Water sampling

Physicochemical parameters such as temperature, pH and electrical conductivity (EC) were measured in situ at all sampling locations, using a portable system (Lutron WA-2017SD). Water samples were taken to perform major elements analyzes. Additionally, a water monitoring sampling has been conducted during five field campaigns (November 2015, April-May 2016, August 2016, February 2017, and March 2018). Surface waters of both lakes CLG and CLC were collected when the shore was reachable as well as water samples from mats sampling sites (Carvallo Maquinchao). Furthermore, hand pits have been dug in and around the Maquinchao River to identify the groundwater flux (^{Alvarez et al., 2021}), and waters from other wells associated with Carvallo Maquinchao point (CLA 2P, CLA 3P) were collected. Pits were dug by hand until reaching the water table, isolated using PVC tubes and closed securely to avoid sediment incorporation. In total five sampling places were defined, including two lake surface waters, one river surface waters, and two pits waters (Fig. 1; Alvarez et al., 2021). Chemical analyses of waters have been performed in the Laboratorio de Suelos de Ecosistemas Patagónicos, from the Instituto Patagónico para el Estudio de los Ecosistemas Continentales and in the Laboratorio de Geoquímica del Centro de investigaciones Geológicas, Argentina. Carbonates, bicarbonates, chlorine, calcium and magnesium have been measured by titration. Sodium, and potassium by photometry and sulfates by UV spectrophotometry. Chemical analyzes were processed and Saturation Indexes (SI) calculated using the software DIAGRAM and PHREEQC, respectively.

Mats sampling and SEM preparation

Mats were sampled during Austral spring 2015 in three different locations A, B and C (Figs. 1 and 3). In location A, the system is composed of several small streams of few centimeter depth (up to 10cm) running west towards CLG. Location B and C are slightly deeper (up to 50cm). Waters are very still, and flow runs toward west in direction to CLG. All samples were fixed in the field using Glutaraldehyde (2.5%). A sample from location A was retrieved in the center of the stream/channel whereas in location B was taken from the top part of a large pebble and both samples were directly put in tubes with Glutaraldehyde. Samples from location C were carried and stored in a larger container for several days before being fixed with glutaraldehyde. All samples were subsequently dehydrated with increasing ethanol concentrations (30%, 50%, 70%, 90%). Back in the laboratory they were either directly mounted on stubs and gold coated for SEM observations or dehydrated using a critical point device. When possible, samples were mounted to allow the observation of both their upper and bottom sides as well as a longitudinal section.

Mats observations and the identification of elemental phases were performed by Scanning Electronic Microscopy (SEM) and Energy Dispersive Spectroscopy (EDS), respectively. The SEM-EDS analyses were carried out using a JEOL 7001 PA Scanning Electronic Microscopy instrument, equipped with an Energy dispersive X-ray analyzer EDS JED2300 at the Department of Earth sciences of the University of Geneva (Switzerland).

Molecular Biology

Samples for microbiological/DNA sequencing study were collected in several pounds (Fig. 1) using sterilized scalpels. All samples were stored in tubes (with ethanol) and kept refrigerated until further processing. Samples from locations A, B and C were divided in three subsamples: the top (1), the center (2) and the lower part (3). In location A, the sample was taken from the border and the center of the channel/stream. A1 and A2 correspond to samples taken from the rim having a jelly texture and being drier, respectively. A3 has retrieved from the center of the stream/channel.

DNA extraction was performed in 0.5 mg of processed sample using the kits MOBIO POWER BIOFILM. DNA amplifications were conducted with a total of ~10 ng of DNA using universal Primers for prokaryotes Pro341F (SEQUENCE 5’-CCTACGGGNBGCASCAG-3’) and Pro805R (5’-GACTACNVGGGTATCTAATCC-3’) for the V4-V5 hypervariable region of 16S rRNA gene (^{Takahashi et al., 2014}). Polymerase Chain Reaction (PCR) Products were then purified with Kit Roche HighPure PCR Protocols beads (Roche) and triplicates were quantified and pooled. Samples were then pooled equimolarly and Qubit-quantified for DNA sequencing.

Sequencing was performed at the University of Geneva (Switzerland), on an Illumina Miseq with 2x250 cycles with settings of 7.5 GB yield (including PhiX), an error rate of 2.5% and Q30 at 75%. The analysis yielded 5.3GB of sequence with an error rate within quality specifications. After barcode and primer trimming were removed using the in-house script, sequences were quality controlled, denoised, paired-ends merged and chimeras were removed using DADA2 version 1.20.0 (^{Callahan et al., 2016}) using R. Sequences were then classified as amplicon sequence variants (ASVs). These ASVs were taxonomically assigned using Silva (138.1 release) (^{Quast et al., 2013}) and figures and data obtained using the phyloseq package after rarefaction to 90% of the least abundant taxa in a sample (^{McMurdie and Holmes, 2013}). R scripts are available on the following links: https://osf.io/ngpce/.

Water chemistry

Waters in the Maquinchao Basin can be divided in three different groups according to their sampling location: (1) CLGs; (2) CLCs; and (3) Carvallo Maquinchao CLA 2P and CLA 3P. A more detailed discussion of these data can be found in ^{Alvarez et al. (2021}). The collective water sampling campaigns allowed to have a water dataset for the different seasons. This is clearly noticeable in the temperature dataset, where the lowest temperatures are registered during the Austral winter (Table 1). Conversely, temperatures appear to remain relatively uniform between surface water in the river and water from the pits regardless of the season. The pH values range from 7.2 to 9.9 being the highest associated with the present lakes. The river values are slightly higher than the one from the pits no matter the season. Conductivity in CLG displays the highest values ranging from 10100 to 631000 μS/cm.

Table 1 Physicochemical parameters of water samples. 

Major elements display differences between the lakes and the river and changes between the different sampling campaigns can be identified. Waters cations abundance are as follow: Na+ >> Mg2+ > Ca2+ > K+ (Table 1). Na+ values are the highest by far in CLG ranging from 117 meq/L in February 2017 to 626.1 in November 2015. In comparison, Na+ values for CLC river and pits are lower, ranging from 4.2 meq/L to 24 meq/L. Most of Na+ values in river and pits are constant throughout seasons and years and are around 10 to 12 meq/L. Concentrations in major elements remain generally constant in river point and its associated pits. The same trend of higher values in CLG compared to CLC, river and their associated pits can be observed for K+ and Mg2+. In comparison, Ca2+ presents a different trend. Firstly, Ca2+ concentrations in CLG are lower (0.2 to 0.6 meq/L) than the ones from CLC, river points and their associated pits. Carvallo river values are homogenous along the different campaigns (from 1.7 to 2 meq/L) whereas values for the associated pits fluctuated during the different sampling seasons. Cl-, SO42- and Na+ present similar variation trends. In addition, sulfates values appear to be higher in pits than in their associated river points. As for anions, CO3- was quasi absent, and only recorded in CLG (0.9 to 17.1 meq/L). Conversely, bicarbonates (HCO3-) showed substantial variations. CLG values are lower during summer periods than during autumn and winter seasons. Furthermore, during these last two seasons differences between river and associated pits are less marked than during summer and spring. Fluctuations of Ca2+ are lower than the ones of HCO3-. River waters present less variations in all ions than its associated pits (Fig. 4). All ion values during Austral spring 2015 are lower especially for CLC compared to the values for the other periods and years (Table 1; Fig. 4). This is especially obvious for HCO3- (Fig. 4)

Figure 4 Saturation Index for calcite vs. HCO3- (uppermost panel) and Ca (lowermost panel). Colors indicate different sampling periods whereas symbols stand for sampling locations. 

Saturation Indexes (SI) of different minerals have been calculated for all sampling points and field campaigns. SI for anhydrite, gypsum and halite delivered numbers below 0 thus proving the undersaturation of the waters in the Maquinchao Basin (Table 2). SIcalcite values range from -0.98 to 1.22, indicating undersaturated to saturated waters for calcite. Both CLG and CLC present strong variations of the SI with period of SI > 1 and SI < 0 (Table 2, Fig. 4). In comparison, SI from river point and its associated pits are always below 1 and usually below 0.8 SI. River SI values are higher than the ones from its related pits. Winter and autumn SI for river points and its connected pits are higher than SI for spring and summer.

Table 2. Variations in saturation indexes (SI) for calcite, gypsum, anhydrite and halite for different sites and seasons of the year.

This dataset indicates that the November 2015 values have different ranges compared to the other campaigns and are slightly lower. There are two different trends, one for lakes and one for the river that are consistent with ^{Pacton et al. (2015}) observations. Variations in both lakes can be strong whereas the river data are more homogenous and stable.

Mats diversity and morphologies

Three different mats have been sampled during the 2015 field campaign in locations A, B and C (Figs. 1 and 3). The sampling site A had a hard substrate and was located in shallow waters. The site B was located in a large pound protected by vegetation that does not dry out (pers. comm. from locals). Monitoring of waters has been carried out in sampling site B at the exact site where mats have been sampled (water sampling point named Carvallo, see Fig. 1). Location C was set in a larger pound affected by water level variations.

The uppermost part of the mat from location A is reddish presenting black spheres also known as Nostoc balls (^{Dodds et al., 1995}). It also contains abundant diatoms decreasing with depth. Three different types of filament assemblages can be distinguished under SEM (Fig. 5). Certain areas are characterized by abundant filaments of variable diameter having an outer sheath (Fig. 5a to g). They can be associated with dense calcium carbonate forming molds around the filaments (Fig. 5a); not embedded but having small crystals forming on top of them; or not associated at all with calcium carbonate (Fig. 5b). In the first case most of the time only the outer sheath is preserved (Fig. 5c). Furthermore, the OM structure is still visible as a very thin matrix in the molds, and all-around molds area (Fig. 5d, e, f). On the other hand, some areas display more diversified microbial morphologies (Fig. 5g, h, i), presenting filaments of different diameter but mostly thinner than in the areas of dense filaments (Fig. 5g), as well as coccoids (Fig. 5h, i), bacilli-like (Fig. 5j) and stacks of rounded microorganisms (Fig. 5h). A very fine matrix is also present (Fig. 5i) that in certain areas is associated with small thin minerals (Fig. 5j). The abundance of microorganisms decreases as calcium carbonate increases, forming clusters of very small crystals. The mineral and OM fractions appear to be tightly linked (Fig. 5k), some areas present an alveolar structure made up of an organic sheath and calcium carbonate (Fig. 5l). Calcite is visible above the sheath as a stack of platelet of rhombohedral crystals (Fig. 5m). Rounded structures of possibly microorganisms can be observed under the sheath (Fig. 4n), and in some areas of crack sheath calcium carbonate precipitates are visible (Fig. 5o).

Figure 5 SEM photomicrographs of samples from location A. Panels A, B, C, D, E and F correspond to samples that have been taken from the center of the stream/channel. Panels G, H, I, J, K, L, M, N, O are from samples that have been taken towards the side of the stream/channel. a) Erect filaments (black arrows) associated with carbonate precipitation. b) Filaments (black arrows) of various widths. c) Preserved outer sheaths (black arrows) encrusted in carbonates. d) Fully preserved encrusted filaments (black arrow) under higher magnification. Sheath appearance suggests mineralization. Organic thin sheath can be observed below the filament. e) View of the remaining tubular molds (white arrows). f) Enlargement (dashed-line rectangle in d) of the remaining tubular molds (white arrows). Preserved sheaths can be observed in between (black arrows). Organic sheath associated with the molds. g) Highly diversified microorganisms including filaments of different length and width, coccoids and bacillis. h) Area rich in coccoid bacteria (light green arrows). i) Microorganisms (light green arrows) below an organic sheath. j) Very small minerals precipitates interconnected by very thin organic strands (white arrows). Coccoid and bacilli microorganisms are observed in the same area. k) Thin organic film (orange arrows) interconnected with mineral precipitates (white arrows). l) Ovoid pores presenting mineralized trim (white arrows) and interconnected by a thin organic film (orange arrows). m) Enlargement showing the interconnection between carbonate precipitates (white arrows) and organic film (orange arrows). n) Organic film covering areas presenting apparent rounded microorganisms. o) Enlargement of the area of the rectangle in N, revealing the presence of patched of small minerals (white arrow) under the thin organic film (orange arrows). 

The sample of location B presents diatoms on the top part of the mat. It is composed of filaments of different diameters, abundant bacilli and a very thin organic matrix (Fig. 6a, b). Below, the mat is still composed of abundant bacilli of different sizes, as well as coccoid, diatoms and to a lesser extent filament remains (Fig. 6c, d). A thin, calcium-rich organic matrix is also visible everywhere and in many areas its density creates an OM film (Fig. 6d, e). This calcium-rich phase associated with calcium carbonate minerals is observed below the film and associated with a Mg-Si phase (Fig. 6e).

Figure 6 SEM photomicrographs and EDS analyses of sample from location B. a) Densely populated area composed of large filaments, thinner filaments and smaller bacteria attached either to the filaments or on the surface of an extensive organic matrix made of very thin strands. b) Enlargement of A, showing the organic matrix (white arrows) associated with the bacteria. c) Area displaying thin filaments and diatoms. d) Dense organic film mostly associated with coccoid than ovoid microorganisms. e) Photomicrograph showing clusters of bacteria (white arrows) behind an organic thin film. White dot indicates the location of EDS analysis showing a Ca-rich mineralization (bottom right). 

The mats in location C are similar to location B since both areas display comparatively higher water depths and present a low-to-no energy environment. In both locations microbial mats were attached to basalt pebbles and presented an uppermost yellowish part, the rest of the mat being greenish (Fig. 3). SEM inspection shows areas of filaments associated with clusters of calcium carbonate minerals (Fig. 7a, b) for location B. Other parts are highly mineralized with platelet and arrow shape calcium carbonate associated with OM film (Fig. 7c, d). Some areas present rounded forms underneath the organic film and abundant diatoms (Fig. 7e, f). Finally, certain zones present a very thin organic matrix associated with calcium carbonate precipitate and a Si phase (Fig. 7g, h, i). EDS analyses indicate the presence of Ca and recurrent Mg, Al and Si phases. The mat in location B is less mineralized in comparison to the samples from sites A and C (Fig. 6). Filaments, however, are recurrent in all three samples.

Figure 7 SEM photomicrographs of sample from location C. a) Filamentous network (white arrows) with nodules of carbonate precipitates (black arrows). b) Enlargement of the dense organic filamentous network. c) Small size crystals (black arrows) merging to form platelets. d) Platelets associated with a thin organic film (white arrows). e) Abundant diatoms (black arrows) observed on top of a thin organic film covering rounded cell shapes. f) Interconnection between the organic film (white arrow) and merging arrow-like precipitates (black arrow). Rounded morphologies of up to 2 μm can be observed under the organic film. g) Platelet minerals (black arrows) attached to thin organic string (white arrows). h) Dense organic strings. i) Nanometer carbonate minerals formed in association with thin organic strands (white arrows). 

Analysis of microbial communities based on 16S rRNA gene sequences

The sequencing using the 16SrRNA of the mats provides information on the prokaryote communities present in the samples. Three samples ca. 1mm thick have been taken from the three different mats. Samples from location B and C without clear laminas were sampled at the surface, 3 to 4 mm depth and 6 to 8 mm depth seeking for possible differences in communities between uppermost, center and lower parts of the mats, respectively. ^{Pacton et al. (2015}) described the presence of an upper layer of filamentous and coccoid Cyanobacteria associated with a carbonate phase whereas the lower layer contained abundant SRB and substantially higher content of carbonate. Hence, we have focused the analyses of the results of 16S rRNA mainly on those two groups. Cyanobacteria were hardly detected in the 16S rRNA gene sequences (only a few ASVs, all lost after rarefaction was applied). This being said, microscopic observations attest of the presence of photosynthetic organisms, including diatoms and green-pigmented filaments. It is however clear that phototrophic prokaryotes did not account for an important fraction of the 16S rRNA gene libraries obtained (Fig. 8). As an example, only 3 ASVs affiliated to Chloroflexi were obtained after rarefaction to 2370 taxa per sample.

Figure 8 Bar chart of the abundances of rarefied ASVs affiliations at the phylum level in each sample for locations A, B and C. 

Desulforbacterota (mainly composed of sulfate reducing organisms; ^{Waite et al., 2020}) are much more common, ranging between 14 % (maximum in A3, bottom of mat from zone A) and 0.8 % (minimum in C3, bottom from zone C) (Fig. 9). They are represented in the samples as Desulfobulbia occurring in all samples, Desulfovibrionia and Desulfobacteria (Fig. 9). Desulfurmonada sequences were also obtained, but in lower read abundance (Fig. 9). No clear distinction could be made based on site or microbial mat zone (top/center/bottom), in line with the limited stratification observed for each mat.

Figure 9 Barplot of rarefied Desulfobacterota ASVs at the class level, for each site A, B and C. 

Implication of physico-chemical extrinsic parameters in carbonate precipitation

A difference in water chemistry is obvious between the lakes and the river sampling sites. The lakes present greater fluctuations in chemistry especially CLC, but overall, they tend to have sulfate-rich waters compared to the Maquinchao River, that presents rather bicarbonate-rich waters. Moreover, as the chemical data from the drilled wells are comparable to the river sampling point, it can be suggested that the lakes greater fluctuation in size is mostly related to the bi-water sources feeding it - underground and meteoric- and, thus highly dependent on rainfall (meteorological impact). Although no quantified in this study, evapotranspiration changes and high wind velocities might also impact the lake levels and their water chemistry. The Maquinchao River is more resilient due to the largest groundwater input compared to meteoric waters. Furthermore, although fluctuating over the monitored interval the SIcalcite results show mostly slight saturation (~0.8) to saturation (~1.8) in both lakes. However, the mineralizing mats described in 2011 and microbial mats in general were only reported in some particular areas of the river. It is unlikely that the chemistry only would promote the growth of microbialite in the river rather than in the lakes. Nevertheless, calcium and bicarbonate are critical for the formation of microbialites (^{Farías et al., 2013}). The brackish to saline waters of both lakes could prevent microbial mats proliferation compared to the more freshwater river as proposed by ^{Pacton et al. (2015}). Additionally, changes in certain factors such as chemistry and water level, associated with turbidity of the water, would be concurrent and limiting factors for the development of microbial mats.

To precipitate calcium carbonate SI must be above 0. However, according to ^{Merz-Preiß (2000}), Cyanobacteria encrustation, also known as biologically-influenced mineralization, happens for SI > 0.8 in the water. SI < 0.8 would lead to more biologically induced processes involved in carbonates formation. Both lakes happen to present values above 0.8 but do not show carbonate precipitation. The absence of microbial mats in both lakes probably prevent the precipitation of carbonates as microbial mats often play the role of preferential nucleation sites (^{Riding, 2011}a). In contrast, Maquinchao River SI have been barely reaching 0.8 during the different sampling periods, suggesting that organomineralization would mainly be the result of biologically-induced processes (Merz-Preiß, 2000).

In 2015, however, the mineralization of the mats, at least in location B and C, was low. Lower values for major elements and especially bicarbonate have been recorded, thus affecting, and giving lower SIcalcite in 2015 compared to the subsequent sampling seasons (Fig. 4, tables 1 and 2). The already arid character of the climate in the Maquinchao area has been harden due to a decrease in rainfall from years 2000 to 2011 affecting the lake levels and river as well as the aquifers. However, the years 2014 and 2015 have had comparatively higher rainfall (380mm in 2014; Fig. 2). This meteoric source could have substantially lowered the concentration of major elements in the Maquinchao River. Meteoric waters are less rich in cations and anions, including Ca2+ and HCO3-, in comparison to underground waters, which would have been enriched by the dissolution of the rock reservoir. Meteoric waters would have overtaken underground inputs leading to lower SI in 2015 and consequently diminishing carbonate mineralization. Physico-chemical data between 2011 and 2015 and onward have substantially changed, making their comparison difficult (tables 1 and 2). However, in the light of the chemical information, pH and Ca2+ seem to be more or less stable over the years and seasons except in 2015. Furthermore, the increasing amount of calcium carbonate minerals in the lower part of the mats in 2011 would be in favor of a low SIcalcite as otherwise mats would tend to be more mineralized especially on the upper green part that would serve as a substrate site for abiotic calcium carbonate mineralization. The increasing alkalinity in the lower part of the mat lead to carbonate precipitation by the metabolic action of microorganisms (^{Pacton et al., 2015}). However, in 2014-2015 even if microbial activity could substantially increase the alkalinity and pH within the mat, the depravation in bicarbonate in the environment and calcium in the Maquinchao River due to the increased meteoric contribution compared to groundwater inputs might have decreased the precipitation of CaCO3 in the mat. An additional factor that cannot be neglected is the Puyehue 2011 and Calbuco 2015 volcanic eruptions that had a strong impact in the region and might explain the lack of organomineralization in 2015 compared to what had been observed in 2011.

Sampling location A presents much more mineralization than locations B and C. This mineralization is largely associated with filamentous organisms. Calcium carbonate precipitation associated with filamentous cyanobacteria has been largely described in many sites (Freytet and Verrecchia, 1998; ^{Merz-Preiß, 2000}; ^{Roche et al., 2019}). Furthermore, the presence of a stream in sampling site A implies a higher energy environment compared to locations B and C. Physical degassing of CO2 may occur in streams/fast flowing waters. This phenomenon is one of the potential external parameters involved in carbonate precipitation in lacustrine microbialites (^{Dupraz et al., 2009}). The presence of flowing water and carbonate formation

around the filaments at location A (Fig. 5) favors a substantial abiotic impact in carbonate precipitation. However, biological uptake of CO2 is also possible and should not be neglected. Photosynthesis is known to uptake CO2 and increase alkalinity leading to favorable conditions for carbonate precipitation (^{Dupraz et al., 2009}). This can be achieved by Cyanobacteria and other primary producers such as diatoms. Here given the absence of 16S rRNA gene sequences associated to Cyanobacteria, we suspect the effect of Cyanobacteria, both for the production of alkalinity and EPS, to be minor. Nevertheless, the biofilm DNA extraction protocol, albeit adapted, may have negatively impacted cyanobacterial ASV relative abundance as these organisms are known for their protective sheath thickness. Additionally, diatoms can play a similar positive role in EPS and alkalinity production through photosynthesis, that cannot be neglected here (Figs. 6 and 7).

Potential microbial impact in carbonate precipitation

There are several differences between mats samples. Those from locations B and C are relatively similar encompassing a thin uppermost yellowish layer containing abundant diatoms that is overlying a greenish layer. In location A the uppermost layer is mostly reddish in color containing diatoms whereas the greenish layer is also thinner than in the previous locations. The structures observed in sample A differ from those described in ^{Pacton et al. (2015}) that presented filamentous and coccoid Cyanobacteria as well as some SRB in the uppermost layer of the 2011 living mats containing few crystals of low Mg calcite. The beige layer in the lower part of the mat presented substantial CaCO3 crystals associated with more abundant SRB.

At location B, no major mineralization has been observed neither associated with the large filaments, as described at location A, nor with other coccoids. The low SIcalcite (<0.2) at the location of the sampling (Carvallo point) might have prevented the formation of CaCO3 by extrinsic parameters. However, certain areas present small minerals as a matrix associated with OM remains (Fig. 6). The presence of this very restricted mineralization suggests a very localized alkalinity increase. EDS analyses of the mineralized area show a Mg-Si phase associated with the Ca signal. Mg-Si phases, if not detrital, have been reported in numerous environments and always related to microbial activity (^{Arp et al., 2003}; Souza- Egipsy et al., 2005). These observations suggest that at location B intrinsic parameters (metabolic activity) may be involved in very moderate carbonate precipitation within the microbial mat. As the filaments are free of minerals and Cyanobacteria could not be detected in the genomic data (Fig. 8), heterotrophic microorganisms and in particular SRB (largely detected in our genomic data, Fig. 9) might be implicated, provoking a localized rise of alkalinity leading to carbonate precipitation and potential Mg-Si phase (^{Pace et al., 2018}). Locations B and C are both positioned in comparable environmental settings and an in situ macroscopic inspection of their microbial mats suggest that they are very similar. They both display filaments although less abundant in location C than in B.

The sample from location A has different features than those from locations B and C. In sample A, the highly filamentous part associated with carbonates is very similar to what has been observed in dry crusts from the Maquinchao Basin. Molds of filaments are also observed (Fig. 5d-f). In many cases, an outer sheath is preserved that is often partially mineralized. Mineralization of filament sheaths has been described in various microbialites (^{Merz-Preiß, 2000}; ^{Wacey et al., 2018}; ^{Gong et al., 2019}; ^{Roche et al., 2019}). Filamentous molds preservation is often observed in systems where abiotic precipitation of carbonate is favored, thus, leading to the precipitation of carbonate on the surface of the microorganism/ sheath and described as encrustation. Apart for incrustation, mineralization in the sheath may be also the result of carbonate impregnation (Merz- Preiß and Riding, 1999; Merz-Preiß, 2000; Riding, 2011b) produced by a rise of supersaturation within the film as a result of metabolic activity (Merz- Preiß, 2000) due to Cyanobacteria uptake of HCO3- for photosynthesis in waters of low pCO2. In the studied mat of zone A, a dense carbonate matrix is visible in between filament molds suggesting encrustation (Fig. 5d-f). The presence of both types may suggest that even if extrinsic parameters play a role in carbonate precipitation, metabolic processes may have a very localized impact. These processes might have a changeable impact over time switching mineral precipitation from biotic to abiotic and vice-versa. Thus, our results indicate that these differences might be indistinguishable in fossilized microbialites throughout the geological record.