some papers

I mentioned a few papers earlier, here are some references.

"Introduction: The phrase "liquid water is not stable
on present-day Mars" introduces many publications
on water-related features. While technically correct, the
statement is misleading. On Earth or Mars, water is
ordinarily metastable, slowly evaporating or freezing.
An exception is found on Earth when the relative humidity
reaches 100%. On Mars, water will crust over
with ice when the atmospheric pressure falls below
approximately 6.1 mbar.
It has long been known that water could conceivably
flow and ice could conceivably melt on Mars as a
transient event. Such an event need not be catastrophic,
as the re-freezing rates are on the scale of hours or
days.
This talk reviews reasonable spatial and temporal
scales for such melting and flowing events, and relates
them to plausible Martian conditions. It is shown that
seasonal accumulation of snow and ice on cold peaks
could melt and flow in the summer sun, explaining
gullies recently observed by Malin & Edgett [1]. Further,
it can be concluded that summer wetting may frequently
occur where seasonal ice is present."

http://www.lpi.usra.edu/meetings/lpsc2001/pdf/1364.pdf

"Introduction: The origin of the layering characteristic
of the Polar Layered Deposits (PLD) of Mars is
generally not thought to arise from the flow of the
water ice presumed (along with dust) to comprise
these layers, since rheological modeling indicates that
Mars is presently too cold to permit substantial ice
flow in the polar regions.
However, Martian obliquity deviates chaotically
from its current Earth-like value of 25, surpassing 45
within the last several Myr. Not only will the polar
regions receive additional insolation at these high
obliquities, but the resulting increase in H2O sublimation
from the ice caps will initiate a water vapor
greenhouse heating effect. Hence, surface and subsurface
temperatures will be elevated at high obliquity,
leading to dramatic increases in ice flow velocities."

http://www.lpi.usra.edu/meetings/lpsc2000/pdf/1571.pdf

"Selective linear erosion on Mars: A remarkable
characteristic of much of the record of "fluvial erosion"
revealed on Mars by Mariner 9, Viking and MGS imaging
is the selective distribution of martian fluvial
features, both on local and regional scales. Small valley
networks for instance display dendritic patterns that
leave relatively large internetwork and interbranch areas
on surrounding upland terrains very little dissected (a
characteristic among a few others often deemed "unusual"
by terrestrial standards), while the valleys themselves
are generally discrete and may display anywhere
from very subdued and poorly graded sections and profiles
to ones that are well-incised and graded. Larger
small valleys winding over greater distances across the
martian landscape, such as Ma'Adim Vallis, also present
a remarkable contrast between their discrete entrenchment
and the relatively sparse, selective, and
generally subdued erosion of surrounding uplands. The
tributary canyons on the north-facing south rim of Ius
Chasma provide yet another example of the apparent
selectiveness of "fluvial" erosion, in this case probably
among other contributions, interpreted on Mars.
This characteristic of martian valleys is important
because it is commonly invoked as a key argument for
favoring the release of underground water as the likely
fluvial erosional agent for the formation of many valleys
on Mars, over precipitation followed by surface
runoff which would be expected to have led to more
widespread, less selective fluvial erosion. We report
here on a suite of observations of landscapes of glacial
selective linear erosion made at and near Haughton
Crater on Devon Island in the Canadian High Arctic
where Mars analog field studies are currently in progress
under the auspices of the NASA Haughton-Mars
Project (HMP) [P1,P2]. Glacial selective linear erosion
on Devon Island has produced landscapes and specific
geologic features, including valley forms and adjacent
landscapes of little or no glacial erosion, that are morphologically
and contextually similar to possible counterparts
observed on Mars. I discuss these features and
will explore whether there might be a viable alternative,
i.e., an explanation at least as plausible as the
ubiquitous release of groundwater, for the pattern of
selective fluvial erosion observed on Mars."

"Landscape of glacial selective linear erosion on
Devon Island:
The landscape of Devon Island, like that of many
islands of the Candian Arctic Archipelago, is in effect a
textbook example of one of glacial selective linear
erosion [D,P3,P4]. Such a landscape is characterized
by localized, deep, and often sinuous troughs, separated
by vast tracks of plateau surfaces presenting only
a relatively sparse distribution of discrete networks of
meltwater channels and otherwise little modification. It
is generally thought that such landscapes develop beneath
ice sheets, with the troughs marking former ice
streams and the intervening plateaux marking areas of
slowly moving or cold-based ice [e.g., BE]. Positive
feedbacks between subglacial topography, ice thickness
and velocity, basal temperature and erosion rates are
ofetn invoked to explain trough location and formation.
The geologic features typifying glacial selective linear
erosion on Devon Island are examined below:
Glacial meltwater channel networks. The glacial
meltwater channel networks of Devon Island have been
considered previously and were discussed as possible
analogs for small valley networks on Mars [P3,P4]. In
the context of selective linear erosion, those reported
on Devon Island likely reflect the preferential erosion of
ice-marginal and/or subglacial bedrock in areas where
streams of glacial meltwater would converge and flow.
Surrounding areas may remain protected by a static ice
cover and hence experience contrastingly little erosion.

Glacial trough valleys.The larger scale trough valleys
of Devon Island form a network of deeply incised,
often winding, U-shaped and V-shaped "canyons"
gouging the surrounding plateau, in many places down
to the crystalline basement underlying the island's top
sequence of Paleozoic carbonate-dominated sediments.

The troughs concentrate along the coastal areas of
Devon, thereby revealing their probable origin as outlet
glacial troughs carved at the periphery of a broader regional
ice cover that rested on Devon Island during at
least the Last Glacial Maximum [D]. That even the Vshaped
winding trough valleys of Devon Island are
fundamentally glacial in nature is supported by the
combined observation of: a) the systematic transition of
the larger troughs to more classic U-shaped forms and

fjords downstream, b) the common lack of any apparent
deep-seated structural control in the distribution of the
troughs, c) the observation that the troughs currently
emerging from underneath the edge of the receding ice
cap in the eastern part of Devon Island are still partially
filled with ice displaying streaming flow lines and, in
the summer at least, often torrential subglacial meltwater
streams, d) the ubiquitous presence of preglacial
forms such as tors along the rim of the troughs, and e)
the frequent absence of any significant feeder stream and
the negligible role of sapping.
Landscapes of little or no glacial erosion.. Surrounding
uplands spared from glacial and meltwater
erosion (protected by the cold-based ice cover) may
display remarkably little evidence of any former glacial
occupation. Vast intervening areas of the Devon Island
plateau present such landscapes of little or no glacial
erosion (Figure 3).

http://www.arctic-mars.org/docs/03c.LPSC.pdf

" Approach: We investigate the possibility that these lakes are controlled
by meteorological rather than
geological factors. The model we adopt is an analog to the Antarctic dry
valley lakes. These lakes are perennially
covered with ice, but never freeze completely. Liquid water is maintained
beneath the ice by glacial melting from
nearby snow accumulations [2]. For this model to work, at least three
conditions must be satisfied: (1) there must
be a source of snow, (2) temperatures must exceed the freezing point at
least several days out of the year, and (3)
meteorological conditions must allow liquid water to be stable long enough
to flow into the lake. All three of
these conditions can be addressed with a general circulation model (GCM).

Interestingly, the places where the model predicts that the stability
conditions are satisfied correlates well those
places where Cabrol et al. find evidence for recent paleolakes. While the
correlation is intriguing, it does not in
itself show cause and effect as the first requirement - a source of snow -
has yet to be met. Certainly, at the present
time the tropics are desiccated [e.g., 6].
So how is it possible to get snow to accumulate in the tropics in relatively
recent times? We explore two
possibilities. First is the redistribution of oceanic water. The case for a
past ocean on Mars has been strengthened
by recent MGS observations [7], and may be an inevitable consequence of
progressive crustal assimilation of an
early surface reservoir [8]. However, the possibility is still controversial
and the timing is very uncertain.

So how is it possible to get snow to accumulate in the tropics in relatively
recent times? We explore two
possibilities. First is the redistribution of oceanic water. The case for a
past ocean on Mars has been strengthened
by recent MGS observations [7], and may be an inevitable consequence of
progressive crustal assimilation of an
early surface reservoir [8]. However, the possibility is still controversial
and the timing is very uncertain.
To investigate the potential fate of oceanic water, we have conducted
simulations assuming an ice covered ocean
exists in the northern plains at and below the -4km level. Preliminary
analysis of these simulations indicate that a
substantial hydrological cycle develops characterized in part by sublimation
over the ocean, southward
atmospheric transport, followed by considerable snowfall throughout much of
the southern hemisphere. Whether
the snow accumulates each year depends on the orbit parameters. For the
present parameters, snow accumulation
in the southern hemisphere occurs only in the polar region. For other
parameters snow can accumulate elsewhere
including Hellas, Argyre, and Tharsis. In the simulations conducted thus
far, it appears possible that an ocean-
driven hydrological cycle can provide enough snow to the lower latitudes to
be a significant source of water for
at least some of the paleolakes."

"The second possibility is based on an idea originally proposed by Jakosky
and Carr [9] where at high obliquity,
water subliming from the north polar cap will precipitate and stabilize at
low latitudes. Since that work, it is now
known that Mars obliquity is chaotic on time scales of 107 years and may
have been as high as 60 [10,11]. For
obliquities this high, the tropics receive less annual insolation than the
poles and ice might therefore be more
stable in the lower latitudes at these times.

We have simulated the water cycle at high obliquities and find that water
ice at the north pole is rapidly sublimed
into the atmosphere, transported southward, and some of it snows out and
accumulates in the low latitudes
confirming the original suggestion by Jakosky and Carr. However, the
accumulations are limited regionally and
do not cover all longitudes. As was the case with the ocean, enough snow
accumulates to provide a significant
source of water for palelolakes. Thus far, the simulated snow accumulation
distribution that best matches the
paleolake distribution is for the case when perihelion occurs close to
northern summer solstice.
Conclusion: There appears to be at least two mechanisms by which large
amounts of ice can accumulate at non-
polar latitudes. One involves the redistribution of oceanic water and the
other involves obliquity variations. It i s
important to note that the latter has a solid theoretical foundation and is
much less speculative than the ocean
hypothesis. Thus, our simulations suggest that at the very least, the
surface water ice inventory on Mars has been
- and will be - cycled between the poles and the tropics in rhythm with its
orbital variations."

http://www.lpi.usra.edu/meetings/lpsc2000/pdf/1509.pdf

"Introduction: Many fluvial features on Mars [1]
indicate past active hydrologic cycle. We investigate
atmospheric water cycle on an idealized 'land planet,' a
planet without oceans on the surface, using a general
circulation model. The main target of this study is to
clarify the condition under which wet surface can be
maintained.
We focus on the effect of change in spin-axis inclination
and the average surface. The inclination of the
present Mars' spin-axis is close to that of the Earth.
However, the Mars' inclination is believed to be
changed very much from 0 to 60[2]. Change of inclination
should have profound effect on climate.'

"Method: We made an idealized land-planet model
by removing the oceans, topographies and vegetations
from the CCSR/NIES AGCM5.4g, which have been
developed for the Earth by the Centre for Climate System
Research, University of Tokyo and the National
Institute for Environmental Research. We assume 1bar
CO2 atmosphere and a circular orbit.
We performed two series of experiments: series A
simulates a 'warm' land planet, that is, the surface temperature
exceeds the freezing point in summer at least
at the subsolar point. On the other hand series B
simulates a 'cold' land planet on which the temperature
is always below the freezing point everywhere on the
planet. In both series, we changed the inclination of
the spin axis from 0 to 60.
In both series, we ignore the surface and underground
water transport. We use a bucket model with
the saturation depth of 10cm for ground water calculation.
The total depth of the bucket is 10m. Ice and
snow albedo model are the same with that for the
Earth. In both series, the initial condition is a steady
state circulation with uniform distribution of ground
water.
Results: Atmospheric circulation and hydrologic
cycle approach steady state within 10 sidereal years. In
the steady state, the annual precipitation corresponds to
the annual evaporation everywhere on the surface.
Figure 1 shows the some typical patterns of zonal-mean
annual precipitation (~the annual evaporation) at the
steady state. The inclination clearly affects the hydrological
cycle. Though inclination has strong influence on the overall planetary
climate, its effect is different
for series A and B.
In series A, the precipitation pattern is clearly different
between the cases with smaller inclinations
(30). In the following,
we call the former as an 'upright regime' and the latter
as an 'oblique regime'. In the upright regime, low latitude
areas are dried because of atmospheric water
transport to high latitude area. Eventually, ground
water is lost and no precipitation occurs in the low
latitude areas. Precipitation occurs only in mid- to
high-latitude areas. On the other hand, in the oblique
regime, the low latitude areas have precipitation in
spring and fall. In addition, the summer surface temperature
is much higher than that of the upright cases
for the same atmospheric mass and annual innsolation.
In series B, the planet is entirely covered by snow
or ice, namely, it is a 'snow ball' planet. In the following
we call this as a 'frozen regime'. The surface temperature
is extremely low owing to strong ice albedo
feedback. Essentially no horizontal transport of water
occurs irrespective of the spin-axis inclination. Both
the precipitation and evaporation are just locally balanced
and have maximum at the subsolar point. Its
magnitude is about one-thousands of the results of series
A.
Thus, we found three different regimes: 1. Frozen
regime for wet and cold cases, 2a. Upright regime for
wet, warm and small inclination cases, and 2b. Oblique
regime for wet, warm and large inclination cases. Determining factors of
three climate regimes:
The difference between the upright and oblique regimes
comes from the relative magnitude of the
Hadley-cell width and the Tropic latitude
(=inclination). The width of the Hadley cell is always
about 30. If the width of the Hadley cell is larger than
the Tropic latitude, the subsolar point is always in the
Hadley cell. Hence, low latitude areas are always
warmer than mid- to high-latitude areas and water is
transported from low to high latitude areas. Thus, low
latitude areas are dried. This is the upright regime. On
the other hand, if the width of the Hadley cell is smaller
than the Tropic latitude, the subsolar point is located
outside the Hadley cell in summer. Hence, water istransported from
high-latitude subsolar point to the
Hadley region in summer. Thus, in the low latitude
area is kept in wet condition. This is the oblique regime."

"On the other hand, in the frozen regime, water circulation
is just locally balanced and insensitive to the
inclination. This is a result of extremely low surface
temperatures. At low temperature the atmosphere
contains very small amount of water because of low
saturation vapor pressure, and the atmospheric motion
does not transport water. Extremely low temperature is
caused by the occurrence of a snow ball planet, which
seems an inevitable consequence of wet and cold case,
because the surface snow/ice is supplied by freezing of
water vapor evaporated from the solar-heated surface.
Thus, even if we artificially remove the snow/ice from
the surface, the surface is soon covered by snow, as far
as the surface is kept wet.
Both the width of the Hadley cell and occurrence of
snowball planet are insensitive to the atmospheric
composition and mass. Therefore, occurrence of these
three regimes should be general features irrespective of
detail of the paleo-atmosphere of Mars.Discussion and Implication:
Inclination of the
spin axis and the average temperature of the planet
affect the precipitation distribution. In particular,
transport of water to the low latitude area occurs only
in the oblique regime.
Cold and wet Mars. If the paleo-Mars was wet and
cold, it should be in the frozen regime and the hydrologic
cycle should be locally closed. Small networks
may be formed, if it is supported by very local circulation
probably controlled by local topography. Also,
ground ice may melt by geothermal heating and flow.
However, this situation does not seem adequate for
sustaining flow in large valley networks.
Warm and wet Mars. If the paleo-Mars was wet and
warm, global scale hydrologic cycle supports precipitation
at mid- to high-latitude. If the inclination of the
spin axis was small, it should be in the upright regime,
and the low latitude areas should be dry out without
precipitation. Under such situation, formation of valley
networks at low latitude areas is unlikely. Large scale
low latitude valley networks can be formed only when
the paleo-Mars was in the oblique regime. It should be
noted that the summer surface temperature for the
oblique regimes is much higher than that of the upright
regimes.
Our results suggest that the high obliquity period is
a good candidate for the era of fluvial feature formation. formation.
Obviously, it does not mean all the fluvial features
are formed during high obliquity period.
One major assumption that may affect the results is
simplification of topography and exclusion of
oceans[3]. Ocean obviously affects the hydrologic
cycle. For example, present precipitation at the equatorial
region of the Earth is sustained by low latitude
oceans. If an ocean exists on northern plain of Mars, it
may induce a Monsoon-like circulation. Thus, in the
future study, we have to investigate the effects of
oceans and topographies."

http://www.lpi.usra.edu/meetings/lpsc2002/pdf/1731.pdf