evidence for amazonian northern mid-latitude regional ... · of ﬁnal retreat and downwasting of a...

Icarus 216 (2011) 23–39

Contents lists available at SciVerse ScienceDirect

Icarus

journal homepage: www.elsevier .com/locate / icarus

Evidence for Amazonian northern mid-latitude regional glacial landsystemson Mars: Glacial flow models using GCM-driven climate results and comparisonsto geological observations

James L. Fastook a, James W. Head b,⇑, Francois Forget c, Jean-Baptiste Madeleine c, David R. Marchant d

a University of Maine, Orono, ME 04469, USAb Department of Geological Sciences, Brown University, Providence, RI 02912, USAc Laboratoire de Météorologie Dynamique, Institut Pierre Simon Laplace, Université Paris 6, BP 99, 75252 Paris cedex 05, Franced Department of Earth Sciences, Boston University, Boston MA 02215, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 15 December 2010Revised 21 March 2011Accepted 14 July 2011Available online 16 August 2011

Keywords:Mars, surfaceMars, climateMars, polar geology

0019-1035/$ - see front matter � 2011 Elsevier Inc. Adoi:10.1016/j.icarus.2011.07.018

⇑ Corresponding author.E-mail addresses: [email protected] (J.L. Fasto

(J.W. Head).

A fretted valley system on Mars located at the northern mid-latitude dichotomy boundary contains lin-eated valley fill (LVF) with extensive flow-like features interpreted to be glacial in origin. We have mod-eled this deposit using glacial flow models linked to atmospheric general circulation models (GCM) forconditions consistent with the deposition of snow and ice in amounts sufficient to explain the interpretedglaciation. In the first glacial flow model simulation, sources were modeled in the alcoves only and werefound to be consistent with the alpine valley glaciation interpretation for various environments of flow inthe system. These results supported the interpretation of the observed LVF deposits as resulting from ini-tial ice accumulation in the alcoves, accompanied by debris cover that led to advancing alpine glaciallandsystems to the extent observed today, with preservation of their flow texture and the underlyingice during downwasting in the waning stages of glaciation. In the second glacial flow model simulation,the regional accumulation patterns predicted by a GCM linked to simulation of a glacial period were used.This glacial flow model simulation produced a much wider region of thick ice accumulation, and signif-icant glaciation on the plateaus and in the regional plains surrounding the dichotomy boundary. Degla-ciation produced decreasing ice thicknesses, with flow centered on the fretted valleys. As plateaus lostice, scarps and cliffs of the valley and dichotomy boundary walls were exposed, providing considerablepotential for the production of a rock debris cover that could preserve the underlying ice and the surfaceflow patterns seen today. In this model, the lineated valley fill and lobate debris aprons were the productof final retreat and downwasting of a much larger, regional glacial landsystem, rather than representingthe maximum extent of an alpine valley glacial landsystem. These results favor the interpretation thatperiods of mid-latitude glaciation were characterized by extensive plateau and plains ice cover, ratherthan being restricted to alcoves and adjacent valleys, and that the observed lineated valley fill and lobatedebris aprons represent debris-covered residual remnants of a once more extensive glaciation.

� 2011 Elsevier Inc. All rights reserved.

1. Introduction (e.g., Head et al., 2010). Furthermore, analysis of the geological re-

1.1. Background and evidence for climate change on Mars

Exploration of Mars over the decades has shown that there aresignificant volumes of water sequestered in various surface depos-its at different latitudes, such as the polar layered terrains (e.g.Thomas et al., 1992; Keiffer and Zent, 1992; Phillips et al., 2008),in the shallow subsurface at high latitudes (e.g., Boynton et al.,2002; Feldman et al., 2002), in the deeper cryosphere (e.g., Clifford,1993), and even in mid-latitude alpine-valley glacial landsystems

ll rights reserved.

ok), [email protected]

cord of Mars shows a rich array of Amazonian and Hesperian-agedfeatures and deposits that involved water, sometimes in hugeamounts (e.g. Clifford, 1993; Carr, 1996; Masson et al., 2001).Noachian-aged surfaces display valley networks that suggest a‘‘warmer and wetter’’ early Mars (e.g., Craddock and Howard,2002; Fassett and Head, 2008). Clearly, the climate has changedduring martian history from a time when liquid water flowedacross the surface in valley networks and pluvial activity may haveoccurred (e.g., Craddock and Howard, 2002; Fassett and Head,2008), to the very cold and hyperarid surface conditions observedtoday (e.g., Zurek, 1992; Head et al., 2003a; Carr and Head, 2010).

Coincident with this long-term climate change has been a tran-sition in the global hydrological cycle from an early Mars when itappears that the hydrological cycle was vertically integrated (e.g.,

http://dx.doi.org/10.1016/j.icarus.2011.07.018

mailto:[email protected]

mailto:[email protected]


http://www.sciencedirect.com/science/journal/00191035

http://www.elsevier.com/locate/icarus

24 J.L. Fastook et al. / Icarus 216 (2011) 23–39

Carr, 1996; Clifford and Parker, 2001; Baker, 2001; Head et al.,2003b) to one in which the hydrological cycle is horizontally strat-ified, with a cryosphere separating the groundwater system fromthe surface system. The surface system then consists of three majorwater reservoirs: (1) surface deposits (currently the polar caps), (2)the atmosphere, and (3) the regolith, and most of the movement ofwater is associated with seasonal and longer-term climate changesrelated to spin-axis/orbital variations (e.g., Laskar et al., 2004).

In addition to these long-term global climate trends, there isabundant evidence for shorter-term climate fluctuations drivingthe surface water cycle related to variations in these spin-axis/orbital parameters, such as obliquity, eccentricity and precession(e.g., Laskar et al., 2004). The discovery and documentation of geo-logically recent gullies (e.g., Malin and Edgett, 2000) and furtherdocumentation of Amazonian tropical mountain glaciers (e.g.,Williams, 1978; Head and Marchant, 2003; Shean et al., 2005,2007; Milkovich et al., 2006; Kadish et al., 2008) are among thegrowing evidence suggesting that variations in orbital parametersmay cause climate excursions resulting in the redistribution ofwater and a Mars very different from that typical of very recenthistory. Indeed, Laskar et al. (2004) predict that throughout muchof its history Mars may have been characterized by an obliquitysignificantly higher than its current value.

One of these high obliquity states modeled by Forget et al.(2006) (for obliquity at 45� with a source of moisture at the presentpoles) produced regions of persistent ice accumulation on thenorthwestern flanks of the Tharsis volcanoes consistent with geo-logical observations (e.g., Head and Marchant, 2003), and glacialflow models of the resulting ice sheets agreed well with detailsof the deposits (Fastook et al., 2008a). Furthermore, geologic anal-ysis of the mid-latitudes of Mars has shown abundant evidence forAmazonian ice-assisted flow of debris (e.g., Squyres, 1978, 1979;Lucchitta, 1981, 1984; Pierce and Crown, 2003; Chuang and Crown,2005; Li et al., 2005).

Fretted terrain and fretted channels, well developed in thenorthern part of Arabia Terra (Sharp, 1973), were clearly identifiedin Viking data as the location of lobate debris aprons (LDA) and lin-eated valley fill (LVF) (Squyres, 1978, 1979). Carr and Schaber(1977) analyzed Viking Orbiter images, concluding that frost creepand gelifluction were the primary processes in LDA formation.Squyres (1978) documented the fact that debris aprons (LDA),characterized by sharply-defined flow fronts and convex-upwardsurfaces, extend from massifs and escarpments outward to dis-tances of up to �20 km. Squyres (1978, 1979) interpreted theaprons to have resulted from mass-wasted debris that floweddue to the presence of interstitial ice, envisioning that during peri-ods of climate change, atmospheric water vapor would diffuse intotalus piles formed at the base of steep topography, and interstitialice would cause mobilization and flow of the talus to produce thelobes. Longitudinal ridges and grooves in surface debris seen on thefloors of the fretted valleys (LVF) were interpreted to form by ice-assisted debris aprons flowing from valley wall talus slopes andconverging in the middle of the valley floor (Squyres, 1978,1979). A number of workers noted that some LVF in the Nilosyrtisregion appeared consistent with down-valley flow; following ear-lier ideas by Kochel and Baker (1981). Kochel and Peake (1984)interpreted some debris aprons as formed by processes similar tothose of terrestrial ice-cored or ice-cemented rock glaciers and fur-ther noted a transition from flow-parallel to flow-transverse ridge-and-furrow topography. Lucchitta (1984) interpreted the LDA andLVF as flow of debris with interstitial ice and showed evidence oflocal down-valley flow of LVF, likening some examples to flow pat-terns in glacial ice in Antarctica. Carr (1996, pp. 116–120) urgedcaution in the general interpretation of a range of landforms asbeing due to glaciation. He cited: (1) the fact that the glacialhypotheses requires major climate change late in Mars history

(e.g., Baker et al., 1991) for which he found little supporting evi-dence and (2) that the low erosion rates in the Amazonian arguesagainst any major climate excursions accompanied by precipita-tion, as envisioned by Baker (2001).

More recently, Pierce and Crown (2003) used new image andaltimetry data and found evidence for a wide range of possibleLDA ice content, interpreting the evidence to be consistent withmultiple models of apron formation (e.g., rock glacier ice-assistedcreep of talus, ice-rich landslides, debris-covered glaciers), withtypical LDA resulting from flow of debris that was enriched inground ice. Mangold (2003) showed that ice content in LDA mayexceed pore space, and Li et al. (2005) compared LDA MOLA pro-files to simple plastic and viscous power law models for ice–rockmixtures, concluding that LDAs were ice-rich rock mixtures withsome perhaps >40% ice by volume. Chuang and Crown (2005), doc-umenting the detailed character of 65 LDA, concluded that theyconsisted of mixtures of debris and ice, but that it was ‘‘difficultto constrain the internal distribution of ice or the method of debrisapron initiation from the current datasets.’’ In summary, all work-ers agree that formation of LDA and LVF involve both talus and ice,but there is disagreement as to the amount of ice: One end-mem-ber calls on formation by ice-assisted creep of talus (often definedas producing a ‘‘rock glacier’’) (e.g., Squyres, 1978, 1979), while an-other end-member calls on formation as debris-covered glaciers,(predominantly glacial ice with a cover of sublimation lag or till)(e.g., Head et al., 2005, 2006a,b).

Head et al. (2010) developed criteria to distinguish betweenthese end-members and investigated a wide range of occurrencesto assess the origin of LDA and LVF. On the basis of the backgroundstudies described above, and using a range of terrestrial analogsapplicable to the recent cold-desert environment of Mars (e.g.,Marchant and Head, 2006, 2007), Head et al. (2010) developedthe following criteria to assist in the identification of debris-cov-ered glacial-related terrains on Mars (features, followed by theinterpretation of each, listed in parentheses): (1) alcoves, theater-shaped indentations in valley and massif walls (local snow andice accumulation zones and sources of rock debris cover); (2) par-allel arcuate ridges facing outward from these alcoves and extend-ing down slope as lobe-like features (flow-deformed ridges ofdebris); (3) shallow depressions between these ridges and the al-cove walls (zones originally rich in snow and ice, which subse-quently sublimated, leaving a depression); (4) progressivetightening and folding of parallel arcuate ridges where abuttingadjacent lobes or topographic obstacles (constrained debris-cov-ered glacial flow); (5) progressive opening and broadening of arcu-ate ridges where there are no topographic obstacles (unobstructedflow of debris-covered ice); (6) circular to elongate pits in lobes(differential sublimation of surface and near-surface ice); (7) largertributary valleys containing LVF formed from convergence of flowfrom individual alcoves (merging of individual lobes into LVF); (8)individual LVF tributary valleys converging into larger LVF trunkvalleys (local valley debris-covered glaciers merging into largerintermontaine glacial systems); (9) sequential deformation ofbroad lobes into tighter folds, chevron folds, and finally into lin-eated valley fill (progressive glacial flow and deformation); (10)complex folds in LVF where tributaries join trunk systems (differ-ential flow velocities causing folding); (11) horseshoe-like flowlineations draped around massifs in valleys and that open in adown-slope direction (differential glacial flow around obstacles);(12) broadly undulating along-valley topography, including localvalley floor highs where LVF flow is oriented in different down-val-ley directions (local flow divides where flow is directed away fromindividual centers of accumulation); (13) integrated LVF flow sys-tems extending for tens to hundreds of kilometers (intermontaineglacial systems); (14) rounded valley wall corners where flowconverges downstream, and narrow arete-like plateau remnants

J.L. Fastook et al. / Icarus 216 (2011) 23–39 25

between LVF valleys (both interpreted to be due to valley glacialstreamlining). Taken together, the occurrence of a number of thesetypes of features in an area is interpreted to represent the formerpresence of debris-covered glaciers and valley glacial systems inthe Deuteronilus–Protonilus region (Head et al., 2010).

Detailed geological analysis of several different areas in thenorthern mid-latitudes (Fig. 1a) has shown strong evidence forthe presence of valley glacial landsystems (Eyles, 1983; Evans,2003) during the Amazonian and the sequestration and preserva-tion of debris-covered glacial ice (Head et al., 2005, 2006a,b,2010; Levy et al., 2007; Kress and Head, 2008; Morgan et al.,2009; Dickson et al., 2008, 2010; Hauber et al., 2008; Marchantand Head, 2008; Baker et al., 2010), interpretations supported bythe detection of extensive buried ice by the SHARAD (SHAllow RA-Dar) instrument on board the Mars Reconnaissance Orbiter (e.g.,Holt et al., 2008a,b; Plaut et al., 2009, 2010). Further atmosphericgeneral circulation modeling by Madeleine et al. (2009) for anobliquity of 35� with the moisture source in the Tharsis region,rather than the poles, produced persistent ice accumulation alongthe mid-latitude region and the dichotomy boundary, also consis-tent with the presence of these geological features indicative ofvalley glaciation documented there.

In this contribution we use the results of the LMD/GCM(Madeleine et al., 2009) that reproduce conditions that favor iceaccumulation in the mid-latitudes. From these predictions, wedevelop glacial flow models to assess the accumulation and flowpatterns of ice to compare to the patterns documented by the

Fig. 1. (a) Northern mid-latitudes dichotomy boundary region showing the locations of keMarchant, 2009; 4: Morgan et al., 2009; 5: Head et al., 2006a; 6: Baker et al., 2010; 7: Diclineated valley fill system in the central Deuteronilus–Protonilus Mensae (DPM) describedBoxes and letters in (b) indicate locations of illustrative images in Head et al. (2006a). (c(narrow arrows) and broad flow trends of LVF on the valley floor (wide arrows). Compilenature of alcoves and LVF features associated with the system shown in (b and c). (d)individual, convex outward lobes extending from each alcove out into the adjacent LVF.deformation, and their convergence with the general trends of the LVF on the valley floolobes from alcoves (bottom left and right); lobes join major LVF of area C (upper right) neright lobe is increasingly compressed until it resembles a tight isoclinal fold. Both lobes u(b). MOC R1702578.

geological observations, to test their validity, to further test thedistinctions between the (1) ice-rich debris creep and (2) deb-ris-covered glacier models for LDA and LVF, and if successful, tolearn more about the nature of interpreted Amazonian valleyand regional glaciation.

1.2. Geological observations

A unique fretted terrain, located along the northern mid-latitude dichotomy boundary (Fig. 1a), has long been held toprovide clues to the processes responsible for the formationand degradation of the dichotomy boundary (Sharp, 1973). Partsof these fretted valleys (Fig. 1b and c) display a multitude ofcharacteristics typical of integrated valley glacial landsystemson Earth (e.g., Eyles, 1983; Evans, 2003), including multipletheater-headed, alcove-like accumulation areas (Fig. 1d); sharparete-like ridges typical of glacial erosion; converging patternsof downslope valley flow (Fig. 1e); valley lineation patternstypical of folding and shear (Fig. 1e); wrap-around features indic-ative of flow around obstacles; and broad piedmont-like lobes asthe valley fill extends out into the northern lowlands (Head et al.,2006a, 2010). These characteristics suggest that the boundaryarea was subjected to large-scale glaciation during the Amazo-nian. Recognition and documentation of this important late-stageprocess provides information critical to reconstructing the origi-nal dichotomy boundary, and to the understanding of Amazonianclimate history.

y published study areas (1: Kress and Head, 2009; 2: Head et al., 2006b; 3: Head andkson et al., 2008; 8: Levy et al., 2007; 1/7: Head et al., 2010). (b) Fretted channel and

in Head et al. (2006a) (box 5 in (a)). Viking Orbiter image mosaic and location map.) Map showing the location and trend of small alcoves and lobate flow-like featuresd using HRSC orbit 1395, THEMIS and MOC images. (d and e) Images illustrating the

Four adjacent alcoves oriented convex-inward toward the plateau (bottom) withNote the convergence of these individual lobate extensions, their compression andr (left). Location is box A in (b). THEMIS VIS V11208010. (e) Outflow of two ridged

ar convergence with B (c). Left lobe is swept westward, forming broad arcuate folds;ltimately merge into the general LVF parallel to the valley walls. Location is box D in


1.3. Interpretation as a valley glacial landsystem

Detailed analysis of a typical fretted valley network at �34�E,41�N in the central region between Deuteronilus and ProtonilusMensae (Fig. 1a, box 5) was undertaken by Head et al. (2006a)(Fig. 1b and c). This set of fretted valleys extends for over 200 kmin a N–S direction, from deep within the upland plateau to the edgeof the continuous upland scarp at �41–42�N. Using MOLA altime-try, and Viking, THEMIS, HRSC, and MOC images, Head et al.(2006a) traced the LVF from the narrow parts of the deeper uplandvalleys progressively toward the northern lowlands. In the mostinland regions of the fretted valleys analyzed (areas A and B inFig. 1b) the valleys are narrow (�10–20 km in width), often havedistinctively cuspate (convex outward) shoulder-to-shoulder al-coves along the walls (region A) or larger tributary valleys oftenforming hanging valleys along the wall margins (region B). In theA/B regions (Fig. 1c) the valley floor is about 1500 m above thenorthern distal end of the deposit.

In contrast to the sharp parallel valley walls and parallel along-valley lineated fill typical of some fretted valleys described else-where (e.g., Squyres, 1978; Carr, 1995, 2001), these alcoves aresources of LVF that can be seen to emerge from individual alcovesas lobes (Fig. 1c and d), descending onto the valley floor, bendingand turning in a downslope direction (Fig. 1c and e), and mergingwith adjacent lobes, deforming in the process (Fig. 1c) ultimatelyto form the distinctive along-strike ridges of LVF. These types of al-coves, the arete-like borders between them, the arcuate-ridgedlobes extending from them, the progressive compression and fold-ing of the lobes as they converge onto the valley floor and lose theiridentity, are all hallmarks of glaciated valley landsystems on Earth(Eyles, 1983; Benn et al., 2003). In the context of such valley glacialenvironments, the alcoves in the DPM area were interpreted torepresent microenvironments where ice and snow accumulate, al-cove walls shed debris onto the ice, and debris-covered glaciersemerge and move downslope into the main valleys, widening thealcoves and creating aretes over time (Head et al., 2006a). Mappedtrends show the emergence of lobate material from these alcoves(Fig. 1c, narrow arrows) and the general trends with which theymerge into the larger-scale LVF (wide arrows). The patterns inareas A, B, and C (Fig. 1c) clearly indicate the convergence typicalof glaciated valley landsystems (Eyles, 1983; Benn et al., 2003)where abundant small debris-covered glaciers contribute to lar-ger-scale valley glaciers. As the generally north-trending LVFemerges from areas A to C (Fig. 1c), the individual LVFs merge intotwo main trends, areas D and E. LVF C merges with B and continues�95 km to the NNW where it bifurcates around a mesa to form LVFF and G, each continuing 50–75 km to the northern lowlands. Wes-tern portions of LVF B turn west into D, merging with LVF A to pro-duce complex fold patterns. Portions of A turn NW and extendalmost 100 km to the northern lowlands. Several basic trends areobserved over this region extending from A–B–C to the northernlowlands (Head et al., 2006a). The elevation of the LVF floor de-creases by �1500–2000 m from the proximal areas (areas A–C)to the end of the LVF in the northern lowlands. LVF slopes tilt to-ward the north and are less than 1� throughout most of the LVFfloor, but increase rapidly to >1–2� in the northern 25–50 km ofthe LVF adjacent to the lowlands. These areas are characterizedby piedmont-like lobes extending into the northern lowlands be-tween mesas. These features are reminiscent of the fronts of terres-trial glaciers in terms of their lobate nature and relatively steepfront, and in many cases are characterized by flow lines, pits andmoraine-like ridges.

In summary, spacecraft data show compelling evidence for anintegrated picture of LVF formation (Fig. 1b–e; Head et al.,2006a) with a significant role being played by regional valley gla-ciation (e.g., Eyles, 1983; Post and Lachapelle, 2000; Benn et al.,

2003) in the modification of the valley systems. There is evidencefor: (1) localized alcoves, which are the sources of dozens of nar-row, lobate concentric-ridged flows interpreted to be remnants ofdebris-covered glaciers; (2) depressions in many current alcoves,suggesting loss of material from relict ice-rich accumulation zones;(3) narrow pointed plateau ridge remnants between alcoves, simi-lar to glacially eroded aretes; (4) horseshoe-shaped ridges up-val-ley and upslope of topographic obstacles; (5) convergence andmerging of LVF fabric in the down-valley direction, and deforma-tion, distortion and folding of LVF in the vicinity of convergence,all consistent with glacial-like, down-valley flow (Fig. 1c); and(6) distinctive lobe-shaped termini where the LVF emerges intothe northern lowlands, with associated pitting and concentric ter-minal ridges.

This assemblage of features can be traced and mapped in anintegrated fashion (Fig. 1b and c) for over 200 km in length, andcovers an area (�30,000 km3), �10 times that of the AntarcticDry Valleys (Marchant and Head, 2007), being more comparablein scale to the remnant North American ice caps on Ellesmereand Baffin Islands in the Canadian high Arctic. This assemblage offeatures suggests that the LVF in this region formed as part of abroad integrated valley glacial system extending for about200 km in the S–N direction. The extensive development of similarlineated valley fill occurrences, often with very similar characteris-tics, in fretted valleys in the DPM region (Fig. 1a; see also Headet al., 2010), suggests that glaciation may have been instrumentalin the modification of the dichotomy boundary over an area aslarge as several million km2. The age of the LVF in different partsof the DPM region has been estimated as Amazonian (e.g., McGill,2000; Mangold, 2003; Morgan et al., 2009; Baker et al., 2010), withrelatively younger alteration by processes of mantle deposition andsurface layer sublimation and modification (e.g., Mangold, 2003;Head et al., 2003a,b; Levy et al., 2009). On the basis of these typesof observations, Head et al. (2006a) (and others; Fig. 1a) inter-preted the process of alpine-style valley glaciation to have beenextensive in this region.

To assess this glacial origin hypothesis for this valley system(Fig. 1b) and the region in general (Fig. 1a), Madeleine et al.(2009) explored the conditions under which GCMs reproduced suf-ficient ice deposition and seasonal preservation to cause glaciation.They found that the LMD/GCM for a dusty Mars atmosphere withobliquity set to 35� and a water source in the Tharsis region couldgenerate ice accumulation in good agreement with these geologi-cal observations. According to Madeleine et al. (2009), thisproposed climate is what one might expect to follow a higher-obliquity excursion (Forget et al., 2006; Levrard et al., 2004) ofthe sort thought to build ice sheets on the flanks of the Tharsis vol-canoes (Head and Marchant, 2003). We now use the predictions ofthe LMD/GCM to formulate glacial flow models to compare to thepatterns of LDA and LVF interpreted by Head et al. (2006a) to bedue to glaciation, and to test the validity of these interpretationsof valley glacial landsystems (Fig. 1a), and to test further thedistinctions between the (1) ice-rich debris creep and (2) debris-covered glacier models for LDA and LVF. We then explore theimplications of the locations of ice depocenters predicted by theLMD/GCM for more regional glacial landsystems.

2. Glacial flow modeling of the Deuteronilus–Protonilus Mensae(DPM) region

2.1. The University of Maine Ice Sheet Model (UMISM)

UMISM is an adaptation for the martian environment (Fastooket al., 2004, 2005, 2006) of a terrestrial ice sheet model used fortime-dependent reconstructions of Antarctic, Greenland, and

2350

2400

2450

2500

2550

2350

2400

2450

2500

2550

-50 0 50

-50 0 50

-3600

-3400

-3200

-3000

-2800

-2600

-2400

-2200

-2000

-1800

-1600

-1400

-1200

-1000

-800

-600

-400

-200

0

200

400

600Positive Mass Balance RegionsAlcove

Fig. 2. DPM lineated valley fill system showing alcove regions of positive mass balance of 1 mm/a; �0.1 mm/a negative mass balance elsewhere. Background is MOLA griddedtopography, color-coded in elevation (m). Red lines show location of positive mass balance regions. Grid is model reference grid in 10 km boxes.


paleo-ice sheet evolution in response to changing climate on Earth(Fastook, 1993). UMISM uses a thermo-mechanically coupled Shal-low-Ice Approximation (vertically-integrated momentum com-bined with continuity) where the dominant stress is internalshear and longitudinal stresses are neglected. Primary input tothe model is the bed on which the ice sheet is to be reconstructedand the net annual surface mass balance (SMB), or accumulationrate. Secondary input includes the mean-annual surface tempera-ture and the geothermal heat flux, used to calculate internal tem-peratures from which the mechanical properties of the ice areobtained. In addition, internal temperatures allow for the possibil-ity that the base of the ice reaches the pressure melting point, atwhich point some sliding criteria can be invoked, a phenomenawe have not observed in any modeled martian glaciers (Fastooket al., 2008a,b).

The fact that with the exception of the bed, these inputs arepoorly constrained for Mars introduces uncertainty into the resultsthat will be presented. However, choice of reasonable values formean annual temperature, geothermal flux, and SMB still allow pro-duction of results comparable to the geologic observations. Giventhese uncertainties the choice of a Shallow-Ice Approximation mod-el is also appropriate, since the considerable computational load of ahigher-order model would not produce more accurate results.

2.2. Modeling: alcove-only accumulation areas

As a preliminary step in the modeling of the DPM interpretedglaciation (Head et al., 2006a), we first examine a simple casewhere accumulation only occurs in the alcoves along the valleywalls. This simple treatment has terrestrial analogs in the Dry Val-leys of Antarctica (Marchant and Head, 2004, 2005, 2006, 2007). In

particular, we follow a model of ice deposition observed in the DryValleys of Antarctica on Mullins Glacier, a debris-covered glacierthat originates in an alcove analogous to the martian alcoves(Kowalewski et al., 2011). Here deposition of ice occurs only in avery limited area at the base of a scarp that also serves as thesource of the surface debris. Without this protective layer of rockdebris, the glacier would sublimate rapidly as it flows down intothe Dry Valleys, an area where normal sublimation removes anyice exposed at the surface. Marchant and Head (2004, 2005,2006, 2007) and Marchant et al. (2010) have shown that buriedice in the Mullins Glacier may be several millions of years old, a sit-uation that would be not be possible without the armoring effect ofthe debris cover (Kowalewski et al., 2006).

In reconstructions of Tharsis Montes ice sheets, we used either aparameterization of the SMB as a function of height (Fastook et al.,2004, 2005) or GCM results (Fastook et al., 2006) as the source ofmass in the continuity equation. Here we chose the expedient ofsimply specifying a positive SMB of +1 mm/year in the alcoves,and a negative SMB 1/10th of that outside the alcoves (Fig. 2). Thesevalues are typical of debris-covered glaciers in the Dry Valleys, Ant-arctica (Kowalewski et al., 2011). The shape of the resulting profile isrelatively insensitive to the magnitude of the SMB. SMB is raised tothe 1/8th power in analytic solutions for uniform SMB elliptical pro-files, implying a thickening of only 10% for a doubling of the SMB.What larger or smaller SMB will produce is different formation timesand resulting velocity magnitudes. Doubling SMB effectively dou-bles velocity because thickness is only increased by 10% and the in-creased flux must be accommodated by increased velocity.Changing SMB does not significantly affect the orientation of theflow field because flow direction is down the surface gradient andsurface changes are small relative to changes in the SMB. These

2350

2400

2450

2500

2550

2350

2400

2450

2500

2550

-50 0 50

-50 0 50

0255075100125150175200225250275300325350375400425450475500525550575600625650675700

Thickness [3006]Time=300000

2350

2400

2450

2500

2550

2350

2400

2450

2500

2550

-50 0 50

-50 0 50

0255075100125150175200225250275300325350375400425450475500525550575600625650675700


2350

2400

2450

2500

2550

2350

2400

2450

2500

2550

-50 0 50

-50 0 50

0255075100125150175200225250275300325350375400425450475500525550575600625650675700


2350

2400

2450

2500

2550

2350

2400

2450

2500

2550

-50 0 50

-50 0 50

0255075100125150175200225250275300325350375400425450475500525550575600625650675700


Fig. 3. DPM lineated valley fill system color-coded thicknesses (m) at 300, 500, 1000, and 1500 Ka. Contour lines indicate surface elevation.

1 For interpretation of color in Figs. 1–10 and 12, the reader is referred to the webversion of this article.


higher or lower velocities will result in faster or slower advance ofthe glacier during the formation, and of course will affect retreattimes during collapse. There are no constraints on either the forma-tion times or the velocity magnitudes, so the simple expedient of+1 mm/a in the alcoves and �0.1 mm/a elsewhere was chosen toyield a 10:1 ratio between accumulation area and ablation area.

Starting with no ice, the model is run for 2 Ma, which delivers aflow pattern (primarily the orientation of the velocity field) thatcan be compared to the Head et al. (2006a) glacial interpretation.Resulting ice thicknesses and ice flow velocities, with superim-posed surface elevations, are shown in Figs. 3 and 4 for four timeintervals during the growth of these ice deposits.

(1) At 300 Ka (Fig. 3, upper left), the ice is thin (less than 300 m;blue1 areas) and the flow from the sides is at low velocity (lessthan �50 mm/a; Fig. 4, upper left). The flow has not yetmerged along the center of the valleys (blue areas in Fig. 3upper left). A configuration such as this would not yet producethe turning flow observed by Head et al. (2006a) (Fig. 1).

(2) By 500 Ka (Fig. 3, upper right), the beginning of a coherentdown-valley flow is observed (green and blue areas in thelower right-hand and central part). The ice flowing from

2350

2400

2450

2500

2550

2350

2400

2450

2500

2550

-50 0 50

-50 0 50

0102030405060708090100110120130140150160170180190200210220230240250

Velocity [3006]Time=300000

2350

2400

2450

2500

2550

2350

2400

2450

2500

2550

-50 0 50

-50 0 50

0102030405060708090100110120130140150160170180190200210220230240250


2350

2400

2450

2500

2550

2350

2400

2450

2500

2550

-50 0 50

-50 0 50

0102030405060708090100110120130140150160170180190200210220230240250


2350

2400

2450

2500

2550

2350

2400

2450

2500

2550

-50 0 50

-50 0 50

0102030405060708090100110120130140150160170180190200210220230240250


Fig. 4. DPM lineated valley fill system color-coded velocities (mm/a) at 300, 500, 1000, and 1500 Ka. Contours indicate surface elevation.


each side of these parts of the valleys has merged in the cen-ter. Note that the velocities (Fig. 4, upper right) are not yetappreciable, with most velocities being less than �50 mm/a.

(3) By 1000 Ka (Fig. 3, lower left), there is a well-established val-ley glacier system extending from the fretted valleys in thesouth to the mouths of the valleys along the dichotomyboundary, with thicknesses exceeding 400–500 m in thesouthern part of the valley system. Velocities (Fig. 4, lowerleft) are almost everywhere higher than 50 mm/a, andlocally in excess of �200 mm/year.

(4) As time progresses to 1500 Ka (Fig. 3, lower right), the valleyglaciers become very well developed in the valleys andextend out of the valleys and onto the northern lowlands.Ice thicknesses are locally in excess of 600 m. Velocities(Fig. 4, lower right) are broadly in excess of 100 mm/a, andlocally exceed 200 mm/a.

Comparison of time steps in Fig. 3 shows that ice accumulationand flow at either 1000 Ka (Fig. 3, lower left) or 1500 Ka (Fig. 3,lower right) would clearly be producing the kinds of landforms

2345

2350

2355

2360

2365

2370

2345

2350

2355

2360

2365

2370

5 10 15 20 25

5 10 15 20 25

0255075100125150175200225250275300325350375400425450475500525550575600625650675700


2345

2350

2355

2360

2365

2370

2345

2350

2355

2360

2365

2370

5 10 15 20 25

5 10 15 20 25

0255075100125150175200225250275300325350375400425450475500525550575600625650675700


2345

2350

2355

2360

2365

2370

2345

2350

2355

2360

2365

2370

5 10 15 20 25

5 10 15 20 25

0102030405060708090100110120130140150160170180190200210220230240250


2345

2350

2355

2360

2365

2370

2345

2350

2355

2360

2365

2370

5 10 15 20 25

5 10 15 20 25

0102030405060708090100110120130140150160170180190200210220230240250


Fig. 5. The region furthest from the dichotomy boundary and the valley mouths (area BC in Fig. 1b, THEMIS VIS V11208010) is characterized by eastward-opening adjacentalcoves with multiple convex-outward concentric-ridged lobes extending, converging, deforming, and merging with the valley floor LVF trends. Here the model at 1000 and1500 Ka shows extremely low surface slopes (surface contours are only 50 m). Ice is 300–500 m thick, and velocities are generally less than �50 mm/a. The flow is directedaway from the alcoves with little indications of turning down-valley.


observed by Head et al. (2006a) and interpreted to be glacial in ori-gin (compare to patterns in Fig. 1b–e).

We can further examine these correlations and test the glacialinterpretation by comparing smaller areas of accumulation andflow orientations to the patterns observed by Head et al. (2006a).High resolution excerpts from the glacial flow model results, forboth thickness and velocity at 1000 Ka and at 1500 Ka (Figs. 3and 4), are compared to patterns in the THEMIS images of regionsBC, F, and G from Fig. 1b (Head et al., 2006a) in Figs. 5–7. Whiledirection of flow is not explicitly shown in Figs. 5–7, the velocityvector is always perpendicular to the surface elevation contoursand aligns well with the interpreted flow feature orientations.

The region furthest from the dichotomy boundary and the val-ley mouths (area BC in Fig. 1b) shown in Fig. 5 is characterized byeastward-opening adjacent alcoves with multiple convex-outwardconcentric-ridged lobes extending, converging, deforming, andmerging with the valley floor LVF trends. Here the model at 1000and 1500 Ka shows extremely low surface slopes (surface contoursare only 50 m). Ice is 300–500 m thick, and velocities are generallyless than �50 mm/a. This matches well with the interpretation ofalcove-sourced debris-covered glaciers merging gradually intothe valley flow system. Velocities are very low and there is littleindication of the flow turning down-valley.

A region further down the valley toward the dichotomyboundary (area F in Fig. 1b) containing an eastward-facing zoneof multiple alcoves and converging lobate flows in the distal re-gions of the system, along the edge of the northernmost largemesa (area G in Fig. 1b) is shown in Fig. 6. Here concentric-out-ward ridges extend from alcoves and are progressively com-pressed, folded and flattened as the ridges deform and becomepart of the valley floor LVF. In this case, model thicknesses at1000 and 1500 Ka are 300–500 m and velocities are locally

�250 mm/a toward the center of the valley, but less than100 mm/a in the alcove and adjacent area. In this example,one can clearly see the low flow rates at the alcove and theabruptly increasing flow rates at the point of the turning asthe glacial lobes merge with the faster along-valley flow.

A region further down the valley near the dichotomy bound-ary (area G in Fig. 1b) is away from the alcoves and in the cen-tral part of the valley where ice flow was interpreted by Headet al. (2006a) to have been down-valley glacial flow that bifur-cated around massifs in the central part of the valley system(Fig. 7). LVF bifurcates (bottom right) and flows around a massifforming a broad up-flow collar and a diffuse, down-flow ‘‘wake’’.LVF in the narrow pass between massifs is compressed. The lo-bate LVF extends into the northern lowlands at top left. Notethe arcuate, piedmont-like configuration of the lobe and the highdensity of pits. The modeled flow clearly narrows as it pinchesbetween the two bedrock highs and accelerates in these areasto values in excess of 200 mm/a.

In summary, the glacial flow and velocity orientation patternsin the ice flow model simulation clearly match many of the fea-tures observed in the lineated valley fill, interpreted by Headet al. (2006a) to be of glacial origin. Particularly significant were:(1) the turning flow observed in the lower valley emanating fromthe alcoves and merging with the general down-valley flow(Fig. 6); (2) the deflection of flow around obstacles (consistent witha relatively thin glacial ice deposit confined to the valley) (Fig. 7);thicker, more extensive ice sheets might follow the terrain lessfaithfully, as the obstacles would be overridden; and (3) the simi-larities in the thicknesses and extents of the modeled ice and theLVF deposits (Figs. 1b, c and 3).

We conclude that the glacial flow models described heresuccessfully reproduce the lineated valley fill patterns and flow

2440

2445

2450

2455

2460

2440

2445

2450

2455

2460

15 20 25 30 35

15 20 25 30 35

0255075100125150175200225250275300325350375400425450475500525550575600625650675700


2440

2445

2450

2455

2460

2440

2445

2450

2455

2460

15 20 25 30 35

15 20 25 30 35

0255075100125150175200225250275300325350375400425450475500525550575600625650675700


2440

2445

2450

2455

2460

2440

2445

2450

2455

2460

15 20 25 30 35

15 20 25 30 35

0102030405060708090100110120130140150160170180190200210220230240250


2440

2445

2450

2455

2460

2440

2445

2450

2455

2460

15 20 25 30 35

15 20 25 30 35

0102030405060708090100110120130140150160170180190200210220230240250


Fig. 6. A region further down the valley toward the dichotomy boundary (area F in Fig. 1b, THEMIS VIS V11208010 with model thickness and velocity at 1000 and 1500 Ka)containing an eastward-facing zone of multiple alcoves and converging lobate flows in the distal regions of the system, along the edge of the northernmost large mesa (area Gin Fig. 1b). Here concentric-outward ridges extend from alcoves and are progressively compressed, folded and flattened as the ridges deform and become part of the valleyfloor LVF. In this case, thicknesses are 300–500 m and velocities are locally �250 mm/a toward the center of the valley, but less than 100 mm/a in the alcove and adjacentarea. In this example, one can clearly see the low flow rates at the alcove and the increased flow rates at the point of the turning of the glacial lobes and the merging with thealong-valley flow.


scenarios interpreted by Head et al. (2006a) to be evidence for al-pine-type valley glacial landsystems, and thus support this inter-pretation by providing a physically-reasonable scenario thatproduces results quantitatively similar to the geological observa-tions. In this scenario, ice accumulates in alcoves and other pro-tected areas, and debris cover derived from the adjacent exposedcliffs records the flow features of the underlying ice even afterthe ice beneath has partly sublimated away (Head et al., 2006a).A conclusion is that we would only expect to see such debris fea-tures where a source of surface debris was available, as is observedwith the Mullins Glacier in the Dry Valleys of Antarctica (Marchantand Head, 2007; Marchant et al., 2010; Kowalewski et al., 2011;Shean and Marchant, 2010) and in the valley walls of the frettedterrain and along the dichotomy boundary. Otherwise, if not cov-ered by debris, the ice at these latitudes would have sublimatedaway in the modern climate (e.g., Hauber et al., 2008) leaving notrace.

A corollary to this is that we would predict that any ice depos-ited at higher levels on the plateau above the valleys, with no high-er scarps or nunataks from which debris could be deposited, wouldleave no record of ice flow. This raises the question: Could the cur-rently observed LVF (Fig. 1b–e) represent not the greatest lateralextent of an alpine-type valley landsystem, but instead representthe waning stages and a remnant of a larger ice sheet preservedonly when the scarps were exposed during retreat (Marchant andHead, 2006, 2007, 2008)? In the following section we examine thespecific accumulation areas predicted by the LMD/GCM to assesswhether the LVF might be the record of the waning stages of glaci-ation, after the regional ice surface in the valleys had dropped be-low the levels of the surrounding scarps, allowing debris toaccumulate on the surface of the glaciers.

3. Glacial flow modeling of GCM-defined accumulation areas

3.1. The Mars GCM of the Laboratoire de Météorologie Dynamique

In order to specify the spatial distribution of the SMB of accu-mulated ice for a glacial flow model, one can arbitrarily choose val-ues, and explore the consequences for specific values andpredictions as we have successfully done above. An alternative ap-proach is to use the results of a GCM that was chosen on the basisof its ability to reproduce the broad conditions that are observedgeologically, as we did in the Tharsis region (e.g., Head andMarchant, 2003; Forget et al., 2006; Fastook et al., 2004, 2005,2008a). The GCM used is the LMD/GCM (Laboratoire de Météorol-ogie Dynamique, Forget et al., 1999); this GCM is able to reproducethe present-day water cycle with good accuracy (Montmessinet al., 2004). Extensive exploration of the parameter space foundthat necessary conditions for persistent ice deposition along thenorthern mid-latitude dichotomy boundary (Fig. 8a) includedmoderate obliquity (25–35�), high eccentricity (0.1) with perihe-lion at Lp = 270�, high dust opacity (1.5–2.5), and a water sourcefrom sublimation of an ice sheet deposited on the flanks of theTharsis volcanoes during a prior period of higher obliquity wheretropical mountain glaciers are observed in the geological record(Head and Marchant, 2003; Shean et al., 2005, 2007; Milkovichet al., 2006; Kadish et al., 2008). The high dust content of the atmo-sphere was necessary to increase its water vapor holding capacity,thereby moving the saturation region to the northern mid-lati-tudes. Precipitation events are then controlled by topographic forc-ing of stationary planetary waves and transient weather systems,producing surface ice distribution and amounts that are consistentwith the geological record. Both lower eccentricity and reversed

0255075100125150175200225250275300325350375400425450475500525550575600625650675700Thickness [3020]

Time=1000000

0102030405060708090100110120130140150160170180190200210220230240250Velocity [3020]

Time=1000000

0255075100125150175200225250275300325350375400425450475500525550575600625650675700Thickness [3030]

Time=1500000

0102030405060708090100110120130140150160170180190200210220230240250Velocity [3030]

Time=1500000

Fig. 7. A region further north near the dichotomy boundary (area G in Fig. 1b, THEMIS VIS V11208010 with model thickness and velocity at 1000 and 1500 Ka) far from thealcoves and in the central part of the valley where ice flow is down-valley. LVF bifurcates (bottom right) and flows around a massif forming a broad up-flow collar and adiffuse, down-flow ‘‘wake’’. LVF in the narrow pass between massifs is compressed. The lobate LVF extends into the northern lowlands at top left. Note the arcuate, piedmont-like configuration of the lobe and the high density of pits. Model results show the flow clearly narrowing between the two bedrock highs and increasing the velocity to valuesin excess of 200 mm/a.


perihelion did not produce persistent ice cover consistent withgeological observations. This use of a GCM tuned to produce a po-sitive SMB where glaciers are interpreted to have existed, informsus about the necessary state of the climate at the time the glacierswere deposited.

3.2. Specific global circulation model results for the northern mid-latitude dichotomy boundary area

Work with the LMD/GCM (Forget et al., 1999; Montmessin et al.,2004; Levrard et al., 2004; Hourdin et al., 1993) has provided aframework so that a map of potential accumulation rates for thedichotomy boundary region can be produced (Madeleine et al.,2009), which can then be used in an ice sheet model to describe apossible ice sheet with associated valley glaciation observed in theglacial geology (Head et al., 2006a; Fastook et al., 2008b). The distri-bution of positive SMB regions for the Madeleine et al. (2009) cli-mate simulation showing best conditions for development of themid-latitude glaciation are illustrated in Fig. 8a. The valley glaciationregion described and tested above (Head et al., 2006a) and furtherdiscussed here lies in the area designated by the arrow in Fig. 8a(see also Fig. 1a, box 5). With peak values of ice accumulation reach-ing 16 mm/a, this pattern of SMB is clearly capable of producing alarge ice sheet along the dichotomy boundary, but will its behavioragree with the geological observations (Fig. 1)?

3.3. Ice sheet modeling results

Clearly such a wide area of positive SMB will create a broad,extensive ice sheet (Marchant and Head, 2008), as opposed to only

the localized valley glaciers observed in the geological record(Fig. 1a; Head et al., 2006a,b, 2010; Levy et al., 2007; Kress andHead, 2009; Morgan et al., 2009; Dickson et al., 2008, 2010; Hauberet al., 2008; Baker et al., 2010). We chose to focus attention againon the Deuteronilus–Protonilus Mensae (DPM) valley system de-scribed in Head et al. (2006a) (Fig. 1b–e). Running the ice sheetmodel at a resolution sufficient to resolve the valley, however, isprohibitive for such a broad area. We utilize instead the embed-ded-grid feature of UMISM. This feature allows us to run abroad-domain, low-resolution grid with a more limited-domain,higher-resolution grid embedded within it. This embedded grid ob-tains boundary condition information from a spatial and temporalinterpolation of the low-resolution grid. This embedded feature al-lows us to nest grids, so that the jump in resolution is not so ex-treme as to produced spurious results.

The grids in the nest used in this model, with topography fromMOLA, are also shown in Fig. 8. Starting with Fig. 8b, the broad gridoutlined by the box in Fig. 8a (with 10,164 nodes and a resolutionof 50 km), the nest progresses to Fig. 8c (with 16,625 nodes and12 km resolution), onto Fig. 8d (with 7521 nodes and 6 km resolu-tion), and finally to Fig. 8e, with the highest resolution (1.6 km, and12,769 nodes). The three outer grids in the nest at 600 Ka areshown in Fig. 9, the time at which growth is stopped and retreatbegins. The three vertical columns contain surface elevation, icethickness, and ice velocity, respectively. The three rows correspondto the three outermost grids (Fig. 8b–d).

On the basis of this glacial accumulation and flow model, it isobserved that the broad pattern of positive SMB in the northernplains (Fig. 8a) builds an ice sheet up to 4 km thick with a volumeclose to 9 million km3 in 600 Ka (Fig. 9, first row, left and middle

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

0 10 20 30 40 50 60 70 80 90 100 110 120

0 10 20 30 40 50 60 70 80 90 100 110 120

0

10

20

30

40

50

60

70

80

0

10

20

30

40

50

60

70

80

0 10 20 30 40 50 60 70 80 90 100 110 120

0 10 20 30 40 50 60 70 80 90 100 110 120

30

35

40

45

50

30

35

40

45

50

15 20 25 30 35 40 45 50 55

15 20 25 30 35 40 45 50 55

30

35

40

45

50

30

35

40

45

50

15 20 25 30 35 40 45 50 55

15 20 25 30 35 40 45 50 55

40

45

40

45

30 35 40

30 35 40

40

45

40

45

30 35 40

30 35 40

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35

-6000

-5500

-5000

-4500

-4000

-3500

-3000

-2500

-2000

-1500

-1000

-500

0

500

1000

1500

2000

2500

3000

(b) (c)

(d)

(e)

(a)

Fig. 8. (a) Mass balance in mm/yr from Madeleine et al. (2009) showing persistent ice deposition along the northern mid-latitude dichotomy boundary. GCM parametersincluded moderate obliquity (25–35�), high eccentricity (0.1) with perihelion at Lp = 270�, high dust opacity (1.5–2.5), and a water source from sublimation of an ice sheetdeposited on the flanks of the Tharsis volcanoes; (b) the outermost of the nested grids with a resolution of 50 km, outlined by the box in (a); (c) with 12 km resolution,outlined in (b); (d) with 6 km resolution, outlined in (c); and (e) with the highest resolution of 1.6 km, outlined in (d).


columns). This is considerably more than our own estimate of aTharsis-region volume of 0.53 million km3 (Fastook et al.,2008a,b), but the Tharsis-region estimate was very conservativeand was constrained to be no greater than the clearly observableglacial deposits mapped in that area. Other estimates of the Tharsisvolumes are considerably higher with much more extensive icesheets. One reconstruction by Kite and Hindmarsh (2007) is 25–54 million km3, a size hard to reconcile with the current volumeof the North Polar Layered Deposits, which is 1.2–1.7 million km3

(Zuber et al., 1998). If we restrict our ice sheet to the next grid inthe nest (Figs. 8c and 9, second row) our volume is close to 2 mil-lion km3, still larger than our estimate for Tharsis, but closer to the

volume of the current North Polar Layered Deposits. We do infact take depletion of the source into account, as that is the rea-son we turn off the accumulation portion of the SMB at 600 Kawhen retreat and collapse of the ice sheet begins, leading tothe patterns of ice thickness shown in the valley complex inFig. 12. Clearly, the actual sources, volumes and transport path-ways of water ice are not fully understood and subject to currentinvestigation (e.g., Mischna et al., 2003; Levrard et al., 2004,2007). Nonetheless, the broad modeled ice sheet covers muchof the adjacent plateau and is particularly thick near the dichot-omy boundary (Fig. 9, middle column); fretted valleys, particu-larly in areas interpreted to be the sites of local valley glacial

20

30

40

50

60

30

-10 0 10 20 30 40 50 60 70

-10 0 10 20 30 40 50 60 70

20

40

50

60

-10 0 10 20 30 40 50 60 70

-10 0 10 20 30 40 50 60 70

20

30

40

50

60

20

30

40

50

60

-10 0 10 20 30 40 50 60 70

-10 0 10 20 30 40 50 60 70

-6 -5 -4 -3 -2 -1 0 1 2 3

30

40

50

20 30 40 50

20 30 40 50

SURFACE [2100] {b}TIME= 600000

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

30

40

50

20 30 40 50

20 30 40 50

THICKNESS [2100] {b}TIME= 600000

-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

30

40

50

30

40

50

20 30 40 50

20 30 40 50

VELOCITY [2100] {b}TIME= 600000

-6 -5 -4 -3 -2 -1 0 1 2 3

40

45

30 35 40

30 35 40

SURFACE [3100] {c}TIME= 600000

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

40

45

30 35 40

30 35 40

THICKNESS [3100] {c}TIME= 600000

-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

45

40

45

30 35 40

30 35 40

VELOCITY [3100] {c}TIME= 600000

-6 -5 -4 -3 -2 -1 0 1 2 3

SURFACE [1500] {a}TIME= 600000

0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

THICKNESS [1500] {a}TIME= 600000

-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

VELOCITY [1500] {a}TIME= 600000

Fig. 9. Surface elevation, thickness, and velocity (columns) for the three outermost grids (rows) of Fig. 8b–d in the DPM lineated valley fill system.


landsystem deposits (e.g., Fig. 1a; Morgan et al., 2009; Headet al., 2006b; Baker et al., 2010; Dickson et al., 2008; Levyet al., 2007), show thick ice accumulations.

The specific DPM valley (Head et al., 2006a; Fig. 1b–e) lies on aGCM grid point where high accumulation rates are predicted (theresolution of the GCM is 5.625� longitude by 3.75� latitude). Thisresults in a ‘‘peninsula’’ of ice that stretches into the highlands ofthe dichotomy boundary with thicknesses approaching 3 km. Note

that in the logarithmic velocity scale, a value of 1 corresponds to avelocity of 10 mm/a. The DPM valley (Fig. 1b–e) lies in a saddle re-gion where ice flow is clearly faster, �100 mm/a, and is channeli-zed in the trunk of the valley.

This channelization is more evident in the highest-resolutiongrid (Fig. 8e). In Fig. 10 the evolution during growth of the valleyice complex (top to bottom: surface, thickness, and velocity) isshown at 100 Ka intervals (left to right: 100–600 Ka). In the surface

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 35-6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

32 33 34 35

32 33 34 35

SURFACE [4001] {d}TIME= 100000

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35-6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

32 33 34 35

32 33 34 35


32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35-6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

32 33 34 35

32 33 34 35


32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35-6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

32 33 34 35

32 33 34 35


32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35-6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

32 33 34 35

32 33 34 35


40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 35-6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35


40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 350.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

40

41

42

43

32 33 34 35

32 33 34 35

THICKNESS [4001] {d}TIME= 100000

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

32 33 34 35

32 33 34 35

VELOCITY [4001] {d}TIME= 100000

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 350.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

40

41

42

43

32 33 34 35

32 33 34 35


32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 350.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

40

41

42

43

32 33 34 35

32 33 34 35


32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 350.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

40

41

42

43

32 33 34 35

32 33 34 35


32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 350.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

40

41

42

43

32 33 34 35

32 33 34 35


32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 350.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35


32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

32 33 34 35

32 33 34 35


32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

32 33 34 35

32 33 34 35


32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

32 33 34 35

32 33 34 35


32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

32 33 34 35

32 33 34 35


40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35


Fig. 10. Growth to 600 Ka, showing surface (m), thickness (m), and velocity (log 10(mm/a)) at 100, 200, 300, 400, 500, and 600 Ka in the DPM lineated valley fill system.

Fig. 11. Ice evolution versus time for a typical point in Deuteronilus (45�N28.1�E).By tracking increasing and decreasing thickness, the accumulation and ablationcomponents of the mass balance can be separated.


figures (Fig. 10, top row), we see a gradual progressive drowning ofthe valley topography. Thickness (Fig. 10, middle row) begins witha uniform mantling, but quickly evolves into thicker ice in the val-leys (4 km) but much thinner over the plateaus (less than 3 km)and aretes (�1 km). Recalling that ice flow directions are downthe surface topographic gradient, perpendicular to surface eleva-tion contours, we see the organization of a coherent flow patternwith maximum velocity (Fig. 10, lower row) in the valleys(�100 mm/a) but one, two, and even three orders of magnitudeless over the more thinly covered aretes and plateaus betweenthe valley trunks. Thus, to a first order, the general distributionof ice accumulation, as predicted by the LMD/GCM, readilyreproduces regional plateau glaciation, and shows that glacial flowis focused into the fretted terrain valleys due to the effect of thepre-existing topography.

What is the fate of the regional ice sheet during its sublimationand retreat? As was mentioned, with the relatively high accumula-tion rates predicted by Madeleine et al. (2009), growth of a signif-icant ice sheet is relatively rapid, and we grew the ice sheet understeady climate conditions for 600 Ka. At this point, we removed thepositive accumulation component of the SMB, leaving only thenegative sublimation component. This might be expected to occur,

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

-6 -5 -4 -3 -2 -1 0 1 2 3


40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35

40

41

40

41

42

43

32 33 34 35

32 33 34 35

40

41

43

32 33 34 350.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

41

32 33 34 35


40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35


40

41

42

43

32 33 34 35

32 33 34 35-6 -5 -4 -3 -2 -1 0 1 2 3


40

41

42

43

32 33 34 35

32 33 34 35

40

41

43

-6 -5 -4 -3 -2 -1 0 1 2 3

SURFACE [4105] {d}TIME= 1100000

40

41

42

43

32 33 34 35

32 33 34 35

32 33 34 35

-6 -5 -4 -3 -2 -1 0 1 2 3

32 33 34 35

SURFACE [4106] {d}TIME= 1200000

40

41

42

43

32 33 34 35

32 33 34 35-6 -5 -4 -3 -2 -1 0 1 2 3

SURFACE [4107] {d}TIME= 1300000

40

41

42

43

32 33 34 35

32 33 34 35-6 -5 -4 -3 -2 -1 0 1 2 3

SURFACE [4108] {d}TIME= 1400000

41

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 350.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

40

41

42

43

32 33 34 35

32 33 34 35


40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 350.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

40

41

42

43

32 33 34 35

32 33 34 35


40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 35

42

32 33 34 35

32 33 34 350.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

32 33 34 35

32 33 34


40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

32 33 34 35

32 33 34 35

32 33 34 35

32 33 34 350.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5

32 33 34 35


40

41

42

43

32 33 34 35

32 33 34 3532 33 340.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5


40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35


40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35


40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

41

43

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43 43

32 33 34 35

32 33 34 35


43

40

41

42

43

32 33 34 35

32 33 34 35

43

40

41

42

43

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

43

40

41

42

43

32 33 34 35

32 33 34 35


40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35-10 -9 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3

40

41

42

43

40

41

42

43

32 33 34 35

32 33 34 35


Fig. 12. Surface (m), thickness (m), and velocity (log 10(mm/a)) at 800, 900, 1100, 1200, 1300, and 1400 Ka in the DPM lineated valley fill system.


for example, if the source region in the Tharsis region becomes de-pleted, or if the obliquity changed (which is to be expected onapproximately this time scale; Laskar et al., 2004). We removethe positive accumulation component by separating the positive(accumulation) and negative (ablation) portions of the SMB pro-vided by the GCM results.

A typical modeled point from the GCM results is shown inFig. 11. The amount of ice equivalent on the surface is shownas a function of time throughout the martian year. Ice amountbegins at zero and climbs to close to 3 kg/m2 (a period of posi-tive accumulation). Ice then begins to thin (negative ablation).We allowed the ice amount to go negative in the LMD/GCM tohave access to the amount of ice that would be ablated, if therewere any there to ablate. After this period of sublimation, a sec-ond period of positive accumulation occurs, ending the year witha net accumulation of �10 kg/m2. It is this number that is re-ported in Fig. 8a. In this analysis, on the other hand, we keeptrack of the separate amounts of positive accumulation (icegrowing) and negative ablation (ice thinning). Our difference be-tween these two components of the SMB corresponds exactly tothe values reported by Madeleine et al. (2009) in Fig. 8a, but wehave a measure of potential negative mass balance (a pointwhere the final endpoint of ice amount was negative) and we

have the possibility to modify either the positive part (turn offthe snowfall) or the negative part (increase or decrease sublima-tion) while keeping the rest of the climate true to the GCM re-sults (the spatial distribution of mass balance and surfacetemperatures).

We now explore the retreat of the regional ice sheet. In thismodel we do not include any effects of debris cover that mightbe produced as the valley and dichotomy boundary scarps be-come exposed and could potentially provide a rock and debriscover to armor the ice and protect it from sublimation. Instead,the retreating ice is assumed to be pure ice. The configurationof the DPM valley at 600 Ka, the point at which retreat is al-lowed to begin by removing the positive portion of the massbalance, is shown in the right-most column of Fig. 10. The evo-lution of the ice sheet as this retreat progresses is shown inFig. 12.

(1) By 800 Ka (Fig. 12, first column), 200 Ka after retreat begins,the surface is dropping and the underlying topography isbecoming more evident as the first aretes emerge as nunat-aks. Velocities are considerably reduced to 1 mm/a in thevalleys and 0.001 mm/a on the plateaus, although stillclearly channelized in the valley proper.


(2) As ice sheet collapse continues, the plateaus begin to emergeby 900 Ka (Fig. 12, second column), and are fully cleared by1100 Ka (Fig. 12, third column).

(3) By 1200 Ka (Fig. 12, fourth column), ice in the valleys is lessthan 500 m thick, while on the adjacent plains ice thicknessis still over 1 km.

(4) By 1300 Ka (Fig. 12, fifth column), the remaining patches ofice separate, and the valleys are ice-free (again, assuming nodebris cover).

(5) Of particular interest is the configuration at 1400 Ka (Fig. 12,right-most column), where only limited patches remain sur-rounding nunataks on the plains outside the mouth of thevalley, a configuration similar to the patterns for lobate deb-ris aprons reported by Ostrach and Head (2008), Ostrachet al. (2008), Holt et al. (2008a,b), Plaut et al. (2008), Safaei-nili et al. (2009) and Kress and Head (2009). These relation-ships predict that the lobate debris aprons (LDA) mightslightly postdate the lineated valley fill (LVF) in their forma-tion. These remnant patches occur along the base of steeptopographic scarps because the ice is thickest where it flowsfrom the high to lower terrain across a step in bed elevation.While not modeled in this experiment, the potential for deb-ris armoring beneath scarps would increase the likelihood ofLDA preservation.

In summary, coupling the regional accumulation rates of theLMD/GCM (Madeleine et al., 2009) with a glacial flow model showsthat the resulting regional plateau and plains glaciation that wouldhave characterized the broad area during the period of glaciation isquite capable of producing results during its collapse that are con-sistent with the alpine valley glaciation interpretation of the LVF.Mapping of the patterns of retreat during the waning stages ofthe glaciation shows that the valley walls are eventually exposedand could readily serve as sources of debris to produce the lineatedvalley fill and related flow patterns seen in numerous places alongthe dichotomy boundary (Fig. 1a–e). Furthermore, the final config-uration of the ice sheet upon its retreat is that of glacial ice sur-rounding massifs in the plains just north of the dichotomyboundary, a pattern very similar to that of lobate debris apronsthat are currently observed to surround the massifs. We are cur-rently using the glacial flow models illustrated in Fig. 12 to modelthe exposure of the rock cliffs and massifs and assess the effectsthat the addition of cliff-derived rock debris will have on the sub-limation rates and flow rates of the lineated valley fill and lobatedebris aprons (Fastook et al., 2010; e.g., Kowalewski et al., 2011).

4. Conclusions

A fretted valley system located at the northern mid-latitudedichotomy boundary and containing lineated valley fill with exten-sive flow-like features interpreted to be glacial in origin (Headet al., 2006a) has been modeled using glacial flow models linkedto atmospheric general circulation models. The glacial flow modelresults are consistent with the geological features observedand with the interpretation of these as alpine valley glaciallandsystems.

In the first glacial flow model simulation, sources were modeledin the alcoves only and were found to be consistent with the alpinevalley glaciation interpretation for various environments of flow inthe system. These results indicated that the currently observed LVFdeposits could be the result of initial ice accumulation in the al-coves, accompanied by debris cover that led to advancing alpineglacial landsystems to the extent observed today, with preserva-tion of their flow texture and the underlying ice during downwa-sting in the waning stages of glaciation.

In the second glacial flow model simulation, the regional accu-mulation patterns predicted by the LMD/GCM were used (Made-leine et al., 2009). This glacial flow model simulation produced amuch wider region of thick ice accumulation, and significant glaci-ation on the plateaus and in the regional plains surrounding thedichotomy boundary (Figs. 1a and 8a). Deglaciation produceddecreasing ice thicknesses, with flow centered on the fretted val-leys. As plateaus lost ice, and scarps and cliffs of the valley anddichotomy boundary walls were exposed, there was considerablepotential for the production of a rock debris-cover that could pre-serve the underlying ice and the surface flow patterns seen today.In this model, the lineated valley fill and lobate debris aprons werethe product of final retreat and downwasting of a much more re-gional glacial landsystem (Marchant and Head, 2008), rather thanthe maximum extent of an alpine valley glacial landsystem.

From these glacial flow models, information can be derivedabout the possible timing of the growth of these systems. In the al-cove-only case, several millions of years are required for the flowsystem to achieve a pattern consistent with the geological record.Because the alcove-only case uses specified areal distribution of icedeposition we cannot constrain its occurrence to any specific timeinterval.

Coupling the ice sheet model with a GCM yields a pattern of icesheet development, which during its collapse also yields a config-uration consistent with the geological observations. Because ofthe high accumulation rates predicted by the LMD/GCM, the icesheet grows much more rapidly than the alcove-only case. As theice sheet collapses, presumably due to the exhaustion of the mois-ture source in the Tharsis region or spin-axis/orbital changes, aconfiguration emerges of a limited valley glacier system compati-ble with the geological observations. Importantly, as the surfacelowers to the point where armoring debris would be available fromthe cliff scarps, the configuration resembles the alcove-only case.At this point, if debris is supplied from the adjacent newly exposedrock cliffs, we would expect the sublimation rates over the debris-covered valley-floor glaciers to be reduced by orders of magnitude(Kowalewski et al., 2006) potentially preserving ice even now inthe valley trunks. This will also be the case with isolated massifsat the mouths of the valleys, leading to the possibility that the lo-bate debris aprons are ice-cored remnants of a much more exten-sive ice sheet that covered the area during climatically appropriateperiods. We are currently exploring these options by modeling theaccumulation of debris once cliffs are exposed, and the evolutionand flow of such debris-covered glaciers (Fastook et al., 2010).

Acknowledgments

We gratefully acknowledge the scientists and engineers con-tributing to the success of the NASA Mars Observer, Mars Odyssey,and Mars Reconnaissance Orbiter, and the ESA Mars Express mis-sions. JWH gratefully acknowledges support from NASA throughthe Mars Data Analysis Program and the Mars Express HRSC Team.J.L.F. gratefully acknowledges support from the NSF for develop-ment of the terrestrial version of UMISM through many grants.D.R.M. and J.W.H. gratefully acknowledge support from the NSFthrough Grant ANT-0944702. Thanks are extended to Anne Côtéfor assistance in manuscript preparation.

References

Baker, V.R., 2001. Water and the martian landscape. Nature 412, 228–236.Baker, V.R., Strom, R.G., Gulick, V.C., Kargel, J.S., Komatsu, G., Kale, V.S., 1991.

Ancient oceans, ice sheets and hydrological cycle on Mars. Nature 352, 589–594.

Baker, D., Head, J.W., Marchant, D.R., 2010. Flow patterns of lobate debris apronsand lineated valley fill north of Ismeniae Fosse, Mars: Evidence for extensive


mid-latitude glaciation in the Late Amazonian. Icarus 207, 186–209.doi:10.1016/j.icarus.2009.11.017.

Benn, D.I., Kirkbride, M.P., Owen, L.A., Brazier, V., 2003. Glaciated valleylandsystems. In: Evans, D.J.A. (Ed.), Glacial Landsystems. Oxford Universitypress, pp. 372–406.

Boynton, W.V. et al., 2002. Distribution of hydrogen in the near surface of Mars:Evidence for subsurface ice deposits. Science 297, 81–85.

Carr, M.H., 1995. The martian drainage system and the origin of networks andfretted channels. J. Geophys. Res. 100, 7479–7507.

Carr, M.H., 1996. Water on Mars. Oxford University Press, New York, NY, 229pp.Carr, M.H., 2001. Mars Global Surveyor observations of martian fretted terrain. J.

Geophys. Res. 106, 23571–23593.Carr, M.H., Head, J.W., 2010. Geologic history of Mars. Earth Planet. Sci. Lett. 294,

185–203. doi:10.1016/j.epsl.2009.06.042.Carr, M.H., Schaber, G.G., 1977. Martian permafrost features. J. Geophys. Res. 82,

4039–4054.Chuang, F.C., Crown, D.A., 2005. Surface characteristics and degradational history of

debris aprons in the Tempe Terra/Mareotis fossae region of Mars. Icarus 179,24–42. doi:10.1016/j.icarus.2005.05.014.

Clifford, S.M., 1993. A model for the hydrologic and climatic behavior of water onMars. J. Geophys. Res. 98, 10973–11016.

Clifford, S.M., Parker, T.J., 2001. The evolution of the martian hydrosphere:Implications for the fate of a primordial ocean and the current state of thenorthern plains. Icarus 154, 40–79.

Craddock, R.A., Howard, A.D., 2002. The case for rainfall on a warm, wet early Mars.J. Geophys. Res. 107, 21–36.

Dickson, J.L., Head, J.W., Marchant, D.R., 2008. Late Amazonian glaciation at thedichotomy boundary of Mars: Evidence for glacial thickness maxima andmultiple glacial phases. Geology 36, 411–414.

Dickson, J.L., Head, J.W., Marchant, D.R., 2010. Kilometer-thick ice accumulation andglaciation in the northern mid-latitudes of Mars: Evidence for crater-fillingevents in the Late Amazonian at the Phlegra Montes. Earth Planet. Sci. Lett..doi:10.1016/j.epsl.2009.08.031.

Evans, D.J.A. (Ed.), 2003. Glacial Landsystems. Edward Arnold, London, 532pp.Eyles, N., 1983. The glaciated valley landsystem. In: Eyles, N. (Ed.), Glacial Geology.

Pergamon Press, pp. 91–110.Fassett, C.I., Head, J.W., 2008. Valley network-fed, open-basin lakes on Mars:

Distribution and implications for Noachian surface and subsurface hydrology.Icarus 198, 37–56. doi:10.1016/j.icarus.2008.06.016.

Fastook, J.L., 1993. The finite-element method for solving conservation equations inglaciology. Comput. Sci. Eng. 1, 55–67.

Fastook, J.L., Head, J.W., Marchant, D., Shean, D., 2004. Ice sheet modeling:Terrestrial background and application to Arsia Mons lobate deposit, Mars.Lunar Planet. Sci. XXXV. Abstract 1352.

Fastook, J.L., Head, J.W., Marchant, D., Shean, D., 2005. Ice sheet modeling: Massbalance relationships for map-plane ice sheet reconstruction: Applications toTharsis Montes glaciation. Lunar Planet. Sci. XXXVI. Abstract 1212.

Fastook, J.L., Shean, D.E., Head, J.W., Marchant, D., 2006. Ice sheet modeling duringhigh-obliquity climates on Mars: Application to Tharsis Montes tropicalmountain glaciation. Lunar Planet. Sci. XXXVII. Abstract 1794.

Fastook, J.L., Head, J.W., Marchant, D., Forget, F., 2008a. Tropical mountain glacierson Mars: Altitude-dependence of ice accumulation, accumulation conditions,formation times, glacier dynamics, and implications for planetary spin-axis/orbital history. Icarus 198, 305–317.

Fastook, J.L., Head, J.W., Marchant, D., 2008b. Dichotomy boundary glaciationmodels: Implications for timing and glacial processes. Lunar Planet. Sci. XXXIX.Abstract 1109.

Fastook, J.L., Head, J.W., Madeleine, J.-B., Forget, F. Marchant, D., 2010. Modeling anice-rich lobate debris apron in Deuteronilus Mensae. Lunar Planet. Sci. XLI.Abstract 1823.

Feldman, W.C. et al., 2002. Global distribution of neutrons from Mars: Results fromMars Odyssey. Science 297, 75–78.

Forget, F. et al., 1999. Improved general circulation models of the martianatmosphere from the surface to above 80 km. J. Geophys. Res. 104, 24155–24176.

Forget, F., Haberle, R.M., Montmessin, F., Levrard, B., Head, J.W., 2006. Formation ofglaciers on Mars by atmospheric precipitation at high obliquity. Science 311,368–371.

Hauber, E., van Gasselt, S., Chapman, M.G., Neukum, G., 2008. Geomorphic evidencefor former lobate debris aprons at low latitudes on Mars: Indicators of themartian paleoclimate. J. Geophys. Res. 113, E02007. doi:10.1029/2007JE002897.

Head, J.W., Marchant, D.R., 2003. Cold-based mountain glaciers on Mars: WesternArsia Mons. Geology 31, 641–644.

Head III, J.W., Marchant, D.R., 2009. Inventory of Ice-related deposits on Mars:Evidence for burial and long-term sequestration of ice in non-polar regions andimplications for the water budget and climate evolution. Lunar Planet. Sci. XL.Abstract 1356.

Head, J.W., Mustard, J.F., Kreslavsky, M.A., Milliken, R.E., Marchant, D.R., 2003a.Recent ice ages on Mars. Nature 426, 797–802.

Head, J.W., Carr, M., Russell, P.S., Fassett, C.I., 2003b. Late Noachian hydrologicalcycle: Summary of the late Noachian hydrological cycle 7. Vernadsky-BrownMicrosymposium 38. Abstract ms032. .

Head, J.W., Marchant, D.R., Agnew, M.C., Fassett, C.I., Kreslavsky, M.A., 2005.Regional mid-latitude late Amazonian valley glaciers on Mars: Origin oflineated valley fill and implications for recent climate change. Lunar Planet.Sci. XXXVI. Abstract 1208.

Head, J.W., Nahm, A.L., Marchant, D.R., Neukum, G., 2006a. Modification of thedichotomy boundary on Mars by Amazonian mid-latitude regional glaciation.Geophys. Res. Lett. 33, L08S03. doi:10.1029/2005GL024360.

Head, J.W., Marchant, D.R., Agnew, M.C., Fassett, C.I., Kreslavsky, M.A., 2006b.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. doi:10.1016/j.epsl.2005.11.016.

Head, J.W., Marchant, D.R., Dickson, J.L., Kress, A.M., Baker, D.M.H., 2010. Northernmid-latitude glaciation in the Late Amazonian period of Mars: Criteria for therecognition of debris-covered glacier and valley glacier landsystem deposits.Earth Planet. Sci. Lett. 294, 306–320. doi:10.1016/j.epsl.2009.06.041.

Holt, J.W. et al., 2008a. Radar sounding evidence for ice within lobate debris apronsnear Hellas Basin, mid-southern latitudes of Mars. Lunar Planet. Sci. XXXIX.Abstract 2441.

Holt, J.W. et al., 2008b. Radar sounding evidence for buried glaciers in thesouthern mid-latitudes of Mars. Science 322, 1235–1238. doi:10.1126/science.1164246.

Hourdin, F., Le Van, P., Forget, F., Talagrand, O., 1993. Meteorological variability andthe annual surface pressure cycle on Mars. J. Atmos. Sci. 50, 3625–3640.

Kadish, S.J., Head, J.W., Parsons, R.L., Marchant, D.R., 2008. The Ascraeus Mons fan-shaped deposit: Volcano ice interactions and the climatic implications of cold-based tropical mountain glaciation. Icarus 197, 84–109.

Keiffer, H.H., Zent, A.P., 1992. Quasi-periodic climate change on Mars. In: Keiffer,H.H. et al. (Eds.), Mars. University of Arizona Press, pp. 1180–1233.

Kite, E.S., Hindmarsh, R.C.A., 2007. Did ice streams shape the largest channels onMars? Geophys. Res. Lett. 34, L19202. doi:10.1029/2007GL030530.

Kochel, R.C., Baker, V.R., 1981. Backwasting and debris accumulation along martianescarpments. In: Proceedings of the Third Colloquium on Planetary Water. StateUniversity of New York, Buffalo, New York, pp. 53–58.

Kochel, R.C., and Peake, R.T., 1984. Quantification of waste morphology in Martianfretted terrain. In: Proc. Lunar Sci. Conf. 15. J. Geophys. Res. Suppl. 89, C336–C350.

Kowalewski, D.E., Marchant, D.R., Levy, J.S., Head, J.W., 2006. Quantifying low ratesof summertime sublimation for buried glacier ice in Beacon Valley, Antarctica.Antarct. Sci. 18 (3), 421–428.

Kowalewski, D.E., Marchant, D.R., Swanger, K.M., Head, J.W., 2011. Modeling vapordiffusion within cold and dry supraglacial tills of Antarctica: Implications forthe preservation of ancient ice. Geomorphology 126, 159–173.

Kress, A.M., Head, J.W., 2008. Ring-mold craters in lineated valley fill and lobatedebris aprons on Mars: Evidence for subsurface glacial ice. Geophys. Res. Lett.35, L23206. doi:10.1029/2008GL035501.

Kress, A., Head, J.W., 2009. Lineated valley fill and lobate debris aprons in theDeuteronilus Mensae region, Mars: Implications for regional glaciation. LunarPlanet. Sci. XL. Abstract 1632.

Laskar, J., Correia, A.C.M., Gastineau, M., Joutel, F., Levrard, B., Robutel, P., 2004. Longterm evolution and chaotic diffusion of the insolation quantities of Mars. Icarus170, 343–364.

Levrard, B., Forget, F., Montmessin, F., Laskar, J., 2004. Recent ice-rich depositsformed at high latitudes on Mars by sublimation of unstable equatorial iceduring low obliquity. Nature 431, 1072–1075.

Levrard, B., Forget, F., Montmessin, F., Laskar, J., 2007. Recent formation andevolution of northern martian polar layered deposits as inferred from a GlobalClimate Model. J. Geophys. Res. 112, E06012.

Levy, J.S., Head, J.W., Marchant, D.R., 2007. Lineated valley fill and lobate debrisapron stratigraphy in Nilosyrtis Mensae, Mars: Evidence for phases of glacialmodification of the dichotomy boundary. J. Geophys. Res. 112, E08004.doi:10.1029/2006JE002852.

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 mantleprocesses. Icarus 202, 462–476. doi:10.1016/j.icarus.2009.02.018.

Li, H., Robinson, M.S., Jurdy, D.M., 2005. Origin of martian northern hemispheremid-latitude lobate debris aprons. Icarus 176, 382–394.

Lucchitta, B.K., 1981. Mars and Earth: Comparison of cold climate features. Icarus45, 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., Millour, E., 2009.Amazonian northern mid-latitude glaciation on Mars: A proposed climatescenario. Icarus 203, 390–405.

Malin, M.C., Edgett, K.S., 2000. Evidence for recent groundwater seepage and surfacerunoff on Mars. Science 288, 2330–2335.

Mangold, N., 2003. Geomorphic analysis of lobate debris aprons on Mars at MarsOrbiter Camera scale: Evidence for ice sublimation initiated by fractures. J.Geophys. Res. 108. doi:10.1029/2002JE001885.

Marchant, D.R., Head, J.W., 2004. Microclimate zones in the Dry Valleys ofAntarctica: Implications for landscape evolution and climate change on Mars.Lunar Planet. Sci. XXXV. Abstract 1405.

Marchant, D.R., Head, J.W., 2005. Equilibrium landforms in the Dry Valleys ofAntarctica: Implications for landscape evolution and climate change on Mars.Lunar Planet. Sci. XXXVI. Abstract 1421.

Marchant, D.R., Head, J.W., 2006. Glacial landsystems on Mars: Integrating landformassemblages, glaciations, and climate cycles. Lunar Planet. Sci. XXXVII. Abstract1422.

Marchant, D.R., Head, J.W., 2007. Antarctic Dry Valleys: Microclimate zonation,variable geomorphic processes, and implications for assessing climate changeon Mars. Icarus 192, 187–222. doi:10.1016/j.icarus.2007.06.018.


http://dx.doi.org/10.1016/j.epsl.2009.06.042




http://dx.doi.org/10.1029/2007JE002897

http://dx.doi.org/10.1029/2005GL024360



http://dx.doi.org/10.1126/science.1164246


http://dx.doi.org/10.1029/2007GL030530

http://dx.doi.org/10.1029/2008GL035501

http://dx.doi.org/10.1029/2006JE002852


http://dx.doi.org/10.1029/2002JE001885



Marchant, D.R., Head, J.W., 2008. Kilometer-thick ice-sheets in the northern mid-latitudes in the Amazonian: Analogs from the East Antarctic ice sheet and theDry Valleys. Lunar Planet. Sci. XXXIX. Abstract 2097.

Marchant, D.R., Mackay, S.L., Head, J.W., Kowalewski, D.E., 2010. Documentingmicroclimate variation and the distribution of englacial debris in MullinsGlacier, Antarctica: Implications for the origin, flow and modification of LDAand LVF on Mars. Lunar Planet. Sci. XLI. Abstract 2601.

Masson, P., Carr, M.H., Costard, F., Greeley, R., Hauber, E., Jaumann, R., 2001.Geomorphologic evidence for liquid water. In: Kallenbach, R. et al. (Eds.),Chronology and Evolution of Mars. Kluwer Academic Publishers, Boston, pp.333–364.

McGill, G.E., 2000. Crustal history of north central Arabia Terra, Mars. J. Geophys.Res. 105, 6945–6959.

Milkovich, S.M., Head, J.W., Marchant, D.R., 2006. Debris-covered piedmont glaciersalong the northwest flank of the Olympus Mons scarp: Evidence for low-latitude ice accumulation during the Late Amazonian of Mars. Icarus 181, 388–407.

Mischna, M.A., Richardson, M.I., Wilson, R.J., McCleese, D.J., 2003. On the orbitalforcing of martian water and CO2 cycles: A general circulation model study withsimplified volatile schemes. J. Geophys. Res. 108, 5062. doi:10.1029/2003JE002051.

Montmessin, F., Forget, F., Rannou, P., Cabane, M., Haberle, R.M., 2004. Origin androle of water ice clouds in the martian water cycle as inferred from a generalcirculation model. J. Geophys. Res. 109, E10004.

Morgan, G.A., Head, J.W., Marchant, D.R., 2009. Lineated valley fill (LVF) and lobatedebris aprons (LDA) in the Deuteronilus Mensae northern dichotomy boundaryregion, Mars: Constraints on the extent, age, and episodicity of Amazonianglacial events. Icarus 202, 22–38. doi:10.1016/j.icarus.2009.02.017.

Ostrach, L.R., Head, J.W., 2008. Volumetric estimates of Amazonian lobate debrisaprons (LDA) in the mid-latitudes of Mars: Support for the presence ofsignificant water–ice. Lunar Planet. Sci. XXXIX. Abstract 1652.

Ostrach, L.R., Head, J.W., Kress, A.M., 2008. Ring-mold craters (RMC) in lobate debrisaprons (LDA) in the Deuteronilus Mensae region of Mars: Evidence for shallowsubsurface glacial ice in lobate debris aprons. Lunar Planet. Sci. XXXIX. Abstract2422.

Phillips, R.J. et al., 2008. Mars north polar deposits: Stratigraphy, age, andgeodynamical response. Science 320. doi:10.1126/science.1157546.

Pierce, T.L., Crown, D.A., 2003. Morphologic and topographic analyses of debrisaprons in the eastern Hellas region, Mars. Icarus 163, 46–65.

Plaut, J.J. et al., 2008. Radar evidence for ice in lobate debris aprons in the mid-northern latitudes of Mars. Lunar Planet. Sci. XXXIX. Abstract 2290.

Plaut, J.J. et al., 2009. Radar evidence for ice in lobate debris aprons in the mid-northern latitudes of Mars. Geophys. Res. Lett. 36, L02203. doi:10.1029/2008GL036379.

Plaut, J.J. et al., 2010. Thick ice deposits Deuteronilus Mensae, Mars: Regionaldistribution form radar sounding. Lunar Planet. Sci. XL1. Abstract 2454.

Post, A., Lachapelle, E.R., 2000. Glacier Ice. University of Washington Press, Seattle,Washington.

Safaeinili, A. et al., 2009. New radar evidence for glaciers in Mars Phlegra Montesregion. Lunar Planet. Sci. XL. Abstract 1988.

Sharp, R., 1973. Mars: Fretted and chaotic terrains. J. Geophys. Res. 78, 4073–4083.Shean, D.E., Marchant, D.R., 2010. Seismic and GPR surveys of Mullins Glacier,

McMurdo Dry Valleys, Antarctica: Ice thickness, internal structure andimplications for surface ridge formation. J. Glaciol. 56 (195), 48–64.

Shean, D.E., Head, J.W., Marchant, D.R., 2005. Origin and evolution of a cold-basedtropical mountain glacier on Mars: The Pavonis Mons fan-shaped deposit. J.Geophys. Res. 110, E05001.

Shean, D.E., Head, J.W., Fastook, J.L., Marchant, D.R., 2007. Recent glaciations at highelevations on Asia Mons, Mars: Implications for the formation and evolution oflarge tropical mountain glaciers. J. Geophys. Res. 112, E03004.

Squyres, S.W., 1978. Martian fretted terrain: Flow of erosional debris. Icarus 34,600–613.

Squyres, S.W., 1979. The distribution of lobate debris aprons and similar flows onMars. J. Geophys. Res. 84, 8087–8096.

Thomas, P., Squyres, S., Herkenhoff, K., Howard, A., Murray, B., 1992. Polar depositsof Mars. In: Keiffer, H.H. (Ed.), Mars. University of Arizona Press, Tucson, pp.767–795.

Williams, R.S., 1978. Geomorphic processes in Iceland and on Mars: A comparativeappraisal from orbital images. Abstr. Programs – Geol. Soc. Am. 10, 517.

Zuber, M.T. et al., 1998. Observations of the north polar region of Mars from theMars Orbiter Laser Altimeter. Science 282, 2053–2060.

Zurek, R.W., 1992. Comparative aspects of the climate of Mars: An introduction tothe current atmosphere. In: Keiffer, H.H. (Ed.), Mars. University of Arizona Press,Tucson, pp. 799–817.

http://dx.doi.org/10.1029/2003JE002051

http://dx.doi.org/10.1029/2003JE002051



http://dx.doi.org/10.1029/2008GL036379

http://dx.doi.org/10.1029/2008GL036379

evidence for amazonian northern mid-latitude regional ... · of ﬁnal retreat and downwasting of a...

Documents