12
Impure marbles of the Lesser Himalaya : another source of continental radiogenic osmium Anne-Catherine Pierson-Wickmann , Laurie Reisberg, Christian France-Lanord Centre de Recherches Pe ¤trographiques et Ge ¤ochimiques, 15 rue Notre-Dame des Pauvres, BP 20, F-54501 VandEuvre-le 's-Nancy, France Received 26 February 2002; received in revised form 3 September 2002; accepted 19 September 2002 Abstract The marked rise of the 187 Os/ 188 Os ratio of seawater over the Cenozoic Era is generally attributed to increased continental input, and particularly to increased erosion of black shales. Thus Os isotopes are thought to trace the weathering of organic carbon. Here we present evidence that impure marbles, lacking organic carbon, provide an important source of radiogenic Os to the bedload of Himalayan rivers. These marbles provide a critical component necessary for balancing the Os isotopic budget of the bedloads. The analyzed marbles have 187 Os/ 188 Os ratios as high as 194 and often contain as much 187 Os as typical black shales. Most of this radiogenic Os is not contained in the carbonate itself, but rather in associated phases. These phases are often of hydrothermal origin, and their Os may ultimately be provided by fluids circulating through black shales. While these results complicate the relationship between the Os isotopic ratio of seawater and erosion of organic-rich rocks, they also provide new information necessary for understanding the marine Os isotopic record. ȣ 2002 Elsevier Science B.V. All rights reserved. Keywords: black shale; osmium; rhenium; hydrothermal alteration; carbonates; Himalayas 1. Introduction Continental weathering rates greatly in£uence the CO 2 content of the atmosphere, and thus in- directly strongly a¡ect the Earth’s climate. It is therefore important to ¢nd geochemical tracers of weathering rates in the past. The dramatic in- crease in 187 Os/ 188 Os of seawater during the Late Cenozoic has been interpreted as a response to increased continental erosion. Because of the tem- poral coincidence, Himalayan erosion has been proposed as the primary source of this increased input of crustal Os [1^3], in analogy with the Sr isotopic system. Other authors have argued against this proposal, on the basis of direct mea- surements of Os in Himalayan river water [4,5]. In either case, the Himalayas remain an ideal labo- ratory for the study of the behavior of the Re^Os system during erosion. 0012-821X / 02 / $ ^ see front matter ȣ 2002 Elsevier Science B.V. All rights reserved. PII:S0012-821X(02)00962-7 * Corresponding author. Present address: Department of Geological Sciences, Queen’s University, Kingston, ON, Canada K7L 3N6. Tel.: +1-613-533-6000, ext. 75659; Fax: +1-613-533-6592. E-mail addresses: [email protected] (A.-C. Pierson-Wickmann), [email protected] (L. Reisberg), c£@crpg.cnrs-nancy.fr (C. France-Lanord). Earth and Planetary Science Letters 204 (2002) 203^214 www.elsevier.com/locate/epsl

Impure marbles of the Lesser Himalaya: another source of continental radiogenic osmium

Embed Size (px)

Citation preview

Impure marbles of the Lesser Himalaya:another source of continental radiogenic osmium

Anne-Catherine Pierson-Wickmann �, Laurie Reisberg,Christian France-Lanord

Centre de Recherches Pe¤trographiques et Ge¤ochimiques, 15 rue Notre-Dame des Pauvres, BP 20, F-54501 VandEuvre-le's-Nancy, France

Received 26 February 2002; received in revised form 3 September 2002; accepted 19 September 2002

Abstract

The marked rise of the 187Os/188Os ratio of seawater over the Cenozoic Era is generally attributed to increasedcontinental input, and particularly to increased erosion of black shales. Thus Os isotopes are thought to trace theweathering of organic carbon. Here we present evidence that impure marbles, lacking organic carbon, provide animportant source of radiogenic Os to the bedload of Himalayan rivers. These marbles provide a critical componentnecessary for balancing the Os isotopic budget of the bedloads. The analyzed marbles have 187Os/188Os ratios as highas 194 and often contain as much 187Os as typical black shales. Most of this radiogenic Os is not contained in thecarbonate itself, but rather in associated phases. These phases are often of hydrothermal origin, and their Os mayultimately be provided by fluids circulating through black shales. While these results complicate the relationshipbetween the Os isotopic ratio of seawater and erosion of organic-rich rocks, they also provide new informationnecessary for understanding the marine Os isotopic record.5 2002 Elsevier Science B.V. All rights reserved.

Keywords: black shale; osmium; rhenium; hydrothermal alteration; carbonates; Himalayas

1. Introduction

Continental weathering rates greatly in£uencethe CO2 content of the atmosphere, and thus in-directly strongly a¡ect the Earth’s climate. It is

therefore important to ¢nd geochemical tracersof weathering rates in the past. The dramatic in-crease in 187Os/188Os of seawater during the LateCenozoic has been interpreted as a response toincreased continental erosion. Because of the tem-poral coincidence, Himalayan erosion has beenproposed as the primary source of this increasedinput of crustal Os [1^3], in analogy with the Srisotopic system. Other authors have arguedagainst this proposal, on the basis of direct mea-surements of Os in Himalayan river water [4,5]. Ineither case, the Himalayas remain an ideal labo-ratory for the study of the behavior of the Re^Ossystem during erosion.

0012-821X / 02 / $ ^ see front matter 5 2002 Elsevier Science B.V. All rights reserved.PII: S 0 0 1 2 - 8 2 1 X ( 0 2 ) 0 0 9 6 2 - 7

* Corresponding author. Present address: Department ofGeological Sciences, Queen’s University, Kingston, ON,Canada K7L 3N6. Tel.: +1-613-533-6000, ext. 75659;Fax: +1-613-533-6592.E-mail addresses: [email protected]

(A.-C. Pierson-Wickmann), [email protected](L. Reisberg), c£@crpg.cnrs-nancy.fr (C. France-Lanord).

EPSL 6427 6-11-02

Earth and Planetary Science Letters 204 (2002) 203^214

www.elsevier.com/locate/epsl

An important source of radiogenic Os fromcontinents is the weathering of old black shales[6], which are enriched in Os and especially inRe relative to the average continental crust [7^9]. Previous studies of Himalayan river sedimentsand source rocks [10^12] have shown that theradiogenic Os isotopic signature of the sedimentsis due in large part to the contribution of ancientradiogenic black shales, more speci¢cally thosefrom the Lesser Himalaya (LH). Nevertheless,mass balance calculations [12] indicate that blackshales alone cannot explain the Os isotopic com-positions of sediments sampled at the out£ow ofthe Himalayan range, and other sources of radio-genic Os must be identi¢ed. The rough positivetrend relating [Os] to total carbonate content insediments (Fig. 1a), as well as the tendency forsediments with higher proportions of dolomite,which are generally derived from the LH, tohave higher 187Os/188Os ratios (Fig. 1b), led usto consider the hypothesis that carbonate-richrocks from the LH may include phases that playan important role in the Himalayan Os and Rebudget. In the river sediments from the LH, thecarbonate fraction can represent up to 25% [13].Although in the past carbonate-rich rocks havenot been considered to be major sources of Osor Re [14], recent results from Levasseur et al.[4] suggest that carbonates may indeed provideimportant quantities of dissolved Os. Levasseuret al. [4] showed that while the Os of Ganges riverwater is mostly derived from erosion of blackshales (70%), carbonates also play an importantrole (25%). Here we present Re and Os concen-trations and 187Os/188Os ratios of marbles andblack shales from the LH, speci¢cally in CentralNepal. These analyses show that impure marbles,lacking organic carbon, sometimes contain largequantities of Re and Os, and that this Os can bevery radiogenic. Thus these lithologies must beconsidered when reconstructing the continentalbudget of Os and Re.

2. Geological setting and samples

The High Himalayan range can be divided intothree main formations (Fig. 2). From North to

South these are: (i) the Tethyan Sedimentary Se-ries (TSS), essentially composed of carbonatesand metamorphosed sediments, including blackshales, of Cambrian to Eocene age; (ii) the HighHimalayan Crystalline series (HHC), composed ofhighly metamorphosed paragneisses and marbles;and (iii) the LH, composed of metamorphosedsediments, marbles and minor black shales of Pre-cambrian age.The rock samples analyzed were mostly impure

dolomitic marbles, though a few LH black shaleswere also studied. An important di¡erence be-tween these rock types is that LH black shalescontain up to 10% of organic carbon [12], whereasthe marbles are nearly completely lacking in or-ganic carbon. The dolomitic marbles, containingfrom 60 to 82% carbonate, were collected fromoutcrops and from a Pb^Zn mine in the LH ofcentral Nepal (see Fig. 2 for sample locations).The marbles collected from outcrops are denotedby the pre¢x ‘AP’ and are of Precambrian age.They are characterized by low Sr concentrationsand radiogenic Sr isotopic ratios (87Sr/86Sr up to0.857) [13]. These samples outcrop as thin layersin the schists and quartzites of the LH, and havelow N

18O values due to £uid interaction duringHimalayan metamorphism [13,15]. The mine sam-ples were collected in or near the Lari mine, lo-cated in the north of the Mailung Khola, near theGanesh Himal and the Trisuli valley, at an alti-tude of about 4250 m. The Pb and Zn mineraliza-tions are con¢ned in a saccharoid dolomitic gang-ue. The presence of phlogopite and sul¢des inmany of the mine samples underscores the impor-tance of hydrothermal circulation. The sul¢desare mostly pyrite, sphalerite, chalcopyrite andpyrrhotite. Some Fe oxides, such as magnetite,were also found. For several of the mine samples,powders were prepared from both hydrothermallyaltered and adjoining, apparently pristine portionsof the same rock. In addition, about 100 mg ofhandpicked minerals (pyrites and phlogopites)from several hydrothermally altered marbleswere analyzed. Carbonate-rich samples B9-, B12-and B15-SU11 and black shale B12-SU6 are fromprospecting cores from Suple, in the vicinity ofLari mine. Marbles GA 70 and GA 39 were col-lected in the mineralized zones of Lari mine. GA

EPSL 6427 6-11-02

A.-C. Pierson-Wickmann et al. / Earth and Planetary Science Letters 204 (2002) 203^214204

39 is characterized by a hydrothermal layer in acarbonate matrix. The hydrothermal vein isunderlined by the presence of carbonate, phlogo-pite, phengite, and pyrite. Black shale GA 38 wascollected in the same mineralized zone as GA 39.The other black shales, GA 7 and GA 8, werepreviously presented in Pierson-Wickmann et al.[16,17].

3. Analytical procedures

In all cases, Os and Re were extracted by themethod of Birck et al. [18]. Bulk powders werespiked with 190Os and 185Re and dissolved in anHF^HBr mixture. The Os was oxidized by addi-tion of a concentrated CrO3 solution, extractedinto liquid Br2, and ¢nally puri¢ed by microdis-tillation [19]. After reduction of the Cr6þ withethanol, Re was extracted from the aqueous resi-due using iso-amylol. Os and Re concentrationswere determined by isotope dilution. Os isotopicratios were measured by N-TIMS [20,21], using aFinnigan MAT 262 mass spectrometer. 187Re/185Re ratios for isotope dilution calculationswere determined by ICPMS (Elan 6000). Totalprocedural blanks were about 0.3 pg for Os and5 pg for Re. This method was also used for the

determination of the Os isotopic composition andRe and Os concentrations of Himalayan samples[12,17], allowing direct comparison with these re-sults.As noted above, sediments at the out£ow of the

range have quite high Os isotopic ratios. In orderto determine whether this radiogenic Os is carriedby carbonate or by other phases, several testswere performed on NAG 48, a typical river bed-load sample containing about 14% carbonate(10% calcite and 4% dolomite; [13]). This samplehas a 187Os/188Os ratio of 3.78 and a 188Os con-centration of 1.8 ppt [12]. After sieving to a 150^300 Wm grain size, heavy liquids and a Franzmagnetic separator were used to produce a frac-tion with high density (bs 2.7 g/cm3) and verylow magnetic susceptibility containing about50% carbonate. The calcite from this fractionwas extracted by leaching in weak (0.16 N) HClfor about 15 min. This leachate, together withH2O rinses of the residue, was centrifuged andthen spiked with 190Os and 185Re. The residuewas then leached with 2.5 N HCl to remove do-lomite. This second leachate, with H2O rinses ofthe residue, was also centrifuged and spiked with190Os and 185Re. The masses of calcite and dolo-mite used to calculate the Os and Re concentra-tions were estimated by the weight loss of the

Fig. 1. (a) Variation of Os concentration as a function of TSS+LH carbonate proportion in river bedloads. (b) Variation of187Os/188Os ratio of river sediments as a function of dolomite content in river bedloads. The Os isotopic data are from Pierson-Wickmann et al. [12]. In the graph, ‘TSS+HHC+LH’ denotes sediments that have crossed the three main Himalayan formationsand were sampled in Central Nepal (squares) while ‘Bangladesh’ indicates sediments sampled in the Indo-Gangetic Floodplain(stars). The circles represent the river bedloads, which crossed only the LH.

EPSL 6427 6-11-02

A.-C. Pierson-Wickmann et al. / Earth and Planetary Science Letters 204 (2002) 203^214 205

residue during each leaching step. The residue,after drying at low temperature, was spiked anddissolved by the same HF^HBr technique used forthe rock samples and previously used by Pierson-Wickmann et al. [12] to analyze the bulk sedi-ment. In addition, a mica separate was preparedfrom the 150^300 Wm size fraction. Heavy liquidsand magnetic separation were ¢rst used to pro-duce a dense fraction of high magnetic suscepti-bility, likely to be rich in mica. Then the phyllo-silicates were separated on the basis of theirelectrostatic attraction for the plastic sample vialwalls, resulting in a nearly pure phyllosilicate sep-arate. This fraction and an aliquot of the remain-ing powder were dissolved in HF^HBr. Finally,an untreated split of the 150^300 Wm size fractionwas also analyzed to examine the e¡ect of gran-ulometry.

4. Results

Whole rock analyses of LH marbles from bothmines and outcrops display a huge range of 187Os/188Os ratios, from 0.4 to 193.7 (Table 1). Total Osconcentrations vary from 9 to 188 ppt, but mostof this variation is due to the wide range in 187Oscontents. 188Os concentrations vary from 0.4 to7.1 ppt, similar to concentrations in LH silicates[12]. Carbonates from outcrops are moderatelyto extremely radiogenic (187Os/188Os= 2.9^193.7),and many have 187Os contents comparable tothose of LH black shales. The mine carbonatesnot directly associated with hydrothermal veinsare much less radiogenic (187Os/188Os = 0.4^0.9),while their non-radiogenic Os concentrations aresimilar to those of the outcrop samples (Table 1).Mine marbles containing evidence of hydro-

N50 km

Annapurnas

Pokhara

Seti

Andhi

Trisuli

Marayandi

Bhuri Gandaki

BhoteKosi

Kathmandu

Alluvions

Siwaliks

TSS

HHC

LH

HHL

LHN

Manaslu

AP 865

KaliGan

daki

Narayani

Tansen

MFT

Dhaulagiri

Bheri GaneshLantang

ShishaPangma

AP 867

NL 449

AP 207

AP 811

GA 7-8Lari

NL 178

T 255

F III

NAG 48

NL 623

Fig. 2. Location of carbonate rocks and black shales analyzed in this study.

EPSL 6427 6-11-02

A.-C. Pierson-Wickmann et al. / Earth and Planetary Science Letters 204 (2002) 203^214206

thermal circulation in the form of phlogopitesometimes, but not always, have higher Os iso-topic ratios than non-hydrothermal samples.Both the carbonate and the hydrothermal por-tions of sample B12-SU11 have low 187Os/188Osratios (V0.4), with similar concentrations. Onthe other hand, sample GA 39, found in the prox-imity of a black shale (GA 38), contains a hydro-thermal vein with a 187Os/188Os ratio of 11.6,while the adjacent marble has a ratio of 2.5.This suggests that a transfer of Os and Re tocarbonate-rich rocks from radiogenic Os rich-rocks, such as black shales, may be at the originof the observed small-scale variation of 187Os/188Os. To better understand the origin of the ra-diogenic signature, we analyzed phlogopite, a typ-ical metamorphic phase in Mg marbles, and py-rite, both separated from the hydrothermal vein in

sample GA 39. The hydrothermal sul¢des are dis-tinguished by a very radiogenic Os isotopic ratio(187Os/188Os= 21.0) and very high Os concentra-tion (1263 ppt total Os; 44 ppt 188Os). Phlogopiteshave Os characteristics (187Os/188Os= 13.0; 252ppt total Os; 12 ppt 188Os) close to that of theaverage vein material (187Os/188Os= 11.6; 153 ppttotal Os; 8 ppt 188Os). In contrast, the carbonatematrix has only a slightly radiogenic Os ratio (2.4)and a low Os concentration (19 ppt total Os; 1.9ppt 188Os). Thus pyrite and phlogopite are themain hosts of radiogenic Os in this rock. It isunclear whether disseminated trace phases, suchas pyrite and phlogopite, are also the main hostsof 187Os in the radiogenic marbles, such as AP 207or AP 865, lacking obvious signs of hydrotherm-alism, or whether the 187Os is carried by the car-bonate itself in these cases. These samples (AP)

Table 1Os isotopic compositions and Os and Re concentrations of carbonates and black shales

Sample Lithology 187Os/188Os [Os] [187Os] [188Os] [Re] 187Re/188Os Os model age 87Sr/86Sra

(ppt) (ppt) (ppt) (ppt) (Ga)

GA 38 black shale 8.076 467 241.2 29.9 ^ ^GA 7 black shale 10.752 871 510.6 47.5 2 158 28.7 17.3GA 8 black shale 9.682 3 200 1794.9 185.2 11 344 38.6 11.9B12-SU6 black shale 7.489 380 189.2 25.3 ^ ^AP 207 dolomite 193.720 102 98.7 0.5 1 120 1 388 7.7 0.85722AP 811 dolomite 15.506 10 6.5 0.4 630 944 0.9 0.74639AP 865 dolomite 19.088 188 135.4 7.1 3 580 318 3.3 0.75872AP 867 dolomite 2.872 16 4.5 1.5 209 86 1.2 0.73950B9-SU11 marble 0.415 11.0 0.6 1.4B12-SU 11-C marble with

phlogopite0.447 16.3 0.9 2.1

B12-SU 11-H marble withphlogopite

0.428 22.4 1.2 2.8 111 24.7 ^

GA 70 marble withphlogopite

0.507 11 0.7 1.4

B15-SU11 whitish marble 0.883 9 1.0 1.1GA 85 mine scraps 0.436 14.4 0.8 1.8 66 22.9 ^GA 39-C hydrothermal

carbonate2.497 19 4.8 1.9 433 143.7 0.5

GA 39-H hydrothermalcarbonate

11.597 153 92.4 8

GA 39-carbonate hydrothermalcarbonate

2.392 17.5 4.2 1.8 392 140.3 0.5

GA 39-pyrite hydrothermalcarbonate

21.050 1 263 924.9 43.9 106 000 1 500 0.7

GA 39-phlogopite hydrothermalcarbonate

12.997 252 158.8 12.2 8 030 414.5 1.7

C: Carbonate-rich fraction of sample. H: hydrothermal fraction of sample.a Data from [13].

EPSL 6427 6-11-02

A.-C. Pierson-Wickmann et al. / Earth and Planetary Science Letters 204 (2002) 203^214 207

are characterized by low N18O values [13] typical

of metamorphosed rocks, suggesting that the re-distribution of Re and Os may have been inducedby £uid circulation during Himalayan metamor-phism.In order to determine whether carbonate is one

of the main carriers of radiogenic Os in sedimentsat the out£ow of the Himalayan range we per-formed a series of tests, detailed above, onNAG 48, a typical carbonate-rich bedload (10%of calcite and 4% of dolomite, [13]) collected inthe Narayani river (Fig. 2). The results (Table 2)can be compared with the data from the bulksample, analyzed twice with good agreement[12]. Simple sieving to a grain size of 150^300Wm has little e¡ect on the Os and Re concentra-tions, though it produces a fraction slightly moreradiogenic (187Os/188Os = 4.21) than the originalpowder (187Os/188Os = 3.78). The carbonate-en-riched aliquot produced from the 150^300 Wm

fraction by density and magnetic separation hasroughly half the 188Os and Re concentrations (1.0and 143 ppt, respectively) and a signi¢cantly low-er 187Os/188Os ratio (2.80). Calcite and dolomiteextracted from this fraction by successive weak(calcite) and strong (dolomite) HCl leaching stepshave even lower 188Os (0.8 and 0.6 ppt) and Re(14 and 20 ppt) concentrations (calculated relativeto the weight loss during each step), with 187Os/188Os ratios of 2.67 and 4.08 respectively. Thelower Os isotopic ratio of the calcite most prob-ably re£ects a larger contribution from the TSSformation in the north of the Himalayan chain,which is known to contain abundant calcite andto have a relatively unradiogenic character (187Os/188OsV0.6^2; [12]), while the dolomite re£ects alarger LH component. The silicate residue afterleaching has a 188Os concentration of 0.57 pptwith a 187Os/188Os ratio of 3.03. A perfect massbalance in Os and Re does not exist between the

Table 2187Os/188Os and Os and Re concentrations in fractions of sediment NAG 48

Sample Treatment Mass fraction 187Os/188Os [Os] [187Os] [188Os] [Re] 187Re/188Os(%) (ppt) (ppt) (ppt) (ppt)

Bulk sedimenta 3.776 20 6.8 1.8 299 103.43.764 20 6.8 1.8

Sieved Sieved to 150^300 Wm 4.210 19.1 6.9 1.6 224 86.0Carbonate-rich fraction Sieved to 150^300 Wm,

carbonate enrichmentby density andmagnetic separation

2.800 10.1 2.7 1.0 143 92.4

Calcite leach 0.16 N HCl leachateof the carbonateenriched-fraction

24.3 2.670 7.9 2.1 0.8 14 11.0

Dolomite leach 2.5 N HCl leachateof the residue after0.16 N HCl leaching

25.6 4.077 7.2 2.5 0.6 20 20.4

Silicate residue Residue of HClleachings

50.1 3.033 6.0 17 0.6 102 113.5

Mica Sieved to 150^300 Wm,mica enrichment bydensity, magnetic andelectrostatic separation

3.111 32.8 9.6 3.1 ^

Mica-poor residue Sieved to 150^300 Wm^ residue afterelectrostatic separationof the mica fraction.

4.218 34.1 12.2 2.9 ^

a Values of sample NAG 48 from Pierson-Wickmann et al. [12]. Mass fractions of the leachate and residue normalized to theoriginal mass of the carbonate-rich fraction used for the leaching tests. Mass fractions calculated from weight loss during leach-ing, after drying the residue.

EPSL 6427 6-11-02

A.-C. Pierson-Wickmann et al. / Earth and Planetary Science Letters 204 (2002) 203^214208

original carbonate rich fraction (187Os/188Os=2.80, [188Os] = 1.0 ppt and [Re] = 143 ppt) andthe weighted average of the residue and leachates(calculated values are 187Os/188Os= 3.17, [188Os] =0.65 ppt and [Re] = 60 ppt). This probably resultsfrom inaccuracies in the masses (calcite and dolo-mite masses are based on weight loss of the resi-due during leaching), sample losses during rinsingand centrifugation, sample inhomogeneity andpossibly loss of Os during residue dry-down.Sorption of the leached Os by the residue doesnot appear to be a major factor, as the residuedoes not have a higher Os concentration than theleachates.These leaching tests indicate that, while calcite

and dolomite have Os isotopic compositionsroughly comparable to that of the bulk sediment,their Os concentrations are much too low to ex-plain the Os abundance and the radiogenic char-acter of this sediment. Results from the minesamples show that hydrothermal phlogopite, asso-ciated with carbonate, can sometimes contain veryhigh concentrations of 187Os. For this reason weprepared a phyllosilicate-rich separate from the150^300 Wm grain size fraction of NAG 48, byelectrostatic separation from an aliquot withhigh densities (bs 2.7 g/cm3) and high magneticsusceptibilities. Microscopic examination of thisseparate showed that it was composed almostcompletely of phyllosilicates, though of widelyvarying color and aspect suggesting a multiplicityof sources. This phyllosilicate separate has a 188Osconcentration of 3.1 ppt with a 187Os/188Os ratioof 3.11. This ratio is lower than that of the bulksediment, re£ecting a large mica contributionfrom the less radiogenic formations (187Os/188OsV0.6^2; [12]). The Os concentration of thephyllosilicate separate is relatively high comparedto that of the bulk 150^300 Wm grain size fraction(188Os = 1.6 ppt). This is consistent with the resultsof a previous study focused on the weathering ofPrecambrian granitoid that showed that biotitewas much richer in Os than the coexisting silicatephases [22]. The residue from the electrostatic sep-aration has a 188Os concentration of 2.9 ppt witha 187Os/188Os ratio of 4.22, making it the mostradiogenic fraction observed in this study. Its187Os concentration is about twice that of the

bulk sediment. This high concentration could re-£ect the presence of hydrothermal sul¢des, suchas pyrite, which can be extremely rich in 187Os(Table 1) and is dense and paramagnetic andthus likely to be concentrated in this fraction. Ifso, it is a bit surprising, as pyrite is probablyquickly oxidized during weathering. The factthat Himalayan rivers often have high SO23

4 con-centrations, thought to result largely from pyriteoxidation [23,24], supports the contention that py-rite is unlikely to survive transport. Anotherheavy phase that might be rich in Os is magnetite,as demonstrated by the study of the Precambriangranitoid cited above [22]. However in that study,the authors found that the magnetite was the leastradiogenic phase, while in our study, the heavy,magnetic fraction is the most radiogenic. Thus theexact identity of the Os-bearing phases remains abit of a mystery. Nevertheless, both the phyllosi-licate and the heavy magnetic fractions are richerin Os than the rest of the sediment.

5. Origin of the radiogenic Os in marbles

The data presented above demonstrate that LHmarbles can have remarkably high 187Os/188Os ra-tios. It is easy to show that closed system evolu-tion cannot explain these very radiogenic ratios.Using the 187Re/188Os ratios, Os model ages rang-ing from 1 to 8 Ga can be calculated for thedolomitic marbles (Table 1). While the youngestof these ages are plausible, the oldest are impos-sible. Thus the radiogenic Os present in most ofthese marbles is not due to simple in situ radio-active decay of 187Re, but instead requires themobility of Os and Re, either by recent loss ofRe, or by input of 187Os. Simple recent loss ofthe Re initially deposited in the carbonate sedi-ments is unlikely in view of the very high Reconcentrations of some marbles (AP 207 and AP865, Table 1). There is no reason why organiccarbon-poor carbonates should be depositedwith Re contents comparable to or even higherthan those currently observed in these samples.However, a more complex process can be envis-aged, in which Re is ¢rst concentrated in hydro-thermal sul¢des formed soon after deposition. Re-

EPSL 6427 6-11-02

A.-C. Pierson-Wickmann et al. / Earth and Planetary Science Letters 204 (2002) 203^214 209

cent oxidation of these sul¢des might then prefer-entially remove Re relative to Os, leaving rockswith high 187Os/188Os ratios unsupported by thecorresponding Re/Os ratios. An analogous pro-cess was proposed to explain the very high butunsupported 187Os/188Os ratios of metalliferoussediments from the Troodos ophiolite [25].Alternatively, the addition of 187Os, accompa-

nied by some Re to explain the high Re concen-trations, could plausibly explain the very high187Os/188Os ratios of some marbles. This Re andradiogenic Os may have been remobilized by hy-drothermal circulation during metamorphicevents. Previous studies of the weathering of blackshales [17,26] have suggested a preferential mobi-lization of Re and 187Os relative to the other Osisotopes. An analogous suggestion was made byDeniel [27] and Edmond [28] to explain very highSr isotopic ratios unsupported by Rb contents ofcertain metamorphosed Himalayan samples. Infact a rough correlation exists between Sr andOs isotopic compositions of the marbles (Fig. 3),suggesting that 87Sr and 187Os may have beenbrought in by the same £uids. This may providean argument favoring 187Os addition over the Reloss mechanism described above, as there is noreason that early precipitation and recent oxida-tion of hydrothermal sul¢des should induce a cor-relation between Sr and Os isotopes. Addition ofradiogenic Os during metamorphism is also sup-

ported by the fact that mine marbles containingevidence (e.g. phlogopite veins) of hydrothermalcirculation sometimes have higher Os isotopic ra-tios. Input of Re and 187Os, relative to 188Os, de-rived from Re-rich rocks provides an explanationfor both the extreme variability and the very ra-diogenic nature of the marbles. The proximity ofradiogenic marbles to black shales argues in favorof a causal relationship. Hence, sample GA 39characterized by a radiogenic Os isotopic ratio isfound in the vicinity of black shale GA 38, whilesample GA 70 located in the same mine, but in amineralized zone lacking in black shale, is Os-poor and has an unradiogenic Os isotopic ratio(Table 1).Even black shales do not have 187Os/188Os ra-

tios high enough to directly provide the extremelyradiogenic Os of certain marbles, especially AP207 with a 187Os/188Os ratio of 194. A preferentialrelease of 187Os during metamorphism of theblack shales seems to be required. Such a scenariois consistent with the black shale data. In a plot of187Os/188Os vs. 187Re/188Os (Fig. 4), about half ofthe black shales plot well above lines representing1 and 2 Ga isochrons, implying that they havelost Re. Some of this preferential Re loss mayhave occurred during recent weathering [17], asshown by the elevated 187Re/188Os ratios ofGanga and Brahmaputra river waters (Fig. 4).However, substantial mobility of Re, accompa-

Fig. 3. (a) Os vs. Sr isotopic ratios of impure marbles. (b) Zoom-in of the rectangle.

EPSL 6427 6-11-02

A.-C. Pierson-Wickmann et al. / Earth and Planetary Science Letters 204 (2002) 203^214210

nied by 187Os derived from Re-decay, probablyalso occurred during Himalayan metamorphism.Incorporation of this 187Os into the marbles dur-ing £uid circulation associated with metamor-phism may explain the extremely radiogenic val-ues observed in some marbles. As noted above,the Re concentrations of these samples, thoughquite high (up to 3580 ppt), are not high enoughto explain their 187Os/188Os ratios by radiogenicingrowth during geologically reasonable time pe-riods. This suggests that Os was incorporatedmore readily than Re into the marble from the187Os and Re rich £uids.

6. The in£uence of radiogenic marbles onerosional products

Regardless of the ultimate source of radiogenicOs in the impure marbles, weathering of theserocks will have an important in£uence on theOs budget of both the solid and dissolved ero-sional products. As mentioned above, the addi-tion of LH black shales alone to silicic sedimentsfrom all three Himalayan formations cannot ex-plain the Os compositions of sediments collectedat the out£ow of the Himalayan chain. Howeverthe addition of about 4% of LH marbles with Oscompositions similar to AP 207 or AP 865, as wellas 0.5^1% LH black shale, reasonable proportionsbased on outcrop observations, allows the Os

characteristics of typical Himalayan out£ow bed-loads to be explained (Table 3). This proportionof marble agrees well with the proportion of do-lomite, which is essentially LH-derived, containedin the river bedload (4% for NAG 48) and withthe proportion of LH carbonates in the bedloadscalculated from the Sr isotopic compositions ofthe bedloads [29]. However it should be stressedthat the radiogenic Os in the impure carbonates isnot necessarily hosted by the carbonate itself, butperhaps by associated trace phases. Also, the pro-portion of LH marbles needed to balance the Osbudget of the bedload depends on which sample ischosen as the mixing end-member. The use of themost radiogenic marbles minimizes the propor-tion of LH marbles required. The choice of aless radiogenic LH marble composition might ne-cessitate an unreasonably high proportion of thiscomponent in the sediment, and thus suggest thatyet another, unidenti¢ed highly radiogenic com-ponent is also present. The processes involvedare also undoubtedly more complicated than sim-ple physical mixing. Nevertheless, the generalagreement found between the simple mixing mod-el results and the observed bedloads suggests thatradiogenic marbles similar to the AP samples doindeed play an important role in the sedimentaryOs budget.The in£uence of impure marbles on the dis-

solved Os load and thus ultimately on the marineOs budget depends on the solubility of the phasesthat carry the radiogenic Os. Our detailed studyof various fractions of the sediment NAG 48 (Ta-ble 2) suggests that the highly radiogenic charac-ter of the bulk sediment is not derived from anysingle phase, but rather from many. Carbonatesand especially dolomite have high 187Os/188Os ra-tios comparable to that of the bulk sediment.Since some of the carbonate is derived from theTSS, which has lower Os isotopic ratios, the aver-age 187Os/188Os of the LH carbonate phase mustbe even higher. Nevertheless the calcite and dolo-mite Os concentrations are low, implying that thecarbonate phases present in the sediment do notcontribute greatly to explaining the radiogenic na-ture of the bulk sediment. These phases, thoughhighly soluble, will also have only a moderatein£uence on the Os composition of the dissolved

marbles

black shales

1 Ga

2 Ga

187Os/188Os

187Re/188Os

0 200 400 600 800 1000 1600

0

10

20

30

40

14001200

BrahamputraGanges

Fig. 4. 187Os/188Os vs. 187Re/188Os for LH carbonates andblack shales. Dashed lines represent 1 Ga and 2 Ga iso-chrons. White and black dots represent the dissolved load ofthe Brahamputra and the Ganges, respectively, based ondata from [5,30,33].

EPSL 6427 6-11-02

A.-C. Pierson-Wickmann et al. / Earth and Planetary Science Letters 204 (2002) 203^214 211

load. On the other hand, the presence of detritalLH carbonate grains may explain why leachatesof sediments in the Bengal Fan are almost alwaysmore radiogenic than their corresponding bulksediments [16]. Hydrothermal phlogopite associ-ated with the carbonates may be an importantcarrier of 187Os but its signi¢cance is di⁄cult toassess as its signal in the phyllosilicate separate ispartially masked by the abundance of mica fromother sources. Indeed the 187Os/188Os ratio of thisseparate (3.1), though much higher than those oftypical Himalayan silicate rocks from all threeformations, cannot alone explain the ratio of thebulk sediment (3.8). The dense, magnetic residue

from the phyllosilicate separation has both a highOs concentration and a high 187Os/188Os ratio(4.2), and thus may represent an important com-ponent of the 187Os in the bulk sediment. It isunclear whether the 187Os in this fraction comesmostly from very radiogenic, Os-rich hydrother-mal sul¢des (similar to those in GA 39) that havesomehow survived transport, or from otherphases.Thus the radiogenic Os signature of bedloads of

rivers £owing out of the Himalayan chain is de-rived from a variety of sources: black shale frag-ments certainly, but also the various phases foundin impure, hydrothermally altered or metamor-

Table 3Calculated river bedload Os compositions based on mixtures of High Himalaya sediments with LH carbonates and black shales

187Os/188Os [Os] [187Os] [188Os] LH carb. LH black shale(ppt) (ppt) (ppt) (%) (%)

AP 207 193.72 102 98.7 0.5AP 811 15.51 10 6.5 0.4AP 865 19.09 188 135.4 7.1AP 867 2.87 16 4.5 1.5Average of black shale 8.6 520 276.6 32.2Initial bedload 1.3 20 2.96 2.28 0 0Initial+carbonatesAP 207 3.0 23.3 6.79 2.21 4 0AP 811 1.4 19.6 3.10 2.20 4 0AP 865 3.3 26.8 8.26 2.47 4 0AP 867 1.3 19.8 3.02 2.25 4 0Initial+black shale

1.8 22.6 4.33 2.43 0 0.52.2 25.1 5.70 2.57 0 1

Initial+black shale+carbonateAP 207 3.4 25.8 8.16 2.36 4 0.5AP 811 1.9 22.1 4.47 2.35 4 0.5AP 865 3.7 29.2 9.63 2.62 4 0.5AP 867 1.8 22.3 4.39 2.40 4 0.5Initial+black shale+carbonateAP 207 3.8 28.3 9.453 2.51 4 1AP 811 2.3 24.6 5.84 2.50 4 1AP 867 2.3 24.8 5.76 2.55 4 1AP 865 4.0 31.7 11.00 2.77 4 1Narayani sedimentsNAG 48 3.8 20.5 6.86 1.82 4 6 1MO 217 3.2 15.6 4.70 1.46 4 6 1

The ‘initial’ bedload at the beginning is an arithmetical average of bedload data from the HHC and those of rivers havingcrossed the TSS+HHC (Table 2 of Pierson-Wickmann et al. [12]). The proportion of LH carbonates in bedloads at the out£owof the Himalayan chain has been calculated to be 4% on average, on the basis of Sr isotopic data from the carbonate fractionsof river sediments (NAG 48), assuming for the TSS, 87Sr/86Sr = 0.712 and [Sr] = 700 ppm and for the LH, 87Sr/86Sr= 0.760 and[Sr] = 50 ppm [13]. Mixtures of 4% of LH carbonates and 0.5 or 1% of LH black shales produce 187Os/188Os ratios similar tothose observed in samples NAG 48 and MO 217.

EPSL 6427 6-11-02

A.-C. Pierson-Wickmann et al. / Earth and Planetary Science Letters 204 (2002) 203^214212

phosed LH marbles. The contribution of each ofthese phases to the solid and dissolved loads de-pends on the abundance and solubility of thephase, as well as on its Os concentration and iso-topic composition. In a more general sense, thecontribution of a given rock type to the dissolvedOs budget depends not only on its volumetricproportion and its Os concentration, but also onits tendency to release Os during weathering,which varies between lithologies. Furthermore,as demonstrated by previous studies [9,30^32],the capacity of both Re and Os to enter intoand remain in solution depends on the redox con-ditions during weathering and transport. Thus thetransfer of radiogenic Os from the continents tothe oceans is very complex process.

7. Conclusions

Impure LH marbles sometimes contain largeamounts of 187Os and thus may represent an im-portant source of radiogenic Os during erosion.However, this radiogenic Os, as well as Re, isprobably not carried only by the carbonate min-erals themselves, but also by associated phases,such as hydrothermal pyrites or phlogopite. Theresults obtained on minor phases in hydrothermalveins indicate that in some cases 187Os was addedduring hydrothermal activity. The direct and ob-vious source of this radiogenic Os is the alterationof nearby radiogenic black shales. In other cases187Os was introduced, or alternatively, Re waslost, during interaction with £uids related to latestage Himalayan metamorphism. The erosion ofsuch hydrothermally modi¢ed carbonates mayplay an important role in the budget of Os enter-ing the oceans, especially when the eroding basinsinclude old cratons dominated by black shalesand carbonate-rich rocks. Thus while the ultimatesource of the radiogenic Os may be organic-richrocks, the temporary storage of this Os in impuremarbles complicates the link between the Os iso-topic record and the erosion of organic carbon.These results provide new information that mayhelp us better understand Re and Os behaviorduring metamorphism and post-depositional pro-cesses as well as the marine Os record.

Acknowledgements

We thank G. Ravizza, D. Colodner and B.Dupre¤ for constructive reviews that greatly im-proved this manuscript. This study was supportedby the Programme National Sol et Erosion(CNRS-INSU). This is CRPG contribution No.1597.[BARD]

References

[1] W.J. Pegram, S. Krishnaswami, G.E. Ravizza, K.K. Ture-kian, The record of sea water 187Os/186Os variationthrough the Cenozoic, Earth Planet. Sci. Lett. 113(1992) 569^576.

[2] G. Ravizza, Variations of the 187Os/186Os ratio of sea-water over the past 28 million years inferred from metal-liferous carbonates, Earth Planet. Sci. Lett. 118 (1993)335^348.

[3] K.K. Turekian, W.J. Pegram, Os isotope record in a Ce-nozoic deep-sea core: its relation to global tectonics andclimate, in: W.F. Ruddiman (Ed.), Tectonic Uplift andClimate Change, Plenum Press, New York, 1997, pp.383^397.

[4] S. Levasseur, J.-L. Birck, C.J. Alle'gre, The osmium river-ine £ux and the oceanic mass balance of osmium, EarthPlanet. Sci. Lett. 174 (1999) 7^23.

[5] M. Sharma, G.J. Wasserburg, A.W. Hofmann, G.J.Chakrapani, Himalayan uplift and osmium isotopes inoceans and rivers, Geochim. Cosmochim. Acta 63 (1999)4005^4012.

[6] B. Peucker-Ehrenbrink, R.E. Hannigan, E¡ects of blackshales weathering on the mobility of rhenium and plati-num group elements, Geology 28 (2000) 475^478.

[7] G. Ravizza, K.K. Turekian, Application of the 187Re-187Os system to black shale geochemistry, Geochim. Cos-mochim. Acta 53 (1989) 3257^3262.

[8] G. Ravizza, K.K. Turekian, The osmium isotopic compo-sition of organic-rich marine sediments, Earth Planet. Sci.Lett. 110 (1992) 1^6.

[9] G. Ravizza, K.K. Turekian, B.J. Hay, The geochemistryof rhenium and osmium in recent sediments from theBlack Sea, Geochim. Cosmochim. Acta 55 (1991) 3741^3752.

[10] W.J. Pegram, B.K. Esser, S. Krishnaswami, K.K. Ture-kian, The isotopic composition of leachable osmium fromriver sediments, Earth Planet. Sci. Lett. 128 (1994) 591^599.

[11] S.K. Singh, J.R. Trivedi, S. Krishnaswami, Re-Os isotopesystematics in black shales from the Lesser Himalaya theirchronology and role in the 187Os/188Os evolution of sea-water, Geochim. Cosmochim. Acta 63 (1999) 2381^2392.

[12] A.-C. Pierson-Wickmann, L. Reisberg, C. France-Lanord,

EPSL 6427 6-11-02

A.-C. Pierson-Wickmann et al. / Earth and Planetary Science Letters 204 (2002) 203^214 213

The Os isotopic composition of Himalayan river bedloadsand bedrocks: Importance of black shales, Earth Planet.Sci. Lett. 176 (2000) 201^216.

[13] A. Galy, C. France-Lanord, L.A. Derry, The strontiumisotopic budget of Himalayan Rivers in Nepal and Ban-gladesh, Geochim. Cosmochim. Acta 63 (1999) 1905^1925.

[14] D.N. Reusch, G. Ravizza, K.A. Maasch, J.D. Wright,Miocene seawater 187Os/188Os ratios inferred from metal-liferous carbonates, Earth Planet. Sci. Lett. 160 (1998)163^178.

[15] C. France-Lanord, Chevauchement, me¤tamorphisme etmagmatisme en Himalaya du Ne¤pal central. Etude isoto-pique H, C, O., Institut National Polytechnique de Lor-raine, Nancy, 1987.

[16] A.-C. Pierson-Wickmann, L. Reisberg, C. France-Lanord,H.R. Kudrass, Os-Sr-Nd results from sediments in theBay of Bengal: Implications for sediment transport andthe marine Os record, Paleoceanography 16 (2001) 435^444.

[17] A.-C. Pierson-Wickmann, L. Reisberg, C. France-Lanord,Behavior of Re and Os during low temperature alterationresults from Himalayan soils, Geochim. Cosmochim. Acta66 (2002) 1539^1548.

[18] J.-L. Birck, M. Roy-Barman, F. Capmas, Re-Os isotopicmeasurements at the femtomole level in natural samples,Geostand. Newslett. 20 (1997) 19^27.

[19] M.R. Barman, Mesure du rapport 187Os/186Os dans lesbasaltes et les pe¤ridotites: Contribution a' la syste¤matique187Re-187Os dans le manteau, Ph. D Thesis, Paris VII,1993.

[20] R.A. Creaser, D.A. Papanastassiou, G.J. Wasserburg,Negative thermal ion mass spectrometry of osmium, rhe-nium, and iridium, Geochim. Cosmochim. Acta 55 (1991)397^401.

[21] J. Vo« lkening, T. Walczyk, K.G. Heumann, Osmium iso-tope ratio determinations by negative thermal ionisationmass spectrometry, Mass Spectrom. Ion Process. 105(1991) 147^159.

[22] B. Peucker-Ehrenbrink, J.D. Blum, Re-Os isotope system-atics and weathering of Precambrian crustal rocks: Impli-

cations for the marine osmium isotope record, Geochim.Cosmochim. Acta 62 (1998) 3193^3203.

[23] A. Galy, C. France-Lanord, Weathering processes in theGanges-Brahmaputra basin and the riverine alkalinitybudget, Chem. Geol. 159 (1999) 31^60.

[24] M.M. Sarin, S. Krishnaswami, J.R. Trivedi, K.K. Shar-ma, Major ion chemistry of the Ganga source WatersWeathering in the high altitude Himalaya, Earth Planet.Sci. Lett. 101 (1992) 89^98.

[25] G.E. Ravizza, J. Blusztajn, H.M. Prichard, Re-Os system-atics and platinum-group element distribution in metalli-ferous sediments from the Troodos ophiolite, Earth Plan-et. Sci. Lett. 188 (2001) 369^381.

[26] L.A. Ja¡e, B. Peucker-Ehrenbrink, S.T. Petsch, E¡ects ofweathering of black shales on the mobility of rhenium,platinum group elements and organic carbon, Earth Plan-et. Sci. Lett. 198 (2002) 339^353.

[27] C. Deniel, Apport des isotopes du Sr, Nd et Pb a' laconnaissance de l’a“ge et de l’origine des leucogranites hi-malayens. Exemple du Manaslu (Himalaya-Ne¤pal),The'se: Etude approfondie de Pe¤trologie-Mine¤ralogie,Univ. de Clermont-Ferrand II, 1985.

[28] J.M. Edmond, Himalayan tectonics, weathering processes,and the strontium isotope record in marine limestones,Science 258 (1992) 1594^1597.

[29] A. Galy, Etude ge¤ochimique de l’e¤rosion actuelle de lacha|“ne himalayenne, Doctorat de l’INPL, Institut Nation-al Polytechnique de Lorraine, 1999.

[30] D. Colodner, J. Sachs, G. Ravizza, K. Turekian, J. Ed-mond, E. Boyle, The geochemical cycle of rhenium a re-connaissance, Earth Planet. Sci. Lett. 117 (1993) 205^221.

[31] J. Crusius, S. Calvert, T. Pederson, D. Sage, Rhenium andmolybdenum enrichments in sediments as indicators ofoxic, suboxic and sul¢dic conditions of deposition, EarthPlanet. Sci. Lett. 145 (1996) 65^78.

[32] J.L. Morford, S. Emerson, The geochemistry of redoxsensitive trace metals in sediments, Geochim. Cosmochim.Acta 63 (1999) 1735^1750.

[33] S. Levasseur, J.-L. Birck, C.-J. Alle'gre, The osmium riv-erine £ux and the oceanic mass balance of osmium, EarthPlanet. Sci. Lett. 174 (1999) 7^23.

EPSL 6427 6-11-02

A.-C. Pierson-Wickmann et al. / Earth and Planetary Science Letters 204 (2002) 203^214214