hydrologic evolotion of gale crater

8/13/2019 Hydrologic Evolotion of Gale Crater

1/11

Icarus 139 , 235245 (1999)Article ID icar.1999.6099, available online at http://www.idealibrary.com on

Hydrogeologic Evolution of Gale Crater and Its Relevanceto the Exobiological Exploration of Mars

Nathalie A. Cabrol and Edmond A. Grin

SETI/NASA Ames Research Center, Center for Mars Exploration, MS 239-20, Moffett Field, California 94035-1000E-mail: [email protected]

Horton E. Newsom

University of New Mexico, Institute of Meteoritics and Department of Earth and Planetary Sciences, Albuquerque, New Mexico 87131

Ragnhild Landheim

SETI Institute/NASA Ames Research Center, Center for Mars Exploration, MS 239-20, Moffett Field, California 94035-1000

and

Christopher P. McKay

NASA Ames Research Center, Space Science Division MS 245-3, Moffett Field, California 94035-1000

Received February 13, 1998; revised January 4, 1999

ThepresenceofanAmazonianimpactcraterlakeintheNoachiancraterGale (located in theAeolisnorthwest subquadrangle of Mars)is indicated by evidence from young oor deposits, streamlined ter-races, layers, and channels observed on the central sedimentarydeposit. Evidence for the lling of this lake by two processes is de-scribed: (a) the drainage of the aquifer in the Aeolis Mensae region,supported byextended mass-wasting and rim sliding in thecrateratthe contact with the mensae and (b) the overspilling of the northernrim byan Amazonian southtransgressionof theE lysium Basin.Thislast hypothesis is supported by hydrologic featuressuch as channelsand channel-likedepressions northof thecraterand bythe crescent-like shape of the central sedimentary deposit. The presence of animpact melt sheet and uplifted central peak may have also gener-ated hydrothermal activity, including an early crater lake, shortlyafter the formation of the crater in the Noachian period. With time,decreasing heat ux, and changing climates Gale may have expe-

rienced transitions in aqueous environments from warm and wetto cold and ice-covered water that could have provided suitableoases for various communities of microorganisms. Preservation of the biological and climatic record may have been favored in thispaleolacustrine environment, which probably occured episodicallyover two billion years. c 1999 Academic P ress

Key Words: Mars; Mars surface; exobiology; impact processes.

INTRODUCTION

This study of Gale crater shows that a late lacustrine episodeoccured in the 170-km diameter crater during the Amazonian.

We propose a bathymetric model based on topographic datduced from paleolacustrine morphologies. Results suggestthe lake formation is related to (a) thesapping of thesurroundaquifer in the Aeolis Mensae region, (b) the rose of the wlevel of the Elysium Basin, and (c) the potential geothermaergy from the central peak region. A lake within Gale cduring the Amazonian provides another piece of evidencewater was present in signicant quantity at the surface of in a relatively recent past.

Amazonian uvial features are observed all across the plThey are not as numerous as during the Noachian/Hespeboundary period, but their size is often much larger. In tion, most of them have a common characteristic: their soof water can be related to ground-ice/heat ux interactioninstance, we will refer to a nonexhaustive list of typical exples: the sources of Maadim Vallis are related to volcanic

faults regions (Landheim 1995, Cabrol et al. 1996, 1997, 11998b, Kuzmin et al. 1998, Landheim et al. 1998), andthe sources of Mangala Valles. There is a plausible relationbecause of the contemporary chronology and a large comboundary, between the formation of the Elysium Basin anElysium Mons activity. The sources of Granicus Valles arcated on the anks of Elysium Mons, and the young bed depoof Dao Vallis in the southern hemisphere of Mars are locatthe anks of the Hadriaca Patera. As a nal example, therstrong chronological correlation between the last outow eration and the last activity on Olympus Mons. These are ofew examples of the numerous large-scale uvial features o


2/11

236 CABROL ET AL.

Amazonian. They strongly suggest a thermally and hydrologi-cally very dynamic Mars during this period, where stored watercould still circulate at depth and at the surface, though probablyprotected by ice covers at the surface because of the climaticconditions.

Theevidence ofmassivewaterbodies, such as the850,000km 3

Elysium Basin (Scott and Chapman 1995) 500 million years agoor less, and the existence of smaller Amazonian basins like theGusev crater contradict the Mars drying out picture in the recentgeological past, and experiencing only isolated water outburstsin the Chryse Planitia region. During the Amazonian, Mars wasable to generate a water body equivalent to the MediterraneanSea (2,000,000 km 2). Moreover, part of the water contained inthe aquifer was still able to circulate to generate lakes in deepcraters like Gale and Gusev. In addition, the depth of young u-vial valley system headwaters are often located a few hundredmeters below the surface, even in the intertropical band of Mars,suggesting that the water table was not necessarily far from thesurface everywhere in recent geologic times. If we list these re-

cent water bodies, we see a distribution over a large area. Forinstance: (1) the lake in Gusev was generated by the Amazonianactivity of Maadim Vallis (Cabrol et al. 1996, Grin and Cabrol1997, Cabrol et al. 1998a, 1998b, Kuzmin et al. 1998, Landheimet al. 1998), (2) Gale, as we propose in the next sections, wasgenerated by aquifer drainage, water inltration, geothermal ac-tivity, and surface drainage in recent geologic times, (3) possiblesources for the Amazonian Elysium Basin range from north-ern source areas (Tanaka and Scott 1986), volcanoground-iceinteraction (Mouginis-Mark 1985, Squyres et al. 1987), andlarge channels originating from the cratered uplands (Scott andChapman 1989, Cabrol etal. 1986,Grin andCabrol 1997, Cabrolet al. 1998a, 1998b).

The 170-km diameter Gale crater is located 5 S/ 222 .5W, inthe northwest part of the Aeolis subquadrangle of Mars, at theboundary of the cratered uplands and the lowlands. The crateris characterized by the presence of a massive streamlined ter-raced and channeled sedimentary structure in its center that maysuggest (a) the reshaping of the ancient sedimentary depositby ice-push from an ice-covered lake, the variation of the levelof the frozen water body being illustrated by the layering of the central sedimentary structure, and/or (b) a possible accu-mulation of lacustrine deposits. Gale crater is bordered by the

Elysium Basin to the north and by the Aeolis Mensae to theeast (Fig. 1). The interpretation of the crater sedimentary mate-rial waspreviously proposed in a broader context of regional andglobalgeological mapping. Forinstance, thecrateroor materialhas been interpreted as lava ows and aeolian deposits by Scottet al. (1978), Greeley andGuest (1987)proposed thepresence of volcanic, aeolian, or uvial sediment, and Scott and Chapman(1995) suggested aeolian, lava, pyroclastic, uvial, and mass-wasted deposits. The present study shows that Gale crater expe-rienced a late lacustrine formation during the Amazonian. Wepropose a bathymetric model based on topographic data deducedfrom paleolacustrine morphologies. Results suggest that the lake

formation is related to (a) the sapping of the surrounding aquin the Aeolis Mensae region and (b) the rise of the water of the Elysium Basin.

1. CRATER MORPHOLOGY AND SEDIMENTDISTRIBUTION

Gale experienced multiple erosional processes. The north

is partly destroyed (Fig. 2.1). Extended mass-wasting occuthe north and east inner ramparts at the contact with theolis Mensae fractured formation (Fig. 2.2). Mass-wasted uare characterized by portions of rims embayed in smoothposits that slide toward the bottom of the crater. Locally, sof the rim debris observed on the crater oor indicate a doward movement of lubricated material over tens of kilome(Fig. 2.2). The rim is partly destroyed in the north, and rempits and blocks are embayed in smooth material.

1.1. Evidence of Water Action

Derived from the topographic data (M 2M I-2121, 1991)central sedimentary structure in Gale crater is 2500 m high frthe crater bottom located at 4500 m according to the MOLA topographic prole No. 03 published by Smith(1998) to the central summit at 2000 m elevation (Figs. 4). We estimate that the central structure corresponds to a2.5 103 km3 volume of material. This structure is creslike in its northern portion. It is strongly asymmetrical, divithe deposit into two almost equal halves: (a) the northern seis the most prominent and diverse in erosional features, (bsouthern sector is almost featureless and covered by an extensmooth deposit observed on the crater oor. At the margithe two sectors, a prominent peak rises above the depositcause of its central position, it is likely that it corresponthe remains of the crater central peak. Streamlined terraceslayers are spectacular in the northern sector and are distinorganized: seven terraces make the junction between the coor and a more round-shaped formation on the summit. Ontop of the structure, the deposit displays a succession of 13izontal to subhorizontal layers (Fig. 5). The terraces surrothe northern sector almost entirely. They are not observed onsouthern part of the crescent-like deposit. They are continuelsewhere, regularly spaced, and locally eroded by perpen

ular drainage systems that cut the different levels. Becaustheir stratigraphic position, we conclude that these drainage stems are the youngest features of the deposit. At the margin wthe southern sector, the terraces are streamlined and paralleachother. Theyare alsointenselyperpendicularlyeroded, foring elongated remnants that may suggest later wind actionsulting in yardang-like formations. Other streamlined remnare observed on the oor of larger channel features on thposit. We estimate the thickness of the terraces and layering shadow-length measurements. The survey was made upixel count on Viking images 631A05, A07, and A08 (10250-m/pxl mean resolution). Both terraces and layers vary f


3/11

HYDROGEOLOGIC EVOLUTION OF GALE CRATER, MARS

FIG.1. Regional view of Gale crater in the Aeolis Mensae regionat the plateau/plain boundary. Thedrainage system of Aeolis Mensae adjacent to the northerrim of Gale ends in a smooth plain. In the debouchment area, the morphology characterized by channels, channel-like depressions, and embayed craters suggean erosional agent coming from both the east and north.

about 100 to 200 m thick and are parallel and regular. Southof the central peak, the crater oor shows a featureless smoothdeposit, locally displaying slight variations in albedo, higher-albedo materialbeing observed at thetransition with theterraced

deposit. The topographic survey shows that this smooth depositlies about 300 m above the oor of the northern sector of thecrater. The main hydrologic feature observed on its surface isthe 250-km long drainage system that successively cuts Galesrim and the oor of the southern deposit, and debouches in thesouthern sector at 4500 m elevation. The stratigraphic positionof this channel demonstrates an activity at least as recent as theoor deposit.

1.2. Stratigraphic Age of Gale and Relative Age of Sediment

The time of emplacement of Gale is a critical parameter inthe discussion of the formation of the lake, as we will show in

the next section. Considering its size and erosional stage, crater appears ancient, though its stratigraphic position cpared to the surrounding geologic units does not allow a posiconclusion. Scott and Chapman (1995) proposed that the c

is buried under Amazonian and Upper Hesperian units. Gley and Guest (1987) proposed that Gale was superimposea Noachian plain unit (Npl 2). Considering these observatour suggestion is that the most likely stratigraphic positioGale would be Late Noachian to Early Hesperian. By conthesedimentary structureobserved on itsoor appears relativyoung. We established a crater survey of the oor to estimprobable age of the last geological event that took place icrater. The surveyed surface area (A) is about 15,400 kmresults show a recent resurfacing for the entire area (centralimentary structure and oor). Only seven craters are obserthree of them are under the statistic limit (

2 km in diam


4/11

238 CABROL ET AL.

FIG. 2. Diverse morphology of the rim of Gale crater. (2.1) View of the north rim showing small sinuous channels (Ch), which suggest a gentle overspilof the rim. (2.2) View of the east rim characterized by long sliding faces (Sr), which have transported rim remnants over tens of kilometers down the slope. Tmorphology is likely explained by the undermining of the rim material by sapping processes. (2.3) View of the south rim corresponding to the less degraded portof Gale. Debris slopes (Ds) covers the bottom of Gale, with possibly some posterosion by water and/or wind of some of the material accumulation, whichlongitudinally shaped. The longest channel debouching in Gale is entering by the south rim and is labeled (Ch). (2.4) West rim of Gale with streamlined erosthat could indicate inux of water coming both from north and west of the crater.

and the larger is about 5 km in diameter. We obtain a normalizedpopulation of 259 112 .4 craters for 10 6 km2. The deduced rel-ative age thus ranges from Early to Middle Amazonian for thecrater population lesser or equal to 2 km in diameter and denedas N(2), and Early Amazonian for N(5). However, the survey forN(5) relies on a unique crater. For superimposed craters only,the normalized results give 194 112 for N(2), and 64 .9 64.9for N(5). The estimate of relative ages are the same as for therst population. Error margins (E) are calculated as

E = N A 10

6 .

These results imply that the last signicant geological evenGale was Amazonian and involved the entire crater oor.

2. ORIGIN AND EXTENT OF THE PUTATIVE LAK

Evidence of water activity is given in the central deposself by the presence of streamlined and parallel terracesby drainage systems. Moreover, the amplitude of the morplogic record cannot be understood without the presence massive water-body inside Gale crater during the Amazoperiod. The northern crescent-like deep hollow and the eroremnants of the northern crest suggest two processes of fortion of the crater lake. For both processes, the origin of w


5/11


FIG. 3. Topographic model of Gale crater based on the Viking 2M topographic Map I-2121 (MC-23NW, 1991), and in the scale-bars, revised altimetry fMGS MOLA data (Smith et al. 1998). The connection between the ooded region and Gale crater has been possible because of the presence of a deep gap entthe crater. One of the most spectacular features of Gale is the crescent-like 2500 m deep depression, which is probably close to the melt layer of the initial

is related to the putative existence of the Elysium Basin paleo-lake during the Amazonian, which is inferred by shorelines, ter-races on the lower member of the Medusae Fossae formation,and/or the existence of active channels in the Aeolis Mensaeformation, described as a potential source area for the ElysiumBasin by Scott and Chapman (1995). The hydrogeologic rela-tionships between the Elysium Basin system and Gale crater are

FIG. 4. Topographic proles of Gale and the surrounding region, extracted from the Viking 2M topographic Map I-2121 (MC-23NW, 1991), and revi

altimetry after MGS MOLA data (Smith et al. 1998).

(a) the Aeolis Mensae source area that borders the Gale crim perimeter over 300 km, (b) the crater bottom that is a2500 m below the Elysium shoreline level, and (c) the elevof the observed gap in the north rim, which is located at a

1000 m elevation.We consider two processes: (a) the rst process is the sapof the rampart of the proximal source area saturated zo


6/11

240 CABROL ET AL.

FIG. 5. Close-up view of the central sedimentary structure. The peak located in the center of the impact crater, labeled CP, is probably the remnant ocrater central peak. Usually eroded in ancient craters like Gale, this peak may have been preserved by an early sedimentary cover. The sediments could hadifferent origins: aeolian, pyroclastic, and uvio-lacustrine. They are spectacularly exposed both in terraces (T) and layers (L). Several channels (Ch) deposiaqueous sedimentary deposits (ASD) at the bottom of Gale. Their stratigraphic position show that they are one of the youngest features in the crater, and indicthe presence of water in Gale during the Amazonian.

Aeolis Mensae and (b) is the supply of water by surface over-spilling, which might be obtained in three cases: (i) the over-spilling of the northern rim (labeled G in Fig. 1) by the releaseof Aeolis Mensae channels that border the plateau/plain bound-ary, (ii) by a channel that drained the plateau south of Gale

(labeled C in Fig. 1), and (iii) by the rising of the Elysium Basinwater level.

The presence of a water-rich aquifer is suggested by uvialvalley systems in the plateau in the vicinity of Gale and by thepresence of smooth deposits on the oor of the Aeolis Mensaeformation thatwere interpreted as Amazonian river-bed materialand channel sediment by Scott and Chapman (1995). Three ob-servations suggest the possibility of a crater drainage by gravity:(a) the recent map of the Elysium Basin by Scott and Chapman(1995) suggests a source area for the Elysium Basin between 0and 1000 m elevation in the Aeolis Mensae region, which dis-playsextendedmorphological evidenceof drainage andooding

that supports the hypothesis of a past water-rich aquifer, (bcording to MOLA (Smith et al. 1998), the elevation of thebottom shows a depth of 5500 m relative to the surrouncratered uplands, and, (c) the undermining of the eastern rimsociated with rim sliding processes occurs at thecontact with

Aeolis Mensae formation. Thus, we assume that the 5500-m dference in elevation between a potentialwater table of thesouarea and the bottom of the crater might be sufcient to gendrainage by gravity. To describe this process, we assume thwater body contained in the crater could have been formegravity drainage of the Aeolis Mensae proximal aquifer elevation). To estimate the water volume, we approximatecrater-bowl volume from existing topographic data (Smith1998) by the volume of successive frustrum cones betweencrater bottom ( 4500 m elevation) up to a conservative levelow the northen gap ( 1000 m elevation). Taking into accthevolume of thecentral peak, this calculation gives 20,500k


7/11


FIG. 6. Bathymetric model of Gale crater constructed from the imageryand the topographic data. The solid line is the graphic representation of thefunction between the volume, the area, and the elevation of the lake. The dottedline shows the relationship between the lake water level and the volume. Thedecrease of thewater-table level(see initial andnal water table) shows thelimitof the sapping process. The overspilling episode may have lled the remnantvolume in the crater.

(Fig. 6). The volume of available underground water is denedby the areal extent of the drainage and by the thickness of thesaturated zone of the unconned aquifer. This thickness is thedistance between the water table (0 m) and the emergence level.The emergence level is located at the base of the crater innerwall at 4500 m elevation. The area where the putative waterlevel is localized in our model is characterized by relics of largemesas that could be the remnants of subsidence surface formedby the underground drainage (Grin and Cabrol 1997). We con-sider that theexpelledwatercausesthedecline of thewater table.The void created by the loss of water generates a compaction of

the aquifer skeleton that is transferred to the surface, where landsubsidence occurs. To estimate the possible volume of expelledwater, and the extent of the drained aquifer, we introduce theconcept of specic storage coefcient as the compressibility of the aquifer media. Thus, we dene the storage coefcient as thecombination of thevolume of waterproduced by thecompactionof theaquifer skeletonandby thedecreaseof thewater pore pres-sure generated by the change of the aquifer thickness (Grin andCabrol 1997). We adopt physical parameters derived from ter-restrial analog materials for the aquifer media and for the water.For the aquifer media the parameters are: the compressibility ,and the porosity n , and for the water: the compressibility , andthe density .

For an aquifer thickness unit b =1000 m, the storativity co-efcient is dimensionlessS = g( +n ) b ,

where g is the martian gravity acceleration. For g =3.7 m/s2 , =1000 kg/m 3 , =1 108 m2N1, =4.8 1010 m2N1 (Domenico and Schwartz 1990), n =0.30. For anaquifer thickness b =4500 m, the calculation gives

S =0.6.

The volume of expelled water for the aquifer thickness4500 m of an area A (km2) is

V = A S =58 103 km3 .

Then, we deduce that A is about 96 103 km2 . This valuresponds to an area depicted by the half-portion of a circ100-km radius centered in Gale (70-km radius). The dra

aquifer is then an area surrounding Gales rims over 15 km. Tradius includes the region of the Aeolis Mensae, which dispextended morphological evidence of drainage and oodssupport the model of a water-rich aquifer.

The survey of the northern rim shows a 4-km wide g

1000 elevation (see Fig. 3), these elevations being comparto the Elysium paleolake shoreline level (Scott and Chap1995). We now consider the contribution of each of the prevprocesses listed above. The Aeolis Mensae region is charaized by a 750-km long drainage network, which headwaterlocated on a Noachian unit on the plateau. The bed oor ochannel is approximately at the same elevation (

1000

thegap in Gales northern rim. This observation suggests thatwater may have percolated through the rampart and also otopped the lowest portions of the crest. In addition, the drainetworks may have generated, and sustained, an undergroaquifer north of Gale. South of Gale, the 250-km long valley network originates in the cratered uplands in a Noaccrater material plain (Nc) and crosses northward HesperianAmazonian plain material (AHc) (Scott and Chapman 19Considering the dimensions of this uvial system, it cahave accounted for the total volume of sediment depositthe crater.

The other contribution by overspilling is related to the rof the Elysium Basin water level. We assume that the rising lof Elysium paleolake might have overtopped the gap. Thisis extended by several 200-km long channel-like depresswhere water may have traveled from the lake into the crater.topographic survey gives additional evidence of water relfrom the north. The crescent shape of the sedimentary deshows the summit of the arch-like feature oriented north, sistent with a north-entering erosional agent. This shape isobserved in the contour lines at 3000 and 4000 m. Coning all the above observations, we propose two comprehenmodels of a lake inside Gale crater that could predict the toraphy and the morphology of the central deposit.

3. DISCUSSION

The considered processes are related to the existence oAeolis Mensae drainage networks and the Elysium Basin rlevel. The downstream part of the Aeolis Mensae region icated at the same 1000-m elevation as Gales rim gap (FThe morphology of this region resembles a ood plain, wcould have been partly drained by small channels througrim of Gale (Fig. 2.1). The elevation of the ood plain nor

Gale is at the same elevation as the Elysium Paleolake prop


8/11

242 CABROL ET AL.

shoreline (Scott and Chapman 1995). The source area in theAeolis Mensae region is both at the same elevation as ElysiumPaleolakeandconnects to the paleolake at 3 .1S/202 .1W. Fromthis location to Gale crater, the plateau/plain boundary is char-acterized by smooth deposits and eroded mesas often displayingchannels and/or channel-like depressions on their top suggest-ing important water erosion. Thus, we propose that instead of adistinct source area, Aeolis Mensae may have been the south-ern limit of an Amazonian transgression of the Elysium Basin.This explanation will be consistent with the terracing observedon the eastern and western borders of the lower member of theMedusae Fossae unit, and it is in agreement with the regular ter-racing of Gales central sedimentary deposit. The morphologyof the inner northern rampart suggests an undermining processresulting from the aquifer drainage. The consequence of suchdrainage would have been the generation of a former lake withinGale prior to the overspilling of the northern rim. The regionalmorphology supports the idea of a gentle and progressive trans-gression rather than a catastrophic outow episode. Because of

the wide area of Elysium Paleolake (2,000,000 km2

), the vari-ation of the water level may have been most likely very slow.This scenario supports observational evidence such as the lack of a large erosional inow system cut in the northern rim thatwould be expected if the crater was empty at the time of the over-spilling. Moreover, we can assume, according to the size of thefew small channels cutting this north rim (Fig. 2.1), that the levelof the former lake in Gale was already about 1500 m, or veryclose to the elevation of the overspilling water outside the crater.

The lifetime of the crater lake depends of the duration of the water supply in the Elysium Basin system, where evidenceof water are found for the Hesperian and Amazonian periods.During this period, Gales lake may have been sustained bytwo ways: in the rst one, the lake could have been sustainedis by assuming drainage by gravity. For instance, we take ter-restrial values of discharge adapted to Mars of 0.01 m 3 perday (3 m3 per year) per square meter of seepage face. For aseepage face averaging 100 m in height, developed over the210-km long inner perimeter of the crater walls, the time neces-sary to ll 20,000 km 3 of the lower volume of the crater lake is3.0 105 years. We suppose that the northern rim of the craterwas surrounded by an active unconned and unfrozen aquiferpermanently sustainedby theAeolisMensaenetwork water. The

drainage by gravity could have been favored by a hydraulic con-ductivity of the aquifer media and a permeability between 1000and 3000 Darcies (Carr 1979, Moore 1995) to allow the waterto travel approximately 20 km.

The second way to ll the remaining volume of Gale crater isby overspilling. We assume a conservative shallow water risingat of 0.02 m/yr, and a ow velocity of 0.5 m/s 1 in the chan-nel, the corresponding discharge rate in this channel would thenequal 40 m 3 /s1. The time required to ll the upper volume of a500-km 3 lake by overtopping is about 5 103 years. The timerequired to ll the lake entirely by sapping and overspilling isthen 8

105 years. We suggest, according to climate models,

that the Elysium Paleolake was covered by a thick ice-sheeaddition, the activity of Elysium Mons at the same period (Mdle to Upper Amazonian, Scott and Chapman 1995) may provided a cover of ash ows, which mixed with ice, and chave prevented rapid sublimation of the ice-sheet cover.

4. HYDROTHERMAL SYSTEMS IN GALE CRATE

The formation of Gale crater was probably associated the formation of hydrothermal systems heated by impact and by geothermal energy from the uplifted basement. Becof the large deep size of Gale, the basin was probably oby groundwater immediately after formation of the crater. ooding of the crater may have been accompanied by somthe erosion and slumping observed in the rim of the crater, dusapping or slope instability from the hydraulic pressure. Tevents represented a stage in the interaction of water with that occurred shortly after formation of the crater, but prithe resurfacing and erosion events reected in the current m

phology of the crater and sedimentary deposits. The evidencthese early events is not obvious because of the apparent duration of uvial and aqueous processes occurring in GHowever, constraints can be placed on the possible magniof the heat sources, the possible evolution of the hydrotheenvironments, and the possible locationsof hydrothermal sprdeposits.

The two major heat sources present after the formatioGale crater are impact melt and geothermal energy from thelifted basement. The amount of impact melt can be estimfrom the diameter of the crater, using 140 km as an estiof the original rim diameter prior to the extensive erosiothe rim deposits. From this diameter, and data compileClifford (1993), a melt volume of approximately 3,900 kmbe estimated. Assuming this melt was spread over a crater with a diameter of 130 km, a melt thickness of approxim670 m is obtained. While very uncertain, this thickness resents a reasonable order of magnitude estimate for considepossible hydrothermal effects. The cooling of large impact sheets has been considered by Onorato et al. (1978), usinnumerical and analytical techniques. The postulated Gale msheet is approximately three times thicker than the 200-m tmelt sheet studied by Onorato et al. (1978). The work of J

(1968) showed that the cooling times are proportional tosquare of the length scale, therefore, the solidication and cing time of the Gale melt sheet will be about 10 times that o200-m melt sheet. The inuence of a melt sheet and othersources on the cooling of a martian impact crater lake was ied by Newsom et al. (1996). Extrapolating from these resuliquid lake with an ice cover could be maintained under cuclimatic conditions for as long as 100,000 years. This assuthat ground water or surface runoff can replace the waterby ablation from the surface of the ice cover. Another imtant assumption is that the cooling would be mostly conductConductive cooling is likely considering the rapid developm


9/11


of a thick crystallized melt rind protecting the molten interiorof the melt sheet (Newsom 1990). In addition, convective cool-ing will be limited due to the burial of the impact melt undermaterial slumped from the walls of the crater, and the commonoccurrence of self-sealing of hydrothermal systems in terrestrialhot spring and mid-ocean ridge systems. Even if the cooling ispartly convective, the lifetime of a liquid lake could be quitelong.

An additional heat source for the lake could come from geo-thermal heat derived from the uplift of the basement due to theformation of the crater. This process has been connected withthe deposition of hydrothermal minerals in the Manson crater(McCarville et al. 1996). In the case of Gale crater, a centraluplift approximately 8 km in diameter is present, and the size of Gale suggests that a peak ring might also be present (Pike 1985).Peak rings often have diameters approximately half the craterdiameter, which would fall beneath the annular central depositon the oor of the crater. If the sediments are partially draped onremnants of such a peak ring, the estimatedvolume of sediments

may be an upper limit. The depth of excavation represented bythe materials of the central peak in Gale could be importantfor obtaining deep samples of the crust. Based on Croft (1985)the depth of excavation is about 0.1 times the transient cavitydiameter, or about 8.5 km for Gale. This depth is still in the crust,because geophysical models suggest that the crust is probably100 to 250 km in thickness (Sohl and Spohn 1997).

As discussed by Newsom et al. (1996), the magnitude of theheating from the uplifted basement can be substantial, but therelative importance compared to the impact melt sheet dependson the time scale on which the heat is released. Because theuplifted basement will be shattered, but not molten, the perme-ability will be much higher than the crystallizing melt sheet,thus convective cooling is more likely. Therefore, the channelsobserved on the young surface of the central deposit are unlikelyto be connected with the initial hydrothermal event. Even if thehydrothermal system involving the central uplift was relativelyshort lived, there is a high probability that hydrothermally al-tered materials and spring deposits occur in the central uplift. If the exposed portions of the central uplift were above the levelof groundwater supplied hydrothermal systems, there is still apossibility of deposits from an acid-sulfate hydrothermal sys-tem supplied by vapor transport from a deeper neutral chloride

system.Even if the supply of water were not sufcient to form ormaintain a lake, a substantial hydrothermal system would arisein the upper part of the melt sheet and the overlying debris. If thesupply of water declined even further, the hydrothermal systemcould have evolved into a vapor-dominated system (Newsomet al. 1998). This consists of a neutral-chloride hydrothermalsystem at the deep groundwater table, heated by the impact meltleading to vapor transportupward, where thevaporcan condenseto form an acid-sulfate hydrothermal system.

Another area where hydrothermal systems and the formationof associated channels may have occurred is on the rim of the

crater (e.g., Brakenridge et al. 1985). Impact melt depon the rim could have supplied heat to ground water or mground ice to form hydrothermal alteration and hot springssubsequent heavy erosion of the rimandpenetration by chanmay have transported fragments of these hydrothermal depand distal impact melt deposits into the sediments withincrater.

CONCLUSION

A lake within Gale crater during the Amazonian providesther evidence that water was present at the surface of Mathe relatively recent past. The evidence of massive water bodsuch as the 850,000-km 3 Elysium basin (Scott and Cha1995) 500 million years ago or less, and the existence of smaAmazonian basins like Gusev and Gale craters are in codiction with the old picture of Mars drying out in the rgeological past and experiencing only isolated water outbin the Chryse Planitia region. During the Amazonian, Mars

able to generate a water body equivalent to the Mediterrasea (2,000,000 km 2). Moreover, part of the water containthe aquifer was still able to circulate to generate lakes in craters, like Gale and Gusev. In addition, the depth of younvial valley system headwaters are often located a few hunmeters below the surface, even in the intertropical band of Msuggesting that the water table was not necessarily far fromsurface everywhere in recent geologic times. If we list thescent water bodies, we see a distribution over a large areainstance: (1) the lake in Gusev was generated by the Amazoactivity of Maadim Vallis, (2) Gale, as proposed in this stwasgenerated by aquifer drainage,water inltration, andsurdrainage in recent geological times, (3) the possible sourcethe Amazonian Elysium Basin range from northern sourceas (Tanaka and Scott 1986), volcanoground-ice interac(Mouginis-Mark 1985, Squyres et al. 1987), and largenels originating from the cratered upland (Scott and Chap1989, Cabrol et al. 1996, Grin and Cabrol 1997), and (4ley systems and outows generated by hydrothermalactivityalso described during this period (Gulick and Baker 1989)Amazonian uvial features are observed all across the plThey are not as numerous as during the Noachian/Hespeboundary, but their size is often much larger. In addition

most all of them have a common characteristic: their soof water can be related to ground-iceheat ux interactioninstance, we will refer to a nonexhaustive list of typicaamples: the sources of Maadim Vallis are related to volcand faults regions (Landheim 1995, Cabrol et al. 19961998a, 1998b, Kuzmin et al. 1998, Landheim et al. 199so are the sources of Mangala Valles. There is a plausibllationship, because of a contemporary chronology and a lcommon boundary, between the formation of the Elysium Band the Elysium Mons activity. The sources of Granicus Vare located on the anks of Elysium Mons and the youngdeposits of Dao Vallis in the southern hemisphere are loc


10/11

244 CABROL ET AL.

on the anks of the Hadriaca Patera. As a nal example, thereis a strong chronological correlation between the last outowgeneration and the last activity on Olympus Mons. These areonly some examples of the numerous large-scale uvial fea-tures of the Amazonian. They strongly suggest a thermally andhydrologically very dynamic Mars during this period, wherestored water could still circulate at depth and at the surface,though probably protected by an ice-covered sheet at the sur-face because of the climatic conditions.

Locally, the morphology and organization of the Gale sedi-mentary deposit provide information about the crater evolutionthrough time and gives an open window on the surrounding hy-drogeologic conditions. The regularity of the spacing betweenlayers in the upper part of the central deposit (Fig. 5), and theircomparable thickness, may suggest a major cycling event and/ora regular decrease of the water level. The major cycling eventcould reect the progression and recession of the Elysium Basinthrough time,with transgression events thatepisodically oodedregions at the plateau/plain boundary. The regular decrease of

thewater levelmayalso be explained by onemajor transgressionand progressive evaporation, sublimation, and inltration of thewater within Gale.

Gale is an ancient impact crater, probably dating back fromthe Noachian-Hesperian boundary. Its last aqueous sedimentaryepisode, as shown in this study, is estimated to be Early to Mid-dle Amazonian. Considering the uvial activity through time inthe Aeolis region, and in the Aeolis Mensae particularly, it islikely that the crater may have been lled with uvio-lacustrinesediment for more than two billion years. The potential seriesof lakes during this extended period of time may have providedvery diversied environments, from warm thermally driven wa-ters to cold-ice-covered waters. In this respect, impact craterlakes are probably the most interesting places to search for lifeon Mars. It is unknown if any potential life on Mars existedon a continuous time scale or was terminated and restarted atmultiple events, as suggested by Maher and Stevenson (1988)and Sleep et al. (1989). Despite this uncertainty, we expect thatevolutionary trends would have allowed organisms to establishnew ecological niches in response to a changing climate and en-vironment. In the case of impact crater lakes, such as Gale andGusev craters, we envision a series of ecological niches havingbeen present from the time following impact until Mars lost its

hydrosphere and the potential to sustain life on the surface. Forsome time following impact, Gale cratermayhavehadtheabilityto sustain life in a thermally driven subaerial or subsea environ-ment. Such terrestrial analog environments have been discussedby Brocks (1967)andBarns et al. (1994). Thepreservation of bi-ological information in thermal spring deposits on Mars, as wellas related sampling strategies were discussed by Walter and DesMarais (1993). Based on the observational evidence of uvialerosion on Gale craters rims, we suggest that potential micro-bial niches may have been developed underwater. The lake bodywould have experienced cooling through time. Various microor-ganisms in the lake may have adapted to physical and chemical

changes induced by the modication of the climate (i.e., temature, pH, salinity). Such changes would in turn have had dimplications on species abundance, diversity, and dominanc

As Gale dried up, resulting playas may have favored theistence of halophiles. Any microbial life that existed in ttypes settings are likely to have been preserved in mineraposits, such as carbonates (Schopf 1983, McKay and Ne1988), and evaporites (Rothschild 1990), and in the scenaria volcanic setting, possibly in silica. Each of these fossilizaagents have different preservation potentials in the fossil recoIn terms of identication, different deposits may be identithrough characteristic spectral signatures. As the temperabecame progressively cooler, microbes may have retreated fthe lake water to subareal habitats and in turn migrated to tected niches in or under rocks on the surface. These typenvironments have been identied in some of theEarths desesuch as the Dry Valleys of Antarctica (Friedmann 1982, 1Friedmann and Koriem 1989). There are also potential mtian analogs represented by the perennially ice-covered lake

Antarctica. These lakes are hosts of microbial mats compprimarly of bacteria, cyanobacteria, and algae that exist witsunlight most of the year (Wharton 1994).

The different environments listed above could all have tained life that may be still present in the fossil record.absence of crustal recycling on Mars opens up for the possity that fossilized life forms from the various ecological nidescribed could be present right at the surface today, whentential extent life should be searched for underground. The meral deposits in which life may be preserved would have a likelihood of being detected with spectral data, such as the Thmal Emission Spectrometer (TES) of the Mars Global Surv(Christensen et al. 1992). Impact crater lakes must be survthrough high resolution imagery and spectral data that can asin further selection of scientically and biologically proming sites for future surface exploration. The possible ecologniches of Gale crater introduced in this section will be furdiscussed in a paper currently in progress.

APPENDIX

The survey of Gale has been undertaken using the following data set:

Viking Orbiter images: 631A03, 631A05-10, 631A25-28, 631631A43, and 631A45.Topographic Series Maps:

1 : 2M Aeolis Northwest subquadrangle, I-1213 (MC-23NW), 191 : 2M topographic seriesAeolis Northwest,I-2121(MC-23NW),1

ACKNOWLEDGMENT

This study is nancially supported by NASAs Cooperative AgreNo. NCC2-1064 for the rst author. We thank both reviewers, Stephen Cliand Kenneth Tanaka, for theirhelpful and constructive remarks and suggestiSpecial thanks also to Mary Chapman forour fruitful discussions about ElysBasin and Gale crater.


11/11


REFERENCES

Barns, S. M., R. E. Fundyga, M. W. Jefferies, and N. R. Pace 1994. Remarkablearcheal diversity detected in Yellowstone National Park hot spring environ-ment. Proc. Natl. Acad. Sci. U. S. A. 91 , 16091613.

Brakenridge,G. R., H. E. Newsom, andV. R. Baker 1985. Ancient hotsprings onMars: Origins and paleoenvironmental signicance of small martian valleys.Geology 13 , 859862.

Brock, T. D. 1967. Life at high temperatures. Science 158 , 10121019.Cabrol, N. A., E. A. Grin, and G. Dawidowicz 1996. Maadim Vallis revisited

through newtopographic data:Evidence for an ancient intravalley lake. Icarus123 , 269283.

Cabrol, N. A., E. A. Grin, and G. Dawidowicz 1997. A model of outow gen-eration by hydrothermal underpressure drainage in volcano-tectonic environ-ment, Shalbatana Vallis (Mars). Icarus 125 , 455464.

Cabrol, N. A., E. A. Grin, and R. Landheim 1998a. Maadim Vallis evolution:Geometry and models of discharge rate. Icarus 132 , 362367.

Cabrol, N. A., E. A. Grin, R. Landheim, R. O. Kuzmin, and R. Greeley 1998b.Duration of the Maadim Vallis-Gusev Crater hydrogeologic system. Icarus133 , 98108.

Carr, M. H. 1979. Formation of martian ood features by release of water fromconned aquifers. J. Geophys. Res. 84 , 29953007.

Christensen, P. R., D. L. Anderson, S. C. Chase, R. N. Clark, H. H. Kieffer,M. C. Malin, J. C. Pearl, J. Carpenter, N. Bandiera, F. G. Brown, andS. Silverman 1992. Thermal emission spectrometer experiment: Mars Ob-server mission. J. Geophys. Res. 97 , 77197734.

Clifford, S. M. 1992. A model for the hydrologic and climatic behavior of wateron Mars. J. Geophys. Res. 98 , 10,97311,016.

Croft, S. K. 1985. The scaling of complex craters. Proc. Lunar Planet. Sci. Conf.15th. J. Geophys. Res. 90 , C828C842.

Domenico P. A., and F. W. Schwartz 1990. Physical and Chemical Hydrogeol-ogy. Wiley, New York.

Friedmann, E. I. 1982. Endolithic microorganisms in the Antarctic cold desert.Science 215 , 10451053.

Friedmann, E. I. 1986. The Antarctic cold desert and the search for traces of lifeon Mars. Adv. Space Res. 6 (12), 265268.Friedmann, E. I., and A. M. Koriem 1989. Life on Mars: How it disappeared (if

it was ever there). Adv. Space. Res. 9 (6), 167172.Greeley, R., and J. E. Guest 1987. Geologic map of the eastern equatorial re-

gion of Mars: U. S. Geol. Survey Misc. Inv. Series Map I-1802-B, scale1 : 15,000,000.

Grin, E. A., and N. A. Cabrol 1997. Limnologic analysis of Gusev crater pa-leolake, Mars. Icarus 130 , 461474.

Gulick, V. C., and V. R. Baker 1989. Fluvial valleys and martian paleoclimates. Nature 341 , 514516.

Jaeger, J. C. 1968. Cooling and solidication of igneous rocks in basalts. InThe Poldervaart Treatise on Rocks of Basaltic Composition (H. H. Hess and

A. Poldervaart, Eds.), pp. 503533. Wiley, New York.Kuzmin, R. O., R. Greeley, R. Landheim, N. A. Cabrol, and J. D. Farmer 1998.

Geological map of the MTM-15182 and MTM-15187 quadrangles, Gusevcrater-Maadim Vallis region, U. S. Geol. Series Map, in press.

Landheim,R. 1995. Exobiologysites for Marsexplorationand geologicmappingof the Gusev crater-Maadim Vallis , Masters Thesis, Arizona State Univer-sity.

Landheim, R., R. Greeley, N. A. Cabrol, E. A. Grin, R. O. Kuzmin, andC. P. McKay 1998. Evolution of the Maadim Vallis-Gusev Crater hydro-geologic system, Mars. Icarus , submitted.

Maher, K. A., and D. J. Stevenson 1988. Impact frustration of the origin of life. Nature 331 , 612614.

McCarville, P. J., and L. J. Crossey 1996. Post-impact hydrothermal alteof the Manson impact structure, Manson, Iowa. In The Manson Impactture: Anatomy of an Impact Crater (C. Koeberl and R. AndersonSpecial Papers of the Geological Society of America, Vol. 302, pp376.

McKay, C. P., and S. S. Nedell 1988. Are there carbonate deposits in the Marineris, Mars? Icarus 73 , 142148.

McKay, C. P., and C. R. Stoker 1989. The early environment and its evoon Mars: Implications for life. Rev. Geophys. Space Phys. 27 , 189.

Moore, J. M., G. D. Clow, W. L. Davis, V. C. Gulick, D. R. Janke, C. P. MC. R. Stoker, and A. P. Zent 1995. The circum-Chryse region as a poexample of a hydrologic cyle on Mars: Geologic observations and theoevaluation. J. Geophys. Res. 100 , A3, 54355447.

Mouginis-Mark, P. J. 1985.Volcano/ground ice interactions in Elysium PlaMars. Icarus 64 , 265284.

Newsom, H. E. 1980. Hydrothermal alteration of impact melt sheets wiplications for Mars. Icarus 44 , 207216.

Newsom, H. E., G. E. Britelle, C. A. Hibbitts, L. J. Crossey, and A. M1996. Impact cratering and the formation of crater lakes on Mars. J. G Res. 101 , 14,95114,955.

Newsom H. E., J. J. Hagerty, and F. Goff 1998. Acid-sulfate hydrothermal and the origin of the martian soil. Lunar Planet. Science XIX , in pre

Onorato, P. I. K., D. R. Uhlmann, and C. H. Simonds 1978. The thermal hof the Manicouagan impact melt sheet, Quebec. J. Geophys. Res. 832797.

Pike, R. J. 1985. Some morphologic systematics of complex impact stru Meteoritics 20 , 4968.

Rothschild, L. J. 1990. Earth analogs for martian life. Microbes in evaponew model system for life on Mars. Icarus 88 , 246260.

Schopf, J. W. (Ed.) 1983. The Earths Earliest Biosphere: Its Origin and Elution. Princeton University Press, Princeton, NJ.

Scott, D. H., and M. G. Chapman 1989. Geologic setting of an uMartian channel: Hypotheses on origin. Lunar Planet. Science XIX382.

Scott, D. H., and M. G. Chapman 1995. Geologic and topographic mapsElysium paleolake basin, Mars. U. S. Geol. Survey Geol. Series Map Iscale 1: 5,000,000.

Scott, D. H., E. C. Morris, and M. N. West 1978. Geologic map of the quadrangle of Mars. U.S. Geol. Survey Misc. Inv. Series Map I-1111,1: 5,000,000.

Sleep, N. H., K. Zahnle, J. F. Kasting, and H. J. Morowitz 1989. Annihof ecosystems by large asteroid impacts on the early Earth. Nature 34142.

Smith, D. E. M.,T. Zuber, H. V. Frey, J. B. Garvin, J. W. Head, D. O. MuhleG. H. Pettengill, R. J. Phillips, S. C. Solomon, H. J. Zwally, W. B. Band T. C. Duxbury 1998. Topography of the northern hemisphere offrom the Mars Orbiter laser altimeter. Science 279 , 16861692.

Sohl, F., and T. Spohn 1997. The interior structure of Mars: ImplicationsSNC meteorites. J. Geophys. Res. 102 , 16131635.

Squyres, S. W., D. E. Wilhelms, andA. C. Moosman 1987. Large-scale volcground ice interactions on Mars. Icarus 70 , 385408.

Tanaka. K. L., and D. H. Scott 1986. The youngest channel system on Lunar. Planet. Sci. Conf. 17 , 865.

Walter, M. R., andD. J. Des Marais 1993. Preservation of biological informin thermal spring deposits: Developing a strategy for the search for foson Mars. Icarus 101 , 129.

Wharton, R. A. 1994. Stromatolitic mats in Antarctic lakes. In PhanStromatolites II (J. Bertrand-Sarfati and C. Monty, Eds.), pp. 5370. KDordrecht.

hydrologic evolotion of gale crater

Documents