A Seismic Origin for Lunar Coldspots

A SEISMIC ORIGIN FOR LUNAR COLDSPOTS. N. C. Schmerr1, E. Frizzell1, C. Hartzell1, H. Bernhardt1, J. Clark1, T. Powell2, and V. Lekic1, 1University of Maryland, College Park, 8000 Regents Dr. College Park, MD 20742, [email protected], 2Johns Hopkins Applied Physics Laboratory. Introduction: Lunar cold spots are regions of relatively lower regolith temperatures (3-10 K cooler than their surrounding) that extend 10-100 crater radii around young (<1 Ma) impact craters and are found over 1% of the lunar surface area [1, 2]. The continuous temperature anomaly of a cold spot surrounds the crater and can decay into spoke-like filaments with discontinuous and radially organized patterns with distance from the central crater (Fig. 1; a-b). The negative temperature anomaly is postulated to originate from the dilation or “fluffing up” of the uppermost 1040 cm of regolith, resulting in reduced thermal inertia and lower heat retention [1]. Although they qualitatively resemble the patterns of ejecta found around impact craters, cold spots do not correlate with visible albedo variations, have no spectroscopic difference in optical maturity or composition from their surroundings, and would volumetrically exceed the materials excavated from the source crater [1, 2]. Given these observations, several mechanisms have been proposed to explain their existence; Bandfield et al. [1] suggested two hypotheses, 1) a series of cascading secondary impacts radiating outward from the primary [3], 2) impact generated gas flow and progressive aggradation features that resemble terrestrial pyroclastic flows. Both mechanisms require some level of granular flow and/or emplacement/disruption of new materials without albedo modifications, and it is unclear if lunar impacts can produce such effects. An alternative hypothesis is that seismic shaking from the impact event could produce the surface disruption patterns and the resulting regolith temperature anomaly. However, this mechanism has often been ruled out under the assumption that seismic waves cannot produce the spoke-like, non-circular geometries of cold spots. Granular wave dilation of regolith has been investigated by [4] as a mechanism for regolith fluffing and can achieve the loft depth and density reduction to depths of 40 cm needed to explain cold spot thermal properties but still does not explain the complex geographic shape of cold spots. Here we explore how a seismic wavefield generated by an impact and subsequently scattered through the heterogeneous lunar megaregolith and crust can reproduce the characteristic patterns of cold spots, and that the size and geometry of the cold spot are linked to the impact source momentum and source dynamics. Our results show that seismically induced ground vibration transmitted from the impact through the ground can produce the disruptions at the surface by granular lofting [5] and density reduction process, making seismic shaking a viable hypothesis for explaining the origin of lunar cold spots. Methods: We created a representative, globally distributed sub-catalog of 136 lunar cold spots that span the range of crater diameters (300-1750 m) and cold spot extents (6-170 km). We measure the extents of the continuous thermal anomaly and the maximum spoke length associated with each cold spot to determine the ratio between the continuous cold spot and region of disturbed radially distributed spokes (Fig. 1c). Notably there is a positive correlation between crater diameter and cold spot continuous extent (R=0.879) suggesting that the size of the cold spot scales with the size of the impact crater. We hypothesize that the transition between the central cold spot and spoke-like region represents the transition of the seismic wavefield below the critical ground motion amplitude needed to fluff up the regolith, and that topographic and structural scattering produce the longest spokes. To test this hypothesis, we generate 3-D seismic simulations of ground motion emanating from the impact crater source. Figure 1. Example lunar crater and surrounding cold spot located at 120.12ºE 29.73ºS, the impact crater is ~1 km in diameter. a) LROC NAC image of the fresh crater, extent of visible albedo changes from ejecta in cyan. b) Diviner bolometric temperature anomaly [6]. c) Crater diameters and measured cold spot extents, cross correlation coefficient indicated on the plot. d) Our scaling of bolide momentum for the measured cold spot database; see text for details. Cyan highlighting indicates the cold spot in parts (a) and (b). Impact source model. Crater diameters provide an estimate for the impact momentum and magnitude of the seismic source. The impact seismic source excitation function can be considered an integration of a directed force across a characteristic impulse duration; the impulse duration is derived from an empirical power law derived from measurements of Apollo-recorded impact seismic events [7]. This is further modified by a seismic efficiency factor that is dependent upon impact angle and ejecta removal from the crater [8]. Scaling of the momentum is achieved using [9] in the strength regime to relate the resulting crater radius to the bolide and target properties capable of producing an impact crater at the observed radius. It is important to note that there are strong tradeoffs in target and bolide properties and the resulting crater. We adopt target properties of lunar regolith, i.e., strengths of 1-65 kPa and densities of 1500-2100 kg/m3 and a range of bolide velocities between 10-30 km/s [10] to define momentum range (Fig. 1d). Seismic wavefield simulations. For a source function we adopt the implementation of [11]; our population of cold spot craters span an estimated range of impact momentums of 5x109–2x1013 Ns, with characteristic impulse durations over 1.5–4 seconds. Figure 2. Simulations of ground motion at a cold spot. a) Peak horizontal acceleration at an impact source for a homogeneous model. b) Peak horizontal acceleration for a 500 m autocorrelation length and a von Karman heterogeneity with 20% velocity variation. c) Decay of ground motion with distance from the source for (a, black), and (b, cyan). d) Isoseismals for the crater diameter distribution in our cold spot database (green). Blue circles indicate the measured extent of the cold spot. The error bar gives the uncertainty in amplitude for the range of impact momentums in Fig. 1d. The minimum ground acceleration at the cold spots is ~0.6 cm/s2. We use a cartesian finite difference code to compute seismic wavefields for our impact source [12] (Fig. 2 ab). The simulations are numerically accurate for wavefields of ~1 Hz and up to 100 km from the source. For background seismic velocity and density structure, we adopt the model of [13]. We test a range of von Karman random distributions of velocity with variable correlation scale lengths from 100-1000 m and velocity and density heterogeneities from 10-25% [14] (Fig. 2b). We fit the resulting simulation data with an isoseismal [15] and compare the resulting pattern of ground motion to the geographic extents of cold spots (Fig. 2 c-d). Finally, we carry out quantitative textural analysis of the cold spot and seismic shaking patterns using scattering transform [16]. Preliminary Results: The seismic simulations with scattering produce peak ground acceleration patterns that reproduce the characteristic spoke-like filaments of lunar cold spots. Wave propagation in heterogenous media leads to channelization of the seismic waves along the edges of large heterogeneities, which are unconstrained for the Moon, but should exist where impacts create variations in the crust via fracturing and ejecta layering [17]. Most importantly, the seismic radiation pattern is texturally similar to the temperature field; for the population of craters modeled, produces ground motions varying over orders of magnitude radially, and at levels capable of altering the lunar regolith thermal inertia through granular dilation [5]. Impact generated seismic waves passing through the regolith are thus a viable hypothesis for explaining cold spot anomalies on the Moon. Acknowledgments: NCS and VL acknowledge support from NASA SSERVI GEODES grant 80NSSC19M0216; HB, TP, and JC by the Lunar Reconnaissance Orbiter Camera team. References: [1] Bandfield, J. L. et al. (2014) Icar. 231, 221 10.1016/j.icarus.2013.12.017. [2] Williams, J. P. et al. (2018) J. Geo. Res.-Plan. 123, 2380 10.1029/ 2018je005652. [3] Chang, Y. R. et al. (2021) Rem. Sens. 13, 19 10.3390/rs13061087. [4] Frizzell, E. S. et al. (2023) Gran. Mat. 25, 35 10.1007/ s10035-023-01363-6. [5] Frizzell, E. S. et al. (2024) Gran. Mat. 26, 31 10.1007/ s10035-024-01460-0. [6] Powell, T. M. et al. (2023) J. Geo. Res.-Plan. 128, 21 10.1029/ 2022je007532. [7] Gudkova, T. V. et al. (2011) Icar. 211, 1049 10.1016/ j.icarus.2010.10.028. [8] Lognonne, P. et al. (2009) J. Geo. Res.-Plan. 114, E12003 10.1029/ 2008je003294. [9] Housen, K. R. et al. (2003) Icar. 163, 102 10.1016/ S00191035(03)00024-1. [10] Marchi, S. et al. (2009) Astro. J. 137, 4936 10.1088/0004-6256/ 137/6/4936. [11] Schmerr, N. C. et al. (2019) J. Geo. Res.-Plan. 124, 3063 10.1029/ 2019je006044. [12] Petersson, N. A. et al. (2010) LLNL Tech. Rep. LLNL-TR-422928. [13] Weber, R. C. et al. (2011) Sci. 331, 309 10.1126/science. 1199375. [14] Saito, T. et al. (2002) J. Geo. Res.-Solid Ear. 107, 10.1029/ 2001jb000264. [15] Watters, T. R. et al. (2024) Plan. Sci. J. 5, 10 10.3847 /PSJ/ad1332. [16] Cheng and Menard (2021) arXiv preprint arXiv:2112.01288. [17] Besserer, J. et al. (2014) Geo. Res. Let. 41, 5771.