Probing Mars' Northern Plains Stratigraphy with Impact Craters

46th Lunar and Planetary Science Conference (2015) 2583.pdf

PROBING MARS’ NORTHERN PLAINS STRATIGRAPHY WITH IMPACT CRATERS. L. Pan1 and B. L. Ehlmann1,2, J. Carter3 , C. M. Ernst4. 1California Institute of Technology (1200 E California Blvd, MC 150-21, Pas- adena, CA, 91125. Email: [email protected]), 2Jet Propulsion Laboratory, 3IAS-Orsay, 4Johns Hopkins University Applied Physics Laboratory

Introduction: The northern plains of Mars pre- serve a geologic record from the Noachian to present, indicated by a heavily cratered Noachian basement [1] overlain by sediments [2] and lavas [3] from the Hes- perian and Amazonian. Using hyperspectral images returned from Compact Reconnaissance Imaging Spec- trometer for Mars (CRISM), we investigate the mineralogy of northern plains stratigraphy to decipher changing aqueous and igneous processes through time. While limited detections of phyllosilicates and hydrated silica are found in Acidalia [4] and Utopia Planitia [5-6] associated with knobs and cones, numerous impact craters within the northern plains have exposed mafic [7] and hydrated [8] mineral signatures in infra- red spectra. Here we report the ongoing survey of CRISM targets on impact craters in the northern plains of Mars. We detect and map the locations of mineral Figure 1: Summary of mineral detections in large craters in phases and use depth-diameter relationships to under- Acidalia and Chryse Planitia. Stars refer to craters >30 km while stand the relationship between mafic and hydrated circles indicate smaller craters. The mineral detection symbols minerals thereby establishing the stratigraphy and geo- for each crater include multiple images around that crater. logic history of the northern plains of Mars. the most diverse spectral features in Acidalia. While Method: We used the Mars global crater database some of the strong olivine signatures on the crater floor [9] to filter craters that are relatively fresh (with degra- are associated with dark sand dunes of mafic composi- dation state greater or equal to 2) into different size tion, other exposures of olivine and hydrated minerals bins and start with craters larger than 30 km in diame- are confined to the central peak region and crater wall ter for the CRISM survey. CRISM image cubes were exposures (Figure 2-A1). converted to CRISM I/F and went through standard At least two different phyllosilicates have been processing as in [10]. As the northern plains dataset identified in Kunowsky (Figure 2-B1): Fe/Mg smectite typically has higher background noise and more dust with 1.4-, 1.9-, and 2.3-µm absorptions and chlo- cover, for particular targets we modified and applied a rite/prehnite phases with 1.4-, 1.9- and 2.35-µm ab- CRISM noise reduction method [11] to produce ratioed sorptions. Other phases include a subtle hydrated fea- cubes and spectral parameter maps for better identifi- ture with a hint of 1.9-µm absorption superposed on an cation of mineralogy. olivine-like slope. These minerals all occur in patches Results: To date, we have surveyed CRISM target- near the central peak. ed images acquired in Acidalia Planitia as of May 30, Davies Crater is smaller in size, with double-layer 2014, with a focus on large craters that likely expose ejecta typical of the “rampart craters” in the northern hydrated minerals. plains. CRISM images of Davies crater typically ex- Summary of crater detections. 61 CRISM targeted hibit weaker absorption features (Figure 2B), and these images, which cover all the craters >30km in the study targets are processed through the CRISM noise reduc- region and a number of smaller candidates, have been tion code as test images. After ratioing and continuum closely studied to detect and map minerals (Fig. 1). 44 removal steps, some 1.9-µm absorptions due to hydra- out of 61 images yield mafic mineral detections (oli- tion can be found. Here, the 3 targets on the central vine or pyroxene) and 25 out of 61 show hydrated peak all show Fe/Mg smectite, olivine, pyroxene and minerals, with a few confirmed Fe/Mg phyllosilicate some other subtle H2O/OH absorptions (Figure 2). One and chlorite/prehnite detections. CRISM target on the first layer of ejecta exhibits a 1.9- Representative observations. Kunowsky Crater is a µm H2O band and 2.28 metal-OH absorption (Figure 62-km-diameter crater in the northern part of Acidalia 2B). One of 3 targets on the crater’s second layer of Planitia (9.3° W, 56.8° N). It is the largest crater with 46th Lunar and Planetary Science Conference (2015) 2583.pdf

other types of transient aqueous events. However, the Fe/Mg phyllosilicates on the ejecta of Davies crater and the cor- responding detections on crater walls are consistent with the expected occur- rence of excavated stratigraphy in impact crater models [16]. In the initial survey, the largest craters have hydrated mineral-bearing units that are not found in smaller craters, which is consistent with the scenario that phyllosilicate-bearing unit is buried deeper than the mafic units. Similar-sized large craters near outflow channels had weaker detections or no detections of mafic or hydrated minerals, in contrast to the large craters in the northern part of the plain. This could be a result of thicker man- tling material atop the phylloslicate and mafic units due to the proximity to the outflow channels. Alternatively, mass-wasted sediments from the high- lands may have subdued the spectral signals by mixing, making minerals of Figure 2. Representative mineral detections from Kunowsky and Davies craters. A1-2) interest harder to detect. More CRISM CRISM footprints with detections; B1) Fe/Mg phyllosilicates and possible chlorite spectra targets in southern Acidalia will be in Kunowsky and Davies crater compared with laboratory spectra; B2) Olivine /pyroxene explored to understand the latitudinal spectra in Kunowsky and Davies crater compared with laboratory spectra. trends. Future work: The application of ejecta also yielded detections of 1.9-µm H2O absorption bands. The central peak detection of Fe/Mg phyl- CRISM noise reduction code proved to be successful losilicates and the ones found on the ejecta share the in enhancing weak mineral signals and will be applied 1.9-µm absorption, but it’s rather ambiguous whether to a larger dataset in Acidalia Planitia and across the the absorption center of the ~2.3 µm feature agrees. northern plains to complete the mineralogy survey. For Impact crater scaling: After careful survey of craters with diverse mineralogy and good exposures, CRISM maps, scaling relationships were applied to we will use the Maxwell-Z model to provide upper and estimate the maximum excavation depth of each crater lower limits to the thickness of different layers. Rela- using different crater models [12-15]. Though there are tionships between crater sizes and latitude will be ana- slight differences, these models predict excavation of lyzed to establish the temporal-spatial variations of materials from depths >2 km for large craters (diame- northern plains geology. ter >30 km) where hydrated minerals are detected, Acknowledgements: This research is supported by the NASA Mars Data Analysis program award NNX12AJ43G. indicating this as the burial depth of the crust contain- References: [1] Frey et al., (2002) GRL 29(10) [2] ing hydrated mineral phases. The small craters (e.g. Tanaka et al., (2003) JGR 108(E4), 8043 [3] Head et al., Yuti) that only show strong olivine detections suggest (2010) JGR 107(E1), 5003 [4] Pan and Ehlmann (2014) GRL. that the mafic rocks are buried to much shallower 41, 1890–1898 [5] Carter et al., (2013) JGR, 118, 831–858; depths < ~1.2 km. [6] Ehlmann and Edwards (2014) Annu. Rev. Earth Planet. Discussion: The preferential detection of hydrated Sci. 42:291–315. [7] Salvatore et al., (2010) JGR, 1344–1345. [8] Carter et al. (2010) Science 32, A74. [9] Robbins and minerals near the central peaks of the large craters (e.g. Hynek (2012). JGR, 117, E05004. [10] Ehlmann et al., Kunowsky, Davies and Bamberg) is consistent with (2009) JGR, 114, E00D08. [11] Carter et al,. (2012) PSS, 76, excavation of the hydrated minerals from depth. An 53-67 [12] Melosh (1989) Oxford University Press. [13] alternative scenario is that the hydrated minerals may Croft (1985) JGR, 90, C828–C842. [14] Holsapple (1993). have formed via hydrothermal alteration after the im- Annu. Rev. Earth Planet. Sci. 21, 333–373. [15] Pan et al., pact, due to an impact-induced hydrothermal system or (2014) Mars 8, #1353 [16] Ernst et al., (2010) Icarus 209, 210-223.