Since landing in Gale crater, the Mars Science Laboratory rover Curiosity has traversed fluvial, lacustrine, and eolian sedimentary rocks that were deposited within the crater ∼3.6 to 3.2 b.y. ago. Here we describe structures interpreted to be pipes formed by vertical movement of fluidized sediment. Like many pipes on Earth, those in Gale crater are more resistant to erosion than the host rock; they form near other pipes, dikes, or deformed sediment; and some contain internal concentric or eccentric layering. These structures provide new evidence of the importance of subsurface aqueous processes in shaping the near-surface geology of Mars.
The Mars Science Laboratory rover Curiosity landed in Gale crater in August 2012, and since then has traversed a sequence of sedimentary rocks deposited within the crater ∼3.6 to 3.2 b.y. ago (Grotzinger et al., 2014, 2015). During this traverse (Fig. 1), the rover has photographed structures that are interpreted as horizontal sections through relatively vertical cylinders or pipes (Figs. 2 and 3; Table 1; additional information about the images is provided in the GSA Data Repository1). These features are too small to detect in orbital images (maximum diameter is 60–70 cm), and they have not been seen in previous landed missions. The objectives of this paper are to describe these structures and to interpret their origin by analogy with similar structures on Earth.
Rocks along the rover traverse are locally crosscut by veins filled with calcium sulfate and occasionally with other unidentified minerals (Nachon et al., 2014). The veins occur throughout the stratigraphic succession and can be particularly prominent in fine-grained facies, where they resemble fracture networks seen in overpressured and hydraulically fractured mudstones on Earth (Schieber, 2014). Brittle fracture of the mudstone host as well as subhorizontal and horizontal fractures suggest a late diagenetic origin, when the associated fluids exceeded lithostatic pressure.
Previous Work on Fluid Upwelling and Cylindrical Structures on Mars
Orbital studies have shown that subsurface fluid flow has been common on Mars. Chan et al. (2010) described a field of knobs in Candor Chasma and considered a variety of origins, including injected sediment pipes. Other mounds, knobs, and ridges have been attributed to movement of subsurface fluids or fluidized sediment (Allen and Oehler, 2008; Oehler and Allen, 2010; Franchi et al., 2014; Siebach and Grotzinger, 2014; Okubo, 2016). These interpretations were based on satellite imagery of geomorphic features 1–3 orders of magnitude larger than the sedimentary features observed on the ground.
Previous rover investigations also suggest that subsurface fluid flow was common on Mars. The rover Opportunity identified post-impact subsurface fluid circulation at Meridiani (Squyres et al., 2012; Arvidson et al., 2014), and the rover Spirit found evidence of ancient groundwater activity in Gusev crater (Arvidson et al., 2008).
Mahaney et al. (2010) suggested that sediment pipes on Mars might be favorable targets for astrobiological exploration because the sediment pipes they studied in Utah had clay mineral coatings on some sand grains that contained fossilized fungi, whereas the host rock did not. The potential astrobiological importance of sedimentary pipes on Mars was also discussed by Dohm et al. (2011), Essefi et al. (2014), and Fairén et al. (2016).
Previous Work on Sedimentary Pipes on Earth
Vertical and subvertical, relatively cylindrical, bodies of sediment on Earth have been attributed to a variety of processes (reviewed by Lowe, 1975; Hunter et al., 1992; Netoff, 2002; Chan et al., 2007; Hurst et al., 2011; Wheatley et al., 2016). The origins can be categorized as (1) upward or downward movement of fluids that formed vertical cavities that were later filled with sediment (Allen, 1961); (2) upward injection of fluidized sediment (Chan et al., 2007; Wheatley et al., 2016); and (3) downward movement of fluidized or brecciated sediment or intact strata (Hunter et al., 1992). In addition to the relatively vertical cylinders or columns that are the focus of this paper, fluidized sediment injectites have a wide variety of geometries including sills, dikes, and irregular bodies (Hurst et al., 2011).
STRUCTURES IN GALE CRATER
The structure at Yellowknife Bay (Fig. 2) has an outer diameter of 7 cm and a rim that is raised above the local bedrock by ∼2–3 cm. The vertical relief is a substantial proportion of the diameter, suggesting that the structure is a cylinder rather than a circular structure confined to a horizontal plane.
Layering within the rim is roughly concentric, with some regions appearing crinkly or irregular. Grain size of the sediment in the structure is generally too fine to resolve in the Mastcam M100 images (each pixel is 0.23 mm wide in Fig. 2B), although the rim contains clasts, mottling, or concretions that are up to a few millimeters in diameter. Adjacent to the cylinder is a more irregular structure with lithology that appears similar to the cylinder rim. The cylinder occurs 2 m from a feature named the “snake” by Grotzinger et al. (2014) that was interpreted as the trace of a clastic dike protruding above the outcrop (Fig. 2A). The cylinder occurs within a few centimeters of the contact between the Sheepbed member and the underlying Gillespie Lake member of the Yellowknife Bay formation (Fig. 2A). Presumably the pipe extends downward into the lower unit, although this cannot be observed in the relatively flat exposure.
The Sheepbed mudstone contains a significant fraction of smectite (22% for the John Klein drill hole and 18% for Cumberland) that was inferred to be diagenetic, along with detrital basaltic minerals and ∼30% amorphous material (Vaniman et al., 2014). Although it was not sampled for mineralogy, the overlying Gillespie Lake sandstone has a similar elemental chemistry and hydration level, suggesting the presence of phyllosilicates, likely as matrix or cement.
Several cylinders and a deformational structure occur near Dingo Gap, where the rover entered a small canyon (Figs. 3A–3E). The largest cylinder has a diameter of 60–70 cm, a raised rim, and a shallow bowl-shaped interior (Fig. 3B). The rim has a lumpy texture due to some larger rounded grains to 3 mm in diameter (Fig. 3C). Sedimentary rocks at this vicinity include conglomerates and cross-stratified sandstones interpreted as fluvial and eolian deposits, possibly including clinoform beds like those 500 m to the southwest at the Kylie outcrop (Grotzinger et al., 2015). A loose rock on the floor of Dingo Gap contains a soft-sediment deformational structure (Fig. 3D).
The elemental chemistry (determined by ChemCam; Wiens et al., 2012; Maurice et al., 2012) of the walls of this cylinder are typical of surrounding sediments with the exception of magnesium, which was elevated (10–14 wt% MgO), and the hydration level, for which the five observation points were consistently among the highest observed by that instrument. Because ChemCam observations do not distinguish compositions of grains and cement, we do not know if these two observed differences in chemistry are due to different grains within the cylinders, or different post-injection cementation.
Curiosity also photographed a cylinder ∼150 m to the west of Dingo Gap (Figs. 3A and 3E); the cylinder is ∼20 cm in diameter, and is noteworthy because it contains internal concentric or eccentric layering.
A pair of structures was observed higher in the Murray formation near Marias Pass (Fig. 3F); they have relatively resistant rims, circular cross sections, and light toned interiors relative to the surrounding sediment, and occur within lacustrine mudstones (Grotzinger et al., 2015). Elemental chemistry appears to parallel the albedo change from interior to walls. The four ChemCam observations that sampled the interior are high in Si (68–72 wt% SiO2) and Ti, and low in Mg and Fe (<5 wt% total Fe as FeO), and show relatively strong hydration. We only have two measurements of the darker wall material, but the mean is lower in silicon (54–66 wt% SiO2) and higher in iron (6–17 wt% total Fe as FeO); magnesium is consistently low in both materials (≤2.1 wt% MgO), while the alkalies are consistently high (3.3–5.6 wt% Na2O, 1.8–3.2 wt% K2O). These compositions reflect those of nearby rocks, which contain silica enrichments apparently due to alteration (Newsom et al., 2016; Frydenvang et al., 2016).
Interpretation of Pipes in Gale Crater
Here we suggest that the cylinders in Gale crater are best interpreted as pipes of injected sediment rather than infilling of vertical cavities. We base this interpretation on two observations. First, where layering is observed in the pipes (Figs. 2B and 3E) it is relatively vertical and lacks horizontal fill. Second, several of the pipes occur near other rare structures that are interpreted to have formed by movement sediment; the cylinder at Yellowknife Bay occurs only 2 m laterally and within a few centimeters vertically of the “snake,” a sand dike (Fig. 2A), and the cylinders at Dingo Gap occur near a rock containing deformed sediment (Fig. 3D). Deformational structures are commonly associated with injectites (Hurst et al., 2011), and Wheatley et al. (2016) used an association with deformed sediment to argue that cylinders on the Colorado Plateau formed by sediment injection.
Although the structures observed at Marias Pass occur as a pair (Fig. 3F), they do not occur near other examples of sediment flow, and the case for their origin is weaker. They are included here to document their occurrence and because their circular planforms and raised rims are consistent with injected pipes.
The geologic characteristics and abundance of pipes on the Colorado Plateau (central USA) make them a good analog for those we describe from Mars. Hurst et al. (2011) compiled a literature review of dozens of occurrences of injectites having diverse geometries; of the five reported occurrences they categorized as vertical cylinders (which they term columnar intrusions) four are from Jurassic deposits of the Colorado Plateau, and one is from South Australia (Mount, 1993). Pipes are also associated with other geometries of injectites at six other locations, including a large complex near Santa Cruz, California (Thompson et al., 2007; Sherry et al., 2012), although these pipes are not necessarily vertical.
Wheatley et al. (2016, their table 1) list 26 published accounts of Colorado Plateau pipes, describing 30 locations (their figure 1), some of which contain multiple swarms, which in turn contain as many as hundreds of pipes; they stated (p. 12), “The Colorado Plateau is known for its abundance of clastic pipes.” Those pipes (Fig. 4), like those we describe from Mars, commonly have preferentially cemented rims, concentric or eccentric layering, association with dikes (Hannum, 1980) and/or soft-sediment deformation, and clustering (although the Mars examples are not as clustered as those of Wheatley et al., 2016). The reported range of host lithologies and facies of the Colorado Plateau pipes (sandstones, shales, siltstones, and limestones deposited in eolian, interdune, fluvial, sabkha, and marine environments) is broad enough to encompass the rocks that contain the pipes in Gale crater (if the marine deposits of the Colorado Plateau are comparable to the lacustrine rocks on Mars). Much of the work on sediment pipes has focused on behavior of the noncohesive flowing sediment, but plugging of the walls with sediment as fine as clay helps to prevent leakage from the pipes; this confines and maintains the flows (Lowe, 1975; Mount, 1993). The finer sediment in more poorly sorted deposits may aid with this plugging. Although we have less information than terrestrial studies of pipes, making it impossible to identify the single best analog, geologic characteristics of the Colorado Plateau pipes, as well as their abundance and publication history, make them an appropriate analog.
An important question regarding origin of sediment pipes on both Mars and Earth is whether sediment movement was dominantly upward or downward. Determining the direction of sediment movement is surprisingly difficult even in the excellent exposures on the Colorado Plateau; determining flow direction from the limited images of the Mars pipes is not possible. Observed downward movement might occur either by net flow into a cavity that was dissolved in evaporites, as described by Hunter et al. (1992) (sulfates are common along the rover traverse), or by late-stage collapse following dominantly upward movement (Chan et al., 2007). Some observations are ambiguous even in good exposures on Earth. For example, upturned beds in the host rock adjacent to a pipe can be attributed to drag by upward flow within the pipe (Wheatley et al., 2016) or to measurable down-dropping of the host rock relative to a pipe supported at its base (Hunter et al., 1992, their figure 12). Interpretations may also be complicated if upward and downward movement of sediment occur simultaneously or where high-permeability pipes focus post-emplacement fluid flow (Martínez-Frías et al., 2007; Sherry et al., 2012; Wheatley et al., 2016).
Hurst et al. (2011) noted that fluidized sediment pipes are merely one kind of sand injectite on Earth, and they can be associated with larger injectite bodies. We have not observed such larger bodies on Mars, but we cannot rule out this possibility.
Mangold et al. (2015) used geochemical data to suggest that during diagenesis of the Sheepbed mudstone, fluids were expelled into the overlying sandstone due to overpressure. Such overpressurization would aid formation of pipes, but, as is the case for many pipes on Earth, we don’t know which processes caused fluidization (earthquakes, impacts, collapse following dissolution of evaporites or carbonates, hydrothermal processes, and/or overpressurization due to burial and compaction).
The rover Curiosity photographed cylinders that resemble fluidized sediment pipes on Earth. These cylinders are elevated above the outcrop surfaces, have preferentially cemented rims and concentric or eccentric internal layering, and are associated with other fluidized sediment features (a dike and deformational structures). They are interpreted to have formed by injection of fluidized sediment, likely triggered by impacts, overpressurized fluids, earthquakes, or by collapse into cavities formed by dissolution of evaporites. These structures are yet another indicator of the importance of aqueous processes in shaping the geology of Mars.
This work could not have been completed without the NASA Mars Science Laboratory (MSL) engineering, management, and operations teams, supported by the NASA Mars Exploration Program. Support for Rubin, Goetz, and Oehler was provided by the NASA MSL Participating Scientist Program. Fairén was supported by the Project icyMARS, European Research Council Starting grant 307496. We thank Margie Chan, Andrew Hurst, David Loope, Massimo Moretti, Jeff Peakall, and James Schmitt for constructive reviews.
- Received 8 July 2016.
- Revision received 3 October 2016.
- Accepted 4 October 2016.
- © 2016 Geological Society of America