The Coevolution of Life and Environment on Mars: An Ecosystem Perspective on the Robotic Exploration of Biosignatures.

Earth's biological and environmental evolution are intertwined and inseparable. This coevolution has become a fundamental concept in astrobiology and is key to the search for life beyond our planet. In the case of Mars, whether a coevolution took place is unknown, but analyzing the factors at play shows the uniqueness of each planetary experiment regardless of similarities. Early Earth and early Mars shared traits. However, biological processes on Mars, if any, would have had to proceed within the distinctive context of an irreversible atmospheric collapse, greater climate variability, and specific planetary characteristics. In that, Mars is an important test bed for comparing the effects of a unique set of spatiotemporal changes on an Earth-like, yet different, planet. Many questions remain unanswered about Mars' early environment. Nevertheless, existing data sets provide a foundation for an intellectual framework where notional coevolution models can be explored. In this framework, the focus is shifted from planetary-scale habitability to the prospect of habitats, microbial ecotones, pathways to biological dispersal, biomass repositories, and their meaning for exploration. Critically, as we search for biosignatures, this focus demonstrates the importance of starting to think of early Mars as a biosphere and vigorously integrating an ecosystem approach to landing site selection and exploration. Key Words: Astrobiology-Biosignatures-Coevolution of Earth and life-Mars. Astrobiology 18, 1-27.

1. Introduction

By the end of this decade, missions will set out to search for biosignatures on Mars (e.g., Meyer and Schulte, 2014; Beegle et al., 2016; Vago et al., 2006, 2017). Yet despite major advances in our understanding of the red planet, the intellectual framework underpinning this new chapter of exploration remains fundamentally the same as the one that guided the characterization of Mars' habitability (e.g., Mustard et al., 2013). Meanwhile, knowledge gaps about the sustainability of habitable conditions (e.g., Ehlmann et al., 2016) severely constrain our ability to evaluate Mars' past and present ecological potential and the possible interactions between environment and life. Critically, these gaps alter our perspective on key questions, such as where, what, and how to search for biosignatures.

These questions are foundational to all exploration. On Mars, they guide science operations and structure the landing site selection process. For Mars 2020 and ExoMars, candidate landing sites must demonstrate physical characteristics consistent with past habitability for early life as we know it, and favorable conditions for the preservation of biosignatures (e.g., Mustard et al., 2013; Vago et al., 2017), along with meeting safety and accessibility criteria. However, implicit assumptions underlie this approach.

One assumption relates to an origin of life and its similarity with terrestrial biology. Past habitability (e.g., Knoll et al., 2005; Des Marais, 2010; Grotzinger et al., 2015a, 2015b) and the discovery of organic chemistry (e.g., Chastain and Chevrier, 2007; Mumma et al., 2009; Freissinet et al., 2015; Webster et al., 2015) make this assumption a viable working hypothesis, and while organic molecules can form from cosmic and geological processes, biology cannot be ruled out (e.g., Mumma et al., 2009; Blamey et al., 2015).

The other assumption relates to biosignatures (e.g., Summons et al., 2011; Hays et al., 2017), which in addition to preservation potential would require that life would have evolved locally or migrated to the proposed sites from its environments of origin. However, causality in this case is not straightforward and currently is strictly inferred from early Earth's environmental history. In effect, the mechanisms or processes by which the early martian environment might have allowed life to evolve and/or migrate from its points of origin to new habitats, that is, biological dispersal (e.g., Nemergut et al., 2013), are barely acknowledged in current exploration strategy or planning, either theoretically or logistically. Early life's evolution on Earth points toward dispersal pathways intimately connected with all aspects of the environment and interactions engaging the lithosphere (including geography and plate tectonics), the atmosphere, the hydrosphere, and their spatiotemporal changes (e.g., Womack et al., 2010; Clobert et al., 2012; Buoro and Carlson, 2014).

On Mars, it is unknown whether life arose (e.g., Jakosky and Shock, 1998; Brack et al., 1999; Brack, 2000; Schulze-Makuch et al., 2005; Jepsen et al., 2007; McKay, 2010; Cockell et al., 2012; Westall et al., 2013; Retallack, 2014; Levin and Straat, 2016; Catling and Kasting, 2017). But if it did, it is expected that the planet's own specific characteristics and environment (e.g., the global dichotomy, tectonic style—two-plate planet or no plate tectonics, and the debated existence of oceans) would have played an equally important, if not overarching, role in the origin and evolution of any indigenous biology; and while early Mars was Earth-like, it was never exactly like Earth.

How much these differences would have affected the evolution of an early biosphere and subsequent dispersal pathways has yet to be evaluated, but as we embark on a search for biosignatures, our vision is still very much anchored in a reductionist approach, where diverse data sets are juxtaposed in snapshots defined by their geological ages. What has yet to happen is to think of Mars in terms of a possible coevolution of life and environment, where data are synergistically pulled together in a dynamic vision of an ecosystem. The merit of such a perspective is to give consideration to what the characteristics of an early martian biosphere, if any, would have been like. Performing this exercise to the best of our knowledge, as limited as it may be today, is not only important at this stage but necessary to the development of a more systemic, substantive, and informed guiding principle about where to search (potential exploration sites), what to search for (types of habitats and biosignatures), and how to explore (scales, resolution, instruments, and detection thresholds).

A number of studies have envisioned spatiotemporal scenarios that lead to the definition of habitable environments, microorganisms, metabolisms, and plausible fossil records on Mars (e.g., McKay and Stoker, 1989; Brack et al., 1999; Fairén et al., 2005, 2010; Schulze-Makuch et al., 2005, 2008; Anand et al., 2006; Knoll and Grotzinger, 2006; Gómez et al., 2012; Westall et al., 2013; Davila and Schulze-Makuch, 2016; Ehlmann et al., 2016). Here, we center the focus on environmental pathways through which life could have dispersed over time on Mars and their significance for biomass repositories and biosignature exploration.

In the following sections, we examine Mars through the lens of the coevolution of life and environment as we understand it from early Earth (Fig. 1). We first consider how spatiotemporal scales and spectral resolution affect our perception of this synergy, but also how they generate some confusion between the notions of planetary habitability and habitat, which in turn affects our perspective on detection thresholds. Then, we explore hypothetical concepts of coevolution, where the factors of a coevolution include what is known today of life's origin and early habitats on Earth—and their timing—projected within the distinct environmental and physical evolution of early Mars. This approach considers the recent results from the Mars Atmosphere and Volatile EvolutioN (MAVEN) mission (e.g., Jakosky et al., 2017) and new evidence that Mars was volcanically active well into the Late Amazonian (references herein). From this comparison, a broad ecological framework emerges, which includes blueprints for early habitats and plausible dispersal pathways. Among the pathways identified, some suggest that differences between the two planets may not have mattered in the first several hundred million years, whereas other pathways would have exposed early life to the pressure of a rapidly declining planetary habitability. Yet others may have sheltered and, thus, sustained life since its origin despite surface changes. All have implications for the distribution and preservation of biosignatures. The final section is, therefore, devoted to science focus, detection thresholds, and exploration strategies.

Framework for a coevolution on Mars—uncertainties about the origin(s) of life on Earth and the early martian environment, and the use of Earth's chronology for early life in the context of Mars, limit any notional coevolution model. Here, the variables are the plausible environments of origin(s) for life on Mars in view of terrestrial theories and what is known of the evolution of habitability on Mars. The fix parameter is life as we know it, with the chronology provided by discoveries made on Earth to date. We assume a timing for the earliest evidence of life consistent with Earth at 4.28 Ga (references in the text) and focus on early life colonization and dispersal pathways from these environments of origin.

2. The Three Earths and the Three Mars

Environmental changes have accompanied life since its inception on Earth either as causes or effects (e.g., Knoll, 2009). Assuming life is active, and/or biosignatures preserved, any planet or moon that develops a biosphere will bear the signatures of its own coevolution as embedded data that sum up all the synergies between life and environment since their formation, and their relative dominance over time. This information may be neither obvious nor equally accessible from one world to the next. Our perception of these synergies is driven by the spatiotemporal scales and resolution used to examine them, and by whether they are analyzed holistically or as discrete sets of variables. Depending on the prism through which we observe them, they either reveal a complex continuum or only tipping points.

For instance, over geological scales and at low spatiotemporal resolution, Earth may be compared to a multitude of distinct planets, alternatively showing similarities with water worlds recently discovered (e.g., Burnham and Berry, 2017): Titan (e.g., Trainer et al., 2006) or Enceladus (e.g., Kargel, 2006). Similarly, the biological Earth can be roughly reduced to three different planets: the prokaryotic Earth of the Archean (>2.5 Ga), the eukaryotic Earth of the Proterozoic (2.5–0.542 Ga), or the Cambrian Earth (<0.542 Ga) with the rise of most major animal phyla.

Although accurate, this reductionist view of Earth's environmental or biological evolution is essentially a caricature, a series of snapshots that gives little information on how these Earths, apparently so distinct from each other, are in fact woven seamlessly together through coevolution. One planetary environment leads to the next through uninterrupted internal feedback loops and external forcing, where every single moment combines all the systemic history that preceded. Comprehending the spatiotemporal complexity of the coevolution of life and environment (i.e., what makes a planet or an environment habitable); how habitats are created and transformed over time; and how life adapts, survives, and changes its environment requires the integration of multiscale data sets from a diversity of scientific fields and an understanding of their relative significance. Obviously, the more gaps in the data, the more difficult the reconstruction of this complexity becomes, and in essence, beyond Earth, this is the primary challenge faced by planetary exploration.

The exploration of Mars best exemplifies this issue. Over the past decades, an integrated program of missions has characterized its geological evolution (e.g., Albee et al., 2001; Chicarro, 2006), past and present climate (e.g., Inada et al., 2007; Jakosky et al., 2017), and ancient habitable environments (e.g., Knoll et al., 2005; Des Marais, 2010; Grotzinger, 2014; Grotzinger et al., 2014). Searching for biosignatures is thus a next logical step, but transitioning from the characterization of ancient habitable conditions to the search for biosignatures requires that, accordingly, we shift the focus of exploration from planetary habitability to the identification of habitats. Although these two notions are often confused outside ecology, they are fundamentally distinct. Habitability designates the general conditions favorable to life at local to global scales (e.g., water, energy, nutrients, shelter), whereas habitats are defined by their inhabitants and, for Mars, focus on potential microorganisms and their spatiotemporal dynamics.

Searching for early signs of life on Mars may involve, at best, the equivalent of exploring the Archean record (e.g., Farmer and Des Marais, 1999; Cady et al., 2003; Schopf, 2006; Summons et al., 2011) with an understanding of the early environment that remains elementary and fundamental questions that are unanswered (e.g., Ehlmann et al., 2016). This understanding relies chiefly on orbital data with spectral range and spatial resolution that have allowed the identification of three distinct geological periods on Mars based on dominant mineral alterations: the Phyllocian (4.5–4.0 Ga), Theiikian (4.0–3.5 Ga), and Siderikan (3.5 Ga to present) (Bibring et al., 2005, 2006). This timescale broadly overlaps with that of impact crater densities and major geological and environmental changes (e.g., Tanaka, 1986), including the Noachian (4.1–3.7 Ga), which is characterized by the heavy bombardment and valley network formation; the Hesperian (3.7–3.1 Ga), which gave rise to increasing aridity, intense volcanism, the formation of outflows, and possibly an ocean; and the Amazonian (3.1 Ga to the present), a time when Mars became the planet we observe today, with residual flood, glacial, and volcanic activity.

In addition to these geological periods, Pre-Noachian Mars (4.5–4.1 Ga) marks the accretion time, the formation of the global dichotomy, the loss of the magnetic field, and changes in the atmosphere (see Section 3 and references therein). Little record is left of this period (e.g., Frey, 2006), but if the chronology of life's origin and early coevolution on Earth has any bearing on Mars, these events alone make the Pre-Noachian a key period in Mars' history given that it established unique initial conditions for a coevolution.

Our knowledge also relies on high-resolution data acquired at Gusev and Gale Craters and Meridiani Planum. From an environmental standpoint, these data are consistent from one site to the next and support the view of a globally declining habitability. In contrast, because of this decline, extrapolating about microbial habitats at scales broader than the individual site should be done carefully, while considering the age of the site and the potential interactions between habitats and environment (e.g., surface, subsurface, deep underground). Within 700 million years of accretion, the surface of Mars had become hostile for life as we know it (e.g., Jakosky et al., 2017), and terrestrial analogues show that local (meter to micrometer scale), not global, factors dominate the evolution of microbial habitats in such environments (e.g., Amils et al., 2007; Cabrol et al., 2007a; Warren-Rhodes et al., 2007a, 2007b; Preston and Dartnell, 2014); see Sections 4–6.

Through the current low-resolution prism, our vision of the evolution of Mars is thus as much of a caricature as the three Earths. As a result, the intellectual framework guiding the search for biosignatures suffers from an internal conflict, whereby the scientific focus (mission goal) is lost to background noise (insufficient data resolution), and the important details (framed hypotheses on microbial ecosystems) are lost to generalization (planetary habitability). However, if current data are not well adapted to characterize Mars at the microbial habitat level yet, they can still help us articulate science questions about early Mars from a different and ecologically dynamic perspective. Importantly, they can support the development of working models and better decision-making tools on whether the types of sites we deem a priority (e.g., Hays et al., 2017) were likely (or not) to produce habitats and preserve biosignatures in Mars-specific sets of environmental circumstances, that is, rather than simply inferring their potential from a strict terrestrial ecosystem analogy.

Examples of reframed science questions include the following:

(1) What role did environmental differences between Earth and Mars play in an early evolution of life on Mars? (2) What was the impact of the unique physical features (e.g., global dichotomy, high obliquity, lost magnetosphere and atmosphere, volcanic and tectonic characteristics) on the formation and spatiotemporal evolution of environmental pathways to biological dispersal? (3) What does a comparison between the timing of early life evolution on Earth and the current environmental models for early Mars suggest about ancient habitable environments, their habitat development potential, biological dispersal, biosignature preservation, and detection thresholds?

The framework defined by these questions allows the exploration of notional models of environmental pathways, which ultimately helps narrow down plausible biomass repositories. The discussion that follows involves geological and environmental thresholds; therefore, the crater density timescale (Noachian, Hesperian, and Amazonian) is being used.

3. Environments for an Origin of Life

How we search for life beyond our planet is guided, and possibly limited, by what we know from Earth, and many questions remain unanswered today about the origin(s) of life. Simple organic compounds may have formed (references herein) and/or were delivered by comets and asteroids (e.g., Chyba and Sagan, 1992; Ehrenfreund and Cami, 2010; Zahnle et al., 2010; Santosh et al., 2017). In the Oparin-Haldane hypothesis, life arose gradually from inorganic molecules in a reducing atmosphere where they reacted with lightning and/or the energy of the Sun to produce amino acids and nucleotides. Over time, proteins and nucleic acids formed into self-sustaining biomolecules capable of replicating (e.g., Armstrong and Brasier, 2004). The test of the Oparin-Haldane hypothesis (e.g., Miller and Urey, 1959) demonstrated that at least some of the building blocks of life could form from simple inorganic compounds in a broad range of environments (e.g., Gilbert, 1986; Lazcano and Miller, 1996; Shapiro, 2000; McCollom, 2013).

3.1. Earth

Theories about the transition of prebiotic chemistry to biology include environments such as deep sea vents (e.g., Baross and Hoffman, 1985; Shock, 1992; Visscher 1993; Wächtershäuser, 2006; Martin et al., 2008; Sleep, 2010; Nisbet and Fowler, 2013; Dodd et al., 2017) on an ocean planet with a few landmasses only (e.g., Cavosie et al., 2007; Ushikubo et al., 2008; Burnham and Berry, 2017); the surface of the sea (e.g., Michaelian, 2011); tidal pools (e.g., Lasaga et al., 1971; Lahav et al., 1978); hydrothermal systems and hot springs (e.g., Shock 1993; Djokic et al., 2017); mineral-water-air interphases (e.g., Hazen, 2012; Yang et al., 2013); and volcanic aquifers (e.g., Washington, 2000). Plate tectonics is viewed as essential to a terrestrial “habitability trinity” defined by the coexistence of ocean, landmass, and atmosphere (e.g., Bercovici and Ricard, 2014; Dohm and Maruyama, 2015; Azuma et al., 2017). It controls rock cycling; the formation and destruction of continents; and the impact of these processes on mineral systems, the environment, and the production of nutrients for life (e.g., Maruyama et al., 2007; Santosh, 2010; Pirajno, 2015; Santosh et al., 2017). Exactly when plate tectonics started is debated (e.g., Condie and Kröner, 2008), with estimates ranging from the early Neoproterozoic (e.g., Stern, 2005) to soon after the formation of the primordial ocean in the Hadean (e.g., Korenaga, 2013). Nutrients may have been produced on primordial continents on a dry Earth after the solidification of the magma ocean, volatiles accreted through impact processes, and conditions occurred that were conducive to the inception of life between 4.37 and 4.20 Ga (e.g., Santosh et al., 2017).

Until recently, the earliest indirect evidence of biological activity was the 3.8–3.7 Ga old metamorphosed sedimentary rocks and minerals from the Isua supracrustal belt in southwest Greenland (e.g., Van Zuilen et al., 2003; Rosing and Frei, 2004). Biogenic carbon may be present in 4.1 Ga rocks from Jack Hills, Western Australia (e.g., Bell et al., 2015), and biogenic structures as old as 4.28 Ga in ancient oceanic floor material of the Nuvvuagittup supracrustal belt, Canada (e.g., Dodd et al., 2017). Metacarbonate rocks in the Isua supracrustal belt contain stromatolites (e.g., Nutman et al., 2016) that developed in a shallow marine environment (3.7 Ga), predating by 220 million years the oldest fossils of the Dresser Formation in Pilbara, Australia (e.g. Walter et al., 1980; Van Kranendonk, 2011; Noffke et al., 2013). Seafloor hydrothermal vents, shallow marine continental margins, and intertidal environments have thus produced the oldest evidence of early life to date (e.g., Klein et al., 1987; Noffke et al., 2006; Westall et al., 2006; Schopf et al., 2007; Noffke, 2009; Van Kranendonk, 2011). Microfossils, microbialites, isotopic and molecular biomarkers also show that prokaryotic life was abundant in these environments ∼3.5–3.4 Ga (e.g., Beraldi-Campesi, 2013). Hot spring pool and vent deposits constitute the earliest evidence (3.5 Ga) of life on land (e.g., Djokic et al., 2017). These biosignatures are close in age to the Strelley Pool fossil record (3.43 Ga) that formed in a shallow sea environment (e.g., Hofmann et al., 1999).

3.2. Mars

The nature of the martian climate and environment during the same time frame (4.3–3.4 Ga) is one of the long-standing questions and outstanding issues having a direct bearing on a coevolution (e.g., Ehlmann et al., 2016; Wordsworth, 2016). We summarize in Table 1 (and references therein) the main theories as they relate to the environmental context for our discussion, and refer to recent reviews for more details.

1.

Mars Early Evolution—Earth-Like, But… (Almost) All Things Being Unequal

	Time (Ga)	Events/Models and implications
Atmosphere	4–5/4.2	Sun T-Tauri stage. No thick primary CO2 atmosphere[1].
	4.3	Catastrophic outgassing of volatiles[2–4].
	4.2–3.8	Secondary CO2 atmosphere from volcanic activity[5] and heavy bombardment[6].
	4.2–3.8	Decrease in solar activity and soft X-ray flux[1].
	4.1–Present	Loss of atmosphere from sputtering by solar wind and radiation[7–8].
Magnetosphere	>4.1	Loss of magnetosphere[9]. Lack of magnetization in Hellas. Termination of magnetic field before the end of the valley network formation[10].
		Differences in magnetic properties between both hemispheres due to the formation of the global dichotomy (see below), or an original single-sphere dynamo[11] from nonuniform temperature across the core-mantle boundary[12].
Global dichotomy	4.3?	Its formation affects geography, climate, hydrology, geology[13], and dispersal pathways for life. Externally driven models: A megaimpact or multiple impacts[14–18] or a southern polar giant impact by a bolide 0.1–1.0 lunar mass[19]. The resulting magma ocean solidifies to form the thicker crust of the southern hemisphere[20].
		Internally driven models: Removal of the basal lowland crust by mantle convection[21–23] or plate tectonics[24].
		The global dichotomy forms during accretion or right after[25], making it the most ancient geological feature of Mars[26–27].
Impact basins	4.1–3.9	Most large impact basins form during the Noachian (Hellas, Argyre, Isidis, Utopia, other)[27–29].
Structure Volcanism	3.8–3.5	Valles Marineris develops and is subsequently modified by erosion and sedimentation into the Amazonian[30–31]. Volcanic activity into recent geological past[32].
	3.8	The rise of Tharsis generates rifting in Valles Marineris and Noctis Labyrinthus.
	3.4–3.1	Elysium develops[33–34], with activity into the Amazonian[35].
Hydrology	4.3–3.8	Hydrological provinces and flow directions structurally defined early with the formation of the global dichotomy and the heavy bombardment[36].
	4.0–3.7	Peak of valley networks [27,37–39] and lake formation[40–42]. Localized activity in the Amazonian. Networks are not fully integrated with the landscape, suggesting only short favorable conditions for their development[39–47].
	3.7>	Outflow channels. Peak formation in the Hesperian with residual activity in the Amazonian[27].
	4.1> (?)	Ocean in the first billion years of history from subpermafrost aquifer modeling[48] and from proposed geomorphic evidence[49–51]. Disputed model[52].
	3.7>	Oceanus Borealis (Hesperian ocean)[53–56]. Cons: Difficulty to reconcile the volumes required to form the outflow channels with known volatile sources and sinks[27], lack of fine-grained sediments[57], evaporites[58], carbonates[59–60], no evidence for Noachian glaciation[61], and already thin Noachian atmosphere[62–63]. Pros: Lack of carbonates explained by destruction by precipitated H2SO4[64], a globally acidic atmosphere[65], or the short lifetime of ocean episodes[54].
		Other potential evidence: The global distribution of Hesperian deltas[66], possibly tsunami deposits[67–69], and shorelines[49–51,70–73]. The existence of shorelines has been debated since originally proposed[53,74]. The water global equivalent layer (GEL) has been used as a pro or con argument depending on models: D/H enrichment factors argue for a large Pre-Noachian water loss (GEL ≥137 m)[75]. Ion-microprobe analysis implies a Pre-Noachian water loss greater than during the rest of the martian history (GEL >41–99 m vs. GEL >10–53 m) and an undetected subsurface water/ice reservoir (∼100–1000 m GEL) that exceeds the current observable inventory of ∼20–30 m GEL)[76–77]. A recent global reevaluation of eroded volumes of valley networks results in an even larger GEL of ∼5 km[78]. Other potential evidence includes polygonal ground in the lowlands[79] and subdued impact craters, whose anomalous depth/diameter ratio could be associated with a Late Hesperian ocean episode.[53,80–83]
Climate and environmental change	4.1>	Episodic periods of warmer and thicker atmosphere after the loss of the magnetosphere contributed by orbital forcing[84–85] and volcanism. Sulfur[86] and sulfates[58,87–88] suggest the release of greenhouse gases[89–92]. Other warming mechanisms (≤10 million years) include carbonate-silicate cycles[93]. Impact cratering provided transient environmental/atmospheric changes[94].
		Excursions of warmer periods on a cold early Mars are supported by the presence of phyllosilicates[95–99] and increasingly more arid conditions from the Hesperian on, with an abundance of sulfates and evaporites[58]. Significant variations in obliquity, eccentricity, and precession[84] have also led to a redistribution of water in the polar ice deposits to lower latitudes to create ice ages regularly during martian history[100].

References in the table: 1Erkaev et al., 2014; 2Elkins-Tanton, 2008, 2011; 3Tian et al., 2009; 4Brasser, 2013; 5Grott et al., 2011; 6Alexander et al., 2012; 7Jakosky and Phillips, 2001; 8Jakosky et al., 2017; 9Acuña et al., 1999; 10Fassett and Head, 2011; 11Stanley et al., 2008; 12Zhong and Zuber, 2001; 13Watters et al., 2007; 14Frey and Schultz, 1988; 15Wilhelms and Squyres, 1984; 16Andrews-Hanna et al., 2008; 17Marinova et al., 2008; 18Nimmo et al., 2008; 19Leone et al., 2014; 20Reese et al., 2011; 21Lingenfelter and Schubert, 1973; 22Wise et al., 1979; 23McGill and Dimitriou, 1990; 24Sleep, 1994; 25Frey, 2003; 26Solomon et al., 2005; 27Carr and Head, 2010; 28Frey, 2008; 29Andrews-Hanna and Zuber, 2010; 30Lucchitta, 2010; 31Watkins et al., 2015; 32Broz et al., 2017; 33Werner, 2009; 34Robbins et al., 2011; 35Greeley and Spudis, 1981; 36de Hon, 2010; 37Carr, 1996; 38Gulick and Baker, 1989; 39Craddock and Howard, 2002; 40Cabrol and Grin, 1999; 41Cabrol and Grin, 2010; 42Fassett and Head, 2008; 43Hynek and Phillips, 2003; 44Stepinski and Collier, 2004; 45Howard et al., 2005; 46Carr, 2012; 47Craddock and Lorenz, 2017; 48Clifford and Parker, 2001; 49Parker et al., 1989; 50Parker et al., 1993; 51Head et al., 1998; 52Carr and Head, 2003; 53Baker, 2001; 54Dohm et al., 2001; 55Dohm et al., 2009b; 56Świąder, 2014; 57McEwen et al., 2007; 58Bibring et al., 2006; 59Ehlmann et al., 2008; 60Wray et al., 2009; 61Wordsworth, 2016; 62Hu et al., 2015; 63Kite et al., 2014; 64Levine and Summers, 2008; 65Greenwood and Blake, 2006; 66Di Achille and Hynek, 2010; 67Rodriguez et al., 2015; 68Rodriguez et al., 2016; 69Costard et al., 2017; 70Thompson and Head, 2001; 71Perron et al., 2007; 72Head, 2007; 73Banfield et al., 2015; 74Malin and Edgett, 1999; 75Villanueva et al., 2015; 76Kurokawa et al., 2014; 77Stuurman et al., 2016; 78Luo et al., 2017; 79Kargel et al., 1995; 80Boyce et al., 2005; 81Mouginot et al., 2012; 82Gulick et al., 1997; 83Kreslavsky and Head, 2002; 84Laskar et al., 2002; 85Mischna et al., 2003; 86Clark et al., 1976; 87Gendrin et al., 2005; 88Ehlmann and Edwards, 2014; 89Postawko and Kuhn 1986; 90Yung et al., 1997; 91Halevy and Head, 2014; 92Ramirez et al., 2013; 93Batalha et al., 2016; 94Segura et al., 2002; 95Carter et al., 2013; 96Poulet et al., 2008; 97Bishop et al., 2005; 98Fairén et al., 2010; 99Head et al., 2003; 100Head et al., 2010.

While an origin of life around deep sea vents is a leading theory for Earth, the existence of a Pre-Noachian to early Noachian primordial ocean is still in question due to the scarcity of the geological record (e.g., Carr and Head, 2003; Frey, 2003, 2006). Its existence and extent hinge on climate models, estimates of volatile sources and sinks, disparate global equivalent layer (GEL) values (i.e., the thickness of an amount of water spread evenly over Mars), subpermafrost aquifer modeling, and proposed shorelines in Hellas (see Table 1 and references therein). It is discussed here as a theoretical environmental end-member.

If a primordial ocean was present in the Pre-Noachian, volcanism distributed globally since crustal formation would have supported the development of a pervasive Pre-Noachian subseafloor aquifer heated by magma. Hydrothermal activity would have been widespread, with energy concentrated in the equatorial regions (e.g., Leone et al., 2014, Table 1). Although decreasing over time, this energy remained available for later oceans that formed episodically across the Hesperian and Amazonian (e.g., Baker, 2001; Carr and Head, 2003; Dohm et al., 2009a, 2009b; Luo et al., 2017). Chemicals released in such environments on Earth set up redox reactions, which would have caused electrical currents to flow naturally around vents and enable early life to use free electrons to synthesize compounds necessary for its evolution (e.g., Roussel et al., 2008; Ishii et al., 2015). Pre-Noachian paleomagnetic stripes are consistent with proto plate tectonics processes in the cratered highlands (e.g., Connerney et al., 1999), and large-scale strike-slip faulting is being interpreted as evidence of a slow, still-evolving, two-plate planet system in Valles Marineris (e.g., Yin, 2012). If these interpretations stand, Pre-Noachian Mars had the elements of a “habitability trinity” (Dohm and Maruyama, 2015).

Pathways and exchanges between the ocean and the deep subseafloor have generated the largest reservoir of prokaryotic life on Earth (e.g., Whitman et al., 1998; Amend and Teske, 2005; Edwards et al., 2012; Jungbluth et al., 2016; Hoshino et al., 2017). The oceanic crustal aquifer represents about 2% of the ocean's volume (e.g., Edwards et al., 2011) and extends several hundreds of meters into deeply buried sediments (e.g., Parkes et al., 1994; Roussel et al., 2008). Thick sediment layers act as a barrier to seawater-basalt exchange in places, while in others, exposed rocky outcrops facilitate fluid exchanges between the deep ocean and the crustal aquifer (e.g., Meyer et al., 2016). The intensity of fluid circulation is a function of geological settings and is more pronounced in heat-driven fluid flow areas, such as deep sea vents (e.g., Fisher et al., 2007).

Dispersal pathways during, and post, primordial ocean times would have depended on global fluid circulation in rocks and groundwater migration, which both have climatic, physical, and chemical components. On Mars, strong cyclic climate swings (e.g., Laskar et al., 2002) would have impacted the long-term stability of unconfined and shallow confined aquifers and their physicochemical properties such as texture, pore size, nutrients, chemistry, which are critical to their ability to support microbial ecosystems and biodiversity (e.g., Ghiorse and Wilson, 1988).

The early formation of the global dichotomy (Table 1), and the theoretical presence of islands and landmasses in a primordial ocean scenario, gives life access to land as early as the Pre-Noachian through subaerial pathways (e.g., shallow ocean margin, coastal regions, streams) and subsurface and underground fluid circulation in a fractured regolith and crust (e.g., Edwards et al., 2012; Fisher et al., 2015). Away from a protective deep water column, Pre-Noachian subaerial habitats, if any, would have been, from that time on, directly exposed to the heavy bombardment, climate variability, a magnetosphere that shut down ∼4.1 Ga, and a dissipating atmosphere.

The existence of a Pre-Noachian ocean has more theoretical than empirical support, and there is no strong geomorphic or mineralogical evidence at current resolution for its survival late into the Noachian. This period corresponds to the heavy bombardment and the formation of the giant impact basins (e.g., Hellas, Argyre, Isidis) by very large asteroids (e.g., Carr and Head, 2010), and the Barberton greenstone belt in South Africa (3.5–3.2 Ga) may offer an analogy to what could have been the fate of a primordial ocean on Mars, with impacts boiling and vaporizing the ocean's top layer (e.g., Sleep and Lowe, 2014).

The presence of a primordial ocean may not have been necessary for an origin of life on Mars. Alternatively (or concurrently) to an ocean, volcanic aquifers have been proposed as robust settings for an origin of life on Earth, with a wide range of physicochemical conditions for the formation of prebiotic organic compounds and RNA (e.g., Washington, 2000). Contrary to a theoretical primordial ocean, there is ample evidence that volcanic aquifers formed on Mars (e.g., Boston et al., 1992; Farmer, 1996; Mouginis-Mark and Christensen, 2005; Cousins and Crawford, 2011; Rodríguez and van Bergen, 2015). Volcanism was globally active since crustal formation (4.3 Ga), with possible recent residual magmatic activity (e.g., Neukum et al., 2004; Broz et al., 2017). Massive volcanic constructs attest to repetitive eruptions over billions of years over hotspots (e.g., McEwen et al., 1999; Werner, 2009; Baratoux et al., 2013), these regions providing relative spatiotemporal durability to martian habitats (e.g., Schulze-Makuch et al., 2005; Westall et al., 2013).

Volcanic aquifers could have sustained long-term subsurface sheltered habitats against changing surface conditions after 4.1 Ga, with cycling of water, energy, and nutrients (e.g., Fairén et al., 2003; Schulze-Makuch et al., 2007), as well as transient subaerial habitats in hydrothermal springs, geysers, and pools (e.g., Walter and Des Marais, 1993; Cady and Farmer, 1996; Farmer, 1996; Walter, 1996; Gulick, 1998; Farmer and Des Marais, 1999; Cady et al., 2003; Allen and Oehler, 2008; Hofmann et al., 2008). Impact cratering produced an additional source of hydrothermal systems lasting up to tens of thousands of years for the largest impacts (e.g., Cabrol and Grin, 2001; Abramov and Kring, 2005; Newsom, 2010; Osinski and Pierazzo, 2012). On Earth, these environments are conducive to hypogene and supergene clay formation (e.g., Hazen et al., 2013). They are associated with weathering in acidic environments near active volcanoes, diagenesis, hydrothermal and other aqueous alterations (Table 1), and with the serpentinization of preexisting minerals (e.g., Schulte et al., 2006; Ehlmann et al., 2010; Chassefière et al., 2013; Greenberger et al., 2015). Serpentinization of basalt and other mafic igneous lithologies coupled with seawater-basalt interactions (if an ocean was present) would have provided an environment stimulating the confinement function for biomolecules and biochemical reactions in clay hydrogels (e.g., Yang et al., 2013).

Assuming these hypothetical environments as conducive to the emergence of life, Pre-Noachian biological dispersal pathways could have engaged

(a) deep crustal aquifers (hundreds to thousands of meters' depth) of volcanic and/or oceanic origin, maintained by the geothermal gradient at global scale and phreato-magmatic activity at regional scale. Unless disrupted by large impact events, these environments could have produced stable deep habitats over hundreds of millions to billions of years, with some possibly still sustained in recent geological times (e.g., Broz et al., 2017);

(b) an active subsurface (1 m to hundreds of meters' depth), with habitats connected by surface and subsurface fluid circulation, phreato-magmatic activity, and local surface/atmosphere interactions, such as topography and slope exposure, moisture penetration, and sediment texture and composition. Climate variability (e.g., obliquity cycles, freeze/thaw cycles and active layers) and stochastic events (e.g., impact cratering, volcanic eruptions) regularly affected habitat evolution (i.e., expansion, reduction, destruction) and connectivity networks; and

(c) an unstable surface environment with a high rate of habitat changes linked to climate variability, atmosphere/surface interactions, stochastic events (e.g., cosmic, environmental), and the state of the magnetosphere, atmosphere, and hydrosphere.

Recent missions have brought a greater focus on the various ways in which these environments would have been affected by early planetary changes (Sections 4–6). MAVEN's 38Ar/36Ar measurements show that Mars rapidly lost over 65% of its atmospheric argon through sputtering by pickup ions (e.g., Jakosky et al., 2017). The inferred loss of CO2 and O suggests an early partial pressure of 1 bar CO2 and a large early atmospheric loss from intense solar extreme UV radiation during the Sun T-Tauri stage (e.g., Rahmati et al., 2015; Airapetian et al., 2016). The strong decline in valley network formation by the end of the Noachian (e.g., Carr and Head, 2010) and changes in mineralogy (e.g., Bibring et al., 2006; Murchie et al., 2009) attest to this atmospheric transition.

4. The Noachian Paradox

Magnetosphere loss ∼4.1 Ga, atmospheric escape, frequent climate swings, and a distinct internal dynamic (Table 1) are the foundational parameters that would have differentiated a martian biosphere from Earth's very early and impacted the spatiotemporal evolution of its microbial biogeography in unique ways (e.g., Fierer, 2008). Species facing changing environments and habitat destabilization need functionally connected networks of habitats to adapt (e.g., Capone and Hutchins, 2003; Jansson and Taş, 2014; Zeglin, 2015), and the success of these networks relies on species that have sufficient dispersal ability to track suitable habitats. The mechanisms and extent of terrestrial microbial dispersal are still debated (e.g., Fierer, 2008; Nemergut et al., 2011, 2013); some propose that high abundance increases the probability of long travels by individuals and the colonization of remote locations by chance (e.g., Fontaneto and Hortal, 2012), while others argue that geographic barriers and physical isolation drive microbial evolution (e.g., Papke and Ward, 2004), with dispersal distances varying between microbial taxa (e.g., Martiny et al., 2006; Jenkins et al., 2007).

On Noachian Mars, passive and active biological dispersal agents were abundant, including gravity, water currents, impact ballistics, and atmosphere for the former and, potentially, microbial self-propulsion for the latter (e.g., Ebrahimi and Or, 2014; Petro et al., 2017). From an ecological standpoint, the heavy bombardment may have equally contributed to habitat formation and destruction, global dispersal, extinction, and biodiversification. It frequently reshaped the structure of the regolith and crust, facilitating fluid circulation in fractured and brecciated materials, thereby favoring microbial migration to new environments (e.g., Osinski and Pierazzo, 2012).

The Noachian is also the first geological period with a robust hydrologic record (e.g., Carr and Head, 2010). Currents and gravity influence aquifer and surface flow characteristics and play a central role in the distribution and types of habitats (e.g., Zeglin, 2015). Large impact basins formed in the Noachian. Many were associated with volcanic, lacustrine, and hydrothermal activity (Table 1) and presented habitable conditions (e.g., Cabrol and Grin, 2001, 2005; Varnes et al., 2003; Irwin et al., 2004; Fairén et al., 2005, 2016; Schulze-Makuch et al., 2005; Jepsen et al., 2007; Fassett and Head, 2008; Newsom, 2010; Cousins and Crawford, 2011; Schwenzer et al., 2012; Osinski et al., 2013; Retallack, 2014; Michalski et al., 2017). It can be expected, therefore, that during high obliquities—given that liquid water could circulate at the surface, through the regolith, and in the crust—microbial ecotones (i.e., transition zones that allow connectivity between microbial habitats [e.g., Borchers et al., 2015]) would have included fractured bedrock and clastic sediments (including their composition, texture, mineralogy, and albedo), the water interface between the unsaturated and saturated zones, and the groundwater/surface water interface (e.g., Griebler and Lueders, 2009; see also Table 2). Lakes, ponds, deltas, and rivers tapped into these aquifers through springs, catastrophic discharges, erosion, and surface water/groundwater interfaces. Residence time of days for the surface, to centuries and millennia for kilometer-deep aquifers, is consistent with at least some subsurface and deep biotic material reaching the surface through these pathways and being subsequently dispersed. On Earth, this process supports pioneer microbial colonies, whose success and stability over time depend on interactions and interrelationships between populations and their adaptation to the environment (e.g., Kelly et al., 2014).

2.

Environmental Elements of Microbial Ecosystems on Mars Through Time

Habitable environments	Habitats	Microbial ecotones	Dispersal pathways
Ocean[1]	Water surface	Shallow ocean margins/land transition	Waves, wind, current, gravity, connectivity networks with estuaries, deltas, streams, and glaciers through subglacial flows
	Water column	Transition zones of temperature, light, pressure, minerals, salts, chemistry, gases	Currents, gravity, fluid fluxes
	Deep sea vents	Temperature, chemistry, gases, minerals	Currents, plumes and fluid fluxes, particulates
	Subseafloor aquifer	Crust/sediment exchange	Fluid circulation, pore pressure and connectivity
Volcanic aquifer[1]	Pools, springs, spring mounds, geysers, mudpots	Surface water/groundwater interface, chemistry, mineralogy, temperature, biofilms	Current, gravity, infiltration, pressure, plumes and fluid fluxes, wind
	Pore spaces (e.g., catenary, cul-de-sac), grain texture, and mineralogy	Water interface: unsaturated/saturated zone	Recharge, underflow, discharge, current, gravity, pore geometry, fluid pressure
	Deep rock structure, texture, temperature	Rock/confining bed (clays)/clastic sediment interface	Pore geometry, pressure, connectivity, fluid pressure
Stream[1]	Water column	Surface water/groundwater interface, water chemistry, temperature, nutrient concentration; deltas, areas where water changes speed	Hydrologic connectivity and water intermittency in ephemeral and perennial streams, current, gravity
	Pore spaces (e.g., catenary, cul-de-sac), grain texture, and mineralogy	Water interface: unsaturated/saturated zone	Pore pressure and connectivity, fluid circulation
	Interstitial pore water in sediment and aquifer	Surface water/groundwater interface, chemistry, mineralogy, temperature	Pore connectivity, fluid circulation
	Subglacial	(See ice)	(See ice)
Lake[1]	Water surface	Shore/river/land transition, surface water/groundwater interface	Wave, wind, current, gravity, connectivity networks with deltas, streams, and glaciers through subglacial flows
	Water column	Temperature, light, pressure, minerals, salts, chemistry, gases	Currents, gravity, evaporation, wind
	Lake bed sediment	Sediment texture, temperature, composition, geochemistry, water chemistry, gases	Currents, gravity, evaporation, desiccation, wind
	Hydrothermal input (impact crater lake, volcanic lake)	(See volcanic aquifer)	(See volcanic aquifer)
	Ice (ice-covered lake)	(See ice)	(See ice)
Ice[2–4]	Crystals in glacier and snow	Pressure, temperature, gases, composition	Drainage from melt, infiltration, wind
	Spring mound, sediment texture	Temperature, composition, rheology	Stream, current, gravity, infiltration, wind
	Debris-rich basal ice	Solid ice crystal/water-filled veins interface, ice/land transition	Ice movement (in glaciers), meltwater, current, gravity, infiltration, wind
Permafrost[5–6]	Ice, regolith, sediment, pore spaces, water lenses	Temperature gradient, ice content, multiphase state of water, mineralogy	Melting, thawing, stream, gravity, active layers
Rock[7–8]	Cracks, pores, vesicles, weathered surfaces, mineral type and optical properties	Mineral, albedo, composition change, face exposure, rock/soil and other transitions (e.g., endolithic, chasmolithic cryptoendolithic)	Pore pressure, fluid circulation, wind, precipitation, thermokarst
Soil/Sediment[9]	Intergranular spaces, pores, grain texture	Soil/sediment transition, gas composition and content, land/water/ice transitions, geochemical fluxes	Fluid circulation, wind, precipitation, saltation
Salt/Evaporite[10]	Saline soils, hypolithic and endolithic habitats, hygroscopic minerals	Soil horizons, transitions in salt density, water content gradients, salt type, rock/sediment texture, composition, transparency	Deliquescence solutions, fluid circulations, surface tension in rock matrix, capillary pressure, porosity, permeability, wind
Cavity/Cave[1]	All habitats above	All ecotones above	All dispersal pathways above

See text, references therein, and Table 1; 2Boetius et al., 2014; 3Montross et al., 2013; 4Lutz et al., 2017; 5Jansson and Taş, 2014; 6Steven et al., 2006; 7Walker and Pace, 2007; 8Blackhurst et al., 2005; 9Groffman and Bohlen, 1999; 10Davila and Schulze-Makuch, 2016.

The early loss of the magnetosphere implies, however, that short UV radiation, that is, UVA, UVB, UVC, and ionizing radiation, increasingly reached the surface as the Noachian progressed (e.g., Hassler et al., 2014; Ehresmann et al., 2016). With a thinning atmosphere, radiation doses substantially increased (e.g., Atri et al., 2013), and radiation from muons would have impacted the subsurface with energetic secondary particle fluxes. When this process started, the probability of maintaining a long-term surface biosphere on Mars became very low regardless of the abundance and effectiveness of dispersal agents (e.g., Cockell et al., 2000; Rontó et al., 2003; Dartnell, 2011; Quinn et al., 2013; Wadsworth and Cockell, 2017).

Microorganisms and habitats are impacted differently by surface exposure (e.g., during catastrophic dispersal) or by accumulation doses (e.g., for a habitat and its biosignature preservation potential). Ultraviolet radiation on the surface has a ∼1000-fold greater biocidal effect on Mars than on Earth (e.g., Beaty et al., 2006) and is rapidly lethal to exposed microorganisms (seconds to tens of minutes depending on microorganisms), although temporary survival is possible under a thin layer (≤ 1mm) of dust or regolith (e.g., Rummel et al., 2014). By contrast, the Radiation Assessment Detector (RAD) on board the Curiosity rover measured a relatively constant and negligible ionizing radiation flux, mostly from galactic cosmic rays, and only one solar particle event over a period of 300 sols (e.g., Hassler et al., 2014). In the galactic cosmic ray case, biomolecules are compromised by dose accumulation, not instant exposure (300 million years down to 10 cm depth) (e.g., Pavlov et al., 2012), and sterilization depth and destruction may vary depending on the mineralogical composition and structure of the subsurface (e.g., Gómez et al., 2010; Ertem et al., 2017).

The Noachian presents, therefore, a paradox whereby primitive life could have occupied diverse and widespread environments and migrated through numerous dispersal mechanisms, but where a surface and subsurface sterilization process became active early and at global scale (e.g., Jakosky et al., 2017). Resolving this paradox is critical to understanding whether any types of durable Noachian subaerial habitats were possible, whether life had enough time to produce biomass in concentration that could be detected by mission payloads only reaching within a few centimeters to a meter at best into the regolith, and whether conditions were ever met to preserve such a record.

Resolving this paradox entails gauging the unique complexity of the martian environment, which goes well beyond understanding the individual impact of the loss of the magnetosphere or that of the declining atmosphere on biological processes. The unique complexity of Mars, even in these early stages, resided in the combination of a multitude of environmental extremes and the variability of their interplay over time. Some parameters declined in time (magnetosphere, atmosphere, energy); others still had distinct spatiotemporal effects depending on obliquity (e.g., water, ice distribution). For example, although the loss of the atmosphere was ultimately linked to that of the magnetic field, weak fields are less of a factor in surface radiation doses than the loss of atmospheric depth, the latter increasing doses by orders of magnitude (vs. folds for the former) (e.g., Atri et al., 2013). Changes in the atmospheric shield were therefore not only a factor of time but also a factor of obliquity (e.g., Jakosky and Phillips, 2001; Kreslavsky and Head, 2005), and this unpredictability in the radiation environment is only one of many factors (e.g., temperature, desiccation, geochemical, and textural changes in sediment) exposed microorganisms would have had to contend with.

To date, we only have a limited understanding of the effect of environmental polyextremes on microorganisms (e.g., Seckbach et al., 2013). This fundamental gap in knowledge has the potential to impair the exploration of planetary biosignatures whether on Mars or beyond, as individual coevolution on planets and moons will be shaped differently by distinct synergistic spatiotemporal changes between environmental parameters and life adaptation. The study of early martian lake analogues in the Andes (e.g., Cabrol et al., 2007b, 2009, 2014) provides some insights into the impact of environmental polyextremes on shallow aquatic habitats relevant to Noachian Mars and reveals complex loops and feedback mechanisms. For instance, although high salinity preserves liquid habitat by depressing the freezing point of water, it allows UV to penetrate the water column and exposes life to radiation. As water evaporates, increased total dissolved solids in lakes amplify UV scattering, thereby attenuating the penetration and impact of radiation. In ice-covered lakes, short UV are screened, but less photosynthetically active radiation reaches the water column (e.g., Cabrol et al., 2009). In these polyextreme environments, the relative spatiotemporal influence of each parameter is a critical element in the evolution of microbial habitats (e.g., Cabrol et al., 2007c) and may vary on very short timescales (days, weeks, seasons). While rapid climate change at the global scale is the fundamental force driving the current retreat of glaciers and lakes in the Central Andes (e.g., Bradley et al., 2006; Rabatel et al., 2013), local factors such as reduced watersheds, energy sources, mineralogy, geochemistry, exposure, sediment texture, water acquisition, and water temperature are what determine the rapidity at which habitats fluctuate and how they respond and evolve. This is reflected both in a hyper-adaptation and low biodiversity of the resident species (e.g., Cabrol et al., 2009; Demergasso et al., 2010), and by contrast, this creates a multitude of highly distinct colonies located only hundreds of meters from each other, thereby somehow augmenting regional biodiversity.

Mars became environmentally polyextreme early. How this might have translated into a biosphere, habitats, metabolisms, and dispersal pathways (Table 2) would depend on whether microorganisms ever resided at the surface, in the subsurface, underground, or occupied a subset—or all—possible niches. If life resided deep underground, surface and subsurface variability would have had only limited impact on habitats (Fig. 2A). Assuming that a martian biosphere temporarily involved subaerial habitats in the Noachian, the picture depicted by the Andean lakes suggests an evolution toward polyextremophiles (Fig. 2B), hyper-adaptable microbial communities, and oases driven by highly localized factors. The footprint of such communities would depend on where habitat conditions are met and, in terrestrial extreme environments, is often as small as a crack in a rock, a slope exposure, or a specific mineralogy (e.g., Navarro-González et al., 2003; Warren-Rhodes et al., 2006; Wierzchos et al., 2006; Cabrol et al., 2007a; Azúa-Bustos et al., 2011; Davila et al., 2013).

Polyextreme environmental evolution of Mars. Mars became polyextreme very early. Its capability to preserve subaerial habitats, ecotones, and microbial dispersal pathways during the transitional period between the Noachian and Amazonian would have depended on fluctuating interactions between multiple environmental extremes and their relative dominance at any given time, which in part depended on time and obliquity. (A) Evolution of the abundance and diversity of microbial habitats, ecotones, and dispersal pathways over time as a function of increase in polyextremes (number and intensity). The impact is most severe for those at the surface (solid black line); it would have been less severe for those in association with the subsurface (dashed black line) and least severe for those of the deep underground (dotted line), where survivability would have depended primarily on the evolution of geothermal energy and deep water circulation. The straight lines show temporal trends, while evolution would have varied with obliquities. (B) Interactions between extreme environmental parameters. White arrows = promote; black arrows = prevent; left-right black arrows = counter; left-right gray arrow = combine; dotted thin arrows show loop cycles.

When atmospheric depth became insufficient on Mars—including during favorable obliquities (e.g., Head et al., 2003; Kahre et al., 2013; Mischna et al., 2013)—the potential for the expansion of a surface biosphere came to a halt. MAVEN and other missions show through changes in atmospheric characteristics, hydrology, and mineralogy that this scenario had already unfolded by the end of the Noachian (Table 1). MAVEN's conclusions suggest that the last part of the Noachian may never have been environmentally stable enough to sustain long-term subaerial habitats. If that is the case, depending on the timing of their formation, many environments currently perceived as favorable for biosignature exploration may only have been the temporary receptacles of ephemeral biological communities cyclically flushed out of sheltered underground habitats, rather than thriving subaerial oases. Ultimately, Mars' capability to retain subaerial habitats, ecotones, and dispersal pathways during the transitional period between the more favorable Noachian and the hostile Amazonian would have depended on how multiple extremes interacted with each other, which ones dominated in time and during any given obliquities, and whether these interactions where beneficial or detrimental to individual ecosystems.

5. The Great Hesperian Dispersal

The Late Noachian signaled an environmental turning point that constrained the subaerial expansion of a biosphere. Yet a coevolution on Mars presents us with possibly another paradox, whereby the Hesperian would have been a highly effective time for biological dispersal. While a primordial ocean remains notional, a Hesperian ocean, or limited shallow seas (3.7–3.4 Ga), although debated (e.g., Carr and Head, 2015; Wordsworth, 2016), has stronger empirical and theoretical support (Table 1 and references therein). Despite the sharp decline in hydrological processes, many outflow channels are Hesperian in age and formed in the equatorial region (e.g., McKenzie and Nimmo, 1999; Baker, 2001; Dohm et al., 2001; Montgomery and Gillespie, 2005; Harrison and Grimm, 2008; Carr and Head, 2010; Fassett and Head, 2011) through the catastrophic release of lakes (e.g., Lucchitta, 2010), ice streams (e.g., Lucchitta, 2001), hydrate deposits (e.g., Kargel et al., 2007), or aquifers (e.g., Carr, 1996; Baker, 2001). The formation of episodic oceans in the northern hemisphere has been attributed to the release of these reservoirs.

With the prospect of subaerial habitats becoming increasingly unlikely after the Noachian, the processes linked to the long-term underground water circulation from the highlands to the lowlands (e.g., Clifford, 1993) could have been one of Mars' dispersal “highways” and produced over time the largest concentration of biomass in equatorial reservoirs near the boundaries of the global dichotomy. As these reservoirs were catastrophically released, and the ocean formed in the northern lowlands, this process could have resulted in a sudden and large-scale biological dispersal. Whether oceanic habitats formed in the Hesperian would have depended on the ability of microorganisms to adapt to new environmental conditions (e.g., Fairén et al., 2003). Theoretically, habitable environments and habitats previously described for a notional primordial ocean (e.g., deep sea vents, deep subseafloor crustal aquifers, Tables 1 and 2) could still have formed in a Hesperian northern ocean. The buildup of Tharsis at the equator in the same time frame was accompanied by intense faulting and volcanic activity. Both Valles Marineris and Tharsis were areas of high concentration and dissipation of energy through volcanic and tectonic activity (e.g., McKenzie and Nimmo, 1999; Yin, 2012; Leone et al., 2014). The connection of any ocean margins either directly or indirectly (through fault systems) to the Tharsis/Valles Marineris and Elysium provinces would have had conditions favorable to the development of hydrothermal activity over hundreds of millions of years (e.g., Farmer, 1996; Schulze-Makuch et al., 2005; Al-Samir et al., 2016) that possibly continue today (e.g., Dohm et al., 2009a; Broz et al., 2017). The Hesperian deltas (e.g., Di Achille and Hynek, 2010) provided connectivity networks back to land through transitional environments (estuarine, shallow marine, subseafloor aquifers). The duration of oceanic habitats and the spatial extent of biological dispersal would have depended on volatile sources and cycling, ocean structure (e.g., currents, bathymetry), extent, composition, microbial adaptation, and the sustainability of energy sources and nutrients (e.g., through hydrothermal sources and vents and the weathering of land on the surface). The development of shallow water habitats ultimately would have depended on the state of the atmosphere and the potential for the water column (e.g., depth, water chemistry) and sediment characteristics to attenuate short UV (e.g., Leavitt et al., 1997; Rhode et al., 2001; Cabrol et al., 2009; Häder and Gao, 2015).

At the opposite environmental extreme, Mars may have always been too cold for any ocean to form (e.g., Wordsworth, 2016). Yet biomass concentration and biological dispersal in the regolith and crust would still have been possible in this scenario through hydrothermal convection (e.g., Travis et al., 2003) and would have depended on depth, structure, and the physical and textural characteristics of the rocks to enable unrestricted pore fluid circulation. There is ample empirical evidence of Hesperian catastrophic floods and surface flows in the subtropical and equatorial regions of Mars (e.g., Carr, 1996, Table 1). Further, missions have provided unambiguous confirmation that lakes formed and water ponded (e.g., Ruff et al., 2014; Grotzinger et al., 2015a, 2015b, and Table 1). With a cold scenario, the question is not so much Mars' continued capability to move biomass during the Hesperian as it is what may have changed, that is, (i) the depth at which dispersal took place, (ii) the environmental conditions for the development of new habitats, and (iii) the preservation of biosignatures.

6. An Amazonian Biosphere

Catastrophic biological dispersal would have repeated through episodic outflow activity during the Middle Amazonian (1.23–0.3 Ga) in the Tharsis region (e.g., Schulze-Makuch et al., 2005; Rodriguez et al., 2015) and possibly much more recently in Western Elysium (e.g., Plescia, 2003; Russell and Head, 2007; Balme et al., 2011) in an environment that was already hyperarid and cold. In theory, this succession of episodes designates the northern lowlands as a region where biomass could have been abundant, concentrated, and protected in repositories covered under thick sedimentary deposits (e.g., Boyce et al., 2005; Mouginot et al., 2012); see also Section 7.1.

Mars has been geologically more quiescent in the past 3 billion years. However, recent volcanic activity (e.g., Werner, 2009; Hauber et al., 2011; Broz et al., 2017) and volcanic/ice interactions (e.g., Mouginis-Mark et al., 1984; Head and Wilson, 2002; Lanz and Saric, 2009; Levy et al., 2017) show that energy and water were still present tens to hundreds of millions of years ago at depth in the equatorial region. Additionally, the availability of water ice within the top meter in the high latitudes (e.g., Levrard et al., 2004; Bandfield, 2007) makes the north polar and subpolar regions areas of interest, especially when associated with evidence of geologically recent volcanic activity (e.g., Hovius et al., 2008). The idea that volcanic/ice interactions could still host a modern-day biosphere in a wide range of latitudes and settings (e.g., Boston et al., 2001; Boston, 2010; Cousins and Crawford, 2011) is not new. However, recent evidence of volcanic activity in the Late Amazonian greatly reinforces the possibility of a still-active martian biosphere. In these regions, and in regions of active layers, biomolecules may be exposed at the surface cyclically in the deposits from recent gullies (e.g., Wynn-Williams et al., 2001), recurring slope lineae (e.g., McEwen et al., 2011), and periglacial and glacial deposits from recent high-obliquity cycles (e.g., Head et al., 2003). These regions currently fall under the labels of Uncertain Regions and Naturally Occurring Special Regions for exploration (e.g., Rummel et al., 2014).

7. Discussion

Potentials and limitations of the analogy we draw between early Earth and Mars are redefined with each new Mars mission, but a simple side-by-side comparison between the two planets viewed from space sends an already powerful message: Over 4 billion years of coevolution have resulted in a deep imprint of the biological processes over the terrestrial landscape, of which biomediated minerals (e.g., Bradley, 2015) and sedimentary bioconstructs (e.g., Walter et al., 1992; Summons et al., 2011) are only a few examples. Many biosignatures are visible from space (e.g., McCabe et al., 2016), and Earth's atmosphere is the first evidence of how influential biology has been in shaping the terrestrial environment (e.g., Lyons et al., 2014).

By comparison, Mars does not show any visible signs that biology ever had a significant impact on surface processes. If it did, it is at a resolution yet to be reached, or in ways we do not understand yet, and this alone is a critical piece of information. Nevertheless, the potential for an origin of life did exist on Mars, and despite all differences, it is possible to theoretically assemble an early martian biosphere from existing data. Yet the data also clearly show that ultimately, while life could have survived the heavy bombardment, as it did on Earth, early subaerial habitats may never have had the opportunity to rebound from an irreversible environmental change, as was often the case for our biosphere. Sometime during the Noachian, the elements of a coevolution of life and environment that involved the surface simply went away on Mars. The next step is thus to understand where this observation takes the search for biosignatures.

7.1. Large biomass repositories

Regardless of where life could have developed (ocean, volcanic aquifer, other), greater habitat stability on Mars consistently resided below the surface over time (Table 2). While early surface environmental conditions were equally extreme on Earth and Mars, the coevolution of life and environment on our planet led to a biosphere capable of expanding and occupying every niche, whereas a coevolution on Mars would have forced life—as we know it—to continuously shrink its surface footprint and ultimately retreat underground. This would not have prevented Mars from developing a deep, diverse, and extended biosphere. Many terrestrial lithoautotrophic microbial communities do not depend on surface organisms for primary production (e.g., Stevens, 1997). They reside in all physical spaces that provide minimum requirements for life (e.g., Jepsen et al., 2007; Harrison et al., 2013; Schulze-Makuch et al., 2015), including at the limit of habitability (e.g., Yanagawa et al., 2017), where their abundance and diversity are constrained by the availability of water, energy/nutrients, and carbon (e.g., Reith, 2011; Rempfert et al., 2017). Microorganisms in dark, nutrient-poor habitats live in biofilms attached to mineral surfaces (e.g., Jones and Bennett, 2017) and have adapted to specific minerals for their biogeochemical processes. Survival, growth, and community structure are linked to mineral chemistry and reactions at the microbe/mineral interface (e.g., Jones et al., 2011), and they survive in a wide range of environments using thermodynamic disequilibria within their geological settings (e.g., Gaidos and Selsis, 2007).

A coevolution approach to Mars shows that, over eons and with expected regional and latitudinal variability, microorganisms would likely have concentrated in the underground, forming deep biomass vaults (Fig. 3). Among those, Arabia Terra stands out as a unique topographic and hydrologic collection area over 4 billion years. It represents a topographic trap for the underground water circulation originating in the southern regions of Mars (e.g., Clifford and Parker, 2001) and in Valles Marineris to the west (e.g., Carr, 1996; Baker, 2001; Dohm et al., 2001; Carr and Head, 2010), where recent volcanism (e.g., Broz et al., 2017) suggests that fluid circulation at depth could still be supported by magmatic activity. Although the origin of methane emissions in Arabia Terra has not been determined yet (e.g., Mumma et al., 2009; Etiope et al., 2013; Michalski and Bleacher, 2013), this specific location is to be expected from a deep biosphere perspective, that is, considering the interactions between lithosphere, hydrosphere, and energy resources (e.g., Chapelle et al., 2002; Schulte et al., 2006; Pirajno, 2009; Amils et al., 2014; Rempfert et al., 2017).

Primary dispersal pathways and biomass repositories—the red arrows represent potential primary biological dispersal pathways; blue arrows are hypothetical shallow dispersal paths along costal marine currents; boxes indicate primary deep biomass repositories. The suggested surface areas are representative only. They include the following: (1) The south polar–Argyre–Chryse trough drainage system (Clifford and Parker, 2001), where Argyre collects polar basalt meltwaters. Surface discharges occurred through the Chryse trough (Parker et al., 1993; Fairén et al., 2016) into west Valles Marineris and an early northern ocean. Surface flow is questioned by Hiesinger and Head (2002), who favored subsurface drainage of a lake. Evidence supports an early habitable environment (e.g., Moore and Wilhelms, 2001; Hiesinger and Head, 2002; Fairén et al., 2016; Williams et al., 2017). (2) Hellas presents similar habitability potential as Argyre (e.g., Gulick, 1998; Schulze-Makuch et al., 2007, and references herein; Wilson et al., 2010). Hellas is a closed basin and the deepest point on Mars at −7152 m. Its seasonal atmospheric pressure ∼89% higher than the surface (Grassi et al., 2007) allows transient surface liquid water episodes and glacial processes (e.g., Haberle et al., 2001) and the release of deep materials to the surface. (3) Sirrenum and Memnonia Fossae are structural troughs connected to Tharsis. They mark the origin of large lakes and channels, including in the Mangala and Ma'adim Vallis regions. The latter is additionally linked to volcanic/hydrothermal systems from impact cratering (e.g., Gusev Crater) and volcanic activity (e.g., Apollinaris Patera). The exploration of Gusev confirmed a Noachian habitable environment, with geomorphic and mineralogical evidence presented as possible bioconstruct analogues (Ruff and Farmer, 2016). (4–5) Sustained volcanic and hydrothermal activity with cyclic accumulation of volatiles in equatorial aquifers makes the Tharsis/Valles Marineris and Elysium regions high-priority areas for deep biomass repositories. (6) Arabia Terra is an outstanding candidate repository. Topographically, it has been a surface, subsurface, and deep underground collection area over the entire history of Mars. Late Amazonian volcanism (Broz et al., 2017) shows modern magmatic processes and a potential for hydrothermal circulation that drains from Valles Marineris toward Arabia. The region is characterized by higher epithermal neutron count (e.g., Feldman et al., 2002) and methane plumes (e.g., Mumma et al., 2009; Oehler and Etiope, 2017). (7) The northern plains repository could theoretically be composed of biomass released from the highland through catastrophic releases of equatorial aquifers and from oceanic habitats. Other biomass repositories and deep dispersal pathways might include the equatorial belt at depth, where the highlands/lowlands flow circulation concentrated underground over billions of years; the planet's overall deep interior (≥500 m)—with water at depth, combined lithostatic pressure and geothermal gradient have maintained conditions to develop possible deep habitats over time since accretion. Microbial organisms migrating from subseafloor and/or volcanic aquifers could have colonized deep aquifers, caves (e.g., from chemical dissolution, ancient magma chambers, lava tubes, and underground rivers), cavities, mineral surfaces, and pore spaces (see Table 2). Mars analog studies have abundantly demonstrated the suitability of these environments for a broad range of microbial communities. Credit: The basemap was prepared by Daniel Macháček.

By definition, many of these deep vaults are located hundreds of meters below the surface and are thus not directly accessible by current exploration means. They could be designated as prime astrobiology investigation sites for future exploration (e.g., Niles et al., 2017), including deep drilling (e.g., Stoker et al., 2008), as current robotic coring systems can only reach a few centimeters to a meter into the subsurface (e.g., McKay et al., 2013; Zacny et al., 2013).

7.2. Ecosystem exploration with rovers and detection thresholds

Environmental evolution suggests that a biosignature “low-hanging fruit” may be rare at the surface of Mars, but two types of records are still theoretically within reach of current exploration means. The first one relates to materials extracted from their original geological settings by impact cratering and hydrogeologic activity (e.g., streams, glacial and periglacial processes, landslides, tsunamis). This record is now haphazardly distributed and out of context, and after ∼4 billion years of environmental exposure might be difficult to decipher with current instrument detection levels. The other potential biosignature record may still be in situ and possibly survived destruction, and ideally, this is the one to be targeted by landing site selection for upcoming missions.

Central to the landing site selection process are the concepts of habitability, biosignature formation, and biosignature preservation potential (e.g., Mustard et al., 2013). They are currently evaluated primarily through geological, geochemical, and compositional evidence, rather than through an ecological and microbiological lens. From a coevolutionary standpoint, however, and with a mission goal to identify samples most likely to have retained biosignatures, it is essential for this approach to be supported by an ecosystemic perspective that focuses on microbial systems and what constitutes evidence (or lack thereof) of interactions between past microbial habitats, their occupants, and the environment (e.g., Cockell, 2010). Over the past decades, an abundance of physical data has been accumulated in support of the formation of habitable environments at the surface of Mars between the Noachian and the Hesperian, but whether these environments were ever inhabited (e.g., Cockell et al., 2016) is really what ultimately matters for the upcoming missions.

The current challenge is not in a lack of understanding as to what these interactions could be or the scale of evidence to search for (e.g., Summons et al., 2011; Ehlmann et al., 2016); it is in the lack of orbital data with adapted spatial scale and spectral resolution to support a robust evaluation of the potential for such interactions in a landing ellipse prior to a mission. Broad-scale environmental habitability potential on Mars is identifiable from orbit with current exploration assets, and its pre-mission characterization has successfully supported previous landing site selections (e.g., Golombek et al., 2005). However, ancient potentially habitable environments might not translate into ancient habitats once at the surface. This data gap increases uncertainty in decision-making with respect to achieving the central mission goal and entirely shifts the burden to rover science operations for the upcoming missions.

Current knowledge gaps will not be filled by the time Mars 2020 and ExoMars land, but biosignature exploration can be supported and uncertainties reduced in several ways. One is to identify missing detection thresholds from orbit to the ground. Through the NASA Astrobiology Institute–funded project Changing Planetary Environments and the Fingerprints of Life, ongoing research in extreme environments (e.g., Cabrol et al., 2015) is quantifying these thresholds for a variety of relevant microbial habitats by using satellite imagery and drones. Preliminary results (e.g., Phillips et al., 2017) show that, to be effective and diagnostic, the orbital resolution of a visible imager should reach ∼1 cm/pixel to resolve fine-scaled geomorphic features in microbialites, spring mounds, and salt habitats, which is beyond the recommendation of ∼10–15 cm/pixel for a future orbital visible imager (e.g., MEPAG NEX-SAG Report, 2015). HiRISE and a future orbiter with ∼10–15 cm/pixel resolution could identify bulk geomorphic features that are consistent with a potential habitat (e.g., Suosaari et al., 2016), but might still not be diagnostic for many of them (e.g., Allen and Oehler, 2008) unless data sets can be improved through new super-resolution techniques (e.g., Tao and Muller, 2016).

Combined with visible data, orbital spectral signatures are critical elements in ancient habitat detection. Interactions between microorganisms and their environment are captured through the presence of biomediated minerals (e.g., Röling et al., 2015). In that, CRISM resolution (e.g., Murchie et al., 2009) brings some support to habitat detection. While the instrument might not be able to resolve microbial habitats individually, the ∼6 m/pixel along-track oversampled hyperspectral imaging resolution (e.g., Kreisch et al., 2017) provides an important strategic tool for the identification of mineralogical transitions associated with potential habitats and ecotones. For example, ∼60 m spatial resolution is already capable of revealing temperature changes in hydrothermal springs reflected in changing mineralogies, and/or pigments of microbial colonies in these environments (e.g., Kruse et al., 2011).

Another means of supporting biosignature exploration prior to, and during, a mission is by integrating survey techniques developed in microbial ecology. Past and recent NASA-funded rover field experiments testing microbial habitat and biosignature exploration strategies at the surface and in the subsurface of the Atacama Desert (e.g., Wettergreen et al., 2005, Cabrol et al., 2007a) led to the development of ecological mapping techniques adapted to autonomous rover operations, and to the generation of predictive maps of habitat potential (e.g., Hock et al., 2007; Warren-Rhodes et al., 2007a, 2007b, and unpublished data; Smith et al., 2007). This is typically an area where advances in detection algorithms and machine learning can bring new support to biosignature detection, including from orbit through, for example, the segmentation of hyperspectral imagery, correlation of orbital and surface spectra, novelty detection, multiscale adaptive sampling, data segmentation and superpixel clustering, and biohint probability mapping, among others (e.g., Thompson et al., 2011; Candela et al., 2017).

In addition to a geological and environmental perspective, developing an ecosystemic view of a region prior to landing is absolutely essential to guide biosignature exploration. Research in terrestrial extreme environments and the exploration of the rare Archean record shows how challenging such an exploration can be with boots on the ground, and on Mars it will be performed within the operational constraints and scientific payload limitations of robotic systems. The support of robust orbital remote-sensing capabilities with relevant spatial and spectral resolution is thus foundational to the development of robotic exploration protocols for the search for biosignatures on Mars, but even more broadly to the development of autonomous biosignature exploration anywhere in the Solar System and beyond. It is fundamental not only to the preparation of a mission and the characterization of a landing ellipse but also to the planning of the future search for life beyond Earth. Tactically, drones supporting local rover exploration along a traverse could substantially increase our ability to develop dynamic spatiotemporal models of ancient ecosystem evolution along a traverse as we go during a mission (e.g., Pergola and Cipolla, 2016; Phillips et al., 2017). Strategically, however, having an understanding of habitat potential at global scale through orbital capabilities (vs. global habitability potential as we currently do) would help us design missions that are best adapted to the evidence we find (or the lack thereof), and optimize a long-term biosignature exploration strategy.

8. Conclusion

Whether a coevolution of life and environment ever took place on early Mars is unknown. However, the factors at play demonstrate the uniqueness of each planetary experiment, regardless of similarities with Earth. They also emphasize the differences between searching for biosignatures and characterizing habitability. Searching for biosignatures involves a systemic understanding of a coevolution, that is, the integration of what we have learned in past decades about Mars' habitability and environmental changes over time, with inferences we make about their impact on hypothetical biological processes and their constraints on habitats. Critically, it requires that we start thinking of Mars in terms of a biosphere, ecosystem, and ecology. It also leads us to explore the closest planetary paleoenvironment to early Earth, which is probably one of its greatest challenges.

In many ways, the obstacles of reconstructing a Noachian coevolution are comparable to those faced when trying to reconstruct the Archean coevolution, and the paucity of the early terrestrial record and its alteration illustrate the complexity of this task. For Mars, we still have to fully understand how regional and local hydrogeochemistry could have affected biogeochemical cycles and life size, composition, metabolism, morphology, and biosignatures (e.g., Summons et al., 2011). In that, the range of martian environments discovered shows both neutral and acidic conditions, which points to an early diversity comparable to Archean Earth. As Mars developed a more acidic hydrosphere in time, and the surface became more hostile, the likelihood for the production of substantial amounts of organic matter and large-scale biostructures in a subaerial context substantially decreased. In addition to a potentially low initial production of organic matter, deposits were subsequently affected by diagenesis, climate, tectonic and impact processes, and other physicochemical processes. This would have contributed to significant alteration of a biorecord and our ability to recognize it, although studies such as Río Tinto show that sedimentary deposits formed in a highly acidic and oxidizing environment record aspects of the physical, chemical, and biological environment of the regional ecosystem. This could indicate that preservation of some indicators through diagenesis and environmental influences might still have been possible on Mars (Amils et al., 2014).

Regardless of initial conditions, subaerial environments may never have been stable enough over extended periods of time, simply due to the high variability of the martian climate, this instability only growing over time with the loss of the atmosphere. Ultimately, the subsurface and the deep underground were always more stable and would have allowed the development of specialized microbial communities that had more time to thrive and accumulate large biomass deposits. Sheltered conditions would also have provided more favorable conditions for preservation.

Understanding how to search for biosignatures on Mars also entails filling in knowledge gaps with continuous research here on Earth. This includes, and possibly chiefly, the identification of detection thresholds; the design, testing, and implementation of autonomous biosignature and habitat detection methods; and the development of a better understanding of the evolution of polyextremophiles in complex environments and their physical, mineralogical, and biological signatures. As we think of Mars as a coevolution, we must also be prepared to envision all possible end-member terms, including the recognition that a martian biosphere may never have involved a surface expression, and anticipate the next steps along that path as well.

Further, we should consider a variety of complementary approaches to the exploration strategies currently deployed. In particular, although our understanding of early Mars in its first billion years remains fragmented, present-day Mars is a reflection of its past 3 billion years of history. We have now developed an understanding of the impact of orbital forcing and past and modern climate cycles at global to regional scales. However, a characterization of the high-resolution spatiotemporal scale of variability of the environment today has yet to be undertaken. Doing so would provide an immediate and uncertainty-free insight into the ecological potential of Mars' surface and near surface over the past ∼75% of its environmental history. While life is not expected to survive near the surface today, understanding present surface conditions, the influence of the landscape (e.g., topography, geology, texture, albedo), and its variability at sufficient resolution to characterize the scale of microbial habitat, will identify the surface expressions of seasonal and perennial environmental “hotspots,” which may be potential windows into subsurface and deeper biomass vaults. Such areas in terrestrial extreme environments are often characterized by unique conditions at multiple scales but particularly at the meter to micrometer scale (e.g., slope exposure, cracks in rocks, cavities in the mud of an ancient lake bed, surface/atmosphere interactions). At these sites, parameters such as temperature, moisture, light, mineralogy, sediment texture, pH, and others generally differ from the average ambient environment and remain stable over time for life to persist. In addition to high spatial and spectral resolution orbiters, high-resolution environmental arrays strategically deployed on Mars would dynamically map the distribution and concentration of these hotspots. Ultimately, this approach could become a cornerstone strategy for the exploration of Mars, as it supports three critical exploration goals: the search for biosignatures; human exploration (climate, weather, and activity planning); and planetary protection through the identification and characterization of environmental hotspots below current resolution and their spatial relationships with respect to present and future missions.