Earth and Planetary Science Letters 294 (2010) 332–342
Contents lists available at ScienceDirect
Earth and Planetary Science Letters
j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / e p s l
Kilometer-thick ice accumulation and glaciation in the northern mid-latitudes of Mars:
Evidence for crater-filling events in the Late Amazonian at the Phlegra Montes
James L. Dickson a,⁎, James W. Head a, David R. Marchant b
a
b
Department of Geological Sciences, Brown University, Providence, RI 02912, USA
Department of Earth Sciences, Boston University, Boston MA 02215, USA
a r t i c l e
i n f o
Article history:
Accepted 25 August 2009
Available online 25 September 2009
Keywords:
Mars
glaciation
climate change
water
a b s t r a c t
Amazonian non-polar ice deposits on Mars record periods and events when the climate differed substantially
from that of today. Particularly evident are examples of ice-rich deposits in the martian mid-latitudes (lobate
debris aprons, lineated valley fill, and concentric crater fill). Uncertain, however, is the amount of ice
remaining in these deposits today, and the thickness of ice that might have existed when they formed. Here,
we use HRSC, CTX and HiRISE imagery and MOLA topographic data to document an occurrence of concentric
crater fill within which the past minimum volume of ice can be constrained. An ~ 8 km impact crater is
superposed on the rim of a ~ 32 km impact crater near the contact between the Phlegra Montes and the
Vastitas Borealis Formation in the northern mid-latitudes of Mars. We find evidence for flow from the larger
crater into the perched smaller crater that indicates an earlier period of significant ice accumulation and
glaciation within this double crater. Lobate ridges observed outside of the perched younger crater suggest
that ice filled and overtopped the crater rim, providing minimum estimates of ice thickness and volume
within the system. Glacial ice must have been at least ~ 1000 m thick to overtop the rims of both craters and
induce gravitational flow onto the surrounding plains, with a minimum volume of ice of ~750 km3. This is
the first volumetric measurement of this kind on Mars for concentric crater fill craters, and the thickness is
comparable to that measured in a lineated valley fill glacial system along the dichotomy boundary at a
similar latitude. We also document late-stage episodes of more localized glacial flow that include ridges on
valley walls that we interpret as late-stage glacial high-stands, and concentric crater fill (CCF) that
characterizes most of the present-day crater floor. Similar deposits in a crater ~ 60 km to the northeast
suggest that such episodes were at least regional in nature. This sequence provides evidence for significant
spin-axis/orbital parameter-driven shifts in the Late Amazonian climate of Mars and suggests that regional
ice sheets may have existed in the mid-latitudes of Mars within the last several hundred million years.
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
Global mapping of the surface of Mars using Viking Orbiter data
revealed a suite of enigmatic Late Amazonian landforms in the midlatitudes of each hemisphere, including lineated valley fill (LVF), lobate
debris aprons (LDA) and concentric crater fill (CCF) (Squyres, 1978,
1979). Morphology indicative of viscous flow and the latitudedependence of these features implicated ice-related processes in their
formation (Squyres, 1979; Lucchitta, 1981), and it was proposed that
small amounts of water vapor condensing within pore space mobilized
dry material to produce the observed landforms (Squyres, 1978;
Lucchitta, 1984).
As higher resolution data have been obtained, these features have
been studied in detail not available at the time of their discovery, adding
insight into their formation and providing implications for the Late
Amazonian climate within which they formed (Pierce and Crown, 2003;
⁎ Corresponding author.
E-mail address: jdickson@brown.edu (J.L. Dickson).
0012-821X/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.epsl.2009.08.031
Head et al., 2005; Li et al., 2005; Head et al., 2006a,b; Dickson et al., 2008;
Holt et al., 2008; Head et al., 2010-this issue; Plaut et al., 2009). Regional
mapping using the High Resolution Stereo Camera (HRSC [Neukum et al.,
2004; Gwinner et al., 2005; Scholten et al., 2005]) on Mars Express
(MEX), together with other complementary data sets, has revealed
valleys along the dichotomy boundary that show evidence for integrated
networks of LVF (Head et al., 2005; Head et al., 2006a,b), consistent with
the flow of glacial ice. Measurements using the SHAllow RADar
instrument (SHARAD) on the Mars Reconnaissance Orbiter (MRO) of
LDAs in each hemisphere detected subsurface reflections, with the LDAs
having dielectric properties consistent with an ice-dominated composition on contemporary Mars (Holt et al., 2008; Plaut et al., 2009). Dickson
et al. (2008) used HRSC elevation data to document an LVF lobe with an
upslope profile that has been stranded within a box canyon at the contact
of Coloe Fossae and Protonilus Mensae along the dichotomy boundary,
with measurements of downwasting suggesting a previous ice thickness
of at least ~920 m. While networks of LVF (Head et al., 2005; Head et al.,
2006a,b) and ice-rich LDAs (Pierce and Crown, 2003; Head et al., 2005; Li
et al., 2005; Holt et al., 2008; Plaut et al., 2009) are observed in each
J.L. Dickson et al. / Earth and Planetary Science Letters 294 (2010) 332–342
hemisphere, the question remains as to whether the glacial high-stand
interpreted by Dickson et al. (2008) was an isolated occurrence or
whether other regions show similar deposits consistent with glaciation
and kilometer-scale ice thicknesses. In this paper, the high resolution and
large spatial footprint of HRSC (Neukum et al., 2004; Scholten et al., 2005)
provides context for the analysis of a second example of a hanging valley
filled with glacial-like lobes of comparable inferred past thickness at the
contact between the Phlegra Montes and the northern lowlands. HRSC
coverage of nearby craters reveals similar landforms that suggest that
such massive glaciation was regional, not local.
2. Context and background
The Phlegra Montes trend north–south from the easternmost extent
of the Elysium rise northwards into the northern lowlands, from ~30°N
to ~52°N (Fig. 1A) (Greeley and Guest, 1987; Tanaka et al., 1992). The
333
flanks of these mountains frequently exhibit LDAs (Head et al., 2010-this
issue; Safaeinili et al., 2009), and impact craters in the region host CCF
(Squyres, 1979; Levy et al., 2009). To the northwest of the Phlegra
Montes, at the contact with the Vastitas Borealis Formation (42.9°N,
157.8°E), is a ~32 km diameter impact crater with a smaller (~8 km
diameter) impact crater superposed on its northeastern rim (Fig. 1B).
The topography is such that the floor of the younger, superposed crater is
elevated with respect to the floor of the larger, older crater (Figs. 1b, 2A).
While the rim of the larger crater is still preserved, the walls and floor
have been heavily modified and exhibit the well-defined concentric
ridged terrain that is characteristic of CCF (Fig. 2). The wall itself is heavily
dissected with broad ~250 m wide valleys incised into it around much of
the crater (Fig. 2). The easternmost wall of the crater shows evidence for
coalescing of flows, constriction between obstacles (Figs. 2, 6), and a
series of convex up lobes on the floor of the crater (Fig. 2). The center of
the crater floor is characterized by cuspate pitting and several “ring-mold
Fig. 1. A) MOLA topography over MOLA shaded relief of part of the northern hemisphere of Mars, centered at 160°E. B) MOLA topography over HRSC image 2841_0000, showing the
study area at the contact between the Phlegra Montes to the east, and the Vastitas Borealis Formation to the west.
334
J.L. Dickson et al. / Earth and Planetary Science Letters 294 (2010) 332–342
Fig. 2. A) CTX mosaic of images P01_001553_2232 and P01_001619_2232, showing the ~ 8 km impact crater superposed on the ~ 32 km crater. The floors of both craters show
evidence for significant modification of the surface. B) MOLA topography (color) over CTX data. MOLA tracks that intersect two craters are shown. C) MOLA track 17237. Circled
numbers indicate portions of the crater described in the text. D) MOLA track 13024.
craters” (Kress and Head, 2008), unusual impact crater structures that
suggest the presence of buried ice at the time of and subsequent to the
impact event.
Mars Orbiter Laser Altimeter (MOLA) profiles show that the floor of
the crater is asymmetric and generally slopes poleward (Fig. 2B–D). We
subdivided the interior of the crater into five distinct segments from
south to north: 1) the steep upper part of the southern interior wall that
descends over 800 m from the sharp rim crest (~−3060 m) to the base
of the wall; 2) a shallow depression, about 4 km wide, that corresponds
to a zone of radial ridges and depressions; 3) a convex-upward portion
of the floor, about 7.5 km wide, characterized by concentric ridges and
troughs, and pitted terrain in its down slope portion (the lowest portion
of the floor, at ~−4300 m, separates this segment from the next); 4) a
slightly convex-upward portion of the floor about 8 km wide and
sloping to the south that is characterized by concentrically textured and
ridged material; 5) the northern inner wall of the crater, about 5 km
wide, rising about 1320 m to the crater rim crest at −2880 m (Fig. 2).
These topographic attributes, along with the morphology of the crater
interior, are all indicative of concentric crater fill as initially described
and mapped by Squyres (1979).
The deposits observed on the floor of the crater represent only the
final phase of modification of the crater interior. To decipher the extent of
modification that occurred before this most recent phase, we examined
the superposed crater on the northern rim of the larger crater, and the
walls and exterior of the larger crater to study the crater filling processes.
3. Detailed geology
3.1. Superposed smaller crater
The floor of the superposed smaller crater (~8 km in diameter)
exhibits a texture similar to that observed in other LVF/LDA terrains
J.L. Dickson et al. / Earth and Planetary Science Letters 294 (2010) 332–342
on Mars (Fig. 3) (Pierce and Crown, 2003; Head et al., 2005; Head
et al., 2006a,b; Dickson et al., 2008; Head et al., 2010-this issue). A
series of concentric, symmetric, arcuate lineations trend from the
southern extent of the crater floor to the northern crater wall. This
crater has been imaged several times by HRSC (Neukum et al., 2004)
and twice by CTX, allowing for the generation of a stereo anaglyph at
~ 6 m horizontal resolution (Fig. 3B). In stereo, these lineations are
seen to define several broad ridges. Three ridges are clearly defined,
with the outermost having a width of ~ 850 m, the middle a width of
~ 420 m, and the innermost a width of ~190 m. Ridge crests are
typically smoother than the troughs that separate them, and smaller
but pronounced northwest–southeast trending ridges that characterize the southern portion of the crater floor are more diffuse at the
distal margins of the ridges. Outside of the ridges, the base of the
crater wall is mantled with smooth material with scalloped and
cuspate margins facing the crater floor. Flow lineations and lobate
margins have been used to identify source regions and flow directions
elsewhere on Mars for LVF (e.g. Head et al., 2005, 2006a,b; Dickson
et al., 2008; Head et al., 2010-this issue). The arcuate shape of the
ridges (trending to the north) on the crater floor suggest that the
majority of flow emanated from the south, where the crater rim has
been removed. This is in contrast to typical crater-filling patterns,
where material is thought to be mobilized from the crater walls and
flow towards the center of the crater to produce CCF patterns
(Squyres, 1979; Levy et al., 2009), as observed in the most recent
deposits within the larger crater. Had this material emanated from the
walls of the perched smaller crater, the arcuate ridges would trend
towards the center of the crater, or one would expect continuous
lineations from the walls of the smaller crater southwards towards the
floor of the larger crater, had there been enough ice to sustain flow to
the south. Evidence for mobilization of material from the crater walls
within the younger superposed crater is absent at HRSC or CTX
resolution. The western margin of the outermost arcuate ridge does
appear to show some evidence of modification (Fig. 3A), possibly from
material coming off of the western crater wall, but these relationships
335
are presently unclear and we interpret the dominant flow pattern
responsible for the concentric ridges to be from south to north.
3.2. Contact between the two craters
To test the interpretation that material flowed from south to north
into the smaller crater, we investigated the contact between the two
craters. This contact is observable in several MOLA tracks as a ~8° southfacing 500 m-scarp (Fig. 2B, D). At HiRISE resolution, the large ridges in
the perched crater are observed to drape over this scarp where it
encounters a thin (meters-across) linear but discontinuous ridge that
runs perpendicular to the ridges (easternmost portion of Fig. 4A).
Downslope of this thin ridge, the terrain transitions to a texture with
fine-scale lineations trending downslope towards the crater floor
(“lineated terrain”). This texture is common along the interior and
exterior walls of the larger crater and is composed of parallel linear
ridges and grooves, no more than ~10 m in width. This lineated terrain
emanates from valleys incised into the scarp face, and exhibits a convexup profile within the valleys themselves. At the base of the slope, the
lineations are superposed by distinct mounds of polygonally-patterned
ground, seen primarily at the mouths of the valleys (Fig. 4A). The
lineated terrain continues further downslope from the polygonallypatterned mounds, becoming less pronounced at its distal ends.
On the walls of these valleys, HiRISE reveals a distinct ridge tens of
meters above the present-day surface of valley-fill material (Fig. 4B, C).
This ridge can be traced along the walls of several of the parallel valleys,
and the wall below the ridge shows a slightly lower albedo than that
upslope of the ridge. Dickson et al. (2008) documented similar
lineations in the Coloe Fossae-Protonilus Mensae region in CTX data,
and a subsequently acquired HiRISE stereo pair (PSP_008809_2215 and
PSP_009455_2215) shows they are narrow ridges similar to those
observed in this study.
Further downslope, the unit upon which the lineated terrain occurs
terminates along a broadly curvilinear and locally cuspate equatorfacing scarp (Fig. 4A). Downslope from this scarp, the terrain is
Fig. 3. A) Part of CTX image P01_001619_2232, showing the superposed ~ 8 km crater. Material appears to have flowed in from the south, filled the crater, and breached the rim in at
least two locations along the northern rim. B) CTXstereo anaglyph comprised of P01_001619_2232 and P01_001553_2232. Note that the two well-defined lobes on the surrounding
plains to the north emanate from the depressed portions of the crater rim. C) Sketch map of the observed features.
336
J.L. Dickson et al. / Earth and Planetary Science Letters 294 (2010) 332–342
Fig. 4. A) Part of HiRISE image PSP_007500_2235 (Context provided in Fig. 3A). Small valleys contain lineated terrain that is found on convex-up material that trends downslope. At
the mouth of each valley is a polygonal mound of material. B). Detail of HiRISE image, showing a marginal ridge on the wall of a valley hosting lineated terrain. C) Detail of HiRISE
image showing a similar ridge in a separate valley.
characterized by small-scale polygonally patterned ground that covers
and embays portions of a broad unit of hummocky terrain on the crater
floor. This hummocky unit is adjacent to the central crater floor, which
contains a complex assortment of rimless depressions, heavily modified
impact craters, locally linear ridges frequently with central troughs
(ridges are ~200 m long at most), linear dune fields, small pit chains, and
a low-albedo deposit on the northeastern portion of the crater floor.
Fresh craters, even at HiRISE scale, are extremely rare, suggesting a
young surface age, easily eroded surface material, or both.
3.3. Surrounding plains
To determine the amount of material that flowed into the
superposed smaller crater from the south, we used HRSC and CTX
image data along with stereo topography (Fig. 3B) to study the
preserved northern walls of the crater and the adjacent plains for
evidence of overtopping. While several MOLA orbits cross the rim of
the crater (e.g. Fig. 2), significant gaps in the track data have prompted
us to use high spatial-resolution (~6 m/px) CTX stereo anaglyphs to
decipher detailed topographic relationships. Unless otherwise stated
in this section, references to topography are derived from this latter
data set.
The northern rim of the smaller crater shows an extensive history
of modification and degradation. MOLA track 13024 (Fig. 2B) shows
the rim to be ~410 m above the crater floor at its northernmost extent,
but a detailed study of the rim using CTX stereo reveals that the rim
shows a wide variance in elevation. Directly to the east of this
measurement, the rim is depressed for a distance of ~3 km (Fig. 3B, C).
J.L. Dickson et al. / Earth and Planetary Science Letters 294 (2010) 332–342
A similar depressed segment is found along the northwestern rim that
extends for ~2.1 km (Fig. 3B, C). The base of the exterior crater wall at
these locations is truncated in comparison to the wall below
preserved portions of the rim (Fig. 3A).
On the surrounding plains at both of these locations, pronounced
north-trending lobate ridges (continuous ridges with an arcuate
planform) are observed emanating directly from the depressed
portions of the crater rim (Fig. 3). For the northern example, a
mantled but well-defined lobate ridge extends radially from the crater
rim for ~ 2.5 km. The lobate ridge is ~ 750 m across at the base of the
crater wall, and the ridges that delineate the lobate margin are 200 m
across and are less well-defined at the lobe's distal margin. Heavily
subdued concentric ridges are observed ~ 700 m outside of the welldefined lobe, and these ridges directly correspond to the margin of the
depressed portion of the crater rim (Fig. 3B). For the northwestern
depressed portion of the rim, another mantled but well-defined lobate
ridge is observed extending to the north. This lobate ridge is ~1.1 km
across at the crater wall, and the ridges that bound the lobe are
~ 250 m across and also are not as well-defined towards their distal
margins. At the terminus of the lobate ridge, a more subdued lobate
ridge is observed ~ 600 m downslope. This lobate ridge is concentric
with the well-preserved lobate ridge.
The correlation between the lobate ridges on the plains and the
depressed portions of the crater rim provides evidence that material
entered the smaller crater from the south, filled the crater and
overtopped the rim in these two location. The alternate hypothesis,
that flow within the system has only been from north to south, does
not account for these features. MOLA tracks were not obtained at
either of the depressed portions of the crater rim, but nearby tracks
provide an upper bound (maximum elevation) of the rim at these
locations. For the northern depressed portion of the rim, nearby track
13024 (Fig. 2D) determined a rim elevation of − 3190 m. For the
northwestern example, track 11867 obtained data directly at the
eastern margin of the depressed portion of the rim, and found a rim
height of −3173 m. During the most recent phase of crater-filling,
material reached this approximate elevation before overtopping the
crater rim and flowing onto the exterior plains.
Using this evidence for crater filling and overtopping in the
superposed crater, we analyzed in detail the margins of the exterior
crater wall for the larger crater to look for evidence of overtopping
and flow on the adjacent plains. On the northwestern rim of this crater
(Fig. 2A,B), individual lobate ridges similar to those observed on the
plains exterior to the superposed crater are observed to trend radially
from the larger crater towards the northwest (Fig. 5). Based upon
high-resolution data acquired by CTX, these lobate ridges also appear
to emanate from depressed portions of the crater rim and extend for
~ 3 km. The lobate ridges are mantled, but still allow for approximate
measurements: the most prominent lobate ridges are ~1.1 km across
at the entire lobe at its widest, and the ridges themselves are ~ 300 m
across. These are nearly identical in size and morphology to the lobate
ridges observed on the outer margins of the superposed crater to the
north. Concentric lobate ridges are not observed beyond the margins
of the proximal lobate ridges though the terrain has been significantly
mantled and subdued. Much of the exterior walls of the larger crater
are characterized by the same lineated terrain that is observed along
the interior crater wall and the scarp that marks the contact between
the two craters (Fig. 4A). The eastern exterior wall, however, shows a
broadly sinuous but locally lobate ridge that extends parallel to the
crater rim for ~ 20 km (Fig. 6). The ridge itself is heavily pitted and
dissected and protrudes through the stippled terrain that characterizes the surrounding plains to the east. This portion of the exterior
rim is on the opposite side of the crater rim from the most welldeveloped lobate flows on the crater interior (Fig. 6). The exterior
crater wall itself is heavily mantled by smooth material and does not
show significant evidence for downslope movement of material,
though several localized units of lineated terrain are observed
337
Fig. 5. Part of CTX image P01_001619_2232, showing ridges on the plains exterior to the
larger, older crater (Context provided in Fig. 2A).
emanating from broad valleys incised into the exterior crater wall.
The crater rim itself is well-preserved, and it remains unclear whether
the sinuous ridge on the exterior plains is (1) evidence for overtopping of the crater rim, (2) a deposit formed from localized
mobilization of material on the exterior crater rim, or (3) remnant
topography from proximal ejecta deposits that has been heavily
mantled.
4. Interpretation
The large crater within this study area shows the classic surface
morphology and topographic asymmetries of concentric crater fill
(CCF) in the mid-latitudes of Mars, interpreted to form from icerelated processes (Squyres, 1979; Levy et al., 2009). The unique
geometry of a perched crater on the northern rim and the interactions
on the surface between these craters provides a window into the
geologic history of this region that precedes the latest episode of CCF
activity. Here, we use these observations and the stratigraphic
relationships among the various units to propose a sequence of
events that appears to require significant amounts of glacial ice to
produce the observed landforms.
4.1. Initial formation of two-crater system
The initial impact event that formed the large underlying crater
occurred at the contact between the Phlegra Montes to the east and
the Vastitas Borealis Formation to the west (Fig. 1). Phlegra Montes is
a region where mapping has shown local concentrations of LVF, LDA
and CCF (Squyres, 1979), all of which are generally interpreted to
arise from ice-related processes, either by ice-assisted creep of relatively dry material (Squyres, 1979; Lucchitta, 1981), or cold-based
debris-covered glaciers in the Late Amazonian (e.g. Head et al., 2005,
2006a,b; Dickson et al., 2008; Holt et al., 2008; Head et al., 2010-this
issue; Plaut et al., 2009). The Hesperian-aged Vastitas Borealis
Formation (Greeley and Guest, 1987; Tanaka et al., 1992) covers the
majority of the northern plains of Mars, and in this region HRSC data
338
J.L. Dickson et al. / Earth and Planetary Science Letters 294 (2010) 332–342
Fig. 6. The eastern rim of the older, larger crater (part of CTX image P01_001553_2232) (Context provided in Fig. 2B). Material along the interior rim converged and constricted
between obstacles to flow onto the crater floor, leaving concentric ridges. The exterior crater wall is characterized by a sinuous ridge that trends parallel to the wall itself.
show a high concentration of pedestal craters, where impact ejecta
has been armored relative to the surrounding plains (Fig. 1B). Analysis
of marginal pits on pedestals in the nearby Utopia Planitia provides
new evidence that implicates ice as the material underlying the
pedestal surface (Kadish et al., 2008), suggesting that ice (of unknown
thickness) has been areally expansive in this region during the
Amazonian period of martian geologic history. The westernmost
ejecta deposits from this impact event are still preserved on the
surface, while all other ejecta has been mantled or removed. Due to
the extensive resurfacing that has occurred within and around this
crater, reliable crater size-frequency information is not available, and
a formation age is not yet known, though it cannot predate formation
of the Vastitas Borealis Formation in the Hesperian.
Superposed on the northeastern rim of this larger crater is an
~ 8 km impact crater. The ejecta deposit of this crater has mostly been
removed, although its distal northeastern margin is observed in CTX
data beneath a thick mantling unit. The southern rim of this crater has
been completely removed, while the rest of the rim remains intact
(Fig. 3).
4.2. Filling of larger crater and flow north into smaller crater
The two craters are presently separated by a ~ 8° south-facing
scarp (Fig. 2). Draped over this scarp are the southernmost extents of
the lobate ridges that trend north onto the floor of the younger crater
(Fig. 3). This draping relationship and the north-trending convex-up
lobes provide strong evidence that material flowed from the larger
crater into the smaller crater, reversed from what would be expected
from the current topographic relationship. For this to occur, material
must have filled the larger crater to the elevation of the floor of the
smaller crater (~ −3700 m). Using MOLA gridded data, we calculated
a volume of the larger crater beneath the −3700 m contour of
~ 265 km3 and a minimum thickness of ~ 640 m of crater-filling
material. We interpret these measurements to be the minimum
thickness and volume of material necessary to induce gravitational
flow into the smaller crater to the northeast.
4.3. Flow of material on the exterior plains
Evidence is found on the external margins of the smaller crater for
overtopping of the crater rim and flow of material onto the adjacent
plains (Fig. 3). These lobate ridges emanate from depressed portions
of the crater rim, suggesting that the material that flowed into the
crater from the south overtopped the rim and became concentrated at
these locations. Similar lobate ridges are observed on the outside of
the northwestern crater rim, also emanating from depressed sections
of the rim (Fig. 5). If this interpretation is correct and material
overtopped the crater rim at multiple locations, this significantly
increases the minimum amount of material that filled this two-crater
system. We used HRSC and high-resolution CTX imagery to map the
crater rims of both craters and, using MOLA data, calculated the
present-day volume of the two-crater system. A volume of ~ 750 km3
of material and a minimum thickness of ~1000 m must have been
present for material to flow in the observed patterns. Provided that
the floor of the larger crater has been significantly modified by CCF,
this volume is interpreted to be a minimum volume necessary and is
likely to have been greater than this measurement suggests.
After overtopping the rims of both craters, conditions at the
surface dramatically changed and large amounts of material were
removed to reveal the two-crater geometry that we observe today.
There is no evidence of meltwater distributary features or deposits
(e.g., fluvial channels, deltas, fans) that would lead to the interpretation of wet-based glacial activity. Instead, we interpret the observed
features to be consistent with cold-based glacial behavior, similar in
nature to that proposed by Dickson et al. (2008) for a similar stranded
lobe of LVF located along the dichotomy boundary in Coloe Fossae/
Protonilus Mensae. As cold-based glaciers are welded to the surface
and deform internally (Benn and Evans, 1998), they are capable of
flowing over terrain without significantly eroding the substrate. As
the climate changed and accumulation waned, the ice receded in
thickness and the rims of the craters became exposed, leading to the
shedding of debris onto the glaciers and the development of
supraglacial sublimation lags or tills. This debris would form the
series of lobate ridges on the exterior plains of each crater, and on the
floor of the smaller crater (Fig. 3).
As the ice thickness decreased and the glacial system underwent
recession, there appears to have been enough ice at certain locations
along the ~ 8° scarp face that separates the two craters to initiate small
amounts of backflow southward towards the center of the larger
crater (Fig. 7). This flow reversal would represent the waning phase of
glaciation as the system as a whole recessed. Along other portions of
the scarp face, the lobate ridges that extend northward into the
J.L. Dickson et al. / Earth and Planetary Science Letters 294 (2010) 332–342
339
Fig. 7. CTX image of the scarp that represents the contact between the two craters (part of CTX image P01_001553_2232; context provided in Fig. 3A). Deposits from the perched
crater are draped over the scarp in some places, while local areas on the eastern portion of the scarp show evidence for flow reversals towards the south.
smaller crater are draped over the scarp face with very little
modification evident (Figs. 3, 4A). The minimum thickness necessary
to fill the two-crater system with ice (~1000 m) is comparable to that
calculated for a similar system at a similar latitude along the
dichotomy boundary by Dickson et al. (2008) (~920 m). While the
dichotomy boundary example occurred within a network of LVF
(Dickson et al., 2008), the constrained nature of these units within a
two-crater system allows for the first volumetric measurements of
minimum ice necessary at this location. The very large amounts of
glacial ice necessary to fill this two-crater system (at least ~ 750 km3)
suggest that significant accumulation occurred in this region and that
Phlegra Montes, this crater in particular, and other localized traps
(Safaeinili et al., 2009) hosted significant reservoirs of water ice
during periods of the Amazonian when surface conditions were
different from today.
4.4. Formation of concentric crater fill and localized episodic flow
Craters exhibiting CCF are common in the Phlegra Montes region
(Squyres, 1979). The larger of the two craters in this study shows
classic landforms consistent with those mapped using Viking data by
Squyres (1979). Concentric ridges and grooves parallel to the crater
rim, often lobate when viewed in detail, characterize most of the
crater floor and extend ~10 km into the crater. This texture is more
apparent in the southern portion of the crater, and the asymmetric
(pole-facing) profile of the crater floor observed by MOLA (Fig. 2)
argues for significant amounts of material transport from the southern
crater rim towards the center of the crater late in the history of the
deposit evolution. Given the lobate nature of these ridges, together
with the supporting evidence for more extensive glaciation in this
region, we propose a model of cold-based glacial flow and deposition
to explain the late-stage CCF within the larger crater. We interpret the
CCF texture to be representative of the waning phase of this extensive
glaciation. As regional ice deposits thinned, the crater rim crests and
walls became exposed, providing rocky debris to cover the glaciers.
Once debris was available to create a protective layer, rates of
sublimation of ice decreased, and the late-stage deposits in crater
interiors were preserved (e.g., Helbert et al., 2008; Head et al., 2010this issue). The presence of ring-mold craters (Kress and Head, 2008)
in the concentric crater fill suggests that buried ice may still remain
below the CCF today.
On the interior eastern rim of the larger crater (Fig. 6), we observe
a ~ 3.25 km wide series of concentric lobes of material that are
oriented towards the crater floor. These lobes are sourced from an
elevated terrace along the crater wall, where CTX data show evidence
for downvalley flow and constriction between obstacles, similar in
nature to the “hourglass” feature on the eastern margin of the Hellas
basin in the southern hemisphere (Head et al., 2005). The concentric
lobes are bounded at their margins by broad ridges with convex-up
profiles consistent with glacial ice. These lobes show evidence for
higher concentrations of ice than the CCF nearby, and shows the
variability in accumulation at various locations within the crater.
Dickson et al. (2008) found evidence for episodes of localized
alpine-like valley glaciation that followed kilometer-thick-scale
glaciation along the dichotomy boundary. Here we find a similar
sequence of glaciation, where localized cold-trapping led to isolated
flow of ice-rich material along the interior and exterior crater walls,
followed by sublimation and retreat. This is most prominently
displayed by the expansive units of lineated terrain (Figs. 3C, 4A).
These units are characterized by downslope-trending ridges that are
found atop convex-up lobes. Downslope lineations are common at the
base of wet-based glacial environments on Earth (glacial scour,
fluting, etc.), but these ridges occur within the unit, not at the base,
and thus are not evidence for modification of the substrate. Also, the
lack of evidence for wet-based glacial processes in this region (eskers,
meltwater channels, etc.) lead us to consider cold-based origins to the
lineated texture. We interpret these ridges to be remnants of internal
flow lines or supraglacial debris, deposited within or on top of the icerich deposit. The valleys from which the lineated terrain are sourced
reveal evidence along their walls for vertical downwasting of the ice
that we interpret to have contributed to the formation of the lineated
terrain. At HiRISE scale, a generally continuous ridge is observed tens
of meters above the present day valley floor (Fig. 4B, C). HiRISE
coverage of these features is limited to a small area at the scarp face
that separates the two craters, and where the ridge is visible it can be
traced along the walls of several valleys at a consistent elevation
above the valley floors. This is identical to the sidewall lineations
observed in a tributary valley at the dichotomy boundary by Dickson
et al. (2008) in CTX data, which they interpreted to be the martian
equivalent of terrestrial “trimlines,” which record the most-recent
high-ice stands along the walls of glacial valleys. On Earth this is
generally represented by either a contrast in albedo above and below
340
J.L. Dickson et al. / Earth and Planetary Science Letters 294 (2010) 332–342
the trimline, or as a contact between vegetated land above the
trimline and bare surface below. On Mars, in both locations where we
have observed this phenomenon, the “trimline” takes the form of a
generally continuous ridge along the wall parallel to the slope of the
valley (Fig. 4), which we interpret to be a marginal moraine that
formed from deposition of debris as ice within the valley underwent
downwasting before receding to its present level.
5. Conclusions
The two-crater system described in this study provides evidence
for glaciation involving a kilometer-scale thickness of ice in the
northern mid-latitudes, followed by recession and more focused
localized debris-covered glacial flow of ice within tributary valleys
and the crater interior to produce the classic CCF texture. This
sequence of events is similar to that inferred from the study of Dickson
et al. (2008) along the dichotomy boundary. Each glacial system
appears to have undergone a similar sequence of evolution (Fig. 8):
1) Kilometer-thick glaciers/ice-sheets fill valleys and regional lows
such that flow into neighboring canyons is induced (Fig. 8A).
2) Environmental conditions change and ice stagnates and undergoes
significant downwasting, leaving perched glacial deposits in the
small crater (Fig. 8B).
3) During and after downwasting to present day topography, ice is
still able to accumulate within local cold traps, producing alpinelike glaciers with convex up profiles and CCF deposits (Fig. 8C).
4) Mars undergoes continued climate change, desiccating the midlatitudes and causing sublimation of these alpine-like glaciers. The
most recent high-stand is recorded in the form of marginal
moraines, or the martian equivalent of terrestrial “trimlines”
(Fig. 4).
What are the implications of the similarity between the deposits
observed along the dichotomy boundary (Dickson et al., 2008) and
those observed here? If kilometer-thick ice was common at these
latitudes, then we may be seeing the surface deposits of regional/
hemispheric ice sheets in the Late Amazonian (Head et al., 2010-this
issue). HRSC data show that nearby craters to the system described
here show similar evidence for large-scale glaciation in the Late
Amazonian (Fig. 9).
These examples may, however, represent more localized or
regional occurrences of accumulation of wind-blown snow during
transitions from high-to low-obliquity. Numerical simulations show
that the present obliquity of Mars (~25°) is low compared to its
predicted mean value over the last 250 million years (~34°) (Laskar
et al., 2004). Therefore, the hyper-arid desert conditions observed in
the mid-latitudes of Mars today may not be indicative of the surface in
the relatively recent past. Precise solutions for the obliquity of Mars
prior to ~20 Myr are not presently available (Laskar et al., 2004), but
geologic evidence suggests that Mars may have undergone obliquity
excursions in excess of 40°–45° within the last several hundred
million years (Kreslavsky and Head, 2006).
What is the water cycle of Mars like under such higher obliquity
conditions? Head and Marchant (2003) documented evidence for
tropical mountain glaciers on Mars in the Late Amazonian, leading
Forget et al. (2006) to try to reproduce the conditions under which these
might have formed. They showed through general circulation models
(GCMs) that at obliquity values of 45°, ice from the northern polar cap
can be mobilized and redistributed within the equatorial zone,
preferentially on the western flanks of the Tharsis Montes (Forget
et al., 2006). As Mars transitions to lower-obliquity conditions, these ice
deposits become unstable and are transported back towards highlatitudes (Levrard et al., 2004). But where are the most likely places for
ice to accumulate in the northern mid-latitudes? Recent GCMs show
that during this transition from high to low obliquity, ice that is sourced
from the Tharsis Montes glaciers would be deposited in the northern
mid-latitudes (Madeleine et al., 2007). While ice can be deposited
widely across the northern mid-latitudes, Madeleine et al. (2007) found
that the most likely sites for significant accumulation of ice are eastern
Protonilus Mensae (near the high-stand of Dickson et al. (2008) and the
deposits interpreted to be of glacial origin (Head et al., 2010-this issue))
and the Phlegra Montes (each accumulating> 12.5 mm/year of ice at
~35° obliquity).
The nature of the topography in this environment is such that the
two-crater system is likely to have acted as a local trap and
microclimate for wind-blown snow in the region, as well. In addition
Fig. 8. Sequence of modification within and around the two-crater system. Base map is a CTX mosaic of P01_001619_2232 and P01_001619_2232.
J.L. Dickson et al. / Earth and Planetary Science Letters 294 (2010) 332–342
341
Fig. 9. Subframe of HRSC orbit h3211_0000, showing a ~ 35 km impact crater ~ 60 km to the northeast of the two-crater system (see Fig. 1B for context). This crater shows evidence
for a similar sequence of crater-filling glaciation followed by more localized accumulation along the crater walls. Remnant lineations indicate that when crater-filling glaciation
occurred, flow was primarily from the south, deflecting around the central peak of the crater.
to being a topographic trap, crater walls provide sheltered slopes that
can shield surface ice deposits from solar insolation and from
sublimation. Additionally, ice that sublimates from equator-facing
slopes can be redeposited directly on pole-facing slopes (Forget et al.,
2008), adding to the annual accumulation rates. This is consistent
with the asymmetric profile of the crater floor and the morphologic
evidence for material preferentially flowing onto the crater floor from
the south (Fig. 2). Nearby craters with similar ice-related deposits
show similar pole/equatorward (Fig. 9).
Mars has had a long and complicated history of ice deposition and
glaciation within the last several hundred million years of the
Amazonian. The glacial high-stand proposed here is consistent with
that of Dickson et al. (2008), providing evidence in multiple places in
the northern mid-latitudes for kilometer-scale glaciation (Head et al.,
2010-this issue; Head and Marchant, 2008). This may not necessitate
regional ice sheets that blanketed either or both of these regions, but
rather could represent localized elevated accumulation of ice. These
results should be incorporated into the array of evidence for glaciation
in each hemisphere to better constrain the amount of ice needed to
produce these features on the global scale, and the environmental
conditions that led to their formation.
Acknowledgements
We appreciate suggestions made by Jeffrey Plaut and an anonymous reviewer. Caleb Fassett significantly aided in the processing
and registration of all data used in this study, and we thank him
accordingly. The features at the core of this study were initially found
in THEMIS VIS data, and we appreciate the work of the science and
engineering teams for THEMIS and Mars Odyssey. Likewise, we
appreciate the efforts of the science and engineering staff for HRSC,
HiRISE, CTX and MOLA. John Huffman of Brown University assisted
with the visualization of stereo data, and we appreciate his contributions. We greatly acknowledge financial assistance from NASA
for support on the ESA Mars Express High-Resolution Stereo Camera
experiment (JPL1237163), for NASA Mars Data Analysis Program
Grants NNG04GJ99G and NNX07AN95G, and for NASA Applied
Information Systems Research Program Grant NNG05GA61G.
References
Benn, D.I., Evans, D.J.A., 1998. Glaciers and glaciation, first ed. Oxford U. Press, New York.
Dickson, J.L., Head, J.W., Marchant, D.R., 2008. Late Amazonian glaciation at the
dichotomy boundary on Mars: evidence for glacial thickness maxima and multiple
glacial phases. Geology 36, 411–414.
Forget, F., Haberle, R.M., Montmessin, F., Levrard, B., Head, J.W., 2006. Formation of
glaciers on Mars by atmospheric precipitation at high obliquity. Science 311,
368–371.
Forget, F., Madeleine, J.-B., Spiga, A., Mangold, N., Costard, F., 2008. Formation of gullies
by local melting of water ice: clues from climate modeling. Workshop on Martian
Gullies, pp. 38–39.
Greeley, R., Guest, J.E., 1987. Geologic map of the eastern equatorial region of Mars.
USGS Misc. Inv. Series, Map I-1802-B.
Gwinner, K., Scholten, F., Spiegel, M., Schmidt, R., Giese, B., Oberst, J., Jaumann, R.,
Neukum, G., the HRSC Co-Investigator Team, 2005. Hochaufloesende Digitale
Gelaendemodelle auf der Grundlage von Mars Express HRSC-Daten. Photogramm.
Fernerkund. Geoinf. 5, 387–394.
Head, J.W., Marchant, D.R., 2003. Cold-based mountain glaciers on Mars: Western Arsia
Mons. Geology 31, 641–644.
Head, J.W., Marchant, D.R., 2008. Evidence for non-polar ice deposits in the past history
of Mars. LPSC 39, 1295.
Head, J.W., Neukum, G., Jaumann, R., Hiesinger, H., Hauber, E., Carr, M., Masson, P.,
Foing, B., Hoffmann, H., Kreslavsky, M., Werner, S., Milkovich, S., van Gasselt, S., the
HRSC Co-Investigator Team, 2005. Tropical to mid-latitude snow and ice
accumulation, flow and glaciation on Mars. Nature 434, 346–351.
Head, J.W., Marchant, D.R., Agnew, M.C., Fassett, C.I., Kreslavsky, M.A., 2006a. Extensive
valley glacier deposits in the northern mid-latitudes of Mars: evidence for Late
Amazonian obliquity-driven climate change. Earth Planet. Sci. Lett. 241, 663–671.
Head, J.W., Nahm, A.L., Marchant, D.R., Neukum, G., 2006b. Modification of the
dichotomy boundary on Mars by Amazonian mid-latitude regional glaciation.
Geophys. Res. Lett. 33, L08S03.
Head, J.W., Marchant, D.R., Dickson, J.L., Kress, A.M., Baker, D.M., 2010. Northern midlatitude glaciation in the Late Amazonian period of Mars: Criteria for the recognition
of debris-covered glacier and valley glacier land system deposits. Earth Planet. Sci.
Lett. 294, 306–320 (this issue).
Helbert, J., Head, J.W., Benkhoff, J., 2008. The Berlin Mars near Surface Thermal Model
(BMST) —surveying the stability of ground water ice in selected areas on Mars.
LPSC 39, 1909.
Holt, J.W., Safaeinili, A., Plaut, J., Head, J.W., Phillips, R.J., Seu, R., Kempf, S.D., Choudhary,
P., Young, D.A., Putzig, N.E., Biccari, D., Gim, Y., 2008. Radar sounding evidence for
buried glaciers in the southern mid-latitudes of Mars. Science 322, 1235–1238.
342
J.L. Dickson et al. / Earth and Planetary Science Letters 294 (2010) 332–342
Kadish, S.J., Head, J.W., Barlow, N.G., Marchant, D.R., 2008. Martian pedestal craters:
marginal sublimation pits implicate a climate-related formation mechanism.
Geophys. Res. Lett. 35, L16104.
Kreslavsky, M.A., Head, J.W., 2006. Modification of impact craters in the northern plains
of Mars: implications for Amazonian climate history. Met. Planet. Sci. 41,
1633–1646.
Kress, A.M., Head, J.W., 2008. Ring-mold craters in lineated valley fill and lobate debris
aprons on Mars: evidence for subsurface glacial ice. Geophys. Res. Lett. 35, L23206.
Laskar, J., Correia, A.C.M., Gastineau, M., Joutel, F., Levrard, B., Robutel, P., 2004. Long
term evolution and chaotic diffusion of the insolation quantities of Mars. Icarus 170,
343–364.
Levrard, B., Forget, F., Montmessin, F., Laskar, J., 2004. Recent ice-rich deposits formed at
high latitudes on Mars by sublimation of unstable equatorial ice during low
obliquity. Nature 431, 1072–1075.
Levy, J.S., Head, J.W., Marchant, D.R., 2009. Concentric crater fill in Utopia Planitia:
History and interaction between glacial “brain terrain” and periglacial mantle
processes. Icarus in review.
Li, H., Robinson, M.S., Jurdy, D.M., 2005. Origin of martian northern hemisphere midlatitude lobate debris aprons. Icarus 176, 382–394.
Lucchitta, B.K., 1981. Mars and earth —comparison of cold-climate features. Icarus 45,
264–303.
Lucchitta, B.K., 1984. Ice and debris in the fretted terrain, Mars. J. Geophys. Res. 89,
B409–B418.
Madeleine, J.-B., Forget, F., Head, J.W., Levrard, B., Montmessin, F., 2007. Exploring the
northern mid-latitude glaciation with a General Circulation Model, 7th Mars, p. 3096.
Neukum, G., Jaumann, R., the HRSC Co-Investigator and Experiment Team, 2004. HRSC:
the high resolution stereo camera of MarsExpress. European Space Agency Special
Publication, ESA SP-1240, pp. 17–35.
Pierce, T.L., Crown, D.A., 2003. Morphologic and topographic analyses of debris aprons
in the eastern Hellas region, Mars. Icarus 163, 46–65.
Plaut, J.J., Safaeinili, A., Holt, J.W., Phillips, R.J., Head, J.W., Seu, R., Putzig, N.E., Frigeri, A.,
2009. Radar evidence for ice in lobate debris aprons in the mid-northern latitudes
of Mars. Geophys. Res. Lett. 36, L02203.
Safaeinili, A., Holt, J., Plaut, J., Posiolova, L., Phillips, R., Head, J.W., Seu, R., the SHARAD
team, 2009. New radar evidence for glaciers in Mars Phlegra Montes region. LPSC
40, 1988.
Scholten, F., Gwinner, K., Roatsch, T., Matz, K.-D., Whlisch, M., Giese, B., Oberst, J.,
Jaumann, R., Neukum, G., the HRSC Co-Investigator Team, 2005. Mars express HRSC
data processing — methods and operational aspects. Photogramm. Eng. Remote
Sensing 71, 1143–1152.
Squyres, S., 1978. Martian fretted terrain — flow of erosional debris. Icarus 34, 600–613.
Squyres, S., 1979. The distribution of lobate debris aprons and similar flows on Mars.
J. Geophys. Res. 84, 8087–8096.
Tanaka, K.L., Chapman, M.G., Scott, D.H., 1992. Geologic map of the Elysium Region of
Mars. USGS Map I-2147.