53-57. [12] Lin R. P. et al. (1998) Science, 281, 1480-1484. fows ad excavated Th-rch, submae material, such a Aristachus, [13] Hood L. L. and Schubert G. (1980) Science, 208, 49-50. Aristillus, and Kepler [8]. Other spots, however, corespond to [14] Hood L. L. and Williams C. R. (1989) Proc. LPSC 19th, 99- surfcial frmations that constitute mainly rough topography associ­ 133. [15] Lucey P. G. et al. (1996) LPS XVI, 781-782. [16] Lucey ated with Imbrium ejecta or circum-Imbrium ring mountans and do P. G. et a. (1995) Science, 268, 1150-1153. [17] Lucey P. G. et al. not necessaly imply the presence of exposed KREEP basalts. (1998)JGR, 103, 3679-3699. [18] Elphic R. C. eta. (1998) Science, The Fra Mauro Formation south of Copricus towad the Apllo 281, 1493-1496. [19] Lewis J. S. (1991) Space Power, JO, 363-372. 14 site and regions of the Alpes Formation southwest of Copercus [20] Taylor L. A. (1994) in Engineering, Construction, and Opera­ [9] in the vicinity of Reinhold lie within the most prominent hotspot; tions in Space IV, Proc. Space '94, pp. 678-686, Albuquerque. here the T concentation is consistent with that fund in the Apollo [21] Wittenberg L. J. et al. (1986) Fusion Technol., 10, 167-178. 14 soils (-12-13 ppm) when the proprions ofFra Mauro Foration [22] Jorda J. L. (1989) in Space Mining and Manufacturing, An­ ad mae basalt ae considered [5]. The aea between Copericus ad nual Invitational Symposium, pp. 665, NASA SERC, Univ. Ari­ Kepler and northwest of Copericus in terra extending to the zona, Tucson. [23] Jorda J. L. (1990) LPJ Tech. Rpt. 90-02, 43-5. Carpathias is similaly enriched in Th. The Apennines fom Era­ [24] Johnson J. R. et al. (1991) JGR, 96, 18861-18882. tosthenes towad the Apllo 15 site contain elevated Th concenta­ [25] Johnson J. R. et al. (1996) Eos Trans. AGU, 77, 446. tions, as does the northwester quadrat of circum-Imbrium tera, [26] Johnson J. R. et a. (1999) GRL,26, 385-388. espcially between (but not including)La Condamine and Plato, ad in the region northwest of the Jura mountains extending southwad past Maira to the Gruithuisen-Domes region. Witn the main fORENRCHNTWITHITHPROCELLARUM topographic rim of Imbrium, the Apnnine Bench frmation south of KREEP TERRANE: THE RECORD IN SURFACE Archimedes appars to have relatively elevated Th concentaion DEPOSITS AND SIGNIFICANCE FOR THERMAL [8,10]. EVOLUTION. B. L. Jollig, J. J. Gillis, ad L. A. Haskin, Comparing the map of Th distibution to a digital-elevaton map Deparent of Earth and Plaetary Sciences ad the McDonnell derived fom Clementine altimet, it appas that most of the aeas Center fr the Space Sciences, Washington University, St.Louis MO richest in Th occur where the surface is elevated relative to the 63130, USA ([email protected]). majority of PKT volcaic plans.Not all rough topgraphy within the PKT ha such elevated Th, however. Based on a aalysis of the 5° Introduction: The neaside-faside structural and composi­ data, ad using the calibration of [11], the mea Th concentrations tional asymety of the Moon was recognized during the ealy days fr manly volcanic-resuaced terain ad rgged terain ae simila of Apllo and the suggestion was made [e.g., 1] that the migration of (-5.5 ppm) [4]. This occurs in part because craters that penetrated matle melts to the neaside would have been fvored by ealy Earth­ mae baalt excavated Th-rich material. Even so, there appea to be Mon orbital dynaics ad nonuniform planetesimal bombadment. extensive aea of volcaic resurfacing that have no obvious extinsic Recent global geochemical mapping by Luna Prospector has pro­ source of Th-rich materia, suggesting that the basalts, themselves, vided additonal data, paiculaly in the Th distribution, that stongly may contain as much as 5-6 ppm Th [5]. High-FeO concentrations supports the noton of global, prefrential melt migration, which led ( 18 to > 20 wt%) indicate that these ae not KREEP baalts but mare in part to the development of the Procell am KREEP Terae (PKT) basalts. If so, this is surprising because most of the Apollo-sampled [2-5]. The surface distibution of Th was then reshaped by basin­ mae basalts have very low-Th concentations (typically <2 ppm frming impacts into the PKT, espcially the Imbrum impact, which [12]). was the last and lagest to strike in that region. The Imbrium event Thorium-rich materials in the Apollo Samples: A vaety of probably excavated material fom a parially molten zone deep in the Th-rch materials occur in the sample collection, paticulaly in the crust ad delivered Th-rich ejecta Moon-wide [2]. A fndamentally saples fom the Apllo 12, 14, ad 15 sites. The most abundant Th­ import but poorly understood aspct of the globa Th distbution rich rok tps are the mac impact-melt breccias, which although is the concentation of Th in the subsurace rocks of the PKT crstal fund at all sites, ae most abundat at Apllo 14, where they secton. For exaple, depending on what assumptons ae mae, the dominate the rock saples and mae up some 40% of the rock PKT crstal section, which is -12% of the crst ad only -1.2% of paicles in the soil. These have Th concentatons ranging up to -30 the whole Moon, may contain as much a 40% of the Moon's entire ppm ad averaging -18 ppm [13]. The Apllo 14 soils contain -13 Th budget [4]. Such a distbuton of Th and related heat-producing ppm Th, refecting the high abundance of ts melt-breccia compo­ elements would have had a pro�ound efect on melting, mixing, ad nent (Fig. 1). At the nearby Apollo 12 site, the rocks consist mainly the thermal evolution of the PKT ad te underlying mantle [2]. In of mae basats, ad these have low-Th concentations, mostly

18 �-. -A-1-.4-I-·-·Av-,= ••-��----�--- --, Implications: Dynamc exchage ofTh-rich (KREEP) residua 16 ,·;- � ·��!:f\?.\ ♦mostlyt&lra between the lower PKT crst ad the underlying matle would have .. ·'.· :,'., · .., , , iimos11yman, · . supplied the heat to continue to generate the youngest( ad unsapled) ,. » •. · '·' .l•mixed neaside basalts, ad would provide material fr deep mixing ad assimilation. Inversion ofmatle cumulates in response to density . ��};:,;'.�;12sds instability, coupled with the exchage of heat-producing elements in the uppr mantle, could provide a frtile source fr the generation of magnesia-suite paental melts and would supply the magnesian olivine compnent needed to satisf chemical mass-balance models [22) fr materials derived within the PKT. Although the mechaism of concentraung KREEP residual melt into the PKT is not under­ stood at this time, the consequences of that concentration appa to 11 13 15 17 19 21 23 F (w%), Clementine have been active crustal magmatism within the PKT, crstal relax­ ° ation fllowing multiple basin impacts, ad extended matle melting Fig. 1. Ferrous oxide vs. Th, Procellarum KREEP terrane, 5 data. ad volcanic resurfcing of the terae. Acknowlegent: Tis reseach is supprted by grants NAG5- to 50 ppm [14). A single fagment ofKREEP baalt fom the Apllo 6784 (BLJ) and NAG5-4172 (LAH). 12 sil contains -50 ppm Th [15). Despite the identifcation of highly Reference: [1] Wood (1973) Moon, 8, 73-103. [2] Haskin Th-enriched lithologic components at the Apllo 12 site, the soils (1998) JGR, 103, 1679-1689. [3] Wieczorek and Phillips (1999) there vary linealy in composition so a to extapolate to a moderate­ JGR, submitted. [4] Jollig et al.(1999) JGR, submitted. [5] Haskin Th KREEP-basalt compnent or a composition like that of the Fra et al. (1999) JGR, submitted. [6] Lawrence et al. (1998) Science, Mauro frmation a refected by Apllo 14 soil( 14 ppm Th at 10 wt% 281, 14841489. [7] Lawrence et al. (1999) LPS XXX, Abstact FeO, Fig. 1) [4). The neaby Lansberg Crater (39-k diameter) is a #2024. [8] Gillis ad Jollig, this volume. [9] Wilhelms ad McCauley likely cadidate to have delivered submae material such a buried (1971) U.S. GeologicalSurveyMapl-703.[10) Etchegaray-Raez Fra Mauro or Alpes material [9] to the Apollo 12 site. The Apollo 15 et al. (1983) Proc. LPSC 13th, in JGR, 88, A529-A543. [11) Gillis site provided appaently endogenous KREEP basalt (-10 wt% FeO et al. (1999) LPS XXX, Abstract #1699. [12) Korotev (1998) JGR, ad 11-12 ppm Th) a well a a variety of impact-melt breccia and 103, 1691-1701. [13) Jollig(1998) Intl. Geol. Rev., JO, 916935. evolved rocks such as 15405 [16). [14] Quick et al. (1977)Proc. LSC 8th, 2153-2189. [15] Laul (1986) The existence of chemically evolved, Th-rich rock types such as Proc. LPSC 16th, inJGR, 91, D251-D261. [16] Ryder(1976)£PSL granite, alkali aorthosite, quartz monzodiorite, and whitlockite 29, 255-268. [17] Shih et al. (1993) GCA, 57, 4827-4841. monzogabbro, some of which have Th concentrations raging fom [18) Lugmair and Carlson (1978) Proc. LPSC 9th, 689-704. 20 to 66 ppm [12), refect magmatic recycling ad diferentiation of [19] Schultz ad Spudis (1983) Nature, 302, 233-236. [20) BVSP KREEP-rich, primay, residua-melt products ofthe magma ocea. (1981) Pergamon, NY [21) Jollif et al. (1991) Proc. LPS Vol. 21, The isotopic systematics ofthe evolved rocks indicate magmatic ages 193-219. [22) Korotev (1999) JGR, submitted. raging fom 4.4 to 3.9 Ga [17) ad derivation fom a early difer­ entiated (4.4 Ga) KREEP reservoir [ 18). The average Th concentra­ tion ofthe PKT crst, however, is probably better represented by the SPACE WEATHERG I THE FI SIZE FRACTIONS main groups of mafic melt breccia, most ofwhich probably derived OF LUNAR SOIS: SOI MTURITY EFFECTS. L. P. fom the Imbrium bain and thus sample the sub-Imbrium crst. The Keller 1, S. J. Wentworth2, D. S. McKay3, L. A. Taylor4, C. Pieters5, average Th concentations ofthese groups rage fom -4 to 18 ppm ad R. V. Moris3, 1MVA, Inc., 5500 Oakbrook Parkway, Norcross [12), and as products of a bain-frming impact, these presumably GA 30093, USA([email protected]), 2Mail Code C23, Lockheed sapled lage aeas of the PKT crst. We note that the main-group Martin, NASA Johnson Space Center, Houston TX 77058, USA, (poiklitic) Apllo 17 impact-melt breccias have -5 ppm Th; if they 3Mail Code SN, NASA Johnson Space Center, Houston TX 77058, originated in the Serenitatis event, this suggests an enriched source USA, 4Plaetay Geosciences Institute, University of Tennessee, at depth, although at the surface, the site lies outside of the PKT a Knoxville T 3 7996, USA, 5Box 1846, Brown University, Providence defned in ( 4]. RI 02912, USA. Although most of the sampled mare baalts have Th concenta­ tions below 2 ppm, a fw have higher concentrations and prhaps Itroduction: The efects of space weathering on the optical other unsapled mare basalts, particularly some of the presumed properties of luna materials have been well documented [1-2). younger ones(e.g., 1-2 Ga [19)) have relatively high Th. Apollo 14 These efects include a reddened continuum slop, lowered albedo, aluminous and very-high-K basalts have 2-3 ppm T [12, 21), ad and attenuated absorption fatures in refectace spcta of luna some Apllo 11 high-K basalts have a much as 4 ppm (20). A soils as compaed to fnely comminuted rocks fom the sae Apllo baaltic impact glas fom Apllo 14 has 5.4 ppm Th and 19. 7 wt% sites. However, the regolith processes that cause these efects ae not FeO [21). The obsesaton that lage aea ofmae volcaic plains in well known, nor is the ptographic setting ofthe products of these the PKT have Th as hgh as 4-7 ppm indicates that te basalts processes fully understood. A Luna Soil Chaterization Consor­ themselves ae enriched in Th. Modest enrichment of T in the tium has been fred with the puose ofsystematically integrating Traquillitatis basalts indicated by Luna Prospctor data [6] implies chemical ad mneraogical data with the optical proprtes of luna a heterogeneous distibution of Th in the sub-PKT ad ajacent soils [3]. Understading space-weathering efects is critca in order matle. to fully integrate the lunar sample collecton with remotely-sensed LP/ Contribution No. 980 33

data fromrecent robotic missions( e.g., Lunar Prospector,Clementine, in 79221 consist of agglutinitic glass and agglutinate fragments, andGalileo). while the remainder are predominantly crystalline mineral grains. We have shown that depositional processes (condensation of The agglutinic glass particles contain abundant nanophase Fe and impact-derived vapors, sputter deposits, accreted impact material, vesicles. Angular particles are rare, with most showing smooth, e.g., splash glass, spherules, etc.) are a major factorin the modifica­ rounded exteriors. Of the mineral grains analyzedthus far,over 90% tion of the optical surfaces of lunar regolith materials [4-6]. In of the grains have amorphous rims that contain nanophaseFe (these mature soils, it is the size and distributionof the nanophasemetal in rims are believedto have formedby vapor deposition and irradiation the soil grainsthat has the major effect on optical properties[7]. In effects [7]). The nanophase Fe in theserims probably accounts fora this report, we compare and contrastthe space-weathering effectsin significant fraction of the increase in ls/FeO measured in these size animmature and a mature soil with similarelemental compositions. fractions (3]. In addition to the rims, the majority of particlesalso Methos: For this study, we analyzed <10 µm sieve fractions show abundant accreted material in the form of glasssplashes and of two Apollo 17 soils, 79221 (mature, Is/FeO = 81) and 71061 spherulesthat also contain nanophase Fe (Fig. 1). In starkcontrast, (immature, ls/FeO = 14). Details of the sieving procedures and the surfacesof the mineral grains in the 71061 sampleare relatively allocation scheme are given in [8]. The results of other detailed prisitine (Fig. 2), asonly -14% of the mineral grains in the sample chemical, mineralogical, and spectroscopic analyses of these soil exhibited amorphous rims. Furthermore, the mineral particles are samplesare reported elsewhere [3]. A representative sample of each more angularand show greater surface roughness thanin the mature soil was embedded in low-viscosity epoxy, and thin sections (-70 nm sample.Accreted material on particle surfaces is rare.Agglutinitic thick) were obtained through ultramicrotomy. The thin sections used material is a major component of the 71061 sample; however, for these analyses typically contained cross sections of up to 500 nanophaseFe andvesicles arenot as well developed as in the 79221 individual grains. The thin sectionswere studied using a JEOL 2010 sample. transmission electron microscope (TEM) equipped with a thin­ window energy-dispersiveX-ray (EDX) spectrometer.An individual thin section was selected from each soil, and foreach grain in the section we determined (1) the elemental composition by EDX; (2) whether the grain was crystalline or glassy using electrondiffraction and darkfield imaging; (3) the presence or absence of rims and accreted material; and (4) the distribution of nanophaseFe where present. The results of these analysesare collected in Table 1. Most of the categories are self-evident; however, we divide the agglutinate­ derivedrnaterialintoagglutinitic glass (glass with approximately the same composition as the bulk soil that contains nanophase Fe with or without vesicles) and agglutinatefragments, which are composed of crystalline grains and agglutinitic glass. Lithic fragments are definedas polymineralicgrains with no glass.Pyroxene grains have been divided into high- and low-Ca groups. Resuts and Discussion: As expected, there are a number of differencesin the petrography of the <10-µm fractionsof 79221 and Fig. 1. A field-emission SEM image of a typical plagioclase grain from 71061 given the great differencein theirrespectivematurities, but we 79221 covered with micropatina showing a rounded, smooth surface with focus here on two major distinctions: agglutinate content and the abundant splash glass (pancakes, etc.). number of grains with micro patina Slightly over 50% of the particles

TABLE 1. Transmission electronmicroscopy modal mineralogy of <10-µm fractionsof 79221 and 71061.

79221,138 71061,170 # % # % Aggutinitic Material 55 51.4 33.0 30.6 Agglutinate Fragments 13 12.1 17 15.7 AgglutinateGlass 42 39.3 16 14.8 Anorthite 20 18.7 28 25.9 Ca-pyroxene 10 9.3 8 7.4 Low-Ca pyroxene 6 5.6 10 9.3 Ilmenite 5 4.7 11 10.2 Olivine 3 2.8 4 3.7 Cristobalite 3 2.8 5 4.6 GlassSpheres 5 4.7 3 2.8 Lithic Fragments 0 0.0 6 5.6 Fig. 2. A field-emission SEM image of a typical mineral grain Total 107 100.0 108 100.0 (pigeonite) from the 71061 sample showing the surface roughness and lack of abundant accreted materials. 34 Workhop on New Views of the Moon I/

Sumary: Ifis now recognized that nanophaseFe is probably sided sectors (Procellarum Ocean-,SPA basin--)separated by two the main agent in modifying the optical properties oflunarsoil grains. differently uplifted ones (+ +, +), one of which (+ +) is the highest The most important result of this study is the observation that in the lunar highland region. Observingthe angular momentum preserva­ finesize fractionsof mature soils, nearlyevery grain has nanophase tion law, the highest sector is composed of anorthosites, andeven of Fe within 100 nm of the particle surface. the less dense Na-rich varieties of this rock. The deepest SPA basin References: [1] Pieters C. et al. (1993) JGR, 98, 20817. sector with an abrupt northern boundary separating it from the [2] Fischer E. (1995) Ph.D. thesis, Brown Univ. [3] Taylor L. A. highest sector (like the Indoceanic sector contacts with the highest et al., this volume. [4] Keller L. P. et al. (1999) LPS XXX, Abstract African one) must be filled with denser rocks than the shallower #1820. [5] Wentworth S. et al. (1999) MAPS, in press. [6] Keller Procellarum ocean sector filled with basalts and Ti basalts. The L. P. and McKay D. S. (1997) GCA, 61, 2331. [7] KellerL. P. et al. Clementine spectral data show a presence of orthopyroxene andan (1998) New Views of the Moon, LPI, p. 41. [8] Taylor L. A. et al. absence of plagioclase [2], favoring some dense ultrabasic rock. The (1999) LPS XXX, Abstract #1859. obvious tendency to approach this type of rock would be to observe it in the Luna 24 samples from also very deep Mare Crisium. In fragments there prevail pyroxene andVLT-ferrobasalts (Mg-poor). THE DEEPEST LUNAR SPA BASIN AND ITS UNUSUAL Unusual melt matrixbreccia with globules and crystalsof Fe metal INFILLING: CONSTRAINTS IMPOSED BY ANGULAR were also found[3 ]. In SPA basinfill some admixture ofFe metal and MOMENTUM CONSIDERATIONS. G. G. Kochemasov, troilite could be also predicted. With this rock in mind we can Institute of Ore Deposits, Geology, Petrology, Mineralogy, and construct a ladder of ascending UB-basic rock densities against Geochemistry, Russian Academy of Sciences, 35 Staromonetny, descending topography: KREEP basalts, low-Ti basalts, high-Ti Moscow 109017, Russia basalts,VLT-Mg-poor ferrobasalts, and pyroxene (with metal) rich rocks. On Earth,the density of basalt floods (their Fe/Mg ratio) also Successful applications of planetary wave tectonics [1] forpre­ increases in the same direction. The lunar and terrestrial sectoral dicting the shapes of small celestialbodies (asteroids, satellites), structures as well as tectonic dichotomies were formed in the very Phobos' rippling, the dumbbell shape of martianspheres, and frac­ beginning of their geological histories. tionated martian crust, allow us to extend this method to lunar References: [1] Kochemasov G. G. (1999) Geophys. Res. tectonics andrelated it to the chemistryof the enigmatic South Pole Abstr., 1, 700. [2] Pieters C. M. (1997) Annales Geophysicae, Suppl. Aitken Basin. III to v.15, 792. [3] Basu A. et al. (1977) Conerence on Lna 24, The accepted origin by many (but not all) planetologists is an Houston, LSI Contribution 304, 14-17. impact hypothesisof the SP A basin; we alternatively,consider it as a part of a global lunar sectoral structure centered in the Mare Orientale. Sectoral structuresof celestial bodies are a result of inter­ NORTH-POLAR LUNAR LIGHT PLAINS: AGES AND ference of standinginertia-gravity waves proceeding in fourdirec­ COMPOSITIONAL OBSERVATIONS. U. Koehler 1,2, J. W. tions (ortho- and diagonal). These warping planetary waves arise in Head 1111, G. Neukum2, and U. Wolf2, 1Box 1846, Department of them as a result of their movements in elliptical orbits with periodi­ Geological Sciences, Brown University, Providence RI 02912, cally changing curvatures and cosmic accelerations. Fundamental USA ([email protected]), 2Deutsches Zentrum ftir Luft- und waves of long 21tR(R = body radius) produce tectonic dichotomy; Raumfahrt,12484 Berlin, Germany. waves of long 1tR(the firstobertone) produce sectoring; and smaller waves length of which is proportional to orbital periods produce Abstract: Varying surface ages of lunar light plains in north­ tectonic granulation. Segments, sectors, andgranulas of differing ern-nearside latitudesindicate anorigin of these smooth terraeunits radius-vectors (risen+ andfallen- tectonic domains) tend to equal­ not exclusively related to Imbrium and/orOrientale impact ejecta and ize their angularmomenta by density ofinfilling matter. That is why subsequent processes. Multispectral data seem to support a more oceanic and mare basins normally are filled with denser material diversified history formany of these plains. (basalts) thanlighter highlands. Introduction and Background: The nature, ages, stratigraphic On Earth one observes six antipodal centers of 1tR-structures position, composition, and mode of emplacement of lunar light (three pairs: (1) Equatorial Atlantic; (2) New Guinea; (3) The plains have been discussed with controversy for more than three Pamirs-Hindukush; (4) Easter Island; (5) Bering Strait; and decades. Covering -5% of the lunarterra surface, the relatively low­ (6) Bouvet Island.)that regularly converge by common algorithm albedo plains arethe most distinctive terralandforms after the more fallen normallyoceanic and risen normallycontinental blocks. Around craterlike ejecta of the freshbasins [1]. Morphological properties, the Pamirs-Hindukush center, for example, areplaced two differ­ like their smoothness, lower crater densities, their superpositionon ently risen sectors (African+ +, Asian+) separatedby 2 differently the "Irnbrian Sculpture," and frequentoccurence as crater fill, are subsided ones (Eurasian-, Indoceanic - -). The six centers form mare-like.Other features, such asrelative ( comparedto marebasalts) vertices of an octahedroninscribed in theterrestrial sphere. The first high-albedo andgeological/stratigraphical setting, are more high­ antipodalpair lies in the equatorial zone, the second in tropics,and land-like.Not surprisingly,light plains were seen as both volcanic the third in the polar ring zones stressing profound connection andimpact related deposits (summarized in [2]). The Cayley Plains, between cosmic positionof a bodyand its internalstructure. a typelocality in the central-nearsidehighlands, hasbeen choosen as On the Moon we now know four antipodal centers of 1tR-struc­ the landing site of the Apollo XVI mission partly to help resolve tures: (1) Mare Orientale;(2) Joliot-Maxwell-GiordanoBruno area; these interpretations. The astronauts collected samples - highly (3) Daedalus-Heaviside; and(4) Ptolemaeus-Flammarion.Around brecciated rocks - andconcluded that the Cayley was indeed of the MareOrientale, like on Earth, aretwo opposite differentlysub- impact origin ([3, and subsequent reports in Al6 PSR]). These L/ Contribution No. 980 35

findings have been extrapolated to stratigraphically similar plains formationpeaked at the time of, or shortly after, the Ori en tale impact units on the nearside,focusing on the Imbrium and Orientale impacts event; (4) light plains formationcontinued a significanttime span as responsibleevents forresurfacing terraeenvironment to formlight after the Orien tale basinimpact; (5) there arelight plains ages that are plains [4]. In addition, theoreticalmodeling mechanismshave been remarkably younger than the age of the Orientale basin, exposing provided that could explain how basin andcrater ejecta were able to surface ages that areonly slightly older thanthe mare-basalt ages of make up for the smoothness of light plains by stirring up local their neighboring units at the eastern margin of Mare Frigoris; and materialthrough secondary-impactrelated processes [5, 6], or mega­ (6) the time span of light-plains surface formation seems to extend impact induced seismic shaking [7]. fromaround 3.95-4.0 Ga to 3.60-3.65 Ga. In addition, there seem However, subsequent age determinations showed that some light to be correlationsbetween ages of light plainsand their geographic plains cannotbe correlatedto the Imbrium or Orientale event [8,9], position; oldest ages>3. 90 Ga occur in the vicinity of the north pole, the lasttwo basin-forming impactsto occur on the Moon that had the in and surrounding the craters Byrd, Main, Scoresby, and north of ability to resurface areas thousands of kilometers away from their Anaxagoras. Plainswithlmbrian-imp actages as young asthe Orientale targetsite. An unknown formof highland volcanism was proposed impact areclustered in andaround craters Meton, Barrow,Barrow as a contributing process in light-plains formation. The question K, Baillaud, east of W. Bond, and Neison. Ages that are clearly remained unansweredwhether processes other than impact-related younger than the Orientale impact can be found in the extended processes were responsible for the formation of these enigmatic smooth plains area west, south, and southeast of crater Gartner, geological units, and how these processes might have worked. Fo­ northeast of Epigenes, east of De Sitter, Barrow K, and the areas cusing on light plainsin the northern-nearsidehighlands, a chrono­ borderingthe easternmargin of MareFrigoris. Except forthe suspi­ logical approach has been chosen to address these questions, and cious correlationbetween geographic occurence andabsolute ages, compositional information frommultispectral data has been included the statisticpattern of light-plains ages is fairly well comparableto to supportour investigations. that of the central-highlands(Cayley Plains)units [14], where espe­ Surface Ages: Mappingthe northern-nearsidelight plains, ear­ cially light plains of ages younger thanthe Orientale impact canbe lier workers have recognized the stratigraphicallyand morphologi­ found- but in that case in no direct contact relations to mareunits. callyobvious bimodal distributionof smooth terraeunits north and MultispectralData: In order to get additionalinformation on northeastof MareFrigoris [10]. Theolder of these plains (lp1; based the light-plains areas that have been selected for crater statistics, on stratigraphic relationsand surface-crater densities) show gradual multispectral data have been investigated from the second Earth/ transitioninto (Imbrium-impact) FraMauro Formation units, whereas Moon encounter of the Galileo spacecraft(December 1992), which the younger unit (Ip2) cannot be related to this relatively nearby­ had excellent viewing geometryof the areaof interest [ 15]. Based on impact event. Instead, it wasproposed that the impact of Orientale, previous work [16] it could be confirmedthat the north-polarplains despite the factit is severalthousand kilometers away, could have indeed show distinctive albedobehavior - being clearlybrighter smoothed these terraeunits with its ejecta stirringup local highland than maria but darkerthan highlandunits and showing less albedo material [4], aninterpretation that seems to be not very compelling, variation thanhighland. Well-exposed mare-basaltunits and well­ as these younger plains show quite homogeneous surfaces over exposed unambigous highland areas have been chosen to extract extended areas.Determining precise surface ages of smooth surface referencespectra in order to get clues to plains affinities. A number units should help with getting a chronology of plains emplacement in these latitudes. Based on the principle of crater-frequencydistribution measure­ 7----,..------...... --- ments andadjusting the cumulative crater-frequencydistributions to Orien�e Imbri a lunar standard distribution [11], and "fixing" the Orientale and 6------Imbrium event with the absolute ages obtainedby radiometric mea­ surements of Apollo samples[12] to this distribution,one is able to s------_ determine reliable absolute-age data forsurfaces after measuring the crater-frequency distribution on it. As the decline in impact fre­ quency is rather steep during the first billion yearsin lunarhistory, 4t------ll----+--➔ the discrimination between surfaces of different age canbe deter­ minedto a relatively high degree of reliability. 3 Forty-fivesmooth plains areas northand northeast ofMare Frigoris have been mappedfor crater frequency.Some light plains have been 2------disturbed by secondarycratering impactto a degree that madededuc­ tion of reliable age data impossible, so only 27 crater counts yielded useful absolute ages. A histogram of how these plains ages are distributed is shown in Fig. 1. For reference, the Imbrium and 3. 90 0 Orientaleevents aremarked attheir respective time of impactat 3.ro Ga, and 3.84 Ga respectively [13]. 3.50 3.70 3.80 3.90 4.00 There areseveral observationsthat can be made in this distribu­ tion: (1) the ages do not cluster around oneor two peaks;instead, the Fig. 1. Frequency (y-axis) of light plains ages (x-axis,bi llion years) in majorityof light plainsformed in a rather broad time spanbetween thenorthern-nearside hemisphere. Note that the twoyoungest ages (blue) 4.0 Ga and3.65 Ga; (2) there aresome light plains that seemed to aresurfaces of mare-basaltunits in easternMare Frigoris, adjacent to light have formed before the Imbrium impact event; ( 3) light plains plainsunits. Tune of Imbriumand Orientale eventsare given as reference. 36 Workhop on New Views of the Moon l

of fresh young craters in the light plains environment have been factorsmake the LMs importantto remote-sensing studies: (1) Most chosen to "sample" the material that is covered by the mature re­ arebreccias composed of regolith or fragmentalmaterial; (2) all are golith. There arethree interesting observationsto report: (I) There rocks that resided ( or breccias composed of material that resided) in arelight plains showing clearhighland affinities; (2) there are light the upperfew meters of the Moon prior to launch [e.g., 15]; and (3) plains that show definite mare affinities; and (3) the occurence of most apparently come from areasdistant fromthe Apollo sites. these two groupings is confined to areas in greater distance from How Many Lunar Locations? At this writing (June 1999), MareFrigoris forthe first,and close to MareFrigoris for the latter there are 18 known lunar meteorite specimens (Table 1). When "spectral" group. In an ongoing process, application of spectral­ unambiguouscases of terrestrial pairingare considered, the number mixing models to Galileo andClementine multispectraldata should of actual LMs reduces to 13. (Terrestrial pairingis when a single help answering fundamental questions on compositional aspects piece of lunar rock entered Earth'satmosphere, but multiple frag­ here. ments were produced because the meteoroid broke apart on entry, Conclusion: Based on the absolute ages of light-plains sur­ upon hitting the ground or ice, or while being transportedthrough the faces,the formationof these geologic units cannotbe attributedto a ice.) uniformprocess. The factthat there are plains both clearlyyounger We have no reason to believe that LMs preferentially derive from andolder than the two largest last basin-forming impact events of any specific region(s) of the Moon; i.e., we believe that they are Imbrium and Orientale,respectively, excludes an impact-ejecta ori­ samplesfrom random locations[ 15]. However, we do not know how gin related to these events, although Imbrium andOrientale ejecta many different locations are represented by the LMs; mathema­ depositionand their subsequent effectsmay certainly have played a tically, it could be as fewas 1or asmany as 13. The actual maximum major role in plains formation.Varying spectral characteristics con­ is <13 because in some casesa single impact appears to have yielded firmthat probably more thana single process is responsible forthe more thanone LM. Yamato 793169 andAsuka 881757 (Table 1, N formingof these terraplains. As forthe younger plains, it could be = 12 and 13) areconsidered "source-crater paired" [15] or "launch suggested that Late lmbrian craters contributed ejecta deposits to paired" [l] because they arecompositionally and petrographically cover existing early marebasalts (making up forthe smoothness of similar to each other anddistinct from the others, and both have these surfaces)in theeastern Frigoris area, but the scarcity of dark­ similarcosmic-ray exposure (CRE) histories[12]. Thesame canbe haloed craters does not supportthis idea-furtherinvestigations are said of QUE 94281 andY 793274 (Table 1, N = 9 and 10) [l]. Thus necessary.The wide timespanof about 300 m.y. in plains ages further the 13 meteorites of Table 1 probably represent a maximum of 11 towardthe north pole alsorequires a more detailed look at compo­ locations on the Moon. sitionalaspects; due to some morphological evidence, unidentified The minimumnumber of likely source craters is debated andin formsof highland volcanism should be kept in mind, asit is known fluxas new data fordifferent isotopic systems areobtained. Conser­ that nonmarevolcanism occured elswhere in the Imbrium viciniy, on vatively, considering CRE data only, a minimum of -5 impacts is the Apennine Bench [17], forminga largeunit of terraplains there. required. Compositional and petrographicdata offeronly probabilis­ References: [1] Wilhelms D. E. and McCauley J. F. (1971) tic constraints.An extreme, but not unreasonable viewpoint, is that U.S. Geological SurveyMisc. Geol. Inv. Map, 1-703. [2) Wilhelms such data offerno constraint.For example, if one were to cut up the D. E. (1987) U.S. Geological Survey Prof Paper, 1348, pp. 20-22; Apollo 17 landing site (which was selected for its diversity) into 215-220. [3] YoungJ. W. et al. (1972)NASAApollo 16Prelim. Sci. softball-sized pieces, some of those pieces (e.g., sample 70135) Rpt., 5-1-5-6. [4) Boyce J. et al. (1974) Proc. LSC 5th, 11-23. would be crystalline marebasalts like Y 793169 whereas others (e.g., [5] OberbeckV. R. (1975),Rev. Geophys. Space Phys., 13, 337-362. sample 7313 I would be feldspathic regolith breccias like MAC 88104/ [6] Oberbeck V. R. etal.(1975) Proc. LSC 5th, 111-136. [7] Schultz 88105. However, nature is not so devious. Warren argues that LMs P. H. and Gault D. E. (1975) Th e Moon, 12, 159-177. come fromcraters of only a few kilometers in diameter [15]. If so, [8] Neukum G. et al. (1975) Th e Moon, 12, 201-229. even though CRE data allow, for example, that ALHA 81005 and [9] Neukum G. (1977) TheMoon, 17, 383-393. [10] LucchittaB. K. Y 791197 (Table I, N = 6 and 7) were launched simultaneously (1978) U.S. Geological Survey Geol. Map 1:5,000,000, 1-1062. from the same crater, the probability is nevertheless low because the [11] Neukum G. et al. (1975) Proc. LSC 4th, 2597-2620. two meteorites are compositionally and mineralogically distinct. [12] Metzger A. E. et al. (1973) Science, 179, 800-803.[13] Neukum Thus, within the allowed range (5-11) forthe number of locations G. (1983) Rabil. thesis, Munich University, 183 pp. [14) Koehler U. represented by the LMs, values at the high end of the range are et al. (1993) LPS XIV, 813-814. [15] Belton M. J. S. et al. (1994) probably more likely. Science, 264, 1112-1115. [16] MustardJ. F. and Head J. W. (1995) Mare Meteorites: Three LMs consist almost entirely of mare LPS XXVI, 1023-1024. [17] Spudis P. D. (1978) Proc. LPSC 9th, basalt. Two, Y 793169 and Asuka 881757 (Table 1, N = 12 and13), 3379-3394. areunbrecciated, low-Ti, crystalline rocks that arecompositionally and mineralogically similar (but not identical) to each other; they probably derive froma single lunar-marelocation [16,12]. The third, LUNAR METEORITES AND IMPLICATIONS FOR EET 87521/96008,is a fragmentalbreccia consisting predominantly COMPOSITIONAL REMOTE SENSING OF THE LUNAR of VL T marebasalt. Thus, these LMs probably represent only two SURFACE. R. L. Korotev, CampusBox 1169, Department of lunarmare locations. The basalticLMs have mineraland bulk com­ Earthand Planetary Sciences, Washington University, St. LouisMO positionsdistinct fromApollo marebasalts. 63130, USA ([email protected]). "Mixed"Meteorites: The petrography of Calcalong Creek hasnot been describedin detail,but compositionallyit is unique in Lunar meteorites (LMs) are rocks found on Earth that were that it correspondsto a mixture (breccia) of about one-halffelds­ ejected fromthe Moon by impact of anasteroidal meteoroid. Three pathic material(i.e., the mean compositionof the feldspathiclunar LP! Contribution No. 980 37

meteorites, below), one-furth KREEP norite, one-fourth VL T mae Among Apllo rocks, the compositions of the FLMs corespond basalt (like EET 87521), and I% Clchondrite. With 4 µgig Th (Table most closely to the fldspathic grculitic breccias of Apollo 16 and I) cd corespondingly high concentations of other incompatible 17 [2,13]. elements, it is the only luna meteorite that is likely to have come fom The FLMs of Table I, consequently and ironically, are especially within the Procellaum KREEP Terce (PKT; see [3] fr discussion importct in tat among lunar samples, they provide the best estimate of the terce concept). we have of the compsiuon of the Feldspathic Highlands Terce [ 6]. Ybato 793274 cd QUE 94281 ae together distinct in being All ae regolith or fagmental breccias cd, b such, ae more likely fagmental breccias containing subequal parts of fldspathic hgh­ to represent the composition of the upper crst in the aea fom which land materal cd VLT me bbalt. Jollig et al. [4] estimate a mare they were launched thc any igneous rock that mght have been to highland rauo of 54:46 fr QUE 94281 and 62:38 fr Y 793274; excavated cd launched fom depth. FMs contain little mae mate­ this digerence is well within the rcge obsesed for soils collected rial (probaly <<5%, on average). All have low concentations of Th only cenumeters apart (in cores) at interface site like Apllo 15 and cd other incompatible elements, indicating that they ae also mini­ 17 [11]. Although the two meteorites were fund on opposite sides mally contaminated (<3%) with materal fom the PKT. Thus, Th of Antacuca, they ae probably launch-paired. Te stongest evi­ concentations of the FLMs provide a reasonable lower limit to dence is that the pyroclbtic glass spherules that occur in both ae of values that might be expected for lage aeb of the luna faside as two compsiuonal groups and the two groups are essentially the detected by Luna Prospector [9]. same in both meteortes [1]. The mean FeO concentauon of the FLMs, 4.45 ± 0.6% (Table 1; Ybato 79 I197 (Table 1, N = 7) is nominally a feldspathic luna ±95% confdence limit) is remakaly simila to the "global mode" meteorte (below), but bong FlMs, it probably contains the highest of 4.5 ± 1.0% obtained fom analysis of Clementine spectal refec­ abundcce of clbts and glbses of mae derivauon. As a conse­ tcce data [10]. (Because the FLMs consist of nea-surface regolith quence, its composiuon is at the high-Fe, low-Mg end of the range material, a signifcant faction of the Fe in the FLMs is of meteoritic fr FLMs cd is not included in the FLM average of Table I. Its origin, lagely fom mcrometeortes. For the purpose of estimating composition is consistent with -10% mae-derived material [8]. the compsition of the upper fw kilometers of fldspathic crst, this Similaly, the two small (Y 82) pieces of Y 82192/82193/86032 compnent must be removed, yielding 4.2% FeO fr the upper (Table I, N = 5) are more mafc than the lage (Y 86) piece, probably crust [ 6]. However, fr comparsons to Clementine results, the FLM b a result of -7% mae-derved material [8] (the mass-weighted mec of Table I is more approprate.) Similaly, the mec TiO2 mean compsition is listed in Table 1). concentation of the FLMs, 0.35 ± 0.15%, agrees well with the vaue Feldspathic Lunar Meteorite (FLMs): All Apollo mssions obtained fom C)ementine, 0.45 ± 1.0% (10], giving credence c low­ went to aeas in or near the PKT, and, consequently, all Apllo TiO2 concentation to the technique used to derive the Clementine­ regolith sbples are contaminated with Th-rich material fom the based estimates. PKT [7]. At the nominally "typica" highlcd site, Apollo 16, -30% The six FLMs ae all similar in composition. Such similarity is of the regolith (

TABLE 1. Lunar meteorites (lunaites): A list in order of increasing Fe concentration. ' N Name Lunar rock type Mass Where Alp3 TiO2 FeO MgO Mg Th (g) found (%) (%) (%) (%) (%) µgig Feldspathic lunar meteorites (FLMs) 1 Dar al Gani 400 regolith breccia 1425 Libya 28.9 0.18 3.8 5.1 71 ? 2 MAC 88104/88105 regolith breccia 61 + 663 Antarctica 28.1 0.23 4.28 4.05 63 0.39 3 Dar al Gani 262 regolith breccia 513 Libya 28.4 0.24 4.35 5.0 67 0.37 4 QUE 93069/94269 regolith breccia 21 + 3 Antarctica 28.5 0.23 4.43 4.44 64 0.53 5 Yamato 82192/82193/86032 fragm.or reg. breccia 37 + 27 + 648 Antarctica 28.3 0.19 4.36 5.23 68 0.20 6 ALHA 81005 regolith breccia 31 Antarctica 25.7 0.27 5.47 8.2 73 0.28 meanFIM 28.0 0.22 4.45 5.3 67 0.35 "Mixed" meteorites 7 Yamato-791197 regolith breccia 52 Antarctica 26.4 0.31 6.1 6.1 64 0.35 8 Calcalong Creek regolith breccia 19 Australia 21.3 0.83 9.7 7.7 59 4.0 9 QUE 94281 regolith breccia 23 Antarctica 17.2 0.59 13.1 9.6 57 1.0 10 Yamato 793274 regolith breccia 9 Antarctica 15.7 0.60 14.2 9.0 53 0.9 Mare lunar meteorites 11 EET 87521/96008 fragmental breccia 31 + 53 Antarctica 13.2 0.97 18.8 6.8 39 1.0 12 Yamato 793169 mare basalt 6 Antarctica 11.4 2.15 21.0 5.7 32 0.75 13 Asuka 881757 mare basalt 442 Antarctica 10.0 2.42 22.6 6.2 33 0.45

Italics= preliminary or highly uncertain values. Mg'= bulkmole % Mg/(Mg+Fe). Datacompiled froma large number of literature sources. 38 Workhop on New Vews of the Moon I

that the FLMs may represent as many as six source craters. Among obtain a better calibrated andmore accurate picture of the Fe abun­ FLMs, ALHA 81005 is distinct in being richer in MgO andhaving dances on the Moon. a correspondingly higher Mg' (Table 1). Although bulk composi­ Data Analysis: To derive Fe abundances, we are using two y­ tions of the other FLMsare generally consistent with their derivation ray lines near 7 .6 MeV. Thesey-rays areproduced by thermalne_utron mainly fromrocks of the ferroan-anorthosite suite [14], the mafic capture. Here, Fe nuclei absorb thermal neutrons,become energeti­ minerals, at least, in ALHA 81005 cannot derive from a source cally excited, and then de-excite with the production of y-rays. dominated by ferroan-anorthosite-suite rocks. Mafic rocks of the Because this process depends upon thermal neutrons, the measured magnesian suite of lunar plutonic rocks are probably all special fluxof 7 .6 Me Vy-rays is proportional not only to the Fe abundances, products of the PKT [7]. Thus ALHA 81005, as well as the high-Mg' but also to the thermalneutron number density. Here, we use mea­ variety of Apollo feldspathic granulitic breccias [13,5], suggest that surements fromthe LP neutronspectrometer (NS) [5] to correctfor there areregions of the feldspathic crustdominated by nonferroan this thermal neutron effect. As seen in [5], this correction is quite (i.e., high-Mg')feldspathic lithologies. Locating such areasshould large as the thermal neutron count rate varies over the Moon by a be a high-priority goal of remote-sensing missions. factor of 3. Many considerations need to be taken into account to Acknowledgents: This work wasfunded by NASA (NAG5- make sure an appropriate correction is applied. These include (1) 4172). converting the measured thermal neutron flux into a true thermal References: [l] Arai T. andWarren P. H. (1999) MAPS, 34, neutron number density; (2) accounting forcomposition effects such 209-234. [2] Cushing J. A. et al. (1999) MAPS, 34, 185-195. as thermal neutronabsorption due to Gd andSm [6]; and (3) equating [3] JolliffB. L. (1999) LPS XXX, Abstract #1670. [4] JolliffB. L. the instrumentsurface resolution forthe GRS and NS. To take care et al. (1998) MAPS, 33, 581--601.[5] Korotev R. L. (1997) MAPS, of most of these considerations, detailed calculations need to be 32, 447-478. [6] Korotev R. L. (1999) LPS XXX, Abstract#1303. carriedout. Yet with some assumptions,a preliminaryestimate of the [7] Korotev R. L. (1999) LPS XXX, Abstract #1305. [8] Korotev thermal neutroncorrection can be obtained. R. L. et al.(1996) MAPS, 31, 909-924. [9] Lawrence D. J. et al. To relate the measuredthermal neutron fluxto the thermal neu­ (1998) Science, 281, 1484-1489. [10] Lucey P. G. et al. (1998) JGR, tron number density, we assume that the thermal neutron energy 103, 3679-3699. [11] MorrisR. V. et al. (1989) Proc. LPSC 19th, distribution is Maxwellian and independent of soil composition. 269-284. [12] Thalmann C. et al. (1996) MAPS, 31, 857-868. With this assumption, the measuredneutron flux, F, is proportional [13] LindstromM. M. and LindstromD. J. (1986) Proc. LSC 16th, to both the neutronnumber density, n, and the squareroot of the mean in JGR, D263-D276. [14] WarrenP. H. (1990) Am. Mineral., 75, lunarsurface temperature, T: 46-58.[15] WarrenP. H. (1994) lcarus, 111, 338-363. [16] Warren P. H. and Kallemeyn G. W. (1993) Proc. NIPR Symp. Antarct. Foe n✓T Meteorites, 6, 35-57. Here, the measuredneutron fluxalso includes epithermal neu­ tronsthat contributeto the overall neutronnumber density. To obtain a global surface temperature, we have used a mean equatorial surface temperature of 250 Kand scaled it as cos(latitude)114• Because this IRON ABUNDANCES ON THE MOON AS SEEN BY THE scaling results in very low temperatures at the poles, a low tempera­ LUNAR PROSPECTOR GAMMA-RAYSPECTROMETER. ture cutoff has been set at 130 K. D. J. Lawrence1, W. C. Feldman1, B. L. Barraclough1, R. C. Elphic1, Results: For this study, we have used the high-altitude GRS S. Maurice2, A. B. Binder3, and P. G. Lucey4, 1 Mail Stop D-466, Los data to obtainthe relative Fe abundances.We correctfor the thermal AlamosNational Laboratory, Group NIS-1, Los Alamos NM 87545, neutron effectusing the low-altitude summed thermal and epithermal USA ([email protected]), 2Observatoire Midi-Pyrenees, neutron data smoothed to the footprintof the high-altitudeGRS data Toulouse, France, 3Lunar Research Institute, Tucson AZ, USA, set. The resulting relative Fe abundance map is shown in Fig. 1. 4University of Hawai'i, USA. A comparison between this map and the published CSR Fe map shows a good correspondence. This correspondenceis shown more Introduction: Measurements of global-Fe abundances on the quantitatively in Fig. 2, which is a scatter plot between the two Moon are importantbecause Fe is a key element that is used in models datasets. Here, the CSR data has been smoothed to the footprint of the of lunarformation and evolution. Previous measurementsof lunar­ GRS data. As seen, formost of the Moon, there is a good correlation Fe abundances have been made by the Apollo y-Ray (AGR) experi­ between the datasets with a correlation coefficient of 0.93. One ment [1] andClementine spectralreflectance (CSR) experiment[2]. region of some disagreement is in the South Pole Aitken (SPA) Basin The AGR experiment madedirect elemental measurements for about (red points), where the CSR measurements correspond to a -20% 20% of the Moon. However, these measurements had large uncer­ higher count rate thanis observedwith the GRS data. tainties due mostly to low statistics [3] and anabsence of thermal In contrast to the correctedGRS data, the blue points show the neutrondata (see below). The CSR-derived Fe data hasmuch better GRS data before the thermal neutron correction is applied. The coverage ( 100%coverage equatorwardof ±70 ° latitude) andspatial correlation coefficient between the uncorrected GRS data andthe resolution (-100-msurface resolution vs. -150-km surfaceresolu­ CSR data is poorer at 0. 73. In addition,the spread of the GRS data tion for the AGR data), but there have been questions regardingthe at a given CSR abundance(-5 channels) is almostas large as the accuracy of these data far fromthe Apollolanding sites [4]. entire dynamic range of the uncorrectedcount rate (-6 channels), Here we present preliminaryestimates of the relative Fe abun­ which implies that there are large uncertainties in the uncorrected dancesusing the Lunar Prospector(LP) y-ray spectrometer(GRS). data. While these data areimportant and useful by themselves, the ultimate Because there aremany assumptions leading to Fig. 1, the uncer­ goalof this study is to combine the LP Fe data with the CSR data to taintiesof these data arenot yet known.For example,it is not clear LPI Contribution No. 980 39

DISCRIMINATION BETWEEN MATURITY AND :a. COMPOSITION FROM INTEGRATED CLEMENTINE :a ULTRAVIOLET-VISIBLE AND NEAR-INFRARED DATA. <31 2 2 [a S. Le Mouelicl, Y. LangevinI, S. Erardl, P. Pinet , Y. Daydou , and 2 �:m S. Chevre1 , 1 Institut d'Astrophysique Spatiale, Centre National de � la Recherche Scientitique, Universite Paris XI, Bat 121, 91405 '° Orsay cedex, France ([email protected]),2 Unite Mixte de Recherche ' 5562, Group forSpace Researchand Geodesy, Centre Nationalde la --- Recherche Scientifique,14 av E. Belin, Toulouse, 31400, France.

Introduction: The Clementine UV-VIS dataset has greatly improved our understandingof the Moon. The UV-VIS camerawas limited to fivespectral channels from 415 to 1000nm.The Clementine near-infrared(NIR) camera was designed to complement this spec­ tral coverage. The NIRfilter at 2000 nm allows the discrimination between olivine and pyroxene within identified mare basalts[l]. In addition, we will show that the integration of Clementine UV-VIS Fig. 1. Global map of relauve Fe abundances fom the LP GRS overlaid and NIR datasets allows a better evaluation of the ferrous 1-µm wit a luna surface features map. (o view Figs. 1 and 2 in color, go to absorption band depth and gives access to the slope of the continuum. http://cass.jsc.nasa.gov/meetngs/moon99/pdf8036.pdf.) The discriminationbetween maturity and FeO compositioncan be achieved by a principal component analysis performedon spectral . parameters. Data Reduction: We selected 952 Clementine UV-VIS and NIRimages to compute a multispectralcube covering theAristarchus Plateau. AristarchusPlateau is one of the most heterogeneous areas ,_ 25 on the Moon. Highland-typematerials, mare basalts, anddark mantle deposits have previously been mentioned [2]. The mosaic represents a set of -500x 600nine-channel spectraUV- VIS filtersat 415, 7 50, 20 900, 950, and1000 nm were calibrated using the ISIS software [3]. '1. We applied the reductionmethod described in [4] to reduce the NIR ...1 15 filtersat 1100, 1250, 1500and 2000 nm. Absolute gain and offset ..d values were refined for the NIR images by using eight telescopic .... spectra acquired by [SJ as references. With thiscalibration test, we ..__,.. 10 Corrected Data were able to reproduce the eight telescopic spectra with a maximum errorof 1.8%. 0 R = 0.925 Effectsof Maturity on Extracted Spectra: The integration of eoJctd UV-VIS andNIR spectralchannels allows the visualization of com­ 5 R = 0.727 i:t: uneoJctd plete low-resolution spectra. In order to investigate the spectral u(/) effectsof the space-weathering processes, we focused our analysis 0 on a small mare crater and its immediate surroundings (see arrow 30 40 50 60 70 80 90 Fig 2a). Acco�ng to the small size of the crater (-2-km) andits IP GRS Fe Counts location on an homogeneous mare·area [2], we can reasonably assume that the content in FeO is homogeneous. The impact event Fig. 2. Scatter plot of LP GRS Fe data vs. CSR FeO data. Uncorected has induced a variation of the maturity of the soil by excavating LP data shown in blue; corected LP dat shown in black. freshmaterial. Figure la displays five absolute reflectance spectra extractedfrom this areaFigure 1 b displays the samespectra divided by a continuum, which is considered to be a right line fitting the spectraat 0.75 and l.S µm. Spectrum1 is extractedfrom the brightest part of the crater interior, and spectrum S is extracted from the surrounding mare material. Spectra2, 3, and 4 areextracted from if the discrepancy seen in SPA is real or is some effect of the intermediate distancesbetween the two areas. simplifieddata-reduction technique. Even so, we aredemonstrating The l-and-2 µm absorption band depths andthe overall reflec­ the ability to compare LP andCSR Fe datasets andshow that they tance increase fromspectrum S (correspondingto a mature area) to may indeed agree quite well. spectrum 1 (the most immature area). Conversely, the continuum References: [1] Davis P.A. (1980) JGR, 85, 3209. [2] Lucey slope decreases fromspectrum 5 to spectrum1 (Fig. 1 a). These three et al. (1998)JGR, 103, 3679. [3] Davis P.A. andSpudis P.O. (1987) spectraleffects of maturityhave also been identifiedon laboratory JGR, 92, E387. [4] Clark P. E. and BasuA. (1998) LPS XX.IX. spectraof lunarsamples [6]. [5] Feldmanet al. (1998) Science, 281, 1489. [6] Elphic etal.(1998) Depth of the 1-JllllFeature and ContinuumSlope: Most of Science, 281, 1493. the lunarsoils exhibit a signature near1 µm.This absorption bandis 40 Workhop on New Views of the Moon I

be mainly dominated by maturity effects, as suggested by the high (a) correlation with the independent evaluation of maturity (OMAT 0.20 parameter ) proposed by Lucey at al.[9]. We have also founda good correlation between the continuum slope and the OMAT parameter on laboratory spectra of lunarsamples of the J. B. Adamscollection (availableat http://www.planetary.brown.edu). 0.15 The discrimination between maturity effects and composition effects can be achieved by using a principal component analysis (PCA) on three spectral parameters,which arethe reflectanceat 0. 75 2 increasing µm, the depth of the 1-µm feature, and the continuum slope. These o.,o maturity parameters are mostly affected by maturity andFeO content. The 3 l effectsof various glass content are assimilated to maturity. The aim " of the PCA is to decorrelatethe FeO content and maturity effectsin 5 the three input parameters.Figure 2a displays the firstaxis of the PCA, which accounts for81.8% of the total variance.This image is 1.5 2.0 0.5 1.0 firstorder comparable with the OMAT parameter fromLucey [9]. *a�le"glh (microm,) at Figure 2b corresponds to the second axis of variation (14.4%) re-

1.10 (b)

E 1.

�0.90 5 4

J' OJI() 2

o.,o .._...... _._�_.__ ...... ___ ...... �--'-....._..._� -..., ..... 0.0 0,5 1.0 1.5 2.0 wa-le"glh (micrors) Fig. 2a. First axis of the principal component analysis. Impact craters Fig. la and lb. Spectra extracted from a fresh mare crater (see arrow and the walls of Valles Schroteri appear symmetrical. This image is at in Fig. 2a). I = crater interior(immature): 5 = surrounding mare (mature). first order very similar to the continuum slope image and is controlled Toe overall reflectance and spectral contrast decrease with increasing by maturity and glass content. maturity, while the continuum slope increases with maturity.

• due to the presence ofFe2+ in maficminerals such as orthopyroxene, (b) clinopyroxene, and olivine. In the case of Clementine UV-VIS data . •· alone, the depth of the 1-µmfeature is evaluated by the 950/750-nm reflectanceratio. This ratio combined to the reflectanceat 750 nm has been used to evaluate the global content in FeO of the lunarsurface .. C w. [7-8]. • Near-infrareddata makes a more precise evaluation of the 1-µm band depth possible by providing the right side of the band. The continuum in the vicinity of the bandcan be evaluated by anarith­ meticmean or a geometric interpolation of both sides of the band, which are taken at 750 and1500 nm. The geometricinterpolation is less sensitiveto residual calibration uncertainties.With this method, the 1-µm absorption banddepth for the AristarchusPlateau can be refined by as much as 10%. The differenceis maximum on Fe-poor, highland-typematerials. Fig. 2b. Second axis of the PCA. This image is very similar to the FeO Similarly,the NIRdata provide the possibility to investigatethe image derived from Lucey et al. [8], with less sensitivity to the continuum slopeof the spectra.The continuum slopeis a key param­ topography. Iron-poor highland type material is detected in the vicinity eter in anyspectral analysis.The continuum slope variationsseem to of the Aristarchus Crater (a), in the cobra head (b), and on Herodotus (c). LP/ Contriburion No. 980 41

vealed by the PCA. This image is very consistent with the FeO image Data: Clementine UV-VIS data at 1-l/pixel spatial resolu­ derived fom Lucey et al. [81, with less improvement due to less tion were placed in a orthographic projection centered on the bain sensitivity to the topography. (55°S, 180°E). Maps of Fe ad Ti abundace were produced using Conclusion: The integration of UV-VIS and NIR dataets the spctal algorithms of Lucey et al. [5]. A map of topogaphy allows fr a better understading of the spectal properties of the derived fom Clementine laer altmeter measurements was reteved luna surface by giving access to key paameters such as the 1- ad fom the Plaetary Data System Web site [61 ad reprojected to 2-µm bad depths ad the continuum slop. The continuum slope match the UV -VIS image ad elemental maps. can be combined with the depth of the mafic 1-µm absorption fature Findings and Discussion: We fnd a inverse corelation be­ ad the refectace at 750 nm to discriminate between maturity ad tween topograhy ad FeO content fr the bain interior (Fig. 1). The composition. NI images of the sample ret stations will be very corelation coefcient between elevation and [1/(FeO)l is about 0.7. interesting to refne absolute FeO content ad maturity evaluations. Figure 2 shows a typical eat-west compositional profle across the References: [11 Le Mouelic et al. (1999) GRL, 26, 1195. bain compared to a model profle generated by a linea fit to the daa [21 McEwen A. S. et al. (1994) Science, 266, 1858-1862. in Fig. 1. [31 McEwen A. S. et al. (1998) LPS XXIX, Abstract #1466. The occurence of greater FeO contents at greater depths (Fig. 1) [41 Le Mouelic S. et a. (1999) JGR, 104, 3833. [5] Lucey P. G. et al. could be explained in several ways. Te magma ocea model fr the (1986) JGR, 82, 34354. [61 Fischer E. M. and Pieters C. (1994) fration of the luna crst predicts a crst that becomes more Fe rich Icarus, 111, 475-488. [71 Lucey P. G. et al. (1995) Science, 268, with depth, so in SP A we may be seeing an exposed crstal column. 1150-1153. [8) Lucey P. G. et al. (1998) JGR, 103, 3679-3699. An undigerentiated melt sheet compsed of such a statifed crstal [9] Lucey P. G. et al. (1998) LPS XXIX, Abstact #1356. material is lower in FeO.

TOPOGRPHIC-COMPOSITIONAL RELATIONSHIPS WITH T SOUTH POLE AITKN BASI. P. G. Lucey, � P.M b itr J. Holtzmann, D. T. Blewett, G. J. Taylor, and B. R. Hawke, 15 Dparent of Plaeta Geosciences, Hawai' i Institute of Geophysics adPlaetology, University ofHawai'i, 2525 Corea Road, Honolulu HI 96822, USA ([email protected]).

Introduction: The South Pole Aitken (SPA) Bain is a im­ mense stcture that dominates the geology of much of the faside of 5 the �oon. Its foor is compsed mostly of impact deposits, though it also ha numerous relatively smal regions of mae basalt. The bain foor exhibits a lower albedo and higher mafc mineral abundace 0 than the suounding highlads [1). The origin of this mafc aomaly Elevation, m is a major queston in luna geology. Hypotheses fr the presence of the mafc anomaly were briefy reviewed in [2] ad include mae Fig. 1. Plot of FeO vs. elevation for the interior of South Pole Aitken depsits mixed ad obscured by basin or crater ejecta (cryptomaia), Basin. An approximate fit of the data is FeO (wt%) = [(-0.0011) a lage impact melt sheet that may have digerentiated, exposed lower (elevation in meters)) + 7.2. crstal material, ad a significat compnent of excavated mantle. A study of mineralogy as reveaed in Clementine UV-VIS imagery fr 20.�-.�- .�- .�-,�-. limited porions of the basin fund a predominatly low-Ca pyroxene (noritic) chaacter [21, ruling out cryptomaria as a importat con­

tibutorto the mafc enhacement. A fw smal cryptomaia, revealed 15 by dark-halo impact craters ad light plains units with high-FeO contents, have been fund in SPA [31; however, it appeas thb extensive crptomaiaae lacking in this basin. The unifrmly noritic lithology within SPA led [2) to fvor expsed lower crust or a homogenized melt sheet as the explaation fr the mafc anomaly. Models of basin frmaton predict that a basin the size of SPA should have excavated though the entire luna crst (assuming non­ oblique impact), ptentaly exposing or mixing a lage compnent of material fom the matle. Compaison of SP A foor FeO ad TiO2 (derved fom Clementine UV-VIS obsesatons) [41 ad aso T (fom Lunar Prospctor) [31 with model-matle chemistries appars E-W Prfile to be consistent with a mixture of approximately equal proportions Fig. 2. Typical compositional profile across SPA. The solid line is an of lower-crust ad mantle material. actual FeO profile. The dashed line is model FeO generatedby applying In the present study, we examine the relationship between the a linear fit of the datain Fig. 1 to the topographic profile. Verticaldotted bain's topography ad compsiton in order to provide further lines show the approximate location of the SPA Basinrim. SPA is -2500 insight on the orgin of the bain foor material. km in diameter. 42 Workhop on New Views of the Moon I

Further speculation. The excavation of the SPA cavity would missing data. Marehave numerous features,that are resolved in fast have been a massive unroofing event that could have triggered neutrons.For instance, the region extending northwest of Aristarchus significantmafic igneous activity through pressure-release melting. (23. 7°N, 4 7.4 °W) is clearlyseparated from Montes Harbinger(27 .0°N, The melts that were generated would tend to eruptand or collect in 41.0°W) by a high-emission channel, and Mare Vaporum(13.3 °N, the low-elevation centralportion of the basin. At higher elevations, 3.6°E) is separated from Sinus Aestuun (10.9°N, 8.8°W) by a low­ progressively smaller quantities of magma and greater impact dilu­ emission area. tion with underlying more-felspathic rocks might lead to the ob­ Moon Composition: We present a new technique to extract servedcorrelation between elevation and Fe content. It is interesting information on soil composition from the fast-neutron measure­ to note that KREEP basalts, which may have originated in a similar ments. The analysis isapplied to the centralmare region. There are manner in the Imbrium Basin [8], possessFeO of roughly 12 wt%, two steps forthe development of the technique. comparableto much of the floorof SPA. 1. For the first step, which has been fully completed [3], we References: [1] Belton M. et al. (1992) Science, 255, 570. assumethat variationsof fast-neutroncounting rates are due solely

[2] Pieters C. M. et al. (1997) GRL, 24, 1903. [3] Blewett D. T. et al. to TiO2 andFeO. Upon this assumption, we correlate Clementine (1999) LPSXXX, Abstract#l438. [4] Lucey P. G. et al. (1998)JGR, SpectralReflectance Fe andTi oxide maps [4] with fastmeasure­ 103, 3701. [5] Lucey P. G. et al. (1999) JGR, submitted. [6] Zuber ments. Above 16.5% of FeO, effects of TiO2 variationsshow in LP M. T. et al., http://pds-geophys.wustl.edu/pds/clementine/. data. Below 6.5% of FeO, Fe cannot be discriminated; this is the [7] Morrison D. (1998) LPS XX.IX, Abstract#1657. [8] Ryder G. region of most highland terrains. _(1994) Geo/. Soc. Am. Spec. Paper 293, 11. Under assumption of only two oxides to modulate the signal, we show that fastcounts are3.2x more sensitive to FeO thanto Ti 02. The resolution in FeO is 1.2 wt% , in TiO2 it is 3.8. These results arevery HIGH-ENERGY NEUTRONS FROM THE MOON. satisfying,specially forthe distributionofFeO. However, they do not 1, 2, 2, 2, S. Maurice W. C. Feldman D. J. Lawrence R. E. Elphic permitreproduction of Clementine TiO2 map fromthe residual of the 3 3 0. Gasnault , C. d'Uston , and P. G. Lucey4, 1Observatoire Midi­ fast counting rates and Clemetine FeO correlation.Particularly, the Pyrenees, Toulouse, France([email protected]), 2Los Alamos discriminationbetween hi-Ti and low-Ti mare is not striking. NationalLaboratory, Los Alamos NM, USA, 3Centred'Etude Spatiale 2. The second step is still under development. We assume that des Rayonnements, Toulouse, France, 4Hawai'i Institute of variationsof fast-neutroncounting rates areprimarily due to FeO, but

Geophysics and Planetology, University of Hawai'i,Honolulu HI, also to Ti 02, Si0 2, CaO, Al203, MgO, and Na2O. Gasnault et al. [5] USA. have calculatedthe number of fast neutrons in our energy band of interest: Introduction: Galactic cosmic rays that impact the lunar soil produce neutronswith energies from fractionsof eV's to -100 Me V. Na 1.28 Mg 1.68 The high-energy band from 0.6 to 8.0 MeV is referred as the "fast Al 2.09 Si 2.26 neutron" band, which is measured by Lunar Prospector (LP) Gamma­ Ca 1.75 Ti 3.02 Ray Spectrometer [1). Fe 2.81 0 0.165 Fast neutronsplay an importantrole in neutron spectroscopy [2] per weight fractionof each element, foran average cosmic-ray flux. that may be summarized as follows: Fast neutrons definethe total This simulation is for a FAN soil. It will have to be refined and neutron input to the moderating process towardlow-energy popula­ iteratively adapted to the soil composition. tions, so that epithermaland thermal neutron leakage currents must With informationon TiO2 and FeOdistributions fromClementine be normalized to the leakage of fastneutrons; they allow the deter­ andthe coefficientabove, we know the globalsoil content in theother minationof the burial depth of H, a measure necessary to understand oxides (weighted by Gasnault et al. coefficients).On the other hand, characteristics of water deposits; they provide information on the we have fromreturn samples estimates of correlationbetween oxide surface content in heavy elements, such as Ti and Fe; and they concentrations. We demonstrate that such processing allows estima­ provide a direct insight into the evaporation process. As discussed tions of SiO variationsin the lunar regolith. hereafter, fast neutrons mayyield informationon other oxides, such 2

as SiO2. Datasets and Data Processing: This study uses fast neutron -BO data fromDecember 20, 1998 to May 22, 1999, when the spacecraft altitude was between 10 and50 km. The data reduction includes - 460 normalization to the cosmic-ray rate, to the spacecraft height and latitude, and finally to the gain drift. .! 440 From all-Moon maps at a 2°x 2°-equal-arearesolution, we learn the following: (1) All mareidentified, from the largestones, Imbrium, 1 420 Frigoris,Fecunditatis, to the smallestones, Undarum,and Spumans, ' are high-neutron emitters; (2) there are extended regions, such as 400 South Aitken Basin, or the area around Schickard, which are me­ dium-to-high-neutron emitters; (3) highlands are definitively low 30 fast neutronemitters, which formthe blue-background of the map; 0 5 10 15 20 and (4) the largestcraters areresolved in fast neutrons. FeO (l) Figure 1 is a map of the centralmare region. White data represent Fig. 1. LP! Contribution No. 980 43

References: [l] Feldma et al. (1999) NIM, 422, 562-566. compositional compaisons allow their source regions to be com­ [2] Feldman et al., Science, 281, 1489. [3] Maurice et al. (1999)JGR, paed. Highly siderophile elements Au and 1r ae more abundat in submited. [4] Lucey et a!. (1998)JGR, 103, 3679. [5] Gasnault et al. the glasses relative to the basalts (Fig. 2). As these elements ae (1999) JGR, submitted. generally incompatible in silicate minerals, crysta factionation ex­ perenced by the basalts will tend to increase the Au ad 1r abun­ daces. Therefre, the glasses may be derived fom a souce enriched THITERIOROFTHMOON,CORFORTION,AN in highly siderophile elements such a the platinum-group elements TH LUNAR HOTSPOT: WHAT SAMLES TELL US. (PGEs) represented by lr, relative to the source of the baalts. Tis C.R. Neal, Department of Civil Engineering and Geological Science, obseration can be acomodated with the baalts being derved University of Notre Dame, Notre Dame IN 46556, USA fom the LMO cumulates ad the glases derived fom a source tat (Neal. [email protected]). represents "primitve Moon" that did not melt and, therefre, did not have its budget of PGEs ad Au reduced through core frmation. Itroduction: Remotely-gathered Lunar Prospctor data have demonstated the existence of a luna "hotspt" on the nea side of the Moon. This hotspt contains relatively high abundances of KEEPy incompatible tace elements (i.e., Th). It is generally ac­ cepted that prmordial KEEP or urKREEPrepresents the residua liquid after the crystallization of a luna magma oea (LMO) [e.g., 1-2]. The crystalline products fom the LMO fred the source 1000 regions fr the mae basalts [e.g., 3]. Luna volcaic glasses canot be genetcally related to the crstalline mae baalts [e.g., 4-5], ad exprimental petology indicates they ae derived fom greater (> km) depths than the mae basalts [e.g., 6]. Questions to be addressed 100 include: (1) What was the extent of LMO meltng (whole vs. patial - Moon melting)? (2) Wha is the compsiton of the core? (3) Are Glasses there distnct geochemical reseroirs in the Moon? (4) Is there ■A Apollo 11 v Apollo 12 evidence of gaet in the lunar interior? (5) What caused the fra­ 10 ton of the luna hotspt? ◊ Apollo 14 + Apollo 15 Whatwas the extent ofLMO melting and what is the composition X Apollo 17 of the lunar core? The scale of the LMO has been suggested to be E VLT Baslts whole Moon melting [cf. 7-9] or only the outer -400 km [cf. 2]. If 1'---����-����---��- whole Moon melting is invoked, then diferentiation of the Moon 0.1 10 100 into a fotaton plagioclase-rich crst, a mafc mineral cumulate Hf (ppm) matle, ad a Fe-rich core is more easily facilitated. However, as pointed out by [9], if the maeral that frmed the Moon came Fig. 1. primaily fom the already-diferentiated Earth matle, there may not be enough Fe to fr a metallic Fe core on te Moon. Tese authors suggested that the luna core is made up of dense, ilmenite-rich, late­ stage cumulates fom the LMO. Ths ca be tested by examining the O A17 Orange Glass (74220) Zr/Hf ratos of mae basalts and, where pssible, the volcanc glases. □ A15 Green Glass Patition coeficients fr Zr ad Hf in ilmenite ae 0.29-0.32 ad ◊ A15 Yellow-Brown Glass ♦ A 11 Basalts 0.4-0.43, respctively [cf. 1O], with Zr being less compatible. Tere­ 10 - ■ A12 Basalts fre, extaction of an "ilmenite" core would have a profund efect , A 14 Basalts • A15 Basalts on the Zr/Hf ratio of urKEEP a ilmenite is a late-stage factionat­ • A 17 Basalts ing LMO phae. Assuming either a "primitive matle" or chondritic 0 0 stang material with aZr/Hf raio of 36-37, ilmenite extacton will ♦ increae this ratio in the residual liquid. Conversely, derivation of a 0 melt fom a source rich in ilmenite will produce a melt of lower Zr/ 0 □ Hf rato. Hughes ad Schmitt [ 11] defned a mea Zr/Hf fr KEEP a 0.1 .... • .... of 41.0 ± 0.4, -39 fr Apollo 15 basalts, ad 30-32 fr Apllo 11, • ■ • 12, ad 17 basalts, wit the decreases in Zr/Hf broadly corelating .... ♦ with Lab. However, literature daa fr Apollo 15 KEEP basalts ♦ ... ■ 0.01 •• ...... ad the KEEP-rich Apllo 14 mae basalts exhibit little varation ... in Zr/Hf fom 36 (Fig. 1), indicatng te KEEP compnent dd not result fom a major factonation of ilmenite ad suggesting that the 0.01 luna core is probably metallic in overall compsition [e.g., 12]. 0.001 0.01 0.1 Are there distinct geochemical reservoirs in the Moon and is Ir (ppb) there evidenceof garnetin the lunar interior? With volcaic glases being unrelated to the mae basats ad derived fom greater depths, Fig. 2. 44 Workhop on New Vews of the Moon /

This can be tested by analyzing mare basalts and glasses for the Prospector mapping has identifiedrelatively high Th abundancesin PGEs. Although analytically challenging, the first PGE patterns in this area, suggesting a large KREEP component is present at or near lunarsamples were reported by [13 J anddemonstrated that the source the surface. LMO "layer cake models" have residual urKREEP regions forthe differentApollo 12 basalts could not be differentiated sandwiched between the maficcumulate mantle and the plagioclase on the basis of PGE budgets (Fig. 3), although the profilesare typical flotation cumulate crust. However, late-stage cumulates and the of silicate melts [e.g., 14]. Analysis of other trace-element data residual liquid will be more dense that the early mafic cumulates indicate that the high-field-strength elements canbe used to differ­ resulting in gravitational instabilities and overturnof the cumulate entiate between high- and low-Ti basalts. Also, the volcanic glasses pile [e.g., 21-23). This could transporturKREEP to the baseof the were derived froma source with a higher Zr/Y ratio (Fig. 4) relative LMO cumulate pile, but above the glass source region. The effect of to the basalts, consistentwith retention of garnetin the residue. If the Earth on the symmetry of the Moon has displaced the low-density glasses were derived from >400km, garnet could be stable [see 15 crust,producing a thicker crust on the farside [e.g., 24-25). This has andreferences therein]. It is concluded that the volcanicglasses were produced anoffset of the center of mass forthe Moon towardEarth. derived froma source that contained garnet,but escapedthe melting It is suggested that the gratitational forces of the Earth pooled the that formedthe LMO. The mare basalts were derived fromthe LMO urKREEP beneath at the base of the LMO on the lunar nearside. cumulate pile. Heating through radioactive decay produced thermal instabilities, Whatcaused the formation of the lunar hotspot? Basalticsamples resulting in a plume of hot, KREEPy material rising adiabatically fromApollo 14exhibitarangeinITE [e.g., 16-18).They also exhibit beneath the Apollo14 site. The oldest Apollo 14 basaltscontain no a range of ages from4.33 Ga to 3.96 Ga with the older basaltsbeing evidence of a KREEPy component [cf. 26), suggesting diapiric rise KREEP-poor and the younger being KREEP-rich [e.g., 19-20]. of the KREEPy plume had not occurredat this time. However, by 4.1 Ga, isotopic andtrace-element systematics of the Apollo 14 mare basalts reflectthe inclusion of a KREEP component, which wasmore evident in those basalts erupted at 3.95 Ga [26). References: [1] Smith et al. (1970) 897. 0.01 ------�-� Proc. Apollo 11 LSC, -.-12009,130 [2] Warren(1985) Annu. Rev. EarthPlanet. Sci., 13, 201. [3]Taylor ---\'.l\�'.2�fall andJakes (1974) 1287. [4) Delano (1986) .! Proc. LSC 5th, JGR, 91, ·;:: --.-Ywr;'.2�asa11 D201. [5] Longhi (1987) JGR, 92, E349. [6] Green et al. (1975) "'0 p;,_,.,;,eBasaH C Proc. LSC6th, 871. [7] Hess and Parmentier(1993) LPS XlV 651. 0 [8] Hess and Parmentier (1994) 541. [9] Hess and .C 0.001 LPS XV, 0 Parmentier(1995) EPSL,134, 501. [10) McKay et al. (1986) JGR, 91, D229. [11) Hughes and Schmitt (1985) JGR, 90, D31. [12] Newsom (1984) Eos Trans AGU, 65, 369. [13] Neal et al. as Ch (1999) LPS XXX, Abstract #1003. [14] Barnes et al. (1985) Chem. Geol., 53, 303. [15) Taylor (1982) Planetary Science: A Lunar 0.0001 Fig.3 Perspective, 481 pp. [16) Dickinson et al. (1985) JGR, 90, C365. [17) Neal et al. (1988) Proc. LPSC 18th, 139. [18] Neal et al. Ir Ru Rh Pt Pd (1989) Proc. LPSC 19th, 147. [19] Taylor et al. (1983)EPSL, 66, 33. [20] Dasch et al. (1987) GCA, 51, 3241. [21] Hughes et al. (1988) Fig. 3. GCA, 52, 2379. [22) Snyder et al. (1992) GCA, 56, 3809. [23] Spera (1992) GCA, 56, 2253. [24] Kaula et al. (1972) Proc. LSC 3rd, 2189. [25] Kaula et al. (1974) Proc. LSC 5th, 3049. [26] Neal andTaylor (1990) Proc. LPSC 20th, 101. 140

120 ■ ASYMMETRICEVOLUTIONOF THEMOON:A POSSIBLE CONSEQUENCE OF CHEMICAL DIFFERENTIATION. 100 ■ E. M. Parmentier1, S. Zhong2, and P. C. Hess1, 1Department of Geological Sciences, Brown University, Providence RI 02912, USA, 80 2Department of Earth, Atmospheric and Plantary Sciences, ■ Massachusetts fustituteof Technology, CambridgeMA 02139, USA. ;;.. 60 Introduction: fuits finalstages of cystallization,a lunarmagma Glasses 40 ■ ocean that creates the anorthositiccrust will also precipitate ilmenite­ V Apollo 1, andpyroxene-rich cumula tes (IC). These cumulates will be signifi­ Apollo l' 20 + cantly denser than earlier olivine-pyroxene cumulates (OPC), and X Apollo li buoyantly driven, thermally activated creep should allow them to 0 sink into underlying cumulates. Since incompatible elements includ­ 0 100 200 300 400 500 600 ing heat-producing U and Th will also concentratein the late-stage Zr(ppm) magma oceanliquids, the gravitational differentiationof the ilmenite Fig. 4. cumulates has potentiallysignificant chemical andthermal implica- LP!Contribution No. 980 45

tions fr the subsequent evolution of the Moon [1,2]. These include The size ofIC bodies in the mixed layer is, however, proportiona a deep, high-pressure origin fr melts that may represent prmitive to the squae root of the IC solidifcation rate. This is importat fr mare baalt magmas [3] fom source materals that should have understading the physica chaacter of the chemically heteroge­ crystallized at much shallower depths in the fna stages of magma neous source of mare baalts [I]. ocean solidifcation [cf. 2]. Transporting some faction of heat­ The wavelength of mixed-layer instability is examined on the producing elements to depth and creating a stable chemical stratfi­ basis of a simple model of Rayleigh-Taylor (RT) instability of a dense cation that traps radiogenic heat in the deep interior may help to layer with viscosity m,ayer surounding a viscous sphere of lower explain the existence of elastic lithosphere thick enough to support density and diferent viscosity Interor- The radius of te sphere ad mascons [cf. 45) and the retenton of heat needed to reduce the surounding denser laer coresponds to the radius of the Moon. The aount of global contaction due to cooling [6]. RT growth time fr a rage of wavelength (spherical haonic Hemispheric Asymmetries: The hemispheric asymmetry of degree) ae shown in Fig. 1 fr a layer thickness of 100 km, a mixed­ 2 1 surface distibution of mae basalts [cf. 7) is also fundaental to layer density coresponding to 20% IC [2], 1nteror =10 Pa-s, ad 3 understading the evolution of the Moon. The appaent corelation a range of m1ayer /m;nterior values. If m1ayer /m;nterior is less tha -10- , between mare ad surface elevation had prompted the idea that mare spherica haonic degree 1 grows more rapidly than the shorter­ basalt floded topgraphically low areas. However, Clementine to­ wavelength, higher-spherical haonic degees. Thus, in frther pographic data [8] indicate that while mare basalts clealy fill low assessing this model we need to undestad the conditons fr which aea, lage areas of low elevaton do not contain mae basalt. A OPC just beneath the IC layer will have a viscosity suffciently lower possible explaation fr the distibution is that it arises fom the tha that of OPC at greater depth. hydrodynamic instability of a chemically dense IC core or layer that The RT growth time, aprximatng the time fr a downwelling 2 becomes buoyant as is heated by radioactivity [13]. to fr, is proportiona to fnteror- Even fr Iienor a hgh as I02 A radial distibution of Th aound the Imbrum Basin is aother Pa-s, a downwelling could fr in times on the order of 10 m.y., important aspect of luna asymety that may be explained by the which may b short compaed to that required fr the fnal stages of excavation ofaKREEP-richlayerfom beneah a regionoftheMoon magma-oean solidifcation. Note that this intoduces a complica­ surounding Imbrium [9,10). Other lage basins do not appea to tion to the simple model above, which asumes that te mixed layer excavate compaable aounts of Th-rich material. High Th aso is already fred as the large-scale downwellig develops. occurs in the baalts of Mare Procellarum, leading to the suggestion Further Iplication fr Lunar Evolution: This model aso KREEP is prefrentally concentated beneath this region of the maes predictions fr the geochemical heterogeneity of the Moon. surface [9,11). Chemical factionations that occur during te solidifcation ofIC will The existence ofrelatively shallow concentations ofKREEP and infuence not only the concentaton of heat-producing elements that the later erption of mare basalts in the same region of the Moon it contains [12), but also the subsequent isotopic evoluton of the suggests a causal connection. Gravitatonal instability of the mixed mare basalt source. It is imprtant to understand if the predictions of layer a a single downwelling might explain this asymety [2]. If this model ae consistent with observed geochemical signatures. this downwelling fred prior to the complete solidifcation of the magma ocea, the residual liquid layer would deepen above the down welling. This would draw residual liquid towad thedownwelling ad may explan a high concentaton of KREEP-rich materials in the Procellam region. After sufcient time fr expasion due to heat­ ing, the buoyat rise of previously downwelled, IC-rich matle could then account fr the presence of mae basalts in this same region. Fluid Dynamics of Digerentiation: For an IC layer only 20- 30 km thick, it is remakale that such a lage-scale downwelling would develop. As the fnal stages of magma-ocea solidifcation proceed, we envision that dense IC solids will collect at the bottom of the residual liquid layer ad sink into the underlying OPC. Tis will create a mixed layer of IC and OPC that is denser tha the underlying OPC. Gravitational instability of this mixed layer, if it is substantially thicker tha a pure IC layer, may then account fr the layer thickness= 100 km long wavelength down welling needed to explain the obsesed asym­ 21 meties [2]. µinterj,= IO p a-s We have previously estimated the size of the sinking IC bpies, } o•I ....,.__....., ____ ..._..__.._.....,_._..._,..._..,_..._..._.,_,. how deep they sink over the time that IC solidifes, ad te scale of 5 IO IS 20 gravitational instability of the mixed OPC and IC layer thus created spherical harmonic degree [2,10]. We fnd that the thickness of the mixed layer is independent of the rate of IC solidifcaton. It is proprtional to the thickness that a IC layer would have if no instability occued ad depends but Fig. 1. Rayleigh-Taylor growth time as a function of spherical harmonic only wealy on the rato of the IC and OPC viscosities. We fnd a degree for a dense layer 100 km thick with viscosity m1,yer at the outer mixed layer that is about 5x t thickess of IC alone, ad which surface of a 1600-km radius sphere with viscosity lilu,1erior- For m1aye/ therefre contains -20% IC. It is thus denser t underlying OPC IIlu.tcrior less than about 10-3, spherical harmonic degree 1 is the fastest ad gravitatonally unstable. growing mode of buoyant instability. 46 Workhop on New Views of the Mon l

References: [1] Ringwood A.E. and KessonS. E. (1976) Prc. lower spectral resolution than Earth-based spot spectra, covered LPSC 7th, 1697. [2] Hess P. C. and Pannentier E. M. (1995) EPSL, large areas of the Moon and used filters at wavelengths useful for 134, 501. [3] Delano J. W. (1986) JGR, 91, 201. [4] Pannentier determining the lithologies present. These spacecraft data have also E. M. andHess P. C. (1998) LPS XXIX, Abstract#1182. [5] Alley been used to determine the abundance of FeO and TiO2 present in K. M. and PannentierE.M. (1998)PEPI, 108, 15. [6] Solomons. C. lunar surface materials. Other products, such as band-ratio maps, and Chaiken J. (1976) Proc. LPSC 7th, 3229. [7] Lucey P. G. et al. ,have been produced, and spectra have been extracted from co­ (1998) JGR, 103, 3679. [8] Zuber M. T. et al. (1994) Science, 266, registered image cubes. 1839. [9] Haskin L. A. et al. (1999) LPS XX, Abstract #1858. Lunar Prospector has collected a large quantity of y-ray and [10] Haskin L. (1998) JGR, 103, 1619. [11] Lawrence D. et al. neutron spectrometer data. While much of the data will require (1998) Science, 281, 1484. [12] ParmentierE. M. and Hess P. C. furtherprocessing beforereliable quantitative interpretationscan be (1999) LPS XXX, Abstract #1289. [13] Zhong S. et al. (1999) LPS made, some data from that mission have already been made available. X, Abstract #1789. In particular,they-ray spectrometer countingdata for Th, K, and Fe [ 4 J can be used to confirmand extend our knowledge of the compo­ sition of the lunarfarside crust. A preliminaryTh distribution map THEDISTRIBUTION OF ANORTHOSITE ON THELUNAR has been produced from the raw data by utilizing ground truthfrom FARSIDE. C. A. Peterson1, B. R. Hawke1, P. G. Lucey 1, G. J. the lunar landingsites [5]. Taylor !, D. T. ·Blewettl, and P. D. Spudis2, 1Hawai'i Institute of Results: Noritic anorthosite and anorthositic norite are the Ge0physics and Planetology, University of Hawai'i, Honolulu HI predominantrock types at the surface of the nearside lunar highlands. 96822, USA, 2Lunar and Planetary Institute, Houston TX 77058, Lesser amountsof anorthosite,norite, troctolite,and gabbroic rocks USA. arealso present [2]. Studies of Earth-basedreflectance spectra ini­ tially revealed the presence of anorthositein isolated outcrops ex­ Introduction: There is much evidence to support the hypoth­ tending in a narrowband from the Inner Rook mountains in the west esis that a giant impact on earlyEarth created the Moon and that a to the crater Petavius in the east [e.g., 6]. More recently, additional magma ocean was present on the young Moon [e.g., 1-2). As the outcrops of anorthosite havebeen identified in the centralpeaks of magma cooled and crystallized, plagioclase flotation could have some craters, such as Aristarchus,and in thenorthern and northeast­ produced the upper partof the Moon's original crust. But how much ern nearside[7,8). In most cases, these anorthosite deposits have of this original crust has survived to the present? Has it been entirely been exposedby impacts that removed a more maficoverburden and disrupted,or do portions remain relatively unchanged?Remote sens­ raised them to the surfacefrom deeper in the crust,for example in the ing studies of the lunarhighlands, combined with analysis of lunar peakrings of the Orientale, Humorum,and Grimaldi Basins[9-11). materials returned fromknown locationson the surface of the Moon, Much of the nearsidelunar highlands likely sharesthis stratigraphic have allowed the determination of the lithologies present in many sequence: a layer of pure anorthosite overlies a more mafic lower locations on the Moon. Our study of the distribution of the various crustand is in tum overlain by a somewhat more maficlayer. Much lunar-highlandrock typeshas revealed large-scalepatterns that sug­ anorthositelikely remains hidden by thissurface layeron the nearside gest the broad outlines of the evolution of portionsof the lunarcrust. today. Our previous efforts have used Earth-based spectra and Galileo On the lunar farside, the giant SouthPole Aitken (SPA) Basin SSI to study the lunarnearside. We have used Clementine UV-VIS shows a maficanomaly that is the dominant compositional featureon data to extend our studies of the lunar highlands to the farside. the farside. At 2500 km in diameter,SPA (centered at 55°S, 180°E) Calibration of the Clementine UV-VIS data is essentially complete, is the largest unambiguously identified impact basin in the solar and FeO andTiO 2 values derived from the Clementine data have system. It is also the oldest identifiedlunar basin except possibly for been derived fromthis well-calibrated data. In addition, Lunar Pros­ Procellarurn. There is a 13-km difference in elevation from the pector data are now available in preliminary formand can add to our interior of the basin to the surrounding highlands[12]. The interior understanding of the composition of highlands units on the farside. exhibitsa 7-10% FeO enrichment relative to the surroundinghigh­ We can use this combined dataset to (1) study the composition of lands,and portions of the interior exhibit enhancedTiO 2 values [13 ]. farside highlandsunits; (2) identifyand determine the distribution of Portions of the interior also display elevated Th and K abundances. anorthositeon the lunarfarside; and (3) investigate the stratigraphy The high-FeO values foundinside SPA drop offwith increasing of the farside crust. distance to the north of the basin. The region between 100°E and Method: The great majority of the Moon's highlandssurface is 100°W andbetween 40°N and 70°N exhibits very low values of FeO composed ofonly afew minerals, andthese areeasily distinguishable andTi 02• Thenew LunarProspector data show very low-Th and-K using reflection spectroscopyat wavelengths from the UV through abundances there as well. FeO maps produced from high spatial visible light and into the near infrared [e.g., 3]. The mafic minerals resolution Clementinedata reveal extremelylow FeOvalues near the pyroxene and olivine contain Fe thatcauses the minerals to absorb crater Fowler (43 ° N, 145° W) in the vicinity of the Coulomb-Sarton light with a wavelength near1 µm. In contrast, plagioclase feldspar Basin. We interpret these data as indicating a region in which pure does not absorb light near 1 µm, although plagioclase can show anorthositeis dominant.Many areas in this region appearto contain absorption of light near 1.25 µm if it has not been highly shocked by nothing but anorthosite. impacts. Between SPA andthe farnorth, the data generally indicate inter­ Through the use of Earth-basedtelescopic reflectancespectra, it mediate FeO values andlow Thvalues. However, some lower-FeO is possibleto determine the lithologies present in the areaobserved, values can befound in this intermediate region. In particular, very typically from 2 to 6 km in diameter. The Galileo and Clementine low FeO valuesare exhibited in the inner rings of the Hertzsprung spacecraftreturned multispectralimages of the Moon that, while of and Korolev Basins. This case parallels the situation at nearside LP/Contribution No. 980 47

basins, where anorthosite was exposed from beneath more mafic What color is the Moon? A reflectancespectrum is essentially a material. measure of how much radiation incident on a surface (solarradiation) Discussion: The strikingdifference betweenthe distributionof is reflectedand how much is absorbed at each wavelength. To the eye lithologic types on the lunarnearside and that on the farside appears the Moon is gray-white,but to photoelectricinstruments it is various to be largely attributableto the enormous SPA impact event. Huge shades of red - that is, it exhibits an increase in brightness with quantities of ejecta would have been deposited outside the basin, wavelength.In the near-infrared there are absorptions diagnostic of thickest nearthe basin rim andtapering off with increasing distance minerals superimposed on the Moon's redness. For the Moon and fromthe basin.This ejecta, especiallythat deposited nearest the basin other rocky bodies such as asteroids,most of the detectable absorp­ rim, must have contained much material from the lower crust and tions arise from ferrous iron in various crystallographic sites. The possibly even some mantle material. If the original plagioclase flo­ wavelength, shape, and strength of these absorptions identify the tation crustwas largelyintact at the timeof the SP A impact event, the minerals present, and allow their abundances to be estimated. great thickness of more mafic ejecta deposited near the basin could Accurately measuring these diagnostic mineral absorptionswith have insulated that originalupper crustalmaterial (anorthosite) from remote detectors requires not only a quality instrument, but also all but the largest subsequent cratering events. Large impacts, such excellent electroniccalibration andeither direct measurementof the as those that produced the Orientale, Hertzsprung, and Korolev light source (the Sun) or a proxy, or a well-known referencestandard Basins, could have penetratedthrough the SPA ejecta deeply enough illuminated by the same light source. In the laboratory a white to expose anorthositein their peakrings. Fartherto the north, where referencesuch as halon (or commercialSpectralon), is used which the thickness of ejecta fromSPA wasmuch less, small basinssuch as in turnhas been extensively calibrated relativeto a known radiance. Birkhoff and Coulomb-Sarton could have removed most of the In space, or at the telescope,a separatereference must be found overburdenof SP A ejecta to revealthe underlying anorthositecrust. to mimic solar radiation and to eliminate instrumental and atmo­ If this scenario is accurate, it has great implications for the spheric effects.Radiation fromstars to a firstorder follows a black geologic history of the Moon. It is conceivable that largenumbers of body spectral curve with multiple emission and absorption lines very large impacts preceded the SP A event and are not seen today superimposed.Because stellar lines vary withspectral type, and very simply because evidence of them has been erased by subsequent fewstars are really solarlike, it is actually quite difficultto use stars impacts. However, that possibility is argued against if most of the asspectral standards. original plagioclase flotation crust was still present on the lunar The Moon is a nearby atmosphereless body that reflects solar farside at the time of the SP A impact event. This has important radiation. Because the Moon's surface itself does not change with bearing on the questionof the fluxof largeimpactors during the early time (at least within our lifetimes),it provides an excellent reference history of the inner solar system. standard.The calibration challenge thenreduces to identifying an The Lunar Prospector data must be further calibrated and re­ area on the Moon whose measurable properties are exceptionally duced, but the preliminaryresults have already added greatly to our well known. The returnof lunar samples allows their properties to be understandingof the Moon. It is encouraging that groundbased data, measured accurately in Earth-basedlaboratories. Since the samples Clementine data, andLunar Prospector data are providing a consis­ were collected from known areas on the surface of the Moon, care­ tent picture of the composition of the lunar surface. The complemen­ fully selected samplescan be used to represent the properties of that tary nature of the datasets enhancesthe value of each set.We eagerly area await the additional insight that may be provided by the full set of For Clementinedata, the Apollo16 site was chosen as a calibra­ Lunar Prospector data. tion targetbecause it is a relatively homogeneous areawith no nearby References: [1) HartmannW. K. et al., eds. (1986) Origin of units of a significantlydifferent material. Since all remote data are the Moon, LPI. [2)HeikenG. H.etal.,eds. (1991)LunarSourcebook, acquired as bidirectional reflectance, we use i = 30°, e = 0° as the CambridgeUniv. (3) Pieters C. M. and Englert P.A. J., eds. (1993) standard geometry.The calibration steps used and the assumptions Remote Geochemical Analysis, Cambridge Univ. [4] Lawrence D. J. made are discussed brieflybelow. etal. (1998) Science, 281, 1484. [5] GillisJ.J. et al.(1999)LPSXX, The absolute spectral calibration procedures forClementine data Abstract #1699. [6] Hawke B. R. et al. (1992) LPS XI, 505. include the followingsteps (see Table 1 foran estimation of errors): [7] Tompkins S. and and Pieters C. M. (1999) MAPS, in press. [8] Spudis P. D. et al. (1997) LPS XXVJ, 1361. (9) Hawke B. R. et (a) Select Apollo 16 site calibrationstandard target: a 12 x 33 pixel al. (1991) GRL, 18, 2141. [10) Hawke B. R. et al. (1993) GRL, 20, areato the west of the landingsite freeof crater rays. 419. (11) Peterson C. A. et al. (1995) GRL, 22, 3055. (12) Spudis (b) Select maturesoil representative of the Apollo16 site: 62231. P. D. et al., Science, 266, 1848. [13) Blewett D. T. et al. (1999) LS ( c) Obtain bidirectionalspectral reflectance measurement of 62231, XXX, Abstract#1438. �2231()1.,i,e), at standardgeometry (i = 30, e = 0): (cl) measure brightnessof 62231 relativeto Halon: B6223 i(.A.,30,0); (c2) obtain/derivecorrection for Halon reference standard: Rh(.A.); THEMOON AS A SPECTRALCALIBRATION STANDARD (c3) R62231(.A.,30,0)=B62231(.A.,30,0)*Rh(A); an average spec­ ENABLED BY LUNAR SAMPLES: THE CLEMENTINE trumis shown in Fig. I. EXAMPLE. C. M. Pieters, Departmentof Geological Sciences, (d) Translate 62231 laboratory spectrum into Clementine 5-filter Brown University, ProvidenceRI 02912, USA. spectrum to obtain Apollo 16 "ground truth" data for Clementine filters, �2231c(A, 30,0): Spectralcalibration of Clementinedata relies on the Apollo16 (dl) obtaineffective Clementine filter transmission curves, F(l); site andlaboratory measurements of mature soil 62231. The process (dl 1) measurethe transmissioncurve of each filter: f(A); produces calibratedspectral reflectance factors. (d12) estimate the solarspectrum: S(A); 48 Workhop on New Views of the Moon I

(d13) estimate detector sensitivity: D(A); ( di 4) F(A)=f(A)*S(A)*D(A)*normalization. 0.35 Ap16 6231 standard (d2) R62231c(A, 30,0)=R62231(A, 30, 0) * F(A). Clementine filters Clementine Bandpasses are shown in Fig. 2 and reflectance values are listed in Table 2. 0.. (e) CorrectClementine DN values forall gain, offset,and flat field !0.25 C corrections(including row, column, andtemperature-dependent vari­ CII 0.2 ables) [1-2]: DNciemCl,i,e). j ! (f) Photometrically calibrateClementine DN values to i=30, e=0 [ 1] .0.15 (the Moon is redder at larger phase angles): DNPC!emCA,30,0)= 0.1 DNcicmCA,i,e)*Cphot (A,i, e). (g) Derive spectral calibration correction factor Cr(A) using the 0.05 laboratory data for soil 62231 and the Clementine calibrated DN measurement of the Apollo 16 site: Cr(A)=R (A.,30,0)/ 6223IC 30 ,o so 600 100 eoo 900 100 1100 DN ( ,30,0). PClemApl6 A. waveengh (nm) (h) Derive absolute Clementine reflectance factor#data R(l, 30, 0) forany lunar area: R(A, 30, 0) = DNPClcmCA,30,0)*Cr(A) #reflectance Fig. 2. Average bidirectonal spectm fr 62231 used as a calibration factor [3] = reflectancerelative to a Lambertian surface under the stndard. Ultaviolet-visible Clementne bandpbses ae superimposed. same illumination.(Same as reflectancecoefficient [3] and radiance

TABLE 2. Spectum fr Apollo 16 soil 62231 resampled at Clementne TABLE I. Estmted value for sources of eror. UV-VIS wavelengths.

Efective Wavelength (nm B62l(A,30,0) 1% Repeatability of labrtory ) measurements (absolute)

Rh(). ) <0.5% Accuracy of halon spectal 414.9 0.1077 corection (largely NBS) 753.3 0.1776 f(A) NIA Measuement accuracy of 898.8 0.1893 flter tnsmission 951.5 0.1941 S(A) <<1% Black bdy estmton fr 1000.4 0.2004 sola flux D(A) NIA Detector responsivity estma- These values are used in the fnal U.S. Geological Surey Clementine mo­ lon saics.

DNClcm(}.) 1% Calibrated Clementine DN values [1-2]

DNClcmpl6(A) <1% Variaton withn site repre- seatng Apollo 16 Cphot (J..,i, e) 1% Photometc model [l] coefficient[4 ], but not to be confused with radiance factor, whichis reflectancerelative to a Lambertian surface illuminated at i = 0 [3- 4]). These calibrations require several assumptions: 0.45 (I) A mature soil represents the properties of the site seen re­ This appearsto be valid 0,4 motely under the same viewing conditions. 62231 average as long as regolith has been allowed to develop and there are no 0.35 nearbyfresh craters. • 0.3 (2) 62231 is representative of mature Apollol6 soils. Spectra of 0.25 several other Apollo 16 soils with comparable maturity alllook ... similarto 62231; deviations are on the samemagnitude as repeatabil­ 0.2 ity of individual measurements. 0.15 (3) Non-lAmbertian propertiesof Halon are insignificant at i = 30, e = 0. Any deviationwould be a small scalarcorrection. 0.1 (4) The detector response properties used in calibration are 0.05 invariant over the two-month period of measurements. There is no 0 500 11D0 1500 2000 2500 3000 guarantee that this is true,but no systematicmonth-to-month varia­ Wavelength {nm} tions aredetected that arenot photometricerrors. (5) Fig. I. Refectance specta fr Apollo 16 soil 62231. Several bi­ Contributionfrom Clementine scattered light is compensated directonal spcta taken at i = 30°, e = 0° are averged. The previov through in-flightmeasur ements and calibration procedures. There "difuse" (directonal hemispheric) spectum by J. Ad is shw fr is also no guaranteethat this is true,but no effectsof scattered light comparison. have been able to be demonstrated in spectralanalyses. LP/Contribution No. 980 49

References: [1] McEwen (1996) LPS XXVII, 841; McEwen et terrainpresent in the area. The PPI procedures point up the spectral al. (1998) LPS XXIX, Abstract #1466. [2] Pieters et al. (1996) variabilityof the region andthe DMDare clearly identified as "pure" www.planetary.brown.edu/clementine/calibration.html. [3] Hapke or more extreme spectra.Both the DMDs present the highest values B. (1993) Cambridge Univ. [4] Hapke B. (1981) JGR, 86, 3039. of the images (white and violet areasranging from 200 to 400 DN) while the adjacent zones present intermediate values (red areas ranging from100 to 200 DN). The mariaand highlands present in the REGIONAL DARKMANTLE DEPOSITS ON THE MOON: images have been classified as green areas (ranging from Oto 100 RIMABODE ANDSINUS AESTUUMANALSYSIS. S. Pinori DN). (To view Figs. 1 and 2 in color, go to http://cass.jsc.nasa.gov/ and G. Bellucci, Consiglio Nazionale Ricerche/lnstituto di Fisica meetings/moon99/pdf/8019. pdf.) dello Spazio Interplanetario, Area di Ricerca di Tor Vergata, Via Rima Bode deposit presents a narrow andlong shape and an del Fosso de! Cavaliere 100, 00133 Roma, Italy extensive red area,while the Sinus Aestuum dark mantledeposit is ([email protected]). heartshaped and shows a not vast red area. Rima Bodedeposit (Fig. 1 a) is more homogeneous than Sinus Aestuum, where some patches Introduction: Regional pyroclasticdeposits on the Moon usu­ (white areas)corresponding to the more extremespectra are present. ally arelocated on marebasins borders, blanketing adjacent high­ Using the analysis the PPI image, we have chosen fivedifferent lands [1-2]. end members for each region to apply the SAM procedure. The We present the study of two DarkMantle Deposits (DMD): Rima results areshown in Fig. 2, andin Fig. 3 the mean spectraof the DMD Bode and Sinus Aestuum. The statistical approach allows us to and the mixing areas classes are presented . identifynot only the dark mantle terrain,but alsosome areasspec­ trallydifferent from the DMDthat we have interpretedas a mixing of DMD and theneighboring soils. Instrument and Data Analysis: The data (Table 1) have been obtained at the 1.5-m telescope of Sierra Nevada Observatory (Granada, Spain) using animaging spectrometerworking in the 0.4- 1.0-µm spectral range, corresponding to 96 spectral bands. The spectra have been normalized to Mare Serenitatis (MS2) average spectrum,a standard area centered at 21° 25' E, 18° 40' N [3]. Two differentstatistical procedures have been employed to ana­ lyze the two regions and to discriminate between the different geo­ (b) 400 logic units, the Pixel Purity Index (PPI) and the Spectral Angle Mapper (SAM). The PPI is a method of finding the most "spectrally pure," or extreme,spectra in multispectralimages [4]. The PPI is computed by 200 repeatedly projecting n-dimensionalscatter plot onto a randomunit vector. The extreme pixels in each projection are recorded and the total number of times each pixel is markedas extremeis noted. In this way it is possible to create a PPI image in which the DN of each pixel correspondsto the number of times that pixel has been recorded as 0 extreme (see Fig. la and lb). The SAM is a method forcomparing the image spectra toa set of spectra chosen as a end members [5-6]. The algorithm deter­ Fig 1. Pixel Purity Index results for (a) Rima Bodeand and (b) Sinuus priori Aestum regions. mines the similarity between two spectra by calculating the "spectral angle"between them, treating them as vectors in a space with dimen­ sionality equal to the number of bands.For each referencespectrum chosen, the spectralangle is determined foreach pixel andthe final classification image shows a map where pixels of the same color represent regions of similar composition (see Fig. 2a and 2b). The (a) DMD borders, obtained from the 0.7-µm albedo image, areover­ lappedon the maps. Identificationof MixingAreas: Our analysisis devoted to the study of the relationshipbetween the DMDand the adjacent terrain (b) in order to better understandthe spectralproperties of the different

TABLE I. Characteristics of t datasets. Lunar region Date Time (U.T.) Coordinates Rima Bode March 19, 1997 20:58 3°W 13'N ° Fig. 2. Classification map obtained from the Spectral Angle Mapper Sinus Aestuum December7, 1998 01:26 7 WS'N technique: (a) Rima Bode deposit, (b) Sinus Aestuum deposit. 50 Workhop on New Views of the Moon I

On the basis of the spectraand of the position we identifythe red We expect that DMDs mixed with the highland soils show a class as highland, the blue class as mare, the yellow class as pure decrease in the UV-VIS slope (2). This effect is visible in both the DMD, and the cyan class as crater ejecta. The last two classes, the spectra corresponding tothe green areas. green and the sea-green, have been interpreted as mixing soils of The analysis pointsup the presence of another mixing area (the pyroclasticproducts andhighland, the green one, andDMD and mare sea-green color in Fig. 2a) in the Rima Bode region between DMD the sea-green. andthe mare.The mean spectrum of this class shows a negativeslope. Sinus Aestuum mean DMD spectrum shows a deep absorption Conclusions: The new statistical techniques allow us to iden­ bandwith the minimum at 0.7 µm in the visible range and symmetri­ tify somemixing areaswith mare or highland soils. These regions are cal wings at the extremes. Rima Bode DMD spectrumpresents less spectrally distinctfrom the DMDs and can help us to better under­ absorption in the visible and a lower NIR-VIS value. The difference standthe. possible evolutionof the eruptionsthat led to the deposits suggests aneffective distinct mineralogicalcomposition: The Sinus emplacement. Aestuum deposit could be dominatedby crystallized beads, while References: (1) Gaddis L. R. et a!. (1985) Icarus, 61, 461-489. Rima Bode could be a mixing of volcanic glasses andblack beads. [2] Weitz et al. (1998) JGR, 103, 22725-22759. [3] McCord T. B. et al. (1972)JGR, 77, 1349-1359. [4] BoardmannJ. W.et a!. (1995) Rima Bde yllow cloea (ONO) FifthJPLAirbome Earth Science Workshop, JPL Puhl. 95-1, 1, 23- ,.,o 26. [5] Yuhas F. A. and Goetz A. F. H. (1993) Proc. 9th Thematic Conf on Geologic Remote Sensing, 503-511. [6] Kruse et al. (1993) 145-163. I,�.00 Remote Sensing of Environment, 44, :I! 1 -� i OJ: usu.______0.4 D.6 O. ,.o -lng (} LUNAR ELEMENTAL ABUNDANCES FROM GAMMA­ RAYAND NEUTRON MEASUREMENTS. R. C. Reedy1 and a r__s_in_ u_Auc; u.m.yl"l.aw c"la"aac(DM�D) 2 1 ,.,o ; ._-- D. T. V animan , Mail Stop D436, Los AlamosNational Laboratory, Los Alamos NM 87545,USA ([email protected]),2 MailStop D462, Los Alamos National Laboratory, Los Alamos NM 87545, USA ([email protected]).

Introduction: The determination of elemental abundances is usu.______0.4 D,6 OJ 1.0 one of the highest science objectives of most lunar missions. Such -Ingf C} multi-elementabundances, ratios, or maps should include results for Sinus Auum green cloaa (mixing) elements that are diagnostic or important in lunarprocesses, includ­ ,.,a ing heat-producing elements (such as Kand Th), importantincom­ patible elements (Th andrare earth elements), H (forpolar deposits i ,� and regolith maturity), and key variable elements in major lunar :ll 1.00 -� provinces (such as Fe and Ti in the maria). Both neutron and y-ray I O.ll:I spectroscopy can be used to infer elemental abundances; the two usu.______. complement each other. 0.4 D,6 OJ 1,D ._..ngt. .. Cam} These elemental abundances need to be determined with high accuracy and precision frommeasurements such as those made by the Rimo Boe graen cloaa (mixing) ,.,a y-ray spectrometer(GRS) [l] and neutron spectrometers (NS) [2) on Lunar Prospector. As presented here, a series of steps, computer 1 i � codes, and nuclear databasesare needed to properly convert the raw :I! 1.00 y-ray and neutron measurements into good elemental abundances, .t ratios, and/or maps. j O.ll:I Lunar Neutron Spectroscopy: Lunar Prospector (LP) is the asa.______0.4 D,6 OJ 1.D firstplanetary mission that has measured neutrons escaping from a ._,,IOngtft v,m} planet other than the Earth[2]. The neutron spectrometers on Lunar Rimo Bo ae green claaa(mixing) Prospector measured a wide range of neutronenergies. Theability to ,.,a measure neutronswith thermal (E < 0.1 eV), epithermal (E - 0.1- 1000e V), andfast (E-0.1-10 Me V) energies maximizes the scien­ 1� tificreturn, being especiallysensitive to both H (using epithermal li ,.co ,t neutrons) andthermal-neutron-absorbing elements [3]. I OJ Neutrons are made in the lunar surface by the interaction of asa..______. galactic-cosmic-ray (GCR) particles with the atomic nuclei in the 0.4 D,6 ( 1.D surface. Most neutronsare produced withenergies above -0.1 MeV. -�(} The fluxof fastneutrons in andescaping fromthe Moon depends on Fig. 3. Mean spectra (normalized to I at 0.56 µm) of the classes the intensity of the cosmic rays (which vary withsolar activity) and obtained by using the SAM classification. the elemental compositionof the surface.Variations in the elemental LPI Contribution No. 980 51

composition of the lunarsurface can affectthe flux of fast neutrons elemental abundances. Such fluxdeterminations were done for analy­ by -25% [2,4J, with Ti andFe emittingmore fastneutrons than light sis of the Apollo y-ray data [SJ. The codes to do such calculations and elements like O and Si. the nuclear data used in such calculations have been improved much Most elements moderate neutronsto thermal energies at similar since then. rates. The main exception is when neutronsscatter fromH, in which Since about 1990, improved computer codes, such as LCS [4], case neutronscan be rapidly thermalized [3J. that numerically simulate the interactions of GCR particles with The cross sections forthe absorptionof thermal neutronscan vary matter have been developed at Los Alamos. Thesecodes have been widely amongelements, with major elements like Ti andFe having well tested using measurements fromlunar samples and meteorites high-capture cross sections. Some trace elements,such as Sm and and at accelerators. The output from LCS includes the fluxes of Gd, have such largeneutron-absorption cross sections that, despite protons and neutrons as a function of depth in the Moon and the their low abundances, can absorb significantamounts of thermal fluxesescaping fromthe Moon. These pivti,clesproduce the nonelas­ neutronsin the Moon [5). tic-scattering y-rays. LCS canalso calculate the rates forthe capture Because the processes affectingneutrons are complicated, good of thermal neutrons by various elements. modeling is needed to properly extractelemental informationfrom These LCS-calculated particle fluxesor capture rates need to be measuredneutron fluxes. The LAHETCode System (LCS) can be converted into production rates forthe y-rays of interest as a function use to calculate neutronfluxes from GCR interactions in the Moon of depth in the Moon's surface using several sets of nuclear data, such [4]. as cross sections fornonelastic-scatteringreactions ory-ray yields for Lunar Gamma-Ray Spectroscopy: The main sources of plan­ neutron-capturereactions. The fractionsof these y-rays that escape etary y-rays are the decay of the naturally occurring radioactive into space without undergoing aninteraction will then be calculated isotopesof K, Th, and U and the interactions of GCRs with atomic to give the expected fluxesof y-rays at the LP GRS. nuclei in the planet's surface. Most "cosmogenic" y-rays are pro­ Preparing and Interpreting Elemental Maps: The elemental duced by fastand thermalneutrons made in the planet'ssurface by abundances and ratios obtained fromy-ray and neutronspectra for GCRs, and their production rates can vary with time. Over 300 various regions and features of the Moon will be compiled. The y-ray lines have been identifiedthat canbe emitted fromplanetary precision andaccuracy of these results need to be examined closely. surfaces by a variety of production mechanisms [5]. There exist Large regions of the Moon with known or well-inferred composi­ nuclear databases that can be used to identify and quantify other tions need to be identifiedthat could be used for"ground truth" in y-ray lines. Use will be made ofy-rays frommajor elements, particu­ testing these elemental results. Ground-truth regions and counting larly those fromSi and 0, that have not been routinelyused in the rates for various energy bands in the LP GRS data for those lunar past. regions canbe used to create additionalelemental maps in a manner The fluxesofy-rays froma given element can varydepending on similar to that done for Th, Fe, andTi from the Apollo GRS. many factorsbesides the concentrationof thatelement. For example, Scientifc Studies: A few of the studies of the Moon that we the fluxesof neutron-capturey-raysin the planetaryregion of interest want to do are noted below. depend on (1) the total cross section forelements to absorb thermal­ Mapping of highland lithologies. Some of the key parametersin ized neutrons[SJ and (2) the H content of the top meter of the surface lunarhighland analysis, such as Ca/Al andAl/Si ratios, areunlikely [e.g., 3). The fluxesof the fastneutrons that induce inelastic-scatter­ to be determined accurately enough fromthe LP data to be useful. ing and other nonelastic-scattering reactions can vary with the com­ However, other questions in highlandstudies arealmost certain to be position of the surface [4]. addressable with elements that will be well determined fromthe LP Data Analysis: There are several key steps in preparing y-ray GRS and NS data. data into a formfrom which accurate elemental abundancescan be The question of lunar granite occurrences is still open.Granitic determined.One needs to identify,quantify, andremove or correct impact products suggest granitic areaslarge enough to contain sig­ forall backgrounds in they-ray spectra. Among the more important nificant impacts without large-scale admixture of nongranitic re­ of these backgrounds are featuresmade by the decay of radioactivi­ goliths or lithologies; we cantest whether these zones of impact are ties made in the GRS by cosmic-ray particles andthe prompt and large enough to be seen by the LP instruments. Regions with high decayy-rays emitted from the material surrounding the active ele­ contents of Si or high Kffhratios might represent granite or other ments of the LP GRS and fromthe LP spacecraft.Gamma-ray spectra rare lunarcomponents. obtained during the cruiseto the Moon or those measuredwhile LP Although KREEP may be principally derived froma localized was at various distancesfrom the Moon can be used to distinguish region beneath the lunarnearside [e.g., l], there is no certainty that featuresin they-ray spectra that arefrom the Moonand thosethat are other lunar incompatible-element enrichments have not been iso­ made on the LP spacecraft. lated at other times and places in the highlands.Maps of K, Th, and Each background-correctedspectrum will be analyzed with exist­ REE across the entire Moon canextend KREEP studies into a more ing y-ray spectral-unfoldingcodes to identifythe energies andinten­ complete understanding of late-stage lunar crustal evolution. sities of all peaks. These peaks will be examined when there are It may be possible to map volatile-elementdepletions fromLunar potential interferences in the analysisof a given y-ray line [5]. Such Prospector data. Silica volatilization is knownfrom a variety of Si­ interferences could be a problem for determining Mg and Al using depleted compositions. The Si/O ratiocould be used to test forthe some oftheirinelastic-scatteringy-rays, such asthe 1.369-MeVy-ray presence of Si-depleted impact melts extensive enough to approach from Mg that is also readily made from Al andSi. the scale of thefootprints forthe LP GRS. The key data needed to get elemental abundancesfrom the fluxes Mappingof mare basalt lithologies. In mappingmare basalts, ofy-rays in the processed spectraare good valuesfor the fluxesofy­ the elements most readily measuredwith the LunarProspector y-ray rays that should be emitted froma given region for knownor nominal instrument(Ti, Fe, Th, K) are particularly suited to recognize and 52 Workhop on New Views of the Moon I

map the principal known varieties of lunar baalts, including "high­ This frmula is IFOV ma/(K-B/H) with IOV max defned a Mai­ Ti" ad "high-K" basalt types. Distnguishing low-Ti fom very-low­ mum Instantaeous Field of View; B/H is the base-to-height ratio Ti (0.4% K20) may be readily identifable, but it will agan be impor­ Altimeter. The Clementine laser altimeter (LIDAR) data were tant to determine to how low a level they-ray K data are reliable fr used previously to produce a global topgraphic model of the Moon mapping. [3]. The model has a vertical accuracy of -100 m ad a spatial Mapping of soil maturit and hydrgen indexes. The Luna resolution of 2.5°. Altimetry data were collected between 79°S ad Prospector neuton data ca provide a new view of regolith maturity 81°N. These data were fltered and then interpolated to fll in the using H content, which should vary directly with regolith maturity. pola regions where the altimeter did not collect data. A global Effects such as enhaced H retenton in ilmenite-rich regolith can be topography model was then derived based on spherical haonic addressed by the examinftiOn of Ti abundaces and Ti/Fe ratos expansion [3]. along with the epitheral (and thermal and fast) neuton data The Image mosaic. A global image mosac of the Moon wa pro­ value of a globa map of lunar soil maturity is immense. Our present duced fom the 750-nm Clementine data [4,5]. The mosaic includes understading of regolith matuty is lagely biased towad mare high-resoluton, oblique, and vertical images. Match points were regions and away fom heavily cratered teraes. This implies an picked to tie the imagery together, ad the caera pinting agles incomplete understading of the most evolved luna regoliths. Ma­ were adjusted to align the imagery. This adjustment used a sphercal ture regolith ha value in resource applicatons ad in tageting future surace, ad the elevation of all pints was held to a constat value, exploration fr understading impact ad sola-emission history in 1737.4 km. Tis produced a seamless image mosaic with lattude ad the vicinity of Earth. longitude inforation but no inforaton on the elevation [4-5]. Acknowlegments: This abstract describes work that we plan Analytical Aerotriagulation: The imagery ad supprt infor­ to do with G. Heiken ad D. Drake as part of NASA's Luna Data maton were downloaded to ou digital photogrametic worksta­ Analysis Program. tion fom the Integated Softwae fr Imagers ad Spctometers References: [1] Lawrence D. J. et al. (1998) Science, 281, (ISIS) system. The support data included the caera location ad 1484-1489. [2] Feldma W. C. et a. (1998) Science, 281, 1489- pinting agles. Match points used to produce the image mosaic were 1493. [3] FeldmaW.C.etal.(1991)GRL 18, 2157-216. [4] Reedy also downloaded. The caera agles were adjusted to account fr the R. C. et a. (1998) MAPS, 33, A127-A128. [5] Reedy R. C. (1978) elevation of the match pints. This wa accomplished with the Mult Proc. LPS 9th, pp. 2961-2984. Sensor Tragulation (MST) softwae fom LH Systems SOCET Set softwae package [6]. The revised camera agles allowed fr the derivation of a digital elevation model (DEM) fom the stereo pairs. LUNAR SOUTH POLE TOPOGRAPHY DERED FROM Initial estimate. The match-point latitude ad longitude fom the CLEMNTINE IMAGERY. M. R. Rosiek, R. Kirk, and global image mosaic are accurate ad used for a initial estimate of A. Howington-Kraus, U.S. Geological Survey, Astrogeology Team, the horizontal position. The elevations of the match points were Flagstaff AZ 86001, USA ([email protected]). estimated fom the altimetry data. The camera angles used in the altimetry processing ad in the creaton of the image mosaic were Introduction: During the Clementine Mission both oblique adjusted independently. Hence, the horizontal position of the altim­ and vertical multispectal images were collected. The oblique ad etry data and the image mosaic are not aligned corectly. Clementine vertical images fom a single spectral bad collected during the same was designed so the altimeter shaed the optical system of the HIRES orbit form a stereo pair that ca be used to derive the topography. camera system. The HIRES ad UV-VIS camera systems were aigned These stereo pairs ae being used to derive the topography of an aea so the HIRES image was centered in the UV-VIS image [IJ. We (90°S to 65°S latitde) surounding the luna south pole. Work on the therefre made an adjustment so that the altimetry points would fall luna north pole topography will start after completion of the south nea the centerline of the UV-VIS imagery. pole topography. This report provides a update on the initial results A DEM was created fom the altimetry data using the adjusted fr the luna south pole topgrphy. psition and the match-point elevations were estimated fom the Clementine Data: In 1994, the Clementine spacecraft adjusted DEM. These points were used in the HA TS softwae, which acquied digital images of the Moon at visible and nea-infaed allows fr the use of weights on the estimated position based on the wavelengths [l ]. Onboad there were fur camera systems ad a laer accuracy of the point. The horizontal positions were given a weight altimeter. During the frst pass, periapsis was at 30°S ad the highest of 1 km, ad the vertical estimates were given a weight of 5 km. resolution images were obtained in the souther hemisphere [2]. Stereo adjustment. 1n frming the Clementine Mosaic, over Over the norther pla aea, a series of oblique ad vertical images 360 images ad 29,00 match pints were used in the souther pla were obtained with the ultaviolet-visible (UV-VIS) camera on each region, a area defned a 6°S or less. Diferent techniques were orbit. During the second pas, periapsis was at 30°N and the image tied in adjusting the images and match pints. Limitations of com­ acquisition staegy was reversed. puter memory required breaking the data into blocks that would be Imager. The UV -VIS camera image size wa 384 x 288 pixels solved sepaately ad then brought back together. Initial results had with fve spctal bands and one broad band. The 750-nm-bad lage elevaton erors in the DEMs derived fom diferent stereo stereo pairs ae the primary image source fr this study. The ground­ pars. This was most likely caused by the fct that although the stereo saple distaces (GSD) fr oblique images rage fom 30 to 400 m. models overlappd, they did not shae simila match pints. One Te GSD fr the vercal images, acquired at the end of a orbit, ae solution fr this problem would require tansfering all 29,00 match slightly lager ad rage fom 325 to 450 m. Using the frula fr points to all the images they fall on within the set of36 images. Tis stereo-height accuracy [2], an estimate of height accuracy is 180 m. would be a temendous amount of work. The solution used was to LP/ Contribution No. 980 53

thin the images to just the stereo pairs (983 images) and to thin the Witin models, the RMS bd bias erors were simila. This is prob­ match pints to cover the aea (973 points.) The MST softwae was ably caused by a ult between the models. The standad deviation of then used to add match points with the criteria that each image should 212 mis close to the expected precision of 180 m. The DEMs, from have nine match pints distibuted throughout the image. This pro­ each model within b orbit, were merged together with b averaging cess added 1181 pints. During editing bd checking the match bd smoothing procedure. The elevation erors in overlapping orbits points, 45 more match points were added. The number of match were looked at next. The bias eror drops to 222 m; since the orbit pints within b image ranges fom 3 to 30 points, with 96% of the DEMs are based on averaged data the tilt in the models wb probaly images having nine or more pints and the average image having 16 averaged out. The stbdard deviations of the elevauon erors increase points. to 336 m. For the adjustment procedure b iterative least-squaes solution The orbit DEMs were merged together bd exported to ArcView is used; this allows the caera angles bd match-pint ground loca­ Geographic Inforauon System (GIS) for analysis bd viewing. tions to chbge during the adjustment. The fnal root meb squae Figure 1 provides a false-color image of the lunar south pole topg­ (RMS)eror of the match points was 0.68 pixels. Generally, a value raphy. (To view fg. 1 in color, go to http://cbs.jsc.nasa.gov/meeings/ of below 1 pixel is acceptable. moon99/pf/806.pdf.) The aea shown covers to 80°S. This will be Elevation Extraction: The SOCET SET softwae provides an expanded to 65°S. Work will begin on the luna north pole topogra­ automated routine to extact elevation data. For every stereo model, phy in September 1999. a corelation pint was determined every 1 km in ground distbce. Reference: [l] Nozette S. et al. (1994) Science, 266, 1835- These patches of teDEM were frst merged along b orbit path, bd 1839. [2]Cook A. C. et a!. (1996)Planet. Space Sci., 44,1135-1148. then theDEM fom adjacent orbits were merged together. The stereo [3] SmithD. E. et a!. (1997) JGR, 102, 1591-1611. [4] Eliason E. M. pairs do not completely cover the enure south ple, and these small (1997) LPS XXVlll, 331-332. [5] Isbell C. E. et al. (1999) LPS gaps could be flled with other images or other techniques such b XXVlll,1812-1813. [6]MillerS. B. bd Walker A. S. (1993)ACSM photoclinomet. ASPRS Annual Convention and Exposition, Technical Papers, Vol. InitialReults: Data were collected stating with the models 3, 256263. closest to the south ple. DEMs were collected fom 108 stereo models, bd the imagery was fom 23 digerent orbits. Erors were sumaized by number of points, RMS, standard deviation, bib, bd INTENTIONAND INTENSIONI THEINTEGRATION OF percentage of pints that were blunders (Table 1). The elevauon LUNAR DATA SETS: THE GREAT INSTAURATION. erors in overlapping models within b orbit were looked at frst. G. Ryder, Luna bd Plbetary Institute, Center fr Advbced Space Studies, 3600 Bay Area Boulevad, Houston TX 77058-1113, USA TABLE 1. Average of errors(Models: n=108, Orbits: n=23). ([email protected]). Points RMS St.Dev. Bias % Blunders Thus have I made as it were a small globe of the intellectual Models 2125 692 m 212 m 636 m 2.7% world, as truly and faithfullyas I could discover. Orbits 8248 447 m 336 m 222 m 1.5% - Francis Bacon, 1561-1626

Languagewas invented so that people could conceal their thoughts from each other. - Chales-Maurice de Talleyrand, 1754-1838

Introduction: In frmulating a luna initiative in the integra­ tion of diverse datasets, we have tended to concentrate on the advan­ tages of the end product, while glossing over the difculties inherent in the process. As a rough guide, we can say that our objective is to obtain as full an understanding of the Moon as we can get by using all sources of data about it in combination in addressing key prob­ lems. Integration, collaboration, and datbets have become increas­ ingly imprtbt concepts over the digital age of te lbt two decades. Integauon to the pint of unity of knowledge was b Enlightenment concept, and hb its modes counterpat in the metaphysical view of the integration of science, social science, bd humaniues [l]. The Geological Society of America, The Ecological Society of America, bd the U.S. Geological Surey have recognized the importance of understbding integration, beginning to discuss the factors that bring about successf oqration of a fncuonal collective across agencies and institutions [2]. But our goal is of a lesser degree thb tat: integration within a fw disciplines of science aound the explbcion -10 •9 ..g .. ; -6 •S -i •l -.2 .J O l 2 J456 Eltvoriouiokm D•11un 1737.4km of a common physical object. The technological advbces that have Fig. 1. Lunar south pole topography. made databases common and accessible allow (bd produce) a mb- 54 Workhop on New Views of the Moon I

sive sea of data whose achiving bd exchange oger unprecedented (intensional defnition). To many remote sensers of the Moon, it opprtunites. At the same tme, there ae problems of documentation merely mebs something (usually an average soil over a fw square (metadata) that infuence identification, suitability, processing, bd kilometers aea) with low Fe, whose chemical composition by infr­ balysis [3). ence approximates that of b bortositic gabbro. Anorthositic gab­ Unfrtunately, the words integration and dataset have also be­ bro in my mind conjures up images of a magma diferentiating in a come buzzwords that ae fequently empty. The meaning of words is lage subterbeb chamber over a period of thousands of yeas, important, bd although a rose by any other name would smell just slowly growing crystals bd eliminating or trapping liquid in be­ a sweet [4 ], attempting to invoke the smell of a rose in someone else tween crstals, bd atoms of europium preferentialy being incorpo­ by calling it a rat would fail; calling something a rose when it smells rated in the growing plagiocase, bd so on. (It is also a chunk of rock like a rat would also fail. Defning what a dataet is (and whether it in a drawer that has a particula look about it.) I do not kow what should be one word or two) is neither a tivial nor a meaningless task. image is conjured up in the mnd of a remote senser whose main Integration and Intention: Integration even within our initia­ interest is not petology. Tis makes it diffcult to produce a simple, tive has diferent levels of meaning. The prosthetic extensions of our adequate, all-usefl glossary of the terminology used in lunar sci­ senses have produced varied kinds and bounts of data about ence. How do I relate the anorthositic gabbro used by remote sensers the Moon. The integraton of these data into concepts about the to the borthositic gabbro used by a geophysicist (which in the luna Moon is not merely a multidisciplinay approach, but needs to be case is probably an borthositic norte byway) bd to the anorthositic intellectually imersive bd synergistic to be successfl. Prely gabbro used by me? multidisciplinay approaches cb probaly provide incremental ad­ Tis problem of intension is imprtant in integration in a cogni­ vances on spcifc problems, but the goal of any signifcant project tive way. For instbce, we ca (bd do, cuently) defne a set of (other thb engineering or technological projects with clearly de­ basaltic rocks a high-Ti mae basalts. These are defned (stipulatve fined tasks) is discovery. defnition) as volcanic, have high Fe (all have more t 15% FeO), Most signifcbt questions ae epigenetic: We now ak questions bd under the Neal bd Taylor [6] defnition have between 9 bd 14% about the role of plagioclae elutiation fom a magma oceb in Ti02 (see accompbying astact fr more critcal discussion). This infuencing subsequent luna evolution, but no such question existed intensive stipulatve defnition leads to a certain set of baalts (the prior to theApolloprograofinvestgationthat eventually produced extension) that have these properties. (It is actually the existence of such a question. Thus a major intent of the integration initative the extensions that has invoked the intension.) But what ae the should be to derive new questions ( or problems) bd not just address ramifcations of this? What else might these basalts have in common? existing ones. Nonetheless, these existing questions ae themselves Are there other chemical fatures that this extensional set also ha (a epigenetic, and remain questions of some importance. We need to see the intensional definition of equilateral triangles also leads to the how using data fom one discipline cb be used in perhaps unex­ sbe extensional set as the intensional definition of equibgular pected ways to addess questions typically thought to belong to some triangles), such as Sc? Or do these basalts all come fom a other discipline, bd not just think taditionally about "common" clinopyroxene-bearing source? Some of these possible chaacters problems. One beauty of science is the recognition of the unity of might be asumed ( even if not demonstrated) by some ptrologists, complex phenomena that to direct obseration appea to be quite yet not even alluded to by a remote senser. sepaate things. An example might be magnetic variations in oceb Datasets: Huge amounts of infration have been generated basins, topography of mid-ocean ridges, and chemisty of Javas in about the Moon, with varied means.Vaious manifestations of groups orogenic chains, all united by plate tectonics. Geologists in fct do of such infrmation are commonly refred to as dataets. Some of this sort of integration continually, both in the feld bd out of it, and these ae sets, others are not. It is important to understand the true at varied scales. Continued unifcation of our lunar understanding is coherency of anything refred to as a dataset. Even coherent, digital one of our intentions. datasets go through vaous tbsfrations, e.g., fom b original Integration and lntensions: One problem of integration across collection of some sort of instrument response with some sort of time disciplines and datasets is communication. Words or terms have both tag (transfrmed into location?) though transductions and manipu­ extensions and intensions (in formal logic; note distinction of inten­ lations into something more directly infrative to a user. For sion bd intention). Using the defnition of Dennett [5], the extension exbple, y-ray data needs to be (fr me to use) in some frm of is the thing (or set of things) to which the word or term refers; the element intensity/location set, preferably deconvolved with geologi­ intension is the particular way the extension is picked out or deter­ cal input. Tis requires sequential datasets, of digerent uses to mined. (As a simple example, Dennett uses the ters equilateral diferent workers, with vaied metadata dependent on the genetic triangles and equiangular triangles, which pick out the same set of database and intended uses. Some of these dataets might be com­ things, bd thus have the sae extension, but clealy do not meb the mensurable, bd thus at one technical level relatively easy to inte­ same thing. Tey have diferent intensions.) Roughly, the intension grate. Some datasets ae much less easy to defne. For exbple, some of a word consists of the properes a thing must have in order to be people have requested producton of, and access to, the luna sample in the extension of that word. There is no doubt that both the chemical dataset. However, luna sample bayses do not frm a intensions bd extensions of paicula terms ae used diferently by coherent set; diferent workers use diferent techiques, with difer­ membrs of diferent disciplines, leading to confsion (identifed by ent element sets, diferent uncertaintes (and diferent outght er­ Bacon fur centuries ago as the direst of erors). This is compunded rors), and vaed sample sizes that quite diferently represent-or fail by the fbt that many of our importbt terms ae fr abstact bd to represent -whole rocks. Mby balyses were collected without vaiedly abitar concepts. Tus "borthositic gabbro" fr me de­ any intention of the baysis being representatve; others have inad­ notes a rock containing easily visible grains, mainly plagioclae bd equate metadata (such as sample size). Thus to be usefl, extensive clinopyroxene with te frmer domnant, bd of igneous orgin metadata ae needed fr such "sets." Some specifc subgoups of data L! Contribution• No. 980 55 might for fairly coherent sets, e.g., neuton activation dalyses of have been given site-specific labels, e.g., Luna 24 VLT, Apollo 11 hi­ luna soils fom one laboratory. These ae signifcant issues to ad­ K, A12 olivine basalts, and Apollo 15 Green Glas C. Tis is classi­ dress in attempting cognitive integration of vared infration in fcation of the frst kind, but is not a usefl classifcation of dy other understdding the Moon. kind. At a minimum, a classifcation is inclusive (all objects have a There is one aspect of data to be considered that results fom place) and exclusive (all objects have only one place). The answer to humd cognition dd evoluton: The natural world contains an "How should rocks be classified?" is far fom tivial, fr it demdds embaasment of riches, dd much incoming infrmation has to be a fndamental choice about nature dd ordering. Classifcation func­ fltered out. The choice of fltering depnds on what is being ad­ tions as a primary tool of prception, opening up ways of seeing dressed by the inspection of that inforation. It is a greater cognitive things and sealing of others. Lacking a classifcation, mae-baalt challenge than most pople consider. petology appas imature with little consensual prception of the Conclusion: A look behind the apparent simplicity of integra­ qubities dd signifcances of the basalts. The appaance may or may tion as an exercise demonstrates that there ae cross-discipline issues not be the reality, but it demonstates a need fr a functoning, that need to be carefully thought abut. These rdge fom the new dd communicatory classifcaton, in paticula fr the dissemination of old themes to be addressed, the terminology used, the nature of ideas and the frtherdce of studies. dbaets, dd the cognitve apects of integration. Combined with the Inconsistency of Nomenclature: Nces ae inconsistent both varied scales that data ae collected on, these promise to mae cong luna roks dd between luna dd terestial rocks. Scples integation a chblenging dd rewarding exercise in retng the ae labeled by elements, chemisty with tags, chemisty cat into Moon to a prominent psition in pldetary science: The Greb mineralogy, or a mineralogical attibute (respctive examples A14 lnstauration. VHK, A17 high-Ti Group Bl, A15 quartz-norative, A-12 pigeo­ References: [1] Wilson E. 0. (1998) Consilience, Alfed Knopf, nite ). Such inconsistency is bound to lead to confsion. Chemical New York. [2] May C. et al. (1999) GSA Today, April 1999, 4. descriptons med diferent things in mildly diferent contexts: A [3] VogelR.L.(1998)EosTrans.AGU,XX,373-380. [4]Shakespae low-K Fra Mauro basalt (not a basalt!) contains slightly more K thd W. (1595?) Romeo and Juliet. [5] Dennett D. C. (1996) Kinds of d Apollo 11 high-K basalt. High-alumina means more thd about Minds: Towardsan Understanding of Consciousness, Har Collins. 11% Al203 fr mare basalts, but 21 % fr highdds ''babts." Vol­ [6] Neal C. and Talor L. A. (1992) GCA, 56, 2177. canic KREEP basalts, -18% Al203, ae not (usually) qualifed with "high-alumina." Yet fr terestib basalts, high-alumina meds more

thd -17% Al203. Further, even very-low-Ti mae basalts have Ti abundances (-0.5-1.5% Ti 02) as great as typical terestial basalts. NAMING LUNAR MARE BASALTS: QUO VADIMUS Thus, paallels between lunar dd terestial nomenclatures are non­ REDUX. G. Ryder, Luna dd Pldetay Institute, Center fr existent (reinfrced by the fact that a mae-baalt composition found Advanced Space Studies, 360 Bay Area Boulevard, Houston TX on Earth would be too ultamafic to nce baalt at all). A separate 77058-1113, USA ([email protected]). type of nce exists fr mae-basalt glases, which are identifed by site, color, and a letter fr any subsequent distinctions, e.g., A15 Introduction: Nealy a decade ago, I noted that the nomencla­ Green Glass C. ture ofluna mae basalts wa inconsistent, complicated, and arcde Arcane Character of Nomenclature: While the inconsisten­ [1]. I suggested that this refected both the limitations of our under­ cies cited above by themselves make nomenclature acde, a greater stdding of the baalts, dd the piecemeal progression made in luna source of difculty is the common use of acronyms such as VHK and science by the nature of the Apollo missions. VLT. Most of these ae partly chemical acronyms, but degrading the Although the word "classifcation" is commonly attached to symbol Ti to T (fr instance) makes them unintelligible and devoid various schemes of mae basalt nomenclature, there is still no clas­ of information even to the intelligent, educated non-expert. sifcation of mae basalts that has any fundcental grounding. We Toward a Classification of Lunar Mare Basalts: Classif­ remain basically at a classification of the firstkind in the terms of cations have functions. A major one must be communication; i.e., a Shand [2]; that is, things have names. Quoting Joh Stuart Mill, name fr a mae basalt provides a common understanding of what the Shand discussed classification of the second kind: "The ends of baalt is. For the small number of suites curently available, the scientifc classfication ae best dswered when the objects ae fred present labels (though ineffcient and insufcient) may work; with into groups respecting which a greater number of propositions can be continued recognition of more basalts, Antarctic meteorite samples, made, dd those propositions more imprtant than could be made orbiter data, sample retus, dd luna base studies, labels will respectng dy other groups into which the sce things could be become increasingly inefcient. Clementne dd Prospctor data distibuted." have made mapping of mare baalts a much more visible btvity thd Here I repat some of the main contents of my discussion fom a it wa, dd increasingly comon ground among scple petologists decade ago, dd add a further discussion based on events of the lat and remote sensers has emerged. decade. To estalish a usable classifcaton, there must be some criteria fr Labels and Classification: A necessa frst step of scple relatonships. Petologists need to decide what the most signifcant studies that ams to understdd luna mae babt processes is to chabters are, dd how these cd be tslated into a clasifcaton. asociate samples with one another as membrs of the sce igneous The common distincton on the bais of Ti (the major element with event, such as a single erpton, lava fow, or diferentaton event. the greatest vaiaton) may or may not be appropriate. It remains to This has been farly successfl, and discrete suites have been iden­ be established whether the use of Ti is of fndamental value both in tfed at all mae sites, members that ae erptively related to ebh relatng baalts to each other dd in communication, or merely d other but not to members of other suites. These erptive members historical bcident or respnse to its vardce. 56 Workshop o; New Views of the Moon II

A great deal of discussion among interested parties will be re­ tral classifcation boundaries on the basis of chemistry ae best quired to arrive at a clear, fnctional, and consistent classifcation of looked upon as equations that ae not of the fr y (or x) = constat. mae basalts. A classifcation would need to be such as could be used Some individual extusive events can produce basalts that cross such by a range of workers including remote-sensing specialists, and thus boundaries. Thus, while the boundaries can be usefl fr comuni­ would need to be hierachical, according to what data are available. caton, there is no indicaton that they are usefl fr petogenetic Acronyms should be eliminated, but some frm of coded classifca­ purpses. tion of use in computer databaes could b a useful supplement to a The classifcation of mare basalts on the basis of Ti has an classifcation. Tere ae severa key questions to address: ( 1) Do we unfortunate consequence: This division is used a inevitable, and know enough about mare basalts to yet frulate a clasifcation, or subsequent investigations are aligned with that diference in mind. is the feld indeed too immature? (2) Should basalts should be Thus baalts ae commonly plotted on diagrams sepaated on the clasifed at the suite scale (presupposing quite a lot of infrmation basis of Ti, ad comparisions ae made within Ti divisions rather than about several saples) or at the hand-sample scale, ad how would across the whole spectum; it may well be that fundamental regulai­ such classifcations satisf remotely sensed infration? (3) How ties ae being missed by this frst cuttng. should chemisty ad texture be balaced in any classifcaton; should Another unfrtunate consequence is that the Ti classifcaton texture merely be a qualifer? (4) Are there natural divisions among gives too much emphais to high-Ti basalts. On the bais of volume mae baalts? Ca abitay divisions still facilitate communication? extuded, as fa as we can tell, high-Ti baalts ae only a small faction Any classifcation must avoid a detailed genetic bae. Obviously of the total. An even greater disproportion is probably given to the the genesis mght be debatable or the consensus change, but more pyroclatic glasses in most minds by their distinction a a group. importantly, there is not "an" origin fr a given mare basat. It ha an Importat though they ae in a petological sense, their absolute origin going back to luna fration, ad combining source produc­ aundance is tivial compaed with the total mas of extant mae lava tion, crystallization, source mixing, partia melting, assimilation, ad References: [I] Ryder G. (1991) LPS XX.II, 1157. [2] Shand so on. It would be difcult to incorrate 5% assimilation into a S. J. (1927) Eruptive Rocks, Hafer Publishing. [3] Neal C. ad genetc classifcaton. Taylor L. (1992) GCA, 56, 2177. [4] Taylor G. J. et al. (1991) The The Current State of Mare Basalt Classifcation, 1999 Lunar SourceBook, Cabridge Univ. [5] Papike J. Jet al. (1998) Edition: All reviews of mae basalts state categorically that the Rev. Mineral., 36. classifcation of mae basalts is based on Ti content [3-5]. Tere is raely ay discussion at all on why that is so, or why paticula divisions ae prefred. Neal ad Taylor [3], fr instance, unilaterally eliminated a feld of interediate-Ti mare basalts. There is one ORIGI AN EVOLUTION OF TH MOON: APOLLO2000 considerable practical advantage to classifying mare basalts on the MODEL. H. H. Schmitt, Deparment of Nuclea Engineering ad basis of Ti contents: Titanium can be detected and meaured by a Engineering Physics, University of WisconsinMadison, P.O. Box wide variety of techniques on samples and fom orbit, ad with the 90730, Albuquerque NM 87199, USA. global-Ti acquisitions fom Clementine ad Prospector a whole feld of mapping opened up. Furthermore, there is no doubt that Ti both Introduction: A descriptive frmulation of the stages of luna vaies widely aong basalt suites ad is abundat enough in luna evolution [I] as a augmentaton of the traditional time-statigraphic mare basalts (uniquely) to be a major element. However, is it the most approach [2] enables broadened multidisciplinary discussions of fundaentally useful in a petrogenetic sense? issues related to the Moon ad planets. An update of this descriptive Titaium would be fndamentally useful if there were defnite frmulation [3 ], integrating Apollo and subsequently acquired data, other chaacters of petogenetic signifcance that varied in parallel provides additional perspectives on many of the outstanding issues with it. It is widely perceived among knowledgeable petrologists at in lunar science. least that there ae such parallels; for instance, many high-Ti mare Stage I: Beginning (Pre-Nectarian) - 4.57 Ga basalts ae also high in other incompatible elements and in Sc. Stage 2: Magma Ocean (Pre-Nectarian) - 4.57-4.2(?) Ga Furthermore, there is a common perception that high-Ti mare basalts Stage 3: Cratered Highlands (Pre-Nectarian) - 4.4(?)- ae derived fom sources containing ilmenite ad clinopyroxene in 4.2(?) Ga contast with lower-Ti basalts containing a low-Ca pyroxene (both in Stage 4: Large Basins - (Pre-Nectaria - Upper Imbrium) - addition to olivine). However, the consistencies here ae not great, 4.3(?)-3.8 Ga espcially when volcanic glasses ae taken into account. There is no Stage 4A: Old Large Basins ad Crustal Strengthening (Pre­ very consistent picture fom isotopic data. There was a common Nectarian) - 4.3(?)-3.92 Ga concept ealy in luna sample studies that high-Ti basats were Stage 4B: Young Large Basins (Nectaa - Lower Imbrium)- derived fom shallower sources tha low-Ti basalts, but further work 3.92-3.80 Ga ha tended to deny that generaizaton too. Stage 5: Baatic Maia (Uppr Imbrium) - 4.3(?)-1.0(?) Ga Neal ad Taylor [3] proposed a new "classifcation" of mare Stage 6: Mature Surface (Coperican and Eratosthenian) basalts, frst baed on Ti diferences, then on alumina digerences, - 3.80 Ga to Present. and then on K diferences. The boundaries of ths classifcation ae Lunar Orign: Increaingly stong indications of a lagely arbita ad have no fndamental petological signifcace (that I undiferentated lower luna mantle ad increaingly constained am awae of). Te division of high-Ti baalts fom low-Ti baalts is initial conditions fr models of an Earth-impact origin fr the Moon set at 9% TiO2, fr instance, and in fct all the bundaries ae [4-7] suggest that luna origin by captue [8] of a indepndently orhogonal in this three-dimensiona chemical space. It is unlikely evolved planet should be investgated more vigorously. Ce that natural divisions in nature would be so orthogonal; most teres- appas to better explain the geohemical ad geophysical details LP/ Contribution No. 980 51

related to the lower mantle of the Moon and to the distribution of Source of Large Basin-fonning Objects: After the cratered elements andtheir isotopes. For example, the source of the volatile highlandsstage and beforethe basalticmaria stage, objects from a components of the Apollo 17 orange glass apparently would have discrete source region formed about 50 large basins on the Moon lain below the degassed and differentiated magma ocean (3) in a over -400m. y. Four possibilities forsources of the impactors of the relatively undifferentiatedprimordial lower mantle.Also, a density large basin stage appearplausible at this time [8,21-23). Of these reversal from3. 7 gm/cm3 to approximately 3.3 gm/cm3 is required at possibilities, the initialbreakup of the original Main Beltplanetesi­ the baseof the uppermantle to be consistent with the overall density mal would appearto be the best present choice as adiscrete impactor of the Moon. Finally, Hf/W systematics allow only a very narrow source. window, if anyat all fora giant impactto form the Moon [3,9,10]. UrKREEP Mobilization: The striking differences between Origin of the Oldest Magnesium-suite Rocks: Continued young, mascon basins (-3.92-3.80 Ga) andold, nonmascon basins accretionary impact activity duringthe crystallizationof the magma (-4.2-3.92 Ga) indicate that the older, isostaticly compensated ba­ oceanwould result in the "splash intrusion" of residual liquids into sins triggeredthe regionalintrusion, extrusion, and solidification of the lower crustof the Moon assoon as the crustwas coherent enough mobile urKREEP-related magmas prior to the formation of the to resist re-incorporation into the magma ocean.For Mg-suite rocks younger, uncompensatedbasins [24]. Thissuggests that the fractur­ with crystallization ages greater than about 4.4 Ga, impact-domi­ ing of the lunar crust by the older basin-forming events permitted nated dynamics of crustal formation resulted in the injection of urKREEP liquids to migrateinto the crust,removing the potential for liquids fromthe magma ocean into the crust. Such a process probably rapid, post-basin isostatic adjustment by urKREEP magma move­ helps to account forthe apparentincreasingly mafic character of the ment at the crust-mantleboundary. crustwith depth [11-12). Cryptomaria: The clear stratigraphiccorrelation of cryptomaria Thermal Requirementfor Rhenium-meltingtheMagmaOcean [25-29) with the old largebasin substage suggests that these units are Cumulates: Creationof a mega-regolith during the cratered high­ related to KREEP basalts or to partiallymelted, low-Ti late cumu­ landstage constituted a necessaryprerequisite forthe later remelting lates of the magma ocean.They underlie ejecta fromthe young large of magma ocean cumulates to produce mare basalt magmas. The basinsand may be represented in the Apollo samples by basaltsof increasingly insulating characterof the pulverized uppercrust would ages clearly greater than 3.92 Ga [30-31) or by KREEP-related slow the cooling of the residual magma ocean. It also would have basaltswith model ages of 4.2-4.4 Ga [32). allowed the gradual accumulation of radiogenic heat necessary to Core Formation: The associationoflunarmagnetic anomalies eventuallypartially remeltthe source regions in the upper mantlethat with the antipodesof post-Nectaris basins [33-34) and the initially produced the marebasalts and related pyroclastic volcaniceruptions. low accrection temperature of the lower mantle suggest that the The reverse wave of heating would proceed downwardinto the upper FexNiyS, liquid separatedfrom the early magma ocean did not coa­ mantlefrom the still molten andsignificantly radio-isotopic urKREEP lesce into a circulating core until about 3.92 Ga. As anomaliesdo not residual liquid zone at the base of the crust. appear to be antipodalto the Nectaris basin, and are apparently of Heterogeneous vs. Homogeneous Early Moon: The potential lower intensity antipodal to Orientale, then a dipole field may have effects of a giant, Procellarumbasin-forming event ca. -4.3 Ga [2] been active only between about 3.92 and 3.80 Ga, the respective andof a geographicallycoincident Imbrium event ca. -3.87 Ga [2] apparent ages of these basins. canexplain the surface concentrationof KREEP-related materials in Vesicles in Crystalline Melt Breccias: Remobilized solar­ the Procellarum region of the Moon [13]. Lunar Prospector y-ray wind H imbedded in the megaregolith of the cratered highlands spectrometer data indicate that the Procell arumevent excavated only probably was the dominant component of the fluidphase that formed relatively small amounts of material related to KREEP. This strongly vesicles in crystalline meltbreccias produced by large basin-forming suggests that urKREEP magmas had yet to move into the Moon's events [35). lower crust. The extensive movement of such liquids across and Vesiclesin Mare Basalt Lavas and Volatiles Associated with possibly along the crust-mantle boundary region to beneath Pyroclastic Eruptions: Remobilized H, derived fromthe decom­ Procellarum, however,may well have occurred in response to the position of primordial water, presumably was the dominant compo­ regional reduction in lithostactic pressure. The coincidental forma­ nent of the fluid phase associatedwith marebasalt vesicles [36] and tion of another large basin, the 1160-km diameter Imbrium basin, pyroclastic eruptions. The total absence of any indication of water nearthe center of Procellarumresulted in the redistributionofKREEP­ associated with this fluid phase demonstrates that all primordial related materials roughly radial to the younger basin. This scenario water in the source materials forthe magma ocean hasbeen lost to may make unnecessary recent proposals of a chemically asymmetric space or decomposed by FexNiyS, liquid separation [37] andmigra­ Moon [14-18) to account forthe surface concentrationof KREEP­ tion. related material aroundImbrium. Hydrogen Concentrationsat the Lunar Poles: The probabil­ Geochemical Dichotomy Between the Lunar Nearside and ity is high that the epithermal neutron anomaly detected over the Farside: The timingof the giant, SouthPole Aitken Basin-forming lunarpoles [38] is largely if not entirelythe consequence of concen­ event at the end of the cratered highland stage (-4.2 Ga.) [3] can trations of solar-wind H rather than cometary water ice [38-39]. account for the lackof both extensive KREEP-related material [13) Hydrogen and other solar-wind volatiles can be expected to be and basaltic maria [19) associated with South Pole Aitken. The concentratedin permanently shadowed areas [41). A continuous absence of anImbrium-size event in South Pole Aitken would have blanketof cometary water ice, unless fortuitouslycovered by protec­ kept hidden anyKREEP-rich crustal province. As would be expected tive ejecta fromlarger but very infrequentimpacts, probablywould with the removal of most of theinsulating uppercrust, relatively little erode [3] andbe lost in a geologicallyshort interval. mare basalt has erupted in South Pole Aitken [19), except possibly References: [1] SchmittH. H. (1991) Am. Mineral. 76, 773- in its northernportions [20). 784. [2] Wilhelms D. E. (1987) U.S. Geological SurveyProf Paper 58 Workhop on New Vews of the Moon l

1348. [3] Schmitt H. H. (1999) in Encyclopedia of the Space when formation of granite-greenstone terranes occurred, and the (H. Marks,ed.), Wiley. [4] Halliday, A. N. and Drake M. J. (1999) Phanerozoic subduction-related magmatism. Science, 283,1861-1863. [S] CameronA. G. W. (1999) LPS XXX, On Earth, the SHMS rocks of 2.5-2.1 Ga formedlarge igneous Abstract#ll50. [6] Agnor C. B. et al. (1999) LPS XXX, Abstract provinces within the Archean granite-greenstone cratons. For ex­ #1878. [7] Jacobsen S. B. (1999) LPS XXX, Abstract #1978. ample, on the Baltic Shield such province evolved on territory [8] Alfven H. and Arrhenius G. (1972) The Moon, 5, 210-225. measuring -1 million km2 and represented by different volcanics [9] Lee D. et al. (1997) Science, 278, 1098-1103. [IO] Palme H. (low-Ti picrites, Mg basalts,high-Al basalts, andesites,dacites, and (1999) LPS X, Abstract #1763.[1 l] Kahn A. and Mosegaard K. rhyolites, where basalts predominated) in riftlike structures, (1999) LS X, Abstract#1259. [12] Pieters C. M. and Tompkins gabbronorite dyke swarms, and largelayered intrusions [l). The S. (1999)LPSXX.X, Abstract#l286. [13] LawrenceD. J. et al. (1998) latter consist of dunites, harzburgites, pyroxenites, norites and Science, 281, 1484-1485. [14] Haskin L. A. et al. (1999) LPS X, gabbronorites, including pigeonite varieties,gabbr o-anorthosites, Abstract#l858.[15] FeldmanW.C.et al. (1999)LPSXX.X,Abstract Mgt gabbronorites, and diorites. The SHMS rocks on their major, #2056. [16]Wieczorek M. A. etal. (1999)LPSXX.X,Abstract#l548. rare,and rare earth elements contents were rather close to the Phan­ [17] Korotev R. L. (1999) LPS X, Abstract#1305. [18] Jolliff erozoic calc-alkaline series related to subduction zones, but on its B. L. et al. (1999) LPS X, Abstract#1670. [19] FeldmanW. C. geological positionthey had within-platetectonic settings andlook et al. (1998) Science, 281, 1489-1493. [20] Blewett D. T. et al. like the continental floodbasalt province. The ENin value of -1 to (1999) LPS XXX, Abstract #1438. [21) Morbidelli A. (1998) Sci­ -2 is characteristic for the studied SHMS-rocks, indicatingthat the ence, 280, 2071-2073. [22) Murray N. and Holman M. (1999) origin of such melts was linked with large-scale assimilation of Science, 283, 1877-1881. [23] FernandezJ. A. (1999) in Encyclo­ crustal rocks by high-temperature, mantle-derived magmas during pedia ofthe Solar System (P.R. Weissman et al., eds.),pp. 554-556, their ascending to the surface. Academic Press. [24] Schmitt H. H. (1989) in Workshop on Moon in Potassic granites andmonzodiorites and K-enriched volcanics, Transition, LPI Tech. Rpt. 89-03, (G. J. Taylor and P. H. Warren, variedfrom low-Ti picrites andtrachybasalts to alkaline andesites eds.),pp.111-112. [25]HawkeB.R.et al. (1999)LPSXX.X,Abstract and dacites, are not uncommonly associated with the SHMS rocks. #1956. [26) BellJ. andHawkeB. R.,JGR, 98, 6899-0910.[27) Clark According to isotopic data, these rocks also had mixed mantle­ P. E. and Hawke B. R. (1991) Earth, Moon, Planets, 53, 93-107. crustalorigin. [28) HeadJ. W. et al. (1993) JGR, 98, 17165-17169. [29) Williams Around 2.2-2.0 Ga, practically on the whole Earth simulta­ D. A. et al.(1995) JGR, JOO, 23291-23299. [30] Taylor L. A. et al. neously , the Fe-Ti picrites andbasalts of differentalka linity, similar (1983) EPSL, 66, 33-47. [31] Heiken G. H. et al. (1991) Lunar to the Phanerozoic within-plate rocks were appeared,and substituted Sourcebook, 209 pp. [32) Heiken G. H. et al. (1991) in Lunar the SHMS rocks as amain type of within-plate magmatic activity [2]. Sourcebook, pp. 218-219. [33) Lin R. P. et al. (1998) Science, 281, The lunar magmatism of the highlands was rather close to the 1481. [34] ] Lin R. P. et al. (1999) LPS XXX, Abstract #1930. early Paleoproterozoic magmatism ofthe Earth. The earliest (4.45- [35) Schmitt H. H. (1973) Science, 182, 682. [36] Schmitt H. H. 4.25 Ga) magnesian-suite magmatism,which intruded the primary et al. (1970) Apollo 11 LSC, 11-13. [37] Agee C. B. (1991) in anorthositic crust, was represented by volcanicsand intrusive rocks Workshop Phys. Chem. Magma Oceans (C. B. Agee andJ. Longhi, [3-4]. Pristineintrusive rocks of this suite arerepresented by dunites, eds.), pp. 11-12, LPI. [38] FeldmanW. C. et al. (1998) Science, 281, harzburgites,troctolites, norites, gabbronorites, and anorthosites 1496-1500. [39) Nozette S. et al. (1999) LPS XXX, Abstract #1665. (ANT-series). Important role in these rocks play orthopyroxene and [41) Watson K. et al. (1961) JGR, 66, 3033. pigeonite; Cr spine!, orthoclase, quartz, apatite, and Ti phases occur in many varieties. On their composition these intrusive rocks rather close to the rocks of the early Paleoproterozoic layered intrusions. As a whole, all these rocks on their major, rare and rare earth MAGMATISM OF THE LUNAR IIlGHLANDS AND THE elements pattern are rather close to the terrestrial early Palaeo­ EARLY PALEOPROTEROZOIC MAGMATISM OF THE proterozoic SHMS. What is more,the ENiT) value in the rocks of the EARTH: SIMILARITIES AND DISTINCTIONS. E. V. lunarmagnesian suite is about -I, which correlateswith data for the Sharkovand 0. A. Bogatikov, Institute of Ore Deposits Geology, terrestrialSHMS rocks (from-Ito -2). The main difference between Petrology, Mineralogy and Geochemistry, Russian Academy of these lunarrocks fromterrestrial rocks is a lower alkali content and Sciences, Moscow 109017, Russia ([email protected]). very subordinate role of acid and intermediate melts. According to Snyder et al. [3-4 ),the origin ofthese rocks was linked with contami­ Important featureof tectonic-magmatic activity on the Moon is it nation of high-temperature mantle-derived magmas •by lunarcrustal closeness to the Earth'searly Palae oproterozoic stage of evolution. material during their ascensionto the surface. The most ancient magmatism of the Moon was begun at highlands Rocks of the 4.3-4.0 Ga KR EEP series are the second type of fromlow-Ti melts of the magnesian suite, which were rather close in magmatismofthehighlands,where they areassociated with the rocks compositionof rocks,their mineralogyand geochemistry to the early of the magnesian suite [3-4]. They arecharacterized by elevated K Palaeoproterozoic terrestrial magmatism of the siliceous high-Mg content, REE and P alkalinity, and are represented by volcanics (boninite-like) series (SHMS). Around 3.8 Ga, it was changed by (mainly basalts) and their intrusive analogs "alkaline series" - basaltic magmatismof maria, where high-Ti varietieswere common. "alkaline" anorthosites, gabbro and gabbronorites, monzodiorites Thistype of magmatismis resembled the oceanic-typemagmatism of andpotassium granites could be finite members of the series. The the Earth(including Fe-Ti basalts of oceanicislands), which firstly rocks of these compositionand tectonic settingsclosely resemble the appeared in the late Palaeoproterozoic ca 2 Ga. On the Moon are enrichedKi n the earlyPaleoproterozoic terrestrialrocks of cratons. lacking both analogs of the Archean type of activity of the Earth, Thus, the lunar magmatism is close to only one (the Palaeo- LP/ Contribution No. 980 59

proterozoic), stage of terrestrialmagmatism evolution. The cause of however, such an anomaly has not been found in the very low-Ti this could be both in some differentprimordial composition of the (VLT) varieties. Traditionally, the presence of such ananomaly is silicate mantlesof these planetarybodies, andthe differentdegree of associatedwith fractionationof plagioclase in magmatic processes. their differentiation during solidificationof their planetary magma However, we suggest that in this particularcase this phenomenon is oceans: it suggests that the depth of the latter in the case of the Moon more likely linked with reduced character of the Moon interior, was-200-300 km, whereas in the Earth it was- 700 km. because in magmatic systems a valueof Eu2•/Eu3+ ratio depends on 2+ References: [l] Sharkov E.V. et al. (1997) Petrology, 5, 448- foz, and in reducing systems all Eu is represented by Eu , which 465. [2] SharkovE. V. andSmolkin V. F. (1997) Precambr. Res., 82, accumulated in the basic plagioclase. In contrast to the terrestrial 133-151. [3] Snyder G. A. et al. (1995) JGR, 100, 9365-9388. basalts, where both Eu isotopes occur and removing one of them [4] Snyder G. A. et al. (1995) GCA, 59, 1185-1 203. smoothed offthrough accumulation of the second, here only Eu2+ occurred and so its fractionation led to strong depletion of the residual melt in Eu. Consequently, appearance of the Eu anomaly could be linked with fractionation of plagioclase in transitional MARE MAGMATISM OF THE MOON AND OCEANIC magmaticchambers (intrusions) in the formof differentgabbroids. MAGMATISM OF THE EARffl: SIMILARITIES AND Because Eu content in basalts quickly risesunder growth of Ti [4), DISTINCTIONS. E. V. Sharkovand 0. A. Bogatikov, Instituteof this process is most effective in the case of the high-Ti melts. Ore Deposits,Geology, Petrology, Mineralogyand Geochemistry, 6. Beneath the lunar maria excess of masses (mascones) occurs Russian Academy of Sciences, 35 Staromonetny,Moscow 109017, and the lunar crustbecomes thinner above them In general, such a Russia ([email protected]). situation resembles a characterof distributionof Earth'scrust above continental rifts.Probably, the mascones are solidifiedmantle-plume 1. Maria with maficbasaltic floorsare the second major type of head, which are known on the Earth as large lens like bodies of morphostructuresof theMoon, which were formedin the late stages anomalousmantle beneath mid-oceanic ridges andcontinental rift of its evolution from 3.8 to 3.2 Ga [l]. They form large rounded zones [5]. depressions (10-km depths, [2]) in the lunarrelief. On their structure 7. So, there areessential similarities in structureand origin of the (presence of mid-maria moundlike rises and local rises), and on lunarmaria andterrestrial oceans. However, instead of the Earth, the basaltic rocks composition they resemble oceanic segments of the ancientlunar crust did not disruptwith formationof the lithosphere Earth with its mid-oceanicridges and within-plate oceanic islands. of oceanictype and further development according to a plate-tecton­ These type� of the Earth's structures also began to formlater than ics model. The process ended on the first stage, analogous to the continental segments of the lithosphere, fromca. 2 Ga [3]. terrestrialcontinental trapsformation. 2. Like on the Earth,two types of basalts aretypical forthe lunar References: [l] Snyder G. A. et al. (1994) GCA, 58, 4795- maria: low-Ti andhigh-Ti, which could be correlatedwith the mid­ 4808. [2] Spudis P. D. (1996) The Once and Future Moon, oceanic ridge basalts (MORB) and oceanic island Fe-Ti basalts Smithsonian Inst., 308 pp. [3) Bogatikov 0. (1999) Magmatism and (OIB) consequently. This type of tectonic-magmatic activity sur­ Geodynamics: Terrestrial Magmatism in the Earth's History, Gor­ vived until now on the Earth; its petrogenesisassociated with ascend­ don and Breech, London, in press. [4] Wilson M. (1989) Igneous ing of the mantle plumes that originated on the core-mantle boundary Petro genesis, A Global Tectonic Approach, Univ. Hyman, London, (CMB), in the layer D". For the terrestrialplumes it is characteristi­ 466 p. [5] Anderson D. L. (1992) Science, 256, 1645-1650. cally a presence of specific fluid components, enriched in Fe, Ti, alkalies,•,••, Zr,•••, etc. Concentration of the fluidcomponents is low in MORB and the highest in the within-plate Fe-Ti basalts. 3. Ocean-island basalt magmatism has a limited range in the MARE BASALTIC MAGMATISM: A VIEW FROM THE Earth'soceans andis often localized on the mid-oceanic ridge slopes SAMPLE SUITE WITH AND WITHOUT A REMOTE­ or farfrom them. With distance fromthe ridges axescomposition of SENSINGPROSPEC TIVE. C. K. Shearer1, J. J. Papike1 , and melts became more alkaline and titaniferous. The same picture is L. R. Gaddis2, 1Institute of Meteoritics, Department of Earth and common for the continental traps provinces, whose origin is linked Planetary Sciences, University of New Mexico, Albuquerque NM with superplumeascension; close to the MORB tholeiitic basalts are 87131, USA, 2U. S. Geological Survey, Astrogeology Program, predominate, and alkaline Fe-Ti basalts present in a limited range. FlagstaffAZ 86001, USA. Traps appearance oftenpreceded oceansopening andin this sense they could be definedas the frrststage of the oceandevelopment. Introduction: It haslong beenrecognized that the lunar-sample 4. By analogyof the Earth, it is suggested that petrogenesis of the suite returnedby the Apolloand Luna missions is biased with regard mare basalts couldbe also linked with plume activity, which were to its representation of lunarmare basalts [l-3]. This samplingbias ascendedfrom the lunar CMB of that time.Instead of the Earth, the is reflectedin both an incorrectportrayal of the volume of mare basalt lunarouter (liquid) core did not surviveuntil present, but had to be types andthe absence of many basalt groups known to exist from activeduring the mria formation. The maindifference of a fluid spectraldata [1-2]. This bias obviously affectsmodels forthe petro­ components of these plumes fromthe Earth's analog was the practi­ genesis of mare basaltsand the interiorof the Moon. Here, we explore

cal absenceof H20 in them; this indicates a mineralcomposition of the implications of this bias and comparemodels forlunar magmatism the mre basalts,where importantrole-play mineral phasesformed that arederived solely from samples with potential models derived under reduce conditions. from combined sample andremote-sensing data We focus on the 5. Another importantdifference is linked withthe lunarhigh-Ti implicationsof these contrastsin several areas: volume, distribution, basalts, which are characterized by a strong negative Eu anomaly; andage of marebasalts, KREEP enrichment on thenearside of the 60 Workhop on New Vews of the Moon I

Moon, heat sources for melting, and depth of mare basalt source of pre-3.9 Ga nonmarebasaltic volcanism [e.g., 7-8]. Most of this regions. record is retained in small clasts fromhighland soils andbreccias or Volume, Distribution, and Age of Mare Basalts: Volume of hasbeen identified through remote sensing. The relationship be­ mare basalt groups. The mare basalt samplesuite indicates that the tween the samples and units identified through remote sensing is TiO2 distributionof crystallinemare basalt samples is bimodal, with speculative. Further identificationand delineation of older episodes a majority of the marebasalts occurring in the rangeof 1.5-5.5 and of volcanism and their relationship to episodes of crustalplutonism

10--13 wt% TiO2. A compositional gap appears to exist between 6 (Mg and alkali suites) is critical to our interpretation of mantle and9 wt% TiO2• Although the population of picritic mare glasses evolution following magma ocean crystallization and prior to the

also exhibits a bimodal distributionwith regardto TiO2, it is domi­ onset of marevolcanism. Combined sample andremote sensing data nated by very low-Ti glasses (

global TiO2 distribution hasbeen interpretedto be continuousin the crust;and (5) melting of deep, hybrid mixed cumulate sources . . maria with no hint ofbiomodality and an abundancepeak between 1 The relationship between mare basalt TiO2 content and age is and3.5wt% TiO2 [3]. These new observationsindicate a mare source subject to controversy.Based on the sample suite, researchershave model in which a small volume oflate,ilmenite-bearing LMO cumu­ concluded that there is no correlation [9-10), a weak correlation lates mixed with a large volume of earlyLMO cumulates in which [11], or a significantcorrelation [7]. On the basisofremotely-sensed

ilmenite was absent. These differencesin models have implications data, researchers have concluded that TiO2-rich basaltstend to be for heat sources for generating mare basalts,the relative depth of older than TiOrpoor basalts, but within individual basins TiO2 various mare basalt sources, and the early dynamics of the lunar content and age are not correlated [6]. A simple time-composition mantle. correlation implies a systematic relationship between sequence of Distribution of mare basalts. Although remote optical andspec­ meltingand source region. Such a relationship suggests a model in tralobservations of the lunarsurface document the concentrationof which melting could be initiated in shallow high-Ti cumulates and marebasalts on the nearside of the Moon [e.g.,l], our limited sam­ progress into the deep lunar mantle. Both the lack of composition­ pling ofbasalts does not definethe overall distributionof marebasalt time correlation within individual basins [6] and a potential deep­ compositions. The relative distribution of high-Ti basalts should mantle origin forall primarymare basalts[8, 12] require a much more shed light on the asymmetry of LMO cumulates andrefine or refute complex process that involves the extent of magma ocean cumulate marebasalt models that require recycling of late, high-Ti cumulates mixing, distribution andcomposition of mantlereservoirs forheat­ into the deep lunarmantle. producing elements, and the mechanism(s) for magma transport to Lunar pyroclasticglasses have been identifiedat all of the Apollo the lunar surface. sites. Based on samples,the compositional variation of these glasses KREEP Enrichment on the Nearside of the Moon: Perhaps is bimodal andtheir overall abundanceis unknown. Understanding the bias of our lunar samples is best illustrated by recent global data both their compositional diversity and distribution is critical to for Th (=KREEP) abundance [13], which implies that the Apollo deciphering how these near-primarymagmas were transportedto the collection sites were all within a large, high-KREEP bulge on the lunar surface fromgreat depths and how they arerelated to crystalline nearside of the Moon. If KREEP distribution on the lunar surface mare basalts. Based on Clementine data andprevious work, more reflectsa real and preferreddistribution of heat-producing elements than 100 lunar pyroclastic deposits have been proposed [4] and in the lunar mantle, this has significant implications for our models potentially a large number consist of high-Ti glass beads [5]. The forlunar magmatism. First, it sugge·sts either an asymmetry in the existence of so many deposits that could consist of near-primary crystallizing LMO or an episode ofKREEP remobilization. Second, basaltic magmas implies that mechanismsfor their transport to the the irregularity of KREEP distribution early in lunar history will surface are not extraordinary and that density contrasts between modify our models forthe generation of both premare, highland,and melts and surrounding mantle do not significantly impede their mare magmatism. For example, if assimilation is an importantpro­ source segregation and movement, at high pressure, to the lunar cess in producing premare magmas,highland magmas outside the surface. KREEPbulge may not have a KREEP signature at all. On the other Age of mare basalt groups. While radiometric age dating of hand,if KREEP was recycled into the lunarmantle andwas instru­ lunar mare basalts provides a precise means of dating individual mental in initiating the mantlemelting that produced the highland samples, when it is combined with relative age relationships deter­ magmas, the highland units such as the Mg suite may largelybe a mined by remote sensing ( e.g., crater counts) it becomes a method for nearsidephenomena reconstructing magmatism on a planetary scale [e.g., 6]. Two ex­ Parental Magas for the Mare Basalts and the Depth ofMare amples where this approach has provided usefulinformation and will Source Regons: Much ofthe debate concerningthe origin ofmare continue to bear fruit are the duration and early history of lunar magmas has focused upon the nature of parental mare magmasand volcanismand the relationship between marebasalt composition and their depth of origin in the lunar mantle.Two possible models are( 1) eruptivehistory. that parental magmas to the mare basalts aresimilar in composition Although the petrologic record has been obscured by the early to the picritic glasses andwere derived from initialmelting in the catastrophicimpact history of the Moon, there is abundant evidence deep lunarmantle (>400 km) or (2) that marebasalts represent the LP/Contribution No. 980 61

crystallization of more Fe-rich, olivine-poor primary magmas that here summarize our current view of crustal structure and identify were produced at shallower depths. Combined analysis of sample some required knowledge to better understandthe origin and evolu­ andremote-sensing data may allow us to address these problems. tion of the lunar crust. First, if crystalline mare basalts andpicritic mareglasses definethe Models of Crustal Structure: Wood et al. [4] attempted to sameranges in TiO2 andoverlap in age, this could be interpreted as estimate the amountof plagioclase in the crust, based on the average indicating that they were produced by melting of similar mantle elevation differencebetween mare andhighlands and some simple assemblagesunder comparableconditions. Second, crystallinemare assumptions about anorthosite and basalt as responsible for the basalts have all the chemicalattributes (lower Ni, Mg', higher Al) principal lunar rock types. Later,more complex models emerged, associated with substantial olivine fractionation.Although picritic involving layered crusts of feldspathic material over more basaltic mare glasses and crystalline mare basalts in most cases were not material [5-6] or a laterally variable crust, with Mg-suite plutons related by shallowfractional crystallization, calculated liquid lines of intrudinga grossly anorthositiccrust [7]. Later models attempted to descent areparallel [14-15]. Therefore,the preservation of primitive reconcile these contrasting styles by incorporating both features magmas (i.e., picritic mare glasses) may be a function of magma [e.g., 8]. In part, crustal structure was inferred by the envisioned transportand eruptive mechanism rather thandramatically different mode of crustalformation. A decade-long debate on the reality of the source regions. Third, is the generation of mare basalts from a lunar"magma ocean,"stimulated by the provocativenotion ofWalker shallow, subcrustalsource between 3.9 and 3.0 Ga reconcilablewith [9] that the Moon never had a magma ocean,and the recognition that the thermalnature of the upperlunar mantle? The liquidus tempera­ the anorthositesand Mg suite probably recorded differentand unre­ tures of mare basalts exceed 1200°C at depths correspondingto -0.5 lated magmatic events [10]. Such a scenario leaves much about Gpa [16]. There is no mechanism to reheat the shallowlunar mantle crustal structure an openquestion, but allows for both lateral and 600 m.y. after the formation of the Moon and to maintain this vertical heterogeneity, thus accommodating both principal crustal temperature forover 700 m.y. Furthermore,the existence and pres­ models. ervationof mascons in multiringed basins that formed at -3.9 Ga New Data fromClementine and Lunar Prospector: Global could indicate a relatively cool and rigid uppermantle incapableof maps of Fe (11], Ti [11-12], andTh [13] both confirmold ideas and producing basalts. create new problems. It is clearthat vastareas of the lunarhighlands References: [1] Pieters (1978) Proc. LPSC 9th, 2825-2849. are extremelylow in Fe (11 ], consistent with a significantamount of [2] Pieters (1990) LPI-LPST Workshop on Mare Volcanism and anorthosite.Such a distribution supports the magma ocean.How­ Basalt Petrogenesis: Astounding Fundamental Concepts (AFC) ever, the average lunar highlandscomposition is, aslong suspected, Developed Over the Last Fieen Years, 35-36. [3] Giguere et al that of "anorthositic norite" [11,14], a mixed rock type,somewhat (1999) LPS XX. [4] Gaddis et al. (1999) LPS X. [5] Weitz et al. similarto manyof the lunar meteorites (e.g., ALHA 81005 [15] and (1990) JGR, 22725-22759. [6] Riesinger et al. (1999) LPS XX. more mafic than pure ferroananorthosite. Anorthosite proper does [7] Snyder et al. (1999) in Origin of the Earth and Moon (R. Canup occur on the Moon; it is found almost exclusively within the inner and K. Righter, eds.), in press. [8] Shearer and Floss(1999) in Origin ringsof multiring basins [16]. These basins span a range of ages and of the Earth and Moon (R. Canup and K. Righter, eds.), in press. distributions [17]. Maficprovinces occur in the central Procellarum [9] Papike et al. (1976) Rev. Geophys. Space Sci., 14, 475-540. region of the frontside and on the floor of the South Pole Aitken [10] BVSP (1981) Basaltic Volcanism on the Terrestrial Planets. Basin. In these areas, the lunarsurface is "highlandbasaltic" compo­ Pergamon,New York. [1 l]Nyquist andShih(1992) GCA, 56, 2213- sition(FeO -9-10 wt%). Additional highland basaltic areas occur in 2234. [12] Delano J. W. (1986) Proc LPSC 16th, in JGR, 91, D201- the vicinity of nearsidebasins, such as Serenitatis. The major lunar D213. [13] Lawrence et al. (1999) LPS XXX. [14] Longhi (1987) "hot spot" of high Th concentration (-10 ppm) occurs within a broad, Proc. LPSC 17th, in JGR, 92, E349-E360. [15] Shearer and Papike oval depression approximately coincident with OceanusProcellarum. (1993) GCA, 57, 4785-4812. [16] Longhi (1992) GCA, 56, 2235- Slightly less elevated amounts (-4 ppm) are associated with the 2251. basalticfloor of SPA Basin on the farside.Aside fromthis, Th highs are isolated and minor. A "New" Crustal Model: On the basis of the new global data, STRUCTURE AND COMPOSITION OF THE LUNAR as well as from our continuing study of the composition of basin CRUST. P. D. Spudis1, D. B. J. Bussey2, and B. R. Hawke3, ejecta to probe the deep crust [17], we have modified slightlyour 1Lunar and Planetary Institute, Houston TX 77058, USA existing crustal model [8] to accommodatethe new findings(Fig. I). ([email protected]), 2European Space Research and We proposea three-layer model of crustalconfiguration. The upper­ Technology Centre, European Space Agency, Noordwijk, The most zone, down to depths of -15-20 km, consists of megabreccia Netherlands,3 Planetary Geophysics Division, University ofHawai' i, of mostly anorthositic norite composition(FeO - 4-6 wt%; Al2O3 Honolulu HI 96822, USA. -26 wt%). This zone is neither laterally or vertically uniform [14], displaying anomalous compositionalzones at scalesof tens to hun­ Since the firstreturn of lunar samples indicated that globaldiffer­ dredsof kilometers, but is remarkablyhomogeneous at planetwide entiation of the Moon had occurred, numerous models of crustal scales. In bulk composition, it resembles the "ferroan anorthositic structurehave been proposed[summarized in l]. With the comple­ norite" suite of mixed rocks described by Lindstormet al. [18] and tionof the firstglobal reconnaissancemapping by Clementine [2] many of the highlands regolith breccias found as lunar meteorites and Lunar Prospector [3], we are now in position to re-evaluate [15]. It is alsosimilar to the average crustalcomposition inferred by crustalstructure and compositionat a global scale.Although this is Taylor (19], on the basisof Apollo granulitic breccias and limited a difficultand complex task, and one requiring significantstudy, orbital chemical data Although some areason the northernfarside some first-orderresults are apparentnow andare quite telling. We appear very anorthositic (11], most areas of the upper highlands 62 Workhop on New Views of the Moon I

resulting regolith is thus slightly contaminated witlrfeldspathicma­ terial, although dominantly basaltic. In the highlands crust, the inverse case holds. The original feld­ spathic, anorthosite crustof the Moon overlies a more mafic,"basal­ tic" lower crust.The massive impact bombardment of this target • produces a megaregolith of dominantly feldspathic, anorthositic composition.However, smallamounts of the underlying, maficlower crustis admixed into the megaregolith, making it slightly more mafic than the anorthositelayer it rests on. As in the caseat the Apollo 11 site, the dominant character of the "regolith" is that of the zone it immediatelyrests on, but is "contaminated" by the contrasting com­ position of the deep layer below it. References: [l] Heiken G. et al. (1991) Lunar Sourcebook, Fi. 1. Poposed moel of the lun cst. Cambridge. [2] Clementine issue (1994) Science, 266, 1777-1916. [3] Lunar Prospector issue (1998) Science, 281, 1475-1500. [4]Wood I. etal. (1970)Proc.Apollo 11 LSC, 965. [5] RyderG. and Wood I. (1977) Proc. LSC 8th, 655. [6] Charette M. et al. (1977) differentfrom this composition aremore mafic,not more feldspathic, Proc. LSC 8th, 1049. [7] James 0. (1980) Proc. LPSC 11th, 365. showing affinitiesto highlandbasalt, with or without KREEP. [8] Spudis P. D. and Davis P.A. (1986) Proc. LPSC 17th, in JGR, The next zone of the crust is foundat depthsbetween 15 and-35 91, E84. [9] Walker D. (1983) Proc. LPSC 14th, in JGR, 88, Bl7. km. It appearsto consist largelyof nearlypure, ferroananorthosite [10) Warren P. (1985) Annu. Rev. Earth Planet Sci., 13, 201. (FeO < 2 wt%; Al203 > 33 wt%). Outcrops of pure anorthosite [11) Lucey P. et al. (1995) Science, 268, 1150; (1998) JGR, 103, principally occur on the Moon in the inner rings of multiringbasins 3679. [12) Blewett D. et al. (1997) JGR, 102, 16319. [13) Lawrence (Fig. 1), which arestructurally uplifted blocks frommidcrustallevels D. et al. (1999) GRL, in press; Tompkins S. andPieters C. (1999) [20], or rarely,as centralpeaks in some selected craters (Alphonsus, MAPS,34,25. [14]WarrenP.etal.(1989)EPSL,91, 245. [15) Hawke Aristarchus [16]. The anorthosite is apparentlyconfined to middle B. R. et al. (1995) LPS XX.VI, 563. [16) Bussey D. B. J. and Spudis levels in the crust;moreover, it appears to be at least partly of global P. D. (1999) JGR, in press. [17] LindstromM. et al.(1986) Proc. extent, as basin rings of anorthositeare found in basinsspanning the LPSC 16th, in JGR, 91, D263. [18] Taylor S. R. (1982) Planetary globe, from Orientale [16-17] to Humboldtianum [21]. Because Science, LPI , 300 pp. [19] Spudis P. D. (1993) Geology Multi-ring anorthosite is most likely to represent the primordial crust [10], we basins, Cambridge Univ.,263 pp. [19a] Bussey D. B. I. and Spudis interpret this global "layer" of anorthosite as the remnant of the P. D. (1996) Eos Trans AGU, 77, F448. [20] Lucey P. et al. (1998) original crust of the Moon. JGR, 103, 3701. [22] Bussey B. et al. (1998) LPS XXIX, Abstract The petrological nature of the roughly halfof the crust below the #1352. [23] Spudis P. D. et al. (1996) LPS XXVII, 1255. anorthositezone (depths of35-65 km) remains obscure, butseveral observationsmay be made about the likely nature ofrocks to be found there. First, where most ofthe upper crusthas been removed by large, basin-formingimpact (such as the floorsof the SPA [22] and Imbrium THE MINERALOGY OF THE YOUNGEST LUNAR Basins [23]), the crustal composition appears to be that of highland BASALTS. M. I. Staid and C. M. Pieters, Box 1846, Department basalt (FeO- 9-12 wt%; Al203- 18-20 wt%). Second, a plot of the ofGeological Sciences, Brown University, Providence RI 02912, total Fe content of basin ejecta (determined fromorbital measure­ USA. ments; [11]) against basin size shows that larger basins excavate more mafic (Fe-rich) material [24]. This relation suggests that the Summary: The last stage of lunar volcanism produced spec­ lower levels of the crust aremore "basaltic" thanmiddle or upper trally distinct basalts on the westernnearside of the Moon, which levels. Finally, basinimpact melts, typifiedby the Apollo 15 and 17 remain unsampled by landing missions. The spectral properties of impact melt breccias [1,8], areof highlandbasaltic composition, and these late-stage basalts are examined using high-spatial-resolution were probably formedby massive melting ofmiddle-to-lower crustal Clementine images to constrain their mineralogic composition. The levels [5,6, 8]. These observationsin total suggest that the lower crust young high-Ti basalts in the western Procellarum and Imbrium is broadly "highlandbasalt" in composition. The petrologicalnature Basins display a significantlystronger ferrousabsorption thanearlier of the lower crust remains obscure, but the high Mg# and KREEP marebasalts, suggestingthat they may be the most Fe-rich deposits trace-element signature of basaltic impact melts from the Moon on the Moon. The distinct long-wavelength shape of this ferrous indicate that the Mg suite must represent a major contributorto the absorption is found to be similar for surface soils and materials bulk composition(Fig. 1). excavated from depth. The pervasive character of this absorption Origin of the Highlands''Megaregolith": The upperzone of featuresupports the interpretation of abundantolivine within these the Moon's crust has been heavily processed by repeated impact late-stage lunardeposits. cratering. In ourmodel, it is analogous to the regolith foundat any Late-StageHigh-Titanium Basalts: Importantdistinctions ex­ Apollo site in a shallow maria,such as Apollo 11 [l]. In that case,a ist between the early-stage easternmaria and the late-stage western thin lava flowoverlies a feldspathic, highlandssubstrate. Regolith basalts, even though both appear to be Ti-rich. For example, the formation by postmare craters mixes minor amounts of highlands western maria are more radiogenic than eastern deposits [1-3]. debris into a ground-up layer of mostly basalticcomposition. The Telescopic spectraof the high-Ti westernmaria also exhibit a unique LPI Contribution No. 980 63

combination of a strong 1-µm fature and a relatively weak or attenuated 2-µm absorption [4]. Pieters et al. [4] concluded that the 0.41/0.75 µm unusual stength cd shape of the 1-µm absortion in westes basalts Ratio Tranquil Ii tatis results fom an additional absorption fom abundct olivine andor .7 Fe-bearing glass. Either mineralogy could produce the stong long­ HDWA wavelength 1-µm bcd, but a glbsy Fe-rch surface could only frm by rapid cooling along the exterior surfaces of fows. Procellarum ClementineImagery: Clementine UV- VIS data of late-stage HDSA basalts ae examined fr regions in Ocecus Procellarum (HDSA) .65 and Mae Imbrium (hDSA). The spectal properties of westes re­ Imbrium gions are compaed to the sampled Apllo 11 basalts in Mae hDSA Trcquillitatis, which contain simila albedos and UV -VIS spctal proprues. Forrefrence, the westes basalts ae also compaed to the low-Ti cd Fe-rich bbats in Mae Serenitatis (mlSP). Serenitatis .60 Serenitati s bbalts have the stongest mafc absorption of cy ebtes neaside mTSP maria in Clementine imagery [5]. Unlike previous Eath-bbed cd Galileo imager, Clemenune data resolve the spectal propries of immature crater depsits small Reflectance @ O. 75 µm .55 ..______enough to sbple individual volccic fows [5-6]. A stategy has _ 6 been developed to reevaluate luna basat typs using Clemenune 10 14 18 Fig. 1: Scatter plot of the Clementine UV-VIS ratio vs. 0.75-µm imagery of such fesh mare craters cd their bsociated soils [5, 7]. To reflectance fr each mae study area. alow direct comparisons between regions, scatter plots of usefl spectal parameters were constucted by sbpling a fxed number of evenly spaced pixels fom each mae region. Scatter plots comparing the mae study areas ae shown in Figs. I cd 2. Since mature soils 1.0/0.75 µm dominate the surfaces expsed, the density distibution of each data Ratio cloud has been presented after a root stretch to enhcce the visibility Tranquillitatis of the less-abundct immature materials. Five-color spectra were HDWA also collected fr all fesh craters within each mae region cd 1.1 • t. grouped according to size. Sercnitati s Results: Titanium content. The UV-VIS ratio has been used l \ ntlSP extensively to estimate Ti in mature soils [8-9] cd plots of this parameter against 0.75-µm refectcce are included fr each mare i\ region in Fig. 1. The UV-VIS ratio coupled with the 0.75-µm paam­ -:;..-,..,., ...- eter hb been applied more recently to estimate Ti content across 1.0. mcy lunar materials [10-11]. High-Ti basalts plot in the upper left Procell arum portion of Fig. I because of their low-albedo and high-UV-VIS ratio HDSA Imbrium values. hDSA Clementine UV-VIS ratio values for the Procellaum HDSA unit ae similar to, but slightly lower thc, HDW A Apollo 11 basalts Reflectance @ 0.75 µm (purple cd dark blue respctively in Fig. 1). These values ae 0.9 '-'------·'--- consistent with previous evaluations of the westes high Ti basalts 6 10 14 18 using telescopic [9,12] and Apollo y-ray data [13], which suggest Fig. 2. Scatter plot of the Clementine mafc ratio vs. 0.75-µm reflectance only a minor difference in TiO2 contents between these mare depos­ its. The Imbrium hDSA cd Serenitatis mISP basalts (light blue cd fr each mare study area. orange in Fig. I) are seen to be progressively less dak cd blue, consistent with the previously noted decreasing bount of weight [10,12,14). percent TiO2 Maficmineralogy. The scatter plot in Fig. 2 captures the I µ absorption stength cd abedo of lage aeas fr each study region opaques (which should subdue absorption fatures) the wester over a rcge of optical maturities. Tis scatter plot allows tends HSA cd hSA mare units also exhibit a stonger mafc ratio than related to maturity to be evaluated [5,15-17]. Materials whose soil the Fe-rich Serenitaus basalts. These comined proprties indicate surfaces hae not acheved optical maturity ae slightly brighter cd c excepuonally high abundance of mafc minerals cd suggest that display a stonger frous bcd. For each bbalt type, the result is a the Eatosthenic deposits within Procellarum may be the most Fe­ roughly parallel rcge of values fr these spctal paameters fr­ rich basalts extuded on the surface of the Moon. It is difcult to ing a disunct "weathering cloud" of data [5]. esumate the FeO content of these young basalts since reted The westes HSA cd hSA bbalts show a much stonger sbples demonstate that all luna soils contain a faction of freign mafc rauo tc the Tranquillitatis basalts fr both matue soils cd materals cd mae sois have a lower weight percent FeO thc thei ime crater materals (Fig. 2). Despite a higher abundance of bsociated basalts [18-19]. We ae in the process of considering such 64 Workhop on New Views of the Moon I

16 Waen P.(1991) in TheLunar Sourcebook, pp. 357-474. [19] Laul J.C. cd Papike J.J. (1980) Proc.LPSC 11th, 1307-1340. 15

14 THE LUNAR ATMOSPHERE AND ITS INTIMATE '# mSP CONNECTION TO THE LUNAR SURFACE: A REVIEW. 13 80302, hSA S. A. Stes, Southwest Reseach Institute, BoulderCO USA 12 ([email protected]). HSA

11 HWA Due to the lack of optical phenomena asociated with the luna atmosphere, it is usually stated that the Moon ha no atmosphere. 10 This is not corect. In fact, the Moon is suounded by a tenuous envelope with a surface number density and presure not unlike that 9 of a cometary coma. However, the calogy ends there: The luna Wavelength (nm) 8 atmosphere is essentially everywhere collisioness, unlike a cometary 400 800 1000 coma, its compsition is quite different fom that of any comet, cd the extct lunar spcies do not create optically bright emissions. Fig. 3: Five-color Clementine spectra of the least weathered materials Since the luna atosphere is in fct c exosphere, one cc think of identified in each study region. (To view this figure in color, go to its various compositional components as "indepndent atmospheres" http://cass.jsc.nasa.gov/meetings/moon99/pdf/8026.pdf.) occupying the same space. Uppr limits derived by Apollo-based instments indicate that sbple infraton cd mixing issues in order to estimate the actua the entire native envelope weighs only -10 t. However, the com­ FeO aundances of the mac-rich westes baats. plexity cd scientfc value of this atosphere ae not commensurate Regions that represent the most immature materials within each with its low mas. mae aea were selected by identfing pixels that corespond to the The luna atosphere contains vital inforation about the loca­ lower-right limit of each mare unit's 1-µmvs. 0.75-µmscatter plot tion of nea-suace volatiles, including water; it also acts as a cloud in Fig. 2. These specta, shown in Fig. 3, allow comparisons of reseroir of gases released fom the interior cd may even mirror the the stong frous absorption fr the most crystalline materias within composition of certain surface-lying mineraogical units. Further­ each basalt type. The shape of the 1-µmfeature is much fatter cd more, the luna atosphere is the most accessible of the sola system's centered at a longer wavelength in the spectra of the westes SBEs, cd ofers a rich variety of physical processes to study as Procellarum basalts compared to the easter Serenitatis and caogs to other SB Es across the solar system. An importct advan­ Tranquillitatis baalts. Spectra of immature materials fom the HDSA tage of the Moon for such studies is that we enjoy abundant surface unit exhibit the strongest downtu in the ferous bcd at 1 µm of al samples cd orbital geochemical data that provide key "bounday the mae units in Fig. 3. conditions" to the physics cd chemistry at work, unlike cy other A strong and long-wavelength mafc absorption was fund to be similarly expsed planetary surface. pervasive throughout the fesh materias excavated by craters of In this review I will summarize the present state of knowledge digerent size groups. Mature soils within the wester high-Ti depos­ about the lunar atmosphere, describe the importct physical pro­ its also have a strong and long-wavelength ferous band [4].The cesses taking place within it, cd provide a compaison of the lunar presence of this absorption throughout the soils cd crater deposits atmosphere to other tenuous atmospheres in the solar system.Par­ of westes high-Ti basalts indicates that this optical property is ticular emphasis will be placed on the intimate connection between inherent to the mineralogy of the basalts.This observation excludes the luna atmosphere and its source, the lunar regolith. a glassy fow surface a a source fr the 1-µmfeature and instead supports the atesate interpretation of abundant olivine [4] within these Fe-rich basalts. APOLLO 17 SOIL CHARACTERIZATION FOR RE­ References: [l] Soderblom L. A. et al. (1977) Proc. LSC 8th, FLECTANCE SPECTROSCOPY. L.A.Taylor 1, C. Pieters2, 1, 1191-1199. [2] Adbs J. B. et al. (1981) BVTP, 438-490. A. Patchen 1, R.V. Moris3, L. P.Keller, S.Wentworth 3, and D.S. [3] Lawrence D. J.et al.(1998) Science, 281, 1484-1489. [4] Pieters McKay3, !Planetary Geosciences Institute, University ofTennessee, C. M. et al. (1980) JGR, 85, 3913-3938. [S] Staid M. l.cd Pieters Knoxville TN 37996, USA ([email protected]), 2Depament of C. M. (1999) Icarus, submitted. [6] Staid M. I. cd Pieters C. M. Geological Sciences, Brown University, Providence RI 02912, USA, (1996) LPS XX.Vil, 1259-1260. [7] Staid M. I. and Pieters C. M. 3MalCode SN, Johnson SpaceCenter, Houston TX 77058, USA, (1998) LPS XX.IX, Abstact #1853. [8]Chaette M. P. et al. (1974) 4MVA, Inc., Oakbrook Pakway, Norcross GA 30093,USA. JGR, 79, 1605-1613. [9] Pieters C. M. (1978) Proc. LPSC 9th, 2825-2849. [10] Lucey P. G. et a. (1998) JGR, 103, 3679-3699. It is the fne factons that dominate the obsered spectal signa­ [11) Blewett 0.T. etal. (1997)JGR, 102, 16319-16325. [12] Johnson tures of bulk luna soil [1-5], and the next to the smalest size J. R. et a. (1991) JGR, 96, 18861-18882. [13) Davis P. A. (1980) factons are the most similar to the overall properties of the bulk soil JGR, 85, 3209. [14] Melendrez et al. (1994) JGR, 99, 5601-5620. [4]. Thus, our Lunar Soil Chaacterzaton Consortium [6-8) has [15) Fischer E. M. and PietersC. M. (1996) JGR, 101, 2225-2234. concentated on understanding the inter-relations of compositional, [16) Lucey P. G. et al. (1995) Science, 268, 1150-1153. [17] Lucey mineralogical, and optcal proprtes of the <45-µmsize facton cd P. G. et al. (1998) JGR, 103, 3679-3699. [18] Hakin L. A. and its compnent sizes (20-44-µm,10-20-µm, and <10-µmsize fac- LP/ Contribution No. 980 65

tions). To be able to generalize our results beyond the particular sample set studied, it is necessary to quantitatively identify the observed effects of space weathering and evaluate the processes involved. For this, it is necessaryto know the chemistryof each size 71061-i, 71501-os 70181-47 79221-tn fraction,modal abundances of each phase, average compositionsof F:16 17A 16 ·154 1= 7. 9 81 8.5 the minerals andglasses, 1 5/FeO values, reflectancespectra, and the 100 ■ physical makeup of the individual particlesand their patinas.This 0lww characterization includes the important dissection of the pyroxene 80 El-GI G3 oa.n minerals into four separatepopulations, withdata on both modes and 60 average chemical compositions. Armed with such data, it should be □- • Pig possible to effectively isolate spectral effects of space weathering 40 !:S� from spectral properties related to mineral and glass chemistry. 1 AC 20 Four maresoils fromthe Apollo 17 site were selected for charac­ terization based upon similarities in bulk composition and their 0 contrastingmaturities, ranging fromimmature to submature to ma­ --""'<11) ---<10 ----<111 ---<111 ture. The methodology of our characterizationhas been discussed previously, [e.g., 9]. Results of the Apollo 17 mare soils, outlined herein, are being preparedfor publicationin MAPS. Fig. 1. Cumulative modal percentage of Apollo 17 mare soil Modal Analyses: As shown in Fig. 1, with decreasing grain components. size, the agglutinitic (impact) glass content profoundly increases. This is the most impressive change for the mare soils. In several soils we have examined,there is an over two-foldincrease in the agglutinitic glass contents between the 90-150-µm and the 10-20-µm size fractions[7-9]. Accompanyingthis increase in agglutinitic glass is {SizeFractlOIIII: 20-44,10•20, u.. A 70111-47 their similarities in FeO and Ti02 content, allowing for direct com­ 14 parisons between evolutions of chemistry betweensize fractionsand ■ 79221-81 -14, •35, -47, ..1, among different maturities of soils. The bulk chemistry of these 12 1./FeO lo<90-150 fractions was determined by EMP analyses of fused glass beads. (5 Aggi. Glass � In contrast to the systematic variations in bulk chemistry dis­ cussed below, the relatively uniform composition of agglutinitic 10 12 14 IG 18 glass with grain size and soil maturity is illustrated in Fig. 2. The A'20, composition of the bulk fraction of each size fractionbecomes more feldspathic with increasing maturity, with the effect being most Fig. 2. Apollo 17 mare soil chemistry. pronounced for the finestfractions. The composition of the agglutinitic glass, however, is relatively invariant and more feldspathic(i.e., rich

in Al203) than even the < 10-µm fraction. This relation not only strengthens the "fusion of the finest fraction" (F3) hypothesis [11- with its single-domain Fe0 component. This would seem to indicate 12], but also highlights the important role of plagioclase in the that at least some of the Fe0 is surface correlated. To illustrate this formation of agglutinitic glass. effect, if it is assumed that the nanophase Fe0 is entirely surface As described by Taylor et al. [8], with decreasing grain size, FeO, correlated, then equal masses of 15-µm and 6-µm spheres should 0 of MgO, and Ti02 contents decrease, whereas CaO, Na20, andAl 203 have -3x as much Fe in the finer fraction. The recent findings (plag components) increase for all soils. This relationship is illus­ Keller et al. [ 13-I 8] of the major role of vapor-deposited,nanophase trated in Fig. 3. These chemical variations would appear to be Fe0-containingpatinas on most soil particlesis a majorbreakthrough coupled with the significant increase in agglutinitic glass and de­ in our understanding of the distribution of Fe0 within agglutinitic crease in oxide(ilmenite), pyroxene, andvolcanic glass. These changes glass andupon grain surfaces. in chemistry do not appear to be due to distinct changes in the Reflectance Spectra: Bidirectional reflectance spectra for a compositionsof individual phases but to their abundances. representativeApollo 17 soil (70181) areshown in Fig. 4. The size Ise Vaue: As shown in Fig. 3, values ofl/FeOincrease separates all have similar albedo in the blue andfollow a regular with decreasing grain size, even though the bulk FeO contents sequence in which the continuum slope increases, ferrous bands decrease. That is, the percentage of the total Fe that is present as weaken, and albedo increaseswith decreasing particlesize. The bulk nanophase Fe0 hasincreased substantially in the smaller size frac­ <45-µm soil is typically close to the 10-20 µm spectrum. It is tion. Note that the increase in nanophaseFe 0 in smaller size fractions importantto note that although the finestfraction ( <10 µm)is close is significantlygreater thanthe increase in agglutinitic glass content, in composition to the abundantagglutinitic glass in each sizefraction 66 Workhop on New Views of the Moon l/

Taylor (1998) New Views, LP! Contribution No. 958, 6. [6] Taylor et al. (1998) LPS XXIX. [7] Taylor et al. (1998) New Views, LP! 71061-14• 71501-35" 70181-47" 79221-11• Contribution No.958, 71. [8] Taylor et al. (1999) LPS XXX. FeO 16.2 17.4 16.4 154 [9] Taylor et al. (1996) 5596. [10] Morris (1977) A Icarus, 124, Proc. TN'\. 7.3 Q� Af .� 3719. [11) Papike et al. (1981) 409. 11 LSC 8th, Proc. LPC 12B, 11 'f:o0 g 14 � [12] Walker and Papike (1981) Proc. LPSC 12B, 421. [13) Keller 12 °" 10 v '°'/ .. and McKay (1997) GCA, 61, 2331. [14] Keller et al. (1998) LSC

12 > XXIX. [15] Keller et al. (1998) New Views, LP! Contribution No. 958, 44. [16) Keller et al. (1999) LPSC XXX. [17] Wentworth et al. 1D . : (1999) MAPS, in press. [18] Keller et al., this volume. • "no.:� � • D.15 GA ,,..-,,- . G.3 .Ht,,,0 GLOBAL GEOCHEMICAL VARIATION ON THE LUNAR 150 ls/FeO SURFACE: A THREE-ELEMENT APPROACH. D.R. 10 Thomsen!, D. J. Lawrence1, D. Vaniman1 , W. C. Feldrnan1 , R. C. SD . l 1 2 3 . _-, Elphic , B. L. Barraclough , S. Maurice , P. G. Lucey , andA. B. ...,.._...,...... ,.._.....,..._ ,_...,.._.....,..._ Binder4, 1Mail Stop D-466, Los Alamos NationalLaboratory, Los • ls/FeOof!I0-150 ---J1J11 [10) -- AlamosNM 87545, USA ([email protected]),2 ObservatoireMidi­ Pyrenees, 14 aveEdBelin, 31400Toulouse,France, 3 Hawai'i Institute Fi. 3. Soil chemistry. of Geophysics and Planetology, University of Hawai'i, Manoa HI, USA, 4 Lunar Research Institute, 1180 Sunrise Drive, Gilroy CA 95020, USA.

Introduction: We present a method fordisplaying the relative abundancesof three importantelements (Th, Fe, andTi) on the same map projection of the lunarsurface. Using Th-, Fe-, andTi-elemental abundances from orbital geochemical data and assigning each ele­ ment a primary color, a false-color map of the lunar surface was created (Fig. I). This approach is similar to the ternary diagram approach presented by Davis and Spudis [1] with some important differences, discussed later. For the present maps, Th abundances were measured by the Lunar Prospector (LP) Gamma-Ray Spectrom­

D ...... ,...._...... _....._...... ,.._._.....,...... _._.,_...... _....._._. eter (GRS) [2 J. The new LP GRS low-altitude dataset was used in this 500 1000 1500 20002500 analysis [13]. Iron and Ti weight percentages were based on Wavelength (nm) Clementine spectral reflectance data [3] smoothed to the LP low­ altitude footprint. Thismethod of presentation was designed to aid in Fi. 4. Bidirectional reflectance spectrum. the location and recognition of three principal lunar compositions: ferroan anorthosite (FAN}, mare basalts (MB), and the Mg suite/ KREEP-rich rocks on the lunarsurface, with special emphasis on the highlands and specific impact basins [1,4-12]. In addition to the (see above), thissize fractionis relatively featureless and does not recognition of these endmember rock compositions, this method is dominate the spectrum of the bulk <45-µm soil. an attempt to examinethe relationship between elemental composi­ Volcanic Glass: It has long been suspected that agglutinitic tions that do not conformreadily to previously accepted or observed glass, to a large extent,is the product of meltingof the finestfraction endmember rocks in various specific regions of interest, including of the soils [11-12], with a dominanceof plagioclase. Given the low easternhighlands regions centered on 150° longitude, anda northern abundance of pyroxene in the finestfractions of each soil (Fig. 1 ), the highlands Th-rich region observed by Lawrence et al. [13). source of the FeO in these Apollo 17 agglutinitic glasses is not fully Method: The LP low-altitude data has fullwidth at half-maxi­ identified.We suspect the abundant volcanicglass in these samples mum spatial resolution of -40 km [13]. The Clementine spectral may be a significant contributor and thishypothesis will be tested reflectance datasets were adapted using an equal-area, gaussian with the suite under study fromother Apollosites. If the abundance smoothing routine to this footprint. In addition, these datasets, re­ of Fe-rich volcanicglass playsa significantrole in the composition ported in weight percent of FeO and of Ti0 2, were adjusted to Fe and of agglutiniticglass in these soils, it may account forthe significantly Ti weight percentages. Each dataset was then assignedone of the lower albedo observed remotely forareas with abundant volcanic three primary colors: blue forTh, red forFe, and green for Ti. For glass. each element, the data range was normalizedto represent the ratio of References: [1] Pieters (1993) Remote Geochem. Anal., Cam­ each point to the maximum in the dataset. This color table provides bridge Univ., 309pp. [2] Fischer and Pieters (1994) Geol. Remote a scheme asfollows: pure red, green, or blue indicatesthe presence Sens., 1287. [3] Fischer and Pieters (1994) Icarus, 111, 475. of the pure form of the respective element; white indicates the [4] Fischer (1995) Ph.D. dissertation, Brown Univ. [5] Pieters and maximum values of all three elements (Th "'10.3 µgig, Fe "'15.6%, LP/ Contribution No. 980 67

andTi =7.5%); black indicates the lack ofor exceedingly low values good compositional indicators of differentmare basalts. Mixed with of each of the three elements. Any other color indicates a different red for Fe, the green for Ti produces a yellow signal in high-Ti mixture of the three. (To view the color table, go to http:// basalts. While universally high in Fe relative to the surrounding cass.jsc.nasa.gov/meetings/moon99/pdf/8033.pdf.) The full range highlands,mare basalts have a diverse rangeof Ti values, making Ti of lunarlongitudes is represented, but due to the lack ofcoverage of concentrationa valuable asset to the classificationand identification the Clementine data forlatitudes > 70° and<-70 °, the data for these of differentbasalt types. Finally, animportant constraint in element regions is excluded. The differences between this approach and the selection is the availability of the global data, both from LP and ternarydiagram approach ofDavis andSpudis [ 1] eliminate some of Clementine results. Data for Th, Fe, and Ti areamong the highest the uncertainty and ambiguity of the ternary diagram approach. qualityof existing lunarremote-sensing data In addition, LP data for Rather than using a ratio of Th to Ti normalizedto CI chondritic Fe and Ti will become available, enabling these data to be incorpo­ ratios,and a ternarydiagram with ternaryapexes located at specific rated into the analysis. end.membercompositional values, elemental compositionswere used Results: Using upper-limit valuesforen dmember rock compo­ independently, eliminating the errorsresulting fromdividing num­ sitions calculated from Korotev et al. [14], attempts were made to bers that canhave high uncertainties,especially at low concentration locate the different end.member compositions of terranes on this [13]. diagram. Most strikingly, ferroan anorthosite (Th� 0.37µg/g; Fe The three elements used in this method of presentation were (wt%)� 2.29; Ti (wt%)� 0.22) [13], which should appear as an chosen forseveral reasons. One reason forthe inclusion ofTh in this almost black, reddish color, does not appear on the diagram at any study is that it is an accurate indicator of KREEP. Iron and Ti noticeable frequency.Based on this analysis,the suggestionof exten­ concentrationsare both low in highlandregolith, causing any small sive FAN regions on the lunarsurface is not strong,especially at the fluctuationsin Th to standout very well. In addition, Fe and Ti are presently accepted valuesfor Fe andTh. However, to make sure this

'Y21 t;:, lfZ.l .-.-��A.,:,zr.� 1.e:� �--�

·t,. :::, ::,,., '!. "'� IJ 1,,_.,;.o;�

N 0

p 0 0

1) Mll> I...:\ TITUDES

F. 1. 68 Workhop on New Views ofthe Moon l

effectis not due to systematic errors,a thorough investigation of the SIDEROPHILE ELEMENT SYSTEMATICS AND THE precision, accuracy, and uncertainties of the Fe, Ti, and Th abun­ MOON'S CORE-MANTLEDIFFERENTIATION HISTORY. dances needs to be carriedout, especially at low concentrations. P. H. Warren, Institute of Geophysics and Planetary Physics, A particular region of interest is anarea of high Th concentrations University of California,Los Angeles, Los Angeles CA 90095-1567, relative to Fe and Ti content north and east of Humboldtianum USA ([email protected]) Crater. First observedby Lawrence et al. [13), this region does not coincide with anyvisible impact structureand comprises one of the The most highly siderophile elements (HSEs), i.e., the most closest approximations to pure blue (high Th, very low Ti and Fe) on sensitive indicators of planetarymetal-silicate differentiation, are the the lunar surface.Such anelemental composi tiondoes not lend itself heavier platinum-group elements: Os, Ir, andPt, andtheir periodic­ readily to classification, and presents something of an anomaly. table neighbors Re andAu. The HSE that hasbeen determined most More detailed analysis of this region is needed to understand its often and most successfullyin lunar rocks is Ir. Even so, previous structure and origin. estimates of the concentrationof Ir in the lunarmantle indicate great There seems to be a longitudinal asymmetryin the Th concentra­ uncertainty, as they rangefrom 0.01 [l] to 4 [2] nglg, or 0.00002to tions of the highlands regolith. High-Th, low-Ti, andFe regions are 0.009 x CI chondrites.This uncertainty stems fromseveral factors, located between 135° and180 ° longitude andbetween-30 ° and+30 ° including the "nugget effect"and other analyticaldifficulties. The latitude. While the Th levels are not high enough to attract attention lunar HSE database is scattered amidst many papers, andin most in a single elemental display, the variation in the abundance of Th cases analysesfor Ir (andother HSE) did not simultaneously deter­ relative to Fe andTi abundancescan be clearlyseen. The composi­ mine compatible-lithophile elements. Previous interpretations of tion that these data suggest is not well represented in the sample­ thesedataintermsofmantleconcentrationshaveinvolvedlittlemore return suite. In addition, these regions were largelymissed by the than comparisonbetween mean concentrationsin lunar vs. terrestrial Apollo orbital ground tracks, which only covered the outer edge of basalts [3). Such interpretationsare inadequate, because basaltsare the areasof interest [15). The LP orbitalTh data represent the first diverse. Even the most extremelysiderophile elements exhibit com­ information about the Th concentrations in these regions of the patible behavior under metal-absent conditions [4). I have compiled highlands. a database that is a comprehensive blend of siderophile data with Regions containingMB and Mg suite/KREEP-rich composi­ compatible-lithophile data foridentical basaltic samples.Over 50% tions arereadily apparent.MB deposits (average composition: Th== of the HSE data are fromthe compilation of Wolf et al.[5]; many of 0.913µglg, Fe == 14.4%, and Ti == 2.5%) appear as yellowish and the other data are from past publications frommy UCLA· lab. This appearin the Serenitatis-Tranquilitatis region, among others. This database manifests correlations between siderophile elements and particular rock type is the most varied in color because of widespread several compatible elements that imply relatively tight constraintson variation in Ti content, and the contamination of these provinces by the lunar mantle siderophile composition. KREEP-rich material. Mafic, KREEP-rich impact melt breccias The notion that FeNi metal may have been a residual mantle phase (average composition of Th rangingfrom 5 mg/g to concentrations that determined the HSE compositions of lunar basalts [3] is pre­ higher than 10 µgig, Fe== 7.18%, andTi== 0.83%) are shown on the cluded by a correlationbetween W (which would be markedly sid­ map as pink regions such asthe ones nearCopernicus and in the Jura erophile in the presence of metal) and U [6). Conceivably residual mountains. This three-elementmapping technique is advantageous sulfidescontrolled the HSE. But textural-mineralogical observations in the visualization of different KREEP-rich terranes because, in preclude S saturation in the low-Ti basalts [7], and experimental data contrast to the single-element Th map, this technique allows Fe and [8] indicate that even the most S-rich high-Ti types erupted -50% Ti compositional variations within KREEP-rich regions to be readily undersaturated in S. seen, aiding in the identificationof underlying rock types and allow­ Nickel, although not quite a HSE like Ir, is a strongly siderophile ing for a better understanding of the processes behind KREEP element with a very extensive database. A plotofMgO vs. Ni (Fig. 1) emplacement. manifests clearly significant, albeit nonlinear, correlation trends. Acknowledgments: This work was supported in part by The lunartrend parallels that for Earth basalts and komatiites (and a Lockheed Martin under contract to NASA and conducted under the similar trend for martian rocks [9]), but is displaced to lower Ni. The auspices of the U.S. Department of Energy. MgO contents of the two mantles differ slightly, but mostly this offset References: [l] Davis P. and Spudis P. (1987) Proc. LPSC reflects an -5x lower (Ni) in the lunar mantle vs. its terrestrial 17th, in JGR, 92, in E387-E395. [2] Lawrence D. J. et al. (1998) counterpart. The terrestrialtrend passesthrough the MgO content of Science, 281, 1484-1489. [3] Lucey P.G. et al. (1999) GRL, submit­ Earth'smantle at Ni==2200 µgig. The implied Ni content of the lunar ted. [4] Duncan A. R. et al. (1976) Proc. LSC 7th, 1659-1671. mantle is - 400µgig. The hypothesis [2] that (Ni) is the samein the [5] McKay G. A. et al. (1978) Proc. LPSC9th, 661-Q78.[6] Longhi lunarand terrestrialmantles is clearlyuntenable. Ni shows similarly J. and Boudreau A. E. (1979) Proc. LPSC 10th, 2085-2105. strongcorrelations vs. Cr and V (except among terrestrial rocks, [7] NormanM. D. andRyder G. (1978) Catalogs of Pristine Non­ where Vis generally incompatible with olivine). mare Rocks, NASA JSC-14565 and -14603. [8] Ryder G. (1979) A plot of MgO vs. Ir (Fig. 2) shows greaterscatter, but the same Proc. LPSC 10th, 561-581. [9] NormanM. D. andRyderG. (1980) basicrelationships. Again, the lunar data forma trendthat parallels Proc. LPSC 10th, 317-331. [l0]Walker D. (1983)Proc. LPSC 14th, the Earth trend,but at a displacement indicating -6x lower (Ir)in the in JGR, 88, B17-B25. [11) Davis P. and Spudis P. (1985) Proc. lunar mantlevs. its terrestrial counterpart.This same inference ex­ LPSC 16th, in JGR, 90, D61-D74. [12) Davis P. and Spudis P. tends at least to Os, which shows anexcellent correlationvs. Ir, at (1986) Proc. LPSC 17th, in JGR, 91, E84-E90. [13) Lawrence D. J. chondritic Os/Ir, amonglunar basalts. The lunarmantle Re/Ir (and et al. ( 1999) GRL,submitted. [14] Korotev R. et al. ( 1998) JGR, 103, Re/Os)is also at least nearlychondritic; Au/Ir,however, gives some 1691-1701. [15) Arnold J. et al. (1977) Proc. LSC8th, 945. indication of being as much as 3x greater than chondritic. The LP Contribution No. 980 69

10 ,------should have been depleted in the lunarmantle to about 0.0002x the bulk-Moon concentration (which in any event is presumably not chondritic forsiderophile elements, given the Moon's gross deple­ tion in Fe metal). For Ni, the metal/silicate D is roughly 25x lower 1000: [12], so core-mantleequilibration forNi should have resulted in!:!,.= 0.005 and Ni/Ir= 25x the bulk-Moon ratio (and the implication regardingNi/Ir is approximately the samefor all finitevalues of c). cii By manymodels [e.g., 2], the Moon formed largely from,and thus 1100 inherited its bulk composition largelyfrom, Earth's mantle.If a core z subsequently formedwithin the Moon, the Moon-mantleIr/Ni should 0 have been depleted by anadditional factor of 25 vs. the Earth-mantle value. Thus, it is remarkable that Ir/Ni is so nearlyidentical in the two o mar, loT mantles. A gross disequilibrium between the mantleand core, i.e., • mar, llthl fgnts ■ mare,high-T accretion of a "late veneer," is most unambiguously implied forthe < mare,pyroclastic o nnmar low-1i Moon, even though the proportion of the veneer (0.1 wt% of CI-like • • + Earth material, based on Ir) is much less forthe Moon's mantlethan often 1.L------�-_.;=::::::;===;;::::::::::I0 4 8 12 16 2 24 28 3 suggested (- 0.7 wt% [e.g., 9]) forthe Earth's mantle. MgO(w) References: [1] Newsom H. E. (1986) in Origin of the Moon, Fig.1. pp. 203-229. [2] Ringwood A. E. (1992) EPSL, 111, 537-555. [3] WolfR. andAnders E. (1980) GCA, 44, 2111. [4] Naldrett A. J. andBarnes S.-J. (1986) Fortschr.Mi neral., 64, 113-133. [5] Wolf 10 �------, R. et al. (1979) Proc. LPSC 10th, 2107-2130. [6] Palme H. and Rammense W. (1981) Proc. LPS 12th, 949-964. [7] BrettR. (1976) GCA, 40, 997-1004.[8] DanckwerthP. A. et al. (1979) Proc. LPSC + 10th, 517-530. [9] Warren P. H. et al. (1999) GCA, in press. + [10] Tschauner 0. et al. (1998) in Origin of the Earth and Moon, pp.

+ 48-49. [11] Konopliv A. S. et al. (1998) Science, 281, 1476-1480. + [12] Jones J. H. (1995) in Rock Physics and Phase Relations: A Handbook of Physical Constants (T. J. Ahrens, ed.), pp. 73-104. [13] Borisov A. and Palme H. (1995) GCA, 59, 481-485.

+ 0.01 THE TRANSPORT OF MAGMA FROM THE MARE o mre, low•T SOURCE TO THE SURFACE. M. A. Wieczorek and R. J. + • mare, litlc fragments ■ mare, highT ◊ mre, pyroclastic Phillips, Campus Box 1169, Department of Earth and Planetary + o nnrre low-Ti Sciences, Washington University, One Brookings Drive, St. Louis � +_E_a_rth 0.001 4-----'i"----.---,---+ �-�-----.. MO 63130, USA ([email protected]). 0 4 8 12 16 20 24 28 32 MgO(wt%) Introduction: Alhough it has been known fromthe beginnings Fig. 2. of the space age that the lunarmare basalts areprimarily located on the nearside of the Moon, the definitive causeof this phenomenon inferred composition implies that Ni/Ir is enhanced in the lunar has remaineda mystery. One popularexplanation is that the hydro­ mantle by nearly the samefactor, -30 x CI chondrites, as in Earth's static pressure of the mare source controls the eruption of mare mantle. basalts. In this scenario, the depth of the mare source controls the In a sense, it is remarkable that the lunar mantle siderophile maximum height that magma can rise in the crust. If the maximum depletions arenot much greater. In the case of the Earth and Mars, the depth of the mare source was globally uniform, then mare basalts hypothesis that siderophiles were added as a veneer afterdifferentia­ would only be able to erupt at the surface below a critical elevation. tion of the core [e.g., 9] remains controversial. In the case of Earth, Following the discovery of the Moon's 2-km center-of-mass/center­ high-pressure dampening ofmetal /silicate partitioningbehavior [e.g., of-figureoffset [1], manyhave suggested that the higher elevations 10] may have played a more important role.Recent LunarProspector of the lunar farside could have prevented farside magmas from results indicate that the Moonpossesses a core amounting to 1-2 reaching the surface due to their lack of the necessary pressure in their wt% of its bulk mass [11]. The depletion factorI::,. implied by an source [e.g., 2-3]. equilibriumbetween the silicate portionof the planetand its core can Recent data obtained fromthe Clementine andLunar Prospector be calculated fromthe simple mass balance missions suggest that this scenario may be a bit simplistic. Using Clementine altimetrydata, the fulltopographic extent of the South I::,.=1/(Dc+[ l-c]) Pole Aitken Basinhas been determined[4]. Even though the lowest where c is the weightfraction of the core andD is the core/silicate elevations of the Moon were foundto occur within this basin'sfloor, (metal/silicate?) distributioncoefficient. With a low-pressure metal/ mareflows in this basinsare volumetrically insignificant when com­ silicate D of the order 5 x 1().5[12] andperhaps much higher[13], Ir pared to the nearsidebasins andOceanus Procellarum.If elevation 10 Workshop on New Views of the Moon I

was the only factor contolling the eruption of basalts, then this basin 0.05 should surely have ben completely flooded [5]. Gamma-ray data fom the Luna Prospector mission [6] also suggest that elevation is 0.0 uQ) not the only factor that contols the erption of mae baalts to the surface. Based on the surface distribution of KREEP, as well b 0.03 results fom previous studies [e.g., 7-9], it has recently been agued u [10-13] that the Pocellaum ad Imbrium region of the Moon is a C: 0.02 unique geochemical crsta province enriched in incompatble ad heat-producing elements (nbed the "Procellam KREEP terae," 0.01 or PKT). Wieczorek ad Phillips [11] have noted that more tha 60% 0.00+--'�-,;;.;:::j,!:.....,.�..l.:,-.,....-,-...- ...... ,..._;;::::::;,-,--1- of the Moon's mare basats reside within this province, ad have � � � 4 4 � 4 0 1 2 3 4 5 6 7 8 9 argued fr a genetic relatonship between the two. Spcificaly, by Topo - Geoid (km) placing a layer of KREEP basalt in ths province, their theral mod­ els predict that the lunar matle should have paially melted only Fig. 2. Histograms of lunar elevation referenced to the geoid for the beneath this province. The erption of mae baalts thus may be entire Moon, the mare inside of the PKT, and the mare outside of the primaily contolled by the distibution of heat sources in the Moon. PKT. In this study, we reexamine some of the factors that contol the espton of mae basats on the Moon. We frst show that the dist­ fows occur above an elevation of 2.6 km, ad the vat majority of bution of mae baalts is indeed limited by elevation, consistent with fows occur below -0.5 km. This is tue fr both mae basats that the hydrostatic-erption hypthesis. Though this obsesation sug­ erpted inside ad outside the confnes of the PKT. This result is gests a shallow depth of the mare basat source (<135 km), the consistent with the hypothesis that hydrostatic frces in the me petologically constaned depth of origin (based on the premise that source contol the erption of baaltc magmas, if the maimum the picritic magmas were multiply saturated with olivine ad pyrox­ depth of the mae source is globaly unifr. ene in their source) suggests a deepr origin (360-50 km). We The maximum elevation of mae basalts ca be used to determne postulate ad discuss three possible resolutions to this appaent pa­ the maimum depth of the mare source by prforming a simple frce adox: ( 1) The erption of mae basalts is not primarily contolled by balace. For mechancal equilibrium, the pressure at the bae of a hydrostatic frces, but rather by their buoyacy; (2) the petologi­ column of magma must be baaced by the Moon's reference hydro­ cally constained depth of origin is either not vaid, or has been static pressure at this depth. We have attempted to maimize the misinterreted; and (3) the petrologically constained depth of origin depth of the mae source in this calculation by taking extreme vaues is corect, but matle paal melts accumulated in shalow subcrstal of the density and bulk modulus of the mare basaltic magmas. These magma chambers befre erupting onto the surface. values were computed at the material's liquidus temperature using The Evidence in Favor of a Hydrostatic Mechanism of Mare the technique of Delco [14] and the data of [15] and [16] for the Emplacement: Figure 1 shows the elevation of the lunar mare mare basaltic reference suite of [17]. The variation in density of this referenced to the geoid. We reference the mare elevations to the geoid magma with depth was determined using the second-order Birch­ for two reasons. First, magma on the surface of the Moon would fow Muaghan equation of state, and gravity was computed as a function down the geopotential gradient (i.e., downhill). Secondly, for a given of depth. hydrostatic pressure in the mare source, the maximum height a If mare basalts only erpt below an elevation of2.6 km, then fom column of magma can extend is a function of elevation referenced to the above considerations, a column of magma in hydrostatic equilib­ the geoid. In Fig. 2 we plot histograms of elevation refrenced to the rium at this elevation would extend to a depth of 135 km below the geoid for the entire Moon, the mae within the confines of the surface. If this column of magma were to extend deeper into the luna Procellarum KREEP terae, ad the remainder of the mare outside mantle, then it would have been possible fr mare basalts to erupt at this terrane. From this plot it is fund that the eruption of mae basalts elevations higher than ae observed. It is natural to interret this does indeed appea to be limited by elevation. In paticular, no mare depth as the maimum depth of the mae source, and as a working hypothesis we assume this to be the case. We note, however, that this interpretation appeas to be inconsistent with the petrologically con­ strained depth of the mae source. Assuming that the picri tic glases were multiply saturated in olivine and pyroxene in their source, the mae baalts (which ae likely derived fom these magmb) should have all been derved fom depths between 360 ad 500 km [e.g., 18]. In addition, if the magmas parentalto the pi critic glases factionated olivine befre erpting, then these depts should be interpreted as a lower limit. In the fllowing secuons, we briefy discuss three possible sce­ narios that may rectf the dispaty between the depth of the mare source based on hydrostauc ad ptologic aguments. The Eruption of Basalts IsControlled by Buoyancy: If the Fig. 1. Elevation of the mare referenced to the lunar geoid. The spatial density of mae baaltic magma was less ta that of the suounding distribution of mare is taken from themapping of [19), and the location crust, then these melts woud hve ben ale to rise to the surface of the mare/highland contacts is estimated to be accurate to within 60 km. baed on buoyacy forces alone. Though the upper aorhositic crst LP/ Contribution No. 980 71

of the Moon is certainly less dense than most marebasaltic magmas, and Wilson L. (1992) GCA, 56, 2155. [4] Zuber M. T. et al. (1994) many have argued[e.g., 20-22] that the lunarcrust becomes progres­ Science, 266, 1839. [5] Smith D. E. et al. (1997) JGR, 102, 1591. sively mafic with increasing depth below the surface. A typical cited [6] LawrenceD. J. etal. (1998)Science, 281, 1484. [7] Metzger A. E. value for the density of the lower crust of the Moon is about 3.1 et al. (1973) Science, 179, 800. [8] WarrenP. H. andWasson J. T. g/cm3, andwe findthat all of the marebasaltic magmas areless dense (1979) Rev. Geophys., 17, 73. [9] Wasson J. T. and Warren thanthis value. P.H. (1980) Icarus, 44, 752. [10] HaskinL. A. (1998) JGR, 103, Geophysical crustal thickness modeling [23] and investigations 1679. (11] Wieczorek M. A. andPhillips R. J. (1999) JGR, submit­ of the compositionof basin ejecta[e.g., 24] suggest that the upper ted. (12] JolliffB. L. (1999) JGR, submitted. (13] Korotev R. L. anorthositic crust may have been completely excavated during the (1999) JGR, submitted. [14] DelanoJ. W. (1990) Proc. LPSC 20th, impact events that formedthe multiring basins. If this is true,then, 3. [15] LangeR. A. and CarmichaelI. S. E. (1987) GCA, 51, 2931. based on buoyancy forces alone, mare basalts could have been able [16] Kress V. C. andCarmichael I. S. E. (1991) Contrib. Mineral. to rise to the surface within the confinesof these basins. The crustal Petrol., 108, 82. [17] BVSP (1981) Basaltic Volcanism on the thickness models of [23] show that marebasalts are indeed present Terrestrial Planets. [18] Longhi J. (1992) GCA, 56, 2235. wherever the upper crustis predicted to be absent (most notably, the (19] Wilhelms D. E. (1987) U.S. Geological Survey Spec. Paper nearside basinsand SPA), consistent with this hypothesis. 1348. [20} Ryder G. and Wood J. A. (1977) Proc. LSC 8th, 655. Though this explanation does not account for the presence of [21] Spudis P.D. andDavis P. A. (1986) Proc. LPSC 17th, inJGR, marebasalts in the AustraleBasin and Oceanus Procellarum,we note 91, E84. [22] Wieczorek M.A. and Phillips R. J. (1997) JGR, 102, that our understanding ofcrustal composition andstructure is incom­ 10933. (23] Wieczorek M. A. and Phillips R. J. (1998) JGR, 103, plete, and that the crustin these regions may in fact be more mafic 1715. [24] Spudis P. D. et al. (1984) Proc. LPSC 15th, in JGR, 89, thanthe geophysicalcrustal thickness models predict. C197. The Petrologically ConstrainedDepth of OriginIs Not Valid or Has Been Misinterpreted: It hascommonly been assumed that the pressure at which the picritic glasses are multiply saturated with olivine and pyroxene represents the hydrostatic pressure of the mare A VIEW OF THE LUNAR INTERIOR THROUGH LUNAR source. Using a relationship between pressure and depth in the LASER RANGE ANALYSIS. J. G. Williams, D.H. Boggs, J.T. Moon, this assumption necessarily constrainsthe depth of origin of Ratcliff, C.F. Yoder, and J. 0. Dickey, Jet Propulsion Laboratory, the marebasalts. California Institute of Technology, Pasadena CA 91109, USA This assumptionmay be flawedin two ways. First,if largedegrees ([email protected]). of melti_ng occurred in the maresource (on the order of 40%), then olivine would have been the only phase present in the crystalline Introduction: Laser rangesbetween observatories on the Earth residuum. The depth of the maresource based on the multiple-phase andretroreflectors on the Moon startedin 1969 and continue to the saturation assumption would, in this case, be an overestimate. present. Recent range accuracies are 2 cm while earliest rangesare an Secondly, it is well known that when rocks melt at constant order of magnitude less certain. Fourretroreflectorsare ranged: three pressure, the volume of the melt is -10% greater than the corre­ located at the Apollo 11, 14, and 15 sites andone on the Lunakhod 2 sponding solid phases [e.g., 15]. If melting were to occur under rover. Accurate analysis of the range data determines a number of constant volume conditions (as may be appropriate in the lunar lunarscience parameters. The lunar interior variables includeafl uid­ mantle), the process of melting would give rise to an excess pressure core parameter. in the maresource. Using the data of [15] and (16], 17% melting of Data Analysis: The Lunar Laser Ranging effort is reviewed in the mantle at a depth of 135 km would give rise to an excess pressure [1]. Many parametersare detected through their influence on rota­ of about 12 kbar.The pressure of the mare source in this case would tion. Also detected are solid-body tides and accurate selenocentric correspond to the reference hydrostatic pressure of the Moon at a reflectorlocations. Determined through the rotation are moment-of­ depth of -400 km. inertiadifferences, gravitationalharmonics, potential Love number, Dikes Do Not Extend All the Way from the Mare Source to the and dissipation effectsdue to tides and molten core. The rotation of Surface: Even though melting may have occurredat depths greater the Moon is not at its minimum energy state; some recently active than 500km in the Moon, there is no guaranteethat open fractures process has caused free librations [2]. The moment differencescon­ will exist from this depth all the way to the surface. For instance, tributed to the recent improvement of the Moon's moment of inertia magma transport in the mantle may have occurred primarily by from the Lunar Prospector gravity field [3]. The Love numbers porous flow. Only when this rising magma reached a rheological provide bulk elastic properties.Future possibilities formeasurement barrier (such as the base of the crust or lithosphere, or a layer of include oblateness of the core-mantle boundary and core moment. depleted mantle) would it begin to accumulate and intrudethe lower Dissipation: A study of dissipation signaturesin the rotation portion of the crust as dikes. As magma continued to accumulate at determinestidal Q vs. frequencyand concludes thatthe Moon has a these barriers, dikes would intrudeto higher levels in the crust, and molten core [4]. At 1 month the tidalQ is 37 and at 1 yr it is 60.The would eventually eruptonto the surface. core radiusis S352 km forFe and S374 km forthe Fe-FeS eutectic. Conclusions: Each of the above scenarios appearsto reconcile The core detection exceeds 3x its uncertainty. The spin of the core is the disparity between the hydrostatic and petrologic depth of the not alignedwith the spin of the mantle andtorque arises from the maresource. Though all threemay play a role, we arecontinuing to velocity difference at the boundary.Yoder's turbulent boundary investigate the relative importanceof each. layer theory (5-6] is used to compute the radii. References: [1] Kaula W. M. et al. (1972) Proc. LSC 3rd, TidalHetg: The present heat generation fromtides and core 2189. [2] Solomon S. C. (1975)Proc. LSC 6th, 1021. [3]Head J. W. interaction is minor comparedto radiogenic heating.The heating for 72 Workshop on New Views of theMoon II

ancienttimes is more interesting. Peale and Cassen[7) investigated He contains a high abundance of the rare isotopic 3He, which has lunar tidal heating while the lunar orbit expanded d�e to tides on been identifiedas a potentiallyvaluable fuel for nuclear-fusion space Earth. Their calculations predate the measurement of Q. andshould power systems. be multiplied by 3.45 to match the lunar-laser-determined Love In order to determine the economic potentialof these lunarvolatiles, number and monthly Q. Tidal-heating computations depend on how we need informationto assess the in situ quantities of these gases and fast the lunar orbit evolved and whether the tidal dissipation is identifythe most abundant sites. In order to acquire such informa­ localized. Neither is known, but plausible assumptions lead to early tion, a large numberof soil samples must be acquired andanalyzed central region temperature increases of several hundred degrees. because it is not known if these volatile gases in the soil vary widely Most of the energy is deposited early in the Moon's history. over the distanceof a fewmeters or several kilometers. In addition, Core Heating: The turbulent boundary layer theory allows a all of the lunar soil samples were analyzed on Earth, after being predictionof energy dissipated at the core-mantle boundary during contaminated by terrestrialair andwater. For these reasons, there­ orbit evolution. Under the assumption that the propertiesof the early fore, a mobile, robotic vehicle has been proposed that would be core arethe sameas at present, the energy dissipated by core-mantle landedon a lunar mariaand assay the volatiles evolved by heating the interaction is about thesame as for tidaldissipation, but it is depos­ indigenous lunarregolith. ited in a smaller volume. This source of energy is capable of promot­ A lunarrover platformwith the sample equipment attached has ing convection in an earlyfluid core and driving a dynamo.This is been designed. This science platformwas conceptionallydesigned to a transient phase with a duration depending on the rate of orbit fiton a small MarsokhodRover (75 kg) with a 100-km range. evolution. Plausible assumptions lead to a duration of a fewhundred Sample Collection and Analysis: The proposedsampling pro­ · million years.Thus, the remnantmagnetization of many lunarrocks tocol would be to collect two samples, nearlyadjacent. If theresults is compatiblewith a brief global magnetic fieldpowered by dynami­ , agreed within the experimental deviation, the rover would proceed cal energy dissipation. 0.5 km along the planned route and select two new samples.The Future Data: Analysisof the lunar laser rangesis providing progress of therover andthe results of the analyseswould be continu­ information on lunar geophysics. Future datawill improve accura­ ously monitored fromEarth so that the sampling protocol could be cies of present solution parameters,and several more interior effects revised as needed. should be detectable. The scientific equipment would accomplish the assay of the References: [1) Dickey et al. (1994) Science, 265, 482-490. regolith samples in the following sequence: (1) retrievea sample of [2) Newhall and Williams J. G. (1997) Celestial Mechanics and regolith fromthe lunar surface by use of a scoop mounted on the Dynamical Astronomy, 66, 21-30. [3] Konopliv A. S. et al. (1998) platform; (2) reduce the sample to -1 g of particles <200 µrn; (3) Science, 281, 1476-1480. [4] Williams et al. (1999) LPS XXX, weigh the sample; (4) characterize the mineral content (TiO2); (5) ° Abstract #1984. [5) Yoder C. F. (1981) Phil. Trans. R. Soc. London heat the sample to 1200 C in a vacuum furnace; (6) collect the A, 303, 327-338. [6] Yoder C. F. (1995) Icarus, 117, 250-286. volatiles; (7) characterizethe volatile products; and (8) transmit the [7] Peale S. J. and Cassen P. (1978) Icarus, 36, 245-269. data. Analytical Equipment: The components of the scientificpack­ age were conceptually designed and arebriefly described: A STUDY OF AN UNMANNEDLUNAR MISSION FOR THE 1. Theheater unit. A 1-g sample of the surfaceregolith would be ASSAY OF VOLATILE GASES FROM THE SOIL. L. J. placed in a ferritic steel crucible 0.8 cm OD x 1.57 cm high. This Wittenberg 1, I. N. Sviatoslavsky2, G. L. Kulcinski, andE. A. Mogahed, container would be placed in a coiled electrical heater inside an 1,2Fusion Technology Institute, University of Wisconsin-Madison, evacuated 1-L container. Heat transfer calculations indicate that the 1500 Engineering Drive, Madison WI 53706, USA sample would attain 1200°C in 14 min with a 25-W heater. For a ([email protected]). high-Ti maria regolith sufficientgases arereleased to create a pres­ sure of 70 Pa at 30°C, which is a sufficient sample for the mass Introduction: The success of a manned lunar outpost may spectrometer. require that indigenous resources be utilized in order to reduce the 2. Determinatioll of metallic elements. Before the sample is requirements for the periodic resupply fromEarth for the human heated, a laser beam delivers 0.45-2.0 Joules per pulse at a wave­ inhabitants. Some indigenous lunar resources do exist. For instance, length oft µm to the surface of the sample. The absorption of the laser studies of the lunar regoliths, acquired by the Apollo and Luna energy vaporizes some of the minerals in the soils. The vaporized missions fromseveral maria, indicate that upon heating in a vacuum, ions are quantitativelydetermined by the mass spectrometer. these soils evolve the volatile gases: helium (He), hydrogen (H2), 3. Mass spectrometer. This instrument must be utilized to char­ carbondioxide (CO2), carbon monoxide (CO), nitrogen (N2), and acterize themineral content of the soil andthe volatile gases, essen­ sulfurdioxide (SO2). The He, H, C, andN were originally implanted tially from 1 to 72 AMU range. A Fourier Transform Mass by.solar wind [1] Spectrometer(FTMS) may be particularly usefulfor this analysis, These gases would be valuable to supply life-support systems. but requires testing.

For instance, the H2 could be used as arocket fuel, or alternatively, 4. Power supply. The initial powersubsystem assumedthe avail­ reacted with the mineralilmenite (FeTiO3), indigenous to the lunar ability of a general purpose heat source, or a RadioisotopeThermo­ soil, to yield water (H20). In an enclosed structureirradiated by solar electric Generator. If the launch of an RTG is forbidden for safety energy, theH 2O, N2, andCO 2 could be utilizedto grow edible plants reasons, alternative power supplies would be considered such as forlunar inhabitants. Alternatively, the H 2O could be electrolyzed, solar-electricor beamed microwave sources. using photovoltaic cells, yielding breathable 0. Theinert gas, He, References: [I] Heiken G. H. et al. (1991) Lunar Sourcebook: would be usefulfor filling inflatable structures. In addition,the lunar A User's Guide to the Moon, CambridgeUniv.