Stabilized impact flux on the Moon at 2.8 Ga constrained by Chang’e-6 low-Ti basalts

Current knowledge of lunar impact flux

Impact craters are the most ubiquitous landforms on terrestrial planetary surfaces, recording crucial information about planetary evolution. Due to its minimal weathering rates, the Moon preserves the most complete impact record, serving as the primary reference for reconstructing the impact history in the solar system.¹ Impact craters formed by extralunar impactors (i.e., primaries) since the emplacement of a geological unit constitute the production population, and their size-frequency distribution (SFD) can be mathematically described by the production function (PF).¹ In combination with crater spatial densities, the radiometric ages of samples returned from the Moon provide a baseline to construct the lunar impact flux, which is termed the crater chronology function (CF).¹ The CF and PF enable the construction of crater density isochrons at diameters of at least 10 m for arbitrary model ages. Vice versa, the model ages of specific crater populations can be derived by fitting their SFD against the CF and PF.¹

The first-order reliability of the lunar crater chronology has been repeatedly verified. For example, the current impact flux on the Moon, derived from newly formed impact craters, is generally consistent with the annual production rate predicted using the canonical Neukum crater chronology (cf. Xiao et al.²); the model ages of the Chang’e-5 landing area, derived from crater statistics, align well with the radioisotope ages of returned local basalts (cf. Xiao et al.²). However, significant uncertainties persist in the canonical crater chronology model, especially in the age interval of ∼1.0–3.2 Ga, which was only recently anchored by Chang’e-5 basalts.² Enhancing the reliability and precision of the lunar crater chronology remains a key objective of ongoing and future lunar exploration missions.²

Despite being calibrated using the same anchor points, different versions of CF exist due to uncertainties in sample ages and crater densities (cf. Xiao et al.²). All available CF models suggest a similar overall trend of lunar impact flux, which features an early sharp decline and a quasi-constant rate at later times. However, the timing of this transition varies from ∼3.7 to 3.0 Ga according to different CF models (cf. Xiao et al.²). In addition, recent studies have revealed episodic short-term spikes of impact flux over the past ∼1 billion years (cf. Xiao et al.²), though their potential effects on the overall stability of impact flux remain unresolved. Deciphering the post-3.7 Ga changes of lunar impact flux will require new samples formed within this time frame.

New calibration point established by Chang’e-6 low-Ti basalts

The Chang’e-6 mission returned the first sample from the lunar far side, providing an invaluable opportunity to establish a new calibration point for the lunar crater chronology. Among the diverse components in Chang’e-6 regolith, low-Ti basalt fragments that originated from the local mare were recognized based on their consistent geochemical characteristics and uniform radioisotope ages of 2,807–2,830 Ma.^3,4 This age further validates the first-order reliability of the canonical crater chronology model¹ because geological context studies revealed that the surface regolith at the landing site is predominantly composed of local mare basalts, and the reported model ages of ∼2.4–3.3 Ga (cf. Luo et al.⁵) are within ∼20% of the measured radioisotope ages.^3,4 More essentially, a new anchor point at ∼2.8 Ga can be established to refine the lunar crater chronology, providing additional insights into the change of lunar impact flux.^3,4

Establishing a new calibration point for CF requires both the sample age that represents the formation time of a geologic unit and the spatial density of primaries formed in the unit. Canonically, this crater density is for craters with diameters equal to or larger than 1 km, i.e., N(1) value, but it can also be obtained by interpolating SFDs of smaller or larger craters using a PF.² However, PFs for the Moon are subject to considerable uncertainties due to issues such as nonuniform effects of topographic degradation and contamination by secondary craters (secondaries) on different-aged crater populations (cf. Xiao et al.²). For the Chang’e-6 landing mare, the reported N(1) values varied significantly from 1.91 × 10⁻³ to 3.26 × 10⁻³ km⁻², which were derived from statistics of craters with D < 1 km using the Neukum PF (cf. Luo et al.⁵). These discrepancies arise from the inherent subjectivity in crater statistics, which involves assigning appropriate counting areas and distinguishing suitable primaries.⁵ This challenge is particularly pronounced in the Chang’e-6 landing mare, which contains abundant impact rays and secondaries of varying sizes.⁵

By comparing metadata of earlier crater statistics performed for the Chang’e-6 landing mare, unfavorable effects caused by solar illumination conditions and secondaries were noticed.⁵ To reduce these effects and also avoid uncertainties of deriving N(1) values using a PF, Luo et al.⁵ designed multiple counting areas and investigated the production populations within specific diameter ranges of impact craters, considering that topography roughness and the proportion of secondaries in visible crater populations are both scale dependent. They identified the production populations for the landing mare at three diameter segments of 75–503, 400–864, and 1,000–1,799 m. The three production populations exhibit a random spatial distribution and follow the same power-law relationship, yielding a direct measurement of N(1) value for the landing mare of (2.01 ± 0.90) × 10⁻³ km⁻².⁵

The production populations in the landing mare have comparable crater SFD to the canonical Neukum SFD isochrons with model ages of ∼2.4–3.3 Ga, but the obtained N(1) based on multiple production populations is independent of any PF.⁵ The same approach of deriving production populations was applied to the lower-Ti mare located to the east of the landing mare, yielding a larger N(1) value of (6.05 ± 2.71) × 10⁻³ km⁻². The production populations in the three different diameter segments of the eastern mare are systematically larger than those on the landing mare, yet they still follow the same power-law SFD (see Luo et al.⁵ for details). This observation is consistent with an earlier stratigraphic investigation indicating an older emplacement age for the eastern mare (cf. Luo et al.⁵).

Stabilized impact flux on the Moon at 2.8 Ga

Using the radiometric age of ∼2.8 Ga and the N(1) of the Chang’e-6 landing mare, Cui et al.³ updated the lunar crater chronology, yielding N(1) = $1.94\times {10}^{-14}\left({\mathrm{e}}^{7.196\mathrm{t}}-1\right)+7.652\times {10}^{-4}t$, where model age t is in Ga (Figure 1). The new crater chronology model provides a better constraint for the impact flux around 2.8 Ga. The overall deviation from the Neukum canonical model is minimal, with the maximum difference in derived model ages being less than 0.25 Gyr for a given N(1) value (Figure 1A). In addition, the updated lunar crater chronology model reveals a similar overall trend of impact history to the other available chronology models, showing that the lunar impact flux experienced a sharp decline in the early history before stabilizing in later periods (Figure 1B).

Figure 1.

A new calibration point of lunar impact flux established by Chang’e-6 low-Ti basalts

(A) Available calibration points (yellow circles) and crater chronology models (solid colored lines) of the Moon. The Chang’e-6 calibration point is represented by a red square. The vertical error bars represent the inherent uncertainties in the observed crater densities, while the horizontal error bars represent the uncertainties in the radiometric ages of lunar returned samples. The inset shows differences of model ages predicted using the Neukum et al.¹ chronology model and the updated model by Cui et al.³

(B) Lunar impact flux derived from the derivative of the crater chronology models shown in (A) with respect to model age.

Thanks to the Chang’e-6 sample, it is now confirmed that the lunar impact flux has remained largely stable since at least ∼2.8 Ga and possibly since ∼3.2 Ga (Figure 1B). It is notable that the eastern lower-Ti mare exhibits an equilibrium diameter of 189 m, corresponding to a model age of ∼3.25 Ga according to the empirical relationship between equilibrium diameter and the model age of the crater populations.⁵ The radiometric age for the emplacement of the eastern mare, which may be recorded in exotic ejecta in the Chang’e-6 regolith (e.g., extremely low-Ti basalts³), shall provide further constraint on the transition time of the impact flux when combined with the N(1) value reported by Luo et al.⁵

Multiple lines of evidence have suggested that the post-2.8 Ga impact flux in the Earth-Moon system has experienced episodic spikes rather than being uniformly constant.² Examples include the concentrated occurrences of terrestrial impact spherule layers between ∼2.0 and 2.5 Ga, the Ordovician breakups of parent bodies of the L-chondrites, and multiple concentrations of radioisotope ages of post-2.03 Ga impact glass spherules in Chang’e-5 regolith (cf. Xiao et al.²). For comparison, the available calibration points of the lunar crater chronology predict an overall stable impact flux since 2.8 Ga (Figure 1B). While the proposed short-term peaks of impact flux do not necessarily correspond to impact events with magnitudes equal to or greater than those represented by N(1), the apparent contradiction can be well accommodated by the fact that the updated crater chronology describes the long-term evolution of impact flux and does not preclude short-term variations, as these operate on different timescales. Essentially, more sample return missions from the Moon are needed to establish additional calibration points for the lunar crater chronology so that the transition times and potential episodic spikes of impact flux can be further deciphered.

Acknowledgments

The authors are supported by the National Natural Science Foundation of China (42241108, 42273040, 42473049, 42402232, and 62227901), the Postdoctoral Fellowship Program of CPSF (grant GZB20240881), and the B-type Strategic Priority Program of the Chinese Academy of Sciences (grant XDB41000000). The editor and two anonymous reviewers provided constructive comments.

Declaration of interests

The authors declare no competing interests.

Published Online: February 21, 2025