Icarus 175 (2005) 58–67
www.elsevier.com/locate/icarus
Mars low-latitude neutron distribution: Possible remnant near-surface
water ice and a mechanism for its recent emplacement
Bruce M. Jakosky a,b,∗ , Michael T. Mellon a , E. Stacy Varnes a , William C. Feldman c ,
William V. Boynton d , Robert M. Haberle e
a Laboratory for Atmospheric and Space Physics, University of Colorado, Boulder, CO 80309, USA
b Department of Geological Sciences, University of Colorado, Boulder, CO 80309, USA
c Los Alamos National Laboratory, Los Alamos, NM 87544, USA
d Department of Planetary Sciences, University of Arizona, Tucson, AZ 85721, USA
e NASA Ames Research Center, Moffett Field, CA, USA
Received 7 July 2004; revised 2 November 2004
Available online 2 March 2005
Abstract
The Mars Odyssey Gamma-Ray Spectrometer/Neutron Spectrometer/High Energy Neutron Detector has provided measurements of nearsurface hydrogen, generally interpreted as resulting from water, in the equatorial and mid-latitudes. Water abundances as great as 10% by
mass are inferred. Although such high abundances could be present as adsorbed water in clays or water of hydration of magnesium salts,
other measurements suggest that this is not likely. The spatial pattern of where the water is located is not consistent with a dependence on
composition, topography, present-day atmospheric water abundance, latitude, or thermophysical properties. The zonal distribution of water
shows two maxima and two minima, which is very reminiscent of a distribution that is related to an atmospheric phenomenon. We suggest
that the high water abundances could be due to transient ground ice that is present in the top meter of the surface. Ice would be stable at
tens-of-centimeters depth at these latitudes if the atmospheric water abundance were more than about several times the present value, much
as ice is stable poleward of about ±60◦ latitude for current water abundances. Higher atmospheric water abundances could have resulted
relatively recently, even with the present orbital elements, if the south-polar cap had lost its annual covering of CO2 ice; this would have
exposed an underlying water-ice cap that could supply water to the atmosphere during southern summer. If this hypothesis is correct, then
(i) the low-latitude water ice is unstable today and is in the process of sublimating and diffusing back into the atmosphere, and (ii) the current
configuration of perennial CO2 ice being present on the south cap but not on the north cap might not be representative of the present epoch
over the last, say, ten thousand years.
 2004 Elsevier Inc. All rights reserved.
Keywords: Mars; Neutrons; Mars Odyssey; Water ice; Mars climate
1. Introduction
The Gamma-Ray Spectrometer/Neutron Spectrometer/
High-Energy Neutron Detector instrument suite (GRS/NS/
HEND) on board the Mars Odyssey spacecraft is able to detect the presence of hydrogen in, at most, the top meter of the
regolith, with a spatial resolution of about six hundred kilo* Corresponding author.
E-mail address: [email protected] (B.M. Jakosky).
0019-1035/$ – see front matter  2004 Elsevier Inc. All rights reserved.
doi:10.1016/j.icarus.2004.11.014
meters. Although hydrogen can be present in many forms,
including bound OH in minerals, H2 O2 produced from atmospheric chemical processes, methane, etc., its most likely
form is as H2 O (Drake et al., 1988; Feldman et al., 2002,
2004b; Boynton et al., 2002; Mitrofanov et al., 2002). This
is especially the case given the relatively large amounts of
hydrogen that are inferred to be present, which may be
inconsistent with other reservoirs. The GRS/NS/HEND is
mapping, in essence, the water content of the uppermost surface layer (Drake et al., 1988; Feldman et al., 2002, 2004b;
Boynton et al., 2002; Mitrofanov et al., 2002).
Mars low-latitude neutron distribution
It was originally anticipated that the high-latitude regolith would contain water in the form of ground ice that
was in diffusive equilibrium with the atmosphere, and that
the low- and mid-latitude-regolith water would be most directly related to adsorbed water in diffusive equilibrium
with the atmospheric water vapor (Drake et al., 1988). Although this expectation appears to have been borne out at the
high latitudes (Feldman et al., 2002; Boynton et al., 2002;
Mitrofanov et al., 2002; Mellon et al., 2004), the water abundances and distribution at low and mid latitudes is not so
simple. A number of possible explanations have been put
forward to explain the low- and mid-latitude water distribution, including the presence of hydrated magnesium sulfate
salts, water-rich zeolites, and adsorbed water (see Feldman
et al., 2004a). Here, we explore the question in more detail,
in order to understand both the constraints on any possible
explanation and alternative physical models that might be
equally or more consistent with the observations.
We examine the relationship between the water in the
regolith and the physical properties that might be playing
a role in controlling the water abundance, in order to see
what physical connections might exist. The apparent lack
of a unique or a strong global relationship then drives us to
consider nonequilibrium and nonsteady-state processes that
might be controlling the water distribution.
2. Relationship between regolith water and physical
properties
We expect to find water in the regolith at low and mid latitudes today. It could result, for example, from the presence
of water that was a part of the original volcanic rock erupted
to the surface. In this case, we might expect it to be present at
an abundance of less than a couple of per cent or less (Crisp,
1984; McSween et al., 2001). It also could be present as adsorbed water, lightly bound to individual regolith grains by
van der Waals forces and in equilibrium with water vapor
in the surrounding pore space. Water will adsorb onto the
surface whenever water vapor is present, and would be expected to be in equilibrium with atmospheric water vapor via
diffusion through the porous regolith (Mooney et al., 1952;
Zent and Quinn, 1995; Jakosky, 1985). For present-day atmospheric water abundances, there could be as little as a
fraction of a per cent adsorbed water by mass for a basaltic
regolith or as much as a few tens of per cent adsorbed water
for a clay-rich regolith (see Jakosky, 1983a, 1983b).
If water is present as an adsorbed species, we would expect its abundance to relate directly to the physical properties of the regolith and atmosphere, as these control the
equilibrium abundance of bound water. One would expect
that surface or subsurface temperature, surface composition,
and atmospheric water abundance each would exert partial
control on the amount of adsorbed water. In addition, other
properties that might relate directly to these include latitude
(which is related to both atmospheric water abundance and
59
subsurface temperature), thermal inertia and albedo (which
control temperature, with albedo also being a good firstorder surrogate for composition), and local elevation (which
affects dynamical transport of water through the atmosphere
as well as the near-surface atmospheric water-vapor number
density).
Figure 1 shows a map of the amount of hydrogen present
in the regolith, expressed as the equivalent amount of water.
As both neutrons and gamma rays that are detected by the
GRS/NS/HEND and which are sensitive to water emanate
from the upper ∼1 m of the regolith, these values are representative of the water content of the uppermost meter or so
of the regolith. Values outside of the high latitudes where ice
is expected to be present range from a low of about 2% to
a high of over 10% (Feldman et al., 2004a). Also shown in
Fig. 1 for comparison are maps of many of the physical properties that might exert some control over the amount of water
in the regolith. There are some intriguing connections, with
boundaries in the water abundance corresponding in some
instances with boundaries in other properties. However, not
every boundary in the regolith water abundance corresponds
to a boundary in one of the controlling parameters. There is
no simple correlation that is able to explain all of the global
variations, and not every component of the regolith water
distribution corresponds to variations in any of the potential
controlling parameters. For example, the high water abundances in the Arabia region (near the middle of the map in
Fig. 1) correspond to the high albedo of this region, but the
local agreement is not compelling and the high-albedo terrain in the Tharsis region and to the west contains a wide
range of water abundances.
We explored the quantitative connections between the regolith water abundance and each of the physical properties
that might be controlling the abundance. We did this by calculating the degree of correlation between the epithermal
neutron abundance (from which the water abundance is derived) and each of the physical parameters in Fig. 1, after
smoothing the latter data sets in order to match the spatial
resolution of the neutron measurements. Data were compared between latitudes of ±45◦ latitude only, in order to
not include the high-latitude regions in which ice in diffusive
equilibrium with the atmosphere is thought to exist. Table 1
shows the results, and an example cross plot between the epithermal neutron abundance and one of the parameters (in
this case, mean regolith temperature) is shown in Fig. 2. The
largest correlation coefficients are between the neutron abundance and the peak annual atmospheric water abundance, the
mean annual subsurface temperature, and the latitude. Each
of these might be expected to show a causal connection to the
neutrons. However, in each case the correlation coefficient
was less than 0.3, indicating only a very weak connection at
best. Clearly, neither a comparison by inspection of Fig. 1
nor a quantitative comparison shows a compelling relationship, and no single parameter is able by itself to explain a
significant fraction of the water distribution.
60
B.M. Jakosky et al. / Icarus 175 (2005) 58–67
Fig. 1. Maps showing the abundance of hydrogen and of different physical properties that might control the water distribution in the regolith. Each map
is between ±45◦ latitude. From top to bottom, the maps are the water-equivalent abundance of hydrogen (Feldman et al., 2004a, 2004b), the annual peak
abundance of atmospheric water vapor as derived from MGS TES measurements (Smith, 2002; M.D. Smith, personal communication, 2003), the mean annual
surface temperature (derived from the models of (Mellon et al., 2004)), the topography (Smith et al., 1999), the mean annual abundance of atmospheric water
vapor as adjusted to a common datum to “correct” for topography (Smith, 2002; M.D. Smith, personal communication, 2003), the thermal inertia of the surface
(Mellon et al., 2000), the mean annual abundance of atmospheric water vapor as measured without a topographic correction, and the bolometric albedo as
measured from MGS TES measurements (see Christensen et al., 2001).
Mars low-latitude neutron distribution
61
Table 1
Correlation coefficient between epithermal neutrons and potentially relevant controlling physical properties (see text for discussion; the data sets are those
shown in Fig. 1)
Neutrons
Max H2 O
Latitude
Temp
Elevation
Topo-corrected H2 O
Thermal inertia
Uncorrected H2 O
Albedo
Neutrons
Max H2 O
Latitude
Temp
Elevation
Topo-corrected H2 O
Thermal inertia
Uncorrected H2 O
Albedo
1
−0.28
−0.263
0.252
0.202
0.087
0.082
−0.072
−0.064
−0.28
1
0.559
−0.284
−0.668
−0.049
0.03
0.39
0.108
−0.263
0.559
1
−0.092
−0.54
0.285
0.005
0.594
0.147
0.252
−0.284
−0.092
1
−0.031
0.327
0.214
0.14
−0.427
0.202
−0.668
−0.54
−0.031
1
−0.115
−0.127
−0.488
0.002
0.087
−0.049
0.285
0.327
−0.115
1
−0.053
0.623
0.247
0.082
0.03
0.005
0.214
−0.127
−0.053
1
−0.041
−0.284
−0.072
0.39
0.594
0.14
−0.488
0.623
−0.041
1
0.286
−0.063
0.108
0.147
−0.427
0.002
0.247
−0.284
0.286
1
Fig. 2. Cross plot between the epithermal neutron abundance measured from
the Mars Odyssey GRS/NS and mean annual temperature. The latter could
hypothetically be controlling the former, but the obvious lack of strong correlation suggests otherwise.
It is possible that a combination of several different physical properties affects the water distribution. One can envision a situation in which each of several controlling properties shows only a small correlation with the observed neutron
distribution, but that a combination of several of them would
correlate more strongly. To explore this possibility, we did
a multi-dimensional least-squares fit of all of the parameters to the neutron abundance. Only elevation, mean temperature, peak water abundance, and latitude showed any
possible contribution, so a least-squares fit also was done
again using only these parameters. Figure 3 shows a cross
plot between the observed neutron distribution and the leastsquares fit based on these four parameters. Again, there is
only a weak correlation at best, and the appearance of the
cross plot suggests that the basic controlling properties have
not yet been accounted for.
These arguments suggest that no combination of relevant
physical parameters is able to explain the current regolith
water distribution. Tempering this conclusion is the fact that
the measurements that refer to the physical properties of
the surface materials all relate formally to different thicknesses of regolith. Albedo refers to the uppermost micrometers, thermal inertia to a layer between 1 and 10 cm thick,
Fig. 3. Cross plot between epithermal neutron abundance and the best-fit
combination of several physical parameters. The best fit is a combination of
the mean annual temperature, latitude, peak water abundance, and elevation.
and the GRS/NS/HEND-derived water abundances to a layer
roughly a meter thick. If properties are not uniform over this
range of depths, these characteristics might not be referring
to the same materials. In this case, it is possible that vertical structure within the subsurface could be complicating
the comparisons. However, the relatively strong correlation
between albedo and thermal inertia, or thermal inertia and
radar properties (which also refer to a thicker layer in some
cases; see Jakosky and Christensen, 1986), suggests that this
may not be the case in general.
Taken in total, the data suggests that it is unlikely that
the geographical distribution of regolith water detected by
the GRS/NS/HEND represents a steady-state equilibrium of
adsorbed water at the present epoch, with the abundance
controlled by the relevant physical properties.
There is another reason why the hydrogen is not likely to
be present as adsorbed water. Basalt is a much poorer adsorber than is clay. Water adsorbed on basalt at Mars ambient
conditions would constitute less than 1% of the material by
mass (e.g., Zent and Quinn, 1995; Jakosky, 1983a). The relatively high water abundances, at levels of up to 10%, are
not likely to occur as adsorbed water unless a significant
fraction of the surface is made of clay minerals. Although
the presence of clays such as montmorillonite or nontron-
62
B.M. Jakosky et al. / Icarus 175 (2005) 58–67
ite has been suggested based on the Viking biology experiments (e.g., Clark et al., 1982), spectral evidence suggests
that basalts dominate and that clays are present only in very
small abundances if at all (e.g., Bell and McCord, 1989;
Bandfield et al., 2000). There is no evidence to suggest that
clays are present in amounts that would be sufficient to account for the high hydrogen abundances that are observed,
or that their spatial distribution would be such as to produce
the observed water distribution. And, the lack of correlation with the albedo (which is a good approximation to the
distribution of materials of like composition) or the thermal
inertia (which shows the presence of centimeters-or-thicker
deposits of dust) supports this view that weathered materials
do not contain the water.
This conclusion drives us to explore more-complicated
nonequilibrium processes and time-dependent solutions.
3. Time-variable processes
Time-dependent processes might explain the observed
water distribution if they can result in greater regolith water
abundances at some time in the past and if the regolith could
have retained at least some of its water up to the present.
One way for this to occur is if the atmospheric water content had been greater at a previous time, creating a situation
in which ground ice could be stable over a larger fraction of
the planet. If this occurred recently enough that any ground
ice that was deposited would not have had enough time to
sublimate and diffuse into the atmosphere, there could be a
residual ground ice present even though it is not stable today.
Ground ice could be stable at low latitudes if the atmospheric water vapor content were greater than it is today. At present-day atmospheric water abundances, the condensation temperature of the atmosphere is around 198 K
(Farmer and Doms, 1979; Mellon and Jakosky, 1993). This
means that, where the regolith temperature is lower than
this, water vapor could diffuse from the atmosphere into the
regolith and condense out as ice. (The situation is a little
more complicated than this due to the seasonal variations
of both atmospheric water content and subsurface temperature; the actual criterion is that the time-dependent vapor
diffusion allow a net condensation of water as ice over the
course of a year, but this is not too different for our purposes
here. See Mellon and Jakosky (1993, 1995) and Mellon et al.
(2004) for detailed discussion of this point.) At the present
epoch, ground ice is stable poleward of about ±60◦ latitude,
at depths that vary between a couple of tens of centimeters
and a meter (Mellon et al., 2004). This distribution matches
up well with the extremely high water content inferred from
the GRS/NS/HEND measurements (Feldman et al., 2002,
2004b; Boynton et al., 2002), and is suggestive of the presence of ground ice in these locations.
An increase in atmospheric water content would result
in a higher condensation temperature. This means that areas that are at lower latitudes, with warmer temperatures,
would then be within the stability region for ice, and ice
would be expected to form there. As the water content of
the atmosphere increases, the boundary separating locations
where ice is not stable from those for which ice is stable
would move to lower latitudes.
This can be seen in Fig. 4, which shows the calculated
regions of ice stability within the regolith for enhanced atmospheric water abundances. These calculations are based
on the models described by Mellon et al. (2004), and include
an approximation to the effects of the seasonally dependent
terms. A doubling of the water content will substantially increase the area in which ice is stable. And a five-fold increase
to 100 pr µm of water allows ice to be stable within the regolith over a large fraction of the planet.
One way to increase the water content is to raise the planet’s obliquity. At higher obliquity, the polar caps tilt more
toward the Sun during summer and are heated up to higher
temperatures, and more water vapor will sublimate into the
atmosphere; this water can be carried to lower latitudes,
and could saturate the atmosphere and deposit onto the
ground or into the subsurface as ice (Jakosky and Carr, 1985;
Mischna et al., 2003). Atmospheric water contents as high as
100 pr µm can occur for obliquities less than 35◦ (Jakosky
et al., 1993, 1995; Mischna et al., 2003) that might have occurred as recently as a few times 105 years ago (Ward, 1992;
Laskar et al., 2002). However, for this model to be correct,
ice would have to have survived at the equator, at depths of
substantially less than a meter, for longer than 105 years. It
is possible that the regolith could be this impermeable to diffusion, especially given the global distribution of cemented
materials (“duricrust”) that might have decreased permeability (e.g., Jakosky and Christensen, 1986); however, while
possible, this scenario seems forced at best.
We examine an alternative scenario that can yield the
same increased atmospheric water contents at much more
recent times.
The present-day atmospheric water content represents a
balance between summertime sublimation of water into the
atmosphere from the north-polar residual water-ice cap, wintertime condensation and deposition onto both the north and
south caps, seasonal exchange with adsorbed water in the
top few centimeters of the regolith, and advective transport through the atmosphere resulting from global winds
(Jakosky, 1983a, 1983b, 1985; Haberle and Jakosky, 1990;
Zent et al., 1993; Houben et al., 1997; Richardson and Wilson, 2002).
The polar caps represent the major reservoirs for supply to and loss from the atmosphere. The peak northern
hemisphere water abundance of about 100 pr µm results
from summertime sublimation from the exposed north-polar
water-ice cap (Haberle and Jakosky, 1990; Jakosky and
Haberle, 1992; Smith, 2002). Much lower southern hemisphere water contents result because the south cap has not
been observed to lose its CO2 ice cover and to expose underlying water ice. As a result, the cap never warms to temperatures that would result in sublimation of significant quan-
Mars low-latitude neutron distribution
63
Fig. 4. Maps showing regions of ice stability for different atmospheric water contents. Each map is between ±45◦ latitude, with white showing where ice
is not stable, and the color indicating the depth to stable ice. Maps are shown for annual average atmospheric water contents of 10, 20, 40, and 100 pr µm.
Calculated using the algorithm of Mellon et al. (2004).
tities of water. This difference in cap behavior is reflected
in the existence of a strong gradient in average atmospheric
water content from north to south, as seen in Fig. 1 (Jakosky
and Farmer, 1982; Smith, 2002).
Water ice is expected to be present beneath the south polar
CO2 ice because it is less volatile than CO2 ice. The covering of CO2 frost acts as a cold trap to remove water vapor
from the atmosphere. If the south cap were to entirely lose its
CO2 ice covering at the present epoch, it would expose this
underlying water ice. With CO2 frost no longer present, the
south cap would heat up to higher temperatures during summer than does the north cap, as the planet is presently closest
to the Sun during southern summer, and significant quantities of water would sublime into the atmosphere. The peak
southern water content could be as great as several hundred
pr µm (Davies et al., 1977; Jakosky et al., 1993, 1995). The
low-latitude atmospheric water content reflects the control
by the polar regions, and lies between the annual average polar amounts; observations and models suggest that it is about
equal to the average of the annual average water vapor abundance over the two polar caps (Jakosky and Farmer, 1982;
Smith, 2002; Richardson and Wilson, 2002). Thus, at a time
when there is a supply of water to the atmosphere from both
summer polar caps, equatorial water abundances should increase substantially. The annual average atmospheric water
content over the north polar cap is about 30 pr µm; the annual
average over an exposed south polar is likely to be closer to
150 pr µm, assuming that it has the same properties as does
the northern water-ice cap (see Davies et al., 1977; Jakosky
et al., 1993, 1995). Thus, it is possible that the annual average atmospheric water content could be 100 pr µm, five
or more times greater than the present observed values. As
Fig. 4 shows, this would result in substantial deposition of
water ice into the low-latitude subsurface.
Can the south cap lose its CO2 covering at the present
epoch? Theoretical models of the seasonal CO2 cycle suggest that the south residual cap has two distinctly different
stable states (Jakosky and Haberle, 1990). In one state, CO2
can be present year round, and the wintertime condensation and summertime sublimation exactly or nearly balance.
In the second state, the CO2 ice can disappear by about
mid-summer, exposing the underlying material. This material then will heat up substantially during the remainder
of summer. The resulting downward-conducted heat con-
64
B.M. Jakosky et al. / Icarus 175 (2005) 58–67
(a)
(b)
Fig. 5. Thermal model for CO2 ice (a) stable year round and (b) disappearing mid-summer. In (a), the albedo was selected to produce a closed cycle
for the condensation temperature chosen, and the figure shows the amount
of CO2 frost present on the surface. In (b), the physical parameters are identical to (a), except that CO2 frost was removed in year 1 just before the end
of summer and the results were followed for 10 years. The figure shows the
wintertime abundance of CO2 frost and the summertime temperature for
years 1, 3, and 10; cap is stable and repeats its behavior by year 10, with the
CO2 disappearing from the cap by mid-summer. From Jakosky and Haberle
(1990).
ducts back up to the surface during the winter, and results in the wintertime condensation of a smaller amount
of CO2 ice; as less CO2 ice is present at the end of winter, it will again disappear entirely during summer. These
two states can be seen in Fig. 5, which shows the results of annual thermal models for each of these scenarios.
In addition, one can envision processes by which the cap
might jump back and forth between these two stable states
(Jakosky and Haberle, 1990). A small perturbation to the
energy balance could remove a small residual covering of
CO2 , if it is thin, and expose the underlying water ice at end
of summer, thereby triggering a jump from a “covered” to
an “uncovered” state. Similarly, deposition of clean water
ice from the opposite pole could raise the albedo, causing a
cooling of the cap and a possible jump to a “covered” state.
(While the changing argument of periapsis might also trigger such a jump, this would occur on much longer timescales
rather than the interannual timescale discussed here.)
There is some evidence that the cap might occupy each
of these stable states at times. Atmospheric water measurements made from Earth in 1969 showed a much greater
southern hemisphere abundance than has been seen subsequently. The water vapor abundance was high enough to be
interpreted at that time as indicating the presence of a residual water-ice cap in the south (Barker et al., 1970; Jakosky
and Barker, 1984). More recently, observations from the
Mars Global Surveyor and the Mars Express spacecraft have
suggested that there are portions of the south cap which
have exposed summertime water ice (Titus et al., 2002;
Bibring et al., 2004). This exposure, if it varies from year
to year, might explain the factor-of-ten interannual variability in southern summer atmospheric water abundance (e.g.,
see Clancy et al., 1992). The behavior of the south-polar
“Swiss-cheese” terrain also suggests year-to-year variability
in the CO2 -ice cover (Malin et al., 2001). The appearance
of the terrain, and the increase in the size of the holes that
takes place over a year, both suggest that the CO2 -ice cover
is variable and that there could be a significant redistribution
of water ice taking place. Whether this redistribution represents an increasing exposure of water ice at the expense of
CO2 ice and we are witnessing an increasing uncovering of a
water-ice cap, or whether it represents a redistribution of the
geographical coverage with no change in the total coverage
is not known.
If the south polar region contained an exposed waterice cap during summer, the resulting enhancement of atmospheric water content might result in water ice being stable over a large fraction of the planet. With ice stable, water
vapor can diffuse into the subsurface and condense out as
ice in a relatively short time (Mellon and Jakosky, 1995).
The time for a significant amount of water ice to condense
within the regolith could be as short as 103 –104 years; this
is short compared to the time in which either the obliquity or
the argument of perihelion would change as the orbital elements evolve. If the cap only recently became covered again
with CO2 frost, sufficient time might not have elapsed for
the low-latitude water ice, now unstable with respect to sublimation, to diffuse back into the atmosphere.
Thus, it is possible that a south-polar-cap configuration
in which it is covered with CO2 year round might not be
representative of the most recent epochs, despite this being
the state observed from spacecraft. Instead, the more repre-
Mars low-latitude neutron distribution
sentative state could be with the south cap losing its CO2
ice covering during summer. This could have been the case
as recently as a few decades ago, a hundred years ago, or a
thousand years ago, at which time water ice could have been
stable quasi-globally even under the present orbital conditions. Ice deposited 100 or 1000 years ago might still be
present as a transient phase even though it would not be stable today.
4. Discussion
Several potential explanations for the high regolith water
content have been put forward. In addition to the possibility
described here involving transient, unstable ground ice, possibilities include the presence of hydrated magnesium sulfate
salts, water-rich zeolites, and adsorbed water (see Feldman
et al., 2004a). Hydrated magnesium salts are especially interesting due to the high sulfur abundances seen at the Viking
and Pathfinder landing sites (Clark, 1978) and the high abundance of magnesium salts identified at the Mars Exploration
Rover Opportunity site (Squyres et al., 2004). The apparent
lack of correlation with remote-sensing-derived indicators of
composition, however, makes these explanations somewhat
problematic.
Is the transient-ground-ice model proposed here consistent with the geographical distribution of regolith water measured from the GRS/NS? The distribution of ground ice predicted from our model at high atmospheric water content
(Fig. 4) is driven in large part by (i) the temperature as controlled by the thermal inertia and albedo, and (ii) the topography, as it affects the modeled near-surface atmospheric water
vapor number density. As none of these parameters match
well with the neutron distribution, it would be expected that
the geographical distribution of ground ice predicted from
the model would not match well either, and this is in fact the
case. Let us explore a couple of the pertinent issues.
The simple model used here assumes a uniform water vapor abundance in the atmosphere, modulated only by the
topography. In reality, the water vapor abundance varies
from this uniform mixing ratio by more than a factor of
two (Jakosky and Farmer, 1982; Smith, 2002). This variation
presumably results from the role of atmospheric transport in
redistributing the water in systematic but nonuniform ways.
A proper regolith model should include this actual variability as a boundary condition. However, neither the map of the
actual nor the topographically adjusted water vapor abundances matches well with the neutron distribution, so that a
model incorporating these as a boundary condition probably
would not match well with the water distribution either.
At the same time, the geographical distribution of atmospheric water vapor would be different if both the north
and the south polar caps were supplying water to the atmosphere. In theory, one could use the modeled atmospheric
transport to produce a modeled distribution of water and
then use this as a boundary condition. However, the signif-
65
icant differences seen between the different models of atmospheric circulation (e.g., Houben et al., 1997; Richardson
and Wilson, 2002) suggest that there still are factors that we
do not understand that are controlling the instantaneous and
net atmospheric circulation and transport.
One interesting point, though, is that the hydrogen distribution shows a zonal wave-number two behavior, with two
maxima and two minima around the equator (see Fig. 1).
This is similar to the wave-number two behavior of the atmospheric circulation, in which the topography drives a net
average upward motion over topographically high regions
and downward motion over low regions (Webster, 1977;
Pollack et al., 1981). This again suggests an atmospheric
phenomenon controlling the distribution, consistent with the
modulation of atmospheric water vapor that might be produced by the atmospheric circulation.
A different question relates to the occurrence of highlatitude ground ice. That ice appears to be in vapor-diffusive
equilibrium with the present-day atmosphere (see Mellon et
al., 2004). Is it possible for the high-latitude ice to be in
equilibrium at the present and for low-latitude ice to be out
of equilibrium? In fact, this is the behavior that one would
expect for a climate that oscillates between two states. At
low atmospheric water abundance, as seen today, ice would
be stable only poleward of about ±60◦ latitude. At high atmospheric water abundance, ice would be stable over larger
areas; at the same time, there would be no change to the
high-latitude ice. If the climate switches between the two
states, then, the high-latitude ice would always be present
while the low-latitude ice would come and go.
At present, none of the possible explanations provides a
satisfying match to the observed distribution of water in the
regolith. Is it possible to further test any of the hypotheses
or to identify analyses that would help us to choose between them? A detection of a decrease in low-latitude water
over time would be strong evidence for the continuing disappearance of transient water. Similarly, evidence related to
the interannual behavior of CO2 ice on the south polar cap
would provide an important boundary condition for the presence and stability of ground ice. On the other side of the
hypotheses, detection of abundant widely distributed magnesium salts or zeolites would be an important discriminator
between models. Absent these, landing on the surface, digging down, and searching for local ground ice might be the
most definitive (albeit the most expensive) test.
Acknowledgments
We thank Mike Smith for providing results from the MGS
TES analysis of atmospheric water vapor, and Bob Reedy
and an anonymous reviewer for useful suggestions regarding
the manuscript. This research was supported in part by the
Mars Odyssey project and the NASA Planetary Geology and
Geophysics Program.
66
B.M. Jakosky et al. / Icarus 175 (2005) 58–67
References
Bandfield, J.L., Hamilton, V.E., Christensen, P.R., 2000. A global view
of martian surface compositions from MGS TES. Science 287, 1626–
1630.
Barker, E.S., Schorn, R.A., Woszczyk, A., Tull, R.G., Little, S.J., 1970.
Mars: detection of atmospheric water vapor during the southern hemisphere spring and summer season. Science 170, 1308–1310.
Bell III, J.F., McCord, T.B., 1989. Mars: near-infrared comparative spectroscopy during the 1986 opposition. Icarus 77, 21–34.
Bibring, J.P., Langevin, Y., Poulet, F., Gendrin, A., Gondet, B., Berthe, M.,
Soufflot, A., Drossart, P., Combes, M., Bellucci, G., Moroz, V., Mangold, N., Schmitt, B., the OMEGA Team, 2004. Perennial water ice
identified in the south polar cap of Mars. Nature 428, 627–630.
Boynton, W.V., Feldman, W.C., Squyres, S.W., Prettyman, T.H., Bruckner,
J., Evans, L.G., Reedy, R.C., Starr, R., Arnold, J.R., Drake, D.M., Englert, P.A.J., Metzger, A.E., Mitrofanov, I., Trombka, J.I., d’Uston, D.,
Wanke, H., Gasnault, O., Hamara, D.K., Janes, D.M., Marcialis, R.L.,
Maurice, S., Mikheeva, I., Taylor, G.J., Tokar, R., Shinohara, C., 2002.
Distribution of hydrogen in the near surface of Mars: evidence for subsurface ice deposits. Science 297, 81–84.
Christensen, P.R., Bandfield, J.L., Hamilton, V.E., Ruff, S.W., Kieffer, H.H.,
Titus, T., Malin, M.C., Morris, R.V., Lane, M.D., Clark, R.L., Jakosky,
B.M., Mellon, M.T., Pearl, J.C., Conrath, B.J., Smith, M.D., Clancy,
R.T., Kuzmin, R.O., Roush, T., Mehall, G.L., Gorelick, N., Bender,
K., Murray, K., Dason, S., Greene, E., Silverman, S., Greenfield, M.,
2001. Mars Global Surveyor Thermal Emission Spectrometer experiment: investigation description and surface science results. J. Geophys.
Res. 106, 23823–23871.
Clancy, R.T., Grossman, A.W., Muhleman, D.O., 1992. Mapping Mars atmospheric water emission at 1.35 cm with the VLA. Icarus 100, 48–59.
Clark, B.C., 1978. Implications of abundant hygroscopic minerals in the
martian regolith. Icarus 34, 645–665.
Clark, B.C., Baird, A.K., Weldon, R.J., Tsusaki, D.M., Schnabel, L., Candelaria, M.P., 1982. Chemical composition of martian fines. J. Geophys.
Res. 87, 10059–10067.
Crisp, J.A., 1984. Rates of magma emplacement and volcanic output. J. Volcanol. Geotherm. Res. 20, 177–211.
Davies, D.W., Farmer, C.B., LaPorte, D.D., 1977. Behavior of volatiles in
Mars’ polar areas: a model incorporating new experimental data. J. Geophys. Res. 82, 3815–3822.
Drake, D.M., Feldman, W.C., Jakosky, B.M., 1988. Martian neutron leakage spectra. J. Geophys. Res. 93, 6353–6388.
Farmer, C.B., Doms, P.E., 1979. Global seasonal variation of water vapor on
Mars and the implications for permafrost. J. Geophys. Res. 84, 2881–
2888.
Feldman, W.C., Boynton, W.V., Tokar, R.L., Prettyman, T.H., Gasnault, O.,
Squyres, S.W., Elphic, R.C., Lawrence, D.J., Lawson, S.L., Maurice, S.,
McKinney, G.W., Moore, K.R., Reedy, R.C., 2002. Global distribution
of neutrons from Mars Odyssey. Science 297, 75–78.
Feldman, W.C., Mellon, M.T., Maurice, S., Prettyman, T.H., Carey, J.W.,
Vaniman, D.T., Bish, D.L., Fialips, C.I., Chipera, S.J., Kargel, J.S., Elphic, R.C., Funsten, H.O., Lawrence, D.J., Tokar, R.L., 2004a. Hydrated
states of MgSO4 at equatorial latitudes on Mars. Geophys. Res. Lett. 31.
Feldman, W.C., Prettyman, T.H., Maurice, S., Plaut, J.J., Bish, D.L., Vaniman, D.T., Mellon, M.T., Metzger, A.E., Squyres, S.W., Karunatillake,
S., Boynton, W.V., Elphic, R.C., Funsten, H.O., Lawrence, D.J., Tokar,
R.L., 2004b. The global distribution of near-surface hydrogen on Mars.
J. Geophys Res. 109.
Haberle, R.M., Jakosky, B.M., 1990. Sublimation and transport of water
from the north residual polar cap on Mars. J. Geophys. Res. 95, 1423–
1437.
Houben, H., Haberle, R.M., Young, R.E., Zent, A.P., 1997. Modeling the
martian seasonal water cycle. J. Geophys. Res. 102 (E4), 9069–9084.
Jakosky, B.M., 1983a. The role of seasonal reservoirs in the Mars water
cycle. I. Seasonal exchange of water with the regolith. Icarus 55, 1–18.
Jakosky, B.M., 1983b. The role of seasonal reservoirs in the Mars water cycle. II. Coupled models of the regolith, the polar caps, and atmospheric
transport. Icarus 55, 19–39.
Jakosky, B.M., 1985. The seasonal cycle of water on Mars. Space Sci.
Rev. 41, 131–200.
Jakosky, B.M., Barker, E.S., 1984. Comparison of groundbased and Viking
Orbiter measurements of martian water vapor: variability of the seasonal
cycle. Icarus 57, 322–334.
Jakosky, B.M., Carr, M.H., 1985. Possible precipitation of ice at low latitudes of Mars during periods of high obliquity. Nature 315, 559–561.
Jakosky, B.M., Christensen, P.R., 1986. Global duricrust on Mars: analysis
of remote sensing data. J. Geophys. Res. 91, 3547–3559.
Jakosky, B.M., Farmer, C.B., 1982. The seasonal and global behavior of
water vapor in the Mars atmosphere: complete global results of the
Viking atmospheric water detector experiment. J. Geophys. Res. 87,
2999–3019.
Jakosky, B.M., Haberle, R.M., 1990. Year-to-year instability of the Mars
south polar cap. J. Geophys. Res. 95, 1359–1365.
Jakosky, B.M., Haberle, R.M., 1992. The seasonal behavior of water on
Mars. In: Kieffer, H., Jakosky, B., Snyder, C., Matthews, M. (Eds.),
Mars. Univ. of Arizona Press, Tucson, pp. 969–1016.
Jakosky, B.M., Henderson, B.G., Mellon, M.T., 1993. The Mars water cycle
at other epochs: recent history of the polar caps and layered terrain.
Icarus 102, 286–297.
Jakosky, B.M., Henderson, B.G., Mellon, M.T., 1995. Chaotic obliquity and
the nature of the martian climate. J. Geophys. Res. 100, 1579–1584.
Laskar, J., Levrard, B., Mustard, J.F., 2002. Orbital forcing of the martian
polar layered deposits. Nature 419, 375–377.
Malin, M.C., Caplinger, M.A., Davis, S.D., 2001. Observational evidence
for an active surface reservoir of solid carbon dioxide on Mars. Science 294, 2146–2148.
McSween Jr., H.J., Grove, T.L., Lentz, R.C.F., Dann, J.C., Holzheid, A.H.,
Riciputi, L.R., Ryan, J.G., 2001. Geochemical evidence for magmatic
water within Mars from pyroxenes in the Shergotty meteorite. Nature 409, 487–490.
Mellon, M.T., Jakosky, B.M., 1993. Geographic variations in the thermal
and diffusive stability of ground ice on Mars. J. Geophys. Res. 98,
3345–3364.
Mellon, M.T., Jakosky, B.M., 1995. The distribution and behavior of martian ground ice during past and present epochs. J. Geophys. Res. 100,
11781–11799.
Mellon, M.T., Jakosky, B.M., Kieffer, H.H., Christensen, P.R., 2000. Highresolution thermal-inertia mapping from the Mars Global Surveyor
Thermal Emission Spectrometer. Icarus 148, 437–455.
Mellon, M.T., Feldman, W.C., Prettyman, T.H., 2004. The presence and
stability of ground ice in the southern hemisphere of Mars. Icarus 169,
324–340.
Mischna, M.A., Richardson, M.I., Wilson, R.J., McCleese, D.J., 2003. On
the orbital forcing of martian water and CO2 cycles: a general circulation model study with simplified volatile schemes. J. Geophys. Res. 108.
Mitrofanov, I., Anfimov, D., Kozyrev, A., Litvak, M., Sanin, A., Tret’yakov,
V., Krylov, A., Shvetsov, V., Boynton, W., Shinohara, C., Hamara, D.,
Saunders, R.S., 2002. Maps of subsurface hydrogen from the High Energy Neutron Detector, Mars Odyssey. Science 297, 78–81.
Mooney, R.W., Keenan, A.G., Wood, L.A., 1952. Adsorption of water by
montmorillonite. I. Heat of desorption and application of BET theory.
J. Am. Chem. Soc. 74, 1367–1371.
Pollack, J.B., Leovy, C.B., Greiman, P.W., Mintz, Y., 1981. A martian general circulation experiment with large topography. J. Atmos. Sci. 38,
3–29.
Richardson, M.I., Wilson, R.J., 2002. Investigation of the nature and stability of the martian seasonal water cycle with a general circulation model.
J. Geophys. Res. 107 (E5).
Smith, M.D., 2002. The annual cycle of water vapor on Mars as observed
by the Thermal Emission Spectrometer. J. Geophys. Res. 107 (E11).
Smith, D.E., 18 colleagues, 1999. The global topography of Mars and implications for surface evolution. Science 284, 1495–1503.
Mars low-latitude neutron distribution
Squyres, S.W., the Athena Science Team, 2004. Initial results from the MER
Athena science investigation at Gusev Crater and Meridiani Planum.
Lunar Planet. Sci. 35. Abstract 2187.
Titus, T.N., Kieffer, H.H., Christensen, P.R., 2002. Exposed water ice discovered near the south pole of Mars. Science 299, 1048–1051.
Ward, W.R., 1992. Long-term orbital and spin dynamics of Mars. In: Kieffer, H.H., Jakosky, B.M., Snyder, C.W., Matthews, M.S. (Eds.), Mars.
Univ. of Arizona Press, Tucson, pp. 298–320.
67
Webster, P.J., 1977. The low-latitude circulation of Mars. Icarus 30, 626–
649.
Zent, A.P., Quinn, R.C., 1995. Simultaneous adsorption of CO2 and H2 O
under Mars-like conditions and application to the evolution of the martian climate. J. Geophys. Res. 100, 5341–5349.
Zent, A.P., Haberle, R.M., Houben, H.C., Jakosky, B.M., 1993. A coupled subsurface-boundary layer model of water on Mars. J. Geophys.
Res. 98, 3319–3337.