Phyllosillicates and Low Calcium Pyroxene-Rich Noachian Crust Exposures in the Walls of Valles Marineris, Mars

41st Lunar and Planetary Science Conference (2010) 1524.pdf

PHYLLOSILLICATES AND LOW CALCIUM PYROXENE-RICH NOACHIAN CRUST EXPOSURES IN THE WALLS OF VALLES MARINERIS, MARS. J. Flahaut1, H.Clenet1 , J. F. Mustard2, C. Quantin1 and P.Allemand1. 1Laboratoire des Sciences de la Terre, UMR CNRS 5570, Université Claude Bernard/Ecole Normale Supérieure de Lyon, 2 rue Raphaël Dubois, 696222 Villeurbanne Cedex, France ([email protected]). 2Department of Geological Sciences, Brown University, Providence, RI 02912.

Introduction: The relative importance of volcan- of the CRISM targeted observations used during this ism versus sedimentary deposition in the early history survey, surperimposed over THEMIS IR Day data. of Mars is still unclear. This history is recorded in the walls of Valles Marineris, which are as deep as 11 km, Hyperspectral and multispectral data from CRISM on and exhibit parts of the upper crust. Previous studies board of Mars Reconnaissance Orbiter were used to [1] show that the upper parts of the walls are likely determine mineralogical compositions. CRISM data made of layered basalts, related to Tharsis volcanism, were processed to remove the atmospheric and photo- in most of the chasmata. Beneath these basalts may be metric contributions, and noise-removal algorithms exposures of Noachian crust [2,3]. were applied, as in [5,6]. CRISM high resolution ob- The present study investigates the walls located in Co- servations (up to 18 m/pixel) allowed us to identify prates Chasma, using CRISM (The Compact Recon- mafic and hydrated minerals within the walls of Valles naissance Imaging Spectrometer for Mars) and Hi- Marineris. High resolution maps of all the minerals RISE (High Resolution Imaging Science Experiment) identified in the walls were generated and merged with data. Coprates Chasma is an approximately 1000 km high resolution images (HiRiSE and CTX) providing a long and 150 km wide linear canyon which extends precise geological context for the detections. from Melas Chasma to Capri Chasma, at the embou- chure of Valles Marineris. As it has not been filled with large interior layered deposits (ILDs), Coprates Morphological observations: Layering in the up- Chasma has been proposed to be a graben that may per walls of Valles Marineris has been well described have opened after that most of the canyons were al- using MOC data by [1,7,8]. Alternating bright and ready emplaced [4]. Therefore Coprates Chasma exhi- dark layers are found over the uppermost 400m of bits the most well-exposed cross-section of the upper- chasma walls. The extent of this layering seems to vary most material of the crust through its walls. from one canyon to another [1,7] implying that the stratigraphy of the walls includes some regional inputs. a In Coprates Chasma itself, we noticed an asymmetry between the northern and the southern walls. Southern walls are less stratified and are affected by more landslides than the northen walls. At lower elevations, layers are then mantled by a thick talus that obscures the wall structure and stratigraphy. Therefore, the stra- tigraphy is not clearly resolved in the middle and lower 200km sections. Rocks outcrop only along the spurs, and with b depth, the layered appearance of rocks transitions to boulders and talus. The very bottom of walls exhibit a dense, fractured bright bedrock, similar to a megabrec- cia. It is urregularly exposed along the walls, as it is not present (or seen) on all the CRISM observations acquired at the same elevations. When terraces inter- sect the walls, especially in the northern rim of Co- prates, the bedrock is underlain by dark stratified lay- ers, which have been exposed by erosion.

Figure 1: a- Localization of the study area (white box) at the east of Coprates Chasma, near the outlet of Identification of signatures: Twenty-six CRISM Valles Marineris, over a MOLA elevation map. b- observations were processed in the eastern part of Co- Close-up on the studied area, which is approximately prates Chasma. Due to the footprint size of the ac- 600 by 200 km wide. Green dots indicate the location quired data, only the half-resolution long hyperspectral data (HRL, 36 m/pixel) cover the full cross-section of 41st Lunar and Planetary Science Conference (2010) 1524.pdf

the wall. Full-resolution targeted images (FRT, 18 Conclusion: These results show that the lower m/pixel) only cover the upper or lower parts of the walls of Valles Marineris show compositions see else- walls. Hydrated minerals are identified with CRISM where in Noachian crust (enrichment in LCP and Fe- by investigating the overtones and combinations of Mg phyllosilicates). The uppermost Hesperian-aged fundamental vibrational absorption features in the 1.0- layers are depleted of any infrared mineral signatures. 2.6 μm interval [5,9]. Mafic minerals are studied with This could back a sedimentary origin [12] rather than a MGM methods over the L-detector (between 1 and 3 volcanic one [8]. Further investigations on the sur- μm) [10]. rounding plateaus [13] and in other chasmata of Valles CRISM observations on the upper part of the wall Marineris should allow us to learn more about the up- are depleted of any diagnostic absorption bands. We permost crustal stratigraphy. More exposures are to be only observed small variations in the reflectance conti- analyzed and should be presented at the conference. nuum slope depending on the angle of observation. Some weak olivine signatures may be present on the Aknowlegdments: The work was supported by the darker top layers. This lack of signatures could be due Région Rhône-Alpes, France. We are also really grate- either to a thick dust cover, or simply means that the ful to the Brown University for their warm coopera- material forming the wall does not contain any spec- tion. We especially want to thank Janette Wilson for trally remarkable minerals. her precious advice. Morphology and composition cannot be discerned in the middle of the walls as they are similarly obscured References: [1] Beyer R. A. and McEwen A. S. (2005), Icarus, 179, 1-23. [2] Roach L. H. et al. by dust. (2009), Icarus, in press. [3] Murchie S. L. et al. The mineralogy of the wall becomes really interest- (2009), JGR, in press. [4] Lucchitta et al. (1994), JGR, ing under -400 m in elevation. A first spectral type is 99, 3783-3798. [5] Mustard J. F. et al. (2008) Nature, identifed with absorptions near 1.4, 1.9, and 2.31µm. 305-309. [6] Flahaut J. et al. (2009) In prep. [7] Wil- Comparison with laboratory spectra are most consis- liams J.-P. et al. (2003), GRL, 30(12). [8] McEwen A. tent with Fe/Mg phyllosilicates [9]. Contrary to the Al- S. et al. (1999) Nature, 397, 584-586. [9] Ehlmann B. phyllosilicates, which have a diagnostic absorption et al. (2009) JGR, in press. [10] Skok J. R. et al. band at 2.20 µm, Fe-OH and Mg-OH bonds are re- (2009) JGR, in press. [11] Mustard J. F. et al. (2005) spectively responsible for vibration around 2.28-2.29 Science, 307, 1594-1597 ; [12] Malin M. C. et al. µm and 2.30-2.31 µm. The exact position of the bands (2000) Science, 290, 1927-1937. [13] LeDeit L. et al. could indicate the presence of a Mg-smectite such as (2009) AGU Fall meeting, Abstract #P13A-1266 . saponite. These detections are generally associated to morphological dark boulders, and sometimes to some dusty talus inherited from the above boulders.

Pyroxenes are also identified with a broad 1.9 µm and a smaller 1.06 µm absorption. The exact position and shapes of the bands argue more for a mixing of Low Calcium Pyroxene (LCP) and High Calcium Py- roxenes (HCP) [11], rather than a pure pyroxene phase. LCP seems dominant with a 50 to 80% ratio according to the CRISM Modified Gaussian Model [10]. They do not seem to be aligned at a constant ele- vation but appear somewhere between -1400 and - 2400m in the northern wall of Coprates Chasma and seem to be associated with the light-toned brecciated outcrops. In the northern wall, this massive bedrock is Figure 2: left: CRISM ratioed spectra for the underlain by thin dark layers, which are highlighted by FRT00009804 observation. Green: pyroxene, Red: terraces that crosscut the walls around -4000m in ele- olivine, Purple: phyllosilicates. Raw spectra were ra- vation. The dark basal layers seem to be enriched in tioed over the same dust spectrum acquired in the olivine. The floor of the canyon is covered by small same column, to remove the noise due to the instru- mounds, dust and dark sand dunes in some places. The ment. Right: spectra from the USGS spectral library plateau on the top of the cross-section shows flat dusty are plotted for comparison. Green: orthopyroxene, spectra. Red: forsterite, Purple: saponite.