Recognition of Sedimentary Rock Occurrences in Satellite and Aerial Images of Other Worlds—Insights from Mars

Abstract

Sedimentary rocks provide records of past surface and subsurface processes and environments. The first step in the study of the sedimentary rock record of another world is to learn to recognize their occurrences in images from instruments aboard orbiting, flyby, or aerial platforms. For two decades, Mars has been known to have sedimentary rocks; however, planet-wide identification is incomplete. Global coverage at 0.25–6 m/pixel, and observations from the Curiosity rover in Gale crater, expand the ability to recognize Martian sedimentary rocks. No longer limited to cases that are light-toned, lightly cratered, and stratified—or mimic original depositional setting (e.g., lithified deltas)—Martian sedimentary rocks include dark-toned examples, as well as rocks that are erosion-resistant enough to retain small craters as well as do lava flows. Breakdown of conglomerates, breccias, and even some mudstones, can produce a pebbly regolith that imparts a “smooth” appearance in satellite and aerial images. Context is important; sedimentary rocks remain challenging to distinguish from primary igneous rocks in some cases. Detection of ultramafic, mafic, or andesitic compositions do not dictate that a rock is igneous, and clast genesis should be considered separately from the depositional record. Mars likely has much more sedimentary rock than previously recognized.

1. Introduction

Sedimentary rocks are records of past surface and subsurface (diagenetic) environments that capture information about the nature and actions of fluids and the histories of associated tectonic, magmatic, and impact events. The foundational steps in the study of the sedimentary rock record of other worlds are to use images acquired by instruments onboard flyby, orbiting, airborne, ascending, or descending platforms to:

• Learn to recognize sedimentary rock occurrences; and
• Identify where they outcrop at the planet’s (or moon’s) surface.

These steps can then lead to detailed studies using existing data, e.g., [1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32], or to target new data collections, or to develop new spacecraft and instrumentation, to obtain additional observations.

This contribution is about learning to recognize sedimentary rock occurrences—for Mars—largely using visible wavelength, reflected light images, to identify sub-meter- to decameter-scale sedimentary structures and other signs that a given geological material, exposed at the planet’s surface, is sedimentary rock. Identification of sedimentary rocks on worlds other than Mars likely requires remote sensing methods that differ from those employed for Mars; for example, finding a sedimentary rock record on Venus is a high scientific priority [33,34,35] that will likely require use of new imaging radar systems and near-infrared airborne or descent cameras [36]. A search for sedimentary rocks on Titan, a world with present-day aeolian, fluvial, and lacustrine processes [37,38,39,40], might also require imaging radar systems and cameras on airborne, descending, and landed platforms such as the planned Dragonfly rotorcraft mission [41].

Sedimentary rocks were speculated to occur on Mars for decades [42,43,44,45,46], although some wondered whether the necessary diagenetic processes had ever been operative [47]. The first unambiguous evidence [48] that Mars has sedimentary rocks came in 1999 [49] from 1.4–4.5 m/pixel images obtained by a (then) new instrument orbiting the planet, the narrow angle camera of the Mars Global Surveyor (MGS) Mars Orbiter Camera system (MOC-NA) [50]. These early MOC-NA observations were soon followed by recognition of lithified landforms that resemble an aeolian dune field (Figure 1a) [51,52] and a river delta (Figure 1b) [53]. While some rocks at the Sojourner rover field site seen in 1997 were speculated to be conglomerates [54,55], the ultimate confirmation that sedimentary rocks occur on Mars came from images of the sandstones of the Burns formation examined using the Opportunity rover in 2004 in Meridiani Planum (Figure 1f) [56,57].

The sedimentary rock record of Mars is important to deciphering the planet’s past. Without sustained, mobile-lid (plate) tectonics [59], the upper Martian crust might have retained much of its earliest sedimentary record [60]. Of course, some early sedimentary rocks would have been subjected to subsequent weathering, erosion, and recycling [61]. With strata time-equivalent to Earth’s Hadean and early Archean eons—older than the oldest sedimentary and metasedimentary rocks on our planet [62,63]—the Martian archive presents opportunities to learn, by analog, about the kinds of environments that might have existed on Hadean Earth, Eoarchean Earth, or similar extrasolar worlds cf. [64]. If Mars never had life, then its sedimentary record also offers an opportunity to examine the products of deposition, diagenesis, weathering, and erosion—on a rocky planet with a silicate crust—in the absence of biological influence. In addition, Martian sedimentary strata might include records of local, to solar-system-wide, astronomical events [65,66], potentially correlative across multiple worlds.

This contribution is a guide to a daunting and unfinished task, the recognition of Martian sedimentary rock occurrences in images acquired by cameras aboard aerial (ascending, descending, or airborne platforms) and satellite (orbiting or flyby spacecraft orbiting another body or leaving the solar system) platforms that are, or were previously, operating (or someday will operate) at Mars. The observations, concepts, and ideas presented here are also applicable to future Mars landing site selection and robotic and human crew exploration planning efforts that make use of orbiter, aerial, or long-distance views from landed cameras. We use the term occurrences because outcrop inadequately limits identification; some examples can be recognized even though they are not strictly outcropping because they are partly (or mostly) covered with aeolian sand or dust [12,67,68].

Our work was highly informed by key observations made over the last decade using images obtained by cameras aboard the Mars Science Laboratory (MSL) Curiosity rover and the 99.97% (as of 1 January 2021) coverage of the planet achieved at ~6 m/pixel by the Mars Reconnaissance Orbiter (MRO) Context Camera (CTX). Tens of thousands of higher-spatial resolution orbiter images in the 0.25–6 m/pixel range, of select areas around the planet, also inform the results presented here.

Our overarching theme is that identification of sedimentary rocks on Mars is not limited to exposures that have geomorphic expressions that mimic their original depositional setting (Figure 1a–c), nor limited to stratified exposures of light tone that have poor impact crater retention (Figure 1d,e), nor is identification of sedimentary rock occurrences always easy. Contrary to understanding a decade ago [48,69], some sedimentary rocks are dark-toned (Figure 2), retain craters as well as do lava flows (Figure 3), and have mafic to felsic compositions [70,71,72,73] that, if observed remotely, could be misinterpreted as signals of igneous rock. One of the key challenges lies in the distinction of sedimentary from igneous rocks, including lava and tuff. This concern applies whether one is examining images acquired by orbiting cameras, airborne cameras, or even images obtained by rover cameras that view rock surfaces at scales of 16–32 µm per pixel (e.g., Figure 4).

In addition to serving as a guide to Martian sedimentary rock identification, this contribution is also an invitation to more deeply explore the vast data sets acquired from Mars over the last >55 years to look beyond the easy and the obvious. For example, are there sedimentary rocks—including those of dark tone and strong resistance to erosion—exposed in the walls of the Valles Marineris? What percentage of the yardang-forming and non-yardang-forming members and facies of the Medusae Fossae Formation [82] are sedimentary rocks rather than tuffs? How much of the upper crust of the Martian heavily cratered terrain is sedimentary, and are there sedimentary rocks beneath the Martian northern plains?

2. Materials and Methods

Pre-dating the 2021 arrival of (and, thus, not including results from) the Perseverance [83] and Zhurong [84] rovers, and the Ingenuity helicopter [85], the observations and ideas presented here come from more than two decades (February 1998–January 2021) of image acquisition planning, targeting, and examination of visible wavelength, reflected-light images obtained by cameras aboard Mars orbiters, descending platforms, and rovers. The data we used are described in the Data Availability Statement. Our data analysis focused on commonly used aerial, satellite, and remotely operated vehicle image interpretation approaches and skills, e.g., [86,87,88].

Most of the tens of thousands of satellite images examined were acquired by the MGS MOC-NA (1.4–12 m/pixel) [50,89], MRO CTX (~6 m/pixel [90]), and MRO High Resolution Imaging Science Experiment (HiRISE; 25–60 cm/pixel [91]) cameras. Prioritized, hypothesis-driven acquisition of the data [50] was strongly informed after 2001 by the geomorphic expressions and thermophysical contrasts of rock and regolith revealed by the ~100 m/pixel images from the Mars Odyssey (ODY) Thermal Infrared Imaging System infrared subsystem (THEMIS-IR [92]) and, after 2006, also by the growing 6 m/pixel coverage of Mars by CTX. High spatial resolution images from the Tianwen-1 High Resolution Imaging Camera (HiRIC [93]), and the ExoMars Trace Gas Orbiter (TGO) Colour and Stereo Surface Imaging System (CaSSIS [94]), were not examined for this study, but their data are relevant to application of the ideas presented here because these cameras provide images at scales that fall within the 0.25 to 6.0 m/pixel range bracketed by HiRISE and CTX.

Images from the cameras onboard the Curiosity rover, operative in Gale crater [95], were also vital and examined as the data were received on Earth. These include images obtained during the August 2012 landing by the Mars Descent Imager (MARDI [96]); images and mosaics acquired by the Navigation and Hazard cameras (Navcams, Hazcams [97]), Mast cameras (Mastcams [96]), and Remote Micro Imager (RMI [98]) on the rover’s remote sensing mast [99]; as well as the Mars Hand Lens Imager (MAHLI [100]) on the turret at the end of Curiosity’s robotic arm [101]. The rover’s mobility system [102] was used to move the cameras to different localities in Gale crater. Images from the Panoramic camera (Pancam [103]), Microscopic Imager (MI [104]), and Descent Image Motion Estimation System (DIMES [105]), onboard the Mars Exploration Rover (MER), Opportunity—which operated in Meridiani Planum (2004–2018 [106])—also informed the results presented here.

3. Results

Identification of sedimentary rock occurrences on Mars can be challenging. This is true even when considering 16–32 µm/pixel images of Martian rocks obtained by rover cameras (Figure 4). First, there is the likelihood that colleagues will disagree on interpretation, typically because they have not first agreed on the nature of the observables. This type of disagreement can occur even on Earth when a group of geologists of differing backgrounds and experience are together, examining an outcrop in the field. On Mars, no one is “in the field” and so the discussion is reliant on remotely gathered data. This section describes features that are sought to make the case that a given geological feature is an occurrence of sedimentary rock.

Sedimentary rock outcrop and landform erosional expressions are a function of rock physical properties and the processes and chemistries that have acted upon the rocks to modify them over time [107,108,109]. Important attributes include depositional texture (e.g., grain size, packing), depositional structure (e.g., bedding), diagenetic features at all scales (e.g., differential cementation, fractures, veins, concretions, etc.), and depositional, basin-scale facies configurations (e.g., interfingering of sandstones and mudstones; transgressive and regressive relations in subaqueous sediments). Deformational structure can also be helpful for sedimentary rock occurrence identification in some cases (e.g., jointed or polygonal fracture patterns).

Identification is aided by standard photointerpretation methods, particularly through the “recognition elements” of Ray [86]: tone (and color, where available), texture (frequency of tonal change as a function of image scale), pattern, shape, size, relation of a feature to associated features, and combinations of these. Solar illumination conditions and seasonal frost patterns are also important to consider when examining aerial and satellite images of Mars. The phase angle can have a profound impact on the visibility of landforms and the perception of tone and color.

Sedimentary rock occurrence identification also benefits from familiarity developed from field experience on Earth, robotic (rover) field experience on Mars, photographic guidebooks, e.g., [110,111], and illustrations of Earth’s sedimentary rock landforms [107,109,112,113,114]. For example, systematic, parallel, or subparallel alternations in tone [115]—characteristic of layered rock outcrops bearing distinct stratal horizons—provide approximate knowledge of the stratigraphic locations of depositional surfaces [116,117].

3.1. Signs That a Geological Material Is Rock

Identification of sedimentary rock occurrences in satellite and aerial images of Mars begins with determination of whether an observed feature is, or has the properties of, an exposure of rock. Rock is a consolidated (lithified) geological material that, on Mars, is distinguished from ice on the same basis that whole worlds are classified as rocky or icy. Rocky worlds are dominated by silicates and metals [118]; ices (e.g., H₂O, CO₂, CH₄) dominate the crust and regolith of colder worlds [118]. Thus, while Mars has H₂O and CO₂ ices [119,120]—and these can cement sediment and regolith at high latitudes [121,122,123] and form ephemeral boulders during mass movement events [124]—ices and ice-cemented sediments on Mars are not generally considered to be rock.

Rock is created through lithification. On rocky worlds, it is produced by freezing of melt (magma, lava, pyroclasts, impact melt), welding of pyroclasts, sintering of clasts via impact-cratering-induced shock metamorphism [125], formation of chemical sediments, or diagenesis of sediment and primary volcaniclastic deposits. Clastic diagenesis can involve cementation, compaction, and cementation, or compaction without cementation [126]. Chemical sediments can also be subjected to diagenesis, including recrystallization. Lithification of clastic sediment can occur in both surface/near-surface, e.g., [127,128,129], and sub-surface environments, e.g., [126,130]. Clastic lithification is distinguished from induration because the latter refers to a broader range of phenomena. Induration includes lithification through sediment diagenesis and welding of tephra, but also refers to case hardening of weathered rock surfaces and thermal alteration of sediment or primary volcaniclastics in contact with lava or magma [131].

Recognition that a material is rock in high spatial resolution (e.g., centimeters to several meters per pixel) satellite or aerial images of Mars can be challenging, and the observables are not always fool proof. For example, even on Earth there remains a lack of consensus as to whether cemented, Quaternary loess (aeolian silt) should be considered to be rock [132]. Dry playa muds present another example; some playa surfaces can be so hard as to permit landing an aircraft or spacecraft on them [133]; such a material—if present on Mars and hardened at the surface/atmosphere interface—could be indistinguishable from bedrock. However, there are several observables that can be sought to help make an interpretation that a material is rock or—like dried playa muds—has properties similar to rock: boulder production; preservation of steep slopes; distinct contrasts in tone or color between strata; sharp expressions of faults, fractures, or joints; resistance to the impact of saltating grains; the presence of yardangs or other aeolian erosional forms; or the presence of an overlying material determined to be rock. These observables do not all have to be evident to make the case that a material is rock. In addition, they might not always be diagnostic.

3.1.1. Boulder and Megaclast Production

Clast production and release can occur as a unit of rock breaks down at a planet’s surface. Clast release is the liberation of particles (grains or concretions) already present in a rock, whereas clast production refers to the breakdown sensu [134,135] of rock into smaller fragments. On Mars, boulders and megaclasts are typically produced when an impact cratering event fragments rock (Figure 5a), when rock is weathered and eroded from steep slopes, or when plateau-, mesa-, or butte-capping rocks break up as the landforms decay (Figure 5b,c). Boulders in crater ejecta are a good sign that the impacted material is rock, although boulders present in an impacted, unconsolidated deposit can also be released by such an event. Lava flows and sedimentary rock outcrops, alike, can exhibit signs of boulder production. Polygonal fracture patterns in rock can also lead to boulder-sized fragment production when fracture-bounded slabs are dislodged and displaced from their original outcrop position (Figure 5d).

Of course, how long a produced boulder or megaclast may persist in the landscape—before being broken down to finer particles that are transported away—will depend upon the rock physical properties and the processes acting on that rock. A boulder produced from an erosionally-recessive rock, for example, will not persist as long as a boulder produced from an erosionally-resistant rock, as evident in the comparison of the landscapes at the Opportunity and Spirit rover sites in Meridiani Planum (sandstone bedrock) and Gusev crater (basalt lava bedrock) [137].

3.1.2. Retention of Steep Slopes

The ability of a geological material to hold a steep slope that exceeds the angle of repose—such as an escarpment, caldera wall, impact crater wall, or canyon wall—is a key indicator that a geological material is likely rock [12,49,138]. Slip faces of whole aeolian sand dunes that have become lithified [51,52,139] can also exceed the angle of repose [139,140]. Among Martian rock outcrops, escarpment-bounded landforms include plateaus, mesas, buttes, and pinnacles (Figure 5b; see definitions of [108]). Overhangs developed above recesses in escarpments also signal the presence of rock and can be rendered visible in satellite or aerial images by their shadows or via off-nadir views (Figure 6a).

3.1.3. Distinct Contrast between Strata

Sharp contrasts in tone or color exhibited by, or a lack of blur between, adjacent strata or stratal packages exposed on a slope are good signals that rock is present (Figure 6b) [49]. One caution is that apparent sharp tonal contrasts between strata can, in some cases, instead be an indicator of modern sediment (aeolian sands; talus) deposited on rock outcrop benches separating strata of similar tone (Figure 6c) [24,25].

3.1.4. Sharp Expressions of Faults, Fractures, and Joints

Sharply distinguished faults and offsets along faults (Figure 6d) [49], as well as fractures and joints (Figure 6e) [8,25,143,144,145], are good indicators of the presence of rock. However, like the aforementioned uncertainty regarding cemented loess [132], the presence of fractures might not always indicate a lithified material.

3.1.5. Resistance to the Impact of Saltating Sands

Wind-mobilized sand is active on Mars in the modern environment [146,147]. Aeolian interdune substrates and dune transport corridors free of aeolian dust and sand accumulations can be good signs of rock exposures on Mars [49]. Saltating sand grains bounce off of hard surfaces [148], such as a bedrock exposure, limiting the accumulation of aeolian dust and sand. Figure 6f shows an example observed using an orbiting camera, and Figure 6g shows an example viewed from the ground in Gale crater.

3.1.6. Kharafish (Wind Erosion Forms)

Signaling the presence of rock—or at least indurated material—kharafish is an Egyptian term that encompasses the full range of aeolian erosional forms [149], from kilometer-scale mega-yardangs [150], to centimeter-scale ridges and grooves [151] (p. 68), and millimeter-scale flutes [152]. Yardang-scale kharafish (Figure 7a) occurs in a variety of forms that go beyond the classical inverted boat hull shape [153,154], including meter- to kilometer-scale scour-remnants that develop in the lee of knobs (Figure 7b) of greater erosion resistance [155] (p. 38). Yardangs form in a variety of lithologies on Earth, including carbonates, tuffs, and clastic sediment [156,157,158,159,160,161]; unusual examples include yardangs formed in weathered basalt [157] and weathered gneiss [155,162]. The term, kharafish, might extend to the “knapped” (Figure 7c) [24,25] erosional expression common to some Martian sedimentary rock outcrops, although the exact origins of these surfaces needs further study. Field studies, e.g., [153,157,163,164] indicate that the morphologies of yardangs and other kharafish are related to material properties such as texture (e.g., grain size), depositional structure (e.g., bedding), and deformational structure (e.g., joints, faults, folds). In addition, wind erosion shapes rock according to its physical properties, such that erosion-resistant heterogeneities in the rock may form protrusive elements (Figure 7d) [165].

3.1.7. Emergence from Beneath a Body of Rock

Another sign that a material is rock is its partial exposure from beneath an overlying material that is determined to be rock based on any of the above observables. Figure 8a shows a case in northern Meridiani Planum in which an impact crater ejecta deposit is partly exhumed from beneath light-toned, sedimentary rock [168]. Because the stratigraphically higher light-toned material in Figure 8a has been lithified, the clastic impact ejecta beneath that rock must also have been subjected to diagenesis and lithification, not unlike examples of lithified impact ejecta on Earth e.g., [169,170]. The other examples in Figure 8 illustrate cases of fluvial and alluvial sedimentary forms (see Section 3.5.2 and Section 3.5.3) emergent from beneath rock or remnants of rock. An exception occurs in cases of lava flow superposition on unconsolidated regolith, sediment e.g., [171], or tephra, provided that the underlying material has not subsequently become lithified through diagenesis. It should also be noted that some sediments emergent from beneath units of rock could be much more poorly lithified than their overburden, such as much of the Fontainebleau Grés (Sandstone) in France, which is overlain by limestones [172,173,174,175].

3.2. Stratification

Commonly expressed in landscapes and rock outcrops as a product of differential resistance to erosion, stratification is “fundamental to” [86] (p. 16) the recognition of sedimentary rock occurrences in satellite and aerial images [86,115,116,117]. Sedimentary strata are time-ordered records of change (i.e., geological history) and their three-dimensional configurations provide information about sediment transport and depositional conditions. Of course, strata are not unique to sedimentary rocks—unconsolidated sediment, primary volcaniclastic deposits, e.g., [178], accumulations of lava, e.g., [179], the perennial Martian polar caps [121,180,181], and even the igneous rocks and impactogenic regoliths of the Moon, e.g., [182,183,184], can all exhibit stratification.

MOC-NA and HiRISE images emphasized the ubiquity of stratified materials—sedimentary, igneous, and otherwise—exposed at the Martian surface [48,115,185,186]. In such images, stratification is identified through “the presence of alternations in brightness,” particularly those that are “laterally traceable for ~10 m or more,” or the observation of “distinct shadowed, shelf-like topographic breaks in slope” [115] (p. 17). Images of higher than HiRISE (~25 cm/pixel) spatial resolution can reveal stratification at finer scales, both in terms of lateral extent and layer thickness [21,187]. The finest strata on Mars—laminae undetectable using orbiter imaging systems sent to Mars thus far—display sub-millimeter thicknesses [188,189,190].

Satellite and aerial images of Martian sedimentary rock outcrops exhibit stratification through landform expression, stratal patterns, and the revelation—by erosional exposure—of major unconformities that reveal paleo-surfaces. Considerable information about the materials can be interpreted from images of Martian sedimentary strata, including (but not limited to) studies that address depositional settings and conditions, e.g., [1,2,5,17,27,191], depositional and deformational structure, e.g., [2,4,8,25], environmental controls on repeated (cyclic, rhythmic, periodic) stratification e.g., [16,48,192], and stratigraphic correlation across distance, e.g., [12,19,28].

3.2.1. Expressions of Stratification

Where large enough to spatially resolve in a given satellite or aerial image, strata exposed on Martian rock outcrop slopes are expressed as terraces or cliff-bench forms, alternating bands of differing color or tone, or through the presence of gaps that parallel or separate strata (Figure 9). Gaps, narrow relative to the scale of the outcrop (Figure 9d), can result from the presence of erosionally-recessed layers or from fractures or joints (fractures of regular spacing) that parallel the bedding. Fractures or joints (as well as bedding plane separations) can result from stress relief caused by overburden removal, e.g., [193] or by undermining. Gaps can also be apparent features, not true gaps, created at erosionally recessed partings on rock slopes, e.g., [194] (pp. 208–209); a parting is a thin stratum or group of strata consisting of a lithology that differs from the thicker, main lithology above and below it [195]. Gaps will be apparent on a rock face if a parting has recessed far enough back that it cannot be seen in available images.

Cliff-bench, stair-stepped, and terraced landform expressions are common, although not universal, among Martian sedimentary rock occurrences (Figure 9a,b). Cliff-bench forms consist of prominent stacks of layered rock exhibiting vertical (cliff) and horizontal (bench) elements. Stair-stepped occurrences are similar, but the vertical height of each cliff, and the width of each bench, are similar and repeated (or nearly so), like a staircase. Terraced hillslopes [109] are another variant in which benches form narrow terraces, cliffs are absent, and slopes may approach that of the angle of repose. The cliffs in cliff-bench sedimentary rock landforms are usually not actually vertical, and the pattern can be one of thin, erosion-recessive and thicker, erosion-resistant couplets [17]. A thin, recessive layer need not always be present; differences in resistance to weathering and erosion resulting from differences between the top and bottom of a given stratum or stratal package (e.g., particle size, sorting, cementation, and/or composition [196]) could also contribute to the patterns observed.

Banding can indicate stratification, and is expressed in orbiter or aerial images of Mars as alternations of strata or stratal packages that exhibit differing color or tone (Figure 9c). Color or tone differences can result from intrinsic tonal or color properties of the rock, or from textures (e.g., rough vs. smooth) that exist on the rock surface (or alluvium accumulated on the rock surface) at a scale below the spatial resolution of the image (see Section 3.3). Banding can also result from accumulation of rock breakdown debris, windblown sand, or windblown dust on benches or terraces on slopes cutting across stratified rock (Figure 6c). Finally, banding might, in some cases, be reflective of stratigraphic properties but is actually a product of diagenesis or weathering. The Burns formation exposure in Endurance crater examined via Opportunity provided an example of diagenetic banding [57]. According to Young et al. [107] (p. 12), outcrops of the Glass Hill Sandstone of the Bungle Bungle range in the Australian Kimberley provide an example on Earth of banding that is only skin-deep. There, colorful bands exhibited by towering sandstone cliffs are caused by weathering and microbial activity acting on minor stratigraphic variations in porosity and clay content.

3.2.2. Patterns of Stratification

Exposures of stratified sediment can exhibit stratal patterns [197] that aid the identification of sedimentary rock and are useful for interpreting depositional settings and conditions [14]. Patterns include the nature of stratum and stratal package thickness, the three-dimensional shape of a stratum or stratal package (e.g., lens, wedge, etc.), repetitious stratification, and sedimentary structures large enough to observe in aerial and satellite images, such as meter- to decameter-scale cross stratification.

In their initial survey of sedimentary rocks in early MOC-NA images, Malin and Edgett [49] described units of thin, thick, and massive stratal packages. “Thin” packages were said to be a few meters to ~200 m thick and contain “one to hundreds” of thinner beds; “thick” packages were described as being ~200–2000 m thick and also contain thinner layers “Massive” packages were a few hundreds of meters to a few kilometers thick and exhibit poor to zero evidence of stratification within them. In describing a stratified rock occurrence observed in a satellite or aerial image, terms such as thin and thick should be accompanied by quantitative definitions [14,198]. Massive, too, is a problematic term—although it remains in use, even for Martian sediments e.g., [23,31,49]—because it has nothing to do with mass; an alternative term is structureless [197]. In this case, structureless refers to depositional rather than deformational structure [197]. The structureless descriptor should be applied relative to observation scale (e.g., the rock unit exhibits no depositional structure observable at an image scale of ~25 cm/pixel). Some of the rock units explored at the Opportunity and Curiosity rover sites can appear to be (or are nearly) structureless when observed in HiRISE images, but exhibit considerable sedimentary structure (e.g., fine laminae, cross bedding) when viewed at higher spatial resolution using rover cameras (Figure 10).

Stratification patterns include those that repeat and those that provide information about depositional forms and conditions (e.g., lenses, channel fills, cross stratification). Repeated patterns include repetition of stratum (or stratal package) thickness, rock properties (as expressed in the landscape by erosion), or the patterning within one stratal package as compared to the next. Repetition (Figure 9a) might be related to reoccurring, periodic, or quasi-periodic climate patterns [16,192]. In clastic rocks, repeated strata also imply that sediment supplies of similar or equivalent particle size, composition, and source are available through each period or cycle. Although spectacular in appearance [16,69,192], repeated (or cyclic) stratal patterns are, apparently [14,48], uncommon on Mars relative to the overall abundance of sedimentary rock occurrences identified (up to the year ~2012) in 0.25–6.0 m/pixel orbiter images.

Cross-stratification and other sedimentation patterns are sometimes evident—albeit very difficult to discern in many cases—in satellite or aerial images of well-exposed sedimentary rocks on Mars. Grotzinger and Milliken [48], for example, noted an outcrop that cut through a feature that they interpreted as a small, filled, buried impact structure. Dromart et al. [2] described a complex suite of stratal patterns in Melas Chasma—a stratal architecture—interpreted to include subaqueous clinoforms and channel fill sediments (Figure 11a). Meter- to decameter-scale cross stratification, interpreted to be aeolian, is seen in a variety of settings on Mars, including several locations in Juventae Chasma (e.g., Figure 11b) [25], and stratal packages exposed in Aeolis Mons in Gale crater [22,191,201]. An important caveat applies here; Fedo et al. [202] noted that, based on an experience with mudstones explored via Curiosity in Gale crater, some apparent cross-bedding patterns might—upon closer inspection—turn out to be a fracture (or fracture and vein) pattern, and determining this might be impossible using orbiter or aerial images.

Another important challenge to interpretation of sedimentary rocks observed on Mars was explored by Day and Catling [30]. That is, some known Martian aeolian sandstones do not exhibit cross-bedding patterns that are large enough to be observed in the presently highest spatial resolution orbiter images (i.e., HiRISE), particularly the Burns formation in Meridiani Planum [57,203,204], and the Stimson formation in Gale crater [136,188].

3.2.3. Unconformities

Unconformities in sedimentary rock sections indicate episodes during which sediments were not deposited and erosion occurred [205]. Previously eroded surfaces, emergent from beneath sedimentary strata, provide the most obvious examples of unconformities that can be recognized in satellite and aerial images of Mars. Unconformities in stratal sections that expose older, partly exhumed terrain include the cratered surface identified by Malin and Edgett [49] about a quarter of the way up the north slope of Aeolis Mons in Gale crater (Figure 12a), and the partly-exhumed surfaces in Figure 8. Unconformities can also be identified in places where a younger stratal package overlies older strata in a crosscutting relation; examples occur in Terby crater [see 11], Melas Chasma (Figure 11a) [2], and Galle crater (Figure 12b) [206].

3.3. Tone and Color

3.3.1. Tone

Global, visible wavelength image coverage of Mars at the highest spatial resolution currently available (6 m/pixel), is grayscale. Acquired by the panchromatic MRO CTX, these grayscale images provide the foundational basis for planet-wide identification of actual and candidate sedimentary rock occurrences. They also provide vital context for the higher spatial resolution views obtained by other orbiter cameras such as MOC-NA, HiRISE, HiRIC, and CaSSIS, as well as aerial imaging. While some of these higher spatial resolution cameras provided color views of Martian sedimentary rock outcrops, the entire planet has not yet been imaged, in color, at high spatial resolution (e.g., 0.25–6 m/pixel). Thus, tone is presently more valuable, globally, for identification, description, and inter-regional comparison of sedimentary rock occurrences. Tone is a qualitative proxy for albedo that is usually discussed in a relative sense, i.e., whether a feature is lighter or darker than adjacent or nearby features in an image or image mosaic.

Many of the early identifications of sedimentary rock occurrences on Mars noted their light tone (Figure 1d,e and Figure 6f) [48,49]. In addition, the sandstones of the Burns formation at the Opportunity rover site were also light-toned (Figure 1f) [56,57]. However, sedimentary rocks of dark tone and “intermediate” tone (between light and dark) were also known [49], but considered less common.

Observations made at the Curiosity, Opportunity, and Spirit field sites have underscored a few important points regarding the tones of sedimentary rock occurrences observable in satellite and aerial images. First, the observed tone can be an intrinsic property of the rock, such as the observation that the sandstones at Meridiani Planum are light-toned (Figure 1f). Second, there is the possibility that tone can be an intrinsic property not of the rock but of a surface coating or weathering rind [207,208,209]. Third, and more importantly, the observed tone might be an extrinsic signal of a combination (in various proportions) of rock surface texture and aeolian dust or sand accumulation (Figure 13). Surface roughness at scales smaller than a given satellite or aerial image pixel can cause a rock surface to appear darker than the intrinsic rock tone because of self-shadowing. In contrast, accumulation of light-toned aeolian dust can make the rock appear to be lighter than its intrinsic tone (Figure 13a,b) [210] (p. 6). Martian aeolian sand is largely dark-toned [211], and saltating sand will generally bounce off of rock surfaces [148]; however, sand grains can accumulate in pits and fractures that are too small to observe in remotely-acquired images (Figure 13c,d) and, thus, can impart an apparent darker tone. It is also important to note that where an entire region is coated with dust, the apparent (extrinsic) tone of a given rock unit will be indistinguishable from the surroundings (e.g., lithified dune field near the Apollinaris Sulci, Figure 1a).

3.3.2. Color

Rock color is like tone. It can be an intrinsic property or it can be (literally) a reflection of sub-pixel scale roughness, superimposed aeolian sediment, or a combination of these. Color is also challenging to establish—what is the “true color” of something? If you have a lawn of green grass, how does that green look to your eye in sunlight? On a cloudy day? In the shadow of a tree? In twilight? Do your eyes perceive green the same way as another person’s eyes? In some cases, satellite and aerial imaging systems use cameras that attempt to capture color close to that which the typical human eye sees (e.g., Bayer pattern microfilters [212]), or they use cameras that capture images at wavelengths not visible to the human eye (e.g., MRO HiRISE near-infrared band [91]).

A key observation from the Curiosity rover mission in Gale crater is that Martian sedimentary rocks exhibit a variety of colors, as evident in Bayer pattern color interpolated images of bare rock surfaces acquired by the MAHLI (e.g., Figure 14). Examples from Curiosity include conglomerates, sandstones, and mudstones of various shades of gray, from nearly black to nearly white; mudstones that are reddish-brown, yellowish brown, or purplish-gray; and vein-forming minerals that are white and various shades of gray (Figure 14). The colors (excluding white) generally result from varied amounts of oxidized versus reduced iron-bearing minerals (e.g., hematite vs. magnetite [213,214]).

3.4. Rock Breakdown and Resistance to Erosion

Landforms and landscapes are temporary, even on the ancient surface of Mars [215] (p. 45). Their configurations result from the competition between the processes and rates at which new rocks are formed, or become exposed at the surface, versus the processes and rates at which rocks disaggregate and the fragments and dissolved ions are transported and deposited elsewhere. On modern and geologically recent Mars, lateral erosion (e.g., undermining, scarp retreat, rock falls, wind abrasion) has been proceeding at a higher rate than vertical erosion (lowering of surfaces); this is evident, for example, in the pace of erosional exposure of rock in Gale crater [216]. Vertical erosion rates may slow as lags develop or erosion-resistant strata become exposed.

Subtle and not-so-subtle differences in rock properties can make a big difference in landscape configuration. In sedimentary rock exposures, these result from differences in lithology, including differing intraformational depositional or diagenetic facies [107]. For example, consider the impact on the landscape of the Petrified Forest National Park in Arizona, USA, of intraformational sandstones and conglomerates in the otherwise mudstone-dominated Chinle formation e.g., [217]. Such erosion-resistant rocks, part of the same formation as the very colorful and much more erodible mudstones [217], led to the creation of plateaus and hills that are capped by sandstones and conglomerates or mantled by—or formerly mantled by—their broken-down debris. Variability in intraformational diagenetic facies in the Murray formation, a largely mudstone-bearing rock unit in Gale crater, provides examples of diagenetic contributions (e.g., variable cementation) to landscape configuration, including Vera Rubin ridge [218,219].

Breakdown (or weathering or decay) [134,135,220], and erosion, of Martian sedimentary rock imparts signals of rock properties into the landscape that can be observed in images acquired by orbiting and aerial instrumentation. Some of these signals are interpretable as rudimentary information about sedimentary rock texture or structure. Others—such as impact crater retention and formation of escarpments—provide information that can be interpreted in terms of relative rock hardness. Resistance to breakdown and erosion can also be a function of the nature of overlying, unconsolidated debris, including lags deposits and aeolian bedforms that protect the bedrock from erosion.

3.4.1. Small Impact Crater Retention

The contribution of impact cratering to Martian rock breakdown and redistributive transport of the resulting debris should not be underestimated [137,221,222]. In this context, small (sub-kilometer diameter) impact crater retention refers to the opposite—the presence and abundance of impact structures, at the Martian surface, as a signal of rock physical properties. In particular, it refers to the erodibility of the rock in geologically recent and modern times, rather than the application of impact crater size-frequency distributions for absolute or relative age dating of geological materials.

Figure 15a illustrates the concept using an example of a lava flow located west of Meridiani Planum that superposes its surrounding terrain. In this case, the flow is younger than the materials it embays, but it retains many more small impact structures per unit area than does the older, surrounding terrain. Figure 15b, showing lavas that embay yardang-forming material in Galdakao crater, provides a similar illustration; in this case, the one large impact structure in the yardang-forming material underscores the greater crater retention age sensu [223] of these rocks, but the relative abundance of very small craters on the lava surface illustrate the point that the lavas are more resistant to erosion and, thus, retain small craters better than the yardang-forming material. Of course, it should be noted that not all lavas are heavily peppered with small impact craters. Figure 15c shows an example in Elysium Planitia that has a young crater retention age, attributed to the youth of these lavas. Vaucher et al. [224] estimated these lavas might be only 2.8 ± 0.5 million years old.

Until Curiosity’s exploration in Gale crater was well underway, it had seemed that the majority of Martian sedimentary rock occurrences have very few superimposed impact craters [48,49,69,90,115,225]. This lack of craters was considered to be a signal of the erodibility of these rocks which, in turn, was seen to be an indicator that they consist of materials that are readily released by weathering and removed by erosion (e.g., clasts of sand size and smaller). In contrast, surfaces heavily peppered with sub-kilometer-diameter impact craters seemed to provide a good signal for distinguishing lava plains in cases for which no vents, edifices, flow fronts, pressure ridges, or other geomorphic evidence of volcanism are present [226]. However, there was a mystery. Among the very first sedimentary rocks identified on Mars—in MOC-NA images in 1999—were those of a lithified aeolian dune field in the Apollinaris Sulci region (Figure 1a) [51,52]. The lithified dune field is dotted by many small impact craters. One and two decades ago, the retention of small craters in a material that is best interpreted as sandstone seemed strange relative to the majority of known sedimentary rock occurrences [226].

The crater-retaining, lithified dune field (Figure 1a) was a signal that, perhaps, not all Martian sedimentary rock occurrences are so easily eroded. However, recognition that this observation is not unusual did not come until the early years of the Curiosity mission. During Curiosity’s descent to the floor of Gale crater in 2012, its MARDI recorded an image of the impact of the MSL heat shield on a surface that retains many sub-kilometer-diameter impact structures (Figure 3b). For months thereafter, the science team considered it possible that this crater-retaining bedrock surface (literally called a “cratered surface unit” [75,76]) might be lava, with the largest crater on its surface (Figure 3a) a possible maar. Indeed, many other cratered surfaces, elsewhere in northern Gale crater, were also considered to be candidate lava surfaces [76]. Ultimately, Curiosity was driven close to the northern edge of the cratered surface unit that was impacted by the heat shield (Figure 3c) and its Mastcam and ChemCam RMI images (e.g., Figure 3d,e) showed that the rock is sandstone [77,78].

Thus, a picture began to emerge, that some Martian sedimentary rocks are sufficiently resistant to erosion that they retain (at least, qualitatively) sub-kilometer-diameter impact craters as well as or nearly as well as do Martian lava flows and plains [227]. Continued investigation via Curiosity has shown that while well-cemented [228] sandstones and conglomerates retain small impact craters [229], so, too, can some mudstones. This is particularly evident on Vera Rubin ridge (Figure 16). Diagenesis caused—in a manner that cuts across stratigraphic boundaries [190]—this particular part of the Murray formation to be more erosion-resistant than other parts [190,218]. While most of the Murray formation consists of recessive mudstones [81], Vera Rubin ridge is a topographic high and retains more craters per unit surface area than typical Murray formation exposures [230].

The number of small (sub-kilometer diameter) impact craters and remnant impact structures retained by a given surface (of a given surface area) on Mars is the integrated product of:

• The total period (which can include multiple episodes of burial and exhumation) during which the surface was exposed to the atmosphere through which impactors penetrate to reach it;
• The impact cratering rate and variability of that rate—owing to changes in atmospheric density, impactor supply, and astrodynamic availability of impactors—during the period(s) of exposure;
• The formation of secondary impact craters and proximity to the primary impact site; and
• The rate and variability of the rate at which impact structures are removed by erosion, including destruction by subsequent impact events.

The removal of small impact structures is, at least in part, a function of rock physical properties—erosion-resistant rocks can retain many small craters [137]. The Curiosity rover mission has shown that such erosion-resistant rocks can include mudstones (Figure 16), conglomerates, and, especially, well-cemented sandstones (Figure 3).

3.4.2. Boulder Production and Retention

Retention of boulders, cobbles, and megaclasts [231] produced from rock breakdown is analogous to impact crater retention; rocks more resistant to erosion produce megaclasts, boulders, and cobbles that are retained longer because they are more resistant to post-production disaggregation and removal of the resultant fines (and vice versa). Boulders, cobbles, and megaclasts can be produced from sedimentary rock by impact events (Figure 5a), fracturing followed by dislodgement and displacement (e.g., mass movement; rock fall) at retreating escarpments (Figure 5b,c) and crater walls, or from fracturing followed by slower dislodgement and displacement processes (Figure 5d).

In MOC-NA images acquired through the middle months of the year 2000, boulders associated with sedimentary rock occurrences were not seen [49]. This observation was considered to be an indicator that the rocks are composed of fine-grained clasts such as clay, silt, and sand [49]. However, this notion changed dramatically once the MOC-NA was operated in cPROTO (compensated Pitch and Roll Targeted Observation) mode for higher spatial resolution views [50]. MOC-NA cPROTO images showed boulders shed from sedimentary rock escarpments in, for example, northern Sinus Meridiani and western Candor Chasma [168,232]. Subsequent imaging by MRO HiRISE and rover cameras has, of course, also shown boulders produced from Martian sedimentary rock (Figure 5). Still, of course, a lack of retained boulders is, in some cases, interpretable as a signal that a rock poorly resists erosion, and that any boulders produced are soon reduced to particle sizes small enough to be removed by wind [49,233].

3.4.3. Friability

Contrary to common usage of the term in Mars geoscience [72,225,234,235,236,237], the friability of a rock cannot be detected using satellite or aerial images. The geological term, friable, refers to rock that crumbles or sheds small fragments (such as sand grains) as a result of rubbing by human touch. No geomorphic observations can be said to indicate friability, although, in the Mars context, poor impact crater retention and kharafish landforms such as yardangs have been invoked as evidence of friability. Where no human hand (or robotic touch of similar pressure) has rubbed the rock, friability cannot be determined. Further, friability is not an indicator of rock hardness or resistance to erosion. For example, Young [238] described friable, yet steep-sloped (in some places, vertical), sandstone outcrops in the Bungle Bungle range of the Kimberley of Australia. In this case, a lack of intergranular cement makes the rock friable, but compaction and interlocking grains make the rock resist erosion [238].

3.4.4. Escarpments and Scarp Retreat

Slopes cut through sedimentary rock strata on Mars can take a variety of forms that reflect (1) local- to deposit-scale (or basin-scale) physical properties related to depositional and diagenetic texture and structure; (2) the landscape-shaping processes that have acted upon the outcrop, including (but not limited to) impact crater formation, faulting, mass movement, fluvial erosion, and aeolian erosion; and (3) the nature of subsequent slope-modifying processes and their relation to the time elapsed since the slope was initially formed. Differential erosion of stratified rock units of differing physical properties can be expressed in landscapes in the form of escarpments and scarp retreat (Figure 5b), leading to formation of the canonical progression of landforms from plateau to mesa to butte to pinnacle and hill [108,112,239]. Lateral erosion in rocks of differing properties may lead to undermining, collapse, disaggregation, and dispersal of erosion-resistant lithologies that overlie those of poorer resistance. Where rock is fractured and groundwater is present, it can lead to dissolution and weathering of the rock even before it is exposed at the planetary surface [108,240]. Where the rock is fractured, water, or ice tables are lowered or not present at all, and the environment is subject to aeolian processes, deflation and abrasion by saltating grains can contribute to scarp retreat and shallow basin formation [241,242].

As noted previously, rover observations made in Gale crater suggest that lateral scarp retreat rates via aeolian processes dominate over vertical erosion rates [216]. Plateaus, mesas, and buttes are generally topped by a capping rock unit that is more resistant to erosion than the rock beneath it. Observations from Gale crater demonstrate that such erosion-resistant rocks can be sedimentary (e.g., the sandstones capping the Emerson plateau and Murray buttes [136,188]). Such capping rocks can be intraformational, such as the sandstones and conglomerates of the Chinle Formation on Earth, noted above [217], or they can consist of a drastically differing lithology (e.g., lava overlying sedimentary rock). In any case, they can be either conformable or unconformable with subjacent strata; an example of the latter is the Salsberry Peak sandstone (Figure 2), which is part of the Pahrump Hills member of the Murray formation in Gale crater [74,80], and examples of the latter include the Stimson formation rocks that overlie the Murray formation also in Gale crater (Figure 17) [136,145,188].

3.4.5. Dispersal of Liberated Clasts or Rock Fragments in the Landscape

As clastic rocks break down, produced lithic fragments and liberated clasts are released into the landscape. In the case of sand-, silt-, and clay-sized particles, they are likely to be subject to mobilization and transport by wind. Distinguishing such materials and linking them directly to their source outcrop in satellite and aerial images is generally quite difficult. It would be ideal to be able to say that a given rock is a sandstone because a satellite image reveals it to be the source of liberated sand that has moved downwind from the outcrop. While this is usually not possible, Chojnacki et al. [244] demonstrated that a combination of satellite image and multispectral remote sensing analysis provided a reasonable indictor that some of the sulfate mineral-bearing sedimentary rocks in the Valles Marineris have been shedding sand-sized material (whether produced fragments or liberated clasts cannot be known from the data) that is contributed locally to aeolian sand dunes.

Larger liberated or produced clasts—boulders, cobbles, and pebbles—are generally likely to remain near their source until they breakdown to finer particles and are removed by wind (exceptions can include ballistic transport by impact events and transport by fluvial action earlier in Martian history). The boulder-sized rock fragments referred to in Section 3.4.2 and Figure 5 provide an example, as do the pebbles liberated from and lying on the ground next to conglomerate outcrops in Gale crater (Figure 18a) [245], and pebbles produced from the breakdown of Jura member (of the Murray formation) mudstones, also in Gale crater (Figure 18b,c) [246].

3.4.6. Burial or Lag Protection

Liberation or production of pebbles from rock breakdown can accumulate and form a pebbly regolith or an aeolian lag [248] that further enhances resistance to vertical erosion of underlying rock. A pebbly regolith or lag can appear to be smooth and featureless at image pixel scales that exceed the size of the pebbles. This is true whether the source of the pebbles is liberation from a conglomerate or breccia (Figure 18a) or breakdown of a finer-grained rock, such as mudstone (Figure 18b). An example of this occurs at the Kimberley field site visited by the Curiosity rover, where the Point Coulomb member of the Kimberley formation [247], a conglomerate bedrock, is overlain by a regolith lag of liberated pebbles (Figure 18d) that give the terrain a smooth appearance in HiRISE 0.25 m/pixel images (Figure 18e). At meter to decameter scales, of course, the rock is not smooth but hummocky, reflecting prior erosion of the conglomerate. In addition to lags, aeolian sand accumulations (i.e., dunes and dune fields) can also bury and protect underlying rock from vertical erosion, at least as long as the sand remains in place.

3.5. Depositional Setting Mimicry

Some sedimentary rock exposures are readily identified on Mars because their geomorphic expression in the modern landscape mimics their original depositional setting or deposit morphology (e.g., Figure 1a–c). Such landforms exhibit properties that indicate they are rock (Section 3.1). This mimicry is a product of the combined attributes of deposit-scale depositional structures, sedimentary facies, and the impact of diagenesis as a function of differing compaction, paleo-porosity, and paleo-permeability.

3.5.1. Bedforms

Lithified bedforms expressed in the modern Martian landscape include those that resemble whole, intact aeolian dunes and dune fields in Melas Chasma (Figure 19a) [50,139,249] and near the yardangs of the Apollinaris Sulci (Figure 1a) [51,140,185]. Preservation of intact dunes and dune fields requires a process or processes that allow the material to become rock, while not disrupting the original dune morphology. On Earth, such preservation sometimes occurs through inundation by lava [171,250]. As no lavas are apparent at the Melas Chasma and Apollinaris Sulci locations, the possibilities for preservation of these dune forms include rapid burial by aeolian dust or tephra to a depth at which the wind could no longer access and mobilize the sand, or through stabilization of the dunes (e.g., wetting, freezing, cementation), followed by—or not—slower, gentle air-fall deposition of dust or fine tephra over some period of time.

Features interpreted by some investigators to be smaller, lithified aeolian bedforms, such as “transverse aeolian ridges” [252,253], have also been identified (Figure 19b) [191]. Additionally, a rock stratum (or stratal package)—emergent from beneath subsequent strata—in Terby crater exhibits large “ripples” of wavelengths of ~90 m, crest-to-crest (Figure 19c). Whether these bedforms were deposited subaerially or subaqueously is unresolved [50].

The subject of Martian bedform preservation observable at satellite and aerial photograph scales intersects with preservation of the opposite—spaces left behind by the removal of a bedform. Day and Catling [251] described barchan dune-shaped depressions in Hellas Planitia (Figure 19d) and Noctis Labyrinthus that they termed “ghost dunes” and interpreted as dune casts (alternatively, they can be called molds because they have negative relief). Formation of such ghost dune hollows required deposition of interdune material that surrounded each dune, followed by erosion that removed mobile sand. Based on an Earth analog in Idaho [254], some amount of aeolian sand is thought to have remained in the Martian ghost dune hollows and could provide a sandstone or meta-sandstone record for future exploration [251]. Day et al. [27] described a different situation, in which similar-oriented, lithified, transverse and crescentic forms in and near Barth crater are the preserved remnants of former aeolian dunes (Figure 19e). Additional dune-like lithic forms are seen in Figure 19b.

Transverse ridge patterns preserved in Martian rock can be signs of former bedforms (Figure 19b)—or they can be faux signals of depositional setting mimicry. There are two varieties of concern. One is a ridge pattern called “periodic bedrock ridges” [255], and the other is a pattern of regularly spaced, low ridges and shallow troughs called “washboard” [201]. Periodic bedrock ridges exhibit patterns similar to aeolian bedforms but are created by the combination of (1) substrate protection from aeolian erosion beneath megaripples or transverse aeolian ridges; (2) aeolian erosion of the substrate between the bedforms; and (3) the eventual removal of the bedforms to reveal ridges of substrate material (e.g., rock) that had been protected beneath them [255,256].

The other transverse ridge and trough pattern, sometimes referred to as a “washboard” texture [191,201,257], is a geomorphic expression observed on rock surfaces as a series of regularly spaced troughs and ridges of decameter-scale wavelength and meter- to decameter-scale amplitude (Figure 20). They are common in Gale crater [191,201] but also occur in other low latitude regions of Mars [257] Some of the rocks in Gale crater that exhibit a washboard pattern have been explored using the Curiosity rover and are found to be aeolian sandstones [188]. However, no linkage between aeolian sedimentary structure (Figure 20d) and the washboard pattern was found [188]. A concern regarding washboard patterns in Gale crater is whether they are an erosional expression of aeolian sedimentary structures or of differential erosion of fractured (jointed) rock. To date, this concern is unaddressed and it is possible that some washboard patterns do reflect aeolian depositional architectures, while others reflect deformational and diagenetic events that occurred during or after lithification.

3.5.2. Deltas and Alluvial Fans

Perhaps the most broadly recognized examples of depositional setting mimicry are the deltaic landforms in Eberswalde and Jezero craters (Figure 1b,c) [53,258]. Many other similar, sediment distributive landforms are evident on Mars in the form of both deltas and fans [13,19,259,260,261,262]. In some cases, the colleagues who investigated these did not address whether the materials are lithified, but some percentage of them most certainly are. For example, Figure 8d,e shows sedimentary fans that are emergent from beneath younger rock; an observation that implies that these examples are lithified. Figure 21 shows additional examples, including a boulder-producing fan crosscut by a graben.

3.5.3. River and Stream Sediment

The initial recognition of the lithified delta in Eberswalde crater included the observation of ridges that resemble former stream courses [53,263]. Similar landforms had been identified on Mars at a few locations before that time [264,265], and these were reminiscent of networked, inverted fluvial forms previously identified on Earth [266,267]. Colloquially referred to as “inverted channels” or “sinuous ridges” [201,235,268,269,270,271], in more recent years, the term “fluvial ridge” has been devised [176,177], especially because not all cases are a product of landscape inversion [13]. However, none of these terms adequately describes the landform [177], as some ridges are not sinuous and the other terms imply knowledge of process (inversion of topography) or material (fluvial sediment).

Ridges (and networks of ridges) that record the locations of former stream courses are quite common on Mars [26,177,272,273]. Figure 22a–d shows examples. On Earth, analogous ridges usually do not consist of 100% fluvial sediment [177]. In some cases, the bulk of the material might be an erosion-resistant igneous rock (e.g., lava) that filled the channel or valley of a former stream course. In other cases, much of the ridge volume might be occupied by rock that originally occurred beneath the stream (e.g., “ghost channels” of Clarke et al. [274]). Thus, a ridge that marks a former stream course might or might not indicate the presence of sedimentary rock.

Most Martian ridges that mark former stream courses are candidates for the presence of preserved fluvial sediment. Preservation of fluvial sediment in ridge form on Earth generally follows one of four pathways [177]:

• The sediment and/or channel-adjacent alluvium become cemented in the surface/near-subsurface environment and becomes a ridge when adjacent, unconsolidated or poorly-cemented alluvium is eroded away;
• The sediment is preserved beneath lava or tuff that filled the channel or valley;
• The sediment becomes lithified by deep burial diagenesis (compaction, cementation) and then returns to the surface (tectonism, erosion) and is subjected to differential erosion and exposure; or The sediment is never lithified, and coarse fluvial sediment grains (boulders, cobbles, pebbles) are left behind as an aeolian lag to protect the material from further removal.

Cementation in the surface/near-subsurface environment, also known as early diagenesis or eogenesis [126], typically involves formation of a ’crete (silcrete, gypcrete, calcrete, etc.), at least on Earth [177].

While ridges—marking the locations of former stream courses—composed of lava are common on Earth [177], very few examples have been (thus far) been identified on Mars [275]. The proportion of the abundant fluvial ridges on Mars [26,177,273] that are the products of deep burial sediment diagenesis, surface/near-surface cementation, or aeolian lag development is not known. A few cases on Mars provide definitive examples of fluvial ridges that are emergent from beneath younger rock (Figure 8b,c). An initial remote sensing study of the thermophysical properties of 20 such ridges suggested that those examined are more likely composed of lithified material than unconsolidated sediment protected by a lag [276].

Mars orbiter imaging has also shown that not all stream channel depositional setting mimicry takes the form of a ridge. The inner channels of Gediz Vallis and Sakarya Vallis, two canyons that cut the stratified Aeolis Mons in Gale crater, contain sediment—interpreted to be lithified [20,49,201]—still located within the original channels (Figure 22e,f).

3.5.4. Lakes and Ponds

No specific attribute(s), such as water surface area or depth, distinguish lakes from ponds. Indeed, “it is difficult to specify an upper size limit to what is called a pond and when it should be called a shallow lake” [277] (p. 716). On Earth, ponds are more common than lakes [278] and are generally regarded as “small” water bodies [279,280] that lie along a continuum between megalakes [281] and bodies of surface area ≤ 100 m² [282,283]. Similar to ponds, pools are distinguished from ponds because they dry out at episodic (“regularly, yearly or every few years,” [277] (p. 716)) intervals [284].

Just as Mars has ridges that mimic stream courses and provide records of past fluvial sediment transport and depositional settings, so too, can some raised landforms (e.g., mesas, hills, mounds) reflect the former presence of a small lake, pond, or pool. The least ambiguous examples (Figure 23) occur as local topographic highs at a terminus of a ridge or ridges that record a former stream course. As with fluvial ridges, the proportion of the volume of these landforms that consists of aqueously deposited sediment is currently unknown, as is the pathway (surface/near-surface cementation, lava, deep burial diagenesis, and aeolian lag formation) by which any given example followed before attaining its present configuration in the Martian landscape. Analogs to these Martian examples do exist on Earth [285,286,287,288,289]. They do not always mimic the original lake or pond shape, but some examples do [290,291].

3.5.5. Impact Crater Ejecta

Impact ejecta deposits consist of (1) material violently released from a given crater and (2) materials eroded from the substrate beneath the ejecta during emplacement [221,293]. Rock formed from impact melt can be locally present in an ejecta deposit [294], and some of the ejected rock fragments are shock-metamorphosed [295]. The bulk of an ejecta deposit, however, is essentially a clastic sediment (breccia) that can, like any other sediment, undergo diagenesis and lithification [169,170,296,297].

Some proximal impact crater ejecta deposits on Mars have been buried (or partly buried) and then partly exhumed from beneath sedimentary rock (e.g., Figure 8a). Their present-day geomorphic form mimics the original ejecta deposit but, because the material has been excavated from beneath subsequently emplaced rock, the ejecta deposit must be inferred to have been lithified.

3.5.6. Mass Movement Deposits

Like any other sediment, landslide deposits [298] can be subjected to burial and diagenesis or surface/near-surface diagenesis without burial. For example, in carbonate-clast-rich cases on Earth, near-surface diagenesis can (in some cases) cement and lithify the brecciated debris [299]. Lithified landslide deposits can mimic their original depositional form if they are exhumed, partly exhumed, or—like the aforementioned carbonate examples [299]—became lithified without burial. We have not yet definitively identified an example of a lithified landslide deposit on Mars, but Figure 24 shows a good candidate that is emergent from beneath subsequently-deposited, erosion-resistant (small impact crater-retaining), sedimentary fans that might also be lithified.

3.6. Distinction of Igneous from Sedimentary Rock

Distinction of igneous from sedimentary rock is one of the most challenging and vital concerns with regard to deciphering the geological history of Mars. Doing so is not always easy, whether using hand-lens-scale images from a rover (e.g., Figure 4), or using satellite and aerial images from Mars. Of course, some cases are not difficult at all; for example, there is little question that Olympus Mons is a volcanic edifice [300], and that repeated strata of repeated thickness and physical properties argue for environmental influences on deposition (i.e., sedimentation) [16,49]. However, with: (1) the recognition of erosion-resistant, dark-toned, mafic sedimentary rocks at the Curiosity field site [70,71,77,188]; (2) pre-Curiosity-landing speculation that some of the upper stratal packages of Aeolis Mons (the 5-km-high mountain in Gale crater) consist of Medusae Fossae Formation material [301,302], considered by some colleagues to be tuffs [303,304,305]; and (3) pre-Perseverance-landing uncertainty, as to whether the bedrock at or near the surface of the floor of Jezero crater would be igneous or sedimentary [31,306]. The concern about distinguishing igneous from sedimentary rocks has become more acute and has real-world consequences, such as landing site selection.

The main concerns for discrimination between igneous and sedimentary rocks center on the diagnostic value (or not) of columnar jointing, wrinkle ridges, boulder production and retention, lobate rock unit margins, the nature and patterning of stratified outcrops, the challenge to distinguish tuffs, and, most importantly, rock outcrop context. In addition to these, first, we briefly depart from the imaging theme to discuss remote spectroscopic detection of rock composition. In particular, while some sedimentary rock occurrences might be dominated by infrared spectral signatures indicative of sulfates, clay minerals, hematite, chlorides, or carbonates, e.g., [29,307,308,309,310], others have compositions similar or equivalent to igneous or impact melt rocks and, therefore, can potentially be misinterpreted.

3.6.1. Remote Sensing of Composition

By itself, information regarding rock composition determined using satellite or aerial remote sensing instruments does not permit distinction between igneous (or impact melt) and sedimentary rock on Mars. Most important to this point is the detection by near- or thermal-infrared remote sensing of igneous rock-forming minerals (e.g., olivine, pyroxene, plagioclase), or a spectrum interpretable as an igneous rock type (e.g., basalt, andesite). Observations made using the instruments onboard Curiosity in Gale crater have, for example, shown that Martian sandstones can be basaltic [71,73,77], and even mudstones can contain considerable, detrital feldspar and pyroxene [141,214]—and, in some cases, olivine [311]. The investigations using orbiter observations of intercrater “bedrock plains” described by Rogers et al. [72] and Cowart et al. [312], and the studies of intracrater dark-toned pyroxene- and olivine-bearing rock and light-toned, feldspar-bearing rock by Irwin et al. [23], underscore the points that sedimentary rocks with igneous rock-like compositions are present on Mars, may be identifiable via satellite or aerial remote sensing, and might even be common.

Ultimately, compositional information (e.g., interpreted from infrared spectra) is informative of composition but it does not, alone, inform us of the genesis of a given rock unit. Indeed, Rogers et al. [72], Cowart et al. [312], and Irwin et al. [23] relied on geomorphic observables to make their interpretations. Such observables included poor retention of small impact craters (indicative of poor erosion resistance), kharafish (aeolian erosion forms, carved into the rock, at a variety of scales), paucity of regolith (indicative of aeolian removal of sand-size and finer rock breakdown debris), a presence of aeolian lags (e.g., small bedforms) and, in some cases, association with fluvially-incised valleys or the inverted forms of former fluvial systems [23,72,312]. A lack of volcanic features such as flows, vents, and edifices was also an important factor in their interpretations [23,72,312].

3.6.2. Columnar Jointing

One possible discriminator between igneous and sedimentary rock observable in satellite and aerial images of Mars is columnar jointing. Columnar jointed rocks definitively occur on Mars [144]. In fact, columnar jointing provides excellent—but not entirely conclusive—evidence for magmatism. On Earth, they are most common in lavas e.g., [179,313] and shallow igneous intrusive bodies like the Devil’s Tower volcanic plug in Wyoming, USA [314]. They also occur in some impact melt sheets [315] and welded [316] and non-welded [317] tuffs. Although unusual, examples of columnar joints also occur in metasedimentary rocks such as thermally altered chalk [318], paleosols baked beneath lava flows [318], and, especially, contact-metamorphosed sandstones [319,320,321,322,323]. Thus, although an excellent signal of magmatic activity, the context of columnar jointed rocks on Mars should be considered to evaluate whether they formed in igneous rock, impact melt, or in contact-metamorphosed sedimentary rock.

3.6.3. Wrinkle Ridges

During the first several decades of spacecraft exploration of Mars, wrinkle ridges were considered a reasonably good signal of the presence of a lava plain [324,325,326]. In some research products, the presence of a wrinkle ridge in a Mariner 9 or Viking orbiter image was the only evidence available to interpret a given inter- or intra-crater plain as volcanic [327,328,329]. However, others cautioned that wrinkle ridges might not be diagnostic of lava plains [330,331] and that Earth analogs include examples occurring in sedimentary rock [331,332].

The term, wrinkle ridge (e.g., Figure 25a), refers to a landform assemblage that generally consists, in cross-section, of a narrow ridge superimposed on a broader arch [332]. They typically range from linear to arcuate as measured along their axes. Their name comes from the viewer impression that near-surface bedrock has been wrinkled or buckled. They can also be called maria ridges, or mare-type ridges, because they were first known from telescopic investigations of the lunar maria [333]. Wrinkle ridges are surface expressions of compressional deformation—thrust faulting and/or folding [334,335]—in competent (typically, erosion-resistant) rock. It is because of their similarity to the wrinkled ridges of the volcanic lunar maria, that some early Mars investigators took their presence to be a good sign of a lava plain.

In recent decades, there has been occasional acknowledgement that Martian wrinkle ridges are not definitive signals of lava plains [186,338,339], and that they can be expressed in competent, erosion-resistant, sedimentary rock. However, they have not been described to occur in the major light-toned layered rock outcrops on Mars, such as Meridiani Planum e.g., [168,340] and the Valles Marineris interior layered materials e.g., [4,8,24]. Figure 25b–g presents a candidate example of a wrinkle-ridged sedimentary rock occurrence in an intercrater basin in Arabia Terra. No known volcanic (e.g., vents, flows) or intrusive (e.g., dikes, plugs) landforms occur in the region. In Figure 25c,g, wrinkle-ridged basin floor material is expressed as erosion-resistant bedrock (bounded by escarpments) that exhibit areas of differing erosional expression (e.g., some surfaces with many parallel ridges, some without). Two ~80–600 m-wide ridges intersect the intra-basin plain and can be traced laterally for about 8–9 km, outward, and then they merge with the plain. These ridges (or, based on Zaki et al. [177], some percentage of their interior materials) are interpreted as occurrences of fluvial sediment; their intersection with the plain is suggestive that the basin floor material consists of fluvial and lacustrine sediment. As shown in Figure 25d–f, the rock exposed on the steeper slopes of the wrinkle ridge, and on the larger fluvial ridge, has been breaking down, producing rhombohedral rock slabs (perhaps by undermining of a recessive rock layer). Boulders are less angular and more reduced in size the further they occur down the slope, away from the ridge from which they came. These observations regarding the setting and rock breakdown features on the wrinkle ridge and fluvial ridge slopes are suggestive that the material deformed by the wrinkle ridge is sedimentary rock rather than lava.

3.6.4. Boulder Production and Retention

When Martian sedimentary rock occurrences were first described in 2000 by Malin and Edgett [49], the MOC-NA images available, at the time, had insufficient spatial resolution to detect boulders produced by the breakdown of the sedimentary rocks that they identified. Such images did, however, show boulders produced at escarpments cut through igneous rock [185,341]. Thus, it seemed possible, for a few years, that boulder production and retention observations provide helpful information to distinguish sedimentary from, at least, competent igneous rocks such as basaltic lavas. However, today, we know that Martian sedimentary rocks produce boulders, and that some of these rocks are competent enough to retain those boulders over some time period (Figure 5). Boulders produced from the breakdown of sedimentary rocks only became apparent in satellite images with the advent, in 2003, of higher spatial resolution via the MOC-NA cPROTO technique [50,232] and the start of the MRO HiRISE imaging campaign in October 2006, e.g., [10,11,15,21,25,210,342]. Today we know that Mars has sedimentary rocks that are so competent that they retain many small impact craters, and produce and retain boulders, as well as do (or similar to) Martian lavas (Section 3.4.1 and Section 3.4.2). The opposite is also likely true—that clastic rocks that poorly retain boulders produced from them (e.g., via impact cratering) could be sedimentary or they could be non-welded tuffs.

Thus, boulder production and retention are important observations to be made regarding rock physical properties, but an abundance of (or lack thereof) produced and retained boulders (or megaclasts, for that matter) is not a signal, by itself, for distinction of sedimentary from igneous rock. However, there is a possibility, yet unexplored, that boulders produced from lava might be larger or more abundant than those produced from competent sedimentary rocks; this notion comes from the fact that abundant, large boulders and megaclasts were indeed observable in volcanic regions of Mars at the lower spatial resolutions available in early MOC-NA images [185,341].

3.6.5. Lobate (or Arcuate) Rock Unit Margins

Lobate (or arcuate) margins of an erosion-resistant, planar to sub-planar bedrock unit are sometimes considered to be evidence for lava e.g., [343]. This interpretation is sometimes correct e.g., [343], but lobate margins, alone, are imperfect observables for the distinction of lava from sediment. Some of the erosion-resistant rock units in northern Gale crater provide illustrative examples (Figure 3 and Figure 26). Only one of these was investigated using the Curiosity rover (Figure 3). The other example (Figure 26) was too far away from the planned and executed rover traverse (Figure 26a). Before Curiosity landed, it would have been reasonable to interpret these crater-retaining, lobate-margined rock units as lavas. The example in Figure 3, addressed in Section 3.4.1, was investigated using the rover and turned out to be mafic sandstone [77,78]. The example in Figure 26 (see west side of Figure 26c) is partly confined to a low area within an impact structure. Extending Curiosity rover observations via HiRISE images into the nearby terrain not visited by the rover, Jacob [76] proposed that the topographically-confined example in Figure 26c,d could be either a lava flow or record the emplacement of clastic sediment via low-energy transport. Given the close proximity to the Peace Vallis fan (Figure 26a) [344]; the conglomerates, sandstones, and mudstones examined by Curiosity [79]; and the recognition that some sedimentary rocks can be resistant enough to erosion as to retain boulders and many small impact craters (Section 3.4.1 and Section 3.4.2), the sedimentary interpretation seems more likely. If it were lava, other geomorphic evidence for flow might be expected (e.g., pressure ridges; nearby vents). Such features—that help distinguish lava—might not always be present at a given location on Mars; it depends, too, on the erosional state (including damage by impact cratering) of the material.

3.6.6. Stratified Outcrops

Another challenge to the distinction between igneous and sedimentary rocks in satellite and aerial images of Mars centers on stratification, particularly for cases in which the only exposure of the rock is in troughs and escarpments or crater walls and central uplifts. While sediments exhibit stratification, so too can accumulations of tuff and lava [178,179], even on the Moon [182,346]. However—given current knowledge that Martian sedimentary rocks can be dark-toned, erosion-resistant, and even have mafic compositions e.g., [71]—visual similarity [347] between layered rock outcrops on Mars and stratified volcanic rocks on Earth or the Moon, is insufficient information to declare a Martian case to be igneous.

Törmänen et al. [348] explored this concern by comparing layers exposed in escarpments cut through definitively volcanic rock (i.e., exposed on volcanoes) versus layers in examples of sedimentary rock on Mars. They examined only a limited set of examples, however. They noted that lava emplacement is generally episodic, local, and confined to volcanic regions (i.e., in which vents, edifices, and flows are common). Their survey suggested that stratified escarpments in igneous rock (generally, lava; they did not address tuff) tend to be very similar to each other from place to place, with slopes exhibiting spur-and-gully morphologies in their upper (approximately) one-third, and talus deposits covering their lower two-thirds. Figure 27a shows an example similar to what they [348] found, except that the proportion of stratified to talus-covered slopes is ~25% to ~75%. They found that layers, in lava accumulations, are best exposed in the spurs, and appear “very rough” and “highly serrated” (Figure 27b). Layered lavas were also described as “highly disrupted and consist of rounded and plate-like blocks” of meters to tens of meters size (e.g., Figure 27d). Over short distances (meters to kilometers), the layering in stratified lavas can have variable thickness and can be non-parallel. Törmänen et al. [348] did not comment on the fact that finding good slope exposures of definitive igneous rock is actually challenging to accomplish; for example, in Tharsis and Elysium, slopes are commonly mantled with sufficient debris, and aeolian sediment, that the rocks are obscured. We also note that a spur-and-gully morphology is not always observed in volcanic rock exposures (Figure 27c), and that much work remains to be done to consider how such exposures compare with sedimentary rocks on similar slopes that have undergone similar erosional processes.

The limited examples of sedimentary rock exposures examined by Törmänen et al. [348] suggested that sedimentary rocks are more likely, than igneous, to exhibit rhythmic layering, cliff-and-bench slope expressions, “smooth” and “gently winding” slopes, and lack talus deposits of comparable extent to those observed at volcanic sites, such as in Figure 27. Regarding rhythmic layering, Annex and Lewis [28] similarly emphasized that packages of repeated stratal patterns, or repeated thickness and physical properties (i.e., rhythmic or cyclic strata), are more likely to be sedimentary than to be tuffs because “volcanic events are stochastic and generally not forced by orbital variation or other cyclic drivers”.

It is important to understand that tuffs, sediments, and regolith can become interbedded with lavas, and that these clastic materials may not be distinguishable on the basis of composition (e.g., presence of olivine, pyroxene, plagioclase) or tone (e.g., dark gray sandstone versus dark gray tuff versus dark gray lava). Beyer and McEwen [349] considered MOC-NA images of stratified rock exposed in the walls of Coprates Chasma, and identified an interbedding of apparent “strong” (erosion-resistant) and “weak” (erosion-recessive) layers (Figure 28). The erosion-resistant strata were seen as boulder producers and interpreted as likely being lava. The more recessive strata were considered more likely to consist of tuff, sediment, paleoregolith, thin lava flows, or some combination of these. The weaker layers were found to be more abundant, occurring in packages that are hundreds of meters thick, interbedded with resistant layers that have (or occur in packages of) tens of meters thickness. Given the current understanding from Gale crater that some sedimentary rocks can form very competent layers (Figure 3), including plateau-capping rocks (Figure 2, Figure 5c and Figure 17) [136,188], it now seems possible that even the “strong” layers investigated by Beyer and McEwen [349] could just as easily be sedimentary, as igneous.

3.6.7. Distinction of Tuffs

Tuffs present a special challenge to the distinction of sedimentary from igneous rocks in satellite and aerial images of Mars. Here, we use the terms volcaniclastic and tuff in the sense proposed by White and Houghton [350], which focuses on clast genesis, a departure from earlier perspectives [351,352,353], which included epiclasts created by weathering of volcanic rocks (and contrasted with the term, plutoniclast e.g., [354,355], for grains created from intrusive rocks). The term, volcaniclastic, now applies only to particulate material produced by or resulting from volcanic eruptions [350,356]. Primary volcaniclastic deposits can include coexistent, co-located facies of reprocessed material—i.e., volcaniclasts deposited by a non-volcanic process such as wind or wave action that occurred simultaneously with an eruptive event [357]. In contrast, secondary volcaniclastic deposits are, essentially, sediments [350]. These sediments form from reworking of unconsolidated primary volcaniclastic deposits after some period of storage [350]; they are not considered igneous and are classified using the same terminology as other clastic sediments and sedimentary rocks [350]. Further complicating the matter is the fact that primary tephra deposits can undergo diagenesis including, compaction and cementation e.g., [358,359,360], just like sediment.

Observations made on the Martian surface at the Spirit field site in Gusev crater, and the Opportunity site in Meridiani Planum, illustrate the challenges of distinguishing tuffs from sediments. At these locations, the effort to discriminate sedimentary rock from tuff relied on rover cameras, rather than aerial and satellite images, yet, whether all examples identified and discussed are actually tuffs remains uncertain. The very best case that Martian tuffs exist comes from the Home Plate plateau in Gusev crater [361,362], but even the strata exposed at Home Plate might include a volume of sand-sized clasts that were reworked by wind to form a sediment [362]. Home Plate was considered to be a tuff, in part, because of an apparent bomb and bomb sag [361], although the “bomb” is only ~4 cm across. Vesicular stones identified at the Spirit site were called scoriaceous but were generally considered to be impact-produced fragments of vesicular basalt, rather than products of explosive volcanism [363]. Other clastic rocks observed in Gusev, composed of sand-size and smaller particles, drape over paleo-landform surfaces on Husband Hill [187,364]. Similar clastic rocks occur in Meridiani Planum on the ridges of the west side of the Endeavour impact structure [365,366]. Spirit and Opportunity team members interpreted these rocks as either airfall volcaniclastics or airfall impactites; but they could not distinguish between these hypotheses [187,367,368,369,370]. Given the challenges for the distinction of primary volcaniclastic deposits at the ground level, even using 31 µm/pixel images [368,370], trying to distinguish them in satellite and aerial data should be expected to be even more difficult.

Explosive volcanism is thought to have been common on Mars because degassing of volatile-bearing magma during ascent to the planet’s low-pressure atmosphere should be quite vigorous [371]. However, primary volcaniclastic deposits and tuffs are very difficult to identify using remote observations of Mars [372]. Those that have been identified are not all universally agreed upon, but one thing they have in common is proximity to an eruptive vent; this is the case whether considering the proposed very youthful tephra deposits among the Cerberus Fossae [373], the thermophysical observations considered suggestive of tephra on volcanoes such as Alba Mons [374] and Tyrrhena Patera [375], or individual cones or rings bearing a summit depression [376,377,378,379].

Some investigators have considered, too, the possibility that fine, Martian tephra has been deposited far from a source vent following aeolian suspension [234,236,380]. The most prominent and persistent interpretation of such materials is that of the yardang-forming rocks of the Medusae Fossae Formation [303,305,380,381,382], which we discuss in Section 4.2.3.

One of the more informative candidate identifications of tuff on Mars, made using orbiter data, comes from Kremer et al. [237] and Mandon et al. [383]. Kremer et al. [237] provided a good model as to how to attempt distinction between airfall tuff, pyroclastic flows, lava flows, and sedimentary rock. Both investigation teams studied stratified, olivine-bearing (some of which is also carbonate-bearing) rock exposures located along the western and southwestern margins of the ancient Isidis impact basin. The airfall tuff interpretation is based on several attributes of the material: (1) It is said to consist of rock of nearly uniform thickness that drapes over preceding topography (Figure 29) of relief, in some places, exceeding 100 m. (2) Layers within the material (where identifiable) exhibit parallelism that also conforms to the superposed paleo-topography (Figure 29). (3) The material is interpreted to be clastic, largely because of evidence for erodibility in the form of yardangs, poor impact crater retention, and a model-dependent, infrared observation-based, estimate of olivine grain sizes of ~1 mm [384]. (4) It exhibits no deformational features—at HiRISE 0.25 m/pixel scale—that would indicate flow (e.g., pressure ridges). (5) It occurs over a great areal extent, appears to have a uniform composition over that extent, and retains the above characteristics over its extent. Finally, (6) it occurs within a regional volcanic context, in this case the circum-Isidis extensional faults that were conduits for Syrtis Major volcanism [385]. The combination features—especially the large areal extent, draping relations, association with a region of known volcanism, and evidence that the material is likely clastic—make these olivine-bearing rocks a strong, although not conclusive, candidate for Martian tuffs.

The apparent overall deposit thickness, the repeated thickness of individual strata (or stratal packages), and the uniform erosional expression of these strata in the example in Figure 29, however, could also be interpreted as signs of sediment draped over the paleo-topography. Although the mafic or ultramafic composition and draping of paleo-topography make the materials in the Nili Fossae region described by Kremer et al. [237] an attractive candidate tuff, we note that Stimson formation outcrops in Gale crater (e.g., Figure 17 and Figure 20b–d) exhibit near-uniform thickness, near-uniform basaltic composition, and drape across hundreds of meters of topography [243], yet are cross-bedded aeolian sandstones [136,188]. The cross-bedding is not observable in HiRISE images, and Day and Catling [251] provided a reminder that we should not expect aeolian cross-bedding to always be visible at the highest spatial scale obtainable by HiRISE. Where the Stimson formation and the rocks described by Kremer et al. [237] differ, besides the details of their composition, is that the Stimson formation rocks do not occur over a large, regional extent, and do not occur within a volcanic (or plausibly volcanic) context.

3.6.8. Importance of Context

Ultimately, what matters most for distinction between sedimentary and igneous rocks on Mars is context [115,186,312,386]. For example, pyroclastic deposits generally thin with distance from their source vents, and flows (whether lava or pyroclastic) move down-slope from a vent or fissure. The presence of flows is not always diagnostic of volcanism; they could be mudflows [387] and might even be mudflows that came from a vent [388]. Another form of context is the regional setting. A flow within an impact crater, Isil, in the region of circum-Hellas basin volcanism [385] is more likely to be a lava flow (Figure 30a) than is a flow that is far from a known volcanic center, such as the proposed debris flows associated with the impact event that formed Hale crater (Figure 30b) [389].

Unknown is whether Mars has very ancient igneous landforms that are not recognized because they were completely buried, or were so battered by impact cratering, or so eroded in the early Martian climate, as to have retained no characteristics recognizable in satellite and aerial images of the planet’s surface. Known Martian igneous rocks are most commonly associated with evidence for magma conduits to the surface—for example, the edifices in Tharsis and Elysium [300], the extensional faults of the Cerberus Fossae and Tempe Terra regions [224,390,391], and impact-induced conduits such as those that are semi-circumferential to the Hellas and Isidis basins [385,392]. Commonly acknowledged volcanic landforms are largely confined to specific regions, such as Tharsis, Amazonis, Elysium, Syria/Sinai, Solis Planum, and the circum-Isidis and circum-Hellas regions. Other volcanoes might exit outside these regions; Xiao et al. [393] inventoried 75 landforms in the heavily cratered terrain of Mars that various investigators have interpreted as possible ancient volcanoes, but whether they actually are volcanoes is less certain. Interpretation of stratified rocks in or very near these regions should trigger caution as to whether the materials in question are igneous or sedimentary. Of course, the two rock types can coexist—for example, consider the conduits, edifices, and flows of the San Francisco volcanic field in northern Arizona, USA, which were cut through and superimposed upon Colorado Plateau sedimentary rocks [394].

Context is relatively easy to discern in regions of Mars where volcanic landforms are indisputably present. One seeks vents, edifices, and flows that emanated from them. One also seeks firm evidence of excavated dikes, sills, and plugs, e.g., [395]. However, it is important to recognize that dark-toned (or dust-covered), erosion-resistant material exposed at the present-day surface of an intercrater plain or a partly-filled impact structure—where no unambiguous flow or vent morphologies are present—is not necessarily volcanic, even if it has infrared spectra interpretable as mafic, and even if it is deformed by a wrinkle ridge.

3.7. Deformation

Deformation of rock is a product of stress, a force applied per unit area. Strain is the change in size, shape, or volume of the rock that results from stress. As stress and strain increase, deformation goes from elastic to ductile until the rock breaks (fractures). Evidence of deformation of Martian bedrock visible at satellite and aerial image scales includes folds, fractures, and faults. Joints are fractures at which displacement parallel to the surface of the rupture is imperceptible. Joint sets consist of multiple parallel, evenly spaced joints, and joint systems consist of intersecting joint sets. Mars exhibits abundant examples of bedrock deformation at sub-millimeter to multi-kilometer scales, e.g., [8,25,145,332,335,396,397,398,399,400,401,402,403,404,405,406]. These deformations are products of crustal tectonism, impact cratering, and smaller, more localized stressors such as thermal cycling, overburden removal, hydraulic fracturing, and gravity (e.g., fracturing of undermined rock, leading toward mass movement).

Faults, fractures, joint systems, deformation bands, and possibly folds have been observed in satellite images of Martian materials interpreted to be sedimentary rocks, e.g., [3,4,8,25,145,332,397,399,400,401,406,407,408,409,410]. The concern, here, is whether any deformational forms are diagnostic signals—or at least encouraging signs—of sedimentary rock. Section 3.6.2 considered columnar jointing, which usually provides a strong signal of magmatism and a likely (but not unique) signal that a rock is igneous. Section 3.6.3 considered “wrinkle ridges,” which have sometimes been cited as the only evidence for interpreting that a given rock occurrence is igneous [326,329].

Regarding the use of deformation features for the distinction of sedimentary rock occurrences on Mars, two observations, in particular, stand out. These are not diagnostic, but can be considered helpful. These are the presence of polygonally fractured rock for which the polygons exhibit similar sizes and patterns as known examples of sedimentary rock examined at rover field sites; and the presence of convoluted structures interpretable as post-deposition, soft sediment deformation. In the latter, deformation occurs before lithification.

3.7.1. Polygonal Fracture Patterns

Polygonal fracture patterns visible in orbiter images of Mars are not unique to sedimentary rocks and can occur in other materials including igneous rocks (Figure 31a) and middle- and high-latitude ice-rich regolith (Figure 31b), e.g., [411,412,413]. Polygonal fracture patterns occur at a wide range of scales, from centimeters to kilometers, and can form through a variety of processes. Relevant scales, materials, and processes are reviewed elsewhere [403,404,414].

Joint systems, also not unique to sedimentary rock occurrences e.g., [158,179,415], create some of the variety of polygonal fracture patterns observed in some sedimentary rocks on Mars (e.g., Figure 6e) [25]. In some cases, such as in Gale crater, polygonally-fractured rocks can become dislodged and displaced, forming boulders or megaclasts that either remain close to where they were produced (e.g., Figure 5d), or can move downslope if a slope is available.

Here we focus on one key aspect of polygonally fractured sedimentary rock. That is, for locations on Mars where the rock is known to be sedimentary—or a very well-reasoned case can be made that it is sedimentary—finding a similar polygonal fracture pattern at another location on Mars might be a signal that the other locality has rocks of similar physical properties as those of the known (or well-reasoned) sedimentary rock case. For example, Figure 32 shows that the pattern and scale of polygonal fractures in rock exposed in northern Gale crater is similar to the pattern and scale of polygons in the Vistula Valles region, south of Chryse Planitia. The polygonally fractured rocks in Gale crater (Figure 32b) were not visited by the Curiosity rover, but a well-reasoned (albeit inconclusive) inference is that the material is sandstone [404]. This interpretation is based on similarity to sandstones that the rover did visit—at its Yellowknife Bay field site [79]—and based on a facies model that the rocks are coeval with the deposition of Peace Vallis fan (Figure 26a) sediments [344]. The inference, here, is that the fractured rocks at Vistula Valles could also be sandstones or, at least, they may be clastic rocks of similar mechanical properties.

3.7.2. Soft Sediment Deformation

Soft sediment deformation cannot actually be used to identify an occurrence of sedimentary rock, because doing so entails circular reasoning. However, deformation can occur in sediments after deposition—before lithification—and, where present, can (at least) be used to strengthen the case that a material is sedimentary. Soft sediment deformation usually occurs during and soon after deposition (i.e., synsedimentary) while sediments are loosely packed and have plenty of intergranular fluids. Hence, some of the most common environments for these structures are deep water basins under the influence of tidal currents, river deltas, and shallow seas affected by storm currents [417]. The most common trigger for deformation is seismicity (including those induced by impact events); however, other triggering mechanisms are known, such as sediment loading, salt tectonics, submarine mass movement, tsunamis, storm waves, tidal activity, and karst dissolution and collapse [417]. Moreover, there are at least 120 different types of soft sediment deformation forms described on Earth, of which some of the common are convolute bedding, slump folds, flame structures, dish-and-pillar structures, load casts, and clastic injections [417].

As might be expected on a world that has a record of aqueous sedimentation, millimeter- to decimeter-scale soft sediment deformation examples have been interpreted to occur in rocks outcropping at the Opportunity [57,418] and Curiosity [419] rover sites. More surprising is the fact that much larger examples are considered to be visible in orbiter images, particularly in the Valles Marineris [407,420] and Capen [7], Terby [421] craters. Further, the prospect that mud volcanism may have occurred on Mars [422] underscores the likelihood that clastic, soft sediment deformation has occurred [423].

3.8. Sedimentary Rock Landforms

Recognition of sedimentary rock occurrences—and even identification of rock type—would be greatly assisted by a thorough understanding of the landforms that exposed, eroded Martian sedimentary rocks produce. Over the last few decades, studies of “sandstone landforms” and “sandstone landscapes” on Earth have proliferated [107,113,424,425,426,427], and the concepts have been extended to “conglomerate landforms” [109], as well. In addition, Danxia landforms in China include both sandstone and conglomerate variants [114]. These publications make the point that sedimentary rock types, when exposed at a planet’s surface, can exhibit common erosional expressions that relate to the physical properties and deformational states of rock at scales ranging from microscopic to entire depositional basins.

The recognition and study of Martian sedimentary rock landforms is in its infancy. The Martian landscape exhibits a wide variety of sedimentary rock landforms and erosional expressions for which rock type is currently unknown. Shaping sedimentary rock landforms on Mars is more than a matter of aeolian erosion in the modern, hyperarid, setting. Many of the landforms were initially shaped more than two or three billion years ago, when groundwater and surface runoff were more common—the fluvial canyons cut into the sedimentary rocks of Aeolis Mons in Gale crater [20,201,428] provide examples.

Connecting sedimentary rock landform expression to rock type on Mars is presently an imperfect prospect because not very many differing examples have been examined in field settings (i.e., rover missions). As illustrated in Section 3.5, some sedimentary rock landforms mimic their original clastic depositional setting. In some cases of depositional setting mimicry, rock type can be interpreted—or, at least, speculated upon—based on knowledge of Earth analog facies. For example, the lithified dune field near the Apollinaris Sulci (Figure 1a) is interpreted as sandstone [52], and the detailed geomorphology of the Eberswalde delta (Figure 1b) [1,15] can be used to infer which facies might contain conglomerates versus sandstones versus mudstones [429]. Here, we consider what is presently known about landforms created in clastic sedimentary rocks and chemical sediments (to the extent that they might exist on Mars).

3.8.1. Sandstone Landforms

We begin with sandstone landforms because they are the most well-studied cases on Earth. Sandstone exposures on our planet form a wide range of visually spectacular landscape features [107,112,113,114,427], including arches [430]. Counterparts of similar geomorphic expression have generally not yet been identified on Mars, but Martian sandstone outcrops do exhibit a range of morphologies, and it is likely that many candidate sandstones are yet to be identified. As shown in Section 3.5.1, depositional setting mimicry in the form of whole, lithified aeolian dunes [51,52,139,140,249], and “ghost dunes” [251], are excellent starting points, but such outcrop expressions are rare. Slightly less rare are bedrock outcrops that exhibit a “washboard” erosional pattern (Figure 20; Section 3.5.1). However, whether washboard patterns always indicate aeolian sandstone, or whether some cases result from deformational structure (e.g., fracture patterns), is not yet fully known (Section 3.5.1).

Most of the sandstones encountered by Curiosity in Gale crater are dark gray, well-cemented, and erosion-resistant (Figure 2, Figure 5c and Figure 17). A few of them, however, are recessive—these include sandstones beneath an overhang at the Cooperstown outcrop (Figure 33a,b) [21,431]; dipping beds between prominent, erosion-resistant strata (Figure 13b) in rock interpreted to be deltaic clinoforms at the Kimberley field site (Figure 33c,d) [80]; and an occurrence of reddish, very fine-grained sandstone bedded with mudstones the Karasburg member of the Murray formation (Figure 33e,f). The clinoform [80] bedding is prominently seen as a “striated unit” [80,247] in HiRISE images of the Kimberley field site in Gale crater (Figure 13a and Figure 18e).

The dark gray, erosion-resistant aeolian sandstones of the Stimson formation in Gale crater are expressed as plateau, mesa, and butte-capping rocks (Figure 5c, Figure 17 and Figure 20) [136,145,188,432]. They are underlain by recessive rocks, largely mudstones [80,81] of the Murray formation. Among the Murray buttes (e.g., Figure 5c), the sandstones occupy the bulk of each butte’s volume [136]. In addition to their capping rock expression, Stimson formation outcrops also exhibit washboard-patterned surfaces of alternating, parallel ridges, and swales (Figure 20). Other erosion-resistant sandstones in Gale form cratered plains (Figure 3), benches that overhang recessive rocks (Figure 33a), or decimeter- to meter-scale escarpments.

Another sandstone visited by a Mars rover, Opportunity, is the Burns formation in Meridiani Planum [57,58,203,204,418]. These largely aeolian and wet interdune facies sandstones form the light-toned bedrock beneath a regolith of dark gray, aeolian bedforms [67]. The plains-forming sandstone is considerably more erodible than the lavas explored by Spirit at Gusev crater [137], and the dark gray, erosion-resistant sandstones in Gale crater. The full lateral extent of the Burns formation is not known; however, rocks of similar properties extend across the Meridiani region [168,340,433] and are exposed in northern Meridiani as mesas, buttes, and pinnacles (Figure 5b).

Other sandstone landform configurations likely occur on Mars. Meter-scale cross-bedding observed in mountainside outcrops of the sulfate-bearing rock unit in Aeolis Mons, Gale crater, have been interpreted to indicate that some this material is aeolian sandstone [191,434]; and two additional (non-sulfate-bearing) meter-scale cross-bedded stratal packages that occur higher in Aeolis Mons stratigraphy have also been interpreted as aeolian sandstones [22,201,435]. Similar meter-scale cross-bedding features occur in three of the four light-toned, sulfate-bearing, layered rock exposures (two of which are mounds) in Juventae Chasma [25]. These features, too, are suggestive of aeolian sandstones [25]. Yardang-forming rock on Mars might also include sandstones, but these do not form exclusively in sandstones on Earth [107,156]. Polygonal fracture patterns can occur in some Martian sandstones but can be absent from others.

3.8.2. Conglomerate Landforms

Migoń [109] recently summarized current knowledge of conglomerate landforms on Earth. Perspectives on Martian conglomerate landforms is limited. Some of the depositional setting mimicry cases, such as alluvial fans and deltas, likely include conglomerates (Section 3.5.2 and Section 3.5.3). Presently, the most extensive experience with in-place Martian conglomerate comes from the Curiosity site in Gale crater. When the rover landed, its terminal descent engines (Figure 34a) exposed conglomerate bedrock (Figure 34b) [245,436]. As viewed in MARDI descent images and HiRISE orbiter images (Figure 34c,d), the terrain on which Curiosity landed was described as having relief elements that are “rugged” and “hummocky” [79]. This terrain retains impact structures (Figure 34d), but not with the abundance of the neighboring, cratered sandstone that Curiosity’s heat shield impacted (Figure 3 and Figure 34d). Prior to landing, no one had experience that could have been used to predict that these observations were a sign of a conglomeratic bedrock. In general, the conglomerate where Curiosity landed is modestly erosion-resistant and pebbles, liberated from the rock, litter the ground (Figure 18a). As described in Section 3.4.6, another example of conglomerate landform configuration in Gale comes from the Kimberley field site, where terrain formed from a conglomeratic bedrock simply appears “smooth” and “hummocky” in HiRISE 0.25 m/pixel images (Figure 18d,e).

3.8.3. Finestone Landforms

Finestones are clastic rocks that consist largely of grains smaller than sand size. The term is not used on Earth, but is preferred here to discuss rocks that encompass a range of clastic sedimentary rock types that consist of particles finer than sand: siltstones, mudstones (also known as shales), loessites, and duststones. Finestone is also a term that could be applied to very fine to extremely fine tuffs, where they cannot be distinguished from their sedimentary counterparts.

The terms siltstone, mudstone, loessite, and duststone are applicable to cases for which data are available to unambiguously determine the depositional conditions of the sediment. For example, to call a rock a mudstone requires knowledge—from sedimentary structures, facies, context, etc.—that the clasts were wet (i.e., mud) at the time of deposition. Identification of loessites requires knowledge that the materials consist of silt-sized grains deposited by wind [437,438]. The presence of loess or loessite does not require that the silt was produced by glacial processes [439], nor that they are composed of quartz. Duststone is a term coined specifically for Mars by Bridges and Muhs [440] that refers to knowledge that the material was deposited by settling out of suspension from the Martian atmosphere. Duststone is a term less restrictive than loessite with regard to particle size distribution, as long as the particles are generally of sizes transported in suspension in the Martian atmosphere. Particles that can travel in long-term aerial suspension [441] on Mars generally consist of silt- and clay-sized particles [442]. No loessite or duststone has been definitively identified at a rover field site on Mars.

A minor finestone facies was observed in the Burns formation at Meridiani Planum [443]. However, most of what is known about finestone landforms comes from examination of the mudstones of the >350-m-thick Murray formation in Gale crater [79,80,189,190,405]. In the highest spatial resolution satellite and aerial images acquired to date, layering in the Murray formation is not readily apparent (Figure 10a). Images acquired from the ground, however, show that the rocks are finely laminated at sub-millimeter to millimeter scales (Figure 10b,c). Polygonal fracturing and local displacement of fractured blocks—visible in HiRISE 0.25 m/pixel images—are common in some parts of the Murray formation (Figure 5d and Figure 33e). Aeolian erosional forms (kharafish) are generally not evident in the existing satellite and aerial image data set, but millimeter- and centimeter-scale forms caused by aeolian abrasion are common in rover camera views [165,444]. One possible exception regarding aeolian erosion at scales visible in satellite images is a suite of parallel ridges interpreted as periodic bedrock ridges [445], which would have formed in Murray formation rocks as saltating grains abraded inter-bedform substrates between former (now gone) aeolian megaripples [255,256].

Murray formation finestones generally have a recessive erosional expression [80,446]. Areas with abundant, erosion-resistant veins can locally increase resistance to erosion and influence outcrop appearance. Although generally recessive, finestones, such as the Murray formation, can occur in positive-relief landforms in cases where an erosion-resistant cap rock has broken down and its debris removed at a pace that exceeded that of the gradation of the finestone beneath it. Further, the Vera Rubin ridge literally stands out as an exception to the recessive nature of the Murray formation (Figure 16). The rocks exposed on Vera Rubin ridge experienced diagenetic changes that differ from most of the other Murray formation mudstones; these changes might have rendered the rock more resistant to erosion [190,218,219,447]. Alternatively, Vera Rubin ridge was previously protected from erosion by a capping rock or coarse debris; this possibility is suggested by the fact that a similar ridge, capped by an erosion-resistant material, occurs west of Vera Rubin ridge (Figure 16d). It is actually possible that both things are true; that the ridge was protected from erosion by a previous capping material and its resistance to erosion was enhanced by late diagenetic alterations that occurred beneath, or at the interface with, this capping material. Surfaces that appear to be “smooth” in HiRISE 0.25 m/pixel images occur in the recessive Glen Torridon valley immediately south of Vera Rubin ridge; rover images show (Figure 18b,c) that this smooth appearance results from a regolith that consists of abundant, pebble-sized mudstone fragments produced by the breakdown of the local bedrock [246].

3.8.4. Chemical Sediment Landforms

Earth has landforms and landscapes influenced by differential erosion of chemical sediment occurrences—and clastic materials cemented by chemical sediments—that formed in surface/near-surface environments in contact with meteoric waters [129]. For example, duricrusts can resist erosion and help leave former streams and ponds standing as inverted topography [177,448]. Whether similar chemical sediments occur on Mars is uncertain, but appears to be likely. The sulfate-bearing sandstones of the Burns formation [57] of Meridiani Planum were, for example, interpreted to consist (at least in part) of sand-sized aggregates of sulfate-bearing muds (containing detrital silicates) formed in ancient, evaporative playa settings [449]. In addition, an apparently stratified, clast-bearing, calcium sulfate pebble was found in a conglomerate boulder at the Curiosity rover site [61]. It was tentatively interpreted as a chemical sediment rock fragment [61]. However, the stratal package that the fragment came from is unknown.

Beyond the Burns formation in Meridiani Planum, and that one, mysterious pebble—and its less-well observed counterparts in the same boulder [61]—in Gale crater, the next best candidates for chemical sediments on Mars occur at the locations identified, using satellite-borne near-infrared or thermal-infrared remote sensing instruments, as having chlorides [310,450,451] or carbonates [29,384,452,453] present. It is unclear whether remote detections of sulfate-bearing rocks, beyond Meridiani Planum, are necessarily indicative of chemical sediments (e.g., evaporites) or diagenetic fluids.

Bridges et al. [454] provided a recent review regarding Martian carbonates. In general, most occurrences are not considered to be sediments [384,452,453,455], but might result from various rock alteration pathways (e.g., weathering, diagenesis, hydrothermal, serpentinization), particularly involving alteration of olivine. “Comanche Class” rocks studied using the Spirit rover in Gusev crater were interpreted as clastic [370] and to contain Mg-Fe carbonates [455]. These carbonates might have formed under hydrothermal conditions [455] or precipitated among silicate clasts in a lacustrine setting [369]. A promising candidate carbonate sedimentary rock occurrence lies at the crater wall/floor interface in western Jezero crater [29]. These Jezero crater examples occur in close proximity to fluvial and lacustrine landforms, including a remnant delta at the terminus of Neretva Vallis [29]. These carbonate-bearing rocks—whether a sediment or not—exhibit low surface relief, are light-toned, fractured, retain some (but not abundant) impact structures, and are described as having “blocky,” and “ridge-forming” morphologies (Figure 35a) [29,31]. Such descriptions are not likely diagnostic of carbonate chemical sediments; appropriate near- or thermal-infrared spectroscopy remains highly recommended to make such detections.

The first satellite-based chloride salt detections resulted from observation by the ODY THEMIS-IR, a multispectral thermal infrared imaging instrument that images Mars at ~100 m/pixel [310]. Additional detections have been made using the near-infrared hyperspectral capabilities of the ~18 m/pixel MRO Compact Reconnaissance Imaging Spectrometer for Mars (CRISM) instrument [451,456]. Hundreds of locales have been identified [450]. They generally occur throughout the heavily cratered terrain in local topographic lows with areal extents ranging 0.33 to 1300 km² [450]. In some cases, channels or fluvial ridges (inverted channels) converge on the low area [450,456,457,458], and in some of these cases, chlorides are detected in the channels or on the ridges [450]. The chlorides occur in association with silicic mineral fines (primary rock-forming minerals and/or clay minerals [310,450,459,460]). Colleagues have argued for and against the ideas that these chloride deposits record ancient groundwater upwellings, playas, or deeper lakes that lasted decades or more [310,450,456,457,458,459,460,461]; the convergence of fluvial channels and ridges at some of these locations argues strongly that some of them, at least, resulted from runoff of surface water (brines).

As with carbonates, we recommend that satellite or aerial near- or thermal-infrared spectroscopy is best employed to detect chloride sites. However, because there are hundreds of them, we can comment on some of their geomorphic elements in the context of chemical sediment landforms. As noted, they usually occur in localized topographic lows in inter- and intracrater settings in the Martian heavily cratered terrain [450]. The chloride occurrences are usually light-toned rocks; Figure 35b–d shows some of the variety of geomorphic expressions of chloride-bearing materials identified on Mars. Some occurrences are found in channel bottoms or compose sinuous ridges (Figure 35b) interpreted as topographically inverted channels [310,450]. Some occurrences include small buttes or hills capped by a more erosion-resistant material (Figure 35c) [450,456] some surfaces are ridged (Figure 35d) [450]; and some are emergent from beneath other rock strata (Figure 35c) [450]. Whether the materials are internally stratified (as best as can be determined in 0.25–6 m/pixel satellite images) has rarely been noted. About 75% of the occurrences are fractured meaning that about 25% are not [450]. Those that are fractured typically display intersecting, open (or aeolian sand-or regolith-filled) fractures that break-up the outcrop into polygonal patterns (Figure 35d) [450,459,461]. However, some polygonal patterns result from erosion-resistant ridges, rather than open fractures [450,459,461]. Polygonal patterns have received considerable attention with regard to interpretations of the desiccation of lake muds or playas [450,459,461], although little or no comparison with the plethora of other polygonally fractured sedimentary rock occurrences on Mars have been made.

4. Discussion

As the preceding section emphasizes, distinction of sedimentary rock occurrences in satellite and aerial images of Mars can be easy, in some cases, and very challenging in others. Presently unrecognized occurrences most certainly exist. In this section, we explore two remaining concerns: (1) the depth of burial at which Martian clastic sediment becomes rock, and (2) the global distribution of Martian sedimentary rocks. Both of these matters are unresolved and important to discuss because they point the way toward future work.

4.1. Clastic Sediment Diagenesis and Depth at which Sediment Becomes Rock

The settings in which clastic sediment lithification occur are very important to deciphering the geological history of Mars. Was a given sediment buried, compacted, cemented, and later returned to the surface as rock, or was it cemented in the surface/near-subsurface environment and never deeply buried? For example, was the delta in Eberswalde crater (Figure 1b) deeply (hundreds of meters to kilometers) buried and then exhumed? How deep was it buried? How much overburden had to be removed before the deltaic form was visible? Where did the overburden go? Alternatively, was the Eberswalde delta never buried, or only buried to a shallow depth, with its clasts cemented in the presence of meteoric waters by materials analogous to Earth’s ferricretes, silcretes, gypcretes, and calcretes?

4.1.1. Perspectives on Burial Diagenesis from Gale Crater

The Salsberry Peak sandstone is a dark gray, erosion-resistant, plateau-forming rock (Figure 2) that occurs within the Pahrump Hills member of the Murray formation [74,80,189]. The Pahrump Hills member outcrops on the lower north flank of Aeolis Mons, the 5-km-high mountain in Gale crater. A boulder produced by breakdown of the Salsberry Peak sandstone was accessible to the Curiosity rover (Figure 2). Named Kanosh, MAHLI images of the boulder surface (Figure 2c) show 130–500 µm-sized sand grains that were cut across by either the fracturing that produced the boulder, aeolian abrasion, or both. The sandstone is very well-cemented; pits occur where grains were removed during diagenesis or weathering, but intergranular pore space is not visible. Rocks coeval with this sandstone, and the other sandstones and mudstones of the Pahrump Hills member [80,189] project southward into lower Aeolis Mons. Above these rocks lies ~4.5 km of additional strata. Thus, the sediments of very well-cemented Salsberry Peak sandstone, and the Pahrump Hills mudstones, are known to have been buried to at least ~4.5 km depth.

In contrast, a stratal package of boulder-producing layered rock occurs at the summit of Aeolis Mons (Figure 36), ~4.5 km above the Salsberry Peak sandstone [16,48]. These rocks are exposed at an elevation ~2.2 km above the crater’s north rim, and lie ~800 m below the highest elevation on its south rim (Figure 36a). Because the layered materials at the summit of Aeolis Mons are lithified, does that mean that they were also, like the Salsberry Peak sandstone, once buried at several kilometers’ depth, so as to become lithified? Knowing whether rocks exposed at the summit of Aeolis Mons in Gale crater were once at some depth below ground would have profound implications for the geological and hydrological history of the region, particularly if all of Gale crater—and the areas around it—were mostly buried.

4.1.2. Burial and the Transition of Clastic Sediment to Rock

The scientific literature on clastic sediment diagenesis on Earth is vast, in part because of the value of this information (particularly regarding porosity and permeability) to the exploration for oil and gas e.g., [126,462,463,464,465]. Less common, however, are observations that show the transition of sediment to rock as a function of depth and diagenetic conditions. One such example comes from a submarine setting captured in a 1927-m-long core extracted off the east coast of South Island, New Zealand [466]. It showed fine-grained siliciclastic and calcareous marine sediments transitioning from uncompacted sediment to rock. In the first ~50 m below the ocean floor, the sediment only exhibited signs of early diagenesis. As depth increased, a transition occurred between ~50 and ~350 m toward sediment compaction. Compacted sediments began to appear at 347 m depth; porosity loss due to compaction became more pronounced below 1300 m, and stylolites (jagged pressure solution features; see example in Figure 37) began to appear around 1600 m depth [466]. Of course, in detail, burial diagenesis involves more than just the pressures imposed by burial. Studies of the sediments deposited in the Gulf of Mexico, for example, provide insight into the complexities regarding the roles of pressure, temperature (which increases with depth), fluid chemistry, and the porosity and permeability of layered sediment of varied compositions and particles sizes, e.g., [130,467].

4.1.3. Forming Clastic Sedimentary Rock with Little or no Burial

While most clastic sedimentary rocks on Earth were formed during deep burial diagenesis, some were not. This is an important consideration for Mars. A key example to contrast with deep burial on Earth is the quartz beach and dune sands of the Oligocene Fontainebleau grés (sandstone) in France [172,173,174,175]. The sands are overlain by limestones and their maximum burial depth was ~100 m. Although emergent from beneath younger rock, most of the Fontainebleau sands are uncemented, except in very tightly-cemented lenses 2–8 m thick located at various levels and sites within the sand body [172,173,174]. These lenses are so well-cemented that they have sometimes been called quartzite [172]. These lenses do not result from the temperature and pressure effects of deep burial diagenesis but, instead, from silicification associated with the influences of climate, geomorphology (e.g., stream and pond locations), groundwater levels, and the geochemistry of meteoric fluids [172,173,174].

Other examples of earthly clastic sedimentary rocks that form or begin to form without deep burial diagenesis include aeolian beach sands cemented by carbonates in the subaerial, littoral environment [469,470], aeolian gypsum sands that begin to cement in the presence of meteoric waters [128], and clastic sediments cemented by ’crete-forming (silcrete, ferricrete, calcrete, dolocrete, gypcrete) mineral or amorphous materials [129]. For example—and although further field investigations are needed to confirm many details—it appears that nearly 60% of the clastic sediments in Earth’s known fluvial ridges (inverted channels) were cemented in surface and near-subsurface environments [177]. Chemical precipitate sediments (e.g., evaporites, tufa, and travertine)—which can, in some cases, contain clastic sediments—also form rock in surface and near-subsurface aqueous settings e.g., [471,472].

4.1.4. Martian Clastic Sedimentary Rock Formed under Conditions of Little to no Burial?

Until analyses from new and on-going exploration at the Perseverance and Zhurong rover sites are reported in the scientific literature, Mars offers at least one additional, important data point regarding sediment diagenesis and depth of burial, the Burns formation studied via the Opportunity rover in Meridiani Planum. There, the rover science team’s interpretation was that sandstones of the Burns formation are largely aeolian dune sediments [57,203,204] with some water-lain (wet interdune) facies [57,58,418]. The wind-transported clasts were considered to be aggregates of basaltic, and perhaps clay-mineral-bearing [473], playa muds cemented by sulfates [449]. Subjected to multiple diagenetic episodes involving groundwater [57,449], the rocks might not have been buried more than ~80 m [474]. This depth assessment is based on the proposed erasure of Hesperian-aged impact craters [474]. Whether the Burns formation was more deeply buried at the time of lithification is not known, but the interpreted setting is consistent with the idea of a shallow or surface/near-surface diagenetic setting.

4.1.5. Clastic Sediment Diagenesis on Mars, Looking Forward

Martian clastic sediment diagenesis, the distinction of diagenetic from weathering influences on sediment composition, and the depths of burial involved, are at the frontier of Mars geoscience. To become lithified, clastic sediments must first be retained. Sediment deposited on a dry hilltop is less likely to be retained than sediment deposited in a wet basin setting. Earth instructs us, and—through Gale crater and Meridiani Planum—Mars has begun to inform us, that the planet likely has a continuum of clastic sediment burial depths and lithification conditions. We should not expect a single “depth of burial” that leads to clastic sediment lithification on Mars. A minimum depth of burial can be known for Martian sediments that are exposed on slopes for which overburden rock remains intact (e.g., the Salsberry Peak sandstone; Figure 2). However—and unlike the submarine core sample extracted near South Island, New Zealand [466]—these Martian burial depths do not tell us about the actual depth at which the sediment first became rock. In addition, identification of clastic sedimentary rocks lithified in the surface/near-subsurface environment—such as the nearly 60% of fluvial ridges on Earth [177]—will likely remain challenging for years to come, although colleagues are trying (and should continue to try) to identify such examples using relevant remote sensing techniques and understanding of Earth analogs [177,276].

4.2. Global Distribution of Martian Sedimentary Rocks

Early maps indicating where sedimentary rocks occur on Mars [48,49,50,115,475] have not been updated to reflect newer observations made over the last decade [476]. Indeed, based on the results presented here—and the general challenges of distinguishing sedimentary rocks from tuffs (Section 3.6.7) [72,312]—mapping the global distribution of sedimentary rocks exposed at the Martian surface is likely to be challenging and take many years to accomplish.

Here, we explore four key questions that, if addressed through new research, would greatly impact understanding of where Martian sedimentary rocks occur and the depth and breadth of the record of past environments available for study: (1) how much of the bedrock of the upper crust of Mars in its heavily cratered terrains consists of sedimentary rock? (2) Are there sedimentary rocks among the strata exposed in the walls of the Valles Marineris? (3) What percentage of the rocks composing the Medusae Fossae Formation [82] consist of sedimentary versus igneous facies? (4) Are there sedimentary rocks beneath the northern lowland plains of Mars?

4.2.1. Sedimentary Rocks of the Martian Heavily Cratered Terrain?

Mariner 4, 6, and 7 Mars flyby images acquired in the 1960s showed that some fraction of the Martian surface is heavily cratered, like the moon, but (as compared with the moon) many of the impact craters were shallowed by material deposited in them and by widening of their walls and lowering of their rims by erosion [477,478]. Even back then, some impact structures appeared to contain fill materials that had been subsequently eroded [479] (p. 19).

The interface between the Martian surface and atmosphere on early Mars would have been a fragment production factory—impact cratering would create rock fragments and solidified melt droplets [294], volcanism would include volcaniclast production [480], and weathering processes would have broken down exposed rock surfaces, altered primary minerals, and carried ions away in solution [481]. Weathering would also have extended into the subsurface in the presence of groundwater, as occurs in cavernous and fractured rock (even siliciclastic rock [240,427]) on Earth. Once produced, the clasts of early Mars would have been subjected to (and, in some cases, sorted by) transport and deposition in aeolian, fluvial, lacustrine, mass movement, and perhaps glacial [233,482] environments. Impact craters, intercrater lows, and tectonically-formed depressions, would have served as depositional basins for these sediments [48,49,324,326].

Large portions of the bedrock of the Martian heavily cratered terrains are likely obscured by regolith, mantled by aeolian dust [483], mantled with presently or previously ice-bearing regolith or dust [484,485], or buried beneath south polar layered deposits and ices [121,486,487]. However, some regions do exhibit exposures of intercrater and intracrater rock [12,72,312,488,489,490]. Among these, and in particular, are portions of western, central, eastern, and northeastern Arabia Terra.

Arabia Terra is an important gift to Mars science. Despite considerable aeolian dust cover [68,483,491], the landscape is one that has been eroded in a manner that reveals the nature of the upper crust of the Martian heavily cratered terrain. Western Arabia Terra was so deeply eroded that it retains few valley networks [492] and impact structures once filled and buried have been excavated and returned to the surface [493]. Fluvial ridges, also known as inverted channels, are common across Arabia Terra [26,292], further attesting to the erosion of this landscape. Erosion throughout Arabia Terra has revealed the layered nature not only of materials deposited in impact structures [17,18,28,49,251,410,494], but also in the intervening, intercrater terrain [9,12,48,68,493,495,496,497]. Some percentage of the stratified rock that composes the upper Martian crust in Arabia Terra is sedimentary (Figure 19e, Figure 22a, Figure 23a,d, and Figure 25b–f) e.g., [12,26,27,292,493], although full understanding of the percentage awaits future study. In general—and modulated by proximity to volcanic regions [312]—the stratified inter- and intra-crater rocks of Arabia Terra could be representative (although specific lithologies may vary) of what lies beneath the surface in all of the other, less deeply eroded, heavily cratered terrains of Mars [185].

4.2.2. Sedimentary Rocks Exposed in the Walls of the Valles Marineris?

Stretching ~4160 km across the southern low-latitudes of Mars, the walls, intra-chasm ridges, and plateaus of the Valles Marineris, and associated Hebes and Juventae chasms, expose (laterally) more than 27,400 km of pre-chasm, Martian upper crust to depths that exceed more than 8 km, in places. The surrounding terrain includes wrinkle-ridged plains, lavas and volcanoes, impact craters, filled and partly filled impact structures, and heavily cratered terrain [327,498,499,500,501,502]. As implied by the very first geologic maps of the region [498,503], where the uplands surrounding the Valles Marineris are not heavily-cratered, the stratigraphy argues that ancient, heavily-cratered bedrock lies beneath the present surface and has been cut and exposed in the chasm walls.

Orbiter images of the walls of the Valles Marineris and associated chasms show that they are stratified [347,349,504]. This is known despite the fact that vast portions of the Valles Marineris walls are covered with material that obscures the bedrock. Some of this obscuration is talus, landslide debris [505], alluvial fans [506], aeolian bedforms, light-toned aeolian dust, and dark-toned mantling deposits of uncertain origin (tephra falls? aeolian silt and sand?). Some sections exhibit packages of erosion-resistant “strong” layers and erosion-recessive “weaker” layers (Figure 28) [349]. Some of the chasm walls at the east end of the Valles Marineris exhibit stratal packaging of layered rocks overlying dark-toned, poorly-layered or structureless (“massive”) boulder-producing materials, which, themselves, overlie lighter-toned, fractured, structureless rock of proposed magmatic origin [504]. These lower rocks are cut by mafic dikes [507,508] and the country rock adjacent to them has been subjected to visible alteration by contact metamorphism [507].

The layered rocks exposed in the Valles Marineris walls are mostly dark-toned and often interpreted as lavas, sometimes with a caveat that some amount of regolith, tephra, impact ejecta, and sediment could be interbedded with these materials [347,509]. Except where Valles Marineris walls intersect and cut through lava flows expressed at the present Martian surface, the case for lava deeper in chasm wall stratigraphy seems to have been a “looks like” argument—that some of the outcrop expressions resemble those of escarpments cut into flood basalt strata on Earth [510]. More evidence would be helpful. For example, a key signal of the likelihood of lava or intrusive magma is columnar jointing (Section 3.6.2). However, no columnar joints, or column-shaped boulders shed from columnar-jointed outcrops, have yet been identified in the strata of the Valles Marineris and associated chasms [24,502]. Of course, columnar joints too small to be identified in HiRISE 0.25 m/pixel images could be present, and candidate columnar jointing has been observed in some of the dikes exposed in the Valles Marineris [508]. Others have suggested that some of the layered rocks in the Valles Marineris might be crystalized, layered plutons [511], and still others suggest some of the “weaker” strata might be tuffs [480].

The distinction of stratified (and structureless at observable scales) sedimentary rocks from lavas, tuffs, sills, buried impact melt rocks, and paleosols (paleo-regoliths) exposed in the walls of the Valles Marineris is not easy and, thus, no examples of dark- or intermediate-toned sedimentary rock have yet been definitively identified [504]. Given the likelihood that, based on observations of Arabia Terra (Section 4.2.1), the sedimentary rocks may be common in the heavily cratered regions of Mars, and that the walls of the Valles Marineris (and associated chasms) had to have cut through heavily cratered terrain bedrock, it seems plausible that some of the rock exposed in the chasm walls has to be sedimentary. Observations from the Curiosity rover site in Gale crater show that sedimentary rocks exposed in the Valles Marineris walls need not be light-toned, can be dark-toned, can even be mafic, and can be sufficiently resistant to erosion so as to resemble the erosion resistance of lavas, sills, and impact melt rocks. In that case, both the “weak” and “strong” layers of Beyer and McEwen [349] have the potential to be sedimentary.

In addition to the darker-toned wall rocks, there is the question of whether light-toned, stratified rocks are exposed in various places in the walls of the Valles Marineris, and whether these are records of very ancient sediments (or tuffs) deposited before the chasms opened up, or deposited in the chasms after they opened up, or were deposited in precursor basins at various stages as the chasms slowly began to open. Based on a survey of the scientific literature, the latter hypothesis is generally favored [10,512,513,514,515,516,517,518]. However, some investigators have concluded that there do exist examples of light-toned rock—generally considered to be sedimentary, but could in some cases be tuffs—emergent from within Valles Marineris chasm wall strata [24,25,49,185,407,519,520]. The most frequently accepted occurrence is at “mound A” in southwestern Juventae Chasma (Figure 38a) [24,25,407,519,520]. Additional candidates have been identified in the ridge that separates north-central Candor Chasma from southwestern Ophir Chasma (Figure 38b–d) [49,185], a feature in Ganges Chasma called “Ganges 5” [342], a light-toned rock exposure called “feature D” in Juventae Chasma [24,25], occurrences in the lower walls of northwest Candor Chasma [519], southern Melas Chasma [519], Hebes Chasma [519], southwestern Candor Chasma [185], the upper walls in a portion of Ganges Chasma [185], and on the Nectaris Montes in Coprates Chasma [185]. Light-toned, stratified rock in the Coprates Montes ridge in Coprates Chasma has been variously interpreted as an exposure of pre-chasm sedimentary rock [185] or as sediments deposited and lithified in a precursor depression that formed during a proposed multi-stage opening of Coprates Chasma [516]. A light-toned, wind-eroded, stratified rock occurrence “perched” [514] in an alcove above northwestern Coprates Chasma has also been variously interpreted as pre-chasm rock [185]—such as sediment filling a buried impact structure—or the eroded remnants of sediment deposited and lithified within a precursor depression during the period of the opening of the chasms [514,521]. In the former model, the alcove exists because of the pre-chasm presence of light-toned rock of relatively poor erosion resistance; as it was exposed by the extensional faulting and mass movement that opened the chasms, the more erosion-resistant overburden was undermined, broken-down, and removed by erosion, forming the alcove.

There has been a general resistance to acceptance of evidence that light-toned, stratified rocks outcrop at some locations in the walls of the Valles Marineris. That resistance discourages investigation of this model and leads to a bias, which assumes that all such occurrences must be erosional remnants of sediment or tephra deposited in the chasms after they opened up, or during successive episodes of their opening and widening [10,498,499,512,513,514,515,516,517,518,522]. The resistance might stem, at least in part, from a misinterpretation of the Malin and Edgett [49] model, which viewed the rocks exposed in the Valles Marineris walls as heterolithic in both a lateral and vertical sense, reflective of the heterolithological appearance of the deeply-eroded terrains of western Arabia Terra [9,12,18,493]. Imagine, for a moment, what western Arabia Terra looked like before the deep erosion that erased some—and topographically inverted other—valley networks [292,492], and exhumed filled and buried impact structures [493]. Now, imagine using some earthmoving equipment to cut a several-kilometers wide and several-kilometers deep trench across that western Arabia paleo-landscape. What would be seen in the walls of that trench? Evident would be the same lateral and vertical lithological diversity, including dark- and light-toned strata of differing resistance to erosion, which we see exposed planimetrically in the region today. Now, imagine what lies buried beneath the terrain cut by the Valles Marineris. Should the walls of the Valles Marineris, and associated chasms, not also reveal a similar lithological diversity that includes rocks of differing tone and resistance to erosion, as well as various filled and buried impact structures and former intercrater terrain?

The question of Valles Marineris (and associated chasms) wall rock lithologies remains unresolved. Many excellent observations have been documented [24,324,347,349,396,502,504,505,507,519,523], but work remains to be done to decipher whether it is true that intrinsically-light-toned rocks of relatively poor erosion resistance are exposed in various patches and locations in these walls, whether “weak” or even “strong” stratal packages include sedimentary rocks, whether columnar joints (other than in magmatic dikes [508]) occur in any of the rocks, and whether filled and buried impact structures and filled, former valleys or channels can be identified in the rocks.

4.2.3. Sedimentary Rocks in the Medusae Fossae Formation?

Like the walls of the Valles Marineris, the Medusae Fossae Formation (Figure 39) presents additional challenges to identifying sedimentary rock occurrences on Mars [48], in part because of four decades of interpretation and assumption that the material is tuff [82,303]. The Medusae Fossae Formation is a suite of yardang-forming [524] and butte-, mesa-, and fractured-plateau-forming [305] rock units that occur in a discontinuous zone that stretches along the equator between western Tharsis and southern Elysium Planitia (Figure 39a) [82]. The rocks are stratigraphically younger than, and partly straddle [327,328], the Martian topographic dichotomy boundary [525]. Most of these rocks occur on the northern, lowland side of the boundary. Where they occur south of the boundary, they unconformably overlie bedrock previously heavily cratered and cut by fluvial valleys. Outlier remnants occur both north and south of the main bodies of the unit, suggesting that it was once more extensive [526,527]. In the northern lowlands, Medusae Fossae Formation rocks in some areas overlie, and in others are embayed (and buried) by, lava flows [224,303,528,529,530,531]. Fluvial sediment and sedimentary fan deposits, exhibiting depositional setting mimicry, occur in terrain adjacent to or emergent from beneath Medusae Fossae Formation rocks in the south-central (e.g., Figure 39d) [301] and western (e.g., Figure 39e), e.g., [13,19,532,533] parts of the region.

In most areas, Medusae Fossae Formation rocks lie beneath a mantle of aeolian dust [483,536] and—in some areas—the rocks are covered by dust-coated or (less commonly) dust-free aeolian sand [528,537,538]. The materials retain some impact craters, but few enough that the rocks were considered for several decades to be Amazonian in age [301,327,328,539]. Later analyses, focused on impact crater counts and embayment or interbedding of Medusae Fossae Formation rocks with lavas, suggested that deposition might have begun during the Hesperian [302,528,540].

Because wind eroded some of the material to create yardangs (Figure 39b,c), and yardangs on Earth usually consist of clastic materials [156], Medusae Fossae Formation rocks have long been considered to be clastic [303,381]. The formation is not, however, monolithic [82]; in some analyses, these rocks have been subdivided into lower, middle, and upper stratigraphic members based on erosional expressions that are, reasonably, assumed to be connected to rock stratigraphic and physical properties [327,328]. High spatial resolution orbiter images show features that suggest these rocks can be further subdivided into more specific units and facies [304,305,528], including some that may be fluvial (e.g., Figure 39e) [13,19,70,235,541]. Whether the fluvial materials are properly part of the Medusae Fossae Formation, or a preceding formation, is debatable; certainly, some of the fluvial sediments, expressed as ridges, might have become ridges before they were buried beneath yardang-forming material (Figure 39e).

A vital question regarding the Medusae Fossae Formation is whether it exhibits a lithological diversity commensurate with the observed geomorphic diversity [304,305], and whether some percentage of the materials are sedimentary rock [48]. A variety of interpretations for the origin of the materials have been proposed over the last > 4 decades, including pyroclastic flows (ignimbrites) [303,305,381], airfall tephra [304,381], aeolian sediment [328,542], marine-rafted pumice [543], marine carbonate platforms [544], and paleo-polar dust and ice mixtures [545].

Interpretations that the Medusae Fossae Formation consists of primary volcaniclastics have generally dominated the discussion [82,303]. The proximity of the rocks to the volcanoes and lava flows of the Tharsis, Amazonis, Elysium, Apollinaris, and Cerberus Fossae regions provide the necessary volcanic context (Figure 39a). Further, models of possible tephra dispersal patterns [236,380] suggest that a relation between the volcanoes and the Medusae Fossae Formation is possible. However, the interpretation that the rocks consist of primary volcaniclastic material conflates two topics that should be considered separately: (1) the depositional processes and settings recorded by the rocks, and (2) the genesis of the clasts that compose the rocks. Interpretations regarding the relatively low density of the rocks [382,546,547] conflate a third topic: (3) the porosity of the rock, which is a product of the physical nature of the primary deposits and their diagenesis, including fracturing. Although the low density was interpreted by some [382,547] as an indicator of primary volcaniclastics, density is not a signal of clast genesis.

As noted, the presence of yardangs in the Medusae Fossae Formation is a signal that the materials are likely clastic [381,524], but interpreting the material as clastic says nothing of clast genesis. Aeolis Mons (Mt. Sharp) in Gale crater illustrates the point. A few investigators have interpreted the yardang-forming rocks—and the other rock units exposed at and just below the mountain summit—to be erosional outliers of the Medusae Fossae Formation [301,302,527]. More recently, investigations of these rocks show that most of them, except the uppermost rock unit (Figure 36) [16,48], exhibit meter-scale cross bedding in rover and orbiter images, suggesting the materials were deposited as aeolian dunes [22,191,434,435]. Other depositional structures observed in some of these rocks might actually be fluvial [434]. Whether the yardang-forming rocks—and other rock units high in the stratigraphy on Aeolis Mons—are truly outliers of the Medusae Fossae Formation, or not, they certainly display depositional structures reflective of sedimentary environments and exhibit no evidence as to clast genesis or any relation to volcanism.

Unless the rocks reflective of fluvial sedimentation in the central and western areas of the Medusae Fossae Formation (e.g., Figure 39d,e) [13,19], and the upper stratigraphic units of Aeolis Mons in Gale crater, are part of the Medusae Fossae Formation, then no one has yet (or otherwise) performed the necessary, detailed analyses of depositional and deformational structures in the majority of the formation. Such studies are now possible with 100% orbiter image coverage at 0.25–6 m/pixel. If some rock units in the formation are ignimbrites [303] then such analyses would help pinpoint vent locations, a key concern regarding the ignimbrite hypothesis [548]. Such mapping would help discern the various depositional facies recorded in the Medusae Fossae Formation. In lieu of such detailed work, Figure 40 shows some features observed in various parts of the formation that appear to be depositional structures; these include an example of an erosional unconformity (Figure 40e). In the end, the degree to which the Medusae Fossae Formation includes sedimentary rocks, versus tuffs, remains to be deciphered, but it is possible that considerable amounts of sedimentary rocks are present.

4.2.4. Sedimentary Rocks Beneath the Northern Plains?

Evidence for sedimentation in the northern lowlands of Mars—especially from the circum-Chryse outflow channels [500]—has been observed and discussed ever since receipt of images from the orbiter missions of the 1970s [551,552,553,554]. Whether the northern lowlands also served as an oceanic sedimentary basin has, likewise, been discussed and debated for several decades [555,556,557,558,559]. Landforms interpreted as mud volcanoes [422,423,560] attest to the possibility of fluid expulsion events over large areas of the proposed sedimentary deposits [561,562]; various flow-like forms and flow fronts suggest intrabasinal sediment transport [552,553,555,563]; and intersecting troughs forming “giant” polygon patterns are suggestive, to some, of compaction of fine-grained sediment resulting from loss of subsurface material such as water or ice [564,565].

The presence of sedimentary deposits in some (or all?) parts of the northern lowlands [566] leads to the question of whether sedimentary rocks are present beneath these landscapes. Burial diagenesis can involve compaction, alone, or compaction and cementation [126]. In addition, the other element of clastic sediment lithification—time—has been plenty; for example, several billion years have elapsed since the circum-Chryse outflow channel sediments were deposited [566]. Salvatore and Christensen [562,567] proposed that fine-grained, light-toned, stratified sediments deposited in the Chryse and Acidalia basins by outflow channel flood events can be identified in the walls of many impact craters in the region. These layered materials exhibit depositional structures suggestive of sedimentation, such as lenses and pinch-outs, and infrared spectra suggest they contain hydrated minerals [562,567]. Geomorphic evidence for fluid expulsion associated with these sediments (e.g., giant polygons and proposed mud volcanoes [561,562]) strengthen the case that compacted sediments are present [567]. Elsewhere in the northern lowlands, Gross et al. [568] identified stratified rocks containing phyllosilicates, exposed in northern plains impact structures; these, they suggested, could be northern plains sediments or could be much older, sub-basin (e.g., buried, heavily cratered terrain) sediments.

As currently understood, we should expect sedimentary rocks to occur beneath the northern lowland plains of Mars, and Salvatore and Christensen [567] have identified excellent candidate examples. Exposures of northern plains sedimentary rocks are likely to be limited to impact structures, boulders and megaclasts in impact ejecta, and perhaps areas where regolith has been removed by aeolian or fluvial erosion, exposing bedrock, if such surfaces exist. There is also value in recalling the surprisingly “hard,” vesiculated, salt-bearing mudstones produced in near-vacuum conditions in a laboratory forty years ago [569]; these were considered among the various candidate explanations for the vesiculated rocks at the northern lowland Viking 2 lander site [570]. On today’s Mars, we also recognize that ice occurs beneath some of the northern lowland plains [571,572,573]. The role of ice in terms of contribution or inhibition of Martian sediment diagenesis is unknown.

4.2.5. Additional Challenges

Most recognized Martian sedimentary rock occurrences are found at equatorial to middle latitudes, but this might be a bias with regard to their visibility. There is no reason to expect, for example, that heavily cratered terrain at middle and high latitudes have fewer sedimentary rock occurrences. For instance, the bedrock beneath most of the troughs of the mid-latitude fretted terrain (sensu [398]) north of Arabia Terra is obscured by materials that have been variously interpreted as ice-bearing or as previously ice-bearing “lineated valley fill” e.g., [574,575,576,577], but there are areas in the Nilosyrtis Mensae fretted terrain in which such landforms are absent and stratified, likely sedimentary rocks, are observed (Figure 41a) [578]. Sedimentary rocks also occur in a similar setting, in the opposite hemisphere, among the mountains and valleys on the west side of the Argyre impact structure (Figure 41b).

Additional challenges to global identification (and mapping) of sedimentary rock occurrences at and very near the surface of Mars include: (1) middle and high latitude mantling deposits that obscure bedrock [484,485]; (2) the aforementioned blanket of aeolian dust over much of Arabia Terra [483,491]; (3) seasonal frost and the fogs and dust storms that accompany seasonal polar cap retreat and limit frost-free viewing opportunities during parts of late winter and spring at middle to high latitudes; and (4) low image contrast at 50° to 70° south latitude during frost-free periods, interpreted to be a result of dust coatings on surfaces, and subsequent vigorous, spaghetti-like streaking of these surfaces by dust devils, that challenge bedrock visibility and identification.

5. Concluding Statements

Geology is the pursuit in which all of the sciences meet to tell the history of worlds. It is also the pathway for identification and management of rock, mineral, and geochemical resources. Both of these notions apply to efforts to locate sedimentary rocks on Mars, Venus, Titan, and other bodies, now and in the future. Two decades after the seminal publication regarding the observation of sedimentary rocks on Mars [49], identification of occurrences exposed at the planet’s surface remains challenging and incomplete. In some cases, care must be taken to develop certainty that a feature observed is truly observed, and that one’s hidden assumptions are identified and addressed. Overreliance on previously published interpretations is rarely wise; one should continuously question one’s certitude, and investigators should clearly separate observations from interpretations (and label speculations as such).

We have been careful in this contribution to not use the words, criteria, and criterion, in the context of identifying Martian sedimentary rock occurrences. There is not—and should not be—a checklist of criteria that guarantee identification (indeed, use of a checklist would be an oversimplification). We hope that what we have done here is to provide some guidance and made it clear that the light-toned, stratified occurrences, the less-common intermediate-toned and “dark mesa-forming” materials [49], and the examples of depositional setting mimicry identified before 2012 [48,69], were a sub-set of a greater whole. We hope we have made it clear that:

• Some Martian sedimentary rock occurrences have physical properties that permit them to resist erosion and retain small impact craters in a manner similar to lava plains;
• Tone (e.g., light-toned rocks) can be a property intrinsic to the rock, intrinsic to a coating or weathering rind, or can be a function of rock surface roughness and aeolian sediment (light-toned, dust; dark-toned, sand) accumulation on its skyward-facing surface;
• Pebble-sized clasts can be liberated from conglomerates and breccias, or produced from the breakdown of erosion-resistant, fine-grained rocks, and these pebbles can form lags that give the terrain a “smooth” appearance at image spatial resolutions of (at least) 0.25–6 m/pixel;
• Martian sedimentary rocks include examples that are dark-toned and have primary igneous rock compositions (e.g., containing glass, plagioclase, pyroxene, olivine), and remote detection of such rock compositions (e.g., basaltic) should not automatically be interpreted as tuffs or lavas; and
• Distinction of sedimentary from igneous rock is not always easy, even when viewed by landed payload cameras at sub-millimeter scales; for cases in which making the distinction is not possible, it is important to acknowledge the observational reason(s) as to why.

A key concern going forward is to separate questions of clast genesis from the processes of clast deposition and depositional setting. The > 4 decades of dialog regarding the Medusae Fossae Formation, for example, teaches us to worry less about clast genesis before being concerned with the geological record of depositional conditions. Over the 1979–1982 period, cases were made on the basis of erosional expression that the Medusae Fossae Formation rocks must be clastic. Most of the discussion, since then, focused on whether the large volume of clasts was produced by explosive volcanism. However, the proposed clasts cannot be seen without in situ investigation, and, thus, their genesis cannot actually be known. There is presently a higher value in encouraging renewed study of the large CTX, HiRISE, and CaSSIS data sets to consider (or reconsider) the depositional (and diagenetic, if present) structures in the Medusae Fossae Formation (e.g., Figure 40) and interpret the depositional facies present.

Although obscured by regolith, aeolian dust, or mid-latitude landforms and materials in some regions, sedimentary rocks might be quite common throughout the Martian heavily cratered terrain—in impact structures, in the spaces between them, and in buried examples of the same. Sedimentary rocks could also have formed from sediments deposited in the northern lowland plains of Mars. With impact events, volcanism, weathering, and erosion making the interface between the early Martian lithosphere, hydrosphere, cryosphere, and atmosphere a clast production factory, and with the wider variety of sedimentary rock landforms now recognized, it is reasonable to anticipate that the upper crust of Mars has a considerably more voluminous sedimentary rock record than previously realized. This record awaits further exploration.