Journal of Geophysical Research: Planets

RESEARCH ARTICLE Geologic History of the Northern Portion of the South 10.1029/2018JE005590 Pole-Aitken Basin on the Moon Special Section: M. A. Ivanov1,2 , H. Hiesinger2 , C. H. van der Bogert2, C. Orgel3 , J. H. Pasckert2 , Planetary Mapping: Methods, and J. W. Head4 Tools for Scientific Analysis and Exploration 1V.I. Vernadsky Institute of Geochemistry and Analyitical Chemistry Russian Academy of Sciences, Moscow, Russia, 2Institut für Planetologie, Westfälische Wilhelms-Universität, Münster, Germany, 3Division of Planetary Sciences and Remote Sensing, Freie Key Points: Universität, Berlin, Germany, 4Earth, Environmental and Planetary Sciences, Brown University, Providence, RI, USA • A geological map of the northern portion of South Pole-Aitken basin was made Abstract We conducted a detailed photogeological analysis of the northern portion of the South • Confirmed and potential volcanic – – materials make up ~8% of the basin Pole-Aitken (SPA) basin (10 60°S, 125 175°W) and compiled a geological map (1:500,000 scale) of this floor region. Our new absolute model age for the Apollo basin, 3.98 + 0.04/À0.06 Ga, provides a lower age limit for • The pre-SPA lunar crust was likely the formation of the SPA basin. Some of the plains units in the study area were formed by distal ejecta from stratified with respect to iron remote craters and basins. The characteristic concentrations of FeO and TiO2 of other plains are indicative of their volcanic origin. The oldest volcanic materials occur near the center of the SPA basin and have an Early Imbrian age of ~3.80 + 0.02/À0.02 Ga. Late Imbrian volcanic activity occurred in and around the Apollo basin. Correspondence to: In total, the volcanic plains cover ~8% of the map area and cannot account for the extensive SPA iron M. A. Ivanov, fl – [email protected] signature. The sources of the signature are the oldest materials on the SPA oor (FeO ~11 14.5 wt%). In contrast, the ejecta composing the SPA rim are significantly poorer in FeO (<7.5 wt%). The signature could be related to the differentiation of the SPA impact melt. However, the spatial segregation of the ancient iron-rich Citation: Ivanov, M. A., Hiesinger, H., van der and iron-poor materials suggests that the SPA iron signature predated the basin. Thus, the signature Bogert, C. H., Orgel, C., Pasckert, J. H., & might be explained by a pre-SPA lunar crust that was stratified with respect to the iron concentrations, so that Head, J. W. (2018). Geologic history of the SPA impact excavated the upper, iron-poorer portion of the crust to form the SPA rim and exposed the the northern portion of the South fl Pole-Aitken basin on the Moon. Journal deeper, iron-richer portion on the oor of the basin. of Geophysical Research: Planets, 123, 2585–2612. https://doi.org/10.1029/ Plain Language Summary We conducted a geological analysis of the northern portion of the South 2018JE005590 Pole-Aitken basin, which is the largest recognized and likely the oldest impact structure on the Moon. Results of our mapping efforts permitted the unraveling of the major sequence of impact and volcanic events Received 28 FEB 2018 that have shaped the basin throughout its evolution and resulted in the discovery of the oldest materials Accepted 13 SEP 2018 Accepted article online 17 SEP 2018 related to the basin formation. Analysis of the distribution and concentrations of iron and titanium in the Corrected 31 OCT 2018 materials of different age within the South Pole-Aitken basin allows the characterization of the structure of Published online 12 OCT 2018 the ancient lunar crust and mantle. These results introduce important constraints on the current models of

This article was corrected on 31 OCT the early evolution of the Moon. 2018. See the end of the full text for details. 1. Introduction We studied a region between 10°–60°S and 125°–175°W that represents the northern portion of the South Pole-Aitken basin (SPA), centered on the Apollo basin (Figure 1a). Several important characteristics of the SPA region and, in particular, its northern extent allow us to address a number of key questions regarding lunar evolution, including the constraints that SPA can place on the formation of the lunar magma ocean, the structure of the lunar crust, the formation and evolution of large impact basins, and the history of post-SPA volcanism and cratering. First, the SPA is the largest impact structure on the Moon with dimensions of ~2,400 by 2,000 km (Garrick-Bethell & Zuber, 2009; Hiesinger & Head, 2004; Shevchenko et al., 2007; Spudis et al., 1994; Stuart-Alexander, 1978) and likely is the oldest recognized lunar impact structure (Hiesinger et al., 2012; ©2018. The Authors. This is an open access article under the Wilhelms, 1987). SPA has a discernible morphologic appearance (Stuart-Alexander, 1978; Wilhelms et al., terms of the Creative Commons 1979) and exhibits a prominent topography (Zuber et al., 1994). These characteristics imply that the SPA basin Attribution-NonCommercial-NoDerivs formed within a solid lunar crust. Thus, constraining and/or establishing the age of the SPA will provide an License, which permits use and distri- bution in any medium, provided the important constraint on the time scale of the evolution of the magma ocean on the Moon. original work is properly cited, the use is – non-commercial and no modifications Model simulations of a large impact event at the angle 30° 60° suggest that an SPA-scale event should or adaptations are made. penetrate through entire lunar crust and excavate mantle materials (Melosh et al., 2017). Yet the region

IVANOV ET AL. 2585 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 1. Study area that shows location of the Apollo basin (white dashed line) near the rim of the South Pole-Aitken (SPA) basin and large impact craters for reference. The background is a mosaic of wide-angle camera (WAC) images with reso- lution 100 m/pixel. Map in the orthographic projection centered at 35oS, 150oW.

inside the SPA does not show strong spectral signatures of olivine (Lucey, 2004), which has been thought to be the major component of the lunar mantle. Recent Kaguya data have observed some local but limited olivine occurrences (Yamamoto et al., 2010). The scarcity of the olivine exposures inside the SPA basin could be explained several ways. It may be that the lunar mantle consists predominantly of low-calcium pyroxene (Melosh et al., 2017). Alternatively, the basin may have been formed during an oblique (30°) impact that involved impactor decapitation (Schultz, 1997; Schultz & Crawford, 2011) such that the impactor was unable to penetrate as deeply into the mantle. An oblique SPA impact and its correspondingly shallower excavation depth seem to be consistent with the presence of a Th anomaly inside the SPA basin (Lawrence et al., 2000). Although the Th concentrations on the SPA floor are lower than within the Oceanus Procellarum KREEP Terrane or the Compton-Belkovich region (Jolliff et al., 2011; Lawrence et al., 2007), they are systematically higher than the typical Th abundance in the feldspathic highlands terrane (Jolliff et al., 2000). This may suggest the presence of a KREEP component in the SPA region. The KREEP-enriched material is thought to be the magma ocean residuum that was sandwiched between the lunar crust and mantle (Jolliff et al., 2000), which potentially could be exposed by the SPA event. Another possibility is the formation and differentiation of a large pool of impact melt inside the SPA basin (Hurwitz & Kring, 2014; Vaughan & Head, 2014). The fractional differentiation of a large volume of melt, if it was undisturbed and fully developed, should cause the separation and sinking of denser olivine cumulates and the concentration of a norite-like component at the top of differentiating body. The SPA region also exhibits a distinct iron signature that has been described using Galileo, Clementine, and Lunar Prospector data (Belton et al., 1992; Pieters et al., 2001; Lawrence et al., 2002). The iron abun- dance within the SPA basin, ~10–12 wt% (Lawrence et al., 2003), is noticeably higher than in its

IVANOV ET AL. 2586 Journal of Geophysical Research: Planets 10.1029/2018JE005590

immediate surroundings and is similar to the FeO concentrations in the Schickard-Schiller and Mare Australe regions where cryptomare deposits were detected (Antonenko & Head, 1994; Antonenko et al., 1995; Mustard et al., 1992; Lawrence et al., 2015; Whitten & Head, 2015). The enhanced concentration of FeO on the floor of the SPA basin was interpreted as evidence for the presence of extensive cryptomaria in this area (Gibson & Jolliff, 2011). However, although the entire SPA interior is a region of very thin crust—thinner even than beneath the major mare basalts on the lunar nearside in Oceanus Procellarum (Wieczorek et al., 2013)—the floor of the SPA shows only a minor extent of volcanic plains. The presence of the small and scattered patches of the low albedo, presumably volcanic plains (Head, 1976; Head & Wilson, 2017; Pasckert et al., 2018; Stuart- Alexander, 1978; Wilhelms et al., 1979; Yingst & Head, 1999) on the floor of the SPA basin (Figure 1) connects our study area with another fundamental problem of lunar geology: the history of volcanism on the Moon. Characterization of associations of iron-rich materials with specific rock stratigraphic units is, thus, important for the assessment of the timing and extent of volcanic activity in the SPA region. Here we also encounter questions about the structure and extent of the lunar crust. The idea that crustal thickness is a primary control on the spatial distribution of lunar volcanic materials (Head & Wilson, 1992) is at odds with the limited number and relatively small occurrences of lava plains within the SPA interior, because this area is among the deepest topographic depressions on the Moon and represents an area of thin- ner crust (Wieczorek et al., 2013). Thus, it is likely that factors other than crustal thickness played a more important role in the asymmetry of volcanic activity. These factors include (1) a higher concentration of the basins on the nearside, as recently documented in high-resolution gravity data from the GRAIL (Gravity Recovery and Interior Laboratory) mission (Miljkovic et al., 2013), (2) a globally asymmetric distribution of the KREEP (potassium, K, rare earth elements, REE, and phosphorous, P) component causing global variations in volcanic activity (e.g., Wieczorek & Phillips, 2000), (3) the absence of deep-seated structures on the lunar farside that would allow migration of magma to the surface, perhaps related to the formation of SPA basin itself (Schultz & Crawford, 2011), and (4) the stripping of crust by the SPA impact that allowed faster cooling of the interior of SPA, compared with the lunar nearside, thus blocking magmatism (Pasckert et al., 2018). Specific questions that we examine in our study and that are of key importance for understanding lunar geology are as follows. What is the nature of the iron and thorium signatures and their association with the geological units in SPA basin? Do they reflect the preexisting structure of the ancient lunar crust or are they related to volcanic activity in the SPA? Why does the ancient, extensive region of the thinner crust inside the SPA show evidence for minor volcanism in comparison with the Oceanus Procellarum region? Can both mare and cryptomare on the floor of the SPA account for the enhanced concentration of iron in this area? What is the age (or range of ages) of volcanism within the SPA? What is the time lag between the formation of the basin and the emplacement of lava plains in its interior and vicinity? In addition, the robust interpreta- tion of the compositional data requires a thorough photogeologic analysis that provides a link between the surface compositional characteristics and specific material units of known stratigraphic position. Thus, the primary goal of our investigation is the compilation of a detailed geological map of the study region (Figure 1) based on new, high-quality images, which shows the distribution of material units in space and time and provide the basis for interpretation of the remote sensing compositional data.

2. Data and Methods We conducted our study of the northern portion of the SPA basin on newly available medium- and high- resolution images, and topographic measurements in combination with other data types, such as spectral data. These include the following: (1) Lunar Reconnaissance Orbiter (LRO) wide-angle camera (WAC) images with a pixel resolution of 100 m (Robinson et al., 2010). A mosaic of WAC images is the main photographic base for our photogeological study. (2) High-resolution topography data from the Lunar Orbiter Laser Altimeter (LOLA) on board LRO (Smith et al., 2010). In our study, we used the gridded topographic data with a resolution of 64 pixel/deg, which corresponds to ~470 m/pixel at the equator and ~390 m/pixel at the mean latitude of the study area (35°S). (3) Spectral data taken by the Clementine mission (Pieters et al., 1994) and

processed into global maps of FeO and TiO2 distributions (Blewett et al., 1997; Lucey et al., 1998, 2000). We used these data to assess the model concentrations of FeO and TiO2 in the units we defined and mapped based on their morphological characteristics.

IVANOV ET AL. 2587 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 2. Assessment of the number of random points that is required for a representative sampling. In all cases, dashed lines show the cumulative frequency distribution of all Lunar Orbiter Laser Altimeter data points in our study region (original population, ~10,000,000 points). Solid lines indicate the distribution of points selected at random from the original population. Number of points in the random samples is shown in the upper right corner of each plot. When number of points in a random sample approaches ~0.1% of the original populations (c), the distribution of points becomes indistinguishable.

To analyze the compositions of different units that we defined, we collected topographic and spectral data for pixels within each mapped unit using the random point function within ArcGIS, which allows the generation of a specified number of points that are randomly distributed within a specified sampling area. One challenge is to determine the number of required points to provide a representative sample for each investigated attribute. Thus, we estimated the necessary number of points experimentally, by first using the LOLA gridded topographic data and collecting all the pixels available within the map area (~10,000,000 points). Second, we sampled this area with an increasing number of random points that com- prise 0.01%, 0.05%, 0.1%, and 0.5% of the total number of points and compared the cumulative distribution of the randomly sampled topography with those resulting from the sampling of all points (Figure 2). The com- parison of the distributions shows that when the number of sampling points reaches 0.5% of the total popu- lation (Figure 2d), both the true and the model distributions become indistinguishable, and a null hypothesis

IVANOV ET AL. 2588 Journal of Geophysical Research: Planets 10.1029/2018JE005590

that they are drawn from different populations can be rejected (Kolmogorov-Smirnov test) at a high confi- dence level (0.999). However, in order to avoid possible biases of random sampling in our study, we increased the number of random points to 20% of the total population of pixels within each of our mapped units. Such a high proportion ensures that our analysis is unbiased and representative.

The estimation of the FeO and TiO2 contents is more complicated, because the Clementine compositional maps were not topographically corrected. In places, the absence of such a correction disturbs the coregistra- tion of the units defined morphologically and their representation on the compositional maps. This effect is more pronounced within higher elevation regions but remains practically negligible in the low-lying areas. To account for the shift between the morphological features in the WAC mosaics and their signatures in the Clementine maps, we used prominent impact craters that are well seen in both types of data. In cases when the locations of the same crater were different in the WAC images and the Clementine maps, we analyzed only those points that occur within the overlapping portions of the WAC mosaics and Clementine maps. We also excluded points that occur within either illuminated or shadowed slopes. Although this selection decreases the populations of the measurement points, their number still remains large enough (more than

10% of the total number) to provide a representative sample of either FeO or TiO2 content in the mapped units. Using the well-established procedures developed in previous photogeological studies of the Moon and the planetary bodies (e.g., Wilhelms, 1987 and references therein), we defined morphologically uniform units and mapped them to compile a geomorphological map of the study region. We then analyzed the stratigraphic relationships of these units and determined their absolute model ages (AMAs) using crater size-frequency distribution (CSFD) measurements (Michael & Neukum, 2010; Neukum et al., 2001). We used the ArcGIS CraterTool (Kneissl et al., 2011) to map count areas and measure primary craters (>0.3 km); obvious clusters of secondary craters were excluded from the count areas. The relative and absolute dating of the units allowed the transformation of the geomorphological map into a geological map (1:500 000 scale) that shows the spatial and temporal distribution of processes that took place in the map region.

3. General Topography of the Study Area The study area, located within the northern part of the SPA basin, extends from its apparent center (SW corner of the map area) to include the northeastern segment (~1,600-km long) of the basin rim (Figures 1 and 3). The total topographic range in this region is ~17.8 km (from about À8.6 km to about 9.2 km), but the difference between the most common topographic levels on the floor and the rim of the basin is ~8.5 km (Figure 3b). The hypsographic curve or hypsogram within the study area is strongly bimodal (Figure 3b) indicating that the entire region consists of two topographic domains: the floor and the rim of the SPA basin. The minimum of the hypsogram corresponds to the zero contour line; thus, the SPA floor domain is below the mean planetary radius (MPR) and the SPA-rim domain is above the MPR. The bimodal topographic distribution is a natural phenomenon for an area that overlays both the floor and the rim of a large impact structure with a relatively steep inner wall. However, the division of the study area into two dif- ferent topographic domains is useful for our study, because both the floor and the rim domains may repre- sent regions consisting of different types of materials. The floor of the SPA basin is slightly tilted toward the basin center at a steady slope of ~0.2° (Figure 3c). High- frequency topographic variations with amplitudes of ~0.5–1 km characterize a large portion of the floor. These features represent remnants of ancient cratered terrains. Plains surfaces that often occur within larger impact craters are much more level with typical topographic variations of less than 100–200 m. The most pro- minent topographic features on the SPA floor are large (100- to 200-km diameter) impact craters that are 2- to 3-km deep (Figure 3c). Besides the largest and deepest impact craters, the most prominent topographic features of the SPA rim domain are segments of relatively narrow (40- to 50-km wide) ridges that are 5- to 6-km high (Figure 3a,c). In our study area, the segments of the ridges are aligned concentrically to the basin center near the transition from the rim to the floor; a broad (100- to 150-km wide) and deep (5- to 6-km deep) U-shaped trough sepa- rates the two ridge segments (Figures 3a and 3c). Both the ridge segments and the trough form a consistent pattern of topographically positive and negative features that outline the northeastern portion of the SPA rim and likely reflect the original configuration of the rim.

IVANOV ET AL. 2589 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 3. (a) Regional topography of the study area. Black dashed line indicates the Apollo basin, solid lines show position of topographic profiles. The background is the Lunar Orbiter Laser Altimeter topographic map, with resolution 64 pixel/deg. Map in the orthographic projection centered at 35°S, 150°W. (b) The hypsogram for the study area shows two principal topographic domains, the South Pole-Aitken (SPA) floor and rim, that were formed after the SPA event. (c) A topographic profile across the floor and the rim of the SPA. Lunar Orbiter Laser Altimeter topography with resolution 64 pixel/deg.

The multiring Apollo basin (~550-km diameter, Figures 4a and 4b) is the most extensive topographic depres- sion within the study area (Figure 1b) (Baker et al., 2017; Guo et al., 2017; Potter et al., 2018). The majority of the basin is within the SPA floor topographic domain, but the northeastern portion of the Apollo basin outer rim overlays the SPA rim domain (Figures 3a and 4a). Thus, the location of the Apollo basin results in a

IVANOV ET AL. 2590 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 4. Topographic profiles across the Apollo basin. (a) The profile running in the direction from west to east. (b) The profile crossing the basin in the direction from north to south. The profiles show the overall asymmetric configuration of the Apollo basin because its northeastern portion is on the South Pole-Aitken (SPA) rim and the southwestern portion of the basin is on the SPA floor. Lunar Orbiter Laser Altimeter topography with resolution 64 pixel/deg.

pronounced difference of the heights of different portions of its outer rim, which is ~3- to 5-km high within the SPA floor domain and ~7–8 km within the SPA rim domain (Figures 4a and 4b). The inner rim of the Apollo basin is seen only in the west and northeast of the basin floor and disappears both morphologically and topographically in the southeast, south, and northwest of the floor (Figures 1 and 3a). In places where the inner ring is prominent, its height reaches ~1–2 km above the surrounding terrains (Figure 4b). The surfaces between the outer and inner rims of the Apollo basin are mostly at about À5-km topographic level and morphologically smooth with small (a few hundred meters) topographic variations (Figures 4a and 4b).

4. Definition of the Morphological Units In the study area, there are two major classes of landforms: (1) impact craters and related features and (2) plain-forming terrains that represent both volcanic activity (lava flooding) and impact cratering (emplace- ment of distal ejecta from impact craters/basins).

4.1. Impact Craters and Related Features In our study, we used the characteristics of impact craters, including the presence or absence of rays, discern- ible ejecta blankets, and chains of secondary craters, as well as the preservation state of the crater rims to classify the craters and crater-related landforms into several morphological categories. 4.1.1. Fresh Crater Landforms Craters with prominent ray systems and ejecta blankets (Figure 5a): These craters have sharp rim crests, ejecta with a visible hummocky texture, and long chains of secondary craters with a crisp morphology. There are only a few tens of such craters in the study area, the largest of which is Crookes (10.4°S, 164.9°W), which has a diameter of ~45 km (Figure 5a).

IVANOV ET AL. 2591 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 5. Examples of the freshest craters in the map area. (a) Craters with prominent ejecta that have prominent ray systems; crater Crookes, ~45-km diameter (see Figure 1 for location). (b) Craters with discernible ejecta but without rays; crater Dryden, ~52 km diameter (see Figure 1 for location). (c) Craters with sharp-crested rims but without discernible ejecta and rays; crater Wilsing C, ~32-km diameter (18.5°S, 152.9°W). All images are portions of the wide-angle camera mosaic with resolution 100 m/pixel.

Craters with discernible ejecta, but without rays (Figure 5b): These are typically large craters with diameters many tens of kilometers; the largest of them is crater Ioffe (14.52°S, 129.10°W, 78 km). These craters are sur- rounded by visible ejecta and sometimes by extensive fields of secondary craters with diameters as large as 10–15 km. The largest craters of this category show sharp rim crests, terraced inner walls, and patches of smooth, light-toned plains on their floors. There are only 10 such craters in the study region. Two of them (Finsen and Alder, Figure 1a) are outside of the map area, but chains of their secondary craters affect the sur- face within the map area. Craters with sharp-crested rims, but without discernible ejecta and rays (Figure 5c): These craters are typically small (diameters are 20–30 km and smaller) and abundant in the study area (we mapped about 600 such craters). 4.1.2. Degraded Crater/Basin Landforms Craters without extensive ejecta blankets (Figure 6a): These are typically larger craters with diameters of sev- eral tens of kilometers that have a prominent, mostly completed rim and exhibit only a proximal ejecta deposit near the rim (e.g., crater Langmuir, 35.80°S, 128.78°W, 85 km diameter; Figure 6a). The largest

IVANOV ET AL. 2592 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 6. Examples of degraded craters, and crater and basin materials in the map area. Specific morphological units are outlined by dotted lines and indicated by arrows. (a) Craters without extensive ejecta blankets; crater Langmuir~85-km diameter (see Figure 1 for location). (b) Massifs of the Apollo basin rim (center of the image is at 36.2°S, 155.6°W). (c) Massifs of the South Pole-Aitken basin rim (center of the image is at 21.7°S, 160.1°W). (d) Undivided, strongly degraded craters (center of the image is at 17.9°S, 172.5°W). All images are portions of the wide-angle camera mosaic with resolution 100 m/pixel.

crater of this category is Oppenheimer (~200-km diameter). These craters are abundant and typically merge with each other. In these cases, their rims and the immediately surrounding ejecta form extensive occurrences whose areas can be as large as a few thousands of square kilometers. Craters of this morphological category make up the majority of the map area (Table 1). Massifs of the Apollo rims (Figure 6b): The Apollo rim massifs occur as elongated and equidimensional ridges with blocky and heavily cratered surfaces. They form the outer and inner rims of the basin. Massifs of the SPA rim (Figure 6c): Morphologically, these features are similar to but more degraded than the Apollo rim massifs. They correspond to the prominent double topographic ridge-like features within the SPA rim topographic domain (Figure 3a). The SPA rim massifs form tall, long ridges with heavily cratered and scal- loped surfaces within the northeastern portion of the SPA rim near its transition to the SPA floor. The SPA rim massifs do not occur on the outer slopes of the SPA rim. Strongly degraded crater materials, undivided (Figure 6d): This morphological unit includes rugged surfaces consisting of chaotically oriented short ridges and equidimensional blocks several kilometers across and hun- dreds of meters high. Both the ridges and blocks likely represent remnants of the oldest craters that have

IVANOV ET AL. 2593 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Table 1 been destroyed during the formation of younger craters. The strongly Areas of Map Units degraded craters are usually exposed between the other crater-related Area Area features and occur within both the SPA rim and SPA floor 2 Unit type Unit (km ) (%) topographic domains. Entire map area Crater materials Cc, Copernican craters 43,097 2.4 4.2. Plain-Forming Units Ec, Eratosthenian craters 117,706 6.5 Within the map area, three types of terrains form patches of UIc, Late Imbrian craters 158,488 8.7 LIo, Early Imbrian Orientale ejecta 20,484 1.1 morphologically smooth plains that are tens to hundreds of kilometers LIc, Early Imbrian craters 57,141 3.1 across. These units have different albedos and characteristic NpNc, Nectarian and 652,624 35.8 topographic features. pre-Nectarian craters Plains Ilp, Imbrian light plains 66,586 3.7 Low-albedo (dark) plains (Figure 7a): These plains have an albedo that is Total 1,116,126 61.3 noticeably lower than that of the surrounding terrains. The dark plains SPA floor domain form relatively small patches several tens of kilometers across that occur Nonplains pNrm_APL, pre-Nectarian 40,966 2.2 on the floors of impact craters. The largest concentration of the dark plains Apollo rim massifs is inside the Apollo basin, within its center, and between the inner and pNm_SPAf, pre-Nectarian 190,381 10.4 SPA floor materials outer rims in the southwestern segment of the basin. The areas covered Total 231,347 12.6 by individual occurrences of the dark plains vary significantly from ~10 Plains UIdp, Late Imbrian dark plains 29,738 1.6 up to 7,000 km2 (center of the Apollo basin). UIlp, Late Imbrian light plains 54,375 3 LIlr, Early Imbrian low-relief 60,086 3.3 Light-toned plains (Figures 7b and 7c): These plain-forming units have the rugged plains same apparent albedo as adjacent crater-related landforms. The light pNlrr_APL, pre-Nectarian 51,682 2.8 plains are scattered throughout the map area and form two types of occur- low-relief rugged, Apollo rences. The most abundant are light plains that fill the floors of small cra- Total 195,881 10.7 ters (Figure 7b). In most cases, they occur within the SPA rim topographic SPA rim domain fi Nonplains pNm_SPAr, pre-Nectarian 208,356 11.4 domain and form equidimensional elds a few tens of kilometers across 2 SPA rim materials and smaller. The typical area of their occurrences is ~100–200 km . Less pNrm_SPA, pre-Nectarian 72,142 4 frequent but typically larger (700–900 km2) patches of the light plains SPA rim massifs (Figure 7c) are concentrated within the SPA floor topographic domain. In Total 280,498 15.4 this region, the light plains are associated with the Apollo basin, occurring Note. SPA = South Pole-Aitken. in the zone between the inner and outer rims and near the southwestern corner of the map area. Low-relief rugged terrain (Figure 7d): The flattened surface of this morphological unit has a rugged morphol- ogy due to numerous low (a few hundred meters) and curvilinear ridges. Some of the ridges are portions of impact crater rims. The low-relief rugged terrain occurs in two localities within the map area: (1) inside the Apollo basin between its inner and outer rims and (2) in the southwestern corner of the map area where this unit shows no associations with specific impact structures.

5. Relative and Absolute Model Ages of the Map Units 5.1. Stratigraphic Relationships The relationships of superposition, embayment, as well as the degradation state of landforms establish con- sistent relationships between the relative ages of our defined morphological units and allow reconstruction of the general sequence of events in the study region. In the next section of the paper, we refine and calibrate the stratigraphic scheme with the estimates of the absolute model ages via CSFD measurements. The crisp morphology and prominent systems of rays and secondary craters indicate that craters possessing these features are the youngest in the study region. Ejecta and secondary craters of these primary craters overlay or cut across all other adjacent landforms (Figure 8a). Traditionally, these craters are defined as the Copernican primary craters (Wilhelms & McCauley, 1971Wilhelms & El-Baz, 1977; Lucchitta, 1978; Wilhelms et al., 1979), unit Cc (Figure 9a). The absence of rays and a higher degree of degradation indicate that the sharp-crested craters (Figure 5c) and craters with subdued ejecta (Figure 5b) are older than the craters with rays. The typically small (10- to 20-km diameter) sharp-crested craters are numerous in the map area, and these craters often superpose the craters with subdued ejecta (Figure 8a). The small sharp-crested craters do not have a flat floor that is

IVANOV ET AL. 2594 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 7. Examples of plains and plain-forming units in the map area. Specific morphological units are outlined by dotted lines and indicated by arrows. (a) Low- albedo plains (center of the image is at 36.2°S, 153.0°W). (b) Light-albedo plains that partially fill small impact craters, preferentially on the South Pole-Aitken rim (center of the image is at 24.4°S, 151.5°W). (c) Light-albedo plains that form extensive occurrences on the SPA floor (center of the image is at 44.7°S, 155.9°W). (d) Low- relief rugged terrain (center of the image is at 40.7°S, 154.0°W). All images are portions of the wide-angle camera mosaic with resolution 100 m/pixel.

large enough for reliable CSFD measurements, and the walls of such craters appear to be steep, which causes fast removal of the younger craters. Thus, the CSFD measurements on these craters cannot be used to define reliable AMAs. They certainly predate the Copernican craters and postdate the craters with subdued ejecta (Figure 8a). In our study, we interpret the sharp-crested craters as structures of Eratosthenian age (unit Ec, Figure 9a) and craters with subdued ejecta as features of Imbrian age. The Imbrian craters can be further divided into two groups that are either younger or older than the strati- graphic marker provided by the ejecta from the Orientale basin, whose ejecta and chains of secondary craters overlay the eastern edge of the study region (Unit LIo, Figure 9b). For example, ejecta from crater Ioffe over- lays the sculptured terrain formed by the ejecta of the Orientale impact (Figure 8b). Chains of Orientale sec- ondary craters are seen on the floor of crater Fizeau (Figure 8c). Both craters belong to the morphological class of the craters with subdued ejecta (Figure 5b) but formed before (e.g., Fizeau) and after (e.g., Ioffe) the Orientale basin. The absolute model age of the Orientale event corresponds to the transition from the Lower to the Upper Imbrian system (e.g., Stöffler et al., 2006, and references therein). The stratigraphic rela- tionships with the Orientale-related features/materials and the CSFD measurements for craters that are

IVANOV ET AL. 2595 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 8. Examples of stratigraphic relationships among key units in the study area. (a) Chains of secondary craters (dashed lines, arrows) from the Copernican crater Das (unit Cc) cut the floor of the Late Imbrian crater Mariotte (unit UIc) and small, sharp-crested Eratosthenian craters (unit Ec). Center of the image is at 27.5°S, 137.6°W. (b) Ejecta from the Late Imbrian crater Ioffe (unit UIc) are localized within elongated lows of the Orientale sculptured terrain (unit LIo). Center of the image is at 16.3°S, 127.3°W. (c) Chains of secondary craters (dashed lines, arrows) from the Orientale basin cut the floor of the Early Imbrian crater Fizeau (unit LIc). Center of the image is at 57.3°S, 134.2°W. (d) Ejecta from the Early Imbrian crater Fizeau (unit LIc) overlay the eastern portion of the older crater Minkowski (unit NpNc). A field of the Late Imbrian dark plains (unit UIdp) covers the central portion of the Minkowski floor. Center of the image is at 56.1°S, 142.6°W. All images are portions of the wide-angle camera mosaic with resolution 100 m/px. Examples of stratigraphic relationships among key units in the study area. (e) Materials of the Late Imbrian light plains (unit UIlp) embay an occurrence of the Early Imbrian rugged plains (unit LIrp). Both types of plains embay rims of the Early Imbrian craters Bose (unit LIc, right) and the Nectarian/pre-Nectarian crater Bose-U (unit NpNc, left). Center of the image is at 53.5°S, 173.6°W. (f) Materials of the Late Imbrian light plains (unit UIlp) are localized within local lows on the surface of the pre-Nectarian rugged terrain (unit pNlrAPL) in the NE sector of the Apollo basin. Materials of the unit pNlrAPL overlay a portion of the Apollo rim massifs (unit pNrmAPL). Center of the image is at 33.4°S, 145.3°W. (g) Materials of the rim of crater Walker (upper portion of image, unit NpNc) overlays the unit of degraded crater materials on the floor of the South Pole-Aitken (SPA) basin (unit pNmSPAf). Materials of the Late Imbrian light plains (unit UIlp) fill local depressions on the surface of the unit pNmSPAf. Center of the image is at 26.6°S, 162.5°W. (g) Materials of the unit of degraded craters on the rim of the SPA basin (unit pNmSPAr) surround an outcrop of the SPA rim massifs (unit pNrmSPA). Center of the image is at 21.6°S, 164.8°W. All images are portions of the wide-angle camera mosaic with resolution 100 m/pixel.

IVANOV ET AL. 2596 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 9. Spatial distribution of distinct stratigraphic units within the map area. (a) Copernican and Eratosthenian crater materials. (b) Imbrian crater materials. (c) Imbrian light plains that partly fill small craters. (d) Imbrian plain-forming units. Spatial distribution of distinct stratigraphic units within the map area. (e) Nectarian and pre-Nectarian crater materials. (f) Units that form the Apollo basin. (g) Pre-Nectarian materials of degraded craters that make up the floor and rim of the South Pole-Aitken basin.

IVANOV ET AL. 2597 Journal of Geophysical Research: Planets 10.1029/2018JE005590

beyond the Orientale ejecta zone allow classification of the craters with subdued ejecta as having the Late and Early Imbrian ages (units UIc and LIc, Figure 9b). Varieties of light plains that preferentially occur within smaller craters on the SPA rim (Figure 7b) do not fill the Eratosthenian sharp-crested craters. Thus, in our stratigraphic scheme, we interpret occurrences of the crater- filling plains as having Imbrian ages (unit Ilp, Figure 9c). Materials of the dark and light plains in the study area embay the surrounding terrains (Figures 7a and 7c) and therefore younger. The AMAs of these plains were determined to correspond to the Upper Imbrian system by (Pasckert et al. (2018; units UIdp and UIlp, Figure 9d). Ejecta from the Early Imbrian crater Fizeau overlays the rim and floor of the crater Minkowski (Figure 8d); thus, Minkowski is older than Fizeau. The morphology of Minkowski crater is typical for the majority of impact structures in the study region. Although these craters usually have a complete (or nearly complete) rim, their ejecta deposits have mostly vanished and only a proximal portion of continuous ejecta is still visible near the crater rims (Figure 6a). The relationships exhibited by crater Fizeau suggest that craters possessing prominent rims, but lacking most of their ejecta, are older than the Early Imbrian craters, so we interpret them as belong- ing to the Nectarian and pre-Nectarian systems (Figure 9e). The plain-like terrains with the rugged surface texture (Figure 7d) occur in the southwestern corner of the map area in association with the extensive fields of light and dark plains (units UIlp and UIdp, Figure 9d), as well as within the Apollo basin. In the first location, the light plains embay the rugged terrain, which, in turn, embays the ejecta of the Early Imbrian crater Bose (Figure 8d). These relationships indicate that the rugged terrain is older than the Late Imbrian light plains and younger than the Early Imbrian Bose ejecta. CSFD measurements on the surface of the plain-forming rugged terrain near craters Bose and Bhabha (see next section) indicate an Early Imbrian age for the rugged terrain (unit LIrp, Figure 9d). Inside the Apollo basin, the rugged terrain is embayed by Late Imbrian light plains and embays sections of the massifs of the Apollo inner rim (Figure 8f). CSFD measurements on the surface of the plain-forming rugged unit in the Apollo basin (see next section) indicate an old, pre-Nectarian age (unit pNlrAPL, Figure 9f). The oldest, certainly pre-Nectarian, terrains in the study area are associated with the primary features of the Apollo and SPA basins. Within the Apollo basin, the oldest materials form the basin rim massifs (unit pNrmAPL, Figure 9f). Within the SPA basin, the oldest materials occur within both topographic domains in the form of a morphological unit consisting of strongly degraded craters (Figure 6d). Within the map area, the Nectarian/pre-Nectarian crater materials (unit NpNc, Figure 8f) overlay this morphological unit, which we have split into two units according to their locations, pNmSPAf (the SPA floor domain, Figure 9g) and pNmSPAr (the SPA rim domain, Figure 9g). Within the SPA rim, materials of the unit pNmSPAr surround the high-standing massifs of the basin rim (unit pNrmSPA, Figures 8h,9g), but the relative ages of these units cannot be established confidently and CSFD measurements on their surfaces do not provide reliable results. The assessment of the relative age relationships among the morphological units allowed the transformation of the morphological map into the geological map (Figure 10a). The map, along with the correlation chart of the stratigraphic units (Figure 10b), shows the distribution of processes that acted in the northern portion of the SPA basin in space and time.

5.2. Absolute Model Ages The correlation chart (Figure 10b) represents a qualitative assessment of the sequence of events that shaped the surface in the study region. In order to calibrate this stratigraphy with absolute model ages, we performed CSFD measurements for key units. We used the lunar cratering chronology developed by Neukum et al. (2001) and Stöffler et al. (2006). It is important to note that the lunar chronology for the pre-Nectarian terrains has major uncertainties (summarized in Stöffler et al., 2006), mostly due to the absence of a robust correlation between the radiometric ages and specific pre-Nectarian landforms. Thus, our AMA determinations of the oldest units within the northwestern portion of the SPA basin represent the model-dependent estimates. The AMA of the SPA basin was estimated to be ~4.2–4.3 Ga (Hiesinger et al., 2012), and the SPA impact was the first major recognizable event in the area of our study. Previous determinations of the AMA of Apollo basin range from 4.14 + 0.02/À0.03 (Orgel et al., 2018), an age determined using the Fassett et al. (2012) data set with a buffered nonsparseness correction, to 3.91 + 0.04/À0.06 (Hiesinger et al., 2012). In our study, we

IVANOV ET AL. 2598 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 10. (a) Geological map of the northern portion of the South Pole-Aitken basin. (b) Correlation chart of the stratigraphic units mapped in the northern portion of the South Pole-Aitken basin.

estimated the AMA of the Apollo basin to be 3.98 + 0.04/À0.06 (Figure 11a). This AMA value is based on our mapping results that allow better determination of the basin boundaries. We also conducted CSFD measurements for a number of larger craters (>1 km) that occur at different states of preservation to establish age markers and, thus, characterize the ages of the stratigraphic units repre- sented by these craters. We typically counted superposing craters on the crater floors within the impact melt deposits because this method would tend to provide the lower age limit of the crater formation age. The only

IVANOV ET AL. 2599 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 11. Size-frequency distribution of superposed primary craters: (a) within the outer rim of the Apollo basin and (b–d) on the floor of the larger craters that predate the Imbrian craters and compose the unit NpNc.

exclusion was the crater Maksutov the floor of which is covered by dark plains. For this impact structure, we did the CSFD measurements on its ejecta blanket. Three craters, Langmuir (35.8°S, 128.7°W, 88 km), Gerasimovich-R (25.3°S, 126.3°W, 48 km), and Stoney (55.7°S, 156.5°W, 45 km) have discernible impact melt deposits on their floors and represent the morphological unit of the degraded craters with complete rims, that is, unit NpNc. The AMAs for these craters vary from 3.91 + 0.05/À0.07 (Langmuir, Figure 11b) to 3.86 + 0.04/À0.06 (Stoney, Figure 11d). The AMA of the crater Gerasimovich-R is estimated to be 3.88 + 0.04/À0.05 (Figure 11c).

IVANOV ET AL. 2600 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Three craters, Fizeau (58.1°S, 133.8°W, 106 km), Bose (54.0°S, 169.3°W, 85 km), and Bhabha (55.4°S, 165.3°W, 68 km), represent the morphological unit of craters with discernible but degraded ejecta that predate the ejecta from the Orientale basin and, thus, are of Early Imbrian age. The AMAs (Figures 12a–12c) of these cra- ters range from 3.85 + 0.02/À0.03 (Fizeau, Early Imbrian-Nectarian transition) to 3.81 + 0.02/À0.03 (Bhabha). Six craters, Ioffe (14.5°S, 129.1°W, 74 km), Maksutov (40.8°S, 168.5°W, 82 km), Mariotte (28.4°S, 139.1°W, 70 × 48 km), Dryden (33.1°S, 156.1°W, 50 km), Finsen (42.3°S, 177.7°W, 66 km), and Paschen-S (14.6°S, 142.6°W, 40 km), represent the morphological unit with craters having visible ejecta that postdate the Orientale basin. The AMAs of these craters (Figures 13a–13f) vary from 3.72 + 0.03/À0.04 (Ioffe) to 3.40 + 0.08/À0.16 Ga (Paschen-S). Occurrences of light plains that partly fill impact craters on the SPA rim (unit Ip) are typically too small for reli- able AMA estimates at the resolution of the WAC mosaic. We were able to perform CSFD measurements only for three large fields of the in-crater light plains (area 1: 19.8°S, 142.9°W, area 2: 23.6°S, 147.0°W, and area 3: 20.3°S, 137.7°W), which give AMAs for that range from ~3.80 to ~3.72 Ga (Figures 14a–14c), which correspond to Early and Late Imbrian ages. The light plains that are localized inside craters are likely the deposits of distal ejecta from remote craters and basins (Muehlberger et al., 1972; Eggleton & Schaber, 1972; Head, 1974; Meyer et al., 2016). The large variations in the AMAs for these plains (Figures 14a–14c) suggest different sources. The low-relief rugged terrain areas have distinctly different AMAs in different locations. In the southwestern corner of the map area (Figure 9d), the AMA is about 3.80 + 0.02/À0.02 Ga (Early Imbrian age, unit LIrp, Figure 14d), whereas the rugged terrain within the Apollo basin (Figure 9f) is significantly older: its AMA is 3.98 + 0.04/À0.06 Ga (pre-Nectarian age, unit pNlrAPL, Figure 14e), which coincides exactly with the AMA for the Apollo basin (Figure 11a). The practically identical, tight spatial association with the Apollo basin, and small topographic variations of the pre-Nectarian rugged terrain strongly suggest that this unit repre- sents Apollo impact melt. The AMAs for the dark plains and many patches of the light plains on the SPA floor, clustered near the ages of ~3.5–3.6 Ga (Figure 15), have been estimated by Pasckert et al. (2018), and in our work we adopted these ages. Figure 15 summarizes our AMAs for key stratigraphic markers such as the Apollo basin and large impact cra- ters in the study region. Practically all occurrences of the plain-forming units are of Imbrian age, and the majority of the dark plains, which are interpreted as volcanic materials (Haruyama et al., 2009; Head, 1976; Stuart-Alexander, 1978; Wilhelms et al., 1979), are of Late Imbrian age.

6. Discussion The morphologic characteristics of the units defined in our map area to a large degree constrain their possi- ble origin. For example, the dark and morphologically smooth surface of the dark plains strongly supports a volcanic origin (e.g., Wilhelms et al., 1979). In contrast, the light plains can be related to both volcanic (Schultz & Spudis, 1979) and impact (Muehlberger et al., 1972) processes, and their morphology is not enough to con- fidently assess their mode of origin.

The compositional maps of FeO and TiO2 concentrations derived from the Clementine mission spectral data (Lucey et al., 1998, 2000) have resolutions comparable to the WAC mosaics and, in contrast to Kaguya com- positional maps (Lemelin et al., 2016; Ohtake et al., 2014; Uemoto et al., 2017), cover the entire area of our study. The Clementine compositional maps provide additional and independent information, which in com- bination with the photogeological analysis, helps to interpret the nature of the mapped units and address several important problems of the regional geology of the SPA basin.

6.1. Iron and Titanium Concentrations of the Stratigraphic Units One of the major issues regarding the history and evolution of the SPA region is the age and extent of vol- canic activity within the basin (Haruyama et al., 2009; Head, 1976; Pasckert et al., 2018; Stuart-Alexander, 1978; Wilhelms et al., 1979). Obvious volcanic plains (dark plains, unit UIdp) make up a small fraction of the map area (Table 1), occur in isolated local lows, and cover relatively small areas (Figure 9d). The other types of the plain-forming units in our study area (i.e., the Imbrian light and rugged plains, units UIlp and LIrp) are more abundant (Table 1 and Figure 9d) but could represent plains of either volcanic (cryptomaria) or impact

IVANOV ET AL. 2601 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 12. Size-frequency distribution of superposed primary craters on the floor of the larger craters that predate the Orientale deposits and compose the unit LIc.

IVANOV ET AL. 2602 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 13. Size-frequency distribution of superposed primary craters on the floor of the larger craters that postdate the Orientale deposits and compose the unit UIc (craters Ioffe, Maksutov, Mariotte, Dryden, Finsen, and Paschen-S).

origin (morphologically smooth ejecta). The problem of the extent of the mare and cryptomare deposits on the SPA floor (Gibson & Jolliff, 2011; Jolliff et al., 2011) is important because they presumably mark the beginning of volcanism in the SPA region, illustrate migration of volcanic activity through time, and, potentially, could cause the iron signature within the SPA region. Thus, we investigated and compared the

FeO and TiO2 concentrations of these units to aid the interpretation of their origin(s).

Both iron and titanium are major components of lunar basalts, in which concentrations of FeO and TiO2 are typically higher than ~17 wt% and ~3 wt%, respectively (e.g., Taylor et al., 1991). Although very low-Ti lunar

basalts (<~1 wt% of TiO2) have been identified among the lunar samples, they are rare (Giguere et al., 2000) and rich in iron, ~20 wt% of FeO (Taylor et al., 1991). In contrast, the lunar highland rocks are generally much

poorer in both iron and titanium, <6 wt% of FeO and <1 wt% of TiO2 (Dymek et al., 1975; Taylor et al., 1991; Warren et al., 1987), although there is a known mafic impact-melt component in the lunar highlands (e.g.,

IVANOV ET AL. 2603 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 14. Size-frequency distribution of superposed primary craters on the light plain-forming units that fill smaller craters in the northern portion of the South Pole-Aitken basin and on the low-relief, rugged plain-forming units.

Ryder & Wood, 1977). It is possible that the nature of the formation of the SPA basin allowed the formation of impact melt deposits with more iron- and titanium-rich compositions (e.g., Moriarty & Pieters, 2018) than that sampled by the Apollo and Luna missions. Thus, our analysis is limited to the comparison between known sample compositions and the global remote sensing data set.

Here we compare the Clementine model contents of FeO and TiO2 (Lucey et al., 1998, 2000) for the map units in our study with those typical of the Serenitatis and Humorum maria. We selected these maria as references

for two reasons. (1) The model FeO and TiO2 concentrations in both regions are similar to the iron and tita- nium distribution in other lunar maria (Figure 16) and (2) the Clementine data for these regions are gapless and show no artifacts related to different orbits. Thus, random points from the Serenitatis and Humorum

maria provide a representative and unbiased sample of the FeO and TiO2 concentrations in the lunar mare basalts.

IVANOV ET AL. 2604 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 15. Summary of the absolute model age estimates within the northern portion of the South Pole-Aitken basin. Ages of the dark- and light-albedo plains are from Pasckert et al. (2018).

Points that show the model concentrations of FeO and TiO2 in the Serenitatis and Humorum volcanic materi- als are completely intermixed in the TiO2-FeO plot (Figure 17a), indicating the absence of systematic differ- ences in the FeO and TiO2 abundance within both mare regions. Another important characteristic of the Serenitatis and Humorum mare basalts is a pronounced positive correlation between the titanium and iron concentrations, with a correlation coefficient of ~0.73, which is significant at any reasonable confidence level for our mare sample that includes ~60,000 data points (Figure 17a). 6.1.1. Dark Plains (UIdp)

The early Imbrium dark plains unit (UIdp; green dots in Figure 17a) exhibits mean values of FeO and TiO2 con- tent that lie inside the cloud of points representing the Serenitatis and Humorum maria and show a correla-

tion between the TiO2 and FeO concentrations that is characteristic of mare basalts. The model concentrations of FeO and TiO2 of the dark plains in the northern portion of the SPA basin are consistent with those of low-titanium type of lunar basalts (Taylor et al., 1991). 6.1.2. Pre-Nectarian Rim Materials (pNrmSPA) In contrast, the pre-Nectarian unit that composes the majority of the SPA rim domain (pNmSPA, yellow dots

in Figure 17a) is characterized by lower concentrations of both FeO (<~8 wt%) and TiO2 (<~1 wt%), com- pared with both the Serenitatis/Humorum maria and the dark plains on the SPA floor. The other characteristic

feature of the SPA rim unit is the absence of a steep, mare-like trend of mutually increasing TiO22 and FeO concentrations. As the model FeO concentrations in the rim units increase from ~4 to ~7 wt%, the TiO2 con- centrations show little variations and remain low, at less than ~1 wt%. As a result, the points of the SPA rim

unit form a flat trend in the FeO/TiO2 diagram (Figure 17a). The contrasting behavior of iron and titanium in the nonmare (rim materials) and mare (dark plains) units suggests that the FeO and TiO2 concentrations can serve as discriminators potentially allowing separation of the volcanic and nonvolcanic materials. 6.1.3. Light Plains (UIlp)

The FeO and TiO2 mean concentrations of the light plains (unit UIlp) form two groups of points (Figure 17b). The first group (yellow dots in Figure 17b) follows the flat TiO2-FeO trend, which characterizes the SPA rim units. The lower concentrations of iron and titanium and the flat trend for this group suggest that they likely represent morphologically smooth crater ejecta from highland materials that are enriched in iron relative to the SPA rim units. The second group (green dots in Figure 17b) has higher iron and titanium concentrations

and appears to follow the steep TiO2-FeO trend typical of the volcanic materials. These light plains probably represent cryptomaria (e.g., Schultz & Spudis, 1979; Whitten & Head, 2015) and are concentrated mostly within the southwestern portion of the map area (light plains marked by arrows in Figure 9d).

IVANOV ET AL. 2605 Journal of Geophysical Research: Planets 10.1029/2018JE005590

6.1.4. Rugged Plains (LIrp) The Early Imbrian rugged plains (unit LIrp), as well as the light plains (unit UIlp), have morphologically smooth surfaces with a higher albedo than the

dark plains. In the FeO-TiO2 diagram, rugged plains points (green dots in Figure 17c) are tightly clustered at the lower end of the volcanic

FeO/TiO2 trend similar to the light plains of the second group (crypto- maria). We interpret both the higher concentrations of iron and titanium and the noticeable correlation of these components as indicative of a vol- canic origin for the rugged plains, and thus, the plains likely represent cryptomaria on the SPA floor. The rugged plains exhibit an Early Imbrian AMA (Figure 15) representing the oldest recognizable plain-forming unit on the SPA floor in our study area. Occurrences of the rugged plains are concentrated in the southwes- tern corner of the map area (Figures 9d and 10a), that is, in the topographi- cally lower central region of the SPA floor (Figure 3c). The younger, Late Imbrian volcanism on the SPA floor present as both dark and light plains, continued to be active near the basin center and in the northern segment of the floor in and around the Apollo basin (Figure 9d). Effusive volcanism dominated in the northern portion of the SPA basin, but in the large floor- fractured crater Oppenheimer to the west of the Apollo basin, pyroclastic activity produced several dark-mantling deposits (Gaddis et al., 2003, 2013) at apparently the same time as the majority of the effusive volcan- ism (Ivanov et al., 2016).

6.2. Origin of the SPA Iron Signature The other major question regarding the SPA basin is the origin of the extensive iron and thorium signature in this region (Lawrence et al., 2000, 2002; Lucey et al., 1998, 2000). In our study, we address only the pos- sible origin of the iron signature, but the close spatial association of both components (Garrick-Bethell & Zuber, 2005, 2009; Jolliff et al., 2000) sug- gests that their enhanced concentrations on the SPA floor may have the same origin. We consider three different possible origins for the signature in the context of our study: (1) volcanic deposits, (2) impact melt deposits, and (3) vertical compositional stratification of the crust/mantle. 6.2.1. Volcanic Origin The existence of the SPA iron signature might be consistent with the pre- sence of volcanic materials on the basin floor (Figure 9d). However, although the volcanic (or presumably volcanic) plains are rich in iron Figure 16. The frequency distribution of FeO and TiO2 concentrations in the (Figure 17), they cannot account for the entire SPA iron signature for two Serenitatis and Humorum mare materials mimic the distribution of these reasons. (1) The model concentrations of FeO in the dark volcanic plains, components in the other lunar maria (excluding Serenitatis and Humorum). possible cryptomaria, and rugged plains are higher (>~12 wt% of FeO) than the typical FeO concentrations of the SPA iron signature (~10–11 wt%). (2) The plain-forming units on the basin floor (even if all occurrences of possible impact- related light plains are considered to be volcanic materials) comprise only ~8%–10% of the map area (Table 1); the iron signature, however, is much more extensive and occupies over ~60% of the study region. 6.2.2. Impact Melt Deposit The fractional differentiation of a large impact melt body formed by the SPA impact provides another possi- ble explanation for the iron signature (Vaughan & Head, 2014). In this model, a very thick and voluminous, (~50 km, ~108 km3) body of impact melt differentiated, producing an uppermost norite-like layer ~12.5- km thick overlying a layer of pyroxenite (~30-km thick) with a volume of dunite cumulates at the bottom (Vaughan & Head, 2014). The fractional differentiation model explains the virtual absence of olivine expo- sures on the SPA floor, the floor iron signature, and the slightly elevated concentrations of thorium in this region by preferential concentration of incompatible elements in the late derivates.

IVANOV ET AL. 2606 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Figure 17. The TiO2 versus FeO plot for the Serenitatis and Humorum mare materials and several key units in the northern portion of the South Pole-Aitken basin. In order to highlight the major pattern of the distribution of FeO and TiO2, the diagrams show the average values for each individual occurrences of a specific unit in the study area (yellow and green dots) with 1-sigma error bars.

A major challenge of the melt-pool differentiation model is the lateral extent of the norite-like lithology that corresponds to the Mg-pyroxene annulus that overlies almost the entire SPA floor (Moriarty & Pieters, 2018). In the framework of the differentiation model, the annulus may represent the roof of the body of differentiat- ing impact melt, the thickest portion of which is confined within the transient crater (Vaughan et al., 2013). The annulus, thus, marks the edges of the transient cavity. The diameter of the annulus is ~1,000– 1,100 km (Moriarty & Pieters, 2018), however, a transient cavity of this size as calculated by Melosh et al. (2017) would cause the emplacement of large quantities of mantle material on the SPA rim. To restrict the emplacement of mantle materials within the SPA floor may require a much smaller transient crater, which would allow formation of a melt pool within a roughly circular (~400- to 500-km diameter) area near the apparent center of the SPA basin (Ohtake et al., 2014; Uemoto et al., 2017). This area approximately

IVANOV ET AL. 2607 Journal of Geophysical Research: Planets 10.1029/2018JE005590

corresponds to the SPA compositional anomaly (Moriarty & Pieters, 2018) and is characterized by abundant Ca, Fe-rich pyroxenes (Moriarty & Pieters, 2015, 2018; Ohtake et al., 2014). Results of our study indicate that the majority of this area is covered by low-Imbrian rugged plains (unit LIrp, Figures 9 and 10) that likely repre- sent cryptomare deposits. The relatively young AMA (3.80 ± 0.02 Ga, Figure 14d) of the rugged plains is inconsistent with their interpretation as the SPA impact melt. Moreover, the dimensions of the area occupied by unit LIrp (400–500 km), which is comparable with the diameter of the Apollo basin (400–450 km), cannot explain both the overall dimen- sions and topographic configuration of the iron signature of the SPA basin. 6.2.3. Compositional Stratification of the Crust/Mantle A key characteristic of the SPA iron signature is its strong correlation with the regional topographic configuration of the basin. Regions that lie below the zero contour line (i.e., the SPA floor domain, Figure 3b) are significantly richer in iron (>~7.5 wt% of model FeO), whereas the area above the zero contour line typically has less than ~7.5 wt% of FeO. Thus, the sources of the signature must be units that occur within the SPA floor topographic domain. The horizontal dimensions of the signature and its intensity require that the source units should make up a large fraction of the basin floor and have an appropriate FeO content (~10–15 wt% FeO).

In our study region, only two units fit these requirements: (1) Nectarian and pre-Nectarian crater materials (NpNc) and (2) pre-Nectarian SPA floor materials (pNmSPAf). Exposures of these units make up ~50% of the SPA floor in the study region, and their typical iron concentrations are ~11– 15 wt% of the model FeO, similar to the iron concentrations in the majority of the SPA signature. Unit NpNc is the result of impact reworking of preexisting materials such that its compositional characteristics were inherited from these older units. For example, the principal pattern of the iron concentrations of unit NpNc coincides with the distribution of the iron content in the underlying mate- rials and repeats the strongly bimodal distribution of iron between the SPA floor and rim domains (Figure 18). Therefore, unit NpNc cannot be a pri- mary source of the SPA iron signature. The only visible unit responsible for the signature thus must be the oldest materials on the SPA floor: unit pNmSPAf (Figure 9g), in which typical concentrations of FeO vary from Figure 18. The frequency distribution of FeO concentration in (a) the pre- ~11 to ~14.5 wt% (Figure 18a). Nectarian units of the South Pole-Aitken floor and rim topographic domains, and (b) unit NpNc, which is the result of impact gardening of pre- The units that form the SPA rim domain in our map area (pNmSPAr existing units. and pNrmSPA) are stratigraphically equivalent to the pre-Nectarian floor materials (Figure 10b) but show significantly lower concentrations of FeO (~3–7 wt%) (Figure 18b). This indicates that the distribution of iron concentrations in the oldest units of the study region is strongly bimodal and coincides with the bimodal distribution of the major topographic features in the northern portion of the SPA basin (Figure 3b). The topographic bimodality of the study area is a characteristic generated by the SPA impact event. It follows that the relationship between the topography and FeO concentrations may also be a result of the impact. In this case, the SPA iron signature would represent an indigenous feature of the basin, as proposed for the SPA thorium signature (Garrick-Bethell & Zuber, 2009; Jolliff et al., 2000) and cannot be explained by the deposition of ejecta from the younger basins such as Imbrium (Haskin et al., 1998) or Serenitatis (Wieczorek & Zuber, 2001). Thus, we consider whether a compositional stratification of the crust/mantle could explain the signature.

IVANOV ET AL. 2608 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Owing to our poor knowledge of the lunar interior, model compositions of the lunar mantle vary from predominantly olivine-rich (see review in Wieczorek et al., 2006) to predominantly low-calcium pyroxene (Kuskov, 1997; Melosh et al., 2017). In any case, the hypothetical exposed mantle material would be strongly different in composition from the lunar anorthositic crust, which is exposed on the SPA rim. Could this dis- agreement be related to the nature of the SPA impact? Numerical simulations of SPA-forming mega impacts (30°–60°) predict large excavation depths and the emplacement of a kilometer-thick layer of mantle materials outside the transient crater (Melosh et al., 2017; Potter et al., 2012), the size of which is an important quantity governing the spatial distribution of the ejecta. However, in scenarios for the SPA impact that give the smallest transient crater (850- to 1,000- km diameter), the model predicts emplacement of a layer (at least a few kilometers thick) of mantle materials well beyond the outer best fit topographic ellipse defined by Garrick-Bethell and Zuber (2009). These mantle materials, thus, should dominate the SPA rim topographic domain that consists of the units pNmSPAr and pNrmSPA (Figures 9 and 10). However, in contrast to the model predictions, these SPA rim materials are char- acterized by lower FeO concentrations (Figure 17) and anorthositic lithology (Moriarty & Pieters, 2018). This result led Melosh et al. (2017) to propose that the lunar upper mantle is composed of low calcium pyroxene rather than olivine. These model results are also consistent with our interpretation that the SPA iron signature is the result of a compositionally stratified crust/mantle, where the iron signature corresponds to exposed upper mantle materials with higher Fe contents (unit pNmSPAf, Figures 9 and 10) than the highlands materi- als deposited on the SPA rim as ejecta (units pNmSPAr and pNrmSPA, Figures 9 and 10).

7. Conclusions Detailed photogeological analyses of the northern portion of the SPA basin (10°–60°S, 125°–175°W) allowed us to map a set of morphologically uniform and distinct material units. Determination of their relative and absolute model ages resulted in the compilation of a geological map (1:500,000 scale) and a correlation chart of the units that illustrate the sequence of the major events in the study region. The first recognizable event was the SPA-forming impact that created the primary topographic configuration of the basin, consisting of SPA floor and rim topographic domains. The characteristic topographic difference between the domains is ~8.5 km. The largest impact structure in the map area that clearly postdates the SPA event is the Apollo basin (~3.98 + 0.04/À0.06 Ga).

Analyses of the morphology and modeled concentrations of FeO and TiO2 in key stratigraphic units helps assess the origins of the map units. In the study area, there are two principal morphological classes of units: the cratered terrains and plain-forming materials. The plains are the result of deposition of both volcanic flows and impact melt breccias. The volcanic plains are characterized by a prominent correlation between their iron and titanium concentrations. The oldest volcanic materials (dark, light, and rugged plains) in the study area are concentrated in its south- western corner near the center of the SPA basin and have Early Imbrian ages of ~3.80 + 0.02/À0.02 Ga. The younger, Late Imbrian volcanic activity in the study area migrated to the northern portions of the SPA floor and occurred in and around the Apollo basin. The volcanic plains cover ~8% of the map area and occur as relatively small patches that fill local topographic depressions of impact craters. The limited extent of the vol- canic plains and their localization within isolated lows indicate that volcanic activity on the SPA floor could not have changed the original depth of the basin significantly or be the principal source of the extensive SPA iron anomaly. Analyses of the variations of the FeO concentrations in different units show that the oldest, pre-Nectarian units in the study region are responsible for the formation of the SPA iron anomaly. The other oldest unit in the study region (pNmSPAr) forms the rim of the SPA basin and is characterized by low concentrations of iron (<~7.5 wt% of FeO). In contrast, the oldest unit that makes up the SPA floor domain (pre-Nectarian crater materials, pNmSPAf) occupies a significant portion of the floor and the most typical FeO concentrations in its materials vary from ~11 to ~14.5 wt%, corresponding to the typical iron concentrations of the SPA anomaly. Unit pNmSPAf likely does not represent the result of the impact reworking of the preexisting vol- canic plains unless these hypothetical volcanic materials are poor in iron. The rugged plains unit, although exhibiting a morphology that might be interpreted as an impact melt unit, is much younger than the SPA

IVANOV ET AL. 2609 Journal of Geophysical Research: Planets 10.1029/2018JE005590

basin and is thus unlikely to represent a differentiated impact melt deposit. Postulated impact melt bodies within SPA are also not areally extensive enough to explain the iron signature. We conclude that the iron signature is consistent with a preexisting crust/mantle compositional contrast that was unearthed by the SPA impact.

Acknowledgments References We gratefully acknowledge very constructive and helpful reviews Antonenko, I., & Head, J. W. (1994). Cryptomaria in the Schiller-Schickard, mare Humorum and western Oceanus Procellarum areas: Studies – provided by Brad Jolliff and an using dark-halo craters. Paper presented at Lunar and Planetary Science Conference 25th, 35 36, Houston, TX. anonymous reviewer that significantly Antonenko, I., Head, J. W., Mustard, J. F., & Hawke, B. R. (1995). Criteria for the detection of lunar cryptomaria. Earth, Moon and Planets, 69(2), – improved the original version of the 141 172. https://doi.org/10.1007/BF00613096 Baker, D. M. H., Head, J. W., Phillips, R. J., Neumann, G. A., Bierson, C. J., Smith, D. E., & Zuber, M. T. (2017). GRAIL gravity observations of the manuscript. M. A. I. was supported by transition from complex crater to peak-ring basin on the Moon: Implications for crustal structure and impact basin formation. Icarus, 292, the German Science Foundation 54–73. https://doi.org/10.1016/j.icarus.2017.03.024 (Deutsche Forschungsgemeinschaft – Belton, M. J. S., Head, J. W., Pieters, C. M., Greeley, R., McEwen, A. S., Neukum, G., et al. (1992). Lunar impact basins and crustal heterogeneity: DFG) grant HI 1410\12-1 and Russian New western limb and far side data from Galileo. Science, 255, 570–576. Science Foundation (grant 17-17- Blewett, D. T., Lucey, P. G., Hawke, B. R., & Jolliff, B. L. (1997). Clementine images of the lunar sample-return stations: Refinement of FeO and 01149). H. H. and C. vd. B. were TiO2 mapping techniques. Journal of Geophysical Research, 102(E7), 16,319–16,325. https://doi.org/10.1029/97JE01505 supported by German Space Agency Dymek, R. F., Albee, A. L., Chodos, A. A. (1975). Comparative petrology of lunar cumulate rocks of possible primary origin: Dunite 72415, (Deutsches Zentrum für Luft- und troctolite 76535, norite 78235, and anorthosite 62237. Proceedings the of Lunar and Planetary Science Conference, 6, 301-341. – Raumfahrt DLR) project 50OW1504 Eggleton, R. E., & Schaber, G. G. (1972). Cayley Formation interpreted as basin ejecta. NASA Apollo 16 Preliminary Scientific Report, 29-7–29-16. and as part of a project that has Fassett, C. I., Head, J. W., Kadish, S. J., Mazarico, E., Neumann, G. A., Smith, D. E., & Zuber, M. T. (2012). Lunar impact basins: Stratigraphy, received funding from the European sequence and ages from superposed impact crater populations measured from Lunar Orbiter Laser Altimeter (LOLA) data. Journal of ’ Union s Horizon 2020 research and Geophysical Research, 117, E00H06. https://doi.org/10.1029/2011JE003951 innovation programme under grant Gaddis, L. R., Staid, M. I., Tyburczy, J. A., Hawke, B. R., & Petro, N. E. (2003). Compositional analyses of lunar pyroclastic deposits. Icarus, 161(2), agreement N°776276 (PLANMAP). C. O. 262–280. https://doi.org/10.1016/S0019-1035(02)00036-2 was funded by DFG SFB-TRR-170 A3. All Gaddis, R. L., Weller, L., Barrett, J., Kirk, R., Milazzo, M., Laura, J., et al. (2013). New volcanic features in lunar, floor-fractured Oppenheimer data relevant to this work are available crater. Paper presented at Lunar and Planetary Science Conference 44th., Abstract 2262, Houston, TX. at: http://www.planetology.ru. The Garrick-Bethell, I., & Zuber, M. T. (2005). An indigenous origin for the South Pole Aitken basin thorium anomaly. Geophysical Research Letters, following link can be used to download 32, L13203. https://doi.org/10.1029/2005GL023142 an archive with all data that are relevant Garrick-Bethell, I., & Zuber, M. T. (2009). Elliptical structure of the lunar South Pole-Aitken basin. Icarus, 204(2), 399–408. https://doi.org/ to our paper: https://brownbox.brown. 10.1016/j.icarus.2009.05.032 edu/download.php?hash=8db5bd78. Gibson, K. E., & Jolliff, B. L. (2011). Correlation of surface units and FeO concentrations in the South Pole - Aitken basin interior. Paper pre- sented at Lunar and Planetary Science Conference 42nd, Abstract 2326, Houston, TX. Giguere, T., Taylor, G. J., Hawke, B. R., & Lucey, P. (2000). The titanium contents of lunar mare basalts. Meteoritics and Planetary Science, 35(1), 193–200. https://doi.org/10.1111/j.1945-5100.2000.tb01985.x Guo, D., Liu, J., Head, J. W., Potter, R. W. (2017). Apollo impact basin: Probing the depth of the lunar farside. Paper presented at Lunar and Planetary Science Conference 48th, Abstract 2833, Houston, TX. Haruyama, J., Ohtake, M., Matsunaga, T., Morota, T., Honda, C., Yokota, Y., et al. (2009). Long-lived volcanism on the lunar farside revealed by SELENE terrain camera. Science, 323(5916), 905–908. https://doi.org/10.1126/science.1163382 Haskin, L. A., Korotev, R. L., Rockow, K. M., & Jolliff, B. L. (1998). The case for an Imbrium origin of the Apollo thorium-rich impact-melt breccias. Meteoritics and Planetary Science, 33(5), 959–975. https://doi.org/10.1111/j.1945-5100.1998.tb01703.x Head, J. W. (1974). Orientale multi-ringed basin interior and implications for the petrogenesis of lunar highland samples. The moon, 11(3-4), 327–356. https://doi.org/10.1007/BF00589168 Head, J. W. (1976). Lunar volcanism in space and time. Reviews of Geophysics, 14(2), 265–300. https://doi.org/10.1029/RG014i002p00265 Head, J. W., & Wilson, L. (1992). Lunar mare volcanism: Stratigraphy, emption conditions, and the evolution of secondary crusts. Geochimica et Cosmochimica Acta, 56(6), 2155–2175. https://doi.org/10.1016/0016-7037(92)90183-J Head, J. W., & Wilson, L. (2017). Generation, ascent and eruption of magma on the Moon: New insights into source depths, magma supply, intrusions and effusive/explosive eruptions (Part 2: Pre dicte d emplacement processes and observations). Icarus, 283, 176–223. https:// doi.org/10.1016/j.icarus.2016.05.031 Hiesinger, H. & Head, J. W. (2004). Lunar South Pole-Aitken impact basin: Topography and mineralogy. Paper presented at Lunar and Planetary Science Conference 34th, Abstract 1164, Houston, TX. Hiesinger, H., van der Bogert, C. H., Pasckert, J.-H., Schmedemann, N., Robinson, M.S., Jolliff, B., et al. (2012). New crater cize-frequency dis- tribution measurements of the South Pole Aitken basin. Paper presented at Lunar and Planetary Science Conference 43rd, Abstract 2863, Houston, TX. Hurwitz, D. M., & Kring, D. A. (2014). Differentiation of the South Pole-Aitken basin impact melt sheet: Implications for lunar exploration. Journal of Geophysical Research: Planets, 119, 1110–1133. https://doi.org/10.1002/2013JE004530 Ivanov, M. A., van der Bogert, C. H., Hiesinger, H., Pasckert, J-H., Bauch, K. (2016). Bracketing the age of lunar pyroclastic deposits in Oppenheimer crater. Paper presented at Lunar and Planetary Science Conference 47th, Abstract 1070, Houston, TX. Jolliff, B. L., Gillis, J. J., Haskin, L. A., Korotev, R. L., & Wieczorek, M. A. (2000). Major lunar crustal terranes: Surface expressions and crust-mantle origins. Journal of Geophysical Research, 105(E2), 4197–4216. https://doi.org/10.1029/1999JE001103 Jolliff, B. L., Wiseman, S. A., Lawrence, S. J., Tran, T. N., Robinson, M. S., Sato, H., & Paige, D. A. (2011). Non-mare silicic volcanism on the lunar farside at Compton-Belkovich. Nature Geoscience, 4(8), 566–571. https://doi.org/10.1038/ngeo1212 Kneissl, T., van Gasselt, S., & Neukum, G. (2011). Map-projection-independent crater size-frequency determination in GIS environments— New software tool for ArcGIS. Planetary and Space Science, 59(11-12), 1243–1254. https://doi.org/10.1016/j.pss.2010.03.015 Kuskov, O. L. (1997). Constitution of the Moon: 4. Composition of the mantle from seismic data. Physics of the Earth and Planetary Interiors, 102(3-4), 239–257. https://doi.org/10.1016/S0031-9201(96)03259-1 Lawrence, D. J., Elphic, R. C., Feldman, W. C., Prettyman, T. H., Gasnault, O., & Maurice, S. (2003). Small-area thorium features on the lunar surface. Journal of Geophysical Research, 108(E9), 5102. https://doi.org/10.1029/2003JE002050 Lawrence, D. J., Feldman, W. C., Barraclough, B. L., Binder, A. B., Elphic, R. C., Maurice, S., et al. (2000). Thorium abundances on the lunar surface. Journal of Geophysical Research, 105(E8), 20,307–20,331. https://doi.org/10.1029/1999JE001177

IVANOV ET AL. 2610 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Lawrence, D. J., Feldman, W. C., Elphic, R. C., Little, R. C., Prettyman, T. H., Maurice, S., et al. (2002). Iron abundances on the lunar surface as measured by the lunar prospector gamma-ray and neutron spectrometers. Journal of Geophysical Research, 107(E12), 5130. https://doi. org/10.1029/2001JE001530 Lawrence, D. J., Puetter, R. C., Elphic, R. C., Feldman, W. C., Hagerty, J. J., Prettyman, T. H., & Spudis, P. D. (2007). Global spatial deconvolution of lunar prospector Th abundances. Geophysical Research Letters, 34, L03201. https://doi.org/10.1029/2006GL028530 Lawrence, S. J., Stopar, J. D., Jolliff, B. L., Robinson, M. S., Sato, H., Hiesinger, H., et al. (2015). Characterizing mare deposits in the Australe region. Paper presented at Lunar and Planetary Science Conference 46th, Abstract 2739, Houston, TX. Lemelin, M., Lucey, P. G., Gaddis, L. R., Hare, T., Ohtake, M. (2016). Global map products from the Kaguya multiband imager at 512 ppd: Minerals, FeO, and OMAT. Paper presented at Lunar and Planetary Science Conference 47th, Abstract 2994, Houston, TX. Lucchitta, B. K. (1978). Geologic map of the north side of the Moon (Map I-1062). USGS. Lucey, P. G. (2004). Mineral maps of the Moon. Geophysical Research Letters, 31, L08701. https://doi.org/10.1029/2003GL019406 Lucey, P. G., Blewett, D. T., & Hawke, B. R. (1998). Mapping the FeO and TiO2 content of the lunar surface with multispectral imagery. Journal of Geophysical Research, 103(E2), 3679–3699. https://doi.org/10.1029/97JE03019 Lucey, P. G., Blewett, D. T., & Jolliff, B. L. (2000). Lunar iron and titanium abundance algorithms based on final processing of Clementine ultraviolet-visible images. Journal of Geophysical Research, 105(E8), 20,297–20,305. https://doi.org/10.1029/1999JE001117 Melosh, H. J., Kendall, J., Horgan, B., Johnson, B. C., Bowling, T., Lucey, P. G., & Taylor, G. J. (2017). South Pole-Aitken basin ejecta reveal the Moon’s upper mantle. Geology, 45(12), 1063–1066. https://doi.org/10.1130/G39375.1 Meyer, H. M., Denevi, B. W., Boyd, A. K., & Robinson, M. S. (2016). The distribution and origin of lunar light plains around Orientale basin. Icarus, 273, 135–145. https://doi.org/10.1016/j.icarus.2016.02.014 Michael, G. G., & Neukum, G. (2010). Planetary surface dating from crater size-frequency distribution measurements: Partial resurfacing events and statistical age uncertainty. Earth and Planetary Science Letters, 294(3-4), 223–229. https://doi.org/10.1016/j.epsl.2009.12.041 Miljkovic, K., Wieczorek, M. A., Collins, G. S., Laneuville, M., Neumann, G. A., Melosh, H. J., et al. (2013). Asymmetric distribution of lunar impact basins caused by variations in target properties. Science, 342(6159), 724–726. https://doi.org/10.1126/science.1243224 Moriarty, D. P., & Pieters, C. M. (2015). The nature and origin of Mafic Mound in the South Pole-Aitken basin. Geophysical Research Letters, 42, 7907–7915. https://doi.org/10.1002/2015GL065718 Moriarty, D. P., & Pieters, C. M. (2018). The character of South Pole-Aitken Basin: Patterns of surface and subsurface composition. Journal of Geophysical Research: Planets, 123, 729–747. https://doi.org/10.1002/2017JE005364 Muehlberger, W. R., Batson, R. M., Boudette, E. L., Duke, C. M., Eggleton, R. E., Elston, D. P., et al. (1972). Preliminary geologic investigation of the Apollo-16 landing site. in: Apollo 16 Preliminary Science Report, p. 6–1. Mustard, J. F., Head, J. W., Murchie, S. M., Pieters, C. M., Belton, M. S., McEwen, A. S. (1992). Schickard cryptomare: Interaction between Orientale ejecta and pre-basin mare from spectral mixture analysis of Galileo SSI data. Paper presented at Lunar and Planetary Science Conference 23rd, 957–958, Houston, TX. Neukum, G., Ivanov, B. A., & Hartmann, W. K. (2001). Cratering records in the inner solar system in relation to the lunar reference system. Space Science Reviews, 96(1/4), 55–86. https://doi.org/10.1023/A:1011989004263 Ohtake, M., Uemoto, K., Morota, T., Yamamoto, S., Nakamura, R., Haruyama, J., et al. (2014). Geologic structure generated by large-impact basin formation observed at the South Pole-Aitken basin on the Moon. Geophysical Research Letters, 41, 2738–2745. https://doi.org/ 10.1002/2014GL059478 Orgel, C., Michael, G. G., Fassett, C. I., Riedel, C., Kneissl, T., van der Bogert, C. H., & Hiesinger, H. (2018). Ancient bombardment of the inner solar system—Reinvestigation of the "fingerprints" of different impactor populations on the lunar surface. Journal of Geophysical Research: Planets, 123, 748–762. https://doi.org/10.1002/2017JE005451 Pasckert, J. H., Hiesinger, H., & van der Bogert, C. H. (2018). Lunar farside volcanism in and around the South Pole-Aitken basin. Icarus, 299, 538–562. https://doi.org/10.1016/j.icarus.2017.07.023 Pieters, C. M., Head, J. W., Gaddis, L., Jolliff, B., & Duke, M. (2001). Rock types of South Pole-Aitken basin and extent of basaltic volcanism. Journal of Geophysical Research, 106(E11), 28001–28022. https://doi.org/10.1029/2000JE001414 Pieters, C. M., Staid, M. I., Fischer, E. M., Tompkins, S., & He, G. (1994). A sharper view of impact craters from Clementine data. Science, 266(5192), 1844–1848. https://doi.org/10.1126/science.266.5192.1844 Potter, R. W. K., Collins, G. S., McGovern, P. J., & Kring, D. A. (2012). Constraining the size of the South Pole-Aitken basin impact. Icarus, 220(2), 730–743. https://doi.org/10.1016/j.icarus.2012.05.032 Potter, R. W. K., Head, J. W., Guo, D., Liu, J., & Xiao, L. (2018). The Apollo peak-ring impact basin: Insights into the structure and evolution of the South Pole-Aitken basin. Icarus, 306, 139–149. https://doi.org/10.1016/j.icarus.2018.02.007 Robinson, M. S., Brylow, S. M., Tschimmel, M., Humm, D., Lawrence, S. J., Thomas, P. C., & Hiesinger, H. (2010). Lunar Reconnaissance Orbiter Camera (LROC) instrument overview. Space Science Reviews, 150(1-4), 81–124. https://doi.org/10.1007/s11214-010-9634-2 Ryder, G., & Wood, J. A. (1977). Serenitatis and Imbrium impact melts: Implications for large-scale layering in the lunar crust. Proceedings the of Lunar and Planetary Science Conference, 8, 655–668. Schultz, P. H., (1997). Forming the South-Pole Aitken basin: The extreme games. Paper presented at Lunar and Planetary Science Conference 28th, 1259–1260, Houston, TX. Schultz, P. H., & Crawford, D. (2011). Origin of nearside structural and geochemical anomalies on the Moon. In W. A. Ambrose, & D. A. Williams (Eds.), Recent advances and current research issues in lunar stratigraphy, Geological Society of America Special Paper, (Vol. 477, pp. 141–159). https://doi.org/10.1130/2011.2477(07) Schultz, P. H., & Spudis, P. D. (1979). Evidence for ancient mare volcanism. Proceedings the of Lunar and Planetary Science Conference, 10, 2899–2918. Shevchenko, V. V., Chikmachev, V. I., & Pugacheva, S. G. (2007). Structure of the South Pole-Aitken lunar basin. Solar System Research, 41(6), 447–462. https://doi.org/10.1134/S0038094607060019 Smith, D. E., Zuber, M. T., Neumann, G. A., Lemoine, F. G., Mazarico, E., Torrence, M. H., & Bartels, A. E. (2010). Initial observations from the Lunar Orbiter Laser Altimeter (LOLA). Geophysical Research Letters, 37, L18204. https://doi.org/10.1029/2010GL043751 Spudis, P. D., Reisse, R. A., & Gillis, J. J. (1994). Ancient multiring basins on the Moon revealed by Clementine laser altimetry. Science, 266(5192), 1848–1851. https://doi.org/10.1126/science.266.5192.1848 Stöffler, D., Ryder, G., Ivanov, B. A., Artemieva, N. A., Cintala, M. J., & Grieve, R. A. F. (2006). Cratering history and lunar chronology. Reviews in Mineralogy and Geochemistry, 60(1), 519–596. https://doi.org/10.2138/rmg.2006.60.05 Stuart-Alexander, D. E. (1978). Geological map of the central far side of the Moon (Map I-1047). USGS. Taylor, G. J., Warren, P., Ryder, P. G., Delano, J., Pieters, C., & Lofgren, G. (1991). Lunar rocks. In G. H. Heiken, D. T. Vaniman, & B. M. French (Eds.), Lunar sourcebook, (pp. 183–284). Cambridge, New York, Melbourne: Cambridge Univ. Press.

IVANOV ET AL. 2611 Journal of Geophysical Research: Planets 10.1029/2018JE005590

Uemoto, K., Ohtake, M., Haruyama, J., Matsunaga, N., Yamamoto, S., Nakamura, R., et al. (2017). Evidence of impact melt sheet differentiation of the lunar South Pole-Aitken basin. Journal of Geophysical Research: Planets, 122, 1672–1686. https://doi.org/10.1002/2016JE005209 Vaughan, W. M., & Head, J. W. (2014). Impact melt differentiation in the South Pole-Aitken basin: Some observations and speculations. Planetary and Space Science, 91, 101–106. https://doi.org/10.1016/j.pss.2013.11.010 Vaughan, W. M., Head, J. W., Wilson, L., & Hess, P. C. (2013). Geology and petrology of enormous volumes of impact melt on the Moon: A case study of the Orientale basin impact melt sea. Icarus, 223(2), 749–765. https://doi.org/10.1016/j.icarus.2013.01.017 Warren, P. H., Jerde, E. A., & Kallemeyen, G. W. (1987). Pristine moon rocks: A "large" felsite and a metal-rich ferroan anorthosite. Proceedings the of Lunar and Planetary Science Conference, 17, E303–E313. Whitten, J. L., & Head, J. W. (2015). Lunar cryptomaria: Physical characteristics, distribution, and implications for ancient volcanism. Icarus, 247, 150–171. https://doi.org/10.1016/j.icarus.2014.09.031 Wieczorek, M. A., Jolliff, B. L., Khan, A., Pritchard, M., Weiss, B. P., Williams, J. G., et al. (2006). The constitution and structure of the lunar interior. Reviews in Mineralogy and Geochemistry, 60(1), 221–364. https://doi.org/10.2138/rmg.2006.60.3 Wieczorek, M. A., Neumann, G. A., Nimmo, F., Kiefer, W. S., Taylor, G. J., Melosh, H. J., .et al. (2013). The crust of the Moon as seen by GRAIL. Science, 339, 671–675, 6120, DOI: https://doi.org/10.1126/science.1231530. Wieczorek, M. A., & Phillips, R. J. (2000). The "Procellarum KREEP Terrane": Implications for mare volcanism and lunar evolution. Journal of Geophysical Research, 105(E8), 20,417–20,430. https://doi.org/10.1029/1999JE001092 Wieczorek, M. A., & Zuber, M. T. (2001). A Serenitatis origin for the Imbrian grooves and South Pole-Aitken thorium anomaly. Journal of Geophysical Research, 106(E11), 27,853–27,864. https://doi.org/10.1029/2000JE001384 Wilhelms, D. E. (1987). The geologic history of the Moon. USGS Special Paper 1348, pp. 302. Wilhelms, D. E., Howard, K. A., & Wilshire, H. G. (1979). Geologic map of the south side of the Moon (Map I-1192). USGS. Wilhelms, D. E., & McCauley, J. F. (1971). Geologic map of the near side of the Moon (Map I-703). USGS. Wilhelms, D.E., & El-Baz, F. (1977). Geologic map of the east side of the Moon (Map I-948). USGS. Yamamoto, S. R., Nakamura, R., Matsunaga, T., Ogawa, Y., Ishihara, Y., Morota, T., et al. (2010). Possible mantle origin of olivine around lunar impact basins detected by SELENE. Nature Geoscience, 3(8), 533–536. https://doi.org/10.1038/ngeo897 Yingst, R. A., & Head, J. W. (1999). Geology of mare deposits in South Pole-Aitken basin as seen by Clementine UV/VIS data. Journal of Geophysical Research, 104(E8), 18957–18979. https://doi.org/10.1029/1999JE900016 Zuber, M. T., Smith, D. E., Lemoine, F. G., & Neumann, G. A. (1994). The shape and internal structure of the Moon from the Clementine mission. Science, 266(5192), 1839–1843. https://doi.org/10.1126/science.266.5192.1839

Erratum In the originally published version of this paper, the caption for figure 18 was published incorrectly. The correct Figure caption is as follows: “The frequency distribution of FeO concentration in (a) the pre-Nectarian units of the South Pole-Aitken floor and rim topographic domains, and (b) unit NpNc, which is the result of impact gardening of preexisting units.” Additionally, the acknowledgments section was published incorrectly. The acknowledgment statement should read: “We gratefully acknowledge very constructive and helpful reviews provided by Brad Jolliff and an anonymous reviewer who significantly improved the original version of the manuscript. M. A. I. was supported by the German Science Foundation (Deutsche Forschungsgemeinschaft – DFG) grant HI 1410\12-1 and Russian Science Foundation (grant 17-17-01149). H. H. and C. vd. B. were supported by German Space Agency (Deutsches Zentrum für Luft- und Raumfahrt – DLR) project 50OW1504 and as part of a project that has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement N°776276 (PLANMAP). C. O. was funded by DFG SFB-TRR-170 A3. All data relevant to this work are available at: http://www.planetology.ru. The following link can be used to download an archive with all data that are relevant to our paper: https://brownbox.brown.edu/download.php?hash=8db5bd78.” These errors have since been corrected, and this version may be considered the authoritative version of record.

IVANOV ET AL. 2612