A global database of Mars-relevant hydrovolcanic environments on Earth with potential biosignature preservation

Hydrovolcanic environments that formed from volcanic eruptions into water or ice are regarded as important for origins-of-life and early planetary habitability research and, as such, should be considered ideal targets for astrobiology (Baross and Hoffman, 1985; Cockell, 2014; Sasselov et al., 2020). The Mars crust is primarily composed of basalt (Zuber, 2001; Nimmo and Tanaka, 2005; Bridges and Warren, 2006; McSween et al., 2009), and its surface preserves evidence of extensive volcanism throughout its history, but especially in the Noachian and early Hesperian periods (Wilson and Head, 2007; Carr and Head, 2010; Smellie and Edwards, 2016; Williams et al., 2017; Brož et al., 2021). The exact nature and history of the Martian hydrosphere is still debated, with studies arguing for a “warm and wet,” “cold and dry,” and/or “cold and wet (icy)” early Mars, depending on atmospheric conditions (Baker et al., 1991; Craddock and Howard, 2002; Fairén, 2010; Wordsworth et al., 2015; Smellie and Edwards, 2016; Galofre and Jellinek, 2016). It is clear that water, either as a pure liquid, ice, or brine on the surface or in the subsurface, affected Mars’ geological evolution and would have influenced its volcanism during eruption and subsequent alteration (Wilson and Head, 2007; Keszthelyi et al., 2010; Brož and Hauber, 2013; Smellie and Edwards, 2016; Brož et al., 2021).

While studies documenting the biology and habitable conditions of seafloor basalts are numerous (e.g., Furnes et al., 2001; Bach and Edwards, 2003; Edwards et al., 2005; Ivarsson and Holm, 2008), less focus has been given to the habitability of terrestrial (land- or lacustrine-based) hydrovolcanic environments for endolithic microorganisms (Cockell et al., 2009b; Cousins and Crawford, 2011; Cousins et al., 2013; Nikitczuk et al., 2016; Pentesco, 2019; Bergsten et al., 2021). These hydrovolcanic environments include tuff cones, tuff rings, maars (Fig. 1), and glaciovolcanic structures (Fig. 2), all of which are more readily accessible for field studies than oceanic floor basalts. Some studies showed that glass-rich rocks found in marine and terrestrial hydrovolcanic environments can host diverse microbial communities that may use photosynthesis, organic material, and/or chemical energy available from rocks (Cockell et al., 2009b, 2009a; Cousins et al., 2013; Bergsten et al., 2021, 2022). In addition, basaltic glass, or sideromelane, which is partially altered to palagonite, was found to contain microtextures that are potentially indicative of microbial alteration (Cockell et al., 2009b; Fisk and McLoughlin, 2013; Nikitczuk et al., 2016, 2022; Fisk et al., 2019; Pentesco, 2019; Izawa et al., 2019). These environments are complementary to impact craters as sites of astrobiological potential, where long-lived hydrothermal systems and abundant glass exist alongside shocked and fractured rock to act as habitats for endolithic microorganisms (Sapers et al., 2015; Osinski et al., 2020).

Phreatomagmatic and glaciovolcanic environments exist around the world, and there is a strong body of literature that documents aspects such as eruption hazard to local communities, geochemistry, and eruption mechanics. However, fewer researchers have focused on the relevance of these environments to habitability, astrobiology, or Mars volcanism (Komatsu et al., 2006; Cousins and Crawford, 2011; Cousins et al., 2013; Smellie and Edwards, 2016; Farrand et al., 2018). Here, we review the data pertaining to all terrestrial hydrovolcanic environments to identify which volcanic fields have the potential to contain microbial alteration signatures in a Mars-relevant context. Our goal is to highlight potential sites of interest for future researchers for lab or field studies focused on Mars-analogue hydrovolcanism and the astrobiological potential of these environments.

A note on language: This paper is presented to audiences from both the planetary science and traditional volcanology and geomicrobiology communities, in which the term “terrestrial” has different meanings depending on context. We use the word “terrestrial” to mean “on land” (or in a shallow lacustrine/nearshore marine environment, as opposed to “marine” or “submarine”). When we contrast the environment of Earth with Mars, we use “Earth-based” in that context rather than “terrestrial.” In addition, we use the term “hydrovolcanic” to encompass both phreatomagmatic eruptions (explosive terrestrial eruptions resulting from magma-water interactions near the surface, as opposed to effusive seafloor eruptions) and subglacial eruptions.

Tuff cones, tuff rings, and maars (Fig. 1) are monogenetic volcanic structures produced by phreatomagmatic eruptions where ascending magma comes into contact with water and explosively erupts as the water flashes to steam, a process known as a molten fuel-coolant interaction (Sheridan and Wohletz, 1983; Németh and Kósik, 2020). This rapidly quenches the magma into fragmented glass, with glass clast sizes partially representative of the energy of this explosive interaction (smaller fragments = higher energy; Smith and Németh, 2017; Németh and Kósik, 2020). The resulting pyroclastic succession and ratio of juvenile pyroclastic material to accidental fragments of country rock reflect a competition between external environmental factors (country rock composition/coherence and groundwater or surface water availability) and internal factors (magma volume, composition, and volatile components; Smith and Németh, 2017).

Depending on the depth of the water-magma interaction, different structures may be formed. Maars form from underground explosions as magma interacts with deep groundwater (Fig. 1A), tuff rings are produced if there is a shallow water table or shallow standing water (Fig. 1B), and tuff cones are formed in slightly deeper standing water as ascending magma nears the water’s surface (Fig. 1C).

As eruptions progress, the eruption style may change from phreatomagmatic to drier explosive Strombolian or effusive Hawai’ian eruptions, depending on the availability of external water as well as other factors including the amount of magmatic volatiles driving the eruption and melt viscosity. As groundwater is drawn down, or access to water is impeded due to the growth of the volcanic edifice, this transition may produce scoria, spatter, lava flows, or a lava lake that forms a flat-topped or agglutinated edifice. Many phreatomagmatic structures are therefore hybridized, initially erupting as maars or tuff rings before being capped by lava or scoria cones (Valentine et al., 2017; Smith and Németh, 2017). The final crater may contain a lava lake or be filled with water.

Tuyas and tindars are hydrovolcanic structures formed from subglacial eruptions (Fig. 2). As magma ascends to ground level and comes into contact with the base of a glacier, a basal pocket of meltwater forms around erupting pillow lavas (Smellie and Skilling, 1994; Smellie, 2001). At the top of the glacier, an “ice caldera” or depression forms directly above this meltwater/volcanic pocket due to the decrease in volume resulting from the transition from solid ice to liquid water at the base. Continuing eruptions at the base form layers of massive to poorly sorted hyaloclastite breccia with a core of pillow basalts. If the heat of the eruption melts the glacier through its entire thickness, an intraglacial lake is created. As the edifice grows, depending on the relative depth of the water, the eruptions may transition to producing more fragmented, bedded tuff deposits draping over the hyaloclastites. The volcanic edifice may grow all the way to the water’s surface, or eruptions may cease before this point, depending on the depth of the lake. Ice thickness and permeability control the depth of the intraglacial meltwater lake, and therefore the height and morphology of the final structure. A thicker ice sheet is less permeable to the basal meltwater created by the volcanic eruption, so a deeper lake is created (Smellie and Skilling, 1994; Smellie, 2001).

A flat-topped tuya (Fig. 2A) forms when the edifice breaches the lake surface and subaerial lava flows cap the tuff and hyaloclastite mound (Smellie, 2001). A ridge-shaped tindar (Fig. 2B) is formed from a fissure eruption beneath a glacier where the eruption does not continue to the point of the edifice growing above the intraglacial lake level (e.g., Smellie and Skilling, 1994; Smellie, 2001; Sæmundsson et al., 2020). Table 1 summarizes the major physical characteristics that differentiate hydrovolcanic structures.

Hydrovolcanic tuff deposits (Fig. 3) contain variations in bedding thickness, sorting, clast size, and clast-versus-matrix support that are indicative of eruptive and depositional conditions. For example, a base-surge deposit is a poorly sorted, wedge-shaped lens of tuff. It forms from a steam-saturated vertical eruption column from a phreatomagmatic eruption that collapses and surges out radially away from the vent (Fisher and Schmincke, 1984b). Base-surge deposits occur as massive, sandwave, and planar bedforms in maars, tuff cones, and tuff rings (Fig. 3F), depending on the degree of turbulence in the base-surge flow (see Fig. 1). Sandwave beds are found closest to the vent, followed by massive beds, and then planar beds at the distal ends of the base-surge deposits. Bedding direction is typically quaquaversal, meaning it dips away from the central vent in all directions (Fig. 3B).

Normally graded beds of 10–100 cm scale are common, reflecting cyclical explosions from fluctuating ratios between external water and magma (Walker and Croasdale, 1971; Fisher and Schmincke, 1984b; Figs. 3D–3F). Ash-sized particles are formed in high-energy explosions with high levels of glass fragmentation. Ash can be distributed either as the fine-grained matrix component of tuffs (that adheres to coarser grains due to the presence of water), as more distal ash-fall deposits, or sometimes in the form of accretionary lapilli (from coalescing around a nucleus in the steam-saturated eruption column; Heiken, 1971; Fisher and Schmincke, 1984b). “Cauliflower” or “breadcrust”- shaped lapilli and bombs are an occasional feature of phreatomagmatic deposits, along with surface cracks and steps on glass shards; these indicate the presence of water during the rise of magma, since the water acts as a cooling agent during eruption (Fisher and Schmincke, 1984b; Németh and Kósik, 2020). Planar beds may exhibit evidence of plastic deformation in the form of “bomb sags,” where lapilli, bombs, or blocks were ballistically ejected from the vent and landed on water-saturated tephra that sags under the weight (Fig. 3C).

The tuffs formed by hydrovolcanic eruptions show a variety of textures at the microscale that indicate initial fragmentation and quenching in the presence of water and subsequent aqueous and putative biotic alteration of glass at varying temperatures (Fig. 4). Products of glass alteration include palagonite and pore-filling precipitates (e.g., zeolites and calcite).

When a silicate magma first encounters external water, it is chilled rapidly below its crystallization temperature, and its melt fraction forms glass, a process known as quenching (Fisher and Schmincke, 1984c). Sideromelane (hydrated basaltic glass) clasts are typically blocky and contain small (<1 mm) or no vesicles; because the magma is quickly chilled by the water, it doesn’t allow for expansion of trapped volatiles prior to solidification (Heiken, 1972; Fisher and Schmincke, 1984c). In thin section, angular glass fragments are observed that contain microphenocrysts, quenched mineral textures that include skeletal or radial grains, and textures marked by contraction or devitrification around fractures and mineral grains (Figs. 4A and 4C–E). Some may contain trichite, an acicular, “fur-like” devitrification texture that is thought to be formed when minerals crystallize along microfractures or planes of weakness as the glass cools and contracts (Ross, 1962; Fig. 4D).

Sideromelane is metastable and prone to devitrification and alteration, especially when exposed to aqueous conditions and higher temperatures (Jakobsson and Moore, 1986; Jackson et al., 2019). The tuffs produced from hydrovolcanic eruptions are exposed to multiple water sources both syn- and post-eruption. They can be water-saturated during eruption and deposition and may erupt into standing water (tuff cones/rings, tuyas, and tindars), or later host a crater lake (maars or tuff rings), in addition to meteoric water exposure. The residual magmatic heat, combined with the presence of circulating fluids, typically induce alteration of the glass, which over time can lead to the induration of tuffs and better preservation of the structures.

Hydrothermal alteration of sideromelane produces clays (typically montmorillonites, smectites, kaolinite, and illite), zeolites, and iron and titanium oxides/oxy-hydroxides. At higher temperatures, the alteration can be rapid (Fisher and Schmincke, 1984a). At variable to lower temperatures, however, an intermediate phase known as palagonite is produced. Palagonite is a yellow-orange to red-brown, waxy to granular amorphous material, and is considered to be the first alteration product of sideromelane that progressively alters to clays and zeolites (Fisher and Schmincke, 1984a; Stroncik and Schmincke, 2001, 2002; Massey, 2017).

Palagonite is found in two forms, which are thought to represent two evolutionary stages of sideromelane alteration to smectites (Stroncik and Schmincke, 2001). Gel palagonite is the first to form next to the unaltered glass surface. Under a petrographic microscope, it appears transparent, isotropic, and often concentrically banded (Figs. 4B, 4C, 4E, and 4F). The interface between fresh sideromelane and gel palagonite is commonly occupied by a thin (~100 µm) zone covered with microchannels or “fingers” that is known as the “immobile product layer” (Stroncik and Schmincke, 2002). Fibro-palagonite, the following stage, is poorly translucent and slightly anisotropic, with a fibrous/lath-like/granular texture (Figs. 4A, 4C, and 4F). Electron microprobe and micro X-ray diffraction studies revealed that palagonite is not in itself a mineral, but an extremely fine-grained mixture of nontronite, saponite, Fe and Ti oxy-hydroxides, and zeolites (Hay and Iijima, 1968; Stroncik and Schmincke, 2001, 2002; Massey, 2017).

Depending on the temperature, pH, and dissolved mineral component of circulating fluids, secondary mineral precipitates commonly form in vesicles and void spaces of hydrovolcanic tuffs (Figs. 4B, 4E, and 4F). These include, for example, zeolites (chabazite, phillipsite, stilbite, analcime, mordenite, and/or heulandite), calcite, and gypsum/anhydrite. The composition and petrographic relationships between these minerals and palagonite or putative biogenic alteration textures (see next section) can help constrain fluid conditions and the timing of alteration events.

Another common sideromelane weathering phase, found both with and without palagonite, is nano-phase iron oxide (npOx). NpOx is ubiquitous on Mars; based on reflectance properties, it is thought to be responsible for Mars’ signature red color (Banin et al., 1993; Bell et al., 1993; Morris et al., 2001; Sklute et al., 2018).

Studies of basaltic glasses from seafloor and terrestrial environments revealed common microtextures, including hollow or alteration-mineral–filled microtubules and granular textures (Metevier, 2011; Fisk and McLoughlin, 2013; Nikitczuk et al., 2016, 2022; Fisk et al., 2019; Izawa et al., 2019; Figs. 4F and 5). These textures are believed to be evidence of microbial dissolution of sideromelane and/or associated palagonite and are referred to here as “putative biogenic alteration” textures (PBAs, as in, e.g., Nikitczuk et al., 2016, and Pentesco, 2019). Investigations into microbial colonization of recent basaltic glass from Iceland (Cockell et al., 2009b, 2009a; Marteinsson et al., 2015) suggest that microbial bio-films that form on exposed glass surfaces create a local microenvironment that increases the reaction rate of glass dissolution. This can speed up the palagonitization and overall alteration process and may produce the granular and tubular textures over time (Fisk et al., 2019).

Microtubules found in terrestrial basaltic glasses from Fort Rock Volcanic Field (Nikitczuk et al., 2016; Ryan et al., 2022), Western Snake River Plain Volcanic Field (Ryan et al., 2022), and Upsal Hogback, USA (Pentesco, 2019); Western Volcanic Zone, Iceland; and Carapace Nunatak, Antarctica (Nikitczuk et al., 2022) are extremely similar in morphology and size distribution to those found in seafloor basalts (e.g., Fisk and McLoughlin, 2013). Some tubule morphology examples from the Fort Rock and Western Snake River Plain volcanic fields are shown in Figure 5. Tubules typically have diameters of 1–5 µm, which concur with the sizes of bacteria or archaea that may be responsible for localized glass dissolution (Thorseth et al., 1992). In addition, tubule length and diameter sizes have a consistent log-normal distribution across the terrestrial examples listed above (for examples, see Figs. 5E and 5F) and in similar seafloor basalts. Log-normal distributions are common in many biological systems (Limpert et al., 2001); thus, this is further evidence for the potential biogenicity of these textures.

Supplemental work is needed to fully confirm the biogenicity of these textures in terrestrial hydrovolcanic glass. Nonetheless, a number of studies of microtubules in submarine glass found associated geochemical, isotopic, and organic molecule evidence for microbial metabolism products associated with them (Banerjee et al., 2011; Bebout et al., 2018; Izawa et al., 2019). These textures, if proven consistently to be biogenic in origin, would be ideal structural biosignatures. They may be preserved for hundreds of millions or even billions of years (Banerjee et al., 2006; Izawa et al., 2019) and would make hydrovolcanic glass an optimal target for astrobiologically focused Mars sample return missions (The MEPAG Next Decade Science Analysis Group, 2008; Nikitczuk et al., 2022).

While the hydrological history of Mars remains contested, it is likely that water has existed on Mars throughout its history, whether in the form of oceans, ephemeral lakes, glaciers, permafrost, groundwater, or brines (Baker et al., 1991; Craddock and Howard, 2002; Carr and Head, 2003, 2010, 2015; Fairén, 2010; Wordsworth et al., 2015; Smellie and Edwards, 2016; Galofre and Jellinek, 2016). With widespread volcanism, hydrovolcanic eruptions may have been common; for example, in the Noachian and early Hesperian periods, as magma ascended into lakes or oceans, or into the Amazonian, as magma interacted with groundwater, subsurface ice, or glaciers (Chapman et al., 2000; Wilson and Head, 2002, 2004; Head and Wilson, 2007; Brož and Hauber, 2013; Smellie and Edwards, 2016; Cassanelli and Head, 2018, 2019; Brož et al., 2021).

Tuff rings, maars, and tuff cones may have formed on early Mars as magma interacted with shallow surface water or groundwater. As an example, a study by Brož and Hauber (2013) looked at a large field of pitted cones along the Martian dichotomy boundary in the Nephentes Planum and Amanthes Cavi region (Nephentes-Amanthes cones). The morphology of the Nephentes-Amanthes cones (crater:base diameter ratio and slope) is very similar to the characteristic wide-cratered, steep-sided tuff rings and maars we see on Earth, although the cones are larger in size due to the lower gravity of Mars compared to Earth. The floors of the Nephentes-Amanthes cones are at or below the surrounding ground level, which is consistent with tuff rings or maars rather than mud volcanoes. Despite the proposed widespread nature of phreatomagmatic eruptions on Mars, it is difficult to distinguish edifices from other similar structures such as mud volcanoes or small impact craters (Brož et al., 2021). More work is needed with high-resolution imagery, remote-sensing data, and morphometric analyses to confirm phreatomagmatic structures in other areas where they are proposed, such as Niger and Dao Valles (Korteniemi and Kukkonen, 2018), Mangala Valles (Wilson and Head, 2004), Argyre Planitia (Williams et al., 2017), and Coprates Chasma (Brož et al., 2017).

Mars’ climactic and volcanic history suggests that volcano-ice interactions were relatively widespread throughout the late Noachian (Wordsworth et al., 2015), and especially during the Hesperian and into the Amazonian (Smellie and Edwards, 2016). Due to the low atmospheric pressure and temperatures at Mars’ surface through the Amazonian period, liquid water is unstable; however, water ice is present at the poles and in large volumes in the subsurface (Boynton et al., 2002; Dundas et al., 2018). Additionally, morphologic and spectroscopic evidence suggests that glaciers or ice sheets may have been present on Mars from the high- to mid-latitudes and in high-altitude, low-latitude regions (Head and Marchant, 2003; Shean et al., 2005, 2007; Forget et al., 2006; Head and Wilson, 2007; Smellie and Edwards, 2016). While interactions between subsurface ice and ascending magma would likely produce phreatomagmatic eruptions (see Phreatomagmatic Eruptions section above), lahars, jökulhlaups, and meltwater outflow channels, subglacial eruptions would form sheets of pillow lavas, tuyas, and tindars just as on Earth, albeit larger in scale due to lower gravity (Wilson and Head, 2007; Smellie and Edwards, 2016).

Numerous landforms on Mars resemble the glaciovolcanic constructs we see on Earth, including tuyas, tindars, and potential lahar flows. Mesa-like features resembling tuyas were identified in Mars’ northern plains, especially in Chryse and Acidalia Planitiae. They range from 1.5 km to 8 km in width and 100 m to 500 m tall and appear to be draped in talus (Martínez-Alonso et al., 2011; Smellie and Edwards, 2016). These areas, along with the northwestern plains near Arsia Mons, also include other evidence of glaciation and glaciovolcanism, such as eskers and tindars (Scanlon et al., 2014, 2015).

An area called Sisyphi Planum in the south circumpolar region includes an undifferentiated late Noachian–late Hesperian-aged unit and the Dorsa Argentea Formation. This area features 21 mountainous edifices called the Sisyphi Montes, identified by Ghatan and Head III (2002) as being potentially glaciovolcanic in origin based on their morphology of broad, flattened summits with steep sides. Remote spectral data analyzed by Ackiss et al. (2018) indicate the presence of smectite–zeolite–iron oxide-dominated mineralogy, which is suggestive of glaciovolcanic hydrothermal alteration and potential palagonitization.

Home Plate is the most extensive exposure of layered bedrock encountered by the National Aeronautics and Space Administration (NASA) Mars Exploration Rover–A (MER-A), Spirit, in Gusev Crater. At 2–3 m high and ~90 m in diameter, it is a prominent feature in the Columbia Hills region (Squyres et al., 2007). Textural characteristics—including a putative bomb sag, accretionary lapilli (rounded lower unit granules), cross and planar bedding, and the presence of basaltic glass (see figs. 2 and 3 of Squyres et al., 2007)—suggest it was formed through an explosive hydrovolcanic process when alkali-rich basaltic magma interacted with an aqueous fluid (Squyres et al., 2007). The structure of Home Plate, with beds dipping toward the center of a roughly circular shape, could be indicative of a tuff ring or maar, or a small impact crater filled with pyroclastic material (Squyres et al., 2007). The upper, fine-grained unit may have formed through aeolian reworking of the lower unit’s pyroclastic deposits.

Geochemical and mineralogical variations across Home Plate indicate post-emplacement hydrothermal activity, possibly in multiple stages (Schmidt et al., 2008, 2009; Morris et al., 2008). Iron phases in the eastern Home Plate section predominantly comprise magnetite and pyroxene, with lesser npOx and hematite. While basaltic glass is found in eastern Home Plate, the npOx that is typically expected as a breakdown product of glass was not identified by Mössbauer spectrometry from Spirit. This could also have devitrified to form magnetite (Schmidt et al., 2009). The higher olivine, basaltic glass, npOx, and lower magnetite component at western Home Plate suggest that it was not persistently saturated with aqueous fluids; otherwise, the olivine would have dissolved. NpOx was likely formed by devitrification of basaltic glass under a low-temperature regime, similar to how palagonite forms (Schmidt et al., 2009).

Opaline silica deposits found in the Eastern Valley below Home Plate (e.g., Elizabeth Mahon) are found in altered volcanic ash deposits. These opaline deposits are light-toned and associated with nodular outcrops (Squyres et al., 2008). Because of their close association with volcanic materials and ferric sulfates, they are interpreted to be hydrothermal in origin (Squyres et al., 2008; Ruff et al., 2011). Ruff et al. (2011) used multiple observations of stratigraphy, geochemistry, and texture of the silica deposits to suggest a sinter origin: precipitation from a near-neutral pH silica solution originating from a hot spring or geyser, which lends further credence to the significance of water in the origins and alteration of Home Plate.

The presence of amorphous silica (tridymite) in Gale Crater, detected by the NASA Mars Science Laboratory (Curiosity rover), was interpreted as being distal or detrital material from volcanic eruptions (Morris et al., 2016; Payré et al., 2022). This suggests that amorphous silica alteration products are not necessarily directly related to volcanism in situ (Yen et al., 2021). These products may result from hydrothermal alteration of basaltic sediments or detrital silicic material (Hurowitz et al., 2017; Rampe et al., 2017; Bedford et al., 2019) without the presence of nearby volcanic activity.

Beginning with the Smithsonian Institution Global Volcanism Program’s (GVP) Pleistocene and Holocene volcanoes lists (Venzke and Smithsonian Institution, 2020), we filtered potentially relevant volcanic fields based on volcano type (cones, complex, pyroclastic cone, tuff cone, etc.) and composition (basalt, basaltic andesite, trachybasalt, etc.). We then examined peer-reviewed literature available through Google Scholar on each potential entry, looking for keywords such as hyaloclastite, tuff cone/ring, phreatomagmatic, palagonite, and/or tuya to determine each entry’s relevance. Sites with enough supporting literature to meet our criteria were entered into our database (Table 2). Volcanic fields with known PBAs were noted in our table.

We mapped each individual hydrovolcanic structure using Esri ArcGIS Pro, with many published maps as references. We compiled available weight percent oxide concentrations for major elements SiO₂, TiO₂, Al₂O₃, FeO^total (or Fe₂O₃ converted to FeO), MnO, MgO, CaO, Na₂O, K₂O, and P₂O₅ in Microsoft Excel. We normalized these data to their totals and created total alkali-versus-silica (TAS) and FeO^total-versus-SiO₂ charts for each geographic region.

For comparison, in situ weight percent oxide compositions from Martian igneous rocks were plotted. We chose in situ rover data from probable volcaniclastic rocks and basalts at Gusev Crater (Ming et al., 2006; McSween et al., 2006), volcaniclastic and float rocks from Gale Crater (Sautter et al., 2015; Cousin et al., 2017; Berger et al., 2020), and basaltic lava samples from Jezero Crater (Farley et al., 2022; Liu et al., 2022; Wiens et al., 2022; Simon et al., 2023). We also included data from shergottite meteorites as compiled in Udry et al. (2020), as they represent likely shallow intrusive or extrusive materials, which makes their compositions relevant to volcanism. Rover data were not normalized, and their errors are reported, while shergottite lab data were normalized to their totals.

File S2 in the Supplemental Material¹ contains all of these data organized by global region. We also searched our collected literature for mentions of alteration minerals and included those results in Table 3.

Figure 6 and Table 2 show all 45 entries in our global database. Table 2 is organized geographically by region underlined, e.g., “Middle East–North Africa” (including Türkiye, Saudi Arabia, and Sudan) and “Southeast Asia” (including Indonesia and the Philippines) and subregion/administrative division in italicized heading, e.g., “United States of America–Oregon,” “United States of America–California,” or “Argentina–Mendoza.” The regional distinctions are primarily based on the “Region” and “Sub-Region” classifications in the original GVP databases. Some entries are administratively part of one country but geographically located in a different region than that country’s mainland. For example, O’ahu, Hawai’I, is part of the United States of America but is included in the “Pacific Islands” geographic region. These regional distinctions were truncated in Table 2 but are expanded in File S6 (see ^{footnote 1}).

Seven of the volcanic fields had documented PBAs: Wells Gray–Clearwater Volcanic Field, Canada; Reykjanes Volcanic System and Western Volcanic Zone, Iceland; Carapace Nunatak, Antarctica (all subglacial); and Fort Rock Volcanic Field, Western Snake River Plain Volcanic Field, and Upsal Hogback, USA (all phreatomagmatic; Cockell et al., 2009b; Nikitczuk et al., 2016, 2022; Massey, 2017; Pentesco, 2019; Ryan et al., 2022). We noted 28 entries in addition to these PBA-containing sites with palagonite and/or sideromelane mentioned or photographed in the literature. The volcanic fields range in age from 8.9 Ma (Icefall Nunatak, Antarctica; Smellie, 2001) to the 1977 phreatic eruption of Taal, Luzon, Philippines (Delos Reyes et al., 2018). An exception is the 180 Ma Carapace Nunatak structure in Antarctica, which falls far outside of our Quaternary age criteria but was shown to contain PBAs and to have erupted in a lacustrine setting (Nikitczuk et al., 2022). In total, we mapped 716 individual volcanic features distributed in 45 volcanic fields; map data for all features are available as an ArcGIS-compatible, point-type feature class in File S3 (see ^{footnote 1}).

Thirty-one of the sites had one or more papers documenting either glass or unaltered bulk-rock geochemical data obtained through a variety of laboratory methods: X-ray fluorescence spectroscopy (XRF), electron probe microanalysis (EPMA), laser-induced breakdown spectroscopy (LIBS), energy-dispersive X-ray spectroscopy (EDS), mass spectrometry (MS), inductively coupled plasma–mass spectrometry (ICP-MS), atomic emission spectroscopy (AES), and wet chemical methods. Some of these methods were or are being used on Mars missions: The Planetary Instrument for X-ray Lithochemistry (PIXL) instrument on NASA’s Mars 2020 Perseverance rover uses XRF, and ChemCam on Curiosity and SuperCam on Perseverance use LIBS (Maurice et al., 2012; Wiens et al., 2017; Allwood et al., 2020). Methods are reported (where available) in File S2. All Earth-based data were normalized to their totals. Mars Alpha Particle X-Ray Spectrometer (APXS) and shergottite data were normalized, while ChemCam, SuperCam, and PIXL data were not. Errors for Mars data (Curiosity APXS, ChemCam; Perseverance PIXL) are reported.

Figures 7 and 8 are two examples of geographic region maps with associated geochemical charts: North America is shown in Figure 7, and Iceland in Figure 8. The remaining 12 map/chart figures are available upon request from the corresponding author.

TAS and FeO^total versus SiO₂ were plotted first in Microsoft Excel as scatter plots, and then charts were imported into Affinity Designer, a vector graphics program, and outlines were drawn around the scattered data points at each volcanic site to create the colored fields displayed in Figures 7B–7E, 8B, and 8C. The fields could be layered on top of one another for easier visualization of compositional intersections between volcanic sites. Moreover, all seven sites with confirmed PBAs were plotted together in their own TAS and FeO^total-versus-SiO₂ charts, and then a “PBA overlap” field was drawn around the area where the scattered points clustered together (Fig. 9) to represent the range of compositions already recognized as being associated with PBAs. These PBA overlap fields were used similarly to the Mars site fields: Individual volcanic site fields were layered over the PBA fields to check for compositional overlap.

The TAS charts were used to provide the volcanic rock classifications based on Na₂O + K₂O (alkali component) on the y-axis and SiO₂ on the x-axis, with field positions drawn from Le Maitre (2002). Most of the volcanic sites plotted within the basalt, trachybasalt, and tephrite/basanite fields, with some stretching into foidite, basaltic trachyandesite, and basaltic andesite fields (see Figs. 7B, 7D, and 8B); most sites also had a mix of alkaline and subalkaline compositions. Data from the Mars rovers show Gusev Crater plotting mainly in basalt with some overlap into trachybasalt; Gale Crater covers a broad range through basanite, tephrite, phono-tephrite, trachybasalt, basaltic trachyandesite, trachyandesite, trachydacite, basaltic andesite, andesite, and dacite. Jezero Crater stretches through foidite, basanite, basalt, trachybasalt, basaltic trachyandesite, and trachyandesite. Jezero Crater compositions contain excess salts (Farley et al., 2022; Scheller et al., 2022), which may impact where they plot within both this and the FeO^total-versus-SiO₂ diagram (see Figs. 7B–D, 8B–C). Shergottite meteorites plot within picro-basalt, basalt, and basaltic andesite fields, with relatively low total alkalis. It is important to note that the Mars data (with the possible exception of Gusev Crater) are not representative of Martian hydrovolcanic compositions but rather volcanic (or volcaniclastic) compositions in general. Other hydrovolcanic eruptions on Mars are likely to have compositions that plot near or within the data shown here. All of our sites overlap the four Mars fields to some degree, especially considering the wide range of compositions represented in Gale Crater, which makes these sites all the more relevant to Mars-analogue research.

FeO^total versus SiO₂ is plotted for two reasons: It represents the most significant compositional difference between Earth and Mars (Dreibus and Wänke, 1987; Treiman et al., 2000; Filiberto and Dasgupta, 2011; Hirschmann, 2022), and iron is a redox-sensitive element that may represent an important energy source for chemolithotrophic microorganisms (Nixon et al., 2012). Nearly all sites (see examples in Figs. 7C, 7E, and 8C) intersect in composition with the Mars fields, and mainly with Gusev and Gale data, as Jezero and shergottites are relatively enriched in iron.

The PBA overlap fields in TAS and FeO^total-versus-SiO₂ spaces represent the ranges of known compositions in volcanic fields where PBAs were found (Fig. 9). Comparing the compositions of the remaining 24 terrestrial volcanic fields with available geochemical data to the seven known PBA sites allows us to narrow down the sites that may be more likely to contain PBAs. In the TAS chart, the PBA overlap field sits within the basalt and trachybasalt fields on both sides of the alkaline–subalkaline dividing line (Figs. 7B, 7D, and 8B). Tuya Volcanic Field, Lake Tahoe, Black Point/Mono Lake, Black Rock Desert, and Potrillo Volcanic Field (Fig. 7D); Vestmannaeyjar Archipelago (Fig. 8B); and Isla Isabel, Serdán-Oriental Basin Volcanic Field, Llancanelo Volcanic Field, Karapınar Volcanic Field, Kamo Volcanic Field, Cheju Do, Taal, Lamongan Volcanic Field, Auckland Volcanic Field, South Auckland Volcanic Field, O’ahu, and Icefall Nunatak all have some overlap with the PBA site fields in TAS space. Eighteen volcanic site fields overlap the PBA field in the FeO^total-versus-SiO₂ chart, which suggests that a range of iron levels may be suitable for PBA formation and preservation. These include Tuya Volcanic Field, Lake Tahoe, Black Point, Black Rock Desert (Figs. 7C and 7E), Vestmannaeyjar Archipelago (Fig. 8C), and others.

Table 3 presents the alteration minerals identified in the 24 volcanic fields in which they are documented, along with the alteration conditions implied by those minerals. Twelve of these entries have named zeolites, which allowed us to use the temperature stability figures from Chipera and Apps (2001) to approximate the formation temperature ranges of these mineral assemblages (Fig. 10). In other cases, post-deposition alteration conditions are described in the literature. The most common minerals include calcite, gypsum/anhydrite, smectites (montmorillonite, saponite, and nontronite), opal/silica, analcime, chabazite, stilbite, and phillipsite, all of which were found on Mars in varying abundances (Bibring et al., 2006; Ehlmann et al., 2009; Carter et al., 2013; Bridges et al., 2015; Rampe et al., 2017; Horgan et al., 2020; Udry et al., 2020; Tice et al., 2022).

In many cases, multiple phases of alteration were identified, indicating neutral to alkaline pH and varying temperatures below 100 °C. The abundance of palagonite at different phases (gel and fibrous) suggests that this lower-temperature hydrothermal regime was more common in these settings than high-temperature alteration, which would result in total alteration to smectites (Stroncik and Schmincke, 2001, 2002). For example, Nikitczuk et al. (2016), writing about Fort Rock Volcanic Field, USA, documented zoned amygdules in plagonitized coarse ash clasts. Vesicles were rimmed on the outer edge (into the glass) with palagonite, while their interiors were either filled completely with calcite or rimmed on the inner edge with chabazite and filled in the center with calcite (see fig. 5 of Nikitczuk et al., 2016). Chabazite forms at pH 9–10 and at temperatures of 10–80 °C; calcite over a range of 0–270 °C; and smectites, the end product of palagonite, in this case formed below 80 °C, based on the lack of olivine alteration (Chipera and Apps, 2001; Nikitczuk et al., 2016). It appears that calcite was precipitating both during and after the palagonitization/zeolitization phases, and most alteration occurred at temperatures ≤80 °C (Fig. 10). At Upsal Hogback, USA, Pentesco (2019) noted abundant calcite-filled voids and vesicles in palagonitized tuff samples, as well as the opposite—palagonitized glass shards contained within carbonate lake-margin tufa deposits—that suggest ongoing fluid circulation and precipitation of calcite. Based on the zeolites present, most alteration occurred between 68 °C and 99 °C (Pentesco 2019; Fig. 10). Risso et al. (2008) showed chabazite crystals in calcite cement and analcime cement between palagonitized glass clasts in base-surge deposits. Notably, zeolites and palagonite were mostly absent from dry fallout deposits between the wet phreatomagmatic beds, which indicates that hot, condensed water from the phreatomagmatic eruptions was the primary agent of alteration (Risso et al., 2008; Fig. 10).

The presence of similar mineral phases in these Earth-based environments compared to those seen at hydrovolcanic environments on Mars indicates that alteration processes on both planets are comparable, with some key differences. For example, the mineralogy at the putative glaciovolcanic landforms in Sisyphi Planum includes smectite–zeolite–iron oxide materials (possibly palagonite), gypsum, and polyhydrated sulfates, which have been interpreted as forming in a volcanic hydrothermal environment (Ackiss et al., 2018). Home Plate, which is the most likely hydrovolcanic structure studied in situ, shows evidence of both high- and low-temperature hydrothermal alteration. The main differences between Home Plate and the sites reviewed here are alteration under both acidic and alkaline conditions (Ming et al., 2006; Ruff et al., 2011), devitrification of glass to npOx as opposed to palagonite, and the importance of brines for the transportation of volatile species (Schmidt et al., 2008, 2009; Yen et al., 2008; Morris et al., 2008; Ming et al., 2008). The igneous rocks of Jezero Crater contain alteration materials that indicate multiple aqueous alteration phases: Fe-Mg carbonate rims around olivines in the Séitah formation imply weakly acidic-to-neutral pH, CO₂-rich circulating fluids with possibly a range of temperatures, while sulfate and perchlorate salt deposits in the Máaz formation could suggest several fluids of different compositions circulating at different times in a near-surface brine evaporation environment (Farley et al., 2022; Scheller et al., 2022).

Our complete database is available in the Supplemental Material and includes the list of all volcanic fields, their associated geochemical and mineralogical data (if available), Google Earth .kml and ArcGIS .gdb files for each hydrovolcano, and geochemistry charts (TAS and FeO^total versus SiO₂) of each global region.

Our work here involves significant literature review and comparison of data from over 100 sources for our final 45 volcanic sites. We mapped every individual hydrovolcanic structure in Google Earth and ArcGIS, extracted geochemical data for comparison among sites, and scoured references for mineralogical and alteration information. This work is of interest to the Mars volcanology community, as we provide a list of accessible, relevant, and well-preserved volcanic sites where composition, alteration, and eruption dynamics like those found throughout Mars’ history can be studied in situ. Our database is also relevant to the geomicrobiology/astrobiology community, as we show a range of compositions and alteration conditions within hydrovolcanoes that preserve potential biosignatures in volcanic glass. Our goals for creating this database are twofold: (1) to compare the compositions and alteration conditions present at terrestrial hydrovolcanic fields and identify the conditions that are common to PBA-bearing sites, and (2) to provide a jumping-off point for future researchers who wish to go to the field to find new PBA-bearing sites. Our results, discussed here, satisfy our first goal; our second goal is to inspire future work by planetary volcanologists and geomicrobiologists.

We combined our analyses of available geochemical and mineralogic/alteration data to produce Table 4, which illustrates the volcanic fields without known PBAs that show the greatest potential to contain PBAs in a Mars-relevant context, based on our scoring criteria. Geochemical overlap with sites that already have documented PBAs is considered a high priority, as glass composition (especially iron content) is likely to have an impact on nutrient availability for lithotrophic microbes; hence, candidate sites with compositions that overlap those of PBA-bearing sites are assigned two points for each compositional space (TAS and FeO^total versus SiO₂). Intersection with Mars composition is assigned one point per Mars site (Gusev, Gale, and Jezero craters; shergottites) in each compositional space. Available data on alteration minerals, and mention of sideromelane/palagonite in the literature, are each assigned one point.

While our current understanding of the timing of PBA formation is incomplete, microbial colonization experiments with 0.8 Ma Icelandic hyaloclastites (Cockell et al., 2009a) indicate slow growth of microbes over months of incubation, and the presence of small PBA textures in the glass. Drill-core data collected in 2017 from the tuff cone of Surtsey, a volcanic island on the Vestmannaeyjar Archipelago, Iceland, which erupted from 1963 to 1967, indicates the presence of diverse microbial communities throughout the length of the drill core. The core is up to 191.85 m depth from the surface (133.85 m below sea level) and 124 °C (Bergsten et al., 2021, 2022). However, while putative biofilm- and filament-like structures were found within the vesicles of some Surtsey samples (Bergsten, 2022), hollow tubular or granular boring textures in the glass have yet to be reported. Therefore, we added the additional constraint of awarding one point to sites with ages >1000 years, which allows more time for PBAs to form.

Five sites had 12/14 points each: Tuya Volcanic Field, O’ahu, Black Rock Desert Volcanic Field, Auckland Volcanic Field, and South Auckland Volcanic Field. Six sites had 11/14 points each: São Miguel, Icefall Nunatak, Karapınar, Black Point, Vestmannaeyjar Archipelago, and Llancanelo Volcanic Field. These top sites are highlighted in light gray in Table 4. We recommend these eleven sites as ideal starting points for future investigations into Mars-relevant habitable hydrovolcanic environments. The availability of already documented geochemical and mineralogical data for these particular sites is what elevates them above the others in our database, as it eliminates a degree of uncertainty when planning a field campaign. However, any of the remaining 22 sites in our database not listed in Table 4 have the potential to contain PBAs, and we recommend further investigation into pre-existing samples from these areas.

Our database entries are based on peer-reviewed literature that confirms the presence of phreatomagmatic/subglacial volcanic structures of the correct age range (Pleistocene–recent) and composition (basaltic) within the given volcanic field. We attempted to additionally establish the presence of palagonitized tuff in each of our sites, and to include complementary geochemical and mineralogical data. The greatest gaps in our database pertain to the post-eruptive hydrological environment and alteration of the glass. Most literature focuses on eruptive mechanisms and physical volcanology, eruption hazards to populations, and lava geochemistry as tools for probing the magmatic sources and tectonic processes. As such, 21 of the sites included in this list lacked documentation of secondary alteration minerals beyond the presence of palagonite, which is critical in determining the post-eruptive conditions. These gaps would be best addressed with study of petrographic relationships among fresh glass, palagonite, and other alteration products, such as calcite and zeolites in tuff samples.

Furthermore, not all volcanic sites listed in the GVP database that satisfied our initial search criteria had adequate literature available in English to review. This was especially common with potential sites in Russia, China, and regions with significant political unrest such as sub-Saharan Africa. This deficiency is reflective of an overall bias in scientific literature toward the “Global North.”

Currently, the literature documenting PBAs in terrestrial environments is scant, despite the potential ubiquity of these features based on how common they are in submarine volcanic glasses. The more we expand the record of PBAs in a terrestrial setting, the more we can constrain the range of conditions in which they form and are preserved. Here, we show the currently known compositions in which PBAs form (Fig. 10), but we do not yet understand why these compositions appear to be preferred. Continuing study of volcanic glasses and their colonization with microbes (such as in Iceland, e.g., Cockell et al., 2009a; Cousins et al., 2013; Bergsten et al., 2021, 2022) may elucidate the factors that lead to their formation (for example, the availability of nutritive elements that contribute to this apparent preference for basaltic glasses). This is an important area of future research.

Moreover, we can compare geochemistry, mineralogy, and alteration textures from terrestrial hydrovolcanoes with PBAs to the submarine PBA-bearing glasses (Metevier, 2011; Bebout et al., 2018; Izawa et al., 2019). This would provide us with more information on how microbes preferentially colonize volcanic glasses. There are also other potentially important factors that correlate with PBAs (such as stable isotopes, trace elements, water temperature, and duration of hydrothermal alteration) that should be examined.

Our database, and future site-by-site comparisons of PBA-bearing environments, will give researchers direction when choosing a potential landing site or sample on an astrobiologically focused Mars mission. Water on Noachian and early Hesperian Mars may have existed predominantly in the form of ephemeral lakes, groundwater, subsurface ice, and glaciers (Ehlmann et al., 2011; Carter et al., 2013; Michalski et al., 2013, 2022; Grotzinger et al., 2014; Salese et al., 2019). Over time, Mars’ hydrology evolved to its current drier state, where water is found as subsurface ice, brines, and glaciers (Michalski et al., 2013; Bramson et al., 2015; Rivera-Valentín et al., 2020; Morgan et al., 2021). Hydrovolcanoes should have been common throughout Mars’ history, and hence would preserve potential biosignatures if Mars were once inhabited by microorganisms. Therefore, understanding the distinct characteristics of terrestrial hydrovolcanoes that preserve PBAs can help us to prioritize sites that are more likely to contain biosignatures. Our study can be compared to complementary research on the spectroscopic, mineralogic, and petrographic properties of similar Mars-analogue materials, as in Farrand et al. (2018).

Additional factors make these sites ideal targets for studying astrobiology and geomicrobiology. Firstly, unlike seafloor basalts, which can only be sampled through expensive underwater drilling methods, terrestrial hydrovolcanoes are accessible at the surface. Their structures can be studied through personal field observations, and samples that represent a range of alteration conditions within a small area can be precisely collected. Secondly, PBAs as structural or textural biosignatures are not subject to the same chemical degradation factors, such as exposure to radiation (i.e., at Mars’ surface), which may damage complex organic molecules. As biological mechanisms for their formation continue to be explored (Cockell et al., 2009b; McLoughlin et al., 2010; Fisk et al., 2019), these textures could be considered strong candidates for potentially accessible Martian biosignatures.

We integrated available literature covering many aspects of Quaternary terrestrial basaltic hydrovolcanism on a global scale: geochemistry, mineralogy, hydrologic environment, alteration history, and physical volcanology. The 45 sites we chose to include in our database have both the features necessary to support microbial incursion into volcanic glass and approximate the probable conditions of Mars’ past. We found palagonite, zeolites, calcite, smectites, gypsum/anhydrite, and silica to be common alteration products across many sites, which are indicative of lower-temperature, alkaline pH aqueous alteration of sideromelane tuffs. Our sites contain tuyas, tindars, maars, tuff cones, and tuff rings. Seven sites have documented putative biogenic alteration textures in sideromelane: Wells Gray–Clearwater Volcanic Field, Canada (Massey, 2017); Upsal Hogback (Pentesco, 2019), Fort Rock Volcanic Field (Nikitczuk et al., 2016; Ryan et al., 2022), and Western Snake River Plain Volcanic Field, USA (Ryan et al., 2022); Reykjanes Peninsula Volcanic Field (Massey, 2017; Nikitczuk et al., 2022) and Western Volcanic Zone, Iceland (Cockell et al., 2009b); and Carapace Nunatak, Antarctica (Nikitczuk et al., 2022).

By comparing available geochemical and mineralogical data from the remaining 38 sites in our database to those known to contain PBAs and to igneous rocks from Mars rover landing sites, we developed a short-list of 11 volcanic fields that we think have the highest relevance for future investigations. These are the Auckland and South Auckland volcanic fields, New Zealand; O’ahu, Hawai’i; Tuya Volcanic Field, Canada; Black Rock Desert and Black Point, USA; São Miguel Volcanic Field, Azores; Icefall Nunatak, Antarctica; Karapınar Volcanic Field, Türkiye; Vestmannaeyjar Archipelago, Iceland; and Llancanelo Volcanic Field, Argentina. We recommend that glassy tuff samples from these sites should be examined for the inclusion of PBAs. Expanding the record of terrestrial hydrovolcanic environments that contain PBAs will help us constrain the conditions under which they form, and hence inform site and sample selection on future Mars astrobiology missions.

C.H. Ryan was funded by the Queen Elizabeth II Graduate Scholarship in Science and Technology for 2021–2023. Additional funding was provided by a Natural Sciences and Engineering Research Council of Canada Discovery grant (no. RGPION-20116–03948) awarded to M. Schmidt. We thank Roberta Flemming for early guidance and overview of this project, and reviewers Candice Bedford, Jim Head, and an associate editor for their time and insightful suggestions. We also thank Elizabeth Sutherland of the Western Libraries Map and Data Centre, London, Ontario, Canada, for her GIS wizardry.