Water On Mars PDF

Water on Mars
Harry Y. McSween Jr.1
The papers in this issue of Elements
W
ater on Mars exists at the poles and in the subsurface. It has
summarize some of the most
interacted with crustal rocks, providing geomorphological, recent and intriguing discoveries
geochemical, and mineralogical insights into Mars’ geological about water on Mars. The investi-
history and inferences about its biological potential. The roles of water are gations described utilize imagery
and remote sensing measurements
revealed through studies of altered materials using orbiting-spacecraft
from orbiting spacecraft, analyses
imagery and spectroscopy, instruments mounted on rovers, and laboratory from rovers scurrying across the
measurements on Martian meteorites. surface, and laboratory studies of
Martian meteorites to try to
KEYWORDS: water, aqueous alteration, spacecraft, rovers, meteorites unravel the mystery of when water
ran and where it hides. This is not
The eroded valley networks pictured on the cover of this a comprehensive catalog of research on the topic, but it
Elements issue demonstrate that flowing water once shaped should allow an appreciation of the diversity of scientific
the face of Mars. Borrowing from an old adage, we might approaches and a glimpse into the excitement of planetary
say that water on Mars “can run, but it can’t hide.” But hide geological exploration.
it does. Although frozen water covers the Martian poles,
In his paper, Baker describes geomorphological features that
liquid water is not stable anywhere on the planet’s surface
were carved by running water. He also postulates that the
under present environmental conditions. Finding where
huge basin that forms the northern hemisphere of Mars was
this elusive substance hides and understanding its role in
once filled with an ocean. Orbiting spacecraft carry the
the geological (and possibly even biological) history of Mars
Thermal Emission Spectrometer (TES) and Thermal Emission
require innovative approaches—as revealed in this collec-
tion of articles.
The history of water on Mars is recorded by its chemical
interactions with rocks. Most of the Martian crust consists
of basaltic lavas (McSween et al. 2003). On Earth, basalts are
readily altered to clays and other hydrated minerals, and
they experience profound geochemical and isotopic
changes when exposed to water. However, remote sensing
of Mars by spectroscopic methods has not yet found clays
in abundance, and geochemical changes in texturally
altered rocks seem to be of limited extent. One spectral fea-
ture that is commonly noted, the ferric absorption edge,
indicates that fine-grained ferric oxides (thought to be
weathering products) are abundant everywhere in soils and
rock coatings (Bell et al. 2000), as are sulfates, chlorides, and
other salts formed by evaporation (Squyres et al. 2004).
These oxides and salts are unambiguous indications of
aqueous alteration, but even so, the extent to which Martian
rocks have interacted with water remains controversial.
Water’s hiding places are revealed by an orbiting spacecraft
carrying an instrument that probes the ground for hydro-
gen. Mineral-bound H2O occurs in areas near the equator,
and subsurface ice is thought to be widespread at higher lat-
itudes. Rovers dig trenches and explore craters, both of
which expose salts once carried by groundwater. Enrich-
ments of the heavy isotopes of hydrogen and oxygen meas-
ured in Martian meteorites disclose a cycle that moves
water from the crust to the atmosphere and back again.
1 Department of Earth and Planetary Sciences

University of Tennessee
Knoxville, TN 37996-1410, USA FIGURE 1 Gulleys in a Martian crater, image width ~3 km. MGS
E-mail: [email protected] MOC Release No. MOC2-1182
ELEMENTS, VOL. 2, PP. 135–137 135 J UNE 2006

Imaging System (THEMIS), which measure the heat emitted
from the planet’s surface. These instruments can identify
minerals and map their distributions. The OMEGA spec-
trometer records light reflected at visible and near-infrared
wavelengths, which provides complementary information
on mineralogy. Wyatt and McSween describe how spec-
trometers on orbiting spacecraft can be used to search for
evidence of aqueous alteration. The Gamma Ray Spectrom-
eter (GRS) reveals the extent of icy permafrost in the sub-
surface over large areas of the planet. Christensen reports
how new imagery from THEMIS and the Mars Orbiting
Camera (MOC) shows the complicated history of ice at the
poles, and high-resolution images of gulleys on the sides of
large craters (FIG. 1) suggest that water has existed, at least
fleetingly, during modern times.
A variety of geochemical and mineralogical tools on the
Mars Exploration Rovers (MER) provide a wealth of new FIGURE 3 ALH84001 Martian meteorite, showing mm-size secondary
carbonate globules
discoveries at two locations. Jolliff and McLennan note that
the Opportunity landing site, in Terra Meridiani, shows
unmistakable evidence for evaporation of a salt-laden sea grand petrological, geochemical, and possibly biological
and flushing with groundwater during diagenesis (FIG. 2). A experiment, complete with water, and its study deepens our
handful of Martian meteorites, thought to be igneous rocks understanding of how our own world works. .
ejected during impacts, contain minor amounts of alter-
ation minerals (FIG. 3) and have unusual isotopic anomalies
in hydrogen, oxygen, and sulfur. Leshin and Vicenzi describe REFERENCES
the evidence for interaction with water in these meteorites. Bell JF III and 23 coauthors (2000) Mineralogic and compositional
properties of Martian soil and dust: Results from Mars Pathfinder.
Journal of Geophysical Research 105(E1): 1721-1756
Hartmann WK, Neukum G (2001) Cratering chronology and
the evolution of Mars. Space Science Reviews 96: 165-194
McSween HY Jr, Grove TL, Wyatt MB (2003) Constraints on the
composition and petrogenesis of the Martian crust. Journal of
Geophysical Research 108(E12): 5135, doi:10.1029/2003JE002175
Squyres S and 49 coauthors (2004) The Opportunity rover’s Athena
science investigation at Meridiani Planum, Mars. Science 306:
1698-1703 .
GLOSSARY OF MARS TERMS
Remote sensing instruments

on current spacecraft (see FIG. 4)
Alpha Particle X-ray Spectrometer (APXS) – Instrument
on Mars Exploration Rovers that measure bulk rock
chemistry using interactions of alpha particles
with the target
Gamma Ray Spectrometer (GRS) – Spectrometer
on Mars Odyssey that measures abundances of H and other
elements by their interactions with neutrons
FIGURE 2 Evaporitic rocks in Endurance crater at the Opportunity Mars Orbiter Laser Altimeter (MOLA) – Instrument on
landing site. Tracks formed when the rover descended
are visible at the top, and (silver dollar–sized) holes formed by grinding Mars Global Surveyor that measures topography using
of the Rock Abrasion Tool (RAT) are seen at various levels. laser reflections from the planet’s surface
Mars Orbiting Camera (MOC) – High-resolution imager on
Mars Global Surveyor
Life, as we know it or can imagine it, requires water, so it is
not surprising that the exploration of Mars focuses on a Microscopic Imager – A 30× microscope on Mars Explo-
search for water, both past and present. A final paper, by ration Rovers used to image rock and soil textures
Knoll and Grotzinger, describes emerging ideas and con- Miniature Thermal Emission Spectrometer (Mini-TES) –
straints on possible Martian organisms and their depend- Spectrometer on Mars Exploration Rovers used to identify
ence on water. minerals by their lattice vibrations
The story of the search for Martian water, which began Mössbauer Spectrometer – Instrument on Mars Exploration
decades ago with images of water-carved landforms, has Rovers used to identify Fe-bearing phases by measuring
become in large part a geochemical and mineralogical nuclear oscillations in crystal fields
investigation. Much of this research involves remote sensing Omega Spectrometer – Visible/near-infrared reflectance
measurements that differ from the conventional laboratory spectrometer on Mars Express used to identify minerals
techniques used in these disciplines, but they are in reality through absorptions due to electronic transitions and
extensions of Earth-based research, and interpretations of lattice molecular vibrations
these data are informed by terrestrial studies. Mars is a
ELEMENTS 136 J UNE 2006

FIGURE 4 Timeline for Mars exploration. Three orbiters and two
rovers are currently operational, another orbiter has been
launched, and several landers are under development. COURTESY OF NASA
Panoramic Camera (Pancam) Martian meteorites – Shergottites, Olivine, pyroxene – Mg, Fe, and Ca
– Color imaging system on Mars nakhlites, chassignites (SNCs) and silicates common in mafic igneous
Exploration Rovers ALH84001—mafic and ultramafic rocks rocks
Thermal Emission Imaging System ejected from Mars by impacts Smectites – Swelling clay minerals rich
(THEMIS) – Imaging system on Mars Palagonite – Amorphous or poorly in Fe or Mg
Odyssey that records heat emitted crystalline clays formed by interaction Zeolites – Hydrated Al silicates, with Na,
from the Martian surface; THEMIS has of volcanic rocks with water K, or Ca
high spatial resolution but low spectral Permafrost – Perennially frozen ground
resolution.
Surface Types 1 and 2 – Global surface Other terms
Thermal Emission Spectrometer (TES) units that are distinguished by thermal
– Thermal infrared spectrometer on Albedo – Percentage of incoming
emission spectra; ST1
Mars Global Surveyor that is sensitive radiation that is reflected from a
is interpreted as basalt and ST2 as
to vibrational motions of atoms in surface
andesite or partly altered basalt.
minerals; TES has high spectral Geomorphology – The nature and origin
Regolith – The particulate surface
resolution but low spatial resolution. of landforms and landscapes
materials on a planet; differs from
soil in that organic matter is absent. Geographic locations – See the map on
Martian materials the following website:
http://pubs.usgs.gov/imap/i2782/
Concretions – Rounded masses formed by Minerals i2782_sh1.pdf
percolating waters
Ferrihydrite, goethite, hematite, Noachian, Hesperian, Amazonian –
Cements – Minerals deposited among rock maghemite, magnetite – Oxides Time periods (oldest to youngest)
particles that bind them together and hydroxides of Fe2+ and Fe3+ defined for Mars’ geologic evolution.
Evaporites – Salts formed by evaporation Gypsum, jarosite, kieserite – Sulfates of Estimated time intervals are roughly
of brines Ca, ferric Fe, and Mg before 3.5, 3.5–2.9, and after 2.9
Iddingsite – Assemblage of minerals Halite – NaCl billion years, respectively (Hartmann
formed by alteration of olivine and Neukum 2005).
Sol – A Martian day, approximately 24
and a half hours

Geomorphological Evidence
for Water on Mars
Victor R. Baker1
CHANNELS, VALLEYS,
M
artian landscapes and landforms indicate episodic activity by water
and ice, extending from the planet’s earliest history up to the present ALLUVIAL FANS,
AND SEDIMENTS
day. Most of the relevant fluvial, glacial, volcano–ice, periglacial,
As recognized early in the era of
lacustrine (even “marine”), and related landforms have direct counterparts spacecraft exploration, channels
on Earth. Moreover, they exist in causally related, holistic associations of space and valleys extensively dissect the
and time that confirm their relationship to a long-term history of water-related surface of Mars. Channels are elon-
gated troughs that display clear
activity. Although strong geomorphological evidence for many of these relation- evidence for large-scale fluid flow
ships has been apparent for 30 years, its scientific importance has only been across their floors and on parts of
recently appreciated because of direct geochemical measurements of water their walls or banks (FIG. 1A).
Immense channels, with widths of
and ice features by surface robotic and orbital instruments.
tens of kilometers and lengths of
KEYWORDS: Mars, geomorphology, landforms, climatic change, hydrology up to a few thousand kilometers,
display a suite of morphological
INTRODUCTION attributes that are most consistent
with genesis by cataclysmic flows of water and sediment
Despite 30 years of accumulating, increasingly abundant
(Baker 2001). On Earth, such flows produced the distinctive
and unequivocal geomorphological evidence, the case for
landforms of the Channeled Scabland (FIG. 2A). An important
past water-related activity on Mars remained immensely
recent discovery is that Martian flood channel activity,
controversial until very recently (Baker 2004). Ingenious
involving outbursts of water and associated lava flows,
models were proposed (e.g. Hoffman 2000) to ascribe non-
occurred in the Cerberus plains region on the order of
aqueous origins to individual Martian landforms that other-
10 million years ago (Berman and Hartmann 2002; Burr et
wise had striking similarities to water-generated features on
al. 2002). The huge discharges associated with these floods
Earth. Nevertheless, this view is profoundly changed
and the temporally related volcanism should have
because of recent developments, notably the nuclear
introduced considerable water into active hydrological
physical measurement of abundant, extant, near-surface ice
circulation on Mars.
(Boynton et al. 2002) and direct chemical analyses of aqueous
minerals associated with sedimentary rocks (Squyres et al. While the huge Martian channels are generally well charac-
2004). Recent dampening of hydrophobic theorizing makes terized by the older imaging systems of the 1970s, it was
it appropriate to reassess the geomorphological evidence for not until a new generation of orbital imaging capabilities in
water on Mars, and this brief overview will emphasize the past several years that major advances occurred in
developments. understanding the nature of valley networks (FIG. 1B).
These valleys dissect the Martian highlands much more
Geomorphology concerns the nature and origins of land-
extensively than was apparent from the earlier images
forms and landscapes. Its application to extraterrestrial
(Hynek and Phillips 2003). Many valleys contain relict
planetary surfaces relies upon known associations of form
channels comparable in their dimensions to the active river
and process on Earth. However, the rationale for ascribing
channels associated with terrestrial valleys (Irwin et
genesis to a newly discovered planetary landform is not
al. 2005). Formed about 3.9 billion years ago, Martian high-
simply a matter of comparative image analysis. Landforms
land craters and impact basins were extensively eroded by
and landscapes exist in complex, interrelated assemblages,
surface runoff processes during episodes with Earth-like
in which the different elements relate to one another in
precipitation (Craddock and Howard 2002).
time and space because of their generation by a unique
sequence of formative processes. Scientifically productive Large alluvial fans occur in ancient highland craters at middle-
alternative explanations must not merely satisfy individual, to-low southern latitudes. They are remarkably similar to
simple-minded, “look-alike” criteria. Instead, they must low-relief terrestrial alluvial fans formed dominantly by
account holistically for the entire, genetically related assem- fluvial, as opposed to debris flow, processes (Moore and
blage, in the same way that the solution of a crime depends, Howard 2005). They probably formed during an episode of
not on a single clue, but on the whole interrelated web of enhanced precipitation that followed the transition from
evidence that is built up through connections of time and the early period of heavy impact bombardment (the
space by a master detective. Noachian Epoch) to the period of lower impact rates after
about 3.9 billion years ago. Relatively high denudation rates
1 Department of Hydrology and Water Resources are inferred for the Noachian, and these are much greater
The University of Arizona than in later periods (Craddock and Howard 2002). The
Tucson, Arizona 85721, USA new observations are consistent with the discovery that
E-mail: [email protected]

A B
C D
FIGURE 1 Examples of Mars landforms indicative of water and ice The water body associated with emplacement of the Vastitas
processes. (A) Cataclysmic paleo–flood channel south of Borealis Formation was approximately contemporaneous
Cerberus (3.8°N, 204.7°W). The 3 km wide image shows a relatively
with the floods responsible for the largest outflow channels,
small streamlined island and a cataract. (NASA MGS Press Release No.
MOC2-866) (B) Portion of the Warrego valley network, 24 km wide and it may have covered as much as 3 million km2 to average
(42.3°S, 267.5°W). (NASA MGS Press Release No. MOC2-868) (C) depths of hundreds of meters. The largest estimates involve
Ancient scroll topography and meander cutoff. This 2 km wide image is as much as 20 to 60 million km3 of water, equivalent to 200
located at 24.3°S, 33.5°W. (NASA MGS Press Release No. MOC2-543)
to 400 meters spread evenly over the whole planet and
(D) Debris flow levees on a poleward-facing dune face in Russell crater
(54.7°S, 347.4°W). (MOC IMAGE M19-01170) comparable to the inferred collective flows from the out-
flow channels (Carr and Head 2003; Boyce et al. 2005), as
proposed by Baker et al. (1991). Other periods of outflow
upper layers of the ancient Martian crust of the highlands channel activity and associated inundations of the north-
contain extensive sedimentary rocks that were emplaced ern plains (Clifford and Parker 2001; Fairén et al. 2003) are
during the intense denudation phases (Malin and Edgett far less certain as to the extent, relative timings, and dura-
2000a). Now confirmed by in situ inspection (Squyres et al. tions of the various inundation episodes.
2004), sedimentary rocks were identified from orbit by their
morphological expression. Imagery from the Mars Orbiter Though the debate over the Martian “ocean” has received
Camera (MOC) of the Mars Global Surveyor (MGS) mission much attention, even more compelling evidence supports
shows that the Martian highlands do not consist of an initial the existence of numerous lakes and seas, which were tem-
lunar-like surface, underlain by an impact-generated porarily extant on the surface of Mars at various times in
megaregolith, as presumed in previous models (e.g. Carr the planet’s history (Cabrol and Grin 2002). The more
1996). Instead, cratering, fluvial erosion, and deposition of ancient lakes occupied highland craters during the heavy
layered materials probably all occurred contemporaneously, bombardment epoch, spilling over to feed valleys such as
leading to a complex interbedding of lava flows, igneous Ma’adim Vallis (Irwin et al. 2004). The floodwater spilled
intrusions, sediments, buried crater forms, and erosional from lakes that held up to several hundred thousand cubic
unconformities (Malin and Edgett 2000a). kilometers of water, covering an area of about one million
cubic kilometers, and flows achieved peak discharges of
millions of cubic meters per second (Irwin et al. 2004).
ANCIENT OCEANS, SEAS, LAKES, AND DELTAS
Abundant crater paleolakes seem to have developed just
Morphological evidence for past large bodies of water after the heavy bombardment, and very large lakes occu-
covering the northern plains of Mars, apparent by the late pied the floors of the impact basins Hellas and Argyre.
1980s, includes the morphological characteristics of sedimen-
tary deposits and, more dramatically, a pattern of sur- Fluvial deltas are commonly associated with the paleolakes.
rounding shorelines (Clifford and Parker 2001). Evidence One complex of ancient meandering alluvial channels,
for the latter was systematically criticized in a global sense comprising a fan-delta partly filling the crater Eberswald or
(Carr and Head 2003), but was supported for local areas by NE Holden (Malin and Edgett 2003; Moore et al. 2003),
very detailed mapping (e.g. Webb 2004). Nevertheless, the displays Earth-like morphologies that can only be explained
general concept of past inundations on the northern plains, by persistent fluvial activity on time scales ranging from
constituting an “Oceanus Borealis,” at least for geologically centuries (Jerolmack et al. 2004) to hundreds of millennia
short episodes, has been found to be generally consistent (Bhattacharya et al. 2005). Paleo–meander bend topography
with considerable geomophological evidence. The distinc- (FIG. 1C) shows that these were laterally accreting, alluvial
tive water-laid sedimentary deposits that cover parts of the rivers of a similar type to the modern Mississippi.
northern plains, known as the Vastitas Borealis Formation,
afford the most convincing evidence (Carr and Head 2003), THE CRYOLITHOSPHERE OF MARS
including (1) margins that roughly mark the surface to What happened to the huge water inventory necessary for
which a body of water would approximate, (2) a distinctive generating channelized megafloods and relatively short-lived
population of impact craters indicating associated ice and lakes and seas? While a variety of atmospheric-loss
sediments (Boyce et al. 2005), and (3) a phenomenally flat processes undoubtedly occurred, the geomorphological
and smooth surface expression, similar to that of abyssal evidence suggests that water, even the “Oceanus Borealis,”
plains in Earth’s ocean basins. was not on the surface for prolonged periods. Instead it
resided nearly all the time, except for brief, sometimes

A B
C D
spectacular episodes, within or beneath semipermanent, ice- FIGURE 2 Examples of Earth water- and ice-related landforms with
distinctive morphologies that have equivalents on Mars.
rich permafrost. The long-term existence of this ice-rich (A) Longitudinal grooves and inner-channel cataract complex eroded
layer, constituting a cryolithosphere about 1–2 km thick in into basalt by the catastrophic Pleistocene megafloods of the Channeled
equatorial areas and 5–6 km thick at the poles, is docu- Scabland in east-central Washington, USA. (B) Ice-wedge polygons
mented by a variety of geomorphological features (Kuzmin formed in permafrost terrain near Barrow, Alaska, USA. (C) Small rock
glacier and adjacent periglacial debris fans in the Altai Mountains of
2005). Most of these have been well known since the 1970s, south-central Siberia. (D) Esker on the Waterville Plateau in eastern
including various types of flow-lobed ejecta blankets (ram- Washington, formed by gravel deposition from meltwater that flowed
part craters), debris flows, lobate debris aprons, and polygo- though tunnels beneath the late Pleistocene Okanogan Lobe of the
nal terrains. A variety of landforms related to volcano–ice Cordilleran Ice Sheet. This esker is about 30 m wide and 3 km long.
interactions (e.g. Chapman et al. 2000) document the occa-
sional short periods of volcanically induced water outbursts near-surface ground ice. The gullies are uncratered, and
from this reservoir of ice and underlying ground water. Fol- their associated debris-flow fan deposits are superimposed
lowing these episodes, surface water seems to have very rap- on both eolian bedforms (dunes or wind ripples) and
idly returned to the cryolithosphere. Thus, despite consid- polygonally patterned ground, all of which cover extensive
erable theorizing (reviewed by Carr 1996), a clear indication areas that are also uncratered (Malin and Edgett 2000b).
of the size of Mars’ mostly hidden global water inventory The patterned ground is itself a very strong indicator of
cannot be gleaned from the isotopic composition of the near-surface, ice-related processes in the active (seasonally
tiny fraction of that inventory which was subject to long- thawed) layer above the Martian permafrost zone (Siebert
term exospheric escape processes. and Kargel 2001).
At the regional scale, gullies occur in high-latitude bands on
RECENT GULLIES, GLACIERS, Mars. They are associated with a variety of other landforms
AND RELATED ACTIVITY that indicate direct emplacement and local degradation of
Recent discoveries from Mars Orbiter Camera (MOC) mantles of ice and dust, possibly even dirty snow, all
images show that Mars displays a diverse suite of excep- derived from the atmosphere (Head et al. 2003). The evidence
tionally young, globally distributed landforms that are consists of small-scale polygonal or patterned terrains, similar
water related. If observed on Earth, these landforms would to the ice-wedge phenomena of Earth’s high-latitude
generally be well understood to have aqueous origins, permafrost regions (FIG. 2B); the mobilization of rocky
involving dynamical hydrological cycling on relatively debris on slopes, similar to the rock glaciers of Earth’s peri-
short time scales (hundreds to thousands of years) in a glacial regions (FIG. 2C); and a sort of regional smoothing of
warmer, wetter, and denser atmosphere than occurs on small-scale topography by deposits, a few to several meters
Mars today. Perhaps the most striking of the recent discoveries thick, that are internally layered and locally eroded (Mustard
made from the high-resolution MOC images is that of et al. 2001). The emplacement of ice-rich deposits at low- to
numerous small gullies (Malin and Edgett 2000b), developed mid-latitudes seems to be consistent with geologically
on hillslopes associated with crater rims and channel or valley recent episodes of higher tilt (obliquity) of Mars’ axis of
walls. Morphological similarity of these hillslope gullies to planetary spin. This would result in warming of the polar
terrestrial, high-latitude, periglacial gullies suggests an caps, thereby increasing the sublimation of ice and migra-
origin by aqueous debris flows, involving the melting of tion of water vapor to the then-cooler lower latitudes.

Another class of very distinctive debris flows occurs on the The new evidence of glaciation is distinguished by its abun-
debris-mantled slopes of large sand dunes (FIG. 1D) (Man- dance, the complex detail of its assemblages, and the com-
gold et al. 2003). Water–sediment mixtures afford the most monly very young geological ages (Kargel 2004). The glacial
likely mechanism for producing these landforms. More landforms of Mars include erosional grooves, stream-
controversial are the abundant dark slope streaks developed lined/sculpted hills, drumlins, horns, cirques, and tunnel
in currently active dust mantles on hillslopes; a case can be valleys; depositional eskers (FIG. 2D), moraines, and kames;
made that these result from local, occasional water activity and ice-marginal outwash plains, kettles, and glaciolacustrine
(Miyamoto et al. 2004). plains. These landforms occur in spatial associations,
proximal-to-distal in regard to past ice margins, that exactly
Glaciated landscapes are some of the most important land-
parallel terrestrial glacial geomorphological settings. Long-
form features documented by the newer high-resolution
recognized areas of past glaciation on Mars include lobate
data. Earlier arguments for extensive glaciation on Mars
debris aprons near uplands surrounding Argyre and Hellas
were severely criticized, in part because glaciation has
(Kargel 2004), lineated valley fills in the fretted troughs of
immense hydrological and climatological implications. The
the highlands/lowlands boundary north of Arabia Terra,
growth and persistence of large glaciers require a dynamic
and the polar regions, where the ice caps were much more
hydrological system that moves large quantities of water
extensive during portions of post-Noachian time. Huge glaciers
from surface-water reservoirs, such as lakes and seas,
marked the western flanks of the Tharsis volcanoes. Debris
through the atmosphere to sites of precipitation. Resistance
aprons at the bases of massifs in eastern Hellas show clear
to the idea of ancient glaciers on Mars is especially curious,
morphological evidence of sublimation, ice-rich substrates,
given that there was a general scientific consensus that
and glacial-like viscous flow (FIG. 3). Geologically recent
Mars displays an immense variety of periglacial landforms,
ice-rich rock glaciers (or debris-covered glaciers) occur at the
most of which require the activity of ground ice. The
base of the Olympus Mons scarp, where they are superim-
periglacial landforms include debris flows, polygonally
posed on older, much larger, relict, debris-covered piedmont
patterned ground, thermokarst, frost mounds, pingos, and
glacial lobes (Head et al. 2005).
rock glaciers. On Earth, most of these landforms develop
under climatic conditions that are both warmer and wetter
than the conditions for the cold-based glacial landforms DISCUSSION AND CONCLUSIONS
now known to abound on Mars (Baker 2001). Theorizing about past water on Mars oscillates between a
hydrophilic view of a wet surface environment and a
hydrophobic view of a dry surface environment. Geomor-
phological evidence from the 1970s and 1980s pointed to a
Mars that was episodically active in a hydrological sense
during post-Noachian time. Hypotheses that explained all
these features as interrelated were extensively criticized and
pronounced inferior to the long-prevailing view that Mars
had been continuously dead and dry since the Noachian.
Though the latter view had considerable theoretical support,
it failed to explain the mounting new data that Mars has
experienced episodic hydrological activity throughout its
geological history (Baker 2001)—most remarkably, even up
to the last several million years. While much of the relevant
new data, like those of the 1970s, is geomorphological, the
change in scientific thinking is occasioned by geochemical
and mineralogical measurements, both remote and in situ.
Geomorphology will continue to contribute to the under-
standing of Mars’ watery past, but this contribution will be
less in the realm of theoretical models and more in the
realm of the unique realities that are distinctly Martian. The
relevance is perhaps best expressed in Stephen Jay Gould’s
“principle of planetary individuality,” described as follows
(Gould 1991, pp 506–508):
The surfaces of planets and moons cannot be
predicted from a few general rules. To understand
planetary surfaces, we must learn the particular
history of each body as an individual object … their
major features are set by unique events—mostly
catastrophic—that shape their surfaces … Planets
are like individual organisms not water molecules;
they have irreducible personalities built by history …
ACKNOWLEDGMENTS
This necessarily brief overview cites only a small sampling
from hundreds of geomorphological publications that doc-
FIGURE 3 Mars Express High Resolution Stereo Camera perspective
view (30X vertical exaggeration) of debris that has flowed ument past aqueous activity on Mars. The author both
from a 4 km high mountain massif, into two adjacent craters. The debris thanks and apologizes to authors of studies not included,
is organized into crater-filling flow patterns of parallel ridges that spread especially to the many who recognized, as early as the
out to a width of 16 km in the foreground. This landform assemblage, 1970s and 1980s, the compelling geomorphological
evidence for Mars’ aqueous past. .
located east of Hellas, is analogous to terrestrial debris-covered pied-
mont glaciers. Numerous pits and depressions show wastage of the gla-
cial ice, which may have been active only several tens of millions of years
ago (Head et al. 2005).

REFERENCES Clifford SM, Parker TJ (2001) The evolution Kuzmin RO (2005) Ground ice in the
of the martian hydrosphere: Implications Martian regolith. In: Tokano T (ed) Water
Baker VR (2001) Water and the martian for the fate of a primordial ocean and the on Mars and Life, Advances in Astrobiol-
landscape. Nature 412: 228-236 current state of the northern plains. ogy and Biogeophysics, Springer,
Baker VR (2004) A brief geological history Icarus 154: 40-79 Heidelberg, pp 155-189
of water on Mars. In: Seckbach J (ed) Craddock RA, Howard AD (2002) The case Malin MC, Edgett KS (2000a) Sedimentary
Origins, Kluwer, Dordrecht, pp 623-631 for rainfall on a warm, wet early Mars. rocks of early Mars. Science 290: 1927-
Baker VR, Strom RG, Gulick VC, Kargel JS, Journal of Geophysical Research 1937
Komatsu G, Kale VS (1991) Ancient 107(E11):5111, doi: 10.1029/
2001JE001505 Malin MC, Edgett KS (2000b) Evidence for
oceans, ice sheets and the hydrological recent groundwater seepage and surface
cycle on Mars. Nature 352: 589-594 Fairén AG, Dohm JM, Baker VR, de Pablo runoff on Mars. Science 288: 2330-2335
Berman DC, Hartmann WK (2002) Recent MA, Ruiz J, Ferris JC, Anderson RC (2003)
Episodic flood inundations of the northern Malin MC, Edgett KS (2003) Evidence for
fluvial, volcanic, and tectonic activity on persistent flow and aqueous sedimentation
the Cerberus plains of Mars. Icarus 159: plains of Mars. Icarus 165: 53-67
on early Mars. Science 302: 1931-1934
1-17 Gould SJ (1991) Bully for Brontosaurus,
W.W. Norton, New York, 540 pp Mangold N, Costard F, Forget F (2003)
Bhattacharya JP, Payenberg THD, Lang SC, Debris flows over sand dunes on Mars:
Bourke M (2005) Dynamic river channels Head JW, Mustard JF, Kreslavsky MA, Evidence for liquid water. Journal of
suggest a long-lived Noachian crater lake Milliken RE, Marchant DR (2003) Recent Geophysical Research 108 (E4): 5027, doi:
on Mars. Geophysical Research Letters ice ages on Mars. Nature 426: 797-802 10.1029/2002JE001958
32, L10201, doi: 10.1029/2005GL022747
Head JW, Neukum G, Jaumann R, Miyamoto H, Dohm JM, Beyer RA, Baker
Boyce JM, Mouginis-Mark P, Garbeil H Hiesinger H, Hauber E, Carr M, Masson P, VR (2004) Fluid dynamical implications
(2005) Ancient oceans in the northern Foing B, Hoffmann H, Kreslavsky M, of anastomosing slope streaks on Mars.
lowlands of Mars: Evidence from impact Werner S, Milkovich S, van Gasselt S, Journal of Geophysical Research 109:
crater depth/diameter relationships. HRSC Co-Investigator Team (2005) E06008, doi:10.1029/2003JE002234
Journal of Geophysical Research 110, Tropical to mid-latitude snow and ice
E03008, doi 10.1029/2004JE002328 accumulation, flow and glaciation on Moore JM, Howard AD, Dietrich WE,
Mars. Nature 434: 346-351 Schenk PM (2003) Martian layered fluvial
Boynton WV and 24 coauthors (2002) deposits: Implications for Noachian
Distribution of hydrogen in the near Hoffman N (2000) White Mars: A new climate scenarios. Geophysical Research
surface of Mars: Evidence for subsurface model for Mars’ surface and atmosphere Letters 30: 2292, doi:
ice deposits. Science 297: 81-85 based on CO2. Icarus 146: 326-342 10.1029/2003GL019002
Burr DM, Grier JA, McEwen AS, Keszthelyi Hynek BM, Phillips RJ (2003) New data Moore JM, Howard AD (2005) Large
LP (2002) Repeated aqueous flooding reveal mature integrated drainage systems alluvial fans on Mars. Journal of
from the Cerberus Fossae: evidence for on Mars indicative of past precipitation. Geophysical Research 110, E04005, doi
very recently extant, deep groundwater Geology 31: 757-760 10.1029/2004JE002352
on Mars. Icarus 159: 53-73
Irwin RP III, Howard AD, Maxwell TA Mustard JF, Cooper CD, Rifkin MK (2001)
Cabrol NA, Grin EA (2002) Overview on (2004) Geomorphology of Ma’adim Evidence for recent climate change on
the formation of paleolakes and ponds Vallis, Mars, and associated paleolake Mars from the identification of youthful
on Mars. Global and Planetary Change basins. Journal of Geophysical Research near-surface ground ice. Nature 412: 411-
35: 199-219 109, E12009, doi 10.1029/2004JE002287 414
Carr MH (1996) Water on Mars, Oxford, Irwin RP III, Craddock RA, Howard AD Seibert NM, Kargel JS (2001) Small-scale
New York, 229 pp (2005) Interior channels in Martian Martian polygonal terrain: Implications
Carr MH, Head JW III (2003) Oceans on valley networks: Discharge and runoff for liquid surface water. Geophysical
Mars: An assessment of observational production. Geology 33: 489-492 Research Letters 28: 899-902
evidence and possible fate. Journal of Jerolmack DJ, Mohrig D, Zuber MT, Byrne S Squyres SW and 18 coauthors (2004) In
Geophysical Research 108(E5): 5042, doi (2004) A minimum time for the situ evidence for an ancient aqueous
10.1029/2002JE001963 formation of Holden Northeast fan, Mars. environment at Meridiani Planum, Mars.
Chapman MG, Allen CC, Gudmundsson Geophysical Research Letters 31: L21701, Science 306: 1709-1714
MT, Gulick VC, Jakobsson SP, Lucchitta doi: 10.1029/2004GL021326
Webb VE (2004) Putative shorelines in
BK, Skilling IP, Waitt RB (2000) Kargel JS (2004) Mars: A Warmer, Wetter northern Arabia Terra, Mars. Journal of
Volcanism and ice interactions on Earth Planet, Springer-Praxis, Chichester, UK, Geophysical Research 109: E09010, doi
and Mars. In: Gregg TKP, Zimbelman JR 557 pp 10.1029/2003JE002205 .
(eds) Deep Oceans to Deep Space:
Environmental Effects of Volcanic
Eruptions, Plenum, New York, pp 39-74

The Orbital Search for
Altered Materials on Mars
Michael B. Wyatt1 and Harry Y. McSween Jr.2
regolith consists of a mixture of
T
he Martian surface is dominated by primary igneous minerals common
bright (high-albedo), globally
in basaltic rocks. Limited chemical alteration exists in fine-grained dust, homogeneous, fine-grained dust and
and is likely in sands and rocks at high latitudes and in the northern dark (low-albedo), locally derived,
lowland plains where materials have interacted with ice and snow. Evidence sand drifts and dunes atop variably
indurated soil, rock fragments, and
for extensive production of secondary phases is revealed at higher spatial
layered bedrock exposures. The
resolutions, where alteration effects of unique, and perhaps time-limited, ancient southern highlands are of
aqueous environments are observed. The distribution of ice on Mars thus Noachian to Hesperian age, whereas
the northern lowlands are com-
appears to have a global influence on the production of alteration materials,
posed of younger Hesperian to
whereas the effects of water are discovered in unique and locally diverse Amazonian materials covering a
geological settings. Noachian basement. The white
line in FIGURE 1A approximates a
40 km crustal-thickness dichotomy
INTRODUCTION separating thinner crust to the north from thicker crust to
Recent observations by the Mars Global Surveyor (MGS) the south (Zuber et al. 2000), and the yellow line marks the
TES, Mars Odyssey (MO) THEMIS, and Mars Express (MEX) boundary of the Vastitas Borealis Formation (VBF) (Tanaka
OMEGA instruments (see glossary on page 138 for et al. 2003). VBF materials have been interpreted as sediments
acronyms in text) have greatly improved knowledge of the reworked by near-surface, in situ volatile-driven processes
abundances and distributions of Martian minerals and have (Tanaka et al. 2003) and as a sublimation residue from
provided insight into past geological environments. This frozen bodies of water (Kreslavsky and Head 2002).
paper summarizes major discoveries about surface composi- There is a bimodal distribution of near-surface H2O ice
tions from orbital experiments, focusing on investigations measured by the Gamma Ray Spectrometer (GRS) (e.g.
of altered materials and their implications for the history of Boynton et al. 2002) and MOC- and THEMIS-observed ice-rich
water on Mars. The search for chemically derived secondary mantles (FIG. 1B). Mantles have been interpreted as meter-
phases on Mars is driven, in large part, by physical evidence thick, ice-rich sediments (Mustard et al. 2001) or remnant
of surface–volatile interactions in the form of ancient, fluvial, snowpacks (Christensen 2003) and are thought to form during
dissected channels and surfaces recently modified by ice. phases of high obliquity. Development of near-surface ice is
The search began over four decades ago with the use of likely connected to surface ice deposition as abundances are
Earth-based telescopes and early robotic explorers. Only too high to be accounted for by vapor diffusion alone (e.g.
recently, however, have observations from TES, THEMIS, Mustard et al. 2001; Christensen 2003; Head et al. 2003).
and OMEGA been used to construct global surface-compo-
sition maps and search for local exposures of altered materials EARLY SEARCHES FOR ALTERATION
at high resolution. The production of secondary phases
depends on a number of factors, including the composi- Investigations of high- and low-albedo surfaces with visible/
tions of precursor igneous minerals and glasses, the nature near-infrared (VNIR) spectra from Earth-based telescopes
of the alteration environment (gas, aqueous, solid), temper- and early Mars spacecraft provided a framework of Martian
ature, pH, and time. Global variations in surface mineralogy compositions (reviewed by Bell 1996). Spectral observations
are crucial to understanding the dynamic interaction in the VNIR are sensitive to electron transitions of metals,
between the atmosphere, hydrosphere, and crust. especially iron, and molecular vibrational transitions in
minerals, which can be used to determine chemical and
mineralogical compositions. FIGURE 2A shows composite,
GEOLOGICAL SETTING
telescopic-orbital (Phobos-2 Imaging Spectrometer for
A globally projected simple-cylindrical image (FIG. 1A) made Mars) VNIR spectra representative of Martian high- and
from MOC and MOLA datasets provides the geological context low-albedo surfaces (Mustard and Bell 1994).
for the distributions of Martian mineral abundances. The
High-Albedo Surfaces
1 Department of Geological Sciences VNIR observations of high-albedo surfaces are consistent
Arizona State University with highly oxidized, fine-grained materials characterized
Tempe, AZ 85287-6305, USA by poorly crystalline iron oxides and a small amount of
crystalline iron oxides (e.g. Morris et al. 1997). Subtle spectral
2 Department of Earth and Planetary Sciences differences in high-albedo surfaces suggest small variations
University of Tennessee
Knoxville, TN 37996-1410, USA in the abundance, and possibly composition, of crystalline
E-mail: [email protected] iron oxide materials. A weak spectral band in the 2.2 µm

region may be indicative of a hydroxyl-bearing phase, while over rates of chemical alteration. This was the beginning of
a strong 3 µm feature suggests an H2O-bearing phase in the an apparent dichotomy between geomorphological evidence
dust. VNIR spectra of high-albedo surfaces on Mars are reason- of extensive surface–volatile interactions and the lack of
ably well matched by some terrestrial palagonitic tephras. compositional evidence to support such observations. How-
ever, during the last decade, instruments on new orbiting
Low-Albedo Surfaces spacecraft have produced an unprecedented amount of data
VNIR observations of low-albedo surfaces are characterized to shed light on the history of water on Mars.
by absorption at wavelengths of ~1 and 2 µm, which are
attributed primarily to the presence of ferrous (Fe2+) iron in MGS-TES GLOBAL SURFACE COMPOSITIONS
the form of pyroxene and hematite (both nanophase and TES is a Fourier Transform Michelson Interferometer that
crystalline). The identification of these features in spectra of measures emitted infrared energy from 1700 to 200 cm-1
dark surfaces in Syrtis Major is interpreted to represent the (~6 to 50 µm) at 10 or 5 cm-1 sampling (e.g. Christensen et
existence of two-pyroxene basalts similar to some Martian al. 2000a). TES began systematic mapping in 1999, and global
basaltic meteorites (Mustard and Sunshine 1995). coverage is now complete. Each spectrum represents the
average emissivity of a 3 × 9 km footprint on the Martian
Implications for Water
surface. Thermal infrared spectra are sensitive to the chem-
Based on these early studies, Mars appeared to be dominated istry and structure of virtually all common rock-forming
by unaltered basalt and moderately altered, fine-grained silicates and exhibit unique signatures that arise from the
dust, consistent with the current cold and dry Martian envi- vibrational motions of atoms in the crystal structure
ronment in which rates of physical weathering dominate
FIGURE 1 (A) Major subdivisions of Martian crust illustrated by using ice (blue) (e.g. Boynton et al. 2002) and mantle deposits (shaded gray
MGS MOC and MOLA datasets. VBF—Vastitas Borealis from 30° to ~60°; poleward of 60°, continuous mantle deposits overlie
Formation (Tanaka et al. 2003). (B) Distribution of near-surface H2O ice GRS-measured near-surface ice). MOC observations of discontinuous
and ice-rich mantles. MOC-MOLA context image with GRS near-surface mantle deposits are shown as white points (Mustard et al. 2001).

A C
(e.g. Si–O stretching and bending in silicates). In complex FIGURE 2 (A) Composite telescopic-ISM VNIR spectra representa-
mixtures, such as rocks, spectral features of component tive of Martian high- and low-albedo surfaces (Mustard
and Bell 1994). (B) TES spectrum representative of high-albedo surfaces
minerals combine linearly to produce a mixed, composite (Bandfield et al. 2003) and ST1 and ST2 spectral end members of low-
spectrum. Conversely, mixtures can be deconvolved into albedo surfaces (Bandfield et al. 2000). (C) TES ST1 and ST2 spectral
abundances of their component minerals (i.e. modal min- end members compared to Meridiani Planum spectrum showing
eralogy) using a spectral end-member library of known hematite absorptions (Christensen et al. 2000b).
mineral compositions (Ramsey and Christensen 1998).
High-Albedo Surfaces Low-Albedo Surfaces

Analyses of TES data reveal a spectral shape that is repre- The bulk variability of low-albedo surface compositions
sentative of all high-albedo regions and is characterized by measured by TES is accounted for by two distinct spectral
a relatively narrow absorption near 830 cm-1 and a broad, end members (FIG. 2B) (Bandfield et al. 2000). The Surface
deep absorption at >1250 cm-1 (FIG. 2B) (Bandfield et al. Type 1 (ST1) spectrum is characterized by a broad, slightly
2003). The overall spectrum is best matched by the frame- square-shaped absorption between 800 and 1200 cm-1.
work silicates plagioclase (Bandfield et al. 2003) and zeolite Absorption from 200 to 500 cm-1 has an overall negative
(Ruff 2004) and is consistent with volume scattering effects slope with decreasing wavenumbers and contains minor,
in fine particulates. The spectrum also exhibits the charac- narrow absorption bands. The Surface Type 2 (ST2) spectrum
teristic absorption band near 1600 cm-1 of mineral-bound is characterized by a more rounded, slightly V-shaped 800
water. This feature is best matched by a similar peak in zeolite, to 1200 cm-1 region of absorption and uniform absorption
whose structure, unlike plagioclase, contains molecular at low wavenumbers.
water (Ruff 2004). The spectral absorption near 1480 cm-1 ST1 is spectrally similar to terrestrial continental flood
corresponds well to the 1450 to 1500 cm-1 bands seen in basalts (Bandfield et al. 2000) and consists of high abun-
anhydrous carbonates and most closely matches that of the dances of plagioclase (~40–50 vol%) and high-calcium
Mg-carbonate magnesite (Bandfield et al. 2003). Only small clinopyroxene (~30–40 vol%) and small amounts of olivine
concentrations of carbonate (~2 to 5 wt%) are required to (McSween et al. 2003). ST2 is spectrally similar to terrestrial
produce the observed spectral signature. arc andesites (Bandfield et al. 2000) and partly altered
basalts (Wyatt and McSween 2002; Morris et al. 2003; Kraft
Implications for Water et al. 2003; Ruff 2004; Michalski et al. 2005). Ambiguity in
Evidence of zeolite and carbonate in fine-grained dust classifying the ST2 lithology arises because a spectral
implies widespread production of secondary phases on component of this unit (~20–30 vol%) can be interpreted as
Mars; however, the amount of volatiles and the extent of volcanic siliceous glass (an abundant phase in andesite) or
alteration are likely limited. On Earth, abundant occur- a combination of secondary phases found in altered basalt
rences of zeolite are often found in tephra deposits altered (smectite, palagonite, amorphous silica–rich coatings, and
by interactions with groundwater. However, zeolites are zeolite). ST2 also contains high amounts of plagioclase
also produced in the cold and dry environment of Antarctica, (~30–40 vol%) and less-high-calcium clinopyroxene
a setting that is perhaps most analogous to current conditions (~10–20 vol%) (McSween et al. 2003).
on Mars (e.g. Gibson et al. 1983). Similarly, carbonates are
found in a variety of environments as primary aqueous Classification of ST2 Lithology
precipitates. However, Mg and Ca carbonates may also form The debate over the ST2 lithology is significant for under-
under low–CO2 pressure conditions, in which liquid water standing the petrogenesis and subsequent alteration of the
is not even stable but is present in a transient state, and Martian crust. Identification of widespread andesite may
should be common, stable weathering products in the present imply an early episode of plate tectonics on Mars, while
Martian atmosphere (Booth and Kieffer 1978). The lack of altered basalt would indicate extensive surface–volatile
any spectral evidence for a carbonate bedrock source, which interactions. The competing spectral interpretations were
may be expected to form if large standing bodies of water addressed by Wyatt et al. (2004) by examining the geological
ever existed on Mars, further supports an interpretation of context (FIG. 1A, 1B) of the ST1 and ST2 global distribution
limited global chemical weathering. pattern illustrated in FIGURE 3.
ST1 Basalt and ST2 Andesite: FIGURE 3 shows no global
systematic relationship between crustal thickness (40 km
dichotomy white line) (Rogers and Christensen 2003) or

age and the distribution of ST1 (green) and ST2 (red), as deposits and the VBF boundary. Thus, alteration of sedi-
would be expected in a basalt–andesite model. Andesitic ments in the northern lowlands may have been enhanced
volcanism on Earth is mostly associated with thick conti- by temporary standing bodies of water and ice.
nental crust; however, on Mars the largest distribution of
ST2 overlies thin crust in the northern plains. Moreover, Implications for Water
the occurrence of ST2 without associated ST1 in the northern Although the dominance of igneous minerals measured by
plains argues against its being derived by fractionation of TES for ST1 and ST2 implies limited chemical alteration on
basaltic magma. Fractionation should produce basaltic and Mars, alteration is nevertheless greater at high latitudes and
andesitic rocks similar in age, but ST2 materials in the in the low-lying northern plains. The Dry Valleys of Antarctica
northern lowlands are significantly younger than southern may be the best terrestrial analogue for weathering on Mars
highlands basalts. Partial melting of an ancient basaltic because of the cold, hyper-arid environment, stable
crust (rather than ultramafic mantle) might conceivably permafrost, and ground ice (Gibson et al. 1983). Basalts in
produce andesitic magmas, especially under wet condi- these environments are dominated by plagioclase and
tions, but the absence of Martian meteorites having pyroxene, with limited abundances of alteration phases
andesitic compositions and appropriate ages argues against similar to those proposed for ST2 materials (palagonites,
this possibility. zeolites, smectites, and silica coatings). All of these secondary
products can be produced on Mars by chemical weathering
ST1 Basalt and ST2 Altered Basalt: FIGURE 3 indicates that ST1
of basalt without an abundance of water.
materials dominate in equatorial and mid-latitude regions
and ST2 materials dominate in the high-latitude northern
lowlands and southern highlands. Wyatt et al. (2004) relate LOCAL SURFACE COMPOSITIONS
this spatial distribution to near-surface ice and ice-rich We now take the orbital search for alteration materials and
mantle deposits (FIG. 1B) and propose both a latitude and evidence of a possibly warmer and wetter ancient Mars climate
topographic influence on the global surface alteration of to local scales using high-spatial-resolution observations
Mars. The lower part of FIGURE 3 is a plot of ST1 and ST2 from MGS TES, ODY THEMIS, and MEX OMEGA. The ODY
normalized abundances from 90° south to 90° north that THEMIS is a multispectral imager that measures nine ther-
are averaged across 0.5° bins of longitude and compared (in mal infrared (TIR) bands centered between 6.8 and 14.9 µm
a histogram) to percentages of dissected ice-rich mantle and five VNIR bands centered between 0.42 and 0.86 µm
deposits (Mustard et al. 2001). The gradual transition from (Christensen et al. 2003). The spatial footprint of TIR
ST1 to ST2 in the Southern Hemisphere correlates well with images is 100 m/pixel while the VNIR is 18 m/pixel. The
the transition from a lack of ice-rich material (0°–25°S), to a MEX OMEGA is a hyperspectral imaging spectrometer that
maximum percentage of dissection (25°S–60°S), to uniform measures 352 channels in the VNIR between 0.35 µm and
mantles of ice-rich deposits (60°S–90°S). This trend is inter- 5.1 µm with a spatial resolution that varies from
preted to reflect increased amounts of chemical weathering 300 m/pixel to 4.8 km/pixel (Bibring et al. 2005).
resulting from basalt interactions with icy mantles. In the
Northern Hemisphere, an abrupt transition from ST1 to ST2 Focus on Meridiani Planum
occurs at ~20° and is correlated both with ice-rich mantle The geological setting of Terra Meridiani is among the most
complex on Mars and includes exposures of ancient, fluvial-
dissected cratered terrains (DCT) and younger, overlying
layered materials and dust mantles (Arvidson et al. 2003).
A
Physical erosion has shaped layered materials into domes,
ridges, and plateaus, referred to collectively as etched terrains
(ET). A smooth plains unit (PH) covers these landforms and
is overlain by younger mantle deposits (MCT) to the north.
A mosaic of THEMIS daytime infrared images of Meridiani
Planum is shown in FIGURE 4A.
TES spectral analyses of the smooth plains unit reveal
prominent absorptions in the 250–500 cm-1 region that are
significantly different from global end members (FIG. 2C)
(Christensen et al. 2000b). These distinguishing spectral
features are caused by gray crystalline hematite mixed with
ST1 abundances of plagioclase and pyroxene. FIGURE 4B
B shows the occurrence of hematite-bearing material in
Meridiani Planum superimposed on a shaded relief image
derived from MOLA data. Hematite abundances vary from
~5% (blue) to ~20% (red) and make an irregularly shaped
unit spanning 500 by 300 km. Comparisons of TES data
with hematite laboratory spectra derived from different
precursor phases suggests low-temperature (~100–300°C)
dehydroxylation of goethite as the most plausible mechanism
of formation (Glotch et al. 2004). Embayment relationships
between hematite-bearing and hematite-poor units suggest
that the formation of hematite was confined vertically and
FIGURE 3 (A) TES image of distribution of ST1 (green) and ST2 that precursor materials were likely deposited in a gravity-
(red) materials and dust (blue). White line approximates driven fluid, rather than as dispersed air fall (Christensen
a 40 km crustal-thickness dichotomy (Zuber et al. 2000), separating
and Ruff 2004). The probable role of water in the formation
thinner crust to the north from thicker crust to the south. Yellow line
marks the boundary of the VBF (Tanaka et al. 2003). (B) ST1 and ST2 of this large hematite deposit led to the selection of this site
normalized abundances from 90° south to 90° north, averaged across for in situ exploration by the Mars Exploration Rover (MER)
0.5º bins of longitude and compared to percentages of MOC observa- Opportunity.
tions (histogram) (Mustard et al. 2001) of dissected ice-rich mantle
deposits (Wyatt et al. 2004).

FIGURE 4 (A) THEMIS daytime infrared
A B mosaic of Meridiani Planum.
DCT: dissected cratered terrains; ET:
etched terrains; PH: smooth plains unit;
MCT: mantle deposits. (B) MOLA shaded
relief image with superimposed TES-
measured hematite abundances (Chris-
tensen et al. 2000b).
Implications for Water A

Opportunity instruments now on the surface of Meridiani A
Planum confirm the presence of coarse crystalline hematite
previously detected in orbit by TES (Squyres et al. 2004).
The layered bedrock from which hematite is derived is rich
in Ca and Mg sulfates (Mini-TES and Alpha Particle X-Ray
Spectrometer) and jarosite (Mössbauer Spectrometer), indi-
cating extensive alteration and fluvial activity during
hematite formation. Sulfate (kieserite) is also identified
regionally in etched terrain materials from spectral analyses
of OMEGA data (Gendrin et al. 2005; Arvidson et al. 2005). B
The mineralogical similarities among these deposits and the B
rocks examined by Opportunity imply that the inferred
ancient aqueous environment extends over large scales and
throughout the several-hundred-meter thickness of etched
terrain deposits. However, the mixing of unaltered basaltic
sands and hematite indicates that chemical alteration has
been limited since hematite formation. These results com-
plement interpretations by Christensen and Ruff (2004)
that a small number of bodies of standing water altered
Meridiani Planum surface materials but that these water
bodies represent only localized phenomena against a back-
drop of a cold, frozen Mars.
C
THE SEARCH CONTINUES
The Martian surface is dominated by primary igneous minerals
common in basaltic rocks. However, an imprint of limited
chemical alteration exists in fine-grained dust and is likely
in sands and rocks at high latitudes and in the northern D
lowland plains where materials have interacted with ice and
snow. These secondary phases do not require abundant
water and indicate that Mars has likely been a cold and
frozen planet for an extensive period of geological time.
However, it is at higher spatial resolutions that we observe
the alteration effects of unique, and perhaps time-limited, FIGURE 5 THEMIS daytime infrared mosaic of Aram Chaos and Ares
environments where water once existed. Further detailed Vallis showing TES-measured (A) olivine (Rogers et al.
2005) and (B) hematite abundances (Glotch and Christensen 2005) and
studies have resulted in the discovery of additional expo- (C) sulfate detected by OMEGA (Gendrin et al. 2005). The three- dimen-
sures of altered materials. sional image of the THEMIS mosaic (D) has been overlain on MOLA
topography.
The discovery of hematite and sulfate in Valles Marineris
and of sulfate in the dark longitudinal dunes of Olympia
Planitia indicates that water played a major role in the and easily altered igneous minerals are mapped in close
formation of interlayered strata and northern circumpolar proximity. Hematite and sulfate occur in the center of the
terrains (Christensen et al. 2000b; Gendrin et al. 2005; flat-floored impact crater Aram Chaos, where possible lake-
Langevin et al. 2005). Phyllosilicates that originate from the bed sediments formed from the release of subsurface water
alteration of mafic and ultramafic igneous rocks have been (Glotch and Christensen 2005; Gendrin et al. 2005).
identified in ancient terrains of Syrtis Major (Bibring et al. Olivine has been identified in bedrock layers within Ares
2005). Each of these new local occurrences represents addi- Vallis, a fluvial-dissected channel, but one where subsequent
tional evidence for environmental change as compared chemical alteration is limited (Rogers et al. 2005).
with the cold and dry Martian climate of today. Such These mineral signatures reveal that Mars has been a
change is clearly illustrated in FIGURE 5, where secondary dynamic and changing world. The orbital search for alteration

minerals has shed light on the role of water and ice in ACKNOWLEDGMENTS
modifying the compositions of surface materials. Ice
The authors wish to thank John Mustard and an anony-
appears to have a global influence on the production of
mous reviewer for helpful comments that improved the
secondary phases, whereas the effects of water are discovered
quality of this manuscript. We would also like to sincerely
in unique and locally diverse geological settings. Future
thank all of those who built and continue to operate the
MGS, Odyssey, and MER spacecraft. .
orbiters and landers will focus on these local settings in the
search for evidence of habitable Martian environments, and
perhaps even signs of extant or present life.
REFERENCES Christensen PR and 21 coauthors (2003) Journal of Geophysical Research 102:

Morphology and composition of the 9125-9133
Arvidson RE, Seelos FP IV, Deal KS, Koeppen surface of Mars: Mars Odyssey THEMIS
WC, Snider NO, Kieniewicz JM, results. Science 300: 2056-2061 Morris RV, Graff TG, Mertzman SA, Lane
Hynek BM, Mellon MT, Garvin JB MD, Christensen PR (2003) Palagonitic
(2003) Mantled and exhumed terrains Gendrin A, Mangold N, Bibring J-P, (not andesitic) Mars: Evidence from
in Terra Meridiani, Mars. Journal of Langevin Y, Gondet B, Poulet F, thermal emission and VNIR spectra of
Geophysical Research 108(E12): 8073, Bonello G, Quantin C, Mustard J, palagonitic alteration rinds on basaltic
doi:10.1029/2002JE001982 Arvidson R, LeMouélic S (2005) Sulfates rocks [abs.]: Sixth International
in martian layered terrains: The Conference on Mars, abstract 3111
Arvidson RE, Poulet F, Bibring J-P, Wolff M, OMEGA/ Mars Express view. Science
Gendrin A, Morris RV, Freeman JJ, 307: 1587-1591 Mustard JF, Bell JF III (1994) New composite
Langevin Y, Mangold N, Bellucci G reflectance spectra of Mars from 0.4
(2005) Spectral reflectance and Gibson EK, Wentworth SJ, McKay DS (1983) to 3.14 µm. Geophysical Research
morphologic correlations in eastern Chemical weathering and diagenesis Letters 21: 353-356
Terra Meridiani, Mars. Science 307: of a cold desert soil from Wright Valley,
1591-1594 Antarctica: An analog of Martian Mustard JF, Sunshine JM (1995) Seeing
weathering processes. Journal of through the dust: Martian crustal
Bandfield JL, Hamilton VE, Christensen PR Geophysical Research 88: A 912-A928 heterogeneity and links to the SNC
(2000) A global view of martian surface meteorites. Science 267: 1623-1626
compositions from MGS-TES. Science Glotch TD, Morris RV, Christensen PR,
287: 1626-1630 Sharp TG (2004) Effect of precursor Mustard JF, Cooper CD, Rifkin MK (2001)
mineralogy on the thermal infrared Evidence for recent climate change
Bandfield JL, Glotch TD, Christensen PR emission spectra of hematite: on Mars from the identification of
(2003) Spectroscopic identification Application to Martian hematite youthful near-surface ground ice.
of carbonate minerals in the Martian mineralization. Journal of Geophysical Nature 412: 411-413
dust. Science 301: 1084-1087 Research 109: E07003, doi: Ramsey MS, Christensen PR (1998) Mineral
Bell JF (1996) Iron, sulfate, carbonate, and 10.1029/2003JE002224 abundance determination: Quantitative
hydrated minerals on Mars. In: Dyar Glotch TD, Christensen PR (2005) Geologic deconvolution of thermal emission
MD, McCammon C, Schaefer MW (eds) and mineralogic mapping of Aram spectra. Journal of Geophysical
Mineral Spectroscopy: A Tribute to Chaos: Evidence for a water-rich Research 103(B1): 577-596
Roger G. Burns. Geochemical Society history. Journal of Geophysical
Special Publication 5, pp 359-380 Rogers AD, Christensen PR (2003) Age
Research 110: E9006, doi: relationship of basaltic and andesitic
Bibring J-P, Langevin Y, Gendrin A, Gondet 10.10129/2004JE002389 surface compositions on Mars: Analysis
B, Poulet F, Berthé M, Soufflot A, Head JW, Mustard JF, Kreslavsky MA, of high-resolution TES observations of
Arvidson R, Mangold N, Mustard J, Milliken RE, Marchant DR (2003) the northern hemisphere. Journal of
Drossart P, OMEGA team (2005) Mars Recent ice ages on Mars. Nature 426: Geophysical Research 108(E4): 5030,
surface diversity as revealed by the 797-802 doi: 10.1029/2002JE001913
OMEGA/Mars Express observations.
Science 307: 1576-1581 Kraft MD, Michalski JR, Sharp TG (2003) Rogers AD, Christensen PR, Bandfield JL
Effects of pure silica coatings on (2005) Compositional heterogeneity
Booth MC, Kieffer HH (1978) Carbonate thermal emission spectra of basaltic of the ancient Martian crust: Analysis
formation in Marslike environments. rocks: Considerations for Martian of Ares Vallis bedrock with THEMIS
Journal of Geophysical Research 83: surface mineralogy. Geophysical and TES data. Journal of Geophysical
1809-1815 Research Letters 30: 2288, doi: Research 110: E05010,
10.1029/2003GL018848 doi:10.1029/2005JE002399
Boynton WV and 24 coauthors (2002)
Distribution of hydrogen in the near Kreslavsky MA, Head JW (2002) Fate of Ruff SR (2004) Spectral evidence for zeolite
surface of Mars: Evidence for subsurface outflow channel effluents in the in the dust on Mars. Icarus 168: 131-143
ice deposits. Science 297: 81-85 northern lowlands of Mars: The Squyres SW and 49 coauthors (2004) The
Christensen PR (2003) Formation of recent Vastitas Borealis Formation as a Opportunity Rover’s Athena science
martian gullies through melting of sublimation residue from frozen investigation at Meridiani Planum,
extensive water-rich snow deposits. ponded bodies of water. Journal of Mars. Science 306: 1698-1703
Nature 422: 45-48 Geophysical Research 107(E12): 5121,
doi: 10.1029/2001JE001831 Tanaka KL, Skinner JA Jr, Hare TM, Joyal T,
Christensen PR, Ruff SW (2004) Formation Wenker A (2003) Resurfacing history
of the hematite-bearing unit in Langevin Y, Poulet F, Bibring J-P, Gondet B of the northern plains of Mars based
Meridiani Planum: Evidence for (2005) Sulfates in the north polar on geologic mapping of Mars Global
deposition in standing water. Journal region of Mars detected by OMEGA/ Surveyor data. Journal of Geophysical
of Geophysical Research 109: E08003, Mars Express. Science 307: 1584-1586 Research 108(E4): 8043, doi:
doi:10.1029/2003JE002233 McSween HY Jr, Grove TL, Wyatt MB (2003) 10.1029/2002JE001908
Christensen PR, Bandfield JL, Smith MD, Constraints on the composition and Wyatt MB, McSween HY Jr (2002) Spectral
Hamilton VE, Clark, RN (2000a) petrogenesis of the Martian crust. evidence for weathered basalt as an
Identification of a basaltic component Journal of Geophysical Research alternative to andesite in the northern
on the Martian surface from Thermal 108(E12): 5135, doi:10.1029/ lowlands of Mars. Nature 417: 263-266
Emission Spectrometer data. Journal 2003JE002175
of Geophysical Research 105(E4): Wyatt MB, McSween HY, Tanaka KL, Head
Michalski JR, Kraft MD, Sharp TG, Williams JW III (2004) Global geologic context
9609-9621 LB, Christensen PR (2005) Mineralogical for rock types and surface alteration
Christensen PR and 15 coauthors (2000b) constraints on the high-silica martian on Mars. Geology 32: 645-648
Detection of crystalline hematite surface component observed by TES.
mineralization on Mars by the Thermal Icarus 174: 161-177 Zuber MT and 14 coauthors (2000) Internal
Emission Spectrometer: Evidence for structure and early thermal evolution
Morris RV, Golden DC, Bell JF III (1997) of Mars from Mars Global Surveyor
near-surface water. Journal of Geo- Low-temperature reflectivity spectra topography and gravity. Science 287:
1788-1793. .
physical Research 105(E4): 9623-9642 of red hematite and the color of Mars.

Water at the Poles
and in Permafrost Regions
of Mars
Philip R. Christensen1
the summer at both poles. The
T
he poles and mid-latitudes of Mars contain abundant water in ice caps,
thickness of the seasonal CO2 ice
thick sequences of ice-rich layers, and mantles of snow. The volume
caps has been estimated from Mars
of the known reservoir is ≥5 x 106 km3, corresponding to a layer ~35 m Global Surveyor (MGS) Mars Orbiter
thick over the planet. Hydrogen in subsurface H2O ice has been detected Laser Altimeter (MOLA) observa-
at latitudes poleward of 50°. Morphological features show downslope flow tions to reach approximately 1.5 m
of ice-rich sediment, and recent gullies have been produced from subsurface near the pole in both hemispheres
(Smith et al. 2001), corresponding
aquifers or melting snowpacks. Variations in Mars’ orbit on timescales of
to ~25% of the total mass of the
50,000 to 2,000,000 years produce significant changes in climate, which Martian atmosphere.
result in the transport of water from the poles, where it currently resides,
As the seasonal caps condense,
to the lower latitudes, where it may play a critical role in surface geology, they incorporate minor amounts
mineralogy, and geochemistry. of dust and H2O ice, which signifi-
cantly affect the sublimation rates
KEYWORDS: Mars, ice, water, polar caps the following spring. Assuming a
water vapor mass fraction of
INTRODUCTION 1 × 10-5 in the condensing atmosphere, the amount of
Water has long been recognized as a major morphological water stored in the seasonal caps is estimated at ~3 × 1010 kg
agent on Mars (Baker this issue), but its present abundance (~3 × 10-2 km3) (the Martian atmosphere contains ~10-1
and location remain enigmatic. A possible reservoir for a km3 of water). Overall, however, the water within the sea-
substantial amount of water is surface and subsurface ice at sonal caps plays a relatively minor role in the global inven-
the poles and mid-latitudes. Recent observations have sub- tory or annual cycle of water on Mars.
stantially improved our knowledge of this reservoir, but
major questions remain as to its volume, age, and history.
Oscillations in the axial tilt, eccentricity, and timing of clos-
est approach to the Sun cause major changes in surface
heating, which produce cyclic changes in Martian climate
on timescales of 105 to 106 years (e.g. Pollack and Toon
1982). These changes redistribute polar ice, transferring it
to lower latitudes as snow and ice during Martian “ice ages”
(e.g. Jakosky et al. 1995). Today the major ice-bearing fea-
tures are the polar ice caps, the layered units that surround
them at both poles, and the mid-latitude permafrost zones
that present morphologies strongly suggestive of subsurface
ice. Each of these has unique properties, water abundances,
and histories, and contributes to the water cycle in
varying ways.
POLAR CAPS
The polar caps of Mars (FIG. 1) have been observed since the
17th century and are assumed to be composed of some
combination of H2O and CO2 ice. The Martian atmosphere
is composed of >95% CO2 with a pressure of only a few
millibars. This fact led to the prediction that CO2 would
accumulate at the poles during winter (Leighton and Murray
1966). This prediction was confirmed by orbital tempera-
ture measurements (e.g. Kieffer 1979), and global mapping
has shown that seasonal CO2 caps grow well into the mid-
latitudes during winter, with perennial ice caps surviving
FIGURE 1 The north polar cap of Mars as seen by Viking. This mosaic of images was
acquired during northern summer when the ice had retreated to its
1 Department of Geological Sciences perennial size. The relatively bright material is H2O ice. The cap has shrunk to essentially
Arizona State University the same location every year that it has been imaged by spacecraft (1971 to present)
Box 871404, Tempe, AZ 85287-1404, USA (James and Cantor 2001). Image width is ~900 km. IMAGE CREDIT NASA/JPL

In their pioneering work, Leighton and Murray (1966)
predicted that CO2 would condense at the poles to sufficient
depth for CO2 to remain throughout the following summer.
Spacecraft temperature observations show that CO2 ice does
survive the summer in the south, but is completely
removed from the northern perennial cap, exposing H2O
ice (Kieffer 1979). Sublimation of this ice releases water
vapor into the atmosphere, which was initially detected by
ground-based observations over 40 years ago (Jakosky and
Barker 1984).
The Viking orbiter Mars Atmospheric Water Detector
(MAWD) instrument provided the first global map of water
vapor and confirmed that large quantities of vapor [~100
precipitable microns (pr µm)] were coming from the northern
FIGURE 2 “Swiss cheese” terrain with quasi-circular depressions,
perennial cap (e.g. Jakosky and Farmer 1982). The MGS
located near 86.9°S, 352.4°E. These pits are several hun-
Thermal Emission Spectrometer (TES) instrument has dred meters across and have a remarkably uniform depth of ~8 m wher-
provided detailed global maps of water vapor over three ever they are observed (Thomas et al. 2000; Malin et al. 2001). Some
Martian years (1997–2004), confirming high water vapor pits have been observed to increase in size over one Martian year (Malin
et al. 2001). A possible explanation is that the upper ~8 m thick layer is
abundances equatorward of the cap in the north, which rise
CO2 ice, which is gradually disappearing and exposing a stable layer of
rapidly to ~100 pr µm in late spring once the CO2 ice has H2O ice beneath it (Byrne and Ingersoll 2003). Image width is about
disappeared (Smith 2004). 3 km. MOC IMAGE R1303615; MGS MOC RELEASE NO. MOC2-695
MAWD data showed no indication of water vapor coming The topographic slopes and gently undulating surface of
from the perennial south polar cap (e.g. Jakosky and Farmer the northern layered materials are consistent with the slow
1982), consistent with measured temperatures that corre- radial flow velocities for H2O, but not CO2, ice rheology.
spond to CO2 ice (Kieffer et al. 2000). A notable exception Assuming that the layered deposits are made of essentially
to this pattern was the ground-based water vapor measure- pure H2O ice, the upper limit for the quantity of water in
ments in 1969 that showed a significant increase in water the northern layered terrains is ~1.6 × 1018 kg (~1.6 × 106
vapor as compared with other seasons or other years. This km3), which corresponds to an equivalent global layer of
has been interpreted to indicate that H2O ice was exposed water ~12 m deep. The areal extent of the southern
that year in the south (e.g. Jakosky and Barker 1984). TES deposits, which have a similar average thickness, is roughly
observations have confirmed the release of water vapor twice that of the northern deposits, suggesting a water
(~45 pr µm) along the edge of the southern perennial cap, inventory (again assuming pure H2O ice) that is roughly
providing conclusive evidence that H2O ice is now being twice that in the north.
exposed on the southern cap (Smith 2004). The presence of
this exposed ice has been confirmed by direct temperature Units within these deposits can be traced for hundreds of
measurements using the Mars Odyssey Thermal Emission kilometers at both poles, suggesting a regional process of
Imaging System (THEMIS) infrared imager (Titus et al. formation (Byrne and Murray 2002; Milkovich and Head
2003) and by near-IR spectral measurements from the Mars 2005). A stratigraphic horizon near the base of the northern
Express OMEGA spectrometer (Bibring et al. 2004). units is interpreted to have been an extensive sand sea that
formed during a period when no icy cap was present (Byrne
A remarkable result from the high-resolution MGS Mars and Murray 2002). The lack of an ice cap would require a
Orbiter Camera (MOC) was the discovery of quasi-circular dramatic climate change and would represent a major event
depressions in the perennial south polar cap that are up to in Martian history. Analysis of layers within the upper units
1 km in diameter and uniformly ~8 m deep (Thomas et al. of the northern layered terrain shows the existence of
2000; Malin et al. 2001) (FIG. 2). Some depressions are ~30 m periodicity, possibly associated with the 50,000 year
expanding at rates of 1–3 m per year (Malin et al. 2001). obliquity cycle (Milkovich and Head 2005); a 100 m unit
They have been modeled as a layer of CO2 ice over a sub- within this sequence lacks this layering and may represent
strate of either H2O ice or high-albedo (dust-free) CO2 ice a recent (0.5–2 Ma) period of ice removal and the formation
(Byrne and Ingersoll 2003). This thin layer of CO2 ice may of a sediment-rich lag. Crater counts also suggest an active
be relatively young and, even if completely sublimated, process, with ages for the upper surfaces of these deposits of
would be a minor contributor to the atmospheric CO2 ~30–100 Ma for the southern and <0.1 Ma for the northern
inventory. In this case, the atmospheric CO2 partial pres- deposits (e.g. Herkenhoff and Plaut 2000). These ages likely
sure, and therefore the atmospheric temperature, would not reflect only the most recent cycle in this process, and cyclic
be much higher than its current value. deposition and erosion may have been occurring in the
polar regions throughout Martian history.
POLAR LAYERED DEPOSITS
Thick stacks of sedimentary deposits extend up to 600 km SUBSURFACE ICE
outward from the poles in both hemispheres. These units
are ~3 km thick at both poles and are layered down to the
Ice Stability Models
resolution of the MOC camera (Malin and Edgett 2001). The stability of subsurface ice depends strongly on the
These layers may have been produced by differences in the porosity, tortuosity, and thermal conductivity of the sur-
amount of airfall dust incorporated into the ice, perhaps as face (Mellon and Jakosky 1995; Mellon et al. 2004). Ice sta-
a result of orbit-driven cyclic changes in climate (e.g. Pollack bility models predict that H2O ice will be stable at all lati-
and Toon 1982; Milkovich and Head 2005). tudes for obliquities >32° but will diffuse outward from the
upper 1–2 m in the equatorial and mid-latitude regions
The volume of the north polar layered deposits is estimated when the obliquity decreases. Mars is currently in an “inter-
to be ~1.2–1.6 × 106 km3 (Zuber et al. 1998). Attempts have glacial” period, with an obliquity of ~25° (Mustard et al.
been made to determine the density, and thus the ice to 2001; Christensen 2003), and near-surface ice is predicted
sediment ratio, of the layered materials using gravity and to be stable only poleward of ~50° (Mellon and Jakosky
topography, but this value has been difficult to constrain. 1995; Mellon et al. 2004).

This prediction is in excellent agreement with the mapping FIGURE 3 Distribution of subsurface H2O ice. This global map was
made using data from the neutron spectrometer that is
of H2O ice abundances by the Mars Odyssey Neutron Spec- part of the Odyssey Gamma Ray Spectrometer suite. The abundance of
trometer, High Energy Neutron Detector, and Gamma Ray hydrogen in the form of H2O ice is >30% poleward of ~60° in both
instruments (Boynton et al. 2002; Mitrofanov et al. 2002; hemispheres. Equatorial water-equivalent hydrogen abundances vary
Feldman et al. 2002). These instruments discovered high from 2% by mass (blue) to 18% (red); polar abundances vary from 0%
(blue) to 100% (red). DATA ADAPTED FROM FELDMAN ET AL. (2002)
hydrogen abundances in the uppermost meter extending
from the pole to 50° in both hemispheres (e.g. Feldman et
al. 2002) (FIG. 3). Assuming that the hydrogen is in H2O ice, snow or frost. In this case, this unit may contain substan-
the ice content of the upper meter is >70% by volume over tially more water than the 1.5–6.0 × 104 km3 initially sug-
large regions in the high latitudes. This great abundance is gested (Mustard et al. 2001). The number of small, fresh
unlikely to have resulted from gas diffusion into soil pores; craters on these mantling units is low, suggesting that these
instead, it more likely represents accumulation as surface mantles are possibly as young as 0.15 Ma but most certainly
snow or frost. less than 10 Ma (Mustard et al. 2001).
Further evidence for ice-rich mantles is found in 1–10 m
Morphological Evidence for Subsurface Ice
thick deposits that preferentially occur on pole-facing
The presence of ice-rich materials in the mid-latitudes has slopes, have features suggestive of flow, and have a distinct,
long been postulated on the basis of (1) lobate, grooved, rounded edge marking the upslope boundary (FIG. 5) (Carr
and ridged textures suggestive of flow on channel, crater, 2001; Christensen 2003). These characteristics suggest ice-
and mesa walls (FIG. 4); (2) unusual lobate crater ejecta possibly rich mantles that were once more extensive but have been
formed by fluidization of ground ice; (3) evidence for vol- removed from all but the cold, pole-facing slopes where
cano–ice interactions; and (4) possible evidence for glacial near-surface ice is stable under solar illumination. Ice-rich
landforms and processes (see review by Clifford et al. 2000). materials have also been suggested in glaciers in numerous
Additional evidence for ground ice comes from a pervasive areas including Hellas and the western flanks of the Tharsis
“basketball” surface texture found between 30° and 50° in volcanoes (e.g. Head et al. 2005).
both hemispheres (Mustard et al. 2001). This unit is 1–10 m
thick and is interpreted to result from the desiccation and MODERN GULLIES
erosion of once ice-rich soils that formed through diffusion Recent gullies are found in the 30–50° latitude range in
of water vapor into soil pore spaces (Mustard et al. 2001). both hemispheres (e.g. Malin and Edgett 2001), and their
This material does not have a hydrogen signature, in agree- origin is the topic of vigorous, ongoing discussion. It has
ment with the predicted desiccation of the upper 1–2 m at been proposed that they form from a range of processes, but
these latitudes (e.g. Mellon and Jakosky 1995; Mellon et al. the most plausible hypotheses are the discharge of liquid
2004). However, the mantle changes poleward to a smooth, water from subsurface aquifers, the melting of pore ice that
unpitted surface with a high hydrogen abundance, suggest- diffused inward from the atmosphere during periods of
ing that it may also have formed by direct condensation of colder temperatures, and the melting of a snow layer
ELEMENTS J UNE 2006

153
A B
FIGURE 4A Ice-rich terrains in the northern mid-latitude region. Ice-rich FIGURE 4B Lobes of ice-rich material flowing off mesas in the north-
soils have flowed down the wall of this valley centered ern hemisphere. THEMIS VIS images colorized using
near 37.6°N, 15.8°E. This image is a mosaic of THEMIS VIS images that nighttime temperatures from the THEMIS IR camera. Lobes of material
has been colorized using the daytime temperatures determined by the are colder at night, and therefore they are finer-grained and less rocky
THEMIS IR camera. Bands of bright material can be traced more than 10 than the substrate over which they are flowing. These distinct differ-
km downslope. They show the flow reaching the bottom of the local ences in surface properties between the lobes and the substrate provide
slope and turning northeast to continue to flow down the valley. Tem- strong evidence that this process is youthful and possibly active,
peratures range from –40°C to –34°C, with colder temperatures (blue because there has not been sufficient time for the homogenization of
tones) associated with darker or rockier surfaces and warmer tempera- the properties of these different surfaces. Mosaic of THEMIS images cen-
tures (reddish tones) related to brighter or dustier surfaces. IMAGE CREDIT tered near 43°N, 27.5°E. IMAGE CREDIT NASA/JPL/ARIZONA STATE UNIVERSITY
NASA/JPL/ARIZONA STATE UNIVERSITY
deposited during periods of higher obliquity when surface

ice was stable at these latitudes (see review by Heldmann
and Mellon 2004).
Of these models, the melting of pore ice does not account
for the fact that as the surface and subsurface temperatures
warm, the upper soil layer will become desiccated before
significant liquid water can be produced (Mellon and
Phillips 2001; Christensen 2003); if ice forms by vapor dif-
fusion, it will dissipate by the same mechanism. Water
released from subsurface aquifers (Malin and Edgett 2001)
explains gully morphology, latitudinal distribution, and
slope position (Heldmann and Mellon 2004). However, this
model does not account for the presence of gullies on iso-
lated knobs and dunes where there is no obvious aquifer
source—the survival and recharge mechanism that would
allow these aquifers to persist to the present—nor their
formation only at latitudes poleward of 30°.
Lee et al. (2001) and Hartmann (2002) suggested that melt-
ing snow might carve Martian gullies, based on analogies
with similar gully morphologies in cold regions on Earth.
This model was developed (Christensen 2003) by noting
the association of gullies with ice-rich, pole-facing slope
mantles (FIG. 5), and by incorporating models for snow
formation at high obliquity (e.g. Jakosky et al. 1995) and a
model of melting within dusty Martian snow (Clow 1987).
In this snowmelt model, water is transported from the poles
to mid-latitudes during periods of high obliquity. Melting Ice-rich mantles and associated gullies on poleward-facing
FIGURE 5
of this snow layer occurs at low obliquity as mid-latitude slopes in the southern hemisphere. This collage shows
temperatures increase, producing liquid water that is stable MOC images from the northwest wall of Dao Valles, between 33° and
35°S. These images show (1) well-developed flow features with com-
beneath the insulating layer of snow. Gullies form within
pressive ridges, which are strongly suggestive of ice (left panel); (2)
and beneath the snow as meltwater seeps into the loose slope mantles of ice-rich material and gullies that are present only where the
materials and destabilizes them. Patches of snow remain mantles are lacking (center panel), and (3) depressions with associated
today on pole-facing slopes, where they are protected gullies, some of which still have mantles, whereas others are free of
mantling material (right panel). These landforms could be explained by
against sublimation by a layer of desiccated dust/sediment.
the melting of a snow mantle, forming gullies that are visible only in
The primary argument against snowmelt is the presence of those locations where the snow has completely disappeared. MOC
gullies on slopes of all azimuths (Heldmann and Mellon 2004). IMAGES LEFT TO RIGHT: M03-04950, M09-02885, AND M0-3-6266

SUMMARY AND OUTSTANDING QUESTIONS 1. What is the age and history of the polar layered
deposits?
The poles and mid-latitudes of Mars contain a large reser-
voir of H2O ice, including ~5 × 106 km3 in polar layered 2. Have the polar ice caps ever been completely removed,
materials, >6 × 104 km3 in mid-latitude mantles and ice-rich and what produced the significant climate change that
sediments, and ~3 × 10-2 km3 in the seasonal ice caps and this would imply?
atmosphere. This known reservoir, if melted, would form a
3. What is the total inventory of subsurface ice?
layer of water ~35 m deep over the entire planet. Portions
of this reservoir appear to move to lower latitudes on 4. What are the source(s) of the water responsible
105–106 year timescales. Aquifers or melting snow or both for forming the modern gullies?
have produced liquid water at the surface in the very recent
past, and these areas hold exciting promise for future explo- ACKNOWLEDGMENTS
ration for past or present life. Many questions remain, and
The author would like to thank Hugh Kieffer, Tim Titus,
among the most intriguing are the following:
Mike Smith, and all those involved with the TES and
THEMIS investigations for many stimulating discussions on
polar processes, and Hap McSween, Vic Baker, and Pierrette
Tremblay for helpful reviews that significantly improved
the manuscript. .
REFERENCES measurements of Martian water Mellon MT, Phillips RJ (2001) Recent

vapor: Variability of the seasonal gullies on Mars and the source of
Bibring J-P and 13 co-authors (2004) cycle. Icarus 57: 322-334 liquid water. Journal of Geophysical
Perennial water ice identified in the Research 106(E10): 23165-23180
south polar cap of Mars. Nature 428: Jakosky BM, Farmer CB (1982) The
627-630 seasonal and global behavior of water Mellon MT, Feldman WC, Prettyman TH
vapor in the Mars atmosphere: (2004) The presence and stability of
Boynton WV and 24 coauthors (2002) Complete global results of the Viking ground ice in the southern hemisphere
Distribution of hydrogen in the near atmospheric water detector of Mars. Icarus 169: 324-340
surface of Mars: Evidence for subsurface experiment. Journal of Geophysical
ice deposits. Science 297: 81-85 Research 87: 2999-3019 Milkovich SM, Head JW III (2005) North
polar cap of Mars: Polar layered
Byrne S, Ingersoll AP (2003) A sublimation Jakosky BM, Henderson BG, Mellon MT deposit characterization and
model for martian south polar ice (1995) Chaotic obliquity and the identification of a fundamental
features. Science 299: 1051-1053 nature of the Martian climate. climate signal. Journal of Geophysical
Byrne S, Murray BC (2002) North polar Journal of Geophysical Research Research 110: E01005,
stratigraphy and the paleo-erg of 100(E1): 1579-1584 doi:10.1029/2004JE002349
Mars. Journal of Geophysical James PB, Cantor BA (2001) Martian north Mitrofanov I and 11 coauthors (2002)
Research 107(E6): 5044, DOI polar cap recession: 2000 Mars Maps of subsurface hydrogen from
10.1029/2001JE001615 Orbiter Camera observations. Icarus the High Energy Neutron Detector,
Carr MH (2001) Mars Global Surveyor 154: 131-144 Mars Odyssey. Science 297: 78-81
observations of fretted terrain. Kieffer HH (1979) Mars south polar spring Mustard JF, Cooper CD, Rifkin MK (2001)
Journal of Geophysical Research and summer temperatures: A residual Evidence for recent climate change
106(E10): 23571-23594 CO2 frost. Journal of Geophysical on Mars from the identification of
Christensen PR (2003) Formation of recent Research 84: 8263-8289 youthful near-surface ground ice.
martian gullies through melting of Nature 412: 411-414
Kieffer HH, Titus TN, Mullins KF,
extensive water-rich snow deposits. Christensen PR (2000) Mars south Pollack JB, Toon OB (1982) Quasi-periodic
Nature 422: 45-48 polar cap spring and summer climate changes on Mars: A review.
Clifford SM and 52 co-authors (2000) The behavior observed by TES: Seasonal Icarus 50: 259-287
state and future of Mars polar science cap evolution controlled by frost
grain size. Journal of Geophysical Smith DE, Zuber MT, Neumann GA (2001)
and exploration. Icarus 144: 210-242 Seasonal variations of snow depth on
Research 105(E4): 9653-9700
Clow GD (1987) Generation of liquid water Mars. Science 294: 2141-2146
on Mars through the melting of a Lee P, Cockell CS, Marinova MM, McKay
CP, Rice JW (2001) Snow and ice Smith MD (2004) Interannual variability in
dusty snowpack. Icarus 72: 95-127 TES atmospheric observations of Mars
melt flow features in Devon Island,
Feldman WC and 12 co-authors (2002) Nunavut, Arctic Canada as possible during 1999–2003. Icarus 167: 148-165
Global Distribution of neutrons from analogs for recent slope flow features Thomas PC, Malin MC, Edgett KS, Carr
Mars: Results from Mars Odyssey. on Mars. Proceedings of Lunar and MH, Hartmann WK, Ingersoll AP,
Science 297: 75-78 Planetary Sciences 1809 (abstract) James PB, Soderblom LA, Veverka J,
[CD-ROM] Sullivan R (2000) North-south
Hartmann WK (2002) Comparison of
icelandic and martian hillside gullies. Leighton RB, Murray BC (1966) Behavior geological differences between the
Proceedings of Lunar and Planetary of carbon dioxide and other volatiles residual polar caps on Mars. Nature
Sciences 33: Abstract 1904 on Mars. Science 153: 136-144 404: 161-164
Head JW and 13 coauthors (2005) Tropical Malin MC, Caplinger MA, Davis SD (2001) Titus TN, Kieffer HH, Christensen PR
to mid-latitude snow and ice Observational evidence for an active (2003) Exposed water ice discovered
accumulation, flow and glaciation on surface reservoir of solid carbon near the south pole of Mars. Science
Mars. Nature 434: 346-351 dioxide on Mars. Science 294: 299: 1048-1051
2146-2148 Zuber MT and 20 co-authors (1998)
Heldmann JL, Mellon MT (2004)
Observations of martian gullies and Malin MC, Edgett KS (2001) Mars Global Observations of the north polar
constraints on potential formation Surveyor Mars Orbiter Camera: region of Mars from the Mars Orbiter
Laser Altimeter. Science 282:
2053-2060 .
mechanisms. Icarus 168: 285-304 Interplanetary cruise through primary
mission. Journal of Geophysical
Herkenhoff KE, Plaut JJ (2000) Surface ages
Research 106(E10): 23429-23570
and resurfacing rates of the polar
layered deposits on Mars. Icarus Mellon MT, Jakosky BM (1995) The
144: 243-253 distribution and behavior of Martian
ground ice during past and present
Jakosky BM, Barker ES (1984) Comparison
epochs. Journal of Geophysical
of ground-based and Viking Orbiter
Research 100 (E6): 11781-11799

Aqueous Processes Recorded
by Martian Meteorites:
Analyzing Martian Water
on Earth
Laurie A. Leshin1 and Edward Vicenzi2
(1998) provide an overview of the
M
artian meteorites have delivered to Earth water molecules and
mineralogy and petrology of these
minerals produced by aqueous processes on Mars. The study of these
samples and review in detail the
samples, using powerful analytical instrumentation, provides a basis evidence for a Martian origin. An
for understanding aqueous activity on Mars. Although most analyses are at up-to-date catalog of descriptions
the scale of microns, the conclusions reached have important implications for of the specimens can be found at
large-scale aqueous processes. Secondary minerals, such as clays, salts, and the Mars Meteorite Compendium
website www-curator.jsc.nasa.gov/
carbonates, are present at some level in all Martian meteorite subtypes and
antmet/mmc/index.cfm
are especially important in the nakhlites and ALH84001. Light element stable
isotope analyses show evidence for mixing between atmosphere and magmatic
fluids into a crustal reservoir, and that this crustal water was not in equilib- EVIDENCE FOR
rium with the host rocks. The mineralogical and isotopic data present a fairly A MARTIAN ORIGIN
consistent picture of the aqueous history of Mars: low levels of aqueous Compelling evidence has been found that
some meteorites originated on Mars:
alteration are generally present but extensive aqueous processes are probably
limited in space and time. ♦ All specimens (except ALH84001) are
KEYWORDS: Mars, meteorites, stable isotopes, igneous rocks with geologically young crys-
tallization ages, suggesting they formed on
carbonates, water–rock interaction
a planet-sized object capable of sustaining
OVERVIEW OF MARTIAN METEORITES geological activity throughout the history
Although it would be of the solar system.
The Martian meteorites (also called SNC, for shergottite– extremely desirable to
nakhlite–chassignite) comprise a group of 36 specimens. study Martian sedimen- ♦ Despite variable bulk mineralogy and ages,
Evidence for the Martian origin of these rocks is discussed tary rocks in laborato- the samples are linked together and to a
in the sidebar. The meteorites are igneous rocks, consistent ries on Earth, such rocks common parent object through common
with the abundant evidence for past volcanic activity on are unlikely to be in geochemical signatures, such as bulk oxygen
Mars. However, the igneous nature of the samples also our hands in the near isotope composition, K/La ratios, and
likely reflects the mechanism by which they were ejected future, as missions to Fe/Mn ratios.
from Mars, during a small number (probably 3–5) of impact return samples from ♦ A direct tie to Mars is provided by the dis-
events over the past ~15 million years. Friable samples, such Mars are expensive and covery, in several samples, of gases such as
as poorly cemented sedimentary rocks, would probably not complex. Therefore it N2, Ar, and Xe trapped in impact-produced
survive this violent process. is imperative to extract glasses. The isotopic compositions and rel-
The meteorites can be separated into subclasses on the basis all possible informa- ative abundances of the gases closely
of their mineralogy or rock type. The four major subclasses tion about water on match the composition of the current Mart-
are the shergottites, the nakhlites, the chassignites, and the Mars from the samples ian atmosphere.
unique sample ALH84001. The shergottites are the most we have in hand.
abundant group and are further subdivided into basaltic, Because of the igneous
lherzolitic, and olivine-phyric types. The nakhlites are nature of the samples, in most cases this information is con-
cumulate clinopyroxenites with up to 10% olivine. Chassig- tained within very rare and/or complex minerals, and thus
nites are dunites (olivine cumulates), and ALH84001 is a the desire to extract knowledge has driven the development
cumulate orthopyroxenite. The meteorites show a range in and application of new, cutting-edge analytical techniques.
age, from 4.5 Ga for ancient ALH84001 to 175–575 Ma for In the following sections, we highlight key mineralogical
the shergottites; the nakhlites and chassignites formed at and isotopic observations from these meteorites that provide
about 1.3 Ga (e.g. Nyquist et al. 2001). Photomicrographs of significant insights into aqueous processes on Mars.
two specimens are shown in FIG. 1. McSween and Treiman
SECONDARY MINERALIZATION IN MARTIAN
1
METEORITES: THE NAKHLITE EXAMPLE
Sciences and Exploration Directorate
NASA Goddard Space Flight Center Unlike rare primary hydrous minerals, such as apatite and
Greenbelt, MD 20771, USA amphibole, crystallized at high temperatures in Martian
E-mail: [email protected] magmas, most hydrous minerals in the SNC meteorites
2 Department of Mineral Sciences formed at low temperatures through secondary processes. If
National Museum of National History the alteration is preterrestrial in origin (i.e. it formed prior
Smithsonian Institution to the meteorite’s entry into Earth’s atmosphere and there-
Washington, DC 20560, USA
E-mail: [email protected] fore presumably on Mars), then by examining secondary

hydrous minerals, we may be able to constrain the volatile FIGURE 1 Photomicrographs of Martian meteorites Zagami (basaltic
reservoirs and processes in the near surface and potentially shergottite) and ALH84001 in cross-polarized light. Thin
sections are approximately 2 cm across. IMAGES FROM A COLOR ATLAS OF
in the ancient Martian atmosphere (Gooding et al. 1991). METEORITES IN THIN SECTION, COURTESY OF DANTE LAURETTA
All Martian meteorite subgroups show some mineralogical
evidence of aqueous alteration. Shergottites and chassignites consistent with the fact that olivine breaks down rapidly in
show the least, with only trace sulfate and carbonate salts, most terrestrial weathering environments compared to
and ALH84001 contains ~1% carbonate. However, some of other magmatic minerals such as pyroxene. Veinlets in
the best evidence for aqueous alteration in Martian mete- olivine are typically several to tens of micrometers across
orites is found in the nakhlites. The near absence of shock and hundreds of micrometers in length. They are most
metamorphism and the relative abundance of alteration often filled with “iddingsite,” a term used to describe an
products make the nakhlites ideal for the study of delicate assemblage of minerals representing the hydrous breakdown
low-temperature minerals that would decompose if subjected products of olivine (FIG. 2A). Iddingsite is reddish brown in
to elevated pressures and temperatures. color, contains Fe3+, is variable in composition, and was
referred to as “iron-rich silicate rust” by Gooding et al.
The alteration materials in the nakhlites constitute only a (1991). The number and types of phases that compose
minor fraction (<1%) of the rock. In conducting studies of iddingsite are not uniform among the nakhlites; however,
meteorite alteration, it is imperative to first determine clay is common to all. Fine-scale mineralogical heterogeneity
whether these minor phases formed on Mars or on Earth. is observed within the veinlets, and transmission electron
Several criteria may be applied to establish the origin of the microscopy is required to properly identify sheet silicates
alteration minerals: such as smectite (Gooding et al. 1991; Treiman 1993). Other
1. Petrography – Are alteration veinlets cut by the phases found in association with smectite are ferrihydrite,
fusion crust, indicating formation before atmospheric maghemite, goethite, and two varieties of amorphous silicates,
heating (Gooding et al. 1991; Treiman 1993)? one rich in silicon and the other in iron (Treiman 1993;
Vicenzi and Heaney 1999; Bridges and Grady 2000). Trace
2. Age – Does the age of the secondary minerals exceed quantities of carbonates, often chemically heterogenous on
the time the meteorite has resided on Earth (Swindle a scale of tens of micrometers, can also be found in veins
et al. 2000)? formed by silicate alteration of olivine (FIG. 3A, B). The
3. Isotopic fingerprinting – Are the stable isotope com- preservation of fine-scale carbonate heterogeneity is inter-
positions of the alteration minerals inconsistent with preted to be an indicator of a low water/rock ratio (Bridges
terrestrial values (Karlsson et al. 1992; Leshin et al. and Grady 2000; Vicenzi and Heaney 2000).
1996)? Mesostasis grain boundaries: In transmitted light, all
If the answer is Yes to these questions, then one can conclude nakhlites appear rich in alteration products within mesosta-
that the alteration formed on Mars, as is the case for the sis—pockets in the rock where the last melt fraction solidified
secondary minerals discussed below. (FIG. 2B). Yet when the crystals in mesostasis are imaged
using high-resolution methods, scant evidence for the
Alteration Types and Mineral Assemblages secondary materials is found, suggesting the alteration is
Olivine-hosted veinlets: Although Fe-rich olivine composes present as a thin surface coating.
only ~10 vol% of most nakhlites, it contains a dispropor- Salt-bearing veinlets and mesostasis: Identification of gypsum
tionate quantity of secondary minerals. This observation is and halite within olivine veinlets constrains the salinity of

A A B
C D
E F
FIGURE 2 Transmitted-light images of low-temperature alteration

products in Martian meteorites. (A) Olivine-hosted veinlet
filled with “iddingsite” (Lafayette meteorite). (B) High-contrast image
showing secondary minerals concentrated in mesostasis (MIL03346
meteorite). FIGURE 3 Electron, X-ray images, and phase maps of Martian mete-
orite alteration. (A) Back-scattered electron (BSE) image
of an olivine-hosted veinlet containing both iddingsite and siderite
alteration brines (Bridges and Grady 2000; Rost et al. 2005). (Lafayette meteorite). (B) False-color phase map of A depicting the com-
Given the high solubility of halite, the material may repre- positional heterogeneity of the carbonate. Blue: silicate–carbonate inter-
growth; yellow: carbonate enriched in Ca; orange: carbonate enriched
sent remobilization of NaCl from an earlier vein-filling
in Fe; red: carbonate enriched in Mn. (C) BSE image of a
event (FIG. 3C, D). Anhydrite is found exclusively as larger pyroxene–olivine grain boundary with an olivine-hosted veinlet filled
grains, sometimes in association with halite within with a halite and clay assemblage (Nakhla meteorite). (D) False-color
mesostasis (FIG. 3E, F). X-ray image overlay of C: Na (blue) and Cl (green), highlighting NaCl
mineralization. (E) BSE image of massive CaSO4 (anhydrite) bounded by
plagioclase, augite (px-1), and low-Ca pyroxene (px-2) in the mesosta-
Temperature and Timing of Alteration Fluids sis of the Nakhla meteorite. (F) False-color X-ray image overlay of E
The limited stability of the alteration phases with respect to showing clays and magmatic phosphate adjacent to anhydrite: S (red),
temperature (e.g. ferrihydrite) suggests that aqueous alter- Ca (green), K (blue), and P (cyan).
ation took place at <100°C (Treiman 1993). Application of
chemistry and mineralogy (and isotopes, see below) indicate
oxygen isotope thermometry, based upon coprecipitated
that the water was likely transient and volumetrically limited.
Fe-carbonate and smectite, yields a temperature range of
–34 to 19°C, consistent with the mineralogical constraint
(Romanek et al. 1998). The high salinity of some aqueous ISOTOPIC EVIDENCE OF AQUEOUS PROCESSES
fluids would lower the freezing point, yet textural relation- The use of the stable isotopes of hydrogen, carbon, and oxygen
ships between carbonate and clay suggest they are associated to determine the nature of aqueous processes has become a
with temporally different pulses of alteration fluid (Vicenzi cornerstone of terrestrial geochemistry. Similarly the appli-
and Heaney 2000). cation of these isotopic techniques to Martian materials has
provided critical information about Martian aqueous
Efforts to determine precisely the age of the secondary
processes. However, in the case of Martian materials, there
minerals, and hence the timing of fluid flow in the Martian
are important challenges. First, the numbers of different
crust, are hampered by the chemical complexity of the
samples and rock types are extremely small—imagine trying
minerals. Despite this complication, age estimates for the
to determine the history of Earth’s volatiles from a few
precipitation of iddingsite determined by K–Ar dating are
dozen igneous rocks! Second, the samples themselves are
~600–700 Ma (Swindle et al. 2000). Hence, low-temperature
typically small and, as discussed above, contain only small
liquid water was available relatively recently for chemical
amounts of alteration minerals. Third, the isotopic “refer-
alteration of near-surface igneous rocks, although the
ence frame” is not always clear. On Earth the widespread
ELEMENTS J UNE 2006
159
and reasonably well-understood reservoir of ocean water
and ocean-produced minerals, as well as the global plate
tectonic cycle, are strong drivers for many isotopic effects.
On Mars, processes recorded in the meteorites are not dom-
inated by processes related to an ocean, and in many cases
the observed isotope variations are found to be large and
complex. Of course, large variations in isotopic ratios make
the signal easier to detect, and give insights into processes
that may not be common on Earth. Below we discuss key
conclusions from three isotopic systems: hydrogen, carbon,
and oxygen.
Hydrogen Isotopes
Ground-based infrared telescopic observations have
revealed that the deuterium-to-hydrogen ratio (D/H) of
water in the Martian atmosphere is ~5 times terrestrial values,
significantly higher than the D/H of any hydrogen on Earth
(Owen et al. 1988; Krasnopolsky et al. 1997). The high D/H
FIGURE 4 δD values and water contents of apatite grains from Mar-
value is a result of preferential loss of H relative to heavier
tian meteorite QUE94201. The data are interpreted to
D from the Martian atmosphere throughout the planet’s represent a mixture of two end members, and most plausibly represent
history. The initial Martian D/H ratio, the timing of water addition (or exchange) of water with an atmospheric D/H signature (δD
loss, and the evolution of the deuterium enrichment ~+4000‰) to minerals that initially uniformly contained water with
remain matters of debate. Nonetheless, the heavy isotope δD of ~+900 ± 250‰. The curve shows the mixing model from which
the initial D/H of the minerals was calculated. FROM LESHIN (2000)
enrichment in the atmosphere can be used as a tracer of the
interaction of the atmospheric volatile reservoir with other minerals analyzed in the QUE94201 shergottite, Leshin
water in the crustal or interior reservoirs. (2000) proposed that magmatic water on Mars has a δD
Martian meteorites record elevated D/H signatures that provide value of about +900‰, approximately twice the terrestrial
direct evidence of surface–atmosphere interaction on Mars value (FIG. 4). This value has been used to argue for an
and offer important windows into Martian hydrologic accretionary history for Mars dominated by small bodies,
cycles. Many Martian meteorites contain D-enriched water, and that the mechanism for water acquisition was different
both in bulk samples (Leshin et al. 1996; Eiler et al. 2002a) from Earth’s (Lunine et al. 2003). An alternative view is sug-
and in individual hydrous minerals (e.g. Watson et al. 1994; gested by the SIMS work of Gillet et al. (2002) and Boctor et
Leshin 2000; Boctor et al. 2003). For example, shergottites, al. (2003). Although Boctor et al. (2003) observed D-rich
nakhlites, and ALH84001 all show D-enriched water in bulk water in all phases analyzed (many of which are nominally
extraction (when bulk samples are heated under vacuum) anhydrous minerals), they also found low δD values, con-
(Leshin et al. 1996), with δD values† approaching +2000‰. sistent with a more Earth-like composition for the low δD
In situ SIMS (secondary ion mass spectrometer) analyses of end member. Gillet et al. (2002) analyzed secondary alter-
individual minerals and glasses in meteorites from all sub- ation in nakhlite NWA817 by SIMS and found only low,
classes, including apatite, amphibole, biotite, feldspathic Earth-like values for δD. It is difficult to rule out terrestrial con-
glass, silica, carbonate, and mafic glasses, all show some tamination when analyzing minerals that are either nomi-
evidence of D-enrichment (Watson et al. 1994; Leshin 2000; nally anhydrous or highly susceptible to contamination
Sugiura and Hoshino 2000; Boctor et al. 2003). However, and recrystallization on Earth. To resolve these complexi-
these analyses also suggest mixing between two compo- ties, more studies that systematically examine hydrous
nents: a D-rich component derived from the atmosphere magmatic phases in Martian meteorites are needed.
(with values up to ~+4000‰) and a D-poor component, Secondary hydrous minerals in Martian meteorites are
probably representing a mantle or magmatic contribution probably the best indicators of the changes in composition
(FIG. 4). It is important to note that many of the analyses of crustal water on Mars through time, but little is known
that show atmospheric signatures were not performed on about their D/H ratio. The only data come from bulk sam-
secondary minerals but on primary magmatic phases (such ple studies by Leshin et al. (1996) and Eiler et al. (2002a),
as apatite and amphibole), suggesting an overprinting and from SIMS work by Gillet et al. (2002) and Sugiura and
caused by post-crystallization circulation of atmosphere- Hoshino (2000). Studying hydrogen in secondary minerals
derived crustal fluids. is especially difficult because these minerals are often
In addition to providing evidence of an atmosphere-derived highly susceptible to terrestrial contamination and H is an
component in igneous and alteration minerals, these studies abundant contaminant on Earth. As discussed above, the
provide constraints on the isotopic composition of the studies provide conflicting evidence about the absolute D/H
magmatic water reservoir on Mars, critical for comparing value of secondary minerals (ranging from Earth-like values
the accretionary histories of Mars and Earth. From careful to values up to ~3× terrestrial) and therefore result in conflict-
correlation of water content and δD of hydrous magmatic ing hypotheses for their origin. Terrestrial contamination is
probably the cause of the wide range of D/H values, and the
most conservative conclusion is that most meteorites show
† Isotope data are often expressed using the δ notation: evidence of elevated D/H values in secondary minerals and
that these values can be traced to interactions with atmos-
phere-derived D-rich groundwaters. This is consistent with the
results of studies of other isotopic pairs, as discussed below.
where R is the ratio of the heavy to the light isotope, x is the sample
and std is a standard. For the elements hydrogen, carbon, and oxygen, Carbon Isotopes
R is D/H, 13C/12C, and 18O/16O (or 17O/16O), respectively. A delta value
is reported in units of ‰ (“per mil”), or parts per thousand deviation As is the case for hydrogen, carbon isotopic signatures in
from the standard. For hydrogen and oxygen, the reference standard Martian meteorites record evidence of contributions from
is Standard Mean Ocean Water (SMOW), and for carbon the standard multiple reservoirs. There have been many reports on the
is a carbonate called Pee Dee Belemnite (PDB).
abundances and 13C/12C values of CO2 extracted from bulk

Martian meteorites and their probable implications for the
evolution of CO2 on Mars (e.g. Carr et al. 1985; Clayton and
Mayeda 1988; Wright et al. 1990; Romanek et al. 1994;
Leshin et al. 1996). Three main carbon-bearing components
are distinguished in Martian meteorites by their temperature
of release (or release by reaction with acid) and their δ13C
values. The first component, released at temperatures below
~500°C, is interpreted as low-temperature carbonaceous
material derived mostly from contamination by terrestrial
organic matter but possibly mixed with extraterrestrial
organics. This component has a δ13C value of –20 to –30‰.
The second component, released between 400 and 700°C in
heating experiments or by reaction of samples with acid,
originates from breakdown of carbonate. This component is
associated with 13C-enriched CO2 in several Martian mete-
orites and gives δ13C values up to ~+42‰. The third com-
ponent consists of carbon released at temperatures above
700°C, has δ13C values of –20 to –30‰, and has been inter-
preted as a “magmatic” component representative of the C
isotope composition of CO2 in the Martian interior. How- Oxygen isotopic compositions—specifically the value of
FIGURE 5
ever, Goreva et al. (2003) analyzed the C isotope composition ∆17O—of waters and secondary minerals from several
of individual magmatic phosphate grains in two shergot- Martian meteorites differ from those of the primary silicates. This indi-
cates that the waters and alteration minerals formed from a reservoir not
tites and found a δ13C value of –15 ± 5‰. In either case, the in equilibrium with the crustal rocks. The elevated ∆17O values are likely
“magmatic” value is lower than the terrestrial mantle C iso- a signature of atmosphere-derived volatiles.
tope composition of ~–5‰. However, the Goreva et al.
(2003) value is isotopically consistent with the bulk of that is displaced from the line for terrestrial rocks by 0.32‰
organic matter in carbonaceous meteorites, supporting the (the so-called ∆17O value‡; Franchi et al. 1999). The mete-
idea that such materials are the likely source of C on Mars. orites have bulk δ18O values between ~+4 and +5.5‰, con-
These results support the hypothesis that the Martian sistent with an igneous origin (e.g. Franchi et al. 1999) and
magmatic C isotope ratio accurately reflects the planet’s similar to the value for terrestrial mafic rocks. However, the
accretionary history rather than tectonic recycling. most important O isotope data with respect to aqueous
processes come from analyses of all three O isotopes (16O,
Near-surface aqueous processes are recorded by carbonates 17O, 18O) in waters, carbonates, and sulfates extracted from
in ALH84001, and these carbonates have been especially
bulk samples (e.g. Karlsson et al. 1992; Farquhar and
well studied due to the controversial hypothesis that they
Thiemens 2000). These data (FIG. 5) show O isotope anom-
preserve Martian fossils (McKay et al. 1996). The environ-
alies (as demonstrated by ∆17O values that differ from those
ment of formation of these secondary minerals has
of Martian meteorite igneous minerals) of more than 1‰,
remained enigmatic despite extensive chemical and miner-
and the anomalies have been used to argue for an atmos-
alogical characterization showing that the carbonates are
pheric chemistry conducive to the production of mass-
strongly zoned in chemical composition (from ankerite to
independent isotope effects (e.g. Thiemens 2002). Such
magnesite) and also despite several in situ O isotope studies
atmospheric volatiles are incorporated into meteoric fluids
that show strong microscale zoning, with δ18O from ~+3 to
and then transferred to secondary minerals and waters in
+25‰, which correlates with mineral chemistry (Leshin et
the meteorites. The ∆17O values also clearly show that the
al. 1996; Valley et al. 1997; Saxton et al. 1998; Eiler et al.
water and alteration minerals in the samples are extraterrestrial
2002b). Extensive in situ C isotope analyses of these car-
in origin, as terrestrial contamination would be expected to
bonates were recently reported by Niles et al. (2005), who
result in ∆17O of zero, equivalent to waters on Earth.
found that the δ13C values are also highly zoned, ranging
from ~+30‰ for Ca-rich carbonates to +60‰ for magne- δ18O values in secondary minerals are generally higher than
site. The isotopically heavy carbon, similar to the heavy car- values in igneous minerals. For example, carbonates in the
bon found in many bulk Martian meteorite samples (dis- nakhlites have δ18O values of ~20–30‰, consistent with
cussed above), is consistent with derivation of fluids from a formation in low-temperature aqueous environments
13C-enriched Martian atmosphere and suggests an origin (Romanek et al. 1998; Saxton et al. 1998). Taken together,
comparable to that of D-enriched waters in Martian mete- the oxygen isotopic data indicate that Martian secondary
orite samples. Hypotheses for carbonate formation that minerals formed at relatively low temperatures from atmos-
account for the large isotopic variation in C and O (see phere-derived fluids. The observation of ∆17O anomalies in
below) and that also consider the mineralogical constraints many different samples supports the idea that waters in the
include mixing of fluids derived from two different envi- Martian crust are not generally in isotopic equilibrium with
ronments (e.g. with different contributions from atmos- the crust. This suggests that intensive hydrothermal circula-
pheric components), a coupled evaporation/CO2 degassing tion of fluids, which would lead to equilibration of the
model, and formation from high-pH fluids. These ideas isotopes and loss of the anomalous ∆17O values, is not
agree particularly well with the environments that are common on Mars.
expected on Mars and suggest that large, abiotic isotopic
variations may be common. For all hypotheses, the elevated SUMMARY: WHAT MARTIAN METEORITES
13C values are thought to result from atmospheric contribu-
TELL US ABOUT WATER ON MARS
tions to the carbon in the fluid(s) that formed the carbonates.
The evidence from Martian meteorites supports the hypothesis
Oxygen Isotopes that the history of volatiles on Mars is complex in space and
time due to atmospheric evolution, the probable lack of
Oxygen isotope data for whole-rock Martian meteorites
long-lived oceans, and the absence of plate tectonics. Still,
demonstrate that the meteorites originated on a common
parent object. This is because, on a 17O/16O vs 18O/16O dia-
gram, the data plot on a line of slope approximately +1/2 ‡ ∆17O (‰) = δ17O – (0.52 × δ18O)

Martian meteorites of all ages show evidence of some inter- search for life on Mars should focus on localized habitable
action with Martian crustal water after crystallization. Thus environments and that these environments should be charac-
aqueous activity, to some degree, is common on Mars. terized intensively in order to interpret correctly any potential
However, none of the samples show evidence of extensive biosignatures in the context of the local geologic setting.
water–rock interaction, which would have resulted in more
abundant and well-formed secondary minerals and equili- ACKNOWLEDGMENTS
brated isotope signatures. Aqueous processes that operate
Helpful reviews by Hap McSween and Rod Ewing, and editing
locally and sporadically, rather than globally and continu-
by the Elements team greatly improved the manuscript. LL
ously, appear to be required. In addition, Martian surface
acknowledges support from the NASA Cosmochemistry
conditions may be conducive to production of large abiotic
Program. EV acknowledges support from the NASA Astrobi-
C isotope effects, which could confound attempts to use C
ology Institute and the assistance of Ms. Elli Pauli and
Dr. Detlef Rost. .
isotopes as a biomarker. These conclusions suggest that the
REFERENCES Krasnopolsky VA, Bjoraker GL, Mumma Saxton JM, Lyon IC, Turner G (1998)
MJ, Jennings DE (1997) High-resolution Correlated chemical and isotopic
Boctor NZ, Alexander CMO, Wang J, Hauri spectroscopy of Mars at 3.7 and 8 zoning in carbonates in the martian
E (2003) The sources of water in µm: A sensitive search of H2O2, meteorite ALH84001. Earth and
martian meteorites: clues from H2CO, HCl, and CH4, and detection Planetary Science Letters 160: 811-822
hydrogen isotopes. Geochimica et of HDO. Journal of Geophysical
Cosmochimica Acta 67: 3971-3989 Research 102 (E3): 6525-6534 Sugiura N, Hoshino H (2000) Hydrogen-
isotopic compositions in Allan Hills
Bridges JC, Grady MM (2000) Evaporite Leshin LA (2000) Insights into martian 84001 and the evolution of the
mineral assemblages in the nakhlite water reservoirs from analyses of martian atmosphere. Meteoritics
(martian) meteorites. Earth and martian meteorite QUE94201. & Planetary Science 35: 373-380
Planetary Science Letters 176: 267-279 Geophysical Research Letters
27: 2017-2020 Swindle TD, Treiman AH, Lindstrom DJ,
Carr RH, Grady MM, Wright IP, Pillinger Burkland MK, Cohen BA, Grier JA, Li
CT (1985) Martian atmospheric carbon Leshin LA, Epstein S, Stolper EM (1996) B, Olson EK (2000) Noble gases in
dioxide and weathering products in Hydrogen isotope geochemistry of iddingsite from the Lafayette
SNC meteorites. Nature 314: 248-250 SNC meteorites. Geochimica et Meteorite: Evidence for liquid water
Clayton RN, Mayeda TK (1988) Isotopic Cosmochimica Acta 60: 2635-2650 on Mars in the last few hundred
composition of carbonate in million years. Meteoritics & Planetary
Lunine JI, Chambers J, Morbidelli A, Science 35: 107-115
EETA79001 and its relation to parent Leshin LA (2003) The origin of water
body volatiles. Geochimica et on Mars. Icarus 165: 1-8 Thiemens MH (2002) Mass-independent
Cosmochimica Acta 52: 925-927 isotope effects and their use in
McKay DS, Gibson EK, Thomas-Keprta KL, understanding natural processes.
Eiler JM, Kitchen N, Leshin LA, Strausberg Vali H, Romanek CS, Clemett SJ,
M (2002a) Hosts of hydrogen in Allan Israel Journal of Chemistry 42: 43-54
Chiller XDF, Maechling CR, Zare RN
Hills 84001: Evidence for hydrous (1996) Search for past life on Mars: Treiman AH (1993) Preterrestrial aqueous
martian salts in the oldest martian Possible relic biogenic activity in alteration of the Lafayette (SNC)
meteorite? Meteoritics & Planetary martian meteorite ALH84001. Science meteorite. Meteoritics & Planetary
Science 37: 395-405 273: 924-930 Science 28: 86-97
Eiler JM, Valley JW, Graham CM, Fournelle McSween HY Jr, Treiman AH (1998) Martian Valley JW, Eiler JM, Graham CM, Gibson
J (2002b) Two populations of meteorites. In: Papike JJ (ed) EK, Romanek CS, Stolper EM (1997)
carbonate in ALH84001: geochemical Planetary Materials, Mineralogical Low-temperature carbonate concretions
evidence for discrimination and Society of America Reviews in in the martian meteorite ALH84001:
genesis. Geochimica et Cosmochim- Mineralogy 36: 6-1-6-53 Evidence from stable isotopes and
ica Acta 66: 1285-1303 mineralogy. Science 275: 1633-1638
Niles PB, Leshin LA, Guan Y (2005)
Farquhar J, Thiemens MH (2000) Oxygen Microscale carbon isotope variability Vicenzi EP, Heaney PJ (1999) Examining
cycle of the Martian atmosphere- in ALH84001 carbonates and a Martian alteration products using in
regolith system: ∆17O of secondary discussion of possible formation situ TEM sectioning; a novel
phases in Nakhla and Lafayette. environments. Geochimica et application of the focused ion beam
Journal of Geophysical Research Cosmochimica Acta 69: 2931-2944 (FIB) for the study of extraterrestrial
105(E5): 11991-11998 materials. Lunar Planetary Science
Nyquist LE, Bogard DD, Shih CY, Greshake XXX: #2005 (CD-ROM)
Franchi IA, Wright IP, Sexton AS, Pillinger A, Stöffler D, Eugster O (2001) Ages
CT (1999) The oxygen-isotopic compo- and geologic histories of Martian Vicenzi EP, Heaney PJ (2000) Multiple
sition of Earth and Mars. Meteoritics meteorites. Space Science Reviews martian fluids: the alteration
& Planetary Science 34: 657-661 96: 105-164 sequence in the Lafayette SNC
Gillet P, Barrat JA, Deloule E, Wadhwa M, meteorite. Meteoritics & Planetary
Owen T, Maillard JP, de Bergh C, Lutz B Science 35 Supplement: 164-165
Jambon A, Sautter V, Devouard B, (1988) Deuterium on Mars: The
Neuville D, Benzerara K, Lesourd M abundance of HDO and the value of Watson LL, Hutcheon ID, Epstein S,
(2002) Aqueous alteration in the D/H. Science 240: 1767-1770 Stolper EM (1994) Water on Mars –
Northwest Africa 817 (NWA 817) Clues from deuterium / hydrogen
Martian meteorite. Earth and Romanek CS, Grady MM, Wright IP, and water contents of hydrous phases
Planetary Science Letters 203: 431-444 Mittlefehldt DW, Socki RA, Pillinger in SNC meteorites. Science 265: 86-90
CT, Gibson EK Jr (1994) Record of
Gooding JL, Wentworth SJ, Zolensky ME fluid–rock interactions on Mars from Wright IP, Grady MM, Pillinger CT (1990)
(1991) Aqueous alteration of the the meteorite ALH84001. Nature 372: The evolution of atmospheric CO2
Nakhla meteorite. Meteoritics 655-657 on Mars: The perspective from carbon
26: 135-143 isotope measurements. Journal of
Romanek CS, Perry EC, Treiman AH, Socki Geophysical Research 95: 14789-
14794 .
Goreva J, Leshin LA, Guan Y (2003) Ion RA, Jones JH, Gibson EK (1998)
microprobe measurements of carbon Oxygen isotopic record of silicate
isotopes in martian phosphates: alteration in the Shergotty-Nakhla-
Insights into the martian mantle. Chassigny meteorite Lafayette.
Lunar and Planetary Sciences XXXIV: Meteoritics & Planetary Science
#1987 (CD-ROM) 33: 775-784
Karlsson HR, Clayton RN, Gibson EK, Rost D, Vicenzi EP, Pauli E (2005) Halite,
Mayeda TK (1992) Water in SNC sulfate, and clay assemblages in the
meteorites: Evidence for a martian Nakhla martian meteorite. Lunar
hydrosphere. Science 255: 1409-1411 Planetary Science XXXVI: #2306
(CD-ROM)

Evidence for Water
at Meridiani
Bradley L. Jolliff,1 Scott M. McLennan,2 and the Athena Science Team
excavated, and exposed rocks to a
T
he Mars Exploration Rover Opportunity has examined sedimentary
variety of depths; it has also exam-
structures in the Burns formation at Meridiani Planum. The materials ined the formation at several loca-
in this formation reflect, in part, subaqueous deposition of reworked, tions between craters on the plains.
sulfate-rich, clastic sediments that likely formed in a playa–interdune setting. Within the ~150 m diameter
Endurance crater, where Opportu-
The chemistry and mineralogy of the sedimentary rocks record an origin by
nity spent six months, these layered
evaporation of sulfate- and chloride-rich brines mixed with a fine, altered, rocks are informally called the
basaltic mud or dust component, prior to reworking. Cementation and postde- Burns formation (FIG. 2) after the
late Roger Burns, who presciently
positional reactions to form hematite-rich concretions and crystal-mold porosity
predicted the occurrence of jarosite
reflect diagenesis in a groundwater-saturated subsurface. More recent dehy- on Mars (Burns 1987; Grotzinger et
dration events are evidenced by polygonal textures in rocks within craters and al. 2005).
exposed on the plains. The timing of formation of fracture fillings that cut Here, we describe some of the fea-
across bedding is not well constrained and may be early postdiagenetic or tures seen by Opportunity’s cameras
and measured with its robotic arm
later. The fracture fillings may have formed by solutions remobilized along
instruments that tell of this watery
zones of weakness. Alteration rinds may reflect more recent interactions past. Such evidence includes
between rock and atmospheric water vapor. preserved primary sedimentary
structures, as well as chemical and
KEYWORDS: Mars Exploration Rovers, Opportunity, water on Mars mineralogical clues, that allow
interpretation of depositional envi-
INTRODUCTION ronments and the subsequent diagenetic history, both during
The Mars Exploration Rover Opportunity has now completed and after burial. The rover has also observed features that
investigations of several craters near its landing site on reflect later remobilization of materials, possibly involving
Meridiani Planum (Squyres et al. 2004a, b; Squyres and water, and features that hold a record of more recent and
Knoll 2005) and is roving southward past the boundary of perhaps ongoing interaction with tenuous atmospheric
its landing ellipse, exploring transitions between major geo- water vapor.
logical and geomorphological terrain types identified from
orbit. This surface mobility is unprecedented in the explo-
ration of Mars and is changing how we think about future
surface exploration.
The present-day surface of the Meridiani plains, as seen during
the first year and a half of this trek, could be described as a
sea of ripples across an ocean of sand (FIG. 1). Ironically this
surface is a cold, arid desert where the ripples are driven by
Martian winds. Yet just about a meter or less beneath the
eolian surface sand sheet lies an ancient, layered, sulfate-
rich rock formation that tells the story of ancient Martian
sediments once saturated with water and of the occasional
emergence—and subsequent evaporation—of stable surface
water (Squyres et al. 2004b; Grotzinger et al. 2005; McLennan
et al. 2005). Fortunately, the aptly named Opportunity
rover has been able to investigate the rocks of this formation
in a series of craters of different sizes that have penetrated,
1 Department of Earth and Planetary Sciences

Washington University
St. Louis, MO 63130-4899, USA
FIGURE 1 Pancam view looking NNW showing rover tracks and
eolian sand ripples on the Meridiani plains (extracted
2 Department of Geosciences from the “Rub al Khali” Panorama, sol 462, L256, approximate true
State University of New York at Stony Brook color). A small exposure of sulfate-rich rocks of the underlying Burns
Stony Brook, NY 11794-2100, USA formation is just barely visible at the arrow. The sheer flatness of the
E-mail: [email protected] Meridiani surface suggests a water-related environment of deposition
for the underlying rock formation. IMAGE CREDIT NASA/JPL/CORNELL

SEDIMENTARY FEATURES
Lamination
Opportunity’s first view of the surface of Mars was of a
“ledge” of outcrop in the interior wall of Eagle crater where
it landed (Squyres et al. 2004a). Although layering was evident
from the beginning, a close approach and increasingly
closer imaging showed that the scale of layering extended
down to millimeter-scale laminations (FIGS. 2, 3). Further
examination has shown that these laminations are pervasive
and reflect variable resistance to wind erosion. Microscopic
imaging has revealed the granular nature of the sediment.
The rock is a sandstone (McLennan et al. 2005) consisting
of grains that appear to be reworked, sulfate-cemented,
altered basaltic mud likely formed from desiccation of a
contemporaneous playa lake (see below).
FIGURE 3 Pancam false color image of Last Chance exhibiting

trough or festoon cross laminations, taken on sol 40 in
Eagle crater (L456, P2541). Rock is ~30 cm across. IMAGE CREDIT NASA/
JPL/CORNELL
diagenesis in a groundwater/brine-saturated environment.

Observed in outcrops at Eagle, Fram, and Endurance craters,
the spherules are dispersed and rather uniform in size,
typically 2–4 mm in diameter (FIG. 4). They constitute
about 1–2% of the volume of the rocks in which they occur,
judging by their abundance in surfaces ground by the rock
abrasion tool (RAT). They are harder and denser than the
sulfate-rich matrix in which they occur; thus they weather
out, drop to the ground, and form a resistant lag deposit
covering a vast area. In exposures abraded by the RAT, cross
sections of spherules appear to be fine grained, homoge-
neous, and lacking internal structure. These characteristics
are consistent with diffusion-limited growth and a concre-
FIGURE 2 Pancam image of the eastern side of Burns cliff,
Endurance crater, investigated on ca. sols 275–295. This
tionary origin (McLennan et al. 2005). In all outcrop exposures
is the type locality for the Burns formation. The image shows the observed to date, spherules have a dispersed distribution
lamination and bedding that characterize the upper parts of the Burns and do not concentrate preferentially along bedding
formation. For a sense of scale, the stratigraphic section exposed at planes, arguing against their formation as accretionary
Burns cliff is about seven meters high, and the large, bright, tabular rock
in the lower left is some 25 cm across. The image is from sol 278,
lapilli or impact spherules.
sequence P2440, L257. IMAGE CREDIT NASA/JPL/CORNELL
Festoon Cross-Bedding
Lamination alone is not necessarily evidence of a watery
depositional environment because planar laminae can also
form in an eolian sedimentary environment. However,
features resembling fine-scale trough or festoon cross-lami-
nation, seen early-on in parts of the Eagle crater outcrop,
indeed suggest the action of surface water in reworking and
depositing sediments. FIGURE 3 shows the rock named Last
Chance in Eagle crater, which has some of the best-exposed
small-scale festoon cross-laminae. These and other primary
depositional features in Last Chance and other rocks are
described and discussed in detail by Grotzinger et al. (2005).
These rocks occur within the “upper” Burns formation, as
exposed at Eagle and Endurance craters, in what is inferred
to be a “wetting upward” sequence. They mark the transition
from eolian dune and sand sheet facies, below, to interdune
or playa deposits in which groundwater occasionally
emerged.
Concretions
One of the great surprises of the mission was the discovery
FIGURE 4 Sol 27 (P2387) L257 portion of the Guadalupe outcrop in
at Meridiani Planum that the surface hematite signature Eagle crater showing bluish concretions embedded in
observed from orbit was largely due to a surface enrichment laminated outcrop rock. Concretions are several millimeters in diameter.
of hematite-rich spherules (Christensen et al. 2004). Litho- Also notable in this exposure are the “bird track” mineral casts formed
logical, mineralogical, and chemical features of the by selective dissolution of lenticular mineral grains. Inset is a cropped
microscopic image of concretions. Width of Pancam image ~20 cm;
spherules indicate that they are concretions, formed during width of inset image ~2 cm. IMAGE CREDIT: NASA/JPL/CORNELL/USGS

Measurements made using the Pancam, Mini-TES, and A
Mössbauer spectrometer confirm the hematite-rich miner-
alogy of the concretions (Bell et al. 2004; Christensen et al.
2004; Klingelhöfer et al. 2004). Analysis of the data indi-
cates that the spherules contain at least 50 wt% and per-
haps as much as 90 wt% hematite. Chemical data from the
Alpha Particle X-ray Spectrometer (APXS) indicate that the
concretions also contain a significant siliciclastic compo-
nent, but determining the exact composition is hindered by
contributions from dust, soil, and rock to most analyses
(Jolliff et al. 2005).
Cements and Secondary Porosity

Sandstones of the Burns formation preserve microtextural
features consistent with diagenetic processes commonly
associated with groundwater recharge and evaporation in
eolian settings. Two temporally distinct episodes of cemen-
tation are recognized: early pore-filling cement, leading to
primary lithification, and post-concretion cement, resulting
from a combination of recrystallization and new growth.
Later cements take the form of millimeter-scale concretion
overgrowths and irregular nodules consisting of cemented B
sandstone.
The rocks contain many millimeter-scale voids that are too
large to be primary depositional features. They are thus
interpreted to be secondary porosity, formed after concretion
growth but prior to the second generation of cements. Two
porosity types are observed. The first is characterized by
crystal-shaped pores interpreted to be crystal molds formed
by dissolution of a highly soluble syndepositional evaporite
mineral. The second consists of elongate to sheet-like pores,
also thought to be groundwater dissolution features. Later
in the diagenetic history, notably during the formation of
distinct stratigraphic horizons such as the “Whatanga”
contact separating the middle and upper units of the Burns
formation, pores were substantially modified and enlarged
due to further groundwater dissolution. FIGURE 5 (A) Pancam L256 false-color image of Fuller, taken on sol
208 (P2565), showing polygonal texture most likely
caused by dehydration or desiccation. Exposure is about 30 cm across.
LATER ALTERATION FEATURES (B) Pancam L257 false-color image mosaic of Wopmay from sol 251
(P2432) in approximate true orientation within Endurance crater. The
Polygonal Texture inferred predominant wind direction is from left to right. Polygonal tex-
Numerous rocks observed in Endurance crater and on the ture occurs on all surfaces of the rock but is more pronounced on the
downwind side (center to upper right). Boulder is about 1.4 m in
Meridiani plains away from craters exhibit polygonal texture.
length. IMAGE CREDIT NASA/ JPL/CORNELL
This is seen in flat-lying rocks close to the ground surface, as
in FIGURE 5A, as well as on rock surfaces in three dimensions
standing above the ground surface, e.g. in Wopmay, a large
Fracture Fillings
rock lying in the interior of Endurance crater (FIG. 5B).
These features cut across bedding structures and thus did Fractures that show evidence of mineralization along their
not form by desiccation contemporaneous with sediment length occur in several locations, notably in rock exposures
deposition. They more likely formed as a result of later within Endurance crater. One such fracture with mineralized
dehydration of originally hydrated salts. Not all exposed linings along its edges is shown in FIGURE 6A. This particular
rock surfaces show these features. A possible explanation fracture is oriented approximately radially relative to the
for this lies in the ease with which the sulfate-rich sediments crater, so one possibility is that it formed as a result of
erode by wind abrasion. Rock surfaces that are exposed to impact. This fracture could also be related to a regional
prevailing winds erode quickly, so that only rocks with a NE–SW-trending fracture pattern. In places where it has
long exposure history but some protection from wind- been analyzed, material making up the lining is composi-
driven abrasion retain this record of dehydration. In the tionally similar to the outcrop rock, so it may represent a
case of the rock Wopmay, the side of the rock facing into relatively minor fluid-assisted remobilization of local material.
prevailing winds is relatively smooth (although it has a Closed fractures with fillings also have been observed in
weak polygonal texture), whereas the downwind side of the rocks, especially one named Ellesmere in Endurance crater
rock exhibits a deep, well-exposed polygonal texture (FIG. (McLennan et al. 2005). Such features reflect an earlier
5B). McLennan et al. (2005) noted the similarity of the period of deformation and fluid mobilization of materials.
polygonal texture of Wopmay to septarian nodules, which Imaging from several meters suggested the possibility of
commonly form three-dimensional polygonal networks radial-fibrous texture, consistent with recrystallization
caused by tensional failure associated with volume changes. along the fracture. In this case, however, the detail of the
In Wopmay, this texture could result from volume decrease texture was at the limit of resolution, and it was not possible
caused by dehydration of sulfate salts and was possibly to position the rover so as to analyze or image these fracture
related to unloading following impact excavation. fillings with the microscopic imager.
ELEMENTS J UNE 2006

165
(Klingelhöfer et al. 2004). One is an octahedral Fe2+ doublet
that could be from pyroxene, but this assignment cannot be
made with certainty; ferrous sulfate or glass are also possi-
bilities. The second component is a ferric doublet that
could represent a variety of ferric oxyhydroxide. From the
Mössbauer spectra, olivine is precluded as a component of
the outcrop rocks. Mini-TES spectra of the outcrop rocks are
consistent with a mixture of hematite, jarosite, Mg- and Ca-
sulfates, a non-crystalline silica-rich phase, feldspar, and pos-
sibly sheet silicates (Christensen et al. 2004; Clark et al. 2005).
Chemical Compositions
Compositional analyses of Meridiani outcrop rocks from
the APXS (Rieder et al. 2004) have shown the rocks to be
extremely rich in S, presumably in the form of sulfate, and
rich in Cl and Br, either as halides or in sulfates. This is the
key bit of evidence regarding the evaporite-rich nature of
A
A
B
FIGURE 6 (A) Pancam false-color image of fracture in Endurance

crater showing erosion-resistant fracture-lining materials
(sol 152, P2380, L257; inset: sol 170, P2598, L257). (B) Detail of Pan-
cam image showing a rock rind in Eagle crater, sol 50, P2578, L257,
B
false-color. The relatively resistant rind is indicated by shadowed over-
hang. IMAGE CREDIT NASA/JPL/CORNELL
Rock Rinds
Weathered rock surfaces, as exposed within craters and on
the smooth Meridiani plains surface, exhibit rinds that differ
in chemical composition and resistance to weathering.
Such rinds are readily distinguished by their morphology
and subtle differences in Pancam color images (FIG. 6B). The
rinds are observed on exposures within craters and on the
plains. Chemically, they are similar to the rocks on which
they form, but they appear to have less sulfate and a higher
proportion of siliciclastic components. These rinds most
likely represent relatively recent alteration and minor redis-
tribution of salts associated with interaction between rock
surfaces and transient water, possibly in the form of frost,
from the thin, H2O-poor atmosphere.
MINERALS, ROCK COMPOSITIONS, FIGURE 7 Chemical compositions (in wt%) of Meridiani materials
(analyzed from sol 1 through sol 250) illustrating the
AND CHEMICAL TRENDS extreme sulfate enrichment of the outcrop rocks. The highest levels
measured previously by the Viking and Pathfinder landers were around
Mineralogy 10 wt% as SO3. Natural rock surfaces and fracture-filling or lining materials
Mössbauer spectra show that the outcropping sulfate-rich have compositions intermediate between those of abraded rock and
lithology in all Meridiani exposures thus far examined con- soils. Two of the soils measured in Endurance crater were especially well
sorted and rich in basaltic components; these are represented by the
tains the Fe-bearing minerals hematite and jarosite (Klingel- black square symbols. Two rock compositions are shown: (1) Bounce
höfer et al. 2004). Fine-grained hematite (other than con- Rock, which is a pyroxene basalt similar to the Martian meteorites
cretions) accounts for some 40 wt% of the Fe. The firm EETA79001 (lithology B) and Zagami, and (2) Barberton, which is a
identification of jarosite indicates the presence of signifi- small cobble that has a Mg-rich composition and is probably a fragment
of a meteorite. (A) total Fe as Fe2O3 versus SO3, showing the Fe enrichment
cant OH in these rocks and formation by precipitation from of the spherule concretions. (B) Al2O3 versus SO3, showing the trend of
acidic, sulfate-rich brine. Two other Fe-bearing components decreasing alumina with increasing sulfate. Alumina extrapolates to zero
have been identified in outcrop from Mössbauer spectra at about 58 wt% SO3.

Burns formation rocks. FIGURE 7 shows the distribution of rich clastic sediments most likely formed by evaporative
compositions in terms of SO3, total iron as Fe2O3, and cementation of mainly eolian-reworked primary evaporites
Al2O3; these plots illustrate some of the important major- and subaqueous, sulfate-rich, basalt-derived muds associated
element compositional relations among different groups of with playa lake or interdune groundwater (brine). The com-
materials at the landing site. In general, alumina, silica, and position of the rocks indicates that the sediments are
other elements associated with the siliciclastic component composed of about half evaporite components and half
vary inversely with SO3, whereas Fe2O3 and CaO decrease very fine-grained siliciclastic components. Following
slightly and MgO remains relatively constant. These relations reworking and deposition, burial and interaction with
result from the association of a portion of the Ca and Fe groundwater brines resulted in several stages of diagenesis.
with S in sulfates and the abundance of Mg sulfate in the The stages include syndepositional formation of tabular
evaporite component. The inferred abundance of Mg sulfate crystals cutting across primary laminations; fine-grained
leaves open the possibility of significant amounts of H2O, cementation and filling of primary porosity; formation of
perhaps as much as 10 wt% or more, if the Mg sulfate is hematite-cemented concretions; selective dissolution of
present as one or more hydrated forms, such as kieserite, early-formed minerals, e.g. ferrous iron sulfates; local sul-
starkeyite, hexahydrite, or amorphous MgSO4•xH2O. fate recrystallization; and dissolution of tabular crystals to
form crystal-mold porosity. Subsequent to diagenesis, these
The observed variations of SiO2 and Al2O3 with SO3 indicate
rocks were affected by the formation of fractures, fracture
the presence of a silica phase, e.g. opaline silica, with the
fillings, and veins, possibly associated with impact and/or
evaporite component, but little or no Al sulfate or aluminous
dehydration. More recently, erosion by wind abrasion of
jarosite. A strong correlation of Na and K with Al2O3 sug-
overlying concretion-bearing strata and removal of fine
gests the presence of a feldspar component. Compositional
materials produced a lag deposit of spherules and possibly
trends for the RAT-ground rock surfaces are consistent with
granular siliciclastic debris, in addition to basaltic eolian
limited variability in the evaporite component and only
sands and dust. Interaction of exposed rock surfaces with
modest variation in the proportion of sulfate to siliciclastic
atmospheric water vapor, and possibly small amounts of
components. This lack of variability is in turn consistent
transient liquid water, produced rinds and generated
with sedimentary reworking and some degree of homoge-
micron- to millimeter-scale redistribution of soluble com-
nization of the original evaporitic deposits. If the rocks had
ponents. The presence of jarosite and inferred Mg- and
formed primarily by evaporation and have remained where
Ca- sulfates holds open the possibility that these sedimen-
they formed until today, we would expect to find more
tary rocks sequestered and still contain a significant reser-
significant and systematic variations in cation ratios, and
voir of water.
evaporite mineral abundances should follow some systematic
evaporation sequence (e.g. Hardie et al. 1985).
ACKNOWLEDGMENTS
IMPLICATIONS Funding for this work was through NASA support of the
MER Athena science team. We thank the MER engineering
Clearly a water-related history is indicated for the sedimentary
and science teams for their dedication and careful work in
origin and diagenesis of the Meridiani sulfate-rich rocks.
guiding the rover through this fantastic exploration and
These rocks may well record one of the late occurrences of
adventure. We especially thank the instrument teams for
surface-water stability on Mars before the atmosphere
their hard work in providing the images and data used in
became cold, thin, and CO2 rich. The survival of soluble
this paper. The manuscript was improved by reviews and
salts in Burns formation rocks, which may be several billion
comments from Matt Golombek, Hap McSween, and Rod
Ewing. .
years old, argues against any significant interaction with
surface water or precipitation in this region following the
formation of craters such as Eagle and Endurance. The sulfate-
REFERENCES Hardie LA, Lowenstein TK, Spencer RJ Squyres S and 49 coauthors (2004a) The
(1985) The problem of distinguishing Opportunity rover’s Athena science
Bell JF III and 39 coauthors (2004) Pancam between primary and secondary investigation at Meridiani Planum,
multispectral imaging results from features in evaporites. 6th Interna- Mars. Science 306: 1698-1703
the Opportunity rover at Meridiani tional Symposium on Salt 1: 11-39
Planum. Science 306: 1703-1709 Squyres SW and 18 coauthors (2004b) In
Jolliff BL and the Athena Science Team situ evidence for an ancient aqueous
Burns RG (1987) Ferric sulfates on Mars. (2005) Composition of Meridiani environment at Meridiani Planum,
Proceedings of the 15th Lunar and hematite-rich spherules: A mass- Mars. Science 306: 1709-1714
Planetary Science Conference, in balance mixing-model approach
Journal of Geophysical Research 92 (extended abstract). Lunar and Squyres SW, Knoll AH (2005) Sedimentary
(supplement, (B4)): E570-E574 Planetary Science 36: 2269 rocks at Meridiani Planum: Origin,
diagenesis, and implications for life
Christensen PR and 26 coauthors (2004) Klingelhöfer G and 18 coauthors (2004) on Mars. Earth and Planetary Science
Mineralogy at Meridiani Planum Jarosite and hematite at Meridiani Letters 240: 1-10 .
from the Mini-TES experiment on the Planum from Opportunity’s
Opportunity rover. Science 306: Mössbauer spectrometer. Science 306:
1733-1739 1740-1745
Clark BC and 23 coauthors (2005) Chemistry McLennan SM and 31 coauthors (2005)
and mineralogy of outcrops at Provenance and diagenesis of the
Meridiani Planum. Earth and evaporite-bearing Burns formation,
Planetary Science Letters 240: 73-94 Meridiani Planum, Mars. Earth and
Grotzinger JP and 19 coauthors (2005) Planetary Science Letters 240: 95-121
Stratigraphy and sedimentology of a Rieder R and 14 coauthors (2004)
dry to wet eolian depositional system, Chemistry of rocks and soils at
Burns formation, Meridiani Planum, Meridiani Planum from the Alpha
Mars. Earth and Planetary Science Particle X-ray Spectrometer. Science
Letters 240: 11-72 306: 1746-1749

Water on Mars and the
Prospect of Martian Life
Andrew H. Knoll1 and John Grotzinger2
favorable medium for the chem-
E
vidence of water on Mars dates back to the first observations of channeled
istry of life. At temperatures typical
landscapes made by Mariner 9 and Viking. More recent images from Mars
of the Earth’s surface, liquid water
Global Surveyor and Mars Express strikingly confirm that fluids have coexists with gaseous CO2, N2, O2,
sculpted the Martian surface at least episodically through its history. The H2S, and NH3; it can accumulate
Mars Exploration Rovers Opportunity and Spirit have added evidence for relatively high concentrations of
extensive rock–water chemical interactions in the regions where these remote CO32-, HCO3-, HS-, SO42-, NH4+,
Fe2+, and other ions in solution; it
geologists landed, while OMEGA and THEMIS have shown that similar processes
carries dissolved organic molecules
took place in many parts of the planet. stably in solution; and it interacts
strongly at polar interfaces. Thus,
Because of the close relationship between water and biological activity on
water provides a medium in which
Earth, such observations have been taken as hopeful signs that Mars, as well, carbon and the other chemical
might once have supported life and, indeed, might still do so in subterranean ingredients of life can interact with
oases. There is, however, much more to consider. Water appears to be neces- one another and, in a biological
sary for life, at least as it exists on Earth and can be contemplated on Mars, world, with organisms.
but it does not, by itself, insure habitability. In this paper, we review the Water may not be unique in this
broader requirements for biological activity as they relate to water and use respect—it has been suggested that
these to constrain astrobiological inferences about Mars. ammonia, formamide, and several
other organic compounds might
KEYWORDS: Mars, Meridiani, water, life, astrobiology function as biopermissive fluids at
the right temperature and pressure
WHY IS WATER KEY TO RECIPES FOR LIFE? (Benner et al. 2004). But such compounds are, at best, triv-
In chemistry, geometry is destiny, and it is the distinctive ially important as fluid environments on Earth and are
molecular geometry of H2O that accounts for many of its doubtfully more relevant to Mars. The fact that water mol-
biofriendly attributes (Finney 2004). In water molecules, a ecules interact in complex and specific ways with enzymes
relatively large oxygen nucleus is flanked by two smaller, (e.g. Rand 2004) might be taken as further
asymmetrically placed hydrogen nuclei, forming a distinctly evidence for the unique fit between H2O and life, but this
polar molecule with a mean HOH angle may tell us only that life on Earth has
of 104.5°. Because of their pronounced evolved to maximize biochemical func-
dipole moment, water molecules interact Water, water, every where, tion in an aqueous milieu (Ball 2005).
strongly with one another via hydrogen And all the boards did shrink;
bonding. Not only do these molecular Water, water, every where, WHY THERE IS MORE
interactions determine the crystallo- TO THE RECIPE
graphic structure of ice, which famously Nor any drop to drink.
(and, from a biological standpoint, use- Nutrients
“Rime of the Ancient Mariner”
fully) floats in water, they also explain Samuel Taylor Coleridge
Water may be necessary for biological
why H2O is liquid at temperatures activity, but as the plight of the Ancient
where most comparably small molecules Mariner illustrates, it is certainly not
are gases. sufficient. Terrestrial life is based on carbon, and collec-
tively, C, H, and O make up nearly 80% of the dry weight
Water ionizes readily, resulting in anomalously high rates
of a bacterial cell. (“Dripping wet,” bacteria are about 70%
of molecular diffusion and a pronounced capacity to conduct
water.) Cells, however, contain many additional elements—
excess protons (Finney 2004). Also, water effectively disso-
more than thirty in typical microorganisms. Other major
ciates ionic species. Thus, water provides a particularly
constituents include N and P, with cellular C:N:P lying near
106:16:1 (what biologists call the Redfield ratio). N and P
1 Department of Organismic and Evolutionary Biology find obligate use in proteins, membranes, and nucleic acids,
Harvard University the fundamental structural, functional, and informational
Cambridge MA 02138, USA molecules of the cell. It is by no means clear that the same
E-mail: [email protected] molecules will characterize life wherever we may find it, but
2 Division of Geological and Planetary Sciences it is hard to conceive of functional and informational
California Institute of Technology macromolecules that do not contain N or P. Sulfur also
Pasadena CA 91125, USA plays key roles in cells, notably in S–S bonds that govern the
three-dimensional conformation of proteins and FeS

clusters that form the functional heart of some ancient and successfully after 1288 ± 271 years (radiocarbon dates) of
critically important enzymes. Indeed, De Duve (1995) has dormancy (Shen-Miller et al 1995).
speculated that thioesters played a crucial role in the origin
Scientists have reported much longer dormancy in bacteria
of life on Earth. For these reasons, life likely can form and
from Pleistocene permafrost and longer still in salt deposits
persist only on planets where N, P, and possibly S are present
formed more than 250 million years ago. Such claims are
in biologically available forms.
controversial—even dormant cells must expend energy
Other elements are required in minor or trace abundances. now and again to repair molecular damage such as sponta-
For example, Fe and, in many organisms, Mo are essential neous DNA breakage. But these reports underscore the fact
cofactors in biological fixation of nitrogen. Mo is required, that we cannot yet place a precise limit on the duration of
as well, by organisms that use nitrate as a source of N for cellular dormancy. Nor do we know with confidence what
proteins, nucleic acids, and other biomolecules. Thus, on biochemical features account for prolonged dormancy in
Earth there could be no biological nitrogen cycle without organisms that exhibit this trait.
metal ions available in solution. Mg occupies the structural
Regardless of this uncertainty, the time scale of return for
and functional center of chlorophyll, while Mn atoms
liquid water must loom as a key variable in ongoing assess-
mediate the extraction of electrons from water by plants
ments of Martian habitability. Playas that receive water
and algae. Similarly, then, without Mg and Mn, photosyn-
once every decade likely persist as habitable environments
thesis as we know it could not exist. Fe plays a central role
indefinitely; those that recharge once every ten million
in hemoglobin, extending the metabolic role of metals to
years may not—unless there exists a reservoir of populations
respiration. The recipe for life, then, is complex, and likely
elsewhere that can recolonize ambient waters whenever
has been from the beginning—metallic cofactors have been
they reappear.
implicated in many prebiotic chemical reactions and in
purportedly primitive biochemical reactions (Williams and This latter point is worth underscoring. On Earth, habitable
Fraústo da Silva 1996). The origin and long-term persistence extremes exist in the context of “normal” environments
of life requires not only water, but adequate supplies of that provide nutrient subsidies and persistent reservoirs of
major, minor, and trace elements, whose availability will be colonizing populations. Maintenance of a viable biota may
determined by source rocks and their chemical weathering, be far more difficult on planets where environments at or
as well as by pH, redox, and other environmental parameters. near the extremes of habitability are the most favorable sites
Thus, while water sparked early astrobiological interest in for life.
Mars, we have come to the point in planetary exploration
where a far more sophisticated understanding of environ- Acidity
mental geochemistry is required. On Earth, the habitability of aqueous environments is
influenced by additional factors, including pH, Eh, and
Water Activity temperature (Knoll and Bauld 1989). Acidity is particularly
Clearly, environments where life can thrive do not contain relevant because geochemical data from Meridiani Planum
pure H2O, but an aqueous solution of ionic and dissolved indicate that sulfuric acid was present when Meridiani sed-
organic constituents. Through their interactions with water imentary rocks formed (McLennan et al. 2005). Places like
molecules, however, these substances limit the availability Rio Tinto, Spain, where strongly acidic waters deposit
of H2O for hydration reactions, and at high concentrations, jarosite and iron oxides, provide insights into acid tolerance
ions and organic molecules can inhibit cellular physiology. (FIG. 1, Fernández-Remolar et al. 2005). Acid-tolerant popu-
Water activity is the effective water content of a solution (in lations thrive in such environments, not because they can
the notation used by microbiologists, aw = n1/(n1 + n2), run their biochemistry at low pH, but because they effi-
where n1 equals moles of water and n2 equals moles of ciently expel protons from their cytoplasm, enabling cell
solute; Grant 2004). Most terrestrial organisms cannot grow chemistry to continue under more or less neutral conditions.
and reproduce at water activities below 0.9. A few bacteria
Acid and desiccation tolerance are not universal attributes
grow where aw is as low as 0.85, and some archaea can live
of terrestrial organisms, and most groups that accommodate
at an aw of 0.75 (the water activity of a halite-saturated solu-
these environmental challenges are descended from ancestors
tion). In terms of this parameter, however, fungi define the
that tolerated them poorly. On Earth, then, life can persist
limits of life: one remarkable species has been shown to
in arid, oxidizing, and acidic habitats, but it might not do
grow at aw = 0.61 (Grant 2004). (Honey has such a long
so if those were the only habitable environments on the
shelf life because its water activity falls below the limit for
planet. Moreover, accumulated data on prebiotic chemistry
microbial spoilage.) The important point is that in brines
suggest that life could not have arisen under such
where aw is persistently low, water may be present but
conditions (Knoll et al. 2005).
uninhabitable.
The Persistence of Water ASTROBIOLOGY AND THE RECORD

Regardless of their ability to persevere at chronically low OF WATER ON MARS
water activities, many organisms—from bacteria to Geomorphological and Sedimentological
animals—can tolerate episodic dryness. Desiccation-tolerant Observations
organisms persist in a dormant state, forming spores, glyco-
Much of the ancient cratered terrain of Mars is dissected by
protein capsules, or other structural phases that retard
small valley systems similar to terrestrial river networks
water loss.
(Baker and Milton 1974). Because significant precipitation
The key variable is the return time of habitable water. and surface runoff are not possible under modern condi-
Endolithic lichens in Antarctic boulders are metabolically tions, some researchers have interpreted the valley networks
active only a few weeks each year, but it is enough for as evidence of warmer, more humid ancient climates (Carr
populations to persist over geological time scales (Fried- 1996). Alternatively, these drainage patterns could have
mann et al. 1993). Cyanobacteria, dried onto herbarium been created through venting of underground water to the
sheets in the nineteenth century, resume photosynthesis surface, events hypothesized to have been vigorous, but
almost immediately upon wetting. Even lotus seeds short-lived (Baker and Milton 1974). More recently, the dis-
preserved in lake beds from China have been germinated covery of a channelized alluvial fan northeast of Holden

crater has provided evidence for deposition of loose sediment Geochemical Observations
under aggrading conditions (Malin and Edgett 2003) within The Mars Exploration Rover Opportunity has discovered
well-developed meandering channels—water flowed freely jarosite [(K,Na,H3O)(Fe3-xAlx)(SO4)2(OH)6, where x < 1] in
across the fan surface. The characteristic time scale for such Meridiani outcrop rocks, while its mechanical twin Spirit
deposits leads to a minimum estimate for the volume found goethite [FeO(OH)] in Gusev crater; both minerals
(900–5000 km3) and duration (50–1000 yrs) of water flow form in the presence of water (Klingelhöfer et al. 2004,
(Jerolmack et al. 2004). This modest estimate does not 2005). Elemental abundances indicate that Ca and Mg sulfates
require precipitation, so long as a local source of water is occur with the jarosite at Meridiani, although their precise
present. mineralogy cannot be ascertained (Clark et al. 2005). Mars
Geomorphological evidence for larger and longer-lived Express has identified gypsum (CaSO4•2H2O), kieserite
water bodies remains controversial (Baker and Milton (MgSO4•H2O), and, possibly, other polyhydrated sulfates
1974). Most recently, high-resolution images of the Elysium on the Meridiani plain and more widely on the Martian sur-
region obtained by Mars Express have indicated the possi- face (Gendrin et al. 2005). Finally, as noted above, Oppor-
ble presence of pack ice preserved beneath a mantle of soil tunity has confirmed the presence of hematite at Meridiani,
(Murray et al. 2005). In this interpretation, a lake or shallow most conspicuously as millimeter-scale concretions that
sea perhaps 50 m deep formed during catastrophic eruption formed during early diagenesis of sulfate-rich sediments
of groundwater from nearby fractures, only to freeze (McLennan et al. 2005). Thus, geochemical measurements
partially, generating pack ice that subsequently became at outcrop level confirm and extend geomorphological and
embedded in a larger ice body. Alternatively, these features sedimentological evidence for water on the ancient surface
may be rafts of frozen lava which floated atop a larger pool of Mars.
of igneous melt that flash-froze to preserve the observed
geometry. Even if the water–ice hypothesis proves correct, Climate Evolution
however, liquid water need not have been stable for long The climatic history of Mars since the end of heavy
intervals (>1000s of years) on the Martian surface. bombardment (ca. 3.8 Ga) is clearly controversial. Some
hold that oceans persisted episodically long after bombard-
Ground-based observations by the Mars Exploration Rovers
ment ended. Others argue that while Noachian oceans
provide considerable insight into the mechanisms by which
existed, the Martian surface froze near the end of heavy
sulfate minerals may have formed on Mars. Opportunity
bombardment and has remained that way ever since.
images provide compelling evidence for the accumulation
Indeed, based on the thermochronology of Martian mete-
of sediment particles—formed of admixtures of sulfate salts
orites, Shuster and Weiss (2005) claim that Mars has not
and silicate minerals—in a variety of sedimentary deposi-
seen temperatures significantly above freezing for the past
tional environments. Eolian strata are capped by interdune
four billion years.
fluvial strata that document shallow overland flows with
moderate flow velocities (FIG. 2, Grotzinger et al. 2005). We can reconcile geochemical and geomorphological
Further evidence for an active water table is found in strati- evidence for liquid water with subzero Martian tempera-
graphically restricted zones of recrystallization and second-
ary porosity, millimeter-scale hematitic concretions, and Blood red waters of Rio Tinto, southwestern Spain. Ferric
FIGURE 1
millimeter-scale crystal molds that cut across primary layering iron colors this highly acidic river (pH 0.9 to 3); the
(McLennan et al. 2005). orange precipitates are comprised of jarosite, schwertmannite, and
other iron sulfate minerals, as well as nanophase iron oxides (Fernández-
Remolar et al. 2005).

tures in a simple way: by invoking antifreeze. Salts, present FIGURE 2 A seven-meter section of sedimentary rocks exposed
in abundance at Meridiani Planum, would lower the freezing within Endurance crater, Meridiani Planum, Mars. These
rocks, the first sedimentary succession ever examined at outcrop scale
temperatures of ambient waters, as would sulfuric acid. on Mars, preserve a record of ancient water. Large-scale cross-bedding
Dilute aqueous solutions of sulfuric acid can depress the visible in the lower left indicates the passage of ancient sand dunes.
freezing temperature of water by as much as 70°C, providing These cross-bedded deposits are overlain by laminated sand sheet
a particularly effective way of reconciling diverse observations deposits also formed by wind-blown sand; however, the overlying sed-
iments include cross-laminated, rippled beds deposited by shallow, sub-
of Martian environments through time (Knoll et al. 2005). aqueous flows. The dark band visible near the top of the succession is
interpreted to have formed during alteration and cementation of the
Considered collectively, and conservatively, there are no
sediments by percolating groundwater.
geomorphological, geochemical, or sedimentological features
yet discovered on Mars that cannot be accounted for by
intermittent, short-term flow of surface water, supplied by be—equally, it could be unusually favorable from an astro-
underground sources and dispersed in a cold, dry climatic biological perspective. All in all, the aqueous deposits of
regime. The simple observation that channels cut early in Meridiani Planum are biologically permissive, but they may
Mars history persist to the present tells us that surface water record the sunset of a habitable Martian surface, not its
flow has been limited for a long time. This does not eliminate beginning.
the possibility of a persistently warm, wet Mars in Noachian
The briny acidic waters of Meridiani Planum would certainly
time, but does emphasize that evidence to support such a
constitute a formidable challenge to the types of prebiotic
model remains meager (Gaidos and Marion 2003).
chemical reactions thought to have played a role in the ori-
gin of life on Earth (Knoll et al. 2005). This is a relevant con-
DISCUSSION sideration because heavy bombardment could have eradi-
Mars today is a forbidding place. Temperature and atmos- cated any surficial life that evolved during Mars’ earliest
pheric pressure lie near the triple point of water—indeed history. One might argue that Mars could have been
liquid water is not stable on the present-day Martian surface (re?)colonized after late heavy bombardment by organisms
(Gaidos and Marion 2003). The surface is also chemically transported by meteorites from Earth. The physical mecha-
harsh and subject to strong radiation. It is doubtful that nism is plausible—the key question, however, is: what is the
organisms thrive today at the Martian surface. probability that terrestrial colonists would have landed in a
Martian environment that could support sustained
From the preceding paragraphs, one might well conclude
metabolism?
that surface environments have been biologically challenging
for most of Mars’ history. The salty dunes and transient The most promising places to look for evidence of surface
interdune streams that covered Meridiani Planum three to life on Mars are probably sedimentary basins that preserve
four billion years ago indicate that while chemical weathering a record of Mars’ earliest history, when water was most
and erosion provided many of the elements required for abundant and persistent and both oxidation and acidity
life, ambient environments were arid, acidic, and oxidizing were least developed. We know relatively little about such
(Knoll et al. 2005). Terrestrial ecology suggests that microor- terrains, but they would seem prime candidates for future
ganisms could survive many aspects of the inferred Meridi- missions aimed at understanding Mars’ environmental
ani environment, but habitability would depend critically history, as well as astrobiology.
on the time scale of water with sufficiently high water
activity to support cell biology—a parameter that is The Subsurface Alternative
currently unknown. Meridiani waters may have been habit- If surface environments on Mars have been challenging for
able upon introduction, but water activity would have life for the past several billion years, what about the
dipped below habitable levels as groundwater, playas, or subsurface? There is inherent skepticism when environ-
both, evaporated to dryness. ments deemed most likely to support life are those least
amenable to observation. Nonetheless, the subsurface was
Whether Meridiani is broadly representative of the Martian
(and may still be) the most likely place on Mars to find per-
surface three to four billion years ago is unknown, but
sistent reservoirs of liquid water. Given a continuing supply
remote sensing from Mars Express suggests that it could

of nutrients (which introduces its own set of challenges), The Future
life in the Martian crust might be sustained by a primordial In future exploration, astrobiologists need to learn about
hydrogen economy—chemical energy in the form of H2 the time scales on which water has persisted on Mars, and
produced by aqueous alteration of basalts. geochemical analyses must be extended to include nutri-
The only Martian rocks known to contain carbonate minerals ents, especially nitrogen in Martian soils. Better models and
are meteorites that preserve iron and magnesium carbon- experiments can sharpen our sense of water activity during
ates precipitated in subsurface cracks flushed by groundwa- the deposition and early diagenesis of Meridiani-type sedi-
ter (McKay et al. 1996; Bridges and Grady 2000). Sulfide mentary rocks, while future orbital and lander missions will
minerals formed as well, indicating that at least some sub- tell us the extent to which such environments were repre-
surface environments were neither acidic nor oxidizing sentative of the early Martian surface. Was ancient Mars
early in Martian history. At present, however, we know little generally arid, acidic, and oxidizing, and if so, when did it
about water activity or persistence in such environments. become that way?
Several laboratories have reported methane emissions from Of course, the biggest hurdle for astrobiology concerns
the Martian surface (Krasnopolsky et al. 2004; Formisano et biology, itself. To what extent can we generalize from obser-
al. 2005), and the argument has been advanced that these vations of the only biological planet we know? That problem
exceed fluxes expected for abiotic methanogenesis will not be solved soon, meaning that the search for evidence
(Krasnopolsky et al 2004). Hydrothermal alteration of crustal of life elsewhere will remain empirical and difficult. But as
rocks may, however, be sufficient to explain the reported future missions provide improved data on the environmen-
fluxes (Lyons et al. 2005). Moreover, the proximal source of tal history of Mars, we may yet learn whether life on Earth
methane (if correctly identified) need not be limited to current is unique or merely uniquely successful in our solar system.
biological or hydrothermal processes. Possibly, current
methane fluxes reflect release from permafrost as it sublimes, ACKNOWLEDGMENTS
decoupling current emanations from physical processes of We thank NASA’s Mars Exploration Rover project for support
formation. and MER’s engineers and scientists for the extraordinary
mission that made this paper possible. .
REFERENCES Gendrin A, Mangold N, Bibring J-P, Malin MC, Edgett K (2003) Evidence for
Langevin Y, Gondet B, Poulet F, Bonello persistent flow and aqueous sedimenta-
Baker VR, Milton DJ (1974) Erosion by G, Quantin C, Mustard J, Arvidson R, tion on early Mars. Science 302:
catastrophic floods on Mars and Earth. LeMouélic S (2005) Sulfates in Martian 1931-1934
Icarus 23: 27-41 layered terrains: The OMEGA/Mars
Express view. Science 307: 1587-1591 McKay DS, Gibson EK, Jr, Thomas-Keprta
Ball P (2005) Water and life: Seeking the KL, Vali H, Romaneck CS, Clemett SJ,
solution. Nature 436: 1084-1085 Grant WD (2004) Life at low water activity. Chillier XDF, Maechling CR, Zare RN
Benner SA, Ricardo A, Carrigan MA (2004) Philosophical Transactions of the Royal (1996) Search for past life on Mars:
Is there a common chemical model for Society B 359: 1249-1267 possible relic biogenic activity in martian
life in the universe? Current Opinion in meteorite ALH84001. Science 273:
Grotzinger JP and 19 coauthors (2005) 924-930
Chemical Biology 8: 672-689 Stratigraphy and sedimentology of a dry
Bridges JC, Grady MM (2000) Evaporite to wet eolian depositional environment, McLennan SM and 31 coauthors (2005)
mineral assemblages in the nakhlite Burns formation, Meridiani Planum, Provenance and diagenesis of the
(martian) meteorites. Earth and Planetary Mars. Earth and Planetary Science Letters evaporate-bearing Burns formation,
Science Letters 176: 267-279 240: 11-72 Meridiani Planum, Mars. Earth and
Planetary Science Letters 240: 95-121
Carr MH (1996) Water on Mars. Oxford Jerolmack DJ, Mohrig D, Zuber MT, Byrne
University Press, New York S (2004) A minimum time for the Murray JB and 12 coauthors (2005)
formation of Holden Northeast fan, Evidence for the Mars Express High
Clark BC and 23 coauthors (2005) Mars. Geophysical Research Letters 31: Resolution Stereo Camera for a frozen sea
Chemistry and mineralogy of outcrops at L21701, doi 10129/2004GL021326 close to Mars’ equator. Nature 434: 352-356
Meridiani Planum. Earth and Planetary
Science Letters 240: 73-94 Klingelhöfer G and 18 coauthors (2004) Rand RP (2004) Probing the role of water
Jarosite and hematite at Meridiani in protein conformation and function.
De Duve C (1995) Vital Dust: Life As a Planum from Opportunity’s Mössbauer Philosophical Transactions of the Royal
Cosmic Imperative. Basic Books, spectrometer. Science 306: 1740-1745 Society B 359: 1277-1285
New York
Klingelhöfer G and 11 coauthors (2005) Shen-Miller J, Mudgett MB, Schopf JW,
Fernández-Remolar DC, Morris RV, MIMOS II on MER—One year of Clarke S, Berger R (1995) Exceptional
Gruener JE, Amils R, Knoll AH (2005) Mössbauer spectroscopy on Mars: from seed longevity and robust growth –
The Rio Tinto Basin, Spain: Mineralogy, jarosite at Meridiani Planum to goethite ancient sacred lotus from China.
sedimentary geobiology, and implica- at Gusev Crater. 36th Lunar and American Journal of Botany 82:
tions for interpretation of outcrop rocks Planetary Science Conference, abstract 1367-1380
at Meridiani Planum, Mars. Earth and pdf 2349
Planetary Science Letters 240: 149-167 Shuster DL, Weiss BP (2005) Martian
Knoll AH, Bauld J (1989) The evolution of surface paleotemperatures from
Finney JL (2004) Water? What’s so special ecological tolerance in prokaryotes. Royal thermochronology of meteorites.
about it? Philosophical Transactions of Society of Edinburgh Transactions 80: Science 309: 594-600
the Royal Society B 359: 1145-1165 209-223
Williams RJP, Fraústo da Silva JJR (1996)
Formisano V, Atreya S, Encrenaz T, Knoll AH and 12 coauthors (2005) An The Natural Selection of the Chemical
Ignatiev N, Giuranna M (2005) Detection astrobiological perspective on Meridiani Elements. Oxford University Press,
of methane in the atmosphere of Mars. Planum. Earth and Planetary Science Oxford, 672 pp .
Science 306: 1758-1761 Letters 240: 179-189
Friedmann EI, Kappen L, Meyer MA, Krasnopolsky VA, Maillard JP, Owen TC
Nienow JA (1993) Long-term productiv- (2004) Detection of methane in the
ity in the cryptoendolithic microbial Martian atmosphere: evidence for life?
community of the Ross Desert, Icarus 172: 537-547
Antarctica. Microbial Ecology 25: 51-69
Lyons JR, Manning C, Nimmo F (2005)
Gaidos E, Marion G (2003) Geological and Formation of methane on Mars by fluid-
geochemical legacy of a cold early Mars. rock interaction in the crust. Geophysical
Journal of Geophysical Research 108(E6): Research Letters 32: L13201,
5055, doi:10.1029/2002JE002000 doi:10.1029/2004GL022161

Water On Mars PDF

Uploaded by

Copyright:

Available Formats

Water On Mars PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Water On Mars PDF

Uploaded by

Copyright:

Available Formats

Water on Mars

Harry Y. McSween Jr.1

The papers in this issue of Elements

1 Department of Earth and Planetary Sciences

ELEMENTS, VOL. 2, PP. 135–137 135 J UNE 2006

GLOSSARY OF MARS TERMS

Remote sensing instruments

ELEMENTS 136 J UNE 2006

ELEMENTS 137 J UNE 2006

ELEMENTS, VOL. 2, PP. 139–143 139 J UNE 2006

ELEMENTS 140 J UNE 2006

ELEMENTS 141 J UNE 2006

ELEMENTS 142 J UNE 2006

ELEMENTS 143 J UNE 2006

regolith consists of a mixture of

ELEMENTS, VOL. 2, PP. 145–150 145 J UNE 2006

ELEMENTS 146 J UNE 2006

High-Albedo Surfaces Low-Albedo Surfaces

ELEMENTS 147 J UNE 2006

ELEMENTS 148 J UNE 2006

Implications for Water A

ELEMENTS 149 J UNE 2006

REFERENCES Christensen PR and 21 coauthors (2003) Journal of Geophysical Research 102:

ELEMENTS 150 J UNE 2006

the summer at both poles. The

ELEMENTS, VOL. 2, PP. 151–155 151 J UNE 2006

ELEMENTS 152 J UNE 2006

ELEMENTS J UNE 2006

deposited during periods of higher obliquity when surface

ELEMENTS 154 J UNE 2006

REFERENCES measurements of Martian water Mellon MT, Phillips RJ (2001) Recent

ELEMENTS 155 J UNE 2006

(1998) provide an overview of the

ELEMENTS, VOL. 2, PP. 157–162 157 J UNE 2006

ELEMENTS 158 J UNE 2006

FIGURE 2 Transmitted-light images of low-temperature alteration

ELEMENTS 160 J UNE 2006

ELEMENTS 161 J UNE 2006

ELEMENTS 162 J UNE 2006

excavated, and exposed rocks to a

1 Department of Earth and Planetary Sciences

ELEMENTS, VOL. 2, PP. 163–167 163 J UNE 2006

FIGURE 3 Pancam false color image of Last Chance exhibiting

diagenesis in a groundwater/brine-saturated environment.

ELEMENTS 164 J UNE 2006

Cements and Secondary Porosity

ELEMENTS J UNE 2006

FIGURE 6 (A) Pancam false-color image of fracture in Endurance

ELEMENTS 166 J UNE 2006

ELEMENTS 167 J UNE 2006

favorable medium for the chem-

ELEMENTS, VOL. 2, PP. 169–173 169 J UNE 2006

The Persistence of Water ASTROBIOLOGY AND THE RECORD

ELEMENTS 170 J UNE 2006

ELEMENTS 171 J UNE 2006

ELEMENTS 172 J UNE 2006

ELEMENTS 173 J UNE 2006

You might also like