Water on Mars now or in the past?

What is the evidence that Mars has water now and had more in the past?  Certainly, the water ice seen at the polar caps, the clouds seen in the atmosphere around the large volcanoes, and the fog and frost that is seen in the low elevation valleys in the Martian morning, and some evidence for gullies in canyons, testify to the presence of some water on Mars today.  But not much water.  Is this all the water Mars has and ever had?

Mars has surface geologic features that have been termed valley networks.  These features look like river valleys on Earth, i.e., fluvial features formed by water flowing for long periods of time, slowly carving out valleys out of the surrounding rock.  Valley networks appear to require a warm, wet Mars, a Mars in which water  flowed on the surface (without freezing or evaporating quickly) for 100s of millions of years.

The other surface feature apparently indicating the presence of water on Mars in the past are the outflow channels.  Outflow channels have been interpreted as the result of catastrophic melting or flooding which consequently carved out deep, wide channels.  This interpretation of outflow channels does not require that water was flowing on Mars for extended periods of time, only that large volumes of water were present and available for flooding when released.

Valley networks are found only on the ancient cratered terrains of Mars (southern hemisphere; ages of 3.5 - 4 BY). The valley networks, first observed by Mariner 9 in 1972, provided the first evidence that Mars had liquid water running across its surface 4 billion years ago.

Outflow channels are largely in the northern lowlands, near the equator, on much less cratered, and therefore much younger, terrain (perhaps of ages 2 BY).

Map of outflow channels and valley networks (from Mike Caplinger, Malin Space Science Systems).  On this map, outflow channels are colored red, and valley networks are colored yellow.  Note that the Valley networks are dominantly in the southern hemisphere and all are in ancient cratered areas.

Outflow Channels

The term outflow channel was coined in 1975 to denote the large channels, several tens of km across,  that start full size at discrete sources.  Outflow channels have few tributaries but often branch downstream (the opposite pattern of a river system).  The term channel implies a conduit for the flow of water that was filled with water.

This large field image shows several outflow channels in the vicinity of Valles Marineris: Kasei Vallis region.  Most outflow channels are more than 100 km wide, and often are 1000s of km long.

The Tiu Vallis region emerged from the region called Hydaspis Chaos, which appears to be a region about 100 km x 200 km that released an enormous quantity of water that then carved out the Tiu valley.

Features known as teardrop islands, here seen in Ares Vallis, are generally thought to be higher elevation regions not carved out by the enormous outflows because the walls of craters protected the land immediately downstream.

The Mangala Vallis outflow channel appears to have started in a deep crack (a graben).

When did the outflow channels form?

All of them date from the Hesperian period on Mars, which postdates the era of the great bombardment and probably covers the time period from about 2-4 BY ago.  They required enormous volumes of water likely locked in permafrost or in underground water reservoirs.  The water would have been released catastrophically by an impact or the breakage of a permafrost seal over a water reservoir by tectonic event.  Whatever the mechanism of release, almost certainly, climactic conditions must have been similar to those on Mars now.

How did outflow channels form?  Probably through the sudden heating and catastrophic release of large quantities of water stored in or near the surface.  The heating might be related to any number of causes, from internal tectonics to externally triggered heating by an impact.  The best example of such an event on Earth is are the Channeled Scablands of the state of Washington.  The Scablands were carved out when the glacier that plugged the Pleistocene era lake, Lake Missoula, collapsed.  Lake Missoula had covered parts of Idaho and western Montana.  Estimates for the volume of discharge necessary to carve out the Martian outflow channels run as high as 1 billion cubic meters per second (for comparison: the average rate of discharge of the Mississippi River is 20,000-30,000 cubic meters per second, so the Martian discharge rate would be equivalent to more than 30,000 times the flow of the Mississippi! the estimated discharge rate for the Channeled Scablands is about 10 million cubic meters per second).  Thus, the floods generated by the formation of the Martian channels may have carried more than 100 times as much water as the largest known floods on Earth.

The minimum estimates for the total volume of water necessary to have carved the Martian outflow channels is at least 6 million cubic kilometers, enough to make a global ocean 40 m deep.

Valley Networks

Valley networks are the most common drainage system on Mars.  The formation of such valley networks normally assumes warm and wet climatic conditions, but this may be much too simplistic.  Valley networks exist in the cratered uplands; they are branching networks in which tributaries converge downstream (the opposite of outflow channels); however, most are far less complex than terrestrial networks.  Most valley networks are no wider than a few km. Most (92%; 759 out of 827 identified) of the valley networks are on parts of Mars called the Noachian, which predates the Hesperian, i.e., they formed before the outflow channels during the great bombardment era.  About 4% (34) may be Hesperian and another 4% may be younger.

Most valley networks may have have formed by the process of groundwater sapping, in which the water source is underground and the valleys from by the collapse of the surface as it is undermined by the movement of water underground.  Thus, it is possible that few or none of these valleys require a warm, wet Mars.

However, the consensus (but not unanimous) view is that the valley networks were cut by streams and rivers and that this demonstrates that Mars was  warm and wet for tens or hundreds of millions of years.

One valley network on Mars is Nirgal Vallis, as seen by Viking and later by the Mars Global Surveyor.  (labeled image)

Another is found in the Thaumasia region.

Terra Meridiani  valley network near the martian Equator. This valley, which is in a heavily cratered region, shows many impact craters. The brightness of the valley is due to the reflectivity of sand deposits.

Viking and MOC images of  Nanedi Vallis

Recent results suggest that the uplift of the Tharsis Bulge is strongly correlated with
the topography of the valley networks.  The study of the Tharsis region indicates that it was already a permanent feature of Mars 4 BY ago.  This study suggests that the eruption an emplacement of the Tharsis bulge requires the equivalent of 10,000 Mauna Loa's worth of magma (300 million cubic km).  Such an eruption would have release a tremendous amount of carbon monoxide (CO) and/or carbon dioxide (CO2), sufficient to produce an atmosphere of 1.5 bars, thereby producing a greenhouse effect, a global ocean 120-m deep, and a warm, wet Mars for a long period of time, at least long enough to carve the valley networks.  Figure 4, found on the web link noted above, shows the general topographic gradients of the valley networks, revealing that they largely flow to the north, as a result of the upward stretching of the southern highlands in response to the Tharsis uplift.

An alternative to water? Could CO2 have carved out some of these channels?

March 30, 2001

Liquid carbon dioxide breakouts rather than water probably created the  martian gullies discovered last summer in high-resolution images from the  Mars Global Surveyor  orbiter camera. Donald S. Musselwhite, Timothy D.  Swindle, and Jonathan I. Lunine of the University of Arizona Lunar and Planetary Laboratory publish their hypothesis in the April 1 issue of Geophysical Research Letters.

Last June scientists announced that gullies seen on some martian cliffs and crater walls suggest that liquid water has seeped down    the slopes in the geologically recent past. Researchers found small channels on slopes facing away from mid-day sunlight, with most channels occurring at high latitudes,near Mars' south pole. The scientists concluded that the relationship between sunlight and latitude may indicate that ice plays a role in protecting the liquid water from evaporation until enough pressure builds for it to be released catastrophically into the surface. If channels are forming today, liquid water may exist in some regions of Mars barely 500 meters beneath the surface, they suggest.

Now UA researchers propose an alternative explanation involving carbon dioxide erosion. They point to several reasons why CO2 is a better candidate than water in gully formation. One reason is that most gullies are found in the southern highlands, the oldest and coldest part of the planet, a place where liquid water is least likely to be stable.

"That's high altitude in a region of low geological activity. It is difficult to invoke some hydrothermal action there," Musselwhite said. "The surface is old but the gullies are new."   Another reason is that the southern hemisphere has more extreme temperature variations throughout the year than does the northern hemisphere, a result of the fact that Mars is closer to the sun during southern summer and farther away during southern winter, Musselwhite said. The gullies are generally on pole-facing sopes where they receive very little or no sunlight for most of the year.   However, Musselwhite said, the most compelling fact is that gullies always start about 100 meters below the top of the cliff. At that depth, the pressure of the rock overhead is just enough for liquid CO2 to be stable, if the temperature is low enough.

"There are many interesting ideas about how to liquid water might carve these things. Still, if the process works in these very special locations where at least during wintertime it is extremely cold, why don't we see the gullies in other places? If you have water cutting these gullies, you should see that everywhere, not just at these specific locations. And where is the water coming from? There is not much of it in the martian atmosphere or on the surface," he said.

It's not liquid carbon dioxide flowing in the gullies. "What's coming out is liquid CO2 that suddenly vaporizes," Musselwhite said.  "As it comes out, it expands very quickly, cools, and actually produces CO2 snow. The snow is suspended in CO2 gas that hasn't solidified yet. Together with rock debris, it forms slurry. Geologists call it a 'suspended flow.' Suspended flow acts like a liquid. It doesn't take very much liquid each time to add to gully formation."

There are analogs on Earth to this process. Martian gullies look almost identical to terrestrial ones found in polar regions and also on cliff walls, where gullies are carved by snow pack. Such channels can also be found on the flanks of Earth volcanoes, carved by a suspended flow of ashes entrained in volcanic gas. And trapped mud and sediment caught in turbidity currents on the ocean floor cut deep channels into the continental shelves, Musselwhite noted.

How do Martian gullies form? The planet's atmosphere is mostly composed of CO2. Under some atmospheric pressure, CO2 condenses from the atmosphere and into Mars' surface. Mars has been pummeled by impacts, so its surface is typically porous, spongy gravel. Gas seeps through the surface and condenses in the pores of rock.  "In wintertime the cliff surface gets so cold that its temperature falls below the freezing point of CO2, which at low pressure goes directly to solid. As the cold wave moves from the surface, the pore space is completely filled in. When spring comes, dry ice warms up and expands. Since all the rock pore space is filled, pressure builds until the ice turns to liquid.  Liquid CO2 takes up more volume than dry ice, so pressure continues to build."

At the same time, the dry ice dam evaporates and thins as temperature rises.  At one point the barrier becomes too thin, and the liquid under pressure bursts out. It breaks through the surface into the atmosphere, where it evaporates very quickly given the sudden drop in pressure. As carbon dioxide vaporizes rapidly, it also cools and entrains the CO2 snow, creating the suspended flow.  Some researchers claim that the gullies are very young and may be currently forming on Mars. They tie gully locations to oscillations in the martian climate caused by varying tilt of the planet's rotation axis, called obliquity. When the obliquity is low.  Mars' axis is almost straight up and the surface near the poles gets less heating all year around. At high obliquity in winter more of the surface would be shaded, but in the summer time it would get much more sunlight than usual.

"If this explanation is correct, gullies are forming today around the south pole," Musslewhite said. "The ones that are farther from the poles are then older. You might expect these to form close to the equator in the period of high obliquity, when the axis is more tilted over. Some may be forming now on a yearly basis."  This idea is supported by evidence that some researchers say suggests that gullies are forming today near the south pole but not closer to the equator. Multiple images of the same gullies are needed to prove that, Musselwhite added.

Ancient Lakes?

Gusev Crater and Ma'adim Vallis. Gusev Crater is approximately 150 kilometers (93 miles) across. Ma'adim Vallis is the nearly straight canyon that enters Gusev Crater from the lower right. (Viking orbiter image).  Close-up MGS image of lower region where Ma'adim Vallis enters Gusev Crater.  Was Gusev Crater an ancient lake?  There is no evidence of ancient shorelines, either because no lake ever existed, because wind erosion has erased the shoreline evidence, or because such features never formed.

The Elysium Basin and Marte Vallis, Mars, Viking image and a close-up Viking image of the region of the MGS images.  There were two competing ideas about the Elysium Basin. One hypothesis held that the depression was once the site of a vast lake approximately 1,500 meters (4,900 feet) deep. Because the floor of Elysium Basin has very few small, fresh impact craters, it was proposed that this lake dried up relatively recently in martian history--that is, the lake would have been younger than most of the volcanoes, craters, and even the Ares Vallis flood channel in which is located the Mars Pathfinder landing site. At some point, the lake in Elysium Basin was thought to have reached such a depth that it began to spill over a rise on its east end. The water spilling out the east end of Elysium Basin was thought to have created Marte Vallis--a channel containing streamlined islands that stretches for hundreds of kilometers (miles) to the northeast. The lake bed and channel, it was proposed, might make good places to land future rovers that could travel around and collect samples that might contain evidence of past martian life.

The other hypothesis held that the Elysium Basin floor was covered with flows that were emplaced as extremely fluid lava (molten rock). It was suggested that a lake of water could have been in the basin long, long ago, but that the most recent geologic events had erupted huge volumes of very fluid lava across the basin floor. Some of this lava was proposed to have even poured out of the basin and traveled down Marte Vallis. In this hypothesis, it was assumed that Marte Vallis--named for the Spanish word for "Mars"--was first carved by water, and then was a conduit for lava from volcanic eruptions. The lavas were proposed to have been very fluid--behaving almost like water. Such fluid lavas are known on Earth to result from molten rock that has a low concentration of silica, a high temperature, and/or a high eruption rate.

 Close-up image shows the textured surface in Elysium Basin.   The texture seen here--one of dark, flat fragments separated by bright cracks--is termed platey, and indicates a surface where the dark areas, or plates, separated and moved apart. Such a texture occurs when the rigid surface of a once-fluid material (for example, ice over water, or the cooled surface of a lava flow over the molten interior) is broken up and moved about by the underlying fluid. In this case, the relationships strongly suggest lava, which today is hardened to rock throughout. If Elysium was an ancient lake, all evidence of materials deposited at the lake bottom appear to have been thoroughly covered over by recent (millions but not billions of years ago)  lava flows.

Or maybe not:
Features on Northern Plains of Mars Are Tectonic Ridges, Not Ancient Ocean Shorelines

 April 4, 2001

What scientists suspect might be ancient ocean shorelines on the northern plains of Mars is actually a network of tectonic ridges related to  dramatic martian volcanism, a University of Arizona planetary sciences graduate student and a collaborating post-doctoral researcher at the Massachusetts Institute of Technology report in the April 5 issue of Nature.

Their new findings don't rule out the possibility that an ancient ocean once did cover the northern half of Mars. However, what previously has been reported to be ancient shorelines apparently are not. The discovery of the network of ridges "opens a new tectonic window into Mars," the authors say.

Paul Withers of the UA and Gregory A. Neumann of MIT analyzed dazzlingly precise new views of Mars' topography from the Mars Orbiter Laser Altimeter (MOLA). The instrument continues an extended mission in orbit around Mars on the Mars Global Surveyor spacecraft. MOLA transmits infrared laser pulses towards the surface of Mars, and the measurements are used to create topographic maps accurate to within a meter of elevation. Viking era topographic maps of Mars were accurate only to about a kilometer.  Withers worked last summer through a graduate student program with members of the MOLA science team at the NASA Goddard Space Flight Center.  He and Neumann analyzed ridges that cover the enigmatic northern plains of Mars. The region is the flattest known surface in the solar system, and a leading theory is that an ocean created such extraordinary smoothness.

Authors of a December 1999 article in Science identified candidate shorelines of the possible ancient ocean based on the new MOLA maps. The topographical profile shows a succession of flat terraces along a linear slope in one case, and in another case a series of slopes in the right relation to be shorelines.  Withers and Neuman specifically re-examined two leading candidate paleoshoreline groups, one group near the Utopia impact basin and the other on the opposite side of the proposed ocean near the Alba Patera volcano.

The details of the ridges near the Utopia basin don't look like paleocoastline, Withers said in an interview. "The morphologies  are  inconsistent with formation by shoreline processes. There are the flat terraces, but the ridges are on what would be the oceanward side. That's difficult to explain if you have an ocean coming in, flattening things smooth over the terrace and then receding again.  He and Neumann conclude that the ridges record a history of enormous tectonic stress and strain that forced the martian crust to form 10-mile-high volcanoes.   "Most ridges appear to be related to obvious stress centres, such as the volcanic Tharsis Rise, the Utopia impact basin and the Alba Patera  volcano," they report in Nature. The direction and shapes of these ridges indicate that they have a tectonic origin.

 The network of ridges is the only tectonic feature in the region.   "In future work, we hope that studying these ridges will reveal how the  huge martian volcanoes formed, what the martian crust and lithosphere were like at the time, and what the northern plains of Mars are like today beneath their blanketing surface layer of martian dust."

Deuterium and Nitrogen on Mars

The deuterium to hydrogen ratio in the atmosphere  of Mars is 8.1 x 10-4  (D/H = 1/1230), which is 5 times larger than the ratio of 1.6 x 10-4 (D/H = 1/6250; we previously rounded this number to 1/10000) on Earth.  This implies that Mars has lost a good deal of its water, as has Venus!

(D/H)mars =  8.1 x 10-4 = 5 (D/H)earth
The current loss rate of H is 170,000 times greater than that of D from Mars!  By one estimate, if the current mass loss rate of H and D was constant over 4.5 BY, then Mars began with the equivalent of a global ocean 3.6 m deep and has lost 94% of it (all but 0.2 m).  Of course, a 3.6 m deep equivalent global ocean is ten to thirty times less water than is estimated as necessary to form the outflow channels.  An obvious simplification - that the mass loss rate has been constant for 4.5 BY - is likely wrong, and if the mass loss rate was periodically higher, the original inventory of water would be much greater, by some estimates 20 times greater, or 72 m, with a remaining amount of 4 m.  Either way, the D/H ratio strongly suggests that Mars may have lost more than 90% of the water in the surface/atmospheric inventory subject to D/H enrichment. Note that water locked into permanent subsurface reservoirs would not participate in the D and H loss nor the D/H enrichment and thus, much of Mars' water reserves could still be present.

Like hydrogen, nitrogen also is found naturally with more than one isotope.   On Mars, the ratio of 15N/14N is 1.62 times larger than the 15N/14N ratio on Earth.

(15N/14N)mars =  1.62 (15N/14N)earth
Why? As with the hydrogen and deuterium, if photochemical processes stimulated by solar ultraviolet light dissociate nitrogen molecules, then mass loss from the atmosphere will favor the escape of 14N and the retention of 15N, leading to an enhanced value of the 15N/14N ratio.  Thus, Mars likely once had much more nitrogen.  At present, Mars has 6.5 x 1017 gm of atmospheric nitrogen, which is 2.7% of the mass of the atmosphere (and 2.7% of 7 mbars is 0.19 mbars atmospheric pressure of nitrogen). Estimates of the total original mass of nitrogen are a few to a few tens of mbars, or a few tens to a few hundred times as much as is present now, a few percent of the total amount of nitrogen in Earth's atmosphere.  Given that the mass of Mars itself is 11% of that of the Earth, this estimate is in line with Mars having a comparable amount of nitrogen, as compared to the mass of the planet, as did Earth.  This is also right in line with the inventory of carbon dioxide on Mars (a few percent of that of Earth) and makes for a convincing case that Mars should have started with as much water as a few percent of the Earth's water inventory.  At present, most of this water is unaccounted for.

What is the importance of nitrogen?  Remember that in Earth's atmosphere, it is the nitrogen that makes the greenhouse gasses so effective (through the process of pressure broadening).    Could nitrogen have done this on Mars, but the loss of nitrogen then contributed to a reduction in the greenhouse effectiveness of the atmosphere, leading to a global cooldown and the end of the greenhouse?  If so, could most of the water still be trapped on Mars, frozen beneath the surface rather than lost?

The Final Word (for now) on Water on Mars

Most planetary geologists believe Mars once had as much as the equivalent of a 0.5 - 1.5 km deep global ocean of water.  If this is combined with a thick CO2 - H2O atmosphere that produced enough of a greenhouse to keep surface temperatures above freezing, one could have a young, warm and wet Mars. Perhaps due to the loss of nitrogen, or perhaps due to gradual fixation of CO2 into carbonates without global tectonic processes to complete the carbon cycle, the greenhouse turned off and Mars froze.

Unfortunately, this simple picture doesn't seem to work very well.  During the first 0.5 BY of Martian history, during the era of the great bombardment, large impacts would do a good job of stripping off a thick CO2 - H2O atmosphere.  In addition, the young sun was cooler than the present Sun by about 25%, thus necessitating an even thicker greenhouse atmosphere to make Mars warm.  This looks unlikely.    In addition, surface rocks do not appear to be thick in carbonate deposits, as one would expect if there were  large oceans precipitating significant carbonate deposits.

Thus, the evidence does not overwhelmingly, let alone unequivocally,  support the theory for a warm, wet Mars, but neither does the evidence make a clear  case for the opposite.

 Complete Mars Image Gallery from the Mars Orbiter Camera
 1997-1998 MOC Images Listed By Theme or Topic
 MGS MOC Captioned Images Listed By Release Date
 Mars Global Surveyor---Mars Orbiter Camera (MOC)