Volume 49, Issue 15 e2022GL099241
Research Letter
Open Access

Survival of Ancient Lunar Water Affected by Topographic Degradation of Old, Large Complex Craters

C. L. Talkington

C. L. Talkington

Department of Geosciences, Auburn University, Auburn, AL, USA

Search for more papers by this author
M. Hirabayashi

Corresponding Author

M. Hirabayashi

Department of Geosciences, Auburn University, Auburn, AL, USA

Department of Aerospace Engineering, Auburn University, Auburn, AL, USA

Correspondence to:

M. Hirabayashi,

[email protected]

Search for more papers by this author
P. E. Montalvo

P. E. Montalvo

Department of Geosciences, Auburn University, Auburn, AL, USA

Search for more papers by this author
A. N. Deutsch

A. N. Deutsch

NASA Ames Research Center, Mountain View, CA, USA

Search for more papers by this author
C. I. Fassett

C. I. Fassett

NASA Marshall Space Flight Center, Huntsville, AL, USA

Search for more papers by this author
M. A. Siegler

M. A. Siegler

Planetary Science Institute, Tucson, AZ, USA

Department of Earth Sciences, Southern Methodist University, Dallas, TX, USA

Search for more papers by this author
S. L. Shepherd

S. L. Shepherd

Department of Geosciences, Auburn University, Auburn, AL, USA

Search for more papers by this author
D. T. King Jr.

D. T. King Jr.

Department of Geosciences, Auburn University, Auburn, AL, USA

Search for more papers by this author
First published: 19 July 2022

Abstract

Lunar water is redistributed by various processes. Topographic diffusion affects the transport of water and contributes to its preservation in subsurface layers. Here, we analyze 16 complex craters (∼3.2 – ∼4.2 Ga) larger than 20 km in diameter near the lunar south pole to quantify their degradation states. The results show that the diffusive rates of late Imbrian craters are similar to those of similarly aged simple craters, while Nectarian and pre-Nectarian craters are degraded less efficiently. Within a complex crater, the mass accumulation by topographic diffusion and ejecta blankets from other complex craters preserves water, ranging up to 1 wt% on average. However, impact mixing and internal heat further limit the stability of ancient water to subsurface regions with ages younger than 3.9 Ga and at depth from a few meters to 10s of meters.

Key Points

  • Topographic degradation of 16 complex craters near the lunar south pole significantly varies with surface age and crater size

  • Mass wasting by topographic diffusion and ejecta blanketing mixes water with regolith, giving a water mass fraction of up to 1 wt%

  • Impact mixing and internal heat limit the existence of unaffected ancient water at 10s of meters in depth and younger than 3.9 Ga

Plain Language Summary

Water in lunar shadowed regions is redistributed by various processes. When small craters form on a large crater's wall, its topography gradually erodes. Surface materials and deposited water move downhill, leading to a mixture of these materials and accumulations on the complex crater's floor. Here, we analyze 16 complex craters older than 3.2 Ga and larger than 20 km in diameter near the lunar south pole to investigate how fast these complex craters are topographically eroded. The results show that the speeds of topographic degradation of craters with an age of ∼3.2 Ga are comparable to those of similarly aged simple craters, while older craters experience significant degradation but are diffused inefficiently. On a complex crater, water may be stored in accumulated mass by topographic diffusion and large ejecta blankets for a long time. Such accumulated regions may host water, ranging up to 1 wt%. However, impact mixing and internal heat compete against the water supply and storage, limiting the water existence in subsurface layers at 10s of meters but a few meters below the surface and younger than 3.9 Ga.

1 Introduction

Various remote sensing observations have shown the existence of water and related species around the lunar poles (Lucey et al., 2021, and references therein). Potential water sources include meteoroid and comet impacts (Ong et al., 2010), solar wind implantation (Hurley et al., 2017), and volcanic outgassing (Head et al., 2020; Needham & Kring, 2017). Water delivery to the lunar surface was likely more active during and before the Imbrian era (≳3.2 Ga) than at present because of a higher rate of asteroid and comet impacts, and enhanced volcanic outgassing. However, competing processes may have redistributed water over time. A key issue is whether the observed surface water indicates recent or ancient deposition. The present study seeks possible interpretations of the existence of ancient water (≳3.2 Ga) in the subsurface areas on lunar south polar complex craters by considering the net effect on water redistribution due to topographic diffusion (Fassett et al., 2017; Fassett & Thomson, 2014), ejecta blanketing from complex craters (Cannon et al., 20202022; Tai Udovicic et al., 2022), impact mixing (Costello et al., 2021), and thermal stability in subsurface layers (Martinez & Siegler, 2021; Paige et al., 2010; Schorghofer, 2008; Siegler et al., 2015).

Lunar water that is cold trapped in permanently shadowed regions (PSRs) via exospheric water transport (Prem et al., 20152019) is thermally stable on a surface over geologic timescales (Figure 1a). However, water deposited on a surface exposed to ejecta blanketing, impact mixing, and internal heat experiences significant disturbance causing redistribution. Ejecta blankets overlay water deposits, possibly keeping them in subsurface regions for a long time (Figure 1c). Numerous impacts mix water with regolith and can generate enough heat to mostly dry out excavated regions (Cannon & Britt, 2020; Costello et al., 2021) (Figure 1d), though volatiles may still survive under impact induced heat (Hirabayashi, 2021; Potter & Deutsch, 2022). Internal heat further limits the existence of water in a subsurface layer (Figure 1d).

Details are in the caption following the image

Water redistribution in a shadowed region. Panel (a) shows water deposition (Prem et al., 20152019). Panel (b) describes impacts occurring on a crater wall. Panel (c) depicts mass wasting and mixing water with regolith over the wall, which overlay water depositions on a crater (Fassett et al., 2017; Fassett & Thomson, 2014). The accumulated layer mixed with water deposits is further buried by ejecta blankets from other complex craters (Cannon et al., 20202022; Tai Udovicic et al., 2022). Panel d illustrates that this mixed region is further affected by impact mixing (Costello et al., 2021) and internal heat (Martinez & Siegler, 2021; Paige et al., 2010; Schorghofer, 2008; Siegler et al., 2015). A series of these processes likely occur iteratively. The legend defines each color unit.

Similar to ejecta blanketing, topographic diffusion, a process that gradually erodes sharp morphological features over a long period by numerous small impacts (Fassett et al., 2017; Fassett & Thomson, 2014; Ross, 1968; Soderblom, 1970), may also contribute to water preservation in subsurface layers (Figures 1b and 1c). Mass wasting over a sloping surface on a complex crater may actively mix water (if it exists) with regolith on the slope, transport such materials to the floor, cover the deposited water on the floor, and create new accumulations containing water. As the accumulation continues, this region may become a harbor of ancient water if not affected by thermal variations and impact mixing. At present, however, topographic diffusion of complex craters is poorly constrained, though some work has attempted to quantify this issue numerically (Riedel et al., 2020).

The major purpose of this work is to (a) analyze topographic diffusion of 16 large craters located near the lunar south pole (Cannon et al., 2020; Deutsch et al., 2020; Tye et al., 2015) and (b) characterize the net effect of the four geologic processes discussed above on the ancient water distribution in PSR subsurface layers on these craters. The results for topographic diffusion of complex craters offer a direct comparison between the degradation rate of complex craters and that of small, bowl-shaped, simple craters (Fassett et al., 2018; Fassett & Thomson, 2014), as well as model predictions of complex crater topographic diffusion (Riedel et al., 2020). The derived diffusion speed is then converted to the magnitude of water-regolith mixing. With these results, we infer a plausible distribution of ancient water by adding the effects of ejecta blanketing, impact mixing, and internal heat to our analysis. Note that our study specifies ancient water as water delivered to a surface during or before the Imbrian. We describe water without stating ice. The reason is that water molecules would coexist as a solid or gas (but not a liquid) under equilibrium vapor pressure in a subsurface layer (Schorghofer, 2008). Finally, we use the term water stability to describe a condition when since deposited on a hosting crater, water has been unaffected by impact mixing and internal heat and has avoided mixing with water of other ages since.

2 Topographic Diffusion on Complex Craters

2.1 Mass Wasting Volume

We first compute the total volume of mass wasting by topographic diffusion over a complex crater. The approach considers a contrast between the modeled crater production and the observed superposing crater population on a crater wall, which provides the total number of craters erased by a series of mass movement events (Tye et al., 2015). Considering the total volume of the erased craters' cavities gives a lower bound of the mass wasting volume over time; larger mass wasting events may easily erase craters on the wall, while smaller mass movements are unsuccessful at erasing them. This volume is also affected by secondary events accompanied with ejecta blankets from other complex craters.

Figure 2 shows the cumulative crater size-frequency distributions (CSFDs) of the walls of 16 complex craters (20.9–107.8 km in diameter) surrounding the lunar south pole (Cannon et al., 2020; Deutsch et al., 2020; Tye et al., 2015). See Figure S1 in Supporting Information S1 for their crater morphologies and counts. Details of crater counting and model age determination are also provided in Texts S1 and S2 in Supporting Information S1. The related data set is available for reference (Data Sets S1 and S2 in Supporting Information S1, publicly available through Talkington and Montalvo (2022)). The Neukum model ages by an age determination approach based on Poisson's statistics (Michael et al., 2016; Michael & Neukum, 2010) are in general consistent with earlier work (Cannon et al., 2020; Deutsch et al., 2020; Tye et al., 2015), though some are not. As both floors and walls formed at the same time, if the geologic conditions are the same, they should have the same crater frequency. These inconsistencies mainly result from the fact that our search regions have different geologic conditions and erosion speeds from those studied by earlier work (Cannon et al., 2020; Deutsch et al., 2020). For each complex crater, among the ages of these different geologic units, we take the older one as the crater age. For example, the youngest crater age is 3.2 Ga (Weichert J). Also, de Gerlache and Nobile are outlier samples, given extremely large craters within them. Nevertheless, we account for these complex craters as part of our statistical samples.

Details are in the caption following the image

Cumulative crater size-frequency distributions (CSFDs) of 16 complex craters. The x axis is the crater diameter of counted craters on the wall, and the y axis shows the CSFD. Red dots with uncertainties show empirically derived data samples in this study, while solid lines show the Neukum-based chronology models. The blue lines are derived based on our analysis, while the green lines are based on earlier work (Cannon et al., 2020; Deutsch et al., 2020; Tye et al., 2015). Three isochrons are also added to panels. Dotted lines are for 0.5 Ga, dot-dashed lines are for 1.25 Ga, and dashed lines are for 2.5 Ga.

The results show that older craters tend to have shallower CSFD slopes at a few kilometers in crater diameter than predicted by the Neukum model. We interpret this as crater depletion as a result of mass wasting. Gradual mass movements driven by smaller impacts (Fassett & Thomson, 2014) or instantaneous events (Dundas et al., 2010) erode the surface topography. On the contrary, the CSFD slopes of craters smaller than 700 m in diameter again become steeper, some of which are as steep as the slope of the Neukum-based production model. Many of these craters are relatively fresh and thus look young, implying that large mass wasting mainly occurred during the early evolution stage of the Moon. In fact, identifying crossing points between isochrons and CSFD slopes reveals that their retention ages are likely as low as 0.5 Ga (Figure 2).

Taking differences between the Neukum production cumulative CSFDs and the counted cumulative CSFDs on crater walls, we obtain an average resurfacing depth of each complex crater over its age by computing the total mass wasting volume and dividing it by the crater area (Text S3 in Supporting Information S1). Figure 3a shows the resurfacing depth variations as a function of the age. Late Imbrian craters (∼3.4 Ga) have a resurfacing depth of ∼10 m, slightly shallower than that of simple craters, which may experience tens of meters over ∼3 Ga (Fassett & Thomson, 2014). This shallow resurfacing depth implies that while the total amount of mass wasting by topographic diffusion may be similar, their large crater sizes make the resurfacing depth shallow.

Details are in the caption following the image

Resurfacing of the 16 complex craters near the lunar south pole. Panel (a) shows the resurfacing depth as a function of age. The color map describes the crater diameter variation. The dashed line is a scaling relationship. Panel (b) gives the crater size dependence of the κt value. The solid line shows the scaling relationship for small, simple craters less than 5 km in diameter (Fassett et al., 2018). The dashed line represents the derived scaling relationship of craters larger than 35 km in diameter. The color map is the Neukum-based model age. Panel (c) shows the variations in the κt value with age. Panel d illustrates the κt value corrected by the size dependency given in Panel (b) The color maps of Panels (c and d) represent the crater diameter. For Panels (c and d), the results by Fassett and Thomson (2014) and Fassett et al. (2018) are only applicable to the ages shorter than 3.8 Ga. In panels, T0 is the reference time defined as 3.4 Ga, while D0 is the reference diameter, which is defined as 35 km. All dashed lines here are derived using the least-square approach.

2.2 Degradation State

We apply the diffusive equation to characterize topographic diffusion and its speed (Fassett & Thomson, 2014). Below, we determine the degradation state, κt, an integral of the diffusivity over time, to quantify the diffusive rate of each complex crater (Text S4 in Supporting Information S1).

The size dependence of κt (Figure 3b) exhibits two distinctive trends. First, the κt values of craters smaller than ∼35 km in diameter are aligned along the scaling relationship for simple, bowl-shaped craters (Fassett et al., 2018). Second, those of larger craters are orders of magnitude higher, and the size dependency slope power in this diameter range is 0.48. These two trends also correlate with surface age. Those having higher degradation states are all Nectarian and pre-Nectarian craters. Those trends infer a transition from a lower degradation trend (solid line) to a higher degradation trend (dashed line).

The derived κt suggests a significantly higher degradation process on complex craters than that predicted for simple craters (Figure 3c). The dimensional κt values are multiple orders of magnitude higher than those for small craters (Fassett & Thomson, 2014). On the other hand, Figure 3d displays the κt corrected by the size dependency described in Figure 3b. Both scaling relationships cross at 3.4 Ga, but the one for complex craters rapidly increases before that age. This finding implies that the complex craters younger than 3.4 Ga are apparently diffused at a similar level to those of simple craters (Fassett et al., 2018), while older craters are efficiently degraded but can still be visible because of their sizes.

3 Water Redistribution at Lunar South Pole

3.1 Mixing of Ancient Water With Mass Wasting From Crater Walls and Ejecta Blankets

After water deposition on a large complex crater (both wall and floor), mainly caused by cold traps of water molecules generated by various processes (Prem et al., 201520182019), topographic diffusion and ejecta blankets induce material transport from its wall to the floor and overlay water deposits on the floor. Topographic diffusion mixes water with regolith in two ways. First, water may be mixed with regolith on the wall when a series of mass wasting occurs. Second, such transported materials overlay accumulated water layers on the floor. This layer is further buried by and mixed with ejecta blankets coming from complex craters (Cannon et al., 20202022; Tai Udovicic et al., 2022). In contrast to impact mixing, which may induce heating and thus water destabilization (Costello et al., 2021), these processes may not significantly heat volatile deposits, and thus regolith layers may cover water layers without sublimation.

To roughly determine the water mass fraction in these accumulated layers, we consider the net depth, that is, the resurfacing depth by topographic diffusion and ejecta blankets. The water mass fraction is defined as a ratio of the amount of water supplied to a cold unit surface to the net depth times material density. From Figure 3a, the resurfacing depth within a time range between 3.2 Ga and 4.2 Ga is given as urn:x-wiley:00948276:media:grl64559:grl64559-math-0001, where T is time in Ga, T0 is 3.4 Ga, and u is given in km. On the other hand, according to Cannon et al. (2020), the mean depth of ejecta blankets emplaced over time is given as urn:x-wiley:00948276:media:grl64559:grl64559-math-0002 (Table S2 and Text S5 in Supporting Information S1).

Full description of how water has been deposited in these polar cold traps is beyond the scope of our work, though earlier work suggested that imparted water molecules might be distributed widely as part of the exosphere over the Moon (Prem et al., 20152019). Such a condition may lead to relatively homogeneous water accumulation over cold regions, if their thermal conditions are met (Siegler et al., 2015). Based on this scenario, we simply compute the water supply density, that is, water deposition over a unit area, by dividing the total water delivery amount by the lunar area, ∼40 Mkm2. While Cannon et al. (2020) employed a Monte Carlo approach to analyze the amount of water deposits, we simply apply the reported estimates below. We account for three sources of water supply in a PSR. The first source is solar wind, and its contribution to water generation over the entire lunar surface over 1 Ga is 5 × 1011 kg/Ga (Hurley et al., 2017), giving its density as 1.38 × 10−2 kg/Ga/m2. The second source is asteroids and comets. The water delivery rate over 1 Ga is 2.5 × 1015 kg/Ga, and its density is 66 kg/Ga/m2 (Ong et al., 2010). The last source is volcanic outgassing, which may have delivered 1 × 1014 kg of water, equivalent to 2.6 kg/m2 in total, to the lunar surface during the Imbrian (Head et al., 2020; Needham & Kring, 2017).

The water fraction is computed using the following assumptions. First, water delivered by solar wind is supplied constantly. This assumption is likely a simplification as earlier work suggested possible variations over lunar history (Borg et al., 1980). Second, volcanic activity-driven water is supplied equally during the Imbrian. This contribution, however, is currently under debate (Aleinov et al., 2019; Mandt et al., 2022). Third, the water supply made by asteroids and comets is proportional to their impact fluxes and so determined by applying the Neukum chronology model (Neukum et al., 2001). However, the comet impact flux should be significantly different from and much less than the asteroid flux (Morbidelli et al., 2018). Nevertheless, because the contribution of asteroids is far more significant than others, these assumptions do not affect our results. The results show that the water mass fraction over the resurfacing depth and the ejecta depth ranges between ∼0.2 wt% and ∼1 wt% over 3.2–4.2 Ga, which is comparable to (or slightly lower than) the water fraction observed on the lunar surface (Hayne et al., 2015; S. Li et al., 2018).

3.2 Internal Heat and Impact Mixing Contributing to Water Distribution

In addition to topographic diffusion and ejecta blankets, which help preserve water in subsurface layers, impact mixing and internal heat may further redistribute water. This section describes how the interaction of these processes contribute to the existence of the ancient water.

First, we incorporate impact mixing over time proposed by analytical work (Costello et al., 2021). We apply the in-site reworking zone, a region above the bottom of saturation excavation and burial by secondaries (Costello et al., 2021). The in-site reworking zone reaches up to 1 m at depth over 3 Ga (Costello et al., 2021). Direct impact cratering may induce enough heat to cause water sublimation within the crater cavity. We follow Costello et al. (2021) to assume that regions within the in-site reworking zone lack the water signature because of this process.

Next, we account for the influence of internal heat on subsurface water and determine the deepest region that water can be thermally stable. We develop a finite element thermal model based on Hayne et al. (2017) but incorporate the thermal conductivity variations in the lunar crust (Wieczorek & Phillips, 2000) and Thorium's decay heat (Kamata et al., 2013; Schubert et al., 2001; Turcotte & Schubert, 2014) to compute the subsurface equilibrium temperature up to 1 km in depth (Figures S2 and S3, Table S3, Texts S6 and S7 in Supporting Information S1). Because our interest is in characterizing whether water has been thermally stable since its deposition during and before the Imbrian, we apply the surface heat flux for the lunar far side 3.2–4.2 Ga ago. It may range between 16 mW/m−2 (3.2 Ga ago) and 24 mW/m−2 (4.2 Ga ago), determined by Laneuville et al. (2013), who provided predictions consistent with the measured surface fluxes by the Apollo missions (Langseth et al., 1976). We compute the subsurface equilibrium temperature distribution given three surface temperatures appropriate for PSRs, 50 K, 75 K, and 100 K (Figure S3 in Supporting Information S1).

Our goal of employing this approach is to find the thermal threshold, the depth that reaches the water stability temperature limit, 145 K, below which water can persist as ice over billions of years (Schorghofer, 2008). We focus on discussing the 16 mW/m−2 heat flux in the main texts to consider the upper end-member of the thermal threshold but provide the 24 mW/m−2 flux (Figure S3 in Supporting Information S1). We do not consider the influence of a different lunar polar orientation on the thermal conditions. If the present day rotation pole was different in ancient times (Siegler et al., 20152016), the ancient water stability may be affected not only by the internal heat but also by differences in solar illumination geometry.

Figure 4 illustrates where ancient water is stable (Software S1 in Supporting Information S1). The net depth of both topographic diffusion and ejecta blankets is an upper bound of the ancient water stability as any regions beneath the original crater floor may have been highly affected by an intense heat during the formation of the hosting complex crater (lower side of the blue lines). A subsurface region deeper than the deepest depth of the in-site reworking zone also allows water to remain stable (upper side of the red lines). However, the subsurface depth for water existence should be shallower than the 145 K thermal threshold to avoid a higher temperature (lower side of the black lines). There is an upper bound of the surface age where ancient water is stable. For example, if the surface temperature is 50 K (Figure 4a), water older than ∼3.9 Ga is unstable and must be transported to other regions where it is mixed with water of other ages. The 75 and 100 K surface temperatures make water unstable in areas older than ∼3.75 Ga and ∼3.55 Ga, respectively (Figures 4b and 4c). Furthermore, if the surface is younger than ∼3.55 Ga (for 50 K), ∼3.45 Ga (for 75 K), and ∼3.3 Ga (for 100 K), the ancient water stability is bounded by the net accumulated depth. If the internal heat is the lower end-member of the thermal threshold, that is, 24 mW/m2, areas with the 50 K surface temperature should be younger than 3.7 Ga to keep ancient water stable (Figure S4 in Supporting Information S1).

Details are in the caption following the image

Existence of ancient water affected by material accumulation owing to ejecta blanketing (Cannon et al., 2020) and topographic diffusion, as well as two disturbing factors, impact mixing and internal heat. For all the panels, the red lines illustrate the depth affected by impact mixing (Costello et al., 2021), while the blue lines describe the net depth driven by ejecta blankets (Cannon et al., 2020) and topographic diffusion. EB stands for ejecta blankets, while TD means topographic diffusion. The black lines describe the 16 m W/m−2 thermal threshold. The cyan regions are areas where ancient water may be preserved. Panels (a, b, and c) show cases when the surface temperatures are 50 K, 75 K, and 100 K, respectively.

4 Discussion

4.1 Low Size-Dependence of Complex Craters' Topographic Diffusion

Earlier work reported that the degradation state may be proportional to D2 (Basilevsky et al., 2018). This slope power defines an ideal situation when the degradation state becomes identical to geometric similarity (Minton et al., 2019). On the other hand, we found that, for craters larger than 35 km in diameter, the degradation state follows a scaling relationship proportional to D0.48, which is shallower than the ideal slope condition, that is, 2. The degradation state slope power less than two implies anomalous topographic diffusion, which means less efficient degradation and directly correlates with the observed variations in their geometries (Kalynn et al., 2013). Our shallow slope power is consistent with other studies (Fassett et al., 2018; Riedel et al., 2020; Xie et al., 2017).

We observed an apparent transition from the degradation state of smaller craters to that of large craters up to 35 km in diameter. If this is indeed a transition, there are two possible explanations. First, it is attributed to the crater size. The transition should occur at a crater diameter ranging between 5 and 35 km, given earlier work that analyzed topographic diffusion of craters with 800 m - 5 km in diameter (Fassett et al., 2018). One possibility may be a complex-simple crater transition on the Moon, which is about 15 km (Pike, 1977). If so, complex crater morphologies are more susceptible to large-scale mass wasting and thus have higher degradation states. However, they can still be visible because of their larger sizes. Second, this transition may simply result from the surface age, that is, the level of bombardments to which the craters have been exposed, which is consistent with the trends of topographic diffusion of small craters (Fassett et al., 2018; Fassett & Thomson, 2014). If this is the case, this transition may suggest a shift of the bombardment trend at some time around 3.4 Ga. Secondary events accompanied by ejecta blanketing from the complex crater formation, which are in general more numerous than primary events (Bierhaus et al., 2018), may enhance topographic diffusion. Alternatively, higher-velocity impacts can induce higher destructive cratering processes, leading to large-scale mass wasting and thus a higher κt. This interpretation supports a hypothesis that higher-velocity impacts were more frequent in an earlier bombardment phase (3.4 Ga to 4.1 Ga) than in a later lunar phase (Marchi et al., 2013; M. Xie et al., 2021). However, given our limited data samples, these discussions still lack a decisive conclusion and need further study.

4.2 Ancient Water Distribution

Our study constrained the conditions of ancient water, given the net effect of ejecta blanketing (Cannon et al., 2020) and topographic diffusion (this work), by accounting for disturbing factors including impact mixing and internal heat (Figure 4). While constraining the spatial distribution of such conditions is beyond our scope, we identified some general trends.

In Nectarian and pre-Nectarian subsurface layers, ancient water deposits are redistributed significantly owing to disturbing factors, that is, impact mixing and internal heat. The 16 mW/m2 (24 mW/m2) heat flux with a surface temperature of 50 K may cause a subsurface layer deeper than 70 m (30 m) to exceed the 145 K thermal threshold (Schorghofer, 2008). This limits the water stability in subsurface layers younger than 3.9 Ga (3.7 Ga). Continuous impact mixing may disturb the top surface layers, likely removing water efficiently (Costello et al., 2021). On these craters, water of this age or older has been significantly redistributed, possibly existing as surface water mixed with water of different ages. On the other hand, late Imbrian craters potentially host ancient water without any influence in a subsurface layer beneath a few meters up to ∼10 m in depth.

Finally, in this work, we do not explore direct correlations of topographic diffusion between water hosting craters and non-water hosting craters. Recent studies have suggested that there may be some topographic variations depending on the water existence (Rubanenko et al., 2019; Deutsch et al., 20212022), though some reports suggest different aspects (Y. Li et al., 2022). It may be possible that the existence of water, depending on its amount, may control the magnitude of mass wasting, varying the resulting topography. However, our limited data sampling challenges capturing such variations.

5 Conclusion

This report investigated topographic diffusion of 16 lunar south polar complex craters. Determining the contrasts between the crater populations observed on their walls and those predicted by the Neukum production model, we estimated the volume of mass wasting by topographic diffusion. With the resulting resurfacing volume of all the craters, our analysis revealed that the degradation state of late Imbrian craters is comparable to that for small, simple craters, while older craters apparently experience significantly higher degradation, despite some still having clear morphologies. However, the degradation efficiency is low, which does not satisfy geometric similarity and is consistent with earlier work (Fassett et al., 2018; Riedel et al., 2020). This finding implies that regardless of the observed high degradation, topographic diffusion acting on these craters is inefficient, compared to the case of geometric similarity. Topographic diffusion and ejecta blankets from other complex craters causes a mixing of water deposited during or before the Imbrian with regolith, leading to a water mass fraction of ∼0.2 to ∼1 wt%. However, impact mixing and internal heat narrowly limit the ancient water stability at 10s of meters in depth (but beneath areas affected by impact mixing) and on a surface younger than 3.9 Ga.

Acknowledgments

This work is performed supported by NASA/VIPER (NNH21ZDA001N/80NSSC22K0534), NASA/Solar System Exploration Research Virtual Institute (SSERVI) under cooperative agreement numbers 80ARC017M0007 (REVEALS), and NASA/EPSCoR (18-EPSCoR R3-0057).

    Data Availability Statement

    The statistics of the crater samples studied in this work are available by Talkington and Montalvo (2022) (https://zenodo.org/record/6706126). 16 csv files include crater counts of all the complex craters Talkington and Montalvo (2022). The file names are provided in Data Set S1 in Supporting Information S1, and the readme file contains the format of each csv file. Software written in Python 3 (3.9 and 3.10, confirmed) for determining the distribution of stable ancient water (Figures 4 and S4 in Supporting Information S1) is available through Hirabayashi (2022) (https://doi.org/10.5281/zenodo.6691603).