LETTERS Ancient groundwater flow in the Valles Marineris on Mars inferred from fault trace ridges
ALLAN H. TREIMAN Lunar and Planetary Institute, 3600 Bay Area Boulevard, Houston, Texas 77058-1113, USA e-mail: [email protected]
Published online: XX Month XXXX; doi:10.1038/ngeoXXXX
1,2 1 Groundwater is a crucial reservoir in Mars’ global water cycle graben-bounding faults, and traces of some of these faults can be 45 3–6 2 and plays an important role in aqueous alteration of bedrock . followed eastward as FTR atop and along the hogback ridge (Fig. 1). 46 3 Understanding groundwater flow is also important for assessing Other similar ridges parallel those that can be followed back to 47 7 4 the possibility of past and present martian life . The Valles plateau fault traces, and are inferred also to be FTR. In some areas, 48 5 Marineris is a series of fault-bounded troughs and chasms FTR are paired (e.g., east of centre on Fig. 2b), like an individual 49 6 stretching over 4,000 km from west to east, with a width of 600 km graben on the plateau. Elsewhere, there are multiple parallel and 50 8 7 and depth of up to 10 km below the surrounding high plains . crosscutting FTR, similar to the complex of faults at the east end 51 8 In this region, ancient groundwater movement is suggested by of the plateau. Not all fault traces on the plateau surface continue 52 9 links between chasmata—deep gashes on the surface of Mars— as FTR on the VM walls (Fig. 1b); there is no obvious difference 53 9 10 and the sources of catastrophic floods [AUTHOR: what are between graben-bounding faults that continue as ridges and those 54 11 these sources?]. Here, I use images obtained by the Viking, that do not. Similarly, individual FTR can be longer than 40 km or 55 12 Mars Express, Mars Reconnaissance Orbiter and Mars Odyssey as short as a few km (Fig. 1). 56 13 missions [AUTHOR: Ok?] to map ridges on the walls of the A crucial observation is that FTR extend over a significant 57 14 Valles Marineris that extend in an east–west direction up depth range, from kilometres deep in the Chasma to essentially 58 15 to the surface of the surrounding high plains from almost the plateau surface. In several places FTR reach elevations 59 10 16 7 km beneath [AUTHOR: Ok?]. These erosion-resistant ridges, above those of graben floors on nearby plateaus (Figs 1c, 3b; 60 17 which can be tens of kilometres long, most likely represent Supplementary Information, Fig. S1). Another observation is that 61 18 cementation of the fault zones and surrounding rock by water- many FTR appear to have lighter tones (i.e., higher albedos) than 62 11,12 19 deposited minerals and suggest that groundwater in this the surrounding rock or dust deposits (Fig. 1c; Supplementary 63 20 region flowed for long distances through major east–west- Information, Figs S1–S3). 64 21 trending fault systems. This interpretation implies that liquid West Candor Chasma and Ophir Chasma are separated by an 65 22 water was stable at (or near) Mars’ surface when the fault E–W trending plateau, Candor Labes, and a hogback or ridge that 66 23 zones were cemented ∼3,500–1,800 Myr ago and that chemical extends it down into the VM (Figs 2,3). Grabens cut the Candor 67 24 deposition from groundwater was regionally significant. Labes surface, and their faults’ traces can be followed along the 68 25 The Valles Marineris (VM) is a series of troughs and chasms hogback ridge as a pair (or complex) of sub-kilometre-wide ridges 69 26 stretching over 4,000 km E–W and 600 km N–S, with depths up (FTR). Where the FTR first appear at the edge of the plateau, their 70 8 27 to 10 km below the surrounding high plains . The walls of the crests are at the elevation of the plateau surface (Fig. 3b). The FTR 71 28 VM thus can present deep exposures into Mars’ subsurface. Long can be traced ∼70 km east and down into the VM, where they 72 29 ridges on the walls of the VM follow the traces of faults that bound are covered by VM interior layered deposits, and farther east as 73 30 grabens (Hesperian age, Set no. 3 of refs 13–15) on the high plains scattered protruding through the interior layered deposits. These 74 31 adjacent to the VM, and are here denoted ‘fault trace ridges’, FTR FTR span the whole elevation range of Candor/Ophir Chasmata, 75 10 32 ( = ‘fault-continuation ridges’ ). The grabens associated with FTR from the plateau surface at +3.25 km to ∼−4 km in the depths 76 33 are typical of others in the area, with depths of a few hundred of Ophir Chasma (Fig. 1, elevations from refs 14,15). In several 77 34 metres, widths of a few kilometres, ‘dog-leg’ offsets (trending places, FTR segments and their down-slope talus may be lighter- 78 35 NE–SW, Fig. 1), and multiple fault strands that weave across the toned than the surrounding rock debris and dust (Fig. 1b). 79 36 grabens. These grabens have been ascribed to extension over a Fault trace ridges also occur on lesser faults across the 80 16 17 37 shallow ductile layer , tension fractures over large dikes , and to whole Valles Marineris, from Noctis Labyrinthus in the west (see 81 18 38 complex structures of deep-seated extension . FTR share many Supplementary Information, Figs S4,S5) to Melas Chasma in the 82 39 geometric features with graben-forming faults, such as occurrence centre (see Supplementary Information, Fig S3) to eastern Coprates 83 40 in pairs and presence of en echelon and ‘dog-leg’ offsets, and Chasma (see Supplementary Information, Fig S6). 84 41 multiple fault strands (Fig. 1). Ridges that are morphologically identical to FTR, but not 85 42 The boundary between West Candor and Melas Chasmata clearly connected with graben fault systems, occur throughout 86 43 (Figs 1,2) is an E–W trending plateau that continues eastward as the VM. The most prominent of these ridge systems supports 87 44 a high ridge or hogback. The plateau is cut by a complex of the central hogback in Ius Chasma, Geyron Montes (Figs 2,4). 88
nature geoscience ADVANCE ONLINE PUBLICATION www.nature.com/naturegeoscience 1 LETTERS
a b c Candor Chasma It
d d FTR
Melas Chasma d 20 km 5 km
Figure 1 Melas & Candor Chasmata. Grabens and fault trace ridges (FTR) between Melas and Candor Chasmata, Valles Marineris. a, HRSC visual image28 north to top, centred: 7.90◦ S, 285.61◦ E. Grabens extend eastward from plateau as FTR (arrows). b, Structural diagram of a; faults and FTR as lines, plateau in gray. ‘d’s denote some dog-leg bends. c, Detail from a, up and left of centre, showing fault traces and FTR near plateau edge. FTR crests have elevations near that of the plateau, distinctly above those of graben floors. Some FTR show lighter tones (lt), than the nearby dusty plateau surface. From THEMIS infrared daytime image I05109001.
3 4 1
20 km Figure 2 Large fault trace ridges. Locations of longest FTR and similar paired ridges in the Valles Marineris, shown as green line overlays on colourized Viking mosaic. Locations of areas in Figs 1,3, and 4 are shown by their respective numbers. North to top. Figure 4 Geyron Montes. The central hogback in Ius Chasma, Geyron Montes, comprises parallel ridges (small arrows) arranged like FTR. Mosaic of THEMIS daytime infrared images, centre at 7.86◦ S, −81.46◦ E. North to top. North and a Ophir Chasma b Candor Labes Ophir Chasma south walls of Ius are at top and bottom of image. Composed from viewer on Candor Labes THEMIS website (http://themis.asu.edu). Inset shows a dog-leg bend in the ridges, from THEMIS visible image V11063002.
Candor Chasma and rock adjacent to fault zones is fractured—both erode more 9 10 km easily than coherent rock. For eroded fault zones to appear as 10 Candor Chasma 5 km ridges like FTR, their gouge and/or fractured aureole must be 11 11,12,19 hardened . The most reasonable mechanism for hardening 12 here is cementation by groundwater—deposition of resistant 13 Figure 3 Ophir & W. Candor Chasmata. Paired fault trace ridges (FTR, arrows) minerals (e.g., Fe oxides, silica, sulphates, carbonates) in the fault 14 between Ophir and W. Candor Chasmata. Graben on plateau of Candor Labes to 19 zone and adjacent fractured rock —which mechanism has been 15 left—their bounding faults connect to FTR along the edges of the Chasmata. North ◦ ◦ invoked to explain similar ridges that cut VM interior layered 16 to top. a, Overview, Viking image 915A11, centre ∼5.0 S, 286.7 E. b, Detail of 11,12 deposits . Other hardening mechanisms, like igneous intrusion 17 transition from graben-bounding faults to FTR in scene a. Northern FTR starts at and fault heating, are unlikely here. Basaltic igneous rock can 18 plateau level; southern FTR has light tone (high albedo?). HRSC image28, centre at ◦ ◦ be emplaced as dikes along faults or tension fractures, and resist 19 4.86 S, 286.24 E. erosion better than surrounding rock. This mechanism seems 20 unlikely for FTR, because: (1) it would require that all the dikes 21 extend to the plateau surface, and thus all represent only the last 22 1 The crest of the hogback is marked by a pair of ridges with emplacement of lava in the area; and (2) it would be inconsistent 23 2 an intervening valley, a few hundred metres lower than the with the light tone (high albedo) of many of the FTR. Mylonite or 24 3 ridge crests. This landform is essentially identical to that of the pseudotachylite in fault zones can be stronger than adjacent rock, 25 4 paired FTR at Candor Labes (Fig. 3). Similar ridges pairs support but both form under significant heat from high strain rates under 26 5 hogbacks throughout the Valles Marineris (see Supplementary pressure, which seem unlikely in the tensional faults of VM graben 27 6 Information, Figs S3,S5–S7). (especially near the plains surface). 28 7 Eroded fault zones are rarely expressed as ridges because For the FTR, cementation by flowing groundwater seems 29 8 material in them is broken and pulverized (i.e., is fault gouge) more likely than by hydrothermal fluids. First, FTR fault zones 30
2 nature geoscience ADVANCE ONLINE PUBLICATION www.nature.com/naturegeoscience LETTERS
1 were hardened both above and below the current graben floors observation indicates that graben-forming, fluid conduit faults 61 26 2 (Figs 1,3), which implies cementation early in the faults’ history and extend to great depth, and possibly are ‘tension fractures’ or 62 18 3 continuing throughout their movement—otherwise, fault zones complex graben structures . Further, abundant water in a fault 63 4 could not have been cemented above their graben floors and FTR system may significantly reduce its effective friction, and thus the 64 27 5 would not extend up to the plateau surface. Fault motion would stresses needed for faulting . 65 6 tend to break earlier-cemented material, so cementation would Finally, FTR show that groundwater interacts chemically with 66 7 have continues throughout the interval of fault movement. Second, the crust it traverses, here affecting mechanical properties of crustal 67 8 it is not clear that hydrothermal systems could operate over the rock. Chemical and physical interactions between martian rocks 68 6 9 whole active lifetimes of these graben-forming faults, nor over their and groundwater are known from martian meteorites and the 69 4,5 10 lengths of >50 km. MER rover sites , and cannot be ignored in models of Mars’ global 70 11 The cements in the FTR are unknown. Many FTR are water cycle. 71 12 indistinguishable in colour and tone from their surrounding rocks 13 and soil, and so may either be cemented by dark material (e.g., Received 23 October 2007; accepted 28 January 2008; published XX Month XXXX. 72 14 Fe oxides) or mantled by surrounding material or dark duricrusts. 15 Crests of some FTR segments and their down-slope talus appear References 73 1. Clifford, S. M. A model for the hydrologic and climatic behavior of water on Mars. J. Geophys. Res. 74 16 lighter-toned than their surroundings (Fig. 3b, THEMIS daily 98, 10973–11016 (1993). 75 17 image 20071018a), suggesting that FTR material is bright. However, 2. Carr, M. H. & Head, J. W. III. Oceans on Mars: An assessment of the observational evidence and 76 18 the apparent lighter tone could represent other geological processes possible fate. J. Geophys. Res. 108, 5042 (2003). 77 11,20 3. Bibring, J. P. et al. Global mineralogical and aqueous Mars history derived from OMEGA/Mars 78 19 (e.g., bleached rock surrounding the FTR, or exposures of later express data. Science 312, 400–404 (2006). 79 20 bright mantling material as in HiRISE image PSP 002551 1700), 4. Squyres, S. W. & Knoll, A. H. Sedimentary geology at Meridiani Planum: Origin, diagenesis, and 80 implications for life on Mars, Mars. Earth Planet. Sci. Lett. 240, 1–10 (2005). 81 21 or photometric effects such as specular-like reflections from slabby 5. Ming, D. W. et al. Geochemical and mineralogical indicators for aqueous processes in the Columbia 82 22 rock or hard-pan surfaces. If the light tones associated with FTR Hills of Gusev crater, Mars. J. Geophys. Res. 111, E02S12 (2006). 83 21 6. Bridges, J. C. et al. in Chronology and Evolution of Mars (eds Kallenback, R., Geiss, J. & 84 23 do reflect mineral cements, they could be of Mg- or Ca-sulphates , Hartmann, W.) 365–392 (ISSI Space Science Series, Vol. 96, Kluwer, Amsterdam, 2001). 85 6 22 24 Mg- and/or Ca-carbonates , silica , or some clays. Of these, silica 7. Beaty, D. et al. Findings of the Mars special regions science analysis group. Astrobiology 6, 86 677–732 (2006). 87 25 is most resistant to erosion, and thus most reasonable. 8. Luchitta, B. K. et al. Mars 453–492 (Univ. Arizona Press, Tucson, 1982). 88 26 The time of FTR formation (i.e., the time at which graben- 9. Palmero-Rodriguez, J. A., Sasaki, S. & Miyamoto, H. Nature and hydrological relevance of the 89 90 27 forming faults became cemented) was likely in the early Hesperian Shalbatana complex underground cavernous system. Geophys. Res. Lett. 30, 1304 (2003). 10. Treiman, A. H. & Spiker, K. Fault-continuation ridges in the Valles Marineris, Mars: Evidence for 91 28 (3,500–1,800 Myr ago). Cementation is inferred (above) to be groundwater circulation. Lunar Planet. Sci. XXVII, 1339–1340 (1996). 92 29 contemporaneous with movement of these graben-forming faults, 11. Okubo, C. H. & McEwen, A. Fracture-controlled paleo-fluid flow in Candor Chasma, Mars. Science 93 315, 983–985 (2007). 94 30 which are of ‘Stage 3’ (radial from a centre in Noctis Labyrinthus) 12. Davatzes, A. & Gulick, V. Evidence for tectonically controlled hydrothermal fluid flow in relay zones 95 13 31 and inferred to be early Hesperian . Cementation of the faults on Mars from early HiRISE images. Lunar Planet. Sci. XXXVIII, Abstract no. 1788 (2007). 96 13. Anderson, R. C. et al. Primary centers and secondary concentrations of tectonic activity through time 97 32 must have pre-dated the excavation of the Valles Marineris, which in the western hemisphere of Mars. J. Geophys. Res. 106, 20563–20586 (2001). 98 8 33 is but generally considered to be Hesperian . The FTR pre-date 14. US Geological Survey Topographic Map of the Ophir and Central Candor Chasmata Region of Mars 99 (MTM 500k–5/287E OMKT). USGS Geologic Investigations Series I-2807 (2003). 100 34 deposition of the VM interior layered deposits, which are inferred 15. US Geological Survey Topographic Map of the West Candor Chasma Region of Mars (MTM 500k 101 8 35 to be mostly late Hesperian to early Amazonian . –05/282E OMKT). USGS Geologic Investigations Series I-2805 (2004). 102 16. Wilson, L. & Head, J. W. III. Tharsis-radial graben systems as the surface manifestation of 103 36 The presence and structures of FTR have important plume-related dike intrusion complexes. J. Geophys. Res. 107, 5057 (2002). 104 37 implications for Mars: the presence and locations of liquid water; 17. Davis, P. & Golombek, M. Discontinuities in the shallow Martian crust at Lunae, Syria, and Sinai 105 38 the importance of fracture permeability; and the importance of Plana. J. Geophys. Res. 95, 14231–14248 (1990). 106 18. Schultz, R. A., Moore, J. M., Grofils, E. B., Tanaka, K. L. & Mege,` D. The Geology of Mars: Evidence 107 39 chemical interactions between groundwater and crust. Although from Earth-based Analogs 371–399 (Cambridge Univ. Press, New York, 2007). 108 40 these major FTR apparently have helped define the shape of the 19. Mozley, P. & Goodwin, L. Patterns of cementation along a Cenozoic normal fault: A record of 109 paleoflow orientations. Geology 23, 539–543 (1995). 110 41 Valles Marineris, they do not appear to be causally related to 20. Ormo,¨ J., Komatsu, G., Chan, M. A., Beitler, B. & Parry, W. T. Geological features indicative of 111 42 the Valles. processes related to the hematite formation in Meridiani Planum and Aram Chaos, Mars: A 112 comparison with diagenetic hematite deposits in southern Utah, USA. Icarus 171, 295–316 (2004). 113 43 The presence of liquid water is important in current ideas 21. Gendrin, A. et al. Sulfates in martian layered terrains: The OMEGA/Mars Express view. Science 307, 114 44 of Mars’ history, and central to Mars’ potential for life. The 1587–1591 (2005). 115 22. Squyres, S. W. & the Athena Science Team. Recent results from the Mars exploration rovers. Bull. 116 45 observation that FTR extend up to the local Hesperian-aged plateau Amer. Astron Soc., Division Planetary Sciences AAS Conf. abstract 14.01 (2007). 117 46 surface implies that cementing agent, groundwater, must have 23. Coleman, N. M. Large-scale permeability in the upper crust of Mars estimated from Earth analogs. 118 47 been present and active at (or near) the ancient plateau surface EOS Trans. AGU, 81, Fall Meet. Suppl., Abstract P62C-11 (2000). 119 24. Tanaka, K. L. & Chapman, M. G. The relation of catastrophic flooding of Mangala Valles, Mars, to 120 48 at that time; a similar inference of near-surface groundwater has faulting of Memnonia Fossae and Tharsis volcanism. J. Geophys. Res. 95, 14315–14323 (1990). 121 4 49 been made for the Meridiani Planum area . If groundwater were 25. Tanaka, K. L. & Chapman, M. G. Kasei Valles, Mars: Interpretation of canyon materials and flood 122 sources. Proc. Lunar Planet. Sci. 22, 73–83 (1992). 123 50 present at Mars’ ancient surface in the VM area, a thick permafrost 26. Tanaka, K. L. & Golombek, M. P. Martian tension fractures and the formation of graben and collapse 124 51 layer could not have been present, which suggests a ‘warm wet’ features at Valles Marineris. Proc. Lunar Planet. Sci. Conf. 19, 383–396 (1989). 125 27. Barnett, D. N. & Nimmo, F. Strength of faults on Mars from MOLA topography. Icarus 157, 126 52 climate when these (Hesperian-age) graben-forming faults initiated 34–42 (2002). 127 53 and grew. 28.
nature geoscience ADVANCE ONLINE PUBLICATION www.nature.com/naturegeoscience 3 Page 1 Query 1: Line no. 1 Author: Please note that the first paragraph has been edited according to style.
Page 3 Query 2: Line no. 87 Author: As per journal style, reference citations should be in numerical order. Therefore references are renumbered from here onwards. OK?