Origins of elements building travertine and tufa: New perspectives provided by isotopic and geochemical tracers

Sedimentary Geology 334 (2016) 97–114

Contents lists available at ScienceDirect

Sedimentary Geology journal homepage: www.elsevier.com/locate/sedgeo

Origins of elements building travertine and tufa: New perspectives provided by isotopic and geochemical tracers P.-A. Teboul a,b,⁎, C. Durlet b, E.C. Gaucher c, A. Virgone c, J.-P. Girard c, J. Curie e, B. Lopez a,c,d, G.F. Camoin a a

Aix-Marseille Université, CNRS, IRD, CEREGE UM34, 13545 Aix en Provence, France Univ. Bourgogne Franche-Comté, CNRS, Biogéosciences UMR6282, F-21000 Dijon, France TOTAL CSTJF, Avenue Larribau, F-64018 Pau Cedex, France d Department of Earth and Environmental Sciences, KU Leuven, Celestijnenlaan 200E, B-3001 Leuven, Belgium e UMR 7041 ArScAn, Université Paris 1 Panthéon-Sorbonne, France b c

a r t i c l e

i n f o

Article history: Received 27 October 2015 Received in revised form 6 January 2016 Accepted 10 January 2016 Available online 2 February 2016 Editor: B. Jones Keywords: Travertine Calcareous tufa Stable isotopes Major elements Trace elements

a b s t r a c t Fluid/rock interaction represents a major process in the formation of calcitic or aragonitic travertine and tufa (CATT). In most cases, CATT is associated to limestone dissolution somewhere along the hydrogeological pathway. However, a wide array of other substratum (basalts, rhyolites, carbonatites, ultramafics, granites, dolomites, evaporites) can act as potential source of elements involved in the formation of CATT. This study reports on the evaluation of potential geochemical tracers linking CATT to their substratum, and unravelling the origin of elements. A large database was established from available literature data as well as new data acquired in the frame of this study for a set of Modern to Recent CATT (Ligurian ophiolites, Italy; the Chaine des Puys, Limagne graben and Paris Basin, France; Reunion Island, Indian Ocean; Jebel Oust, Tunisia). Four most reliable tracing methods are identified (1) δ13C and δ18O cross-plot allows distinguishing epigean (minδ13C = − 27.2 ‰, maxδ13C = 0.9 ‰, meanδ13C = − 12.3 ‰ for N = 314) from hypogean systems (minδ13C = − 4 ‰, maxδ13C = 11.7 ‰, meanδ13C = −2.87 ‰ for N = 198). Very low δ13C values (b−12 ‰) and δ18O N−4 ‰ associated to negative δ13C values are specifically indicative of an ultramafic source rock. (2) Barium and strontium cross-plot helps to discriminate different groups of source rocks amongst the hypogean CATT: (i) source rocks composed of mixed limestones, evaporites, and dolomites are characterised by low barium (b 100 ppm) and high strontium (N400 ppm) contents, (ii) mafic and granitic source rocks are undifferentiated and display similar barium (from 15 to 930 ppm) and high strontium (N 200 ppm) contents, (iii) the carbonatite group is characterised by its exceptional high barium and strontium values. In epigean CATT, a pure limestone source rock usually relates to very low barium and strontium contents (b 50 ppm and b 70 ppm respectively), whereas mixed limestone, evaporite and dolomite source rocks generally display low strontium content (b580 ppm) with higher barium content (N50 ppm). (3) Relatively high beryllium content (N 30 ppm) in CATT seems to indicate a pure granitoid source. (4) High chromium concentrations (N20 ppm) are systematically documented in Modern CATT located on an ultramafic substratum. The definition of diagnostic compositional fields for actively forming or recently formed CATT is influenced by many factors including water composition, water temperature, dissolved gas composition and concentration, biological activity, position in the sedimentary body and early diagenesis, in addition to substratum lithology. However, the results of this study illustrate that, despite these many factors, the combined use of Ba, Sr, Be, Cr, δ13C, and δ18O may be valuable to discriminate the rock lithology prevailing in the hydrogeological or palaeo-hydrogeological reservoir of CATT. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Travertine and tufa are continental carbonate deposits frequently associated with limestone dissolution in superficial (epigean) or deep (hypogean) hydrogeological reservoirs (Fouke et al., 2000; Garnett et al., 2004; Ihlenfeld et al., 2003; Gandin & Capezzuoli, 2008; Kele ⁎ Corresponding author at: Aix-Marseille Université, CNRS, IRD, CEREGE UM34, 13545 Aix en Provence, France. E-mail address: [email protected] (P.-A. Teboul).

http://dx.doi.org/10.1016/j.sedgeo.2016.01.004 0037-0738/© 2016 Elsevier B.V. All rights reserved.

et al., 2011; Koltai et al., 2012; Makhnach et al., 2004). However, a large variety of other rocks can occur as substratum and source of elements building these carbonates. In Modern and Ancient environments, many studies have shown that igneous rocks (basalts, rhyolites, carbonatites, ultramafics, syenites, granites) and other sedimentary rocks (dolomites, evaporites, marls) may constitute derivative sources for calcium and other elements required for travertine and tufa buildup. Such derivative contributions are generally easy to identify in Modern systems where source rocks are usually known and where fluids (gas and water) can be analysed (Fouke et al., 2000; Crossey et al.,

98

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

2009). On the other hand, it is more difficult to identify and demonstrate the contribution of a non-calcareous supplier in fossil travertine and tufa systems, especially when the lithology of the surrounding or underlying reservoir is poorly known. In such fossil cases, parent fluids of travertine and tufa cannot be analysed. Only primary fluid inclusions preserved in primary calcitic precipitates may provide information on the nature of original fluids (El Desouky et al., 2015). In the last decade, one of the most debated cases has been the deeply buried Lower Cretaceous travertine-like deposits occurring in the South Atlantic, off-shore Brazil, Angola and Congo (Della Porta, 2015; Wright & Barnett, 2015; Harris et al., 2013; Wright, 2012). These deposits belong to the so-called Presalt deposits and are composed of non-marine carbonates accumulated in a rift to post-rift context, on top of rocks of variable nature including ultramafics, basalts, volcaniclastic deposits, lacustrine limestones, claystones and marls (Jones & Xiao, 2013). In this complex buried fossil system, wells rarely reach the underlying substratum of the travertine deposits and numerous hypotheses may be made regarding the lithological origin of the elements used for travertine building. A number of multidisciplinary studies have been performed since the 70s to characterise processes leading to travertine and tufa deposition (Pola et al., 2014). These studies show that tufa and travertine macro-/microfacies are not always characteristic of the lithologies prevailing in their hydrogeological basin. In fossil situations, macro/ microfacies features, mineralogical evidences, plant remains, fossils and sedimentary architectures are sometimes available to constrain factors such as cold or hot waters (Pentecost, 2005), presence of gas bubbles (Chafetz et al., 1991) and hyperalkaline waters (Barnes et al., 1982). But these factors depend much more on the nature of hydrogeological processes, i.e. epigean versus hypogean (Lopez et al., in press), than the lithologies involved in the system. Sedimentological and structural approaches are usually used to unravel the processes at stake. Potentially, geochemical tracers could help to. However, studies on major, trace and rare-earth element concentrations in travertine and tufa deposits are largely lacking. The aim of our study was therefore to establish a geochemical database by compiling major, minor, trace elements and stable isotope concentrations for Modern/Recent travertines and tufas. Our investigation was limited to very young (b50.000 years B.P.) active and inactive deposits in order to avoid late diagenetic over-print. Travertine and tufa have been considered according to the nomenclature of Capezzuoli et al. (2014). According to these authors travertine deposits are related to abiotic processes, and characterised by high depositional rates (cm to m/yrs), calcitic to aragonitic mineralogy and elevated δ13C signature (N−1 ‰), whereas tufa deposits are linked to biotic processes, and typified by low depositional rates (mm to cm/yrs), mainly calcitic mineralogy and low δ13C signature (b 0 ‰). Capezzuoli et al. (2014) emphasise the uncertainties regarding the interpretation of coeval deposits in cooled thermal waters, where tufa-like deposits are precipitating. Only Modern to Recent calcitic and aragonitic travertine and tufa, here referred as CATT, containing more than 90% CaCO3 were taken into account. Travertine-like siliceous deposits of sinter settings were excluded. The reported database includes data from published literature and new data acquired as part of this research work. The later includes analysed Modern and Recent non-marine CATT derived from the alteration of rocks with various lithologies (ultramafic, mafic, granite, evaporite, limestone) located in the Ligurian ophiolites (Italy), the Chaîne des Puys (France), the Limagne graben (France), the Paris Basin (France), Réunion Island (Indian Ocean) and the Jebel Oust (Tunisia). Each of these occurrences has been classified according to geological setting and investigated for specific geochemical tracers that may help identify the substratum rocks forming the hydrogeological basins of the studied CATT systems. The term “Modern CATT” will be used below as a simplification for currently active or inactive young (Recent) CATT.

2. Geological settings 2.1. Voltri Massif of the Ligurian ophiolites (Italy) The Voltri Massif in the Ligurian Alps is located in the southernmost part of the Western Alps and to the north of the Apennines system. According to Brouwer et al. (2002), the Voltri Massif represents the relict of a subducted and exhumed lithosphere associated to the PiedmontLigurian Ocean, as part of the Mesozoic Tethyan realm. This ophiolitic body is described as the largest of the Alp-Apennine system and is composed of three main units (Chavagnac et al., 2013a,b; Borghini et al., 2007; Hermann et al., 2000; Hoogerduijn Strating, 1991; Fig. 1(A)): (1) a metasediments and metavolcanics unit (Voltri-Rossiglione Unit); (2) a serpentinite unit with highly altered peridotite (Beigua Unit); and (3) a serpentinised lherzolite and harzburgite unit (Erro-Tobbio Unit). Several hyperalkaline springs with travertine deposits have been reported in the Voltri Massif (Chavagnac et al., 2013a,b; Schwarzenbach et al., 2013; Cipolli et al., 2004). The hyperalkaline waters originate from percolation of meteoric water through fractures and water/rock interaction with ultramafic rocks. The travertine precipitation process is described as being related to atmospheric CO2 uptake and neutralisation of the hyperalkaline waters (Chavagnac et al., 2013a,b). The spring/travertine systems studied here are mainly located in the Beigua Unit, but near the contact with the Voltri-Rossiglione Unit (Fig. 1(A) and Table 1). One exception is the Gorzente spring/travertine system, which is located directly on lherzolithes of the Erro-Tobbio unit. These springs are always associated to fractures in the bedrock. Field measurements show high pH values (up to 11.2 at Acquasanta and Rio Branaga; Table 1) and low water temperatures ranging from 13 to 23.7 °C. Several authors highlighted high calcium, sodium, and abiotic methane contents associated to low silica and transition metal contents in these waters (Chavagnac et al., 2013b; Boulart et al., 2013; Cipolli et al., 2004; Schwarzenbach, 2013). Also, alkalinity measurements show low values (0.91 to 2.99 meq/l). The carbonate precipitate consists mainly of thin calcium carbonate crusts with micro-terracettes that form relatively small travertine bodies on the ultramafic bedrocks (Fig. 2(A), (B)). 2.2. Réunion Island (Indian Ocean) The Réunion Island is located in the western Indian Ocean, 700 km east of Madagascar. The island constitutes the subaerial part of an oceanic shield volcano, and is composed of two volcanoes: the dormant Piton des Neiges and the active Piton de la Fournaise. The central part of the Piton des Neiges exhibits three major depressions (Fig. 1(B)) referred to as the cirques of Cilaos, Mafate and Salazie and originating from collapses, landslides and intense erosion (Oehler et al., 2008). Frissant et al. (2003) highlighted the presence of different volcanic units (Fig. 1(B)). The most common volcanic rocks are composed of olivine basalts showing zeolitisation features (Demange et al., 1989). The edge of the three cirques also shows basalts, hawaiites and mugearites, whilst the bottom part locally exhibits trachytic deposits, gabbros, and syenitic intrusive bodies (Frissant et al., 2003). Thermal springs in the Réunion Island mainly occur in the Salazie and Cilaos cirques, within a few kilometres around the highest point of the island. Some of them are associated with CATT, reported in several studies (Frissant et al., 2003, 2004; Marty et al., 1993; Rançon et al., 1988) but never studied regarding their petrography or geochemistry. In contrast, hydrochemistry of the thermal waters and associated gas has been well documented in studies dealing with volcanology, hydrochemistry, and geothermal energy (Frissant et al., 2003; Moulin et al., 2002; Join & Coudray, 1997; Kluska, 1997; Marty et al., 1993; Demange et al., 1989). CATT analysed in the present study come from three thermal springs located in the Cilaos cirque and two in the Salazie cirque (Table 1, Fig. 1(B)). These thermal springs are often located in association with

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

99

Fig. 1. Study areas and sampled sites: (A) simplified geological map of the Voltri Massif in Ligurian ophiolites (modified after Vignaroli et al., 2009 and Chavagnac et al., 2013b); (B) simplified geological map of the Cirques of Salazie, Mafate and Cilaos, Réunion Island, Indian Ocean; (C) simplified geological map of the Chaîne des Puys and Limagne graben, Metropolitan France; (D) simplified geological map of the Burgundy High, Metropolitan France (E) simplified geological map of the Jebel Oust regional setting, Tunisia (modified after Curie, 2013).

100

Sorgente Sulfurea

Samples (N)

δ O

Crystalline crust, lithoclast–breccia

Cal

1

–6.7

–12

Dam

Crystalline crust

Cal, Arg, Q, Ant

2

–18.4 & –6.3

Dam

Crystalline crust

Cal

3

Dam

Crystalline crust

Cal

CATT lithotypes

Dam, cemented clast

18

Water analysis (from vent)

13

δ C Ca

Mg

Si

Al

Fe

Sr

Ba

Be

Cr

T (°C)

pH

Cond. (mS/cm)

Alk (meq/l)

370,449

7901

3552

847

1088

439

633

<1

30

16.5

8.6

256

2.99

–24 & –11

272,906 70080 27906

1588

4353

373

435

<1

80

22.0

11.0

452

1.8

–19.2 to –11.8

–26.5 to –17.9

386,146

1086

12044

512

1632

293

95.3

<1

30

20.1

10.5

465

1.7

1

–16.4

–21.4

289,199 31060

8227

3864 10882

598

461

<1

300

23.7

10.9

359

1.17

Dam

Crystalline crust

Cal

1

–16.6

–23.5

366,090

5006

11312

1640 11659

446

209

<1

170

13.2

11.3

419

1.53

Dam, spring mound

Crystalline crust

Cal, Mg–cal

1

–13.5

–18.6

348,653

2654

1870

1429

3731

240

93

<1

70

16

9.8

175

0.91

Dam, cemented clast Dam, cemented clast

Crystalline crust, lithoclast–breccia Crystalline crust, lithoclast–breccia

Cal, Mg–cal

1

–20.4

–23.9

357,014

4704

5376

2858

3809

521

231

<1

80

18.1

11.2

525

2.01

Cal

1

–18.4

–25.9

382,525

1146

4487

582

777

401

220

<1

<20

15

11.2

515

2.88

Stream crust

Crystalline crust, shrub

Arg, Cal, Q

2

–9.1 & –8.3

4.4 & 5.2

370,984

1809

4604

106

9600

3708

23

<1

<20

n.d.

n.d.

n.d.

n.d.

Stream crust, cascade

Crystalline crust, shrub, reed

Arg, Cal, Q

2

–9.3 & –9.6

4.5 & 4.6

367,519

844

12784

212

8900

3748

25

<1

<20

31.1

6.6

1980

34.31

Exhaust pipe

Crystalline crust

Arg, Cal, Q

1

–9.4

4.2

368,305

1146

3693

106

10960 3647

15

<1

<20

n.d.

n.d.

n.d.

n.d.

Cascade

Reed

Cal

1

–7.2

–4

354,299

9107

25942

3758

6063

548

18

<1

50

21.6

8.6

319

3.65

Cemented clast, spring mound

Crystalline crust, lithoclast–breccia

Cal, Arg, Q

3

–7.3 to –5.7

4.9 to 8

357,824 10454

7791

2682

6426

1331

85

<1

65

n.d.

n.d.

n.d.

n.d.

Manouilh red

Spring mound

Reed

Cal, Mn–cal, Arg, Q

1

–9.2

3.1

340,650

3739

25238

4393

5830

4914 370

<1

40

30.5

6.8

1417

13.47

Manouilh yellow

Cemented clast

Reed

Cal, Mg–cal, Arg, Q

2

–8.6 & –5.8

4 & 6.3

330,788

6393

14397

2329 35173 1733 209

<1

25

27.7

6.8

1436

13.8

Manouilh spring

Cascade

Reed

Cal, Arg, Q

1

–6.1

4.5

306,278

6393

31318
<1

70

21.9

8.9

451

3.96

Ponte Arma Acquasanta Rio Branega

Microsyenites

Old hotel Old thermal baths

Bachelier spring

Olivine basalts

Pissa spring

Hypogean

Thermal pipes Réunion Island

Other lithologies

Bedrock lithology Serpentinized lherzolites

Maddalena

Metasediments/metavolcanics

Rio Leone

Epigean

Gorzente 2

CATT mineralogy

CATT morphology

1235

38

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

Voltri massif of the Ligurian ophiolites

Gorzente 1

Mean elemental composition (ppm)

Serpentinized metagabbros

Localities

Hydrologic regime

Table 1 Summary of CATT morphologies, macrofacies, mineralogical associations, bulk rock geochemistry and main hydrochemical parameters of the associated waters of the Ligurian ophiolites, Réunion Island, Chaîne des Puys, Limagne, Burgundy High and Jebel Oust springs. CATT morphologies classification is from Pentecost and Viles (1994) and CATT lithotypes classification from Guo and Riding (1998). Mineralogy: Ant: antigorite; Arg: aragonite; Cal: calcite; MHC: monohydrocalcite; Q: quartz. δ18O and δ13C are expressed in ‰PDB.

Tête de lion

Gra– nites Limestone Limestone

Jebel Oust

Epigean

Mâlain

Hypogean

Burgundy

Sainte Marguerite spring

Lantenay

Anthropic channel

Samples (N)

Other lithologies Marls & mafics

Saurier spring

Anthropic channel

δ18O

Crystalline crust, shrub, reed

Cal, Arg

2

–10.2 & –10

Crystalline crust

Cal

1

Cal, Q

Crystalline crust, shrub, pisoid, paper thin raft Crystalline crust, shrub, paper thin raft

Water analysis (from vent)

δ13C Ca

Mg

Si

Al

Fe

Sr

Ba

Be

Cr

T (°C)

pH

Cond. (mS/cm)

Alk (meq/l)

4.2 & 4.3

378,917

3076

22,460

53

6724

2115

98

57

<20

30.1

6.2

4900

n.d.

–10.5

2.1

223,313

9016

373

900

14,116 3330

364

596

<20

35

6.3

n.d.

n.d.

4

–9.1 to –7.6

4.4 to 5

368,787

8745

2466

238

1011

7042

386

41

<20

23.2

6.6

9670

71.6

Cal, Q

2

–10.6 & –10.2

3 & 3.8

373,164

9016

748


1088

4625

335

64

<20

27.6

n.d.

9180

n.d.

Spring mound

Crystalline crust, shrub, paper thin raft, reed

Cal, Q, MHC

2

–7.7 & –5.4

4.7 & 6.2

356,800

6152

3225

1852

6685

1825

255

18

<20

13.8

6.5

6370

n.d.

Anthropic channel

Crystalline crust, shrub, pisoid, paper thin raft

Cal, Arg

2

–9.2 & –8.2

3.8 & 4.6

376,701

6694

351

53

505

2091

68

19

<20

35.5

6.38

4480

42.04

Cal

2

–8.3 & –7.7

4.7 & 5.3

375,701

9559

10844

238

1360

1833

107

13

<20

n.d.

n.d.

n.d.

n.d.

Spring mound Spring mound

Basanites, hawaites, tephrites

Limestones

Hypogean

Limagne

Ceix spring

Granites

Bard spring

La Salet spring

Exhaust pipe

Meta– morphics

Fontaine Pétrifiante

Gimeaux spring

Spring mound

CATT mineralogy

Crystalline crust, shrub, pisoid, paper thin raft, reed Crystalline crust, shrub, pisoid, paper thin raft, reed

Cal

1

–8.2

5.7

342,079 13148

5516

2752

4470

1265

24

38

<20

13.2

6.4

6420

79.2

Cal, Arg, Q

1

–7

6.9

350,797 11157

5376

794

1866

5069

456

6

<20

13.0

n.d.

5840

n.d.

Crystalline crust, shrub, pisoid, paper thin raft, reed Crystalline crust, shrub, pisoid, paper thin raft, reed

Cal, Arg, Q

2

–7.2 & –6.9

7.8 & 8.3

344,223

9831

14911

159

4780

4564

521

21

25

14.4

7.4

4930

n.d.

Cal, Arg

1

–10.1

5.6

368,591

3649

5072

132

3070

6271

233

36

<20

27.1

6.5

8130

76.21

Cascade

Reed

Cal, Q

1

–6.6

–10.4

376,308

603

1145

3281

2176

44

23

<1

<20

10.2

7.5

578

n.d.

Cascade

Reed

Cal, Q

1

–6.4

–9.6

324,643

1930

7853

1138

6529

64

44

<1

30

9.6

8.12

644

n.d.

Mound, anthropic channel, cascade

Crystalline crust, shrub, paper thin raft

Cal, Arg, Q

28

–0.5 to 3.4

–11.7 to –7.2

>?

1334

n.d.

838

2090

4484

25

50

6.2

20940

n.d.

Spring mound Cascade Spring mound

Crystalline crust, reed

n.d. 4.27

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

Grotte du Cornador

CATT lithotypes

Volcano –clastics Granite

Lefort thermes

Hypogean

Chaîne des Puys

Pré dimanche spring

Mean elemental composition (ppm) CATT morphology

Mafics

Bedrock lithology

Hydrologic regime

Localities

101

102

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

syenitic intrusive bodies that induced deep faulting in the bedrock (Fig. 1(B)). These CATT mainly consist of stream crusts, spring mounds, and cascade deposits. Near spring emergence, the CATT is composed of dense to porous laminated red-to-white or yellowish travertines (Fig. 2(C), (D)), whilst in the distal parts porous tufas with bryophytes-rich phytofacies predominate in relation with topography and water flow. Water temperatures are between 21 and 31 °C, and pH between 6.6 and 8.6. The alkalinity is highly variable with a minimum of 3.65 meq/l and a maximum of 34.3 meq/l (Table 1).

or on Quaternary colluviums and alluvions. The architecture of CATT is highly variable (Table 1 and Fig. 2(E), (F)), with spring mounds, cascade tufas, and thick crusts in artificial channels. CATT lithotypes are also extremely variable (see Table 1) including crystalline crusts, microterracettes, paper-thin rafts, coated bubbles, porous phytofacies and spherulitic accumulations. Field measurements show that waters have near neutral pH values, alkalinity between 40 and 80 meq/l and temperatures ranging from 23 to 35 °C in the Chaîne des Puys and 13 to 35.5 °C in the Limagne graben (Table 1).

2.3. Chaîne des Puys and Limagne graben (France)

2.4. Burgundy High (France)

Located in the northern part of the Massif Central in Metropolitan France, the Chaîne des Puys and the Limagne graben are linked to a complex tectonic history including Hercynian orogeny, postHercynian structures (Mesozoic sedimentary covers) and Tertiary basins (Maury & Varet, 1980; Ziegler, 1990; Wattinne, 2004). Most of the Massif Central is however composed of granitic basement and metamorphic rocks of Hercynian age (Fig. 1(C); Autran & Peterlongo, 1980). According to several authors (Merle et al., 1998; Dèzes et al., 2004; Wattinne et al., 2010), the Limagne graben belongs to the European Cenozoic Rift System. Wattinne, (2004) highlights four Oligocene and Miocene lithologies in the Limagne graben: (1) fluvial deposits (conglomerates and sandstones), (2) lacustrine carbonates (dolomites and limestones with stromatolites), (3) lacustrine eutrophic marls and black shales, and (4) carbonate to siliciclastic palustrine deposits. The Chaîne des Puys is a granitic horst partially covered by relatively young volcanic products. The Palaeozoic granitoids ranges from leucogranite to grano-diorite of mostly intrusive allochthonous origin (Feuga, 1987; Casanova et al., 1999). Most volcanoes, basaltic lava flows, and pyroclastic deposits are younger than 200,000 years BP, the last eruption being dated at 7600 years BP (Boivin et al., 2009). The volcanoes are aligned over a deep N–S trending crustal fracture parallel to the aborted Oligocene Limagne graben (Ziegler, 1994; Merle & Michon, 2001). According to Hamelin et al. (2009), the Chaîne des Puys lavas are represented by typical alkali intraplate volcanic suite (alkali basalts/basanites to hawaiites, mugearites and silica-oversaturated benmoreites and trachytes). The studied hydrothermal springs from the Massif Central are located on the western edge of the Limagne graben and on the eastern edge of the Chaîne des Puys where their occurrence is closely related to major faults (Fig. 1(C)). These springs are systematically associated with discrete to intense CO2 degasing (Risler, 1974; Serra et al., 2003) and can emerge from either pure granitic bedrocks, Oligocene limestones and marls, or Pleistocene to Holocene fluvial deposits. Many of the thermal springs located on pure volcanic or granitic bedrocks are not associated with CATT but only with poorly developed iron-rich hard-to-soft deposits. Those associated with CATT are usually located near Oligocene limestones and marls. Two exceptions are the Pré Dimanche and Grotte du Cornador thermal springs, both located on granites and far away from the Oligocene marls and limestones, but very close to volcaniclastic and basaltic rocks. Published work on Modern thermal systems from the Massif Central essentially deal with fluid (water and gas) characterisation (Battani et al., 2010; Baubron et al., 1978; Baubron et al., 1992; Bertin & Rouzaire, 2004; Charguéron et al., 2003; Gal et al., 2012; Négrel, 2004; Pauwels et al., 1997; Serra et al., 2003). Few data are available on the petrography and geochemistry of associated CATT deposits. One notable exception is the study by Casanova et al. (1999) describing bioprecipitation processes in proximal iron-rich CATT of the Sainte Marguerite spring in the Limagne graben. CATT analysed in this study come from seven thermal springs from the Limagne graben and four thermal springs of the Chaîne des Puys (Fig. 1(C) and Table 1). Those from the Limagne mostly form directly on the lacustrine Oligocene marls and limestones meanwhile those from the Chaîne des Puys precipitate directly on Palaeozoic granites,

Located in eastern France, the Burgundy High is composed of faulted limestone plateaus at the southeastern periphery of the Paris Basin. Unfolded Middle Jurassic limestone strata host epigean karstic systems (Delance, 1988) with many emergences and active calcitic tufa bodies (Caudwell, 1983; Caudwell et al., 2001). In the present study, we analysed tufa precipitates associated to perched karstic springs (Mâlain and Lantenay) located 13 and 16 km westwards of Dijon (Fig. 1(D)). We selected these springs because they drain only pure micritic and oolitic Bathonian limestones. These limestones exhibit a very low siliciclastic content and low concentrations in magnesium, iron, strontium, barium, and other trace elements and REE (Purser, 1975; Javaux, 1992, Vincent et al., 1997). In situ water measurements show pH values ranging from 7.5 to 8.2, conductivity near 600 μS/cm, and temperatures averaging 10 °C (Table 1). These CATT occur as bryophyte rich tufas that form fans near the emergences (Fig. 2(G)). Such deposits are very common in karstic areas of Western Europe, and are documented in numerous petrographic and geochemical studies (see synthesis in Andrews, 2006), except for trace element and REE contents (see discussion). 2.5. Jebel Oust (Tunisia) Jebel Oust is located in the northwestern part of Tunisia, 30 km south-west of Tunis (Fig. 1(E)). The site is well known for hot springs used as thermal baths during Classical Antiquity. It is located on a NE– SW dissymmetric horst-anticlinal structure formed during the Eocene Atlasic phase and dominates the alluvial plain of the river (wadi) Miliane. According to Jauzein (1967), the edge of the anticline is mainly composed of Cretaceous marls and limestones, whilst its core is composed of faulted Liassic calcareous units overlying Triassic clays, dolomites, and evaporites. Curie (2013) linked the occurrence of springs on the Jebel Oust to a complex history of faulting and fracturing affecting Triassic and Jurassic deposits. The spring waters are brines, which salinity relates to the presence of Triassic gypsum and salt-rich layers at the base of the anticline (Curie, 2013). Field water measurements show average pH values of 6.22, conductivity of 21 mS/cm, and temperatures around 50 °C (Table 1). These CATT occur mainly in the Roman structures (Fig. 2(H)), starting near the hot spring vent down to around 100 m downwards, showing a great diversity in geometries, morphologies and facies, as those described in Modern travertines from Central Italy (Guo and Riding, 1998, 1999). 3. Material and methods 3.1. Literature survey and database Over 400 published studies have been reviewed in order to compile information on elemental composition and isotopic signature about Modern CATT. There is abundant data available on carbon- and oxygen-isotopes but data on trace elements and REE are lacking. In addition, crucial information about source rock lithology, hydrologic regime, and regional geology is also lacking. A total of 26 studies covering 29 localities have been selected for our database. They are

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

103

Fig. 2. Field photographs of some studied active CATT: (A) hyperalkaline spring diluted by stream water in Gorzente 1, Ligurian ophiolites, Italy; (B) close-up view of calcitic and aragonitic microterracettes in Rio Leone, Ligurian ophiolites, Italy; (C) distal cascade deposits in the Old Thermal Baths spring, Réunion Island, overseas departments of France; (D) small travertine fan cementing a fractured basaltic substrate, Manouilh yellow spring, Réunion Island, overseas departments of France; (E) proximal deposits of a small travertine fan, Pré-Dimanche spring, Chaîne des Puys, Metropolitan France; (F) distal slope of a seasonally active travertine to the right and distal fossil travertine lobe to the left, La Salet spring, Limagne graben, Metropolitan France; (G) small tufa mound in a karstic area, Mâlain spring, Burgundy High, Metropolitan France; (H) small travertine mound in an Ancient Roman Bath, Jebel Oust, Tunisia.

considered as best representing the diversity of geological contexts and lithologies that may promote CATT precipitation in non-marine settings. For each locality, the database (Table 2) includes the hydrologic regime, lithologies of the source rocks, and hydrogeological reservoir, dating, analytical methods, CATT morphology, microfacies, and mineralogies.

3.2. Field analyses and sampling In order to overcome the lack of data in the literature, a field campaign was conducted in 2014 in the Ligurian ophiolites, Réunion Island, the Chaîne des Puys, the Limagne graben and Burgundy, and in 2010 in

104

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

Table 2 Compilation of studies analysing CATT mineralisation, CATT morphologies classification from Pentecost and Viles (1994) and CATT lithotypes classification from Guo and Riding (1998). Mineralogy: Ant: antigorite; Arg: aragonite; Bar: barite; Cal: calcite; Cel: celadonite; Chl: chlorite; Dol: dolomite; Gyp: gypsum; Hl: halite; K-fld: Potassic feldspars; Liz: lizardite; Mc: Mica; MHC: monohydrocalcite; Phyl: phyllosilicates; Plg: plagioclase; Q: quartz; Str: strontianite; Tr: trona; Yuk: yukonite. Authors

Localities

Hydrologic Bedrock regime lithology

Age (yr B.P.)

Benson et al., 1996

Lahontan basin (USA)

Lacustrine

Reed, microbialite-like

n.d.

δ18O, δ13C

Mass spectrometer

Bonny & Jones, 2008

Canada's Northwest Territories

Epigean

10,960 to Lacustrine 2100 crust and reef Modern Spring mound

Crystalline crust

Cal, Bar, opal

Ba, Sr, Mg, Ca

Electron microprobe

Chafetz & Lawrence, 1994

Bridgeport (USA)

Hypogean

Modern

Cal.

δ18O, δ13C

XRD mass spectrometer

La Zitelle (Italy)

Hypogean

Crystalline crust, shrub Crystalline crust, shrub, reed Crystalline crust, shrub

Limestone, volcanics, granites Limestone, dolomite, cherts Volcanics

Main CATT Main CATT morphology lithotypes

Clark & Fontes, 1990

Semail ophiolites (Oman)

Epigean

D'alessandro et al., 2007

Mt. Etna (Italy)

Hypogean

Volcanics, limestone

Fissure ridge Modern Spring mound 9900 to Dam, 830 cemented clast 24,000 to Spring 5000 mound

Dekov et al., 2014

Abhé Lake Lacustrine (Djibouti/Ethiopia)

Volcanics

820

Fouke et al., 2000

Angel terrace (Yellowstone, USA)

Hypogean

Limestone

Modern

Garnett et al., 2004

Wateringbury (UK)

Epigean

Limestone

Horvatinčić et al., 2003 Ilhenfeld et al., 2003

Dinaric Karst (Slovenia/Croatia) Gregory river (Australia)

Epigean

Kato & Kraml, 2010

Koltai et al., 2012 Leybourne et al., 2009

Rwenzori mountains (Uganda) Jandarma spring (Turkey) Karahayit, kirmizi su (Turkey) Pamukkale, Beltes-2 (Turkey) Mecsek Mts (Hungary) Manitoba tufa (Canada)

Li et al., 2008

Salton basin (USA)

Lacustrine

Lojen et al., 2004 Makhnach et al., 2004 Matsuoka et al., 2001 Mervine et al., 2014

River Krka (Croatia) Ptich (Central Belarus) Shirokawa (SW Japan) Semail ophiolites (Oman)

Epigean

Limestone

Epigean

Limestone

Epigean

Limestone

Epigean

Ultramafic

Pentecost, 1999

Matlock Bath tufa (UK)

Epigean

Limestone, dolomite

Modern

Spring mound

Pisarskii et al., 1998

Songwe river (Tanzania)

Hypogean

Carbonatite

Modern

Spring mound

Pola et al., 2014

Veneto (Italy)

Hypogean

Limestone, dolomite

34,000 to Spring Recent mound

Kele et al., 2011

Volcanics, limestones Ultramafics, limestone

CATT Available CATT mineralogy chemistry

Analytical methods

Arg, Cal

Folk, 1994 18

13

14

Cal, Arg

δ O, δ C,

n.d.

Cal, Dol, Arg

δ18O, δ13C, Mg, Ca, Mn, Fe, Sr, Ba

Lacustrine crust and reef

Crystalline crust, shrub, microbialite-like

Cal

Cascade, dam

10,330 to Paludal tufa 6190

Crystalline crust, Arg, cal shrub, reed, pisoid, paper thin raft, coated bubble Crystalline crust, Cal shrub, reed

δ18O, δ13C, major and trace elements, Sr-Nd-Pb–U-Th-Ca isotopes δ18O, δ13C, Ca, Mg, Mn, Sr, Fe, Mn, Na, Ti, S, Ba, V, P, 87Sr/86Sr, SO4, δ 34S δ18O, δ13C, 14C, Ca, Mg, Sr

Limestone

Holocene n.d.

n.d.

Cal

δ18O, δ13C

Epigean

Limestone, dolomite

30 to 19

Fluvial tufa

Crystalline crust, shrub

Cal

δ18O, δ13C, Mg, Sr, Ba, U, Th

Hypogean

Carbonatite, Modern gneiss, granites Marble, Modern limestone

Spring mound

n.d.

n.d.

δ18O, δ13C

Cascade, dam

Crystalline crust, Cal, arg, Q, shrub, reed, Mc, Chl, pisoid, paper thin K-fld, Plg raft, coated bubble

XRD mass δ18O, δ13C, major and trace elements spectrometer, ICP-MS

Limestone, dolomite Limestone, dolomite, anhydrite Carbonates, granites

Fluvial tufa

Reed

δ18O, δ13C

Hypogean Hypogean Hypogean Epigean Epigean

Modern Modern

Spring mound

n.d.

C

20,500 to Lacustrine 1310 crust and reef Modern Fluvial tufa

Reed, microbialite-like

Cal

Microbialite-like

Cal

δ18O, δ13C

12,800 to ? Modern

n.d.

Cal

δ18O, δ13C, 14C

Crystalline crust, shrubs, reed Crystalline crust, shrub, lithoclast-breccia, reed Reed

Cal

δ18O, δ13C

Cal, Arg, Dol, clays

δ18O, δ13C, 14C

Cal

δ18O, δ13C, Ca, Mg, Sr, Na, K, Pb, Zn

Cal, Arg, Dol, Str, Tr, Hl Cal

Major, minor elements

n.d.

Crystalline crust, shrub

XRD, IC/ICP-MS, mass spectrometer XRD, ICP-OES/−MS, mass spectrometer, TIMS XRD, mass Friedman, spectrometer, 1970 TIMS, electron probe microanalysis Mass spectrometer, ICP-AES, EGTA titration Mass spectrometer XRD, ICP-MS, mass spectrometer Mass spectrometer

13

Cal

Spring mound 45,748 to Dam, 350 cemented clast

XRD, mass spectrometer

Mass spectrometer ICP-ES/-MS, δ O, δ C, major and trace elements mass spectrometer XRD, mass δ18O, δ13C, spectrometer 87Sr/86Sr, 14C 18

n.d.

Fluvial tufa

Other references

XRD, mass spectrometer XRD, mass spectrometer Mass spectrometer XRD, mass spectrometer

XRD, atomic absorption, mass spectrometer n.d.

δ18O, δ13C, U-series Mass spectrometer

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

105

Table 2 (continued) Authors

Localities

Hydrologic Bedrock regime lithology

Age (yr B.P.)

Prado-Pérez et al., 2013

Alicun de las Torres (Spain)

Hypogean

Limestone

21,900 to Spring 540 mound

Crystalline crust, shrub, reed

Schwarzenbach Ligurian ophiolites et al., 2013 (Italy)

Epigean

Ultramafic

Modern

Crystalline crust, lithoclast-breccia

Stoffers & Botz, 1994

Lacustrine

Volcanics

Modern

Epigean

Limestone

Recent

Surmelihindi, 2013

Lake Tanganyika (East-central Africa) Aspendos aqueduc (Turkey)

Main CATT Main CATT morphology lithotypes

Dam, cemented clast Lacustrine crust and reef Anthropic channel

the Jebel Oust (Tunisia). Sets of 18 to 110 CATT samples were collected at each site, in order to integrate facies changes along proximal-distal depositional profiles and to study the impact of source rock lithology. At each locality, CATT samples have been taken along proximal-to-distal transects. Samples from Réunion Island, the Chaîne des Puys, the Limagne graben show similar organisations along depositional transects. They can be divided into five settings and facies according to their location with respect to the vent: (1) spring vent facies: generally dense laminated Fe-oxide-rich travertines with a common occurrence of aragonitic botryoides; (2) apron and channel facies showing dense to porous laminated travertine; (3) terrace facies composed of a lamination or alternations of shrubs, paper thin rafts, and calcified bubbles; (4) gentle slope facies showing an accumulation of reworked CATT clasts, crystalline crusts, shrubs, calcified bubbles, spherulites, and plants; (5) cascade tufas: highly porous phytohermal CATT, generally composed of cemented bryophyte or higher plants. Travertines from the hyperalkaline springs of Ligurian ophiolites are mainly brownish/ yellow and white carbonate crusts, formed in close vicinity to the vents or brownish to white micro-terracettes formed along the flow path. The samples from the Burgundy High are composed of bryophyte-rich highly porous tufas. The Jebel Oust samples display a great variety of facies from the spring vent to the Roman baths located further down: generally dense Fe-oxide travertines in the proximal section (Roman sanctuary), paper thin rafts in the proximal slope (Roman channel) and dense and well-laminated, crystalline crust, and spongy fabric travertines in the distal part (Roman baths). In situ measurements of specific physicochemical parameters (T°C, pH, electric conductivity) in spring waters were completed at a majority of sampling points using HACH-LANGE HQ40D instruments. Alkalinity analyses were performed in the field by titrimetric methods (Dickson, 1981) in the Ligurian ophiolites and in Réunion Island. 3.3. Analytical methods Whole rock bulk mineralogy was determined by powder X-ray diffraction (XRD) on 155 samples at the Biogéosciences lab. (University of Burgundy, Dijon, France) using a Brucker D4 Endeavour (stepping 0.0399° from 2.5° to 28.5° 2θ and using a copper X-rays generated at 40 kV and 25 mA). A total of 52 representative CATT samples were selected out of the 345 samples collected in the field. This selection covers the diversity of geological sites (bedrock lithology, hydrologic regimes, proximal versus distal precipitates). All measurements have been performed on whole rocks. CATT samples that were contaminated by detrital particles or non-carbonate precipitates were excluded based on XRD results and petrographic observations under binocular. δ13C and δ18O measurements were done at the Biogéosciences lab. Using a Multiprep carbonate preparation line connected to an Isoprime mass spectrometer, following sample digestion in 100% phosphoric acid at 90 °C. δ13C and δ18O values are reported relative to the Vienna Pee Dee Belemnite standard (V-PDB) by relating δ13C and δ18O values to the NBS19 and NBS18 international

Crystalline crust, shrub, microbialite-like Crystalline crust, shrub

CATT Available CATT mineralogy chemistry Cal, Q, Phyl, Cel, Gyp Cal, Liz

Analytical methods

Other references

δ18O, δ13C, U-series XRD, mass spectrometer, δ18O, δ13C

XRD, mass spectrometer

Cal, Arg, Mg-cal

δ18O, δ13C

XRD, mass spectrometer

n.d.

δ18O, δ13C

Mass spectrometer

standards. Analytical uncertainty is ±0.02 ‰ (1σ) for carbon-isotopes and ±0.05 ‰ (1σ) for oxygen-isotopes. Stable isotope analyses of the Jebel CATT samples were performed at the Leibniz Laboratory (Kiel University, Germany) following a similar procedure and using a Finnigan MAT 251 dual inlet mass spectrometer (Curie, 2013). Major element concentrations of whole rock selected samples were determined at Activation Laboratories, Ltd. (ACTLABS, Ontario, Canada) and at LISA Laboratory (Paris-Diderot University, France) by inductively coupled plasma optical emission spectroscopy (ICP-OES) following fusion of the samples with lithium metaborate/tetraborate and digestion in a nitric acid solution. Trace and REE element abundances were determined in the same laboratories by inductively coupled mass spectrometry (ICP-MS). Information on methods and detection limits are reported in the ACTLABS official website (www.actlabs.com) for the lithium metaborate/tetraborate fusion — ICP method. Precision and accuracy of measurements were controlled by analysis of international standards and replicate analysis. Errors are of ±100% for values relative at the detection limit, ±25% for values relative at 10 times the detection limit, ±10% for values relative at 100 times the detection limit. For the Jebel Oust CATT, the analytical method and the detection limits of the LISA Laboratory are reported and discussed in Monna et al. (2008). 3.4. Statistical analysis Examination of the way lithologies involved in the geological reservoirs affects the CATT elemental concentrations may be investigated by a statistical approach as in Ohta et al. (2005). In order to choose between parametric (i.e. ANOVA) or non-parametric analysis (i.e. Kruskal–Wallis), the assumptions of a normal distribution of the data subsets, of homoscedasticity (i.e. all the variables have the same finite variance), and of similar data subset sample sizes needed to be tested (Ohta et al., 2005). A multiple comparison procedure (i.e. Holm procedure) has also been performed (Holm, 1979). 4. Results The unpublished mineralogical, elemental and isotopic compositions are presented in Table 1. Those extracted from published literature are summarised in Table 2. The detailed general dataset is available as a Supplementary data file #1. 4.1. New δ18O and δ13C data Fig. 3 shows the δ18O and δ13C variations amongst the new results and the literature data. Values from the Ligurian ophiolites show δ18O and δ13C values ranging from −20.4 ‰ to −6.3 ‰ and from −26.5 ‰ to − 11 ‰ (n = 11) respectively, i.e. extremely depleted values (Fig. 3). Highest values (−11.98 ‰ and −11.08 ‰) are measured in 2 samples from Gorzente that formed directly on lherzolites from a mixture of hyperalkaline spring water with stream water. Burgundy High values are very close to these values. δ18O and δ13C from Réunion Island

106

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

range from −9.6 ‰ to −6.1 ‰ and from −4 ‰ to 8 ‰ (n = 13) respectively. Aragonite rich samples from Réunion Island tend to be depleted in 18O compared to calcite-rich ones. One sample from a bryophyterich distal cascade tufa fed by a mixture of thermal waters (Pissa spring) and stream waters is significantly depleted in 13C (−4 ‰). The isotopic ranges of the Chaîne des Puys, Limagne Basin, and Jebel Oust are very similar to those from Réunion Island samples (Fig. 3), although δ13C values tend to be slightly higher (up to + 3.3 ‰) in Limagne graben and slightly lower in Jebel Oust. 4.2. Elemental composition data (statistical analysis) In order to depict the impact of source rock lithology and hydrologic regimes on elemental composition of Modern and Recent CATT, the whole dataset has been divided into 8 CATT subsets for the statistical analysis: (1) CATT from carbonatites, (2) hypogean CATT from granites, (3) hypogean CATT from mafics, (4) lacustrine CATT from multilithological sources, (5) hypogean CATT from limestones and granites, (6 and 7) hypogean and epigean CATT from a mixture of limestones, dolomites, evaporites, siliciclastics or marbles, and (8) epigean CATT from ultramafics. An additional CATT subset, the epigean CATT from pure limestones, has not been included in the statistical approach due to the scarcity of elemental composition data that are available in the literature. Twenty-four elements have been used (Si, Al, Fe, Mn, Mg, Ca, K, Ti, P, Be, V, Ba, Sr, Zr, Cr, Ni, Ge, As, Rb, Cs, La, Ce, Pr, Nd). A Holm multiple comparison procedure has been performed in order to evidence element concentrations that could be significantly representative of the

different data subset listed above. The detailed statistical analysis is presented in the Supplementary data file #2. Table 3 shows the result of the Holm procedure comparison tests at 0.05-confidence level. As suggested by Ohta et al. (2005), the Holm procedure has been used due to the violated homoscedasticity assumption, the large number of combinations and the variable sample sizes (Supplementary data file #2). This method allows the adjustment of Pvalue and limits the family wise error rate (Holm, 1979). This direct pairwise comparison contains 8C2 = 28 possible combinations (Table 3) for the eight levels of geological source rocks. From this result and from Figs. 4, 5, several observations can be made. Barium and strontium are very common discriminant elements in the Holm multiple comparison test, with 15 and 13 occurrences respectively. As a consequence, the Ba/Sr cross plot (Fig. 4) shows strong concentration differences according to the source rock and hydrologic regime. For example, barium contents of the Jebel Oust samples globally tend to be lower than in other localities (b 96 ppm). Strontium and barium of Réunion Island samples display contents ranging from 548 to 4914 ppm and from 15 to 370 ppm respectively. Aragonite-rich samples tend to be enriched in strontium with low barium concentrations whereas calcite-rich samples display high barium content and lower strontium concentrations (Fig. 4). Samples from the Ligurian ophiolites show strontium and barium contents respectively from 240 to 598 ppm and from 93 to 633 ppm. Strontium and barium contents of the Burgundy High samples have not been included in the statistical analysis (only two samples) but they appear low compared to all other data (Fig. 4). Fig. 4 also highlights a major non-linear demarcation between

Fig. 3. Combined δ18O (‰PDB) and δ13C (‰PDB) plot for Recent to Modern calcitic or aragonitic travertine and tufa. Plain markers are from the dataset acquired for this study and dotted fields represent data from the literature. Coloured fields are representatives of CATT implying the same sort of hydrogeological and hydrodynamic contexts. Data source: Epigean CATT from carbonate rocks: a: Makhnach et al. (2004); b: Koltai et al. (2012); c: Garnett et al. (2004); d: Sürmelihindi et al. (2013); e: Ihlenfeld et al. (2003); f: Lojen et al. (2004); g: Horvatinčić et al. (2003); h: Pentecost (1999); i: Matsuoka et al. (2001). Epigean CATT from ultramafics: j: Mervine et al. (2014); k: Clark and Fontes (1990); l: Schwarzenbach et al. (2013). Lacustrine CATT: m: Dekov et al. (2014); n: Stoffers and Botz (1994); o: Benson et al. (1996); p: Li et al. (2008). Hypogean CATT from carbonates and/or igneous rocks (except carbonatites and ultramafics): q: Fouke et al. (2000); r: Chafetz and Lawrence (1994); s: Pola et al. (2014); t: Kele et al. (2011); u: D'Alessandro et al. (2007); v: Prado-Pérez et al. (2013). Hypogean CATT from carbonatites: w: Kato and Kraml (2010).

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

107

Table 3 Results of the Holm multiple comparison procedure at a 0.05 confidence interval. Groups 1 and 2 are relative to the data subsets defined accordingly to hydrogeological settings. In white: Meangroup 1 N Meangroup 2 and in grey: Meangroup 1 b Meangroup 2. Discriminant elements for the purpose of this study are in bold.

Granites

Mafics

Lacustrine CATT from multilithological sources

Fe

Hypogean limestone and granites

Hypogean limestone with dol., evp., slc. or mbl.

Epigean limestone with dol., evp., slc. or mbl.

Group 1 Ultramafics Group 2

Ca

Ca

Mg

Ba, Sr

Ba, Sr, Mn

Sr, Si, Al, Ti

Cr, Al, Ti, La

Carbonatite Ba, Sr, Si, Al, Ti

Ba, Sr, Al, Mg, K

Ba, Si, Al, Mn, Mg

Ba, Sr, Al, Ti

Ba, Sr, Al, Mn, Mg, Ti,

Si, Al, Mg, Ti

Cr ,V, Zr, Ni, Ge, Rb, Cs, La, Ce, Pr, Nd

Al

Al, Ti

Fe, Mg, Ti, As

Be, As

Be, Cr, Ba, Fe, Mn, Mg, As

Cr, V, Zr, Ni, Rb, Cs, La, Ce, Pr, Nd

Be, K

Be, As

Si, Fe, Mn, Mg, As

Granites Be, Sr, Mn, P, As

Cr Mafics

Mg, Ti

Ba, Cr, Fe, Mn, Mg, K

Sr, Fe, Mg, K

Sr,Mn, Mg, K, P

Si, Fe, Mn, Mg, As

Si, As

Si, Mg

Si, Fe, Mg

Cr, V, Zr, Ni, Rb, Cs, La, Ce, Pr

Ba, Cr, V, Zr, Ni, Rb, Cs, La, Ce, Pr, Nd

Sr

Sr, Cr, V, Zr, Ni, Rb, Cs, La, Ce, Pr, Nd

Ti

Cr, Ti

Be, Cr, Ba, Fe, Mn, Mg, Rb

Sr

Sr, Mn, Be, P

Ba, Mn, Mg

Cr, Ba, Fe, Mg, Ni

Sr

Sr, P, As

Sr, Mg

Mn

hypogean and epigean CATT, with high strontium values (N 200 ppm) associated to hypogean systems. Beryllium appears to be a discriminant element for CATT that partially derived from granitic source rocks (Table 3). This is merely suggested by the relatively high Be content (from 31 to 596 ppm) in the Chaîne des Puys CATT (Fig. 5). In addition, Be content of the Limagne Basin shows a narrow range, from 6 to 38 ppm, which tends to be poorer than in the Chaîne des Puys CATT but higher than at all other sites (Fig. 5). Extremely high Chromium content (N 70 ppm) only appears in CATT derived from ultramafic source rocks, as on the Ligurian ophiolites with a maximum content of 300 ppm at Maddalena (Fig. 5). Aluminium frequently appears in the Holm comparison procedure test (Table 3). It could be a clue of CATT contamination by clays or

Lacustrine CATT from multilithological sources

Hypogean limestone and granites

Hypogean limestone with dol., evp., slc. or mbl.

Epigean limestone with dol., evp., slc. or mbl.

other non-carbonate minerals, thus introducing a bias in our results. In order to assess this possibility, Ba, Sr, Be, and Cr contents have been plotted with Al (Supplementary data file #2). No particular positive correlation has been established except for the high chromium value displayed at Maddalena (Ligurian ophiolite). This value must be considered with care. 5. Discussion 5.1. Hypogean versus epigean signatures and source rock lithological classes According to Pentecost (2005), hypogean CATT are easily distinguishable from epigean CATT based on δ13C signature. This statement

108

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

Fig. 4. Combined barium (ppm) and strontium (ppm) logarithmic plot. Plain markers represent the dataset acquired for this study and empty markers represent data from the literature. Ba and Sr detection limits for the new dataset are respectively of 3 and 2 ppm. Error bars are included in marker size.

is confirmed by most data from the current study, e.g. Jebel Oust, Chaîne des Puys, Limagne and Réunion Island. The high δ13C signature of hypogean CATT (N−4 ‰ in this study) reflects non-soil carbon sources. In most cases, the origin of the CO2 is associated with decarbonation of carbonate, water–rock exchange reactions with volcanic and plutonic rocks or to magmatic CO2 from active volcanism. However, one sample from Réunion Island (Ravine Pissa area) is relatively depleted in 13C (−4 ‰) and displays δ18O values of −7.2 (Fig. 3). This low δ13C value can be explained by its distal position, about 200 m downstream of the hydrothermal vent, and its bryophyte-rich fabric. This sample represents a good example of tufa precipitating in a mixture of surface waters and distal hydrothermal waters. δ18O exhibits a wide range of values (Fig. 4). The CATT δ18O signature should reflect water temperature and δ18O of the parent water at the time the CATT formed if isotopic equilibrium is fulfilled (O'Neil et al., 1969 and Friedman & O'Neil, 1977). According to Sunand Liu (2010), the climatic and geographical characteristics are non-negligible factors in the δ18O signatures of CATT as they control the δ18O values of local rainfall. Finally,

several authors have shown that isotopic equilibrium is rarely reached (Friedman, 1970; Fouke et al., 2000; Kele et al., 2011), due to the rapid CO2 degassing of water and elevated precipitation rate of calcite, and preventing complete isotopic equilibration between dissolved CO2− 3 H2O. This is particularly true for hypogean travertines located proximal to hydrothermal vents. Consequently, the CATT δ18O values of such travertine deposits may reflect rather the δ18O signature of dissolved bicarbonates (HCO− 3 ) in the hydrothermal water (Kele et al., 2011). In order to illustrate the possible influence of source rock lithology on the basis of CATTδ18O andδ13C signatures, five source rock lithological classes have been highlighted in Fig. 3: (1) epigean CATT from carbonate source rocks, (2) lacustrine CATT (from various source rocks), (3) epigean CATT from ultramafic source rocks, (4) hypogean CATT from carbonatite source rocks and (5) hypogean CATT from carbonate and/or igneous source rocks (except carbonatites and ultramafics). It can be seen that the five different substratum lithologies are distinguishable in Fig. 3, although they show significant overlapping. Andrews (2006) worked on a synthesis of δ18O andδ13C values in

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

109

Fig. 5. Beryllium and Chromium content in ppm according to specific hydrogeological contexts. Plain markers represent the dataset acquired for this study. Be and Cr detection limits for the new dataset are respectively of 1 and 20 ppm. For readability, only the error bars of minimum and maximum values are represented. For Be N60 ppm, error bars are included in marker size.

epigean CATT from carbonate source rocks; these data are not discussed in this paper. Moreover, lacustrine CATT systems will not be extensively discussed due to their highly complex settings. Indeed, these systems are under the influence of a large variety of source rocks located in the drainage basin. Barium versus strontium cross-plot illustrates some global trends according to the hydrodynamic regime and CATT's source rock lithology (Fig. 4). This figure shows that high strontium values (N200 ppm) tend to be associated to hypogean CATT. Also, the boundary between hypogean and epigean systems on the barium versus strontium cross-plot is non-linear due to the geochemical imprint of the lithologies from the hydrogeological basins (see sections below). 5.2. Geochemical signature of CATT derived from ultramafic source rocks Occurrence of Modern CATT derived from ultramafic source rocks has been reported in many settings including Oman (Clark & Fontes, 1990), California (Barnes & O'Neil, 1969; O'Neil & Barnes, 1971), New Caledonia (Launay & Fontes, 1985; Monnin et al., 2014) and Italy (Chavagnac et al., 2013a,b). Barnes and O'Neil (1969) have studied fluid compositions in ultramafic bedrock undergoing serpentinisation processes. They highlighted the removal of CaO component from peridotites supporting the hypothesis of an ultramafic calcium supplier. According to Clark and Fontes (1990) and O'Neil and Barnes (1971), CATT derived from ultramafic source rocks are relatively easy to distinguish using stable isotopes. According to Fig. 3, CATT with a δ13C lower than −12 ‰ is undoubtedly indicative of ultramafic source rocks. In a pioneer work, Barnes and O'Neil (1969) have shown that the serpentinisation process due to the high water/rock ratio does practically not influence the isotopic composition of coexisting fluids. In this context, CATT precipitation is widely associated to the interaction of low DIC Ca–OH waters with atmospheric CO2. Experimental results presented by Clark et al. (1992) show that the 13C depletion of the Oman travertine is related to extreme kinetic fractionation during CO2 hydroxylation. Schwarzenbach et al. (2013) emphasised the isotopic similarity between the Oman ophiolite CATT and the Ligurian ophiolite CATT and ascribed to the latter the same fractionation processes. Furthermore, if a CATT δ18O value is higher than −4 ‰ and associated to negative δ13C values, then the CATT is likely to originate from ultramafics (Fig. 3). However, some overlapping with the hypogean and lacustrine CATT, if δ13C value is around 0 ‰, adds uncertainty. The large δ13C variability of CATT that precipitated in Oman ophiolites has been

explained as a consequence of variable (i) inputs of soil zone groundwaters and (ii) respiratory activities of C3 and C4 plants (Clark & Fontes, 1990). Schwarzenbach et al. (2013) assigned the control of the isotopic composition to the extent of equilibration with atmospheric CO2 and thus, to the flow rate, the residence time in a pool, temperature and time of equilibration with air. Most CATT from the Ligurian ophiolites have been sampled near the springs and display very low δ13C values. The two samples from Gorzente 1 and 2 displaying δ13C of − 6.7 and − 6.3 ‰ are located either in the distal part of the system, or under the influence of stream water. These values have a similar δ13C signature than many epigean CATT related to limestone source rocks (Fig. 3). The use of barium versus strontium cross-plot (Fig. 4) appears to be a useful way to discriminate the origin of elements that build CATT. Indeed, strontium and barium contents, ranging from 240 to 600 ppm and from 90 to 630 ppm respectively, can be used directly to discriminate CATT associated to ultramafic source rocks from the other CATT. One exception however is the epigean CATT deriving from a mixture of limestone, dolomite and evaporite source rocks. Strontium behaviour during carbonate precipitation in travertine and tufa systems is relatively well constrained. Strontium content in CATT is directly linked to the availability of the element in the waters (Turi, 1986). The strontium content in Ligurian hyperalkaline spring waters varies largely, from 12.6 to 29.6 μmol/l (respectively Rio Leone and Gorzente 2; Chavagnac et al., 2013b). These concentrations are consistent with the global strontium content in Ligurian ophiolite CATT, but are not sufficient to explain the variability between several samples from the same spring. It has been shown in previous studies that strontium content tends to be higher near the spring vents where aragonite is the main carbonate component (Pentecost, 2005). Actually, aragonite precipitates more likely from hot waters (Folk, 1994) and LMC more from cooler waters. However, aragonite to LMC neomorphic processes occur easily and quickly (Pentecost, 2005). It yields to probable strontium partitioning amongst the carbonate body. The syn- to postprecipitation partitioning of strontium explains differences in strontium content observed on the three samples from Rio Leone and the two samples from Gorzente 2. Pentecost (2005) reports a few studies on the barium incorporation into CATT. These studies have shown that barium does not easily substitute for calcium in the calcite lattice due to differences in their ionic radii (0.99 Å for Ca2+ and 1.36 Å for Ba2+) but is more likely to enter the aragonite lattice. An experimental study carried out by Tunusoğlu et al. (2007) shows similar results and explains this behaviour by the orthorhombic crystal structures of both aragonite

110

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

and barium sulphates/carbonates. Moreover, Sturchio (1990) reported values of barium partition coefficient of respectively 0.63 and 0.16 in aragonitic and calcitic travertine from Mammoth Hot Spring, USA. Aragonite neomorphism to calcite would induce a reduction of barium concentrations. Pentecost (2005) also reported similar to higher levels of barium in hypogean and epigean CATT, but did not provide any explanation. In agreement with these previous studies, the present dataset demonstrates that barium content is not necessarily higher in hypogean travertine, as similar concentrations are observed in the hypogean Chaine des Puys and Limagne basin CATT and in epigean Ligurian ophiolite CATT (Fig. 4). Barium content tends to be low in ultramafic rocks but Bocchio (2007) reported the presence of barium-rich micas in eclogites from the Ligurian ophiolite. These could represent substantial barium sources and the spatial distribution of the eclogite blocks from the Ligurian ophiolite region could explain the low barium content of Ponte Arma and Rio Leone samples (Fig. 1(A)). Unfortunately the coupling of δ18O and δ13C in addition to Sr versus Ba is not always sufficient to differentiate ultramafic CATT from epigean CATT derived from limestone, dolomite or evaporite. The statistical analysis of all other analysed elements reveals that the chromium content tends to be higher in ultramafic CATT (Table 3, Fig. 5). Lelli et al. (2013) stressed the common chromium contamination of groundwater associated with local mafic-ultramafic and serpentinite rocks. Chromium occurs mainly within oxides and silicate phases. XRF analysis on serpentinite samples from the Ligurian Nappe has shown significant total chromium content ranging from 1435 to 2700 ppm (Lelli et al., 2013). The Ligurian ophiolite CATT case studies show the highest chromium concentrations in Gorzente 2, Maddalena and Sorgente Sulfurea (Table 1). In addition, the large chromium variability displayed by the Ligurian ophiolite CATT cannot be explained because of the lack of published data concerning the chromium content in the studied spring waters. Therefore, the use of chromium to discriminate the ultramafic CATT from epigean CATT deriving from limestone, dolomite and evaporite needs further investigations and data. 5.3. Geochemical signature of CATT derived from carbonatite or mafic volcanic source rocks Relatively few published studies report geochemical data on Modern CATT formed through the alteration of pure volcanic source rocks. However, an important distinction must be made between mafic source rocks and carbonatite source rocks. Geochemistry of Modern CATT from carbonatites is very poorly documented. One published study provides δ18O and δ13C signatures (Kato & Kraml, 2010) and one other study provides barium and strontium contents (Pisarskii et al., 1998). Both studies deal with hypogean travertines that are derived, at least partially, from alteration of Modern carbonatites from the Ol Doinyo Lengai volcano, Tanzania and from the carbonatites from the Rwenzori springs, Uganda. Pisarskii et al.'s study is of great interest for the purpose of the current work since it demonstrates the extremely high barium and strontium values in CATT deriving from carbonatites compared to those recorded in CATT deriving from mafic source rocks (Fig. 4). Carmody (2012) shows that fenitisation, i.e. hydrothermal alteration of carbonatites, leads to Ba, Rb, Sr, and V enrichments in draining waters of the Ol Doinyo Lengai volcano in Tanzania. These enrichments are directly linked to carbonatitic fluids and to regular Ba and Sr substitutions within carbonate minerals. The same author highlights the natural high abundance of Sr, Ba, P and light rare earth elements in erupted material from this active carbonatitic volcano. Kato and Kraml (2010) provide isotopic data for Ugandian travertines, showing relatively depleted carbon-isotopes compared to values reported in hypogean CATT deriving from other source rocks (Fig. 3). However, this is based on only four samples. More geochemical work on CATT derived from alteration of Modern to Ancient carbonatites is needed to valid their specificδ13C, Sr and Ba values, as suggested in Figs. 3, 4.

Modern CATT developed on mafic volcanic bedrocks can be divided into two categories defined by the influence or not of subsurface carbonate rocks. The first category is quite common and includes many CATT developed in Modern hypogean CATT in Italy (Chafetz & Lawrence, 1994; D'Alessandro et al., 2007), Djibouti (Dekov et al., 2014) and USA (Chafetz & Lawrence, 1994; Benson et al., 1996). It seems likely that some of the calcium incorporated in CATT comes from the dissolution of regional carbonate rocks. In the second category, i.e. those developed in areas where carbonate rocks are lacking, the alteration of volcanic minerals is the most likely source of elements building CATT, as it is the case for hypogean CATT from Réunion Island (this study), the Kamchatka peninsula (Sheshkanova, 2006), and Gran Canaria island (Rodríguez-Berriguete et al., 2012). However, even in a situation dominated by oceanic volcanoes, such as in the Réunion Island system, it is often difficult to demonstrate that Ancient carbonate rocks, potentially present below the volcanic material (i.e. Plio-Pleistocene reef limestones in the Réunion case), are not involved as a source of elements building CATT. The use of strontium isotope geochemistry could be considerably useful in this case (Crossey et al., 2006, 2015). Sallstedt et al. (2014) have however shown that the source of calcium in CATT and speleothems that develop on dolerites in Northern Sweden come from the breakdown of Ca-rich plagioclases such as anorthite (CaAl2Si2O8) without any occurrence of carbonates in the neighbouring areas. Independently of the presence or absence of subsurface carbonates, the δ18O versus δ13C cross plot (Fig. 3) does not allow distinction between hypogean CATT that develop on single or partial mafic context and hypogean CATT related to other lithologies. This problem is due to several abiotic and biological factors controlling both sources and processes of oxygen and carbon incorporation into hypogean CATT (Pentecost, 2005; Fouke et al., 2000; Kele et al., 2011). In contrast to O and C isotopes, the Ba versus Sr cross plot (Fig. 4) shows that Modern hypogean CATT with mafic rocks in their hydrogeological reservoir can be distinguished from hypogean CATT derived from carbonatites and from all epigean CATT. According to Salminen et al. (2005), barium is mainly concentrated in K-feldspars, plagioclases, pyroxenes, amphiboles, apatite and calcite. Based on a comprehensive literature review of Turekian and Wedepohl (1961), barium contents in igneous rocks are of b200 mg kg−1 for ultramafic rocks, 330 mg kg−1 for basaltic rocks, 420 to 840 mg kg−1 for granitic rocks and of 1600 mg kg−1 for syenites (Mielke, 1979; Salminen et al., 2005). The same authors assigned barium concentration of 10 mg kg−1 in carbonates and sandstone and of 580 mg kg−1 in shales. These barium concentrations in sedimentary rocks are directly related to K-feldspars, clay minerals, hydrous iron, and manganese oxides. Moreover, barium can also occur in common minerals such as barite (BaSO4) and witherite (BaCO3), either as authigenic minerals in the source rock or as cements in hydrothermal veins (Paradis & Lavoie, 1996; Canet et al., 2005; Taghipour et al., 2010). Pirajno (2009) documented the occurrence of authigenic tubular barite in submarine hydrothermal chimneys as the result of interactions between hot fluids and cold seawater. Fisher and Puchelt (1972) showed that barium content in hydrothermal fluids is mainly controlled by water/rock interactions and is favoured by interactions with alkaline water. It should be noted that barite is not necessarily a good barium supplier as it is relatively resistant to alteration after burial (Griffith & Paytan, 2012). In addition, evaporite minerals such as anhydrite and gypsum are often associated with authigenic barite located in the lower part of the evaporitic sequences (Einsele, 2000). Nevertheless, barite can be dissolved in zones of sulphate reduction inducing a release of barium into pore water (Riedinger et al., 2006). Excluding possible barite minerals as sources of barium, the relationship between barium content and source rock lithology is well confirmed. Barium concentrations tend to be high in CATT from igneous source rocks and low for CATT from sedimentary source rocks. CATT from limestone, evaporite, and dolomite source rocks generally show low barium content (b100 ppm), whilst barium

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

content in CATT from mafics and granites ranges from 15 to 930 ppm (Fig. 4). In the new data presented here, most of the low barium values of CATT samples from mafics are associated with aragonitic samples from Réunion Island. In addition, the aragonitic and calcitic Limagne graben samples tend to contain more barium than the pure calcitic samples from the same locality. This discrepancy between the two localities may be associated with the availability of barium in the mineralising fluids. Moreover, most of the low barium content found in CATT from Réunion Island is associated with samples from Cilaos. Samples from Salazie display greater amount of barium independently of CATT mineralogy. 5.4. Geochemical signature of CATT that develop on granitoids As mentioned previously, using δ18O versus δ13C and barium versus strontium methods do not allow a clear distinction between CATT originating from mafic rocks and CATT that develop on granitoid bedrocks. This is probably because alteration of granite alone, either in hypogean or in epigean contexts, is insufficient to promote CATT formation. It seems that CATT does not develop on purely granitic continental areas. In the present dataset, all the epigean CATT from the Limagne graben and the Chaîne des Puys are associated not only to granites, but also to other lithologies. In the Limagne graben (i.e. Grotte du Cornador and Fontaine Pétrifiante) dissolution of local Cenozoic limestones is undoubtedly involved in CATT formation (Fig. 1). In the Lefort and PréDimanche springs, which are distant from limestone outcrops, CATT form on two-mica granites in close proximity to Palaeozoic sedimentary and volcaniclastic rocks (Risler, 1974). Despite the limited contribution of granites to CATT formation, the statistical analysis demonstrates the exclusive occurrence of beryllium in CATT samples under the influence of granitic source rocks (Table 3, Fig. 5). The Limagne graben samples display a beryllium content ranging from 6 to 38 ppm, and those from Chaîne des Puys are between 31 and 596 ppm. According to Ure and Berrow (1982), beryllium becomes concentrated in evolved granites (i.e. late-stage products of strongly fractionated granites) and pegmatites due to instability of beryllium complexes at high temperature. Beryllium has not been evaluated in previous CATT studies, and was most likely overlooked or not considered. Due to the economic value of beryllium, several studies have been conducted in order to investigate potentially significant deposits. Barton and Young (2002) considered the occurrence of beryllium deposits associated to hydrothermal activity as a quite common process. They also showed that beryllium-rich deposits occur in a variety of Mn-rich hydrothermal systems and in Fe–Mn-oxide-rich deposits in hot spring systems such as Butte (USA), Silverton (USA), Golconda (USA) and Långban (Sweden). In addition, beryllium deposits are relatively abundant when they are associated with metaluminous to peraluminous granites (Barton & Young, 2002). Warner et al. (1959) related the beryllium occurrence in Golconda hot springs, USA, to the presence of intrusive rocks. Beryllium content in CATT seems to constitute a possible proxy of granitic and/or mafic influence on CATT geochemistry. 5.5. Geochemical signature of hypogean CATT derived from limestones, dolomites and evaporites source rocks Modern CATT including evaporites and/or dolomites in their hydrogeological system are relatively common (Ilhenfeld et al., 2003; Bonny & Jones, 2008; Leybourne et al., 2009; Pentecost, 1999; Koltai et al., 2012; Pola et al., 2014), and are also commonly associated with limestones. Some of them are clearly associated with mafic material and volcanic thermal flux (Fouke et al., 2000), whilst others are not (Kele et al., 2011; Prado-Pérez et al., 2013; Pola et al., 2014). As it can be seen in Fig. 3, besides the possibility to differentiate the hypogean from the epigean CATT, the single use of δ18O signature is unpractical to detect evaporite and dolomite as a source of elements for

111

CATT development. For example, the Jebel Oust travertines, which are undoubtedly influenced by the underlying Triassic evaporites (without mafic rocks), exhibit a similar δ18O signature than those from Réunion Island where evaporites are absent. In contrast, the δ13C signature is slightly different, lower for the Jebel Oust travertines (mean = 2.1 ‰) than that for Réunion Island or Chaîne des Puys (mean = 4.2 ‰). However, this is insufficient to be truly discriminant, and their signatures do overlap with those of Alicun de las Torres (Prado-Pérez et al., 2013), where evaporites are absent. The use of barium and strontium signatures gives better results. Samples from the Jebel Oust contain less barium than other hypogean CATT associated to the occurrence of igneous rocks (Fig. 4). This is confirmed by Kele et al.'s dataset for the Jandarma spring, Turkey. The low barium content is consistent with the low availability of barium in carbonate rocks (see Section 5.3). We note, however, that Fouke et al.'s study showed high barium content in samples from Angel terrace at the Mammoth hot spring, possibly due to the close vicinity of volcanic rocks nearby. 6. Conclusions and perspectives Chemical elements participating in the formation of calcitic or aragonitic travertine and tufa deposits (CATT) may originate from the alteration of a large variety of source rocks, in association with either nonhydrothermal (epigean) or hydrothermal (hypogean) hydrogeological systems. An exhaustive database of the geochemical characteristics compiled from published literature or acquired in the course of this project was built in an attempt to identify geochemical signatures diagnostically tracing the source rocks in Modern CATT. Several potentially useful geochemical tracers were found and are summarised below: (1) As indicated in previous works, δ13C and δ18O cross-plots can be used to discriminate epigean (minδ13C = −27.2 ‰, maxδ13C = 0.9 ‰, meanδ13C = −12.2 ‰ for N = 313) from hypogean systems (minδ13C = − 4 ‰, maxδ13C = 11.7 ‰, meanδ13C = − 2.87 ‰ for N = 198). In addition, very depleted δ13C values (b−12 ‰) are exclusively indicative of an ultramafic origin of elements. Also, δ18O values N−4 ‰ associated to negative δ13C values are indicative of CATT originating from ultramafics. (2) The barium–strontium cross-plot also allows the distinction between epigean and hypogean CATT. In the case of hypogean CATT, it helps to distinguish different source rock groups: (i) the “limestones, evaporites and dolomites” group typically exhibits low barium (b 100 ppm) and high strontium (N400 ppm) contents, (ii) the “mafic and granite” group displays barium contents ranging from 15 to 930 ppm and a high strontium content (N 200 ppm), (iii) the “carbonatites” group shows extremely high barium and strontium values. In epigean CATT, very low barium and strontium contents (respectively b 80 ppm and b100 ppm) are indicative of a pure limestone source rock, whereas low strontium content (b450 ppm) with higher barium content (N 50 ppm) are indicative of mixed limestone, evaporite and dolomite source rocks. CATT from ultramafic source rocks shows similar barium content (from 90 up to 600 ppm) than CATT from limestone, evaporite, and dolomite but their strontium values are statistically higher (between 240 and 600 ppm). (3) A good correlation between ultramafic occurrences in the basin and chromium enrichment in CATT is observed. Chromium being commonly abundant in ultramafic rocks, this element represents a good marker. (4) Based on the Limagne basin and Chaîne des Puys samples, elevated beryllium concentrations, respectively 6 to 38 ppm and 31 to 596 ppm, seem to be indicative of a granitic origin of CATT elements. In non-granitic areas, beryllium content is systematically low (under the detection limit), therefore the use of beryllium as a marker of a granitic contribution in the basin seems relevant.

112

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

Other geochemical tracers, such as strontium and sulphur isotopes, could also be considered. The main purpose of this study was to identify and assess reliable tracers for unravelling geochemical links between CATT and rock lithologies of their hydrogeological reservoir. Making such proxies available is critical for the understanding of the origin of fossil CATT such as those in Presalt deposits in the Southern Atlantic Ocean for instance. It remains, however, to ensure that the proposed geochemical markers/proxies are not significantly affected by diagenetic alteration. This work focussed on Modern CATT. The applicability of the proposed approach to fossil CATT still needs to be demonstrated, in particular when burial diagenesis or telogenetic alteration occurred. This implies to target geochemical investigations onto unaffected primary minerals, such as non-recrystallised early calcite precipitates, by use of micro-sampling or in situ methods. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.sedgeo.2016.01.004. Acknowledgements Thanks are expressed to TOTAL E&P for funding the project and for granting permission to publish. Contribution to the field work is acknowledged to M. Adams, Université de Bourgogne. The Grotte du Cornador direction and staff are thanked for granting permission to sample. Field sampling at the Jebel Oust has been possible thanks to the “Mission Franco-Tunisienne en charge de l'étude du site de Jebel Oust (Tunisie)”. The authors would like to thank T. Cocquerez and L. Bruneau for the XRD and stable isotope analysis, at the Université de Bourgogne. We are grateful to Pr. Laura Crossey and an anonymous reviewer for their useful comments and suggestions. References Andrews, J.E., 2006. Palaeoclimatic records from stable isotopes in riverine tufas: synthesis and review. Earth Sci. Rev. 75, 85–104. Autran, A., Peterlongo, J.M., 1980. France: Massif Central. In: Lorenz, C. (Ed.), Géologie des pays européens : FranceBelgique. Luxembourg. Dunod, Paris, pp. 4–123. Barnes, I., O'Neil, J.R., 1969. The relationship between fluids in some fresh alpine-type ultramafics and possible modern serpentinization, Western United States. Geol. Soc. Am. Bull. 80, 1947–1960. Barnes, I., Presser, T.S., Saines, M., Dickson, P., Koster van Groos, A.F., 1982. Geochemistry of highly basic calcium hydroxide groundwater in Jordan. Chem. Geol. 35, 147–154. Barton, M.D., Young, S., 2002. Non-pegmatitic deposits of beryllium; mineralogy, geology, phase equilibria and origin. Rev. Mineral. 50, 591–691. Battani, A., Deville, E., Faure, J.L., Jeandel, E., Noirez, S., Tocqué, E., Benoit, Y., Schmitz, J., Parlouar, D., Sarda, P., 2010. Geochemical study of natural CO2 emissions in the French Massif Central: how to predict origin, processes and evolution of CO2 leakage. Oil Gas Sci. Technol. 65, 615–633. Baubron, J.C., Bosch, B., Degranges, P., Leleu, M., Marcé, A., Rissler, J.J., Sarcia, C., 1978. Recherches géochimiques sur les eaux thermales de la bordure Ouest de la Limagne. BRGM Report 78 SGN 566 MGA 25 pp. Baubron, J.C., Mercier, F., Rouzaire, D., 1992. Eaux minérales de Sainte Marguerite (Puy de Dôme) — Prospection géochimique in situ des gaz des sols. BRGM Report R-36492AUV-4S-92 35 pp. Benson, L., White, L.D., Rye, R., 1996. Carbonate deposition, Pyramid Lake subbasin, Nevada: 4. Comparison of the stable isotope values of carbonate deposits (tufas) and the Lahontan lake-level record. Palaeogeogr. Palaeoclimatol. Palaeoecol. 122, 45–76. Bertin, C., Rouzaire, D., 2004. Amélioration de la connaissance des ressources en eau souterraine des sites thermaux en Auvergne — site du basin de Vichy (03). BRGM Report RP-53095-FR 169 pp. Bocchio, R., 2007. Barium-rich phengite in eclogites from the Voltri group (northwestern Italy). Periodico di Mineralogia 76, 155–167. Boivin, P., Besson, J.C., Gourgaud, A., Miallier, D., Thouret, J.C., 2009. The eruption of the Lake Pavin maar 6900 yr ago: a review. International Meeting, Lake Pavin and other Meromictic Lakes. Besse et St Anastaise, France, pp. 16–19. Bonny, S.M., Jones, B., 2008. Controls on the precipitation of barite (BaSO4) crystals in calcite travertine at Twitya Spring, a warm sulphur spring in Canada's Northwest Territories. Sediment. Geol. 203, 36–53. Borghini, G., Rampone, E., Crispini, L., De Ferrari, R., Godard, M., 2007. Origin and emplacement of ultramafic–mafic intrusions in the Erro–Tobbio mantle peridotite (Ligurian Alps, Italy). Lithos 94, 210–229. Boulart, C., Chavagnac, V., Monnin, C., Delacour, A., Ceuleneer, G., Hoareau, G., 2013. Differences in gas venting from ultramafic-hosted warm springs: the example of Oman and Voltri ophiolites. Ofioliti 38, 143–156. Brouwer, F.M., Vissers, R.L.M., Lamb, W.M., 2002. Metamorphic history of eclogitic metagabbro blocks from a tectonic melange in the Voltri Massif, Ligurian Alps, Italy. Ofioliti 27 (1), 1–16.

Canet, C., Prol-Ledesma, R.M., Torres-Alvarado, I., Gilg, H.A., Villanueva, R.E., Cruz, R.L.-S., 2005. Silica-carbonate stromatolites related to coastal hydrothermal venting in Baia Concepcion, Baja California Sur, Mexico. Sediment. Geol. 174, 97–113. Capezzuoli, E., Gandin, A., Pedley, M., 2014. Decoding tufa and travertine (fresh water carbonates) in the sedimentary record: the state of the art. Sedimentology 61, 1–21. Carmody, L., 2012. Geochemical characteristics of carbonatite-related volcanism and subvolcanic metasomatism at Oldoinyo Lengai, Tanzania (PhD. thesis) University College London, London, England 338 pp. Casanova, J., Bodénan, F., Négrel, P., Azaroual, M., 1999. Microbial control on the precipitation of modern ferrihydrite and carbonate deposits from the Cézallier hydrothermal springs (Massif Central, France). Sediment. Geol. 126, 125–145. Caudwell, C., 1983. Les rivulariacées actuelles: interprétation possible de la structure zonée des concrétions stromatolitiques à Rivularia haematites. Geobios 16, 169–177. Caudwell, C., Lang, J., Pascal, A., 2001. Lamination of swampy-rivulets Rivularia haematites stromatolites in a temperate climate. Sediment. Geol. 143, 125–147. Chafetz, H.S., Lawrence, J.R., 1994. Stable isotopic variability within modern travertines. Géog. Phys. Quatern. 48, 257–273. Chafetz, H.S., Rush, P.F., Utech, N.M., 1991. Microenvironmental controls on mineralogy and habit of CaCO3 precipitates: an example from an active travertine system. Sedimentology 38, 107–126. Charguéron, C., Bertin, C., Rouzaire, D., Vigouroux, P., 2003. Amélioration de la connaissance des ressources en eau souterraine des sites thermaux en Auvergne — Site de Saint-Nectaire (63). BRGM Report RP 52469-FR 158 pp. Chavagnac, V., Ceuleneer, G., Monnin, C., Lansac, B., Hoareau, G., Boulart, C., 2013a. Mineralogical assemblages forming at hyperalkaline warm springs hosted on ultramafic rocks: a case study of Oman and Ligurian ophiolites: mineral precipitate at Alkaline Springs. Geochemistry, Geophysics. Geosyst. 14, 2474–2495. Chavagnac, V., Monnin, C., Ceuleneer, G., Boulart, C., Hoareau, G., 2013b. Characterization of hyperalkaline fluids produced by low-temperature serpentinization of mantle peridotites in the Oman and Ligurian ophiolites: hyperalkaline waters in Oman and Liguria. Geochem. Geophys. 14, 2496–2522. Cipolli, F., Gambardella, B., Marini, L., Ottonello, G., Vetuschi Zuccolini, M., 2004. Geochemistry of high-pH waters from serpentinites of the Gruppo di Voltri (Genova, Italy) and reaction path modeling of CO2 sequestration in serpentinite aquifers. Appl. Geochem. 19 (5), 787–802. Clark, I.D., Fontes, J.C., 1990. Palaeoclimatic reconstruction in northern Oman based on carbonates from hyperalkaline groundwaters. Quat. Res. 33, 320–336. Clark, I.D., Fontes, J.-C., Fritz, P., 1992. Stable isotope disequilibria in travertine from high pH waters: laboratory investigations and field observations from Oman. Geochim. Cosmochim. Acta 56, 2041–2050. Crossey, L.J., Fisher, T.P., Patchett, P.J., Karlstrom, K.E., Hilton, D.R., Newell, D.L., Huntoon, P., Reynolds, A.C., de Leeuw, G.A.M., 2006. Dissected hydrologic system at the Grand Canyon: interaction between deeply derived fluids and plateau aquifer waters in modern springs and travertine. Geology 34 (1), 25–28. Crossey, L.J., Karlstrom, K.E., Springer, A.E., Newell, D., Hilton, D.R., Fischer, T., 2009. Degassing of mantle-derived CO2 and He from springs in the southern Colorado Plateau region-neotectonic connections and implications for groundwater systems. GSA Bull. 121 (7–8), 1034–1053. Crossey, L.J., Karlstrom, K.E., Dorsey, R., Pearse, J., Wan, E., Beard, L.S., Asmerom, Y., Polyak, V., Crow, R.S., Cohen, A., Bright, J., Pecha, M.E., 2015. Importance of groundwater in propagating downward integration of the 6–5 Ma Colorado river system: geochemistry of springs, travertines, and lacustrine carbonates of the Grand Canyon region over the past 12 Ma. Geosphere 11 (3), 660–682. Curie, J., 2013. Les travertins anthropiques, entre Histoire, Archéologie et Environnement: Etude géoarchéologique du site antique de Jebel Oust, Tunisie (PhD. thesis) Université de Bourgogne, Dijon, France 325 pp. D'Alessandro, W., Giammanco, S., Bellomo, S., Parello, F., 2007. Geochemistry and mineralogy of travertine deposits of the SW flank of Mt. Etna (Italy): relationships with past volcanic and degassing activity. J. Volcanol. Geotherm. Res. 165, 64–70. Dekov, V.M., Egueh, N.M., Kamenov, G.D., Bayon, G., Lalonde, S.V., Schmidt, M., Liebetrau, V., Munnik, F., Fouket, Y., Tanimizu, M., Awaleh, M.O., Guirreh, I., Le Gall, B., 2014. Hydrothermal carbonate chimneys from a continental rift (Afar Rift): mineralogy, geochemistry, and mode of formation. Chem. Geol. 387, 87–100. Delance, J.H., 1988. Le karst de Bourgogne. Kardiologiia 11-12, 7–16. Della Porta, G., 2015. Carbonate build-ups in lacustrine, hydrothermal and fluvial settings: comparing depositional geometry, fabric types and geochemical signature. In: Bosence, D.W.J., Gibbons, K.A., Le Heron, D.P., Morgan, W.A., Protchard, T., Vining, B.A. (Eds.), Microbial Carbonates in Space and Time: Implications for Global Exploration and Production. Geological Society, London Special Publications 418. First published online March 3, 2015. Demange, J., Chovelon, P., Puvilland, P., 1989. Geothermal model of the Salazie Cirque (Réunion Island). Volcanic and structural implication. J. Volcanol. Geotherm. Res. 36, 153–176. Dèzes, P., Schmid, S.M., Ziegler, P.A., 2004. Evolution of the European Cenozoic Rift system: interaction of the Alpine and Pyrenean orogens with their foreland lithosphere. Tectonophysics 389, 1–33. Dickson, A.G., 1981. An exact definition of total alkalinity and a procedure for the estimation of alkalinity and total inorganic carbon from titration data. Deep sea research part A. Oceanogr. Res. Pap. 28 (6), 609–623. Einsele, G., 2000. Sedimentary Basins, Evolution, Facies and Sediment Budget. SpringerVerlag, Berlin 792 pp. El Desouky, H., Soete, J., Claes, H., Özkul, M., Vanhaecke, F., Swennen, R., 2015. Novel applications of fluid inclusions and isotope geochemistry in unravelling the genesis of fossil travertine systems. Sedimentology 62, 27–56.

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114 Feuga, B., 1987. Le système géothermal du cézallier: cadre géologique général et reconnaissance par sondages. Géol. Fr. Mém. 2, 3–16. Fisher, K., Puchelt, H., 1972. Barium. In: Wedepohl, K.H. (Ed.) Handbook of Geochemistry vol. II-3. Springer, Berlin 56-A-1-56-O-22. Folk, R.L., 1994. Interaction between bacteria, nanobacteria, and mineral precipitation in hot springs of central Italy. Géog. Phys. Quatern. 48, 233–246. Fouke, B.W., Farmer, J.D., Des Marais, D.J., Pratt, L., Sturchio, N.C., Burns, P.C., Discipulo, M.K., 2000. Depositional facies and aqueous–solid geochemistry of travertine depositing hot springs (Angel Terrace, Mammoth Hot Springs, Yellowstone National Park, USA). J. Sediment. Res. 70, 565–585. Friedman, I., 1970. Some investigations of the deposition of travertine from hot-springs; the isotopic chemistry of a travertine depositing spring. Geochim. Cosmochim. Acta 34, 1303–1315. Friedman, I., O'Neil, J.R., 1977. Compilation of stable isotope fractionation factors of geochemical interest. Data of Geochemistry 6th, Geological Survey Professional Paper 440–KK. Frissant, N., Lacquement, F., Rançon, J.P., 2003. Etude du potentiel hydrothermal de la zone amont de la rivière du Bras Rouge (Cirque de Cilaos). BRGM Report RP52673-FR 31 pp. Frissant, N., Jossot, O., Cruchet, M., 2004. Quantification des volumes d'eau souterraine mobilisable sur la micro-région Est sur la base des données existantes et avis sur le franchissement des rivières par l'ouvrage destiné à transférer les eaux. Rapport d'avancement. BRGM report SP 53459-FR 27 pp. Gal, F., Brach, M., Braibant, G., Bény, C., Michel, K., 2012. What can be learned from natural analogue studies in view of CO2 leakage issues in carbon capture and storage applications? Geochemical case study of Sainte-Marguerite area (French Massif Central). Int. J. Greenhouse Gas Control 10, 470–485. Gandin, A., Capezzuoli, E., 2008. Travertine versus calcareous tufa: distinctive petrologic features and stable isotope signatures. Italian. J. Quat. Sci. 21, 125–136. Garnett, E.R., Andrews, J.E., Preece, R.C., Dennis, P.F., 2004. Climatic change recorded by stable isotopes and trace elements in a British Holocene tufa. J. Quat. Sci. 19 (3), 251–262. Griffith, E.M., Paytan, A., 2012. Barite in the ocean — occurrence, geochemistry, and paleoceanographic applications. Sedimentology 59, 1817–1835. Guo, L., Riding, R., 1998. Hot-spring travertine facies and sequences, Late Pleistocene, Rapolano Terme, Italy. Sedimentology 45, 163–180. Guo, L., Riding, R., 1999. Rapid facies changes in Holocene fissure ridge hot spring travertines, Rapolano Terme, Italy. Sedimentology 46, 1145–1158. Hamelin, C., Seitz, H.M., Barrat, J.A., Dosso, L., Maury, R.C., Chaussidon, M., 2009. A low δ7Li lower crustal component: evidence from an alkalic intraplate volcanic series (Chaine des Puys, French Massif Central). Chem. Geol. 266, 205–217. Harris, P.M., Ellis, J., Purkis, S.J., 2013. Assessing the extent of carbonate deposition in early rift settings. AAPG Bull. 97, 27–60. Hermann, J., Müntener, O., Scambelluri, M., 2000. The importance of serpentinite mylonites for subduction and exhumation of oceanic crust. Tectonophysics 327, 225–238. Holm, S., 1979. A simple sequentially rejective multiple test procedure. Scand. J. Stat. 6, 65–70. Hoogerduijn Strating, E.H., 1991. The evolution of the Piemonte–Ligurian ocean: a structural study of ophiolite complexes in Liguria (NW Italy). Instituut voor Aardwetenschappen der Rijksuniversiteit Utrecht, Utrecht 127 pp. Horvatinčić, N., Krajcar-Bronic, I., Obelic, B., 2003. Differences in the 14C age, δ13C and δ18O of Holocene tufa and speleothem in the Dinaric Karst. Palaeogeogr. Palaeoclimatol. Palaeoecol. 193, 139–157. Ihlenfeld, C., Norman, M.D., Gagan, M.K., Drysdale, R.N., Maas, R., Webb, J., 2003. Climatic significance of seasonal trace element and stable isotope variations in a modern freshwater tufa. Geochim. Cosmochim. Acta 67 (13), 2342–2357. Jauzein, A., 1967. Contribution à l'étude géologique des confins de la dorsale tunisienne (Tunisie septentrionale). Ann. Min. Géologie 22 475 pp. Javaux, C., 1992. La plate-forme parisienne et bourguignonne au Bathonien terminal et au Callovien — Dynamique sédimentaire, séquentielle et diagénétique, Place et création des réservoirs potentiels. Mémoires géologiques de l'Université de Dijon, Dijon, France 342 pp. Join, J.-L., Coudray, J., 1997. Using principal components analysis and Na/Cl ratios to trace ground water circulation in a volcanic island. Réunion Island example. J. Hydrol. 190, 1–18. Jones, G.D., Xiao, Y., 2013. Geothermal convection in South Atlantic pre-salt lacustrine carbonates: developing diagenesis and reservoir quality predictive concepts with reactive transport models. AAPG Bull. 97 (8), 1249–1271. Kato, V., Kraml, M., 2010. Geochemistry of Rwenzori hot spring waters. Proceeding of the African Rift Geothermal Conference (ARGeo) 3, pp. 172–183 November 22-25. Kele, S., Özkul, M., Fórizs, I., Gökgöz, A., Baykara, M.O., Alçiçek, M.C., Németh, T., 2011. Stable isotope geochemical study of Pamukkale travertines: new evidences of lowtemperature non-equilibrium calcite–water fractionation. Sediment. Geol. 238, 191–212. Kluska, J.-M., 1997. Evolution magmatique et morpho-structurale du Piton des Neiges au cours des derniers 500,000 ans (PhD. thesis) Université Paris XI, Orsay, Paris, France 229 pp. Koltai, G., Kele, S., Bárány-Kevei, I., 2012. Preliminary studies of freshwater tufa deposits in Mecsek Mts., Hungary. Act. Climatol. Chorologica, Universitatis Szeged. 46, 143–151. Launay, J., Fontes, J.C., 1985. Les sources thermales de prony (nouvelle calédonie) et leurs précipités chimiques. Exemple de formation de brucite primaire. Géol. Fr. 1, 83–100. Lelli, M., Grassi, S., Amadori, M., Franceschini, F., 2013. Natural Cr(VI) contamination of groundwater in the Cecina coastal area and its inner sectors (Tuscany, Italy). Environ. Earth Sci. 71, 3907–3919. Leybourne, M.I., Betcher, R.N., McRitchie, W.D., Kaszycki, C.A., Boyle, D.R., 2009. Geochemistry and stable isotopic composition of tufa waters and precipitates from the

113

Interlake region, Manitoba, Canada: constraints on groundwater origin, calcitization, and tufa formation. Chem. Geol. 260, 221–233. Li, H.-C., Xu, X.-M., Ku, T.-L., You, C.-F., Buchheim, H.P., Peters, R., 2008. Isotopic and geochemical evidence of palaeoclimate changes in Salton Basin, California, during the past 20 kyr: 1. δ18O and δ13C records in lake tufa deposits. Palaeogeogr. Palaeoclimatol. Palaeoecol. 259, 182–197. Lojen, S., Dolenec, T., Vokal, B., Cukrov, N., Mihelcics, G., Papesch, W., 2004. C and O stable isotope variability in recent freshwater carbonates (River Krka, Croatia). Sedimentology 51, 361–375. Makhnach, N., Zernitskaja, V., Kolosov, I., Simakova, G., 2004. Stable oxygen and carbon isotopes in Late Glacial–Holocene freshwater carbonates from Belarus and their palaeoclimatic implications. Palaeogeogr. Palaeoclimatol. Palaeoecol. 209, 73–101. Marty, B., Meynier, V., Nicolini, E., Griesshaber, E., Toutain, J.P., 1993. Geochemistry of gas emanations: a case study of the Réunion hot spot, Indian Ocean. Appl. Geochem. 8 (2), 141–152. Matsuoka, J., Kano, A., Oba, T., Watanabe, T., Sakai, S., Seto, K., 2001. Seasonal variation of stable isotopic compositions recorded in a laminated tufa, SW Japan. Earth Planet. Sci. Lett. 192, 31–44. Maury, R.C., Varet, J., 1980. Le volcanisme tertiaire et quaternaire de France. 26 Int. Geol. Congr (Paris), BRGM 3. 142pp. Merle, O., Michon, L., 2001. The formation of the est European rift: a new model as exemplified by the Massif Central area. Bull. Soc. Geol. Fr. 172 (2), 81–89. Merle, O., Michon, L., Camus, G., De Goer, A., 1998. L'extension oligocène sur la transversale septentrionale du rift du Massif Central. Bull. Soc. Geol. Fr. 169 (5), 615–626. Mervine, E.V., Humphris, S.E., Sims, K.W.W., Kelemen, P.B., Jenkins, W.J., 2014. Carbonation rates of peridotite in the Samail ophiolite Sultanate of Oman, constrained through 14C dating and stable isotopes. Geochim. Cosmochim. Acta 126, 371–397. Mielke, J.E., 1979. Composition of the Earth's crust and distribution of the elements. In: Siegel, F.R. (Ed.), Reviews of Research on Modern Problems in Geochemistry. International Association of Geochemistry and Cosmochemistry And UNESCO, Paris, pp. 13–37. Monna, F., Puertas, A., Lévêque, F., Losno, R., Fronteau, G., Marin, B., Dominik, J., Petit, C., Forel, B., Chateau, C., 2008. Geochemical records of limestone façades exposed to urban atmospheric contamination as monitoring tools? Atmos. Environ. 42, 999–1011. Monnin, C., Chavagnac, V., Boulart, C., Ménez, B., Gérard, M., Gérard, E., Pisapia, C., Quéméneur, M., Erauso, G., Postec, A., Guentas-Dombrowski, L., Payri, C., Pelletier, B., 2014. Fluid chemistry of the low temperature hyperalkaline hydrothermal system of Prony Bay (New Caledonia). Biogeosciences 11, 5687–5706. Moulin, M., Lebon, D., Frissant, N., 2002. Synthèse hydrogéologique du cirque de Salazie (Ile de La Réunion). BRGM Report RP-51450-FR 91 pp. Négrel, P., Petelet-Giraud, E., Serra, H., Millot, R., Kloppmann, W., 2004. Caractéristiques hydrogéochimiques et isotopiques d'eaux thermo-minérales du Massif Central. Inventaire du potentiel géothermique de la Limagne (projet COPGEN). BRGM Report 53597-FR 165 pp. Oehler, J.-F., Lénat, J.-F., Labazuy, P., 2008. Growth and collapse of the Réunion Island volcanoes. Bull. Volcanol. 70, 717–742. Ohta, A., Imai, N., Terashima, S., Tachibana, Y., 2005. Application of multi-element statistical analysis for regional geochemical mapping in Central Japan. Appl. Geochem. 20, 1017–1037. O'Neil, J.R., Barnes, I., 1971. 13C and 18O compositions in some fresh-water carbonates associated with ultramafic rocks and serpentinites: western United States. Geochim. Cosmochim. Acta 35, 687–697. O'Neil, J.R., Clayton, R.N., Mayeda, T.K., 1969. Oxygen isotope fractionation in divalent metal carbonates. J. Chem. Phys. 51, 5547–5558. Paradis, S., Lavoie, D., 1996. Multiple-stage diagenetic alteration and fluid history of Ordovician carbonate-hosted barite mineralization, Southern Quebec Appalachians. Sediment. Geol. 107, 121–139. Pauwels, H., Fouillac, C., Goff, F., Vuataz, F.D., 1997. The isotopic and chemical composition of CO2-rich thermal waters in the Mont-Dore region (Massif-Central, France). Appl. Geochem. 12, 411–427. Pentecost, A., 1999. The origin and development of the travertines and associated thermal waters at Matlock bath, Derbyshire. Proc. Geol. Assoc. 110, 217–232. Pentecost, A., 2005. Travertine. Springer-Verlag, Berlin, Heidelberg 445 pp. Pentecost, A., Viles, H.A., 1994. A review and reassessment of travertine classification. Géog. Phys. Quatern. 48, 305–314. Pirajno, F., 2009. Hydrothermal Processes and Mineral Systems. Springer, Berlin 1250 pp. Pisarskii, B.I., Konev, A.A., Levi, K.G., Delvaux, D., 1998. Carbon dioxide-bearing alkaline hydrotherms and strontium-bearing travertines in the Songwe River Valley (Tanzania). Russ. Geol. Geophys. 39 (7), 941–948. Pola, M., Gandin, A., Tuccimei, P., Soligo, M., Deiana, R., Fabbri, P., Zampieri, D., 2014. A multidisciplinary approach to understanding carbonate deposition under tectonically controlled hydrothermal circulation: a case study from a recent travertine mound in the Euganean hydrothermal system, northern Italy. Sedimentology 61, 172–199. Prado-Pérez, A., Huertass, A.D., Crespo, M.T., Martin Sanchez, A., Del Villar, L.P., 2013. Late Pleistocene and Holocene mid-latitude palaeoclimatic and palaeoenvironmental reconstruction: an approach based on the isotopic record from a travertine formation in the Guadix-Baza basin, Spain. Geol. Mag. 150 (4), 602–625. Purser, B.H., 1975. Sédimentation et diagenèse précoce des séries carbonatées du Jurassique moyen de Bourgogne (PhD. thesis) Orsay, Paris 450 pp. Rançon, J.-P., Rocher, P., Auge, T., Nativel, P., 1988. Pétrologie de deux complexes intrusifs lités de l'île de La Réunion. Implications volcano-structurales. Géol. Fr. 2-3, 135–156.

114

P.-A. Teboul et al. / Sedimentary Geology 334 (2016) 97–114

Riedinger, N., Kasten, S., Groger, J., Franke, C., Pfeifer, K., 2006. Active and buried authigenic barite fronts in sediments from the Eastern Cape Basin. Earth Planet. Sci. Lett. 241, 876–887. Risler, J.J., 1974. Description et classification géologique des sources minérales et thermales du Massif Central. BRGM Report 74 SGN 418 MCE 37 pp. Rodríguez-Berriguete, A., Alonso-Zarza, A.M., Cabrera, M.C., Rodriguez-Gonzalez, A., 2012. The Azuaje travertine: an example of aragonite deposition in a recent volcanic setting, N Gran Canaria Island, Spain. Sediment. Geol. 277, 61–71. Sallstedt, T., Ivarsson, M., Lundberg, J., Sjöberg, R., Vidal Romani, J.R., 2014. Speleothem and biofilm formation in a granite/dolerite cave, northern Sweden. Int. J. Speleol. 43 (3), 305–313. Salminen, R. (Chief-editor), Batista, M.J., Bidovec, M. Demetriades, A., De Vivo, B., De Vos, W., Duris, M., Gilucis, A., Gregorauskiene, V., Halamic, J., Heitzmann, P., Lima, A., Jordan, G., Klaver, G., Klein, P., Lis, J., Locutura, J., Marsina, K., Mazreku, A., O'Connor, P.J., Olsson, S.Å., Ottesen, R.-T., Petersell, V., Plant, J.A., Reeder, S., Salpeteur, I., Sandström, H., Siewers, U., Steenfelt, A., Tarvainen, T., 2005. Geochemical atlas of Europe. Part 1 — Background Information, Methodology and Maps. Geological Survey of Finland, Espoo, Finland. Retrieved from: http://weppi.gtk.fi/publ/foregsatlas/. Schwarzenbach, E.M., Lang, S.Q., Früh-Green, G.L., Lilley, M.D., Bernasconi, S.M., Méhay, S., 2013. Sources and cycling of carbon in continental, serpentinite-hosted alkaline springs in the Voltri Massif, Italy. Lithos 177, 226–244. Serra, H., Petelet-Giraud, E., Negrel, P., 2003. Inventaire du potentiel géothermique de la Limagne (COPGEN). Synthèse bibliographique de la géochimie des eaux thermales. BRGM Report RP-52587-FR 84 pp. Sheshkanova, K.O., 2006. New data on composition of precipitates from Grifon Ivanova Nalychevskaya hydrothermal system, East Kamchatka. Vestnik Kraunc. Earth Sci. 8, 201–205. Stoffers, P., Botz, R., 1994. Formation of hydrothermal carbonate in Lake Tanganyika, Eastcentral Africa. Chem. Geol. 115, 117–122. Sturchio, N.C., 1990. Radium isotopes, alkaline earth diagenesis, and age determination of travertine from Mammoth Hot Springs, Wyoming, U.S.A. Appl. Geochem. 5, 631–640. Sun, H., Liu, Z., 2010. Wet-dry seasonal and spatial variations in the δ13C and δ18O values of the modern endogenic travertine at Baishuitai, Yunnan, SW China and their paleoclimatic and paleoenvironmental implications. Geochim. Cosmochim. Acta 74, 1016–1029. Sürmelihindi, G., Passchier, C.W., Baykan, O.N., Spötl, C., Kessener, P., 2013. Environmental and depositional controls on laminated freshwater carbonates: an example from the Roman aqueduct of Patara, Turkey. Palaeogeogr. Palaeoclimatol. Palaeoecol. 386, 321–335. Taghipour, B., Moore, F., Mackizadeh, M.A., 2010. Stable isotope evidences of jarosite–barite mineralization in the Rangan rhyolitic dome NE Isfahan, Iran. Chem. Erde 70, 377–384.

Tonusoğlu, Ö., Shahwan, T., Eroğlu, A.E., 2007. Retention of aqueous Ba2+ ions by calcite and aragonite over a wide range of concentrations: characterization of the uptake capacity and kinetic of sorption and precipitate formation. Geochem. J. 41, 379–389. Turekian, K.K., Wedepohl, K.H., 1961. Distribution of the elements in some major units of the Earth's crust. Geol. Soc. Am. Bull. 72, 175–192. Turi, B., 1986. Stable isotope geochemistry of travertines. In: Fritz, P., Fontes, J.C. (Eds.), Handbook of Environmental Isotope GeochemistryThe Terrestrial Environment, 2B. Elsevier, Amsterdam, pp. 207–238. Ure, A.M., Berrow, M.L., 1982. The chemical constituents of soils. In: Bowen, H.J.M. (Ed.), Environmental ChemistryRoyal Society of Chemistry. Burlington House, London, pp. 94–202. Vignaroli, G., Faccenna, C., Rossetti, F., 2009. Retrogressive fabric development during exhumation of the Voltri Massif (Ligurian Alps, Italy): arguments for an extensional origin and implications for the Alps-Apennines linkage. Int. J. Earth Sci. 98, 1077–1093. Vincent, B., Emmanuel, L., Loreau, J.P., Thierry, J., 1997. Caractérisation et interprétation de cycles géochimiques sur la plate-forme bourguignone (France) au bajocien — bathonien. Compte Rendus de l'Académie des sciences – Series IIa –. Earth Planet. Sci. 325 (10), 783–789. Warner, L.A., Holser, W.T., Wilmarth, V.R., Cameron, E.N., 1959. Occurrence of Nonpegmatite Beryllium in the United States. 318. Geological Survey Professional Paper, U.S. 198 pp. Wattinne, A., 2004. Evolution d'un environnement carbonaté lacustre à bioconstruc-tions, en Limagne bourbonnaise (Oligo-Miocène, Massif Central, France) (PhD. thesis) Museum National d'Histoire Naturelle, Paris, France 279 pp. Wattinne, A., Quesnel, F., Mélières, F., Guillou, H., Bourdillon, C., Barbarand, J., Baele, J.-M., Yans, J., 2010. Upper Cretaceous feldspars in the Cenozoic Limagne Basin: a key argument in reconstructing the palaeocover of the Massif Central (France). Palaeogeogr. Palaeoclimatol. Palaeoecol. 298 (3–4), 175–188. Wright, V.P., 2012. Lacustrine carbonates in rift settings: the interaction of volcanic and microbial processes on carbonate deposition. In: Garland, J., Neilson, J.E., Laubach, S.E., Whidden, K.J. (Eds.), Advances in Carbonate Exploration and Reservoir Analysis. Geological Society, London, Special Publications 370, pp. 39–47. Wright, V.P., Barnett, A., 2015. An abiotic model for the development of textures in some South Atlantic early Cretaceous lacustrine carbonates. In: Bosence, D.W.J., Gibbons, K.A., Le Heron, D.P., Morgan, W.A., Protchard, T., Vining, B.A. (Eds.), Microbial Carbonates in Space and Time: Implications for Global Exploration and Production. Geological Society, London Special Publications 418. First published online February 26, 2015. Ziegler, P.A., 1990. Geological atlas of Western and Central Europe, second and completely. Shell International Petro. Maatschappi B.V, Revised ed. Geological Society Publishing House 239 pp. Ziegler, P.A., 1994. Cenozoic rift system of western and Central Europe: an overview. Geol. Mijnb. 73, 99–127.