ArticlePDF Available

Formation of magnesium-smectite during lacustrine carbonates early diagenesis: Study case of the volcanic crater lake Dziani Dzaha (Mayotte - Indian Ocean)

August 2018
Sedimentology 66(3)

DOI:10.1111/sed.12531

Authors:

Vincent Milesi

SRK consulting

Mathieu Debure

Geological and Mining Research Bureau

Show all 12 authorsHide

The volcanic crater lake of Dziani Dzaha in Mayotte is studied to constrain the geochemical settings and the diagenetic processes at the origin of Mg‐phyllosilicates associated with carbonate rocks. The Dziani Dzaha is characterized by intense primary productivity, volcanic gases bubbling in three locations and a volcanic catchment of phonolitic/alkaline composition. The lake water has an alkalinity of ca 0.2 mol∙L−1 and pH values of ca 9.3. Cores of the lake sediments reaching up to one metre in length were collected and studied by means of carbon–hydrogen–nitrogen elemental analyzer, X‐ray fluorescence spectrometry and X‐ray powder diffraction. In surface sediments, the content of total organic carbon reaches up to 20 wt.%. The mineral content consists of aragonite and hydromagnesite with minor amounts of alkaline feldspar and clinopyroxene from the volcanic catchment. Below 30 cm depth, X‐ray diffraction analyses of the <2 μm clay fraction indicate the presence of a saponite‐like mineral, a Mg‐rich smectite. The saponite‐like mineral accumulates at depth to reach up to ca 30 wt.%, concurrent with a decrease of the contents of hydromagnesite and organic matter. Thermodynamic considerations and mineral assemblages suggest that the evolution of the sediment composition resulted from early diagenetic reactions. The formation of the saponite‐like mineral instead of Al‐free Mg‐silicates resulted from high aluminum availability, which is favoured in restricted lacustrine environments hosted in alkaline volcanic terrains commonly emplaced during early stages of continental rifting. Supersaturation of the lake water relative to saponite is especially due to high pH values, themselves derived from high primary productivity. This suggests that a genetic link may exist between saponite and the development of organic‐rich carbonate rocks, which may be fuelled by the input of CO2‐rich volcanic gases. This provides novel insights into the composition and formation of saponite‐rich deposits under a specific geodynamic context such as the Cretaceous South Atlantic carbonate reservoirs. This article is protected by copyright. All rights reserved.

Content uploaded by Mathieu Debure

Content may be subject to copyright.

This article has been accepted for publication and undergone full peer review but has not

been through the copyediting, typesetting, pagination and proofreading process, which may

lead to differences between this version and the Version of Record. Please cite this article as

doi: 10.1111/sed.12531

DR. VINCENT MILESI (Orcid ID : 0000-0003-0873-2049)

Article type : Original Article

Formation of magnesium-smectite during lacustrine carbonates early

diagenesis: Study case of the volcanic crater lake Dziani Dzaha

(Mayotte – Indian Ocean)

VINCENT P. MILESI*1, DIDIER JÉZÉQUEL*, MATHIEU DEBURE†, PIERRE CADEAU*,

FRANÇOIS GUYOT‡, GÉRARD SARAZIN*, FRANCIS CLARET†, EMMANUELLE

VENNIN§, CARINE CHADUTEAU*, AURÉLIEN VIRGONE¶, ERIC C. GAUCHER¶ AND

MAGALI ADER*

*Institut de Physique du Globe de Paris, Sorbonne Paris Cité, Univ Paris Diderot, UMR 7154

CNRS, F-75005 Paris, France (E-mail: vmilesi@asu.edu)

†BRGM, French Geological Survey, Orléans, France

‡IMPMC, Sorbonne Université, MNHN, Paris, France

§Bourgogne University, Dijon, France

¶Total, EP CSTJF, Pau, France

1Present Address - ASU Tempe Campus, School of Earth and Space Exploration, ISTB4 –

BLDG75, 781 E Terrace Mall, Tempe, AZ, USA, 85287-6004

Associate Editor – Jim Hendry

Short Title – Mg-smectitie formation during diagenesis

ABSTRACT

The volcanic crater lake of Dziani Dzaha in Mayotte is studied to constrain the geochemical

settings and the diagenetic processes at the origin of Mg-phyllosilicates associated with

carbonate rocks. The Dziani Dzaha is characterized by intense primary productivity, volcanic

gases bubbling in three locations and a volcanic catchment of phonolitic/alkaline

composition. The lake water has an alkalinity of ca 0.2 mol∙L-1 and pH values of ca 9.3.

Cores of the lake sediments reaching up to one metre in length were collected and studied by

means of carbon–hydrogen–nitrogen elemental analyzer, X-ray fluorescence spectrometry

and X-ray powder diffraction. In surface sediments, the content of total organic carbon

reaches up to 20 wt.%. The mineral content consists of aragonite and hydromagnesite with

minor amounts of alkaline feldspar and clinopyroxene from the volcanic catchment. Below

30 cm depth, X-ray diffraction analyses of the <2 µm clay fraction indicate the presence of a

saponite-like mineral, a Mg-rich smectite. The saponite-like mineral accumulates at depth to

reach up to ca 30 wt.%, concurrent with a decrease of the contents of hydromagnesite and

organic matter. Thermodynamic considerations and mineral assemblages suggest that the

evolution of the sediment composition resulted from early diagenetic reactions. The

formation of the saponite-like mineral instead of Al-free Mg-silicates resulted from high

aluminum availability, which is favoured in restricted lacustrine environments hosted in

alkaline volcanic terrains commonly emplaced during early stages of continental rifting.

Supersaturation of the lake water relative to saponite is especially due to high pH values,

themselves derived from high primary productivity. This suggests that a genetic link may

exist between saponite and the development of organic-rich carbonate rocks, which may be

fuelled by the input of CO2-rich volcanic gases. This provides novel insights into the

composition and formation of saponite-rich deposits under a specific geodynamic context

such as the Cretaceous South Atlantic carbonate reservoirs.

Keywords Authigenesis, early diagenesis, lacustrine carbonates, magnesian clays, organic

matter decomposition, saponite, volcanic lake.

INTRODUCTION

Magnesium-rich clay minerals, referred to as ‘Mg-silicates’ in this study, are increasingly reported

associated with ancient and modern carbonates rocks (e.g. Hover et al., 1999; Bristow et al., 2009;

Wright, 2012; Zeyen et al., 2015, 2017; Pace et al., 2016; Gérard et al., 2018). In particular, Mg-

silicates have been reported in lacustrine carbonate reservoirs of the South-Atlantic (e.g. Bertani and

Carozzi, 1985a, 1985b; Rehim et al., 1986; Wright, 2012) and their dissolution during diagenesis was

proposed to explain the high porosity of these carbonate rocks (Wright and Barnett, 2015; Tosca and

Wright, 2015). These phases seem to form either through direct precipitation from water or during

very early diagenesis (e.g. Tosca, 2015) and record specific chemical conditions, especially high pH

values. Thus, they have a strong potential to help in palaeoenvironmental reconstruction of continental

carbonates deposits and their evolution through the diagenesis (Chamley, 1989; Weaver, 1989;

Bristow et al., 2009).

The variety of geological environments allowing Mg-rich clays to form is unclear. In

Phanerozoic rocks, the presence of Mg-silicates has been attributed to different depositional

environments including weathering profiles developed on Mg-rich rocks such as basic or ultrabasic

igneous rocks (Huang et al., 2013), hydrothermal settings where fluids interact with Mg-rich igneous

material (Setti et al., 2004), alkaline lakes and evaporative basins (Van Denburgh, 1975; Yuretich and

Cerling, 1983; Darragi and Tardy, 1987; Banfield et al., 1991; Hay et al., 1991; Hover et al., 1999;

Martini et al., 2002; Bristow et al., 2009; Zeyen et al., 2015, 2017; Pace et al., 2016; Gérard et al.,

2018). Spring inflows into lakes hosted in basic volcanic terrains could also contribute to create the

right set of conditions, i.e. high carbonate alkalinity, high Si, Mg and Ca concentrations, for the

development of lacustrine carbonates associated with Mg-silicates (Wright, 2012). The Doushantuo

Formation, which hosts fossil records of the early Ediacaran period (635 to 551 Ma), illustrates the

controversy about the formation of Mg-rich clays. In this carbonate formation, the presence of

saponite, a Mg-rich and Al-rich smectite, was interpreted as either an early diagenetic phase formed in

situ in an alkaline and non-marine environment (Bristow et al., 2009), or as a detrital phase resulting

from local to regional weathering of adjacent continental areas dominated by mafic to ultramafic

volcanic rocks (Huang et al., 2013). In this respect, the aluminum content of Mg-silicates may be of

key interest for such reconstructions because it has been proposed to provide insights into the nature

and the amounts of detrital silicate inputs in lacustrine environments (e.g. Millot, 1970).

Present-day sedimentary features can provide information about the palaeoenvironments

where Mg-rich clays formed. Sediments developed in lakes characterized by waters with high pH

(>9), and either relatively high silica concentrations (Van Denburgh, 1975; Yuretich and Cerling,

1983; Darragi and Tardy, 1987; Banfield et al., 1991; Hay et al., 1991; Zeyen et al., 2015, 2017;

Gérard et al., 2018) or high magnesium concentrations (Hover et al., 1999; Martini et al., 2002), were

found to host Mg-silicates. To date, however, the specific geochemical conditions controlling their

formation remain unclear. In particular, the role of volcanic catchment and spring inflow is uncertain,

limiting palaeoenvironmental and early diagenesis interpretations. To tackle this issue, the tropical

volcanic crater lake Dziani Dzaha in Mayotte (Indian Ocean) was chosen as a possible

palaeoenvironmental analogue for the formation model of Mg-rich silicates associated with

continental carbonates. The lake waters are known to be alkaline and the catchment is composed of

volcanic rocks of phonolitic/alkaline composition. Magnesium-silicate minerals were recently

identified in the stromatolites of the lake (Gérard et al., 2018), suggesting that they may also be

present in the soft carbonated sediments. The mineralogy of the first metre of the sediment column

was studied to identify possible occurrences of Mg-silicates and provide insights on the processes

leading to the formation of carbonates associated with Mg-silicates during early diagenesis.

GEOLOGICAL SETTING AND LIMNOLOGY OF THE DZIANI DZAHA

The Dziani Dzaha (which means volcanic crater lake in Mahorian) is a tropical volcanic crater lake on

the island of Petite Terre (Pamanzi) located within the Mayotte island complex in the Northern

Mozambique Channel. Mayotte is the most southerly and oldest island complex of the Comoros

Archipelago. It is made of two main volcanic islands; Grande Terre and Petite Terre (Fig. 1A and B).

The initial volcanism at the origin of the Mayotte complex island is estimated at about 10 to

15 Ma (Nougier et al., 1986); however, a late period of volcanic activity occurred during the last 2.5

Myr (Pelleter et al., 2014). The alkaline volcanism at the origin of the Comoros Archipelago was

suggested as deriving from the partial melt of lithospheric mantle metasomatized by CO2-rich fluids

beneath an ocean–continent transitional crust (Nougier et al., 1986; Mougenot et al., 1989; Bertil and

Reynoult, 1998; Coltorti et al., 1999; Pelleter et al., 2014). In the Mayotte island complex, the island

of Petite Terre was emplaced during the late stage of the volcanic activity. It is mostly made of

pyroclastic rocks of phonolite/alkaline composition resulting from the crystal fractionation of magmas

derived from the partial melting of an oceanic lithosphere enriched in CO2 (Pelleter et al., 2014).

The Dziani Dzaha is located in a volcanic crater formed by a phreatomagmatic eruption. The

pyroclastic rocks rest over a coral reef dated from 9 ka, which surrounds Grande Terre, the largest

island in the Mayotte island complex. As the most recent volcanic event is dated at 4 ka (Zinke et al.,

2003), the age of the Dziani Dzaha may range from 9 ka to 4 ka.

The lake (Dziani Dzaha) is a roughly elliptical shape with a length of 640 m and a width of

470 m. The lake watershed is very restricted. The volcanic crater is ca 1.2 km long and 50 m high on

average. The flanks consist of indurated layers of volcanic ashes and pumices. The lake bottom is

approximately at sea-level. The lake has an average depth of 3 m, with a maximum depth of 5 m,

except for a singular depression which reaches a depth of 18 m (Figs 1 and 2). This depression is

probably related to the phreatomagmatic eruption at the origin of the Dziani Dzaha. Magmatic

degassing is still active in the island of Petite Terre, as evidenced by the four bubbling areas of

volcanic CO2 identified at the lake surface. Microbialites, i.e. organo-sedimentary structures formed in

close association with micro-organisms, thrive in the shallow waters of the lake and are mainly

composed of aragonite and calcite (Gérard et al., 2018). The distribution of the microbialites is patchy

and mostly limited to the lake shore (Fig. 1) such that below a few tens of centimetres of water depth,

most of the sediments are composed of gelatinous and non-indurated material.

The physical, chemical and biological features of the lake are very unique (Leboulanger et al.,

2017). The salinity of the lake water reaches up to 52 psu (practical salinity unit), i.e. 1.5 times higher

than sea water. The alkalinity is ca 0.2 mol∙L-1, with pH values ranging from 9.1 to 9.4. The mean

daily surface temperature varies between 28°C and 36°C. The lake ecosystem is dominated by

prokaryotes with a dense and perennial bloom of filamentous cyanobacteria accounting for more than

90% of the primary producer biomass. Biological activity is significant all year round, with a primary

productivity of ca 8 gC/m2/day, close to the maximum for tropical and subtropical lakes proposed by

Lewis (2011).

MATERIALS AND METHODS

Nomenclature

An example of the scheme used for naming samples follows. For sample DZ14-10 C4, ‘DZ’ indicates

Dziani, ‘14-10’ indicates the year (2014) and the month (October) of the survey and C4 refers to the

sediment core number, which was 4 (see location in Fig. 1). As the sediment cores were numbered

according to their sampling order, they are named in Fig. 1 and in the text only as CX, with X the

number of the core from 1 to 13. Sediment cores C3, C5, C7, C8 and C12 were not dedicated to the

study of the sediment composition and are not used in this study, except for the water content of

sediment core C12. The samples are listed in Table S1 in Supporting Information.

Sampling, sample treatment and water content of sediment

Sediment cores were collected at different sites within the lake during the surveys performed in April

2012, April 2014, October 2014 and August 2016 (Fig. 1; Table S1). Data from nine sediment cores

are presented in this study: C1, C2, C4, C6, C9, C10, C11, C12 and C13. A Uwitec core sampler (90

mm diameter PVC tubes) (Uwitec, Mondsee, Austria) was used to collect sediment from the water–

sediment interface to a maximum depth of 1 m. The sediment cores were subsampled, as much as

possible, on a colour basis in the form of 3 to 5 cm thick slices, which were preserved under anoxic

conditions in a cold room. The samples were freeze-dried, before being rinsed with deionized water,

centrifuged three to four times, and crushed down to <80 µm. In the sediment core C12, the water

content was estimated for 29 samples by weighing fresh samples before and after drying at 60°C. The

precipitation of halite (NaCl) was corrected assuming salinity similar to that of the water column, i.e.

of 52 g∙kg-1 (Leboulanger et al., 2017). Due to the large number of samples, the analytical procedures

presented below could not be performed on all samples.

Total carbon content and carbon isotope composition

The total carbon (TC) content and carbon isotope composition were analysed on 102 samples of seven

sediment cores. The TC content was quantified with a carbon-hydrogen-nitrogen (CHN) elemental

analyzer (Shimadzu TOC V CSH; Shimadzu Corp, Kyoto, Japan) on bulk samples with a precision of

±10%. The isotopic composition of total carbon was analysed with Flash-EA1112 elemental analyzer

coupled to a Thermo Finnigan Deltaplus XP mass spectrometer via a Conflo IV interface (Thermo

Fisher Scientific, Waltham, MA, USA). Four organic laboratory standards were analysed for

concentration and isotopic calibration; VH1, CAP, MX33 and LC. Carbon isotopic compositions are

reported relative to the Vienna Pee Dee Belemnite (VPDB) standard with a precision of ±0.5‰. In

some samples, carbonates or organic matter were removed to measure the content and the isotopic

composition of the carbon end-members. Decarbonatation was performed on 45 samples with HCl

6N. The content of total organic carbon (TOC) and the carbon isotope composition of organic matter

were measured with the same instruments as total carbon. For most samples, significant loss occurred

during the centrifugation step of the decarbonatation process, which precluded accurate quantification

of the TOC content. The measured TOC content is only reported for sediment cores C6 and C9, for

which samples were freeze-dried before centrifugation, which allowed for the limitation of mass loss.

For 19 samples, organic matter was removed using low-temperature oxygen-plasma ashing in a

POLARON PT7160 RF system (Polaron Equipment Limited, Watford, UK) and the carbon isotope

composition of carbonates was measured with a GasBench coupled to a Thermo Finnigan Deltaplus XP

mass spectrometer (Thermo Fisher Scientific). Three laboratory carbonate standards were used for

both carbonate quantification and for isotopic calibration; Across, Merck and Rennes II (details in

Assayag et al, 2006; Lebeau et al., 2014). Isotopic compositions are reported relative to the VPDB

standard for carbon with a precision of ±0.1‰. All carbonate and organic laboratory standards for C-

isotope composition were calibrated independently against NBS19 and IAEA CO 1 international

standards.

X-ray powder diffraction and energy dispersive X-ray fluorescence analyses

X-ray powder diffraction (XRPD) analyses were performed on 128 samples of seven sediment cores

and 12 samples of the volcanic crater with an Empyrean (PANalytical, Almelo, The Netherlands)

diffractometer equipped with a multichannel PIXcel 3D detector (Malvern PANalytical, Malvern,

UK) and a Cu Kα X-ray source (λ = 1.541874 Å). Each pattern was recorded in the θ-θ Bragg-

Brentano geometry, in the 5° to 90° 2θ range (0.0131° for 70 sec). Analysis of X-ray patterns and

mineralogical identification was performed using the High Score Plus software (Malvern

PANalytical), the Crystallography Open Database (COD) and the Inorganic Crystal Structures

Database (ISCD). The reference intensity ratio (RIR) method was used for semi-quantifications of

mineral phases (Visser and de Wolff, 1964); however, reference intensity ratios for clay minerals

were frequently missing from the databases.

Elemental analyses were further conducted by energy dispersive X-ray fluorescence

(EDXRF) on 99 samples of six sediment cores and four samples of the volcanic catchment with an

Epsilon 3XL (PANalytical) spectrometer equipped with a Au X-ray tube. A 10 min measurement

repeated six times for each sample yielded accurate quantitative analysis.

Scanning electron microscopy and transmission electron microscopy analyses

Scanning and transmission electron microscopy were performed on crushed (<80 µm) samples after

freeze-drying and rinsing. Crushed samples were used for microscopic analyses because the

preservation of textural relationships among different components was rendered difficult due to the

gelatinous texture of the sediment (Description of lake sediment textures section). Scanning electron

microscopy (SEM) was performed on 19 samples with a Zeiss Ultra 55 FEG microscope (Carl Zeiss

AG, Oberkochen, Germany) operated at 10 kV and a working distance of 7.5 mm using a

backscattered electron detector for imaging. Transmission electron microscopy (TEM) was performed

with a JEOL FEG 2100F (JEOL Limited, Tokyo, Japan) operated at 200 kV on 1 clay-rich sample

selected on the basis of the XRPD analyses. Both SEM and TEM were equipped with an energy

dispersive X-ray spectroscopy (EDXS) system. The SEM was equipped with a Quantax EDS system

XFlash 4010 (Bruker, Billerica, MA, USA) with a silicon drift detector equipped with an ultrathin

polymer window. The TEM was equipped with a built-in JEOL EDS system with a Si(Li) detector

equipped with an ultrathin polymer window.

Clay fraction

Clay fraction analysis was performed on two clay-rich samples selected on the basis of the XRPD

analyses of bulk samples. Carbonates were removed using the acetic acid-acetate buffer method

(Ostrom, 1961). Organic matter was removed at 50°C by adding small aliquots of H2O2 to the

suspension until gaseous emission had ceased. The <2 µm fraction was separated using gravity

sedimentation of particles in water (Stoke’s law), and then saturated with a 1 M NaCl solution for 24

h at room temperature. Oriented preparations of the Na-saturated <2 µm fraction were carried out on

glass slides and dried at room temperature to obtain an air-dried (AD) preparation. Ethylene glycol

(EG) solvation was achieved by exposing the oriented slides to ethylene glycol vapour for 12 h.

Glycerol (GL) solvation was carried out by saturation of samples with a 2 M MgCl2 solution for 24 h

at room temperature and then exposure to glycerol for three days at 80°C. The air-dried oriented slides

were also heated at 500°C for 90 min and then solvated in EG for 16 h as described in detail by

Christidis and Koutsopoulou (2013). X-ray diffraction patterns were acquired on the oriented slides

over the 2° to 50° angular region, with 0.03° 2θ angular steps, using a Bruker D8 diffractometer

(Bruker, Billerica, MA, USA) with Cu Kα radiation (λ = 1.5418 Å). Counting time was adapted to the

low amount of sample available and was 1.65 sec per step; the two Soller slits were 2.5°. Analysis of

XRD patterns and mineralogical identification was performed using Diffracplus EVA software

(Bruker) and the ICDD database (International Center for Diffraction Data).

Electron microprobe analyses (EMPA) were conducted after coating of the <2 µm clay

fraction with carbon. Spot analyses were performed using a CAMECA SXFive electron microprobe

(CAMECA, Paris, France) with an accelerating voltage of 15 kV, a beam current of 12 nA, and a 1

µm beam diameter. More details on the analytical methods are found in Lerouge et al. (2017).

Cation exchange capacity (CEC) was measured using the hexamine cobalt(III) chloride

method developed by Hadi et al. (2013) and adapted from Orsini and Remy (1976). The CEC is

deduced from two cross analyses; the amount of hexamine cobalt(III) that disappears from the

solution (adsorbed on the clay), which is directly measured by spectrophotometry at 475 nm, and

analysis by ionic chromatography [high performance liquid chromatography (HPLC) Dionex

Corporation, Sunnyvale, CA, USA] of the major cations present in the supernatant (Na+, K+, Ca2+ and

Mg2+) which also indicates the relative populations of desorbed exchangeable cations.

Quantification of sediment composition

Based on XRPD and XRF analyses, bulk carbon contents and carbon isotope compositions,

calculations were performed to quantify the contents of the different mineral phases and of the organic

matter. The organic matter content was calculated in two steps. First, the content of total organic

carbon (TOC) was calculated using an isotope mass balance calculation. The total carbon content is a

mixture between organic and inorganic carbon, for which the isotopic compositions were measured in

this study. The contribution of organic carbon x in the total carbon content can be calculated with the

equation:

δ13CTC = x * δ13CTOC + (1-x) * δ13CTIC (1)

with δ13C of TC, TOC and TIC being the isotopic composition of total carbon, total organic carbon

and total inorganic carbon, respectively. Then the TOC content is given by:

TOC = x * TC (2)

Second, the content of organic matter was calculated considering a TOC to organic matter weight

ratio of 1.9 (Broadbent, 1953). The mineral content is deduced by simple mass balance. For the

sediment core C13, as the carbon content and the carbon isotope composition were not measured, the

content of organic matter is considered to be the same as in C6, which is a reasonable assumption

since the two cores were collected in the 18 m deep depression of the lake.

The content of each mineral phase was calculated using the XRF analyses, which requires

fixing the chemical formula of each phase. Empirical formulas were used for most minerals (Table 1).

For saponite, the structural formula has been determined in this study (Thermodynamic model

section). On the basis of SEM-EDS analyses of alkaline feldspars and volcanic pumices in lake

sediment samples, the empirical formula of anorthoclase is considered for these phases. Then, the

content of each mineral is calculated based on a given element.

Thermodynamic calculation

Mineral stability domains in the system Mg-Ca-C-O-H and Mg-Al-Si-O-H were calculated at 30°C

and 1 bar with the Thermoddem database (Blanc et al., 2012) and compared with lake water

compositions (Leboulanger et al., 2017). The speciation of dissolved species was calculated with the

Phreeqc software (Parkhurst and Appelo, 2013).

RESULTS

Description of lake sediment textures

Three main sediment textures have been observed within the lake (Fig. 2A). In the sediment cores

collected at water depths ranging from 1 to 5 m; which represents most of the lake area, the sediment

texture is highly gelatinous and characterized by horizontal laminae (Fig. 2B). The uppermost

sediment is markedly laminated with colours ranging from light to dark brown. The laminae are

mostly <1 mm thick but seem to be organized, on a colour basis, in laminae packages of few

centimetres in thickness, and were used when possible to describe the slices cut from the sediment

cores. In most cases, grain size was indistinguishable to the eye, except in rare mudstone layers which

contain milimetre-sized grains. Below ca 50 cm depth, the laminae tend to disappear and the sediment

changes to a more massive and darker texture. The water/rock weight ratio decreases from ca 33 in

surface sediments to ca 4 at 20 cm depth and ca 2 at 1 m depth (Fig. 2C). Thus, water is by far the

dominant phase of the sediment and most of the sediment compaction occurs in the first 20 cm below

the sediment surface. This sedimentary facies is very homogeneous and no lateral variation has been

observed in the area of 1 to 5 m of water depth.

The sediment cores C12 and C13 collected in the depression of the lake show a similar

sediment texture (highly gelatinous, laminated and with very fine grain size); however, deformation of

the laminae is observed below ca 30 cm deep (Fig. 2D). In the sediment cores close to the shoreline in

shallow (<1 m) water depth (C9 and C11), the sediment is greyish and grain size can increase up to

millimetres in size. In C9, the sediment texture is close to that of a mudstone and remains relatively

gelatinous and laminated. In C11, the sediment texture is close to a packstone and laminae are not

visible. The thickness of the sediment column is unknown; however, attempts to collect longer

sediment cores (C3, C4 and C5) around 4 to 5 m of water depth suggests that it is thicker than 2.3 m

in this area.

Total carbon content and isotopic composition of the lake sediments

The content of total carbon decreases from ca 20 wt.% in the uppermost sediment to ca 10 wt.% at

depth, except for the sediment cores collected in the depression of the lake (C6) and close to the

shoreline (C9 and C11) (Table S2). In the depression of the lake (C6), the TC content ranges between

20 wt.% and 25 wt.% and is relatively constant with depth. The TC content in C9 and C11 is lower

with values mostly below 10 wt.%. The sediment core C2 shows a local decrease of TC content down

to 3 wt.% at 7 to 14 cm depth. The TOC content measured in the sediment core C6 is high. Values

range mostly between 10 wt.% and 20 wt.% with the highest values in the uppermost sediment.

The isotopic composition of total carbon varies between ca -15‰ and ca +15‰ depending on

the sediment core (Table S2). The highest values are found in C9 and C11. C4 and C6 show the

lowest values. The variations of δ13C values of total carbon between the sediment cores reflect

variable proportions of organic and inorganic carbon. The isotopic compositions of the two carbon

end-members are relatively constant with depth, with average δ13C values of ca -14‰ for organic

carbon and ca +16‰ for inorganic carbon (Table S2). These isotopic compositions are used together

with the isotopic composition of total carbon to determine the TOC content using an isotopic balance

calculation (Quantification of sediment composition section). For sediment core C6, the calculated

and measured TOC content is perfectly consistent, which validates the isotope and mass balance

results (Table S2). Sediment cores C4, C6 and C10 show high TOC contents up to 24 wt.%, which

decrease with depth. The TOC content is lower in the sediment cores close to the shore line (C9 and

C11), with values mostly below 5 wt.%.

Chemical and mineralogical characterization of the sediment cores

Bulk analysis of the sediment cores (XRF)

The mineral content consists mostly of silicon, magnesium, calcium and aluminum (Table S3). Except

for the sediment cores collected at the shoreline (C11) and in the lake depression (C6 and C13), all

sediment cores (i.e. C4, C9 and C10) show a decrease in magnesium content with depth and a

concurrent increase of aluminum and silicon contents, while the calcium content remains relatively

stable. In sediment core C11, the evolution of the chemical composition with depth is approximately

the same, but the aluminum and silicon contents are higher than in C4, C9 and C10. In sediment cores

C6 and C13, the calcium, magnesium, aluminum and silicon contents are relatively stable with depth.

X-ray powder diffraction analyses of sediment cores

Aragonite, hydromagnesite and diopside or clinoenstatite, referred to hereafter as clinopyroxene, are

detected in the uppermost sediment (see Table S4). This mineral assemblage is relatively constant up

to ca 20 cm depth and occurs in all sediment cores collected under a water depth of >1 m (i.e. C2, C4,

C6, C10 and C13). Semi-quantifications with the RIR method indicate contents of aragonite,

hydromagnesite and clinopyroxene of 40 wt.%, 40 wt.% and 20 wt.%, respectively. Minor phases

such as magnetite, dolomite and quartz are sometimes identified with abundances of ca 5 to 10wt.%.

Halite is also occasionally detected but results from precipitation during freeze-drying. Some

sediment cores present a peculiar assemblage; high amounts of alkaline feldspars in C11 close to the

shoreline, calcite and alkaline feldspars between 7 cm and 14 cm depth in C2, magnesite in the

uppermost sediment of C9 and pyrite and dolomite at depth.

In addition to the aragonite–hydromagnesite–clinopyroxene assemblage, clay minerals occur

at depth, characterized by a diffraction peak close to 14.5 angstroms (Å), consistent with the basal

spacing (001) of smectites, chlorite or vermiculite (Fig. 3; the precise characterization of clay

minerals is presented later in the paper). Overall, clays were detected in all sediment cores below ca

10 to 15 cm, except for the sediment cores collected at the shoreline (core C11; 30 cm long) and in the

depression of the lake (cores C6 and C13), in which clays may be absent or below the XRD detection

limit. Concurrent to the occurrence of clays, the content of hydromagnesite decreases with depth until

complete disappearance in the sediment cores sampled between 2 cm and 5 m of water depth (C2, C4,

C9 and C10). In the other cores (C6, C11 and C13), the content of hydromagnesite remains relatively

constant with depth.

Scanning and transmission electron microscopy

Observations with a scanning electron microscope and chemical mapping show Mg-rich minerals

forming platelets, typical of hydromagnesite, together with Ca-rich minerals corresponding to

aragonite (Fig. 4A and B). Associated with the carbonates, a mineral phase rich in silicon, magnesium

and aluminum is recognized (Fig. 4A to C). Observations with transmission electron microscope of

minerals of similar chemical composition highlight the sheet structure of the phase (Fig. 4D and E).

Significant amounts of volcanic glass containing silicon, aluminum and lower amounts of sodium and

potassium, is also observed (Fig. 4F and G). The chemical composition is close to that of anorthoclase

(Na0.7K0.3AlSi3O8) (Fig. 4H). This is consistent with the XRF and XRPD analysis of volcanic pumices

and ash layers of the volcanic catchment (see Appendix S1, Fig. S1, Tables S5 and S6).

Characterization of the clay fraction

The <2 µm clay fraction was extracted from two samples: (i) one from 27 to 30 cm deep in sediment

core C4 which corresponds to a layer rich in clays (Table S4); and (ii) one from 67 to 72 cm deep in

sediment core C6. The XRD patterns acquired on oriented slides are similar for the two samples. It

may be observed that although the clays were Na-saturated, the 001 reflection was at ca 14.5 Å

instead of 12.5 Å as expected for relative humidity conditions normally occurring in the laboratory

(Brindley and Brown, 1980) (Fig. 5A and B). As illustrated further in the manuscript, CEC

measurement carried out on this Na-saturated fraction indicates that the exchanger still contains a

significant amount of magnesium, which therefore explains this value. The swelling of the (001) from

14.5 Å in the air-dried state towards 17 Å in the ethylene glycol state (Fig. 5B) indicates the presence

of smectite or vermiculite (Brindley and Brown, 1980). Powder XRD indicates that the (060) peak

position is 1.53 Å (Fig. 5A), which is characteristic of trioctahedral phyllosilicates (Moore and

Reynolds, 1997). Based on EPMA analysis (for example, no Li and Zn that ruled out hectorite and

sauconite, presence of Al that ruled out stevensite) and the fact that the phyllosilicate is trioctahedral,

the remaining possibilities are saponite and vermiculite (Brindley and Brown, 1980; Debure et al.

2016). Distinguishing between smectite and vermiculite is not easy. However, glycerol solvation on

Mg-saturated samples, led to swelling with a basal reflection at 16.5 Å (Fig. 5C). This is consistent

with the presence of saponite as vermiculite is not supposed to swell under these conditions (Walker,

1958; Brindley, 1966; Suquet et al., 1975; Tosca et al., 2011). The CEC of ca 87.1 meq/100 g is also

consistent with saponite, whereas values of ca 120 meq/100 g are reported for vermiculite (Brindley

and Brown, 1980). In addition, the fact that when heated to 500°C, the sample still showed a 001

reflection expanding to 17.4 Å after ethylene glycol solvation (Fig. 5B) further supports the saponite

hypothesis (Christidis and Koutsopoulou, 2013).

Cation population obtained from CEC measurements is consistent with EPMA results (Table

2). The two methods led to similar Na+ content values, which is the main interlayer cation. Ca2+ and

K+ values are a bit higher by EPMA because part of K+ present in the inner sphere complexes might

be difficult to extract with the hexamine cobalt method (Hadi et al., 2013). Magnesium content

obtained by EPMA was ten times higher than the value obtained by CEC. The structural Mg content

was deduced from the difference between EPMA and CEC measurements.

The mean structural formula was determined assuming that Al was preferentially tetrahedral,

the other cations being octahedral, and that Na, Ca and K were preferentially in the interlayers

(Lerouge et al., 2017). The calculation revealed that Fe was not in tetrahedral sites but in the

octahedral ones. In order to equilibrate the charge, Fe was only in its ferric form (Fe3+). Based on all

of those considerations, the mean structural formula is

K0.01Mg0.12Ca0.02Na0.28(Mg2.22Fe0.18Al0.28□0.32)(Si3.61Al0.39)O10(OH)2 where □ is a vacant site. Due to its

trioctahedral character, this formula is not in agreement with the clay mineralogy nomenclature

defined by Guggenheim (2006) for saponite; however, it is consistent with literature data that reported

vacancies (from 0.15 to 0.56 depending of the studies) in the octahedral position of ferric saponite

(Post, 1984; Kodama et al., 1988; Vincente et al., 1996; Hicks et al., 2014). In addition, the calculated

charge per formula unit (0.57) is below 0.6 which is considered in the nomenclature as the boundary

between saponite and vermiculite. The differences observed in the structural formula could be related

to the occurrence of minor amounts of other clay minerals or minor non-clay impurities (amorphous

and/or crystalline phases) in the samples that have not been identified. The terminology ‘saponite-like

mineral’ will be used in this paper to be consistent with the nomenclature and the characterization.

Quantification of the sediment content

Results of the calculation of the evolution of mineral composition and organic matter content with

depth are shown in Fig. 6 (Table S7). Apart for the sediment core C11, organic matter represents more

than 25 wt.% of the uppermost sediment, with a maximum of 45 wt.% for sediment core C6. In

sediment cores C4 and C10, hydromagnesite decreases with depth while the saponite-like mineral

accumulates to reach up to 33 wt.% at ca 70 cm deep in C10. Concurrently, the content of organic

matter decreases. The content of aragonite is relatively constant around 25 wt.% whereas the

proportion of primary silicates, dominated by alkaline feldspar, increases with depth to reach up to ca

25 wt.%. The sediment core C11 shows much higher amounts of alkaline feldspars than the others

sediment cores, up to 60 wt.%. Compared with C4 and C10, the sediment cores sampled in the

depression of the lake (i.e. C6 and C13) have lower amounts of saponite-like mineral at depth and

relatively stable contents of hydromagnesite.

The calculation should fulfil an internal consistency such that the elemental abundances

corresponding to the calculated mineral contents should equal the measured elemental abundances.

For instance, the sum of silicon content in alkaline feldspar and in the saponite-like mineral, which is

calculated independently using the measured content of potassium and aluminum, respectively, should

not exceed the total measured content of silicon. For sediment cores C4, C6, C10 and C13, the

calculated silicon content exceeded by 10 mol % the measured silicon content for only ca 10% of

samples. These low errors validate the calculations for these sediment cores. In contrast, the

calculation for C9 is not shown due to a lack of internal consistency. Core C9 was collected close to

the shoreline, where detrital influence is high. In this case, considering only alkaline feldspar and

pyroxene as detrital inputs might be an oversimplification.

Sensitivity tests have been performed on several parameters including the TOC to organic

matter weight ratio (±0.1 unit), the empirical formula of clinopyroxene (enstatite or diopside) and the

chemical composition of the saponite-like mineral (±0.1 mol/mol of mineral on the Al, Si and Mg

content). Most of these parameters modify the percentage content of each phase by less than 10%,

except for aragonite (up to 20%).

Thermodynamic model

The chemical compositions of lake waters (Table S8) (Leboulanger et al., 2017; Gerard et al., 2018)

are compared to the stability domains of carbonate and brucite in the system Mg-Ca-C-O-H (Fig. 7A).

The pore waters are close to metastable equilibrium between aragonite and hydromagnesite, and

supersaturated relative to dolomite.

The stability domains of Mg-silicates are represented in the system Mg-Al-Si-O-H,

considering that aluminum behaves conservatively between the Al-bearing phases (Fig. 7B). Chlorites

are not considered in the calculation as their formations are kinetically limited at ambient temperature

and pressure (Meunier, 2005). High pH values together with high activities of Mg2+ and dissolved

silica favour the stability of saponite compared to other aluminosilicates. The chemical composition

of lake water is close to equilibrium with quartz and supersaturated relative to saponite, kerolite, talc

and sepiolite.

DISCUSSION

Sediment composition and distribution

Three main sedimentary facies were observed within the lake, which respond to a shore to basin

central lake transect. Close to the shoreline, the greyish facies dominated by millimetre-size grains is

relatively poor in organic matter and rich in primary silicates, alkaline feldspars and pyroxene, which

highlights high detrital input from the volcanic catchment. In the area of the lake, 1 to 5 m in depth,

the gelatinous facies is dominated by organic matter and indicates low lateral variation. The detrital

components decrease with distance from the lake shore. The laminated uppermost sediment is mostly

formed of aragonite and hydromagnesite whereas, at depth, the sediment texture becomes massive and

dark, and the hydromagnesite is replaced by the saponite-like mineral. The layers containing coarser

grains and enriched in calcite and alkaline feldspar with less organic carbon (sediment core C2)

reflects a local higher energy detrital input to the lake margin. In the central lake depression, soft-

sediment deformation is interpreted as slump events (Alsop and Marco, 2011). In this area, the

apparently undeformed uppermost sediment is also rich in organic matter, aragonite and

hydromagnesite but hydromagnesite persists at depth and the saponite-like mineral occurs in much

lower amounts. No clay mineral has been detected in the uppermost part of sediment cores suggesting

low amounts of detrital clays inherited from the weathering of the surrounding volcanic rocks.

Overall, the sediment deposited below 1 m of water depth is singular by its extremely high

content of organic matter (up to 45 wt.%), a consequence of intense primary productivity in the lake,

and by the anomalously positive carbon isotope signature of both carbonates and organic matter,

which probably resulted from a combination of volcanic CO2 input and methanogenesis. The

sediment collected between 1 m and 5 m of water depth is the most representative of the sedimentary

deposits observed within the lake. It is used hereafter as a case study to develop a better understanding

of the co-occurrences of carbonates and Mg-silicates.

Thermodynamic and kinetic controls on carbonates and saponite-like mineral formation

Lake water composition is close to equilibrium with aragonite and hydromagnesite and the presence

of these two carbonates in the uppermost sediment indicates that they precipitate relatively rapidly

from lake water, either in the water column or at the sediment–water interface. The occurrence of

aragonite is consistent with a Mg/Ca molar ratio of >10 of the lake water, which has been shown to

favour the formation of aragonite rather than calcite (Simkiss, 1964; Taft, 1967; Debure et al., 2013a

and b; Zeyen et al., 2017) Although thermodynamics suggests that magnesite and dolomite are the

most stable Mg-bearing carbonates at surface temperature and pressure, kinetics exerts a strong

primary control on their formation (e.g. Sayles, 1973; Arvidson and MacKenzie, 1999; Gautier et al.,

2014), which may explain why they are rarely observed.

Despite supersaturation of the lake water relative to saponite, kerolite, talc and sepiolite, none

of these phases were detected in the mineral assemblage of the uppermost sediments, suggesting that

they do not precipitate from the lake water. Precipitation of Al-free Mg-silicates from lake water is

nonetheless possible, as exemplified by the magnesium depletion in Lake Turkana attributed to the

precipitation of stevensite (Yuretich and Cerling, 1983; Cerling, 1996). Tosca (2015) proposed a

critical supersaturation allowing Al-free Mg-silicates to overcome kinetic barriers and to precipitate

directly from the water column. In the Dziani Dzaha, the removal of magnesium by hydromagnesite

precipitation likely hinders the achievement of this critical supersaturation, preventing the

homogeneous nucleation of Al-free Mg-silicates in the water column. Regarding the Mg-

aluminosilicates, they are expected to form mostly through heterogeneous nucleation (e.g. Tosca,

2015; Tosca and Wright, 2015), because of the low solubility of aluminum in most cases and

especially in alkaline lakes (2.2 10-6 mol L-1 at pH 9). In this process, detrital Al-bearing silicates act

as mineral precursors, lowering the supersaturation required for the formation of aluminosilicates. In

the Dziani Dzaha, alkaline feldspars and volcanic glasses are the likely Al-bearing precursors,

allowing the formation of the saponite-like mineral in the sediment.

Achievement of high pH values in the water column is one of the chief factors that enable

supersaturation relative to saponite, aragonite and hydromagnesite. Consistently, Mg-silicates

associated with carbonate sediments have been widely reported in alkaline lakes and evaporative

basins of pH values >9 (Van Denburgh, 1975; Yuretich and Cerling, 1983; Darragi and Tardy, 1987;

Banfield et al., 1991; Hay et al., 1991; Hover et al., 1999; Martini et al., 2002; Bristow et al., 2009).

As for other alkaline and hypereutrophic lakes (Lòpez-Archilla et al., 2004), the intense primary

productivity in the Dziani Dzaha is the most likely mechanism responsible for these pH values,

according to the simplified reaction:

HCO3- + H2O = “CH2O” + OH- + O2 (3)

The development of massive primary productivity could be linked to the input of volcanic CO2 within

the lake, which probably supplies primary productivity with carbon.

In addition to pH, magnesium and silica concentrations are critical for the formation of

saponite. The magnesium concentration of ca 4 mM is comparable to the ones of other alkaline lakes

producing Mg-silicates (Van Denburgh, 1975; Yuretich and Cerling, 1983; Darragi and Tardy, 1987;

Banfield et al., 1991). However, due to the competition for magnesium between Mg-silicates and Mg-

carbonates, silica might have a greater influence than magnesium on the formation of saponite (Zeyen

et al., 2017). The silica concentrations (ca 0.3 mM) are around two orders of magnitude higher than

the usual concentration of surface oceans (<2 µM; Treguer et al., 1995), where the efficiency of

diatoms in precipitating silica drastically lowers the concentration (e.g. Siever, 1992; Racki and

Cordey, 2000). Because no spring inflows have been identified, these high silica concentrations are

attributed to the weathering of volcanic terrains in a diatom-poor lacustrine environment.

Early diagenetic evolution of sediment composition

The evolution of the sediment composition with depth is the likely result of early diagenetic reactions.

The appearance of the saponite-like mineral only at depth, in spite of its supersaturation in the lake

water, can be explained by the intrinsic kinetics limit of saponite precipitation (e.g. Tosca, 2015) and

by the kinetics of dissolution of the Al-bearing detrital phase. The mineralization of organic matter is

evidenced by the decrease of TOC content with depth. In the right proportions, the CO2 produced by

the mineralization of organic matter could account for a decrease of pH destabilizing hydromagnesite

while allowing both aragonite and the saponite-like mineral to remain supersaturated. In addition, a

genetic relationship may exist between hydromagnesite and the saponite-like mineral and affect the

pH. Calculation of mineral stability domains (Fig. 7B) shows that the precipitation of saponite from

supersaturated water tends to decrease the pH values, favouring the destabilization of hydromagnesite,

which, in return, could supply with Mg the formation of saponite (Fig. 8). In addition to organic

matter mineralization, CO2-rich volcanic gases seeping from the volcanic basement could contribute

to the pH decrease of pore waters.

Changes of palaeoenvironmental conditions throughout the lake history are less consistent

than early diagenetic reactions to explain the evolution of the sediment composition. If the decreasing

amount of hydromagnesite with depth was due to lower Mg concentrations in the palaeo-lake waters

and/or lower pH values than currently observed, saponite supersaturation would also be lower, which

is not compatible with the increasing amount of saponite-like mineral at depth. The diagenetic

scenario is thus the most likely, even though it remains to be validated. This would require the

assessment of the saturation state of pore waters and the quantitative evaluation, with reactive

transport modelling, of the diagenetic reactions occurring in the sediment, including saponite

precipitation, hydromagnesite destabilization and organic matter mineralization.

Implications for the formation and distribution of Mg-silicates of the South Atlantic rift basins

Magnesium-silicates (stevensite, kerolite and talc) have been identified in lacustrine Cretaceous

carbonate reservoir rocks of the South Atlantic (Bertani and Carozzi, 1985a, 1985b; Rehim et al.,

1986; Wright, 2012; Tosca and Wright, 2015); their formation was suggested to occur in lakes where

volcanic terrains predominated (Cerling, 1994; Wright, 2012). However, the source of Si and Mg to

account for the amount of Mg-silicate found in those rocks remains under discussion. Wright (2012)

suggested that spring inflows, in addition to the weathering of volcanic terrains, might have been

required to enable the formation of both carbonate and Mg-silicates.

The study case of the Dziani Dzaha suggests that high silica and magnesium concentrations

can be reached in lakes through the weathering of volcanic alkaline terrains, without necessarily

involving external inputs such as spring or hydrothermal inflows. The development of an intense

primary productivity, sustained both by the magmatic CO2 and by the nutrients supplied by the

volcanic terrains (such as iron, potassium and phosphorous), is most likely to be responsible for the

high pH values. Thus, a genetic link may exist between volcanic alkaline terrains, organic-rich

sediment and Mg-silicates. In the Doushantuo Formation (635 to 551 Ma), saponite is associated with

TOC contents reaching up to 4 wt.% (Bristow et al., 2009). Using a carbon to organic matter ratio of

1.9 (Broadbent, 1953) and considering 66 wt.% loss of initial organic carbon by thermal maturation

(Tissot and Welte, 1984), the initial content of organic matter in the Doushantuo sediments may have

been of ca 23 wt.%, which is comparable to the Dziani Dzaha sediments. The deposition of lacustrine

carbonates containing more than 20 wt.% of organic matter could have required a significant source

of carbon in addition to the atmospheric CO2. In the Dziani Dzaha, the volcanic CO2 may have played

a major role by fuelling both the primary productivity and carbonate formation. Similarly, the

Cretaceous lakes at the origin of the South Atlantic carbonate reservoir rocks could have been

influenced by CO2-rich magmatic inflows during the continental rifting, as the continental crust

becomes thinner and the influence of mantle fluids increases.

The absence of Al-free Mg-silicates in the Dziani Dzaha sediment, whereas they prevail in the

carbonate reservoirs of the South Atlantic, supports two mutually compatible hypotheses as to the

factors controlling the presence or absence of Al in Mg-silicates. At the local scale, this supports the

‘clay-zoning’ scenario of Millot et al. (1970), who suggested that the presence of detrital Al-bearing

silicates at the basin margin allows Al-rich clay minerals to form, whereas, towards the basin centre,

lower amounts of detrital materials indirectly promote the homogeneous nucleation of Al-free Mg-

silicates (Weaver and Beck, 1977; Jones and Weir, 1983; Jones, 1986; Calvo et al., 1999; Deocampo

et al., 2009; Galán and Pozo, 2011). Accordingly, the occurrence of a saponite-like mineral in the

Dziani Dzaha is consistent with the presence of detrital component in the sediments and the

prevalence of Al-free Mg-silicates in the South Atlantic carbonate reservoirs would rather indicate

large lake systems with open water lake deposits free from detrital inputs.

At the regional scale, the formation of saponite or Al-free Mg-silicates may also reflect the

nature of the volcanic terrains hosting the lacustrine environments. At the early stage of continental

rifting, low mantle uplift results in low partial melting, producing alkaline magma (e.g. White and

McKenzie, 1989; Wilson, 1992). Below the thick continental lithosphere, the slow cooling of magmas

generates highly differentiated basalts enriched in alkaline elements, aluminum and silicon, i.e. a

phonolite/alkaline composition comparable to the one of magmas at the origin of the Mayotte Islands.

In contrast, during later stages of continental rifts, higher partial melting and lower magmatic

differentiation produce tholeiitic basalts with lower aluminum content. The relatively low availability

of aluminum in the Cretaceous lakes at the origin of the South Atlantic carbonate reservoirs may

indicate lacustrine environments formed during the late stage of continental rifting. This is consistent

with the suggestion of Tosca and Wright (2015) that the water chemistry of these lakes may have been

influenced by the serpentinization of exhumed mantle at the transition between continental and

oceanic crusts.

CONCLUSIONS

The Dziani Dzaha sediment exhibits a particular evolution of composition in the first metre of

sediment, which can be explained by early diagenetic reactions within the lake sediments; these are

mainly gelatinous and exhibit homogeneous facies distribution with local increases in detrital inputs

close to the shore. Mineralization of organic matter within the sediment, possibly associated with

inputs of volcanic CO2, can explain the destabilization of hydromagnesite and the decrease of organic

matter content with depth, while aragonite remains stable. Kinetic limitation prevents the formation of

the saponite-like mineral in the uppermost sediment, such that it only accumulates below 15 cm depth.

Reactive transport modelling is now needed to allow a quantitative evaluation of the role of all

diagenetic processes occurring in these sediments: saponite precipitation, hydromagnesite

destabilization, organic matter mineralization and CO2-rich gas inputs.

The formation of the saponite-like mineral in the Dziani Dzaha results from the high silica

activities and the high pH values of the lake waters. This water chemistry is attributed to the delivery

in a restricted lake system of chemical elements, among which nutrients, produced by the weathering

of volcanic terrains, which favours the development of an intense primary productivity. The bubbling

of magmatic gases in the Dzaini Dzaha suggests that magmatic CO2 may fuel the primary productivity

and the carbonate precipitation. Thus, this study suggests that a genetic link may exist between

volcanic contexts and carbonate sediments rich in organic matter and Mg-silicates, which may be key

in understanding their formation throughout the fossil record. As an example, the migration through

the crust of mantle-derived fluids during continental rifting could have influenced the formation of the

Cretaceous South Atlantic carbonate reservoir rocks. In these rocks, the formation of Al-free Mg-

silicates rather than saponite may be indicative of low aluminum availability, favoured in large

lacustrine environments hosted in poorly differentiated volcanic terrains commonly emplaced during

late stages of continental rifting.

ACKNOWLEDGEMENTS

This work was supported by IPGP, TOTAL (project FR00008189), Agence Nationale de la Recherche

(France) (ANR DZIANI, grant number ANR-13-BS06-0001), Total Corporate Foundation (project

755 DZAHA) and one INSU-INTERRVIE grant (grant number AO2013-785992). The Deep Carbon

Observatory community is thanked here for several informative discussions on the importance of

studying environments related to magmatic CO2 emissions. The authors also wish to thank their

colleagues (C. Leboulanger, C. Bernard, P. Got, E. Fouilland, M. Bouvy, E. Le Floch, V. Grossi and

D. Sala) for their support and assistance during sampling campaigns on Dziani Dzaha. Last but not

least, they thank the Air Austral Airline Company and Alexandra and Laurent at the ‘Les Couleurs’

Guest House in Mayotte for their valuable assistance and support.

REFERENCES

Alsop, G.I. and Marco, S. (2011) Soft-sediment deformation within seismogenic

slumps of the Dead Sea Basin. Journal of Structural Geology, 33, 433-457.

Arvidson, R.S. and Mackenzie, F.T. (1999) The dolomite problem; control of

precipitation kinetics by temperature and saturation state. American Journal of Science, 299,

257-288.

Assayag, N., Rivé, K., Ader, M., Jézéquel, D. and Agrinier, P. (2006) Improved

method for isotopic and quantitative analysis of dissolved inorganic carbon in natural water

samples. Rapid Communications in Mass Spectrometry, 20, 2243-2251.

Banfield, J.F., Jones, B.F. and Veblen, D.R. (1991) An AEM-TEM study of

weathering and diagenesis, Abert Lake, Oregon, II. Diagenetic modification of the

sedimentary assemblage. Geochimica and Cosmochimica Acta, 55, 2795–2810.

Bertani, R.T. and Carozzi, A.V. (1985a) Lagoa feia formation (Lower Cretaceous),

Campos basin, offshore Brazil: rift valley stage lacustrine carbonate reservoirs-I. Journal of

Petroleum Geology, 8, 37-58.

Bertani, R.T. and Carozzi, A.V. (1985b) Lagoa feia formation (Lower Cretaceous),

Campos basin, offshore Brazil: rift valley type lacustrine carbonate reservoirs-II. Journal of

Petroleum Geology, 8, 199-220.

Bertil, D. and Regnoult, J.M. (1998) Seismotectonics of Madagascar. Tectonophysics,

294, 57-74.

Blanc, P., Lassin, A., Piantone, P., Azaroual, M., Jacquemet, N., Fabbri, A. and

Gaucher, E.C. (2012) Thermoddem: A geochemical database focused on low temperature

water/rock interactions and waste materials. Applied Geochemistry, 27(10), 2107-2116.

Brindley, G.W. (1966) Ethylene Glycol and Glycerol Complexes of Smectites and

Vermiculites. Clay Minerals 6, 237-259.

Brindley, G.W. and Brown, G. (1980) Crystal Structures of Clay Minerals and their

X-Ray Identification. Mineralogical Society Monograph 5, Mineralogical Society of Great

Britain and Ireland, London, 495 pp.

Bristow, T.F., Kennedy, M.J., Derkowski, A., Droser, M.L., Jiang, G.Q. and Creaser,

R.A. (2009) Mineralogical constraints on the palaeoenvironments of the Ediacaran

Doushantuo Formation. Proceedings of the National Academy of Sciences of the USA, 106,

13190–13195.

Broadbent, F.E. (1953) The soil organic fraction. Advances in Agronomy, 5, 153-183.

Calvo, J.P., Blanc-Valleron, M.M., Rodriguez-Arandia, J.P., Rouchy, J.M. and Sanz,

M.E. (1999) Authigenic clay minerals in continental evaporitic environments. In:

Palaeoweathering, Palaeosurfaces and Related Continental Deposits (Eds M. Thiry and R.

Simon-Coinçon R), International Association of Sedimentologists, Special Publications, 27,

129–151.

Cerling, T.E. (1996) Pore water chemistry of an alkaline lake: Lake Turkana, Kenya.

In: Limnology, Climatology and Paleoclimatology of the East African Lakes (Eds T.C.

Johnson and E.O. Odada), pp. 225–240. Gordon and Breach, Amsterdam,.

Chamley, H. (1989) Clay Sedimentology. Springer,-Verlag, Berlin, Heidelberg, 623

pp.

Christidis, G.E. and Koutsopoulou, E. (2013) A simple approach to the identification

of trioctahedral smectites by X-ray diffraction. Clay Minerals, 48, 687-696.

Coltorti, M., Bonadiman, C., Hinton, R.W., Siena, F. and Upton, B.G.J. (1999)

Carbonatite metasomatism of the oceanic upper mantle: evidence from clinopyroxenes and

glasses in ultramafic xenoliths of Grande Comore, Indian Ocean. Journal of Petrology, 40,

133-165.

Darragi, F. and Tardy, Y. (1987) Authigenic trioctahedral smectites controlling pH,

alkalinity, silica and magnesium concentrations in alkaline lakes. Chemical Geology, 63, 59–

72.

Debure, M., De Windt, L., Frugier, P. and Gin, S. (2013a) HLW glass dissolution in

the presence of magnesium carbonate: Diffusion cell experiment and coupled modeling of

diffusion and geochemical interactions. Journal of Nuclear Materials, 443, 507-521.

Debure, M., Frugier, P., De Windt, L. and Gin, S. (2013b) Dolomite effect on

borosilicate glass alteration. Applied Geochemistry, 33, 237-251.

Debure, M., De Windt, L., Frugier, P., Gin, S. and Vieillard, P. (2016) Mineralogy

and thermodynamic properties of magnesium phyllosilicates formed during the alteration of a

simplified nuclear glass. Journal of Nuclear Materials, 475, 255-265.

Deocampo, D.M., Cuadros, J., Wing-Dudek, T., Olives, J. and Amouric, M. (2009)

Saline lake diagenesis as revealed by coupled mineralogy and geochemistry of multiple

ultrafine clay phases: Pliocene Olduvai Gorge, Tanzania. American Journal of Science, 309,

834–858.

Galán, E. and Pozo, M. (2011) Palygorskite and sepiolite deposits in continental

environments. Description, genetic patterns and sedimentary settings. In: Developments in

Palygorskite–Sepiolite Research (Eds E. Galán and A. Singer), pp. 125-173. Elsevier,

Amsterdam.

Gautier, Q., Bénézeth, P., Mavromatis, V. and Schott, J. (2014) Hydromagnesite

solubility product and growth kinetics in aqueous solution from 25 to 75°C. Geochimica et

Cosmochimica Acta, 138, 1-20.

Gérard, E., De Goeyse, S., Hugoni, M., Agogue, H., Richard, L., Milesi, V., Guyot,

F., Lecourt, L., Borensztajn, S., Leclerc, T., Joseph, M., Sarazin, G., Jezequel, D.,

Leboulanger, C. and Ader, M. (2018) Key role of Alphaproteobacteria and Cyanobacteria in

the formation of stromatolites of Lake Dziani Dzaha (Mayotte, Western Indian Ocean).

Front. Microbiol. 9, 796. doi: 10.3389/fmicb.2018.00796.

Guggenheim, S., Adams, J.M., Bain, D.C., Bergaya, F., Brigatti, M.F., Drits, V.A.,

Formoso, M.L.L., Galán, E., Kogure, T. and Stanjek, H. (2006). Summary of

recommendations of nomenclature committees relevant to clay mineralogy: Report of the

Association Internationale pour l’Etude des Argiles (AIPEA) Nomenclature Committee for

2006. Clay Minerals, 41, 863-877.

Hadi, J., Tournassat, C., Ignatiadis, I., Greneche, J.M. and Charlet, L. (2013)

Modelling CEC variations versus structural iron reduction levels in dioctahedral smectites.

Existing approaches, new data and model refinements. Journal of Colloid and Interface

Science, 407, 397-409.

Hay, R.L., Guldman, S.G., Matthews, J.C., Lander, R.H., Duffin, M.E. and Kyser,

T.K. (1991) Clay mineral diagenesis in core KM-3 of Searles Lake, California. Clays and

Clay Minerals, 39, 84-96.

Hicks, L.J., Bridges, J.C. and Gurman, S.J. (2014) Ferric saponite and serpentine in

the nakhlite martian meteorites. Geochim. Cosmochim. Ac., 136, 194-210.

Hover, V.C., Walter, L.M., Peacor, D.R. and Martini, A.M. (1999) Mg-smectite

authigenesis in a marine evaporative environment, Salina Ometepec, Baja California. Clays

and Clay Minerals, 47, 252–268.

Huang, J., Chu, X., Lyons, T.W., Planavsky, N.J. and Wen, H. (2013) A new look at

saponite formation and its implications for early animal records in the Ediacaran of South

China. Geobiology, 11, 3-14.

Jones, B.F. (1986) Clay mineral diagenesis in lacustrine sediments. In: Studies in

Diagenesis (Ed. F.A. Mumpton), United States Geological Survey Bulletin, 1578, 291–300.

Jones, B.F. and Weir, A.H. (1983) Clay minerals of Lake Abert, an alkaline, saline

lake. Clays and Clay Minerals, 31, 161-172.

Kodama, H., De Kimpe, C.R. and Dejou, J. (1988) Ferrian saponite in a gabbro

saprolite at Mont Mégantic, Quebec. Clays Clay Miner., 36, 102-110.

Lebeau, O., Busigny, V., Chaduteau, C. and Ader, M. (2014) Organic matter removal

for the analysis of carbon and oxygen isotope compositions of siderite. Chemical Geology,

372, 54-61.

Leboulanger, C., Agogué, H., Bernard, C., Bouvy, M., Carré, C., Cellamare, M.,

Duval, C., Fouilland, E,m Got, P., Intertaglia, L., Lavergne, C., Le Floc’h, E., Roques, C. and

Sarazin, G. (2017) Microbial Diversity and Cyanobacterial Production in Dziani Dzaha

Crater Lake, a Unique Tropical Thalassohaline Environment. PLoS ONE 12, e0168879.

doi.org/10.1371/journal.pone.0168879

Lerouge, C., Gaboreau, S., Grangeon, S., Claret, F., Warmont, F., Jenni, A., Cloet, V.

and Mäder, U. (2017) In situ interactions between Opalinus Clay and Low Alkali Concrete,

Physics and Chemistry of the Earth, 99, 3-21.

Lewis Jr., W.M. (2011) Global primary production of lakes: 19th Baldi Memorial

Lecture. Inland Waters, 1, 1-28.

Lòpez-Archilla, A.I., Moreira, D., López-García, P. and Guerrero, C. (2004)

Phytoplankton diversity and cyanobacterial dominance in a hypereutrophic shallow lake with

biologically produced alkaline pH. Extremophiles, 8, 109-115.

Martini, A.M., Walter, L.M., Lyons, T.W., Hover, V.C. and Hansen, J. (2002)

Significance of early diagenetic water-rock interactions in a modern marine

siliciclastic/evaporite environment: Salina Ometepec, Baja California. Geol. Soc. Am. Bull.

114, 1055–1069.

Meunier A. (2005) Clays. Springer-Verlag, Berlin, Heidelberg, 472 pp.

Millot G. (1970) Geology of Clays. Springer-Verlag, Berlin, Heidelberg, 430 pp.

Mougenot, D., Hernandez, J. and Virlogeux, P. (1989) Structure et volcanisme d'un

rift sous-marin; le fosse des Kerimbas (marge nord-mozambique). Bulletin de la Société

Géologique de France, 2, 401-410.

Moore, D.M. and Reynolds, R.C. (1989) X-ray Diffraction and the Identification and

Analysis of Clay Minerals. Oxford University Press, Oxford, 332 pp.

Nougier, J., Cantagrel, J.M. and Karche, J.P. (1986) The Comores archipelago in the

western Indian Ocean: volcanology, geochronology and geodynamic setting. Journal of

African Earth Sciences, 5, 135-145.

Orsini, L. and Remy, J.C. (1976) Utilisation du chlorure de cobaltihexamine pour la

détermination simultanée de la capacité d'échange et des bases échangeables des sols. Sci.

Sol, 4, 269-275.

Ostrom, M.E. (1961) Separation of clay minerals from carbonate rocks by using acid.

Journal of Sedimentary Research, 31, 123-129.

Pace, A., Bourillot, R., Bouton, A., Vennin, E., Galaup, S., Bundeleva, I., Patrier, P.,

Dupraz, C., Thomazo, C., Sansjofre, P., Yokoyama, Y., Franceschi, M., Anguy, Y., Pigot, L.,

Virgone, A. and Visscher, P.T. (2016) Microbial and diagenetic steps leading to the

mineralisation of Great Salt Lake microbialites. Scientific Reports, 6:31495. doi:

10.1038/srep31495

Parkhurst, D.L. and Appelo, C.A.J. (2013) Description of input and examples for

PHREEQC version 3: a computer program for speciation, batch-reaction, one-dimensional

transport, and inverse geochemical calculations. U.S. Geological Survey Techniques and

Methods, Book 6, A43, 497 pp. Available at https://pubs.usgs.gov/tm/06/a43

Pelleter, A.A, Caro, M., Cordier, C., Bachélery, P., Nehlig, P., Debeuf, D. and

Arnaud, N. (2014) Melilite-bearing lavas in Mayotte (France): An insight into the mantle

source below the Comores. Lithos, 208-209, 281-297.

Post, J.L. (1984) Saponite from near Ballarat, California. Clays Clay Miner., 32, 147.

Racki, G. and Cordey, F. (2000) Radiolarian palaeoecology and radiolarites: is the

present the key to the past? Earth-Science Reviews, 52, 83-120.

Rehim, H., Mizusaki, A.M.P., Carvalho, M.D. and Monteiro, M. (1986) Talco e

estevensita na formação lagoa feia da bacia de campos--possíveis implicações no ambiente

deposicional. Anais do XXXIV Congresso Brasileiro de Geologia. Goiânia, Goiás, 1986, 1,

416-425.

Sayles, F.L. and Fyfe, W.S. (1973) The crystallization of magnesite from aqueous

solution. Geochimica et Cosmochimica Acta, 37, 87-96.

Setti, M., Marinoni, L. and Lopez-Galindo, A. (2004) Mineralogical and geochemical

characteristics (major, minor, trace elements and REE) of detrital and authigenic clay

minerals in a Cenozoic sequence from Ross Sea, Antarctica. Clay Minerals, 39, 405–421.

Siever, R. (1992) The silica cycle in the Precambrian. Geochimica et Cosmochimica

Acta, 56, 3265-3272.

Simkiss, K. (1964) Variation of the crystallization form of calcium carbonate from

artificial sea water. Nature, 201, 492-493.

Stoessell, R.K. (1988) 25°C and 1 atm dissolution experiments of sepiolite and

kerolite. Geochim. et Cosmochim. Ac. 52, 365-374.

Suquet, H., de la Calle, C. and Pezerat, H. (1975) Swelling and Structural

Organization of Saponite. Clays and Clay Minerals, 23, 1-9.

Taft, W.H. (1967) Physical chemistry of formation of carbonates. In: Carbonate

Rocks – Physical and Chemical Aspects (Eds G.V. Chilingar, H.J. Bissell and R.W.

Fairbridge), pp.151-167. Elsevier, Amsterdam.

Tissot, B.P. and Welte, D.H. (1984) Diagenesis, catagenesis and metagenesis of

organic matter. In: Petroleum Formation and Occurrence, 2nd edn, pp. 69-73. Springer-

Verlag, Berlin, Heidelberg.

Tosca, N.J. (2015) Geochemical pathways to Mg-clay formation. In: Magnesian

Clays: Characterization, Origins and Applications (Eds M. Pozo and E. Galán). AIPEA

Special Publications, 2, 283–329.

Tosca, N.J., Macdonald, F.A., Strauss, J.V., Johnston, D.T. and Knoll, A.H. (2011)

Sedimentary talc in Neoproterozoic carbonate successions. Earth and Planetary Science

Letters, 306, 11-22.

Tosca, N.J. and Wright, V.P. (2015) Diagenetic pathways linked to labile Mg-clays in

lacustrine carbonate reservoirs: a model for the origin of secondary porosity in the Cretaceous

pre-salt Barra Velha Formation, offshore Brazil. In: Reservoir Quality of Clastic and

Carbonate Rocks: Analysis, Modelling and Prediction (Eds P.J. Armitage, A.R. Butcher, J.M.

Churchill, A.E. Csoma, C. Hollis, R.H. Lander, J.E. Omma and R.H. Worden), Geological

Society, London, Special Publications, 435, 33-46.

Tréguer, P., Nelson, D.M., Van Bennekom, A.J., DeMaster, D.J., Leynaert, A. and

Quéguiner, B. (1995) The Silica Balance in the World Ocean: A Reestimate. Science, 268,

375-379.

Van Denburgh, A.S. (1975) Solute Balance at Abert and Summer Lake, South-central

Oregon. Professional Paper 502-C. US Government Printing Office, Washington, 28 pp.

Vicente, M.A., Suárez, M., López-González, J.D.D. and Banares-Munoz, M.A. (1996)

Characterization, surface area, and porosity analyses of the solids obtained by acid leaching

of a saponite. Langmuir, 12, 566-572.

Visser J.W. and de Wolff P.M. (1964) Absolute Intensities. Report 641.109,

Technisch Physische Dienst, Delft, Netherlands.

Walker, G.F. (1958) Reactions of expanding-lattice clay minerals with glycerol and

ethylene glycol. Clay Minerals Bulletin, 3, 302-313.

Weaver, C.E. (1989) Clays, Muds, and Shales. Elsevier Science, Amsterdam, 818 pp.

Weaver, C.E. and Beck, K.C. (1977) Miocene of the S.E. United States: A model for

chemical sedimentation in a peri-marine environment. Sedimentary Geology, 17, 1-234.

White, R. and McKenzie, D. (1989) Magmatism at rift zones: the generation of

volcanic continental margins and flood basalts. Journal of Geophysical Research: Solid

Earth, 94, 7685-7729.

Wilson, M. (1992) Magmatism and continental rifting during the opening of the South

Atlantic Ocean: a consequence of Lower Cretaceous super-plume activity? In: Magmatism

and the Causes of Continental Break-up (Eds B.C. Storey, T. Alabaster, R.J. Pankhurst),

Geological Society, London, Special Publications, 68, 241-255.

Wright, V.P. (2012) Lacustrine carbonates in rift settings: the interaction of volcanic

and microbial processes on carbonate deposition. In: Advances in Carbonate Exploration and

Reservoir Analysis (Eds J. Garland, J.E. Neilson, S.E. Laubach and K.J. Whidden),

Geological Society, London, Special Publications, 370, 39-47.

Wright, V.P. and Barnett, A.J. (2015) An abiotic model for the development of

textures in some South Atlantic early Cretaceous lacustrine carbonates. In: Microbial

Carbonates in Space and Time: Implications for Global Exploration and Production (Eds

D.W.J. Bosence, K.A. Gibbons, D.P. Le Heron, W.A. Morgan, T. Pritchard and B.A.

Vining). Geological Society, London, Special Publications, 418, 209–219.

Yuretich, R.F. and Cerling, T.E. (1983) Hydrogeochemistry of Lake Turkana, Kenya:

Mass balance and mineral reactions in an alkaline lake. Geochimica and Cosmochimica Acta,

47, 1099–1109.

Zeyen, N., Benzerara, K., Li, J., Groleau, A., Balan, E., Estève, I., Tavera, R.,

Moreira, D. and López-García, P. (2015) Formation of low-T hydrated silicates in modern

microbialites from Mexico and implications for microbial fossilization. Frontiers in Earth

Science, 3, 64. doi: 10.3389/feart.2015.00064

Zeyen, N., Daval, D., Lopez-Garcia, P., Moreira, D., Gaillardet, J. and Benzerara, K.

(2017) Geochemical Conditions Allowing the Formation of Modern Lacustrine Microbialites.

Procedia Earth and Planetary Science, 17, 380-383.

Zinke, J., Reijmer, J.J.G. and Thomassin, B.A. (2003) Systems tracts sedimentology

in the lagoon of Mayotte associated with the Holocene transgression. Sedimentary Geology,

160, 57-784.

FIGURE CAPTIONS

Fig. 1. Location map of the study site (A) and (B) and bathymetric map of the Dziani Dzaha

collection locations (red triangles).

Fig. 2. (A) North–south topographic profile of the lake, showing an ideal representation of

lake sediments. The location of sediment cores is projected onto the topographic profile

which crosses the 18 m deep depression. The insert (white box) shows the topographic profile

of the entire volcanic crater. Vertical exaggeration is 4x. (B) Picture of C12 at 15 to 38 cm

depth showing varves composed of dark and light brown layers. No disturbance of the

sediment pile is observed. (C) Picture of C13 at 44 to 61 cm depth showing deformation

structures. (D) Water/rock weight ratio as function of depth in C12.

Fig. 3. X-ray diffraction pattern of C9 at 54 to 58 cm depth (a.u. – arbitrary unit). Only the

most representative diffraction peaks of aragonite (dark blue), hydromagnesite (light blue),

clinopyroxene (orange) and clay minerals (red) are shown. Values of d-spacing (Ångström)

are inserted for clay minerals.

Fig. 4. Secondary electron (SE) image (A) and chemical composition map of Mg, Si and Ca

(B) and Al (C) showing hydromagnesite (blue), aragonite (yellow) and clay minerals (pink-

blue) of C1 at 22 to 23 cm depth. Transmission electron microscopy (TEM) image (D) and

energy dispersive X-ray fluorescence (EDX) spectrum (E) of clay minerals of C4 at 24 to 27

cm depth. SE image (F), EDX spectrum (G) and quantitative chemical composition (H) of

volcanic glass of C9 at 10 to 14 cm depth. The red crosses indicate the position of the EDX

analyses. Correlation between the Na and Cl signals in the EDX spectrum (E) indicates NaCl

contamination.

Fig. 5. X-ray diffraction (XRD) patterns of the <2 µm clay fraction extracted from C4 at 27 to

30 cm depth acquired on (A) an un-oriented slide; (B) oriented slide in the air-dried state

(black line), ethylene glycol state (blue line), after heating at 500°C and ethylene glycol

salvation (red line); and (C) after Mg saturation (green line) coupled to glycerol solvation

during three days (purple line).

Fig. 6. Mineral and organic matter content (wt.%) as function of depth in C11, C4, C10, C6

and C13. The location of sediment cores is shown on the topographic profile of the lake (Fig.

2A).

Fig. 7. Mineral stability domains in the Ca-Mg-C-O-H and Mg-Si-O-H systems at 30°C and 1

bar. The chemical composition of Dziani Dzaha waters (Leboulanger et al., 2017; Gérard et

al., 2018) is represented with a black box. (A) Ca-Mg-C-O-H system. Magnesite is not taken

into account in the calculation. Aragonite is considered rather than calcite, consistent with the

observations. The grey dotted lines indicate metastable equilibria between dolomite, huntite,

aragonite, hydromagnesite and brucite. The full black lines indicate metastable equilibria

between brucite, aragonite and hydromagnesite when dolomite and huntite are not considered

in the calculation. (B) Mg-Si-O-H system. The full black lines represent metastable equilibria

between Mg-aluminosilicates, when chlorites are not considered in the calculation. The

dotted lines are fluid compositions at equilibrium with talc, sepiolite and kerolite, i.e. three

Al-free Mg-silicates. The thermodynamic data of kerolite are from Stoessell (1988). The

dashed-dotted line corresponds to the critical supersaturation for homogenous nucleation of

Mg-silicates suggested by Tosca (2015). Silica activities at equilibrium with quartz and

amorphous silica are shown (vertical dotted lines).

Fig. 8. Model of the early diagenetic formation of the saponite-like mineral. Dashed arrows

show the possible feedback effect between the destabilization of hydromagnesite and the

precipitation of the saponite-like mineral.

Table 1. Method for quantification of mineral phases.

Table 2. Refined composition of saponite-like mineral in C4 (in weight %). EMPA – electron

microprobe analysis; CEC – cation exchange capacity

Fig. S1. Volcanic catchment: (A) ash layers; (B) fragments of solid/igneous rocks in ash

layers.

Mineral

Chemical formula

Calculation*

Halite

NaCl

Halite = [Cl]

Alkaline

feldspars

Na0.7K0.3AlSi3O8

Alk feld = [K]/0.3

Saponite

K0.01Mg0.12Ca0.02Na0.28(Mg2.22Fe0.18Al0.28)

(Si3.61Al0.39)O10(OH)2

Sap = ([Al] – Alk feld)/0.67

Clinopyroxene

(CayMg1-y)SiO3

Px = [Si] – 3*Alk feld – 3.61*Sap

Hydromagnesite

Mg5(CO3)4(OH)2·(H2O)4

Hydromag = ([Mg] – 2.34*Sap –

(1-y)*Px)/5

Aragonite

CaCO3

Ara = ([Ca] – y*Px – 0.02*Sap)

Pyrite

FeS2

Pyr = [S]/2

Magnetite

Fe3O4

Mag = ([Fe] – Pyr – 0.18*Sap)/3

*[X] are the concentrations in mol.% of Cl, K, Al, Si, Mg, Ca, S and Fe measured with XRF.

Sample

EPMA results

CEC results

3.65

0.20

2.01

0.07

11.62

0.07

0.70

20.78

0.65

0.02

0.04

0.15

Diagenetic formation of stevensite by replacement of diatom frustules in the sediments of the alkaline Lake Alchichica (Mexico)

Article

Dec 2022

Better understanding the conditions of formation of authigenic Mg‐silicates and their reactivity is key to interpret the palaeoenvironmental message carried by the sedimentary record and evaluate the effect of reverse weathering, a process involved in long‐term climate evolution. Microbialites from most alkaline crater lakes in Mexico contain Mg‐silicates except those in Lake Alchichica, where concentration of orthosilicic acid is low (<26 μM). This study investigated the first metre of sediments in Lake Alchichica in order to check how their mineralogy compared with that of shoreline microbialites. The mineralogy and chemistry of the sediment column were determined, together with the pore water chemistry, providing insights on the processes occurring during early diagenesis. Below ca 3 cm in depth, diatom frustules are progressively pseudomorphized into Al‐poor Mg‐silicates with a composition corresponding to stevensite. This diagenetic process is massive and the resulting silicate represents between 30 and 53 wt.% of the sediment content at all depths. This observation questions the possibility to infer lake palaeochemistry from the presence/absence of Mg‐silicates in the sedimentary record. Moreover, it allowed refinement of the conditions under which Mg‐silicates authigenesis occurs: the saturation of the solution should be higher or equal to the solubility of a Mg‐silicate phase close to that of ‘amorphous sepiolite’. Although the solubility of authigenic silicates is a key parameter of reverse weathering modelling during geological times, it is still debated. In this study, a solubility constant deduced from a natural system is proposed that should be considered when modelling the formation of Mg‐silicates in a natural environment. The proportion of reverse weathering associated with this solubility constant could be higher than previously predicted based on experiments and thus have a greater impact on climate stability over geological timescales.

Early stages of Type I-S kerogen formation revealed by Rock-Eval® 7S analysis of sediment from a modern halo-alkaline lake (Dziani Dzaha, Mayotte)

Article

May 2024
ORG GEOCHEM

Authigenic clay mineral constraints on spatiotemporal evolution of restricted, evaporitic conditions during deposition of the Ediacaran Doushantuo Formation

Article

Full-text available

Jan 2024
EARTH PLANET SC LETT

The carbonate-rich shelf facies of the Ediacaran Doushantuo Formation (South China) have produced a broad array of microfossil and isotope proxy records which underpin much of our understanding of environmental change and biospheric evolution during this key time period. Recent reports of locally abundant authigenic, Mg-rich saponite clay in the Yangtze Gorges Area hint at potentially widespread evaporitic conditions in a lagoonal setting. Despite the implications for interpretation of proxy records and the environmental setting of early metazoans, the spatiotemporal extent of restricted, evaporitic conditions across the Yangtze Block remains largely unexplored. Here we use mineralogical and petrographic techniques to document the spatial and strat-igraphic distribution of authigenic clay minerals across seven sites, representing a transect from shallow prox-imal shelf to deep basinal settings. Our results demonstrate the widespread and persistent occurrence of authigenic saponite in proximal shelf settings, whereas talc is identified at more distal shelf sites, consistent with Mg-clay authigenesis in an evaporitic lagoon with spatially variable detrital aluminosilicate input. Slope and basinal sites contain no Mg-clays and are instead characterized by abundant authigenic illite, consistent with an open marine setting. Stratigraphically, illite was found in the basal cap carbonate and uppermost Doushantuo Formation, with Mg-clays only present in the interval in-between, suggesting gradual development of an extensive carbonate rimmed lagoon over the Yangtze shelf that restricted seawater exchange, followed by stronger marine influence due to erosion of the carbonate rim and subsequent marine transgression. We conclude that restricted, evaporitic conditions were much more expansive than previously assumed, potentially providing favorable conditions for the Doushantuo Biota. Further, the widespread presence of authigenic clays across the shelf supports the case for enhanced reverse weathering before the advent of biosilicification.

The composition of gas emissions at Petite Terre (Mayotte, Comoros): inference on magmatic fingerprints

Article

Full-text available

Sep 2022

The Comoros archipelago is an active geodynamic region located in the Mozambique Channel between East continental Africa and Madagascar. The archipelago results from intra-plate volcanism, the most recent eruptions having occurred on the youngest island of Grande Comore and on the oldest one of Mayotte. Since 2018, the eastern submarine flank of Mayotte has been the site of one of the largest recent eruptive events on Earth in terms of erupted lava volume. On land, the most recent volcanic activity occurred in Holocene on the eastern side of Mayotte, corresponding to the small Petite Terre Island, where twomain and persistent gas seep areas are present (Airport Beach, namely BAS, and Dziani Dzaha intracrateric lake). The large submarine eruption at the feet of Mayotte (50 km offshore; 3.5 km b.s.l.) is associated with deep (mantle level) seismic activity closer to the coast (5-15 km offshore) possibly corresponding to a single and large magmatic plumbing system. Our study aims at characterizing the chemical and isotopic composition of gas seeps on land and assesses their potential link with the magmatic plumbing system feeding the Mayotte volcanic ridge and the recent submarine activity. Data from bubbling gases collected between 2018 and 2021 are discussed and compared with older datasets acquired between 2005 and 2016 from different research teams. The relation between ³He/⁴He and δ¹³C-CO2shows a clear magmatic origin for Mayotte bubbling gases, while the variable proportions and isotopic signature of CH4is related to the occurrence of both biogenic and abiogenic sources of methane. Our new dataset points to a time-decreasing influence of the recent seismo-volcanic activity at Mayotte on the composition of hydrothermal fluids on land, whose equilibriumtemperature steadily decreases since 2018. The increased knowledge on the gas-geochemistry at Mayotte makes the results of this work of potential support for volcanic and environmental monitoring programs.

Weathering and cementation of historic kimberlite residues from South Africa: Implications for residue stabilization and CO2 sequestration

Article

Oct 2024

Constraints on dolomite formation in a Late Palaeozoic saline alkaline lake deposit, Junggar Basin, north-west China

Article

Full-text available

Jul 2023

Alkaline lakes (pH > 9) are among the few modern sedimentary environments that are hydrochemically favourable for low-temperature dolomite formation. While Mg-clays and Mg-evaporites also form more easily in alkaline environments, few studies have focused on how the kinetically inhibited dolomite wins the competition for Mg 2+. Here, a basin-wide survey of the distribution, paragenesis and stable C, O and Mg isotopes of main Mg-rich minerals in the Late Palaeozoic saline alkaline lake deposit of the northwest Junggar Basin, northwest China, is conducted to study the influence of the formation and diagenesis of eitelite, northupite and Mg-clays on dolomite formation. Large, isolated dolomite crystals (20 to 70 μm in diameter), show positive δ 13 C values (ranging from +1 to +7‰) and a restricted distribution in the mudstones of the lake-transitional zone. These crystals have been interpreted as organogenic dolomite driven by methanogenesis via fermentation of organic substrates. The δ 18 O values of dolomitic mudstones (from À7.4 to +3.4‰), calcitic mudstones (from À15.1 to À3.3‰) and bedded Na-carbonate evaporites (from +0.08 to +3.7‰), together with their Mg isotopic compositions, suggest that dolomite was not enriched in the most concentrated environments or during stages with most Mg sources, but in the organic-rich deposits containing few other authigenic Mg-rich minerals. Dolomite is at a competitive disadvantage for Mg 2+ ions compared to Mg-evaporite and Mg-clay minerals due to its slow crystallization rates and the deficiency of micritic calcium carbonate precursors. However, it can nucleate and progressively grow into large crystals (>20 μm) if bacterial methanogen-esis could effectively lower porewater pH (<8.5) and induce the dissolution of generated eitelite, northupite or Mg-clays. These findings suggest that high salinity and/or high alkalinity are not always favourable conditions for dolo-mite formation and highlight the active role of pH fluctuations in inducing low-temperature dolomite formation.

Fluorine speciation and origin of Early-Middle Triassic bentonite deposits in Sichuan Basin, South China

Article

Full-text available

Jan 2023

Fluorine (F) is one of the most important environmentally harmful elements released by volcanic activity, and the bentonite deposits that formed from volcanic ashes are potentially harmful to the environment. However, the mechanisms governing F-rich bentonite formation and its F speciation composition remain enigmatic. The F-rich bentonite deposits are widely distributed in the Early-Middle Strata of the Sichuan Basin, South China. Detailed mineralogical and geochemical studies were conducted on the bentonite deposits from five sections of the Sichuan Basin. X-ray diffraction (XRD) analyses indicate that the F-rich bentonites mainly contain quartz, carbonates (calcite and dolomite) or gypsum, and clay minerals, while the clay minerals are dominated by illite and illite/smectite (I/S). Clay mineralogical studies suggest that bentonites were transformed from volcanic ashes during diagenesis by smectite illitization. The major and trace element distribution in F-rich bentonite deposits altered from volcanic ashes is most likely derived from felsic magmas, and alteration of the parent rocks (e.g., rhyolites) to bentonite is associated with leaching and subsequent removal of F. The total fluorine content ( F TOT ) of the bentonite samples ranged from 1,162 mg/kg to 2,604 mg/kg (average = 1773 mg/kg), well above the average F TOT contents of soils in the world. The results of the sequential extraction experiments show that the highest content is residual-fluorine ( F res ) , followed by carbonate-fluorine ( F car ) with a mean value of 1,556 mg/kg and 186 mg/kg, indicating carbonate is an important F sink in bentonite deposits. The average fluorine value of organic fluorine ( F o r ) , Fe/Mn oxide-fluorine ( F f m ) and exchangeable fluorine ( F e x ) are relatively low with an average value of 17.5 mg/kg, 6.8 mg/kg and 4.1 mg/kg, respectively. However, water-soluble fluorine ( F w s ) has a mean value of 4.0 mg/kg, which is higher than the corresponding average value in soils in an area susceptible to endemic fluorosis in China. Based on the characteristic of fluorine speciation, the fluorine in bentonite deposits may pose a risk to the environment. This study makes an important contribution to a better understanding of the characteristic of fluorine speciation in bentonites and the formation mechanism that governs fluorine enrichment in bentonites.

MODELO CRONOESTRATIGRÁFICO DO CAMPO DE ATAPU: UMA NOVA ABORDAGEM DE INTERPRETAÇÃO SÍSMICA E RESTAURAÇÃO GEOLÓGICA 3D A PARTIR DO CÁLCULO DO TEMPO GEOLÓGICO RELATIVO (RGT)

Thesis

Full-text available

Aug 2023

In Brazil, pre-salt reservoirs gained considerable attention after their discovery in the second half of the 2000s and have shown to be very promising up to the present day in terms of production. Inserted in the main pre-salt exploration area, the Outer High of the Santos Basin, the Atapu Field is currently the sixth largest oil producing field in the pre-salt in the country and represents the area studied in this work. The complexities associated with these reservoirs, such as the great depths, the thick and deformed layer of salt and the heterogeneities inherent to the permoporous quality of carbonate rocks, make the study and characterization of this type of reservoir difficult, since these factors affect the resolution of the seismic data. One of these difficulties is the recognition of intra-Alagoas unconformity (IAU). Characterized by a reflector of low lateral continuity that is often eroded by the salt base, manual interpretation may encounter difficulties. These challenges, therefore, faced new technologies that help the steps involved in the exploration cycle. In this sense, this work adopted a methodology to semi-automatically recognize IAU, through the generation of a tectonostratigraphic model and the Relative Geological Time (RGT) calculation. The first stage of the flowchart consists of building the tectonostratigraphic model of the Atapu Field and calculating the first information about the RGT. Then, the recognition stage of the stratigraphic surfaces internal to the Barra Velha Formation began, the patches, inserting the seeds that, later, propagated the data amplitude information by the seismic volume. At this stage, three intra-formational horizons were recognized, including the one corresponding to the IAU, presenting the characteristics mentioned. However, the result of this automatic recognition was not successful in half of the studied area. So, it was necessary to manually insert the seeds and generate new patches in these regions where it failed. At the end, the horizon still showed areas without data, but which, through seismic sections, it was possible to observe that these were regions where the IAU was eroded. Once the IAU surface was obtained, the calculated RGT was updated considering the unconformity. This allowed the refinement of the relative timelines and the generation of the Wheeler diagram, through the paleospace calculatation. In this way, the methodology allowed the understanding of the temporal relations of the depositional systems and their relationship with the erosive surfaces, besides the understanding of how the pre-salt formations developed through time and the role that the intra-Alagoas unconformity had on the deposition of the Barra Velha Fm.

Lacustrine dolomite in deep time: What really matters in early dolomite formation and accumulation?

Article

Sep 2023
EARTH-SCI REV

Impact of the seismo-volcanic crisis offshore Mayotte on the Dziani Dzaha Lake

Article

Full-text available

Dec 2022

Since May 2018, an unexpected long and intense seismic crisis started offshore Mayotte (Indian Ocean, France). This ongoing seismic sequence is associated with the birth of a newly discovered submarine volcano 50 km east of Petite Terre. Here, we investigate the indirect impact of this crisis on the stability of an atypical ecosystem located in Mayotte, the Dziani Dzaha Lake. This lacustrine system presented physical, chemical and biogeochemical characteristics that were distinct from those classically observed in modern lakes or seawater, e.g. high salinity (up to 70 psu), lack of nitrate, sulfate content below 3 mM, and permanent water column anoxia below ca. 1.5 m depth (2012–2017 period). Based on three surveys conducted in 2020 and 2021, we documented an episode of water column oxygenation, a significant pH decrease and an impressive change in the carbon isotope signatures. These preliminary data suggest that the functioning of biogeochemical cycles in the Dziani Dzaha Lake is impacted by increased CO2 inputs and the changes in the lake mixing dynamics, which is an indirect consequence of the ongoing seismo-volcanic crisis.

Structure et volcanisme d'un rift sous-marin; le fosse des Kerimbas (marge nord-mozambique)

Article

Full-text available

Mar 1989

On the continental margin of the northern Mozambique, a flat-bottomed trough, the north-south trending Kerimbas deep graben, disrupts the continental slope alongside an alignment of seamounts forming the Davie Ridge. Water-gun seismic reflection lines, implemented throughout the region during MD40/MACAMO cruise, show that the graben, cut into Tertiary sediments, corresponds : eastward, to a former discontinuity several times reactivated and located above an abrupt uplift of the Moho; westward, to a Miocene tilting or a recent normal fault. On both sides and at its southern end, the graben supports alkaline basalt intrusive into Neogene sediments. Miocene deformation, shown by flexures and onlaps above reflector A, (half-graben stage), can be explained by the tilting of two domains located on both sides of the western scarp of the Davie Ridge. But the upturning of Miocene strata by faulting denotes a later deformation originated by an oblique opening of this trough (graben stage). The Kerimbas graben, seaward extension of the East African Rift, is currently propagating at both ends of the trough, as indicated by recent faults and rather intense sismicity.

Key Role of Alphaproteobacteria and Cyanobacteria in the Formation of Stromatolites of Lake Dziani Dzaha (Mayotte, Western Indian Ocean)

Article

Full-text available

May 2018

Lake Dziani Dzaha is a thalassohaline tropical crater lake located on the “Petite Terre” Island of Mayotte (Comoros archipelago, Western Indian Ocean). Stromatolites are actively growing in the shallow waters of the lake shores. These stromatolites are mainly composed of aragonite with lesser proportions of hydromagnesite, calcite, dolomite, and phyllosilicates. They are morphologically and texturally diverse ranging from tabular covered by a cauliflower-like crust to columnar ones with a smooth surface. High-throughput sequencing of bacterial and archaeal 16S rRNA genes combined with confocal laser scanning microscopy (CLSM) analysis revealed that the microbial composition of the mats associated with the stromatolites was clearly distinct from that of the Arthrospira-dominated lake water. Unicellular-colonial Cyanobacteria belonging to the Xenococcus genus of the Pleurocapsales order were detected in the cauliflower crust mats, whereas filamentous Cyanobacteria belonging to the Leptolyngbya genus were found in the smooth surface mats. Observations using CLSM, scanning electron microscopy (SEM) and Raman spectroscopy indicated that the cauliflower texture consists of laminations of aragonite, magnesium-silicate phase and hydromagnesite. The associated microbial mat, as confirmed by laser microdissection and whole-genome amplification (WGA), is composed of Pleurocapsales coated by abundant filamentous and coccoid Alphaproteobacteria. These phototrophic Alphaproteobacteria promote the precipitation of aragonite in which they become incrusted. In contrast, the Pleurocapsales are not calcifying but instead accumulate silicon and magnesium in their sheaths, which may be responsible for the formation of the Mg-silicate phase found in the cauliflower crust. We therefore propose that Pleurocapsales and Alphaproteobacteria are involved in the formation of two distinct mineral phases present in the cauliflower texture: Mg-silicate and aragonite, respectively. These results point out the role of phototrophic Alphaproteobacteria in the formation of stromatolites, which may open new perspective for the analysis of the fossil record.

Geochemical Conditions Allowing the Formation of Modern Lacustrine Microbialites

Article

Full-text available

Dec 2017

Interpreting the environmental conditions of ancient microbialites rely on comparisons with modern analogues. Yet, we lack a detailed reference framework relating the chemical and mineralogical composition of modern lacustrine microbialites with the physical and chemical parameters prevailing in the lakes where they form. Here we performed geochemical analyses of water solutions and mineralogical analyses of microbialites in 12 Mexican crater lakes. We found a large diversity of microbialites in terms of mineralogical composition, with occurrence of diverse carbonate phases such as magnesian calcite, monohydrocalcite, aragonite, hydromagnesite, and dolomite as well as authigenic magnesium silicate phases. In parallel, the chemical compositions of the lakes differed particularly by their alkalinity, their concentration of ortho-silicic acid (H4SiO4) and their Mg/Ca ratio. From this study, we infer a minimum alkalinity value for the formation of lacustrine microbialites, as well as several constraints given by the presence of mineral phases on the chemical composition of the lakes in which microbialites formed. Finally, we observe a general correlation between the alkalinity and the sodium content of the lakes. This may relate to variations in evaporation intensity and provide a historical model for lacustrine microbialite formation: microbialite start forming only when the lake is sufficiently old/evaporated.

In situ interactions between Opalinus Clay and Low Alkali Concrete

Article

Full-text available

Jan 2017
PHYS CHEM EARTH

A five-year-old interface between a Low Alkali Concrete (LAC) formulation (CEM III/B containing 66% slag and 10% nano-silica) and Opalinus Clay (OPA) from a field experiment at Mont Terri Underground Rock Laboratory in Switzerland (Jenni et al., 2014) has been studied to decipher the textural, mineralogical and chemical changes that occurred between the two reacting materials. Reactivity between LAC concrete and OPA is found to be limited to a ∼1 mm thick highly porous (ca. 75% porosity) white crust developed on the concrete side. Quantitative mineralogical mapping of the white crust using an electron microprobe and infrared spectroscopy on the cement matrix provides evidence of a Mg-rich phase accounting for approximatively 25 wt % of the matrix associated with 11 wt % of calcite, calcium silicate hydrate (C-S-H) and other cement phases. EDX analyses and electron diffraction combined with transmission electron microscopy of the Mg-rich phase provide evidence for a tri-octahedral 2:1 phyllosilicate with mean composition: (Ca0.5±0.2) (Mg2.0±0.4, Fe0.2±0.1, Al0.5±03, □0.3±0.3) (Al0.9±0.2, Si3.1±0.2) O10 (OH)2, where □ represents vacancies in the octahedral site. Apart from this reactive contact, textural, mineralogical and chemical modifications at the contact with the LAC concrete are limited. OPA mineralogy remains largely unmodified. X-ray micro-fluorescence and EPMA mapping of major elements on the OPA side also provides evidence for a Mg-enriched 300–400 μm thick layer. The cation exchange capacity (CEC) values measured in the OPA in contact with the LAC concrete range between 153 and 175 meq kg⁻¹ of dry OPA, close to the reference value of 170 ± 10 meq kg⁻¹ of dry OPA (Pearson et al., 2003). Changing cation occupancies at the interface with LAC concrete are mainly marked by increased Ca, Mg and K, and decreased Na. Leaching tests performed on OPA with deionized water and at different solid to water ratios strongly suggest that Cl and SO4 have either conservative behaviour or are constrained by the solubility of a precipitated sulfate phase. The Cl and SO4 concentrations measured at 2 cm from the interface are close to concentrations of undisturbed OPA pore waters (SO4: 4.5 ± 1.5 mmol kg⁻¹ of dry OPA; Cl: 7.5 ± 2.1 mmol kg⁻¹of dry OPA), and increase towards the interface with the concrete. The SO4 to Cl ratio also increases towards the interface, suggesting that the increasing anion concentrations are not related to porosity variations but rather to a concentration gradient and sulfate phase precipitation near the interface.

Microbial Diversity and Cyanobacterial Production in Dziani Dzaha Crater Lake, a Unique Tropical Thalassohaline Environment

Article

Full-text available

Jan 2017
PLOS ONE

This study describes, for the first time, the water chemistry and microbial diversity in Dziani Dzaha, a tropical crater lake located on Mayotte Island (Comoros archipelago, Western Indian Ocean). The lake water had a high level of dissolved matter and high alkalinity (10.6–14.5 g L⁻¹ eq. CO3²⁻, i.e. 160–220 mM compare to around 2–2.5 in seawater), with salinity up to 52 psu, 1.5 higher than seawater. Hierarchical clustering discriminated Dziani Dzaha water from other alkaline, saline lakes, highlighting its thalassohaline nature. The phytoplankton biomass was very high, with a total chlorophyll a concentration of 524 to 875 μg chl a L⁻¹ depending on the survey, homogeneously distributed from surface to bottom (4 m). Throughout the whole water column the photosynthetic biomass was dominated (>97% of total biovolume) by the filamentous cyanobacteria Arthrospira sp. with a straight morphotype. In situ daily photosynthetic oxygen production ranged from 17.3 to 22.2 g O2 m⁻² d⁻¹, consistent with experimental production / irradiance measurements and modeling. Heterotrophic bacterioplankton was extremely abundant, with cell densities up to 1.5 10⁸ cells mL⁻¹ in the whole water column. Isolation and culture of 59 Eubacteria strains revealed the prevalence of alkaliphilic and halophilic organisms together with taxa unknown to date, based on 16S rRNA gene analysis. A single cloning-sequencing approach using archaeal 16S rDNA gene primers unveiled the presence of diverse extremophilic Euryarchaeota. The water chemistry of Dziani Dzaha Lake supports the hypothesis that it was derived from seawater and strongly modified by geological conditions and microbial activities that increased the alkalinity. Dziani Dzaha has a unique consortium of cyanobacteria, phytoplankton, heterotrophic Eubacteria and Archaea, with very few unicellular protozoa, that will deserve further deep analysis to unravel its uncommon diversity. A single taxon, belonging to the genus Arthrospira, was found responsible for almost all photosynthetic primary production.

Key role of Alphaproteobacteria and Cyanobacteria in the formation of stromatolites of Lake Dziani Dzaha (Mayotte, Western Indian Ocean).

Article

May 2018

Lake Dziani Dzaha is a thalassohaline tropical crater lake located on the "Petite Terre" Island of Mayotte (Comoros archipelago, Western Indian Ocean). Stromatolites are actively growing in the shallow waters of the lake shores. These stromatolites are mainly composed of aragonite with lesser proportions of hydromagnesite, calcite, dolomite, and phyllosilicates. They are morphologically and texturally diverse ranging from tabular covered by a cauliflower-like crust to columnar ones with a smooth surface. High-throughput sequencing of bacterial and archaeal 16S rRNA genes combined with confocal laser scanning microscopy (CLSM) analysis revealed that the microbial composition of the mats associated with the stromatolites was clearly distinct from that of the Arthrospira-dominated lake water. Unicellular-colonial Cyanobacteria belonging to the Xenococcus genus of the Pleurocapsales order were detected in the cauliflower crust mats, whereas filamentous Cyanobacteria belonging to the Leptolyngbya genus were found in the smooth surface mats. Observations using CLSM, scanning electron microscopy (SEM) and Raman spectroscopy indicated that the cauliflower texture consists of laminations of aragonite, magnesium-silicate phase and hydromagnesite. The associated microbial mat, as confirmed by laser microdissection and whole-genome amplification (WGA), is composed of Pleurocapsales coated by abundant filamentous and coccoid Alphaproteobacteria. These phototrophic Alphaproteobacteria promote the precipitation of aragonite in which they become incrusted. In contrast, the Pleurocapsales are not calcifying but instead accumulate silicon and magnesium in their sheaths, which may be responsible for the formation of the Mg-silicate phase found in the cauliflower crust. We therefore propose that Pleurocapsales and Frontiers in Microbiology | www.frontiersin.org 1 May 2018 | Volume 9 | Article 796 Gérard et al. Alphaproteobacteria and Cyanobacteria in Stromatolites Alphaproteobacteria are involved in the formation of two distinct mineral phases present in the cauliflower texture: Mg-silicate and aragonite, respectively. These results point out the role of phototrophic Alphaproteobacteria in the formation of stromatolites, which may open new perspective for the analysis of the fossil record.

Geology of Clays

Article