On the origin of gypsum in the Mars north polar region
Kathryn E. Fishbaugh,1 Franc?ois Poulet,2 Vincent Chevrier,3 Yves Langevin,2
and Jean-Pierre Bibring2
Received 16 November 2006; revised 26 January 2006; accepted 29 March 2007; published 13 July 2007.
[1] We describe the distribution and concentration of the largest Martian gypsum deposit
discovered to date by the Mars Express OMEGA (Observatoire pour le Mineralogie,
l?Eau, les Glaces et l?Activite?) imaging spectrometer, its relationship to the late
Amazonian-aged north polar dunes in which it is found, and its likely origin. Gypsum
has not been discovered anywhere within the north polar region outside of the Olympia
Undae dune sea. In the areas of highest gypsum a concentration, 35% pure gypsum
grains of a few tens of micrometers in size, mixed with 65% millimeter-sized gypsum
grains containing, dark, spectrally featureless inclusions best fit the OMEGA
observations. The gypsum-rich dunes contain no significant average albedo, temperature,
or morphological anomalies. We propose that water emanating from nearby channels,
carved during melting of the polar layered deposits, infiltrated the eastern end of the
polar dune sea, percolating through the dunes. Deposits of gypsum resulted from a
combination of direct, in situ alteration of sulfide- and high-calcium-pyroxene-bearing
dunes and from formation of evaporitic gypsum crystals in the pore spaces of these
dunes. This gypsum deposit formed in a unique local environment and is disconnected
from sulfate-forming events elsewhere on Mars which are thought to have occurred much
earlier, during the late Noachian and Hesperian, by various means. Sulfates have not
been discovered in any other collection of dunes on Mars.
Citation: Fishbaugh, K. E., F. Poulet, V. Chevrier, Y. Langevin, and J.-P. Bibring (2007), On the origin of gypsum in the Mars north
polar region, J. Geophys. Res., 112, E07002, doi:10.1029/2006JE002862.
1. Introduction and Background
[2] The Observatoire pour le Mineralogie, l?Eau, les
Glaces et l?Activite? (OMEGA) on Mars Express has recently
detected the largest gypsum deposit on Mars within the
Amazonian-aged north polar Olympia Undae sand sea
[Langevin et al., 2005] (Figure 1). OMEGA is a visible and
near-infrared (NIR) imaging spectrometer (0.35 to 5.1 mm)
which has a nominal spatial resolution in the polar regions of
a few kilometers per pixel, but data can also be collected at a
resolution of
1 km/pixel. Perhaps themost intriguing aspect
of the polar gypsum discovery is the fact that liquid water is
needed to form gypsum. While liquid water has carved
gullies [e.g., Malin and Edgett, 2000] and even outflow
channels (review in the work of Carr [1996]) during the
Amazonian (3.05 bya to present), creation of sulfates is
another story. On the basis of the ages of materials revealing
rare outcrops of phyllosilicates [Poulet et al., 2005a] and
sulfates [Squyres et al., 2004; Arvidson et al., 2005; Gendrin
et al., 2005], Bibring et al. [2006] have created a preliminary,
alteration-based timescale of Mars. Phyllosilicates formed
during the ??Phyllosian?? and sulfates during the ??Theiikian,??
with a boundary between the two epochs at 
4 bya. The late
Hesperian to Amazonian (referred to as the ??Siderikian??) has
been dominated by slow weathering, producing anhydrous
ferric oxides. Apparently, the discovery of north polar gyp-
sum belies the general aqueous alteration history of Mars
inferred from the rest of the OMEGA observations. We will
show that geologic conditions unique to the north polar
region allowed an apparently anachronous formation of
gypsum. First, we describe what gypsum is and in what types
of environments it is commonly formed.
1.1. Gypsum and its Occurrence on Mars
[3] Gypsum (hydrated Ca sulfate, CaSO4.2H2O) is a
common sulfate-alteration/evaporation product on Earth
and is formed in a wide variety of depositional environ-
ments [e.g., Smoot and Lowenstein, 1991] which lead to its
occurrence in a wide variety of settings, for example, in
mine waste weathering [Sracek et al., 2004], evaporite
assemblages [e.g., Sinha and Raymahashay, 2004; Vysotskiy
et al., 2004], and hydrothermal deposits [e.g., Martinez-
Fr??as et al., 2004], among many others. This mineral can
range from translucent to opaque but is always light in
color. Because of its softness (1.5?2 on the Mohs scale),
gypsum is easily susceptible to physical weathering.
[4] In White Sands, New Mexico, USA, nearby evaporite
deposits have fed a collection of pure, white gypsum dunes
which cannot migrate far from their source due to the
JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 112, E07002, doi:10.1029/2006JE002862, 2007
Click
Here
for
Full
Article
1International Space Science Institute, Bern, Switzerland.
2Institut d?Astrophysique Spatiale, Universite? Paris-Sud XI, Orsay,
France.
3W. M. Keck Laboratory for Space and Planetary Simulations,
University of Arkansas, Fayetteville, Arkansas, USA.
Copyright 2007 by the American Geophysical Union.
0148-0227/07/2006JE002862$09.00
E07002 1 of 17
physical weakness of the mineral; gypsum sand grains
quickly break down into smaller grains which can no longer
saltate. The current area of the White Sands dune field is
about 710 km2, much smaller (by an order of magnitude)
than theMartian gypsum-rich dune area. However, theWhite
Sands dune field is much younger (
before 5840 year ago to
present) than the north polar erg, its source is being depleted,
parts of the dune field are still advancing [Langford, 2003],
Figure 1. (a) Map of the 1.94-mm absorption feature (diagnostic of gypsum) using the 1.93-mm
OMEGA band (modified from the work of Langevin et al. [2005]). White arrows indicate main near-
surface wind directions as mapped by Tsoar et al. [1979]. Meltwater channels are traced in black. Boxes
outline locations of other figures. (b) Portion of 64 pixel/degree MOC wide-angle image mosaic. The area
containing gypsum is outlined in orange. (c) Map of pyroxene as detected by OMEGA. The pyroxene is
identified using the band in the 1- to 1.3-mm range. The area containing gypsum is outlined in white.
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
2 of 17
E07002
and the size of a sand sea depends on local conditions such as
topography, wind speed, and sand supply.
[5] Gypsum also has a low solubility compared to other
sulfates (log(K) = 4.58 versus 1.88 for epsomite,
MgSO4.7H2O, or 2.19 for melanterite, Fe2+SO4.7H2O
[Tosca et al., 2005]) so any evaporating water containing
dissolved gypsum cannot carry it far before the gypsum
precipitates.
[6] On Mars, in addition to the north polar deposit, the
OMEGA team has tentatively identified gypsum as one
constituent in the layered deposits of Juventae Chasma
[Gendrin et al., 2005]. The wide range in elevations of
sulfate-containing outcrops in Valles Marineris has lead the
authors to suggest that the gypsum in Juventae Chasma did
not likely originate as an evaporite deposit laid down within
a basin, but rather as an alteration-evaporation product;
sulfur-bearing water interacted with Ca-bearing minerals,
in situ, leaving gypsum as the water evaporated. The MER
Opportunity rover has detected concentrations of up to
25 wt.% sulfate salts at Meridiani Planum, which are
overlying middle to late Noachian material [Reider et al.,
2004; Squyres et al., 2004] and which may have been created
by evaporation of fluids involved with weathering of basalts
[Tosca and McLennan, 2005]. The MER Spirit rover has also
detected sulfates within the Hesperian-aged Columbia Hills
outcrops, at concentrations of up to 40 wt.% [Ming et al.,
2004]. While Thomas et al. [1999] have proposed that some
bright dunes on Mars, which do not form regional dune seas,
may be composed of sulfate grains, the north polar sand sea
is the only concentration of dunes on Mars in which sulfates
have been definitively identified.
1.2. North Polar Dunes
[7] The north polar sand sea (Olympia Undae), the largest
on Mars, consists for the most part of transverse dunes
[Tsoar et al., 1979]. Transverse dunes advance perpendicular
to the direction of the dune crest by erosion on the windward
side and deposition on the leeward side, with one major
primary wind direction and other minor wind directions.
Where sand supply is relatively low, dunes take the shape of
barchans (a crescent-shaped subtype of transverse dune), and
where sand supply is higher, they merge into transverse
dunes. Longitudinal dunes, which advance along the direc-
tion of the dune crest, are also found in rare locations [Lee
and Thomas, 1995]. Tsoar et al. [1979], analyzing Viking
images, considered the north polar dunes to be in a young
stage of development and to be active based mainly on the
following four observations: dune slip faces appear to
reverse between images taken during different seasons;
horns extending from some barchan dunes extend on top
of ephemeral ice; a field of barchan dunes appears to be
migrating onto polar ice; and the dunes have a near perfect
form, something accomplished on Earth only by active
dunes. However, more recent observations indicate that
dunes all over Mars may be inactive. No movement of
Martian dunes in particular case studies has been observed
either between the Viking and Mars Global Surveyor (MGS)
missions or during the MGS mission [Edgett and Malin,
2000]. Furthermore, the rounded, elongated barchan dunes
in the Chasma Boreale region may be indurated, on the basis
of both modeling of dune induration and on morphological
similarity with terrestrial indurated dunes [Schatz et al.,
2006]. Whether or not the dunes within Olympia Undae,
in particular, are currently active has not been tested with
post-Viking data. Preliminarily, we would suggest that some
of the dunes have an eroded rather than fresh appearance
indicating inactivity and possible induration (Figure 2). On
the other hand, the apparently low thermal inertia of these
dunes compared to dunes elsewhere on Mars [Herkenhoff
and Vasavada, 1999] does not support induration. Regard-
less of whether or not they are active, the existence of only
one crater within the Olympia Undae dunes argues strongly
for a late Amazonian age [Tanaka et al., 2005] (0?0.45 By
old [Hartmann and Neukum, 2001]).
[8] These young dunes compose some of the darkest
materials on the planet [Thomas and Weitz, 1989] (TES
lambert broadband albedo of 
0.12 (J. Bandfield, personal
communication Infrared Thermal Mapper (IRTM)) and
Viking, 2007 albedo of 0.1?0.2 [Paige et al., 1994]). In Mars
Odyssey Thermal Emission Imaging System (THEMIS)
data, the dunes have a color similar to the polar basal unit
[Fishbaugh, 2005], which lies stratigraphically beneath the
polar layered deposits (PLD) and from which the dune-forming
material is believed to have eroded [Byrne and Murray,
2002; Edgett et al., 2003; Fishbaugh and Head, 2005].
[9] While a thorough compositional study of the Olympia
Undae dunes, in particular, has not yet been produced, the
polar sand seas display the best example of the MGS
Thermal Emission Spectrometer (TES) type II terrain
(J. Bandfield, review?s comments), which is generally
acknowledged to be a surface dominated by intermediate
plagioclase and high-Si glass with perhaps some pyroxene
(OPX or CPX) [Bandfield, 2002]. The plagioclase and high-
Si glasses are transparent in the NIR wavelength region
covered by OMEGA.
[10] Besides gypsum, OMEGA has preliminarily detected
low concentrations of oxides and/or olivine (difficult to
distinguish), areas with no clear spectral absorption features,
Figure 2. Example of dunes within Olympia Undae that
appear eroded and possibly indurated, rather than fresh.
Frost highlights features such as pitting on top of the dunes.
Portion of MOC image R15/00954. Arrow indicates
illumination direction. LS = 3.00. See Figure 1 for location.
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
3 of 17
E07002
and the presence of unidentified, hydrated minerals in most of
the area north of 60
?60
N [Poulet et al., 2005b]. Small
amounts of pyroxene, mostly outside the gypsum-rich area
(Figure 1c), and polyhydrated sulfates are also present, the
latter being confined to small, isolated dune patches to the
east of the main dune sea margin. While pyroxene and
hydrated sulfates have been detected within north polar dunes,
these minerals are not found in dunes elsewhere on Mars.
[11] Herkenhoff and Vasavada [1999] conjectured that the
Olympia Undae dunes may be composed of filamentary
sublimation residue (FSR) particles, ash and dust held
together in sand-sized aggregates formed during sublima-
tion, in this case, of the polar layered deposits (PLD). This
hypothesis was especially attractive when it was thought
that the source of dune material was the sediments within the
PLD [Thomas and Weitz, 1989] rather than within the polar
basal unit; if particulates within the PLD originated as
airborne material, then sand-sized grains could not be erod-
ing from these deposits to form dunes. However, the polar
basal unit, discovered only with the advent of MGS Mars
Orbiter Camera (MOC) data, is thought to be composed, at
least partially, of sand-sized particles so that invoking FSRs
is no longer necessary to explain the presence of the dunes.
Nevertheless, the apparently low thermal inertia has yet to be
explained unless it is due to previously unconsidered factors
such as the effects of slope on apparent temperature. It is
within this unique sand sea that the OMEGA instrument has
discovered the largest gypsum deposit on Mars.
1.3. Previous Studies Concerning North Polar Gypsum
[12] Langevin et al. [2005] have used OMEGA data to
identify gypsum within Olympia Undae mainly by its
absorption feature at 1.94 mm, which is associated with
several features typical of gypsum in the 1- to 2.5-mm range.
By taking a ratio of a spectrum from gypsum-rich dunes to a
spectrum of gypsum-poor dunes, the authors emphasized
that the resultant absorption features much more closely
match those of pure laboratory gypsum than those of other,
less-hydrated calcium sulfates, such as bassanite (Figure 3).
We discuss the concentrations and distribution of this
gypsum deposit in section 2.
[13] While not discussed in detail, Langevin et al. [2005]
favored the following two possible origins for this deposit,
acting either alone or in combination: interaction of Ca-rich
minerals with H2SO4-bearing snow resulting from extended
volcanic activity nearby or formation as an evaporite
deposit within a pool produced by outflow from the PLD
(either the current or any previous PLD). As pointed out by
the authors themselves, a major difficulty with the first
hypothesis is the small area to which the gypsum is
confined. One would expect larger-scale alteration of more
circumpolar materials (with the same Ca-rich composition)
unless this acidic snow somehow fell only in a confined
area.
[14] More recently, in a preliminary study, Tanaka [2006]
has suggested that the north polar gypsum deposit resulted
from circulation of hydrothermal groundwater associated
with magmatism at Alba Patera during the early Amazonian
(
1.75?3.05 bya) [Hartmann and Neukum, 2001]; later,
eolian activity excavated this deposit. Expulsion of ground-
water from graben elsewhere on Mars, in the Elysium and
Tharsis regions, has formed such features as major channels
and lahar deposits (which are not observed in the north polar
region) but no exposed sulfate salt deposits [Christansen,
Figure 3. Comparison of the spectral signature of the gypsum-rich region with spectral libraries. The
black spectrum is a spectral ratio at LS 136.2
 between the region at 240.2
E, 79.9
N which presents the
largest hydration signature at 1.94 mm and a region at 244
E, 77
N where this signature is much weaker.
The squares correspond to the spectral sampling. The absorption features are strongly diagnostic of
calcium sulfates. The blue spectrum corresponds to bassanite (CaSO4, 0.5 H2O), and the red spectrum
corresponds to gypsum (CaS04, 2H2O) from the USGS spectral library [Clark et al., 1993]. The relative
reflectance scale for the Mars spectra (right) has been expanded with respect to that of the laboratory
spectra (left). The position and fine structure of the 1.445-mm absorption, the position of the 1.75-mm
absorption, the position and shape of the 1.945-mm absorption, the presence of an absorption at 2.21 mm,
and the sharp decrease in reflectance at 2.4 mm provide a highly consistent and unambiguous
identification of gypsum with respect to less-hydrated forms of calcium sulfates.
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
4 of 17
E07002
1989; Tanaka and McKinnon, 1989; Skinner and Tanaka,
2001; Russell and Head, 2003]. Timing could also present a
problem in this scenario. Either the gypsumwas created before
the dunes, has been exposed by wind erosion only at the
locations where dunes now exist (i.e., nowhere else in the
northern plains), and is now incorporated within the dunes or
the dunes are actually early Amazonian in age, belying their
crater-counting age.
1.4. Goals of This Study
[15] The north polar gypsum deposit is unique on Mars in
comparison to other sulfate deposits in terms of albedo
(relatively dark), spectral signatures (deep versus shallow),
morphology (dunes versus outcrops), and exposed area
(much larger). The main goal of this study is to investigate
the origin of the north polar gypsum using mainly MGS
MOC images, Mars Odyssey THEMIS data, MGS Mars
Orbiter Laser Altimeter (MOLA) data, and OMEGA data.
The achievement of this goal involves characterization of
the geologic setting of the gypsum deposit, describing the
relationship of the gypsum with the dunes in which it is
found, tentatively identifying the source of Ca2+ and SO4
2
needed to create the gypsum, and finally building a descrip-
tion of the geologic sequence of events which lead to its
formation. While we concentrate on the OMEGA observa-
tions of this gypsum deposit and how this unique deposit
fits into the local geology, we also perform preliminary
geochemical analysis.
2. Observations in the North Polar Region
2.1. Gypsum Concentrations
[16] In Figure 1a, we map the strong 1.94-mm gypsum
absorption feature using the 1.93-mm OMEGA channel in
order to reduce contamination by atmospheric CO2; exclud-
ing the grain-size effect (explained below), it is essentially a
map of gypsum concentration. As discussed in section 1.3,
this gypsum has been identified by spectral ratios so that,
while a strong absorption feature at 1.94 mm is also
diagnostic of other sulfates and clays, in this region, it is
diagnostic of gypsum (Figure 3). The reflectance at the
bottom of the gypsum band, at locations of maximum
concentration (red in Figure 1a, centered around 160
W),
is nearly 25% lower than the continuum, half the value for
pure, laboratory gypsum powder, grading to 
5% band
strength at the low concentrations toward the west (purple
in Figure 1a). The detection limit of the OMEGA instru-
ment, in terms of abundance, depends on grain size as
follows: 
15 vol.% for grain size 5 mm and 5 vol.%
for grain sizes 10 mm [Poulet et al., 2007]. This means
that OMEGA can easily detect small amounts of sand-sized
(hundreds of microns) gypsum but cannot so easily detect
small amounts of fine-grained gypsum (a few microns).
[17] The extraction of quantitative information about
composition and physical state (for example, the volume
percentage of gypsum) from spectra requires accurate
models of radiative transfer within the surface materials.
In this work, we use the approach adopted by Shkuratov
et al. [1999]. This approach, on the basis of approximation
of geometrical optics (as is the more familiar Hapke model
[Hapke, 1981]), is used to transform optical constants into
a reference spectrum and provide the spectral albedo of
powdered surfaces. The compositional model for this
study must satisfy the following two major constraints:
the observed low albedo of the gypsum-containing dunes
and the observed gypsum band depths. We use the gypsum
optical constants derived by Roush et al. [2006]. Dark and
spectrally featureless (DSF) grains act as darkening agents
[Poulet et al., 2003; Poulet and Erard, 2004].
[18] For an intimate (??sand??) mixture wherein gypsum
sand grains are mixed with DSF grains, the best fit to our
compositional model (Figure 4a) is as follows: 55% DSF
mixed with 45% gypsum. We use a grain size of 100 mm?
1 mm for the gypsum component, and a few tens of
microns for the DSF component. Note that the grain size
of the DSF (and thus the abundance) is not well con-
strained because there is no absorption feature. Increasing
the grain size will decrease the abundance of the DSF
phase so that the 55% reported above corresponds to an
upper limit. Estimates of average Martian dune grain size
range from 500 ? 100 mm [Edgett and Christensen,
1991] to several millimeters [Christensen et al., 2003].
Edgett and Christensen [1991] calculated a minimum size
for saltation of 210 mm; hence most of the gypsum grains
within the range of sizes we use would be theoretically able to
saltate. However, an intimate mixture is inconsistent with
TES thermal infrared data which do not indicate a presence of
large amounts of coarse-grained oxides.
[19] In addition to common intimate mixtures, we can
test an ??intramixture?? model in which larger grains house
smaller inclusions [Poulet and Erard, 2004]. If, during its
precipitation, the gypsum incorporated grains of darker
material, such as DSF, drastic effects both on albedo and
band depth would ensue. In this case (Figure 4b), we find
that 35% pure gypsum grains of a few tens of microns in
size mixed with 65% millimeter-sized gypsum grains,
contaminated by small inclusions of DSF, best fits the
OMEGA observations. These grain sizes are averages from
a distribution of possible sizes. The millimeter size is
required to fit the deep absorption features of gypsum,
while the smaller size is mostly necessary to increase the
overall albedo. Within the saltating intramixture, there
should be almost no albedo contrast since the dark inclu-
sions considerably lower the apparent albedo of the sur-
rounding gypsum matrix. Indeed, an intramixture provides
a better fit to our compositional model than does an
intimate mixture (Figure 4b). This type of mixture presents
several major additional advantages as follows: it is con-
sistent with the presence of a large component of spectrally
featureless material, inferred by OMEGA observations to
exist within the dune sea; oxides, which are dark and
spectrally featureless, are a significant secondary phase
expected from the weathering of Fe-bearing high-calcium
pyroxene (HCP) during the formation of gypsum (as
explained in section 3.2.2); and the intramixture model
appears to be more consistent with TES thermal infrared
spectra, because the derived abundance of oxides is under
the detection limit of TES (TES has not observed oxides in
this dune sea).
2.2. Gypsum Distribution
[20] The area of highest gypsum concentrations lies at
the eastern edge of the main north polar dune sea,
covering an area of 
7500 km2 (Figure 1a). This area
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
5 of 17
E07002
of the sand sea is geologically different from the western-
most portions because small, sinuous valleys terminate just
to the southeast (outlined in Figure 1a). Gypsum concen-
tration decreases toward the west, in the same direction as
the main near-surface winds as mapped by Tsoar et al.
[1979] using dune morphology. Beyond 
210
W, little to
no gypsum appears, though the main dune sea extends to

240
W.
[21] At first glance, the gypsum in Figure 1 appears to be
confined to the main dune field (low-albedo area labeled in
Figure 1b). Indeed, this observation is borne-out within
higher-resolution MOC images. Other than small hills
within the dune sea and the extreme eastern edge of the
dune sea where dune coverage is more sparse, we find
almost no areas where gypsum exists without dunes. Even
the small gypsum patches outside of the main dune field, to
the east of the high-concentration area, lie within dunes
(purple patches in Figure 1a, Figure 5). No gypsum has
been detected within the polar cap or circumpolar materials.
2.2.1. Gypsum Within the Dunes
[22] Further investigation of the relationship between the
Olympia Undae dunes and the gypsum reveals that there are
Figure 4. Comparison of OMEGA spectrum from the gypsum-rich dune region (solid line) with (a) a
spectrum generated by our intimate mixture model (dashed line) and with (b) a spectrum generated by our
intramixture model (dashed line). The intramixture model proves a better fit to the OMEGA data.
Figure 5. (a) Example of a small patch of gypsum lying
within dunes outside of the main north polar dunes field
(portion of MOC image R14/00210, arrow indicates illumina-
tion direction). (b) Location of MOC image with respect to the
gypsum concentration map. See Figure 1 for context.
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
6 of 17
E07002
few anomalous characteristics of the gypsum-bearing dunes
in comparison to those in which gypsum has not been
detected. In MOC images, we find no correlation of gypsum
concentration with dune morphology. Both gypsum-rich
and gypsum-free areas contain a variety of barchan and
transverse dune morphologies (Figure 6). This is not
entirely unexpected as dune morphology on Earth generally
has no correlation with sand composition. The gypsum
dunes of White Sands, New Mexico comprise the following
four major dune types whose morphologies are related to
sand supply and wind direction, not to their gypsum
content: parabolic dunes, barchan dunes, transverse dunes,
and dome-shaped dunes [e.g., McKee, 1966]. Additionally,
the area of highest gypsum concentration has only a slightly
higher albedo on average (0.16) than the dunes that lack
gypsum (0.14), and no apparent color anomalies manifest
themselves in THEMIS images (Figure 7). THEMIS IR data
indicate no temperature anomalies in the gypsum area either
(Figure 8).
[23] In MOC images taken during northern summer, we
have found instances of bright patches of material, which
have collected mostly in low areas between dune crests and
are scattered throughout the area covered by gypsum-
bearing dunes, though not ubiquitously (Figure 9). These
patches have not yet been found (in MOC images) within
the western portion of the dune sea in which gypsum has
not been detected. However, we have observed bright
patches in a Mars Reconnaissance Orbiter (MRO) High
Resolution Imaging Science Experiment (HiRISE) image
within the dune field bordering the mesa near the mouth of
Chasma Boreale. No gypsum has been detected here by
OMEGA, but perhaps the grain size is too fine and the
amounts too low (in section 3.4.2, we discuss gypsum in
association with Chasma Boreale). The HiRISE image
Figure 6. Examples of dunes with different gypsum concentrations. (a) Example of complex linear
dunes within the highest gypsum concentrations. Note that all examples were taken during northern
winter and therefore may contain frost. Portion of MOC image R13/03611 (LS = 338.14
). (b) Example
of complex transverse dunes within the highest gypsum concentrations. Portion of MOC image R14/
00532 (LS = 344.68
). (c) Example of complex transverse dunes which contain no gypsum. Portion of
MOC image R14/00706 (LS = 345.28). Arrows indicate illumination direction. Scale bars are 1 km. See
Figure 1 for locations. There is apparently no correlation of gypsum concentration with dune
morphology. Additionally, there appears to be no surficial crusts or ubiquitous loose deposits of bright
gypsum on the dunes with high gypsum concentrations.
Figure 7. The gypsum-rich dunes exhibit no color anomalies in relation to the low-gypsum dunes. The
gypsum-free area is reddish relative to the bluish dunes, likely due to the presence of dust. THEMIS false
color visible image V04067010. Blue = Band 1 (0.425 mm), Green = B2 (0.540 mm), and Red = B3
(0.654 mm). LS = 94.885. Arrow indicates illumination direction. See Figure 1 for image location.
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
7 of 17
E07002
(Figure 9d) reveals eolian ripples on the surfaces of the
patches; this, along with the summer season in which this
image and the MOC images were taken, make it unlikely
that the patches are composed of frost. It may also be
possible that this bright patch observed by HiRISE is not the
same type of bright patch observed by MOC within the
main polar dune sea. Note that these patches are different
from a few exposures between dunes (in the eastern,
gypsum-free portion of the dune sea) of bright substrate,
which is often characterized by polygons.
[24] While OMEGA cannot resolve the bright patches,
the MRO Compact Reconnaissance Imaging Spectrometer
for Mars (CRISM) can. While no data or publications have
yet been released concerning these bright patches, a recent
press release concerning one CRISM image (13 December
2006, http://www.nasa.gov/mission_pages/MRO/multimedia/
pia09094.html) indicates that the bright patches are relatively
low in gypsum compared to the dunes themselves and that the
crests of the dunes exhibit the highest gypsum concentrations.
It is yet to be seen whether or not these observations are
representative of the entire gypsum-rich area, but we suspect
that grain-size effects similar to those of OMEGA data play a
role in the relatively low gypsum concentration of the patches;
the possibly small grain sizes within these patches may make
detection of gypsum more difficult. Additionally, underlying
material may outcrop within the bright patches, creating a
mixed spectral signature.
[25] Figure 10 displays the gradual transition from high
gypsum to lower values to the surrounding gypsum-free
plains in the context of a THEMIS IR image. Again, the
gypsum-rich areas lie almost exclusively within dune-
covered regions. Transitions from high- to low-gypsum
Figure 8. (a) Derived brightness temperature for Band 9 of THEMIS infrared image I04067009 (LS =
94.885; local time = 3.382; incidence angle = 71.256
). Projected temperature data courtesy THEMIS
Processing Web Interface (THMPROC): http://thmproc.mars.asu.edu. Arrow indicates illumination
direction. See Figure 1 for location. While brightness temperature is not equal to the physical
temperature, it is a reasonable approximation since most natural surfaces have an emissivity close to 1 at
IR wavelengths. In this example, no significant temperature anomalies are apparent within the gypsum-
rich dunes. (b) Portion of the gypsum concentration map shown in Figure 1 with the main features of the
THEMIS image in Figure 8a outlined for reference.
Figure 9. Examples of dunes with associated bright patches which may be gypsum. Both images were
taken during northern summer. (a) Portion of MOC image M01/05259, LS = 147.21
. Black arrows point
out examples of the bright patches. White arrow indicates illumination direction. (b) Portion of MOC
image M02/04467, LS = 161.91
. Arrow indicates illumination direction. (c) Portion of MOC image E01/
00916, LS = 117.32
. Arrow indicates illumination direction. See for Figure 1 for image locations for
Figures 9a, 9b, and 9c. (d) Portion of HiRISE image TRA_000861_2580_RED, illuminated from the
lower left, LS = 114.9
. This image lies within the dune field bordering the mesa beyond the mouth of
Chasma Boreale (see Figure 14).
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
8 of 17
E07002
concentrations and between gypsum-containing and gyp-
sum-free areas do not express themselves in the regional
morphology. Put another way, regional morphologic varia-
tions do not appear to have affected gypsum concentration.
Note that small amounts of gypsum appear to be associated
with part of a hill in the middle of Figure 10. Given that the
appearance of any gypsum anywhere other than within dunes
is a quite rare occurrence, we speculate that the gypsum,
apparently associated with this hill, is due to the imperfect
registration of OMEGA data to MOLA data. Note that most
of the hill is entirely gypsum free.
2.2.2. Gypsum in the Basal Unit?
[26] Several authors have identified the north polar basal
unit, lying stratigraphically beneath the polar cap, as the
main, if not sole, source for the north polar sand sea [Byrne
and Murray, 2002; Edgett et al., 2003; Fishbaugh and
Head, 2005]. High-resolution OMEGA data (at 1-km pixel)
reveal a gap between areas containing high gypsum con-
Figure 10. Illustration of the regional, morphological context of the gradual transition from high
gypsum concentrations to low concentrations to surrounding gypsum-free plains. (a) Portion of the
gypsum map from Figure 1a overlying THEMIS IR image in Figure 10b. Note that OMEGA data and
THEMIS data have not been perfectly coregistered. (b) Portion of THEMIS IR image I03750002. Arrow
indicates illumination direction. See Figure 1 for location. The changes from high gypsum concentrations
to lower concentrations do not appear to be controlled by (i.e., do not mimic) changes in regional
morphology.
Figure 11. Illustration of the lack of gypsum in the north polar basal unit (main source of the sand
sea). (a) Portion of normal-resolution gypsum map in Figure 1 as context. (b) High-resolution OMEGA
data (1 km/pixel), spanning the polar cap margin, a gap in gypsum, and part of the highest gypsum
concentrations. Grey area indicates ??no data.?? Rectangle shows location of full-MOC image on the
right. No MOC images were available exactly in the location covered by the high-resolution OMEGA
data. (c) Portion of MOC image E02/01976 showing that the gap in gypsum corresponds with the presence
of the dark, layered north polar basal unit. Arrow represents illumination direction. See Figure 1 for
MOC image location.
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
9 of 17
E07002
centration and the PLD (Figure 11). This gap is occupied by
a dark, layered material previously identified as part of the
basal unit. Thus it appears that the basal unit, while being
the source for the majority of the dune material, is probably
not the source for the gypsum within the dunes. This
observation can be tested further after the collection and
better georeferencing of more high-resolution OMEGA data
as well as dedicated observations by CRISM, which has a
spatial resolution of 18 m/pixel.
3. Discussion
3.1. Association With Sand Dunes
[27] Several observations indicate that the bulk of the
observed gypsum consists of sand-sized grains (with dark
inclusions) which are incorporated into dunes rather than
forming surficial crusts or fine-grained deposits. (1) No
significant, ubiquitous color, thermal, or average albedo
anomalies are associated with the gypsum-bearing dunes.
(2) Gypsum has been detected almost exclusively within
dunes. (3) Gypsum concentration decreases westward, in
the same direction as the main near-surface winds (see
Figure 1). It appears that the decreasing amount of gypsum
in this direction is partly due to the fact that most of the
gypsum-containing dunes haven?t migrated that far.
[28] There are several reasons why gypsum would be
detected within sand dunes. (1) The existence of a collection
of dunes indicates that deposition of wind-blown material is
taking or has taken place there. In areas without dunes or
other sediment deposits, wind-blown material is either being
net removed or just not being deposited. Gypsum would be
expected to accumulate in an area of net deposition (for
example, a dune sea) rather than in an area of net erosion or
nondeposition. (2) The saltation process would clear the
surface of fines, removing any dust which could mask a
gypsum signature. Of course, this is operating in the polar
dune sea only if the dunes are active. (3) At the wavelengths
used to detect gypsum, OMEGA is most sensitive to sand-
grain-sized particles. Small amounts (
15% maximum) of
undetectable, finer-grained gypsum may exist elsewhere.
(4) As discussed below, formation of gypsum would most
likely occur where Ca-bearing sediment (the dunes; see
section 3.2.2) is present and where a large exposed surface
area makes alteration to gypsum easier (within collections
of loose sand grains).
[29] Given the difference in hardness of the gypsum and
DSF components, a physical segregation between grains
mostly composed of gypsum and of mostly DSF and other
components should occur due to the saltation process. The
farther the dunes migrate, and hence the farther the indi-
vidual sand grains saltate, the more the soft gypsum would
break down. Erosion during saltation would reduce the size
of the gypsum particles until they become too small to
saltate. Over time and saltation distance, more and more
gypsum would end up as fine-grained deposits blown by the
wind and collecting in interdune spaces and the leeward side
of dunes but would not be mixed-in with the saltating dark
sand. If the dunes eventually become inactive, then fine-
grained gypsum could also collect on dune crests. The
eolian-rippled, bright patches (that are not exposures of
polygonally-patterned underlying material) observed in
MOC images scattered throughout the gypsum-rich dunes
may comprise the fine-grained (tens of microns), pure
gypsum component of our intramixture and the gypsum
which has been mechanically weathered during saltation.
Since these bright patches do not markedly increase in
spatial density with distance from the highest gypsum
concentrations, the decrease in gypsum concentration to-
ward the west probably has more to do with the gypsum not
having yet saltated that far than with a decrease in grain size
lessening the gypsum signature.
3.2. Source of the Ingredients Necessary to Form
Gypsum (Ca2+, SO4
2, and H2O)
[30] The following discussion concerning the geochem-
istry needed to form the observed gypsum is based upon the
available information to date about the composition of the
north polar dune sea and the polar layered deposits. Studies
dedicated to thorough, detailed compositional mapping of
the entirety of this particular dune sea have not yet been
performed, making it extremely difficult to build a compre-
hensive chemical/thermodynamical model concerning the
formation pathways of gypsum. This discussion is thus, at
this point, somewhat speculative.
[31] Throughout this section, we assume that the water,
which has emanated from the polar deposits, is pure because
the exact composition of the polar deposits is unknown and
because Mars Express Mars Advanced Radar for Subsur-
face and Ionosphere Sounding (MARSIS) data imply a
nearly 100% water composition [Plaut et al., 2006]. Note
that if any dust (which likely contains sulfates [Hamilton et
al., 2005]) was present in the water, it would not have
increased the acidity of the water. Regardless, sulfate has not
been detected by OMEGA within the polar deposits, indi-
cating that it is likely present in only minor amounts if at all.
3.2.1. Sulfur
[32] Generally, the sulfate deposits observed on Mars are
attributed to the following two possible origins: alteration
in the presence of SO2 [Fairen et al., 2004; Tosca et al.,
2004] and alteration in the presence of sulfides (like
pyrrhotite) [Burns and Fisher, 1990b; Burns, 1993]. The
first origin implies dissolution in water of SO2 produced by
volcanic activity [Postawko and Kuhn, 1986; Waenke et al.,
1992]. The resulting strong acid (H2SO4) can easily alter
surrounding silicates. However, the source of SO2 must be
relatively close to the sulfate deposit since the high reac-
tivity of H2SO4 implies fast kinetics of alteration and
sulfate precipitation.
[33] The second origin implies the concomitant weather-
ing of silicates and sulfides in the presence of water, in
oxidative conditions. Previous experiments have demon-
strated that alteration of diopside and pyrrhotite in a
simulated Martian atmosphere produces mostly Fe sulfates
but also produces gypsum in small amounts; as soon as even
trace amounts of calcium come in contact with sulfur,
gypsum precipitates [Chevrier et al., 2004, 2006]. This
hypothesis presents an advantage in simplicity in that a
nearby volcanic source is not required; only sulfide, water,
and calcium are required. Similarity of some Martian
meteorites and primary rocks with terrestrial komatiites
has long suggested the possible presence of abundant
sulfides in the Martian crust [Clark and Baird, 1979; Burns
and Fisher, 1990a]. In addition, sulfides are two to three
times more abundant in SNC shergottites than in their
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
10 of 17
E07002
terrestrial equivalents [Lorand et al., 2005]. The evolution
of these sulfides in oxidative conditions is hypothesized to
have led to sulfate deposits, such as jarosite or gypsum, at
various places on Mars [Burns and Fisher, 1990b, 1993].
The presence of sulfide in the Olympia Undae dunes would
provide a ready source of local sulfur without needing to
invoke late Amazonian volcanism.
[34] It should be noted that, while the ubiquitous, Martian
dust does contain sulfur (and hence could be considered as a
sulfur source), that sulfur is inferred, for the most part, to
already be in the form of sulfates [Hamilton et al., 2005].
Therefore dust could not be the source of the sulfur for a
gypsum deposit containing no other sulfates unless all of the
sulfate within the dust were gypsum, which is not thought to
be the case. Additionally, since sulfates are stable on the
surface of Mars, there is no particular reason to assume that
the dust would destabilize to form a gypsum deposit.
3.2.2. Calcium
[35] Identifying possible sources of calcium is a bit
trickier. Considering the primary mineralogy identified on
Mars, the main potential source of Ca2+ is clinopyroxene
(high-calcium pyroxene (HCP), Ca(Mg, Fe)Si2O6) which is
widely distributed on the surface of Mars (according to
OMEGA observations) [Bibring et al., 2005; Mustard et al.,
2005]. HCP is present in minor amounts at various locations
within the north polar region, including near and within the
sand sea (Figure 1c). An important issue in considering
HCP as a calcium source is the behavior of primary iron and
magnesium within the HCP. The Fe3+ ion forms easily from
Fe2+ present in primary phases and is extremely insoluble
under high-oxidation levels and moderate acidity (pH 

3). In the case of this gypsum deposit, the necessary
presence of SO4
2 in the water indicates high-oxidation
levels, but the pH of the water is unknown; we will assume,
for the purposes of this discussion, that the pH was above 3.
If the primary pyroxene contained iron, iron oxides (any
evolved crystalline product from the precipitation of ferri-
hydrite 
Fe(OH)3, which includes hematite, Fe2O3, goe-
thite, FeO(OH), and magnetite, Fe3O4) or even Fe-bearing
sulfates (jarosite) should be present as secondary phases
[Sracek et al., 2004]. OMEGA observations infer the
presence of a dark, spectrally featureless (DSF) phase mixed
with the gypsum, a description which can be attributed to an
oxide. Using inclusions of oxide as darkening agents in the
gypsum grains not only provides an excellent fit to our
compositional model, as explained in section 2.1, but is also
consistent with alteration of Fe-rich HCP.
[36] The case for Mg2+ is different. Alteration of Mg-
bearing HCP forms rather soluble salts, such as kieserite,
also observed on Mars within Valles Marineris [Gendrin et
al., 2005]. Thus, if the primary HCP contained magnesium
but not iron, it is more likely that spatial fractionation of
minerals would occur and that gypsum would precipitate
first (closest to the water source) while more soluble Mg-
rich salts would be carried further by the water. In the case
of in situ alteration, gypsum would precipitate first and the
more soluble salts later, on top of or near the gypsum, but
not mixed with it. Nevertheless, no Mg sulfate deposit has
been observed close to the gypsum dunes.
[37] Alteration of wollastonite (CaSiO3) or anorthite
(CaAl2Si2O8), instead of HCP, can form gypsum alone with
no associated secondary phases (except for silica, SiO2).
Considering the high content of magnesium and iron on the
Martian surface, alteration of wollastonite or anorthite
remains less likely than alteration of HCP. Yet, the obser-
vation of quartzofeldspathic minerals in the northern plains
[Banfield et al., 2004] suggests that evolved materials
enriched in alkaline elements are present on Mars, even if
apparently disconnected (at the surface at least) from the
northern gypsum deposits. Andesine-labradorite is present
within TES surface types I and II [Bandfield et al., 2000]
and could also serve as a calcium source. However, if
the gypsum within the dunes was created in situ (see
section 3.3), then the aluminum must have also remained
within the dunes, forming secondary phases which have
thus far not been observed. Note that plagioclase itself
cannot be detected by OMEGA.
[38] Of these possibilities, the most likely source of
calcium appears to be Fe-bearing HCP. HCP has been
identified within and near the Olympia Undae dune field,
as well as in many other locations on Mars, and the oxides
which would result as secondary phases might well be the
dark inclusions we suggest exist within most of the gypsum
grains.
3.2.3. Water
[39] Since gypsum is a soft mineral, it cannot saltate long
distances. Thus one would expect the gypsum source region
to lie close to or within the highest gypsum concentrations.
As shown in the map in Figure 1 and in the close-up in
Figure 12, sinuous valleys extending from the polar cap
terminate just to the east of the high-gypsum area. We
hypothesize that the most likely origin of these valleys is
fluvial erosion resulting from the melting and outflow
event which most likely initiated the formation of Chasma
Boreale to the east [Fishbaugh and Head, 2002]. An
alternate or additional water source could be melting due
to the formation of the nearby impact crater in the PLD.
Note in Figure 12a that small amounts of gypsum (associ-
ated with dunes) surround the region where the channels
terminate. Currently, younger, higher-albedo material (likely
ice/frost and dust) covers this area (Figure 12c), masking the
signature of any gypsum potentially existing within it. The
proximity of these channels and the highest concentrations
of gypsum implies a causal relationship.
[40] Having identified the chemistry needed to produce
the gypsum within the limits of what is currently known
about the composition of the polar dune field, the question
now becomes, ??What was the depositional environment of
the gypsum???.
3.3. Possible Depositional Environments
[41] Evaporite deposits form when salt minerals precipi-
tate out of evaporating saline water. This process can occur
in several main types of depositional environments: lacus-
trine, fluvial, eolian, springs, and saline soils [Smoot and
Lowenstein, 1991]. While evaporative processes may often
involve leaching of Ca by acidic water and transport of
dissolved salts to another location where the evaporation
occurs, this situation is unlikely for the Martian north polar
deposits. Liquid may not be able to remain stable at the
surface long enough to allow distant transport.
3.3.1. Lacustrine
[42] The largest terrestrial gypsum deposits typically
form in closed basins. As different minerals precipitate
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
11 of 17
E07002
out of standing water enclosed in the basin, the water
chemistry changes, creating a sequence of mineral forma-
tion wherein the most soluble minerals precipitate last.
Classically, this can translate into the development of
different layers of salt, with the bottom layer consisting
of the least-soluble mineral. Usually, carbonates will pre-
cipitate before gypsum, but Catling [1999] has theorized
that gypsum may form early, if not first, in an evaporite
sequence on Mars if SO4
2/Ca2+ > 2 and if the atmospheric
partial pressure of CO2 is greater than 3 bar. Such a
situation could possibly account for the almost complete
lack of other salts in association with the polar gypsum
deposit. However, we cannot know what the particular
value of SO4
2/Ca2+ is in the case of the Martian north
polar gypsum, and the associated atmospheric pressure is
500 greater than the current pressure. Additionally, there
is no evidence for a basin outside the current margin of the
polar cap unless it has since been modified by deposition or
erosion or it has since been covered by southward extension
of the PLD margin.
[43] On the other hand, saline mudflats, a lacustrine
subenvironment, do not necessarily require an obvious
basin [Smoot and Lowenstein, 1991]. These mudflats con-
sist of fine-grained sediment which comes into contact,
perhaps intermittently, with saline water. In this environ-
ment, salt minerals grow within the sediment by evapora-
tion, forming layers of salt, aggregates of crystals, or
individual crystals which can incorporate grains of other
material. Commonly, saline mudflats are covered in crusts
of relatively insoluble minerals, like gypsum. Salts may also
form in dry mudflats [Smoot and Lowenstein, 1991] that
come into contact with even less water, but the quantity
produced is even smaller than in saline mudflats. Perhaps
the area surrounding the channel mouths served as such a
mudflat. The largest obstacle to this scenario is the fact that
all of the observed gypsum is now incorporated into dunes.
So where is the original evaporite deposit, unless it is
undetectable due to the dust/ice/frost covering the channel
mouth region (as stated in section 3.2.3 and seen in
Figure 12)?
[44] We find a lacustrine origin to be unsatisfactory
because it is difficult to either prove or disprove. We can
find no evidence of an evaporite basin or of a gypsum
source deposit, even if there are various scenarios that could
explain this lack of evidence.
3.3.2. Fluvial
[45] In terrestrial fluvial environments with the appropri-
ate geochemistry, one may find evaporites at the toe of an
alluvial fan, in abandoned stream meanders, and in flood-
plains (which are really a saline soil environment, see
section 3.3.5) [Smoot and Lowenstein, 1991]. In all cases
but the floodplain, the amount of evaporitic material pro-
duced is small and not always euhedral (i.e., does not
always have a fully developed crystal form), often forming
crusts; such a scenario is not conducive to eventual incor-
poration into dunes, nor is it consistent with the large
amount of gypsum observed in the Martian polar dune sea.
3.3.3. Eolian
[46] Evaporites may also form in terrestrial interdune
spaces that intermittently host saline groundwater or act as
playas [Smoot and Lowenstein, 1991]. In this type of
environment, the evaporites will form crusts or even crystals
which sometimes grow around sand grains. If such a
processes had taken place within the Olympia Undae dune
sea, then we would expect almost the entirety of the
observed gypsum to exist within the interdune spaces,
making these areas brighter than the dunes themselves. As
explained in section 2.2.1, bright patches do exist in
interdune areas, but they are not ubiquitous, and are small
in total area compared to the dunes themselves.
3.3.4. Springs
[47] Saline springs may also produce evaporites. We will
not discuss this possibility in detail here since we have no
particular reason to call upon groundwater being brought to
Figure 12. (a) Close-up of the gypsum source region. See Figure 1 for context. (b) MOLA-shaded relief
of same area. (c) Portion of MOC wide-angle mosaic. Valleys visible in the MOLA data are outlined in
Figure 12a. We interpret the valleys as channels formed by outflow of water either from the Chasma
Boreale melting event [Fishbaugh and Head, 2002] or by melting due to the nearby impact into the polar
cap. Note that the area around the channels has a slightly higher albedo in the MOC mosaic resulting
from a cover of frost/ice and dust.
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
12 of 17
E07002
the surface by hydrostatic pressure when we have already
identified a more obvious water conduit, the channels.
Invoking spring activity adds unnecessary complexity to
the story. From where did the groundwater come, and why
was this putative groundwater under hydrostatic pressure?
3.3.5. Saline Soils
[48] If saline water percolates downward through soil,
evaporites could develop within the pore spaces, with
gypsum and halite being the most common such evaporites
in terrestrial soils [Smoot and Lowenstein, 1991]. Euhedral
crystals can form in this environment and may even
poikilitically trap soil grains within them.
[49] We can translate this environment to the Olympia
Undae dune field on Mars. Water emptied from the channels
to the east. Some of this water entered the eastern margin of
the polar dune sea nearby. Since the dunes themselves
contained Ca-rich pyroxene and sulfide, the dune grains
were altered into gypsum as water encasing the grains
evaporated; additionally, evaporative gypsum crystals were
created in the pore spaces as the water percolated through
the dunes. Whether or not direct alteration of the materials
just beyond the channel mouths occurred is unknown since
younger materials cover this area. In any case, alteration of
loose sand grains would take place more easily than
alteration of other deposits, given the large surface area
involved. In both direct alteration of basaltic sand and in
formation of evaporative gypsum crystals, dark basaltic
grains could easily end up encased within the gypsum
grains, lowering their albedo. As explained in section 3.2,
dark oxide inclusions, produced as secondary phases, would
also darken the grains and are implied to exist by our
compositional modeling. In situ alteration of dune material
and creation of evaporative gypsum crystals by percolation
of water delivered from nearby channels through the dunes
provide a simple explanation of the creation of this unique
gypsum deposit in an environment unique to this location.
Water, sulfur, calcium, and easily alterable sand grains are
provided locally, without the need to bring materials from
outside the region of the north PLD, without the need to
invoke large-scale, distant volcanic activity to provide
hydrothermal circulation or sulfur, and without requiring
an explanation for the lack of evidence for an evaporite
basin or for a primary evaporite deposit. However, while
this hypothesis best follows Occam?s Razor, several impor-
tant issues remain.
3.4. Further Analysis and Open Questions
3.4.1. Amounts and Rates
[50] Given the many uncertainties involved, it is extremely
difficult to calculate how much calcium, sulfur, and water
were required to form the amount of gypsum observed, how
long the weathering process took, or even to estimate errors in
any such calculations. However, here, we attempt first-order
estimates of these values. We first make a rough estimate of
the volume of dune material within the high-gypsum area by
breaking up this area into rectangles to calculate the surface
area (
7500 km2), as shown in Figure 13a. We then assume
that the typical elevation variations of 10 m, as observed in a
profile of MOLA gridded data across the area (Figure 13b),
represent the thickness of dune deposits and that the regional
slope along the profile is due to subdune topography (for
example, hidden layers of the basal unit). Hence this is a
conservative volume estimate. The resulting volume of dunes
corresponding to the area of highest gypsum concentration
(red in Figure 1a) is 
75 km3. In all of the calculations
below we assume that the concentrations of gypsum calcu-
lated in section 2.1 are representative of the entire 75-km3
volume.
[51] Assuming a porosity of the dunes of 66% and a
composition of 100% gypsum (thus ignoring the volume
taken up by dark inclusions within the gypsum grains), we
get 25 km3 (5.8  1013 kg) of pure crystalline gypsum. This
mass of gypsum corresponds to a mass of CaO of 1.9 
1013 kg. If all Martian basalts contain 10 wt.% CaO, as they
do at Gusev [Gellert et al., 2004], then one needs 1.9 
1014 kg of basaltic material. If all of the necessary sulfur
comes from sulfide, then since 1 mol of Ca2+ must react
with 1 mol of SO4
2, 2.9  1013 kg of sulfide is required.
However, if the typical amount of sulfide in Martian basalts
is equivalent to the 0.5 wt.% value in shergottites [Lorand
et al., 2005], then only 9.5  1011 kg of sulfide would be
present in our 1.9  1014 kg of basaltic material, a factor of
30 too little. Obviously, large uncertainties on the amounts
of CaO or sulfur in Martian basalts can modify this rough
calculation. For instance, increasing the amount of sulfur in
the basalts by a factor 10 to be in better agreement with the
amount of sulfur found in Martian soils by in situ measure-
ment [Yen et al., 2005] could provide enough sulfur to
explain the amount of gypsum.
[52] If the 0.5 wt.% value for sulfide in basalts is correct,
then it is possible that alteration of basaltic material alone
could not produce the amount of gypsum observed. We are
left with three explanations. First, the alteration process
consumed all of the calcium, and an excess of sulfur was
present from, for example, a hidden sulfide deposit, an
enrichment of sulfide in the basal unit which is the source
for most (if not all) of the dunes, and/or a nearby volcanic
eruption. While volcanic features have been noted near the
region of Chasma Boreale mouth [Hodges and Moore,
1994; Garvin et al., 2000], their age is unknown, and their
connection to this gypsum deposit cannot be definitively
established. The presence of a large hidden sulfide source
(for example, komatiitic rocks) or enrichment of the basal
unit in sulfide cannot be proven. Second, all of the sulfide
was consumed in the presence of an excess of calcium,
requiring even more basaltic material (and hence more
magnesium and iron) and the consequent likely presence
of secondary mineral phases (like Mg and Fe sulfates). Such
secondary phases have not been observed, so that, if this
scenario is the case, these minerals must have left the
system, given their higher solubility. Third, the amount of
gypsum observed by OMEGA is not typical of the entire
bulk volume of dune deposits in which it is observed.
[53] Mass balance issues concerning the sulfide also
remain to be solved. Sulfide will liberate SO4 into solution,
but it will also liberate Fe, so one may expect to see Fe
sulfates in association with the gypsum. Such sulfates have
not been observed. Some of this mass balance problem can
be solved if the DSF phase within the gypsum is magnetite.
Magnetite is one of the oxides expected as a secondary
phase in the alteration of basaltic material, and it can act as a
geochemical sink for the Fe. Even with an equal starting
amount of Fe and S, the resulting gypsum/magnetite ratio
will be 
10. Since the reflectance properties of the dune
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
13 of 17
E07002
material are largely sensitive to the volume ratios of the
components, magnetite would effectively hide the ??missing??
iron.
[54] On the basis of the hydration of gypsum, an absolute
minimum amount of requisite water is 6 1012 kg (60 km3).
Estimating the amount of water delivered by the nearby
channels for comparison is not straightforward. We cannot
know how many outflow events occurred or by how much
subsequent geologic activity has altered their morphology
(for example, by burying portions of the channels). The best
estimate that we can obtain of water delivered is to calculate
the volume of currently visible channels (as outlined in
Figure 13) using MOLA data. This volume comes to only
12 km3. Thus some of the channels have been infilled, the
channels were filled more than once (up to 
5 times), most
of the water was delivered by the long, sinuous channel
connecting these smaller channels to the Chasma Boreale
region, and/or, again, the amount of gypsum observed is not
typical of the bulk dune deposits in the gypsum-rich area.
[55] Finally, we come to the timing. Previous experiments
have shown that sulfates can form within 1 year [Chevrier
et al., 2004, 2006]. Therefore, in the case of weathering in
the presence of SO2 or of sulfides, the formation process
of gypsum could be relatively fast compared to geologic
timescales. Exact numbers for this particular gypsum
deposit are impossible to calculate. Clearly, the water
needed to have remained liquid long enough to form the
gypsum. The question of how long the water needed
to remain liquid is nearly impossible to calculate. Specula-
tively, we suggest that the dune material acted as a kinetic
barrier and the dissolved salts as a chemical barrier to
evaporation while the water was circulating within the
dunes. Of course, this brings up the further question of
exactly how, physically, the water circulated within the
dunes, something which is beyond the scope of this paper
but can be examined in detail in a future study.
3.4.2. What About Chasma Boreale?
[56] If melting of the PLD initiated the formation of
Chasma Boreale [Clifford, 1987; Benito et al., 1997;
Fishbaugh and Head, 2002], then why is gypsum present
in only extremely minor amounts in the Chasma Boreale
region? OMEGA data indicate only trace amounts of
gypsum within the chasma (Figure 14), and the presence
of bright patches that may be composed of gypsum
(section 3.1) within the dune patch bordering the mesa
beyond the Chasma mouth.
[57] Several factors may have precluded the formation or
detection of gypsum here. As described above, formation of
detectable amounts of gypsum by direct alteration of mafic
materials will most easily take place where a deposit of
loose grains exists (for example, a dune field), and dunes
may also be needed to slow the evaporation process enough
to allow formation of significant quantities of evaporitic
grains. Fishbaugh and Head [2002] postulated that when
the formation of Chasma Boreale was initiated by melting,
water initially flowed beneath the ice. There is not neces-
sarily any reason to believe that sediment deposits or dune
fields existed beneath the ice at that time. Additionally, any
gypsum which did form could have been redissolved or
mechanically broken down during transport. Finally, eolian
Figure 13. (a) Portion of the map in Figure 1a showing eastern edge of dune sea in black and white.
Boxes indicate rectangles used for calculating surface area of the high-gypsum area. Dashed line shows
location of MOLA profile in Figure 13b. (b) Profile of MOLA gridded data through high-gypsum area of
dunes. Thickness of dune deposit is taken to be magnitude of topography variation (as indicated by grey
arrows), ignoring regional slope. This thickness is about 10 m. (c) Close-up of channels outlined in
Figure 12. Black lines indicate channel lengths used in calculation of the amount of water delivered by
the channels. Gray scale MOLA topography over MOLA-shaded relief. Other than the polar cap margin
exposed in the lower right corner, lower elevations are darker. Channels are each 
10 m deep.
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
14 of 17
E07002
erosion and deposition within the Chasma since its forma-
tion has certainly affected its present appearance [Howard,
2000; Fishbaugh and Head, 2002], possibly stripping away
some of any meltwater-related deposits.
4. Conclusions
[58] OMEGA has detected gypsum in the north polar
region only within the Olympia Undae sand sea and a few
nearby, isolated dune patches, not within the circumpolar
materials, PLD, or basal unit. In the most gypsum-rich
areas, the best fit to our compositional model yields con-
centrations of 35% pure gypsum grains of a few tens of
microns in size mixed with 65% millimeter-sized gypsum
grains containing inclusions of a dark, spectrally featureless
material. Eolian-rippled, bright patches (distinguished from
bright patches of underlying material), observed in MOC
images between dunes and scattered throughout the gyp-
sum-bearing area, might account for the fine-grained, pure
gypsum and for the mechanical break down of gypsum
during saltation; but preliminary CRISM observations indi-
cate lower gypsum concentrations in these patches com-
pared to the dunes themselves. Nomorphological, significant
average albedo, color, or temperature anomalies manifest
themselves within the gypsum-rich dunes.
[59] We propose that water from nearby channels perco-
lated through the eastern end of the Olympia Undae dune
sea. The gypsum most likely formed as a result of the
creation of evaporative gypsum grains and of direct alter-
ation of sand dune grains bearing high-calcium, Fe-rich
pyroxene and sulfide. Such alteration would produce oxide
as a secondary phase; secondary oxides, such as magnetite,
could easily act as the spectrally featureless darkening agent
(the inclusions) implied by our compositional modeling to
exist within the gypsum grains. Creation of evaporative
crystals could easily involve trapping of basaltic grains
within them which would also lower the albedo of the
crystals.
[60] Liquid water was produced by the Chasma Boreale
melting event and/or by melting due to the nearby impact
into ice. The presence of a large expanse of dune material
(which can be easily altered and which provides plenty of
pore space in which evaporative crystals can grow) near the
mouths of meltwater channels conspired to create a unique
situation wherein gypsum was able to form on Mars, even
within the late Amazonian. No distant volcanic activity is
required, and the confined distribution of the gypsum is
easily explained by this hypothesis. An evaporative deposit
(similar to those created in saline mudflats or on alluvial
fans) may exist near the channel mouths, but its existence is
impossible to prove with orbital spectrometer data since this
area is now covered by dust and frost/ice.
[61] Rough estimates, based on geochemistry, of the
amount of sulfur and calcium required to form the amount
of gypsum observed indicate that roughly 30 times more
sulfide could be needed than what likely exists within the
dunes. Although we are aware that these calculations are
based on values with large uncertainties, we propose that
the possible ??missing sulfide?? could stem from the fact that
the estimated concentration of gypsum detected by OMEGA
does not represent the concentration within the entire
three-dimensional volume of dune material in which it is
found. More complex explanations include the presence of
a hidden sulfide source, late Amazonian volcanism as a
supply of sulfur, or enrichment of sulfide in the basal unit
(the source for most, if not all, of the dunes, but not of the
gypsum).
[62] Dedicated observations of the Olympia Undae dunes
by CRISM could shed even more detailed light on the north
polar gypsum story. One may be able to assess the compo-
sition of small-scale structures such as small collections of
the rippled, the bright patches observed within some dunes,
and hills and knobs within and near these dunes which may
Figure 14. Part of OMEGA gypsum concentration map overlain on MOLA-shaded relief showing
close-up of Chasma Boreale region. Even when contrast enhanced, gypsum concentrations in this region
appear minor to nonexistent on this map. Band depth scale is the same as in Figure 1a.
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
15 of 17
E07002
contain or be blanketed by gypsum. Detailed compositional
mapping of the dune field would also be a key step in the
building of a more thorough understanding of the geochem-
ical pathways taken by the gypsum. Another future step of
this analysis could consist of initiating a study using
OMEGA results to complement those of thermal instru-
ments, MGS TES and Odyssey THEMIS, sensitive to non-
Fe-bearing minerals.
[63] Acknowledgments. Josh Bandfield (Arizona State University)
and Nicolas Tosca (SUNY Stony Brook) provided thorough, constructive
reviews which improved the quality of this paper. The authors wish to thank
the staff of THMPROC (Arizona State University), especially Josh Band-
field, and the USGS for crucial help in processing THEMIS data and
importing data into ESRI Arcmap. Patrick Russell (University of Bern) also
provided helpful comments.
References
Arvidson, R., F. Poulet, J.-P. Bibring, M. Wolff, A. Gendrin, R. Morris,
J. Freeman, Y. Langevin, N. Mangold, and G. Bellucci (2005), Spectral
reflectance andmorphologic correlations in Eastern TerraMeridiani, Mars,
Science, 307(5715), 1576?1581.
Bandfield, J. (2002), Global mineral distributions on Mars, J. Geophys.
Res., 107(E6), 5042, doi:10.1029/2001JE001510.
Bandfield, J., V. Hamilton, and P. Christensen (2000), A global view of
Martian surface compositions from MGS-TES, Science, 287, 1626?
1630.
Banfield, J., V. Hamilton, and H. McSween (2004), Identification of quartzo-
feldspathic materials on Mars, J. Geophys. Res., 109, E10009, doi:10.1029/
2004JE002290.
Benito, G., F. Mediavilla, M. Fernandez, A. Marquez, J. Martinez, and
F. Anguita (1997), Chasma Boreale, Mars: A sapping and outflow
channel with a tectono-thermal origin, Icarus, 129, 528?538.
Bibring, J.-P., et al. (2005), Mars surface diversity as revealed by the
OMEGA/Mars Express observations, Science, 307, 1576?1581.
Bibring, J.-P., et al. (2006), Global mineralogical and aqueous Mars history
derived from OMEGA/Mars Express data, Science, 312 5772, 400?404.
Burns, R. (1993), Rates and mechanisms of chemical weathering of ferro-
magnesian silicate minerals on Mars, Geochim. Cosmochim. Acta, 57,
4555?4574.
Burns, R., and D. Fisher (1990a), Evolution of sulfide mineralization on
Mars, J. Geophys. Res., 95(B9), 14,169?14,173.
Burns, R., and D. Fisher (1990b), Iron-sulfur mineralogy of Mars: Mag-
matic evolution and chemical weathering products, J. Geophys. Res.,
95(B9), 14,415?14,421.
Burns, R., and D. Fisher (1993), Rates of oxidative weathering on the
surface of Mars, J. Geophys. Res., 98(E2), 3365?3372.
Byrne, S., and B. Murray (2002), North polar stratigraphy and the paleo-erg
of Mars, J. Geophys. Res., 107(E6), 5044, doi:10.1029/2001JE001615.
Carr, M. (1996), Water on Mars, 229 pp., Oxford Univ. Press, New York.
Catling, D. (1999), A chemical model for evaporites on early Mars: Possi-
ble sedimentary tracers of the early climate and implications for explora-
tion, J. Geophys. Res., 104(E07), 16,453?16,470.
Chevrier, V., P. Rochette, P.-E. Mathe?, and O. Grauby (2004), Weathering
of iron-rich phases in simulated Martian atmospheres, Geology, 32(12),
1033?1036.
Chevrier, V., P. Mathe?, P. Rochette, O. Grauby, G. Bourrie, and F. Trolard
(2006), Iron weathering products in a CO2 + (H2O or H2O2) atmosphere:
Implications for weathering processes on the surface of Mars, Geochim.
Cosmochim. Acta, 70, 4295?4317.
Christansen, E. (1989), Lahars in the Elysium region of Mars, Geology, 17,
203?206.
Christensen, P. R., et al. (2003), Morphology and composition of the sur-
face of Mars: Mars Odyssey THEMIS results, Science, 300(5628),
2056?2061.
Clark, B., and A. Baird (1979), Is the Martian lithosphere sulfur rich?,
J. Geophys. Res., 84(B14), 8395?8403.
Clark, R., G. Swayze, A. Gallagher, T. King, and W. Calvin (1993), U.S.
Geological Survey Digital spectral library version 1: 0.2 to 3.0 microns,
U.S. Geol. Surv. Open File Rep. 93?592, http://speclab.cr.usgs.gov.
Clifford, S. (1987), Polar basal melting on Mars, J. Geophys. Res., 92,
9135?9152.
Edgett, K., and P. Christensen (1991), The particle size of Martian eolian
dunes, J. Geophys. Res., 96(E5), 22,765?22,776.
Edgett, K., and M. Malin (2000), New views of Mars eolian activity,
materials, and surface properties: Three vignettes from the Mars Global
Surveyor Camera, J. Geophys. Res., 105(E1), 1623?1650.
Edgett, K., R. Williams, M. Malin, B. Cantor, and P. Thomas (2003), Mars
landscape evolution: Influence of stratigraphy on geomorphology in the
north polar region, Geomorphology, 52, 289?297.
Fairen, A., D. Fernandez-Remolar, J. Dohm, V. Baker, and R. Amils (2004),
Inhibition of carbonate synthesis in acidic oceans on early Mars, Nature,
431, 423?426.
Fishbaugh, K. (2005), Characterization of Martian north polar geologic
units using Mars Odyssey THEMIS data, paper presented at Lunar
and Planetary Science Conference 36, Lunar and Planetary Institute,
Houston, Abstract 1335.
Fishbaugh, K., and J. Head (2002), Chasma Boreale, Mars: Topographic
characterization from Mars Orbiter Laser Altimeter data and implications
for mechanisms of formation, J. Geophys. Res., 107(E3), 5013, doi:10.1029/
2000JE001351.
Fishbaugh, K., and J. Head (2005), Origin and characteristics of the Mars
north polar basal unit and implications for polar geologic history, Icarus,
174(2), 444?474.
Garvin, J., S. Sakimoto, J. Frawley, C. Schnetzler, and H. Wright (2000),
Topographic evidence for geologically recent near-polar volcanism on
Mars, Icarus, 145(2), 648?652.
Gellert, R., et al. (2004), Chemistry of rocks and soils in Gusev Crater from
the Alpha Particle X-ray Spectrometer, Science, 305(5685), 829?832.
Gendrin, A., et al. (2005), Sulfates in Martian layered terrains: The OMEGA/
Mars Express view, Science, 307, 1587?1591.
Hamilton, V., H. McSween, and B. Hapke (2005), Mineralogy of Martian
atmospheric dust inferred from thermal infrared spectra of aerosols,
J. Geophys. Res., 110, E12006, doi:10.1029/2005JE002501.
Hapke, B. (1981), Bidirectional reflectance spectroscopy: 1. Theory,
J. Geophys. Res., 86, 4571?4586.
Hartmann, W., and G. Neukum (2001), Cratering chronology and the evo-
lution of Mars, in Chronology and Evolution of Mars, edited by
R. Kallenbach and A. Vasavada (1999), Dark material in the polar layered
deposits and dunes on Mars, J. Geophys. Res., 104(E7), 16,487?16,500.
Herkenhoff, K., and A. Vasavada (1999), Dark material in the polar layered
deposits and dunes on Mars, J. Geophys. Res., 104(E7), 16,487 ?
416,500.
Hodges, C., and H. Moore (1994), Atlas of volcanic landforms on Mars,
U.S. Geol. Surv. Prof. Pap., 1534.
Howard, A. (2000), The role of eolian processes in forming surface features
of the Martian polar layered deposits, Icarus, 144, 267?288.
Langevin, Y., F. Poulet, J.-P. Bibring, and B. Gondet (2005), Sulfates in the
north polar region of Mars detected by OMEGA/Mars Express, Science,
307(5715), 1584?1586.
Langford, R. (2003), The Holocene history of the White Sands dune field
and influences on eolian deflation and playa lakes, Quat. Int., 104(1),
31?39.
Lee, P., and P. Thomas (1995), Longitudinal dunes on Mars: Relation to
current wind regime, J. Geophys. Res., 100(E3), 5381?5395.
Lorand, J., V. Chevrier, and V. Sautter (2005), Sulfide mineralogy and
redox conditions in some shergottites, Meteorit. Planet. Sci., 40(8),
1257?1272.
Malin, M., and K. Edgett (2000), Evidence for recent groundwater seepage
and surface runoff on Mars, Science, 288(5475), 2230?2335.
Martinez-Fr??as, J., R. Lunar, J. Rodr??guez-Losada, A. Delgado, and F. Rull
(2004), The volcanism-related multistage hydrothermal system of El
Jaroso (SE Spain): Implications for the exploration of Mars, Earth
Planets Space, 56(7), v?viii.
McKee, E. (1966), Structures of dunes at White Sands National Monument,
New Mexico (and comparison with structures of dunes from other
selected areas), Sedimentology, 7, 1?69.
Ming, D., et al. (2004), Geochemical and mineralogical indicators for
aqueous processes in the Columbia Hills of Gusev crater, Mars, J. Geophys.
Res., 111, E02S12, doi:10.1029/2005JE002560.
Mustard, J., F. Poulet, A. Gendrin, J.-P. Bibring, Y. Langevin, B. Gondet,
N. Mangold, G. Bellucci, and F. Altieri (2005), Olivine and pyroxene
diversity in the crust of Mars, Science, 307, 1594?1597.
Paige, D., J. Bachman, and K. Keegan (1994), Thermal and albedo map-
ping of the polar regions of Mars using Viking thermal mapper observa-
tions: 1. North polar region, J. Geophys. Res., 99(E12), 25,991?925,959.
Plaut, J., et al. (2006), MARSIS Subsurface Sounding Observations of the
Polar Deposits of Mars, paper presented, Eos Trans. AGU, 87(52), Fall
Meet. Suppl., Abstract P13-D06.
Postawko, S., and W. Kuhn (1986), Effect of greenhouse gases (CO2, H2O,
SO2) on Martian paleoclimate, J. Geophys. Res., 91(B4), D431?D438.
Poulet, F., and S. Erard (2004), Nonlinear spectral mixing: Quantitative
analysis of laboratory mineral mixtures, J. Geophys. Res., 109, E02009,
doi:10.1029/2003JE002179.
Poulet, F., N. Mangold, and S. Erard (2003), A new view of dark Martian
regions from geomorphic and spectroscopic analysis of Syrtis Major,
Astron. Astrophys., 412, L19?L23.
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
16 of 17
E07002
Poulet, F., J.-P. Bibring, J. Mustard, A. Gendrin, N. Mangold, Y. Langevin,
R. Arvidson, B. Gondet, and C. Gomez (2005a), Phyllosilicates on Mars
and implications for the Early Mars History, Nature, 481, 623?627.
Poulet, F., Y. Langevin, J. P. Bibring, B. Gondet, R. Arvidson, and T. O.
Team (2005b), Mineralogy of the northern high latitude regions of Mars,
paper presented at Lunar and Planetary Science Conference 36, Lunar
and Planetary Institute, Houston, Abstract 1828.
Poulet, F., C. Gomez, J.-P. Bibring, Y. Langevin, B. Gondet, P. Pinet,
G. Belluci, J. Mustard, and the OMEGA team (2007), Martian surface
mineralogy from OMEGA/MEx: Global mineral maps, J. Geophys. Res.,
doi:10. 1029/2006JE002840, in press.
Reider, R., et al. (2004), Chemistry of rocks and soils at Meridiani Planum
from the alpha particle X-ray spectrometer, Science, 306, 1746?1749.
Roush, T., F. Esposito, G. Rossman, and L. Colangeli (2006), Gypsum
optical constants in the visible and near-infrared: Real and imagined,
paper presented at Lunar and Planetary Science Conference 37, Lunar
and Planetary Science Institute, Houston, TX, Abstract 1188.
Russell, P., and J. Head (2003), Elysium-Utopia flows as mega-lahars: A
model of dike intrusion, cryosphere cracking, and water-sediment release,
J. Geophys. Res., 108(E6), 5064, doi:10.1029/2002JE001995.
Schatz, V., H. Tsoar, K. Edgett, E. Parteli, and H. Herrmann (2006), Evi-
dence for indurated sand dunes in the Martian north polar region,
J. Geophys. Res., 111, E04006, doi:10.1029/2005JE002514.
Shkuratov, Y., L. Starukhina, H. Hoffmann, and G. Arnold (1999), A model
of spectral albedo of particulate surfaces: Implications for optical proper-
ties of the Moon, Icarus, 137, 235?246.
Sinha, R., and B. Raymahashay (2004), Evaporite mineralogy and geo-
chemical evolution of the Sambhar Salt Lake, Rajasthan, India, Sediment.
Geol., 166, 59?71.
Skinner, J., and K. Tanaka (2001), Regional emplacement history of the
Utopia and Elysium plains deposits, paper presented at Lunar and Pla-
netary Science Conference 32, Lunar and Planetary Science Institute,
Houston, Abstract 2154.
Smoot, J., and T. Lowenstein (1991), Depositional environments of non-
marine evaporites, in Evaporites, Petroleum, and Mineral Resources,
edited by J. Melvin et al., pp. 165?196, Springer, New York.
Squyres, S. W., et al. (2004), In situ evidence for an ancient aqueous
environment at Meridiani Planum, Mars, Science, 306, 1709?1714.
Sracek, O., M. Choquette, P. Gelina, F. Lefebvre, and R. Nicholson (2004),
Geochemical characterization of acid mine drainage from a waste rock
pile, Mine Doyon, Quebec, Canada, J. Contam. Hydro., 69, 45?71.
Tanaka, K. (2006), Mars? north polar gypsum: Possible origin related to
early Amazonian magmatism at Alba Patera and aeolian mining, paper
presented at Fourth International Mars Polar Science and Exploration
Conference, Lunar and Planetary Science Institute, Davos, Switzerland,
Abstract 8024.
Tanaka, K., and D. McKinnon (1989), Release of Martian catastrophic
floods by fracture discharge from volcanotectonic regions, paper pre-
sented at 4th International Conference on Mars, Univ. of Ariz. Press,
Tucson.
Tanaka, K., J. Skinner, and T. Hare (2005), Geologic map of the northern
plains of Mars, U.S. Geol. Surv., Sci. Invest. Ser., Map 288.
Thomas, P., and C. Weitz (1989), Sand dune materials and polar layered
deposits on Mars, Icarus, 81, 185?215.
Thomas, P., et al. (1999), Bright dunes on Mars, Nature, 397, 592?594.
Tosca, N. J., and S. M. McLennan (2005), Evaporites at Meridiani Planum
and implications for surficial processes on Mars, Eos Trans. AGU,
86(52), Fall Meet. Suppl. Abstract P12A-07.
Tosca, N., A. McLennan, D. Lindsley, and M. Schoonen (2004), Acid-
sulfate weathering of synthetic Martian basalt: The acid fog model revi-
sited, J. Geophys. Res., 109, E05003, doi:10.1029/2003JE002218.
Tosca, N., S. McLennan, B. Clark, J. Grotzinger, J. Hurowitz, A. Knoll,
C. Shroeder, and S. Squyres (2005), Geochemical modeling of evapora-
tion processes on Mars: Insight from the sedimentary record at Mer-
idiani Planum, Earth Planet. Sci. Lett., 240, 122?148.
Tsoar, H., R. Greeley, and A. Peterfreund (1979), Mars: The north polar
sand sea and related wind patterns, J. Geophys. Res., 84(B14), 8167?
8180.
Vysotskiy, E., A. Makhnach, T. Peryt, and S. Kruchek (2004), Marine and
continental Lower Permian evaporites of the Prypiac? Trough (Belarus),
Sediment. Geol., 172, 211?222.
Waenke, H., G. Driebus, E. Jagoutz, and L. Mukhin (1992), Volatiles on
Mars: The role of SO2, paper presented at Lunar and Planetary Science
Conference 23, Lunar and Planetary Science Institute, Houston, TX,
Abstract, 1489?1490.
Yen, A., et al. (2005), An integrated view of the chemistry and mineralogy
of Martian soils, Nature, 436, 49?54.

J.-P. Bibring, Y. Langevin, and F. Poulet, Institut d?Astrophysique
Spatiale, Universite? Paris-Sud XI, Orsay, France.
V. Chevrier, W. M. Keck Laboratory for Space and Planetary
Simulations, MUSE 202, University of Arkansas, Fayetteville, AR
72701, USA.
K. E. Fishbaugh, International Space Science Institute, Hallerstrasse 6,
CH-3012, Bern, Switzerland. (fishbaugh@issibern.ch)
E07002 FISHBAUGH ET AL.: MARS NORTH POLAR GYPSUM
17 of 17
E07002