Early Diagenesis by Modern Acid Brines in Western Australia and Implications for the History of Sedimentary Modification on Mars
Published:January 01, 2012
- PDF LinkChapter PDF
Brenda B. Bowen, Kathleen C. Benison, Stacy Story, 2012. "Early Diagenesis by Modern Acid Brines in Western Australia and Implications for the History of Sedimentary Modification on Mars", Sedimentary Geology of Mars, John P. Grotzinger, Ralph E. Milliken
Download citation file:
Mineralogical and geochemical data collected from multiple sites on Mars suggest that acid saline surface waters and groundwater existed there in the past. The geologic context and sedimentology suggest that these acid saline waters were associated with groundwater-fed ephemeral lakes. Ephemeral acid saline lakes in southern Western Australia (WA) are some of the few known natural systems that have the same combination of extreme acid brine chemistry and lacustrine depositional setting as is observed on Mars. Thus, the WA acid saline environments provide a modern analog for understanding past depositional and diagenetic processes that may have occurred on Mars. Here, we examine surface sediments and sedimentary rocks that have been in contact with acid (pH down to ∼1.5) and saline brines (total dissolved solids up to ∼32%) in southern Western Australia. Through sedimentological, mineralogical, geochemical, and petrographic analyses, we identify the impacts of early diagenesis in and adjacent to eight acid saline lakes and evaluate the processes that have been important in creating these deposits. The combination of extreme chemistry, spatial variability, arid climate, and reworking by winds and floods contributes to make spatially complex depositional products that are a combination of siliciclastics and chemical sediments. Important syndepositional and very early diagenetic processes in these settings include the chemical precipitation of minerals from shallow groundwaters to form displacive crystals and cements, dissolution/partial dissolution of chemical sediments, replacement/partial replacement of some minerals, cracking due to repeated wetting and drying, and the formation of iron-oxide concretions. Minerals observed in these sediments include a variety of chlorides, sulfates, iron oxides, and phyllosilicates, many of which have textures and mineral associations that suggest authigenic formation. These observations are supported by the chemistry of the modern acid brines, which appear to be supersaturated with respect to these minerals. The range of early diagenetic products, compositions, and textures that are apparent in the WA acid saline lake sediments may provide insights into the processes that influenced the sediments on Mars and the timing of sedimentary formation processes on Mars.
Geologic Setting, Surface Processes, and Extreme Water Chemistry
Modern shallow saline lakes are abundant throughout the Archean Yilgarn Craton in southern Western Australia (WA). The lake waters and shallow groundwaters range in pH from 1.5 to ∼9 and from nearly fresh to hypersaline (Benison et al. 2007; Fig. 1). They are Na-Cl to Na-Mg-Cl-SO4 brines with variable yet locally high Ca, Br, Al, Fe, and Si (Bowen and Benison 2009; Fig. 2). Many of the lakes are seasonally dry, and even when they do have standing surface water, they are very shallow, with water depths typically less than 1 m deep. Surface waters cover areas as small as 1 km2, although some lakes occupy basins up to ∼700 km2. Many of the lakes occupy closed depressions within buried fluvial inset valleys that were carved into the Archean basement (Fig. 3). Multiple mechanisms have been proposed for the initial formation of the valleys, including erosion from Permian glaciers (McArthur et al. 1989) and/or Jurassic to Middle Eocene fluvial incision (Woodward 1897, Gregory 1914, Clarke et al. 1996, de Broekert and Sandiford 2005). The Yilgarn Craton is characterized by subtle topography and highly weathered regolith, and it is dominantly an erosional vs. a depositional system (Anand and Paine 2002). Some of the lakes occur above paleovalley-hosted basins that locally contain marine, fluvial, and lacustrine sediment accumulations up to 100 m thick (McArthur et al. 1991, Clarke 1993), although other lakes occur directly on top of Archean basement outcrops and contain nearly no sediment (Fig. 3). Ephemeral lacustrine deposition may have existed in some of the basins since the mid-Miocene or earlier, and since that time, this region has experienced relative tectonic inactivity and increasing aridity (De Deckker 1983).
All of the WA lakes are ephemeral, saline, and fed by a combination of rainwater and diffuse influx of regional acid saline groundwaters. The range in pH in the lakes is due to varying contributions of acidity that originates in groundwaters from subsurface oxidation of pyrite and precipitation and dissolution of iron and aluminum hydroxides (e.g., Mann 1983, McArthur et al. 1991, Peiffer et al. 2009). Precipitation/evaporation ratios on the Yilgarn Craton are very low (P/E = 0.026 to 0.22), which is reflected in the high salinity of the lakes, although stable isotope data demonstrate that dissolution of evaporites with the influx of fresh meteoric water generates the highest salinity lake waters (Bowen and Benison 2009). The low-relief topography in the region leads to extremely low hydrologic gradients and the formation of the highly evolved acid groundwaters (Gray 2001).
The surface processes and modern depositional facies that exist in and around these lakes typically include lake, mud flat/sand flat, ephemeral channel, and dune facies (Benison et al. 2007; Fig. 4). In addition, many of the lakes have lithified and partially lithified sedimentary rocks exposed around the fringes and just below the surface. The detrital surface sediments appear to be locally derived, and eolian transport is one the dominant processes for moving sediments intra- and interbasinally. All of these sediments and rocks are exposed to extreme and fluctuating acid saline waters. The conditions that these environments experience change drastically through the year as the lakes go through multiple flooding, evapoconcentration, and desiccation cycles.
Importance of Western Australia Acid Brine Environments as a Mars Analog
Recent data from surface sediments and sedimentary rocks on Mars provide intriguing evidence of extraterrestrial weathering, depositional, and diagenetic processes. Orbital spectral data and ground-based rover data suggest that while water was certainly involved in the formation of many of these deposits, at least some of the fluids had an extreme acid saline chemistry (Grotzinger et al. 2005, McLennan et al. 2005, Squyres et al. 2009). Similar to Western Australia, the generation of acidity on Mars can be explained by surface oxidation of Fe2+-rich groundwater (Hurowitz et al. 2010). The salinity may be a result of long-term weathering of basement rocks and subsequent weathering of the weathering products (e.g., Altheide et al. 2010). While the presence of water has long been considered a prerequisite for the existence of extraterrestrial life, the extreme chemistry of these waters has called into question the likelihood of life existing and/or being preserved in environments exposed to these hostile conditions (Sumner 2004, Tosca et al. 2008).
Interpretation of sedimentary data from Mars requires knowledge of sedimentary processes on Earth. Terrestrial analogs provide ground truth to constrain interpretations of the type of processes that have been important on Mars. This type of comparative sedimentology (using modern to understand ancient, outcrop to understand subsurface, experimental to understand theory, or terrestrial to understand planetary) was achieved for many depositional environments decades ago, facilitating a uniformitarian approach for interpreting sedimentary records. However, acid saline lake systems are rare environments that have only recently been recognized in both the modern and ancient rock record (McArthur et al. 1991, Benison and Goldstein 2002, Risacher et al. 2002, Benison et al. 2007). Although many of the physical processes are analogous to those in other saline lake settings (e.g., influence of rock type, climate, and degree of evaporation), the acid brines and associated sedimentary deposits evolve via geochem-ical pathways distinctly different from other saline systems (e.g., Long et al. 2009). While lakes with varying geochemistry within the same geographic location are not uncommon (e.g., Last 1994), the range in salinity and pH in WA and the resulting variations in elemental concentrations are quite unusual and affect the resulting sedimentary records. The early diagenetic processes, that is, the mineralogical, textural, and geochemical modifications that occur within pore space and around detrital grains as they interact with the acid saline surface and pore fluids, result in an assemblage of Fe-, Al-, and Si-rich minerals not typically expected in a saline system.
As some of the only known naturally acidic and saline modern depositional and diagenetic systems on Earth, the abundant ephemeral lakes in WA provide important analogs for past acid saline environments on Mars (Benison and LaClair 2003, Benison and Bowen 2006, Bowen et al. 2008, Baldridge et al. 2009, West et al. 2009). In addition to the acid saline chemistry, the depositional environment in WA shares many characteristics with those that likely existed on Mars in the past. In WA, groundwater-fed interdune lakes desiccate and erode to supply the surrounding dunes with reworked evaporite grains; this scenario is also envisioned for the formation of many sedimentary deposits on Mars (Grotzinger et al. 2005, Squyres and Knoll 2005, Andrews-Hanna et al. 2007, Fishbaugh et al. 2007). Both systems contain a unique assemblage of chemical sedimentary minerals, including Ca- and Mg-sulfates (Baldridge et al. 2009), chlorides (Clark and Van Hart 1981, Bell et al. 2000, Osterloo et al. 2008), Fe-minerals including smectites, hematite, and jarosite, and Al/Si-precipitates such as alunite and kaolinite (Story et al. 2010).
Of course, there is no perfect terrestrial analog for the unique conditions on Mars, and it is important to understand the differences between the acid saline systems in WA and on Mars. One of the most significant differences between the WA analog and Mars is the nature of the basement material. While Mars is dominated by mafic igneous rocks, the WA acid lakes are hosted by a variety of rock types, most commonly felsic, but some rocks of intermediate and mafic composition as well (Table 1). This difference in starting compositions influences all later weathering products, precipitates, and fluid chemistry, and no doubt leads to significant differences in the specific geochemical reactions that occur. For example, in WA, Al-phyllosil-icates tend to be the most common phyllosilicate mineral (Story et al. 2010) due to the dominance of highly weathered felsic bedrock. In contrast, some work has suggested that Mg-phyllosilicates are the most common phyllosilicate mineral on Mars due to the dominance of less weathered mafic volcanic bedrock (Baldridge et al. 2009). Other significant differences include the overwhelming long-term effects of tectonics, a comparatively warm climate, and the influence of biology in WA. Despite these differences, sediments exposed to extreme acid saline fluids in terrestrial ephemeral interdune lacustrine environments exhibit some characteristic early diagenetic features that may help to elucidate similar processes that have been important on Mars and highlight the potential for preservation of biological materials in these types of settings.
By studying the surface sediments in contact with acid brines in WA, we can investigate the complex relationships among detrital lithology, authigenic mineralogy, and water chemistry. Here, we can examine the spatial relationship between typical evaporite minerals such as halite and gypsum, and their relationship to the minerals specific to acid environments such as iron oxides and iron sulfates. It can be very difficult to definitively distinguish between detrital and authigenic components in highly weathered settings with complex mineralogy such as this (e.g., Story et al. 2010), and research in this setting can help to define characteristics that may help to define mineral provenance in sedimentary systems on Mars. In WA, some of the greatest clues for deciphering the relationships between mineral phases are the mineral associations or assemblages and the spatial relationships between minerals that can be observed petrographically.
Our previous work in WA included detailed studies of the variety of extreme geochemical conditions of ∼60 lakes and their associated shallow groundwaters, depositional processes and products, and mineralogy, all of which demonstrate temporal and spatial complexity and diversity (Benison et al. 2007, Bowen and Benison 2009, Story et al. 2010). Previous work has also revealed the presence of diverse microbiological communities within these lakes (Mormile et al. 2009), and the potential for biological material to be entrapped and potentially preserved within the rapidly precipitating minerals in these systems (Benison et al. 2008). However, questions still remain as to whether these are chemolithotrophic organisms and the role that they play in modifying the chemistry of the brines and the sediments. Past work has also established these environments as important Mars analogs (Benison and LaClair 2003, Benison and Bowen 2006, Bowen et al. 2008), yet since these publications, many new discoveries have been made that further signify the potential importance of acid saline lacustrine systems in the geologic past on Mars.
Our objective is to identify the early diagenetic processes that are important in creating and modifying these deposits and to consider the way in which they may be specific to acid saline lacustrine settings. Understanding the formation of these sediments and rocks and variations in WA will help us to interpret deposits that have formed in potentially similar depositional systems and from similar fluids on Mars.
Materials and methods
Field Observations and Sample Collection
We observed a wide range in aqueous geochemistry, sedimentology, and surface processes at ∼60 saline lakes and pans in WA during five field excursions from 2001 to 2009 that covered winter, summer, flooding, evapoconcentration, and desiccation conditions (Benison et al. 2007, Bowen and Benison 2009). In this study, we focus on new petrographic, mineralogic, and geochemical characterization of shallow sediments and sedimentary rocks from eight acid saline lake systems, which include some of the most acidic and well-characterized sites from our previous work (Table 1; Benison et al. 2007, Bowen and Benison 2009, Story et al. 2010). The three western sites (Lake Brown, Walker Lake, Lake Magic) occupy the paleo–Yilgarn River drainage catchment, while the five eastern sites (Lake Aerodrome, Prado Lake, Lake Gilmore, Hobby Lake, and the Twin Lakes) occupy the paleo– Lefroy River drainage catchment and depressions in the surrounding Archean bedrock (Fig. 1). Representative sediments and sedimentary rocks used in this study were collected from surface environments (e.g., lakes, sand flats/mud flats, ephemeral channels, dunes). In addition, shallow subsurface samples were collected by manually trenching pits down to 2 m with shovels and taking shallow manual cores down to 50 cm with polyvinyl chloride (PVC) pipes (Fig. 5). At select sites (Aerodrome, Brown, Prado, Twin Lake), samples were collected in detailed transects across lakes (Fig. 6). At sediment and sedimentary rock sample sites, we also measured lake water and groundwater geochemistry in the field (temperature, pH, salinity) and collected water samples for detailed isotopic, elemental, and compound analyses in the laboratory (Bowen and Benison 2009; Table 2; Fig. 2). Therefore, sediment and sedimentary rocks samples can be paired with local water geochemistry (e.g., Fig. 6). Sediments and sedimentary rocks were sealed in air-tight sample bags and plastic containers and shipped back to the USA for subsequent analyses.
Mineralogy and Geochemistry
Representative bulk sediment and sedimentary rock samples were analyzed by X-ray diffraction and visible to near-infrared reflectance spectroscopy (350-2500 nm) to identify mineralogy. Additionally, the clay-size (<2 μm) fraction of select samples was analyzed for mineralogy via X-ray diffraction (XRD) (Story et al. 2010). The amount of Cl− in the sediments was quantified with Mohr titrations (Story et al. 2010). Weight percent of major oxides, trace elements, rare earth elements, and some important compounds (e.g., SO4) in 26 representative sediment samples were measured by a commercial laboratory (Activation Laboratories, Ltd.). For these analyses, major oxides and trace elements were measured via lithium metaborate/tetraborate fusion ICP (inductively coupled plasma mass spectrometry) and TD-ICP (total digestion inductively coupled plasma mass spectrometry) respectively, and SO4 testing was performed via combustion/infrared analysis. The geochemistry of lake waters and shallow groundwaters spatially associated with the sediment samples was also characterized in detail (Bowen and Benison 2009). Optical petrography was also used to identify some minerals in thin section.
The geochemistry of waters associated with the sediments being analyzed was evaluated using Geochemists Workbench (Bethke and Yeakel 2009). These water data are a subset of previously published data (Bowen and Benison 2009). Aqueous species and mineral saturation indices were calculated using the thermo.dat thermodynamic database based on measured values of pH, total dissolved solids, and measured concentrations of Na+, Mg2+, Cl−, SO2−4, K+, Ca2+, Br−, Si, Al3+, Fe3+, HS−, Sr2+ Mn2+, and Cu2+ from seven of the lakes (Table 2). Initial conditions were set to 1 kg of water, T = 25°C, surface pressure, assumed equilibrium between atmospheric oxygen and the oxygenated surface water (fO2 — 0.21; Hem 1985), and measured pH and solute concentrations. Redox couples for Fe, Cu, and S were decoupled in SpecE8 (GWB), and the modeled reactions were charge balanced using Cl− concentrations. Minerals saturation indices were calculated for each individual water sample and then compared to list only those minerals that were common to all the water samples at a specific lake (Table 3).
Representative samples from 16 shallow PVC pipe cores and surface sedimentary rocks were prepared into thin sections for petrographic analyses (n = 84; Fig. 5). These samples included unconsolidated and semilithified surface sediments as well as lithified sedimentary rocks. The thin sections were prepared with techniques appropriate for samples sensitive to water and heat, so that the saline minerals and textures were preserved. The samples were vacuum-impregnated with blue epoxy (blue = porosity in micrographs). The thin sections were examined petrographically under both transmitted and reflected (and a combination of both) light at up to 400× magnification.
Geochemistry and Mineralogy
The sediments that have been in contact with geochemically complex acid saline fluids in southern WA exhibit a wide range of compositions and textures that suggest high degrees of early syndepositional/diagenetic modification. Because of the extreme variability in water chemistry and sediment composition, it is difficult to describe a “typical” acid saline sediment (Fig. 6). The sediments contain a complex mixture of both detrital and authigenic components. The sediments contain up to ∼20% Cl, but this may be misleading due to chloride salts that may have formed from surface or groundwater brines after collection. Bulk major oxide and elemental analyses show that SiO2 is the dominant phase (average ∼60 wt% for samples that have >1% SiO2), and that Al2O3 accounts for up to ∼20% of the sediment mass (Fig. 7).
SO4 is also a significant phase in some of the sediments (up to ∼65 wt%), with a distinct separation between sulfate-rich and sulfate-poor sediments. Comparisons of %S vs. %SO4 show four populations of sediments: gypsum-dominated chemical sediments (with S/SO4 ∼ 0.15), SO4- and S-poor halite-dominated chemical sediments, and two groups of clastics with varying amounts of S (Fig. 7B). Samples collected from the surface and in contact with lake waters have higher S amounts, while samples from the shallow subsurface (few to tens of centimeters deep) have S/SO4 values more consistent with what would be expected if all of the S mass was in the form of sulfate.
Fe2O3 constitutes up to ∼15% of the sediment mass and shows corresponding increases in the amount of many trace elements, including U, Ti, and Ni (Fig. 7C). Ni vs. Fe2O3 concentrations show two distinct populations that correspond to specific lake settings with higher Ni in hematite and lower Ni in goethite based on spectral classification of the iron-oxide mineralogy. Detailed spectral data show that ferric iron exists in these sediments as a mixture of hematite, goethite, and jarosite (Fig. 8). Some sediment grain-size fractions are dominated by specific iron-oxide mineralogy. For example, the <2μm fraction of sediments analyzed from Twin Lake (e.g., Story et al. 2010) has an absorption feature minima wavelength indicative of hematite, while the >62μm fraction of these same samples is indicative of goethite (Fig. 8). The samples with the deepest ferric iron absorption features have minima indicative of more mixtures. These interpretations are supported by XRD and petrographic data as well.
Calculated mineral saturation indices suggest that lake waters and groundwaters are supersaturated with respect to many of the observed minerals, including alunite, amorphous silica, goethite, hematite, heulandite, illite, jarosite, kaolinite, smectite, and K-feldspar (Table 3).
The minerals in the lakes and associated sand flats/mud flats can be classified by their four main origins: detrital grains, chemical sediments, reworked chemical sediments, and diagenetic phases. The detrital grains were transported into the environments by wind or sheetfloods. The chemical sediments precipitated directly from the lake waters, typically during evaporation. Some chemical sediments were first precipitated in lakes, and then they were entrained, transported, and redeposited, usually by wind. Finally, diagenetic minerals are those that either grew from or are the alteration products of shallow groundwater. While it can be very difficult to definitively distinguish between these different types, the textures observed petrographically support these interpretations, and our observations are supported by the associated geochemistry (e.g., mineral saturation indices, sediment geochemistry). The significant role of the extreme and varying salinity and acidity in the lakes and groundwaters accounts for a diverse range of chemical sediment and diagenetic mineral types. The physical dynamics of the environments, including flooding, evaporation, and desiccation, and resulting variations in water depths, along with strong and variable-direction winds, cause the various mineral types to intermingle. Therefore, the diagenetic minerals in southern WA can be best understood in the context of the entire sedimentary system.
This siliciclastic component is commonly thinly bedded sand and contains grains that have experienced eolian reworking. These grains originated from weathering of basement material as well as reworked surface chemical sediments (such as gypsum and hematite). The most common detrital mineral is quartz. Other silicates, such as feldspar (both K- and Na/Ca-feldspar), amphibole (e.g., riebeckite Na2[Fe,Mg]5Si8O22[OH]2), and micas (e.g., Fe-muscovite), are represented, but in lesser quantities. Other minerals that may also include a detrital component include Fe- and Al-phyllosilicates (Story et al. 2010).
Field occurrence, textures, and mineralog-ical associations suggest that multiple mineral types precipitate as chemical sediments directly from the acid saline lake water. Here, chemical sediments are defined as those that precipitate from lake waters and form surface beds that can be millimeter to centimeter scale, vs. diagenetic phases that precipitate from pore fluids and form cements. Like diagenetic phases, these are authigenic precipitates that form in situ. Chemical sediments forming from the acid brine lake waters in WA include halite, gypsum, bassanite, iron oxides (hematite and goethite), and kaolinite/halloysite. At the extremely low pH values that are encountered at these lakes (pH commonly down to ∼2), the solubility of Si is more than twice that of a fluid with a pH of 7–8 (Iler 1979). Similarly, Al solubility increases at these low pHs. The availability of these ions allows for supersaturation of phyllosilicates in these fluids, which is supported by equilibrium thermodynamic calculations (McArthur et al. 1991, Marion et al. 2009). When the pH fluctuates slightly with seasonal changes in meteoric water input and flooding-evaporation-desiccation cycles, all of these minerals appear to undergo episodes of both precipitation and dissolution.
Halite precipitates from lake water during evapoconcentra-tion and forms cumulate crystals and chevron crystals. Cumulate crystals grow at the air–water interface or within the water column. Some cumulates grow together to form rafts on the lake surface. Cumulate crystals eventually sink to the lake bottom, where they form beds. Cumulate halite beds typically contain randomly oriented, small (commonly less than 1 cm) halite cubes. If halite growth continues once the cumulates have fallen to the lake floor, chevrons, larger (up to 4 cm) halite crystals pointing upward, form. In WA, these crystals form subaqueous halite beds up to at least 48 cm thick.
Petrographic observations show textures that are related to the sensitivity of halite to dissolution and variability in the rate of mineral formation (Fig. 9). Rapid growth promotes the formation of many fluid inclusions along growth bands, and slower growth forms clear halite growth bands. Fluid inclusions in saline minerals from these lakes have been shown to be important in entrapping and potentially preserving organic remains in these environments (Benison et al. 2008), and they may have unique characteristics specific to the acid chemistry (Jagniecki and Benison 2010).
Gypsum precipitated from lakes takes three forms in the acid saline lakes in WA. There are bottom-growth gypsum crystals, many of which are swallow-tail twinned crystals. These crystals form at some lake bottoms at the water–sediment interface (e.g., Fig. 4A). They exhibit competitive crystal growth, widening upward. The largest bottom-growth gypsum crystal we measured is 12 cm long, and they make beds up to 15 cm thick. Each individual bottom-growth gypsum crystal contains alternating orange, white, and clear growth bands, representing periods in which the lakes precipitated hematite/goethite (orange) and kaolinite/halloysite (white) along with the gypsum, or precipitated gypsum alone (clear) (Benison et al. 2007). A less common type of lake-precipitated gypsum is in the form of clear, needle-shaped crystals that grow on and around solids in the lake. Common substrates are wood fragments from eucalyptus trees and salt bush. These crystals are rarely larger than 1 cm in length. During the latest stages of evapoconcentration, as a lake becomes desiccated, tiny (less than 1 mm long) gypsum needle-shaped crystals and cubic halite crystals grow on the desiccated lake bed. This forms an efflorescent crust that coats the underlying lake salts or muds. After desiccation of all the lake water, some groundwater may be pulled up to the surface by evaporation and precipitate more efflorescent salts. In this way, these efflorescent crusts are surface precipitates of both end-member lake water as well as shallow groundwater. Bassanite forms as yellow rounded crystals atop gypsum crystals during evapoconcentration in at least one of the acid saline lakes.
Iron oxides and phyllosilicates
Hematite, goethite, and kaolinite/halloysite precipitate from acid saline lake waters as mud-sized crystals. They are found in lake facies sediments interbedded with halite and gypsum beds, forming mud drapes over halite and gypsum crystals or forming a hematite/goethite or a kaolinite/halloysite mud bed up to approximately 6 cm thick (Fig. 4F).
Reworked Chemical Sediments
All of the chemical sediments that form in the shallow lakes are exposed during desiccation, eroded, and reworked by wind into sediments in the lakes, sand flats/mud flats, and surrounding sand dunes. Some reworked chemical sediments (rounded gypsum grains) have oolitic textures suggestive of reworking in shallow waters where they are coated by layers of iron oxide and clay (Fig. 10).
Early diagenetic minerals are those that, by field and petrographic occurrence, show they were not deposited either as detrital grains or chemical sediments but were formed after those depositional minerals. Diagenetic minerals in this study have formed from acid saline groundwaters at or just below (centimeter to meter depths) the surface. Early diagenetic minerals observed include common saline minerals such as halite and gypsum, as well as hydrated sulfates such as hydrobasaluminite (Al4[SO4][OH]10·12-36H2O), and metal sulfate salts such as rozenite (Fe+SO4*4H2O), alunite (KAl3[SO4]2[OH]6), and jarosite (KFe3+3+[OH]6[SO4]2). Other diagenetic phases include hematite (Fe2O3), goethite (FeO[OH]), anata0se (TiO2), gibbsite (Al[OH]3), and Al-phyllosilicates (e.g., kaolinite Al2Si2O5[OH]4, dickite, halloysite; see Story et al. 2010). For some phyllosilicate minerals that have been identified in these sediments, such as chlorite-smectite, it is very difficult to confirm whether it is a detrital or a diagenetic constituent (or both). Past work has attempted to identify the authigenic component based on distinct differences in grain size and morphology, but the populations of the sediment fraction that was considered as detrital contain many of the same minerals as the presumed authigenic component (Story et al. 2010). Geochemical modeling supports the hypotheses that the minerals described here as showing textures and mineral associations suggestive of diagenetic/authigenic precipitation would be expected to be supersaturated in these extreme fluids (Table 3; McArthur et al. 1991, Marion et al. 2009).
Because much of the mineralogy of the diagenetic minerals overlaps with that of the chemical sediments and reworked chemical sediments, fieldwork and petrographic examination of thin sections are especially important to distinguish between depositional and diagenetic textures and to identify spatial relationships between minerals. The morphology and spatial context of the minerals relative to the grains can be used to determine the minerals and textures that are diagenetic (i.e., Chilingarian and Wolf 1988, McIlreath and Morrow 1990). Preliminary macroscopic observations of early diagenetic features, such as surface and near-surface cements, were made in the field, and they demonstrate the complex spatial associations between mineral types (e.g., Fig. 4B, C, E). Examination of thin sections allowed for more detailed study of diagenetic features and evaluation of the microscopic associations between detrital and diagenetic phases, the relative timing of these phases, and the nature of any biological preservation within the sediments (Figs. 9–14).
In the surface and shallow subsurface sediments in these acid salt lakes, there are zones with relatively simple pore-filling intergranular isopachous cements, intergranular meniscus cements, and complex multigenerational cements (Fig. 11). Cements are composed of multiple mineralogies, including soluble salts such as gypsum and halite, acid-sulfate minerals such as jarosite, alunite, and rozenite, phyllosilicates (e.g., kaolinite), and iron oxides (Fig. 11). The texture of the cement can reveal details about the conditions under which they formed. For example, cement crystals that are thicker on the bottom side of grains or that are concentrated near the grain-to-grain contacts are called pendant and meniscus cements, respectively, and are evidence of cementation in the vadose zone (Fig. 11). On the other hand, isopachous cements, having equal thickness around individual grains or pore spaces, form in the phreatic zone. In some samples, the cements predate syndepositional to early postdepositional features such as desiccation cracks, illustrating the very early nature of the cement formation (Fig. 12). Concretionary cements are common in these sediments, with massive spheroidal zones of iron-oxide mineralization. While these have been observed at several lakes, they are particularly abundant at Lake Brown (Bowen et al. 2008). The concretions encase detrital grains including reworked chemical sediments such as gypsum (Fig. 10).
Desiccation cracks that display a vertical tapering “v” in cross section and a polygonal pattern on bedding planes are a temporary feature in the lake and mud flat/sand flat facies. Where thick halite crusts have remained on lake beds after desiccation, continued surface evaporation and wicking up of saline groundwaters produce buckled, positive-relief salt along the polygon patterns on salt crusts.
Cracking patterns typically associated with paleosols also form in the shallow subsurface. These cracks include horizontal cracks, vertical cracks, and crazed plane cracks. In addition, we have also observed autoclastic breccias and circumgranular cracks; i.e., rounded cracks that enclose individual large sand or gravel grains and clumps of finer sediment. Cracking strongly suggests repeated wetting and drying. In these sediments, the cracking both predates and postdates other diagenetic features (Fig. 12).
During flooding stages, the chemistry of the surface water changes considerably as freshwater is introduced into the environment. Initially, flooding serves to lower the salinity and slightly raise the pH of the fluids. Halite chevron and cumulate crystals undergo dissolution; fresher waters dissolve the tops of upward-pointing chevron cubes and leave a flat dissolution surface (Fig. 9). These same floodwaters also dissolve “pipes”; i.e., vertically oriented, millimeter- to centimeter-scale holes in bedded halite. These vugs typically are filled with a coarse, clear halite cement during the next evapoconcentration stage (Fig. 9). All of this salt dissolution increases the salinity of the lake water significantly (Bowen and Benison 2009), which could then lead to dissolution of gypsum. Other minerals also show evidence of dissolution. For example, iron oxide–, clay-, and jarosite-rich samples have porosity with textures suggestive of dissolution of these phases (e.g., Fig. 11G). The flooding-evapora-tion-desiccation cycles not only influence the typical saline minerals, but they also change the water chemistry in ways that allow for both precipitation and dissolution of the other acid-related minerals precipitated by these fluids. Although the low pH values in these fluids are needed to concentrate the elements to precipitate the Fe, Al, and Si minerals observed here, when pHs are extremely low (∼<3), some of these minerals would likely go through dissolution.
Many of the diagenetic cements that are observed in these sediments exhibit textures suggestive of displacive precipitation (Fig. 13). The sediments commonly contain displacive halite and gypsum crystals. These are euhedral, millimeter- to centimeter-scale crystals that are randomly oriented in the siliciclastic sand or mud matrix and are found in the shallow subsurface below lakes and sand flats (e.g., Fig. 2J in Benison and Bowen 2006). Displacive gypsum can be distinguished from depositional gypsum crystals by their needle, lath, and rosette shapes, random orientation within host sediment of different composition, and their fairly uniform size. Displacive halite crystals have the same occurrence as gypsum displacive crystals, but they are common and smaller in size. Some sand beds contain both displacive halite and gypsum; some others contain only gypsum or halite. While commonly observed in the field, these features are not well represented in the core samples studied here, likely due to their ephemeral nature. In addition to displacive gypsum and halite, we also observe iron oxides and clays that precipitated in situ and exhibit relationships with clastic host grains suggestive of displacive growth (Fig. 13).
Trace fossils and body fossils are observed in some of the sediments, encased within fine-grained sediments and associated with diagenetic cements (Fig. 14). They occur within very fine-grained phyllosilicate and iron-oxide–dominated sediments, as well as within sand beds. Biological materials are also observed encased within evaporite minerals (Benison et al. 2008).
Evolution of Acid Brines and Paragenetic Sequence of Events in WA
Multiple different physical and chemical processes occur in the acid saline sediments in WA very early after deposition of the siliciclastic grains (Fig. 15). These include (in a general paragenetic sequence based on field and petrographic observations): (1) syndepositional modification and reworking of detrital grains by winds and water, (2) precipitation of chemical sediments from lake waters, (3) solution/partial dissolution of soluble minerals, (4) displacive growth of cements from groundwater, (5) intergranular cement precipitation from groundwater (including iron-oxide concretions), (6) cracking due to drying (and repeated wetting and drying), and (7) precipitation of crack- and pore-filling cements. Many of these processes are concurrent and syndepositional.
Changing aqueous geochemical conditions cause these various processes to occur and are related to seasonal (or shorter timescale) surface processes such as evaporation and desiccation. Subtle changes in water geochemistry occur on these timescales, but major fluid fluxes that lead to wholesale changes in chemistry do not generally seem to have been necessary to cause a transition from one of these processes to the next. For example, flooding caused by local rainstorms is localized and temporally unpredictable. Flooding causes runoff, resulting in sheetfloods that carry meteoric water, sediment grains, and vegetative debris into the lakes. Lake waters are diluted and dissolve soluble chemical sediments and diagenetic minerals such as halite. Evaporation concentrates solutes in lake and shallow groundwater, resulting in the precipitation of halite, gypsum, iron oxides, and clays from lake waters. Some diagenetic mineral precipitation in the shallow subsurface is also driven by evaporation during arid weather. Desiccation causes cracking of sediments and further mineral precipitation by evapoconcentration. Finally, winds promote reworking of siliciclastic grains, chemical sediments, and diagenetic minerals. These surface conditions greatly affect the shallow subsurface and, in concert with the extreme geochemistry of the groundwaters, lead to an unusual assemblage of early diagenetic features.
Multiple different geochemical processes have influenced the chemistry of the brines and the complex mineral assemblage that occurs within these sediments. These processes are related to the composition of the bedrock, the history of weathering, the stratigraphy of the paleovalley fill sediments, the influence of seaspray aerosols, and multiple stages of fluid–sediment interaction. While a complete discussion of the geochemical history of this region is beyond the scope of this study, consideration of some of the most important processes elucidates the significance of these events on diagenesis in these sediments.
The generation of acidity is likely related to a combination of processes, including oxidation of sulfides, iron, and precipitation of various minerals that release H+ into the brines. Metal sulfides are abundant in the subsurface of this region, both in ores and as pyrite associated with Eocene lignites and coals. Weathering of these sulfides, due to interactions with water, oxygen, and potentially bacteria, leads to oxidation and the generation of acidity according to the reaction (Lottermoser 2007):
In addition, the seasonal fluctuations in aqueous conditions and drying cycles likely result in dissolution of Fe- and Al-sulfate salt minerals such as jarosite and alunite, which can also generate acidity according to:
In most terrestrial acid-generating systems, the hydrogen that is produced is then consumed through buffering reactions related to the weathering of silicates, carbonates, and hydroxides (Lottermoser 2007). In WA, lakes with higher pH are likely a result of acid groundwater neutralization by subsurface limestones. Alternatively, the higher-pH rainwater-fed lakes may remain separate from regionally acidic groundwater by localized impermeable sediments (e.g., clay-pans). Extreme and varied waters can coexist within meters of each other (e.g., Fig. 6). For example, a lake with a pH of 8 may have shallow groundwater with a pH of 3 in adjacent sand flats (Benison et al. 2007). The extremely acidic conditions in the lakes and shallow groundwaters cause dissolution of Fe, Al, and Si constituents through reactions that can consume hydrogen and help to neutralize the acid brines. For example, ferric oxides and hydroxides that have precipitated can redissolve in acid conditions, consuming hydrogen ions:
The dissolution of kaolinite consumes acidity and produces silicic acid, which may precipitate as opaline silica or chalcedony.
Geochemical modeling of the surface brines and shallow subsurface groundwater shows that these acid fluids are saturated with respect to multiple mineral groups, including silica, feldspar, iron oxides, manganese oxides, sulfates, clays, smectite, mica, zeolite, and sulfur and sulfides (Table 3). While there are many potential problems associated with modeling geochemical equilibrium in these complex acid brines, these results are encouraging in that they generally agree with the minerals we observe with XRD, spectroscopy, and petrography, and they suggest that many of these minerals may in fact be authigenic precipitates within these sediments. The thermo.dat database, which is used here, is known to be inaccurate when applied to high-ionic-strength solutions such as these; however, the thermodynamic databases that more accurately model high-ionic-strength solutions (e.g., PITZER, etc.) do not typically include Al, Si, and Fe3+, which are important components in this system. Previous workers have also modeled saturation indices for acid brines from this area and have found comparable chemistry to the fluids observed in this study (e.g., Fig. 2) utilizing different modeling parameters, which may account for the high ionic strength of the solutions (e.g., PHREEQE [McArthur et al. 1991] and FREZCHEM modified to include Fe, Al, and Si [Marion et al. 2009]). These models also showed that the acid brines should be expected to precipitate multiple minerals, including iron oxides, jarosite, quartz, kaolinite, alunite, and gypsum (McArthur et al. 1991, Marion et al. 2009).
Comparisons with Mars and Lessons from an Analog
Our understanding of the history of sedimentary processes on Mars has changed dramatically in the last decade (Grotzinger et al. 2011). Significantly, multiple different types of evaporite minerals that record details about the physical and chemical characteristics of the environment where they are formed have been identified on Mars (Gendrin et al. 2005, Langevin et al. 2005, Tosca and McLennan 2006, Osterloo et al. 2008). Saline systems in general can evolve via multiple chemical pathways (e.g., Hardie and Eugster 1970). The addition of extreme acidity to these systems introduces an additional level of complexity, resulting in a unique suite of Fe, Al, and Si authigenic precipitates from the low-pH brines. Understanding the complex evolution of the acid brines and associated sediments in WA can help us to predict the processes that have been important in modifying acid saline sediments on Mars, and the potential that these types of sediments may have for preserving astrobiological signatures.
Sedimentary rocks in the Meridiani Planum region were extensively examined by the Opportunity rover and were dubbed the “Burns Formation” after Roger Burns, the exemplary planetary geologist who promoted the notion of acid fluids on the surface of the red planet (e.g., Burns 1987, 1989, 1993; Burns and Fisher 1990). The Burns formation shares many significant similarities with the sediments in WA. The Late Noachian or Early Hesperian (age ca. 3.5 Ga or older) Burns formation deposits are composed of an evaporitic sandstone that exhibits a transition from a dry dune facies to a sand sheet facies. These rocks contain wind-ripple laminations, evidence of water-table fluctuations, and an interdune facies that includes evidence of some subaqueous deposition, including festoon cross-bedding and tepee salt-ridge structures (Grotzinger et al. 2005). The grains are composed of reworked “dirty” evaporites that include a chemical precipitate from a fluid (either groundwater or surface water) around some weathered host basaltic volcaniclastic sediment (McLennan et al. 2005, Clark et al. 2005). While there is not clear evidence of clay minerals within the Burns formation, geochemical data at Eagle and Endurance crater do show Si in excess of Al, suggesting aluminosilicate minerals that are interpreted to possibly be phyllosilicates (e.g., kaolinite) in the fine-grained component of the outcrop matrix (Clark et al. 2005).
The Burns formation also contains evidence of groundwater diagenesis, such as diagenetic fronts represented by abrupt secondary color changes (e.g., the Whatanga contact; Grotzinger et al. 2005), elongate vugs, and hematite spherules interpreted as concretions within the sandstone. The presence of concretions in the sedimentary rocks on Mars suggests that there was not only episodic surface water in the past on Mars, but that liquid water also saturated the sediments or sedimentary rocks of the region (Chan et al. 2004). This implies a well-developed hydrologic groundwater system on the surface of Mars. The hematite mineralogy of these concretions suggests changing acid and/or redox conditions to facilitate iron mobility and precipitation.
Spheroidal hematite concretions are observed actively precipitating within sediments (<3000 years old) composed of siliciclastics and intrabasinally reworked evaporite grains in WA (Bowen et al. 2008). Nickel enrichment in hematite concretions and iron-rich sediments in WA has been observed to have Fe/Ni ratios that range from 800 to over 10,000, and it appears to be influenced by the type of iron-oxide precipitate, with the higher Ni concentrations in the hematite-bearing sediments and lower concentrations in goethite-bearing sediments (Fig. 7). The sediments with the highest amounts of Ni occur in lakes over paleovalleys with significant sedimentary fill, while the sediments with less Ni are all from a lake that occurs directly on Archean basement that has little sediment cover. Mars sedimentary rocks are also enriched in Ni. The Burns formation has 400–1000 ppm Ni, and it has been estimated (Mclennan et al. 2005, Morris et al. 2005) that the concretions have up to 1800 ppm Ni. The significance of Ni enrichment on Mars is not yet understood, but it may relate to the associated iron-oxide mineralogy and the amount of iron cycling that has occurred within the sediments.
Hematite concretions on Mars postdate syndepositional to early diagenetic minerals now represented by crystal molds and pore-filling cements, but they appear to predate secondary vugs of moldic crystal porosity. The excellent preservation of the eolian sandstone sulfate grains shows that subsequent surface water (depositing the upper Burns formation) and groundwaters did not dissolve or significantly alter these grains, implying that the fluids were all saturated with respect to the sulfate minerals in the grains.
The mineral jarosite has been identified in the Burns formation with Mossbauer spectroscopy, alpha particle X-ray spectrometry, and infrared spectroscopy (Christensen et al. 2004, Klingelhöfer et al. 2004). On Earth, jarosite is stable under low-pH, oxidizing, and arid conditions, and it commonly forms during acid-sulfate alteration of volcanic rocks, or during sulfide oxidation in acid-mine drainage settings. In WA, jarosite forms from the low-pH, sulfate- and iron-rich brines as an evaporite mineral (e.g., Long et al. 1992). Acid-sulfate minerals such as jarosite occur in other settings such as acid bog lake sediments and hydrothermal acid-sulfate alteration systems (McCollom and Hynek 2005), but they do not coexist with iron oxides and evaporite minerals in these types of settings.
Comparisons between the Burns formation and modern acid saline sediments in WA show that both systems contain a unique mineral assemblage indicative of acid brines, in particular co-occurrence of halite, gypsum, hematite, jarosite, and phyllosilicates. In WA, we observe that these minerals are all intimately mixed at microscopic scales, and while they may not necessarily represent an equilibrium assemblage, they persist together in the acid saline conditions. Our petrographic observations illustrate how these minerals co-occur: as a complex combination of detrital grains, reworked chemical sediments, and diagenetic cements. As on Mars, early diagenetic features in WA include displacive sulfate crystals and hematite concretions as well as jarosite cement. Our observations in WA illustrate the short timescales (thousands of years and less) that are needed to form these diagenetic features and the complex spatial heterogeneity that can be expected in acid saline diagenetic systems (e.g., Fig. 6).
On Earth, modern dune and interdune environments are commonly considered deflational vs. depositional, and they typically have a low preservation potential. In contrast, the Burns formation exhibits excellent preservation, which may be a result of the influence of the acid saline fluids. In WA, we observe very early or syndepositional precipitation of iron oxides and evaporitic cements that aid in preservation of these ephemeral and potentially soluble deposits. In addition, the rapid growth of chemical sediments such as iron oxides and evaporites suggests the possibility of microfossil entrapment and preservation within acid saline sediments (Benison et al. 2008, Fernandez-Remolar and Knoll 2008). In WA, early iron-oxide grain coatings armor the more soluble grains (Fig. 10) and prevent them from dissolving when waters freshen with meteoric input, allowing minerals with varying solubility to coexist in a single assemblage. The presence of soluble and insoluble mineral phases together and the presence of hematite concretions and evaporite crystal molds on Mars have led to diagenetic interpretations requiring multiple distinct fluids (McLennan et al. 2005). However, in WA, we have observed this same combination of diagenetic minerals and textures forming very early in shallow sediments from the same acid saline groundwater that undergoes minor changes in chemistry on seasonal scales as the result of flooding evapoconcentration-desiccation cycles. Similar seasonal-scale diage-netic processes should be considered for Mars.
With increasing spectral analysis of surface mineralogy, several more settings have been identified on Mars beyond the Burns formation that suggest a history of acid saline lacustrine systems analogous to the environments in WA. For example, intracrater clay-sulfate deposits were identified by Swayze et al. (2008), Wray et al. (2008), and Wray et al. (2009) in two closed-basin craters in Terra Sirenum. Of the proposed crater lakes, Columbus crater hosts the most diverse mineralogy and is thought to have been one of several groundwater-fed paleolakes in the Terra Sirenum region of the Martian southern highlands (Wray et al. 2011). Minerals identified in Columbus crater include gypsum, interbedded kaolinite, hydrated Fe/Mg-sulfates, montmorillonite, alunite, jarosite, and ferric oxide or hydroxide (e.g., Milliken et al. 2008, Weitz et al. 2010, Wray et al. 2011), which combine to form a similar suite of minerals as that observed in WA. Interestingly, the alkaline and/or high-temperature secondary minerals prevalent elsewhere on Mars (e.g. carbonates, chlorides, zeolites, and/or prehnite; Ehlmann et al. 2008, 2009; Osterloo et al. 2008) have not yet been detected in the Terra Sirenum craters (Wray et al. 2011). The proposed formation mechanisms for these deposits on Mars include mineral precipitation from groundwater to form extensive cements, precipitation from surface springs, or chemical sedimentation from acid saline lakes (Baldridge et al. 2009; Wray et al. 2009, 2011). The types of acid saline groundwater-derived cements precipitated in the Terra Sirenum system may be analogous to those observed in WA. Modeling data suggest that the origin of the aqueous deposits within Columbus crater likely coincided with an inferred period of Late Noachian groundwater upwelling and evaporation; however, the proposed paleolake is hypothesized to have been ∼900 m deep (Wray et al. 2011).
Similar to the sedimentary deposits on Mars, the sedimentary deposits in terrestrial acid saline systems contain a combination of siliciclastic and chemically precipitated sediments that can be difficult to differentiate. Acid saline mineral assemblages include jarosite, alunite, and hematite together with chloride and sulfate minerals such as halite and gypsum, as well as several types of phyllosilicates that exhibit complex spatial associations and heterogeneity on macro- and microscopic scales (e.g., Story et al. 2010). The occurrence of phyllosilicates is influenced by multiple factors, including the composition of the source rocks throughout the catchment, the hydrochemistry of fluids in contact with these rocks, and the climate, which determines the temperature and water availability (Fernandez-Remolar et al. 2010). Phyllosilicates found on the surface of Mars have been interpreted as remnants of a pH-neutral water-rich era with extreme weathering (Bibring et al. 2006), requiring a fluid and climate regime distinctly separate from that which would facilitate precipitation of sulfate- and iron-oxide–rich sediments. However, in WA, we observe sediments with all of these constituents, suggesting that an extreme change in climate and chemistry may not have been needed to produce these varied deposits on Mars. Other extreme acid systems, including Rio Tinto, also exhibit alternating phyllosilicate–acid-sulfate minerals due to seasonal changes (Fernandez-Remolar et al. 2010). All of these extreme acid terrestrial analogs suggest that seasonal variations in pH and fluid chemistry may be sufficient to explain the occurrence of these types of sulfate, phyllosilicate, and iron-oxide–rich acid saline sedimentary deposits on Mars.
Mineralogical, geochemical, and petrographic studies of sediments and sedimentary rocks from acid saline lakes and adjacent environments in Western Australia show that multiple different diagenetic processes occur very early after deposition. These processes include: (1) syndepositional modification and reworking of detrital grains by winds and water, (2) precipitation of chemical sediments from lake waters, (3) solution/partial dissolution of soluble minerals, (4) displacive growth of crystals from groundwater, (5) intergranular cement precipitation from groundwater (including iron-oxide concretions), (6) cracking due to drying (and repeated wetting and drying), and (7) precipitation of crack- and pore-filling cements. Many of these processes are concurrent and syndepositional. The acid saline systems in Western Australia provide a process analog for acid saline deposits on Mars, and while there are some important differences between the two systems, they share the same suite of minerals and likely have experienced many of the same processes. This work demonstrates the complex relationships among acid brine chemistry, mineral weathering, and authigenic precipitates within sedimentary deposits. Geo-chemical and mineralogical data can be best interpreted in conjunction with careful petrographic analysis, which illustrates spatial associations and crosscutting relationships between minerals that only co-occur in acid saline systems, such as halite, gypsum, iron oxide, jarosite, and phyllosilicates. Acid saline sediments in Western Australia do entrap and potentially preserve biological remains, suggesting that there is some likelihood that sediments associated with similar extreme fluids on Mars could host astrobiological materials.
Partial funding for this study was provided by National Science Foundation grants EAR-0719822, EAR-0719838, 621, and EAR-0719892 (to F.E. Oboh-Ikuenobe, B.B. Bowen, and K.C. Benison), and EAR-0433044 622 and EAR-0433040 (to M. Mormile, F.E. Oboh-Ikuenobe, and K.C. Benison). Acknowledgment is made to the donors of the American Chemical Society–Petroleum Research Fund for partial support of this research. We thank F.E. Oboh-Ikuenobe, M. Mormile, B-Y. Hong, E. Jagniecki, J. Knapp, D. LaClair, and M. Hein for field assistance. Thanks also go to R. Bobick, M. Williams for laboratory assistance.
Figures & Tables
Sedimentary Geology of Mars
Often thought of as a volcanically dominated planet, the last several decades of Mars exploration have revealed with increasing clarity the role of sedimentary processes on the Red Planet. Data from recent orbiters have highlighted the role of sedimentary processes throughout the geologic evolution of Mars by providing evidence that such processes are preserved in a rock record that spans a period of over four billion years. Rover observations have provided complementary outcrop-scale evidence for ancient eolian and fluvial transport and deposition, as well as surprisingly Earth-like patterns of diagenesis that involve recrystallization and the formation of concretions. In addition, the detection of clay minerals and sulfate salts on Mars, coupled with large-scale morphologic features indicative of fluvial activity, indicate that water-rock interactions were once common on the martian surface.