Microbially Induced Sedimentary Structures Recording an Ancient Ecosystem in the ca. 3.48 Billion-Year-Old Dresser Formation, Pilbara, Western Australia

The fossil record of the earliest life on Earth is sparse, and reconstructing the most ancient biota is challenging. Current interpretations of the diversity of Earth's earliest life come predominantly from stromatolites (e.g., Lowe, 1980; Walter et al., 1980; Byerly et al., 1986; Hofmann et al., 1999; Allwood et al., 2006, 2007, 2009, 2010; Hickman, 2012), organic microfossils of prokaryotes and biofilms (e.g., Awramik et al., 1983; Walsh and Lowe, 1985, 1999; Walsh, 1992; Hofmann, 2004; Schopf et al., 2007; Schopf and Bottjer, 2009; Sugitani et al., 2010; Wacey et al., 2011, 2012; Hickman 2012), and isotopic signatures of carbon and sulfur (e.g., Shen et al., 2001, 2009; Ueno et al., 2006, 2008; Wacey et al., 2010). Microbially induced sedimentary structures (MISS) provide an additional way to decode life in ancient sediments (review in Noffke, 2010).

Microbially induced sedimentary structures are created by microbial mats colonizing most aquatic environments, including shelves, tidal flats, lagoons, riverine shores, lakes, dune fields, and sabkhas (e.g., Gerdes and Krumbein, 1987; volume by Hagadorn et al., 1999; Eriksson et al., 2000; Prave, 2002; volume by Schieber et al., 2007; Beraldi-Campesi et al., 2009; volume by Noffke, 2009; Noffke, 2010; volume by Noffke and Chafetz, 2012). In contrast to stromatolites, MISS arise exclusively from the response of microbiota to physical sediment dynamics. Syngenetic mineral production within the organic matrix of the microbial mat provided by the extracellular polymeric substances (EPS) as seen in stromatolites (e.g., Reid et al., 2000; Dupraz et al., 2009; Decho et al., 2011) does not commonly take place (Noffke and Awramik, 2013). Occasionally, mineral precipitates do occur but are only of a temporary nature (e.g., Kremer et al., 2008). The formation of MISS is divided into two steps, first the primary shaping by physical sediment dynamics and second the diagenetic mineralization of organic material. This will be explained in the following (see Noffke, 2010, for details).

The MISS-forming microbial mats react to sediment dynamics (erosion, deposition, and latency) in four various ways (quantification of primary processes in Noffke and Krumbein, 1999; Noffke, 1999). (i) Erosive stress by water currents passing over the mat-overgrown sedimentary surface triggers biostabilization (Paterson et al., 1994). If affected by erosive stress, the microbial filaments arrange parallel to the depositional surface; the network of filaments interweaves the sediment grains, and the EPS switch, in a fraction of a second, their biomolecular structure to be flexible and ductile (e.g., Stoodley et al., 2002). These microbial effects prohibit the removal of sediment grains by currents. This biostabilization increases the resistance of a sedimentary surface by up to 12 magnitudes (epibenthic microbial mats), 3–5 magnitudes (endobenthic microbial mats), and around 0.02 magnitudes (biofilm-type overgrowth) (Noffke and Krumbein, 1999). (ii) If a microbial mat is exposed to deposition of sediment, the filaments rearrange and orient themselves perpendicular to the sediment-mat surface. They reach into the supernatant water, causing microzones of turbulent disturbance of the water current. This effect is called “baffling” and induces the fall-out of sediment grains suspended in the water. The microbial mat then builds the grains into the mat matrix, either by filaments migrating around them or by sticky EPS gluing the grains together and fixing them in their positions. Baffling and trapping are active sediment accumulation processes. Dependent on the mat types, grain size selection or heavy mineral enrichment either takes place or not. (iii) Binding is the initial organization of randomly distributed microbial cells and trichomes into a highly structured biofilm community. The microbes communicate with each other and move into a position within the sediment that allows both optimal access to light and/or nutrients and cooperative interaction with the metabolism of neighboring microorganisms. This active arranging takes place during times of quiet sediment dynamic conditions. (iv) Growth, which does not play a role in binding, is a process dependent on availability of nutrients and/or light and not on sediment dynamic conditions; it is biomass enrichment by cell replication and the production of EPS. As a result of their differing modes of genesis, MISS show different morphologies from those of stromatolites (Noffke and Awramik, 2013). Seventeen main types of MISS ranging from centimeter squared to kilometer squared scales have so far been distinguished and both their genesis and resulting morphologies quantified (Noffke, 2010; see also volumes edited by Hagadorn et al., 1999; Schieber et al., 2007; Noffke and Paterson, 2008; Noffke and Chafetz, 2012). Common examples are given in Fig. 1.

While MISS are formed by these primary processes, they are preserved by secondary, diagenetic processes. MISS are lithified by rapid in situ mineralization of the organic matter of the mats, including the microbial cells, trichomes, and filaments, as well as the EPS. In thin sections perpendicular to ancient mat layers, the network of fossil microbes is commonly visible as intertwined and bent laminae. The network includes fossil EPS and (formerly) allochthonous sedimentary grains that once were bound by the mat during its lifetime. However, the appearance of fossil mat texture in MISS differs fundamentally from that of organic microfossils in chert. This difference arises from the preservation of the MISS. Microbial mats in siliciclastic sedimentary rocks are preserved by replacement minerals, not by impregnation as is the case for organic microfossils or organic biofilms in primary chert (e.g., Walsh, 1992; Walsh and Lowe, 1999; Tice and Lowe, 2006; Schopf and Bottjer, 2009; Wacey, 2009). In contrast to such organic material so precisely preserved in chert, the mat texture-forming laminae in MISS have no discrete outline (Noffke 2000; Noffke et al., 2002, 2003, 2006a, 2006b, 2008). In MISS, the laminae appear “cloudy” with diffuse borders. The reason for this appearance is that during the diagenetic alteration of the mat, the chemical compounds of the organic matter released by microbial decomposition diffuse away from their original microsite (e.g., Krumbein et al., 1979; Knoll et al., 1988; Beveridge, 1989; Urrutia and Beveridge, 1994; Konhauser et al., 1994; Schulze-Lam et al., 1996; review on these studies in Noffke, 2010). They react with chemical compounds prevailing in the surrounding water, resulting in initial hydrated “amorphous” mineral precipitates. These initial hydrated mineral precipitates accumulate at nucleation sites and gradually dehydrate and shrink in the course of diagenesis. The dehydration and recrystallization lead to mineral phases of higher crystallinity. At the microscopic scale, the minerals that replace microbial mats therefore are arranged as irregular clots lining the original shape of the laminae. This is in stark contrast to microfossils preserved in chert, where the organic matter was entombed rapidly by synsedimentary silicification and impregnation, and detailed preservation of cells was possible (e.g., Cady and Farmer, 1996). For this reason, very high spatial resolution analytical techniques such as transmission electron microscopy (TEM) or secondary ion mass spectrometry (SIMS and Nano-SIMS) are not particularly useful for resolving morphological characteristics of MISS-forming microbial cells. To acknowledge these differences in preservation, the MISS literature uses the terms “laminae” and “filament-like texture” instead of “filament” or “trichome.” In the fossilized examples from younger Archean, Proterozoic, and Phanerozoic time periods, the ancient mat fabrics may still include some original carbon, but the original organic matter is largely replaced by pyrite, hematite, and goethite (review in Noffke, 2010).

A set of seven biogenicity criteria for MISS has been developed and tested in numerous comparative studies (overview in Noffke, 2010). The first four criteria describe the depositional habitat of MISS occurrence as follows. MISS occur in rocks of not more than low-grade metamorphosis. In stratigraphic sections, MISS occur at turning points of regression-transgression. MISS occur in the depositional “microbial mat facies.” The distribution pattern of MISS reflects the average hydraulic pattern in a depositional area. The last three criteria describe the MISS themselves as follows. The fossil MISS resemble strongly or are identical with geometries and dimensions to modern ones. The MISS include microtextures that represent, or were caused by, or are related to, ancient biofilms and microbial mats. Geochemical analyses support the interpretation as MISS.

Using this set of criteria, more than 14 studies have systematically explored MISS from modern to ancient sites, comparing structures in equivalent environmental settings from the modern right back to the early Archean (e.g., Noffke 1999, 2000; Noffke and Krumbein, 1999; Noffke et al., 2001a, 2002, 2003, 2006a, 2006b, 2008). This suite of studies has assembled a data set that enables the evolution of MISS-prokaryota to be monitored throughout Earth's geological record. The investigation of MISS is divided into four steps: (i) detection, (ii) identification, (iii) confirmation, and (iv) differentiation (Noffke, 2010). (i) Detection is the visual reconnaissance during a geological survey of candidate sediments and sedimentary rocks. (ii) A candidate structure (e.g., for erosional remnants and pockets) is measured with respect to geometry and dimension, and indices such as the MOD-I (modification index) are determined (see later in this article). The assembled data that have arisen from these systematic studies now allow the quantitative comparison of any candidate structure with other MISS from other periods in Earth history including the modern. (iii) Analyses on the mineralogy and geochemistry of the candidate structure are conducted (e.g., presence of carbon in laminae). (iv) A comparison with similar but abiotic phenomena is made, if any similar but abiotic phenomena exist. As demonstrated, this set of biogenicity criteria compiles various lines of evidence and allows for identification of fossil structures with a high probability. This study on possible MISS in the Dresser Formation employs this methodological approach as well.

Microbially induced sedimentary structures are listed as one target for the Mars Exploration Rover Program (Committee on an Astrobiology Strategy for the Exploration of Mars, 2007). On Mars, sabkha settings are well known, where sedimentary surface structures and rock beds record the former existence of fluid water (e.g., Grotzinger et al., 2005; Metz et al., 2009). Because of the similarity of the early history of Earth and the early history of Mars, the knowledge on the sabkha habitats of MISS and the criteria of biogenicity of MISS may assist in the exploration of Mars.

To date, the oldest microbial mat communities that formed MISS are from the 3.2 Ga Moodies Group, South Africa (Noffke et al., 2006b; Heubeck, 2009). Here, two wrinkle structures and one roll-up structure were detected in tidal deposits. The 2.9 Ga Ntombe Formation, Pongola Supergroup, South Africa, includes eight wrinkle structures that record fossil microbial mats on shallow shelf settings (Noffke et al., 2003). The isochronous shelf of the Brixton Formation of the Witwatersrand Supergroup shows 28 wrinkle structures, two sedimentary surfaces yielding erosional remnants and pockets, and one bedding plane with polygonal oscillation cracks (Noffke et al., 2006a). In thin sections from all study sites, textures are visible that resemble degraded microbial mat fabrics. Cyanobacteria have been suggested to be the constructing agents; however, no unambiguous evidence has been documented (see discussions in Noffke et al., 2003, 2006a, 2006b). Highly diverse microbial mat ecosystems from ancient tidal and sabkha environments are recorded by MISS in the 2.9 Ga Pongola Supergroup, South Africa (Noffke et al., 2003, 2008). Of significance is that these ancient MISS resemble strongly the MISS in equivalent modern settings. Because modern MISS allow the detailed quantification of the hydraulic and sediment-dynamic interaction with biofilms, the gained information assists in the interpretation of ancient MISS, the environment they are situated in, and prokaryote evolution. This similarity of MISS over 3 billion years of Earth history is in contrast to stromatolites, where many early Archean species are unlike the modern ones (Noffke and Awramik, 2013).

The objective of this study was to take the comparative investigation one step further and search for MISS in rocks older than 3.2 Ga. This contribution describes a series of distinct types of macrostructures and microstructures from sedimentary rocks of the ca. 3.48 Ga Dresser Formation.

The Dresser Formation is located in the East Pilbara granite greenstone terrane, Western Australia (Fig. 2a). It was chosen for this study because it contains some of Earth's oldest and best-preserved volcanic and sedimentary rocks (Fig. 2b; Barley et al., 1979; Hickman, 2012). The age of this rock succession is determined as 3.481±3.5 Ga [Australian Stratigraphic Units Database (2012), Dresser Formation, Stratigraphic Number 36957]. The formation is geographically restricted to a ca. 25 km² area in the North Pole Dome and consists of bedded chert, carbonate, and siliciclastics, plus pillow basalt and dolerite (Van Kranendonk et al., 2008). The sedimentary rocks were originally micritic carbonates and evaporites deposited under shallow-water, low-energy (sabkha-type) conditions, interbedded with sandstone and conglomerate deposited during periods of growth faulting and tectonic activity (Lambert et al., 1978; Buick and Dunlop, 1990; Van Kranendonk, 2006; Van Kranendonk et al., 2008). Repeated episodes of growth faulting associated with volcanic activity during carbonate-evaporite sedimentation permitted circulation of hydrothermal fluids that overprinted much of the original sedimentary mineralogy. Carbonate, organic matter, and gypsum were largely replaced by pyrite, hematite, barite, and silica (Van Kranendonk, 2006; Van Kranendonk et al., 2008).

In the Gibson and Great Victoria deserts in central Australia and in the Carnarvon Basin at the west coast Eocene to Oligocene, weathering is common (van der Graff, 1983). Here, the Tertiary tropical climate induced the formation of a soil catena, composed of blocky silcrete at the modern topographic heights, and deep red, silty laterite in topographic lows. Typical structures in the otherwise unconsolidated soils include irregular and steep-sloped cones up to 35 cm in height, pisolites from 0.5 to 5 cm in diameter, and karst pipes that facilitate subsurface water flow. However, none of these soil structures were observed in our North Pole study location. Quite to the contrary, the sedimentary rocks in the Dresser Formation are highly consolidated. Stromatolites, ripple marks, ripple cross beds, and other primary sedimentary structures are well preserved and do not display disturbance by any deep soil or silcrete formation.

In addition to the macroscopic phenomena, widespread evidence also remains for the primary mineralogy of the ancient sediments, including relic carbonate rhombs and rhombic voids in silica (Lambert et al., 1978), patches of dolomitic chert in surface outcrop (Walter et al., 1980), peloidal and oncolitic grains (Buick and Dunlop, 1990), extensive carbonate in unweathered drill core material (Van Kranendonk et al., 2008), and observations in this current study of carbonate in many of our thin sections.

Geological mapping, millimeter-scale sedimentological stratigraphic logging, and quantitative analysis of morphologies of sedimentary macrostructures were undertaken in the North Pole Dome, Pilbara, Western Australia, in June 2011. Petrological and mineralogical analyses were conducted on standard geological thin sections and polished rock chips. We used the equipment at Old Dominion University, including Olympus BX51 and Olympus SZX12 microscopes equipped with Q Colour 3 Olympus digital cameras. Raman analyses were carried out at the University of Bergen with a Horiba LabRAM HR800 integrated confocal Raman system and LabSpec5 acquisition and analysis software. For Raman, samples were standard uncovered geological thin sections that allowed optical and chemical maps to be superimposed. The laser was focused ∼2 μm below the surface of the thin sections to avoid surface polishing effects (Fries and Steele, 2011). Since the thin sections were ∼30 μm thick, there was no possibility of obtaining a Raman signal from the mounting resin. All analyses were carried out by using a 514.5 nm laser, 100 μm confocal hole, 1800 mm/line grating, and ×50 objective lens. Dual acquisitions were taken from each analysis point, each with an acquisition time of 4 s. Maps were acquired with a 1.5 μm spatial resolution by using selected peaks from the Raman spectra.

Seven stratigraphic sections were analyzed at millimeter-scale resolution, covering a total stratigraphic thickness of 39.8 m (locations shown in Fig. 2c; detailed profiles in Fig. 3a–3h). The lithology and sedimentary structures record an ancient coastal sabkha in which a subtidal zone, an intertidal zone, a lower supratidal zone, a lagoon, and a barrier were distinguished. Our observations are in line with those of Buick and Dunlop (1990), who described the paleoenvironment of the entire Dresser Formation in detail. In the present time, sabkhas are colonized by ubiquitous microbial mats forming MISS.

In the following, we describe the fossil sedimentary structures of the Dresser Formation and discuss the possibility of biological origin by comparing them with the modern MISS in equivalent sabkha settings. The fossil sedimentary structures form discrete morphologies that display sharp transitions with their surroundings. The structures do not resemble cones, pisolites, or vertical pipes caused by any weathering and soil-forming processes currently known.

A series of studies systematically comparing modern with ancient MISS showed a great consistency of MISS throughout geological time (Noffke, 2000, 2010; Noffke et al., 2001a, 2002, 2003, 2006a, 2006b, 2008). MISS assemblages record diverse microbial mat ecosystems as early as the Mesoarchean Era, as demonstrated by Noffke et al., (2001a, 2008) and Noffke (2010).

The sedimentary structures here described from the Dresser Formation are interpreted as MISS based on several lines of evidence. First, their morphologies are very similar to those of a variety of more recent fossil and modern MISS. Interpretations of these extensively studied MISS examples are based both on qualitative morphological characteristics and on numerical data that quantify the morphologies of the structures and allow quantitative comparison among MISS, as well as between MISS and nonbiological sedimentary structures. In general, transitional forms between MISS and surrounding sedimentary features are not observed, neither in modern nor fossil sequences; that is, all MISS are distinct. Variations within one morphotype of MISS occur and are considered by the quantitative data; transitional forms with intermediate morphological characteristics, however, do not exist between MISS and any abiotic type of sedimentary structure.

A second important line of evidence is provided by the close association of microbially induced sedimentary structures in the Dresser Formation. These MISS display the same associations that are known from modern as well as from fossil sabkhas.

Third, many of the sedimentary structures include microscopic biotextures resembling strongly the textures known from the younger (Archean and Proterozoic) fossil record of MISS.

Fourth, geochemical and petrological analyses are consistent with the typical mineral associations found in fossil microbial mats from other Archean sites previously described. In particular, carbon is intimately bound to the microtextures. The four complementary lines of evidence individually and collectively support the interpretation of biogenicity for MISS in the Dresser Formation.

Weathering of rock surfaces during the Eocene (Lower Tertiary) has been described in detail from other parts of Australia (van der Graff, 1983). Is there the possibility that surface weathering might mimic some of the proposed MISS features of the Dresser Formation? Tertiary weathering has produced unconsolidated soils that include irregularly shaped and steep-sloped cones, pebble-sized pisolites, and vertical karst pipes. None of these phenomena were found in the stratigraphic sections studied by us in the Pilbara. Moreover, the morphologies of the weathering structures differ significantly from those of the sedimentary structures in the Dresser Formation described as MISS. In addition, the distribution of the MISS in the Dresser Formation correlates with specific sections of the stratigraphic profiles. They are not related to the modern geomorphological topography of the study area as is the case with the weathering structures in the other parts of Australia (van der Graff, 1983). Also, the fact that the MISS form specific associations related to adjacent tidal zones differs significantly from any weathering or erosive origin, as does the identification of thermally mature carbon intimately associated with many of the Dresser structures. Finally, if the sedimentary structures we interpret as MISS were of Eocene weathering origin, then the question would be why identical MISS and MISS associations are found in modern environments of the present time that, obviously, have not experienced Eocene weathering.

We conclude that the Dresser Formation records a complex ecosystem of microbial mats. But which prokaryotes might have formed these Early Archean microbial mats? It must be underlined that modern MISS and the modern MISS-forming microbial mats may only serve as analog models for the Dresser examples. The genetic information of individual microbial groups such as modern cyanobacteria is highly variable, differing today even within meters of the same setting. Therefore, any conclusion on the existence of certain groups in the fossil record, let alone in the fossil record at 3.48 Ga, must be speculative. Also, microorganisms occur as biofilms, not as individual cells or groups. A biofilm is a microbial community attached to a solid substrate, in which all members of the community interact to foster light and nutrient harvesting, EPS production, and so on (Stoodley et al., 2002; Noffke et al., 2013). Therefore, we can with certainty state that the fossil MISS are the structural expression of biofilms formed by microbes interacting in similar fashion with the shallow, photic zone sedimentary habitat as both younger fossil and modern microbial mats do (Noffke, 2010). The main message of the Dresser Formation MISS is that microbenthos existed and was able to construct coherent, carpetlike microbial mats. The mats were able to withstand erosion, to respond to deposition, and to withstand semi-arid climate conditions. Using modern MISS and microbial mats strictly as models, we conclude that the ancient MISS-forming microbial mats in the Dresser Formation were dominated by microbes mimicking the behavior of modern cyanobacteria. It is important to note that cyanobacteria are one major group of microbenthos capable of producing the abundant amount of EPS necessary to allow high biostabilization effects (Linda L. Jahnke, frdl. pers. comm. 2013; Paterson et al., 1994; Noffke and Paterson, 2008; Noffke, 2010). Non-transparent wrinkle structures (Fig. 9b) record such high amounts of EPS (Noffke et al., 2002). Detailed studies on the interaction of stromatolites (Beukes and Lowe, 1989) and of MISS (Noffke et al., 2008) with their hydraulically affected sedimentary environment suggested the presence of ancient cyanobacteria already in the 2.9 Ga Pongola Supergroup, South Africa. Cyanobacteria are known to be the first oxygen-producing organisms in the fossil record. However, it is important to keep in mind that photosynthesis comes in two types, oxygenic and anoxygenic. While the abundant evidence for an anoxic atmosphere in the Early Archean (Farquhar et al., 2000; Hazen et al. 2008; Sverjensky and Lee, 2010) suggests that oxygenic photosynthesis was not established until perhaps the Neoarchean Era, it does not exclude the existence of early cyanobacteria of anoxygenic photosynthesis. However, if the stromatolite- or MISS-forming microbial mats in the Pongola Supergroup were indeed cyanobacteria with the capacity to produce oxygen, then this would be supportive to the latest findings of paleosols in the Pongola Supergroup that point toward an oxygen-rich atmosphere around that time (Crowe et al., 2013). With respect to the Dresser Formation microbiota, we note that, while a number of modern filamentous cyanobacteria are capable of anoxygenic photosynthesis, using H₂S instead of H₂O as the electron donor, Chloroflexus or sulfur- or iron-oxidizing bacteria such as Beggiatoa would also be capable of this pathway (e.g., Bailey et al., 2009). Both groups also are capable of forming substantial microbial mats (e.g., Bailey et al., 2009). Pulling all these thoughts together, a conservative interpretation of the ancient Dresser microbenthos would be that the biofilms (microbial mats) of the Dresser sabkha behaved in similar fashion to modern microbenthic biofilm communities found in sabkha settings today.

In summary, the sedimentary structures preserved in the coastal sabkha paleoenvironment of the ca. 3.48 Ga Dresser Formation are here interpreted as MISS. The MISS form assemblages that are shown to be typical for sub-environments of sabkhas through geological time. Using modern MISS in equivalent sabkha settings as analog models, we conclude that the MISS in the Dresser Formation record a complex microbial ecosystem, hitherto unknown, and represent one of the most ancient signs of life on Earth.

We thank Kath Grey, Martin Van Kranendonk, and the Geological Survey of Western Australia for the support of our field work. Andy Knoll has provided valuable comments on an earlier version of this manuscript. Funding for this study was provided to N.N. by the National Science Foundation NSF Paleobiology and Sedimentary Geology Program, NASA's Exobiology Program, and by the NASA Astrobiology Institute. D.C. was granted student research support by the NASA Astrobiology Institute. D.W. is funded by the Bergen Research Foundation, the University of Bergen, and the Australian Research Council Centre for Core to Crust Fluid Systems. R.M.H. received support from the NASA Astrobiology Institute, the Deep Carbon Observatory, and the Carnegie Institution of Washington.

EPS, extracellular polymeric substances; MISS, microbially induced sedimentary structures; MOD-I, modification index.