1. An Interpretation of the Geology of Gale Crater & Mount Sharp, with Implications for the History & Habitability of Mars

    by David E. Palmer 
              In this essay I present a new model (“artesian hydrant”) for the geology of the Gale Crater/Mount Sharp complex, and (re)introduce a lacustrine model (lake-bed sedimentation) for the strata comprising the Lower Formation of Mt. Sharp, which is something that has recently fallen out of favor (the aeolian or “SWEET” model for the development of Mt. Sharp is in vogue, but I find numerous faults with that model, and believe I address all criticisms of the lacustrine model).  I also argue for a near-surface aquifer and geologically-recent groundwater outflow in the complex, and I anticipate my model to be tested when the Curiosity rover reaches the base of Mt. Sharp in late 2014 or early 2015.  Finally, I add my thoughts on two tangental but equally controversial subjects, that of the true colors of Mars and its sky, and the potentials and limitations the planet has to offer any future human colonization effort (I am pro-colony, but argue that any terraforming effort would be disastrous).


              Gale Crater is a large (96 mile diameter) impact crater, located just south of the Martian equator, on the edge of the crustal dichotomy between the Southern Highlands and the Northern Lowlands (Borealis Basin), and believed to have formed 3.6 to 3.8 billion years ago (at the tail end of the “Late Heavy Bombardment,” and approximately at the end of the ”Noachian” era on Mars and the start of the ”Hesparian” era).  The bolide that produced it is believed to have been several miles across, comparable to the one that produced the Chicxulub impact on Earth at the end of the Cretaceous period, and the reason that Gale is smaller than the 150-mile-diameter Chicxulub crater, is because relative orbital and impact velocities at the higher orbit of Mars are generally lower.  What makes Gale Crater different from most on Mars, and a target for the current Curiosity rover, is that it has a mountain in the center (Mount Sharp, a.k.a. Aeolis Mons) that consists largely of a very thick (3.4-mile-high) stratigraphic section of sediments that gives a record of deep history on Mars, a section over twice as thick as the famous Grand Canyon stratigraphic section.  It is therefore hoped that, since the Grand Canyon section records nearly two billion years of Earth’s history, the section represented in Mt Sharp may record much of the history of Mars, including the earlier wet period, the transition to the later dry regime, and the biological implications of these diverse environments.  And indeed, the lowest sediments of the column (to around one 1.5 miles above the crater floor in the area to be explored by the Curiosity rover) show unmistakable spectroscopic signatures of water-deposited or water-modified minerals, namely clay and sulfate-rich layers, which then transition (through an erosional unconformity) to what appear to be cross-bedded aeolian sediments (wind-deposited dunes) with minimal signs of involvement with water.
              On face value, the structure of Gale crater and Mt. Sharp suggests that this crater was originally filled with sediments (from at least two depositional periods, separated by an erosional period), with the cumulative layers ultimately reaching a height greater than that of the crater rim, and with subsequent climate change and a switch to a strongly erosional environment, such that most of the sediments occupying the crater were eroded away (primarily, if not entirely, by the wind), with the notable exception of the surviving Mt. Sharp section in the center.  And so Gale could be seen as a paleo-crater, an ancient or fossil crater that has been entombed and then re-exposed. 
              The biggest problem with this developmental scenario is that it begs the question of why only the outlying sediments filling Gale Crater would have eroded away, leaving the central area comparatively unmolested.  And as a corollary, there is the question of why the outlying portion of the crater fill, in the northern half of Gale Crater, was eroded down to a level approaching the original crater floor, about one mile below the grade of the surrounding landscape (whereas erosive processes normally don’t create large basins, but instead tend to fill in pre-existing ones).  
              The riddle of how Mt. Sharp came to be, is considered the single greatest mystery involving Gale Crater, with an additional and broader question being where the water in Gale (and other ancient Martian locales) came from, and it is hoped that the Curiosity rover can shed light on their solution through a geomorphological and geochemical survey of the crater and mountain.
              However, I will suggest in this paper that the evolution of Gale Crater and Mt. Sharp can (in broad outline) be understood with the information we already have on hand, and by considering familiar terrestrial geologic phenomena that undoubtedly have (or had) analogues on Mars, such as artesian aquifers and the water-related lithification of sediments (and such an interpretation can provide pointers for Curiosity’s investigations, with predictions that are field-testable by the current rover and/or by future missions).  And I would suggest that the reason that no direct Earth analogue to Gale Crater exists, is that the paucity of non-degraded impact craters on Earth simply does not allow enough opportunities for a Gale and Sharp-like complex to have been created by the chance confluence of the necessary geologic forces and conditions.
              Below is a NASA 3-D topographic map of Gale Crater, with north being up (and with brown representing highest elevation, blue lowest, and green medium elevation).  The originally planned landing ellipse of Curiosity is also shown (the actual landing site was on target, being just to the lower right of the center of the ellipse):  
               As can be seen in the above figure, the highest portion of Mt. Sharp actually consists of northern and southern “twin peaks,” separated by about 14 miles, with the southern (and highest) peak believed to be the original “rebound” central peak, surviving from the time of Gale’s creation, whereas the northern (broader and slightly lower) peak is an eroded stack of later sediments, referred to as the “upper mound.”  Below the upper mound we have the western lobe and eastern lobe of the “lower mound,” and below those in turn we have the “southern tableland” between the central peak and southern crater rim, and still lower the “northern floor” of Gale Crater, with its “moat” just north of the mound of Mt. Sharp and representing the lowest point in the Gale/Mt. Sharp complex.  
              I have identified these features below, as well as the “Southwest channel,” which I will be discussing later:
              Below is an approximate true-color orbital image of the northwest quadrant of Gale Crater and Mt. Sharp, which is the area of primary interest to us in this paper, and which includes the area to be explored by the Curiosity rover.  This image was obtained by combining a medium-resolution black-and-white photo with a low-resolution color photo (courtesy NASA/Emily Lakdawalla et al.):

              If Gale Crater was at one point filled to over-topping with sediments, why would the mass of these sediments in the crater center preferentially survive the later erosional regime?  Although the central peak consists of a much harder and more erosion-resistant rock (basalt) than the later-emplaced sediments, and would have potentially provided something of a wind break against erosion, an examination of the above topographic map demonstrates that this cannot have been an important factor, as the thick and laterally extended columns of sediment persist to a radius of 30 miles on most sides the central peak (plus the high point of the sedimentary deposits is 14 miles removed from the central peak).
              Rather, I would hypothesize that the central peak played a far different and far more important role in the genesis of Gale’s geology: namely, that it served as a natural artesian hydrant for a deep aquifer, and in so doing served to lithify (though fluidic action and emplacement of cementing minerals) the otherwise loose, friable aeolian sediments in the center of the crater, thus making them far more resistant to wind erosion.  In a sense then, Mt. Sharp can be seen as consisting of a giant concretion, which has been weathered out of the surrounding matrix!
              The way that the central peak could have served as an artesian hydrant can be understood by considering a cross-section of Gale and the underlying crust (dimensional relationships are for illustration purposes and are not to scale):   

              At a depth of several thousand feet, we may expect to encounter the primordial basaltic crust of Mars, the original crust that solidified out of the magma ocean present just after the planet formed (which is more or less a global feature of the planet, since the crust has not been re-worked by plate tectonics as it has on Earth), and this primordial crust is in turn covered by later sediments and lava flows.  And if this primordial crust is heavily fractured and brecciated, as it is expected to be as a result of the heavy asteroidal and cometary bombardment occurring on the early Mars, this unit can be expected to have been a highly effective aquifer, both trapping liquid water and allowing it to percolate horizontally and vertically (and is very likely an active aquifer to this day, as we will see below).
              When the impact that formed Gale occurred, it punched deep into this basaltic layer, and shattered basaltic debris and lavas upwelled to fill the crater and form the “rebound” central peak, and this central peak would be expected to be both heavily fractured and to commute directly with the deep aquifer (and since the impactor penetrated what was presumably an active aquifer, all affected rock would be expected to be all the more fractured and fragmented due to hydraulic fracturing and steam explosions).  And we may expect the bowl of the crater to have been filled with a lake of ”impact melt,” which subsequently solidified into a solid mass of basalt [1].  And in addition, as a consequence of the surface weathering that was occurring on the early “wet” Mars, it is reasonable to assume that there would be a mantle of clay and other fine-grained sediments deposited in and around Gale Crater, especially in the newly-created sump of the crater floor.  Additionally, the subsurface hydrothermal activity that would be expected to last for thousands to hundreds of thousands of years after an impact of this size, can be expected to produce clays that would tend to fill in cracks that developed in the solidifying impact melt of the crater floor, given that the chemical weathering processes generating clays cause the source material to swell as water is introduced into the reorganized crystalline structure.  And collectively, this marbled layer-cake of igneous and sedimentary rocks and alteration products, would have served as an effective ”aquatard” that trapped the underground water, and allowed such water to reach the surface only in a natural hydrant that projected above those layers, such as a crater central peak.
              And we may also expect that the water in that aquifer would have been under tremendous pressure, more than adequate to force it out of any available conduit in the Gale complex, since Gale is in a near-perfect position to have a hydraulic “head” created by the surrounding topography.  Not only is it one of the lowest locations on Mars, but it is surrounded on three sides by far higher terrain, which would serve to pressurize the aquifer and produce a very effective head at the location of Gale: it lies on the northern edge of the vast Southern Highlands, but it is also located in a southward indentation of the northern margin of those Highlands, and bordered directly on the west by a northward extension of the Highlands (Nepenthes Mensae), plus there is the Elysium volcanic bulge to the north (see topographic map below, courtesy NASA/USGS, where red represents higher and blue lower elevations, with Gale the blue-floored crater centered in the map section).  So if the planet were to have an active aquifer, Gale would be in one of the best positions to tap into it, plus as one of the larger impact craters in its region (and thus involving one of the deeper holes dug into the crust), it may be one of the few spots in its region where a conduit was opened to that deep aquifer (or had an opening that survived being subsequently covered by volcanics and sediments, thanks to Gale’s exceptionally tall central peak jutting over 3 miles high).  In addition, the regional aquifer may have been pressurized by deep magmatic activity associated with the Elsium volcanic complex, which is centered 1,000 miles away but whose outskirts are only 500 miles away.
              As an anecdote, as a young man (and amateur geologist) I was fascinated by something that happened near my home near Auburn, California, in the Sierra Nevada foothills: in a narrow valley with steep hills to the north and south and the massif of Sugar Pine mountain to the east (with the Sierra Nevada mountain range still further east), someone drilled a well for a new home, and it turned out to be an artesian well, where in the middle of the hot and drought-sticken summer of California, water was shooting out of the ground by about 6 feet, and kept doing so for the months to come, until it was capped.  And that topography was very similar in principle to that surrounding Gale Crater, and demonstrates how Mt. Sharp could also have supported its own artesian outflow!  Which also exemplifies the parallels between the geology of Earth and Mars, and how you can use the one to gain insights and understanding of the other, which is something I find quite fascinating.
              And so we can visualize a regime in the early history of Gale Crater where water was flowing out of fissures in the central peak, even shooting upwards as towering fountains, and filling the crater floor with a lake, which is consistent with the evidence of water-deposited and water-altered minerals in the lower sedimentary levels of Mt. Sharp.  And it was also likely during this the time that the Northern Lowlands of Mars hosted an ocean (Oceanius Borealis, or as I prefer to call it, “The Great Northern Ocean”), and rainfall (or snow melt) was carving river channel and valley networks in the equatorial region of Mars.  And in fact, at their highest level, the Borealis Ocean and the Gale Crater lake may have merged to form a continuous waterway, overtopping the northern crater rim at approximately the minus 6,000-foot level (below Mars datum, or the global average elevation), with Gale serving as an inlet or bay of that ocean (so technically, some of the sediments in Mt. Sharp may be marine, not just lacustrine).  Or alternatively, the northern crater rim may have been high enough in that time to serve as an effective dike, and equalization of the water levels between Gale and the nearby ocean was accomplished by groundwater movement.
              Because of the prevailing colder temperatures of Mars as compared to the Earth, even in an earlier phase of the planet when a much denser atmosphere provided more of a greenhouse effect, the Gale Crater lake may have been frozen over much of the time, with only sporadic periods of warmer climate featuring open water.  In addition, because of the lower gravity of Mars as compared to the Earth, we can visualize the draping of the central peak with massive, elaborate, yet delicate ice-palace-like structures of the sort that could never be supported on Earth.  
              But eventually, probably around 3.5 billion years ago (not long after the start of the “Hesperian” era on Mars), major changes occurred on the planet: the Late Heavy Bombardment had ended, the intensity of Mars’ volcanic activity was gradually tapering off, and the volcano-related and impact-related recharging of the atmosphere with water and greenhouse gases had decreased, such that the climate of the planet dried out, the Martian rains forever ended, and the atmosphere became far thinner,  a time which has been referred to as “The Great Desiccation Event,” after which surface water became the exception rather than the rule.  And at this point, the rate of evaporation from this lake would have exceeded the input, and the lake would have dried up, at about the same time that the same fate befell the ocean that presumably occupied the Northern Lowlands.  Subsequently (after a period of wind erosion), Gale Crater and its lacustrine sediments were covered over by aeolian sediments, which apparently grew to a thickness that exceeded the height of the crater rim, and which may have consisted largely of volcanic ash, possibly with the addition of loess (wind-blown silt) brought in from the newly-dried-up oceanic basin of the Northern Lowlands.
              However, the buried central peak would have still had fissures from which water was flowing into an underground network of channels, and this water would have soaked the sediments in the vicinity of the central peak, both carrying minerals that were in solution in the water of the subsurface aquifer, and minerals dissolved from the above-ground sediments as the water percolated through them (plus, new mineral species, such as clays, may have been produced in situ by the water-saturated conditions, if those conditions lasted long enough).  And when these sediments ultimately dried out, the minerals would have precipitated between the sand and silt grains, cementing them together and lithifying them, and making them (at least relatively) resistant to erosion: 



              In addition to clays, candidates for cementing agents would include iron oxides and silica, which often serve this function on Earth and which are also known to be present on Mars, but even more likely on Mars would be salts such as chlorides and sulfates, as they are not only known to be very abundant on the planet, but are stable (are solids) under the predominately dry conditions there (unlike typical Earthly conditions).  Calcite and other carbonates are also common agents on Earth, but are comparatively rare on Mars as a result of the prevailing acidic conditions, and it is not known whether they would be available in sufficient quantities in the area of Gale Crater to have served a role (only trace quantities have been detected as yet).  And in this hypothetical scenario, the higher-level aeolian sediments in the center of Gale would not have been deeply soaked, or for a long enough time, to have altered their mineralogy in any obvious way that would be spectroscopically detectable from orbit (such as creating large amounts of clay), but only long enough to dissolve and precipitate cementing minerals.
              And when the regional Martian climate eventually changed (for unknown reasons) to a strongly erosional rather than depositional regime, the aeolian sediments in the outlying areas of the crater (consisting of friable ash falls and loose silt) that had never been soaked with water (and thus never lithified), would have been easily blown away (in a time I will refer to as “The Great Erosive Event”, likely around the middle of the Hesparian era), leaving the central mound of lithified material encasing the central peak, the complex now known as Mt. Sharp.  And when the aqueous-deposited “Lower Formation” layers in the outlying portions of the crater no longer had the aeolian-deposited “Upper Formation” layers capping and protecting them, they also fell prey to erosion, and in the northern half of Gale Crater’ floor (the topographically lower half, which has been christened Aeolis Palus), the sedimentary layers may have been stripped down almost as far as the basaltic floor of the crater, apparently stabilizing by the beginning of the “Amazonian” era (around 3.0 billion years ago), and exposing a surface subsequently subject to only minor erosional and depositional episodes, as an alluvial plain in what was then a predominately dry regime, much like the closed basins in the desert terrain of the American West and Southwest.
              This cycle of aeolian deposition and subsequent exhumation and dissection has also left its record in the “fretted terrain” of the Aeolis Mensae region to the immediate east and north of Gale Crater, where deep (apparently aeolian) sediments were deposited on top of the ancient cratered and valley-network-incised terrain of the Noachian era, only to be deeply carved and largely stripped away by subsequent wind erosion (and quite possiblly glacial activity) during the Hesparian era [2, 3].  This erosion also greatly lowered the northern crater rim of Gale and obliterated Gale’s impact ejecta originally lying to the north, east, and west of the crater, whereas a wide blanket of ejecta survives to the south and southeast (and is important in dating the age of Gale, on the basis of crater counts).
              Below (courtesy of Wikipedia) is the most widely-accepted geologic time-scale for Mars, with numbers representing millions of years (later I will provide evidence that this scale needs to be somewhat adjusted):
              Below, as photographed by the Curiosity rover, we see an archetypal slice of the Mt. Sharp of today, the product of billions of years of interplay between natural forces of construction and destruction:



              An important piece of evidence supporting my hypothesis of a natural artesian hydrant in Mt. Sharp, and a feature that would otherwise appear to be inexplicable, is the fact that there are three large outflow channels (each one to two miles across at their widest), coming off the west side of Mt. Sharp, all from the same source region, and obviously caused by running water on the basis of their structure (although I believe these channels to be of immense significance, and in fact the key to understanding the geology and history of Gale Crater and Mt. Sharp, they are often glossed over in discussions of the Gale/Mt. Sharp complex, perhaps because no one has had a clue as to what they are doing there).  All three channels originate from a large, very roughly amphitheater-shaped section that has been cut out of the western side of Mt. Sharp’s upper mound by erosional collapse, a section approximately 15 miles long on a north-south axis and 7 miles wide on an east-west axis and around 1 mile high, with the floor or “stage” area on the west, and the “stands” represented by a steep slope on the back face that exposes striking “cliff-bench” or terrace-like groups of layers in cross-section (although this complex is not ideally modeled after an amphitheater, in that the back is convex rather than concave, plus the stage area is downward-sloping towards the northwest, just not as sharply sloping as the back face). 
              There is a north-trending channel, which I will dub “The Northern Channel” (see photo below, taken from the Mars Reconnaissance Orbiter, looking straight down and facing North, centered at 4 degrees 50 minutes south latitude, 137 degrees 26 minutes east longitude, and spanning an elevation drop of around 6,000 feet from the lower right to upper left corners of the picture, over a linear distance of approximately 7 miles, as the crow flies).  In its upper reaches, the channel has a downward gradient of around 12 degrees, and what look like small boulders in the lower channel, are in fact concretionary-like outcrops or buttes the size of warehouses:


              Here is a perspective facing east:

              Below is a simulated oblique view just above and looking down the channel, courtesy of Google Mars:



              There is also an initially west-trending channel which then curves to the north (this is the largest and deepest of the three and has been dubbed by NASA as “The Grand Canyon of Gale Crater,” and which I will refer to as “The Grand Canyon GC” for short). 
              Additionally there is a southwest-trending fluvial erosive feature, and I will refer to it as the “Southern Channel,” even though it might be questioned whether it merits the “channel” designation, as it is far shallower and far less developed than the other two (even though it is of similar width), apparently not experiencing flows as erosive as the other two, likely because the flow rate was lower and/or didn’t last for as long a time, and also because the slope gradient along most of its length is far lower than for the other two channels.  It also flows onto the tableland in the southern part of the Mt. Sharp/Gale Crater complex, rather than directly onto the crater floor, and it may have been significantly modified by wind erosion since it formed, but I mention it here because its presence is important for understanding the dynamics of what was occurring on Mt. Sharp (as I will explain below). 
              The image below shows all three channels (with main features identified), and the next one is a color rendition of the complex:    


              All three channels have their headwaters on the “floor” or “stage” of the amphitheater-like complex, at an elevation of approximately 7,000 to 8,000 feet below the summit of Mt. Sharp (or around 5,000-6,000 feet below Mars datum), and follow paths towards the lower reaches of Gale Crater (the northern floor of which is up to 10,000 feet below the “stage” of the amphitheater complex).  And this is a classic example of the sort of landform we have (on both Earth and Mars) as a result of mass wasting caused by groundwater sapping, where a spring or springs in the side of a hill will undercut it at the point where they exit, causing the material on top to collapse and erode away.
              Below I have drawn in arrows showing the directions of aqueous flow, which serves to show how that flow was derived from a common source area:    

              As an additional anecdotal aside, I am an avid amateur paleontologist, and one of my favorite collecting localities is the sea cliffs around Newport in Oregon, and at my most favored spot (which I have christened “Pecten Beach,” actually the south end of what is officially named Mullock Beach), there is material constantly sliding down the cliff due to that cliff being undercut by sapping (a small stream exits at the base of the slide zone), and while it is causing homes to be lost on the marine terrace above the cliff, with the top edge of the cliff rapidly receding inland, it also provides excellent fossil collecting with lots of “colluvium” to sort through on the beach below.  And being familiar with that locale and how it looks from both the surface and from orbit with Google Earth, it was an “AHA!” moment when I looked at the Mars Reconnaissance photos of Mt. Sharp, it being obvious that the same sort of process had occurred there, again an example of how the same geologic principles apply to both Earth and Mars.
              The unconformable (erosion-produced) boundary between the Upper and Lower formations occurs at the “stage” area of the complex (which also serves as the dividing line between the upper mound and lower mound), and besides sapping from groundwater discharge, another major factor favoring the collapse of the west side of Mt. Sharp was the angular unconformity separating the Upper and Lower formations, the surface of which dips at an angle of 12 degrees to the northwest, and which is the plane down which the block of the Upper Formation collapsed (through either sliding or flowing).  And a characteristic of the Lower Formation is that, apparently being lake-deposited, it consists of fine-grained materials, such as silts and clays, that should be relatively impermeable to the percolation of water, whereas coarser-grained, less-lithified aeolian sediments (such as those immediately adjacent, in the Upper Formation) are more permeable to water.  And so I would surmise that springs were emerging from the side of the mountain at the bottom of the Upper Formation, since water would have had a relatively easy time percolating down through the aeolian sediments, but would tend to be blocked by the Lower Formation strata (and/or the water would have more easily worked its way through the weak zone between the two formations), and these springs would have begun life as minor seepage and gradually grew to a trickle and then to a stream, whereupon a relentless course towards a major collapse was set into motion. 
              The plane between the (relatively permeable) Upper Formation and the (relatively impermeable) Lower Formation would be expected to undergo a pooling of water across its surface, and not only would this water (combined with the fine-grained sediments at the top of the Lower Formation) have facilitated movement by serving as a lubricant, but the 12-degree downward slope would have acted to power such movement, with the net result being that gravity would have been quite adequate to bring about the collapse of the Upper Formation on the western side of the mountain (either as a initially unitary block that slid, or as a possibly multi-stage flow, depending on its state of saturation, its internal cohesiveness, the level of friction at the contact between the formations, etc).  So on the basis of the composition and geomorphology of Mt. Sharp, it is inevitable that the west side would have experienced catastrophic mass wasting and yielded the landform we see today, given a source of abundant liquid water within the mountain, which I am suggesting the underground springs flowing from the region of the central peak provided.
              Below, courtesy of Google Mars, we see Mt. Sharp as viewed from its base, looking south, up the course of the Northern Channel:  
             And below we have the same perspective, with a line drawn to show the missing material, which (I am suggesting) collapsed off the west side of the Upper Mound (the debris from which was subsequently removed by wind erosion, as well as additional fluvial action):

              Another respect in which this landform on Mars parallels geology on Earth that I am well acquainted with, and whose common principles I immediately saw at work on Mars, involves (yet again!) the geology at Newport, Oregon.  There, the Astoria formation (of early Miocene age) overlies the Nye Mudstone (of late Oligocene/early Miocene age), with the bedding plane dipping at close to 30 degrees towards the sea, and wherein the Astoria sandstone is highly permeable to water but the Nye (fine-grained silt and clay) is not, and the water that accumulated served to lubricate the contact area, resulting in the Astoria block sliding into the ocean, and causing the loss of a number of structures at the “Jumpoff Joe” locality.  This is described (on a popular level) by local paleontologist Ellen J. Moore in her book, ”Fossil Shells from Western Oregon” (Chintimini Press, 2000), which includes the two photos below, showing Nye Beach before and after the slide (which occurred in the 1990s).  And while this is a more extreme case than the more gently-dipping Mars strata, it illustrates the mutual principles involved:
              Returning now to Mars: looking at the photos of the “stage” area, we see a triangular-shaped mesa about 2 miles wide, which is an eroded remnant of the approximately 700-foot-thick stack of layers originally lying above the one(s) now occupying the majority of the surface area of the “stage,” and it is likely that the lion’s share of water was seeping (or pouring) out of the mountain at the exposed edge of the bedding plane defined by the top of that mesa, an erosive flow which subsequently deflated the surrounding terrain to the “stage” surface we see today (but we shouldn’t expect to see any outflow openings in orbital surveys, as they would have since collapsed and/or been covered by talus slope debris and aeolian deposits).  It appears that this mesa is composed of the top-most layers of the Lower Formation (rather than being a surviving block of the Upper Formation), as the upper surface of the mesa dips to the NW, implying that the angular unconformity between the formations is defined by the upper surface of this mesa.  Here we see a high-resolution photo of this feature, and also note the 2,000-foot-wide circular structure at the lower left corner of the mesa, which is an exhumed, eroded “paleo-crater” or fossil crater (more on the significance of this later):

              I would surmise that the Grand Canyon GC’s flow was emerging from the side of the mountain (as a spring) for quite some time before the collapse occurred, and was the factor primarily responsible for that collapse, and it’s likely that it had already carved much of the depth of its downstream canyon prior to the collapse.  The Northern Channel was a later feature, probably formed after the collapse had occurred (I will provide evidence for that later), and the Southern Channel probably only flowed around the time of the collapse (not only on the basis of its less-developed features as a channel, but also because it was at the uphill end of the collapsing block, and once the collapse occurred, the lion’s share of the water would have drained downhill towards the NW, i.e. towards the Grand Canyon and what was to become the Northern Channel). 
              Interestingly, the Northern Channel ends (or more precisely, radically changes) at a point around 12,493 feet below Mars datum, which is still about 2,800 feet above the bottom of the slope it traverses (i.e. towards the low point of the crater floor of Gale).  It suddenly transitions into a roughly fan-shaped deposit, but this is not an alluvial fan (such as the Peace Vallis fan on the opposite side of Gale Crater), because an alluvial fan forms when a channel dumps its load onto a nearly level surface, and this location is on the slope of Mt. Sharp, in an area that has a 7.5-degree inclination (under conditions of Martian gravity, channels don’t spread out to form alluvial fans until the slope is about one-fifth this value, as exemplified by the Peace Vallis alluvial fan that is sourced from the northern crater rim).  Which can mean only one thing, namely that this was a delta formed as the channel dumped its load into a standing body of water, specifically a lake!  We may observe that the elevation difference between the head of the fan and its furthest northwest extent (its toe) is around 1,300 feet, over a distance of almost 2 miles, which is the pattern we would expect if sediment were being dumped into a standing, quiescent body of water above the bed of a lake with a moderately sloping shoreline, whereby the velocity of the incoming water collapses when it hits the standing body of water, turbulence subsides, and the flow can no longer hold the sediment load it is carrying, which then settles to the bottom of the lake in a fan-shaped deposit, with the thickest section just below the contact with the lake surface (at around the minus 12,600 foot mark in this instance), feathering out (to a negligible thickness and eventually disappearing) the further we move away from the inlet (towards the northwest in this case).  In contrast, an alluvial fan would have only a modest elevation difference from head to toe, and there is no conceivable alternative explanation involving a “dry” obstruction to the downward flow of the channel on the slope, such as a (now-eroded and long-gone) rock outcrop or ridge, which would either cause the channel to change course around it, or to back up the flow and cause a “reverse fan” where the thickest part of the sediment load would settle out against the obstacle, at the furthest northwest extent of the deposit (and neither condition is in evidence here).  Below is a figure from B. J. Thompson et al. [4], highlighting the delta (the numerical designation is the elevation at which the fan starts, 3,808 meters, or 12,493 feet, below Mars datum):

              Below is a closeup photo of the channel where it transitions to the delta, and where south is up:
              It can be observed that the (now filled) channel continues on through most of the visible length of the delta, and I would surmise that the channel originally followed a path towards the bottom of the slope (below the current delta deposit), but once the water level in the lake rose towards its maximum value, the channel successively filled the lower reaches of itself with its sediment load (which at that point were underwater).  And today, this channel fill has been exposed by wind erosion, and has developed into an “inverted channel” that stands above the outlying portions or the delta and the surrounding landscape (which have been deflated by wind erosion, being more subject to that erosion than the resistant channel fill).
              Gale Crater’s Grand Canyon also has channel fill in its lower reaches and a fairly massive delta deposit, but due to interactions with a far more complex and irregular topography, it does not exhibit as idealized a fan structure.  The channel fill zone joins with the fill from a shorter branch canyon and widens to form a delta at around minus 12,700 feet, somewhat analogous to the behavior of the Northern Channel at the head of its fan, and consistent with its last major outflow occurring at the same time as the Northern Channel’s, during the lifetime of the associated lake (although the flow of the Grand Canyon may have greatly diminished just before the lake reached its highest level, at a time when the lion’s share of the flow may have transitioned to the Northern Channel, based on the slightly different heights of the respective channels’ delta heads).  In the photo below, north is up:

              In the perspective below, the arrows define the point at which the channel fill zone broadens, at around minus 12,700 feet, apparently signifying the fan development process:


              Returning now to the Northern Channel: the photo below shows its position in relation to the landing location of Curiosity, as well as the rover’s brief detour to the “Glenelg” location (at “Yellowknife Bay”).  The future path planned for Curiosity will take it into the immediate area of the delta deposit, whose general region is considered the primary mission target of the rover.  On either side of the head of the delta, can be seen what appears to be a erosional terrace formed by wave action and/or ice push, where the slope of the hillside suddenly decreases (from about 12 degrees to about 7.5 degrees), where the water was apparently lapping against the side of the mountain (or ice pushing against it) during a period of temporary lake-level stability, and which also represents the highest level that the water reached during the timeframe of the Northern Channel’s flow:
              We would expect that such a lake would have also left its shoreline’s mark on the opposite side of the crater, along the inside northern rim, and indeed this is the case: there are a number of spots where a bench or terrace can be seen in topographic maps [5] at around the minus 12,400 foot mark, which is also the level of the top of the delta (with associated terrace) at the end of the Northern Channel.  Also on the crater rim, we see many examples of smaller fluvial features which transition from channels to fans at approximately the same level as the Northern Channel does, with the same pattern in that there is no reduction of slope gradient that could explain it, and likewise they appear to represent the entry point to a lake surface.  Below is a raw photo including several of these fans (directly to the north of Mt. Sharp), and below that is a figure from Thomson et al. [4] with the deltas and channels highlighted (the fans are now of inverted profile, as a result of wind erosion and subsequent deflation of the surrounding landscape, as in the case of the Northern Channel’s fill and delta).  The elevation designations in meters indicate where they transition from channels to fans, which roughly corresponds to the elevation (3,808 meters) where the same thing happens to the Northern Channel.


              The above observations in turn imply that the Gale Crater lake at that time reached a maximum depth of 2,800 feet (from its surface to its deepest point, which would have been at the lowest point of Gale Crater’s northern floor, at minus 15,300 feet), with the average depth of the lake being a fraction of that, at around 600 feet.  Such a lake would have spanned a width of about 15 miles between the central mound and the rim of the crater, and stretched about 50 miles on either side, making up a crescent-shaped lake with a surface area of around 2,000 square miles and a volume of approximately 220 cubic miles!  Although this was likely a lake that was mostly frozen over, in order to outpace evaporation and build up to such a level, even if it were during a time when atmospheric pressure were higher.  
              Below is a representation of how such a lake would have appeared, from a presentation by W. E. Dietrich at the 44th Lunar and Planetary Science Conference in 2013 [5].  I will refer to this lake as the “Second Generation Gale Crater Lake,” to differentiate it from the “First Generation Lake” that existed in the far earlier history of Mars, shortly after Gale Crater formed, and within which the Lower Formation was presumably deposited.

              There is further information to be gleaned from the visible structure of the Northern Channel’s delta deposit: it is notable that there are no channels incising the deposit, which implies that the delta deposit represents the LAST major outflow of the Northern Channel, as any significant subsequent flows would have cut through and eroded the pre-existing deposit.  And it appears that the lake level never rose significantly higher than the 2,800 foot mark defined by the delta (at least during the timeframe of the channel’s flow), as there are no similar deposits further up the channel (as well as there being no higher terrace).  So we have an apparent temporal correlation between the highest lake level on the one hand, and a major outflow (also the last major outflow) of the Northern Channel (and likely the Grand Canyon as well, as its headwater derives from the same complex).  Which seems unlikely to be a coincidence, and suggests that the channel complex of Mt. Sharp was the primary source of water filling the lake (at the very minimum, those channels are the source that topped it off to the 2,800 foot mark).  
              In the northwest corner of Gale Crater, at the crater rim opposite the area being investigated by the Curiosity rover, there is a fairly large inflow channel (Peace Vallis), which has deposited a 10-mile-long alluvial fan on the floor of the crater and is believed to be the source of the water-rounded pebbles of the streambed-produced conglomerates discovered by Curiosity shortly after it landed.  Peace Vallis arises from a watershed in the crater rim hills, and in addition to surface runoff, it may also have been fed by springs flowing from a “perched” aquifer in the hills of the crater rim, which may be the only way to explain the relatively steady flow over a period of thousands of years which is believed necessary to produce the well-rounded river gravels which have been observed near the Curiosity landing site.  Below we see an orbital photo showing the entire course of Peace Vallis, from headwaters to the southern end of the alluvial fan: 
              However, there is no reason to believe that the flows of Peace Vallis correlated with the formation of the channels of Mt. Sharp and the filling of the Second Generation Lake.  Indeed, the alluvial fan of this channel extends very far into the area that would have been occupied by the apparent 2,800-foot-deep lake, and the fan begins at an elevation about 1,000 feet below the water level of that lake, which implies that the Peace Vallis fan deposits were laid down either before of after the lifespan of the lake (and below I will present evidence that Peace Vallis exclusively flowed PRIOR to the creation of the Second Generation Lake, which is perhaps unexpected and runs counter to the most widely-accepted chronology).  Additionally, the Peace Vallis channel is only about half as wide as the channels of Mt. Sharp, so even if it had flowed during that timeframe, its contribution to the Second Generation Lake would have likely been much smaller.
              In the southwest corner of Gale Crater, there is a far larger channel (4 km wide in places, but so far unnamed to my knowledge) that breached the crater rim from the outside (and which could be designated as the “Southwest Channel” or “SW Channel,” not to be confused with the “Southern Channel” on Mt. Sharp):
              This channel drained a large region to the south of Gale (encompassing many tens of thousands of square miles), and it deposited two large fans in Gale Crater:
              One of these fans is an arc-shaped feature on the edge of the tableland in the southwest quadrant of Mt. Sharp, with morphology consistent with deposition in a standing body of water (namely, an extremely flat surface with no incised channels, and a steep scarp at its distal extent):

              And below is the second fan, on the crater floor to the west of Mt. Sharp, the lower reaches (northern or upper portion) of which are now covered by a large, dark sand dune field (the channel and both of its associated fans can also be seen in the plan-perspective color-coded 3-D view of Gale crater near the beginning of this essay):  
              However, both of these fans, and the activity of the SW channel in general, appears to predate the Second Generation lake and the Mt. Sharp channel complex: the arc-shaped fan lies at an elevation about 1,500 feet above the delta of the Northern Channel, so it appears to signify a filling of Gale Crater with water at a time well before, and to a higher level, than the activity associated with the Mt. Sharp channel complex.  And the fan to the west of Mt. Sharp begins at a level approximately 500 feet above that of the Second Generation Lake (and likely ends at a lower level, although that area is obscured by the large dune field), and based on topgraphic contours, that western fan appears to have an erosional bench cut into it at what would have been the shoreline of the Second Generation lake (at about the minus 12,400 foot level, which is at the distal northern end of the shallow channel that incises the upper surface of the fan shown above), so again, this activity pre-dates the activity of primary interest to us in this essay (where I am focusing on activity involving Gale crater as a semi-closed hydrologic system, where the water making up a lake had its source within the crater complex).
              There are very serious constraints on the aquifer that could have fed the channels of Mt. Sharp, and apparently provided the lion’s share of water for the Second Generation lake.  This is because Mt. Sharp is a relatively small isolated mountain, not the edge of a large tableland or larger mountain complex that could have fed it a water supply from a large enclosed aquifer.  The one-to-two-mile-wide channels/canyons of Mt. Sharp would have required a huge volume of water to cut, far more than could have been originally contained in the sediments making up the collapsed western portion of the upper mound, either as ice or liquid water or hydrated minerals.  Much less could the collapsed section have provided enough water to fill a lake: the total volume of the section that has fallen away from Mt. Sharp is approx. 20 cubic miles, less than a tenth of the estimated volume of the lake at its highest, and of course the water component of the eroded sediments would have been only fractional, and so the net contribution to the lake from that entrapped water would have been fairly insignificant.  And if the water in question had somehow drained from within the currently-intact portion of Mt. Sharp (such as from the melting of sediment-entrapped ice), any quantity remotely approaching such a channel-carving and lake-filling volume would have caused the vertical collapse of the entire Mt. Sharp complex, much like the “choatic terrain” found elsewhere on Mars at the headwaters of large outflow channels.
              Thus the water source must have been from outside of Mt. Sharp, and since a “horizontal” source is out of the question, the only possible sources of such a large channel-carving and lake-filling volume of water would be vertical: either downward-vertical (from liquid water percolating into the mountain from precipitation) or upward-vertical (from an artesian hydrant).  But the timeframe in which the channel complex of Mt. Sharp formed is not the one in which rain-fed channel and valley networks were forming on Mars, that was during an earlier time (3.5 to 4.0 billion years ago) when the First Generation Gale Crater lake was laying down its sediments, and the Martian atmosphere was sufficiently dense and warm and moist to support rainfall.  The Mt. Sharp channel complex was a later development, likely a billion or more years later, during the Amazonian era and long after the so-called “Great Dessication Event”, and during a time characterized by only the occasional formation of large outflow channels on Mars, as from underground aquifers or melting permafrost sources, and with no evidence of rainfall, as the atmosphere was too thin and dry and cold at this time to support such a phenomenon.  And the fact that the channel complex is geologically young, and thus very comfortably fits within this “post-pluvial” era of Mars, is also demonstrated by its intact status as a shallow “drape” formation in an environment with a high erosion rate (the flank of Mt. Sharp), and I will have more to say below about the age of the channel complex.
              And even if there had (somehow) been a sufficiently massive quantity of rainfall to supply the outflow channels, by soaking into the mountain and exiting through springs that undercut the west side of the mountain, why would there be no sign of surface-runoff gullies and channels on the mountain, either on the sides of the mountain opposite and adjacent to the “amphitheater”, or at higher elevations, which would inevitably have resulted from such a large quantity of rainfall?  There are many erosional features on Mt. Sharp, especially wind-carved yardangs, but no trace of runoff gullies or other surface drainage patterns.  And there seems to be no accounting for why such (hypothetical) precipitation would only leave its mark 7,000 feet below the summit, only by soaking into the mountain and discharging through springs in a highly restricted area, and effectively by-passing the surface.
              It might be suggested that, if the surface of Mt. Sharp were extremely porous (such as that of a sand dune), then perhaps precipitation could entirely soak in, and not run off, and so exit the side of the mountain at a lower point (such as after seeping downward till it reached a relatively impermeable layer), but this is ruled out by the crisp surface features shown by high-resolution photos of the mountain, including yardangs, ridges, steep topographic gradients such as scarps and cliff-bench layering, impact craters, and boulders in colluvium piles, all of which imply a hardened, indurated surface and well-lithified constituency to the substrate, which would be expected to produce runoff gullies in response to precipitation, whereas a highly friable and porous sand surface would not be able to support such observed features, and instead would readily flow and collapse in response to any stress (much as the way our footprints quickly disappear in the surface of a sand dune).
              And it could NOT be cogently argued that there were originally runoff gullies on Mt. Sharp that have since been worn away by wind erosion (and thus that major precipitation could have occurred, adequate to drive the formation of the channel complex), because such a level of systematic, intense, mountain-wide erosion would also have obliterated any thin, delicate “drape” formation such as the Northern Channel’s delta deposit (which it obviously has not).  Wind erosion is also very fickle and spotty in its pattern, with “lows” on any irregular surface, protected spots of low wind speed that will be less subject to scouring and erosion, and which will host “survivor” features that are obliterated from the higher-profile or more erosion-susceptible locations, and drainage gullies and channels will tend to be so sheltered by having cut downward into the substrate, and yet there are no signs of them on Mt. Sharp.
              Below is a photo from orbit of the highest area of Mt. Sharp, above the “stage” area of the “amphetheater” (i.e. showing the “stands” of the complex), across the summit and part-way down the opposite side (with north facing up), showing the ”cliff-bench” or terrace-like geometry (where a succession of alternating “benches” and cliffs are seen as we go up or down slope) and an abundance of yardang-like ventifacts, but no sign of aquaous drainage patterns.  And the same is characteristic of the lower reaches of Mt. Sharp, which similarly have no sign of small to medium scale fluidic drainage features (whereas these would be expected to occur if there had been significant precipitation on Mt. Sharp).  So in other words, the outflow channels on Mt. Sharp basically come out of “nowhere”, with no supporting infrastructure or tell-tale surface features that would accompany them if the source of their water were from the surface of Mt. Sharp, either directly (via runoff) or indirectly (via infusion into the subsurface and subsequent outflow in springs at lower elevations).    


              Here is a higher-magnification view, showing the dramatic cliff-bench erosional morphology, the yardang-like ventifacts, as well as the swirls of what appears to be lithified aeolian cross-bedded (dune) strata exposed in cross-section by wind erosion (of which I will have more to say later), but no sign of fluvial erosion:


              Snow and ice melt are also not plausible sources for the large flow rate required to carve multiple one-to-two-mile-wide channels, as the infusion rate into the subsurface would be too low, given that the surface of the upper levels of the west side of Mt. Sharp, above the channels and prior to the collapse of the west face, only involved an area measuring about 15 by 10 miles.  A thick column of snow/ice rapidly melting might provide a sufficient infusion rate, and even be roughly equivalent to a heavy rain falling on that watershed, but such is an implausible scenario for the cold surface environment of Mars, and would not be a normal occurrence even for the warmer Earth, where such high flow rates from melting snow/ice normally require a far larger watershed to draw from, or else an extraordinary event such as a volcanic eruption on a snow-capped mountain, neither of which are conditions satisfied by Mt. Sharp.  And in any case, such massive and rapid (and implausible) snow/ice melt would also have resulted in the formation of (the missing) surface runoff gullies and channels.
              It has been suggested by some authors, specifically by Le Deit et al. [6] and Fairen et al. [7], that melt water from permafrost or snow or rock glaciers might have provided the water needed to carve the channels, but they offer no plausible, detailed mechanism to explain how this could have worked in practice.  Because once again, there are none of the inevitable small-to-medium-scale surface runoff features that would characterize a drainage system….and in addition, there is no topography on Mt. Sharp (neither suitable landforms or sufficient areal extent) that could have allowed the formation of an ice dam (in the same vein as Earth’s glacial lake Missoula) which could have accumulated and subsequently released the massive volume and flow rate of water that would have been required to carve the channels, transport huge boulders (which I will discuss later), form massive fans/deltas, and fill (or at the very least, top off) a deep and massive lake in Gale Crater.  In short, any melting ice on Mt. Sharp could not have provided more than a relative trickle, which would be totally inadequate to explain the landforms we are seeing, and there is no evidence of even that limited a flow from surficial sources.
              And so I conclude that the source of the water that carved these channels must have been a deep aquifer below Mt. Sharp, which had sufficient hydrostatic pressure and flow capacity to support artesian springs, and that early in the history of the Gale/Sharp complex, these springs served to fill the crater lake and deposit the over-one-mile thick Lower Formation of clay and sulphate-bearing sediments, next they served to lithify the aeolian sediments covering them (the Upper Formation), and then in the most recent phase, the springs were uncovered by erosion and (during a limited timeframe or timeframes) once again served as a source of flowing surface water, which carved the channels/canyons shown in the above photos, and created a temporary lake once again filling much of Gale Crater (at one point to a depth of 2,800 feet, as shown by the location of deltas and terraces….as indicated earlier, I will refer to this as the “Second Generation Lake,” whereas the original lake in Gale Crater would be the “First Generation Lake”).  I would assume that the First Generation Lake was FAR longer-lived than the second, and also larger and deeper, based on the over-one-mile thick column of sediment that was laid down, whereas the Second left less of a record and was a rather transitory event, occurring as a result of a limited-duration outflow in a later period of Mars’ climactic evolution (drier and with lower atmospheric pressure) that did not favor the presence of long-lived surface bodies of water (hence this lake deposited insufficient sediments to obscure the earlier impact craters on the floor of Gale Crater, and what sediments that were deposited, were mostly removed by subsequent wind erosion….thus there was no obvious or large-scale ”resetting” of the three-billion-year-old “clock” defined by the crater size vs. frequency distribution of the visible floor of Gale).
              To my knowledge, this is the first (and only) systematic hypothesis that has been put forward in an attempt to explain the origin of the Mt. Sharp channels.  It has some features in common with a model put forward for Gale Crater by Andrews-Hanna et al. [8, 8A], whereby it is suggested that regional groundwater systems provided upwelling of groundwater at low-elevation locations such as Gale, and which in turn was involved in the deposition and sediment cementation of the Lower Formation.  However, their model is non-specific and does not provide a mechanism for the introduction of water into Gale that can explain the development of Mt. Sharp as an isolated sedimentary mound, let alone the development of the channel complex, whereas my artesian hydrant model provides a very specific, naturally-engineered mechanism for exploiting subsurface hydrostatic pressure, in order to power the construction of the complex that we see today.
              As regards the age of the channels (or more importantly, the age of their most recent activity), they have a fairly pristine appearance overall, with no sign of having been seriously degraded by the erosion that has been systematically decimating the Gale/Sharp stratigraphic column for a very long time, and they have also not experienced any significant burial by more recent sediments.
              The channel complex also has a low density of impact craters, with few to none in the channels themselves.  And although there are several craters in the “stage” area of the “amphitheater,” most of them are highly degraded or incomplete, and in the majority of cases are obvious paleo-craters that predate the higher layers of Mt. Sharp (as well as the more recent erosion associated with the channel formation, which has served to re-expose these “fossil” craters).  Below is a clear example of such an exhumed paleo-crater, with a diameter of about 2.5 miles, at the head of the Northern Channel and in a high state of degradation (I have outlined its shape for clarity):
              I previously cited another paleo-crater in the “stage” area, about 4 miles to the south of the above pictured one, located just to the SW of the mesa.
              In the southern portion of the “stage” (specifically, the headwater area of the Southern Channel), there is one crater, around three-quarters of a mile in diameter, which is fairly pristine in appearance and obviously post-dates the collapse of the west side of the mountain (plus there are several small nearby craters, one of which is in the upper right of the photo, and which may be associated with the same impact, in that a single bolide may have fragmented prior to impact, something which is suggested by the tight clustering of these craters), but the central and northern portions of the “stage” are virtually free of any obviously recent (non-paleo) craters.  And the aforementioned 3/4-mile-diameter crater has a sharp rim and largely intact ejecta blanket in a very highly wind-erosive environment (see below), so it must be fairly young, presumably no more than several tens of millions of years old, and beyond this, it provides no real information on the timeframe in which the western face of the mountain collapsed, and does not imply a more antique timeframe for that collapse than the past several million to tens of millions of years.
              Also noteworthy is the fact that wind-carved “yardangs” (linear ridges oriented along the axis of the prevailing wind direction) are comparatively subdued in the “stage” area, whereas they are highly prevalent on most other surfaces of Mt. Sharp.  There are some yardangs in the area of the headwater channel of the Grand Canyon GC, and in the “stands” above that point, but most areas of the “stage” are free of them, or only exhibit them in rudimentary form, which shows that the “stage” must be a comparatively young surface (which hasn’t had a chance to form many mature yardangs).  This argument for youthfulness is especially compelling when we consider the fact that the wind erosion rate (and hence, presumably, the rate of yardang formation) on Mt. Sharp is VERY high (as we shall see below).
              These properties of the channel complexes (fresh appearance, a comparative shortage of more recently-formed impact craters, and a comparative shortage of wind-carved yardangs) not only indicate that the channels (or at least their most recent fluvial activity) are geologically young, but also that they were carved into the mountain LONG after the vast majority of material filling Gale Crater had been removed by erosion, and in a time-frame where the mountain essentially looked as it does today.  They are obviously not ancient “fossil” rivers or “paleorivers” that were covered up by sediments and that have now been uncovered by erosion, as shown by the lack of cross-cutting of the channels by erosion, and the lack of overlying (uneroded) sediments draped over major sections of these channels.  And in any case, there is no plausible scenario that could have given rise to a “primordial” Mt. Sharp that featured these outflow channels, a mountain that exactly matched the dimensions of the current Mt. Sharp, which was subsequently covered by sediments, and which has now been implausibly re-exposed to perfection by erosion, including its drainage channels and depositional fans, effectively down to the under-one-meter-scale resolution of the photographs.
              A related consideration that speaks to the recent origin (or activity) of the channels, is the fact that the impact cratering density on Mt. Sharp in general is not only low in relation to values commonly found on Mars, but is very dramatically so in comparison with the adjacent crater floor that surrounds Mt. Sharp (with small craters in particular, which are normally the most common size range by far, being comparatively absent on Mt. Sharp, which indicates a high erosion rate, since small craters fall prey to erosion much sooner than large ones).  This can be readily seen in a topographic map or orbital imagery of the Gale/Sharp complex, and once again, most of the craters that are visible on Mt. Sharp can be shown to be paleo-craters created prior to more recent deposition (and subsequent erosion), which dictates that they must be excluded from any crater counts used to estimate the age of the currently-exposed surface (meaning that the “legitimate” crater count, for purposes of dating surface exposure, is much lower than it appears to be at first blush).  
              Below we see a cluster of craters south of the Grand Canyon GC, most or all of which are paleo-craters, as shown by their irregular shapes and on-lapping lobes of country rock.  These craters formed during the time when the Lower Formation was being laid down, next they were covered by higher Lower Formation layers, and are now being uncovered by erosion.  So they represent “fossil” craters entirely encased within a single formation, rather than defining erosional unconformities between formations separated in time.  And perhaps a bit counter-intuitively, this instance of a terrain with a high cratering density, lacking as it does craters that are diagnostically of recent geologic origin (non-paleo), actually speaks to the youthfulness of the exposed surface of that terrain:

              And some examples directly to the northeast of the Grand Canyon GC:
              Collectively, these observations show that the visible surface of Mt. Sharp must be a recent erosion-created surface, and that if the more delicate features of the channel complex (such as the channel fill and delta at the end of the Northern Channel) were not of geologically recent origin, even startlingly recent, they would have been worn away long ago.  But just how recent?
              In the past, it was generally believed that the erosion rate on Mars is very low, and that the visible sand dunes (which on Earth would normally signify dynamic motion of sand grains through “saltation,” and active erosion of any exposed bedrock through sand-blasting) were mere “fossils” from an earlier time of Mars when the atmosphere was denser (it was also believed that the Martian atmosphere is currently too thin and the winds too weak to support saltation of sand grains, even with the comparatively weak Martian gravity).  All of this has changed in recent years, however, as a result of several lines of evidence. 
              First of all, high-resolution photos of the surface (such as those by the Mars Reconnaissance Orbiter) have shown that many areas of the planet are almost totally lacking in small craters, which indicates highly active on-going erosion, as the impacts that produce such craters have been shown to be actively occurring, at the rate of approximately 200 per year globally [9], a figure derived from MRO observations of fresh craters that have appeared during several years of observation.  In fact, on the basis of the relative shortfall in the number of small craters (which are created in far higher numbers than large craters, but are eroded away far faster) it has been estimated that parts of the planet are being effectively resurfaced, on the scale of even 2,000-meter-wide craters, on a timescale of under 9 million years [10].
              Secondly, observations of sand dunes by the MRO have shown that they are actively moving [11,12], at a speed comparable to similarly-sized sand dunes on the Earth [13].  This implies the active (and erosive) saltation of sand, likely to be basaltic sand in most cases (unlike the more typical quartz sand on Earth), and close parallels have been drawn between Martian conditions and those of Antarctic dry valleys on the Earth, where there are basaltic sand dunes migrating at a similar speed and causing erosion of basaltic bedrock at the rate of 30 to 50 micrometers (millionths of a meter) per year, and models of the saltation of sand grains under Martian conditions of atmosphere and gravity suggest erosion rates, where active dune migration has been observed (measuring approximately .5 meter of travel per Earth year), of 1 to 10 millionths of a meter per Earth year on a flat, level basaltic surface and 10 to 50 millionths of a meter per year on a vertical basaltic rock face [13], with much higher erosion rates to be expected with softer materials (such as most sedimentary rocks).
              These observations, together with the “NASA Ames 3-D general circulation model for Mars” (GCM), have led to estimates of globally-averaged Martian wind erosion rates as high as 100 millionths of a meter per year [14].  Which is one ten-thousandth of a meter per year, and when phrased that way may not sound like much, but these figures accrue to huge sums over geologic time:  being very conservative and taking a value just one-twentieth (5%) of this global estimate (just 5 millionths of a meter per year), there would be an average of five meters (15 feet) of erosion per million years, and 15,000 feet per billion years!  So while the erosion rate on Mars may be lower than the Earth’s, there is increasing evidence that it is not as far behind as has been generally believed, and in any case is almost infinitely higher than that of an airless body such as the Moon.
              And the implications of this are extremely interesting when applied to the delta deposit of the Northern Channel: First off, let us (rather arbitrarily, for purposes of illustration) assume that the flank of Mt. Sharp is subject to erosion at the conservative global average I provisionally adopted above, of 15 feet per million years.  Secondly, let us also assume that the typical Earthly ratio of the width to the depth of a channel also applies to Mars, namely around ten to one (which would seem to be a reasonable assumption, based on the general observation that Martian and Earthly river channels are similar in morphology).  And since we observe that the width of the end of the Northern Channel (before it spreads into a fan) is around 2,000 feet, we then arrive at a depth of around 200 feet.  And this channel is filled with old sediment and currently eroding, and if we assume for the sake of simplicity that 50% (on average) of the original thickness has been eroded away, then we get 100 feet of erosion….which, based on the assumed erosion rate of 15 feet per million years, means that the channel fill is only about seven million years old, since that amount of erosion constitutes just under seven of the 15-foot (per million year) increments! (It is possible that the channel has been widened by wind erosion since it formed, but if that is the case, our estimated channel fill thickness is less, and the estimated age even lower.)
              But is there any reason to believe that the erosion rate in the foothills of Mt. Sharp even remotely approaches this value?  After all, there are also wide areas of Mars that have a much LOWER than average erosion rate, especially the heavily cratered Southern Highlands (otherwise, all but the largest or most recent craters would be long-since worn away).  However, it can be readily demonstrated that the flank of Mt. Sharp does indeed have a high erosion rate, based on the dearth of (non-paleo) craters, both in relation to the adjacent floor of Gale Crater and to global frequency distributions, and is in fact analogous to terrains elsewhere on the planet that have been categorized as very young on the basis of crater counts.  And such a high erosion rate is also consistent with the observed dune migration rates on the adjacent floor of Gale Crater, amounting to as high a speed as .4 meter per Earth year [15,16].
              To date, one of the most important results from the Curiosity rover has been the measured erosion rate on the floor of Gale Crater, at the “Yellowknife Bay” locale, of approximately .75 meter (2.5 feet) per million years, for a vertical mudstone rock face [17].  Reducing this figure to one-sixth for a flat, level surface (based on the average ratios in the above-cited study [13] of basaltic sand migration), but compensating for the fact that the delta deposit is in fact on a sloping surface (a 7-degree grade), a reasonable approximation might be that the erosion rate for the delta deposit is .5 foot per million years, given the same wind regime as at Yellowknife Bay (and the same susceptibility of the local lithology to wind abrasion) .  This would then result in an estimated age of about 200 million years for the channel fill (which would be young in relation to the generally-accepted chronology for Gale Crater, but far older than the figure I arrived at earlier).  However, there is every reason to believe, on both theoretical and empirical grounds, that the wind erosion rate in the foothills of Mt. Sharp is FAR higher than on the floor of Gale.  First of all, based on terrestrial analogues, wind speeds (and hence expected erosion potentials) should rise sharply with elevation, since friction with the ground (and resultant loss of velocity) is greatly reduced.  And secondly, we empirically observe that there are very few recent (non-paleo) craters on the flank of Mt. Sharp at the elevation of the delta deposit, whereas they occur with variable but universally fairly high densities in all Gale Crater-floor locales…..and since craters are by comparison almost completely lacking on the flank of Mt. Sharp, a very rough (but plausible) visual estimate would that they are over an order of magnitude lower in frequency than on the crater floor where Curiosity measured the erosion rate (under one-tenth of Yellowknife Bay’s cratering density), which implies that the erosion rate on the side of Mt. Sharp should be over ten times higher than at Yellowknife Bay, perhaps on the order of 5 to 15 feet per million years for a flat, wind-facing surface with a 7-degree slope.  Which brings us into the temporal range of perhaps 7 to 20 million years for the age of the delta deposit, which overlaps with my original channel-fill age estimate of seven million years (and this time, the figure is based on empirical data, not values arbitrarily selected for the purpose of example).
              However, this method of estimating the age of the delta deposit, does suffer from a couple of weaknesses which impair its precision, namely that the total amount of erosion experienced by a given stratum is rather arbitrarily selected from a possible range, plus I am assuming that all sedimentary materials will erode at the same rate in a given wind regime.  And later in this paper I will present an alternative (and more rigorous and precise) method for estimating the age of the delta deposit based on empirically-measured erosional deflations, and which also takes account of the variable rates of erosion of different lithologies.
              Below we see a picture (courtesy Google Mars) of a one-mile-wide swath of land immediately to the west of the delta deposit.  Note the almost complete lack of impact craters, with those present being highly degraded, almost to the point of being unrecognizable as such (and in most or all cases likely to be exhumed paleo-craters):
              Now we see a one-mile-wide swath of ground centered on the location where Curiosity measured the erosion rate as being 2.5 feet per million years (just inside the edge of Yellowknife Bay, which lies to the upper right of center).  Note that all nearby terrains are densely peppered with impact craters, in various states of erosional degradation, for which a rough visual estimate indicates a frequency in excess of one order of magnitude higher than for the delta deposit region, even in the more rapidly eroding (hence lower-crater-density) Yellowknife Bay strata:
              Here is a closeup of Yellowknife Bay, which features a conspicuous 130-foot-diameter crater in its center (arrowed, and fairly fresh in that it retains a rim and significant bowl depth), as well as several smaller craters in various stages of erosional degradation.  This demonstrates that Yellowknife bay, even though it has a relatively high rate of erosion and a lower crater count per unit area than its immediate surroundings, has a much longer crater retention age than the area of the Northern Channel’s delta deposit, hence the delta deposit’s environment must be experiencing an erosion rate that is far higher still.


              As a cross-check of the above results, we can also compare the terrain encompassing the delta deposit to one which, on the basis of its low crater count, has been independently dated as being between 2 and 30 million years old: the Cerberus Fossae region, about 1,000 miles northeast of Gale, where eruptions of both lava and liquid water from fissures (such as the one shown here) completely resurfaced the terrain.  Below we see a typical one-mile-wide swath, and it can be seen that there are very few craters, perhaps on a par with the area adjacent to the delta deposit, but with those craters looking considerably crisper than those near the delta deposit (and obviously being non-paleo craters).  So while we are dealing with a different mechanism here, lack of craters due to deposition of new surface material vs. lack of craters due to ongoing surface erosion of old materials, the similar dearth of craters is suggestive that the effective ages are similar, for the purpose of dating events that have left their mark on these respective surfaces (and below, I will suggest that there may be a direct causal linkage between the Cerberus Fossae eruptions and the geologically recent fluvial activity in the Gale/Sharp complex).
              So the startling conclusion from these observations is that the channel complex of Mt. Sharp (or more precisely, its most recent activity and the Northern Channel’s delta deposit) may date from under 30 million years ago, and Gale Crater was largely filled with a 2,800-foot-deep lake within that timeframe!  Which will require a sea change in how we view Martian evolution and chronology, since such features (and environmental conditions that would accompany them and support them) are commonly assumed to be a hundred times older than they actually are, and it turns out that Mars is still a very dynamic planetary environment, both geologically and climactically, and able to offer more than a few surprises!  And the implications for what Curiosity may find when it reaches the delta deposit are truly huge, and I will have more to say on this matter later in this paper.
              Furthermore, this youthful estimate of the age of the Mt. Sharp channels provides robust support for the artesian hydrant hypothesis of Mt. Sharp, as there is no other conceivable mechanism that could produce such a flow of water in the geologically recent history of Mars.
              Some researchers, such as Le Deit et al. [6], assume that the channels/canyons on Mt. Sharp date from the early history of Gale, perhaps 3.5 billion years ago, but this is utterly inconceivable in light of the erosiveness of the local environment, which is demonstrated by the low crater count.  And based on the erosion rate measured in-situ by Curiosity (and the local and global erosion rates were likely much higher in the distant past, when the atmosphere of Mars was thicker), any shallow “drape” feature such as the delta deposit would have been completely worn away ten or even a hundred times over, if it were of such an antiquity.
              Fairen et al. [7] allow a somewhat more recent timeframe (Hesparian and Early Amazonian) for the occurrence of fluvial activity within Gale, but they dismiss the possibility of Late Amazonian activity, apparently without taking account of the evidence posed by the Northern Channel’s delta deposit (which they don’t discuss, although I consider it to be the single most important feature in the entire Gale/Mt. Sharp complex, for the purpose of understanding that complex’s history).  And fortunately, the delta deposit is precisely where Curiosity is headed, and should arrive there towards the end of 2014 or early into 2015!

              I would offer the following interpretation: the Mt. Sharp artesian hydrant has been inactive most of the time during recent geologic history (mostly dormant for the past three billion years), but due to a rare confluence between a warmer surface climate and an increase in the subsurface temperature gradient arising from deep magmatic activity, the ice plugging the springs of Mt. Sharp melted and surface access for the aquifer was re-opened, at the same time that the aquifer itself was revitalized and pressurized by regional volcanic activity.  And these factors synergistically combined to cause liquid water to begin working its way out of the mountain, presumably in a weak area between layers or through fractures, with what was initially minor seepage creating ever-larger openings, and this water began the process of carving the outflow channels, until the western face of Mt. Sharp’s upper mound was undermined and collapsed, opening an even more direct route for the aquifer to discharge its contents.  And it was likely that the Gale Crater was largely filled by a deep and massive lake at that time.  But eventually the aquifer drained sufficiently to greatly reduce its head, and what was originally a torrent had become a trickle, and even that was subsequently shut off when movement toward colder conditions froze solid the openings of the springs. 
              It should be noted that below-surface magmatic activity (and associated surface volcanic activity) could very well coincide with a warmer planetary climactic regime, because increased volcanic activity leads to outgassing of greenhouse gases and higher temperatures and atmospheric pressures, greatly increasing the ability of the surface to support liquid water, in both channels and lakes.  So the collapse of the west face of Mt. Sharp’s upper mound, the creation of the channels, and the filling of a Second Generation Lake, may not only have occurred at the same time, but coincided with a peak in the volcanic activity of Mars, thanks to a one-two-three punch that could have been delivered by igneous activity: 1) increased subsurface heat that would have melted large areas of deep subsurface ice and revitalized and pressurized the regional aquifer, 2) outgassing an atmosphere that increased surface temperatures and helped to melt through shallow subsurface ice that may have been sealing off the artesian hydrant, and 3) creating a (temporary) temperature and pressure regime that once again allowed liquid water to be stable on the surface.  And below I will have more to say about the timing of geologically recent Martian volcanic activity.
              However, the history of the channel complex is apparently not a simple one involving a single event or relatively short period of activity, and at least two discrete periods of activity, separated by a large gap of time, were evidently involved.  This is demonstrated by the apparently young age of the Northern Channel’s delta deposit, whereas there is no obvious colluvium or lag deposits still extant from the collapse of the west face of the upper mound.  In addition, the amount of material in the delta of the Grand Canyon GC is just a small fraction of the huge volume of sediments that were excavated in the formation of that canyon.  So the collapse event, and the formation of at least the ancestral Grand Canyon GC, obviously occurred long ago, probably in early to mid Amazonian time, long enough ago for wind erosion to have degraded and dispersed the aforementioned debris.  And yet if all the channel activity had occurred that long ago, and a corresponding amount of erosion had occurred since the last outflow, we may expect that the Northern Channel’s delta deposit would have been entirely worn away, and the channels in general would look far less fresh than they do today, even to the point of being mostly eroded away.  So my conclusion is that the complex is old, and has been generally dormant but not extinct, and that there was a recent burst of activity that freshened up the channels and deposited new sediments, as well as filling Gale with a 2,800-foot-deep lake.  And it is possible that additional outflows occurred within the interval bracketed by the first and last outflows, whenever conditions were optimal for a revitalization of the Mt. Sharp artesian hydrant. 
              In addition, the Northern Channel was likely a fairly late feature to develop, and may have formed during the most recent period of activity.  This is suggested by the observation that a delta begins at around minus 12,500 feet in the case of the Northern Channel and at approximately minus 12,700 feet in the case of the Grand Canyon GC, which I interpret as meaning that the Grand Canyon GC largely dried up prior to the high point of the lake, and that the Northern Channel took over as the primary conduit for flow, topping off the lake.  And this is consistent with the fact that the head of the Northern Channel is at a lower elevation than the Grand Canyon GC’s (the floor of the “stage,” approximately representing the unconformity between the Upper and Lower Formations, is tilted towards the NW), which would have favored the flow of water into the Northern Channel, once whatever remaining barrier to the northward flow of water had been eliminated by erosion or collapse (or else a separate subsurface-to-surface conduit opened that supplied the water that formed the Northern Channel….in either case, this was long after the primary collapse of the west face of Mt. Sharp’s upper mound, but before the “stage” area had been fully eroded).  This interpretation is also consistent with the funnel-shaped erosion pattern on the north end of the “stage”, graphically showing how the water was being directed into the Northern Channel from a broad swath of the surface area of the “stage,” once such a pathway had been made available.  It is also consistent with the fact that the Northern Channel did NOT involve the excavation of a huge amount of material (unlike the Grand Canyon GC, which has hundreds of times the volume of the Northern Channel), which not only suggests that the Northern Channel is far younger, but also explains how there can be a lack of colluvium or other debris (aside from the delta deposit) downhill from the Northern Channel’s source, even if it is a young feature, whereas there must have been an enormous pile of material dumped from the end of the Grand Canyon GC by the time that its primary erosion was completed (close to one cubic mile), which would have required a very long span of time to erode away through wind action, likely hundreds of millions of years at the minimum (perhaps the large dune field beyond the end of the Grand Canyon GC is largely the end product of the breakdown of the debris that was excised in forming the canyon?).  And so I would also interpret the visible channel fill and delta of the Grand Canyon GC as simply the product of the most recent discharge from the artesian hydrant, a far less voluminous flow than the one that carved (at least the lion’s share of) the Grand Canyon GC (although this later-stage flow was still powerful enough to move 30-foot-wide boulders).
              In any case, the apparent youth of the delta deposit and of the current surfaces of the channels, have broad implications for the quest for water (and thus current microbial life) on Mars: if the hydrant of Mt. Sharp has been recently active, this implies that a stable liquid aquifer still exists at depth (probably with the benefit of dissolved salts lowering its melting point).  Which would greatly increase the chances of life on Mars, as a stable long-lived aqueous environment is infinitely more favorable than the scarce near-surface moisture on Mars that may remain frozen for thousands to millions of years between melt cycles.
              An ongoing debate has involved whether Mars still has liquid groundwater, at any depth.  For unless it has enough water to virtually saturate the crust (i.e. to mostly fill the available pore spaces through most of the vertical column), whatever water is there will, by now, have been locked away in the “cold trap” of the upper crust (the “cryosphere”), leaving nothing at depth to form an active aquifer, even if the temperature there is above the melting point of water (as it must be at some depth, even if that point is several miles down).  But this evidence from Mt. Sharp, suggesting geologically recent outflows, weighs in on the side of there being an active aquifer (the recent discovery of “recurrent slope lineae” has no bearing on this issue and also cannot explain the Mt. Sharp channels, as the lineae are believed to be shallow surface phemonena, overlying permafrost, and resulting from “deliquescent” salt deposits absorbing moisture which melts at warmer times of the year, creating brines that seep downhill, but only at a low flow rate, and with little erosive effect).
              Further evidence supporting a positive “water budget” for Mars (and hence a present-day active aquifer) is the recent analysis of igneous Martian meteorites, evidently derived from material in the upper mantle of Mars, which show that Mars has (or had) at least as much water in its mantle as the Earth does, which greatly increases the odds that the planet has had plenty of water to play with [18].
              Additionally, striking evidence for a currently-active aquifer in Mars (and one still capable of producing artesian springs) can be had from a location about 1,000 miles to the northeast of Gale, namely Athabasca Valles, a water-carved outflow channel complex which, together with associated volcanic flows, are believed to be among the most recent volcanic and large-scale aqueous features on Mars, estimated to have an age of just 2 to 30 million years, which is a blink of an eye in Martian geologic time, and should provide us with a good snapshot of what the current aquifer is like (the uncertainty in the age of the outflows is a result of the small number of available craters to count in calculating its age, uncertainties in the Martian cratering rate over time, and suspicions that more than one outflow event occurred).  The water that carved Athabasca Valles emerged from cracks in the Martian crust (Cerberus Fossae) from which lava was also extruded, and the prevailing interpretation is that a rising plume of magma cracked the crust and the associated “cryosphere” of Mars (the vertical thickness of the crust in which any pore and fracture-residing water is frozen), allowing water to escape from a pressurized aquifer below the cryosphere.  And it is estimated that, to create the necessary flow rate to carve the observed channels, the aquifer had an effective head at the surface corresponding to a vertical column of water nearly 3 miles high [19], mainly as a result of the nearby Elysium rise and volcanic complex (where the tallest point, Elysium Mons, towers 11 miles above the outlet for the flood).  But the fissure from which the water erupted is only 2,000 feet lower than the headwaters of the channels on Mt. Sharp, PLUS the Gale/Sharp complex is in an even better position to generate a head from a pressurized aquifer, thanks to the adjacent Southern Highlands.  So there should be a copious liquid water source that could still be tapped today by the Mt. Sharp artesian hydrant, the only limitation being the plugging of the conduit(s) with ice due to the prevailing low surface temperatures at this time.
              In fact, the approximate confluence between the estimated age of the Cerberus Fossae eruptions on the one hand, and my estimate of the age of the most recent Mt. Sharp channel activity and the formation of the Second Generation Lake on the other (all occurring perhaps 30 million years ago or even more recently), strongly suggests that there may have been a relationship between the two events.  First off, the fissure eruptions of magma and water, by providing carbon dioxide and water vapor, would have temporarily thickened the Martian atmosphere and thus raised the surface temperature.  And secondly, it is entirely possible that the two locations share a common aquifer across the relatively flat, one-thousand-mile sub-surface of Elysium Planitia, in which case the additional pressurizing of the aquifer by the heat of the magma plume could transfer to increased pressure at Mt. Sharp.  And thirdly, the thermal pulse involved with a rising magma plume may have had a regional effect on the subsurface geothermal temperature gradient, melting portions of the cryosphere from below and increasing its permeability beneath Gale Crater (while warmth at the surface simultaneously acted on the top of the cryosphere).  And if these events happened to correspond with a warmer climactic cycle on Mars as a function of orbital cycles, the cumulative effect could be very significant, with the atmospheric effects of the “orbital forcing” synergistically combining with the atmospheric effects of the volcanism.
              Indeed, there is strong evidence of a warmer, wetter climatic regime during and after the time of the Cerberus Fossae aqueous and volcanic outflows (whether such a climate was due to the outflows or simply coincident with them, or a combination thereof): there are a number of features in the Athabasca Valles area which suggest a “periglacial” landscape (much like the arctic regions of Earth), in that there are extensive polygonal structures in the surface suggesting there was ice at shallow depth (permafrost) experiencing repeated (presumably seasonal) freeze-thaw cycles, plus thermokarst-like depressions and pingo-like mounds, extensive channel networks that cross-cut pre-existing terrains and which suggest melting of ice and drainage of liquid water, and extensive cirque and spur complexes apparently produced by intense sapping and mass-wasting, and together these features suggest that the water released by the fossae did not simply evaporate or soak into the deep crust, but resided long enough at the surface (presumably for at least centuries) to produce these landforms, implying a warmer, wetter, and higher-surface-pressure regime during that time [20].  But because the lava flows are very closely associated with the aqueous flood features, both spatially and temporally, and the lavas apparently overlaped and covered the flood features in many cases, a great deal of controversy has been generated over the interpretation of specific features, as to whether they are water-caused or lava-caused, given that both types of flow can sometimes produce similar-looking features.  However, given the fact that (even under a predominately volcanic interpretation of the landforms) there are numerous “rootless cones” caused by lava flowing over permafrost and generating explosive eruptions of steam, it is demonstrated that surface or near-surface ice was stable at that latitude at that time, which is not true today, so regardless of the interpretation of specific Athabasca Valles features, it appears that there was a more Earthlike climactic and atmospheric regime several million to several tens of millions of years ago on Mars, and one in which we were far more likely to encounter equatorial surface ice, and even liquid water.
              The aqueous flows associated with Athabasca Valles also yielded a topographic feature that cannot be produced by lava flows and is unequivocal proof of a periglacial landscape with far higher temperatures than today, namely “sorted stone circles” which are produced by repeated freeze/thaw cycles in polygonially cracked ground, and are identical in morphology to terrestrial examples in Arctic regions.  Furthermore, it has been estimated that, in order to support such processes, temperatures at this locale would have had to be 40 to 60 degree Celsius warmer than at present [21].  And given that this location was only 1,000 miles from Gale, and Gale is at a lower elevation and latitude and hence apt to be warmer, the implications are that temperatures at Gale could have been well above the melting point of even pure water, averaging between day and night and seasonally, and with daytime highs approaching 100 Fahrenheit!  And if such conditions had persisted for a significant period, they would have served to melt at least the upper portions of the ice blocking the artesian hydrant of Mt. Sharp (see below for a further explanation of this process).
              Not only in Elysium Planitia, but in the plains of Utopia Planitia to the northwest of Gale Crater, there are landforms highly suggestive of very recent (very late Amazonian) fluvial and periglacial activity [22], including thermokarst-like depressions apparently left behind by ponded surface water that was stable for long enough periods to form terraces, signifying surface temperatures and pressures very substantially higher than those of today.
              There are also unequivocal, recent aqueous features at Lyot crater, where there are no associated volcanics.  Lyot is a 147-mile-diameter impact crater located in the Northern Lowlands but at almost the opposite side of the planet from Gale, and at the fairly high latitude of 50 degrees, where surface ice is currently stable over appreciable periods.  And at Lyot there are what appear to be currently active glaciers, as well as gullies and river channels and depositional fans produced by glacial meltwater that has apparently flowed as recently as several million years ago [23].  And while the mechanism for this water generation is different from that of Athabasca Valles, in that it is not from a deep aquifer but rather surface or near-surface melt on top of a uncompromised cryosphere (much like what occurs in spring and summer in Alaska, on top of permafrost), as such it serves to argue even more strongly for the occurrence of globally warmer, wetter, and higher-atmospheric-pressure episodes in the geologically recent history of Mars.
              There are a number of additional locations that have experienced glacial, periglacial, and fluvial activity in the last few million years, all believed to be powered by Martian obliquity cycles, whereby a given terrain will periodically serve as a cold trap that accumulates snow and ice, and is subsequently subjected to a ”thermal excursion” that melts this ice (see below).  Examples in the southern hemisphere are Gregg crater [24], and the Argyre impact basin [25], as well as several recently-formed, mid-sized, mid-latitude craters showing fluvial-related activity as recently as several hundred thousand years ago, including debris lobes, channels with fans, braiding, terraces, and natural levies [26].
              The Cerberus Fossae floods also provide support to my hypothesis that the ancient sediments in Gale Crater were soaked (and subsequently lithified) by artesian flows from the central peak of Mt. Sharp, as the estimated hydraulic head in the Cerberus Fossae event(s) is several thousand feet higher than the highest surviving layer of the Upper Formation, and thus it could not be argued that there would have been insufficient hydrostatic pressure in the regional aquifer to emplace water where required by my hypothesis.  And we may expect that the aquifer over 3 billion years ago would have been at least as active as it has been in the recent geologic past, plus the upper crustal milieu just over 3 billion years ago would have been sufficiently evolved to have featured a cryosphere under which the water would have been trapped and subject to pressurization, and able to exit the surface in only a few select locations (only where open fractures commuted from the surface to the bottom of the cryosphere).

              It is generally accepted that, due to variations in the axial tilt (obliquity) and orbital eccentricity of Mars, the planet experiences periodic warming and cooling cycles, at least on a regional basis (the Martian equivalent of the Earth’s “Milankovitch cycles”, which influence the glacial and inter-glacial cycles of the Earth, only much more extreme in the case of Mars, due to the lack of a large stabilizing moon), and this has been empirically supported by the observation that fresh permafrost has apparently been emplaced in the recent geologic past (on a scale of several hundred thousand years) over wide areas of Mars, as shown by the near-pure water ice excavated by meteor impacts that have occurred during the time in which the planet has been under orbital observation.  Also providing evidence is the ”fresh” ice directly excavated by NASA’s Phoenix lander, ice that had seemingly been emplaced during a warmer and wetter period well under one million years ago, and which must be of such a “recent” origin as shown by the fact that it lies only a few inches below the surface (if it were older than a few hundred thousand years, it would have evaporated/sublimed down to a greater depth).  And it is conceivable that, being near the equator, Gale Crater warms up enough during some climatic cycles (but probably requiring the random confluence of regional subsurface magmatic activity) to melt through the ice that might otherwise be plugging the conduits of Mt. Sharp, allowing a temporary outflowing of those artesian springs, until the climate chills enough to re-freeze the plumbing.
              The diagram below illustrates how this could work, where dry crust is brown, ice-containing crust (cryosphere) is light blue, and liquid aquifer is dark blue: 

              Due to the good thermal insulating properties afforded by the mass of Mt. Sharp (not just due to its thickness but also to porosity, as it consists of sediments of lower density and thermal conductivity than either the typical crustal basalt or of water-saturated sediments), beneath the mountain we may expect an uptick in the level of the underground contact between the liquid and frozen aquifers.  Simultaneous with that, we likely have a relative depression of the top of the frozen aquifer within the mountain in relation to the surface directly above (i.e. a cryosphere upper surface having a slight upward bulge below the mountain, but lying at a greater distance below the ground surface than for the surrounding, lower-elevation terrain), since we may expect that the mountain would have run relatively “dry” (due to gravity-driven outflow) the last time its internal temperature was high enough for water to be liquid (similar to how we have to drill down deeper to reach the aquifer when we are on top of a mountain on the Earth, even though there is an upward bulge in the aquifer under that mountain).  So the net result is that the cryosphere should be unusually thin inside/under the mountain, and if we have an extended period of surface temperatures above freezing (quite likely coinciding with an increased subsurface thermal gradient), we may expect the cryosphere to be exceptionally subject to melting at this location (assuming the temperatures held long enough for significant warmth to soak into the core of the mountain, perhaps hundreds or thousands of years).  In fact, a warmer regime that would not even begin the melting of the broader cryosphere (due to the cryosphere’s high thermal inertia resulting in turn from its thickness/mass AND from the high latent heat required for the phase change of such a thick column of ice), could well melt through the thin portion of the cryosphere inside Mt. Sharp, a process which would also be favored by the equatorial and below-datum location of Gale Crater, which raises the average temperature in relation to most portions of the Martian surface, and which would result in an unusually-thin cryosphere to begin with in the Gale area.
              Brine (that is, water that is saturated with salt) has a freezing point far lower than that of pure water, specifically minus 21 Celsius in the case of sodium chloride brine, but minus 50 Celsius in the case of calcium perchlorate, which in fact is a temperature below the current day-to-night average recorded by Curiosity (which is around minus 40 Celsius), so even a perchlorate concentration below saturation would permit liquid water under the current surface temperature regime (and with depth below the surface, the temperature is going to do nothing but rise).  And calcium perchlorate salt is known to be a very abundant compound in the Martian soil, making up .5 to 1 % by weight, so we may expect this highly water-soluble compound to be a significant “contaminant” in any liquid water we find on or within Mars (or any frozen water on or below the surface that was once in a liquid state, but NOT atmospherically-precipitated snow and frost on the surface, or glacial flows resulting from those, as such compacted snow would consist primarily of distilled water).  But wherever such contamination occurs, it will serve to greatly lower the melting point, depending on the concentration (indeed, it appears that droplets of perchlorate brine were seen on the legs of the Phoenix polar lander). 
              It has recently been determined that the globally-averaged thermal gradient within the crust and mantle of Mars is lower than previously believed, meaning that the interior of Mars is colder than previously modeled, and that (all else being equal) the cryosphere may reach twice the depth previously believed [27].  However, the recent relevation concerning the abundance of perchlorate “antifreeze” on Mars and its presumed “contamination” of surface and sub-surface water has largely negated this projection of a deeper cryosphere, and it may be that at low-elevation equatorial locations (such as Gale), the top of the liquid aquifer is only a mile or two below the surface.
              And if the central peak of Mt. Sharp and the subsurface of the crater floor were a nexus of hydrothermal activity early in its history (as is believed to happen in all large impact craters), then we may well have an unusually high salt concentration in subsurface fractures which could also be conduits for water, and so it could be that the portion of the cryosphere below Mt. Sharp is currently composed of slushy, semi-frozen ”salt water” that is almost warm enough, under current conditions, to permit upward migration, and any substantial and prolonged period of warmth could send it above that threshold.  In fact, conditions could normally be at a “tipping point” such that any increase in subsurface temperature, as little as a degree or two, from either top-down (surface climatic) or bottom-up (deep magmatic) sources, or better yet from a combination of the two, could send conditions over the threshold that would result in a reactivation of the artesian hydrant.
              Because of the thermal inertia of a planetary crust and mantle, any subsurface thermal excursion caused by a rising magma plume will take millions of years to develop and dissipate, and it may be that during any time of activity in the Elysium or Cerberus Fossae complexes, conditions at Mt. Sharp are for an extended time ”primed” for an outflow, but the final ingredient that is needed to allow a new aqueous discharge is a radical thermal excursion in the surface climatic regime, so as to melt the near-surface ice blocking the plumbing in Mt. Sharp, and this is something that has happened only rarely in the Amazonian era (possibly as few as two times).
              Increased subsurface warmth by itself might be inadequate to re-activate the Mt. Sharp artesian hydrant, because the top of the cryosphere would still be more-or-less exposed to the frigid surface conditions.  And conversely, surface warmth (by itself) would likely not last long enough to melt through even a thin cryosphere, hence my suggestion that it is only during a rare confluence between regional subsurface geothermal activity and a strong surface thermal excursion, that we experience a reactivation of the artesian hydrant.
              But what would happen to such water from such a source if it reached the surface, would it simply boil away as is generally expected for water exposed to the surface of Mars?  The “triple point” for pure water occurs at a pressure of 6 millibars (the conditions of pressure and temperature where frozen, liquid, and gaseous water all exist in equilibrium, and representing the lowest pressure at which a liquid phase is supportable), and since Gale Crater is significantly below the Martian “datum”, the atmospheric pressure there is typically about 8 or 9 millibars, which would permit a liquid range (for pure water) of several degrees.  But having the water in the form of a perchlorate solution could result in a liquid range of at least 50 degrees under the current pressure regime of Gale Crater.  And in a high-obliquity stage of Mars’ orbit (where the planet is almost on its side, and each polar cap would point almost directly at the sun for much of the year) we would expect to vaporize all the dry ice at the poles, which would nearly double the atmospheric pressure [28], increasing the liquid range still further (and also providing an increased greenhouse effect).
              Indeed, surveys of the current north polar cap have shown it to be only 4 million years old, and to consist of periodic layers of ice and dust laid down from the start of that timeframe [28], permitting age-dating much like the rings of a tree.  So just prior to that timeframe (and probably on a number of separate occasions in Amazonian times), the volatiles of Mars were elsewhere on the planet, likely doing more interesting things, and presumably constituting carbon dioxide and water vapor in the atmosphere, and water ice (and even limited liquid water) on the surface.  And not coincidentally, that 4 or 5-million-year-ago range is also believed to be when Mars experienced its most extreme cycle of obliquity in recent geologic history, with an axial tilt of approximately 40 to 50 degrees, which would have caused the “melting” of the polar caps and prevented their redevelopment until the obliquity had greatly lessened [30,31].  And it is further believed that such extreme cycles happen every several million years on average, and in some cases have involved an axial tilt as high as 80 degrees [32].
              So during portions of such a climatic cycle, we can expect conditions at Gale to permit the melting of (at least salty) water and its flow over the surface, and even pooling for a limited time (and although the evaporation rate would be fairly high, that is likely to be mitigated by the formation of an ice cover, which would be encouraged by the fact that evaporation is a cooling process).  And if we had a coincidence with major volcanic activity (or an asteroid or cometary impact, all of which could increase the surface pressure and thus temperature), then all the better for generating conditions which could not only support stable surface water, but also bring about substantial melting of the thin cryospheric layer inside Mt. Sharp, which if assisted by increased warmth from below, could once again permit direct outflow from the pressurized aquifer.  And so I am suggesting that one such orbitally-forced warmer and wetter cycle coincided with the Cerberus Fossae surface eruptions and associated deep magmatic activity, all of which combined synergistically to allow a revitalization of the Mt. Sharp artesian hydrant, as well as conditions that permitted the existence of a relatively short-lived but fairly deep lake in Gale Crater (in other words, during the radical thermal excursion and presumed higher atmospheric pressure that is recorded in the periglacially-derived landscape of the nearby Athabasca Valles).
              And so it may be that we have the misfortune of visiting Mars during one of its less-interesting phases, and there have been comparatively recent times when the planet was warmer and wetter and hosted a thicker atmosphere, and even featured flowing rivers and playa lakes at some spots on the planet (especially low, near-equatorial locales such as Gale Crater), and in so doing boasted more than a passing resemblance to Earthly conditions.  And while such phases will likely occur again, sadly it is also the case that they are becoming more and more infrequent (and less intense) over geologic time, as the internal heat of the planet gradually winds down.
              In any case, the (antiquated) notion of Mars currently being a “dead” planet no doubt had its genesis in the disappointing photos that were returned from the first Martian flybys, which only captured ancient heavily-cratered terrain, giving the impression that Mars was “Moon-like” and that it shared a similar history.  And even though a more dynamic geology was discovered by Mariner 9, the impression (or bias) persisted that such features (such as fluvial channels) were simply flukes or exceptions to the general rule, and that all the visible surface features of Mars, including all signs of active liquid water, were to be dated in the billions of years ago, and that (like the Moon) the planet had been (mostly) frozen in time since shortly after its formation, a position that has been disproven by recent observations, and a viewpoint that I have argued against on the basis of the features of the Gale/Sharp complex.  And although there were a number of fascinating locations on Mars vying to be the landing spot for Curiosity, it is now obvious that NASA could not have picked a better spot! 
              Mars is continually losing atmosphere to space, and it needs to be periodically replenished by volcanism, and we can assume that the atmospheric pressure fluctuates wildly and, all else being equal, is more-or-less in proportion to the level of recent volcanic activity.  And during extended lulls in volcanic activity, the planet may reach a point of having even less atmosphere than what we are seeing today (also, it is suspected that when the tilt of the axis is near zero, as it is every few hundred thousand years, that the polar caps are even more efficient cold traps for carbon dioxide and water than they are now, and the atmosphere ”collapses” to a surface pressure only a fraction of its already-meager value).
              It is commonly stated that Mars has suffered a high rate of atmospheric loss over time because of not having a magnetic field, which (it is stated) results in the top of the atmosphere being directly exposed to the solar wind, which continually strips it away.  But while this may be a significant factor, it is not the primary one….if lack of a magnetic field destroys atmospheres, how then does Venus (which lacks a magnetic field) have an atmosphere almost 100 times as dense as the Earth’s?  And why has not the Earth lost its atmosphere over geologic time due to the periodic reversals in its magnetic field, where the field strength drops to zero?  And in fact, the magnetic field is a double-edged sword, because it serves to funnel charged particles into the atmosphere at the poles (hence the aurorae), putting that part of the atmosphere under even greater stress than it would otherwise be.
              The primary reason that Mars has had difficulty maintaining an atmosphere, is the small mass of the planet (1/10 that of Earth) and the resultant lower escape velocity (5.0 km/sec vs. Earth’s 11.2), causing Mars to have a smaller separation between the average speed of upper atmospheric atoms and the planet’s escape velocity.  And in a gas (in accordance with the Maxwell–Boltzmann distribution), plotting a graph of the number of atoms vs. the speed produces a bell-shaped curve, where at a given temperature, atoms cluster around one speed (which in fact IS their temperature, which is a measure of average atomic/molecular speed), but this also means that there is a small fraction of atoms (due to random collisions and absorption of photons) that greatly exceed the average speed, and will in fact exceed the escape velocity of the planetary body, and so are lost to space (see graph below).  This is referred to as “Jeans Escape,” and Earth also loses some atmosphere by this means, but at such a negligible rate that its atmosphere has a half-life of many billions of years, whereas the half-life of the Martian atmosphere would likely be measured in the hundreds of millions of years (here’s another way of looking at it: kinetic energy increases with the square of the velocity, and Earth’s escape velocity is over twice that of Mars, so it is over four times harder for an atom to escape from the Earth than Mars, and so it rarely happens on Earth, whereas it’s happening all the time on Mars).  And so if Mars had not experienced periodic volcanic activity over geologic time, which serves to (at least partially) replenish the atmosphere, by now it would be almost as airless as the Earth’s Moon.   
              There have been previous estimates that Mars is currently losing atmosphere to space at the rate of approximately 100 tons per day, but the figure should be refined by the current MAVEN mission.
              Another recently popular idea is that Mars lost most of its atmosphere because of the huge impact (from a Pluto-sized body) that created the Northern Lowlands, since such an impact would have blasted most of the primordial atmosphere into space.  But this is irrelevant for current conditions, because even if that collision had not taken place, most of the original atmosphere would have been gradually lost to space anyway, and the planet would have still needed to depend on volcanic activity to maintain what atmosphere it has.  In fact, it has been previously estimated that, given the amount of volcanism that has left a visible record on Mars, a total of 1.5 bars of carbon dioxide has been outgassed over time (1.5 times Earth atmospheric pressure equivalent), assuming that the gas composition of Martian lavas is similar to that of Hawaiian basaltic lavas….but this is an estimate that is likely to only rise in the future, now that a number of ancient craters that were assumed to be impact craters have recently been reinterpreted as supervolcano calderas [33].
              Mars may very well have acquired microbial life early in its history, either originating in situ, or else seeded from space, inside meteorites blasted from the Earth or other bodies.  But because of the harsh surface conditions, principally due to the thin atmosphere, and with relatively habitable episodes being increasingly fleeting over time, it is unlikely that life has maintained a foothold on the surface of Mars, and if there is life on the planet, it resides deep underground in the safe haven of the aquifer, where conditions could be stable for billions of years.  And in fact, I would be surprised if there is NOT currently live bacteria in the Martian aquifer, as there was presumably an exchange of meteoric material in the early solar system that would be capable of transplanting viable microbes, and some of these would no doubt be the anaerobic, “rock eating” type that derive energy and nutrients from the inorganic minerals in the deep subsurface, and are found living at depths of up to 3 miles on the Earth, under conditions likely very similar to the deep subsurface of Mars, and thus would be a “natural” species to transplant between planets.
              Although it awaits further analysis and confirmation, the recent announcement of a possible discovery of biologically-related carbon-rich spheres and microtunnels in the 1.3-billion-year-old Martian meteorite Yamato 000593 is consistent with this scenario, in that the microscopic features observed could conceivably be the product of ”rock-eating” bacteria.  As stated by the authors, “… textural and compositional similarities to features in terrestrial samples, which have been interpreted as biogenic, imply the intriguing possibility that the Martian features were formed by biotic activity” [34].
              Below, from White et al. [34], we see a scanning electron microscope image of a layer in the meteorite, composed primarily of the water-generated clay mineral iddingsite, with the spheres in the red circle showing twice the carbon abundance as the background level (circled in blue).

              Next we see a thin polished section of the meteorite, showing the microtunnels:

              In fact, it has been cogently argued that life on Earth may very well have originated on Mars by means of such a vector, as the smaller planet would have cooled off faster and been able to support liquid water on its surface many millions of years sooner than the Earth, and with the high flux of asteroidal debris throughout the early solar system, causing much material to be rapidly exchanged between at least the inner planets, such life would have no doubt been seeded to Earth just as soon as conditions were cool enough to support it, giving such life a jump-start over any of indigenous Earthly origin.  And taking this line of thinking a step further, it has even been argued that our lineage of life may have first appeared on Ceres, a water-rich “dwarf planet” which would have cooled off to the point of supporting liquid water even before Mars, and which would have been a perfect candidate, both in terms of its orbital position and low escape velocity, to have viable biological material blasted from it that would subsequently seed the planets of the inner solar system!
              Even given the current minimal Martian atmosphere (or less if the planet were to go for a long period without volcanism), the cold trap of the upper crust prevents most of the planet’s water from being lost to space, at least for a very long time span (much as Jupiter’s moon Europa operates, with a liquid water ocean protected from the cold vacuum of space by the icy crust above), a fact that demonstrates how important it is that we obtain a sample from the deep subsurface if we are to assess the biological status of the Mars.  And since heavy drilling equipment is out for the foreseeable future (equipment that could very well need to drill down several miles), our only option is to take advantage of a natural shortcut, which I am proposing Mt. Sharp may very well be.
              And if we do obtain samples of intact biology from Mars, either live or very well preserved, it will be of the utmost interest to decipher its chemistry and map its genetics, to see if we can answer the question of whether it and Earthly life have a common origin!                                                
              It has been proposed by a team of researchers from Princeton University and the California Institute of Technology [35] that Mt. Sharp is basically a big dune, having been created by wind rather than water (and with the outer portion of the crater volume NEVER having been filled with sediments).  Their “SWEET” hypothesis (Slope Wind Enhanced Erosion and Transport hypothesis), which has recently received considerable press coverage, states that a convection cell develops in response to the daily warming cycle experienced by the crater floor, which causes katabatic winds to sweep down the inner crater rim and pick up material from the outlying parts of the crater, whereupon they rise as they move towards the center of the crater, depositing their load in the stagnant-air zone in the center, and over a span of several billion years this process gradually built up the mountain we see today.  And this hypothesis may indeed take account of some of the anomalous features of the Gale/Mt. Sharp complex (namely the existence of Mt. Sharp and the fact that the outlying parts of the crater floor are well below the grade of the surrounding landscape), but I will argue that there are clearly-visible features of the complex that are incompatible with their model, and that there are alternative explanations for what they present as evidence in support of their model. 
              First off, there is at least a prima facie implausibility to what they are proposing: looking at Mt. Sharp, there are no signs of ongoing deposition, in fact there are no signs of deposition (save for channel fill and some relatively small aeolian bedforms) in any time-frame but the very distant past, and instead there is a landscape basically carved out by erosion (looking at first blush much like the landscape of the America’s desert Southwest).  So where is the hand of that process today, which they envision?  Why is not deposition continuing to occur?
              The authors of the study state that their model predicts that, once the mountain reaches a certain critical size and shape, the mountain modifies the winds such that they become a net erosive force, and what we are seeing today is the late end stage of that process.  But would we not expect that, if the erosive winds reduce the mountain to a considerably smaller size, that the depositional process would resume at some point, mimicking its pattern of growth when the mountain was initially of that smaller size?  In natural processes of the sort proposed, what we normally see when they “max out” is a movement towards a steady-state equilibrium, either asymptotical in nature, or else oscillating about a point of equilibrium, where forces (eventually) balance each other out (in this case, the forces causing erosion and deposition).  But what we see instead in the Gale/Sharp complex, is a landscape that has been deeply and systematically ravashed by erosion for a VERY long time, with graphically exposed layers that have been very deeply cut, carving out buttes and hummocks and yardangs and ridges and scarps and canyons, with no indication that this process has slowed down or ever will slow down (let alone reverse itself), as the sparse impact cratering on the flanks of Mt. Sharp show that it is a surface experiencing ongoing intense erosion.  And considering the depth of this relentless erosion, and its asymmetrical nature as can be seen in a topographic map, with massive chunks of the mountain missing, it would come as no surprise if the total volume of Mt. Sharp has been reduced to a fraction of what it once was….and yet there is no sign of any recent wind-generated deposition except for a thin dust cover on the upper slopes.  So I would argue that the burden of proof must be on the authors, to explain how such a “counterintuitive” scheme could work in the real world, which is at odds with what we normally observe across a broad range of natural physical and geological processes.
              Below we see two fine examples of the graphic erosion-carved landscapes of Mt. Sharp.  The first is from the Mars Reconnaissance Orbiter, showing a spur defined by dramatic cliff-bench scarps on the east side of Mt. Sharp, and the second is an extreme zoom shot by Curiosity, of the heavily-dissected, butte and knoll-riddled foothills of Mt. Sharp in the area of the Northern Channel, taken while the rover was still several miles away (note the shadowed canyon of the Northern Channel in the upper left portion of the landscape):

              As far as any direct evidence that the mountain was once far larger, there is indeed an apparent ”outlier” of the Lower Formation, an eroded low-relief layered exposure, in association with a medium-sized crater (quite possibly an exhumed paleo-crater), and largely covered by sand dunes, which extends to a point about 3 miles from the western margin or Mt. Sharp.  Which not only shows that the mountain was far wider at one point in time, but that, by extension, the strata making up Mt. Sharp likely extended clear to the crater wall (the furthest westward extent of this outcrop is only about 6 miles from the first foothills of the crater rim in this portion of Gale Crater).  Below is a pair of figures from Anderson and Bell [36]:

              There are also stratigraphic features of Mt. Sharp which are difficult, if not impossible, to reconcile with the SWEET hypothesis.  The Upper Formation is of a totally different character than the Lower, both in terms of composition (it has spectral characteristics similar to globally-distributed Martian dust, not showing the hydrated clay and sulfate signatures of the Lower Formation), and also in terms of its depositional history: the entire Upper Formation was evidently laid down in a very short time, probably only a few million years for the entire 6,000-foot-thick column, on the basis of the almost complete lack of exhumed paleocraters (in other words, sedimentation was occurring so fast that there wasn’t a chance for any significant number of impacts to occur, while a comparable thickness of Lower Formation strata has many paleocraters, and evidently the older formation took at least a hundred million years or so to be laid down).  Whereas, based on the SWEET model, we would expect a more “uniformitarian” pattern of deposition at all levels of Mt. Sharp.
              The visible bedding morphology is also very different between the two formations.  The Upper Formation beds exhibit strikingly obvious cross-bedding when exposed by erosion, as we see below in high-resolution photos from MRO of the terrace-like or cliff-bench layers of what I call the ”stands.”  So the conclusion is that the Upper Formation consists of a stack of lithified dune fields, which is consistent with the fact that they were deposited above the minus-6,000 foot mark, which is believed to represent the high-water mark of the former Northern Ocean (and thus the maximum water level in Gale Crater):




              Whereas the beds of the Lower Formation are more-or-less horizontal, are laterally continuous over an extended distance, and lack obvious cross-bedding in most facies (and where cross-bedding does occur, I would interpret it as the product of subaqueous turbidity currents, as I will discuss later), and these are properties consistent with being lacustrine in origin.  In fact, such a lacustrine origin is strongly suggested by the placid, orderly character they exhibit when exposed in cross-section by erosion, as we see below in another shot of the area near the Northern Channel.  And yet the SWEET authors are claiming that these two radically different morphologies formed through the same mechanism (aeolian deposition) in the same setting. 

              It is also unclear how the angular unconformity between the Lower and Upper Formations can fit within the SWEET model, which again proposes a relatively steady-state mound-growth process controlled mostly by the structure and dynamics of the Gale Crater complex itself, and not so much by regional and global climactic changes and disruptions that could radically alter, and in fact temporarily reverse, the Gale depositional milieu.
              Additionally, the apparent cross-bedding of the Upper Formation beds suggests active sand dunes, and it is difficult to picture how a dune field could occupy the top of a mountain (whereas, in all previous known examples, dune fields occupy horizontal and sub-horizontal terrains), much less how sand would make it up to that elevation…..Martian winds are capable of transporting sand-sized grains, as we see in the active dune fields, but this is horizontal transport, and even on Earth, we don’t normally see such large and heavy particles being transported upslope over extended distances (only fine dust, which does not characteristically form dune-shaped bedforms).  Le Deit et al. [6] have incorporated SWEET within their models, and state that at one stage of the development of Mt. Sharp, ”Remobilized deposits such as sands accumulate at the top of Aeolis Mons forming the (Upper Formation).”  However, simply stating that this happened does not make it so, and provides no mechanism to explain how such granules could have been transported en masse to such a location.
              Furthermore, cross-bedded dunes imply the presence of fairly strong horizontal winds, which are not only incompatible with the basic model of SWEET (which postulates that the area occupied by Mt. Sharp is a stagnant air zone in which particles fall out of the air and accumulate), but such horizontal winds would also serve to drive any such loose sediment off the top and down the side of the mountain, preventing any such accumulation.  But the presence of cross-bedded dunes is perfectly compatible with the model of Gale Crater being totally filled with sediments, where during the deposition of the Upper Formation, the upper exposed surface of the crater fill was a 90-mile-wide “sand sea” consisting of air-fall material (such as volcanic ash) brought in from outside of Gale, as well as material eroded from the crater rim and central peak.
              But the single biggest problem with the SWEET hypothesis, is that it is incapable of accounting for the existence of the large channels/canyons on the west side of Mt. Sharp, as there is no conceivable mechanism in their model for generating a flood of water inside, or on top of, what is basically a big sand dune!  Heavy rain over an extended period might do the trick, but as we saw earlier, there is not believed to have been rain on Mars during the timeframe in question (we’re more than a billion years too late), and even if there were such rain, it apparently didn’t fall on Mt. Sharp, as the surface of the mountain is sufficiently indurated to have, of necessity, developed tell-tale surface runoff gullies (which it lacks).  And as we saw earlier, similar arguments hold against ice/snow-melt being the water source.
              A related problem with the SWEET hypothesis is that the authors state that Mt. Sharp is currently in a late-stage erosive regime, but if this is so (and for the sake of argument, we grant the plausibility of the switch from wind-generated deposition to wind erosion that they hypothesize), how then do we account for the delicate, easily-erodible channel features, such as the delta at the end of the Northern Channel?  Such features, in their model, should have been worn away long ago…..being shallow surface features and “drape” formations, they are by default the most recent features of Mt. Sharp, and if the mountain were to erode, they would be the first to go…..and in their model they certainly are not positing recent conditions that could have created them (which don’t merely involve small-scale flows of water, but ones sufficient to carve a “Grand Canyon,” transport boulders several meters wide, and form a 2,800-foot-deep lake!).
              I won’t delve too deeply into the ongoing debate as to whether the Northern Lowlands were occupied by an ocean at the time that Gale Crater formed, except to state that I find the evidence of paleo-shorelines and of paleo-deltas that debouched into that ocean to be fairly convincing, much as the similar evidence of deltas and topographic benches within Gale demonstrate the existence of the Second Generation Lake (plus, an intriguing line of evidence has recently been brought to light [66], the apparent discovery of scour marks from oceanic currents in the Northern Lowlands).  And if the prior existence of such an ocean is proven beyond any doubt, it would strongly count against the SWEET hypothesis, as Gale Crater would be expected to share a common water level with such an ocean, and certainly not be dry (or nearly dry) as SWEET proposes, since the shoreline of that ocean would either overlap with Gale or else be so close that groundwater movement would certainly fill Gale (even if my proposed artesian hydrant didn’t), as the lowest point of Gale is about 9,000 feet below the putative water level of that ocean.
              Even neglecting the oceanic hypothesis, how plausible is it that Gale Crater, with a floor up to 15,300 feet below Mars datum and one of the lowest points on the planet, and with a surrounding topography highly favorable for the formation of a powerful hydraulic head, would be dry or nearly dry (as SWEET hypothesizes), when massive floods on the planet have been sourced from aquifers located at much higher levels, as late as mid to late-Amazonian times (at least 2 billion years after Gale formed)?  Surely seepage (even if there was no direct inflow, which seems improbable) would have filled Gale in a timeframe (late Noachian) that has left very extensive records of surface aqueous activity on Mars, when the planet had far more internal heat, and a very thin cryosphere coupled with a near-surface aquifer that surely would have intersected the walls of a crater such as Gale.
              Above is the theoretical argument for the existence of a large and deep lake in the early history of Gale Crater, but in addition we have the argument from historical precident: namely, the fact that there is compelling evidence for at least one massive lake in Gale, in the relatively recent geologic past (as shown by the Northern Channel’s delta deposit and other features in the Gale/Mt. Sharp complex), and it is difficult to imagine how, if the hydrological system of the Gale/Mt. Sharp complex were that powerful, at that late a date, that it would have been incapable of generating at least as massive a lake in the distant past, when conditions for generating (and sustaining) such a lake were far more favorable.
              The authors of the SWEET study argue that the fact that layers at the outskirts of Mt. Sharp appear to be inclined by several degrees, is evidence in support of their hypothesis, in that their model of wind-generated deposition predicts such an inclination (and they also claim that this inclination is incompatible with any alternative depositional model).  See below picture, from Kite et al. [35]:
               However, I do not believe that the inclination of the beds supports their SWEET hypothesis, for two reasons.  First of all, the measured beds are precisely the ones that are the poorest choice for fitting within their model or lending support to their model, because they are the ones that appear to be (on the basis of their mineralological makeup) water-deposited in a lacustrine environment, not wind-deposited (as SWEET’s proposed mechanism for the creation of all of Mt. Sharp’s layers involves).  If the authors are measuring the tilt of beds, they should instead sample ones that are unambiguously wind-deposited (as are the layers above around 9,000 feet from the base of the stratigraphic column).  But given the vagaries of slope inherent in wind deposition and cross-bedded dune formation, we are unlikely to have any consistent pattern of bed inclination in the case of aeolian processes, either within the context of SWEET or more traditional models.  This is unlike deposition in standing water, where gravity is the primary determinant of bedding-plane formation, and unless other factors are coming into play, results in horizontally-oriented beds….whereas aeolian bedding is almost never horizontal, and can lie at any angle between that and the maximum angle of repose of 30 to 40 degrees.  And indeed we see this at work in the bedding planes of the Upper Formation layers, which are tilted by 5 to 10 degrees to the north-northeast, which is inconsistent with the predictions of SWEET (which predicts an approximate 3-degree radially-outward tilting of the mound layers), an observation which has no apparent explanation within the context of their hypothesis, and in fact constitutes an additional stratigraphic anomaly with respect to SWEET.  So it would appear that, in principle, Mt. Sharp is incapable of providing any beds that have the potential of lending support to their hypothesis. 
              The second reason I do not believe the measured inclination of the beds supports the SWEET hypothesis, is that the authors have overlooked an important factor that likely explains the tilting of the Lower Formation layers: namely, the presence of a central peak “island” in the middle of the lake, a geographic feature that appears to have been far larger than conventional models of central peaks (based primarily on Lunar examples) would predict. Our indication that this is the case, is the fact that what appears to be the central peak, is jutting out the top of Mt. Sharp to a height of 3.4 miles from the crater floor, and exceeding the height of most portions of the crater rim by several thousand feet:


              There is an obvious anomaly here, because this is several times higher than we would expect on the basis of other craters of similar size [37], and so there is definitely something unusual about the central structure of Gale Crater.  Gale may be a variation of “complex crater” that features not only a central peak, but associated uplift features in the central area of the crater floor, and in fact the central peak may be just the tip of the iceberg, representing the high point of a surprisingly broad and massive central uplift, or possibly a central peak surrounded by a raised ring or rings, and which in turn serve as the core of Mt. Sharp, with the sedimentary deposits constituting a relatively thin veneer on this basement (at least near the outer edges of Mt. Sharp).  And I would suggest that the structure of Gale may resemble that of Utah’s Upheaval Dome, which features concentric uplift rings around the (highly eroded) central peak.  Uplift Dome was previously believed to be a salt dome, but it has now been established as an impact feature, thanks to the discovery of shocked quartz and impact melt (photo courtesy of Earth Impact Database, University of New Brunswick):


              I would suggest that the difference between Upheaval Dome and Gale Crater, on the one hand, and “classic” Lunar craters on the other, may be that the former two involved impacts in a planetary crust with an active and saturated aquifer, rather than a “dry” crust like the Moon’s, and the large amount of water resulted in extensive hydraulic fracturing, and uplift due to steam explosions.  And this is supported by the fact that Upheaval Dome is only 3 miles across, a size that would be insufficient to generate a multi-ringed “complex” crater in a Lunar setting. 
              In addition, at least in the case of Gale, the large amount of thermal energy absorbed by the vaporizing water may have allowed more of the rock surrounding the central peak area to remain solid rather than melting to form a magma lake, which could contribute to a high and broad central uplift.  Obviously, further study of the geophysics of crater formation is in order, to define how radical differences in crustal compositions and water content affect the resulting impact structures, but of course there is a relative scarcity of potential examples to study on the Earth due to the high erosion rate.
              In any case, if there is a major uplift in the center of Gale Crater (as suggested by the height of the central peak), we may expect that the sedimentary beds would have been laid down on a surface that was sloping towards the outer portion of the crater, and so it is natural that we would see the bed inclinations that we do.  Deposition in a shallow playa lake will of necessity result in horizontal beds, but if deposition occurs subaqueously in a deep lake that has a gently inclined floor, the resultant sedimentary beds will share that same inclination.
              Also, if the central peak and uplift constituted a large (both high and broad) island rising out of the Gale Crater lake, we may expect it to have provided considerable eroded material to add to the sediments, and we may expect this to result in a thicker section of such sediments near the shoreline of the island, with the thickness of the beds gradually pinching out towards the center of the lake’s waterway, generating the illusion of greater bed inclination than is actually present (but which in itself may be considerable, due to a radially-directed outward tilt of the uplifted basement).
              Plus, the presence of the central peak and island could have served as a catch point for wind-blown sediments, whereby the physical barrier to airflow would have caused a local drop in windspeed, in turn causing the winds to dump much of their load, and thus providing additional sediment to the central area of the crater lake.
              In addition, if the waterway between the island and the crater rim was sufficiently wide, and the wind speed sufficiently high, we may have had wind-generated waves, which would have served to push sediment up against the shoreline, tending to produce a wedge-shaped depositional cross-section for near-shore beds.  And an interesting feature of waves when encountering an island, is that they tend to refract around both sides, so we may expect a consistent pattern of onshore waves on all sides of such an island, regardless of the initial wind and wave direction (but of course, in such a mileau we may also expect appreciable sediment movement in the reverse direction, due to turbidites carrying material downslope from the island’s shoreline).
              So as with any shoreline, we may anticipate an appreciable slope (both upwards from the lake’s surface onto the dry land of the island, and downward below the water’s surface, much like a continental slope in miniature), with both basement and overlying depositional factors playing their parts, and we may naturally expect these factors to impart an angled slope to the bedding as we progress away from the island, towards the center of the waterway.
              It is also possible that the initial central peak may have augmented itself by serving as a hydrothermal vent, since the residual underground heat of impact is believed to persist for hundreds of thousands of years in the case of large impacts, and this heat could yield mineral-rich brines which would be discharged from the surface and deposit their load, building up “spring mounds.”  This process is believed to have occurred on a more limited scale in Mars’ Vernal Crater, and occurs on Earth in volcanic settings, and if it occurred in Gale, would also be expected to have added to the slope of the bedding, both directly and indirectly (Rossi et al. [38] proposed that the entire mound in Gale crater is a spring mound, but this is unlikely based on the lateral extent and horizontal continuity of the beds making up Mt. Sharp, which instead suggest deposition in a crater-spanning lake, although a spring mound may lie underneath the visible bedding of Mt. Sharp, and of course my hypothesis also involves springs, but ones with a sufficiently large discharge volume to fill the crater with water, and not simply provide evaporate deposits).
              Simply extrapolating the downward slope of the visible central peak (and ignoring the possibility of a broader uplift and/or concentric rings) results in a projected island with a width of at least 20 miles (over a quarter of the bowl width of the crater), whereas in the model presented by the SWEET authors, this feature is totally ignored, and their diagrammatic cross-sections of Gale and the formation of Mt. Sharp show the central peak as a tiny, short, insignificant feature, a small fraction of the height of the crater rim and of the actual central peak height and width (and so the SWEET model actually misrepresents the structure of Gale Crater).  Whereas I would argue that this feature needs to be factored into any model, and that it is highly unlikely to be a coincidence that this anomalous central peak size and structure is paired with a unusual central mound of sediments.  The linkage, I am postulating, being that the enhanced central structure of the crater is the result of an impact in an active aquifer, and the central structure that was so produced, became an outlet for that very same aquifer.
              In the catalog of Martian craters, Gale is of a size that is transitional between craters typically hosting only a “simple” central peak, and ones with a outer peak ring approximately half the diameter of the crater, and if it does host such a ring, it may well define the curved northern edge of the Mt. Sharp sedimentary mound, and account for the radially-outward tilt of the beds at that edge.  And as pointed out by Allen, Dapremont, and Oehler [39], projecting such a hypothetical ring around the entire circumference of gale (see figure below), shows an alignment with an arcuate line of hills in the southeast quadrant of the crater (and, I might point out, in the SW quadrant as well, to a lesser extent), which may be outcrops of just such a ring:
              In fact, there is another piece of evidence that supports a model of there being a central uplift in Gale almost as wide as Mt. Sharp itself: to the NW and SE of the central peak, there are chaotic zones of large fractured blocks (please refer to the earlier plan photo of the central peak), somewhat of the character of shorter secondary peaks flanking the main central peak, and which have been interpreted as “megabreccia” (breccia consisting of kilometer-scale blocks) that are a remnant from the original formation of Gale Crater [1], when a fractured basaltic basement upwelled to form the central peak complex (but did not melt, although it was surrounded by a lake of impact melt which solidified to form the crater floor).  And if this is the case, then here we see a sampling of an uplifted surface that is considerably larger than even the central peak.  In the case of the better-exposed elevated region, the one to the SE of the central peak, the outer visible face experiences an elevation drop of about 3,000 feet over a linear distance of 18,000 feet (with a starting point 9,000 feet above the northern floor of Gale Crater, located 45,000 feet east-southeast from the summit of the central peak), and this surface slopes radially outward from the central peak, at an angle of about 9 degrees.  And if this slope is characteristic of the broader basement beneath Mt. Sharp, projecting this angle yields an uplifted basement approximately 35 miles wide, the best part of the width of Mt. Sharp itself (which measures about 55 miles wide).  And since, in  most cases, mountains and other uplifts decrease in their slope towards their outer margins, it is quite possible that this uplift is nearly as wide as Mt. Sharp, tapering down to a slope of just several degrees in its outer portions, in which case it could be adequate in and of itself to explain the several-degree radially-outward inclination of the beds near the edge of Mt. Sharp.  
              The pattern of convective, katabatic winds in Gale Crater that the SWEET authors hypothesize may indeed occur, and be responsible for the crater floor being scoured and eroded over time, rather than steadily accumulating sediment (especially in the “Yellowknife Bay” area and in the broader basin or “moat” directly to the north and NW of Mt. Sharp), and it may also account for the layer of dust that covers the upper reaches of Mt. Sharp (whereby fine sediments picked up from the crater floor are carried upslope and deposited on the mountain), but it cannot account for the formation of the mountain’s stratigraphic sequence, and the empirical observation it claims as evidence in support of itself (the inclination of the lower beds) could just as readily be explained in the context of a competing model.  
              The greatest usefulness of the SWEET hypothesis lies (ironically enough) in the fact that it has provided a wind-generated mechanism for explaining how the outlying portions of an original crater-filling mass of sediments could have been eroded away, which heretofore was a major difficulty for the older (and competing) model that Gale was originally filled with a crater-spanning lake that deposited crater-spanning and crater-filling lacustrine sediments.  Specifically, it provides a mechanism for eroding a basin to an ever-deeper level (in this case, a large paleo-crater to a point near its original floor), rather than the processes of erosion (and subsequent deposition) filling in a basin (which is the more typical outcome).
              In itself, my proposed explanation for the erosion of the outer portions of the Upper Formation (the fact that they were never lithified, due to being physically removed from the central peak’s source of lithifying water), does not fully account for the deep erosion experienced by the outer portions of the Lower Formation.  Although being exposed to the wind after losing their protective cap would have caused the lower formation layers to be incised, there is prima facie no accounting for the depth of their ultimate deflation (to a point one mile below the grade of the adjacent landscape outside the crater rim).
              However, once the outer portions of the Upper Formation had been removed, we can expect an erosive convection cell to be set up, basically in line with the one described in the SWEET hypothesis, as we would have a ring-shaped depression surrounding a central mountain, and we may expect daytime heating to cause katabatic winds to rise from the lower-elevation region, carrying entrained sediment, and if upper-level winds were to shear off the top of such a convection cell, there would be a net removal of material from the floor of the crater (and from the Gale crater complex more generally), which would likely produce very substantial deflation over geologic time intervals.
              Below we have a figure from NASA showing the wind patterns they predict for Gale Crater, and they are similar to the ones proposed by the SWEET hypothesis.  Although they are not identical to SWEET, in that they show daytime katabatic winds RISING against the inner crater wall (rather than descending), plus they show lateral movement of wind over the surface of the crater floor, which would present a highly enhanced erosive potential for the crater floor.  But in common with the SWEET hypothesis, they show winds rising over the slope of Mt. Sharp, and if these winds hold entrained sediment that has been eroded from the floor of the crater (which is empirically supported by the observation that the flank of Mt. Sharp has been subject to intense ongoing wind erosion, as demonstrated by the dearth of recent small impact craters and the high prevalence of yardangs, which implies in turn that the upslope winds are strong and have a high level of entrained erosive particulate matter), then there would be a net removal of material from Gale Crater over time, if these upslope winds typically encounter a horizontal wind shear after cresting Mt. Sharp.
              And such a mechanism, once it was initiated, would begin a positive feedback loop that would increase the strength of the erosive wind regime over time, as the floor of the crater would be progressively deflated and the differential elevation increased (and thus the differential diurnal heating, and the convective wind strength).  Plus, these phenomena were primarily occurring in a time when the Martian atmosphere was much denser than today, and thus potentially far more erosive.
              So I conclude that we have a readily-available mechanism to explain the erosion of sediments (originally filling Gale Crater) to a point well below grade, a mechanism (inadvertently) proposed in large part by the competing SWEET hypothesis, where SWEET was originally fielded in an attempt to provide an alternative explanation for Mt. Sharp’s genesis, in the face of what originally seemed to be intractable difficulties with the more traditional crater-filling lacustrine model of sedimentation.  Which is a rather interesting example of the old philosophical notion of “delectical reasoning,” whereby the original idea (the “thesis”) leads to its “antithesis,” which are subsequently combined into a new “synthesis.”
              Besides the tilted beds (for which we have alternative explanations) and the erosion-below-grade problem (also apparently explained), the other seemingly intractable problem with the traditional model of crater-filling sedimentation, was the survival of the center of that sedimentary mass (in the form of Mt. Sharp), which I have explained by means of my artesian hydrant model (and which, as a side benefit, also neatly accounts for the geologically-recent canyons eroded in the flank of Mt. Sharp, as well as the existence of a Second-Generation Lake).
              There are also some examples on Mars of filled craters which are at an initial to intermediate stage of exhumation, and which demonstrate that the developmental sequence that I am claiming for Gale does in fact occur on Mars, and by implication there is no need for invoking an ad hoc process to explain Mt. Sharp, such as SWEET’s formation of a giant sand dune in the center of a crater.  The prime example is Asimov, a 52-mile-diameter crater located at 47 degrees south latitude and 355 degrees west longitude, which was obviously filled to at least the crater rim, but which is now being exhumed, mostly along the outer edge of the bowl, and if it follows a path like I believe Gale followed, it should eventually be left with an isolated mound in the center (image courtesy NASA/USGS):
              Gale likely looked similar to this in the mid Hesparian era, shortly after the “Great Erosive Event” began, forming pits along the inside of the crater rim in the less-lithified outlying sediments, with those pits progressively enlarging and deepening until they formed a continuous ring, and with the convective wind pattern set up by that topography serving to feed the further development of erosive winds, in an ever-accelerating process.  And in fact Gale’s ring is still expanding to this day, with Mt. Sharp growing ever smaller, although the rate of erosion has greatly diminished due to the thinning of the Martian atmosphere over billions of years.
    End of Part 1.

  1. galecratergeology1 reblogged this from galecratergeology1