SEARCH

SEARCH BY CITATION

Keywords:

  • pyroclastic deposits;
  • radiative transfer modeling

Abstract

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[1] While over 100 lunar pyroclastic deposits have been identified remotely, their compositions remain poorly constrained. In this work, we determine the compositions of three lunar regional pyroclastic deposits which are likely to contain large amounts of glass and for which we have high-quality ground-based spectra: the Aristarchus Plateau, Mare Humorum, and Sulpicius Gallus. We use radiative transfer theory and employ measured optical constants of glasses to predict the bidirectional reflectance of a particulate glass surface as a function of Fe and Ti concentration, particle size, and maturity in order to find the best spectral match to the remotely observed deposits. Tools are not available to model the optical effects of the unusual geometries of the ilmenite laths in the black beads, so we address their effects on spectra of the regional pyroclastic deposits using computational mixing. We find that model spectra of pure glass (as opposed to devitrified black beads) provide good matches to all three regions. Radiative transfer modeling predicts iron contents of 21, 20, and 17 wt% FeO for Aristarchus, Humorum, and Sulpicius Gallus, respectively, and suggests that all three regions are low in titanium, a result supported by Lunar Prospector neutron spectrometer data. However, we find that a moderate Ti glass mixed with a small fraction of black beads cannot be ruled out for the Sulpicius Gallus region.

1. Introduction

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[2] Approximately 100 pyroclastic deposits have been identified remotely on the lunar surface on the basis of characteristics such as low albedo, diffuse margins, mantling of preexisting topography, and extremely low returns of radar reflectivity [Pieters et al., 1973; Zisk et al., 1974; Hawke et al., 1979; Gaddis et al., 1985, 2003]. Most of these pyroclastics are small localized deposits presumed to have formed in vulcanian-style eruptions, where local rock caps a magma body until the accumulation of volatiles causes catastrophic failure of the cap and blasts a mixture of local rock and juvenile basaltic material around an endogenic crater [Head and Wilson, 1979; Hawke et al., 1989; Coombs et al., 1990]. Eleven (and possibly several more) of the identified deposits are large regional pyroclastic deposits. These are formed in explosive Hawaiian-style or fire fountain eruptions where volatiles rapidly exsolve as magma ascends to the surface, and bubbles coalesce and expand to fragment the magma and blast it from the vent. The low gravity and lack of an atmosphere on the Moon result in extremely rapid expansion of volatiles, highly fragmented magmas, and widespread deposits of submillimeter pyroclastic beads covering thousands of km2 [Wilson and Head, 1981].

[3] On the basis of samples collected by Apollo astronauts and inferred from remote sensing, regional pyroclastic deposits are thought to consist of two types of materials: glasses and their compositionally identical devitrified equivalent (also known as black beads). The type examples of these are the high-Ti Apollo 17 orange glass and black beads. In the case of the black beads, a slower cooling rate causes the crystallization of blade-like laths of ilmenite along with fine-grained olivine, spinel, and metal. Low-Ti glasses that experienced slower cooling might also form tiny crystals of olivine, spinel, and metal, but would be much less noticeable because of the lack of ilmenite, which causes the black appearance [Papike et al., 1998]. This work focuses on three regional pyroclastic deposits that are likely to contain large amounts of glass: Aristarchus, Humorum, and Sulpicius Gallus.

[4] Chemical analyses and high-pressure experiments on lunar pyroclastic glasses show that their parent magmas originated deep in the lunar mantle (>400 km) and experienced little fractionation during ascent [e.g., Delano, 1986; Papike et al., 1998]. Consequently these glasses are considered to be the best candidates for primary magmas, representing the composition of the region of lunar mantle in which they originated. Twenty-five categories of pristine glasses have been identified in samples collected by Apollo astronauts. However, apart from the Apollo 15 green glass and the Apollo 17 orange glass, these glasses are minor components of the regolith, and their provenance is not known, even if their compositions are known in detail. In contrast, the chemical compositions of the precisely located remotely observed regional pyroclastic deposits are largely unknown. The ultraviolet, visible and near-infrared (UVVIS and NIR) spectra of these deposits have the potential to reveal aspects of their compositions.

[5] The UVVIS-NIR spectral behavior of glass is relatively straightforward. Known lunar pyroclastics have basaltic silicate compositions, with the main variable being titanium content. The detailed shape and mean reflectance of spectra of glasses of this composition are a combination of four main components: iron content, titanium content, particle size, and maturity. The spectral effects of iron in a glass are similar to its spectral effects in pyroxene, the major iron-bearing phase in lunar rocks and soils. Crystal field absorption bands at ∼1 and 2 μm appear and increase in strength with increasing iron content [Bell et al., 1976]. The gross similarities between spectra of iron-bearing glasses and pyroxenes are due to the fact that much of the Fe2+ in glass is located in distorted octahedral sites, similar to the M2 site in pyroxene [Wells and Hapke, 1977]. However, Fe2+ is also found in tetrahedral coordination and while these crystallographic sites exist, glass has only short-range order, resulting in glass absorption bands that are broader and shallower than pyroxene absorption bands [Bell et al., 1976; Wells and Hapke, 1977].

[6] The effect of the second component, titanium, on glass spectra is not similar to the spectral effect of ilmenite, the major titanium-bearing phase in lunar rocks and soils. The reflectance of glass containing titanium decreases sharply toward UV wavelengths, because of strong Fe2+–Ti4+ charge transfer absorption bands between 0.3 and 0.6 μm [Bell et al., 1976; Burns et al., 1976; Wells and Hapke, 1977]. At wavelengths beyond 0.7 μm, outside the influence of the charge transfer absorption, titanium has a minimal effect on the reflectance of glasses or on the intensity of absorption features in their spectra. Thus with increasing amounts of Fe and Ti, glasses become dark in the ultraviolet and more spectrally red in the UVVIS region. Ilmenite, in contrast, causes generally red lunar spectra to become less red (or more blue), due its dark and neutral character. Accordingly, the widely used relationship of increasing UV/VIS ratio with increasing TiO2 content observed by Charette et al. [1974] does not apply to glasses, and in fact the opposite is true [Wells and Hapke, 1977].

[7] The third component, particle size, affects the overall reflectance of particulate surfaces composed of glass and the intensity of absorption bands. For particles that are transparent or weakly absorbing, like glass, a decrease in particle size results in higher reflectance because of an increase in the number of particle boundaries relative to material traversed through the volume, and thus an increase in the chance of a photon being scattered toward the observer before it is absorbed. Conversely, surfaces composed of larger particles have lower albedos because more absorbing material is traversed between scatters, increasing the probability of a photon being absorbed rather than escaping the surface [e.g., Adams and Filice, 1967].

[8] Maturity, the last major influence on lunar glass spectra, is the term applied to the effects of a material being exposed on the lunar surface to the space environment, where it is bombarded with micrometeorites, solar wind, and galactic cosmic rays, processes collectively called “space weathering.” Space weathering acts to comminute particles to a finer size, produce glass-welded aggregates called agglutinates, and produce coatings of amorphous silica containing submicroscopic metallic iron (SMFe) on particles [Keller and McKay, 1993, 1997; Pieters et al., 2000; Hapke, 2001]. The glass in agglutinates derives from local melting by micrometeorite impacts, and this glass incorporates previously produced glass, mineral fragments, and fragments of SMFe-rich coatings. During the remelting, the SMFe appears to anneal to larger sizes, resulting in the agglutinate spectra being dark and neutral [Pieters et al., 1993; Britt and Pieters, 1994; Noble et al., 2001]. Materials exposed to the space-weathering environment “mature” with increasing exposure, and their spectra become increasingly dark, red, and show reduced spectral contrast because of the SMFe and agglutinates [Fischer and Pieters, 1994]. Because of these strong optical effects, in addition to composition, spectral studies of lunar pyroclastic glasses must take into account the effects of particle size and space weathering.

[9] Several studies have estimated the composition of lunar regional pyroclastic deposits using ground-based telescopic spectra [Pieters et al., 1973, 1974; Hawke et al., 1983; Gaddis et al., 1985; Lucey et al., 1986; Pieters and Tompkins, 2005] and Clementine five-band spectra [Weitz et al., 1998; Gaddis et al., 2003]. Telescopic spectra of the Aristarchus Plateau, Mare Humorum, and Sulpicius Gallus regions show broad 1 and 2 μm absorptions that have been interpreted as being indicative of the presence of Fe-bearing volcanic glass [Hawke et al., 1983; Gaddis et al., 1985; Lucey et al., 1986]. The Taurus-Littrow deposit has been interpreted as being largely composed of crystalline, Fe-rich, and Ti-rich material on the basis of shared spectral features with the Apollo 17 black beads [Adams et al., 1974; Pieters et al., 1974]. Weitz et al. [1998] used Clementine spectral ratios of pyroclastic deposits to estimate the glass/crystalline bead ratio, while a large survey of Clementine data by Gaddis et al. [2003] also interpreted spectral ratios in terms of crystallinity, Fe, and Ti content. In this work, we use radiative transfer modeling employing measured optical constants of glass to constrain the compositions of telescopic spectra of lunar pyroclastic deposits. We address the effects of black beads on the spectra of pyroclastic deposits using computational mixing because the spectral effects of the unusual geometry of ilmenite laths in a mostly glass matrix has not been modeled.

2. Methods

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

2.1. Telescopic Spectra

[10] We have analyzed the spectra of pyroclastic deposits in the three regions for which the highest-quality ground-based spectra exist: the Aristarchus Plateau, Mare Humorum, and Sulpicius Gallus (for locations, dates see Gaddis et al. [1985]). The spectra were acquired with the indium antimonide infrared circular variable filter (CVF) spectrometer mounted on the University of Hawaii 2.24-m telescope at Mauna Kea observatory. The spectrometer collected 120 wavelength channels between 0.6 and 2.5 μm; we have only used data up to 2.0 μm because of the presence of significant thermal emission beyond this wavelength. The spectra were converted to absolute reflectance by normalizing the spectra to the reflectance of each location derived from data from Clementine at 0.75 μm. We also extended the spectra to the UV by using the 0.415 μm band from Clementine data (Figure 1).

image

Figure 1. Telescopic spectra of the three regional pyroclastic deposits. Each spectrum was normalized to the value of its reflectance at 0.75 μm as determined from Clementine images. The 0.415 μm reflectance was also ascertained from Clementine data. Error bars represent 1 standard deviation of the mean of several independent telescopic measurements, except at 0.415 μm where they are the standard deviation of the mean of nine Clementine pixels.

Download figure to PowerPoint

2.2. Radiative Transfer Model

[11] In order to determine the compositions of the Aristarchus, Humorum, and Sulpicius Gallus pyroclastic deposits, we apply radiative transfer modeling to the remotely obtained spectra. The model is based on the bidirectional reflectance theory of Hapke [1981, 1993, 2001] as employed by Lucey [1998]. The radiative transfer model uses measured optical constants, also known as the complex indices of refraction, to predict the bidirectional reflectance of a particulate surface. Optical constants are fundamental properties of materials that are independent of particle shape or size, and their variations with wavelength control the spectral features observed, together with the physical state of the particles and the viewing geometry. The complex index of refraction of a material has two components, n, the real index of refraction (which is equivalent to the index of refraction in relatively transparent materials such as silicates and glasses) and k, the imaginary index of refraction (mathematically related to the absorption coefficient). The optical constants of a material and Hapke's treatment are used to calculate single scattering albedo at a specified particle size. The single scattering albedo is the probability of a photon surviving an encounter with a material, and the single scattering albedos of components in an intimate mixture add linearly in proportion to their abundance. The single scattering albedo is then used with Hapke's methods to calculate the bidirectional reflectance of the medium at specified incidence and emission angles.

2.3. Determination of Optical Constants

[12] To compute spectra of a lunar-like particulate glass surface, optical constants are required for the range of Fe and Ti contents found in lunar pyroclastic glasses. In this work we use directly measured optical properties of glasses of different compositions, and use the correlations between composition and refractive index to interpolate the constants to any iron and titanium concentration.

2.3.1. Imaginary Index of Refraction

[13] We used absorption coefficients measured by Bell et al. [1976] of synthetic glasses of seven compositions designed to address lunar glass properties, ranging from a high iron and titanium glass to end-members with no iron and no titanium. Bell et al. provided data from 0.5 to 2.5 μm for all seven glasses, and from 0.4 to 2.5 μm for four of the glasses. The absorption coefficient (α) is related to the imaginary index of refraction by

  • equation image

where λ is the wavelength. We used this relationship to convert the absorption coefficients of the seven glasses to k. Bell et al. indicated that the UV absorption was controlled by the sum of Fe and Ti, but did not present how the relative importance Fe and Ti vary with wavelength. The attribution of the UV absorption to a Fe-Ti charge transfer confined to the UV suggests that there should be variations in the relative contribution of the two key elements to the imaginary index of refraction with wavelength, where titanium would strongly influence k in the UV but have little effect at longer wavelengths. To quantify this, we measured the correlation of k with FeO and TiO2 as a function of the equation FeO + xTiO2, where x is a weighting factor used to vary the relative influence of TiO2 on the k spectrum. The weighting factor was varied between −1 and 2, because a reconnaissance test indicated that it did not vary outside these limits. A higher value for the weighting factor would indicate titanium increases the ability of the glass to absorb photons at that wavelength relative to a Ti-poor or Ti-free glass, whereas a negative weighting factor would indicate that titanium adversely affects the ability of the glass to absorb photons at that wavelength, relative to a glass with less titanium. The correlation of k with FeO + xTiO2 was examined at each weighting and wavelength (Figure 2) and the weighting factor that resulted in the maximum correlation at each wavelength was selected (Figure 3). The optimum TiO2 weighting factor was near zero through most of the VIS and NIR wavelengths, showing that Ti has little direct influence on glass spectra in this range, but increased to greater than 1.5 shortward of ∼0.7 μm, consistent with the presence of an Fe-Ti charge transfer absorption. The maximum correlation between k and FeO + xTiO2 increased very slightly at long wavelengths with a TiO2 weighting factor of up to 0.4. However, as we know of no mechanism whereby titanium would affect the spectrum at these wavelengths, we have set the TiO2 weighting factor to zero at wavelengths greater than 0.78 μm. Using these TiO2 weighting factors results in a high correlation between k and FeO + xTiO2 (Figure 4). The optimum correlation is greater than 0.9 at all wavelengths, and greater than 0.99 near 1 μm.

image

Figure 2. Imaginary index of refraction k versus the FeO + xTiO2 for the Bell et al. [1976] glasses is plotted at three wavelengths, where x is the optimum TiO2 weight factor at those wavelengths. A high correlation results when the optimum TiO2 weight factor is used.

Download figure to PowerPoint

image

Figure 3. Optimum TiO2 weighting factor (which resulted in the highest correlation of FeO + xTiO2 and k) versus wavelength. Dashed line indicates where we set the weighting factor to zero.

Download figure to PowerPoint

image

Figure 4. Solid line shows the correlation between FeO + xTiO2 and k when the optimum TiO2 weighting factor is used. Dashed line shows the correlation when the TiO2 weighting factor is set to zero past 0.76 μm.

Download figure to PowerPoint

[14] With the proper relationship between Fe, Ti, and k in hand, we fit k against the FeO + xTiO2 of the seven Bell et al. [1976] glasses at each wavelength to develop coefficients allowing derivation of k at any Fe and Ti content. From this linear fit, we attained a slope and an offset at each wavelength that relates the k value to the weighted sum of Fe and Ti. This can be used to calculate the complex index of refraction of a glass of an arbitrary iron and titanium composition at each wavelength by multiplying TiO2 by the determined optimum weighting factor, adding FeO, and multiplying this sum by the determined slope and adding the offset

  • equation image

where mλ is the slope and bλ is the offset, as given in Table 1. The resultant k spectrum is then incorporated into the model and together with particle size and viewing geometry used to calculate the reflectance spectrum as described in section 2.2.

Table 1. Offsets, Slopes, and TiO2 Weight Factor Used to Calculate Optical Constants of Glass With Equations (2)(4)
Wavelength, μmn Offset, bλn Slope, mλk Offset, bλk Slope, mλTiO2 Weight Factora
  • a

    Analysis in text set TiO2 weight factor to zero at wavelengths past 0.76 μm.

0.400.940160.044987---
0.4150.950120.037586−3.4779E-031.9462E-041.58000
0.420.953440.035119---
0.440.964320.026991---
0.460.973360.020202---
0.480.980970.014462---
0.500.987450.009559---
0.520.993010.005333---
0.540.997830.001660−7.0494E-044.4773E-051.61029
0.561.00203−0.001555−5.2397E-043.5300E-051.58129
0.581.00572−0.004388−3.5937E-042.7031E-051.52323
0.601.00897−0.006898−2.3324E-042.1068E-051.41692
0.621.01186−0.009133−1.3906E-041.7058E-051.25175
0.641.01444−0.011135−6.4481E-051.4317E-051.02870
0.661.01674−0.012934−2.7412E-061.2568E-050.73477
0.681.01881−0.0145584.0806E-051.1979E-050.44383
0.701.02068−0.0160306.9283E-051.2316E-050.21805
0.721.02238−0.0173688.7194E-051.3563E-050.08011
0.741.02391−0.0185899.4871E-051.5634E-050.01574
0.761.02531−0.0197061.0223E-041.7997E-05−0.02664
0.781.02659−0.0207311.0474E-042.0974E-05−0.03843
0.801.02776−0.0216741.1127E-042.3993E-05−0.05756
0.821.02883−0.0225441.1470E-042.7202E-05−0.06128
0.841.02981−0.0233481.1318E-043.0736E-05−0.05144
0.861.03072−0.0240931.1446E-043.4261E-05−0.04326
0.881.03155−0.0247841.0334E-043.8012E-05−0.01538
0.901.03232−0.0254289.6173E-054.1710E-050.00035
0.921.03303−0.0260271.0154E-044.4678E-05−0.00830
0.941.03368−0.0265871.0138E-044.7903E-05−0.00980
0.961.03429−0.0271111.0504E-045.0817E-05−0.01708
0.981.03486−0.0276011.1189E-045.3433E-05−0.02823
1.001.03538−0.0280621.1261E-045.5921E-05−0.03132
1.021.03586−0.0284951.0471E-045.8307E-05−0.02215
1.041.03631−0.0289021.0139E-046.0089E-05−0.01952
1.061.03673−0.0292851.0099E-046.1318E-05−0.01996
1.081.03712−0.0296479.9296E-056.2301E-05−0.02069
1.101.03748−0.0299899.0319E-056.3216E-05−0.01376
1.121.03782−0.0303137.9573E-056.3791E-05−0.00196
1.141.03813−0.0306197.5498E-056.3897E-05−0.00031
1.161.03842−0.0309096.8968E-056.3822E-050.00598
1.181.03869−0.0311855.8509E-056.3684E-050.01697
1.201.03894−0.0314474.8772E-056.3470E-050.02570
1.221.03917−0.0316963.2901E-056.3007E-050.04194
1.241.03939−0.0319331.2404E-056.2701E-050.06385
1.261.03959−0.0321593.6309E-066.1848E-050.07571
1.281.03977−0.032374−8.7629E-066.0920E-050.09238
1.301.03994−0.032580−2.1241E-056.0127E-050.10631
1.321.04010−0.032777−3.0514E-055.9249E-050.11781
1.341.04024−0.032964−2.9297E-055.8011E-050.11371
1.361.04037−0.033144−3.3371E-055.7140E-050.11496
1.381.04049−0.033316−3.7854E-055.6311E-050.12040
1.401.04060−0.033481−4.1242E-055.5537E-050.12424
1.421.04070−0.033640−5.3123E-055.5259E-050.13663
1.441.04079−0.033792−5.8179E-055.4880E-050.13999
1.461.04087−0.033937−6.3795E-055.4676E-050.14385
1.481.04094−0.034078−7.7351E-055.4931E-050.15702
1.501.04100−0.034213−7.9438E-055.4951E-050.15597
1.521.04105−0.034342−7.5133E-055.4890E-050.14783
1.541.04110−0.034467−8.1713E-055.5256E-050.15531
1.561.04114−0.034588−8.5841E-055.5630E-050.15995
1.581.04117−0.034704−9.0845E-055.6272E-050.16446
1.601.04119−0.034816−9.5996E-055.7006E-050.17040
1.621.04121−0.034924−9.8918E-055.7754E-050.17313
1.641.04122−0.035029−1.0948E-045.8854E-050.18528
1.661.04123−0.035129−1.1432E-045.9768E-050.19070
1.681.04123−0.035227−1.1436E-046.0619E-050.19002
1.701.04122−0.035321−1.1455E-046.1537E-050.18924
1.721.04121−0.035412−1.2130E-046.2654E-050.19369
1.741.04119−0.035501−1.3081E-046.3801E-050.20104
1.761.04117−0.035586−1.3241E-046.4561E-050.19974
1.781.04114−0.035669−1.3426E-046.5228E-050.20005
1.801.04110−0.035750−1.3518E-046.5717E-050.20005
1.821.04107−0.035827−1.3611E-046.6082E-050.19977
1.841.04102−0.035903−1.3860E-046.6489E-050.20090
1.861.04097−0.035976−1.3484E-046.6559E-050.19622
1.881.04092−0.036047−1.3227E-046.6681E-050.19258
1.901.04087−0.036116−1.3732E-046.6954E-050.19796
1.921.04080−0.036184−1.3091E-046.6696E-050.19124
1.941.04074−0.036249−1.4022E-046.6941E-050.20143
1.961.04067−0.036312−1.3775E-046.6675E-050.19923
1.981.04060−0.036374−1.3957E-046.6669E-050.20191
2.001.04052−0.036434−1.4856E-046.6798E-050.21268
2.021.04044−0.036492−1.5735E-046.6883E-050.22231
2.041.04035−0.036549−1.6724E-046.7063E-050.23250
2.061.04026−0.036604−1.7541E-046.7022E-050.24138
2.081.04017−0.036658−1.7157E-046.6580E-050.23720
2.101.04007−0.036710−1.6478E-046.5906E-050.23266
2.121.03997−0.036761−1.8219E-046.6174E-050.25118
2.141.03987−0.036810−1.8352E-046.5901E-050.25210
2.161.03976−0.036859−2.0232E-046.6165E-050.27297
2.181.03965−0.036906−2.0816E-046.6039E-050.28040
2.201.03954−0.036951−2.0831E-046.5786E-050.27968
2.221.03942−0.036996−2.1177E-046.5616E-050.28091
2.241.03930−0.037039−2.2081E-046.5460E-050.29134
2.261.03917−0.037082−2.2857E-046.5273E-050.30147
2.281.03904−0.037123−2.3363E-046.4912E-050.31003
2.301.03891−0.037163−2.3521E-046.4658E-050.31141
2.321.03878−0.037202−2.5433E-046.4820E-050.33080
2.3.2. Real Index of Refraction

[15] While spectral features in reflectance spectra of glasses in the UVVIS-NIR are dominated by variations in k, n, the real index of refraction also changes with both composition and wavelength, and affects the reflectance spectrum. Though in previous studies lacking alternatives n was assumed to be equal to the average visible refractive index [Hapke and Wells, 1981; Hapke, 1993; Lucey, 1998], its variation is significant in the UV, where we have shown above that titanium has its strongest influence. Therefore we have used the following steps to calculate n over the UVVIS-NIR spectral region for a glass of any iron and titanium composition (because data is available over such a limited wavelength range, calculations using dispersion theory are not possible).

[16] First, we used the equation of Church and Johnson [1980] that relates the refractive index of a silicate glass to its chemical composition

  • equation image

where WF denotes the weight fraction of the oxides present. The wavelength for which this refractive index equation is valid was not stated. However, the refractive index measurements were performed using immersion oils, whose standard refractive index is generally specified at the mercury e line (0.55 μm), thus we have designated this refractive index as n0.55. To calculate n0.55 from equation (3), we input the desired FeO and TiO2 weight fractions, and set the weight fractions of the remaining oxides equal to their average value as calculated from the lunar pyroclastic glass data of Delano [1986] (Table 2).

Table 2. Average Oxide Weight Fractions Used in Equation (3) as Calculated From Delano [1986] Data for Lunar Pyroclastic Glassesa
OxideAverage Weight Fraction
  • a

    Weight fractions were renormalized to sum to 1 with input iron and titanium fractions.

SiO20.4135
Al2O30.0696
Fe2O30.0000
MnO0.0026
MgO0.1543
CaO0.0627
Na2O0.0005
K2O0.0012

[17] To understand the variation of n with wavelength, we used the refractive index data of Ghosh [1998] for 32 commercial glasses between 0.4 and 2.3 μm. These glasses all exhibit an increase in refractive index with decreasing wavelength, with the sharpest increase between ∼0.4–0.7 μm (Figure 5a). The shape of the n spectrum is correlated with the value of n0.55: with increasing n0.55, the variation in n from the UV to the NIR increases (Figure 5b). To incorporate this into our model, so that we can use the n0.55 calculated with equation (3) to determine the overall n spectrum, we looked at the variation in n/n0.55 versus n0.55 at each wavelength. The correlation between n/n0.55 and n0.55 is high at all wavelengths; correlations range from 0.77 to 0.88, and are greater than 0.8 everywhere but wavelengths longer than ∼2.0 μm. At each wavelength, we fit a first-order polynomial to n0.55 versus n/n0.55. This provided us with a slope and offset at each wavelength that can be used to calculate the n spectrum depending on its n0.55 value

  • equation image

Slope and offset values are given in Table 1.

image

Figure 5. (a) Real indices of refraction (n) of synthetic glasses from Ghosh [1998]. (b) Real indices of refraction of glasses in Figure 5a normalized to 1 at 0.55 μm.

Download figure to PowerPoint

3. Results

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

3.1. Relative Effects of FeO and Particle Size

[18] To first order, FeO and particle size show similar effects; increasing FeO or particle size decreases reflectance and increases band depth, raising the possibility that these parameters cannot be separated. To study this problem, we first modeled each of the telescopic spectra as a series of pure glasses with fixed iron contents and particle sizes (0–30 wt% FeO, at increments of 0.25 wt% FeO, and 2–60 μm, at increments of 0.5 μm). For each of the approximately 14,000 combinations of FeO content and particle size, a gradient descent algorithm was applied, allowing titanium content, agglutinate abundance, and amount of SMFe to vary until the best possible spectral fit was achieved at that fixed combination of iron and particle size, and the quality of the fit was recorded. We defined the quality of the fit as the total of the sum of the difference between the telescopic and modeled spectra (to record the differences in absolute reflectance), plus the sum of the difference between the telescopic and modeled spectra divided by their mean (to emphasize how well the overall shape of the spectra matched).

[19] The results of the modeling show that iron content and particle size can indeed mimic each other to some extent. Of the best model fits to each of the telescopic spectra, it is seen that a decrease in iron content can be nearly compensated for with an increase in particle size and vice versa without significantly changing the goodness of the model fit (Figure 6). However, there is a clear minimum where a particular combination of particle size and iron best match the telescopic spectra. For Aristarchus this is an iron content of 21 wt% FeO and a particle size of 8 μm, for Humorum we find 20 wt% FeO and 6 μm, and for Sulpicius Gallus an iron content of 17 wt% FeO and a particle size of 6 μm fit best (Figure 7, Table 3). These iron contents are within the range of sampled pyroclastic glasses [Delano, 1986], but their particle sizes are lower than the mean value of sampled lunar pyroclastic glasses (40 μm by mass) [McKay et al., 1974]. However, the optical properties are dictated by the size fraction that with the largest cross-sectional area, not mass (see discussion below).

image

Figure 6. Goodness of fit of the model spectrum compared to the Aristarchus spectrum at each iron content/particle size combination. Lower numbers indicate a better fit. Results for Humorum and Sulpicius Gallus were similar.

Download figure to PowerPoint

image

Figure 7. The three telescopic spectra (as in Figure 1) and their best model fits.

Download figure to PowerPoint

Table 3. Model Parameters That Resulted in the Best Match to the Spectral Reflectance of the Three Regional Pyroclastic Deposits
 FeO, wt%Particle Size, μmTiO2, wt%Fraction GlassFraction AgglutinatesSMFea, Mass Fraction in Coating
  • a

    SMFe is submicroscopic metallic iron.

Aristarchus20.757.5−1.80.580.420.017
Humorum20.255.5−1.70.660.340.018
Sulpicius Gallus17.256.00.40.510.490.021

3.2. Titanium and the UV

[20] The titanium content was determined by selecting the best fit spectrum from the grid search described above, where TiO2 was allowed to vary at each point to match the telescopic spectra. The UV portion of the spectrum (<0.7 μm) is most strongly affected by TiO2 content (as evidenced by the high-TiO2 weighting factors at short wavelengths in Figure 3), but iron also affects this region to a lesser extent because of the Fe-Ti charge transfer. Thus the best fit for the TiO2 content is linked to the FeO content to some degree (Figure 8). However, a comparison of the goodness of fit in Figure 6 and the titanium content in Figure 8 shows that all of the best spectral fits have TiO2 contents that are within ∼1% of each other. The titanium abundances that produced the best-fit model spectra are −1.8 wt% TiO2 for the Aristarchus deposit, −1.7 wt% TiO2 for the Humorum pyroclastics, and 0.4 wt% TiO2 for the Sulpicius Gallus region (Table 3).

image

Figure 8. Modeled titanium content for the best model fits of the Aristarchus spectrum.

Download figure to PowerPoint

3.3. Crystallinity

[21] Previous studies have suggested that many of the regional pyroclastic deposits contain a significant portion of black beads [Adams et al., 1974; Pieters et al., 1974; Zisk et al., 1977; Weitz et al., 1998; Gaddis et al., 2003]. We therefore tested whether a mixture of glass and black beads would provide a better fit to the telescopic spectra. We used the reflectance spectrum collected by Adams et al. [1974] of black beads from sample 74001 as the devitrified component. We converted the black bead reflectance spectrum to single scattering albedo using Hapke theory. This allows the reflectance of a mixture of black beads and glass to be calculated from the sum of their single scattering albedos weighted by their relative abundances. We repeated our gradient descent fitting routine at the same fixed iron abundances and particle sizes, this time allowing the program to vary the abundance of black beads along with TiO2 content, SMFe, and agglutinate abundance. We found that when the abundance of black beads was allowed to vary in the gradient descent routine, the best fits for all three regions contained negative fractions of black beads.

3.4. Validation

[22] For comparison we also modeled the Apollo 15 green glass. A spectrum of the green glass in sample 15401 taken by Adams et al. [1974] is the best example of a pyroclastic glass without contamination by other materials. We fit this sample at its known iron and titanium contents (19.7 wt% FeO and 0.4 wt% TiO2), allowing particle size, SMFe, and agglutinates to vary to find the best spectral match of reflectance. We can match the spectrum reasonably well using the known iron and titanium concentration, except at wavelengths shorter than ∼0.7 μm, where the model predicts a lower reflectance than is observed (Figure 9). The best fit results in a particle size of 18.7 μm. The mean particle size of this sample is ∼62 μm by mass [Graf, 1993]. However, the size fraction that dominates the optical properties of a sample is the fraction that dominates the cross-sectional area [Hapke, 1981]. We normalized the measured size distribution to the cross-sectional areas of the size bins and find that the 16 μm size fraction is the peak of the size frequency distribution. Our result of 18.7 μm for the particle size of the green glass is consistent with this. The model fit contains no SMFe and a small fraction of agglutinates (20%). Sample 15401 is immature, and these low values for the space-weathering parameters suggest that the model is accurately determining maturity.

image

Figure 9. Reflectance of the green glass from sample 15401 [Adams et al., 1974] and a model spectrum produced at the iron and titanium contents of this sample (19.7 wt% FeO and 0.4 wt% TiO2), with a particle size of 18.7 μm.

Download figure to PowerPoint

[23] For an independent look at the titanium values of our three regions, we examined the TiO2 values derived from Lunar Prospector neutron spectrometer data [Elphic et al., 2002]. This data is of relatively low spatial resolution (0.5° × 0.5° bins), but the pyroclastic deposits are large enough that they dominate the field of view. The neutron spectrometer-derived titanium values for the three areas where the telescopic spectra were collected are all low: 0.5, 1.2, and 2.6 wt% TiO2 for Aristarchus, Humorum, and Sulpicius Gallus, respectively.

4. Discussion

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[24] The best observational and laboratory data available, coupled with our radiative transfer model, suggest that the Aristarchus, Humorum, and Sulpicius Gallus spectra represent iron-rich, low-titanium glass deposits. We find iron values (17–21 wt% FeO), which fall within the range of values indicated by glasses in the sample collection (16–24 wt% FeO) [Shearer and Papike, 1993]. The uncertainty of the titanium values is higher, as the model reflectances at short wavelengths, where effects of titanium are strongest (<0.7 μm), are the most in error. Our model spectrum of the Apollo 15 green glass composition (Figure 9) demonstrates that the model produces lower reflectance values than observed at short wavelengths. The model also predicts a lower reflectance than is observed at 0.415 μm for all three remote spectra, which resulted in slight negative values of TiO2 for two of the deposits. Because the reflectance in this region is controlled not just by titanium, but also by iron, we find that no combination of iron and titanium contents can provide an arbitrarily good fit to the reflectance at 0.415 μm, and the reflectance of the rest of the spectrum. The most likely cause for this error is that the optical constants and their variation with titanium content as calculated from the Bell et al. [1976] and Ghosh [1998] synthetic glass data are in error at short wavelengths. The imaginary index of refraction at 0.415 μm is based on just four synthetic glasses from Bell et al. [1976] compared with seven at all other wavelengths, increasing the uncertainty at that wavelength. For example, it is unlikely that the titanium weighting factor used to calculate k (Figure 3) actually decreases at 0.415 μm compared to longer wavelengths. Additional error could come from the ambiguities in matching telescopic observation photographs to select the correct region in the Clementine images to determine the 0.415 μm value. Despite these errors, the best-fit spectra produced by radiative transfer modeling suggest that the observed regions are composed of iron-rich low-titanium glasses (classified as very low or low Ti [Giguere et al., 2000]), a finding supported by the low-TiO2 values for the three regions from Lunar Prospector neutron spectrometer data [Elphic et al., 2002]. Other workers have also attempted to estimate the titanium content of the Aristarchus pyroclastics. This supports the suggestion of Gaddis et al. [2003] that the Aristarchus pyroclastics are lower in TiO2 than the Apollo 17 orange glass and the Pieters and Tompkins [2005] finding of low-Ti glass for the Aristarchus plateau on the basis of the shape of the 1 μm band. Our results indicate that the iron and titanium compositions of these deposits are similar to the Apollo 15 green glass. This is, however, are at odds with the suggestion of Weitz et al. [1998] that all of the observed regional pyroclastic deposits are high in titanium, because deposits with compositions similar to the green glass would have a high albedo and thus would not be detectable. Our modeling shows that the telescopic spectra for three deposits are consistent with a high-iron, low-titanium glass exposed to significant, but typical, degrees of space weathering. Orbital observations support space-weathering effects on pyroclastics. Lucchitta and Schmitt [1974] reported that small, fresh craters and steep slopes exposed material that was brighter (and visually more orange or red) than the dark surroundings in the Sulpicius Gallus pyroclastic deposit. They interpreted this to be a stratigraphic effect of black beads overlying glass, but it is equally plausible that these were areas of recently exposed immature pyroclastic material surrounded by dark mature material.

[25] The best model fits indicate that Aristarchus, Humorum, and Sulpicius Gallus have particle sizes of 6–8 μm. The mean size fraction of the Apollo 17 orange glass and black beads is 40 μm by mass, and is 62 μm by mass for the Apollo 15 green glass [McKay et al., 1974; Graf, 1993]. The size fraction that dominates the optical properties of these samples is ∼16 μm (determined by normalizing the particle size distribution to the cross-sectional area), approximately a factor of 2 larger than our modeled particle sizes. The sampled pyroclastics are among the most fine-grained soils in the sample collection, despite being immature. Mature pyroclastic soils like three remotely observed deposits would be expected to be dominated by an even smaller size fraction, as the evolution of particle sizes with maturity is well established [McKay et al., 1974]. However, it is also possible that there is a systematic particle size error in our modeling, as there has not been an objective test to determine with what fidelity the equations of Hapke [1981, 1993, 2001] relate reflectance and optical constants.

[26] The telescopic spectra are best matched by pure glass, without a significant amount of devitrified black beads. However, there are several important caveats to our modeling of crystallinity. First, the black beads we are mixing in are high iron and titanium (23 wt% FeO; 9 wt% TiO2), which might not be appropriate for these regions. Our modeling suggests these deposits are low in titanium, and a crystalline version of this glass would likely have optical properties significantly different from those of the black beads. Second, as we are not directly modeling the black beads, we cannot change their particle size or their degree of space weathering. Sample 74001 contains black beads with very little space weathering, and hence its spectrum shows much stronger absorptions than black beads that would be found on the surface in these regions. These factors, along with possible errors in the model, make it difficult to rule out a mixture of moderate titanium glass and black beads. It has been suggested that the Sulpicius Gallus region in particular could be a mixture of glass and devitrified black beads [Lucchitta and Schmitt, 1974; Zisk et al., 1977; Weitz et al., 1998], and a model spectrum with up to 5 wt% TiO2 glass mixed with 5 vol% black beads can provide a fit to the Sulpicius Gallus spectrum that is only marginally poorer than the black-bead-free fits. There is also the possibility that other model parameters are able to compensate for the fact that weathered black beads are not included in our model. For example, the best fits of the Sulpicius Gallus spectrum have the highest fraction of agglutinates and highest amount of SMFe among the three modeled deposits (Table 3). It is possible that the increase in these space-weathering-related parameters is due to the model attempting to make up for the lack of a weathered black bead component. These different possibilities imply that on the basis of the spectrum alone, Sulpicius Gallus could be either pure glass or have a small component of black beads mixed in.

[27] Our modeling also enables us to address previous efforts to determine pyroclastic compositions with spectral ratios. Zisk et al. [1977] suggested that on the basis of the 0.61/0.37 μm ratio, the Sulpicius Gallus region is a mixture of 1 part orange glass to 4 parts black beads, but our modeling indicates that any component of black beads is significantly smaller. They also estimated the Aristarchus deposit as composed of 1 part orange glass and 2 parts black beads, a finding which is not supported by our results of a low-Ti pure glass for this region. Gaddis et al. [2003] plotted the 0.95/0.75 μm ratio versus the 0.415/0.750 μm ratio from Clementine data for 75 pyroclastic deposits in order to interpret iron, titanium, and crystallinity variations. We find that model glass spectra that differ only in their iron and titanium content can cover nearly the entire spectral space occupied by the lunar pyroclastics on this plot (Figure 10), and any variations in particle size or space weathering between deposits would only obfuscate possible trends due to composition or crystallinity. Weitz et al. [1998] proposed that these two ratios can be used to determine the relative crystallinity of lunar pyroclastic deposits, assuming iron and titanium content is constant. However, if iron or titanium content varies between deposits, which is almost certain on the basis of variation in glass compositions in the lunar samples, any trend due to crystallinity would be obscured.

image

Figure 10. Model ratios for glasses of 13–26 wt% FeO and 0–18 wt% TiO2 (the approximate range of sampled glasses). Shaded area shows approximately where Gaddis et al. [2003] pyroclastic deposits plot.

Download figure to PowerPoint

[28] This study shows the utility of using radiative transfer modeling to determine compositions of lunar glasses from hyperspectral data. It also shows the need for high-precision, high spectral resolution data. Because of the spectral interplay of composition, particle size, and maturity, it is difficult to determine composition with high levels of confidence without this kind of high-quality data. High spatial resolution spectra would also significantly improve our ability to determine composition, by allowing the targeting of small, fresh craters. Because space weathering decreases the strength of absorption bands and reduces the differences between spectra of different compositions, spectra of immature surfaces would increase the ability to confidently determine composition. Spectrometers with UVVIS and NIR coverage on future lunar missions, such as the spectral profiler aboard the Japanese Aerospace Exploration Agency's (JAXA) SELENE mission and the Moon Mineralogy Mapper on the Indian Space Research Organisation's (ISRO) Chandrayaan-I mission, should greatly enhance the presently limited reservoir of lunar hyperspectral data. Reflectance data from both of these spectrometers down to 0.4 μm, along with improved measurements of the optical constants of glass will be important for refining calculations of the compositions of pyroclastic deposits.

5. Conclusions

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[29] 1. We have successfully used radiative transfer modeling with measured optical constants to produce model spectra that match telescopic spectra of the Aristarchus, Humorum, and Sulpicius Gallus lunar pyroclastic deposits.

[30] 2. We find that model spectra of pure glass (as opposed to devitrified black beads) provide the best matches to all three spectra.

[31] 3. We find iron contents of 21, 20, and 17 wt% FeO for Aristarchus, Humorum, and Sulpicius Gallus, respectively.

[32] 4. Radiative transfer modeling suggests that all three regions are low in titanium, a finding supported by Lunar Prospector neutron spectrometer TiO2 values.

[33] 5. The best match to the Sulpicius Gallus spectrum, a region previously interpreted to be a mixture of glass and black beads, is pure low-Ti glass. However, a moderate Ti glass mixed with black beads provides a fit that is only marginally poorer, and cannot be ruled out.

[34] 6. Multispectral data cannot be used to definitively distinguish variations in glass composition from variations in crystallinity.

[35] 7. New high-precision, high spectral and spatial resolution data expected from upcoming missions will allow improved modeling of pyroclastic materials, especially if immature deposits like those observed by Lucchitta and Schmitt [1974] can be resolved.

[36] 8. Improved measurements of optical constants and their variation with composition would enhance future studies. This is especially true at short wavelengths (<0.7 μm) where there appears to be the largest error.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information

[37] We would like to thank Lin Li and Gary Hansen for thoughtful and thorough reviews, which improved this manuscript. This work was supported in part by NASA Planetary Geology and Geophysics grant NNG05GJ51G and grant NNG05GM94G. This is HIGP publication 1438 and SOEST publication 6765.

References

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information
  • Adams, J. B., and A. Filice (1967), Spectral reflectance 0.4 to 2.0 microns of silicate rock powders, J. Geophys. Res., 72, 57055715.
  • Adams, J. B., C. M. Pieters, and T. B. McCord (1974), Orange glass: Evidence for regional deposits of pyroclastic origin on the Moon, Proc. Lunar Sci. Conf., 5th, 171186.
  • Bell, P. M., H. K. Mao, and R. A. Weeks (1976), Optical spectra and electron paramagnetic resonance of lunar and synthetic glasses: A study of the effects of controlled atmosphere, composition, and temperature, Proc. Lunar Sci. Conf., 7th, 25432559.
  • Britt, D. T., and C. M. Pieters (1994), Darkening in black and gas-rich ordinary chondrites: The spectral effects of opaque morphology and distribution, Geochim. Cosmochim. Acta, 58, 39053919.
  • Burns, R. G., K. M. Parkin, and B. M. Loeffler (1976), Further characterization of spectral features attributable to titanium on the Moon, Proc. Lunar Sci. Conf., 7th, 25612578.
  • Charette, M. P., T. B. McCord, C. M. Pieters, and J. B. Adams (1974), Application of remote spectral reflectance measurements to lunar geology classification and determination of titanium content of lunar soils, J. Geophys. Res., 79, 16051613.
  • Church, B. N., and W. M. Johnson (1980), Calculation of the refractive index of silicate glasses from chemical composition, Geol. Soc. Am. Bull., 91, 619625.
  • Coombs, C. R., B. R. Hawke, P. G. Lucey, P. D. Owensby, and S. Zisk (1990), The Alphonsus region: A geological and remote-sensing perspective, Proc. Lunar Planet. Sci. Conf. 20th, 161174.
  • Delano, J. W. (1986), Pristine lunar glasses: Criteria, data, and implications, Proc. Lunar Planet. Sci. Conf., 16th, Part 2, J. Geophys. Res., 91, suppl., D201D213.
  • Elphic, R. C., D. J. Lawrence, W. C. Feldman, B. L. Barraclough, S. Maurice, P. G. Lucey, D. T. Blewett, and A. B. Binder (2002), The Lunar Prospector neutron spectrometer constraints on TiO2, J. Geophys. Res., 107 (E4), 5024, doi:10.1029/2000JE001460.
  • Fischer, E. M., and C. M. Pieters (1994), Remote determination of exposure degree and iron concentration of lunar soils using VIS-NIR spectroscopic methods, Icarus, 111, 475488.
  • Gaddis, L. R., C. M. Pieters, and B. R. Hawke (1985), Remote-sensing of lunar pyroclastic mantling deposits, Icarus, 61, 461489.
  • Gaddis, L. R., M. I. Staid, J. A. Tyburczy, and B. R. Hawke (2003), Compositional analyses of lunar pyroclastic deposits, Icarus, 161, 262280.
  • Ghosh, G. (1998), Handbook of Thermo-Optic Coefficients of Optical Materials with Applications, 325 pp., Elsevier, New York.
  • Giguere, T. A., G. J. Taylor, B. R. Hawke, and P. G. Lucey (2000), The titanium contents of lunar mare basalts, Meteorit. Planet. Sci., 35, 193200.
  • Graf, J. C. (1993), Lunar Soils Grain Size Catalog, NASA Ref. Publ. 1265, 466 pp., Natl. Aeronaut. and Space Admin., Washington, D. C.
  • Hapke, B. (1981), Bidirectional spectroscopy: 1. Theory, J. Geophys. Res., 86, 30393054.
  • Hapke, B. (1993), Theory of Reflectance and Emittance Spectroscopy, 455 pp., Cambridge Univ. Press, New York.
  • Hapke, B. (2001), Space weathering from Mercury to the asteroid belt, J. Geophys. Res., 106, 10,03910,073.
  • Hapke, B., and E. Wells (1981), Bidirectional spectroscopy: 2. Experiments and observations, J. Geophys. Res., 86, 30553060.
  • Hawke, B. R., P. G. Lucey, and T. B. McCord (1983), Spectral studies of the Aristarchus region: Implications for the composition of the lunar crust, Lunar Planet. Sci., XIV, 289290.
  • Hawke, B. R., C. R. Coombs, L. R. Gaddis, P. G. Lucey, and P. D. Owensby (1989), Remote sensing and geologic studies of localized dark mantle deposits on the Moon, Proc. Lunar Planet. Sci. Conf. 19th, 255268.
  • Hawke, B. R., D. MacLaskey, T. B. McCord, J. B. Adams, J. W. Head, C. M. Pieters, and S. H. Zisk (1979), Multispectral mapping of the Apollo 15 Apennine region: The identification and distribution of regional pyroclastic deposits, Proc. Lunar Planet. Sci. Conf. 10th, 29953015.
  • Head, J. W., and L. Wilson (1979), Alphonsus-type dark halo craters: Morphology, morphometry and eruption conditions, Proc. Lunar Planet. Sci. Conf. 10th, 28612897.
  • Keller, L., and D. McKay (1993), Discovery of vapor deposits in the lunar regolith, Science, 261, 13051307.
  • Keller, L., and D. McKay (1997), The nature and origin of rims on lunar soil grains, Geochim. Cosmochim. Acta, 61, 23312340.
  • Lucchitta, B. K., and H. H. Schmitt (1974), Orange material in the Sulpicius Gallus formation at the southwestern edge of Mare Serenitatis, Proc. Lunar Planet. Sci. Conf. 5th, 223234.
  • Lucey, P. G. (1998), Model near-infrared optical constants of olivine and pyroxene as a function of iron content, J. Geophys. Res., 103, 17031714.
  • Lucey, P. G., B. R. Hawke, C. M. Pieters, J. W. Head III, and T. B. McCord (1986), A compositional study of the Aristarchus region of the Moon using near-infrared reflectance spectroscopy, J. Geophys. Res., 91, 344354.
  • McKay, D. S., R. M. Fruland, and G. H. Heiken (1974), Grain size and evolution of lunar soils, Proc. Lunar Planet. Sci. Conf. 5th, 887906.
  • Noble, S. K., C. M. Pieters, L. A. Taylor, R. V. Morris, C. C. Allen, D. S. McKay, and L. P. Keller (2001), The optical properties of the finest fraction of lunar soil: Implications for space weathering, Meteorit. Planet. Sci., 36, 3142.
  • Papike, J. J., G. Ryder, and C. K. Shearer (1998), Lunar samples, in Planetary Materials, edited by J. J. Papike, pp. 5.15.234, Mineral. Soc. of Am., Washington, D. C.
  • Pieters, C. M., and S. Tompkins (2005), Remote sensing of lunar mineralogy: The glass conundrum, Lunar Planet. Sci., XXXVI, abstract 1346.
  • Pieters, C. M., T. B. McCord, S. H. Zisk, and J. B. Adams (1973), Lunar black spots and the nature of the Apollo 17 landing area, J. Geophys. Res., 78, 58675875.
  • Pieters, C. M., T. B. McCord, M. P. Charette, and J. B. Adams (1974), Lunar surface: Identification of the dark mantling material in the Apollo 17 soil samples, Science, 183, 11911194.
  • Pieters, C. M., E. M. Fischer, O. Rode, and A. Basu (1993), Optical effects of space weathering: The role of the finest fraction, J. Geophys. Res., 98, 20,81720,824.
  • Pieters, C. M., L. Taylor, S. Noble, L. Keller, B. Hapke, R. Morris, C. Allen, D. McKay, and S. Wentworth (2000), Space weathering on airless bodies: Resolving a mystery with lunar samples, Meteorit. Planet. Sci., 35, 11011107.
  • Shearer, C. K., and J. J. Papike (1993), Basaltic magmatism on the Moon: A perspective from volcanic picritic glass beads, Geochim. Cosmochim. Acta, 57, 47854812.
  • Weitz, C. M., J. W. Head, and C. M. Pieters (1998), Lunar regional dark mantle deposits: Geologic, multispectral, and modeling studies, J. Geophys. Res., 103, 22,72522,759.
  • Wells, E., and B. Hapke (1977), Lunar soil: Iron and titanium bands in the glass fraction, Science, 195, 977979.
  • Wilson, L., and J. W. Head III (1981), Ascent and eruption of basaltic magma on the Earth and Moon, J. Geophys. Res., 86, 29713001.
  • Zisk, S. H., G. H. Pettengill, and G. W. Catuna (1974), High resolution radar map of the lunar surface at 3.8-cm wavelength, Moon, 10, 1750.
  • Zisk, S. H., C. A. Hodges, H. J. Moore, R. W. Shorthill, T. W. Thompson, E. A. Whitaker, and D. E. Wilhelms (1977), The Aristarchus-Harbinger region of the Moon: Surface geology and history from recent remote sensing observations, Moon, 17, 5999.

Supporting Information

  1. Top of page
  2. Abstract
  3. 1. Introduction
  4. 2. Methods
  5. 3. Results
  6. 4. Discussion
  7. 5. Conclusions
  8. Acknowledgments
  9. References
  10. Supporting Information
FilenameFormatSizeDescription
jgre2175-sup-0001-t01.txtplain text document5KTab-delimited Table 1.
jgre2175-sup-0002-t02.txtplain text document0KTab-delimited Table 2.
jgre2175-sup-0003-t03.txtplain text document0KTab-delimited Table 3.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.