Journal of Paleolimnology 13: 267-283, 1995. @ 1995KluwerAcademicPublishers. Printedin Belgium.

267

Paleolakes on M a r s R o b e r t A. W h a r t o n , Jr., J a m i s o n M. C r o s b y 1, C h r i s t o p h e r R M c K a y 2 & J a m e s W. Rice, Jr. 3 ~Desert Research Institute, P.O. Box 60220, Reno, Nevada 89506, USA 2NASA-Ames Research Center, Moffett Field, California 94035, USA 3Department of Geography, Arizona State University, Tempe, Arizona 85287, USA Received20 June 1994;accepted6 February1995 Key words: Mars, martian paleolakes, lacustrine sediments, ice-covered lakes

Abstract

Observational evidence such as outflow channels and valley networks suggest that in the past there was flowing water on Mars. The images of fluvial features on Mars logically suggest that there must exist downstream locations in which the water pooled and the sediment load deposited (i.e. lakes). Sediments and morphological features associated with the martian paleolakes are believed to occur in Valles Marineris, and several large basins including Amazonis, Chryse and Elysium planitia. As Mars became progressively colder over geological time, any lakes on its surface would have become seasonally, and eventually perennially ice-covered. We know from polar lakes on Earth that ice-covered lakes can persist even when the mean annual temperature falls below freezing. Thus, the most recent lacustrine sediments on Mars were probably deposited in ice-covered lakes. While life outside of the Earth's atmosphere has yet to be observed, there is a general consensus among exobiologists that the search for extraterrestrial life should be based upon liquid water. The inference that there was liquid water on Mars during an earlier epoch is the primary motivation for considering the possibility of life during this time. It would be of enormous interest from both an exobiological and paleolimnological perspective to discover lakes or the evidence of former lakes on another planet such as Mars. Limnology would then become an interplanetary science. Introduction

Within our solar system it appears that the Earth is the only planetary body capable of maintaining liquid water today. But has this always been the case? Is there any possibility that in the past liquid water existed on other planets? Based upon observational evidence it does appear that liquid water was abundant in the past on Mars. Indeed, there may have been enough water on Mars to maintain large lakes and possibly even oceans. Current understanding of the geological and hydrological history of Mars is based primarily on the images returned from the Mariner and Viking spacecraft missions. Perhaps the most interesting class of features found in these images of Mars are dendritic drainage patterns and other evidence suggestive of flowing liquid water. These features include: outflow channels, valley networks, tear drop-shaped islands, deltas, ter-

races, and dissected wrinkled ridges (Baker & Milton, 1974; Baker, 1979; Baker & Kochel, 1979; Carr, 1979, 1983, 1986; Rice et aL, 1988; Squyres, 1989; Rice 1992, 1993; Rice & Scott, 1993; Rice & De Hon, in press). Recent research has considered if this water could have formed lakes and even possibly oceans (Nedell et al., 1987a; Rice et aL, 1988; Goldspiel & Squyres, 1989; Parker, 1989; McKay & Davis, 1991; Scott et al., 1991a, b; De Hon & Pani, 1992; Parker et aL, 1993; Rice & De Hon, in press;) and whether these lakes could have served as an abode for life (Rice et aL, 1988; McKay & Stoker, 1989; 1992; Wharton et aL, 1989a; Baker et aL, I991; Goldspiel & Squyres, 1991; Scott etaL, 1991, 1992; De Hon, 1992; Rice, 1992). If they do indeed exist, martian paleolake sites are extremely important because they may hold a record of the physical and chemical conditions that prevailed in the past on Mars as well as evidence of former biological activity.

268 In this paper, we discuss the evidence for paleolake sites on Mars. We will first consider relevant aspects of the history of water and climate on Mars. We will then discuss the specific evidence indicating that there were lakes on Mars and consider the possibility of ice-covered lakes on Mars as well as the potential for deposition of sediment and carbonate in these lakes. We conclude with a brief discussion of terrestrial analogs of the martian paleolakes and future possibilities for paleolimnological studies of Mars.

Liquid water on Mars

There is considerable evidence that, at various times throughout its history, liquid water was present on the surface of Mars. The Viking and Mariner missions provided visual evidence of fluvial features that appear to have been carved by low viscosity fluids. These features are broadly classified into two categories: the valley networks and outflow channels (Baker, 1979; Carr, 1981). The valley networks are quasi-dendritic systems believed to have been produced by fluvial erosion in the ancient cratered terrain of the Southern Highlands (Fig. 1). The morphology of the valley networks suggests that they represented relatively stable, long-term flows. The outflow channels by comparison appear to have been generated by the rapid, perhaps catastrophic, release of large quantities of liquid which because of the rapid rate of discharge need not have been in equilibrium with the ambient temperature and pressure conditions (Fig. 2). The liquid responsible for both the outflow and valley networks is generally accepted to be liquid water. Other possible mechanisms including formation by faulting (Schumm, 1974), erosion by lava flows (Can', 1974a; Schonfeld, 1977), wind (Cutts & Blasius, 1981), liquid hydrocarbons (Yung & Pinto, 1978), ice (Lucchitta, 1980; Lucchitta et al., 1981b), and liquid CO2 (Sagan et al., t973) have been suggested, but are not considered as likely as water to explain the morphology and widespread distribution of the features. The most compelling piece of evidence for liquid water in Mars' past is the observation of surface features similar in appearance to terrestrial river valleys. Several types of these features have been described (Baker & Milton, 1974; Sharp & Malin, 1975; Baker, 1978; Carr, 1979, 1986; Pieri, 1980; Carr & Clow, 1981; Rice, 1993). Originally know as runoff channels, these features have since been renamed valley

networks to reflect the recognition that they are valleys and not channels (Pieri, 1976, 1980; Carr & Clow, 1981). The valley networks typically start small, have tributaries, and increase in size downstream (Fig. 1). It has been proposed that the valley networks formed primarily from groundwater sapping (Sharp & Malin, 1975; Pieri, 1980; Squyres & Kasting, 1994). However, some of the networks (Fig. 1) have morphologies that clearly represent the drainage of a distributed source of water - either rain or snow - thus implying a hydrological cycle on early Mars. An additional explanation for the valley networks involves the action of hydrothermal sources. In such a scenario, a subsurface heat source would melt ground ice and force liquid water to the surface (Gulick & Baker, 1989). Fluvial features could be formed this way even if the surface temperature on Mars had been quite cold, however the surface pressure would still have to have been well above the triple point. The valley networks are most often associated with heavily cratered regions of Mars indicating that many of them are contemporaneous with the end of the late bombardment some 3.8 Gyr ago (Pieri, 1976, 1980; Carr & Clow, 1981). However, there are network channels found in locations on Mars that are considerably younger (Scott et al., t991). In particular, Gulick & Baker (1989) have reported channels on the slopes of Alba Patera with an expected age of much less than 2 Gyr. Masursky et al. (1977) presented evidence of interlayering of channels and lava flows at Kasei Valles suggesting that the liquid water flows have been episodic. The valley networks constitute reasonably substantive proof of a different past climate on Mars because their smaller width (cf. outflow channels) suggests that they were formed by a slow meandering flow of liquid water. And although temperatures above freezing may not have been required, temperatures and pressures would have had to have been much greater than the present values to allow flowing water (Wallace & Sagan, 1979; Carr, 1983; Breckenridge et aI., 1985; Rice 1993). The presence of valley networks has been used as an argument for a thicker atmosphere in the past on Mars (Masursky, 1973; Pieri, 1976, 1979; Pollock 1979; Carr & Clow, 1981; S c o t t e t a l . , 1992). In contrast to the valley networks, the outflow channels start abruptly, have few if any, tributaries, and are generally of enormous size, some being over 1500 km in length (Fig. 2). Outflow channels are believed to have formed by large catastrophic floods possibly associated with the rapid drainage of ice-dammed under-

269

Fig. 1. Finelydissected valley network (indicated by arrow) in highly fractured old terrain at 48 ° S, 98 °W. This is one of the densest networks observed on Mars. The picture measures 250 km across (USGS image).

ground reservoirs (McCauley et al., 1972; Masursky, 1973; Milton, 1973; Baker & Milton, 1974; Masursky et al., 1977; Baker, 1979; Baker & Kochel, 1979; Carr, 1979; Rice & De Hon, in press). The existence of outflow channels is not sufficient to prove the former existence of a wetter climate on Mars because the rush of water associated with them is so intense that they could form even under current climatic conditions (Wallace & Sagan, 1979; Carr, 1983). The outflow channels seem to have occurred throughout an extend-

ed period of Martian history perhaps extending until the present time - albeit at a reduced rate (Masursky et al., 1977). If the fluvial features described above are indeed evidence of liquid water, then we can conclude that earlier in Martian history there was a period during which liquid water was prevalent and stable. The duration of this epoch is uncertain but it is clear for the absence of significant erosion on the ancient cratered terrain that an extensive hydrological cycle did not last

270

Fig. 2. A 20-km-wideoutflowchannel at 1 °S, 42 °W. The channel emergesfroma 40-km-diameterdepression enclosing chaotic terrain and

continues eastwardoff the picture to connect with Simud Vallis (USGS image).

for long (< 1 Gyr) after the termination of the late bombardment (Carr, 1981). Liquid water flows continued throughout martian history, most probably in the form of outflow channels discharging subsurface aquifers (possibly relics from an early hydrological cycle). In addition, there is evidence for the occurrence of episodic valley networks (Scott et al., 1992; Rice, 1993). The inference that there was liquid water on Mars during an earlier epoch is the primary motivation for considering the possibility of life during this time (McKay & Stoker, 1986; McKay et al., 1990a; McKay & Davis, 1991). Liquid water is the most critical environmental requirement for life on Earth. The general similarity between early Earth and Mars allows the hypothesis that life on Mars would be similar in this basic environmental requirement (Wharton et al., 1989a).

Climate on early Mars

Fluvial features indicative of the slow, stable flow of liquid water are not consistent with the present martian climate. This is primarily due to the low atmospheric pressure (6-10 hPa), which is close to the triple point pressure of water (6.1 hPa). A liquid state cannot exist at pressures below the triple point of a substance (this is the reason that CO2 at Earth sea level pressure is a 'dry' ice). For pressures slightly above the triple point, a liquid is only marginally stable since the boiling point is effectively close to the freezing point. At martian surface pressures of 6.1 to 10 hPa water will boil at 0 and 7 °C, respectively. Kahn (1985) has shown that if liquid water were present on the surface of Mars it would lose heat rapidly by evaporation, even if it was

271 at 0 °C, since it would be close to its boiling point. This loss of heat would cause the water to freeze since it is also close to the freezing point. The amount of energy input required to maintain the liquid state is larger than the solar constant on Mars for pressures less than about 10 hPa, and becomes extremely large as the pressure approaches 6.1 hPa. The presence of stable liquid water on early Mars requires that the climate must have been different in the past. In particular, the pressure must have been well above the triple point pressure. Several lines of evidence, including geomorphological observations and theoretical arguments, suggest that a denser atmosphere was in fact plausible during Mars' earlier history (Walker, 1978; Pollack, 1979; Cess et aI., 1980; Pollack & Yung, 1980; Toon etal., 1980; Kahn, 1985). Warmer temperatures may have been associated with this thicker atmosphere. It is important to note, however, that liquid water could exist under a thick atmosphere even at low temperatures. Understanding how Mars might have been warm early in its history is difficult since models of stellar evolution propose that the sun was 30% dimmer 3.5 Gyr ago and that in the absence of any greenhouse effect the surface temperature of Mars was - 7 5 °C (Moroz & Mukhin, 1977). It has been suggested that warmer temperatures were engendered by a greatly enhanced greenhouse due to a thick CO2 atmosphere (Moroz & Mukhin, 1977; Owen et aI., 1979; Pollack etal., 1987). However, recent modeling results suggest that the formation of CO2 condensation clouds would have severely limited the efficacy of the CO2 greenhouse effect on early Mars (Kasting, 1993). Squyres & Kasting (1994) suggest that no atmosphere composed of only CO2 and H20 appears capable of producing mean planetary temperatures close to 0 °C. Furthermore, using a 2-D climate model of CO2 on Mars, Haberle et af. (1993) found that the temperatures during the early period never approach freezing; the maximum mean annual temperature in their model in the first Gyr of martian history is - 3 0 °C. However, Squyres & Kasting (1994) argue that modest warming could be provided by even a low pressure CO2 atmosphere if it was supplemented with small amounts of CH4, NH3, or SO2. Liquid water flows and lakes on Mars may have formed while the planet had a thick CO2 atmosphere (over several hundred hPa) or possibly even when the planet had a relatively thin CO2 atmosphere (supplemented with other greenhouse gases) but still have remained quite cold.

The duration of a thick CO2 atmosphere is estimated to have been from 108 to 109 years (Pollack et at., 1987; McKay & Davis, 1991). In the presence of liquid water, chemical weathering of silicate rocks would have converted atmospheric CO2 into carbonate rocks. Early in Mars' history, volcanic heating (Pollack et al., 1987) and shock processes caused by impacts (Carr, 1989) would have provided a CO2 recycling mechanism decomposing the carbonate rocks and re-releasing the CO2 into the atmosphere. On Mars, impact events were drastically reduced at about 3.8 Gyr ago (Can', 1989). Over time, volcanism and other tectonic activity would have declined as well. Mars is a one-plate planet and without plate tectonics (Solomon et aI., 1979). Lacking a long-term recycling mechanism the initial thick CO2 atmosphere could not be maintained. The roles of temperature and pressure in controlling the availability of liquid water is illustrated by beginning with a Earth-like conditions on Mars approximately 4 Gyr ago and considering the transition to the present conditions. Following this approach, McKay et al. (1992) divided the history of water on the martian surface into four epochs based on key values of atmospheric temperature and pressure (Table 1). In Epoch I, during which the primordial CO2 atmosphere was maintained by volcanic recycling and impact degassing, mean annual temperature is believed to have been above freezing, with pressures greater than one atmosphere, and liquid water widespread on the surface. From an biological perspective, it is during this period when life could have arisen. Epoch II was characterized by mean annual temperatures below freezing, but peak temperatures which exceed freezing. Any lakes that may have been present during this time period would probably have been ice-covered lakes. In Epoch III, both mean annual and peak temperatures fell below freezing and any liquid water at the surface would have been short-lived. Finally, in Epoch IV, as pressures dropped to near the triple point pressure of water, liquid water could have no longer existed on the surface of the planet. The scenario outlined above assumes that the only source of surface warming is sunlight. As previously discussed, one hypothesis for the formation of the valley networks on Mars invokes subsurface heat flows (with a higher pressure but still cold atmosphere). Such systems could have persisted through Epochs I, II and III.

272 Table 1. Epochs of water on the surface of Mars t

Epoch number

Possible duration gyr ago

I

Thermodynamic conditions

State of liquid water

Biological analog

4.2-3.8

P_>500kPa (5 atm) T>0 °C

Liquidwater abundant

Origin of life possible

II

3.8-3.1

T<0 °C Tv~ak>0 °C

Ice-covered lakes

Microbialmats in lakes

III

3.1-1.5

Tpeak< 0 oC P>>6.1 hPa (1 hPa= 1 rob)

Liquid water in porous rocks

Life inside rocks

IV

1.5-Present P-.~6.1hPa

Pressure at triple point, no liquid water

No life on the martian surface

I Adapted from McKay et aL (1992).

Table 2. Calculated areas, depths, and volumes of sev-

ern martian paleolakest

Paleolake Elysium Amazonis Chryse Min Prob.

Surface area (km2)

Depth Volume (kin) (km3)

18.0x 105 12.8x105

1.0 1.0

9,1 × 105 5.9×105

13,0x 105 22.7x 105

1.0 1.0

9.2x 105 13.4× 105

1 From Scott, D. H., J. M. Dohm, & J, W. Rice (unpublished)

Evidence for the existence of paleolakes The images of fluvial features on Mars logically suggest that there must exist downstream locations in which the water pooled and the sediment load deposited (i.e. lakes). Presumably, these paleolake sites contain the material removed as the observed channels were carved as well as any precipitates of atmospheric gas such as carbonate (from COz) and nitrate (from lightning produced NO). Several important pieces of evidence, such as layered deposits, ancient shorelines, and spillways, are available to identify Martian paleolake sites. Layered deposits in the Valles Marineris were first observed in imagery obtained from the Mariner 9 mission and

have been subsequently studied using data from that and the Viking mission by a number of investigators (Malin, 1976; McCauley, 1978; Lucchitta, 1981a, 1982; Peterson, 1981; Lucchitta & Ferguson, 1983). Malin (1976) proposed that some of the layers may have been remnants of plateau materials. McCauley (1978), on the other hand, claimed that only deposition in a low-energy water environment could explain the observed features. He further postulated the existence of lakes that formed from melted ground ice ponded in depressions created by tectonic subsidence. Lucchitta (1981a, 1982) also concluded that the deposits' morphology was consistent with deposition in an aqueous environment. The Vatles Marineris consists of a large number of tectonic grabens extending for about 2500 km across the planet's surface (Fig. 3). Individual canyons are about 200 km wide and up to 7 km deep. The fine, rhythmically layered, nearly horizontal deposits are as thick as 5 km and laterally continuous over large areas (Fig. 4). Deposits of this kind have been identified in Ophir, Candor, Melas, Gangis, Eos, and Hebes Chasmas (Nedell e t al., 1987b). Nedell e t at. ( I 9 8 7 b ) calculated that the total of the visible deposits in the canyons contain a minimum of ,-~ 105 km 3 of material. Nedell e t al. (1987b) conducted a detailed analysis of Mariner and Viking imagery and proposed four hypotheses for the origin of the Valles Marineris layered deposits. The possible origin of the deposits are:

273

Fig. 3. The Hebes(top), Ophir (middle), and Candor (bottom)Chasmasin the centralsection of VallesMarineris. Layereddeposits are present in each of the canyons but are best seen in Hebes Chasma. Iamge is approximately 1125 and 890 km from left to right and top to bottom, respectively. (USGS image Map MC-I8 NW).

aeolian, remnants of the material that make up the canyon walls, material from explosive volcanic eruptions, and lacustrine. Nedell et al. concluded that the deposition in standing water (i.e., lacustrine) hypothesis is the most likely scenario because it is most consistent with the deposits' observed features including their lateral continuity, horizontality, thickness, and stratigraphic relationship with other units in the canyons. In addition to studies of Valles Marineris, a substantial amount of geomorphological investigation has been directed toward the large outflow channels and their drainage areas (i.e. potential paleotake sites) in the Elysium, Amazonis, and Chryse Basins (Baker & Kochel, 1979; Carr, 1979; Lucchitta & Ferguson, 1983; Scott, 1983; Carr, 1986; Rice et al., 1988; Scott et al., 1991; De Hon, 1992; Rice, 1992; Rice

& Scott, 1993; Rice & De Hon, in press). Figure 5 shows the generalized outlines of basins on Mars with closures exceeding 105 krn 2 in areal extent as originally described by Scott et aI. (1991). We now discuss the evidence for paleolakes in Elysium and Chryse Basins. The Elysium Basin (Fig. 6) extends more than 2500 km along the equatorial lowland plains between the Elysium volcanic rise and the cratered terrain of the southern highlands. The Elysium Basin and its potential as a site of paleolake deposits has been described in detail by Scott & Chapman (t991). With the possible exception of the Chryse Basin, it is the only known depositional basin of regional extent on Mars where direct evidence of an outflow of water from the basin has been found. In a preliminary report, Plescia

274

Fig. 4. Vikingorbiter image of Hebes Chasma(0 °S, 75 °W), a box canyonabout 280 km long (USGS image). The mesa in the center of the canyonshows layered sedimentsthat are believedto have been deposited in standingbodies of water (Nedell et at., 1987). The lakes that may have existed in these canyonsmay actuallydate to the period after the initial warm epoch that formed the dendriticchannelsand thus are of interest exobiologicallyas a possiblehabitat for life after ambientconditionson Mars had become inimicalto life.

(1990) presents an alternative interpretation of Elysium Basin paleolake sites as young lava plains. However, Scott & Chapman (1991) argue the water appears to have flowed into Elysium from many small and some moderate-size channels in the highlands, from the rise around Elysium Mons, and possibly through subsurface aquifers. A former water level at an elevation of about - 1 0 0 0 m can be recognized in places where

shorelines and terraces have been eroded in relatively soft rocks of the Medusae Fossae Formation. Spillways on the eastern, and possibly western end of the paleolake, are also at an elevation of about - I 0 0 0 m. Streamlined bars and islands within a large outflow channel clearly show that water drained eastward from the lake into the Amazonis Basin. The calculated surface area (Table 2) for paleolake Elysium was based

275 180 o

180 °

90 °

\

J 0o

0o NORTH POLE

SOUTH POLE North

120 ° I

180 °

60 °

0o

300 °

I

I

240 °

| 80 °

~

I

/

!

,S ~.

n,,n.z

--

.60 °

-

180 °

-

30 °

4'

/

.30 o

60 °

I

/

0° -

.

\,

4

¢

,,~

s:59 14

-30 °

-60 ° I

I

120 °

60 °

I

I

I

0°

300 °

240 °

180 °

South

Map of Mars showing locations of potential paleolake basins: 1) Amazonis, 25 °N, 157 o; 2) Diacria, 60 °N, 165 o; 3) Aciadalia, 60 °N, 35 o; 4) Utopia, 35 °N, 250 o; 5) Cebernia, 60 °N, 235 o; 6) Isidis, 15 °N, 270 o; 7) Hellas, 40 °S, 295 °; 8) Argyre, 50 °S, 42 o; 9) Chryse, 20 oS, 38 o ; 10) Elysium, 7 °S, 185 o ;11) Aurorae, 8 oN, 35 o ; 12) Ismenius, 65 oN, 310 o ; 13) Prometheus, 85 oS, 300 o ; 14) Sinai, 20 °S, 77 o; 15) Pyramus, 70 °N, 270 o). Arrows indicate proposed waterways between the basins. The dashed line indicates the martian Highland-lowland boundary. Modified from Scott et al. (1991). Fig. 5.

on shoreline indicators together with the area approximately enclosed by the - 1000 m contour line and successively lower surfaces were interpolated at - 1 2 5 0 , - 1500 and - 2 5 0 0 m contours to calculate lake v o l u m e (Scott & C h a p m a n , 1991). C h a p m a n (1994) has recently suggested that there is e v i d e n c e for a frozen paleolake bed in U t o p i a Planitia as recently as 1.8 Gyr ago and based on g e o m o r p h o l o g i c a l relationships determ i n e d the ice to be about 180 m thick.

W h e t h e r the E l y s i u m Basin contained an intermittent or a persistent lake system is not k n o w n . However, the channels that flow into the basin range from Late Hesperian to M i d d l e A m a z o n i a n in age (Fig. 7) suggesting perhaps the recurrence of lake fillings through these geologic periods. Water has apparently overflowed from E l y s i u m into A m a z o n i s . C h a n n e l s from the highlands, including M a n g a l a Valles, entered the basin in several areas along its southern margin. Strandline indicators are recognizable along its western and

276 25 °

20 °

10 °

0o

-10 °

-15 ° 225 °

220 °

210 °

200 °

190 °

180 °

170 °

Fig. 6. Topographicmap of Elysium Basin (Chasmata) showing approximateshoreline of paleolake. Highland areas are shaded. Modified

from Scott & Chapman(1991).

eastern shores, particularly at Acheron Fossae (Scott et al., 1992). Elongate mounds resembling offshore bars lie along the projected shoreline between the plains and the rugged, faulted upland. To the south, a sharp boundary extends along the - 1000 m contour line for 1200 km and separates rough upland terrain deposits of Olympus Mons from the smooth, generally featureless plains of the lowlands. Ambiguity exists, however, farther south where the shoreline closely follows the - 2 0 0 0 m contour. Elsewhere, at lower elevations on the basin floor, etched and eroded materials, sinuous and bifurcating albedo markings, and ridges having whorled, festoon-shaped patterns resembling ice-shoved ridges are closely associated with scars, scrolls, and loops of river meander relics (Scott, 1983). It seems likely that the paleolake in Amazonis Basin may have been reduced during its waning stages to shallow isolated pools and mud flats with winding channels over much of the basin floor.

Apparently, torrential floods of water carved seven huge outflow channels that poured into the Chryse Basin (Fig. 8) during the Hesperian and Amazonian Periods. Carr etal. (1987) estimated that 6.3 × 106 km 3 of water was required to cut these channels, assuming that the floodwaters carried a maximum load of 40% material by volume. The volume of water that must have flowed into the Chryse Basin is much greater than that estimated for any of the other martian basins. Paired shorelines shown by terraces along the base of the Tempe Terra highlands and a low-relief shelf along western Acidalia Planitia may be the erosional vestiges of a long, relatively narrow waterway that partly drained the basin on the northwest. The western and southern sides of Chryse Basin are highly dissected by the large outflow channels and their many branches; in these areas a shoreline is tentatively placed along the zone where streamlined channel bedforms merge with the featureless valley floor. The northeastern and

277 !

4F.--CO2/N 2 A t m o s p h e r e # L o s s

of Atmosphere-b,

Formation of I Tharsis 4 ~ l - - - - D i m i n i s h i n g

Volcanic Activity

Polar///'" Layered

d

Terrain

Valley Networks m

Outflow Channels Canyon Formation

Period of Warm Moist Conditions i i i 4.5 ~"

4.0 Noachian

3.5 r==~

3.0

2.5

2.0

Hesperian

~,.~

I

I

I

1.5

1.0

.5

Amazonian

Present ~

TIME (G yr)

Fig. 7. Geological time scale for Mars showing Noachian (4.5-3.5 Gyr ago), Hesperian (3.5-1.8 Gyr ago), and Amazonian time (1.8 Gyr ago - present) periods. Modified from McKay & Stoker (1989).

eastern shore boundaries are more difficult to place. Parker et aI. (1993) have described examples of possible shoreline features within the fretted terrain of west Deuteronilus Mensae. Their studies also show curvilinear ridges between knobs in eastern Chryse Planitia that are similar in appearance and scale to terrestrial coastal ridges and tombolos connecting islands. These features may be related to a possible shoreline along the east side of the Chryse Basin where narrow terraces on small hills in the plains could also be associated with a stillstand of ponded water. The shoreline indicators along the western and eastern edges of the Chryse Basin in some places closely follow the - 1000 m contour line. However, the northern and southern margins of a former lake in the basin are ill defined and the trends of their estimated locations are uncertain. A probable minimum value for lake size was obtained by closing the - 1000 m contour line at about lat 30 °N and using this datum, adjusted to coincide with outflow channel terminals, as the shoreline limit. As shown in Table 2, the minimum water volume for the paleolake in Chryse is estimated to be 9.2 x 105 km 3, whereas a more probable volume is 13.4 x l0 s km 3.

Ice-Covered lakes

Lakes on Mars were probably ice-covered during the later stage of their history (Wharton et al., 1989a; Scott et al., 1991). Using a simple climate model of Mars, McKay & Davis (1991) suggested that the most signifi-

cant determinant of the duration of martian ice-covered lakes would have been the availability of meltwater to recharge ablation losses. This in turn would have been determined by the existence of peak seasonal temperatures above freezing. McKay & Davis (1991) suggest that if there was a source of meltwater, liquid water habitats could have persisted under relatively thin ice covers until temperatures fell below - 4 0 °C, which may have occurred up to 700 million years after mean global temperatures fell below freezing. McKay et al. (1985) created an energy balance model that predicts the thickness of ice on Antarctic lakes. It assumes that the sources of heat for an Antarctic lake (Fig. 8) are sunlight penetrating the ice, geothermal heat flux from the bottom of the lake (small in Antarctica), sensible heat from the meltwater entering the lake (also small), and latent heat released as meltwater freezes at the ice water interface. The primary mechanism of heat loss is through conduction up through the ice. The main determinant of ice thickness is the rate of ablation at the ice surface. Applying this model to the purported ice-covered martian paleolakes, McKay et al. (1985) predict that ice thicknesses of 3.4, 11, and 19 m would have existed when the mean temperature was - 2 3 , - 3 3 , and - 4 3 °C, respectively. Clow (1987) demonstrated that transitory melting of surface snowpacks could have occurred in Mars' past history when low pressures and temperatures prevailed. This is important in terms of paleolakes because melted snow, along with groundwater, could have provided a source of liquid water to feed and sustain the lakes. As snowmelt and groundwater flowed into the

278

Fig. 8. Largelake basins and drainagenetworksin ChrysePtal~itia.ModifiedfromDe Hon (1992).

lakes, they would have carried C O 2 and cations dissolved from the regolith. As the water froze at the ice-water interface, dissolved materials would have been forced into the water column where they would become concentrated and enhance carbonate precipitation (McKay & Nedell, 1988). From a biological perspective, the presence of an ice-covered lake extends the time period for life on Mars (Fig. 10). An ice-covered lake could have provided a relatively clement habitat and may have served as one of the last places where organisms could have existed as outside environmental conditions deteriorated. An ice cover insulates a lake from outside temperatures allowing temperature stability. Another important feature of an ice cover is that it allows for the concentration of biologically important atmospheric gases such

as 02 and N2 (Wharton et al., 1989a, b) on Earth and C O 2 and Nz on Mars.

The possible presence of ice-covered lakes on Mars raises an interesting question as to the mechanism(s) of sediment deposition into these lakes. How are sediments deposited in an ice-covered lake? Nedell et al. (1987b) consider three possible mechanisms for sediment deposition in ice-covered martian lakes. Sediments could have been deposited down through the ice cover, up fTom the lake bottom, or in from the lake margins. We now briefly discuss each of these mechanisms. Nedell et al. (1987b) suggested four hypotheses to explain the transport of sediment down through the ice cover of a take on Mars: (1) sediment particles were warmed by the sun allowing them to melt through the ice, (2) sediment worked its way down through

279 ",..SUNLIGHT "

//

~

ATMOSPHERIC . &SENSIBLE AND "%/~ ~ LATENT HEAT \~ ICE "TCONDUCTION COVER........

X ~....

MEELTWATER NSIBLE AND LATENT HEAT

LAKE

DISSOLVED GASES GEOTHERMAL

" ".r/// /,vV//.,.//////// Fig. 9. Schematic of the physics of an ice-covered lake showing the buffering of temperature and the concentration of atmospheric gas that occur in perennially ice-covered lakes. Thermal buffering results from the presence of an ice cover that insulates liquid water from the atmosphere. The ice cover also serves to restrict the exchange of gases between the lake water and the atmosphere. Gases produced in the lake as a result of biology and physical processes will become concentrated with time. On Earth, such lakes are found in the dry valleys of Antarctica and may have also occurred on Mars during Epoch II (see Table 1), when mean global temperatures were below freezing.

-'-""~

30

Ice-free lakes

(oc)

•

~

Ice-covered lakes

-35

-60 4

TIME

(Gyr)

0

Fig. 10.

Hypothetical plot showing the decrease in mean annual surface temperature with time. Ice-covered lakes would have been present on the martian surface when temperatures fell below 0 °C. At temperatures below - 4 0 o C, the input of summer meltwater ceases and the take would eventually ablate away - maintaining the same ice cover thickness.

the ice via vertical water filled channels, (3) sediment deposited on the ice cover was thick enough to cause the ice to founder, dumping the sediment into the lake, and (4) sediment deposited on the ice cover led to a Rayleigh-Taylor instability where sediment diapairs penetrated downward through the ice. The first hypothesis was considered by Simmons et al. (1986) and Wharton et al. (1989b) who con-

cluded on the basis of studies of the ice-covered lakes in Antarctica, that sand-sized particles will not heat up the ice enough to melt through even if the ice is within a degree of melting and all other variables are optimal. On Mars, surface temperatures and solar flux are lower than on Earth, and equilibrium ice thickness may have been greater than in Antarctica. Therefore, the possibility of sediment melting through the ice is fairly remote. Vertical melt channels (hypothesis 2), like the ones that form in Antarctic lakes as a result of gas bubble formation in summer, also do not appear to be a feasible means of transporting sediment into icecovered lakes on Mars. For this mechanism to have worked on Mars, ice covers would have had to have been thinner than the currently believed values of 3.4 to 19 m (McKay et at., 1985) and liquid water at the surface of must have been stable for periods of days to weeks during the summer to permit the melt channels to form. Foundering (hypothesis 3) does not appear to be a viable method for sedimentation. As sediment accumulated on the ice it would prevent ablation and sublimation and cause the ice cover to thicken to an extent that all but the largest amounts of sediment could not cause foundering (Nedell et aL, 1987b). According to Nedell et al. (1987b), Rayleigh-Taylor instabilities (hypothesis 4) appear to be the only reasonable hypothesis explaining sedimentation down through the ice. These types of instabilities occur when a dense fluid overlays a less dense fluid. At the boundary between the two layers localized protrusions of the denser fluid (diapirs) begin to intrude into the lower layer. These protrusions grow because of the unstable density stratification. The final result is that the denser fluid flows through and under the less dense fluid. This can be applied to a lake surface by treating the sediment as a dense fluid (valid if the particles are sufficiently small) and the ice layer as the less dense fluid (valid since ice deforms in a fluid manner). Nedell et aL (1987b) suggest subaqueous volcanism as a second mechanism for providing sediments to an ice-covered lake. This mechanism allows for the addition of sediments to the lake up through the lake bottom, therefore, without having to remove or penetrate the ice cover. The low energy liquid water environment would still be available for the control of the sedimentation process. The efficiency of a subaqueous environment to equally distribute material could help explain why unequivocal volcanic vents have not been found. Underwater volcanism might also

280 help explain the lack of deposition on the surrounding uplands (Nedell et aL, 1987b). The third mechanism advanced by Nedell et aI. (1987b) for sedimentation in ice-covered lakes is by transport of sediment from the canyon walls. However, this hypothesis is problematic because the volume of sediment eroded from the former canyon walls is not sufficient to account for all of the sediment in the deposits. An additional mechanism for the formation of the layered deposits was suggested by McKay & Nedell (1988). They proposed that the deposits could be at least partially composed of carbonate material precipitated from the atmosphere into lakes when the CO2 partial pressure was high. McKay & Nedell (1988) calculated that if all of the layered deposits found in the Valles Marineris were composed of some form of carbonate, then it would represent an atmospheric abundance of CO2 equal to 3000 Pa, which does not exceed estimates of past CO2 abundance in Mars' early atmosphere. This suggests that carbonate precipitation is not only a feasible mechanism for the creation of the layered deposits, but is also able to account for the volume of material estimated by Nedell et aL (1987b) to exist in the deposits. In order to gather empirical support for their hypothesis, McKay & Nedell (1988) reanalyzed Mariner 6/7 infrared spectrometer data in the wavelength region between 2 and 6 #m to try to positively identify the presence of carbonates in the paleolake region. McKay & Nedell did not observe carbonates in their analysis. However, the inability to detect carbonates may be due to factors other than the actual absence of carbonates; for example, the lack of spectral footprints directly overlying the deposits and the possibility that the deposits may be camouflaged by an aeolian 'dust' mantle.

Terrestrial analogs Martian lakes that existed before the climate drastically cooled may have been ice-free or seasonally icefree surfaces. In this case, some of the best terrestrial analogs may be the Pleistocene lakes such as lakes Bonneville and Lahontan in western North America. Non ice-covered lakes and their associated landforms in western North America have been studied for more than a century (Gilbert, 1890). These classical lacustrine landforms which require open water and wave action include: wave cut platforms, strandlines, ter-

races, beach ridges, cusps, spits, bars, tombolos, and sea stacks (Reeves, 1968; Currey, 1980). In contrast, perennially ice-covered lakes lack most of these features because the ice cover acts as a geomorphically protective agent (Rice, 1992, 1994). Martian lakes may have undergone several episodes of ice-covered and non ice-covered regimes. However, as discussed previously, it is probable that the last vestiges of any lakes that existed on the martian surface would have been ice-covered. Much of what is theorized about former martian ice-covered lakes derives from research conducted in Antarctica (Wharton et aL, 1989a; Rice, 1992, 1994). Environmental conditions in the relatively ice-free (oasis) areas of Antarctica strongly mimic those of Mars, with both locations being very cold, windy, and arid (McKay & Stoker, 1989). The Antarctic contains perennially icecovered lakes (Fig. 11) with physical, chemical, and biological properties influenced primarily by the presence of their ice covers (Wharton et aI., 1993). As discussed previously, antarctic lakes are proposed to be analogs of the latter stages of martian paleolakes - a time when the surface of Mars became colder (McKay & Davis, 1991). Results from research of antarctic lakes is being used to help identify possible paleolake sites on Mars and focus geological and exobiological investigations during future Mars missions (McKay et al., 1990a; Rice, 1992). From a geomorphological perspective, the most notable landforms produced by ice-covered lakes are ice-shoved ridges (Rice, 1992). These features form discrete segmented ramparts of boulders and sediment pushed up on the shores of a lake (or sea). The shoreline surfaces are generally planated with the ramparts defining the inner edge of the shoreline. The extensive planated nature of the shoreline surface is due to the bulldozing effects of lake ice push and expansion. The ice-shoved ridges can have a veneer of boulders, gravel and sand mantling an ice core. The ice core normally melts and leaves behind its mantle of material in the form of irregular, discontinuous ridges. The sediment mantle covering the ice core acts as an insulating blanket and can preserve the ice core for several years (Hume & Schalk, 1964). Ice-shoved ridges observed in the Bunger Hills of East Antarctica (66.5 °S, 100 °E) are up to 80 m long, 2 m high, and 4 m wide (Rice, 1992). Some ice-shoved ridges up to 305 m long and 9 m high have been reported along beaches in the Arctic (Hume & Schalk, 1964). However, most terrestrial ice-shoved ridges are usually no more than 1 m high (French, 1976). The

281

Perenniallyice-coveredLake Hoare in TaylorValley, one of the dry valleys in Southern Victoria Land, Antarctica. Lake Hoare is about 4 km long.

Fig. l l.

emplacement of ice-shoved ridges along the shore can be attributed to either thermal contraction and expansion of the ice cover or wind-induced waves acting upon the separate ice plates driving them shoreward (Tricart, 1970). Other unique features associated with polar shorelines are pitted beaches, frost cracks and mounds, gelifluction deposits, patterned ground, pingos, ice rafted erratics, poorly rounded stones, coastal striated bedrock, and ventifacts (Rice, 1994).

Opportunities for paleolimnological studies on Mars Future missions to Mars will provide more definitive data regarding paleolake sites. Of primary importance is high resolution imaging. Current plans call for the systematic mapping of the entire surface of Mars at moderate resolution (30 m per pixel) with selected areas obtained at high resolution (3 m per pixel). In addition, near infrared spectroscopic and thermal emis-

sion instruments in Mars orbit can identify mineralogical signatures indicative of sedimentary and precipitory deposits- in particular carbonates and nitrates. With the failure of the Mars Observer mission these objectives have been delayed, but they are still the current emphasis for future Mars exploration. Follow-up missions could include landers targeted to specific paleolake sites. Investigations into the nature of the sediments on such a lake bed could be conducted with a small rover equipped with a drill and sampling arm that can obtain samples from a few meters depth. Mineralogical and organic analyses could be conducted on a transect from the landing site at the center of the lake across to a inflow channel and then up the channel bed. Future terrestrial studies of Mars analog paleolakes should be continued in the Arctic and Antarctica. The geomorphology of cold region paleolakes and marine beaches should be studied with respect to both large scale landforms and sedimentology. Studies of contemporary polar lakes are also important to more fully

282 understand the geomorphological and sedimentological processes associated with these lakes. On Earth, studies of lake and marine deposits have yielded information about the biological and climatological features of the planet from as long ago as the Archean (3 Gyr ago). While life outside of the Earth's atmosphere has yet to be observed, there is a general consensus among exobiologists that the most productive search strategy for life would be based on locating liquid water. Therefore, the search for other planets or celestial bodies that may be habitable necessarily involves the search for liquid water. The discovery of lakes, or the sediments of former lakes on another planet - Mars being the likely candidate in our solar system - would be of protbund scientific interest and would begin interplanetary Limnology.

Acknowledgments This work was supported by grants to RAW from the National Aeronautics and Space Administration (NAGW-1947, NCA2-799) and the National Science Foundation (OPP-9211773). The participation of JMC was made possible by the Research Experience for Undergraduates (REU) program at the University of Nevada, Reno, which was sponsored by the National Science Foundation (EAR-9200194). We are grateful to A. Z. Butt and M. D. Wharton for their assistance with the manuscript.

References Baker, V. R., 1979. Erosional Processes in channelized water flows on Mars. J. Geophys. Res. 84: 7985-93. Baker, V. R. & R. C. Kochel, 1979. Martian channel morphology: Maja and Kasei Valles. J. Geophys. Res. 84: 7961-83. Baker, V. R. & D. J. Milton, 1974. Erosion by catastrophic floods on Mars and Earth. Icarus 23: 27-41. Baker, V. R, R. G. Strom, V. C. Gulick, J. S. Kargel, G. Komatsu & V. S. Kale, 1991. Ancient Oce,~s, Ice sheets, and the hydrological cycle on Mars. Nature 352: 589-94. Breckenridge, G. R., I-I. E. Newsome & V. R. Baker, 1985. Ancient hot springs on Mars: Origin and paleoenvironmental significance of Martian valleys. Geology 18: 859-862. Carr, M. H., 1974. The role of lava erosion in the formation of lunar rilles and Martian channels. Icarus 22: 1-23. Carr, M. H., 1979. Formation of Martian flood features by release of water from confined aquifers. J. Geophys. Res. 84: 2995-3007. Cart, M. H., 1981. The Surface of Mars. Yale University Press, New Haven, Connecticut. Cart, M. H., 1983. Stability of streams and lakes on Mars. Icarus 56: 476-495. Cart, M. H., 1986. Mars: A water rich planet? Icarus 68: 187-216.

Can', M. H., 1989. Recharge of the early atmosphere of Mars by impact-induced release of CO2. Icarus 79:311-327. Carr, M. H. & G. D. Clow, 1981. Martian channels and valleys: Their characteristics, distribution, and age. Icarus 48:91-117. Cess, R. D., V. Ramanathan & T. Owen, 1980. The Martian paleoclimate and enhanced carbon dioxide. Icarus 41: 159-165. Chapman, M. G., 1994. Evidence, age, and thickness of a frozen paleotake in Utopia Planitia, Mars. Icarus 109: 393-406. Clow, G. D., 1987. Generation of liquid water on Mars through the melting of a dusty snowpack. Icarus 72: 95-t27. Cutts, J. A. & K. R. Blasius, 1981. Origin of Martian outflow channels: the eolian hypothesis. J, Geophys. Res. 86: 5075-5102. De Hon, R. A., 1992. Martian lake basins and lacustrine plains. Earth, Moon, and Planets 56: 95-122. French, H. M., 1976. The Periglacial Environment, Longman Group Ltd., London. Gilbert, G. K., 1890. Lake Bonneville, U.S. Geol. Surv. Monograph 1. Goldspiel, J. M. & S. W. Squyres, 199t. Ancient aqueous sedimentation on mars. Icarus 89: 392--410. Gulick, V. C. & V. R. Baker, 1989. Fluvial valleys and martian paleoclimates. Nature 341: 514-16. Haberle, R. M., D. Tyler, C. P. McKay & W. C. Davis, 1993. A model for the evolution of CO2 on Mars. In Mars Past, Present, and Future, Results from the MSATT Program. LPI Technical Report 93-06: 19. Hume, J. D. & M. Schalk, 1964. The effects of ice push on arctic beaches. Am. J. Sci. 262: 267-273. Kahn, R., 1985. The evolution of carbon dioxide on Mars. Icarus 62: 175-190. Lucchitta, B. K., 1980. Martian outflow channels sculpted by glaciers. In Lunar and Planetary Science XI. Lunar and Planet. Inst., Houston. Lucchitta, B. K., 1981. Valles Marineris - faults, volcanic rocks, channels, basin beds. In Reports of the Planetary Geology Program. NASA-TM 84211: 419-421. Lucchitta, B. K., 1982. Lakes or playas in Valles Mariners. In Reports of Planetary Geology Program, NASA-TM 85127: 233-234. Lucehitta, B. K., D. M. Anderson & H. Shoji, 1981. Did ice streams carve Martian outflow channels? Nature 290: 763-959. Lucchitta, B. K. & H. M. Ferguson, 1983. Chryse basin channels: Low gradients and ponded flows. J. Geophys. Res. 88: 553-568. Malin, M. C., 1976. Nature and origin ofintercrater plains on Mars. Ph.D. thesis. California Institute of Technology, Pasadena. Masursky, H., 1973, An overview of the geologic results from Mariner 9. J. Geophys, Res. 78: 4037-4047. Masursky, H., J. M. Boyce, A. L. Dial, G. G. Schaber & M. E. Strobell, 1977. Formation of Martian channels. J. Geophys. Res. 82: 4016-4038. McCanley, J. E, 1978. Geologic map of the Coprates Quadrangle of Mars. U.S. GeoL Survey Misc. Inv. Series Map 1-897, scale 1:5 000 000. McCauley, J. F., M. H. Cart, J. A. Cutts, W. K. Hartmann, H. Maanrsky, D. J. Milton, R, R Sharp & D. E. Wilhelms, 1972. Preliminary Mariner 9 report on the geology of Mars. Icarus 17: 289-327. McKay, C. R, G. D. Clow, R, A. Wharton, Jr. & S. W. Squyres, 1985. Thickness of ice on perennially frozen lakes. Nature 313: 561-562. McKay, C. R & W. L. Davis, 1991. Duration of liquid water habitats on early Mars. Icaras 90: 214-221. McKay, C. R, E. L Friedmann, R. A. Wharton, Jr. & W. L. Davis, 1990a. History of water on Mars: A biological perspective. Adv. Space Res: 1-7.

283 McKay, C. R, R. L. Mancinelli, C. R. Stoker & R. A. Wharton, Jr., 1992. The possibility of life on Mars during a water-rich past. In Keifer, H. H. et aL (eds.), Mars. Univ. Arizona Press, Tucson: 1234-1245. McKay, C. R & S. S. Nedell, 1988. Are there carbonate deposits in the Valles Marineris, Mars? Icarus 73: 142-148. McKay, C. E & C. R. Stoker, 1989. The early environment and its evolution on Mars: Implications for life. Rev. Geophys. 27: 189-214. Milton, D. J., 1973. Water and processes of degradation in the Martian landscape. J. Geophys. Res. 78: 4037-4047. Moroz, V. I. & L. M. Mukhin, 1977. Early evolutionary stages in the atmosphere and climate of the terrestrial planets. Cosmic Res. 15: 769-791. Nedell, S. S., D. W. Andersen, S. W. Sqnyres & E G. Love, 1987a. Sedimentation in ice-covered Lake Hoare, Antarctica. Sedimentology 34: 1093-1106. Nedell, S. S., S. W. Squyres & D. W. Anderson, 1987b. Origin and evolution of the layered deposits in the Valles Mariners, Mars. Icarus 70: 409-441. Parker, T. J., D. S. Gorscine, R. S. Saunders, D. C. Pieri & D. M. Schneeberger, 1993. Coastal geomorphology of the martian northern plains. J. Geophys. Res., 98: 11061-11078. Parker, T. J., R. S. Saunders & D. M. Schneeberger, 1989. Transitional morphology in the west Deuteronihis Mensae Region of Mars. Icarus 82:111-45. Peterson, C. R, 1981. A secondary origin for the central plateau of Hebes Chasma. Lunar and Planetary Science Conference 12B, Proceedings: 1459-1471. Pieri, D. C., 1976. Martian channels: Distribution of small channels on the Martian surface. Icarus 27: 941-952. Pieri, D. C., 1979. Geomorphology of Martian Valleys. Ph.D. Thesis, Cornell University, Ithaca, N.Y. (Reprinted in NASA TM-81979: 1-160). Pieri, D. C., 1980. Martian valleys: Morphology, distributions, age, and origin. Science 210: 895-897. Plescia, J. B., 1990. Young flood lavas in the Elysium region, Mars. Lunar Planet. Sci. XXI: 969-970 (abstract). Pollack, J. B. & Y. L. Yung, 1980. Origin and evolution of planetary atmospheres. Ann. Rev. Earth Plan. Sci. 8: 425-487. Pollack, J. B., 1979. Climatic change on the terrestrial planets. Icarus 37: 479-553. Pollack, J. B., J. E Kasting, S. M. Richardson & K. Poliakoff, 1987. The case for a wet, warm climate on early Mars. Icarus 71: 203-224. Rice, J. W., Jr., 1992. Antarctic Lakes (Above and Beneath the Ice Sheet): Analogs for Mars, Workshop on the Polar Regions of Mars (Proceedings). MSATT LPI Tech. Report 92-08: 23-24. Rice, J. W., Jr., 1993. The Martian Valley Networks: Origin by Niveo-Fhivial Processes, Workshop on Early Mars (Proceedings). MSATT LPI Tech. Report 93-03: 22-23. Rice, J. W., Jr., In Press. Polar Beach Processes. In Proceedings Lunar Planet.Sci. Conf. 25. Rice, J. W., Jr., S. M. Archibald & R. A. De Hon, 1988. Lower Maja Valles, Lunar Planet. Sci. Conf. 19: 76-77. Rice, J. W., Jr. & R. A. De Hon, in press. Geologic Map of the Darvel Quadrangle, Masa Valles Region, Mars, U.S. Geol. Surv. Misc. Inv. Series Map 1-2432, scale 1:500 000. Rice, J. W., Jr. & D. H. Scott, 1993. Martian Deltas: Morphology and Distribution, Mars. In Past, Present, and Future-Results from the MSATT Program. LPI Tech. Report 93-06: 45-46. Sagan, C., O. B. Toon & R J. Gierasch, 1973. Climate change on Mars. Science 181: 1045-1049.

Schonfeld, E., 1977. Martian volcanism. In Lunar Science VIII. Lunar Planet. Inst., Houston: 843-845. Schumm, S. A., 1974. Structural origin of large Martian channels. Icarus 22: 371-384. Scott, D. H., 1983. Meander relics: direct evidence of extensive flooding on Mars. In Conference on Planetary Volatiles, R. O. Pepin and R. O'Connell (eds.). LPI Tech Rept. 83-01: 157-166. Scott, D. H. & M. G. Chapman, 1991. Mars Elysium Basin: Geologic/volumetric analysis of a young lake and exobiologic implications. In Proceedings of Lunar and Planetary Science Conference 21: 669-77. Scott, D. H., M. G. Chapman, J. W. Rice & J. M. Dohm, 1992. New evidence of lacustrine basins on Mars: Amazonis and Utopia Planitiae. In, Proceedings of the Lunar and Planetary Science Conference 22: 53-62. Scott, D. H., J. W. Rice, Jr. & J. M. Dohm, 1991. Martian paleolakes and waterways: exobiological implications. Orig. Life Evol. Biosph. 21: 189-198. Sharp, R. P. & M. C. Malin, 1975. Channels on Mars. Geol. Soc. Am. Bull. 86: 593-609. Simmons, G. M., Jr., R. A. Wharton, Jr., C. P. McKay, S. S. Nedell & G. D. Clow, 1986. Sand/Ice interactions and sediment deposition in perennially ice-covered Antarctic lakes. Antarct. J. U.S. 21: 217-220. Solomon, S. C., J. W. Head, R. P. Comar, 1979. Thickness of martian lithosphere from tectonic features. NASA Tech. Memo 80339: 60-62. Squyres, S. W., 1989. Urey Prize Lecture: Water on Mars. Icarus 79: 229-288. Squyres, S. W., D. W. Andersen, S. S. Nedell & R. A. Wharton, Jr., 1991. Lake Hoare Antarctica: Sedimentation through a thick perennial ice cover. Sedimentology 38: 363-379. Squyres, S. W. & J. F. Kastings, 1994. Early Mars: how warm and wet? Science 262: 744-749. Toon, O. B., J. B. Pollack, W. Ward, J. A. Bums & K. Bilski, 1980. The astronomical theory of climatic change on Mars. Icarus 44: 552-607. Tricart, J., 1970. Geomorphology of Cold Environments, MacMillan and Co. Ltd, London. Walker, J. G. C., 1978. Atmospheric evolution on the inner planets. In Comparative Planetology, Academic Press, New York: 141163. Wallace, D. & C. Sagan, 1979. Evaporation of ice in planetary atmospheres: Ice-covered rivers on Mars. Icarus 39: 385-400. Wharton, R. A., Jr., G. M. Simmons, Jr. & C. P. McKay, 1989b. Perennially ice-covered Lake Hoare, Antarctica: Physical environment, biology, and sedimentation. Hydrobiologia 172: 306320. Wharton, R. A., C. P. McKay, G. D. Clow & D. T. Andersen, 1993. Perennial ice covers and their influence on Antarctic lake ecosystems. In Physical and Biogeochemical Processes in Antarctic Lakes, Green, W. and I. E. Friedmann (eds.), Antarctic Research Series Vol. 59, American Geophysical Union, Washington, D.C.: 53-72. Wharton, R. A., Jr., C. P. McKay, R. L. Mancinelli & G. M. Simmons, Jr., 1989a. Early Martian environments: The Antarctic and other terrestrial analogs. Adv. Space Res. 9: 147-153. Yung, Y. L. & J. P. Pinto, 1978. Primative atmosphere and implications for the formation of channels on Mars. Nature 273: 730-732.