Lunar mare deposits associated with the Orientale impact basin: New insights into mineralogy, history, mode of emplacement, and relation to Orientale Basin evolution from Moon Mineralogy Mapper (M3) data from Chandrayaan-1
 Moon Mineralogy Mapper (M3) image and spectral reflectance data are combined to analyze mare basalt units in and adjacent to the Orientale multiring impact basin. Models are assessed for the relationships between basin formation and mare basalt emplacement. Mare basalt emplacement on the western nearside limb began prior to the Orientale event as evidenced by the presence of cryptomaria. The earliest post-Orientale-event mare basalt emplacement occurred in the center of the basin (Mare Orientale) and postdated the formation of the Orientale Basin by about 60–100 Ma. Over the next several hundred million years, basalt patches were emplaced first along the base of the Outer Rook ring (Lacus Veris) and then along the base of the Cordillera ring (Lacus Autumni), with some overlap in ages. The latest basalt patches are as young as some of the youngest basalt deposits on the lunar nearside. M3 data show several previously undetected mare patches on the southwestern margins of the basin interior. Regardless, the previously documented increase in mare abundance from the southwest toward the northeast is still prominent. We attribute this to crustal and lithospheric trends moving from the farside to the nearside, with correspondingly shallower density and thermal barriers to basaltic magma ascent and eruption toward the nearside. The wide range of model ages for Orientale mare deposits (3.70–1.66 Ga) mirrors the range of nearside mare ages, indicating that the small amount of mare fill in Orientale is not due to early cessation of mare emplacement but rather to limited volumes of extrusion for each phase during the entire period of nearside mare basalt volcanism. This suggests that nearside and farside source regions may be similar but that other factors, such as thermal and crustal thickness barriers to magma ascent and eruption, may be determining the abundance of surface deposits on the limbs and farside. The sequence, timing, and elevation of mare basalt deposits suggest that regional basin-related stresses exerted control on their distribution. Our analysis clearly shows that Orientale serves as an excellent example of the early stages of the filling of impact basins with mare basalt.
 Despite significant advances in all of these areas, several of the most fundamental questions in mare basalt petrogenesis and emplacement remain unresolved [e.g., Jolliff et al., 2006; Jolliff, 2008]. Mare basalts are known to occur preferentially on the lunar nearside, where both the crust [e.g., Ishihara et al., 2009] and lithosphere [Solomon and Head, 1980] are thinnest, and to be concentrated within and adjacent to impact basins, such as Imbrium, Serenitatis, Crisium and Humorum. The following are among the outstanding questions.
 2. What is the influence of crustal thickness and lithospheric thickness on the emplacement of mare basalts? Are mare basalts inhibited from ascending to the surface by buoyancy traps at the base of the crust or lithosphere [e.g., Head and Wilson, 1992]? Or are there other reasons for the differences in the nearside/farside distribution of mare basalts unrelated to these factors, such as fundamental differences in deep thermal structure and thus mare basalt genesis [e.g., Zhong et al., 2000]?
 Recent acquisition of high spatial and spectral resolution spectrometer data for the Orientale region of the Moon by the Moon Mineralogy Mapper (M3) instrument on the Chandrayaan-1 mission permits further characterization of mare basalt deposits in this region, and enables an analysis of their mineralogy, ages, affinities, and their relationships to Orientale impact basin deposits, topography and evolution. In this analysis, we utilize a range of M3 and related data (e.g., Lunar Orbiting Laser Altimeter, LOLA) to address these outstanding questions concerning (1) the relation of mare basalt origin and emplacement to impact basins and (2) crustal/lithospheric thickness differences.
 The purpose of this paper is to investigate the characteristics of volcanic features and deposits in the Orientale Basin using the latest available data, including (1) the ages of the various volcanic deposits, the melt sheet, and the basin event itself; (2) the areal extent and volume of the identified mare deposits; (3) the spectral characteristic and compositional implications of Mare Orientale and other volcanic deposits; (4) their altimetry, including both elevation and local or regional tilts; and (5) the volcanic features associated with the basin, including sinuous rilles, domes, and vents. Combining all of this new information can assist in further defining the relationship between large impact basins and volcanism on the lunar surface in a nearside/farside transition region.
 We begin with a description of the new M3 data and other data and techniques in order to provide background on how we have utilized these data to bring new insight into the problems outlined above.
2. The Moon Mineralogy Mapper (M3) Experiment, Related Data, and Analysis
2.1. Background: The M3 Instrument
 Within the last several years, petabytes of new data have been returned to Earth from instruments on several lunar orbiters (e.g., Kaguya, Chang' E, Chandrayaan-1, Lunar Reconnaissance Orbiter). Among these instruments is the Moon Mineralogy Mapper (M3), a NASA imaging spectrometer onboard the Indian Chandrayaan-1 spacecraft. M3 was designed to operate in two modes, a global mode and a targeted mode. The data used in this paper were obtained in global mode, which collected reflectance spectra in 85 bands, ranging from 400 to 3000 nm with a spatial resolution of 140 m/pixel from a 100 km orbit [e.g., Green et al., 2007; Pieters et al., 2009]. Data from the M3 instrument were used in crater counting, measurement of the area and volume of old and newly identified mare ponds, determination of the length and width of sinuous rilles, and the definition of the spectral composition of Mare Orientale and related deposits.
 The Lunar Orbiting Laser Altimeter (LOLA), one of the instruments carried by the NASA Lunar Reconnaissance Orbiter (LRO), provides a precise global lunar topographic model and geodetic grid by pulsing a laser instrument during orbit [Zuber et al., 2010; Smith et al., 2010]. A 64 pixel per degree (∼473 m/pixel) global gridded data set was used for basin profiles and mare pond elevation analyses in this study. In addition, LOLA profile data were used to determine the widths and depths of the identified sinuous rilles.
2.2.2. Impact Crater Size-Frequency Distributions
 Crater counting was conducted using a M3 mosaic of Orientale Basin at ∼140 m/pixel [Boardman et al., 2011], as well as the CraterStats program [Michael and Neukum, 2010]. The chronology and production functions of Neukum et al.  were used to calculate the ages from analysis of the new M3 data; crater ages were obtained by plotting unbinned data. Specific count regions of similarly sized area were defined for both the Maunder Formation (melt sheet deposit) and the Hevelius Formation (basin ejecta deposit) to obtain model ages for the basin-forming event. Each count area for the Maunder Formation was between 1,000 and 3,000 km2 and count areas for the Hevelius Formation ranged from 190,000 to 250,000 km2. For dating the basin-forming event only primary craters larger than 5 km were counted, whereas for the mare ponds primary craters greater than ∼0.75 km in diameter were counted. Only if the crater population above 0.75 km was too small to give statistically significant ages were craters from 0.5 km in diameter considered. We found no obvious flow fronts and no distinct differences in mare composition in the individual mare ponds in Lacus Veris and Lacus Autumni; thus, we assume that individual mare ponds are the result of single eruptive events [Yingst and Head, 1997, 1999]. There is some indication of subtle differences within individual mare pond composition, which could either be due to lateral and vertical mixing of feldspathic material with mare soils or actual mineralogical differences [Kadel, 1993; Staid, 2000]. Further investigation is needed to assess these trends in mixing and apparent mineralogical differences.
2.2.3. Spectral Properties and Background on Lunar Near-Infrared Reflectance Spectroscopy
 The mineralogic composition of the lunar surface has been investigated for decades using the techniques of visible, near-infrared (VNIR) spectroscopy, owing to the abundance of diagnostic mineral absorptions in this region of the electromagnetic spectrum [e.g., McCord et al., 1981; Pieters et al., 1996; Ohtake et al., 2009; Isaacson et al., 2011]. These diagnostic absorptions in the VNIR are the result of molecular electronic transitions that are specific to crystallographic sites in mineral crystals. The characteristics of these crystallographic sites vary between different minerals and mineral compositions [Burns, 1993]. Thus, the shape and location of mineral absorptions are indicative of composition. The two most common mafic minerals on planetary surfaces are olivine and pyroxene, both having diagnostic absorptions that vary with composition. Pyroxenes have two diagnostic absorptions near 1 and 2 μm [e.g., Adams, 1974; Cloutis and Gaffey, 1991], and olivines have a combination absorption feature near 1 μm, composed of three separate overlapping mineral absorptions [e.g., Burns, 1970, 1974; Sunshine and Pieters, 1998].
 Oftentimes the continuum slope of spectra are described as either “blue” or “red,” referring to the slope of the visible part of the electromagnetic spectrum. Blue light is the shortest wavelength and red light is manifested at longer wavelengths. Thus, when a spectrum is described as “blue” that means the shorter wavelength region of the electromagnetic spectrum is dominating and the spectrum has a flat slope; “red” refers to positive continuum slopes controlled by the longer wavelength regions of the electromagnetic spectrum. This continuum slope is controlled by composition of a geologic material, whether it is mixed with another spectrally distinct geologic material, and the maturity or degree of space weathering experienced by the surface and the composition of the material [e.g., Staid, 2000; Staid and Pieters, 2000].
 Approximately 120 M3 spectra from Mare Orientale were collected to characterize basalt compositions. Optical period 1B (OP1B) [Boardman et al., 2011] data strips from two regions of Mare Orientale, previously identified as uncontaminated with feldspathic basin material [Staid, 2000], were utilized for data collection. To avoid complications resulting from highland contamination, spectra were collected from craters < ∼1 km in diameter that showed an immature signature in a 0.989 μm/0.750 μm ratio parameter map. Craters have an excavation depth that is approximately 1/10th of the diameter [Melosh, 1989]; thus, collecting spectra from craters < ∼1 km in diameter ensures that only the surface deposit is being sampled. The 0.989 μm band, located at approximately the center of one of the main absorptions for mafic minerals, and the 0.750 μm band, located outside of the absorption on the continuum part of a spectrum, provide a good indication of the strength of the 1 μm absorption in mafic minerals. The closer the value of the ratio is to 1, the weaker the mineral absorption.
 It was necessary to take spectral measurements of the more freshly exposed areas to avoid any effects of maturity or space weathering, such as decreased depth of mineral absorption bands, increased red slope of the spectra, and a decrease the overall albedo of the surface [e.g., Pieters et al., 2000; Anand et al., 2004; Noble et al., 2007]. Calibration of the data is explained in detail by R. O. Green et al. (The Moon Mineralogy Mapper (M3) imaging spectrometer for lunar science: Instrument, calibration, and on-orbit measurement performance, submitted to Journal of Geophysical Research, 2011). Separate from general corrections, a mild spectral correction was applied to the data in order to suppress residual artifacts in M3 data. A detailed explanation of this correction is provided by Isaacson et al. .
 In this analysis, only the 1.0 μm absorption band was investigated due to factors, such as thermal effects, becoming increasingly important to the 2.0 μm absorption, which is in the process of being resolved in the ongoing calibration of the new M3 data. Features of the 1 μm absorption that were studied include the band center and the band strength. It has been shown in numerous studies that absorptions centered around 1.0 μm and 2.0 μm shift to longer wavelengths as Fe2+ and Ca2+ substitute for Mg2+ in pyroxenes [e.g., Adams, 1974; Cloutis and Gaffey, 1991; Pieters et al., 1996; Klima et al., 2007]. Measurements of the position of the 1 μm absorption band center were made using the methods of Cloutis and Gaffey . In these calculations, it was assumed that the absorptions were the result of a single electronic transition absorption, where spectroscopic mineral absorptions are actually composed of several overlapping electronic transition absorption bands. This more basic approach was used in place of the modified Gaussian model (MGM), which can derive much more specific compositional information by deconvolving individual absorption bands [e.g., Sunshine et al., 1990; Pieters et al., 1996; Klima et al., 2007], in order to reach initial conclusions about the composition of basalts in Mare Orientale.
3. Review of Models of Mare Basalt Genesis Related to Impact Basin Formation/Evolution and Global Crustal Thickness Variations
 Early in the modern study of the Moon, basic models for the formation of mare basalts were postulated that involved internal radiogenic heat sources and sequential partial melting of a layered mantle [e.g., Runcorn, 1974; Solomon, 1975]. Subsequent petrological evidence suggested a wide range of source depths for mare basalts. The documentation of increasing complexities in the ages and petrogenesis of mare basalts led to the proposal of several major new ideas for their generation (see summary by Shearer et al. ). Models for the generation, ascent and eruption of mare basalts (as observed in the distribution of mare deposits) can be classified into five types, as follows.
3.1. Model 1: Nearside/Farside Mare Basalt Asymmetry Due to Crustal Thickness Differences
 One of the more fundamental characteristics of the Moon is the nearside-farside asymmetry in the distribution of mare basalts [e.g., Head, 1976]. Early models accounted for this by calling on differences in nearside-farside crustal thickness [e.g., Solomon, 1975; Head and Wilson, 1992]: for globally heterogeneous mare basalt source regions at depth, dikes rising to a constant level were much more likely to erupt onto the surface in the thinner nearside crust than dikes rising in a thicker farside crust. The general paucity of farside mare basalts and the concentration of most farside basalts in the relatively deep South Pole–Aitken basin seemed consistent with this hypothesis.
 Global altimetry obtained by the Clementine mission [e.g., Zuber et al., 1994; Neumann et al., 1996] revealed, however, that the depth of the South Pole–Aitken basin was comparable to mare basalt elevations in basins on the nearside, and yet the SPA basin was not as extensively flooded by mare basalts as the nearside basins [e.g., Head et al., 1993; Greeley et al., 1993; Yingst and Head, 1997, 1999; Pieters et al., 2001]. These observations cast doubt on the assumption of global mare basalt source symmetry and the role of crustal thickness in explaining nearside-farside mare basalt asymmetry, and pointed out the necessity of adding additional factors to models of magma ascent and eruption [e.g., Wieczorek et al., 2001]. This led to the development of several alternative models for the emplacement of basalts in the South Pole–Aitken basin and for the Moon as a whole, raising the questions (1) were there fundamental differences in the nearside-farside source regions established early in lunar history [e.g., Zhong et al., 2000; Parmentier et al., 2002]; (2) did the Procellarum KREEP Terrain (PKT) play a fundamental role in mare basalt generation [e.g., Wieczorek and Phillips, 2000]; (3) was there a basic difference in the nearside-farside thermal gradient that influenced both basin relaxation and mare basalt generation, ascent and eruption [e.g., Solomon and Head, 1980]; and (4) could the formation of a basin the size of SPA have induced sufficient convection to have stripped away a subsurface KREEP layer, and thus to have inhibited the formation of mare basalts below the basin [e.g., Arkani-Hamed and Pentecost, 2001]?
 Model 1a (correlation with crustal/lithospheric thickness) thus predicts a close correlation between these factors, and model 1b (no correlation) predicts not only no correlation, but differences in the nature and distribution of mare basalts in space and time due to nearside-farside mantle source region differences (Table 1).
Table 1. Summary of the Various Models of Mare Basalt Genesis
Nearside/farside asymmetry due to crustal thickness differences
1a) A close correlation exists between crustal thickness and distribution of mare basalts. 1b) No correlation exists between these two variables, instead differences in the nature and distribution are due to nearside/farside mantle source region differences.
Pressure release melting and associated secondary convection
Large quantities in situ pressure release melting are produced instantaneously and located in the basin center. Smaller quantities of basalts are produced as a result of adiabatic melting induced by convection and persist for extended periods of time after impact (i.e., 350 Ma). The lowest-titanium content mare basalts erupt last.
Enhanced KREEP layer induces volcanism in the Western nearside Procellarum KREEP Terrain (PKT)
The presence of KREEP and possible derivatives (i.e., "red spot" volcanism such as the Gruithuisen Domes) might control the distribution of mare basalts.
Large-scale overturn causes eruption of mare basalts
High-titanium basalts erupt earliest, followed by a suite of other types of basalts. Model is independent of basin formation.
Mare basalt emplacement is related to global thermal evolution and basin evolution
Early mare volcanism is located in the center of the basin, and the youngest deposits can be found on the outside rim of the basin. This model is not necessarily related to basin-forming impacts.
3.2. Model 2: Impact Basin Pressure-Release Melting and Associated Secondary Convection
 Pressure release melting is known to be an important mechanism for basalt generation, but the lunar pressure gradient, combined with the composition of the crust and the apparent depth of origin of mare basalts, has led to this mechanism generally not being favored for mare basalt formation. Recently, Elkins-Tanton et al.  have reassessed the magmatic effects of large basin formation, and introduced a two-stage model for melt creation beneath lunar basins triggered by basin formation itself. In the initial stage, crater excavation depressurizes and uplifts underlying mantle material so that it melts in situ instantaneously, forming large quantities of melt below the basin (in addition to impact melt in the cavity). This model thus predicts huge quantities of in situ pressure release melt (98–100% of the melt created) produced instantaneously and available to be extruded into the impact basin as lunar mare basalts.
 In the second stage, the cratered region rises isostatically, warping isotherms upward and inducing convection, at which time adiabatic melting can occur. This second stage is miniscule in terms of melt produced (1–2% of the total) but can last for a longer period of time, up to ∼350 Ma. In the Elkins-Tanton et al.  model, mafic mantle melts can be generated from depths of 150–560 km, depending on mantle potential temperature. Assuming that 10% of the melt generated erupts, Elkins-Tanton et al. [2004, Figure 5] find that the volumes of magma reported for basins are similar to their predictions. Elkins-Tanton et al.  also model the origin and emplacement of high-alumina, high-TiO2, KREEP-rich, and picritic magmas and predict an order of eruption, with the most primitive, lowest-titanium magmas last.
 In a more recent treatment of impact-induced convection, Ghods and Arkani-Hamed  used a suite of numerical models to show that this mechanism might be able to account for the formation of mare basalts, the range of depths of their source regions, the observed delay between impact basin formation and initiation of basaltic volcanism, and the long duration of emplacement of mare basalts. These models treat basins of different sizes (e.g., ranging from Orientale, through Imbrium, to South Pole–Aitken), make different predictions about the record of mare basalt emplacement for each, and can be readily tested against the geological record of mare basalt volcanism.
 Models for pressure release melting predict initially large volumes of mare basalts in the basin center, followed by smaller amounts emplaced over several hundred million years, and the lowest-Ti content lavas last (Table 1). Key tests thus involve the timing, duration, volumes, styles and the comparison of the record in different sized basins.
3.3. Model 3: Enhanced Sub-Procellarum KREEP Layer
 Although clearly outside the Orientale Basin region, more than 60% of the mare basalts by area occur within the boundaries of the Procellarum KREEP Terrain (PKT), which makes up only ∼16% of the surface of the Moon. This observation was one of the major factors that led Wieczorek and Phillips  to propose that there was a cause and effect in terms of the occurrence of KREEP and the generation of mare basalts. They postulated that the enhancement of heat-producing elements implied by the elevated KREEP levels significantly influenced the thermal evolution of the region, causing the underlying mantle to partially melt over much of lunar history to generate the observed basaltic volcanic sequence. The thermal model of Wieczorek and Phillips  predicts that melting occurs only directly below the PKT. Partial melting of the mantle begins immediately after the model is started at 4.5 Ga and continues to a lesser degree to the present. Melting initiates immediately beneath the KREEP basalt layer and becomes deeper with time, with the maximum depth of melting being ∼600 km, and the KREEP layer is kept above its solidus for most of lunar history.
 Does this hypothesis account for the origin of mare basalts in terms of the timing, duration, areal distribution, volumes, and changes in depth with time? If not, is it related to any other volcanism within the PKT region, such as the “red spot” volcanic domes and related deposits? Hess and Parmentier  pointed out several difficulties with this model in terms of (1) petrogenetic evolution, (2) geophysical evidence against the long-term duration of a near-liquid KREEP layer, and (3) the fact that such a layer might form an impenetrable barrier to the eruption of mare basalts (e.g., how can denser mare basalt liquids penetrate through a >30 km thick layer of less dense, KREEP-rich liquid?). This model predicts (Table 1) that the presence of KREEP and possible derivatives, such as “red spots,” might control the distribution of mare basalts, and thus we examine the new data for such evidence.
3.4. Model 4: Large-Scale Overturn of Initial Unstable Stratification
 In model 4, crystallization of the lunar magma ocean (LMO) forms a chemically stratified lunar interior with an anorthositic crust separated from the primitive lunar interior by magma ocean cumulates. Dense, ilmenite-rich cumulates with high concentrations of incompatible radioactive elements are the last magma ocean cumulates to form, and underlying olivine-orthopyroxene cumulates are also stratified with later crystallized, denser, more Fe-rich compositions at the top. These layers are gravitationally unstable. Rayleigh-Taylor instabilities cause the dense cumulates to sink toward the center of the Moon and to form a dense core [Hess and Parmentier, 1995]. Subsequently, the ilmenite-rich cumulate core undergoes radioactive heating and this heats the overlying mantle, causing melting. The source region for high-TiO2 basalts is thus envisioned to be a mixed zone above the core-mantle boundary containing variable amounts of ilmenite and KREEP and involves deep, high-pressure melting, delayed for a period of time subsequent to LMO formation and overturn. Thermal plumes rise into chemically stratified surroundings of the mantle (chemically less dense but colder) above the core and cause mixing and homogenization. The resulting lower thermal boundary layer may be partially to wholly molten depending on mineralogy and the range of input parameters.
 Melting at the top of the mixed layer to produce mare basalt magmas must occur at low enough pressure for melt buoyancy and at high enough pressure to satisfy the depth indicated by phase equilibria. The onset time of mare volcanism is constrained by bulk core radioactivity, and TiO2-rich mare basalt liquids must be positively buoyant enough to form dikes rather than sink. This model predicts (Table 1) early high-Ti mare basalts followed by a suite of other types and is completely independent of impact basin formation.
3.5. Model 5: Nature and Location of Source Regions and Emplacement Styles Related to Global Thermal Evolution and Basin Evolution but Independent of Basin Formation
 In model 5, mare basalt emplacement is not necessarily related to the formation of the impact basin in which the basalts reside. Rather, the shape of the basin, the sequence and geometry of fill and the resulting distribution of loading-induced stresses (influenced by global thermal evolution and the net state of stress in the lithosphere) controls the location of eruptive vents and the style of emplacement [e.g., Solomon and Head, 1979, 1980; McGovern and Litherland, 2010]. Using linear tectonic rilles and wrinkle ridges, Solomon and Head  established that there are two stress systems, one local and another global, that affect the volcanism that occurred in the large lunar basins. The local stresses are the result of lithospheric loading by the basalt fill of the basins and global stresses originate from the thermal evolution of the Moon. The onset of mare volcanism in each basin is favored by the presence of the local extensional stresses in the lithosphere. Loading by central mare fill induces flexure and favors migration of vents to the margins of the basin.
 This hypothesis leads to the prediction (Table 1) that early mare volcanism will be focused in the basin center and that the youngest mare volcanism will have occurred on the edges of the basin. Once the global compressive stress, from global cooling of the Moon, overrides the local extensional stresses, ascent and eruption of mare basalts becomes more difficult, and ultimately terminates.
3.6. Tests of the Models
 Although these five models treat different aspects of the origin, evolution and emplacement of mare basalts, each makes different predictions (Table 1) and can be tested with the further characterization of mare basalts in the Orientale Basin. Our analysis of mare basalts in Orientale thus has the following elements: (1) When and where were Orientale deposits first emplaced? (2) When and where did the terminal stages of volcanism in Orientale basin occur (what is their age, what is their mode of emplacement)? (3) What was the flux of mare basalt volcanism? What is the volume of mare basalt deposits as a function of time, when did it peak, were there multiple peaks, and what was the shape of the decline toward the present? (4) How do the Orientale mare deposits compare to the global distribution of mare basalts as a function of time? (5) How does the mineralogy of the Orientale mare deposits compare to the global distribution of mare basalt types? (6) How do eruption styles compare to the global distribution of mare basalt vent types and implied eruption conditions (e.g., J. W. Head and L. Wilson, Lunar volcanic vent types, landforms and deposits: A synthesis and assessment of modes of emplacement, manuscript in preparation, 2011)? (7) How do the individual and total volumes of Orientale mare basalts help constrain the total volume of mantle melting that has occurred (using estimates of extrusion to intrusion ratios, and depths of origin, to provide an order of magnitude assessment of the total melting that occurred, updating earlier estimates)? (8) How does the distribution of pyroclastics in Orientale compare to the nature and global distribution of mare pyroclastic deposits [e.g., Gaddis et al., 2003], vent characteristics, eruption conditions, and associated mare basalt types? (9) How do the model ages and model age ranges of Orientale mare basalt emplacement relate to the general thermal evolution of the Moon and to the loading and subsidence observed in other basins [e.g., Solomon and Head, 1980]? (10) Is there any evidence in Orientale for the type of “red spot” extrusive domes and related deposits [e.g., Chevrel et al., 1999; Wagner et al., 2002, 2010; Wilson and Head, 2003] that may relate to suspected KREEP basalt extrusions [e.g., Spudis et al., 1988] or to upland and Cayley plains that might represent aluminous basalts or cryptomaria [e.g., Antonenko et al., 1995; Antonenko, 1999]?
4. The Orientale Basin: Background and Setting
 The Orientale multiring impact basin, located on the western limb of the Moon (19°S and 93°W), is the youngest lunar basin, dated to the Upper Imbrium period of lunar history [e.g., Wilhelms, 1987; Kadel, 1993] (Figure 1). Orientale is particularly interesting because, unlike most lunar basins, its interior has not been completely filled with mare deposits, permitting investigations into the volcanic processes involved in the evolution and emplacement of mare basalts [e.g., Head, 1974; Greeley, 1976; Spudis et al., 1984; Head and Wilson, 1992; Bussey and Spudis, 1997, 2000; Yingst and Head, 1997]. The preservation of the Orientale ring topography distinctly separates its various mare deposits and allows for a closer examination of their ages, compositions and modes of emplacement. The nature, ages, and location of volcanic vents and deposits within multiringed basins provide evidence for the role the basins play in the generation of volcanism and provide important information on the thermal evolution of the lunar interior.
 There are four main rings associated with Orientale Basin (Figure 1a) [Head, 1974; Moore et al., 1974; Scott et al., 1977; Wilhelms, 1987; Spudis, 1993; Head et al., 1993; Head, 2010]. The ∼930 km Cordillera Mountain ring is an inward facing mountain scarp that defines the basin. The second largest ring, the Outer Rook ring, is an interconnected network of massifs spanning ∼620 km in diameter. The next innermost ring is the Inner Rook ring, ∼480 km in diameter. It is composed of many isolated massifs resembling central peaks in complex craters and together, these resemble a ring of peaks as seen in peak ring basins. The most interior ring is a central depression ∼320 km in diameter that has been interpreted to have formed by ∼3 km of thermal subsidence [Bratt et al., 1985a]. The preservation of this ring topography separates the various mare deposits and allows for a closer examination of their ages, compositions, duration of volcanism and modes of emplacement.
 The three large previously defined mare deposits in Orientale Basin include (Figure 1) Mare Orientale, Lacus Veris, and Lacus Autumni. Mare Orientale is located in the center of the basin, almost entirely encompassed by the rim of the central depression, and is the largest of the mare deposits. Included in the defined volcanic units of Mare Orientale is the polygonal mare deposit directly to the southwest of the center. Next in areal extent is Lacus Veris, located between the Inner Rook and Outer Rook rings. This mare deposit is composed of five large ponds along with several smaller scattered ponds all oriented in an arcuate belt, extending from ∼NNW to E. Similar in broad location within the basin interior is Lacus Autumni, positioned between the Outer Rook and Cordillera mountains between ∼ENE to E. Only three relatively shallow ponds comprise Lacus Autumni. The deposits within each set of rings become progressively smaller in area with distance from the center of the basin. A large mafic ring occurs in the SSW part of the Orientale interior, centered on the Outer Rook ring (Figure 1c).
 In addition to the volcanic deposits located between the basin rings, there are four main deposits related to the underlying Orientale Basin: the Hevelius Formation, the Montes Rook Formation, the Maunder Formation and related plains material [Scott et al., 1977]. The Hevelius Formation consists of the radial basin ejecta blanket located outside the Cordillera mountain ring. Interior to that deposit is the Montes Rook Formation, between the Cordillera and Outer Rook rings. It consists of hummocky knobs situated in a rough textured matrix. The outer facies of the Maunder Formation has a highly fractured, undulating and corrugated surface and is interpreted to be composed of melt sheet material. Last, the inner facies of the Maunder Formation consists of higher-albedo plains material that has a smooth surface texture, embays topography and is broken by distinct linear fracture patterns; this facies is interpreted to be more pure impact melt lacking the abundance of admixed clasts that result in draping and cracking of the outer part of the Maunder Formation [Head, 1974; Moore et al., 1974; Spudis et al., 1984].
5. Distribution, Areas, and Volumes of Mare Basalts and Related Volcanic Features
 Using M3 data, we first examined the Orientale region for any evidence of previously undetected mare patches. The increased spatial and spectral resolution of M3 has allowed for the identification of several additional mare deposits to the west and south of the basin interior (Figure 2), although the major distribution of mare from the southwest to the northeast of the basin remains unchanged. Orientale Basin itself is ∼930 km in diameter and covers an area of ∼700,000 km2 [Head, 1974]. Within the basin, the largest of the mare deposits is Mare Orientale, which covers an area of 52,700 km2 (Table 2). Determination of the volume of Mare Orientale requires information on the thickness of the deposit. Previous estimates of the thickness of Mare Orientale were less than ∼1–2 km [Head, 1974; Solomon and Head, 1980], and perhaps up to ∼1 km thick [Greeley, 1976; Scott et al., 1977].
Table 2. Areas, Volumes, and Elevations of Mare Ponds in Orientale Basin
 We used M3 spectral reflectance data to assess postmare impact craters that penetrate into and possibly through the mare deposit, and stratigraphic relationships, to assess the range of mare thickness and to derive an average thickness estimate (Figure 3a). Craters that impacted into the mare and showed mare signatures in their interior and ejecta provided minimum thickness estimates, while impacts into mare that excavated nonmare highlands material provided maximum estimates of the thickness of fill. These craters were chosen based on their location in the mare deposits, although, the appearance of the crater and ejecta spectra were important as well. It was necessary for the geologic materials to still have an identifiable spectrum in order to differentiate between the basaltic mare and more feldspathic highland signatures. Mare spectroscopic signatures contain mineral absorptions at both 1 and 2 μm, owing to the presence of olivine and pyroxene in the mare basalt, and have lower albedos (lower reflectance values), as compared to the highland material. Spectral signatures of highland material lack absorption features, but have higher albedo values. This is due to the fact that the highlands are thought to be composed largely of anorthosite, which is usually spectrally neutral on the Moon. Anorthosite has an absorption around 1.2 μm if Fe2+ is included in the mineral structure [e.g., Bell and Mao, 1973; Adams and Goullaud, 1978] and if the rock has not been subjected to sufficiently high shock pressures to erase evidence of this signature [e.g., Adams, 1979; Johnson and Hörz, 2003].
 Crater impacts throughout lunar history have prompted lateral and vertical mixing of geologic materials over the entire surface of the Moon. Mixing of feldspathic highland material and mafic mare material can influence estimates of the thickness of mare deposits as well as the composition of mare basalt. Highland material could contaminate mare material and change the shape of the spectra, including the strength of the absorptions, and the location of absorption band centers [e.g., Staid, 2000]. Several researchers have investigated the mixing relationships along mare-highland boundaries, investigating the importance of lateral and vertical transport of materials [e.g., Fischer and Pieters, 1995; Mustard and Head, 1996; Mustard et al., 1998; Li and Mustard, 2000]. Mustard and Head  investigated the effects of three contact geometries and found lateral mixing to be the dominant transport mechanism when considering boundaries with highland massifs bordering mare deposits.
 The ∼13 km diameter crater Il'in, in northwest Mare Orientale, lacks a highland feldspathic signature in its impact ejecta and interior [see also Staid, 2000]. Based on the depth/diameter relationships defined by Pike , Il'in excavates between ∼0.8 km [Stöffler et al., 1975] and ∼1.3 km [Melosh, 1989] into mare material (Figures 3a and 3b). The crater Il'in, however, is superposed on a mare wrinkle ridge (Figure 3b), and on the basis of the interpreted three-dimensional structure of these contractional features, it is possible that the mare basalts in which the arch and ridges formed have been thickened by the folding and thrusting thought to accompany their formation [e.g., Lucchitta, 1976; Sharpton and Head, 1988].
 The crater Hohmann, ∼100 km east of Il'in, has a nonmare feldspathic signature on its rim and is partially filled with mare material. Hohmann is ∼17 km in diameter, and based on the relationships defined by Stöffler et al.  and Melosh  is estimated to have sampled between ∼1.1 km and ∼1.7 km into the crust (Figures 3a and 3c). Hohmann crater rim is composed entirely of highland material, making this crater the maximum limit for the depth of Mare Orientale.
 Interspersed throughout Mare Orientale are several smaller craters with diameters on the order of 4–5 km excavating feldspathic basin material. Diameters of this size indicate sampling depths of < ∼300 m (Figure 3a). In addition, mare deposits along the margins of Mare Orientale show evidence for lateral mixing, where highlands ejecta has been thrown in from the adjacent basin deposits, and vertical mixing, caused by material being excavated from below thin mare deposits and incorporated into the regolith. For instance, a good example of predominantly lateral mixing is observed in the rectangular mare deposit to the southwest of Mare Orientale. It is mantled with feldspathic soils while most of the craters show a mafic signature. However, the larger of the craters (∼2 km in diameter) have a faint feldspathic signature on their floor, similar to the soil signature in the polygon. In comparison, small craters on the order of several hundred meters in diameter have only a mafic signature. This suggests that locally this mare unit is very thin, less than 40 m thick (Figure 3d).
 Furthermore, there are numerous kipukas (islands of basin deposits surrounded by, and protruding through, the maria) whose spectral properties are consistent with the Maunder Formation protruding through Mare Orientale (Figures 1c and 3e). The abundance of these preexisting topographic features in the mare, combined with the shallow slopes from mare margins suggested by the decrease in kipuka density (Figure 3e), support the interpretation that Mare Orientale is shallow in these specific areas. For example, there is a line of preexisting basin topography through the center of the basin protruding through the surface of the mare deposit (Figures 3a and 3e). Thus, it appears that the center of the basin, in a north-south direction, is much shallower than adjacent mare to its east and west. The juxtaposition of all these features demonstrates that the basin topography underlying Mare Orientale varies significantly, especially from east to west. On the basis of these observations and data, our best estimates for the thickness and volumes of Mare Orientale are as follows: small crater depth of sampling data suggest a typical thickness of ∼200 m (Figure 3a), and kipuka distribution data (Figure 3e) are consistent with this. Applying this as an average thickness to the area of Mare Orientale provides an estimated volume of ∼10,440 km3 (Table 2).
 Volumes and areas of terrestrial continental flood basalts have been calculated by many researchers [e.g., Richards et al., 1989; Camp et al., 2003; Reidel, 2005; Jay and Widdowson, 2008]. For instance, estimated volumes for the Deccan Traps are between 0.5 and 2 × 106 km3 over an area of 1 × 106 km2. In comparison to the Deccan Traps, all of the flows that make up Columbia River Basalts have both a smaller calculated volume (∼2 × 105 km3) [Jay and Widdowson, 2008] and areal extent (>2 × 105 km2) [Reidel, 2005]. Mare Orientale has an area of ∼52,700 km2 and a volume of ∼10,400 km3, much less in extent and volume than the Columbia River basalts.
 Lacus Veris (Figure 1) is the next largest mare deposit, with each of its five ponds varying considerably in area. The largest of the Lacus Veris ponds is ∼8,890 km2 and the smallest pond is ∼145 km2. Using the largest craters to constrain the thickness of these deposits [Yingst and Head, 1997] gives minimal volume estimates ranging from ∼10 km3 to ∼1,385 km3 (Table 2). The largest pond in Lacus Veris appears to vary in thickness throughout its extent in a manner similar to Mare Orientale. In the northern part of the pond a ∼7 km diameter crater excavates into mafic material, but approximately 30 km southeast a ∼5 km diameter crater situated between kipukas excavates feldspathic material (Figure 3f).
 Lacus Autumni is composed of several large ponds as well. The largest of its three ponds covers an area of ∼2,060 km2 and the smallest is ∼815 km2. These mare ponds are some of the shallowest in the basin, as evidenced by the abundance of kipukas. Based on current crater investigations, ponds in Lacus Autumni have volumes between ∼65 km3 and ∼115 km3 (Table 2).
Yingst and Head  previously investigated the areas and volumes of Lacus Veris and Lacus Autumni to understand better the mode and rates of emplacement for these deposits. Both volume and area values were recalculated in the current analysis because of the availability of higher-resolution image, compositional and topographic data, allowing for more precise identification of ponds and small craters. The new total calculated area and volume for ponds 1–24 from this study (19,350 km2 and 2,400 km3) is smaller than the previous estimates of Yingst and Head  (23,400 km2 and 9,700 km3). The total (minimum) volume of mare in the Orientale Basin is ∼46,000 km3, throughout an area of ∼700,000 km2. The smaller areas and volumes are very likely due to higher-resolution data used in defining the margins of the deposits, the spectral data that helped define the mare nature of smooth mare areas and the high spatial and spectral resolution data that permitted much better estimates of the thickness of the mare basalts.
Yingst and Head  found that South Pole–Aitken contains ∼153,000 km3 of mare over its areal extent (∼4,000,000 km2). Comparing the basin areas and mare areal extent of Orientale and South Pole–Aitken shows that, proportionally, Orientale Basin has more mare covering its interior than South Pole–Aitken, when taking Mare Orientale into account. In contrast to these two incompletely filled basins, most other lunar nearside basins have volumes between 200,000 km3 (e.g., Tranquillitatis and Fecunditatis) to 1,000,000 km3 (e.g., Crisium, Serenitatis) [Bratt et al., 1985b], 4 to 20 times more mare than Orientale.
 We identified several ponds (1, 4, 5, 6, 8) using M3 data (Figure 1). Pond 4 (see identification scheme in Figure 2) is ∼80 km2 and lies just inside massifs on the western side of the Inner Rook ring. Ponds 2 and 3, ∼230 km2 and ∼935 km2, have been noted previously in Zond and Galileo data [Scott et al., 1977; Kadel, 1993; Head et al., 1993] and both are located adjacent to the Outer Rook ring, in a similar position to the ponds of Lacus Veris. Ponds 1, 5, and 6, having areas of ∼25 km2, ∼160 km2, and ∼160 km2, lie along the scarp of the Cordillera mountain ring and mirror the occurrence of the ponds of Lacus Autumni. Placing volumetric constraints on these identified ponds is more difficult because they have very few impacts owing to their small size. Table 2 shows minimum estimates of these pond volumes derived from craters with mafic signatures.
 Other small deposits have been identified around and between previously known mare deposits. Pond 20, covering ∼105 km2, has been identified to the east of the Outer Rook ring. Unlike the other deposits identified by M3, this mare pond appears to mantle the preexisting topography instead of ponding in a topographic low. Many other small deposits have been identified between the ponds in Lacus Autumni and spread around the eastern base of the Outer Rook ring (Figure 2). Despite their small size, these deposits identified with M3 data show that volcanism in Orientale was active throughout the entire basin area and not simply confined to the eastern part of Orientale.
 A large, 154 km diameter dark annular ring was first discovered and documented in Soviet Zond 8 images [Lipsky, 1975], and is located in the south-southwestern part of Orientale Basin centered on the Outer Rook ring (Figure 4). Earlier interpretations concluded that the dark mantle ring deposit (DMRD) consisted of a large number of individual vents producing an annulus of dark mantle pyroclastic deposits; the mantle was interpreted to have erupted from vents marking the location of a 175 km diameter pre-Orientale Basin crater, thus explaining its circular nature [Schultz and Spudis, 1978]. This interpretation had a significant influence on attempts to locate the Orientale Basin transient cavity rim crest, since the presence, and thus preservation, of such a large crater would tend to place the transient cavity rim crest inside its location, at or inside the Inner Rook ring [Schultz and Spudis, 1978].
 Clementine data revealed the presence of a 7.5 km wide by 16 km long elongate depression located at the center of the DMRD [Weitz et al., 1998; Head et al., 2002]. This elongate depression, similar to features seen in association with other dark mantle deposits (e.g., Sulpicius Gallus [Lucchitta and Schmidt, 1974]), had no obvious adjacent deposits of dark mantle or mare, but on the basis of its location in the approximate center of the DMRD, Weitz et al.  and Head et al.  investigated the possibility that it could be a source crater for an eruption producing the DMRD. They proposed that the dark ring is the manifestation of a pyroclastic eruption originating at a fissure vent (the elongate depression) and forming an Ionian-like eruption plume. Head et al.  outlined a scenario in which the event producing the eruption began with a dike rapidly emplaced from subcrustal depths to within ∼3–4 km of the surface. The dike stabilized and degassed over ∼1.7 years to form an upper foam layer, which then penetrated to the surface to cause an eruption, lasting ∼1–2 weeks. This eruption produced an almost 40 km high symmetrical spray of pyroclasts into the lunar vacuum at velocities of ∼350 to ∼420 m/s; as a result of this eruption, the pyroclastic material accumulated in a symmetrical ring around the vent to produce the DMRD. Head et al.  attributed the paucity of pyroclastic rings of this type on the Moon to the low probability of a dike stalling at just the right depth (∼3–4 km) to create these eruption conditions.
 In general, lunar volcanic glasses can be identified in spectra by broad absorptions at 1.0 and 2.0 μm. The amount of titanium in the glasses can be determined by an iron-titanium charge transfer absorption feature in the visible region of the electromagnetic spectrum. The higher the Ti content in the glasses, the stronger the absorption edge in the visible [Bell et al., 1976]. Black beads that represent the crystallized equivalents of the orange glasses have a strong, broad absorption centered at ∼0.6 μm that is due to the presence of ilmenite.
 Galileo images provided the first multispectral images of the DMRD region and the dark halo deposit itself; Pieters et al.  showed that the ring deposit is somewhat brighter than nearside mantling deposits, has a very weak 1 μm absorption band, and has an ultraviolet-visible (UV/VIS) ratio that is relatively high but lower than that seen in nearside deposits of black spheres. They concluded that these characteristics could be consistent with contamination of the deposit by highland material (as suggested by Greeley et al. ). They also proposed an alternative interpretation, that the actual spectral properties of the pyroclastics themselves dominate the measured spectral properties. In the first case the deposits could be interpreted as similar to ilmenite-rich dark mantling material such as the black beads sampled at Apollo 17. In the second case, however, the weak 1 μm band could indicate that the deposits are not homogeneous glass but are in a crystallized form. The low UV/VIS ratio, relative to black beads seen elsewhere, could be due to the lower abundance of ilmenite, the opaque component that causes darkening. In the latter case, the Orientale ring deposit might have affinities with the local medium-Ti deposits identified in the Orientale region [Greeley et al., 1993].
 The Clementine UV/VIS spectral data for the Orientale DMRD show a slight absorption at 0.9–1.0 μm, similar to that seen in Aristarchus DMD spectra [Weitz et al., 1998]. On the basis of these characteristics, the beads that compose the Orientale DMRD were interpreted to be dominated by glasses, rather than the crystallized beads typical of the Taurus-Littrow DMD deposit. The implications are that the DMRD beads cooled rapidly after eruption at the surface, preventing crystals from forming. Low optical density in the eruption plume, due to either a high gas content or a low mass eruption rate, is required to enhance rapid cooling and explain the dominance of glasses in the deposit. This implies relatively rapid cooling times for the eruption products, consistent with the Head et al.  scenario of a dike stalling and buildup of volatile foams prior to eruption. Although analysis is not yet complete, M3 data suggest that the pyroclastic deposits have been weathered and are glass-rich only at small fresh craters. Although the age of these deposits is not known due to the difficulty of dating friable mantle deposits, there is no reason to believe that they lie outside the range of model ages for the mare deposits.
 The total duration of volcanic activity in Orientale spans 0.85–1.50 Ga (see section 7), but despite this lengthy duration, the ponds have relatively low volumes [Yingst and Head, 1997]. In comparison to other large impact basins on the Moon, Orientale contains a very small total volume of mare material, ∼46,000 km3. Bratt et al. [1985b] calculated total volumes for basins on the nearside of the Moon, which include the volume of the topographic depression and mare fill. For comparison, these volumes are taken as the maximum volume of mare. Serenitatis and Nectaris, two basins similar in size to Orientale, have total basin volumes of 1 × 106 km3 and 7 × 105 km3, respectively. According to Bratt et al. [1985b], Orientale has a total basin volume of 7 × 105 km3, meaning that there is ∼6.55 × 105 km3 that was not filled with mare. Consequently the mare load in the center of Orientale Basin is not as massive as the initial central loads that are likely to have characterized the deeper basin interior in other lunar basins. For example, Solomon and Head  estimated the thickness of Mare Orientale to be 1–2 km in modeling the mare load and its influence on flexure. The mantle plug that was uplifted into the excavated cavity when it collapsed [Wieczorek and Phillips, 2000], and its superisostatic state may thus be of more significance in terms of a cause of the local extensional stress along the edge of the basin.
6. Craters Kopff and Maunder
 Two large craters in the interior of Orientale, Maunder and Kopff (Figure 5), have attracted attention for years due their similar sizes but contrasting morphology. A first-order examination of the 55 km diameter Maunder crater indicates that it is a classic example of an impact crater, having a central peak, a flat floor, terraced walls, a raised rim and continuous ejecta deposit, and a system of radial rays [El-Baz, 1974; Pike, 1980]. In contrast, the 42 km diameter Kopff crater has no central peak, no wall terraces, an unusual rim shape, and unusual smooth crater ejecta deposits and secondaries. Several authors hypothesized that Kopff crater was volcanic in origin based on its subdued morphology [McCauley, 1968; Guest and Greeley, 1977], with one hypothesis favoring an explosion caldera [Pai et al., 1978]; other researchers favored an unusual impact event, for example, into partially molten material [Wilhelms and McCauley, 1971; Guest and Greeley, 1977; Spudis et al., 1984]. Still others hypothesized that Kopff was a volcanically altered impact crater [Schultz, 1976; Wilhelms, 1987].
 The rim crest of Kopff is very irregular, with multiple undulations along its circumference. The largest crater rim-to-floor relief is ∼1.7 km. This is less than half of the crater rim crest-to-floor measurement of Maunder, ∼4 km deep. The depth of Kopff is too shallow to be a typical fresh impact crater and its floor diameter of ∼36 km is too large [Pike, 1980] (Figures 6a and 6b). Both these characteristics, in addition to the observed floor flooding, indicate that the crater has undergone at least some volcanic modification. Measurements of morphologic features that would not be affected by volcanic modification, such as rim height and width, indicate that Kopff has characteristics consistent with an impact crater (Figures 6c and 6d). Its rim height value is on the lower end of collected measurements, but this could be the result of impact into partially molten material. In comparison, the rim height of Maunder is slightly above the trend, likely to be the result of its location on preexisting topography along the edge of the inner depression.
 An impact crater ∼3 km in diameter located on the eastern edge of the mare-covered floor of Kopff has excavated anorthositic material (<2 wt % FeO) [Bussey and Spudis, 2000], confirmed here with M3 data (Figure 5b). In the standard M3 color composite (Figure 5b) the rim of Kopff has a distinctly feldspathic signature, suggesting that it is composed largely of premare basin material. The depth of excavation of the ∼3 km superposed crater is ∼0.2 km, constraining the mare deposit volume to be <145 km3, using techniques described by Pike , Stöffler et al. , and Yingst and Head . If Kopff was a typical impact crater, its depth should be on the order of ∼3.5 km (Figure 6a).
 In order to evaluate processes of modification for Kopff, Maunder crater was “flooded” to an elevation of approximately −2.6 km from its rim crest (comparable to the depth of the floor of Kopff below the rim crest) to evaluate its morphology (Figure 7). Comparing the still uncovered portion of Maunder with the morphology of Kopff could help to determine if the flooding scenario is possible. Assuming an initial depth of 3.5 km for the floor of Kopff, this flooded elevation is similar to the amount of flooding Kopff is assumed to have experienced. At this flood level (∼2.2 km), Maunder is filled with a volume of ∼1900 km3. Using the various morphometric impact crater measurements for Kopff [Pike, 1980] then yields a volume of ∼1700 km3. This calculated volume is significantly different from that determined from compositional information. The ∼3 km diameter crater on the east of Kopff floor that excavated nonmare material may have impacted into an uplifted crater floor, or possibly a buried slump terrace.
 The floor of Kopff is very asymmetric due to a small depression in its center and a slight bulge on its eastern side, superposed by a floor fracture. The uneven nature of the mare floor indicates that after flooding, some type of volcanic activity continued to occur and eventually warped the surface. The bulge and fracturing of the floor could result from a small mare intrusion forming beneath the crater floor, creating a floor-fractured crater [Schultz, 1976].
 In addition to the unusual morphology of Kopff, crater count statistics provide additional evidence concerning its origin. The Maunder Formation has been dated at ∼3.64 Ga on average, and Kopff formed at ∼3.63 Ga, based on model ages derived from its smooth ejecta deposit (Table 3). The Kopff impact occurred soon after the emplacement of the melt sheet. If the central uplifted area of the Orientale Basin and the overlying melt sheet had not thermally equilibrated after ∼10 Ma, perhaps the thermal state of the substrate influenced the formation and early modification of the morphology of Kopff [e.g., Spudis et al., 1984]. Dates on the volcanic flooding of the interior of Kopf are ∼3.36 Ga, significantly later than the formation of the crater itself. This favors volcanic modification of an unusual crater type.
Table 3. Model Ages of Orientale and Datable Mare Ponds Thereina
Model ages calculated using the CraterStats program [Michael and Neukum, 2010]; data are unbinned and both the 1983 and 2001 Neukum chronology and production functions are used. Error estimates are only a function of count statistics and do not incorporate any systematic uncertainties in the absolute age calibration.
Numbers in parentheses correspond to pond identification numbers from Figure 2.
 Crater production and chronology functions [e.g., Neukum, 1983; Ivanov et al., 1999, 2001; Hartmann et al., 2000; Neukum et al., 2001] have been developed that permit the assignment of ages from crater size-frequency distributions for the surfaces of the terrestrial planets. Crater counting is very useful for dating lunar surface units due to the high degree of preservation of impact craters. This dating method can be employed to determine the timing of large basin-forming impacts, ages of different volcanic units or flows, and the duration of volcanism. Important considerations when dating volcanic surfaces include the effects of resurfacing events, secondary craters and endogenic craters [e.g., Hiesinger et al., 2000; Greeley and Gault, 1971; Oberbeck and Morrison, 1974], all of which can produce ages that differ from the ages of the actual surface units.
 Crater count analyses have been conducted on Mare Orientale, Lacus Veris and Lacus Autumni by previous researchers [e.g., Greeley et al., 1993; Kadel, 1993; Morota et al., 2010]. Initially, Hartmann and Yale  interpreted early crater counts to mean that Lacus Veris and Autumni occurred soon after the Orientale event, followed later by infilling of Mare Orientale. Using Lunar Orbiter photographs, Greeley et al.  determined that the oldest mare was emplaced in south central Mare Orientale at ∼3.70 Ga, followed by the emplacement of maria in western and southeastern Mare Orientale at ∼3.45 Ga. Moving further from the center of the basin, Greeley et al.  showed that Lacus Veris had an average model age of ∼3.50 Ga, and Lacus Autumni is the youngest deposit with a model age of ∼2.85 Ga. Even younger model ages have been reported for central Lacus Veris, at ∼2.59 and ∼2.29 Ga [Kadel, 1993]. Based on these findings, mare volcanism in Orientale was estimated to have lasted between 0.85 and 1.50 Ga [Greeley et al., 1993], beginning ∼100 Ma after basin formation.
 Using high-resolution M3 data at 140 m/pixel, as compared to the Lunar Orbiter ∼3.5–7.6 km/pixel resolution used by Greeley et al. , crater counts were conducted on the ejecta deposit of Orientale Basin, parts of Mare Orientale, the Maunder Formation [see Head, 1974; Bussey and Spudis, 2000] and the identified mare ponds in Lacus Veris and Lacus Autumni of sufficient size to produce reliable model ages (Table 3). Previous estimates of the model age of the Orientale Basin event are ∼3.8 Ga [Baldwin, 1974; Nunes et al., 1974; Schaeffer and Husain, 1974; Neukum, 1983; Baldwin, 1987; Wilhelms, 1987; Kadel, 1993]. Updated crater age equations produce a model age of ∼3.68 Ga for the Orientale impact event (Figure 8). The older Neukum  production and chronology functions yield a model age of ∼3.8 Ga, using the counted regions from this study. This difference in calculated ages arises from an update of Neukum production and chronology equations. New counts on impact craters in and around Orientale Basin permitted a reestimation of the size-frequency distribution curve for craters between 1 km and 20 km [Neukum et al., 2001]. Earlier basin model age estimates made using the Hartmann production function could differ as a result of the inherent differences between the two approaches for deriving a production function. Neukum uses an 11th-order polynomial, whereas Hartmann uses a log incremental equation to approximate the size-frequency distribution curve. These two functions differ most at diameter bins between 2 and 20 km [Neukum et al., 2001, Figure 8].
 Our data suggest that the melt sheet material of the Maunder Formation solidified shortly after the basin forming event, ∼3.64 Ga. In theory, the melt sheet and basin ejecta ages should be the same, since the melt is produced instantaneously with the impact, and comes to rest during the short-term modification stage, the terminal stage of the cratering event. The discrepancy between ages is likely to be due to differences in the degree of crater preservation, or possibly due to differences in material properties [e.g., van der Bogert et al., 2010]. Melt sheet material tends to drape topography, creating smooth surfaces that allow for the preservation of smaller craters (∼<10 km). Ejecta blankets tend to have rougher textures that may lead to poorer preservation of smaller craters. Thus, the melt sheet age is typically the more reliable of the two for the age of the impact event. However, we counted craters > 5 km on the Hevelius Formation and craters > 0.75 km on the Maunder Formation in an attempt to count the most preserved craters on each deposit. Our calculated crater retention ages are almost the same, ∼3.68 Ga for the ejecta and ∼3.64 for the melt sheet; this suggests our efforts were successful. Mare Orientale was emplaced shortly after the Maunder Formation ∼3.58 Ga (Figure 8 and Table 3). This delay of ∼60–100 Ma between basin formation and volcanism in Mare Orientale, and later extended volcanism in Lacus Veris and Lacus Autumni argues against Orientale impact pressure release melting (Table 1) being a significance factor in mare basalt production [e.g., Elkins-Tanton et al., 2004], at least for the last volcanic deposits that reset the age of these surfaces.
 Impact craters were counted, size-frequency distributions were compiled and ages were calculated for the five largest ponds in Lacus Veris; this yielded an age range of ∼3.20 to ∼3.69 Ga (Figures 9a–9f and Table 3). These model ages are relatively consistent with those calculated from previous studies, but these are differences in grouping and separation of the various ponds [e.g., Greeley et al., 1993; Kadel, 1993]. Greeley et al.  grouped several of the ponds in Lacus Veris together, while Kadel  separated individual ponds into separate units.
 The largest discrepancy in model ages between our study and previous work occurs in the ponds of Lacus Autumni. Our crater counts give an age range between ∼3.47 and ∼1.66 Ga for the ponds (Figures 9j–9l; note the model age uncertainties). The young model age of ∼1.66 Ga falls well within the calculated time range of mare volcanism occurring on the lunar nearside [Hiesinger et al., 2000] (Figure 10). The new model ages calculated for ponds 22 and 23 in Lacus Autumni do not wholly agree with ages reported by previous researchers; Kadel  calculated a model age of 3.00 Ga for pond 23 and 2.85 Ga for pond 22, differences of 0.57 Ga and 1.19 Ga, respectively. This discrepancy in the model ages could be the result of sizable differences in the definition of the count area in these two ponds, or from count statistics. However, the fact that most of the other ponds in Lacus Veris and Autumni have similar model ages in both studies suggests that generally, the two crater counting methodologies do not appear to differ significantly.
 The time of impact for both Maunder and Kopff craters was calculated by dating their continuous ejecta deposits. These two large craters are of particular interest because though similar in location and size, they are very different in morphology. The mare floor of Kopff is dated at ∼3.36 Ga and its ejecta at ∼3.63 Ga, supporting the idea that Kopff might have been produced by impact into material with different thermal or mechanical properties [e.g., Wilhelms and McCauley, 1971; Guest and Greeley, 1977]. Maunder ejecta is dated ∼2.88 Ga. This younger model age for Maunder crater is supported by stratigraphic relationships, such as the superposition of Maunder ejecta on Kopff's mare floor.
 The range of newly calculated model ages for the mare deposits coincides with the climax of nearside lunar volcanism. The majority of our model ages plot around ∼3.5 Ga (Figure 10a), similar to nearside flow units dated by Hiesinger et al. . Thus, Orientale mare deposits were emplaced in the same era of significant volcanism that occurred elsewhere on the Moon.
8. Mineralogy of Mare Basalts
 The mineralogy of Orientale Basin mare basalts is of particular interest, especially the question of how they compare with the compositional range seen in nearside lunar basalts. Morphological analysis of Orientale basalts has been interpreted to mean that mare ponds were emplaced in single eruptive events [Yingst and Head, 1997], which has implications for the size of the source region beneath this part of the Moon. Several researchers [Greeley et al., 1993; Kadel, 1993; Staid, 2000] have investigated the composition of the mare basalts in Orientale Basin, but none have been able to establish the actual mineralogy due to limitations in instrument spatial and spectral resolution. Previous studies have found that Mare Orientale and other mare ponds within the basin are highly contaminated by the local highland material [Spudis et al., 1984; Staid, 2000]; despite this, Spudis et al.  concluded that the basalts in Mare Orientale, Lacus Veris and Autumni are probably of similar composition to nearside basalts.
 Many researchers have focused on the TiO2 contents of the mare ponds in Orientale [Greeley et al., 1993; Kadel, 1993]. Using Galileo Sold Sate Imaging (SSI) 0.41/0.56 μm spectral reflectance ratio images, Greeley et al.  found that units in Mare Orientale are medium-high-Ti basalt soils (3–7 wt % TiO2), with lower-Ti signatures in the northeast and west central Mare Orientale. Lacus Veris contains medium- to high-Ti basalt soils and Lacus Autumni has medium- to high-Ti basalt soils in the north and medium-Ti basalt soils (<4 wt % TiO2) in the south. The authors pointed out that contamination from Orientale feldspathic materials could be influencing the observed compositions. Using Galileo EM-1 multispectral images, Kadel  determined that Mare Orientale had a titanium content of 3–7 wt % TiO2. With Clementine data, Staid  calculated titanium values of ∼4% TiO2 using the methods from Charette et al. . Currently M3 data is being used to investigate nearside lunar basalts with varying TiO2 contents [e.g., Staid et al., 2011; Dhingra et al., 2010]. Determining the TiO2 content of mare basalts across the lunar surface is an ongoing project. However, generally M3 data does not appear to negate previous TiO2 estimates.
 From crater ages, Greeley et al.  concluded that volcanism in Orientale spanned ∼0.85 Ga and that for such a long time period the narrow range of mare compositions was unusual compared to the nearside mare. They concluded that this could indicate multiple eruptions from a relatively homogeneous source or multiple sources with subtle compositional differences [Greeley et al., 1993]. It remains unclear as to whether multiple flows of different mineralogy exist in Mare Orientale, or whether the mineralogy is generally similar, but appears different due to mixing with adjacent and underlying Orientale Basin material. The two most uncontaminated regions in Mare Orientale (as defined by Staid ) were investigated in depth using the M3 1.0 μm band centers and band strength values to determine whether multiple flows exist.
 Evidence for vertical mixing through thin basalts can be seen in the central portions of the interior deposits as well as in the ejecta of craters near mare boundaries. Elevated kipukas of basin materials, including the rim of Hohmann crater (17 km), are embayed by subsequent mare flows, indicating that large areas of the central basalt deposits are relatively thin. Lateral mixing has also brightened and altered the spectral properties of the Orientale basalts over large areas of the northern portion of the interior deposits and near mare boundaries [Staid, 2000]. In the northern portion of the basin, dark halo craters are observed to penetrate through ejecta from Maunder crater (55 km), exposing spectrally bluer and darker or less contaminated basalts from beneath brighter and redder ejecta-contaminated mare surfaces.
 In the northwest region, no distinct pattern emerges in the band center data (Figure 11a). The average band center in this region is 0.997 μm, with values ranging from 0.981 μm to 1.019 μm. The strength of the 1 μm absorption is another parameter that could reveal additional existing compositional differences. In theory, collecting spectra from a number of small fresh craters should eliminate any significant differences in maturity in our sample population. Thus, any observed differences in the absorption strength are inherent in the minerals themselves. In the northwest study region (Figure 11c), the majority of the strong 1 μm absorptions occur to the south, while the weaker absorptions are concentrated in the north. However, as was the case with the band center data, an obvious correlation does not exist. One possible explanation for this might be the impact of Il'in crater (13 km), whose ejecta material blanketed a significant area in this region of Mare Orientale. M3 spectral data (Figure 1c) indicates that ejecta from Il'in crater is dominated by mare basalt signatures, is less weathered than surrounding soils and lacks a detectable component of the underlying feldspathic basin materials, indicating that basalts are thicker in this portion of the basin.
 Additionally, redder spectral properties that may result from the presence of lower-Ti basalts are observed in some regions of Mare Orientale. TiO2 rich mare basalts have a flatter, or bluer, continuum slope than lower-Ti basalts [Staid and Pieters, 2000]. These red regions appear to have lower UV/VIS ratios and higher albedos than the bluest regions of Mare Orientale, while retaining stronger mafic signatures than areas which have obvious nonmare mixing (e.g., regions south of Maunder crater). Such areas include deposits in southwestern Mare Orientale and small regions north and east of Il'in crater. However, due to the large amounts of nonmare contamination throughout the Orientale deposits, confirmation of such potential variations in emplaced basalt compositions will require more detailed analysis of associated mare craters in another study using lower illumination angle M3 data acquired in later optical periods of the mission. Future research will need to be conducted to quantify how much mixing contributes to spectral signatures in Orientale Basin mare deposits.
 Comparing the 1 μm characteristics of the northwest study region with the southeast region clearly shows that the northwest lacks the diversity in band center position found in the southeast (Figures 11a and 11b). The southeast region has spectra band centers at both longer and shorter wavelengths; absorption band centers range from 0.978 μm to 1.053 μm, with an average of 1.003 μm. Proximity to the edge of the melt sheet does not appear to have an effect on band center, as the range in values is quite large for craters located along this boundary. Surface albedo differences and distance from kipukas do not appear to affect the band center values either. In addition, the 1 μm band strength data was investigated to determine if any mineralogical differences exist in the mare composition (Figure 11d).
 As more Fe2+ and Ca2+ substitutes for Mg2+ in the pyroxene structure the absorption centers shift to longer wavelengths, which is reflected in the spectra collected from the northwest and southeast regions in Mare Orientale. Typically, orthopyroxenes have their first electronic transition absorption centered around 0.9 μm; another absorption of roughly equal strength is centered about 1.9 μm [e.g., Adams, 1974; Cloutis and Gaffey, 1991; Klima et al., 2007]. High-Ca clinopyroxenes exhibit absorption bands at longer wavelengths, centered near 1 μm and 2.3 μm. The collected spectra have 1 μm band centers that are shifted to longer wavelengths, from 0.978 μm to 1.053 μm, indicating that the pyroxene mineralogy is dominated by a high-Ca clinopyroxene. From this study alone it is difficult to state exactly which elements comprise the minerals and in what proportions they are present. While this study is simply a first look at the composition of basalts in Mare Orientale, further spectral analysis, such as MGM or radiative transfer modeling, is needed to better quantify the compositions of pyroxenes throughout Orientale Basin.
 In Lacus Veris and Autumni the mare basalts appear to have similar spectral properties as those observed within the deposits of Mare Orientale (Figure 1c). Mare soils in these regions are relatively blue, but contain significant and variable nonmare highland material contamination, leading to reddening and brightening of surfaces. The surfaces of the older Lacus Veris basalts north of Orientale appear redder, on average, than younger Lacus Veris deposits to the east. Nonmare contamination appears to be at least partially responsible for these properties as spectrally bluer dark halo craters are observed within several areas of these northern deposits. Lower-Ti basalts [Kadel, 1993; Staid, 2000] in the Lacus Veris deposits north of Orientale, if present, may also contribute to the redder spectral properties of surface soils. However, distinguishing lower-titanium deposits from contaminated higher-Ti soils within northern Lacus Veris will require a more detailed examination of nonmare mixing in this region.
 The deposits in Lacus Autumni also have UV/VIS spectral properties that are similar to basalts in eastern Lacus Veris and most of Mare Orientale, consistent with relatively high-Ti contents (3–7 wt % TiO2). Although these deposits appear to be homogeneous in terms of titanium content, small variations in 1 and 2 μm band strength are observed in thermally corrected M3 integrated band depth images. These variations appear to result both from different degrees of nonmare mixing as well as actual differences in emplaced basalt mineralogy. One region that contains relatively strong variations in 1 and 2 μm mafic band strength occurs as a spectral boundary between western and eastern deposits in Lacus Autumni. These differences are not associated with obvious differences in UV/VIS slope or brightness and, therefore, are likely to result from differences in mineralogy rather than feldspathic contamination. Similar spectral boundaries occur in the M3 standard color composite in other areas of Lacus Veris, but are more subtle and, in some cases, may result from different levels of nonmare mixing. Many of the spectral variations observed for mare soils within these composite images are nonunique and more detailed analysis of mare crater spectra associated with these spectral boundaries will be necessary to reveal potential variations in mafic mineralogy within these deposits.
 It is unlikely that all the diversity is due to feldspathic contamination. There are indications of some variations in UV/VIS slope, 1.0 versus 2.0 μm band strength and band shape that are associated with the basalts themselves. However, the differences appear to be relatively small in most areas, such that we are unable to map obvious spectral units given the large amount of mixing and the high phase angles of the data investigated thus far. Overall, these are relatively homogeneous basalts considering the spatial and temporal extent of their emplacement and the full range of spectral properties observed in the M3 global mosaics.
 Our analyses are consistent with a basaltic composition for Mare Orientale that contains several pyroxene minerals with various amounts of Fe2+ and Ca2+. Most variation in the deposit is likely due to lateral and vertical mixing with feldspathic basin materials. Lacus Veris and Autumni appear similar in composition to Mare Orientale, but certain regions display spectral characteristics indicative of either feldspathic mixing or slight differences in basalt composition. Further investigation is needed to determine the actual cause in apparent spectral differences. All mare basalts in Orientale Basin appear to have approximately the same titanium content, a medium-high value between 3 and 7 wt % TiO2. Calculations of titanium content using the Charette et al.  method indicate that compared to other lunar basins (excepting Tranquillitatis and Western Procellarum), Orientale has a relatively high-Ti content [Staid, 2000]. Using the Lucey et al.  method, Orientale has a titanium value that is generally comparable to the majority of nearside lunar mare basin deposits [Staid, 2000].
9. Volcanological Features Associated With Mare Basalt Deposits
 Lunar mare deposits are characterized by a wide variety of volcanological features [Guest, 1971; Head, 1976], including flow fronts, lava channels, collapsed lava tubes, sinuous rilles, domes and cones, vents and collapse pits. Also observed are multiple flow units distinguished on the basis of spectral contrast and impact crater size-frequency distributions [e.g., Hiesinger et al., 2000, 2008, 2011]. Source regions observed in lunar nearside mare deposits include linear fissures formed by dikes propagating to the surface and erupting along the fissure to form a fissure vent. These features are often characterized by a linear array of cones and small domes, and they commonly have a larger edifice or pit that represents the widest part of the dike, where effusion was concentrated following the initial eruptive phase [Wilson and Head, 1988]. Lava distributary features observed in nearside mare regions include lava channels within broader lava flows (centralized channels defined by marginal levees). These are generally considered to be constructional features, formed by the dynamics of the flow process within the flow itself [e.g., Greeley, 1971; Hulme, 1974]. Channels can also be roofed over, forming an insulating carapace and tube-fed flows; rows of pits and channel segments provide evidence for the presence of tube-fed flows. Fissure vents or point source vents form sinuous rilles, features that are thought to be erosional in nature. Sinuous rilles may form by thermal [e.g., Hulme, 1973; Williams et al., 2000] and mechanical erosion [e.g., Gregg and Greeley, 1993] of the substrate, in contrast to the constructional style of channel-fed flows. The characteristics and distribution of these various features provide important information on the styles of eruption and processes of magma ascent and eruption. Several of these described features are observed in the Orientale Basin.
Greeley  conducted an analysis of the volcanism in Orientale to determine the modes of emplacement of the mare deposits. Initially investigating basaltic terrains on Earth, Greeley  defined three terrain types he considered most important for interpreting basaltic regions on other terrestrial planets: flood basalts, shield volcanoes, and basaltic plains. Using this classification scheme, Greeley  interpreted the modes of emplacement for the various mare deposits in the Orientale Basin. Based on the lack of preserved flow features, areal extent and large implied volume, Mare Orientale was classified by Greeley  as flood basalt style volcanism. Lacus Veris is much different in character, having an abundance of sinuous rilles (Figure 2), thought to be lava channels or collapsed lava tubes, and several small coalescing shield-like structures [Greeley, 1976]. Lacus Autumni has a large number of sinuous rilles as well. As a result of implied fissure vents and rille point sources, Lacus Veris and Lacus Autumni were classified by Greeley  as basaltic plains style volcanism.
 The high spatial and spectral resolution M3 data add new insight into the modes of mare emplacement in Orientale. The absence of (1) multiple, distinctively different flow subunits distinguished by spectral data; (2) morphologically prominent flow fronts; or (3) areal differences in impact crater size-frequency distributions (as is the case in Mare Imbrium) [Bugiolacchi and Guest, 2008], all suggest that Mare Orientale may have been emplaced in a single eruptive phase, consistent with a flood basalt style [Greeley, 1976; Yingst and Head, 1997]. On the other hand, several prominent mare dome-like structures exist in the central part of the basin, and these could potentially have been point sources for some of the plains within Mare Orientale (Figure 12).
 One such dome feature (∼20 km in diameter) is located west of Hohmann crater among the chain of kipukas in the center of the basin (Figures 12a and 12b). There is a crater on the summit region measuring ∼1.6 km in diameter. Based on crater diameter/basal dome diameter relationships established by Head and Gifford  through the evaluation of over 200 lunar mare domes, crater diameter is 19–26% of the basal dome diameter. The summit crater of this Orientale dome does not fit this relationship, suggesting that this mare dome is not a lunar volcano. In addition, the visible outline of the dome does not appear distinct in LOLA data (Figure 12c) nor does it have a distinct spectral signature in M3 data (Figure 12b). Instead, this dome may be categorized more appropriately as a class 5 dome [Head and Gifford, 1980]. It is irregular in shape, fits within the delineated diameter range, does not have a real typical summit crater and is closely associated with highland units. This dome is likely to have originated when the center of Orientale Basin was flooded and mare draped over preexisting topography.
 Another group of mare domes is located in pond 14 of Lacus Veris (Figures 12d, 12e, and 12f). They are ∼6–8 km in diameter and look like small terrestrial shield volcanoes rising ∼170–230 m high. However, despite their volcano-like appearance, all of these mare domes lack a distinct summit crater. The peaks of several of these domes have a higher albedo than their slopes that display a feldspathic signature in the M3 data. Thus, this group of domes is likely to be the result of highland material being draped with mare material (a class 5 dome according to Head and Gifford ). Their proximity to the edge of the mare pond, coupled with M3 spectral data, supports this interpretation.
 In east central Mare Orientale, a dome-like complex with a superposed fracture (∼7.5 km long) is interpreted to indicate continued volcanic activity after solidification of a mare deposit (Figures 12g–12i). It stands ∼150–280 m above the mare surface and is ∼29 km in diameter. The dome complex lacks distinct morphologic and spectral signatures to distinguish it from Mare Orientale basaltic material. This complex may have formed during late stage volcanism, perhaps draping of mare basalt over the substrate during cooling and contraction or upwarping by a subsurface magma chamber.
 Another volcanic feature associated with Mare Orientale is a sinuous rille that is cut into the southern rim of the basin interior. Rille A (Figure 13a), just to the south of the southern margin of Mare Orientale, was previously thought to be a channel draining impact melt sheet material [Greeley, 1976]. The new M3 data, however, clearly indicate that the channel was a conduit for mare material. Mafic signatures are seen both at the head of the rille and halfway down the rille, in a location where mare material ponded before continuing to flow into the basin floor (Figures 1c and 13b). This feature is the longest sinuous rille in Orientale (Table 4), originating in the Maunder Formation and extending into Mare Orientale for a distance of 80 km. Its flow path has been significantly influenced by fractures in the Maunder Formation, making the rille less sinuous than others in the basin. Narrowing of the channel along its length and other morphologic characteristics are consistent with thermal erosion of the coherent Maunder Formation, interpreted as impact melt; erosion is predicted to be greatest near the vent (where the lava is at the highest temperature) and decrease down rille, at least in the absence of other changes in setting (background slope or eroding substrate) [Hulme, 1973].
Table 4. Characteristics of Identified Sinuous Rilles in Orientale Basin
 A combination of high-resolution M3 and LOLA data has permitted the identification of several new sinuous rilles (e.g., rilles T, P, and K; Figures 2 and 13). Other previously identified sinuous rilles [Greeley, 1976] have not been confirmed in the new high-resolution data (Figure 2; compare with Figure 2 of Greeley ). Areas reported to have sinuous rilles on the basis of previous data were examined first with M3 data to confirm the existence of sinuous structures. If a definitive determination could not be made, LOLA profiles were analyzed to assess whether or not a trough existed. Several sinuous rilles identified by Greeley  did not appear in LOLA profiles, suggesting that sinuous rilles do not exist in these locations. Table 4 enumerates the different rille characteristics observed (length, width and depth). Most of the rilles in Orientale Basin have shorter lengths than those on the lunar nearside; however, the average calculated length of 30 km is consistent with previous measurements (Figure 14). Schubert et al.  measured lengths of sinuous rilles on the lunar nearside in regions covered by the Orbiter IV high-resolution photographs, finding that the average length was 34 km long, though a few rilles were greater than 300 km in length (Figure 14b).
 Generally in Orientale, sinuous rilles or parts of sinuous rilles that cut into feldspathic material have greater widths compared to rilles in the volcanic mare deposits. This increase in width is potentially the result of mechanical erosion by lava flowing over more poorly consolidated material. Those that cut into both mare and feldspathic material vary in width from ∼0.3 to ∼0.5 km, becoming narrower after transitioning onto mare material. With the current resolution data it is difficult to state whether or not changes in slope play a role in this observed change in width.
 Differences in channel width and type of material being eroded could have significant implications for the dominant type of erosion, either thermal or mechanical [e.g., Carr, 1974; Hurwitz et al., 2010a]. Rilles that have large depressed heads are proposed to be indicative of thermal erosion: as the volcanic material is being erupted, the hottest material lands closest to the vent and causes the most erosion [e.g., Wilson and Head, 1980; Head and Wilson, 1980; J. W. Head and L. Wilson, manuscript in preparation, 2011]. Along with other characteristics, we believe a depressed head suggests that thermal erosion is likely the dominant process in these particular rilles.
 In comparison to rille A (Figure 13a), rille S (Figure 13d) is one of the more sinuous of the lava channels observed in Orientale, likely related to the small elevation difference in pond 19. The walls of rille S have a strong mafic signature (Figure 13e), and it by far the widest sinuous rille (Figure 13f) extending entirely through mare material. Since the width remains fairly consistent along the entire rille it is difficult to say which end is the source. Based on these characteristics we suggest that thermal erosion was the dominant process because it seems likely that excess energy is required to carve a channel that large and sinuous through gradually sloped volcanic material.
 Rille P originates in feldspathic basin material and, although it does not end in a mare deposit, its tail trends toward a small mare deposit east of pond 22 (Figures 13g and 13h). There is little evidence for a residual mafic signature on any observable walls of the rille, likely due to its location on feldspathic material (Figure 13h). The tail end of rille P appears to either disappear into a dark depression, potentially the opening to a roofed part of the lava channel, or end abruptly at the base of the slope (Figure 13g); it is difficult to state definitively without higher-resolution data. Mechanical erosion may have played a larger role in the formation of this sinuous rille, as feldspathic material on the slopes of Orientale massifs and mountain rings appears unconsolidated and prone to mass wasting.
 Another new rille, T, is associated with one of the newly discovered mare deposits (pond 6). This sinuous rille transitions from feldspathic to mafic material, corresponding to a change in its width from ∼1.1 km to ∼0.60 km (Figures 13j–13l). This change in width is gradual, making it difficult to say if it is a product of a change in material or whether other variables, such as a change in slope, may be a controlling factor. The bottom half of the rille still has a strong mafic signature in NIR data (Figure 13k). Compared to other rilles and ponds this system is unique. It is very isolated from other Orientale mare deposits, suggesting that the entire mare pond has been deposited by the rille. Further analysis of rille T will facilitate the placement of constraints on the flux of material through the rille. The formation process for rille T is more difficult to predict. Narrowing and shallowing of the rille as it enters the mare pond could be caused by thermal erosion, where dissipation of heat results in an inability to efficiently melt and remove material, or the observation could be explained by mechanical erosion, with the geologic material at the end of the rille being more resistant to this process. Future analyses, which focus on high-resolution topography and morphology [e.g., Hurwitz et al., 2010b] and spectral data, will help resolve many of the remaining ambiguities associated with sinuous rille formation.
 Our current study broadly agrees with the classification scheme of Greeley , but the new high-resolution M3 data indicate that Mare Orientale contains several local sources, including a sinuous rille in the southern part of the basin as well as additional complexities observed in domes and related structures. Sinuous rilles and lava channels at least partially contributed to the formation of Lacus Veris and Lacus Autumni. More detailed analysis is necessary to distinguish between constructional lava channel and sinuous rille formation processes, and between the relative roles of thermal and mechanical erosion.
10. Altitude Distribution of and Deformation of Mare Basalts
 The LOLA instrument onboard LRO is providing the highest-resolution altimetric data yet obtained for the Moon [Smith et al., 2010]. Coupling this topographic data with age and compositional data can aid in deciphering the relationship between multiring basin formation and volcanic activity. The average elevations of the mare ponds (Table 2) accentuate the stair-like shape of the basin from Mare Orientale out to the Cordillera Mountain ring ∼8 km above (Figures 1a, 15, and 16). Mare Orientale is the lowest lying basaltic deposit, followed by the ponds of Lacus Veris and, lastly, those of Lacus Autumni. A plot of the average pond elevations versus their ages (Figure 17) shows the order of emplacement as it relates to the basin morphology. A very broad trend shows the youngest mare deposits being higher in elevation, which corresponds with greater distances from the center of the basin. Mare Orientale was the first and lowest basalt deposit, followed generally by the basalts of Lacus Veris, by and large being emplaced from the west to the east and eventually succeeded by the ponds of Lacus Autumni. Notably, Lacus Veris has ponds that span a large portion of the duration of volcanism, while Mare Orientale only erupted at the beginning of the basin history and Lacus Autumni erupted near the end of volcanism in the basin.
 The surfaces of each of the ponds were investigated for any suggestion of a regional tilt by measuring topographic profiles from a LOLA DEM (Figure 15). Current data show no indication of a significant radial tilt inward toward the center of the basin or an east-west or north-south regional tilt. The only consistent trend was found in the ponds in Lacus Autumni, which were found to all tilt toward the center of the basin, but the scale of the tilt was so small, on the order of thousandths of a degree, that it appears insignificant. Examination of suspected small-scale radial tilts in the mare ponds showed that many were the result of local topography, for example, from mass wasting and impact cratering. Thus, to a first order, the interior of the Orientale Basin did not appear to undergo regional subsidence and tilting subsequent to the emplacement of Lacus Veris and Autumni.
 An interesting feature observed in the LOLA data is the profile of the mare polygon to the southwest of Mare Orientale. Its northern end dips in the direction of Mare Orientale (Figure 15). The mare appears to have been deposited within the fractured polygon, and later tilted toward the center of the basin. This tilt is likely to be the result of the thermal subsidence that created the inner depression in Orientale Basin. The formation of compressive features (e.g., wrinkle ridges) and uneven cooling underneath the center of the basin could result in some areas subsiding more than others.
 Mare deposits within Mare Orientale, however, have been significantly deformed by a series of wrinkle ridges (Figure 15) extending across the central part of the basin. There is a series of 2–3 ridges in the south central part of the basin that parallel preexisting basin topography for ∼130 km. Near the northern end of this group of ridges another ∼190 km long wrinkle ridge begins in the western part of Mare Orientale. These features are superposed on topography associated with the modification of the surface of the Maunder Formation. Scarps and fractures indicate that cooling of the impact melt sheet, thermal equilibration of the raised isotherms and impact-related heat, together produced up to ∼3 km of subsidence in the period following the basin-forming event (Figure 15). The superposition of the wrinkle ridges on Mare Orientale indicates they were produced after that mare deposit was emplaced. Lack of prominent wrinkle ridges in Lacus Veris and Autumni suggest that this compression in the center of the basin occurred before their emplacement. Based on our crater retention ages, this dates the compressional event to between 3.58 and 3.47 Ga.
 Outside of Mare Orientale, there is one wrinkle ridge (∼10 km long) in the southern portion of pond 22. Similar to those in Mare Orientale, this wrinkle ridge had to occur after the mare in pond 22 was emplaced, which in this case was ∼1.66 Ga. This little ridge might be the manifestation of compressional stresses on the outside edge of the basin, created just after the last phases of volcanism in the basin.
 A ∼340 km long circumferential graben occurs just inside of Lacus Veris. This extensional feature can be seen where it cuts into the Maunder Formation and is embayed by the mare deposits of Lacus Veris. These stratigraphic relations indicate that the graben formed before the emplacement of pond 14 in Lacus Veris, but after the Maunder Formation cooled. Remnants of another graben can be seen in pond 12. Here a similar situation exists, where the graben is cutting into the feldspathic basin material while being covered in certain areas by mare material. Thus, based on stratigraphic relationships, these graben likely formed between 3.64 Ga and 3.36 Ga.
11. Evaluation of Models for Formation and Evolution of Mare Basalts
 Analyzing the various characteristics (e.g., ages, area, volumes, elevations, spectral compositions and associated volcanic features) of Orientale Basin has assisted in further understanding the relationship between the formation of lunar impact basins and the onset of mare volcanism. In comparison to the other lunar basins, Orientale offers great insight into the initial stages of mare emplacement due to its incomplete flooding. With the latest surge of lunar data, from instruments such as M3 and LOLA, new information has been obtained to help clarify this relationship between large lunar basins and volcanism. We now use this synthesis data to assess models of mare basalt genesis, ascent and eruption.
11.1. Model 1
 Model 1 of the origin and evolution of mare basalts assumes that the nearside-farside asymmetry in the distribution of mare basalts is due to differences in crustal thickness (Table 1). Orientale Basin provides a view into the importance of the crustal thickness differences since it is situated on the boundary between the thick lunar farside and the thinner nearside crust. Despite the discovery of several new deposits to the south and southwest of Orientale, the general trend in the abundance of mare deposits, increasing from the southwest to the northeast is still prevalent. Our data show that the genesis of mare basalt was symmetric over the lunar surface. Differences in the distribution of mare lies in the difficulty of getting melt products to the surface due to thermal and buoyancy barriers.
11.2. Model 2
 In model 2 [Elkins-Tanton et al., 2004], basin-forming impacts induce pressure release melting and associated secondary convection to explain the distribution of lunar mare basalts (Table 1). The first stage predicts large volumes of mare basalts in the center of the basin (98–100% of the total melt), followed by smaller amounts of basalts (1–2% of the total melt) emplaced over a much longer period of time. Calculated volumes of mare basalts in Orientale Basin suggest Mare Orientale contains ∼94% of the total mare by volume and the rest (∼6%) resides in Lacus Veris and Autumni. Despite this general agreement in the locations and their relative volumes of mare, dating the basaltic surfaces in Orientale has shown that this process of pressure release melting is not consistent with the predicted timing. The basin and melt sheet are dated between 3.68 and 3.64 Ga, while central Mare Orientale was not emplaced until 3.58 Ga, a time gap of ∼60–100 Ma (Table 3). This time gap shows that the extrusion of Mare Orientale was not instantaneous and therefore, was not the product of impact induced pressure release melting.
11.3. Model 3
 Another model (model 3) [Wieczorek and Phillips, 2000] suggests that the presence of an enhanced sub-Procellarum KREEP layer explains the existence of mare basalts and viscous domes in the western nearside PKT (Table 1). Signifiers of this influence on the thermal evolution of the Moon are manifested as “red spot” volcanism, which has been associated with KREEPy rock compositions [Malin, 1974; Chevrel et al., 1999]. “Red spot” volcanism on the lunar nearside has been identified as a postbasin, premare event. This nonmare, spectrally distinct expression of volcanism appears to be absent from both the edges of Orientale Basin and the mare pond deposits, and no radioactive elemental anomalies have been detected. All volcanic features and deposits appear basaltic in composition.
11.4. Model 4
 Instabilities in the lunar mantle, caused by large-scale overturn are another mechanism suggested to control the emplacement and evolution of mare basalts [Hess and Parmentier, 1995] (Table 1). Unlike previous models, model 4 is completely independent of basin formation processes. The presence of early high-Ti basalts, superposed on basalts with lower titanium values would support this model. However, in Orientale the composition of the mare basalts appears to be fairly homogeneous despite the eruption of basalts over a period of ∼1.9 Ga. The basalts in Mare Orientale have medium- to high-Ti contents, and Lacus Veris and Autumni have relatively high titanium values. In the case of Orientale Basin it does not appear that any low-Ti deposits were emplaced after the medium-to high-Ti (3–7 wt % TiO2) deposits that are observed.
11.5. Model 5
 In model 5, the emplacement of mare basalts is related to both the global thermal evolution and basin evolution (Table 1). Rather than relying on the basin-forming event, this model depends on local and global lunar stress fields as well as the geometry of the lunar basins. Based on the model predictions, local stresses control the location of deposits in the lunar basins. Early mare volcanism will be focused in the center of the basin and the youngest deposits should occur along the edges. As the lunar lithosphere cools, its thickness increases with time. The increasing density of the cooling, thickening lithosphere increases the load on the center of the basin, consequently creating the local extensional stresses on the exterior of the basin. Unloading-related extensional stresses are greatest immediately after the basin-forming impact and decrease with time as the lunar lithosphere thickens and evens out underneath the entire basin. Eventually global compressive stresses, from cooling of the Moon, dominate local extensional stresses and lead to the termination of the ascent and eruption of mare basalts.
 Model age dates from this study confirm that the earliest volcanism did occur in the center of the basin (Table 3) and was independent of the basin-forming impact, evidenced by the time lag between basin formation and emplacement of Mare Orientale. The younger deposits in Lacus Veris and Lacus Autumni show a weak positive correlation between age and increasing distance from the center of the basin; deposits young as they are emplaced further from the basin center. In addition, the areal extent of the deposits decreases with distance from the center of the basin, possibly because there is more crust to transect (Figures 15 and 16).
 Analysis of the character and distribution of mare deposits in the Orientale Basin has provided key information on the relationship between the formation of large lunar impact basins and mare basalt volcanism. Results from our tests of the five models indicate the following.
 1. The limited distribution of mare basalts in Orientale provides clues to the initial stages of mare basalt filling. Mare filling began in the center of the basin ∼3.58 Ga, approximately a hundred million years after basin formation (∼3.68 Ga), assuming this fill is the result of a single volcanic event. The central deposit erupted as a flood basalt and was followed closely by later pulses of volcanism that produced the mare ponds in Lacus Veris and lastly those in Lacus Autumni (Table 3). Estimated flow unit volumes range from 30 to 7,700 km3, averaging between 590 and 940 km3 [Hiesinger et al., 2002]. This time sequence correlates well with the theory that early flooding is dominated by larger deposit thicknesses over a smaller area, and later flooding is opposite in character, with shallower deposits that cover a larger area [Head, 1982]. The maximum difference of 50–130 Ma between the basin forming impact event and the average age of Mare Orientale represents a significant time gap and argues against near instantaneous depressurization melting related to the impact event.
 2. Model age dating has shown that the youngest mare deposits occur on the edges of the basin (i.e., Lacus Autumni), varying in age from 3.47 to 1.66 Ga. A few ponds in Lacus Veris have crater retention ages within this range as well. Generally, lava flow emplacement occurred first in Lacus Veris and finally in Lacus Autumni, with some overlap. Their abundance of sinuous rilles indicates they represent a basaltic plains style volcanism dominated by both point sources (e.g., vents and rilles) and fissures. The trend between the model ages and elevations of mare ponds (Figure 17) shows how the youngest ponds are both the furthest from the center of the basin and at the highest elevations. This observation supports the idea that extensional stresses within the basin propagate radially outward over time. The correlation with the postbasin state of stress in the lithosphere favors dike emplacement in the ring regions where annulus-normal basin stresses are likely, as the significant remaining basin topography adjusts thermally and mechanically [Bratt et al., 1985a].
 3. The model age distribution of Orientale mare basalts shows that they span nearly the entire range of nearside model ages but that the volumes are considerably smaller than for the mare filling typical nearside impact basins.
 4. Model age data indicate that Orientale mare basalts were erupting onto the surface over a significant time period (∼1.9 Ga; Table 3). The majority of them were deposited during the highest frequency of eruptions on the lunar nearside, around 3.5 Ga (Figure 10) [Hiesinger et al., 2011].
 5. Mare Orientale was determined to be the result of a single eruptive phase, lacking any compelling compositional evidence for the presence of multiple distinct flow units. Lacus Veris and Autumni appear similar in composition to Mare Orientale, but display spectral characteristics that could be indicative of both feldspathic mixing and slight differences in basalt composition. Most of the mare deposits in Orientale appear to have a medium- to high-Ti content (3–7 wt % TiO2 from Kadel ).
 6. Mare eruption locations are focused along the margins of the different basin rings, mostly in the form of vents and sinuous rilles. Most of the obvious vents and all of the rilles are located on the eastern side of the Orientale Basin. No rilles exist on the southwestern side of Orientale possibly because the crustal thickness there was too great to allow for the propagation of dikes to the surface, hindering the opening of vents and extrusion of significant quantities of mare basalt over a sustained time period. Thus, the minor deposits present to the west and southwest of Orientale are likely the result of small fissures or vents that have since been covered up by the deposits themselves.
 7. There is no evidence in Orientale Basin for “red spot” extrusive domes and related deposits. Thus, there does not appear to be any KREEPy basalt extrusions or aluminous basalts.
 8. Kopff crater, previously interpreted as a volcanic caldera [e.g., Pai et al., 1978], is shown to be a volcanically modified impact crater. However, the unusual character of its morphology suggests an impact into central basin material of unusual thermal properties.
 M3 is supported as a NASA Discovery Program mission of opportunity. Both the science results and the science validation are supported through NASA contract NNM05AB26C. Thanks are extended to members of the M3 team for assistance in the preparation of this manuscript. Additionally, the M3 team is grateful to ISRO for the opportunity to fly as a guest instrument on Chandrayaan-1 and is grateful to all on the ISRO team that enabled M3 data to be returned. We gratefully acknowledge the valuable input to this project by the public release of LRO Laser Altimeter data.