Available online at www.sciencedirect.com

www.elsevier.com/locate/gca

Geochimica et Cosmochimica Acta 72 (2008) 5070–5089

Magnesium isotope systematics of the lithologicallyvaried Moselle river basin, France

Agnes Brenot *, Christophe Cloquet, Nathalie Vigier, Jean Carignan,Christian France-Lanord

CRPG, Centre de Recherches Petrologiques et Geochimiques, 15 rue Notre Dame des Pauvres, 54501 Vandoeuvre les Nancy, France

Received 24 October 2007; accepted in revised form 27 July 2008; available online 20 August 2008

Abstract

Magnesium and strontium isotope signatures were determined during different seasons for the main rivers of the Mosellebasin, northeastern France. This small basin is remarkable for its well-constrained and varied lithology on a small distancescale, and this is reflected in river water Sr isotope compositions. Upstream, where the Moselle River drains silicate rocks ofthe Vosges mountains, waters are characterized by relatively high 87Sr/86Sr ratios (0.7128–0.7174). In contrast, downstream ofthe city of Epinal where the Moselle River flows through carbonates and evaporites of the Lorraine plateau, 87Sr/86Sr ratiosare lower, down to 0.70824.

Magnesium in river waters draining silicates is systematically depleted in heavy isotopes (d26Mg values range from �1.2 to�0.7&) relative to the value presently estimated for the continental crust and a local diorite (�0.5&). In comparison, d26Mgvalues measured in soil samples are higher (�0.0&). This suggests that Mg isotope fractionation occurs during mineral leachingand/or formation of secondary clay minerals. On the Lorraine plateau, tributaries draining marls, carbonates and evaporites arecharacterized by low Ca/Mg (1.5–3.2) and low Ca/Sr (80–400) when compared to local carbonate rocks (Ca/Mg = 29–59;Ca/Sr = 370–2200), similar to other rivers draining carbonates. The most likely cause of the Mg and Sr excesses in these riversis early thermodynamic saturation of groundwater with calcite relative to magnesite and strontianite as groundwater chemistryprogressively evolves in the aquifer. d26Mg of the dissolved phases of tributaries draining mainly carbonates and evaporites arerelatively low and constant throughout the year (from �1.4& to �1.6& and from �1.2& to �1.4&, respectively), within therange defined for the underlying rocks. Downstream of Epinal, the compositions of the Moselle River samples in a d26Mg vs.87Sr/86Sr diagram can be explained by mixing curves between silicate, carbonate and evaporite waters, with a significantcontribution from the Vosgian silicate lithologies (>70%). Temporal co-variation between d26Mg and 87Sr/86Sr for the MoselleRiver throughout year is also observed, and is consistent with a higher contribution from the Vosges mountains in winter, interms of runoff and dissolved element flux. Overall, this study shows that Mg isotopes measured in waters, rocks and soils,coupled with other tracers such as Sr isotopes, could be used to better constrain riverine Mg sources, particularly if analyticaluncertainties in Mg isotope measurements can be improved in order to perform more precise quantifications.� 2008 Elsevier Ltd. All rights reserved.

1. INTRODUCTION

Magnesium and calcium are intimately linked to the car-bon cycle. Both are related to ocean pH, and the weathering

0016-7037/$ - see front matter � 2008 Elsevier Ltd. All rights reserved.

doi:10.1016/j.gca.2008.07.027

* Corresponding author. Present address: BRGM, 3 avenue C.Guillemin, BP 39009, 45060 Orleans, France.

E-mail address: a.brenot@brgm.fr (A. Brenot).

of Ca and Mg silicates is a significant long-term sink ofatmospheric CO2 (Walker et al., 1981; Berner et al.,1983). Quantifying the sources of Ca and Mg in the dis-solved phases of rivers, as well as the relative proportionsof silicate and carbonate weathered is thus of primaryimportance for understanding atmospheric CO2 consump-tion rates (Galy and France-Lanord, 1999; Millot et al.,2003).

Magnesium isotope systematics on the Moselle river basin 5071

Many studies use Ca/Na, Ca/Sr ratios and the Sr iso-tope composition of rivers and sediments to distinguishstream water solute fluxes derived from silicate and carbon-ate weathering (e.g. Wadleigh et al., 1985; Galy et al., 1999;English et al., 2000). This approach assumes that carbonatedissolution occurs congruently, and/or that dissolved Ca,Sr and Mg behave conservatively during transport. How-ever, some studies document significant excesses in riversof dissolved Mg and Sr relative to Ca compared to thechemical composition of limestones present in the basin(Kotarba et al., 1981; Sarin et al., 1989; Galy et al., 1999;Jacobson et al., 2002). Most of these studies suggest thatan excess of dissolved Mg and Sr may result from therecrystallization of calcite or gypsum from saturated sur-face waters. In such environments, dissolved Ca displaysnon-conservative behaviour. Consequently, the chemicalcompositions of rivers that drain carbonates can be signif-icantly different from the source rocks. Similarly, as a resultof weathering incongruency, the geochemical compositionof rivers draining silicates does not reflect the compositionsof rocks drained at the basin scale. In mixed lithology ba-sins it is therefore difficult to identify signatures inducedby silicate weathering alone.

Over the last few years, new potential isotopic tracers ofsilicate weathering, generally based on elements that arehighly enriched in silicate rocks, such as Li, Mg and Si,have been developed (e.g. Huh et al., 1998; Georg et al.,2006). Also, studies of small monolithological basins havebeen undertaken (e.g. Louvat and Allegre, 1997; Ziegleret al., 2005). New generation inductively-coupled plasmamass spectrometers (ICP-MS) have allowed the develop-ment of precise isotopic measurements for various elements(e.g. Luais et al., 1997; Marechal et al., 1999; Halicz et al.,1999; Rehkamper and Halliday, 1999; Galy et al., 2001;Rouxel et al., 2002; Beard et al., 2003; Cardinal et al.,2003; Wombacher et al., 2003). This has provided a newmeans of studying geological and environmentalprocesses.

Magnesium isotope studies of terrestrial materials arestill at an early stage (Galy et al., 2002; Chang et al.,2003; Carder et al., 2004; De Villiers et al., 2005; Tipperet al., 2006a,b). Galy et al. (2002) and Tipper et al.(2006a) reported significant differences in the Mg isotopiccompositions of silicate (�0.6 to 0&) and carbonate rocks(�4.5& to �1.1&), highlighting the strong potential ofMg isotopes as a lithological source tracer. However, untilnow, few data have been available for river waters, andthe existing data are mainly derived from large riversdraining mixed lithology basins (Tipper et al., 2006a,b).In this study, we test the potential of Mg isotopes to tracelithological sources and weathering processes by analyzingriver waters, rocks and soils of the Moselle River basin(Northeastern France). This small basin is remarkablefor its well-constrained and varied lithology, with silicaterocks upstream and carbonate/evaporite rocks down-stream. We have analyzed major and trace elements andSr and Mg isotopic compositions for the dissolved phasesof the main rivers. Soils and parent rocks, representativeof the main lithologies drained by these rivers, were alsoanalyzed.

2. HYDROLOGY AND GEOLOGY

The Moselle River basin (surface area = 3080 km2) is lo-cated in northeastern France (Fig. 1). The Moselle Riverflows from its source in the Vosges mountains to its outletat Pont Saint Vincent, 150 km downstream. The basin dis-plays strong contrasts in lithology and topography on asmall scale. Upstream of the city of Epinal, silicate rocksform the Vosges mountains (at altitudes between 1300and 400 m), whereas downstream of Epinal, a carbonateplatform corresponds to the Lorraine plateau (400–200 m)(Fig. 1). Mean annual rainfall varies from 1730 mm in theVosges mountains down to 1000 mm on the Lorraine pla-teau. The mean annual temperature is 6 �C in the Vosgesand 10 �C on the Lorraine plateau, with an annual ampli-tude of �16 �C. Forests are dominant in the Vosges moun-tains, while on the Lorraine plateau agricultural soilspredominate.

Near its source, the Moselle River flows over gneisses,granites, microgranites and greywacke schists of theVosges mountains. Based on chemical composition andmineral abundances, Nedeltcheva et al. (2006) havesubdivided these granitoids into three principal types (1)Type I: K-Feldspar and muscovite granites; (2) Type II:Plagioclase predominating over K-feldspar and muscovitegranite (main granitoid type of the sub-basin ending atMaxonchamp (Fig. 1), representing 18% of the surfacearea); (3) Type III: Biotite and hornblende bearing gran-ites (main granitoid type of the Moselotte sub-basin, rep-resenting 45% of the surface area). This part of the basinis characterized by low substratum permeability, andgroundwater systems are limited to Quaternary morainesand granitic sands (Dadi, 1991). Thus, most of the precip-itation water is transferred to the Moselle River by surfacerunoff (Jung, 1927; Gagny, 1959; Dadi, 1991). In contrast,the Permian and Buntsandstein sandstones located in thecentral part of the basin correspond to higher permeabilitylayers and to the main aquifer system of the area, and in-puts to the Moselle River mainly occur by infiltration(Periaux, 1961; Dadi, 1991). In the silicate part of the ba-sin, the main tributaries of the Moselle are the MoselotteRiver (�18% of the Moselle discharge at Epinal), whichflows mainly over granites, and the Vologne River(�33% of the Moselle discharge at Epinal), which alsodrains some sandstones. The diversity of silicate rocks isentirely integrated by the Moselle River at the city ofEpinal, the location of the major lithological boundary.Downstream of Epinal, the Moselle flows over carbonateand evaporite formations of the Lorraine plateau, whichcan be divided into three sedimentary units: (1) theMuschelkalk formation, comprising marls, dolomites,limestones and some gypsum deposits; (2) the Keuper for-mation, comprising marls with gypsum and anhydrite lay-ers, and (3) the Lias and Bajocian formations with marlsand limestones. The Madon River, the main Moselle trib-utary downstream Epinal, drains marls, limestones andsome evaporites (Fig. 1). The other tributaries are theDurbion, which mostly drains the Muschelkalk formation,and the Euron, draining mainly evaporite layers from thelower Keuper formation, as reflected in its specific and

Fig. 1. Geology, sampling points and river flow monitoring stations of the Moselle River basin (Nancy location: 48.42N; 06.12E).

5072 A. Brenot et al. / Geochimica et Cosmochimica Acta 72 (2008) 5070–5089

constant S isotope signature over the course of the year(Brenot et al., 2007).

River flows are recorded continuously at some gaugingstations and are available from the ‘‘Banque HYDRO”,databasis of the DIREN: French Regional Directory inEnvironment (http://www.hydro.rnde.tm.fr) (see monitor-ing stations on Fig. 1). At Epinal, the annual discharge ofthe Moselle River is 39 m3/s, which represents 70% of theannual river discharge of the Moselle at Pont Saint Vincent,located downstream in the carbonate plateau (Fig. 1). Thecontribution of the Durbion and Euron (0.95 m3/s) to the

Moselle in terms of water masses is relatively small com-pared to the Madon (10.7 m3/s).

3. MATERIALS AND METHODS

3.1. Sampling

The Moselle River and its main tributaries were sampled10 times, all points for each sampling within a single day,between February 1999 and January 2004. The samplingstrategy was mainly based on river flow and lithological

Magnesium isotope systematics on the Moselle river basin 5073

diversity, and samples were collected upstream and down-stream of the confluence of the main tributaries (Fig. 1).Whenever possible, the sampling points were located closeto Water Agency river flow monitoring stations. Riverwater was sampled from bridges, in the middle of eachchannel, and filtered through a 0.20 lm nylon Milliporemembrane and stored at 4 �C prior to chemical and isotopeanalyses. Samples for cation concentration and isotopemeasurements were acidified with distilled HNO3. pH val-ues were measured directly in the field.

Soils and parent rocks representative of the main lithol-ogies (granite, limestone, dolostone, gypsum evaporite)drained by the rivers were also sampled (Fig. 1). Leachingexperiments were performed on a few sedimentary rocks,in order to dissolve the calcium carbonate phase only.12 ml of 0.3 M HCl for �20 mg of rock powder wasused, and leachates were separated by centrifugation, forCa/Mg and Mg isotope measurements.

3.2. Major and trace element analyses

Major cation concentrations were measured in acidifiedwater samples by ICP-AES IRIS Thermo Elemental at theSARM (French National Facilities, Nancy). Uncertaintieswere better than 2%. Sr content was analyzed by ICP-MSSCIEX/Perkin Elan 6000, following the procedure reportedin Yeghicheyan et al. (2001). Corresponding uncertaintieswere better than 10%. Accuracy and reproducibility weremonitored by repeat analyses of SLRS-4 reference material(international reference for river water, NRC-CNRC,Canada). Analyses of anion concentrations were performedat the LIMOS laboratory (Nancy). Cl�, NO3

� and SO42�

were measured by ion chromatography on non-acidifiedwater samples using a DIONEX TM series 4000I instru-ment with an AG-9HC/AS-9HC column and conductivitydetection associated with an anion self-regeneratingsuppressor. The eluent used was 9 mM Na2CO3 with a rateof 1.5 ml/min. Uncertainties were better than 0.5% for allthe anions measured. Alkalinity was determined usingcharge balance calculations.

Rock and soil samples were fused with LiBO2 and sub-sequently dissolved in diluted HNO3. Major elements wereanalyzed by ICP-AES as reported by Govindaraju andMevelle (1987). Related uncertainties were better than2%. Trace elements were analyzed by ICP-MS, followingthe method reported in Carignan et al. (2001). Uncertain-ties were better than 6% for Sr and better than 15% forTh. Accuracy and reproducibility were monitored by repeatanalyses of rock reference materials (Carignan et al., 2001).

3.3. Sr and Mg isotope analyses

For Sr and Mg isotopic analyses, between a few micro-liters (for rivers flowing on carbonates) and 60 ml (for riversflowing on silicates) were evaporated in Teflon� beakers.The residues were then dissolved in 1 N HCl and centri-fuged. The residue after centrifugation was treated withHF, evaporated and re-dissolved in 1 N HCl with boric aciduntil complete dissolution, prior to Sr or Mg chemicalseparation.

For Mg isotopic analysis, a few mg of powderedrocks and soils were dissolved with HNO3 and HF.After complete dissolution, the solution was evaporatedat 80 �C in a Teflon� beaker and the residue dissolvedin 10 N HCl.

3.3.1. Strontium isotope measurements

The standard chemical separation technique used Ei-chrom Sr-SpecTM resins (Horwitz et al., 1992), followinga method inspired by Pin and Bassin (1992). Proceduralblank levels were lower than 300 pg, which is negligible inrelation to total sample Sr and the precision of the isotopicmeasurements. The Sr isotopic compositions were mea-sured using a Finnigan Mat 262 thermal ionization massspectrometer either in static mode or in dynamic modeand normalized to 86Sr/88Sr = 0.1194. Analyses of the inter-national reference material NIST-NBS987 measurementsyielded 87Sr/86Sr = 0.710182 ± 0.000040 (2r, N = 20) instatic mode and 87Sr/86Sr = 0.710218 ± 0.000020 (2r,N = 4) in dynamic mode. Uncertainties for individual87Sr/86Sr measurements were �25 � 10�6 (2SD).

3.3.2. Magnesium isotope measurements

For Magnesium isotopic measurements by MC–ICP–MS (multiple-collector inductively coupled plasma massspectrometry), separation by ion chromatography is a pre-requisite in order to avoid interferences and matrix effects.Galy et al. (2001) showed that the presence of Ca, Al andNa can induce an important instrumental isotopic fraction-ation of more than 1& relative to a pure Mg solution. Sev-eral authors (Galy et al., 2002; Chang et al., 2003) havedeveloped a technique for Mg separation with a yield of100%, and which is suitable for high precision analysis ofMg isotopes by MC–ICP–MS. Following the same princi-ple, a method modified from James and Palmer (2000),using an 8.5 cm tall bed of AG50W-X12 cation exchangeresin, was developed. After an initial pass through the col-umn, Mg was completely separated from Na and Ca, and atthis stage, the Mg fraction contains �15% of the K, 1.5% ofthe Al and �30% of the Mn of the sample. Water samples,dissolved in 0.5 ml 1 N HCl, were passed twice through thiscation exchange resin. Magnesium fractions eluted fromriver water samples consistently had K/Mg ratios lowerthan 0.025, sufficient to rule-out any possible matrix effect(Galy et al., 2001). Solid samples, dissolved in 10 N HCl,were passed first through an anion exchange resin (AG1X8) for eliminating Fe, and then six times through the cat-ion exchange resin. The total procedural blank, after sepa-ration by ion chromatography and measured by ICP-AES,was between 4 and 22 ng, representing a maximum Mg con-tribution of 0.09%.

We verified that our separation technique did not induceany significant isotopic fractionation by using syntheticsolutions with compositions typical of silicate-drainingand carbonate-draining river waters (SMS and SMC,respectively) (see Table 2 for corresponding compositions).Magnesium was taken from the NIST SRM980CRPG refer-ence solution (an internal isotope reference solution obtainedat CRPG by dissolution of the NIST SRM980 referencematerial). In addition, we analyzed four geological reference

Table 1MC–ICP–MS operating conditions during Mg isotope measure-ments in hard extraction

Parameters

Special facilitiesTeflon� nebuliser Flow rate = 100 ll/mnCyclonic chamber

Inductively coupled plasmaRF power 1350Plasma Ar flow rate 13.6 l/mnIntermediate Ar flow rate 1 l/mn

Mass spectrometerSampler cone NickelSkimmer cone NickelHexapole collision cell He collision gas; flow rate = 10 ml/

mnData aquisition

parametersFaraday collectors

usedL3: tailing peak of 23Na; L2: 24Mg;Axial: 25Mg; H2: 26Mg; H3: 27Al

One-peak zerosprocedure

120 s acid blank acquisitionsubtracted on line

Standard-Sampleacquisition

25 cycles of 10 s each;6–9 brackets/sample

Cleaning procedure HNO3 0.6 N for 60 sand HNO3 0.05 N for 5 min

Table 2d26Mg and d25Mg values relative to DSM3 and their associated uncertaintfully replicated samples

Sample Session

Reference material DSM3 1: Marc2: May3: Octob4: Decem5: Janua6: May

Camb-1 1: Marc3: Octob5: Janua6: May

SLRS-4 6: MayCal-S 7: JanuaBEN 6: MayDRN 8: Janua

SRM980CRPG All sessi

SRM980CRPG through chemistry SRM980CRPG pure 4: decemSMS-a* 4: DecemSMS-b* 4: DecemSMS-c* 4: DecemSMC** 1: Marc

River water full replicate AB0903-3a 5: JanuaAB0903-3b 5: JanuaAB0903-4a 5: JanuaAB0903-4b 6: May

*SMS Synthetic solution with NIST-SRM980CRPG, imitating the composCa = 3.9 (molar).**SMC Synthetic solution with NIST-SRM980CRPG, imitating the comK = 0.5: Ca = 5: Li = 0.5 Sr = 0.2 (molar).

5074 A. Brenot et al. / Geochimica et Cosmochimica Acta 72 (2008) 5070–5089

materials: a river water (SLRS-4, NRC-CNRC river waterreference, Canada), a carbonate rock (CAL-S, CRPG refer-ence limestone, Yeghicheyan et al., 2003), a basalt (BEN)and a diorite (DRN), both reference rocks from Geostan-dards (http://www.crpg.cnrs-nancy.fr/Geostandards/).CAL-S and DRN are from the Lorraine plateau and theVosges mountains, respectively.

Magnesium isotope ratios were measured using an Isop-robe MC–ICP–MS (ex-Micromass, now IsotopX) atCRPG. Instrument operating conditions and set-up are de-scribed in Table 1. A standard-sample bracketing techniquewas used to correct for instrumental mass bias (Galy et al.,2001). 27Al was monitored during measurement in order toverify that no aluminium was present in the Mg fraction, aswell as mass 23.6. All Magnesium fractions were dissolvedin 0.05 N HNO3 and adjusted to match (within ±10%) theMg content of the bracketing reference solution (3.5 ppm).Acid blank signal intensity for 26Mg, 25Mg and 24Mg neverexceeded 1% of the sample signals and were systematicallydistributed.

Since the NIST SRM980 reference material has beenshown to be heterogeneous (Galy et al., 2003; Carignanet al., 2004), all results are reported relative to DSM3(Dead Sea Magnesium, a reference material distributedby Cambridge University) in the conventional deltanotation:

ies (i.e. twice the standard deviation, 2r) for reference materials and

d26Mg 2r d25Mg 2r N

h 2004 0,0 0,09 0,0 0,08 22004 �0,1 0,12 0,0 0,14 3er 2004 0,0 0,13 �0,1 0,28 3ber 2004 0,1 — 0,0 — 1

ry 2005 0,1 0,13 0,0 0,11 42005 0,0 — 0,1 — 1h 2004 �2,8 — �1,4 — 1er 2004 �2,7 0,41 �1,4 0,27 3ry 2005 �2,7 0,14 �1,3 0,01 22005 �2,7 0,05 �1,4 0,13 62005 �1,0 0,25 �0,5 0,15 3ry 2006 �4,3 0,18 �2,2 0,16 82005 �0,1 0,24 0,0 0,07 3ry 2007 �0,5 0,04 �0,3 0,04 4

ons �4,0 0,14 �2,1 0,05 18

ber 2004 �4,0 0,40 �2,1 0,14 2ber 2004 �3,8 0,12 �2,0 0,03 2ber 2004 �4,1 0,30 �2,0 0,06 3ber 2004 �3,9 0,07 �1,9 0,03 3

h 2004 �4,1 0,41 �2,0 0,31 2

ry 2005 �1,0 0,23 �0,5 0,25 3ry 2005 �1,2 0,26 �0,6 0,15 4ry 2005 �1,0 0,05 �0,5 0,07 22005 �0,8 0,15 �0,4 0,13 3

ition of typical river water on silicates: Mg = 1: Na = 7 : K = 1.1:

position of typical river water on carbonates: Mg = 1: Na = 5:

Magnesium isotope systematics on the Moselle river basin 5075

δ = −

Mg

Mg

Mg

Mg

Mg

sample

DSM 3

1 100 ð1Þ

δ = −

Mg

Mg

Mg

Mg

Mg

sample

DSM 3

1001 ð2Þ

Delta values for the NIST SRM980CRPG reference solutionare d26MgDSM3 = �4.00& (±0.14&) and d25MgDSM3=�2.05& (±0.05&), based on 18 measurements of DSM3.

Errors for d26Mg and d25Mg are given as twice the standarddeviation (2SD) throughout the text. 2SD were 0.14& and0.10& during d26Mg measurements of DSM3 and Camb-1(n = 4), respectively (not passed through the chemistry).External uncertainty, estimated using pure and dopedSRM passed through the chemistry, was 0.22& (2SD) ford26Mg and 0.14& (2SD) for d25Mg (Table 2).

Delta values for the Cambridge-1 pure Mg solution(internal reference material of Cambridge University) wered26MgDSM3 = �2.73& (±0.10&, 2r) and d25MgDSM3 =�1.38& (±0.07&) (Table 2), in close agreement withpublished values (d26MgDSM3 = �2.60 ± 0.14&, �2.57 ±0.13& and �2.58 ± 0.14&, obtained by Tipper et al.(2006a); Pearson et al. (2006) and Galy et al. (2003), respec-

Fig. 2. Magnesium three-isotope plot showing the 62 samples analyzedriver waters, soils and rocks) and their associated 2r errors. The slopcalculated using Isoplot linear regression of Ludwig (2001); All measuremDSM3.

tively). d26MgDSM3 measured for SLRS-4 is �0.98 ±0.25&, for Cal–S is �4.31 ± 0.18& and for BEN is�0.14 ± 0.24&. These values are all consistent with valuesgiven in Wombacher et al. (2006) (�0.97 ± 0.05&,�4.38 ± 0.31& and �0.41 ± 0.19&, respectively).

Linear regression of all the data in a d25Mg0 vs. d26Mg0

diagram (62 samples, Fig. 2), taking into account the uncer-tainty for each sample, yields a slope of 0.509 ± 0.023, inagreement with theoretical and published values (e.g.0.518 ± 0.038 from Galy et al., 2001; 0.5165 ± 0.0005 fro-mYoung and Galy, 2004).

4. RESULTS

4.1. Dissolved elements

4.1.1. Concentrations in rivers upstream of Epinal (silicate

area)

In the silicate part of the basin, river pH ranges between6.2 and 7.4. Concentrations of dissolved elements in theMoselle River and its main tributaries (Moselotte andVologne Rivers) range between 30 and 103 lmol/l forMg, 19 and 90 lmol/l for K, 57 and 523 lmol/l for Siand 0.15 and 0.53 lmol/l for Sr (Table 3, Brenot et al.,2007), with alkalinities of less than 850 lmol/l. Two smallisolated streams of an upper catchment (Grosse Pierreand Pont Martin, sampling points #21 and #22) yield con-centrations of Mg (27 and 34 lmol/l), K (14 and 7 lmol/l)and Si (69 lmol/l) that are within the ranges of the largerrivers (Table 3, Brenot et al., 2007). Concentrations of

(SRM980CRPG through chemistry, Camb-1 (Cambridge-1), Cal-S,e of the regression line and associated error (0.509 ± 0.023) wereents were made relative to SRM980CRPG, and then renormalized to

ble 3eochemical analyses for the Moselle River and tributaries sampled between 1998 and 2003

o. River Location Distance to thesource (km)

Surfacedrained(km2)

Collectiondate

Cl(lmol/l)

SO4

(lmol/l)NO3

(lmol/l)Alk(lmol/l)

Ca(lmol/l)

Mg(lmol/l)

K(lmol/l)

Na(lmol/l)

Si(lmol/l)

Maxomchamp 31 153 26/02/03 255 49 59 280 151 38 25 290 7813/06/03 398 67 77 373 206 50 34 437 9625/06/03 298 55 48 711 222 55 90 523 90

Epinal 73 1220 08/07/98 409 — — — 202 62 56 534 115?/02/01 429 — — — 110 42 33 368 6326/02/03 330 75 74 376 181 59 44 407 10713/06/03 411 79 69 525 228 67 56 517 10225/06/03 484 114 49 657 265 77 70 664 108

Girmont 87 26/02/03 351 100 79 495 240 70 50 457 10525/06/03 632 444 38 1172 620 139 93 1119 129

Portieux 100 08/07/98 508 — — — 498 144 77 854 101?/02/01 460 — — — 211 74 83 405 6926/02/03 345 163 84 898 428 141 51 463 10225/06/03 614 455 33 1515 788 233 90 941 110

Bayon 122 08/07/98 516 — — — 601 224 63 777 100?/02/01 437 — — — 279 106 71 420 11126/02/03 358 224 85 1018 516 175 52 477 9125/06/03 667 611 — — 952 326 86 933 91

Messein 144 08/07/02 511 — — — 861 410 66 750 82?/02/01 513 — — — 373 160 69 458 9126/02/03 379 480 96 1486 825 345 55 527 10025/06/03 651 1142 — — 1356 603 87 893 85

Pont StVincent

148,3 3080 08/07/98 733 — — — 1588 799 96 888 75

?/02/01 530 — — — 627 252 71 457 8826/02/03 749 1670 200 5800 3371 1263 71 750 9725/06/03 913 2281 18 12270 5249 2751 129 1634 176

Le PontMartin

Vosges 06/12/04 33 8 6 112 29 27 14 42 69

Grossepierre

Vosges ?/05/01 — — — — 32 34 7 27 —

Moselotte St Ame 38 469 08/07/98 688 — — — 115 48 67 964 12326/02/03 364 53 48 286 131 54 34 399 9713/06/03 452 66 52 553 150 61 71 696 9825/06/03 619 67 36 691 159 66 87 941 106

Vologne Jarmenil 61 355 08/07/98 465 — — — 241 79 63 636 12326/02/03 346 79 88 441 202 69 50 442 13213/06/03 494 85 66 800 306 83 62 690 12425/06/03 500 88 60 813 313 92 71 668 124

5076A

.B

reno

tet

al./G

eoch

imica

etC

osm

och

imica

Acta

72(2008)

5070–5089

TaG

N

4

7

9

10

11

21

22

2

3

6D

urb

ion

Pal

legn

ey93

26/0

2/03

271

1523

130

4800

2801

1170

6723

712

425

/06/

0331

023

4928

4874

3387

1364

8632

012

58

Eu

ron

Fro

vill

e10

891

,926

/02/

0349

125

8429

112

558

5719

3259

103

450

151

25/0

6/03

613

5625

155

1426

278

4748

4713

775

524

012

Mad

on

Po

nt

St

Vin

cen

t14

894

008

/07/

9812

89—

——

4242

2335

117

1233

85

?/02

/01

598

——

—19

9365

371

444

8726

/02/

0386

820

1824

977

5243

4916

6174

813

9525

/06/

0316

5045

24—

—27

4713

6811

212

1311

913

Mad

on

Har

ou

e26

/02/

0310

3724

7422

583

5846

7520

7475

995

9825

/06/

0318

2943

7968

9033

5789

3043

125

1898

198

17G

itte

Tat

ign

eco

urt

15/0

3/03

329

1045

135

8489

3491

1841

6831

163

14/0

5/03

318

1255

9493

2338

1020

6291

409

115

30/0

9/03

438

4223

119

5265

4482

2364

9847

772

25/1

1/03

483

3314

408

3057

3725

1391

9924

911

1

Un

cert

ain

ties

wer

eb

ette

rth

an0.

5%fo

ran

ion

san

db

ette

rth

an2%

for

maj

or

cati

on

s.<

L.D

.L

ow

erth

anth

ed

etec

tio

nli

mit

.

Magnesium isotope systematics on the Moselle river basin 5077

dissolved Ca in this two streams (29 and 32 lmol/l) are low-er than in the Moselle River (72–356 lmol/l) but the rangedocumented for dissolved Ca in small streams in the Vosgesmountains is relatively large (28–117 lmol/l; Probst et al.,2000; Nedeltcheva et al., 2006). In contrast with other ma-jor elements, concentrations of Na (159–964 lmol/l) and Cl(112–688 lmol/l) are much higher in the Moselle River andits tributaries than in the two streams studied (42 and27 lmol/l for Na; 33 lmol/l for Cl), or in other Vosgianstreams documented by Probst et al. (2000) and Nedeltchevaet al. (2006). This strongly suggests that significant Cl andNa are acquired from anthropogenic inputs, most probablyfrom NaCl road-salts. Indeed, the total amount of Cl andNa spread per year as road-salt in the basin would represent�80% and �40% of the annual flows of Cl and Na, respec-tively in the Moselle at Epinal. Thus Cl concentrations can-not be used in this area to correct the dissolved river loadfor atmospheric inputs (Stallard, 1980; Stallard and Ed-mond, 1981; Meybeck, 1983; Negrel et al., 1993). The atmo-spheric contribution was therefore calculated using thecomposition of mean open field precipitation sampled inthe Vosges mountains (Cl = 15 lmol/l; Probst et al.,2000), assuming 38% evapotranspiration (calculated forthe silicate part of the basin using Meteo France data andBenichou and Le Breton (1987)). According to these calcu-lations, atmospheric inputs contribute less than 15% of thedissolved Mg and less than 20% of the dissolved Sr in riverwater.

4.1.2. Concentration in rivers of the Lorraine plateau

(carbonate dominated area)

Alkalinities of rivers draining the Lorraine plateau aregreater than for the silicate area, and range between 3300and 11000 lmol/l, with pH ranging between 7.2 and 8.5. Tak-ing into account the local composition of mean open fieldprecipitation, atmospheric inputs were calculated to repre-sent less than 10% of dissolved Mg and less than 6% of dis-solved Sr in river water. All samples were collected fromacross the entire watershed within a single day, allowing fora direct comparison of the dissolved concentrations. An in-crease in the concentrations of Ca (193–2584 lmol/l), Mg(60–1214 lmol/l), and Sr (0.3–17.7 lmol/l) in the MoselleRiver between Epinal and Pont Saint Vincent (the outlet ofthe studied area) is systematically observed for all samplingdays, and can be related to the dissolution of carbonatesand evaporites present on the Lorraine plateau (Brenotet al., 2007). Indeed, tributaries (Durbion, Euron, Gitte,Madon) draining mainly sedimentary units of the Lorraineplateau display very high concentrations of SO4

2� (561–9718 lmol/l), Ca (1754–7847 lmol/l), Mg (653–5039 lmol/l), and Sr (5–94 lmol/l) (Table 3, Brenot et al., 2007). Cl con-tents (318–3045 lmol/l) are also significant, especially for theMadon tributary where dissolution of halite, present in thelocal lithology, is a relatively important source of Na andCl (Brenot et al., 2007). For a sampling campaign in February2002 of the Madon River along its course, the Ca, Mg and Srcontents of the Madon River also increased significantlydownstream, from 21 to 4175 lmol/l for Ca, from 30 to2023 lmol/l for Mg and from 0.1 to 27 lmol/l for Sr. TheMadon River chemistry before its confluence with the

5078 A. Brenot et al. / Geochimica et Cosmochimica Acta 72 (2008) 5070–5089

Moselle (at Pont Saint Vincent) seems therefore to bebuffered by weathering of carbonates and evaporites,when compared with the chemical signal inherited fromBuntsandstein sandstones upstream.

4.1.3. Sr and Mg isotopic compositions

The section of the Moselle River and its tributaries thatflow over silicate rocks yield radiogenic Sr isotopic compo-sitions, with 87Sr/86Sr ratios ranging between 0.7128 and0.7174. The two streams from the upper catchment (GrossePierre and Pont Martin) yield more radiogenic composi-tions (0.72080 and 0.72489). The Sr isotopic compositionof the Moselle River is less radiogenic downstream ofEpinal (down to 0.70824), where the river also drainscarbonates, marls and evaporites of the Lorraine plateau.All tributaries draining only these formations (Durbion,Euron, Gitte, Madon) have low 87Sr/86Sr (from 0.70815to 0.70862).

Thirty four water samples from the Moselle River basinwere analyzed for Mg isotopic composition (Table 4). Anoverall variation of 0.9& in d26Mg is found, representing�4.1 times the estimated external uncertainty (Section3.3.2). Magnesium dissolved in the Moselle River and trib-utaries in the silicate part of the basin display d26Mg valueshigher than �0.9& except for two values (d26Mg values in-cluded in the range �1.2& to �0.7&). The streams with themost radiogenic Sr isotopic compositions also yield thehighest d26Mg values (d26Mg = �0.7&, Table 4). For com-parison, other small basaltic and granitic rivers also displayd26Mg greater than �0.9&, ranging between �0.86& and�0.31& (Tipper et al., 2006b). De Villiers et al. (2005) re-ported highly depleted d26Mg, between �3.8& and�2.0&. for rivers draining granite, sandstone and lava inSwaziland (d26Mg values were recalculated relative toDSM3, based on reported seawater analyses in this studyand seawater value (d26MgDSM3 = �0.8&) published inYoung and Galy (2004) and Carder et al. (2004)). AtEpinal, the Moselle River d26Mg is constant throughoutthe year (�0.8&). From Epinal to Pont Saint Vincent, onthe Lorraine carbonate plateau, the d26Mg of the MoselleRiver ranges from �0.8& to �1.4&. For comparison,rivers draining Himalayan mixed lithology catchments dis-play d26Mg values ranging between �1.8& and �0.8&

(Tipper et al., 2006a). In larger rivers, d26Mg values rangebetween �1.7& (for the Mackenzie) and �0.5& (for theNile) (Tipper et al., 2006b).

The Madon River, draining carbonates and evaporites,displays relatively constant and homogeneous d26Mg, rang-ing between �1.6& and �1.4&. The Euron, drainingmainly evaporites, similarly displays low and relatively con-stant d26Mg (from �1.4& to �1.2&). All tributaries drain-ing carbonate and evaporite formations display d26Mgvalues lower than �1.1&.

4.2. Rocks, soils and leachates

Chemical and Mg isotopic compositions were deter-mined for a Vosgian diorite (DRN), sedimentary rocks,and associated soil samples (Table 5). Soils developed onsilicates have greater LOI (loss on ignition) values (3.9–

30.0%) than their related parent rocks (0.9–1.5%). SinceCa contents are broadly the same in rocks and soils (Table5), this suggests enrichment in secondary phyllosilicate min-eral phases in the soils. Mg/Al ratios are broadly similar insilicate rocks and soils (�0.01–0.9).

Soils developed on limestones display a strong depletionin carbonate phases as indicated by significantly lower Cacontents (CaO = 15.2% and 1.0% for two soil samples,and 50.4% and 53.0% for their respective parent rocks)(Table 5). These soils are also greatly enriched in silica(SiO2 = 63.7 and 62.8%) relative to their parent limestones(SiO2 = 5.3% and 2.8%). This indicates that limestones con-tain some silicate phases, mainly phyllosilicate minerals,and that, during weathering, the calcium carbonate phaseis preferentially dissolved. In contrast, Mg contents arebroadly similar in source limestones (MgO = 1.2% and0.6%) and their associated soils (MgO = 1.7 and 1.1%).Ca/Mg ratios are also similar in sedimentary rocks and inacid leachates (not shown).

The DRN Vosgian diorite displays a d26Mg of �0.53&,close to published values for a High Himalayan paragneissand a loess (�0.42& and �0.6&, respectively; Young andGaly, 2004; Tipper et al., 2006a). One gypsum sample wasanalyzed, for which d26Mg is �0.8&, identical to the pres-ent-day seawater value of d26Mg = �0.8& (Young andGaly, 2004; Carder et al., 2004; De Villiers et al., 2005).d26Mg is 0& for the soil developed on granites. This is with-in the range obtained for Himalayan soils (�0.11& to0.02&, Tipper et al., 2006a).

Two limestones, both characterized by low Si, Mg andAl concentrations, display significantly different d26Mg val-ues (�4.5& and �1.0&). The dolomitic limestone (R4,12.4% MgO, Table 5) yields d26Mg of �1.4&. These valuesare within the published range for similar rocks (Galy et al.,2002; Carder et al., 2004; De Villiers et al., 2005). Soilsdeveloped on limestones are enriched in heavy isotopes rel-ative to their source rocks, with d26Mg of 0.6& and �1.3&.Acid leachates, aimed at preferentially dissolvingcarbonates from the limestones and the dolomitic lime-stones, display d26Mg similar, within analytical uncertainty,to the corresponding rocks (Table 5).

5. DISCUSSION

5.1. Silicate Vosgian rivers

In the silicate (Vosgian) part of the basin, river watersare characterized by 87Sr/86Sr ratios (0.7128–0.7249) thatare intermediate between values for mean local precipita-tion (0.7102 and 0.7128; Probst et al., 2000; Chabauxet al., 2005) and values for mineral separates and granitesfrom the Vosges mountains (0.71612–5.865, Bonhomme,1967; France-Lanord, 1982; Probst et al., 2000; Aubertet al., 2001; Aubert et al., 2004). Fig. 3 is a plot of Sr isoto-pic compositions vs. Mg/Sr molar ratio for rivers flowingover silicates compared to values reported for mineral sep-arates, bulk bedrock and initial 87Sr/86Sr ratio of localgranites. As described in Section 4.1.1, atmospheric inputsrepresent a maximum of 20% of the dissolved Sr in theserivers. The dominant input of Sr in the silicate part of the

Table 4Sr concentrations (Uncertainties 10%) and Sr and isotopic composition for the Moselle River and tributaries sampled between 1998 and 2004

No. River Location Collection date d26Mg (&) 2r d25Mg (&) 2r N Sr (&) 87Sr/86Sr 2r/10�6

1 Moselle Maxomchamp 11/03/02** — — — — — 0,19 0,71349 1318/09/02** — — — — — 0,31 0,71372 3310/01/03** �0,7 0,3 �0,4 0,3 3 0,20 0,71320 1104/05/03** �0,9 0,3 �0,4 0,2 2 0,26 0,71275 2013/06/03 — — — — — 0,30 0,71343 2802/09/03** �0,9 0,3 �0,4 0,2 3 0,33 0,71372 13

4 Epinal 08/07/98 — — — — — 0,43 0,71504 11911/03/02** — — — — — 0,25 0,71528 2318/09/02** — — — — — 0,38 0,71456 1610/01/03** �0,8 0,3 �0,5 0,2 2 0,23 0,71521 1304/05/03** �0,8 0,4 �0,4 0,4 3 0,31 0,71483 4413/06/03 — — — — — 0,37 0,71459 1702/09/03** �0,8 0,3 �0,4 0,2 3 0,44 0,71451 17

5 Girmont 12/03/02** — — — — — 0,33 0,71374 2818/09/02** — — — — - 0,64 0,71123 2710/01/03** — — — — — 0,28 0,71425 2004/05/03** — — — — — 0,50 0,71233 4302/09/03** — — — — — 0,83 0,71176 25

Portieux 18/09/02** �0,9 0,3 �0,4 0,2 2 1,17 0,71020 2210/01/03** �0,8 0,3 �0,5 0,2 2 0,44 0,71191 1704/05/03** �1,1 0,3 �0,5 0,2 2 1,09 0,71030 2802/09/03** �1,4 0,3 �0,7 0,2 2 1,52 0,70999 11

9 Bayon 08/07/98 �1,3 0,3 �0,7 0,2 2 3,05 0,7091818/09/02** — — — — — 2,75 0,70915 2910/01/03** — — — — — 1,17 0,70959 804/05/03** — — — — — 2,11 0,70920 1502/09/03** — — — — — 2,91 0,70917 39

10 Messein 08/07/02 �1,4 0,3 �0,7 0,2 2 7,06 0,70868 21?/02/01 — — — — — 2,49 0,70940 18612/03/02** — — — — — 3,54 0,70891 1610/01/03** �1,0 0,3 �0,5 0,2 2 2,14 0,70909 1604/05/03** �1,3 0,3 �0,7 0,2 3 4,69 0,70867 2302/09/03** �1,3 0,3 �0,7 0,2 2 6,37 0,70864 18

11 Pont St Vincent 08/07/98 — — — — — 14,37 0,70840 14?/02/01 — — — — — 3,86 0,70894 3612/03/02** — — — — — 3,54 0,70891 3318/09/02** — — — — — 14,36 0,70843 1410/01/03** — — — — — 8,16 0,70845 3104/05/03** — — — — — 36,74 0,70824 2602/09/03** — — — — — 23,24 0,70830 22

21 Le Pont Martin Vosges 06/12/04*** �0,7 0,3 �0,4 0,2 2 0,10 0,72080 1122 Grosse pierre Vosges ?/05/01 �0,7 0,4 �0,3 0,3 2 0,08 0,72495 20

St Ame 11/03/02** — — — — — 0,16 0,71737 1518/09/02** — — — — — 0,26 0,71655 1510/01/03** �0,8 0,3 �0,4 0,2 2 0,15 0,71705 1104/05/03** — — — — — 0,24 0,71586 3613/06/03 — — — — — 0,25 0,71665 3002/09/03** �0,7 0,3 �0,3 0,2 3 0,28 0,71695 9

Jarmenil 18/09/02** — — — — — 0,38 0,71454 2610/01/03** �0,9 0,3 �0,5 0,2 3 0,28 0,71539 1204/05/03** �1,1 0,3 �0,6 0,2 3 0,42 0,71387 1713/06/03 — — — — — 0,43 0,71424 2502/09/03** �1,2 0,3 �0,6 0,2 4 0,53 0,71403 18

6 Durbion Pallegney 12/03/02** — — — — — 7,60 0,70844 2418/09/02** — — — — — 13,01 0,70824 2110/01/03** — — — — — 4,55 0,70857 1104/05/03** — — — — — 15,84 0,70820 1602/09/03** — — — — — 18,98 0,70815 14

8 Euron Froville 12/03/02** �1,3 0,3 �0,6 0,2 1 46,70 0,70825 2618/09/02** — — — — — 79,00 0,70829 2110/01/03** �1,4 0,3 �0,7 0,4 3 32,22 0,70827 1104/05/03** �1,2 0,4 �0,6 0,2 3 71,56 0,70819 9

(continued on next page)

Magnesium isotope systematics on the Moselle river basin 5079

Table 4 (continued)

No. River Location Collection date d26Mg (&) 2r d25Mg (&) 2r N Sr (&) 87Sr/86Sr 2r/10�6

12 Madon 08/07/98 — — — — — 41,17 0,70827 27Pont St Vincent ?/02/01 — — — — — 8,17 0,70842 14

18/09/02** — — — — — 40,27 0,70826 926/02/03 — — — — — 20,02 0,70826 4404/05/03** — — — — — 39,02 0,70830 3025/06/03 — — — — — 18,58 0,70858 3502/09/03** �1,5 0,3 �0,7 0,2 3 47,02 0,70836 33

13 Madon Haroue 18/09/02** — — — — — 44,08 0,70826 1110/01/03** �1,4 0,3 �0,7 0,3 5 10,70 0,70843 604/05/03** �1,6 0,3 �0,7 0,2 2 42,99 0,70823 1902/09/03** �1,5 0,3 �0,7 0,2 2 55,83 0,70824 13

17 Gitte Tatignecourt 15/03/03 �1,3 0,3 �0,6 0,2 2 17,13 0,70830 1114/05/03 �1,1 0,3 �0,6 0,2 2 24,07 0,70835 2630/09/03 �1,2 0,3 �0,5 0,2 2 32,25 0,70829 725/11/03 �1,4 0,4 �0,7 0,3 4 15,07 0,70862 8

Major concentrations of these samples are reported in Table 3 and in Brenot et al. (2007). <L.D. Lower than the detection limit.** Major elements concentrations published in Brenot et al. (2007).

5080 A. Brenot et al. / Geochimica et Cosmochimica Acta 72 (2008) 5070–5089

Moselle River basin is therefore the weathering of silicaterocks. The positive correlation between 87Sr/86Sr and Mg/Sr ratios reflects preferential weathering of Mg-rich andradiogenic mineral phases such as biotite(87Sr/86Sr = 2.293–5.865, Bonhomme, 1967; Probst et al.,2000; Aubert et al., 2001) (Fig. 3). A similar correlation isobserved between 87Sr/86Sr and K/Sr, which also supportsthe preferential weathering of other mineral phases suchas K-feldspars (87Sr/86Sr = 0.7827–0.7975, Probst et al.,2000; Aubert et al., 2001). The range of 87Sr/86Sr shownby the silicate rivers of the Moselle basin might be best ex-plained by the disparities in silicate mineral compositionsand abundances within the sub-basins (Fig. 3). Neverthe-less, according to the literature, local precipitations are ex-pected to display intermediate Mg/Sr molar ratio and87Sr/86Sr values between local rocks and separated minerals(Fig. 3). Thus mixing line in Fig. 3 could also be partly dueto mixing with precipitation.

The magnesium isotopic compositions of Vosgian riv-ers (d26Mg = �1.2& to �0.7&) are significantly lowerthan the value estimated for the continental crust andVosgian diorite (� � 0.5& this study, Young and Galy,2004; Tipper et al., 2006a) (Fig. 4). In contrast, thed26Mg of silicate soil samples are enriched in heavy iso-topes, with values close to 0& for both the Moselleand Himalayan soils (Tipper et al., 2006a) (Fig. 4b). Thissuggests that Mg isotopes fractionate during silicate weather-ing, resulting in the preferential release of light isotopesinto the aqueous phase. This observed fractionation ofMg isotopes between source rocks and dissolved phasescan not be an ‘‘artifact” due to anthropogenic inputs ofdissolved Mg in river systems. Indeed the previous studyof Brenot et al. (2007) considered that only road salt con-tributions may alter the natural geochemical signature ofriver water in the silicate (Vosgian) part of the basin,where forests are dominant. Potential contributions ofMg from road salt leaching would represent less than4% of the annual flux of dissolved Mg in the Moselleat Epinal. Such a contribution would generate a maxi-mum decrease of 0.04& for d26Mg values in river water.

This is not significant compared to the difference of valuesobserved between source rocks and dissolved phases(d26Mgrock-water between �0.4& and �0.7&).

A number of different processes may be responsible forthe observed Mg isotopic fractionation between sourcerocks and dissolved phases, such as secondary mineral for-mation, mineral leaching, isotope exchange, uptake byplants, or organic litter decomposition. In the simplest case,it might be assumed that Mg isotopes do not fractionateduring mineral leaching, and that only clay formation sig-nificantly fractionates Mg isotopes. Following a Rayleighlaw, soil d26Mg (0&) associated with d26Mg in waters rang-ing between �1.1& and �0.7&, would be consistent withd26Mgclay-water isotope fractionation ranging between0.4& (for <40% of Mg uptake by clays) and 0.8& (for�60% of Mg uptake by clays). However, the Mg/Sr molarratios of the dissolved phases are broadly similar to theMg/Sr ratios measured in local rocks and soils. TheMg/Al ratios are also similar in silicate rocks and soils. Thissuggests that little Mg is involved in clay neoformation inthe Vosges. Moreover, in river waters, the highest 87Sr/86Srvalues are associated with the highest d26Mg values (Table4). Mg and Sr leaching from source rocks and mineralstherefore also appears to be a key process, leading to Mgisotope fractionation. Following a Rayleigh law, the corre-sponding d26Mgrock-water is estimated to range between�0.4& and �0.7& (with a corresponding fraction of Mgof less than �70% released in the aqueous phase). Riverwaters have d 26Mg lighter than host rocks. Black et al.(2006) have shown that plant material prefers light Mg iso-topes, thus uptake by plants is certainly not a dominantprocess explaining isotopic fractionation.

More data and experiments are needed in order todetermine and calibrate the processes that fractionateMg isotopes. However, overall, these data show thatd26Mg of silicate rocks cannot be used directly as anend-member for natural waters, and highlight the poten-tial of Mg isotopes as a tracer of processes such as clayformation, weathering intensity or recycling by plants ata watershed scale.

Table 5Major, trace element concentrations and Mg isotopic signatures of selected rocks and soils

No. Location Nature d26Mg(&)

2r d25Mg(&)

2r SiO2

(%)Al2O3

(%)Fe2O3

(%)MnO(%)

MgO(%)

CaO(%)

Na2O(%)

K2O(%)

TiO2

(%)P2O5

(%)LOI**

(%)Total(%)

Sr(ppm)

CIA*** (molarratio)

R1a Col de Grossepierre

Granite* 0,2 0,50,0 0,2 63,7 13,9 4,4 0,1 4,1 2,7 2,5 6,5 0,7 0,5 0,9 100,0 327 46,46

R1b Soil on granite 0,5 0,40,2 0,3 63,3 13,9 4,3 0,1 2,4 1,0 1,4 6,3 0,7 0,5 6,3 100,1 305 55,57R1c Soil on granite 0,4 0,30,1 0,2 59,6 14,3 5,6 0,1 4,9 2,8 1,7 5,8 0,9 0,6 3,8 100,3 283 50,18R2a Zainvillers Granite 0,5 0,30,5 0,2 70,4 15,4 2,1 <L.D. 0,6 0,4 1,5 7,5 0,4 0,2 1,5 99,9 147 57,84R2b Soil on granite 0,0 0,20,0 0,1 57,3 12,8 3,1 <L.D. 0,6 0,1 0,9 4,8 0,4 0,3 20,0 100,3 119 65,04R3a Adoncourt Limestone* �1,0 0,1�0,5 0,3 5,3 1,3 0,7 0,0 1,3 50,4 0,1 0,4 0,1 <L.D. 40,8 100,4 351 —R3a-leacheda �1,2 0,1�0,7 0,1R3b Marly soil 0,6 0,20,2 0,1 63,7 13,9 4,1 0,1 1,7 15,2 0,2 3,9 0,6 0,1 17,4 99,8 89 —R4 Girmont Dolomitic

limestone�1,4 0,2�0,7 0,1 16,3 5,1 3,3 0,4 12,4 25,2 0,1 1,7 0,2 0,1 35,6 100,4 138 —

R4-leacheda �1,3 0,3�0,7 0,1R5a Zincourt Limestone* �4,5 0,3�2,4 0,1 2,8 0,6 <L.D. <L.D. 0,7 53,1 <L.D. 0,3 <L.D. <L.D. 42,3 99,7 528 —R5b Clayey soil �1,3 0,3�0,6 0,4162,8 14,1 6,9 0,4 1,2 1,1 0,4 3,7 0,7 0,4 8,8 100,4 105 —R6 Circourt Gypsum �0,8 0,2�0,4 0,1 2,9 0,6 0,1 <L.D. 0,6 32,4 <L.D. 0,1 <L.D. <L.D. 60,7 1205 —R7 Bayon Clayey marl — — — 63,8 13,4 3,8 0,0 2,2 2,6 0,1 6,2 0,8 0,4 6,4 99,7 50 —R7-leacheda 0,4 0,20,3 0,3R8a Rozelieure Dolostone — — — 2,1 0,6 0,3 0,0 20,8 30,5 <L.D. 0,2 <L.D. <L.D. 45,7 100,1 101 —R8a-leacheda �1,5 0,2�0,7 0,2R8b Marly dolomitic

soil— — — 42,6 9,5 3,8 0,1 7,3 10,4 0,3 3,0 0,5 0,2 21,7 99,3 88 —

R8c Dolostoneweathered

— — — 0,7 0,2 0,5 0,0 21,3 30,3 <L.D. 0,1 <L.D. <L.D. 46,7 99,7 106 —

R8d Dolostoneweathered

— — — 3,7 1,2 0,6 0,1 20,3 28,5 0,1 0,3 <L.D. <L.D. 44,9 99,5 93 —

R8e Calcareousdolomitic soil

— — — 21,8 6,8 2,5 0,1 13,9 19,2 0,1 1,9 0,3 <L.D. 33,3 99,7 89 —

12 Madon at PontSaint Vincen

bed load sediments — — — 28,5 2,7 12,1 — 1,0 28,4 0,2 0,6 — — — — 449 —

Uncertainties were better than 2% for major elements and better than 6% for Sr. <L.D., Lower than the detection limit. Fe2O3 = 0.1; MnO = 0.03; Na2O = 0.15; TiO2 = 0.03; P2O5 = 0.03.* Samples submitted to full replicates.

** LOI, Lost in ignition, volatilized elements (mass percent) loss after heating at 980�C.*** CIA, chemical index of alteration using molecular proportions: CIA = [Al2O3/(Al2O3 + CaO + Na2O + K2O)] � 100.

a Leaching experiments consist in 12 ml 0.3 HCL for 20 mg rock powder and separation of leachates by centrifugation.

Magn

esium

isoto

pe

systematics

on

the

Mo

selleriver

basin

5081

Fig. 3. 87Sr/86Sr vs. Mg/Sr molar ratio for rivers flowing over silicates compared to values reported for mineral separates, bulk bedrock andinitial 87Sr/86Sr ratio of local granites (from ‘‘Granite des Cretes” and ‘‘Granite de Brezouard” documented by (a) Bonhomme (1967); (b)France-Lanord (1982); (c) Probst et al. (2000); (d) Aubert et al. (2001)). Biotite 87Sr/86Sr = 2.293–5.865; albite 87Sr/86Sr = 0.7420–0.7743;apatite 87Sr/86Sr = 0.71612; bulk granite 87Sr/86Sr = 0.7970–0.8384; initial 87Sr/86Sr = 0.7080–0.7137. Simulated silicate weathering end-member of Probst et al. (2000) = 0.73629–0.74203. Mean open field precipitation 87Sr/86Sr = 0.7102 and 0.7128, as documented by (c) Probstet al. (2000) and (e) Chabaux et al. (2005), respectively. 87Sr/86Sr of mean throughfall precipitation = 0.7124, as reported in (c) Probst et al.(2000).

5082 A. Brenot et al. / Geochimica et Cosmochimica Acta 72 (2008) 5070–5089

5.2. Mg excess in rivers draining carbonates

Rivers draining the Lorraine plateau formations displayhigh alkalinities and high Ca contents, mainly attributed tocarbonate weathering. However, Ca/Mg molar ratios (1.5–3.2) are significantly lower than those of local carbonates(29–59 for limestones, Fig. 5). Similarly, the range of Ca/Sr is lower (�80 to 400) than estimated for carbonaterocks (�370 to 2200), and a positive correlation betweenCa/Sr and Ca/Mg can be highlighted in the Moselle basinwaters (Fig. 5). Low Ca/Mg ratios in dissolved river loadshave already been documented by Meybeck (1984) forwaters draining marls and evaporites of the ‘‘Marnes iris-ees Inferieures” formation (Upper Keuper formation) ofthe Lorraine plateau. Ca/Mg ratios for these waters aresimilar to, or even lower than, Ca/Mg ratios for riversdraining silicate bedrocks. Usually, carbonate dissolutionis considered to be a congruent process, with no chemicalfractionation. Consequently, Ca/Mg and Ca/Sr ratios areexpected to be similar in the dissolved phase and corre-sponding carbonate solid phase, as shown by our leachingexperiments (Table 5). The observed dichotomy in theMoselle basin highlights a significant Mg excess and, toa lesser extent, a Sr excess (relative to Ca). The Mg excessis even more important if we consider that dissolved Ca isalso partially derived from the dissolution of evaporite(gypsum), as shown in Brenot et al. (2007). In the litera-ture, only a few cases of Mg and Sr excess relative toCa are highlighted and explained (Kotarba et al., 1981;Sarin et al., 1989; Galy et al., 1999; Jacobson et al.,2002). On a more global scale, most carbonate rivers dis-play low Ca/Mg ratios, between 0 and 5, with 87Sr/86Sr

values between 0.705 and 0.720 (Fig. 6; Sarin et al.,1989; Petelet et al., 1998; Gaillardet et al., 1999; Galy,1999; Roy et al., 1999; English et al., 2000; Jacobsonet al., 2002; Millot et al., 2003).

First, we consider whether silicate minerals present asimpurities in limestones could explain this Mg excess. In-deed, these waters also display 87Sr/86Sr values that areslightly more radiogenic than pure Triassic carbonates(87Sr/86Sr � 0.707, Koepnick et al., 1990; Martin andMacdougall, 1995; Korte et al., 2003), and this differenceis best explained by some contribution from silicatematerial, either from the Vosges mountains (for the MoselleRiver), or present within the carbonate formations. For theMoselle River, simple calculations show that Sr originatingfrom the weathering of the silicate phase would represent atmost �16% of the dissolved Sr (assuming that phyllosili-cates are inherited from erosion of the local Paleozoic gran-ite, and using the mean composition of the Moselle River atEpinal, 87Sr/86Sr = 0.7149). Moreover, the extremely lowCa/Mg ratio measured for dissolved phases of the Lorraineplateau is difficult to explain by a significant contributionfrom the detrital phase of marls and carbonate formations.Indeed, it is unlikely that phyllosilicates have the capacityto weather easily, compared to calcite, and to release signif-icant amounts of Mg into water.

On the Lorraine plateau, dominated by agricultural landuse, Mg-fertilizers could potentially modify natural Ca/Mgratios and the d26Mg signature of river water. The Mg-fer-tilizers that are potentially used in this area are derivedfrom dolostones and thus they display low Ca/Mg ratio(1.05 for R8a, Table 5) and they are expected to haved26Mg signatures similar to local dolostone rocks

Fig. 4. (a) d26Mg vs. Ca/Mg ratio for the dissolved phase, rocksand soils of the silicate part of the Moselle River basin. The Ca/Mgrange for silicate-draining rivers and silicate rocks of the sameregion is taken from Probst et al. (2000) and Nedeltcheva et al.(2006). The displayed analytical uncertainty corresponds to theestimated external error (0.22&, see Section 3.3.2). (b) Histogramcomparing the d26Mg range for granitic/gneissic river waters, bulkrocks of the continental crust, and silicate soils (this study, Tipperet al. (2006a,b)). N is the number of samples.

Fig. 5. Ca/Mg and Ca/Sr molar ratios for the Moselle River andtributaries flowing over silicates and carbonates compared to Ca/Mg and Ca/Sr for silicate rocks, sedimentary rocks of the Lorraineplateau, and Madon bedload sediments (Table 5).

Magnesium isotope systematics on the Moselle river basin 5083

(d26Mg = �1.5& for R8a, Table 5). As shown in a previousstudy (Brenot et al., 2007), major contributions of anthro-pogenic inputs from fertilizer leaching occur on theLorraine plateau mostly during the high river flow period.Thus Mg-fertilizer contribution is also expected to occurpreferentially during this period. However Ca/Mg ratiosin rivers draining carbonates are lower during the low-flowperiod (1.5–2.6, Table 3, Brenot et al., 2007), when leachingof fertilizers in soils is of least importance (Brenot et al.,2007). Furthermore d26Mg values of the Madon River,draining the most cultivated part of the Moselle basin, donot vary with sampling period, while other proxies clearlyaffected by fertilizers, such as S isotopes, correlate with riverflow (Brenot et al., 2007). These results suggest that Mg-fer-tilizer is not a significant source of dissolved Mg to riverwater. Thus observed Mg excess on the Lorraine plateaucan not be explained by anthropogenic inputs of dissolvedMg.

Dolostones are also present naturally in the lithologiesof the Lorraine plateau. Their Ca/Mg ratios range between1 and 2 (Minoux et al., 1978; Table 5) and could explain thecomposition of the dissolved load and the observed Mg ex-cess. Gaillardet et al. (1999) documented an average Ca/Mgratio of 2.5 (±21) for rivers flowing mainly on carbonates.This ratio is close to the values obtained of the Lorraineplateau rivers. Assuming congruent dissolution, the corre-sponding average carbonate should contain between 3%and 8% MgO. For comparison, ‘‘pure” limestones, withnegligible amounts of Si (Cal–S, Yeghicheyan et al., 2003)have 0.4% MgO while dolostones have, by definition, morethan 10% MgO. This simple calculation implies a significantcontribution from dolostone for the Lorraine plateau andother world-wide rivers. This is very unlikely because dolo-stones form a minor lithological component of the plateauLorraine and of continents (e.g. Berner and Berner, 1996).River d26Mg does not allow a better determination of thiscontribution since the range of d26Mg for Lorraine lime-stones is large and includes dolostone d26Mg values (Table5; Tipper et al., 2006b).

Low Ca/Mg in river waters could also be due to the re-moval of Ca by calcite precipitation, as inferred forHimalayan rivers (Kotarba et al., 1981; Sarin et al., 1989;Galy et al., 1999; Jacobson et al., 2002; Tipper et al.,2006a). In the Moselle River basin, the calcite saturation in-dex (Kempe, 1984) indicates that tributaries on the carbon-ate part (Durbion, Euron, Madon) are supersaturated withrespect to calcite (0.1 < logXc < 1.7). Theoretical condi-tions are thus fulfilled to allow calcite precipitation. It ishowever difficult in a carbonate soil to determine whethercalcite precipitation actually occurred in situ and, accordingto the literature (Jacobson and Usdowski, 1975; Dandu-rand et al., 1982; Suarez, 1983), waters supersaturated withrespect to calcite are sometimes stable with respect to cal-cite. On the Moselle basin, no apparent re-precipitationof carbonates can be seen in the river beds. As speculatedby Tipper et al. (2006a) for Himalayan rivers (on the basis

Fig. 6. 87Sr/86Sr vs. Ca/Mg ratios for dissolved loads documented in the literature for world wide rivers and for the Moselle River basin.

5084 A. Brenot et al. / Geochimica et Cosmochimica Acta 72 (2008) 5070–5089

of Ca isotope data), this calcite precipitation could possiblyoccur in groundwater and soil porewater. The relationshipbetween dissolved Ca/Mg ratios and the calcite saturationindex, logXc (Fig. 7), may suggest a non-conservativebehavior of Ca. Furthermore, most of the carbonate/evap-orite draining tributaries are also supersaturated with re-spect to magnesite (MgCO3) and strontionite (SrCO3).Local groundwaters in carbonates have also been demon-

Fig. 7. Ca/Mg vs. calcite saturation index (logXc) for Vosgiansilicate rivers, for the Moselle River downstream of Epinal, and fortributaries draining carbonates/evaporites of the Lorraine plateau.

strated to be supersaturated with respect to these species,thus our speculative explanation is that there is continuousbulk rock dissolution coupled with re-precipitation of cal-cite early on and magnesite and strontianite later on. As aconsequence, repetitive dissolution/mineral precipitationcycles during the groundwater flow path would be the mostplausible mechanism to achieve observed enrichment in dis-solved Mg and Sr in river water compared to the parentcarbonate rocks (Fig. 5). Based on knowledge of carbonategeochemistry, Ca isotope analysis of these waters may helpto further constrain and understand the mechanisms in-volved in attaining the observed Mg and Sr excesses. None-theless, repetitive dissolution/mineral precipitation cyclesare not expected to significantly change the Mg isotope sig-nature of waters (Tipper et al., 2006a) because the Mg par-tition coefficient between inorganic calcite and fluid is low(e.g. Rimstidt et al., 1998). This is consistent with Mg isoto-pic data observed for dissolved Mg (Table 4) and carbon-ates rocks (Table 5) in this study.

5.3. Assessing Mg isotopes as a source proxy in the Moselle

basin

In the silicate part of the basin, river samples having thehighest 87Sr/86Sr ratios also have the highest d26Mg values(Table 4, Fig. 8). This observation could be explained bya coupled leaching of Mg and Sr from primary rocks andminerals and should be of fundamental importance tounderstand the Mg isotope fractionation. Possibly the

Fig. 8. (a) d26Mg vs. 87Sr/86Sr for waters of the Moselle basin. ‘‘Silicate discharge” corresponds to water flowing through the silicate part ofthe catchment, and is represented by the Moselle River at Epinal. ‘‘Carbonate/evaporite discharge” corresponds to water flowing through thecarbonate/evaporite part of the catchment and is represented by the Madon River at Pont Saint Vincent, and ‘‘evaporite discharge” isrepresented by the Euron River (see text for further details). Stars display mean values. The solid line corresponds to mixing between thesilicate discharge and the carbonate/evaporate discharge, considering annual means for Mg and Sr contents. The shaded area displays thevariation between low-flow and high-flow seasons. The dashed line corresponds to mixing between the silicate discharge and the evaporitedischarge. The displayed analytical uncertainty corresponds to the estimated external error (see Section 3.3.2) (b) d26Mg measured in theMoselle River water at Portieux and Messein (two cities located downstream of Epinal) throughout one year. (c) 87Sr/86Sr measured in theMoselle River water at Portieux and Messein for the same period. (d) histogram comparing d26Mg values for rivers draining the silicateVosges (in grey), and rivers draining carbonates and evaporites of the Lorraine plateau (in white). 85% of the values for waters drainingsilicates are comprised between �0.7& and �0.9&. In contrast, Moselle tributaries draining carbonates and evaporites display d26Mg6�1.1&.

Magnesium isotope systematics on the Moselle river basin 5085

leaching of primary magnesian minerals, having high87Sr/86Sr (e.g., biotite), could explain high d26Mg valuesin river water if these magnesian minerals display theappropriate complementary Mg isotope composition (i.e.,after leaching). An alternative explanation is that clay for-mation has not progressed significantly in the little-weath-ered Vosges headwater catchments, such that radiogenicSr would be associated with Mg having d26Mg close to thatof the granitoid source rocks. These different points havenot been documented yet and should be investigated inthe future through the analysis of d26Mg values for separateminerals, in particular biotite and clay minerals.

The Moselle is an ideal river basin to test the impact oflithological sources on Mg isotopes. Downstream of thecity of Epinal, located on the lithological boundary, the

composition of Moselle River water is expected to reflecta mixing of waters draining the three lithologies: silicates,carbonates and evaporites. As the diversity of silicate rocksis integrated by the Moselle River at Epinal, the composi-tion at this locality is considered to be representative ofthe silicate discharge of the Moselle River basin. The com-position of the Moselle at Epinal is fairly constant through-out the year (d26Mg = �0.8&; 87Sr/86Sr = 0.71451–0.71521), supporting the choice of this sampling point tomodel the composition of silicate-draining river. The Ma-don River at Pont Saint Vincent (the outlet of the studiedarea) drains all the carbonate and evaporite formations ofthe Lorraine Plateau. Thus we consider that the composi-tion of the Madon at Pont Saint Vincent, provides a suit-able estimate of the carbonate/evaporite discharge of the

5086 A. Brenot et al. / Geochimica et Cosmochimica Acta 72 (2008) 5070–5089

Moselle River basin. The composition of Euron River isused to represent the evaporite discharge as this river drainsalmost exclusively evaporitic rocks (see Section 2, Fig. 1). Inorder to verify if the Mg isotope composition of the MoselleRiver downstream of Epinal can be explained by a simplemixing process, the three corresponding end-members (sili-cate discharge, carbonate/evaporite discharge and evapo-rite discharge) are examined in a Mg isotope vs. Srisotope diagram (Fig. 8). In this Figure, we included onlythe data necessary to discuss the mixing of water in theMoselle river after the main change in the lithology atEpinal.

The isotope composition of the Moselle River at Epi-nal is distinct from the composition of the Madon River(carbonate/evaporite discharge), which is also relativelyconstant throughout the year (d26Mg = �1.6& to�1.4&; 87Sr/86Sr = 0.70823–0.70843). The difference ind26Mg between both river end-members (water flowingon monolithological sub-catchment) is small but can notbe neglected (0.6–0.8&, Fig. 8d). This highlights, at leastfor this French basin, the potential of Mg isotopes as asource tracer, especially if analytical uncertainties duringMg isotope measurements can be improved. The EuronRiver (evaporite discharge) displays d26Mg ranging from�1.4& to �1.2& throughout the year. The isotope com-positions of the Moselle River downstream of Epinal,sampled at different places and during different seasons,all plot between the river end-members, and can be ex-plained by simple mixing processes, with a significant con-tribution from the silicate lithologies (>70%, Fig. 8a).This observed trend (Fig. 8) remains convincing after add-ing all the data from Table 4. This contribution is, how-ever, not large enough to explain the Mg excess describedin Section 5.2.

Temporal co-variations between d26Mg and 87Sr/86Sr forthe Moselle River downstream of Epinal further suggestthat lithology is an important influence on river Mg isotopecompositions (Fig. 8b and c). During the high flow season,the Moselle on the Lorraine plateau is characterized byhigher 87Sr/86Sr ratios (from 0.7102 to 0.7119 at the cityof Portieux), and also by higher d26Mg values (from�1.4& to �0.8&). This is consistent with a higher contri-bution, at the outlet of the basin, of both runoff and dis-solved elements from the Vosges mountains in winter(Table 3). This is also compatible with more intense soilleaching in winter, perhaps favouring phylosillicate weath-ering in impure carbonate soils.

6. SUMMARY AND CONCLUSION

This study investigates the potential of Mg isotopes asa source tracer in the Moselle River basin (northeasternFrance). This small basin is remarkable for its well con-strained and contrasted lithology at a small distancescale, and this is well reflected in river water Sr isotopecompositions. Upstream from Epinal, the Moselle Riverdrains silicate rocks of the Vosges mountains, and ischaracterized by relatively high 87Sr/86Sr ratios (0.7128–0.7174). Downstream of the city of Epinal, the MoselleRiver flows through carbonates and evaporites of the

Lorraine plateau and 87Sr/86Sr ratios are lower, downto 0.70824.

Magnesium in river waters draining silicates is systemat-ically depleted in heavy isotopes (d26Mg = �1.2 and�0.7&) relative to the value estimated for the continentalcrust and a local fresh diorite (�0.5&). In comparison,d26Mg measured in soil samples are higher (�0.0&). Thissuggests that Mg isotope fractionation occurs during sili-cate weathering. Different processes may be involved suchas leaching, secondary mineral formation, and elementalredistribution between soils and biota. The lack of knowl-edge regarding Mg isotope fractionation as a result of theseprocesses limits, at present, deconvolution of the keyparameters.

Significant Mg and Sr excesses, relative to Ca, are high-lighted in waters draining carbonate and evaporite litholo-gies. We propose that early thermodynamic saturation ofgroundwater with calcite relative to magnesite and stron-tianite during groundwater chemical evolution in the car-bonate aquifer is the most likely cause of Mg and Srexcesses for the Moselle basin. This study also shows thatMg excesses are not restricted to the Himalayas and that re-moval of Ca by calcite precipitation is likely to occur innumerous river basins.

All samples collected downstream of Epinal can be ex-plained by mixing of Mg derived from silicates, carbonatesand evaporites. A temporal covariation between d26Mg and87Sr/86Sr throughout the year is highlighted and is consis-tent with a higher contribution of runoff and dissolved ele-ments from the Vosges mountains in winter.

This study illustrates that Mg isotopes measured inwaters, rocks and soils, coupled with other tracers such asSr isotopes, can be used to better constrain riverine Mgsources. However, analytical uncertainties need to be im-proved if Mg isotopes in river waters are to be used as aprecise lithological tracer. Even if the d26Mg determinedfor granitic rocks are higher than that of carbonates, Mgisotope fractionation during silicate weathering inducesthe preferential release of light isotopes in waters, resultingin low dissolved d26Mg, closer to the carbonate end-mem-ber. Thus, this process clearly limits the potential of Mg iso-topes to trace sources of dissolved Mg. However, significantimprovements in the knowledge and quantification of thecarbon cycle may be made through identification of thecontrols on low temperature Mg isotope fractionation insilicate and carbonate environments.

ACKNOWLEDGMENTS

We are most grateful for the analytical assistance of C.Fournier, E. Bolou Bi Bolou, C. Guilmette, D. Merlet, D.Yeghicheyan and C. Zimmerman (CRPG, Nancy). We thankD. Foissy (INRA) for collecting river water near Mirecourt.We particularly thank N. Angeli and T. Nedeltcheva for provid-ing comparisons with pristine streams in the silicate part of thebasin. We also thank N. Angeli, M. Benoıt, E. Dambrine andT. Nedeltcheva for judicious discussion. A. Williams (CRPG)is thanked for English corrections. Two anonymous reviewersand A. Galy, are thanked for their detailed reviews that signifi-cantly improved a first version of the manuscript. We also thankR. James, P. Pogge von Strandmann and T. Bullen, reviewers of

Magnesium isotope systematics on the Moselle river basin 5087

the second version of the manuscript and the Editor M. Rehk-amper whose comments were crucial to improve the final versionof the manuscript. This research was supported by a PhD-engi-neer (BDI) grant from the Region Lorraine and the Centre Na-tional de la Recherche Scientifique (CNRS). This study is part ofa multidisciplinary research program on the Moselle River(ZAM). This is CRPG contribution 1927.

REFERENCES

Aubert D., Stille P. and Probst A. (2001) REE fractionation duringgranite weathering and removal by waters and suspended loads:Sr and Nd isotopic evidence. Geochim. Cosmochim. Acta 65,

387–406.

Aubert D., Probst A. and Stille P. (2004) Distribution and origin ofmajor and trace elements (particularly REE, U and Th) intolabile and residual phases in an acid soil profile (VosgesMountains, France). Appl. Geochem. 19, 899–916.

Beard B. L., Johnson C. M., Skulan J. L., Nealson K. H., Cox L.and Sun H. (2003) Application of Fe isotopes to tracing thegeochemical and biological cycling of Fe. Chem. Geol. 196, 43–

56.

Benichou P. and Le Breton O. (1987) Prise en compte de latopographie pour la cartographie des champs pluviometriquesstatistiques. La Meteorologie 7, 23–34.

Berner K. E. and Berner R. A. (1996) Global Environment—Water,

Air, and Geochemical Cycles. Prentice-Hall, Upper SaddleRiver.

Berner R. A., Lasaga A. C. and Garrels R. M. (1983) Thecarbonate silicate geochemical cycle and its effect on atmo-spheric carbon dioxide over the past 100 million years. Am. J.

Sci. 283, 641–683.

Bonhomme M. (1967) Ages radiometriques de quelques granitesdes Vosges moyennes. Bull. Cate Geol. Als. Lorr. 20, 101–106.

Black J. R., Yin Q.-Z. and Casez W. H. (2006) An experimentalstudy of magnesium-isotope fractionation in chlorophyll-aphotosynthesis. Geochim. Cosmochim. Acta 70, 4072–4079.

Brenot A., Carignan J. and France-Lanord C. (2007) Geologic andland use controls on d34S and d18O of river dissolved sulfate: theMoselle river basin, France. Chem. Geol. 244, 25–41.

Carder E. A., Galy A. and Elderfield H. E. (2004) The magnesiumisotopic composition of oceanic water masses. Geochim. Cos-

mochim. Acta 68, A329.

Cardinal D., Alleman L. Y., Jong J., Ziegler K. and Andre L.(2003) Isotopic composition of silicon measured by multicol-lector plasma source mass spectrometry in dry plasma mode. J.

Anal. Atomic Spectrom. 18, 213–218.

Carignan J., Hild P., Mevelle G., Morel J. and Yeghicheyan D.(2001) Routine analyses of trace elements in geological samplesusing flow injection and low pressure on-line liquid chroma-tography coupled to ICP-MS; a study of geochemical referencematerials BR, DR-N, UB-N, AN-G and GH. Geostandards

Newslett. 25, 187–198.

Carignan J., Cardinal D., Eisenhauer A., Galy A., Rehkamper M.,Wombacher F. and Vigier N. (2004) A reflection on Mg, Cd,Ca, Li and Si isotopic measurements and related referencematerials. Geostandards Geoanal. Res. 28, 139–148.

Chabaux F., Riotte J., Schmitt A-D., Carignan J., Herckes P.and Pierret M-C. (2005) Variations of U and Sr isotoperatios in Alsace and Luxembourg rain waters: origin andhydrogeochemical implications. Comptes Rendus Geosci. 33,

1447–1451.

Chang V. T.-C., Makishima A., Belshaw N. S. and O’Nions R. K.(2003) Purification of Mg from low-Mg biogenic carbonates for

isotope ratio determination using multiple collector ICP-MS. J.

Anal. Atomic Spectrom. 18, 296–301.

Dadi S. (1991) Qualite des eaux de la Moselle a la prise d’eaudu dis-trict de l’agglomeration naceienne. Analyse desdonnees pour la periode 1973–1988. PhD thesis. INPLNancy, 214 p.

Dandurand J. L., Gout R., Hoefs J., Menschel G., Schott J. andUsdowski E. (1982) Kinetically controlled variations of majorcomponents and carbon and oxygen isotopes in a calciteprecipitation spring. Chem. Geol. 36, 299–315.

De Villiers S., Dickson J. A. D. and Ellam R. M. (2005) Thecomposition of the continental river weathering flux deducedfrom seawater Mg isotopes. Chem. Geol. 216, 133–142.

English N. B., Quade J., DeCelles P. G. and Garzione C. N. (2000)Geologic control of Sr and major element chemistry inHimalayan Rivers, Nepal. Geochim. Cosmochim. Acta 64,

2549–2566.

France-Lanord (1982) Utilisation d’un litage magmatique pour ladatation Rb–Sr du granite des Cretes. DES Sciences Naturelles.

Nancy I, 28.Gagny C. (1959) Contribution a l’etude des granites de l’associa-

tion des cretes dans le bassin de la haute Moselle. PhD thesis.Paris, 83 p.

Gaillardet J., Dupre B., Louvat P. and Allegre C. J. (1999) Globalsilicate weathering and CO2 consumption rates deduced fromthe chemistry of large rivers. Chem. Geol. 159, 3–30.

Galy A. (1999) Etude geochimique de l’erosion actuelle de la chaınehimalayenne. PhD thesis. INPL Nancy, 464 p.

Galy A. and France-Lanord C. (1999) Weathering processes in theGanges–Brahmaputra basin and the riverine alkalinity budget.Chem. Geol. 159, 31–60.

Galy A., France-Lanord C. and Derry L. A. (1999) The strontiumisotopic budget of Himalayan Rivers in Nepal and Bangladesh.Geochim. Cosmochim. Acta 63, 1905–1925.

Galy A., Belshaw N. S., Halicz L. and O’Nions R. K. (2001) High-precision measurement of magnesium isotopes by multiple-collector inductively coupled plasma mass spectrometry. Int. J.

Mass Spectrom. 208, 89–98.

Galy A., Bar-Matthews M., Halicz L. and O’Nions R. K. (2002)Mg isotopic composition of carbonate: insight from speleothemformation. Earth Planet. Sci. Lett. 201, 105–115.

Galy A., Yoffe O., Janney P. E., Williams R. W., Cloquet C., AlardO., Halicz L., Wadwha M., Hutcheon I. D., Ramon E. andCarignan J. (2003) Magnesium isotopes heterogeneity of theisotopic standard SRM980 and new reference materials formagnesium-isotope-ratio measurements. J. Anal. Atomic Spec-

trom. 18, 1352–1356.

Georg R. B., Reynolds B. C., Frank M. and Haliday A. (2006)Mechanisms controlling the silicon isotopic compositions ofriver waters. Earth Planet. Sci. Lett. 249, 290–306.

Govindaraju K. and Mevelle G. (1987) Fully automated dissolu-tion and separation methods for inductively coupled plasmaatomic emission spectrometry rock analysis. Application to thedetermination of Rare Earth Elements. J. Anal. Atomic

Spectrom. 2, 615–621.

Halicz L., Galy A., Belshaw N. S. and O’Nions R. K. (1999) High-precision measurement of calcium isotopes in carbonates andrelated materials by multiple collector inductively coupledplasma mass spectrometry (MC–ICP–MS). J. Anal. Atomic

Spectrom. 14, 1835–1838.

Horwitz E. P., Chiarizia R. and Dietz M. L. (1992) A novelstrontium-selective extraction chromatographic resin. Solvent

Extract. Ion Exchange 10, 313–336.

Huh Y., Chan L.-H., Zhang L. and Edmond J. M. (1998) Lithiumand its isotopes in major world rivers: implications for

5088 A. Brenot et al. / Geochimica et Cosmochimica Acta 72 (2008) 5070–5089

weathering and the oceanic budget. Geochim. Cosmochim. Acta

62, 2039–2051.

Jacobson and Usdowski (1975) Geochemical controls on a calciteprecipitation spring. Contrib. Mineral. Petrol. 51, 65–74.

Jacobson A. D., Blum J. D. and Walter L. M. (2002) Reconcilingthe elemental and Sr isotope composition of Himalayanweathering fluxes: insights from the carbonate geochemistryof stream waters. Geochim. Cosmochim. Acta 66, 3417–3429.

James R. H. and Palmer M. R. (2000) The lithium isotopecomposition of international rock standards. Chem. Geol. 166,

319–326.

Jung J. (1927) Contribution a la geologie des Vosges Hercyniennesd’Alsace. Mem. Serv. Carte geol. Als.-Lorr. 2, 481.

Kempe S. (1984) Sinks of anthropogenically enhanced carbon cyclein surface fresh waters. J. Geophys. Res. 89(D3), 4657–4676.

Koepnick R. B., Denison E. E., Burke W. H., Hetherington E. A.and Dahl D. A. (1990) Construction of the Triassic and Jurassicportion of the Phanerozoic curve of seawater 87Sr/86Sr. Chem.

Geol. 80, 327–349.

Korte C., Kozur H. W., Bruckschen P. and Veizer J. (2003)Strontium isotope evolution of Late Permian and Triassicseawater. Geochim. Cosmochim. Acta 67, 47–62.

Kotarba M., Sokolowski A. and Bogacz W. (1981) Hydrogeolog-ical investigations in the Kali Gandaki thermal springs area(Nepal Himalayas). Bull. Polish Acad. Sci. Earth Sci. 29, 283–

291.

Louvat P. and Allegre C. J. (1997) Present denudation rates on theisland of Reunion determined by river geochemistry: basaltweathering and mass budget between chemical and mechanicalerosions. Geochim. Cosmochim. Acta 61, 3645–3669.

Ludwig K. R. (2001) Isoplot/Ex version 2.49. A geochronologicaltoolkit for Microsoft Excel. Berkeley Geochronology Center.

Special publication 1a. Nov., 20.Luais B., Telouk P. and Albarede F. (1997) Precise and accurate

neodymium isotopic measurements by plasma-source massspectrometry. Geochim. Cosmochim. Acta 61, 4847–4854.

Marechal C. N., Telouk P. and Albarede F. (1999) Precise analysisof copper and zinc isotopic compositions by plasma-sourcemass spectrometry. Chem. Geol. 156, 251–273.

Martin E. E. and Macdougall J. D. (1995) Sr and Nd isotopes atthe Permian-Triassic boundary: a record of climate change.Chem. Geol. 125, 73–99.

Meybeck M. (1983) Atmospheric inputs and river transport ofdissolved substances. In Proceedings of Hamburg Symposium,vol. 141. IAHS Publication, pp. 173–192.

Meybeck M. (1984) Les fleuves et le cycle geochimique deselements. PhD thesis. Paris 6, 493 p.

Millot R., Gaillardet J., Dupre B. and Allegre C. J. (2003)Northern latitude chemical weathering rates: clues from theMackenzie River Basin, Canada. Geochim. Cosmochim. Acta

67, 1305–1329.

Minoux G., Marotel C. and Guillaume C. (1978) B.R.G.M.: Cartegeologique au 1/50.000eme de Mirecourt. N� 304.

Nedeltcheva T., Piedallu C., Gegout J.-C., Stussi J.-M., Boudot J.-P., Angeli N. and Dambrine E. (2006) Influence of granitemineralogy, rainfall, vegetation and relief on stream waterchemistry (Vosges mountains, North eastern France). Chem.

Geol. 231, 1–15.

Negrel P., Allegre C. J., Dupre N. and Lewin E. (1993) Erosionsources determined by insertion of major and trace elementratios in river water: the Congo Basin case. Earth Planet. Sci.

Lett. 120, 59–76.

Pearson N. J., Griffin W. L., Alard O. and O’Reilly S. Y. (2006)The isotopic composition of magnesium in mantle olivine:records of depletion and metasomatism. Chem. Geol. 226, 115–

133.

Periaux J. (1961) Contribution a la geologie des Vosges greseuses.Mem. Ser. Carte geol. Als-Lorr. 18, 236.

Petelet E., Luck J.-M., Othman D. B., Negrel P. andAquilina L. (1998) Geochemistry and water dynamics ofa medium-sized watershed: the Herault, southern France1. Organisation of the different water reservoirs asconstrained by Sr isotopes, major, and trace elements.Chem. Geol. 150, 63–83.

Pin C. and Bassin C. (1992) Evaluation of a strontium-specificextraction chromatographic method for isotopic analysis ingeological materials. Anal. Chim. Acta 269, 249–255.

Probst A., ElDh’mari A., Aubert D., Fritz B. and McNutt R.(2000) Strontium as a tracer of weathering processes in a silicatecatchment polluted by acid atmospheric inputs, Strengbach,France. Chem. Geol. 170, 203–219.

Rehkamper M. and Halliday A. N. (1999) The precise measure-ment of Tl isotopic compositions by MC–ICP–MS: applicationto the analysis of geological materials and meteorites. Geochim.

Cosmochim. Acta 63, 935–944.

Rimstidt J. D., Balog A. and Webb J. (1998) Distribution of traceelements between carbonate minerals and aqueous solutions.Geochim. Cosmochim. Acta 62, 1851–1863.

Rouxel O., Ludden J., Carignan J., Marin L. and Fouquet Y.(2002) Natural variation of Se isotopic composition determinedby hydride generation multiple collector inductively coupledplasma mass spectrometry. Geochim. Cosmochim. Acta 66,

3191–3199.

Roy S., Gaillardet J. and Allegre C. J. (1999) Geochemistry ofdissolved and suspended loads of the Seine river, France:anthropogenic impact, carbonate and silicate weathering.Geochim. Cosmochim. Acta 63, 1277–1292.

Sarin M. M., Krishnaswami S., Dilli K., Somayajulu B. L. K. andMoore W. S. (1989) Major ion chemistry of the Ganga–Brahmaputra river system: weathering processes and fluxes ofthe Bay of Bengal. Geochim. Cosmochim. Acta 53, 997–1009.

Stallard R. F. (1980) Major element chemistry of the AmazonRiver system. PhD thesis. MIT/Woods Hole OceanographicInst., WHOI-80-29.

Stallard R. F. and Edmond J. M. (1981) Geochemistry of theAmazon. Precipitation chemistry and the marine contributionto the dissolved load at the time of peak discharge. J. Geophys.

Res. 86, 9844–9858.

Suarez D. L. (1983) Calcite supersaturation and precipitationkinetics in the Lower Colorado River, All-American Canal andEast Highline Canal. Water Res. Bull. 19, 653–661.

Tipper E. T., Galy A. and Bickle M. J. (2006a) Riverine evidencefor a fractionated reservoir of Ca and Mg on the continents:implications for the oceanic Ca cycle. Earth Planet. Sci. Lett.

247, 267–279.

Tipper E. T., Galy A., Gaillardet J., Bickle M. J., Elderfield H. andCarder E. A. (2006b) The magnesium isotope budget of themodern ocean: constraints from riverine magnesium isotoperatios. Earth Planet. Sci. Lett. 50, 241–253.

Wadleigh M. A., Veizer J. and Brooks C. (1985) Strontium and itsisotopes in Canadian Rivers: fluxes and global implications.Geochim. Cosmochim. Acta 49, 1727–1736.

Walker J. C. G., Hays P. B. and Kasting J. F. (1981) A negativefeedback mechanism for the long-term stabilization of Earthssurface-temperature. J. Geophys. Res. Oceans Atmos. 86, 9776–

9782.

Wombacher F., Rehkamper M., Mezger K. and Munker C. (2003)Stable isotope compositions of cadmium in geological materialsand meteorites determined by multiple-collector ICPMS. Geo-

chim. Cosmochim. Acta 67, 4639–4654.

Wombacher F., Eisenhauer A., Heuser A. and Weyer S. (2006)Combined chemical separation of Mg, Ca and Fe from

Magnesium isotope systematics on the Moselle river basin 5089

geological reference samples for stable isotope determinationby MC–ICP–MS and double-spike TIMS.EGU, abstract06219.

Yeghicheyan D., Carignan J., Valladon M., Bouhnik-Le Caz M.,Le Cornec F., Castrec-Rouelle M., Robert M., Aquilina L.,Aubry E., Churlaud C., Dia A., Deberdt S., Dupre B., FreydierR., Gruau G., Henin O., De Kersabiec A. M., Mace J., MarinL., Morin N., Petitjean P. and Serrat E. (2001) A compilationof silicon and thirty one trace elements measured in the naturalriver water reference material SLRS-4 (NRC-CNRC). Geostan-

dards Newslett. 25, 465–474.

Yeghicheyan D., Carignan J., Valladon M., Bouhnik Le Coz M.,Samuel J., BenBakkar M., Bruguier O., Keller F., Pin C.,

Pourtales L., Henin O., Mace J., Morin N., Guilmette C. andMarin L. (2003) The new carbonate reference material Cal–S:preliminary results. Abs. Geoanal. 2003, 146.

Young E. D. and Galy A. (2004) The isotope geochemistry andcosmochemistry of Magnesium. Rev. Mineral. Geochem. 55,

197–230.

Ziegler K., Chadwick O. A., Brzezinski M. A. and Kelly E. F.(2005) Natural variations of d30Si ratios during progressivebasalt weathering, Hawaiian Islands. Geochim. Cosmochim.

Acta 69, 4597–4610.

Associate editor: Mark Rehkamper