The crater retention ages of the mare deposits within the Orientale multi-ring impact basin are investigated using 10-m resolution images obtained by the SELENE (Kaguya) spacecraft, in order to constrain the volcanic history of the Moon around the nearside-farside boundary. Precise crater-counting analyses reveal that mare deposits in the Orientale region are much younger than previously estimated: ∼2.9 Ga mare basalt in the eastern part of Mare Orientale and ∼1.8–2.2 Ga mare deposits in Lacus Veris and Lacus Autumni, maria along the northeastern rings of the basin. The latter age estimates indicate that the Orientale region experienced volcanic activities ∼2 billion years after the basin-formation impact. The dominance of a uniform surface age across the mare deposits in the peripheral regions strongly suggests that these volcanic eruptions are contemporary with the elevated volcanic activity episode proposed for the Procellarum KREEP Terrane on the lunar nearside at ∼2 Ga and that this activity peak is much more widespread than previously estimated. The longevity of mare volcanism in the Orientale region further suggests high initial temperatures and/or high content of heat-producing elements in the underlying mantle of this region.
 The Orientale basin is a multi-ring basin located at the western limb of the Moon (19°S and 93°W) (Figure 1a). The basin is surrounded by the Inner Rook, the Outer Rook, and the Cordillera rings in a concentric arrangement. The largest mare occurs at the center of the basin (Mare Orientale, ∼90000 km2), and other smaller mare deposits called Lacus Veris (∼12500 km2) and Lacus Autumni (∼5000 km2) mainly occur along the northeastern part of the rings.
 Recent crater counts on nearside mare basalts revealed that mare volcanism continued until ∼1.5 Ga [Hiesinger et al., 2000, 2003; Morota et al., 2011], while the magma activity on the farside ceased 2.5–3.0 Ga [e.g., Haruyama et al., 2009], showing an apparent dichotomy in the timing of mare emplacement. The latest volcanism on the nearside occurs at the center of the Procellarum KREEP Terrane (PKT), where the surface abundances of radioactive elements are high [Jolliff et al., 2000]. However, the existence and extent of such late volcanic activity and any relationship with the abundances of heat-producing elements remain unclear. Orientale mare deposits, located at the nearside-farside boundary, are ideal for constraining the spatial extent of the possible volcanic activity peak period at ∼2 Ga in the PKT [Hiesinger et al., 2003; Morota et al., 2011].
 Several previous studies focused on the age of mare deposits associated with the Orientale basin by using crater size-frequency measurements [Greeley et al., 1993; Kadel et al., 1993; Whilhelms, 1987; Whitten et al., 2011]. In these studies, the basin formation and the following Mare Orientale emplacement were dated at ∼3.8 Ga and ∼3.5 Ga, respectively. They reported even younger model ages for Lacus Veris (∼2.29 Ga [Kadel et al., 1993]) and Lacus Autumni (∼2.85 Ga [Greeley et al., 1993]). Recently, Whitten et al.  investigated crater chronology and spectroscopic characteristics of mare deposits around the Orientale basin, finding that a ∼1.65 Ga mare basalt occurred in Lacus Autumni. However, these studies' age estimates for young volcanic eruption events are highly uncertain; error bars are about +/−1 billion years, mainly due to the low spatial resolution of the data used. Indeed, the spatial resolutions of M3 data and the Lunar Orbiter photographs, used by Whitten et al.  and Greeley et al. , were 140 m/pix and 60–100 m/pix, respectively. This large range in age-determination error prevents lunar scientists from making accurate estimations about the temporal and spatial distribution of post main-phase mare volcanism. Thus, an accurate age estimate for this region is highly valuable for understanding the global evolution of lunar volcanism.
 In this study, we conducted crater counts from the images obtained by the Terrain Camera (TC), a stereo camera that globally obtained images of the lunar surface with 10 m/pix resolution [Haruyama et al., 2008], onboard the Selenological and Engineering Explorer (SELENE) spacecraft, in order to investigate the ages of mare deposits. These high-resolution observations permit improved crater count statistics, thus ensuring highly reliable model age determinations. Based on these new age estimates, we will also discuss the duration and the timing of mare volcanism in the Orientale region and the temporal relationship with PKT volcanism.
2. Crater Counting
 Crater counting is a well-established technique for the dating of planetary surfaces. Based on the straightforward idea that older surfaces accumulate more craters, we can estimate the relative and absolute ages of surfaces by measuring the crater size-frequency distribution (CSFD) with image data. Here we will just briefly present the procedure, because the technique has been described in detail in a number of papers [Greeley et al., 1993; Hiesinger et al., 2000, 2003; Morota et al., 2011; Neukum, 1983].
 In this study, we performed crater size-frequency measurements manually with images map-projected in a transverse Mercator projection. Obvious secondary craters are eliminated from crater counts on the basis of their morphological characteristics (i.e., the clusters or the chains of multiple craters) [Melosh, 1989] (Figure S1 in the auxiliary material). The regions selected for this study's crater counting were: two regions from Mare Orientale (Southwest and East), a polygonal mare unit located 150 km southwest of Mare Orientale (Southwest Polygon), two regions from Lacus Veris (North and South), and three regions from Lacus Autumni (North, Middle and South) (Figure 1a). TC albedo data and the Clementine FeO map [Lucey et al., 1998] were used to define these units. We separated Mare Orientale SW from Mare Orientale East on the basis of relatively high albedo and low FeO content. Mare Orientale East has been considered to have occurred as a single massive event because the morphological data and the spectral data show no definitive evidence of multiple flow units [Whitten et al., 2011]. For crater counting, we carefully chose areas not contaminated by ejecta or secondary craters.
 The model ages of the mare deposits are determined with the Craterstats program [Michael and Neukum, 2010], which provides model isochron fittings. In this study, we used the cratering chronology and crater production functions by Neukum  to compare our results with previous ones. The crater production function of Neukum et al. was also used to evaluate the model dependence of crater retention ages. The observed size distributions for craters with diameter >250 m are fitted to the production functions to estimate the ages. The high-resolution data increased the number of craters more than one order of magnitude compared with that of the previous studies, leading to improved count statistics and more accurate age determination. Model fits are applied for >400 m craters for Mare Orientale SW and SW Polygon, since the maria are so old that the number of small craters reaches the saturation level. Craters larger than 2 km are used to determine the timing of Orientale basin formation. It is noted that errors in model age reflect only uncertainties in counting statistics and do not include any systematic uncertainties in the absolute age calibration of these functions.
3. Crater Retention Ages
Figure 2illustrates observed crater size-frequency distributions and production function fit plots. Model ages obtained in this study are summarized inTable 1, along with the values obtained in the previous works. The spatial distribution of the model ages for the units defined in Figure 1a is shown in Figure 1b. Our counting results on the Orientale basin impact melt indicate that the basin was formed 3.79 (+/−0.02) billion years ago (Figure 2a), which is consistent with several previous studies [Greeley et al., 1993; Whilhelms, 1987; Whitten et al., 2011].
 High-resolution observations revealed that the eastern part of Mare Orientale occurred at ∼2.9 Ga, which is much younger compared with previous estimates (∼3.45 Ga [Greeley et al., 1993]; ∼3.58 Ga [Whitten et al., 2011]) (Figure 2b). The elimination of secondary craters made possible by the high-resolution observations may account for the younger age. The southwestern part of Mare Orientale shows an age of ∼3.77 Ga, suggesting multiple eruption episodes within Mare Orientale. Southwest Polygon is the same age as the SW part of Mare Orientale. Note that the age of the Orientale event is calculated to be younger than those of the southwest region of Mare Orientale and the southwestern polygonal region, which is obviously contradictory to the stratigraphic relations shown when the model byNeukum et al.  is employed. Such a problem may be attributed to uncertainties in the production function.
Figure 2c shows that all the three measured regions of Lacus Autumni are consistently dated at ∼2 Ga, ranging from 1.5 to 2.5 Ga when errors are considered. Whitten et al.  pointed out that Lacus Autumni is as young as ∼1.66 Ga, although the model ages suffer from large errors (i.e., 1.65 + 0.83/−0.96 Ga for Lacus Autumni Middle and 2.38 + 0.88/−1.50 Ga for Lacus Autumni South). Additionally, the discrepancy in model ages with those of the adjacent mare deposits in Lacus Autumni (3.47 + 0.11/−0.68 Ga for Lacus Autumni North) was as large as ∼2 Ga. In contrast, we obtained consistent and uniform model ages of the mare basalts in Lacus Autumni (Figure 2c). Our results also yield model ages systematically younger than those estimated by Greeley et al.  (2.85 + 0.37/−0.67 Ga) and Kadel et al.  (2.85 + 0.28/−0.57 Ga), with our positive error bars slightly overlapping with their negative error bars. Since the CSFD shows no evidence of a resurfacing event (see next paragraph for details), we might attribute these apparent discrepancies to the inadequate crater statistics and/or differences in count areas. Indeed, we are able to reproduce similar ages when we conduct a production function fit in the larger crater diameter range (i.e., >400 m).
 We then investigated two separate regions of Lacus Veris and found that they have somewhat complicated stratigraphy among three lava units, with the ages of the youngest units at ∼2.2 Ga for both of the two separate regions (Figure 2d). On the other hand, older model ages have been reported for Lacus Veris by previous studies: 3.20 + 0.13/−0.30 Ga [Whitten et al., 2011] and 3.50 + 0.05/−0.08 Ga [Greeley et al., 1993]. Indeed, our crater counting analyses also yield ages (∼3.3 Ga and ∼3.5 Ga) for the same larger diameter ranges consistent with those reported in these previous studies. Since resurfacing by a lava flow selectively erases small craters, making the deflection of CSFD at the diameter corresponding to its thickness [Hiesinger et al., 2002], we interpret this data to mean that this region is covered by a thin lava flow, which has not been found previously with lower resolution images, on the older units. The thickness of the lava flow is estimated to be 10–15 m, on the basis of the relation between rim height and diameter [Pike, 1980], since the deflection of the CSFD occurs at around a diameter of 350 m.
4. Volcanic History in the Orientale Region and Its Implications
 On the basis of the estimated ages for mare deposits, we can construct the following sequence for the volcanic history in the Orientale region. (1) A 3.79 Ga impact formed the Orientale basin. (2) Magma eruptions in Mare Orientale started soon after the Orientale event (3.77 Ga). (3) Subsequently, the large area of the basin center was resurfaced at ∼2.9 Ga. (4) Lacus Veris may have erupted ∼3.4–3.6 Ga, but the latest eruption occurred at ∼2.2 Ga after an interval of 1 billion years. (5) Lacus Autumni also erupted at 1.8–2.2 Ga, which is about 2 billion years after the basin-forming impact. These recent mare eruptions coincide with the elevated volcanic activity episode proposed for the PKT on the nearside at around ∼2 Ga, suggesting a possible correlation with the volcanic activities in the PKT (Figure 3) [Hiesinger et al., 2003; Morota et al., 2011].
 We have shown that the duration of mare volcanism in the Orientale region is longer than previously estimated: ∼1 Gyr in Mare Orientale and ∼2 Gyr in the basin rim regions with the mare deposits. Possible volcanic mechanisms to account for such young eruptions are briefly discussed below, with one of the candidate mechanisms showing no obvious contradictions with current observations.
 Based on a simple thermal conduction model, Wieczorek and Phillips suggested that heat-producing elements concentrated in the lower crust of the PKT could cause direct partial melting in the underlying mantle immediately after magma ocean crystallization, and that the melting zone would have remained under the central region of the PKT over much of the lunar history. Since the Orientale basin is located in the southwest region of the PKT, it is possible that a radioactive element-rich layer extends underneath the northeastern part of the Orientale basin, which might account for the apparent distribution of mare deposits concentrated in the northeastern rims. However, gamma-ray data obtained by orbiting satellites have revealed that the concentration of radioactive elements appears low in the Orientale region (e.g., <0.5 ppm of U, <1 ppm of Th) in the current spatial resolution (130 km/pix) [Yamashita et al., 2010] despite the expectation that the lower crust would have been excavated (∼50 km in depth [Wieczorek and Phillips, 1999]) and exposed by the basin-forming impact. Such values are much lower than those of the PKT (5–10 ppm of Th) and comparable with those of the Feldspathic Highland Terrane (FHT, <1 ppm of Th).
 On the basis of numerical simulations with 3D mantle convection models, Ziethe et al.  pointed out that a megaregolith layer with low thermal conductivity could prevent the melted layer from cooling until ∼2 Ga. This mechanism might explain the young mare emplacement in the Orientale region, although it should be noted that the estimated thickness of the regolith layer in the Orientale region is not significantly different from that in other areas [Petro and Pieters, 2008].
 Other possible explanations are that a partially molten layer was maintained by extremely high initial mantle temperature, and/or by the high abundance of heat-producing elements underlying the mantle [Spohn et al., 2001]. Although the initial temperature of the Moon and the concentration of the heat-producing elements in the lunar mantle have not been fully constrained at present, such interpretations do not contradict the current observations of the lunar surface. That being the case, a thermal evolution model to reconcile the hot mantle with the large gravity anomaly of the basin [Namiki et al., 2009] is required.
 In this study we conducted crater counting analyses using 10 m/pix resolution images obtained by the SELENE spacecraft, in order to elucidate the volcanic history of the Orientale basin. We found widespread young mare deposits in the Orientale region: ∼2.9 Ga at the eastern Mare Orientale, ∼2.2 Ga at Lacus Veris and ∼1.8–2.2 Ga at Lacus Autumni. The mare basalt emplacement on Lacus Veris and Lacus Autumni is contemporary with young mare volcanism found at the PKT on the lunar nearside. The uniform age (∼2 Ga) of the mare deposits around the Orientale basin rims, which are rather far away from the PKT, suggests that the suspected peak of volcanic activity at around 2 Ga was not limited to the regions with enhanced radioactive elements at the surface (PKT) but extended to other regions. The longevity of mare volcanism in the Orientale region suggests high initial temperatures and/or the high content of heat-producing elements in the underlying mantle of this region.
 We wish to thank N. Petro and J. Fritz for their careful reviews and constructive comments for improving the paper. This research greatly benefited from the 4th Laboratory Training Course on Data Analysis of Planetary Exploration, sponsored by the Center for Planetary Science (CPS) under the MEXT Global COE Program, “Foundation of International Center for Planetary Science,” a joint project between Kobe University and Hokkaido University. We greatly appreciate the proofreading/editing assistance from the GCOE program for improving the paper. Y. Cho was supported by a Grant-in-Aid from the Japan Society for the Promotion of Science (JSPS) Fellows.
 The Editor thanks Noah Petro and Joerg Fritz for assisting in the evaluation of this paper.