Tsiolkovsky crater' A window into crustal processes

Transcription

1 JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 104, NO. E9, PAGES 21,935-21,949, SEPTEMBER 25, 1999 Tsiolkovsky crater' A window into crustal processes on the lunar Carl M. Pieters farside Department of Geological Sciences, Brown University, Providence Rhode Island Stefanie Tompkins Science Applications International Corporation, Chantilly, Virginia Abstract. Mineralogy and distribution of rock types in the Tsiolkovsky region of the lunar farside are evaluated in terms of crustal stratigraphy and evolution. Calibrated multispectral images from five orbits of Clementine data provide compositional information at a scale that allows diverse geologic features to be analyzed and compared. The entire region is seen to be highly anorthositic. with outcrops of relatively pure anorthosite quite common. The crater itself has excavated blocks of anorthosite and noritic anorthosite from the near-surface highland crust, often in a mixed jumble. The central peaks of Tsiolkovsky, which have exhumed much deeper material, however, are composed of relatively pure anorthosite with olivine-rich zones concentrated near the ridge crests. Boundaries between the anorthosite and olivine-rich zones are sharp, but spatial relations suggesthe coherent olivine-rich zone is relatively thin (<1 km). The composition of the mare material filling Tsiolkovsky is shown to be low-titanium basalt. Smaller exposures of mare basalt occur about one crater radius from the rim. The crustal structure to -25 km in this region prior to the Tsiolkovsky impact appears to be (top to bottom) a few kilometers of anorthositic breccias comprising the megaregolith, a zone of crystalline anorthosite with pockets of noritic anorthosite, a zone of anorthosite devoid of iron-bearing minerals, and a zone of relatively pure anorthosite containing olivine-rich material, possibly emplaced as intrusions. 1. Introduction that intrusive equivalents to basaltic material may thus occur commonly on the farside [Head and Wilson, 1992, 1994]. Although the nearside of the Moon is familiar to both scien- The composition and diversity of the farside crust are of partists and nonscientists alike, the lunar farside has remained mysticular interest not only with respecto thermal evolution, but terious and enigmatic. It is only through limited spacecraft observations that the character of the farside of the Moon has been also because they are key to determining the bulk composition of the lunar crust. An accuratestimate of bulk composition is necstudied. In contrast to the nearside, the farside exhibits signifiessary to test several models for the origin of the Moon and its recantly less evidence of mare volcanism, has a thicker crust, and lation to Earth's mantle [e.g., Hartmann et al., 1986]. Initial has a geologic history that is strongly linked to the formation of the enormous South Pole-Aitken Basin that dominates the farside analyses of Clementinc data suggest farside highland surfaces, especially the northern highlands, to be richer in anorthosite than southern hemisphere. There are several long-standing issues of lunar science that can be addressed only with a better understand- the nearside [Lucey et al., 1995]. Tompkins and Pieters [1999] ing of the lunar farside. Recent data such as the global multispecsurveyed the central peaks of large impact craters across the Moon, which are believed to exhume crustal materials from betral imagery returned from the Clementinc mission [Nozette et neath much of the megaregolith, and found anorthosite abunal., 1994] and low-resolution global chemistry data from the Ludance at depth to be comparable in a general sense to that estinar Prospector mission [Binder, 1998] finally expose the lunar farside to more detailed compositional studies. mated for the surface by Lucey et al. However, the deep-seated One of the most obvious features of the lunar farside is its lack crustal materials excavated by such craters across the Moon also of extensive mare volcanism. An assessment of the type and exhibit notable and distinct compositional diversity on the scale abundance of volcanic and plutonic materials has been an ongo- of a few kilometers. Much more detailed analyses are required to ing quest [Head, 1976; Soloman and Head, 1979, 1980; Greeley document and understand the important links between geologic et al., 1993] in formulating a thermal evolution model for the history and stratigraphy on a local scale in order to address Moon. Most models account for the dearth of basalts on the lu- global-scale issues of geologic evolution and the bulk composition of the crust. nar farside as the result of a thicker farside crust. An importap t aspect to consider is a density difference between the basaltic The crater Tsiolkovsky is an excellent reference point for exmelt and the thick overlying crustal rock, which suggests to some amining the details of crustal composition on the lunar farside. Tsiolkovsky is large enough to have exhumed materials from significant depths within the crust (-30 km), it formed in a region Copyright 1999 by the American Geophysical Union. of typical farside crustal thickness, and it presents one of the few Paper number 1998JE farside examples of mare volcanism. In the discussion presented /99/1998JE here we reexamine the compositional character of the farside 21,935

2 ,936 PIETERS AND TOMPKINS' CRUSTAL PROCESSES ON THE LUNAR FARSIDE '. "';?..,,.4".. ;:..;.. <.'.a'; ':..*'". - ',?/4'.".*.:,"X*'...V,,-...,.,,..,..::: ;;,.... * A":*:'"'½' '" -..,-..,.. ' : ;x....:,..:'?.,. '::?re*: '"....,('..;,.,.?..' ; s.½,.:... "T;.' Figure la. Lunar Orbiter image (LO III 121-M) of the southwest lunar farside. Tsiolkovsky is the prominent 180- km mare-filled crater. crust at Tsiolkovsky with Clementine multispectral images. The geologic setting of Tsiolkovsky and the principal data processing approach for this region of Clementine data are summarized first. The basis for mineralogical interpretations with the five-color Clementine data is then discussed. Compositional results are presented for near-surface highlands, mare materials, highland crust below the megaregolith, and deep-seated highland crust. This mineralogical information is used to recreate the compositional stratigraphy of the highland crust in the Tsiolkovsky region. 2. Geologic Context Tsiolkovsky is a prominent, 180-km Late Imbrian age crater on the lunar farside (20øS, 129øE), filled with dark mare deposits. Because maria are not extensive on this side of the Moon, Tsiolkovsky is a notable feature from orbit. The crater and its surroundings are shown in Figure 1 as imaged under low-sun lighting conditions by Lunar Orbiter 3 and much later under high-sun conditions by Clementine. During Apollo 15, several exquisite photographic images of the crater were obtained from orbit. A mosaic of Apollo 15 high-resolution images centered on Tsiolkovsky central peaks is shown in Figure 2. The general characteristics of Tsiolkovsky have been described by Guest and Murray [1969] and mapped by Wilhelms and El- Baz [1977]. Wilhelms [1987] mapped the geology of the Tsiolkovsky region within a global context. The crater is located in what might be an ancient 700-km pre-nectarian basin, Tsiolk- ovsky-stark [Wilhelms, 1987]. The continuous ejecta deposits of Tsiolkovsky appear to be asymmetric, with more extensive deposits to the south-southeast [Guest and Murray, 1969]. The central peak is located slightly north of the center and is surrounded by low-albedo mare. Although the lunar crust is generally thicker on the farside, with the extra thickness believed to restrict basalt emplacement, the size of the crater and a possible preexisting basin environment may help account for this rare farside mare fill. The elevation of Tsiolkovsky mare [Zuber et al., 1994] is higher than most, but not all, nearside maria. Crater counts of the mare indicate that the Tsiolkovsky mare deposits are of an age comparable to those of the eastern nearside maria [Wilbur, 1978; Wilhelms, 1987]. The darkness of the mare deposits initially suggested to some that they might be Ti-rich basalts comparable to those of Apollo 11 and 17 of similar age [Wilbur, 1978; Wilhelms, 1987].

3 PIETERS AND TOMPKINS: CRUSTAL PROCESSES ON THE LUNAR FARSIDE 21,937.: %?.... Figure lb. Clementine 750-nm albedo mosaic of the Tsiolkovsky region. Figure 2. Mosaic of Apollo 15 photographic images of Tsiolkovsky central peaks. The original images (AS , 1308, and 1309) were digitized and mosaicked.

4 21,938 PIETERS AND TOMPKINS: CRUSTAL PROCESSES ON THE LUNAR FARSIDE 3. Clementine Data Calibration and Processing Processing of calibrated data includes first coregistering each filter to a standard reference frame (in this case, the image ac- The Clementine spacecraft spent 2 months in an elliptical orbit quired with the 750-nm filter). Experience showed that registraaround the Moon and produced a global map of the surface with tion to the nearest 0.1 pixel was required for color ratio images two cameras [Nozette et al., 1994]. Five-color data from the and science analyses, creating a five-color image cube, or frame UVVIS (ultraviole to visible) camera are discussed here. Data set. With the return of the initial images from the Moon, it was from the near-infrared camera, with six additional filters, require found that data for the first filter (415 nm) were substantially deextensive calibration [Lucey et al., 1998a]. The band passes of graded in spatial resolution, presumably an effect of the camera the UVVIS narrowband filters are shown superimposed on reflec- optics and detector. Color ratio image composites involving the tance spectra of key lunar minerals in Figure 3. The wavelengths 415-nm image require the second image (usually 750 nm) to be of the filters were chosen to be sensitive to the iron-bearing min- degraded to match the resolution of the 415-nm image prior to erals that distinguish lunar rock types [e.g., McEwen and Robin- dividing one image by the other. Multiple frame sets are then son, 1997]. mosaicked along the orbit. Two of the orbits crossing Tsiolk- The UVVIS sensor is a framing camera with an active area of ovsky occurred in the first month of mapping, and the remaining 384 x 288 pixels. For the Tsiolkovsky region, the pixel size var- three occurred in the second month. Data from the two months ies from the first month to the second (approximately are at a different spatial resolution but include some east-west m/pixel). Images of the surface are taken sequentially through overlap with independent data. In order to maintain the maxithe five filters. The data are acquired with 8-bit precision, but to mum amount of original information we have not mosaicked the expand the effective data number (DN) range available, two dif- data east to west here. Fully mosaicked five-color data will be ferent exposures were taken for each filter. Initial data calibra- available from the Clementine team through USGS in tion and processing consist of several steps performed with float- Mosaics for the five-orbits of images centered on Tsiolkovsky ing point precision. Details of data processing are discussed by are shown in Figure 4 and Plate 1. The 750-nm reflectance mo- McEwen et al. 1998] and C. M. Pieters et al. ( plane- saic of Figure 4 has been photometrically corrected to i = 30 ø and tary.brown.edu/clementine/calibration.html). The basic calibra- e = 0 ø and can be used to provide an albedo overview of surface tion includes several offset corrections, one of which is scenedependent. The Tsiolkovsky data were acquired at gain 1, so no features. Fully mosaicked and geometrically corrected 750 nm albedo data products sampled to 100-m resolution are available gain corrections were required. Preflight flat field information through the Planetary Data System for the entire Clementine covwas in error, requiring corrections to be made using data derived erage [USGS, 1997]. The color composite mosaic of Plate 1 is from in-flight flat fields produced by the U.S. Geological Survey particularly sensitive to distinguishing between highland and (USGS). The calibrated images were photometrically corrected mare lithologies. The red-green-blue color assignment is a slight to a standard phase angle of 30ø; initially, no corrections were modification of the familiar color composite used for Galileo lumade for small wavelength-dependent variations in photometry nar data [Belton et al., 1992, 1994]. The red channel (750/415 (see below). Laboratory bidirectional reflectance measurements nm) represents the overall character of the continuum from the of Apollo 16 returned lunar soil were used to radiometrically blue through the visible. Mature highland soils have a relatively calibrate the data to standard spectral reflectance factors, defined red continuum and appeared. The green channel (750/950 nm) [Hapke, 1993] as the reflectance of the surface relative to a Lam- is sensitive to the strength of the ferrous absorption of mafic minbertian surface under the same bidirectional illumination (in this erals. Mare soils and craters containing abundant iron-rich mincase, i = 30% e = 0ø). erals have a distinct yellow or green tone, while highland lithologies with a detectable mafic component appear turquoise. Last, the blue channel in this color composite carries albedo information (750 nm), with contrast set to highlight the brightest materials. Much of the bright, fresh crater material in the highlands ap- Clementine Bandpasses o Plag<250 pears deep blue in the color composite (high albedo, no mafic ab < Lunar Minerals I... Ciera dvvls Cpx <45 I sorptions, and a relatively flat continuum). The phase angle for ß '... '... ' ' ) Opx <500 I- l all Tsiolkovsky data is <35 ø, so shadowing in craters is mini- 0.7 o mized (although noticeable in some locations). 0.6 Even though a general photometricorrection was performed 0.,5 with these data to adjustheir reflectance to the standard geometry (i=30 ø, e=0ø), there is also a well-known gradual change of color with phase angle [Lane and Irvine, 1973]. The surface of 0.4. the Moon becomes redder with increasing phase angle. For 0.:3 Clementine UVVIS wavelengths, the relative color between I ' ' and 750 nm is most strongly affected with color gradients about %/deg for a 415/750 nm ratio. The variation in phase angle of the Tsiolkovsky data ranges from 21 ø to 35 ø north to south for orbits 252, 253, and 254 and from 13 ø to 31 ø north to south 0 for orbits 120 and 121. This results in about a 6-7% north to Wavelength (nm) Figure 3. Bidirectional reflectance spectra of lunar mineral separates (i = 30ø; e = 0ø). The shapes of the band passes for the five UVVIS filters are superimposed. south variation in 415/750 nm color across the Tsiolkovsky image due simply to viewing geometry. This regular variation can be seen as a variable gradient of the red channel in the color composite of Plate 1, with the two high-resolution orbits (120 and 121) being most strongly affected.

5 ..... ß -.,. PIETERS AND TOMPKINS' CRUSTAL PROCESSES ON THE LUNAR FARSIDE 21,939 ::.: ;::.:... ß ::::;i:.::i' :. -<:-:::.'? Figure 4. Mosaic of five orbits of photometrically corrected reflectance images at 750 nm centered on Tsiolkovsky. The standard geometry is i=30 ø, e--o ø. The orbits used for these data are (left to right) 254, 121, 253, 120, and 252. Fields of view for these orbits are 60 km (254, 253, and 252) and 40 km (121 and 122). Plate I. Color composite mosaics of the five orbits of Clementine UVVIS data centered on Tsiolkovsky shown in Figure 4. The color assignment is red = 750/415 nm, green = 750/950 nm, and blue = 750 nm. Mature highland soils appear red, anorthosites appear deep blue, and both noritic anorthosites and olivine/troctolites appear turquoise; basaltic soils appear yellow, and fresh basalt appears bright yellow.

6 21,940 PIETERS AND TOMPKINS' CRUSTAL PROCESSES ON THE LUNAR FARSIDE 0.45 [ [ ] [ [ [ ] at [- Rock and Soil Types at Tsiolkovsky Anorthos,te t [- _ ', (:>-.--o -]- A -='",,,,,Ol v,ne ] t J, e-' <)" l'roctol,t j_ [- '"'.' T ' [ oc, H,ghland oe, -- So,, -:'.7. _-..O'-.',... *.....gap r' ' r < o.o..ø.-ø-'ø"; o '. ".'.-' u Basalt... thered t -I T 0.10 '''...- -'"' '"' '" Basalt Soft _.-.r.-: m-... ]_ st I'... i' I I I I I! e---shigh Cr 1 ß SHigh Cr 2 I I I I ---e -SHigh Cr 3 Highland Craters C -SHigh Cr4 -SHigh Cr 5 N & S of Tsiolkovsky ----e, - SHigh Cr Npkl.- Npk2 = Nhigh cr -- --NHigh cr (ffc),--e--n highlands b // '"- - ß : "..O... Highland so,i Wavelength (Fm) Wavelength (Fm) 1' Tsiolkovsky Mature Soils N, S H ghlands -,.--..,,.....,,,-; Ap16... e '= '-""'"- -e-- - ef,..., / Mare _ >;... +MS, %...½ :... Mare sod - I I I I I I I Wavelength (pm) Mare Craters in Tsiolkovsky and mare ponds ' Wpondl - n- o.15- H ghland sod I I I I I I I Wavelength (Fm) Figure 5. Five-color spectra of rock and soil types in the Tsiolkovsky region. Each spectrum represents the average of a 4x4 pixel area. Error bars are one standardeviation. Shown are (a) rock and soil types, (b) mature soils, (c) highland craters, and (d) mare craters. For reference, the same highland soil (N highlands b) is shown in all plots. 4. Spectroscopic Identification of Soil and Rock Types at Tsiolkovsky The color composite mosaics of Plate 1 allow the general diversity of rock and soil types in the Tsiolkovsky region to be discemed. In this image, undisturbed mare soils (yellow) are clearly distinguished from general highland soils (red), which appear approximately the same everywhere. In contrast, most of the highland craters (which have exposed fresher, more lithic material) have distinct, but similar color and albedo (blue). Several areas along the crater walls and central peaks of Tsiolkovsky, however, exhibit diverse properties, indicating significant heterogeneity. Many areas appear turquoise in the color composites, indicating the presence of iron-bearing minerals. The color mosaics do not provide a unique characterization of these areas, but as discussed below, the full five-color data do allow important distinctions be- tween iron-bearing lithologies. A variety of five-color spectra for each of these materials is presented and discussed in subsequent sections. Representative five-color spectra for the principal rock and soil types are shown in Figure 5a, and their data values are presented in Table 1. Distinguishing features and compositional interpretations are discussed here as background. The most diagnostic absorption features of lunar minerals occur as the result of electronic transitions of ferrous iron that resides within well-defined crystallographic sites. Such crystal field absorptions, lack thereof, distinguish the primary minerals of lunar rock types and representativexamples of such minerals can be seen in Figure 3. For the iron-bearing minerals such as pyroxene and olivine, ferrous iron occurs in octahedral sites of the mineral structure. For plagioclase, the site that can accommodate small amounts of iron is typically more complex (eightfold to twelvefold coordination). The shape of an absorption band Table 1. Clementine Values for Typical Tsiolkovsky Five-Color Spectra in Figure 5a Wave An-1 s.d. An-2 s.d. Olv s.d. N-An s.d. H Soil s.d. Ap16 Basalt s.d. B Soil s.d Abbreviations are wave, nominal wavelength; s.d., standar deviation; An-1, anorthosite-l' An-2, Anorthosite-2; Olv, olivine; N-An, noritic anorthosite; H soil, highland soil; Ap16, Apollo 16 soil (62231); basalt, unweathered basalt; and B soil, basaltic soil.

7 PIETERS AND TOMPKINS: CRUSTAL PROCESSES ON THE LUNAR FARSIDE 21,941 and the energy of the transition in the near infrared are determined by the specific characteristics (dimensions, symmetry) of the mineral site [Burns, 1993]. Although the most diagnostic properties that occur near 1 and 2 lam (Figure 3) cannot be characterized with just five filters, the Clementine filters are shown to be well placed for distinguishing the bulk properties of many of the principalithologies of interest here Basaltic Materials The Tsiolkovsky maria exhibit classic properties of basaltic materials. They are dark, and their spectra contain a prominent mafic (Fe +2) absorption near 1 gm, most readily seen in unweath- The relatively red highland soils of this area are generally uniform. They are very similar in bulk properties to well-known feldspathic soils of Apollo 16, excepthat the Tsiolkovsky highland soils are slightly brighter (Figure 5a). Freshly exposed highland materials, or highland features with topography that does not allow a mature soil to be developed, are all significantly brighter than surrounding feldspathic soil. All highland areas with such immature soils have a somewhat flatter (bluer) continuum (low 750/415 nm) and appear blue or blue-green in Plate 1 because of their higher albedo. From the five-color spectral properties of these immature areas, two different highland lithologies can be discerned (anorthosite and noritic anorthosite, discussed below), although there also may be gradations between Anorthosite Plagioclase, the dominant mineral of the lunar highlands, is a bright aluminum-rich silicate which normally occurs in the form of anorthite (high Ca, low Na) on the Moon. Plagioclase can accommodate minor amounts of ferrous iron in its structure, with the iron typically substituting for Ca in a site with 8-12 surrounding ions. Crystalline plagioclase with % FeO exhibits a broad but distinctive absorption near 1.25 lam (see Figure 3). Moderately shocked plagioclase, however, becomes disordered and loses this band, becoming featureless throughouthe nearinfrared [Adams et al., 1979]. At higher shock pressures, plagioclase becomes diaplectic glass [Ostertag, 1983] with distinct but weak glass absorption features [Pieters, 1996]. Using high spectral resolution near-infrared telescopic spectra, shocked lunar an- orthosites have previously been identified by spectra that are bright with no evidence of absorptions due to iron-bearing minerals [e.g., Pieters, 1986;Hawke et al., 1991, 1995; Peterson et al., 1996]. However, the higher spatial resolution data from Clementine allowed McEwen et al., [1994] to clearly distinguish a 1.25-gm absorption in the bright central peaks of Aristarchus, thus identifying the presence of crystalline unshocked anorthosite. In general, for the wavelengths covered by the Clementine UVVIS camera, anorthosite in the crystalline state should be bright, should have no significant mafic mineral absorption near 1 gm, but should exhibit a slight turndown longward of 0.9 lam as an indication of the beginnings of the 1.25-gm feature (see Figure 3). Spectra of craters with these characteristic features measured with the Clementine filters are common in the ered basalt exposed at fresh mare craters (Figure 5a). The character of this absorption in Clementine data is fully consistent with Tsiolkovsky region, and examples are shown in Figure 5a. Such the presence of abundant high-ca pyroxene, the dominant mafic mineral of basalts, which exhibits an absorption centered between areas appear bright blue in the color composite of Plate and 1.00 lam (Figure 3). In the five-color Clementine spectra, the absorption for unweathered basalt appears to flatten be Noritic Anorthosite tween 0.95 and 1.00 gm, with no suggestion of a long- In contrast, the presence of small amounts of iron-bearing wavelength turn-up (which would be characteristic of pyroxenes minerals in highland lithologies can be readily recognized by a with a lower Ca content). The basalt soils within Tsiolkovsky have reflectance values between 11.5 and 13% at 750 nm (or an albedo between 8.5 and 10.5% if linearly extrapolated to 550 nm). Tsiolkovsky soils are dark, but not as dark as some of the Ti-rich soils on the nearside [Pohn and Wildey, 1970]. The matic-rich mare soils appear yellow in Plate 1 (low albedo, strong 1-1am absorption, relatively red continuum). The fresh craters in the mare are brighter than the mare soils, but are still darker than most highland soils and many of the highland craters, as would be expected from their more mafic character. The overall albedo of freshly exposed basaltic material at 750 nm approaches that of well-developed highland soils (Figure 5). notable decrease in reflectance near 1 gm, indicative of the diagnostic absorption of ferrous iron in octahedral sites of pyroxene or olivine (see Figure 3). Noritic anorthosite is the most common highland rock type seen on the surface of the lunar nearside [Pieters, 1986]. This rock type is bright and exhibits a weak but noticeable absorptionear 1 lam (usually_<0.95 lam). In the lunar context, the brightness attests to abundant plagioclase feldspar with few opaque phases, and the strength and wavelength of the absorption indicate the presence of minor amounts of low-ca pyroxene (see Figure 3). The five-color Clementine spectra of such immature noritic areas are bright and exhibit a weak mafic mineral absorption with a curvature that suggests a shorterwavelength center than that observed for the basalts. The actual 4.2. Highland Soils center, however, cannot be estimated accurately without additional information about the continuum (which requires meas- urement near 1.5 gm). These mafic-bearing highland areas appear turquoise in Plate Olivine-Rich Materials When iron-bearing minerals are relatively abundant and their absorption bands are strong, noritic (low-ca pyroxene) and gabbroic (high-ca pyroxene) lithologies can be readily distinguished based on the wavelength of their absorption, with the low-ca species exhibiting absorptions at a shorter wavelength. Distinguishing between pyroxene and olivine in a mixture is more difficult and normally requires high spectral resolution near-infrared spectra in order to evaluate the symmetry and shape of both sides of the absorptio near 1 gm [cf. Gaffey et al., 1993; Pieters et al., 1996]. When prominent absorptions are present, the five-color spectra of Clementine can nevertheless distinguish between lithologies that are dominated principally by pyroxene and those dominated principally by olivine. As mentioned above with respect to spectra of basaltic terrains, the effect of pyroxene is to create a strong 1-1am absorption feature, with perhaps an upward turn near 1 gm. In the Clementine five-color spectra, a lithology for which olivine is the dominant mafic mineral exhibits a deep absorption continuing into the near infrared with no indication of a minimum (concave spectrum) since olivine has a much broader absorption centered at a longer wavelength than pyroxene (see Figure 3). Olivine-rich lithologies have been identified and occur principally in the central peaks of Tsiolkovsky, an example of which is shown in Figure 5a. In the simple color composite of

8 21,942 PIETERS AND TOMPKINS: CRUSTAL PROCESSES ON THE LUNAR FARSIDE a b C,-.. d Plate 2. Color composites of the central peaks and southern rim of Tsiolkovsky. Width of each frame is 60 km. Shown are (a) color composite of the central peaks using the same assignment as shown in Plate 1 (red- 750/415 nm, green = 750/950 nm, blue = 750 nm), (b) color composite of Tsiolkovsky south wall with the same color assignment as in Plate 2a and Plate 1, (c) enhanced color ratio composite of the central peaks combining information in four bands to highlight distinctions between mafic-rich areas (750/900 nm = red, 750/950 nm = blue, 750/1000 nm = green), and (d) same as Plate 2c except red = 750 nm (albedo).

9 PIETERS AND TOMPKINS: CRUSTAL PROCESSES ON THE LUNAR FARSIDE 21,943 Plate 1, however, olivine-rich lithologies cannot be easily distin- In Plate 1 they appear blue trending to turquoise with a weak guished from other mafic-bearing rock types. component of the green (750/950 rim) channel. For highland lithologies expressed at immature surfaces, the The anorthositic character of near-surface materials in the Tsimineralogical assignments described above are directly compara- olkovsky region of the farside is pervasive. Even material exc qble to those used in the survey of central peaks by Tompkins and vated as central peaks of the---40-km crater just north of Tsiolk- Pieters [1999]. In the central peak study a method was deve l- ovsky is anorthosite (N pk 1 (east peak) in Figure 5c) or noritic oped to quantify two important parameters describing the ferrous anorthosite (N pk 2 (west peak)). Other local variations of a maband (strength and shape, or curvature). These parameters al- tic-bearing lithology within the dominantly anorthositic region lowed identification of specific highland rock types that occur can be detected at numerous small craters (see Plate 1), but no across the central peaks, thus allowing global comparisons of large-scale (3100 km) mafic-bearing unit is detected in this highmineralogyø land region. While small-scale (<10 km) mafic-bearing areas clearly occur, the observationsuggest that the highland crust in the Tsiolkovsky region is predominantly anorthosite. 5. Character of Near-Surface Highland Materials As noted in the discussion of the color composite mosaic, the spectral properties of most highland soils in the Tsiolkovsky re- 6. Character and Distribution of Mare Materials gion are all quite similar. Representative spectra of several undisturbed, or mature, highland soils are shown in Figure 5b. There is nothing notable about these spectra, except they are slightly brighter than Apollo 16 soils. This is an indication that The mare filling Tsiolkovsky is clearly basaltic in character. Small craters in the maria excavate materials rich in iron-bearing minerals (bright yellows in Plate 1). Five-color spectra of such mare craters (shown in Figure 5d) exhibit characteristics indicathe Tsiolkovsky highland soils have a slightly lower abundance tive of abundant Ca-rich pyroxenes: a strong ferrous absorption of iron. The anticorrelation of soil brightness with iron content centered near m. The soils developed on the mare (or positive correlation between AI and albedo) has been recognized since Apollo [Adler et al., 1972, 1973], as has the additional anticorrelation between albedo and maturity, or cumulative exposure, of lunar soils [Adams and McCorc 1971a, b, 1973]. Analyses of lunar optical properties using multispectral image data have derived useful spectral parameters with which to ac- (Figure 5b) are slightly brighter and redder than soils at the standard area in Mare Serenitatis (MS2 in Figure 5b) [cf. Pieters, 1978], indicating that the Tsiolkovsky basalts are low-titanium in composition, <2% TiO2 [e.g., Pieters, 1978; Pieters et al., 1993]. Roughly contemporaneous with basalts of the nearside, the Tsiolkovsky basalts do not appear to be distinct in composition; commodate these maturity effects, allowing the iron abundance they are quite comparable to other 1ow-Ti basalts. Although mulof highland soils to be estimated using the depth of the 1-txm band [Fischer and Pieters, 1994] and/or the albedo [Lucey et al., 1995, 1998b; Fischer and Pieters, 1995, 1996; Blewett et al., 1997]. For the Tsiolkovsky region, these various approaches are consistent with the inference that the Tsiolkovsky highland soils are comparable to typical farside soils: relatively feldspathic and lower in FeO than Apollo 16. In addition, it should be noted that tiple episodes of basalt emplacement have been suggested [Wilbur, 1978], no characteristics are seen in the Clementinc data that might suggest significantly different compositions of basalt within the crater (e.g., no significant differences in TiO2 (from the 415/750 nm ratio) or iron-bearing mineral abundance (from strength of the ferrous band estimated by the 750/950 nm ratio)). To the south of Tsiolkovsky, the crater Waterman contains the Lucey et al. [1995, 1998b] method suggests that the northern two small low-albedo areas identified as mare ponds [Wilhelms farside highlands are even lower in FeO than these Tsiolkovsky and El-Baz, 1977]. Five-color spectra from the western Watersoils. Most of the immature highland material excavated by craters of different sizes in the Tsiolkovsky region is notably anorthositic. Representative five-color spectra of small craters and mountains or massifs in the region not directly part of Tsiolkovsky crater are shown in Figure 5c. These include areas both within and beyond the mapped extent of Tsiolkovsky's ejecta blanket [Guest and Murray, 1969]. The majority of such areas man pond are shown in Figures 5b and 5d (labeled Wpond). As can be seen in Figure 5b, the mare soils of these small ponds have a stronger ferrous absorptio near 1 txm than surrounding highland soil. Although these dark soils are somewhat brighter than the mare soils within Tsiolkovsky, they are clearly also basaltic in character based on the shape of the prominent ferrous absorption for small craters in the pond (Figure 5d). Detailed mixing analyses [e.g., Mustard and Head, 1996; Mustard et al., 1998; (which appear deep blue in Plate 1 and are shown with long- Staid et al., 1996] could be used to determine whether their dashed lines in Figure 5c) are very bright, exhibit no ironbearing mineral component near 1 ptm, but do exhibit an inflecsomewhat higher albedo could be due to minor highland contamination. The strong 1-txm absorption and low albedo allow tion longward of 0.9 txm, indicating that the primary lithology small, similar basaltic mare ponds to be easily identified in Plate present at these locations is plagioclase, or anorthosite. On the other hand, a few areas, typically less bright, do not exhibit the inflection near 0.9 tm but are also blue in Plate 1 (shown with dash-dotted line in Figure 5c). Because the inflection near throughout this part of the highlands by their yellow appearance in this color composite. For example, in addition to the Waterman ponds, another smaller occurrence of basalt (previously unidentified) occurs to the north of Tsiolkovsky at a roughly symgm is present in crystalline plagioclase with a 1.25-txm ferrous metric position to the crater as the Waterman ponds in the south. absorption, the lack of such an inflection, and lack of evidence for any mafic absorptionear 1 txm, suggesthese areas are the more heavily shocked form of anorthosite [Adams et al., 1979]. A smaller number of small, fresh highland areas exhibit a weak ferrous absorptio near tm. Two such areas are shown with solid lines in Figure 5c. This ferrous absorption, presumably due to the presence of minor amounts of pyroxene, is North of Tsiolkovsky, just to the northwest of this northern basaltic area, is a 25-km floor-fractured crater. A Lunar Orbiter image of this crater is shown in Figure 6 along with the Clementinc 750 image. As seen in the color composite of Plate 1, the soils of the region are typical of these farside highlands. A small, bright crater in the upper center exhibits anorthositicharacteristics (Nhigh cr (ffc) in Figure 5c). A 2-km crater in the very weak. These areas are interpreted to be noritic anorthosite. northwest quadrant of the floor-fractured crater, however, exhib-

10 ,944 PIETERS AND TOMPKINS' CRUSTAL PROCESSES ON THE LUNAR FARSIDE ": '<---' - : 4..:'!-" :.'-!i :, ::½. '";";"'"'"' :":='; ß ' ß ;'.:.; '.:.:.:.½ U......,,.:!; :;.:-::..':, "4': "i. ' '*"?'::'" ':' ':E"; :::.::..':: Figure 6. Floor fractured crater north of Tsiolkovsky: (a) Clementine image at 750 nm and (b) Lunar Orbiter I- 136H image. Locations of areas for which spectra were acquired are indicated with arrows. its a distinctly lower albedo material. A five-color spectrum of units of the Tsiolkovsky region. This more focused color comthis dark soil (ffcr-m in Figure 5b) exhibits a prominent 1-gm ab- posite display (Plate 2a) suggests the presence of plagioclase as sorption and strongly suggests an affinity with mare basalt soils. well as a mafic-bearing lithology for the central peaks The type Although this feature is darker than surrounding highland soils, of iron-bearing mineral present, however, cannot be determined its higher albedo compared with other mare soils is likely a result with this simple color composite display. The three images used of significant contamination with surrounding highlands (mixing for the color composite are shown in Figures 7a-7c. In these imof basalt with high -albedo material raises the albedo but does not ages, the presence of a mafic-bearing lithology is easily noted eliminate the prominent ferrous absorption features). The asso- (bright areas in the 750/950 ratio of Figure 7b), but the mineralciation of basalts with this highland floor-fractured crater is con- ogy cannot be more explicitly identified. Five-color spectra for sistent with the suggestion [Brennan, 1975; Schultz, 1976] that pixels taken along traverses across several of the central peaks intrusive volcanic material emplaced beneath the rubble constitut- and for blocks along the southern crater wall are shown in Figure ing the floor accounts for the fracture pattern observed in such 8. The direction for the three peak traverses is shown in Figure craters. 7a, and each spectrum in the sequence represents a 4x4 pixel av- Given the very anorthositic nature of the Tsiolkovsky region, erage. it is perhaps rather surprising that multiple examples of basaltic The central peaks of Tsiolkovsky exhibit striking composimaterial (other than the mare filling Tsiolkovsky itself) occur tional boundaries consisting of a relatively simple mineral asthroughouthe region. If it is found that this association is semblage of various amounts of plagioclase and olivine (anorunique to this area, it could simply be a result of the fractured thosite, troctolite, and perhaps dunite). From the high-resolution crust formed by the Tsiolkovsky impact creating pathways for in- images of the peaks obtained during Apollo 15 (Figure 2), outterior melt. On the other hand, if these small, very localized oc- crops of rocky material can be seen along the mountain crests. currences of basaltic material are typical of farside highland The traverse of the western peak (Figure 8a) starts at the top in a crust, such observations fit the model of Head and ilson[ 1992, plagioclase unit, traverses the crest into a sharply delineated oli- 1994], which predicts the preponderance of stalled intrusions of vine-bearing lithology, and follows the olivine-bearing compobasaltic melt on the farside [Yingst and Head, 1998, 1999]. nent downslope to the south, where it becomes more diluted with additional plagioclase and more weathered soil. The sharp reflectance drop between Wpk6 and Wpk7n is where the traverse be- 7. Character of the Deep-Seated Highland Crust gins the descent on the shaded si. de of the mountain. In the secat Tsiolkovsky ond traverse, the centrally located peak ridge that is almost or- Because of its size (180 km), Tsiolkovsky has clearly exca- thogonally oriented to the western ridge exhibits the highest convated well into the crust of the Moon at this location on the far- centration of olivine at this resolution (Figure 8b). The third side. The peaks represent the deepest material, which could have traverse across the eastern ridge shows it to be largely plagioclase originally occurred at 20- to 35-km depth (depending on the but containing a distinct and very narrow outcrop of the olivinecratering model)[melosh, 1989; Cintala and Grieve, 1994]. bearing lithology. Clementine mosaics of the central peaks and a representative sec- For comparison to the olivine lithology seen at Tsiolkovsky, tion of the wall are shown in Figure 7 and Plate 2. The overview similar Clementine data for the central peaks of Copernicus are color composite display of Plate 1 helped to identify the general shown in Figure 8d. The solid and short-dashed lines are for

11 PIETERS AND TOMPKINS: CRUSTAL PROCESSES ON THE LUNAR FARSIDE 21,945 "7ii 750 Figure 7. The central peaks of Tsiolkovsky (two-frame mosaic of low-exposure images). Width of each frame is 60 km. Shown are (a) albedo image at 750 nm (arrows indicate traverse shown in Figure 8), (b) 750/950 nm ratio image with normal stretch, (c) 750/415 nm ratio image with normal stretch, (d) 750/900 nm ratio image with enhanced (12%) stretch, (e) 750/950 nm ratio image with enhanced stretch (12%), and (f) 750/1000 nm ratio image with enhanced stretch (12%). Figures 7b and 7c are the same ratio image with different contrast stretches. The standard color composite (Plate 2a) uses Figures 7a, 7b, and 7c. A color composite of enhanced stretched images Figures 7d, 7e, and 7f is shown in Plate 2c. "peak 3," the centermost of the central peaks. This central peak was the target for which the high spectral resolution, but lower spatial resolution, near-infrared telescope data clearly identified the distinctive broad ferrous absorption of olivine [Pieters, 1982, 1993]. Spectra indicated with long-dashed lines are for smaller peaks to the east not resolved by the telescope data. The spectra of these small peaks clearly indicate plagioclasessentially devoid of iron-bearing minerals also make up some of the peaks of Copemicus. The shape of the five-color spectra provides strong evidence that the mafic mineral present in the peaks of Tsiolkovsky is olivine. For both Tsiolkovsky and Copemicus, the central peaks exhibit extensive anorthosite with narrow zones of abundant olivine. At higher resolution, some zones could even prove to be dunite. From the total lack of concave curvature <1.0 Ism, no suggestion of pyroxene is detected (see Figure 3), and it is important to reiterate that pyroxene is much more optically active than olivine. When the two minerals coexist, pyroxene dominates, causing the spectrum of the mixture to exhibit prominent pyroxene characteristics near both 1 and 2 Ism. Because of this, it is easy to hide some olivine with pyroxene in a spectrum, but it is very hard to hide even small amounts of pyroxene with olivine. At Copernicus, high spectral resolution near-infrared spectra were used to verify the absence of any 2-Ism pyroxene absorption, confirming the lack of pyroxene at that site [Pieters, 1982, 1993]. Using the Copernicus data as a guide, a reasonable limit to low-ca pyroxene abundance for both the anorthosite and olivinerich lithologies discussed here is estimated to be <5%. The olivine-dominated unit in Tsiolkovsky's central peaks exhibits sharp contacts with anorthosite. Although high spectral resolution data are normally required to characterize olivinebearing rocks, the simplicity of the mineralogy of Tsiolkovsky peaks (olivine without pyroxene) allows the olivine-rich litholo- gies to be distinguished from pyroxene-rich lithologies that are elsewhere in the region. Specially processed multispectral images allow the spatial distribution of olivine to be examined in further detail. As discussed above (and evident in Figures 5 and 8), at Clementine wavelengths the ferrous absorption of olivinedominated lithologies exhibits a continuous decrease in reflectance from 750 to 1000 nm, whereas absorptions due to pyrox-

12 21,946 PIETERS AND TOMPKINS' CRUSTAL PROCESSES ON THE LUNAR FARSIDE o.4 I 'Wpk2 I L. ñ Wpk5 : Wpk7 r _ a Wpk6 I ] Wpk lo ß w pk 11 --e--w pk ß ' --.--W pk W pk 14 -,--W pk [ 15[ [ Tsiolkovsky We, t Pea[k Traverse o.15 [- ' -'... o o o 8 o Wavelength (pm) e -Cpkl [] Cpk2 )( C pk I I Tsiolkovsky Central Peak Traverse I I I I I Wavelength ( m) I I I I I I, I c J 0.45 I I I I I I I d Tsiolkovsky East Peak Traverse 0.40 Copernicus Central Peaks --B--E pk 2 1 ] '-'Epk3 z Epk41 I --v''e ksi '("". -' ] T 8 --B--E pk " ";'; I /,, _c --_',,,.- -',' 1 T.... { o.o I ½:-!½:h"",,,, 0.15 I I I I I I I Wavelength (p,m) Wavelength (p,m) m Tsiolkovsky --m S. Crater Wall Blocks j ---x /.o ß ' rr rr I I I I I I I 0.05 I I I I I I I Wavelength (Fro) Wavelength (Fro) Figure 8. Five-color spectra of the central peaks and walls of Tsiolkovsky and Copernicus. Error bars are one standard deviation for 4x4 pixel averages. Shown are (a) Tsiolkovsky west peak traverse, (b) Tsiolkovsky centermost peak traverse, (c) Tsiolkovsky east peak traverse, (d) Copernicus central peaks, (e) Tsiolkovsky south wall blocks, and (f) Copemicus south wall enes are slightly concave over the same wavelengths. This important distinction can be seen in controlled contrast stretches of spectral images ratioed with the 750-nm image. Shown in Figures 7d, 7e, and 7f are the 900-, 950-, and 1000-nm ratio images, all stretched the same 12% (1-1.12) to emphasize areas with strong mafic absorptions. Note that Figures 7b and 7e are the same image ratio (750/950) stretched to display different aspects of the image. In these highly contrasted ratio images (Figures 7d- 7f), fresh craters in the mare as well as the outcrops of mafic material in the central peaks all have high ratio values (strong absorption). The strength of the olivine absorption for central peak areas, however, clearly increases with increasing wavelength, whereas the pyroxene absorption in the maria is approximately equal in strength at 950 and 1000 nm. When these images are presented in color composite form (Plate 2c), the olivine-rich (green) and the pyroxene-rich (purple) materials are easily distinguished and mapped based on the character of the marie absorption (independent of albedo). Substituting an albedo image (750 nm) for one of the ratio images allows the spatial relations of the prominent olivine outcrops to be seen in geologic context, as illustrated in the composite shown in Plate 2d. Although the best exposures of olivine-rich material can be seen to occur along the ridge crests of the peaks, not all of the blocks seen in the high-resolution image of Figure 2 are olivine bearing. In particular, a large part of the actual crest of the western peak exhibits unshocked anorthosite properties,

13 PIETERS AND TOMPKINS: CRUSTAL PROCESSES ON THE LUNAR FARSIDE 21,947 with much of the olivine-rich outcrop occurring slightly downslope on its southern flank until the point where the western crest joins the olivine-rich crest of the orthogonal centermost peak. If this plagioclase-olivine distribution along the western ridge is related to the small ridge crest crater (seen in Figure 2), the coherent olivine-rich zone may be relatively thin (<1 km) to be disrupted by such a small crater. 8. Character of the Highland Crust Below the Megaregolith In contrast to the distinctive plagioclase-olivine assemblage seen in the deep-seated material of the central peaks, the composition of the walls and rim of Tsiolkovsky is more complex, with anorthosite and noritic anorthosite the most commonly observed rock type. Although the additional presence of olivine-bearing species in some locations along the wall cannot be ruled out, it has not been readily identified. These rim and wall deposits should represent material at Tsiolkovsky from a higher stratigraphic zone than those of the central peaks. Some areas along CUS. central peaks were found to be entirely anorthosite, and another 45% contained anorthosite in conjunction with more mafic lithologies. Furthermore, at least six (including Tsiolkovsky) were found to contain plagioclase devoid of iron-bearing minerals and an olivine-bearing specie similar to that found at Tsiolkovsky. Therefore any explanation for the origin of an olivineanorthosite association should apply to all instances and be of global significance. The origin of olivine in central peaks has been under investigation since its original detection in Copernicus [Pieters and Wilhelms, 1985]. In the early study, two principal hypotheses were proposed: (1) excavation of a mantle that had been uplifted by earlier events and (2) excavation of a differentiated layered pluton emplaced in the crust. At that time the hypothesis of a layered pluton in the lower crust was preferred. Two additional hypotheses more recently discussed for evolution of the lunar crust must also be considered: (3) excavation of a large differentiated melt sheet (described by Grieve et al. [1991]) and (4) excavation of nonequilibrium rock types involving Mg-rich olivine and anorthosite perhaps resulting from early mantle overturn (described by Hess [1994] and Hess and Parmentier [1995]. Resolution of the issue about the origin of the olivine-rich zone found in selected central peaks is beyond the scope of this paper but is sum- the wall are principally one lithology, such as the southeast wall dominated by anorthosite (note the high albedo and extensive deep blue region in Figure 4 and Plate 1). Most of the wall and rim, however, exhibit a neighboring jumble of distinct lithologies. A typical example is the south wall shown in Plate 2b. Five-color spectra for several nearby locations within this image marized by Pieters and Tompkins [1999]. The new pulse of data for the Moon has allowed compositional issues about unexplored lunar regions to be seriously studare presented in Figure Be; such spectra are used to characterize ied for the first time in decades. The 180-km crater Tsiolkovsky the rock types exposed along the wall. No coherent pattern is ob- is particularly interesting because it is on the farside and is in a served among the knobs and blocks. Adjacent features can have unrelated compositions with blocks of anorthosite (deep blue in Plate 2b) commonly seen next to blocks of noritic anorthosite Tsiolkovsky Region (turquoise). The excavation and emplacement of these materials (-3.9 AE) during the Tsiolkovsky impact event appear to have maintained Km coherent blocks on the several kilometers scale, but such blocks are randomly mixed on the scale of tens of kilometers. Megaregolith A similar scale of lithologic diversity within the walls of a large crater was seen at Copernicus during early analysis of Clementine data [Pieters et al., 1994]. The improved calibrations \ currently in use confirm this diversity and allow more direct Noritic compositional interpretations of the spectral character of ob- Anorthosite served wall blocks (shown for comparison in Figure 8f). Like 10- Tsiolkovsky, there is much anorthosite within the walls of Copernicus. The anorthosite spectra (shown with dash-dotted lines), however, are not quite as bright and do not exhibit any turndown in the last few filters, which would indicate crystalline plagio- Anorthosite clase. These Copernicuspectr are thus interpreted to represent more heavily shocked plagioclase. Neighboring these anorthositic blocks in the south wall of Copernicus are materials with a much lower albedo and strong 1-gm ferrous absorption. These dark wall blocks are interpreted to be remnants of the thin surface basalt, or possibly feeder dikes, of the original target at Coperni- Olivine (Dunite/Troctolite) Figure 9. Sketch of crustal stratigraphy Tsiolkovsky (prior to 9. Discussion and Summary the crater formation). The upper 2 km of megaregolith is a very Although each large crater has its own story tell, the sharp- anorthositic rubble pile with minor components of noritic anor- ness of distinct lithologies across the central peaks such as those seen in Tsiolkovsky is not uncommon among central peaks on the Moon. In a global survey of 109 large craters with central peaks, Tompkins and Pieters [1999] found that 41 (38%) exhibited multiple lithologies and that sharp boundaries were observed between different materials for about half of the craters with multiple lithologies. Of the 109 craters studied, 17% of the deep-seated thosite. The upper crust consists principally of anorthosite with local areas of noritic anorthosite of unknown origin. Intermediate depths contain more abundant noritic components. These noritic and anorthositic zones are later exposed along the walls of Tsiolkovsky. Depths approaching the excavation depth of Tsiolkovsky (-20 km?) exhibit a mafic-free anorthosite zone. Olivinerich material is emplaced within the deep anorthosite and later excavated to form the central peaks.

14 21,948 PIETERS AND TOMPKINS: CRUSTAL PROCESSES ON THE LUNAR FARSIDE region beyond the topographic and gravity anomalies associated Brennan, W. J., Modification of premare impact craters by volcanism and with the huge South Pole-Aitken Basin which dominates much tectonism, Moon, 12, , Burns, R. G., Mineralogical Application of Crystal Field Theory, 2nd ed., of the farside. The detailed relations among compositional data 551 pp., Cambridge Univ. Press, New York, observed for Tsiolkovsky provides a window into the lunar far- Cintala, M. J., and R. A. F. Grieve, The effects of differential scaling of side crust and suggests the following partial story for evolution of impact melt and crater dimensions on lunar and terrestrial craters: the crust in this environment. Some brief examples, in Large Meteorite Impacts and Planetary Evo- Prior to the impact event that created Tsiolkovsky crater (-3.9- lution, edited by B. O. Dressier, R. A. F. Grieve, and V. L. Sharpton, 3.8 eons ago), the stratigraphy of the farside highland crust in that Geol. Soc. Am. Spec. Pap., 293, 51-59, Fischer, E., and C. M. Pieters, Remote determination of exposure degree part of the Moon is heterogeneous with depth, as described in and iron concentration of lunar soils using VIS-NIR spectroscopic methods, Icarus, 111, , Fischer, E. M., and C. M. Pieters, Lunar surface aluminum and iron concentration from Galileo solid state imaging data, and the mixing of mare and highland materials, d. Geophys. Res., 100, 23,279-23,290, Figure 9. The overall composition is very anorthositic, with pockets of more noritic-rich material. The noritic components disappear with depth, but near -25 km a zone occurs that contains olivine-rich material embedded within anorthosite (either as intrusive dikes perhaps resulting from mantle overturn or as a pyroxene-free layered pluton). The crater-forming event excavated through this column, bringing the olivine/anorthosite material to the surface to form the central peaks, emplacing abundant anorthosite along the southeast wall, and distributing blocks of anorthosite and noritic anorthosite around the wall of the 180 km cra- ter. Subsequently, low-titanium mare basalt flooded and filled the floor of the crater. Small amounts of additional basaltic material also reached the surface around the periphery of the crater (within one crater diameter), but some basaltic melts never quite made it to the surface, sometimes filling spaces beneath loose rubble of a crater floor. Although anorthosite is seen to be abundant in this region, the evolution of the crust has also clearly involved intrusive events. The process responsible for the olivine-rich material now occurring in the central peaks, however, was earlier than and different from that which produced the mare basalts filling the crater-3.85 eons ago. Acknowledgments. The authors appreciate the thoughtful reviews by B. Jolliff and P. Spudis. The innovative assistance of George (Guoqi) He in data processing of the Clementine multispectral images was a significant contribution. Digitization and careful production of the photographic images by Peter Neivert is much appreciated. NASA support for this research is gratefully acknowledged: grants NAGW-4910, NAG (C.M.P.), and NASW (S.T.). References Adams, J. B., and T. B. McCord, Alteration of lunar optical properties: Age and compostion effect, Science, 171, , 1971a. Adams, J. B., and T. B. McCord, Optical properties of mineral separates, glass, and anorthositic fragments from Apollo mare samples, Proc. Lunar Sci. Conf. 2nd, , 197 lb. Adams, J. B., and T. B. McCord, Vitrification darkening in the lunar highlands and indentification of Descartes material at the Apollo 16 sites, Proc. Lunar Sci. Conf 4th, , Adams, J. B., F. H6rz, and R. V. Gibbons, Effects of shock-loading on the reflectance spectra of plagioclase, pyroxene and glass (abstract), Lunar Planet. Sci., X, 1-3, Adler, I. J., et al., The Apollo 15 X-ray flourescencexperiment, Proc. Lunar Sci. Conf 3rd, , Adler, I. J., et al., Results of the Apollo 15 and 16 X-ray experiment, Proc Lunar Sci. Conf 4th, , Belton, M. J. S., et al., Lunar impact basins and crustal heterogeneity: New western limb and far side data from Galileo, Science, 255, , Belton, M. J. S., et al., Galileo multispectral imaging of the north polar and eastern limb regions of the Moon, Science, 264, , Binder, A. B., Lunar Prospector: Overview, Science, 281, , Blewett, D. T., P. G. Lucey, B.R. Hawke, and B. L. Jolliff, Clementine images of the lunar sample-return stations: Refinement of FeO and TiO 2 mapping techniques, J. Geophys. Res., 102, 16,319-16,325, Fischer, E. M. and, C. M. Pieters, Composition and exposure age of the Apollo 16 Cayley and Descartes regions from Clementine data: Normalizing the optical effects of space weathering, J. Geophys. Res., 101, , Gaffey, S. L., L. A. McFadden, D. Nash, and C. M. Pieters, Ultraviolet, visible, and near-infrared reflectance spectroscopy: Laboratory spectra of geologic materials, in Remote Geochemical Analysis: Elemental and Mineralogical Compositton., edited by C.M. Pieters, and P.A.J. Englert, pp , Cambridge Univ. Press, New York, Greeley, R., et al., Galileo imaging observations of lunar maria and related deposits, d. Geophys. Res., 98, 17,183-17,205, Grieve, R. A. F., D. St6ffier, and A. Deutsch, The Sudbury structure: Controversial or misunderstood?, J. Geophys. Res., 96, 22,753-22,764, Guest, J. E., and J. B. Murray, Nature and origin of Tsiolkovsky crater, lunar farside, Planet. Space Sci., 17, , Hapke, B., Theory of Reflectance and Emitrance Spectroscopy, 455 pp., Cambridge Univ. Press, New York, Hartmann, W. K., R. J. Phillips, and G. L. Taylor, Origin of the Moon, 781 pp., Lunar and Planet. Inst., Houston, TX, Hawke, B. R., P. G. Lucey, G. J. Taylor, J. F. Bell, C. A. Peterson, D. T. Blewett, K. Horton, and G. A. Smith, Remote sensing studies of the Orientale region of the Moon: A pre-galileo view, Geophys. Res. Lett., 18, , Hawke, B. R., C. A. Peterson, P. G. Lucey, and D. T. Blewett, Remote sensing studies of lunar anorthosite deposits (abstract), Lunar Planet. Sci., XXVI, , Head, J. W., Lunar volcanism in space and time., Rev. Geophys., 14, , Head, J. W., and L. Wilson, Lunar mare volcanism: Stratigraphy, eruption conditions, and the evolution of secondary crusts, Geochim. Cosmochim. Acta, 56, , Head, J. W., and L. Wilson, Lunar graben formation due to near-surface deformation accompanying dike emplacement, Planer Space Sci., 41, , Hess, P. C., Petrogenesis of lunar troctolites, J. Geophys. Res., 99, 19,083-19,093, Hess, P. C., and E. M. Parmentier, A model for the thermal and chemical evolution of the Moon's interior: Implications for the onset of mare volcanism, Earth Planer Sci. Lett., 134, , Lane, A. P., and W. M. Irvine, Monochromatic phase curves and albedos for the lunar disk, Astron. X, 78(3), , Lucey, P. G., G. J. Taylor, and E. Malaret, Abundance and distribution of iron on the Moon, Science, 268, , Lucey, P. G., et al., Calibration of the Clementine near infrared camera: Ready for prime time (abstract), Lunar Planet. Sci., XXIX, 1576, 1998a. Lucey, P. G., D. T. Blewett, and B.R. Hawke, Mapping the FeO and TiO2 content of the lunar surface with multispectral imagery, J. Geophys Res., 103, , 1998b. McEwen, A. S., and M. S. Robinson, Mapping of the Moon by Clementine, Adv. Space Res., 19, , McEwen, A. S., M. S. Robinson, E. M. Eliason, P.G. Lucey, T.C. Duxbury, and P. D. Spudis, Clementine observations of the Aristarchus region of the Moon, Science, 266, , McEwen, A., et al, Summary of radiometri calibration and photometric normalization steps for the Clementine UVVIS images [CD-ROM], Lunar Planet. Sci., XXIX, Melosh, H. J., Impact Cratering.' A Geologic Process, 254 pp., Oxford Univ. Press, New York, 1989.

15 PIETERS AND TOMPKINS: CRUSTAL PROCESSES ON THE LUNAR FARSIDE 21,949 Mustard, J. F., and J. W. Head, Mare-highland mixing relationships along Pohn, H., and R. L. Wildey, A photoelectric-photographic study of the the southwestern shores of Oceanus Procellarum, d. Geophys Res., normal albedo of the Moon, Map 599-E, U.S. Geol. Surv. Prof Pap., 101, 18,913-18,925, Mustard, J. F., L. Li., and G. He, Nonlinear spectral mixture modeling of Schultz, P. H., Floor fractured lunar craters, Moon, 15, , lunar multispectral data: Implications for lateral transport, d. Geo- Solomon, S.C., and J. W. Head, Vertical movement in mare basins: Relaphys. Res., 103, 19,419-19,425, tion to mare emplacement, basin tectonics and lunar thermal history, Nozette, S., et al., The Clementine mission to the Moon: Scientific over- d. Geophys. Res. 84, , view, Science, 266, , Solomon, S.C., and J. W. Head, Lunar mascon basins: Lava filling, tec- Ostertag, R., Shock experiments on feldspar crystals, Proc. Lunar Planet. tonics, and evolution of the lithosphere, Rev. Geophy., 18, , Sci. Conf. 14th, Part 1, d. Geophys Res., 88, suppl., B364-B376, Peterson, C. A., B. R. Hawke, P. G. Lucey, G. J. Taylor, D. T. Blewett, Staid, M. I., C. M. Pieters, and J. W. Head, Mare Tranquillitatis: Basalt and P. D. Spudis, Anorthosite on the lunar nearside and farside. Lunar emplacement history and relation to lunar samples, d. Geophys. Res., Planet. Sc., XXVII, , , 23,213-23,228, Pieters, C. M., Mare basalt types on the front side of the Moon: A sum- Tompkins, S., and C. M. Pieters, Mineralogy of the lunar crust: Results mary of spectral reflectance data, Proc. Lunar Planet. Sct. Conf 9th, from Clementine, Meteorit. Planet. Sci., 34, 25-41, , U S. Geological Survey, Clementine digital image model UVVIS 750- Pieters, C. M., Copernicus crater central peak: Lunar mountain of unique nm basemap, USANASAPDSCL , Reston, Va., composition, Science, 215, 59-61, Wilbur, C. L.,Volcano-tectonic history of Tsiolkovskij (abstract), Lunar Pieters, C. M., Composition of the lunar highland crust from near infrared Planer Sci. IX, , spectroscopy, Rev. Geophys., 24, , 1986 Wilhelms, D., The geologic history of the Moon, U.S. Geol. Surv. Prof Pieters, C. M., Compositional diversity and stratigraphy of the lunar crust Pap. 1348, 302 pp, derived from reflectance spectroscopy, in Remote Geochemical Analy- Wilhelms, D. E., and F. E1-Baz, Geologic map of the east side of the sis.' Elemental and Mineralogical Composition., edited by C. M. Piet- Moon, Map 1-948, U.S. Geol. Surv. Reston, Va., ers and P. A. J. Englert, pp , Cambridge Univ. Press, New Yingst, R. A., and J. W., Head, Characteristics of lunar mare ponds in York, Smythii and Marginis Basins: Implications for magma transport Pieters, C. M., Plagioclase and maskelynite diagnostic features, Lunar mechanisms, d. Geophys. Res., 103, 11,135-11,158, Planet, Sci. Conf XXVII, , Yingst, R. A. and J. W. Head, Geology of mare deposits in South Pieters, C. M., and S. Tompkins, The distribution of lunar oli- Pole-Aitken Basin as seen by Clernentine UVVIS data, J. vine/troctolite outcrops: Mineralogical evidence for mantle overturn?, Geophys. Res., 104, 18,957-18,979, Lunar Planer Sci. XXX,, 1286, Zuber, M. T., D. E. Smith, F. Lemoine, and G. A. Neumann, The shape Pieters, C. M., and D. E. Wilhelms, Origin of olivine at Copernicus, and internal structure of the Moon from the Clementine Mission, Sci- Proc. Lunar Planet. Sci. Conf 15th, Part 2, J. Geophys. Res., 90, ence, 266, , suppl. C415-C420, Pieters, C. M., et al., Crustal diversity of the Moon: Compositional analyses ofgalileo SSI data, d. Geolvhys. Res., 98, 17,127-17,148, Pieters, C. M., et al, A sharper view of impact craters from Clementine C.M. Pieters, Department of Geological Sciences, Brown University, data, Science, 266, , Providence, RI (pieters mare.geo.brown.edu) Pieters, C. M., J. F. Mustard, and J. M. Sunshine, Quantitative mineral S. Tompkins, Science Applications International Corporation, 4501 analyses of planetary surfaces using reflectance spectroscopy, in Min- Daly Drive, Suite 400, Chantlily, VA eral Spectroscopy: A Tribute to Roger G. Burns, edited by M.D. Dyar, C. McCammon, and M. Schaefer, pp , Geochem. Soc. (Received November 30, 1998; revised June 2, 1999; Slveci. Pub. 5, , accepted June 3, 1999.)