Lakes on Mars

USA

Searching for lakes on Mars: Four decades of exploration

Nathalie A. Cabrol, Edmond A. Grin

SETI CSC/ NASA Ames, Space Science and Astrobiology Division, Moffett Field, CA, USA

Keywords

Aquifers

Lakes

Mars

Power law distribution

Runoffs

Acknowledgment

The authors wish to thank Victor Baker and Keith Harrison for their in-depth reviews that helped improve the manuscript. The analysis comparing the terrestrial and Martian lake distributions used the data published in Table 1 by Fassett and Head (2008) and Table 2 by Downing et al. (2006), respectively with permission of the Icarus and Limnology and Oceanography journals.

Abstract

Morphological and mineralogical evidence has accumulated over the past decade to confirm the presence of lakes early in the history of Mars. The ability of lakes to keep a record of climatic, hydrological, and biological cycles make their preserved sediments primary targets for exploration. As a result, ancient lake basins now represent a significant proportion of the candidate landing sites for future missions. The history and evolution of such terrain are intimately associated with the history of water on Mars and there are many uncertainties as to the climatic conditions under which they formed. Although a range of temperature-dependent mineral assemblages have been found in Martian paleolakes, a broad range of physical environments can account for these observations. While warmer conditions may have prevailed at some point in time, they did not last. However, the ability of water to circulate and pond does not necessarily require lasting warm conditions. Another important issue is how representative the seemingly low residual lake record is of the original population. When applied to Mars, advances in terrestrial modeling show that nearly 12,000 lakes could have been produced from the present incomplete record alone. Their size-frequency distribution is predicted by a power law function. A constant production rate 60% that of Earth is compatible with sustained climate conditions accompanying their formation. The reconstructed spatial distribution of these lakes is bimodal, with one mode in each hemisphere and maximum concentration in the southern mid-to-high latitude. The inferred resurfacing of ancient lake morphologies appears to be related both to the size of the lakes and to their latitude. These results show that despite a seemingly low residual record, lakes were abundant in early Mars, reinforcing their role as potential primary habitats for life.

1.1 Introduction

Lakes are time capsules. On Earth, they are considered sentinels of climate change (Williamson et al., 2009) and may have played the same role on early Mars (e.g., Baker, 2001; Cabrol and Grin, 2002, 2005). Their basins capture the record of geological and environmental fluctuations over a wide range of temporal and spatial scales (Figure 1.1). Terrestrial lakes host a diversity of habitats where life’s adaptability can be pushed to the edge in often unstable environments (e.g., Bronmark and Hansson, 1998; Cabrol et al., 2009; Herbst, 2001; Hollibaugh et al., 2001; Oren, 2001; Price, 2000; Stivaletta et al., 2009; Warwick and Laybourn-Parry, 2008; Wynn-Williams and Edwards, 2000; see also Chapter 13). They preserve the evidence of ancient life as sedimentation rapidly entombs dead organisms and generates anoxic conditions favoring the formation of fossils (e.g., Beaty et al., 2005; Des Marais and Farmer, 1995; Des Marais et al., 2003, 2008; Farmer, 1999, 2003; Hoffman et al., 2008). This makes them prime candidates for exploration.

Figure 1.1 The existence of ancient lakes on Mars is now supported by converging evidence provided by both mineralogy and sedimentary records. Here, the Holden basin near the mouth of Uzboi Vallis has become one of the primary candidate sites for MSL (see Chapter 12 ). HiRISE image PSP_001666_1530, NASA/JPL/University of Arizona.

The existence of lakes on ancient Mars is now widely accepted but that was not always the case. The history of science shows that knowledge on any scientific question is shaped by the means of exploration and those means are molded by what we think the world is. Prior to MGS of the late 1990s, the relatively low resolution of orbital imagery made it difficult to confirm Martian paleolakes by direct observations, though their existence was inferred because valley networks had already been identified on Viking and Mariner 9 images. Interpretation rested on ambiguous morphological evidence at 200 m/pixel on average with only localized coverage at higher resolution. Yet, physical conditions measured at Mars and early modeling (Baker et al., 1991; Clifford, 1993; Haberle and Jakosky, 1990; Haberle et al., 1993; Jakosky and Phillips, 2001; Kasting, 1997; McKay and Davis, 1991; Sagan and Mullen, 1972; Sagan et al., 1973; Toon et al., 1980) supported the hypothesis that channels and valleys had been carved by water early in Martian history. Under such conditions, lakes could have formed in topographic lows unless water had completely evaporated and infiltrated before reaching basins. However, topographic uncertainties made for difficulty in identifying with confidence depressions that might have hosted paleolakes. The exception, of course, was the obvious basin morphology created by impact craters, and this explains why much of the early Mars paleolacustrine research emphasized these features (Cabrol, 1991; Cabrol and Grin, 1995; De Hon, 1992; Forsythe, 1990; Forsythe and Blackwelder, 1998; Forsythe and Zimbelman, 1995; Goldspiel and Squyres, 1991; Grin and Cabrol, 1997; Kuzmin et al., 2000; Newsom et al., 1996; Scott et al., 1991, 1995).

Although valley networks were identified, they appeared poorly integrated (see Section 1.2.2 and Chapters, 2 and, 8), and their origin by water was questioned. This set the stage for other hypotheses, such as lava (Carr, 1974; Greeley, 1973; Schonfeld, 1977), liquefaction of surface material (Nummedal, 1978), glaciers (Lucchitta, 1982), CO2 (Hoffman, 2000; Sagan et al., 1973), or the dissociation of clathrates (Baker et al., 1991; Kargel and Lunine, 1998; Max and Clifford, 2000; Mellon, 1996; Miller and Smythe, 1970; Milton, 1974; Musselwhite and Lunine, 1995; Peale et al., 1975; Yung and Pinto, 1978). Each of these hypotheses bears considerable limnological and astrobiological implications. As our understanding of Mars grows, especially with regards to the mineralogical nature of sedimentary deposits, the hypothesis that valley networks and channels were formed by CO2 and/or clathrates seems less probable, although it cannot be completely dismissed (Kargel et al., 2000). It remains a plausible mechanism for the formation processes associated with outflow channels in the equatorial regions (Baker et al., 1991; Kargel et al., 2007; Komatsu et al., 2000).

In a planet strongly influenced by volcanic processes for much of its history, it is not surprising that the study of high-resolution imagery has confirmed in greater detail the major role played by lava in the formation of many channels (Basilevskaya et al., 2009; Keszthelyi et al., 2006; Sakimoto and Zuber, 1998; and others). Lava lakes formed in basins and calderae during volcanic activity (Gregg and Lopes, 2007; Lang and Farrell, 2009; Mouginis-Mark and Robinson, 1992; Mouginis-Mark and Wilson, 1999) and, as it was the case in the Viking era (Milton, 1973), lava is still at the heart of the controversy when discussing the nature of the Martian channels, valleys, lakes, and ponds (Bleacher et al., 2010; Chapman, 1999; De Hon, 1992; Jaeger et al., 2008; McEwen et al., 2002; Scott and Chapman, 1991; and others). The putative paleolakes hosted by the Hellas and Elysium basins are the most recent examples of such ongoing debate (Chapters, 6 and, 7) and show the constraints and limitations of orbital exploration even with high-resolution imagery and multispectral data (Figure 1.2). However, there are now converging morphological, geological, and mineralogical lines of evidence that reasonably support the existence of ancient standing bodies of water on Mars. These are the features discussed in this chapter and defined as lakes.

Figure 1.2 Elysium Planitia platy flows have been interpreted either as evidence for a frozen sea (see text and references therein) or lava flow. HiRISE image TRA_000854_1855, NASA/JPL/University of Arizona.

1.2 Conditions at the time of lake formation

Whether the early climate of Mars was much warmer and wetter in the Noachian compared to the later geological epochs is still the subject of ongoing debate. Many lakes do not require warm conditions to form (Chapter 4). On Earth, 60% of them are located at high latitudes in the northern hemisphere (ILEC web-based database), many inherited from the last deglaciation. But lakes require water to form, and water balance drives their evolution and duration. Therefore, uncertainties about climate and water on Mars weigh on our understanding of the Martian limnology. Current theories about the acquisition and evolution of water and early Martian climate are presented by Carr and Head in Chapter 2 (and references therein), thus are not developed here. The premise of our discussion is the following: The existence of standing bodies of water on Mars required that at some point in its history, possibly repeatedly, physicochemical and environmental conditions allowing water to circulate and to pond were met. Any conditions allowing lake formation will be referred to generically as favorable conditions in the remainder of this chapter. In this section, we examine the collective mineralogical and morphological evidence, assess the range of plausible environments consistent with lake formation, and determine the significance of these environments in terms of relative water abundance and climate at the time of lake formation.

1.2.1 Clues from mineralogy

Discrete ranges of physical environments can account for most of the temperature-dependent minerals and mineral assemblages that have recently been identified on Mars (Figure 1.3). Data suggest that, in addition to the largely mafic bedrock of the Martian surface, there are components of both (i) a mobile dust fraction and (ii) a chemically altered in situ bedrock (Bibring et al., 2006). Iron oxides (Christensen et al., 2000, 2001; Glotch and Rogers, 2007) and hydrated minerals have been detected in outcrops (Bibring et al., 2005, 2006; Mustard et al., 2008). Their stratigraphic relationships suggests that mineralization of phyllosilicates essentially took place in the Noachian (Bibring et al., 2005; Mustard et al., 2008; Poulet et al., 2005). At Nili Fossae, carbonate-bearing rocks are present in layered exposures (Elhmann et al., 2008) associated with phyllosilicates and olivine-rich strata of comparable ages. The mineralogical sequence inferred from stratigraphy is consistent with alkaline to neutral waters during the early Noachian and acidic weathering during the Hesperian (Bibring et al., 2005, 2006; Mustard et al., 2008). The preservation of carbonates through the early Hesperian (Ehlmann et al., 2008) may indicate local to regional exceptions.

Figure 1.3 Diverse mineralogy on a plateau near in Juventae Chasma, Valles Marineris 4.7°S, 296.4°E. The deposits contain opaline silica and iron sulfates, consistent with low-temperature, acidic aqueous alteration of basalt that could be the result of surface runoff and fluvial deposition during the Hesperian ( Mustard et al., 2008 ). HiRISE image PSP_003579_1755, NASA/JPL/University of Arizona.

The primary process of clay formation as hydrothermal processes through volcanism and impact cratering may have included the weathering of Fe/Mg mafic minerals, as inferred from their dominance relative to plagioclase (Mustard et al., 2008). From an environmental standpoint, the presence of clays is consistent with, but does not necessarily require, sustained warm and wet surface conditions on early Mars (Catling, 2007). The existence of a volatile-rich subsurface during this early period is supported by the presence of lobate ejecta craters (Barlow et al., 2000 and others) and ancient water activity associated with hydrothermal activity (Bishop et al., 2008; Gulick, 1998; Neukum et al., 2004; Squyres et al., 2007), which provided a mechanism for clays to form through interaction between the Martian crust and hydrothermal waters.

It is also possible that the current locations for some of the deposits may not be relevant to the source areas for their contained minerals. The relevant depositional histories might include (i) fluvial erosion from distant source areas and sediment transport, (ii) in situ precipitation from solutions in lakes, or (iii) reactions of amorphous material in response to in situ conditions. Current deposits may be also evidence that conditions remained stable long enough for the minerals to remain inert in the particular environment where they were discovered. Alternatively, they may have been exposed to the surface only recently. Linking the recent discovery of clays and their geographical location with ancient surface conditions and global processes is, therefore, not straightforward. Their relation to an early warm and wet climate must be assessed together with morphological evidence on an individual basis.

Sulfate deposits have been identified from orbit (Bibring et al., 2006; Gendrin et al., 2005) and from the ground by both MER rovers (Arvidson et al., 2006; Grotzinger et al., 2005; McLennan et al., 2005; Squyres et al., 2004; Tosca et al., 2005). Their abundance and distribution represent an arguable case for widespread evaporitic environments involving surface to near subsurface water processes extending into the Hesperian. Their formation is interpreted to be the result of sulfur and water release during volcanic activity, production of H2SO4 from oxidation of sulfur in the atmosphere, and precipitation (Bibring et al., 2006). Liquid water circulation at the surface could have been enabled by volcanism-induced greenhouse effects through sulfur dioxide climate feedback (Haveli et al., 2007), while increased regional heat flow might have facilitated subsurface water circulation. By raising temperature, sulfur dioxide climate feedback may have contributed to the wet deposition of sulfur through precipitation (e.g., Fujita et al., 2003). Climate models predict the likelihood of snow precipitation at the equator and in the highlands (Forget et al., 2006; Haberle et al., 2001), and these results are consistent with regional studies that combine geological evidence and modeling (Baker et al., 1991; Carr, 2003; Clifford and Parker, 2001; Gulick et al., 1997; Haberle et al., 2001; Moore et al., 1995) and with studies of runoff production rates that imply rainfall (Irwin et al., 2005).

1.2.2 Clues from morphology

The proposition that lakes existed on ancient Mars finds key support from the physical connection of various paleolakes to channels and valley networks (Figure 1.4). Although, the contribution of volcanism and impact cratering to the Martian hydrographic system is well-supported and generally accepted as evidence for local to regional water release, both are climate-independent. Their only causal relationship to climate is their potential to temporarily alter it. In contrast, valley networks and the deltaic termini of channels are central to the issue of climate on early Mars.

Figure 1.4 Channel entering the Jezero impact crater (18.9°N, 77.5°E) in the Nili Fossae region to form a delta. Topographic correlation between delta and terraces in the basin argues for the presence of a body of water. THEMIS image V16660006, NASA/JPL/Arizona State University.

Valley networks

The origin, morphometry, ages of (re)activation, and duration of valley networks have been the subject of numerous studies reviewed in Chapters, 2 and, 8 (and references therein), whose conclusions have significant implications for the formation of lakes and their duration. The immaturity of the Martian hydrographic system is a key observation (Aharonson et al., 2002; Baker and Partridge, 1986; Irwin et al., 2002). With Viking, valley networks were found to be diffuse (Pieri, 1980), with irregular tributary junction angles (Cabrol, 1991; 1993) and large undissected intervalley areas showing no compelling evidence that precipitation were involved (Carr, 1996; Malin and Carr, 1999). Instead, their formation was linked to headward extension due to basal sapping (Pieri, 1980).

At high resolution, lower tributary orders are observed and networks appear more integrated. Drainage systems remain, however, immature if formed by runoff (Craddock and Howard, 2002; Grant, 2000; Gulick, 2001; Irwin et al., 2005) with possible exceptions (Hynek and Phillips, 2003). The lack of tributaries smaller than 100 m across have been attributed to either a lack of precipitation, or a lack of runoff due to high infiltration rates. It has also been suggested that non-fluvial erosive or depositional processes might have obliterated all valleys up to a maximum width (Hartmann and Neukum, 2001). High erosion rates (Craddock and Howard, 2002) and the occurrence of valley formation during heavy bombardment are consistent with both a greater supply of water and with warmer conditions, but these deviations from the present state did not have to be drastic (Chapter 2). An important clue may have been found by CRISM in the relative paucity of kaolinite (Mustard et al., 2008), which points toward limited hydrologic activity through space and time, and burial of deposits that have been shielded from further precipitation-related aqueous alterations.

Hecht (2002) showed that flow would have been possible on Mars at nearly freezing temperature. Conditions supporting hydrologic and limnologic activity in a cold Mars environment have been investigated through field analogs in polar regions (Doran et al., 2010), and at high altitude (Cabrol et al., 2009; Chapter 13) using Terrestrial and Martian data. Ice-covered streams and lakes can survive over extended periods of times (Morgan and Head, 2009). Ice provides thermal insulation from outside temperature fluctuations and protection from evaporation (McKay et al., 2005; Wallace and Sagan, 1979). Subglacial rivers are sustained in non-turbulent flows, until ablation removes the ice or until flow stops. Eutectic solutions of single salts with modest freezing point depression such as NaCl could allow sporadic liquid water flow anywhere on Mars (e.g., Clark et al., 2005; Grimm and Stillman, 2008; Hecht, 2002; Hecht et al., 2008). In that respect, the Phoenix mission (Smith et al., 2009) has shown physical and thermodynamical evidence that such brine solutions can form and remain liquid mostly anywhere even under current conditions (Fairén et al., 2009; Hecht et al., 2008; Renno et al., 2009; Seinfeld and Pandis, 2006). Snow precipitation (Smith et al., 2009) was also observed for the first time on Mars by Phoenix.

Deltas

Climatic and environmental conditions in the watershed affect discharge, erosion, transport, and deposition. Terminal basins are particularly sensitive to those changes and record them in their sedimentary deposits. Before MGS, large sediment accumulation at the termini of channels and valley networks provided a rare opportunity to evaluate their origin and the climate conditions associated with their formation. Case studies concluded that many of the deposits were short-lived, possibly reactivated, alluvial fans (Moore and Howard, 2005). Other works at global to regional scale proposed the existence of deltas, some interpreted to be inherited from warm and wet conditions (De Hon, 1992; Di Achille et al., 2006; Fasset and Head, 2005; Irwin et al., 2005; Kleinhans et al., 2009; Malin and Edgett, 2003; Mangold and Ansan, 2006; Ori et al., 2000) and others from cold environments (Kraal et al., 2008, and Chapter 8). Mars has seasons and an atmosphere, thus a latitude and elevation-dependent climate much like Earth, accentuated by high-obliquity cycles. Therefore, a wide range of morphology is actually expected and may reflect geographical location and local to global environmental conditions at a given time.

Comparison with Earth shows that vast and complex modern deltas can be constructed in cold environments. Those forming in cold continental to arctic conditions commonly have winter temperatures as low as –70°C, fluctuating between –40°C and –30°C in January and rising up to +10°C in July (Berezovskaya et al., 2005). This range is not too dissimilar to current-day Mars in the equatorial regions or to the temperatures in the intertropical belt during summer.

Delta development is affected by climatic, hydrologic, and geologic factors and can last from tens to thousands of years. Regardless of climate conditions, then, the lifetime of a delta represents only a small fraction of the Noachian’s half-billion-year duration. Over such vast period, the Martian deltaic record (not alluvial fans) appears limited to a few tens of landforms with relatively small areal development compared to their terrestrial counterparts (Chapter 8). Their original number, as that of valley networks, may have been affected by burial, impact gardening, and erosion but this remains speculative. If they are representative of the original record, this is a significant observation for the interpretation of the early Martian environment. This suggests either (i) short-lived favorable conditions or (ii) localized, accidental water releases. Stepped delta morphology recognized recently on Mars (Kraal et al., 2008) requires only decades to form and can result from a large single discharge (Figure 1.5). On the other hand, Gilbert deltas (Ori et al., 2000) can form over thousands of years. Their duration is discharge-dependent and may indicate favorable conditions either continuously or episodically present for a total of a few thousands of years.

Figure 1.5 This stepped delta at 8°S, 159°E has been interpreted as evidence of a single basin-filling event that lasted about 10 years (Kraal et al., 2008). CTX image P02_001644_17B, NASA/JPL/Arizona State University.

1.2.3 Outstanding questions

Ancient lake deposits are similar to time-lapse cameras integrating snapshots of conditions prevailing during their lifetime. In lakes dominated by overland flow from precipitation, this record captures, to a large extent, any global-scale changes in climatic conditions that occur while the lake is active. For limnologic systems essentially relying on groundwater, linking lake water volume, mineralogical deposits, and climate becomes a more complicated proposition.

Long groundwater flow paths

This complexity is illustrated by lakes in the Central Andes (4300–6000 m elevation), which is one of the regions exposed to some of the most rapid and severe climate change on Earth (Giese et al., 2002; IPCC, 2001; Montecinos and Aceituno, 2003; Warren et al., 2006; see also Chapter 13). Temperature has increased between 0.013°C/year and 0.02°C/year from 1932 to 1992 (Rosenblüth et al., 1995) while precipitation rates have fallen 50–75%, mainly between 1979 and 1988 (Legates and Willmott, 1990). Temperature increases are shifting the equilibrium line altitude of glaciers upward. As a result, accumulation areas are shrinking and/or the net accumulation is becoming progressively smaller (Bradley et al., 2006). Lakes between 18°S and 24°S reached high stands in the past 12,500 to 11,000 years (Grosjean et al., 2001) following the glacial retreat at the end of Pleistocene, when they were supplied by glacial melt, 500 mm/year precipitation, groundwater, and hydrothermal activity. Isotopic studies (Aravena, 1995) show that the last two major aquifer recharges occurred between the last deglaciation and the onset of arid conditions with the beginning of the Holocene. Precipitation in the south Bolivian and northern Chilean altiplano for the past 8 years has fluctuated between 30 mm/year and 120 mm/year (Cabrol et al., 2009) with a water balance of –1500 mm/year (evaporation/precipitation) for many lakes (Hock, 2008). However, the effect of climate and precipitation regimes is buffered in other local lakes dominated by long-distance groundwater flow paths from 12,500-year-old aquifers (Nester et al., 2007; Valero-Garcés et al., 2004). In their case, precipitation regimes lack relevance to understanding present fluctuations. Such residence times are not uncommon for terrestrial aquifers (Kazemi et al., 2006), but more knowledge of prior Martian climates and subsurface aquifer parameters will be necessary before we can fully understand their impact on